Hypotonic stress-induced dual Ca 2 responses in bovine aortic endothelial cells

Am J Physiol Heart Circ Physiol 279: H630 H638, 2000. Hypotonic stress-induced dual Ca 2 responses in bovine aortic endothelial cells MASAHIRO OIKE, CHIWAKA KIMURA, TETSUYA KOYAMA, MIYUKI YOSHIKAWA, AND YUSHI ITO Department of Pharmacology, Graduate School of Medical Sciences, Kyushu University, Fukuoka 812-8582, Japan Received 30 June 1999; accepted in final form 7 February 2000 Oike, Masahiro, Chiwaka Kimura, Tetsuya Koyama, Miyuki Yoshikawa, and Yushi Ito. Hypotonic stress-induced dual Ca 2 responses in bovine aortic endothelial cells. Am J Physiol Heart Circ Physiol 279: H630 H638, 2000. We have investigated the effects of hypotonic stress on intracellular calcium concentration ([Ca 2 ] i ) in bovine aortic endothelial cells. Reducing extracellular osmolarity by 5% to 40% elicited a steep Ca 2 transient both in normal Krebs and Ca 2 -free solutions. The hypotonic stress-induced Ca 2 transient was inhibited by phospholipase C inhibitors (neomycin and U-73122), a P 2 -receptor antagonist (suramin), and an ATP-hydrolyzing enzyme (apyrase), suggesting that the hypotonic stress-induced Ca 2 transient is mediated by ATP. A luciferin-luciferase assay confirmed that 40% hypotonic stress released 91.0 amol/cell of ATP in 10 min. When the hypotonic stress-induced fast Ca 2 transient was inhibited by neomycin, suramin, or apyrase, a gradual [Ca 2 ] i increase was observed instead. This hypotonic stress-induced gradual [Ca 2 ] i increase was inhibited by a phospholipase A 2 inhibitor, 4-bromophenacyl bromide. Furthermore, exogenously applied arachidonic acid induced a gradual [Ca 2 ] i increase with an ED 50 of 13.3 M. These observations indicate that hypotonic stress induces a dual Ca 2 response in bovine aortic endothelial cells, i.e., an ATP-mediated fast Ca 2 transient and an arachidonic acid-mediated gradual Ca 2 increase, the former being the predominant response in normal conditions. mechanical stress; adenosine 5 -triphosphate; arachidonic acid. VASCULAR ENDOTHELIUM is continuously exposed to mechanical stress by blood flow and blood pressure, and these mechanical stresses have been known to modulate various endothelial functions such as the production of vasoactive agents, gene expression, and the alteration of endothelial alignment (for a review, see Ref. 6). Because of the dependence of these functions on intracellular Ca 2 concentration ([Ca 2 ] i ), investigation of mechanical stress-induced Ca 2 mobilization in the endothelium would have significant importance in vascular biology. However, the details of mechanical stress-induced Ca 2 mobilization are not fully understood. Oike and colleagues (15) have previously reported that hypotonic stress induces a gradual increase in [Ca 2 ] i in human umbilical vein endothelial cells. Other mechanical stresses such as membrane stretch or shear stress also produced a gradual increase of [Ca 2 ] i. Because all of these Ca 2 transients were inhibited by phospholipase A 2 inhibitors (4-bromophenacyl bromide or cyclosporin A), Oike et al. (15) concluded that mechanosensitive Ca 2 transient was mediated by arachidonic acid in human umbilical vein endothelial cells. Fluid shear stress is also reported to induce a fast Ca 2 transient (22) or activate Ca 2 - permeable cation channels (18) in the vascular endothelium. Furthermore, it has been reported that ATP, substance P, and acetylcholine, which could induce Ca 2 mobilization, are released from the endothelium in response to the increased flow (14). In the present study, we investigated the effects of hypotonic stress on [Ca 2 ] i in the aortic endothelium. Because plasma osmolarity is strictly controlled in a narrow range in vivo, hypotonic stress would be rarely applied to the endothelium in a physiological environment. However, reported hypotonic stress-induced endothelial responses share some common characteristics with shear stress-induced ones. For instance, transient reorganization of the actin cytoskeleton induced by hypotonic stress (17) is quite similar to that by shear stress (12). Furthermore, it has been reported that shear stress activates chloride current, a current that is like hypotonic stress-induced volumeregulated chloride current, in endothelial cells (2). Therefore, we consider that the investigation of endothelial responses to hypotonic stress would provide some information about the physiological responses in the endothelium. We used hypotonic stress as a mechanical stress rather than shear stress in the present study, because cellular responses to hypotonic stress were highly reproducible. Furthermore, besides its physiological relevance, it is also possible that swelling of endothelial cells would occur during cell division. Because endothelial proliferation is closely related to angiogenesis (8), endothelial responses to Address for reprint requests and other correspondence: M. Oike, Dept. of Pharmacology, Graduate School of Medical Sciences, Kyushu Univ., Fukuoka, 812-8582 Japan (E-mail: moike@pharmaco. med.kyushu-u.ac.jp). The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. H630 0363-6135/00 $5.00 Copyright 2000 the American Physiological Society http://www.ajpheart.org

HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM H631 hypotonic stress may have a pathophysiological significance. We found that aortic endothelium has dual mechanosensitive Ca 2 mobilizing machinery. The possible significance of each mechanism for endothelial function will be discussed. MATERIALS AND METHODS Cell culture. Bovine thoracic aortas of 1-year-old calves were obtained from the local slaughter house. Endothelial cells were cultured in Dulbecco s modified Eagle s medium with 10% fetal bovine serum as previously described (16). Cells were grown on coverslips or 96-well culture plates for the measurement of [Ca 2 ] i or ATP concentration, respectively. Measurement of [Ca 2 ] i. We measured [Ca 2 ] i from single bovine aortic endothelial cells. [Ca 2 ] i was measured from isolated, nonconfluent cells to avoid the influence from the neighboring cells through cell-to-cell connections. Cells were loaded with 2 M of the acetoxymethyl ester form of fura 2 (fura 2-AM, Dojindo, Kumamoto, Japan). The coverslip with fura 2-loaded cells was placed in a chamber of 0.5 ml volume and mounted on an inverted microscope (Diaphot TMD, Nikon, Tokyo, Japan). The cell was excited with two alternative excitation wavelengths, 340 and 380 nm, applied by a spectrometer (Spex, Edison, NJ). The fluorescence ratio (R), F 340 / F 380, was calculated from the fluorescent intensity measured at 505 nm (F 340 and F 380, respectively) after subtraction of the background fluorescence. We calculated apparent [Ca 2 ] i using the equation Ca 2 i K d S f2 R R min S b2 R max R where K d is the dissociation constant between Ca 2 and fura 2, S f2 and S b2 are the fura 2 fluorescence emitted by 380 nm at zero Ca 2 and saturating Ca 2, respectively, and R min and R max are the fluorescence ratios at zero Ca 2 and saturating Ca 2, respectively. These constants were obtained by using the same optics as the experiments. Because a precise in vivo calibration of [Ca 2 ] i was difficult to perform, it should be noted that the calculated value is not actual intracellular concentration. Luciferin-luciferase bioluminescence assay. Extracellular concentration of ATP ([ATP] o ) was measured by using luciferin-luciferase bioluminescence. Cells were seeded on a 96- well plate and cultured for 2 days before use. Each well contained 4,000 cells on average before the experiment. After culture medium was carefully removed, 50 l of isotonic or hypotonic Krebs solution containing 10 mg/ml luciferaseluciferin (Wako, Osaka, Japan) was added to each well. The plate was then immediately put in a dark box, and illuminated photons were counted for 10 min by a luminescence detection system (Argus-50/2D luminometer, Hamamatsu Photonics, Hamamatsu, Japan). Obtained data were analyzed with Argus-50 software (Hamamatsu Photonics). To convert the photon counting into [ATP] o, photon-[atp] relationships were examined for each solution, because the catalytic efficiency of luciferase is largely influenced by monovalent cations (see Fig. 4B). Cytotoxicity assay. The effects of hypotonic stress on cytotoxicity were examined by measuring the leakage of lactose dehydrogenase (LDH) into an extracellular solution with a commercial kit (LDH-Cytotoxic test, Wako). Cells were cultured on a 96-well culture plate at a constant density. Culture medium was replaced with each experimental solution, and the plate was kept at room temperature for 20 min. Conversion of lactic acid into pyruvic acid by the released LDH was then measured as an alteration of 560 nm absorbance with a microplate reader (model 550, Bio-Rad, Hercules, CA) according to the manufacturer s instruction. Measurement of ethidium bromide fluorescence. Uptake of ethidium bromide into the cell was measured as an indicator of microlysis. Solutions containing 1 g/ml ethidium bromide (Sigma) were perfused to the cell, and light at 490 nm wavelength was applied every 30 s. The emitted fluorescence of 510 nm was then measured using the same equipment as Ca 2 measurement. Solutions and drugs. The standard extracellular solution for the measurements of [Ca 2 ] i was a modified Krebs solution (1.5 mm Ca 2 solution) containing (in mm) 132 NaCl, 5.9 KCl, 1.2 MgCl 2, 1.5 CaCl 2, 11.5 glucose, and 11.5 HEPES; ph was adjusted to 7.3 with NaOH. Ca 2 -free Krebs solution was made by substituting the CaCl 2 of the Krebs solution with 1 mm EGTA. Hypotonic solutions were made by adding the appropriate amount of distilled water to normal Krebs solution. In some experiments, 40% hypotonic solution was made by reducing the NaCl concentration to 72 mm, and the corresponding 20% and isotonic solutions were made by adding mannitol to obtain the osmolarity of 270 and 300 mosm, respectively (measured by an osmometer; model OM802, Vogel, Giessen, Germany). The bath was perfused continuously with these solutions at a rate of 1.5 ml/min unless specially mentioned. ATP and thapsigargin (Sigma, St. Louis, MO) were used to release Ca 2 from the intracellular Ca 2 store sites. All other drugs were also from Sigma except for suramin (Bayer, Germany) and U-73122 (Research Biochemicals, Natick, MA). Data analysis. Pooled data are given as means SE, and statistical significance was determined using Student s unpaired t-test. Probabilities less than 5% (P 0.05) were regarded as significant. RESULTS Effect of hypotonic stress on intracellular [Ca 2 ] i in bovine aortic endothelial cells. First of all we examined the effect of hypotonic stress on [Ca 2 ] i in bovine aortic endothelial cells. As shown in Fig. 1A, perfusion of hypotonic Krebs solution ( 20%) elicited an oscillatory increase in [Ca 2 ] i. The delay between the onset of hypotonic challenge to the first Ca 2 transient was 101.1 8.1 s in the case of continuous perfusion (n 25). More rapid exchange of the isotonic Krebs with the hypotonic solution, which was obtained by infusing solution with a syringe for a few seconds, also induced Ca 2 oscillation (Fig. 1A, inset), with a shorter latency of 61.3 8.7s(n 20), suggesting that the response depends on hypotonic cell swelling. Hypotonic Ca 2 - free solution also showed a transient increase in [Ca 2 ] i, suggesting that Ca 2 was released from intracellular Ca 2 store sites (Fig. 1B, n 7). Hypotonic Ca 2 -free solutions ( 5, 10, and 20%) were sequentially applied to a cell (Fig. 1C, inset). In this particular cell, the threshold reduction in osmolarity to elicit a Ca 2 transient was 10%. After the intracellular Ca 2 store sites were reloaded by perusing the cell with a Ca 2 -containing Krebs solution, a procedure that had been shown previously to refill the store sites almost

H632 HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM Fig. 1. Effect of hypotonic stress on intracellular Ca 2 concentration ([Ca 2 ] i ) in bovine aortic endothelial cells. A: hypotonic solution ( 20%) was perfused to the cell (rate: 1.5 ml/min). Ca 2 -containing Krebs solution was used. Inset: more rapid exchange from isotonic to hypotonic solution also induced Ca 2 transient but with shorter latency. Dotted line, basal [Ca 2 ] i level in Ca 2 -containing Krebs solution before application of hypotonic stress. B: hypotonic Ca 2 -free solution ( 20%) was applied. A little elevation of [Ca 2 ] i was elicited after reapplication of Ca 2. C: concentration-response relationship of hypotonic stress and Ca 2 transient. Hypotonic solutions ( 5, 10, and 20%) were applied sequentially to a cell (inset). After the small Ca 2 transient in 10% hypotonic solution was measured, Ca 2 was perfused to the bath for 5 min to reload the store sites (16). Peak [Ca 2 ] i value induced by 20% hypotonic stress was normalized as 1.0 for each cell (n 5 7 cells). D: concentration-response relationship of hypotonic stress and Ca 2 oscillation., Number of oscillation peaks/5 min (calculated from number of cells in parentheses);, maximal peak amplitude. Solution was made by adding corresponding volume of water to Krebs solution. Oscillation frequency increased in an osmolarity-dependent manner, whereas peak amplitude reached maximal value at 20%. and Œ, Oscillation frequency and maximal amplitude in mannitol-reduced hypotonic solutions, respectively, obtained by reducing NaCl to 72 mm and 20% hypotonic solution and isotonic solutions were made by adding mannitol to it. completely (16), the solution with lower osmolarity ( 20%) elicited a larger increase in [Ca 2 ] i. The corresponding osmolarity-response relationship is shown in Fig. 1C. This Ca 2 release-reload protocol, as in Fig. 1C (inset), was performed in each experiment. The peak amplitudes of Ca 2 transients evoked by 2.5%, 5%, and 10% hypotonic solutions were 0, 0.033 0.033 and 0.472 0.201 relative to that of the 20% hypotonic solution, respectively (n 5 6). Further reduction of osmolarity up to 40% did not induce larger Ca 2 transients but showed Ca 2 oscillations with a higher frequency (Fig. 1D). The number of Ca 2 peaks in 5 min induced by 10, 20, 30, and 40% hypotonic solutions were 1.1 0.3 (n 8), 1.8 0.3 (n 14), 3.4 0.5 (n 7) and 4.9 0.5 (n 7) times, respectively (Fig. 1D). Hypotonic solutions ( 20 and 40%) made by reducing the NaCl concentration (not by addition of water) also elicited a Ca 2 oscillation (n 14 16) in which the frequency and maximal amplitude were not significantly different from the results of water-added hypotonic solutions (Fig. 1D). This suggests that the hypotonic response depends on the tonicity of the solution but not on its ionic composition. We therefore performed the following experiment using water-added hypotonic solution. Effects of inhibition of phospholipase C on hypotonic stress-induced Ca 2 release in bovine aortic endothelial cells. We then tried to clarify the intracellular mechanisms of the hypotonic stress-induced Ca 2 transient. When cells were preincubated for 30 min at 37 C with 1 mm neomycin, an inhibitor of phospholipase C, the

HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM H633 Fig. 2. Effects of the inhibition of phospholipase C on hypotonic stress-induced Ca 2 transient. A: effect of hypotonic Ca 2 -free solution ( 20%) on the neomycin-treated endothelial cell. The cell was pretreated with 1 mm neomycin for 30 min at 37 C. Only a small Ca 2 transient was elicited. The following application of 1 M thapsigargin induced [Ca 2 ] i increase. B: hypotonic stress ( 20%) was applied to the U-73122-treated endothelial cell. The cell was pretreated with 10 M U-73122 for 15 min at 37 C. Dashed line, basal [Ca 2 ] i level before application of hypotonic solution. Inset, pretreatment of the cell with 10 M U-73343, an inactive analog of U-73122, for the same period did not affect hypotonic stress-induced Ca 2 transient. C: effect of 20% hypotonic stress on neomycin- and 4-bromophenacyl bromide (pbpb)-treated cell. The cell was incubated with Krebs solution containing 1 mm neomycin and 10 M pbpb for 30 min at 37 C. No apparent [Ca 2 ] i increase was observed. hypotonic solution ( 20%) did not induce a steep Ca 2 transient (Fig. 2A) but induced a gradual increase in [Ca 2 ] i in all cells examined (n 6). A subsequent application of 1 M thapsigargin induced, however, an increase of [Ca 2 ] i, suggesting that intracellular Ca 2 store sites were intact after the pretreatment with neomycin. We also examined the effects of U-73122, a phospholipase C inhibitor that is also known to inhibit phospholipase A 2 (5), on the hypotonic stress-induced Ca 2 transient. As shown in Fig. 2B, hypotonic stress did not evoke any increase in [Ca 2 ] i in the U-73122- treated cells but gradually reduced [Ca 2 ] i (n 7). On the other hand, U-73343, an inactive analog of U-73122, did not inhibit Ca 2 oscillation (Fig. 2B, inset). To confirm this dual action, we treated cells with neomycin and 4-bromophenacyl bromide (pbpb), a phospholipase A 2 inhibitor, together. After the combined incubation of the cell with neomycin and pbpb, hypotonic stress reduced [Ca 2 ] i, as was the case for U-73122 (Fig. 2C). The net [Ca 2 ] i increase by 20% hypotonic stress in the neomycin-treated cells in the absence and presence of 10 M pbpb were 34.1 5.8 nm (n 6) and 0.5 2.8 nm (n 7), respectively (P 0.05). On the other hand, the hypotonic stress-induced [Ca 2 ] i increase in neomycin-treated cells was not affected by inhibiting the downstream metabolism of arachidonic acid; i.e., indomethacin and 5-nitro-2-(3- phenylpropylamino)-benzoic acid did not inhibit the [Ca 2 ] i increase (data not shown). These observations indicate that the hypotonic stress-induced fast Ca 2 transient is mediated by inositol (1,4,5) trisphosphate [Ins (1,4,5) P 3 ], generated either directly by hypotonic cell swelling or by some Ins (1,4,5) P 3 -generating agonist. Furthermore, activation of phospholipase A 2 is also involved in the hypotonic stress-induced Ca 2 transient, especially when the Ins (1,4,5) P 3 -induced fast Ca 2 transient is inhibited. Possible involvement of extracellular ATP in hypotonic stress-induced Ca 2 transient in bovine aortic endothelial cells. We then investigated the detailed mechanism of the hypotonic stress-induced, Ins (1,4,5) P 3 -mediated fast Ca 2 transient. Endothelial cells have been reported to produce various biologically active mediators such as ATP or substance P in response to fluid stress (14), so the results above raise the possibility that one of these substances might also mediate hypotonic Ca 2 responses. So we examined the effect of suramin, an inhibitor of P 2 purinergic receptors, and apyrase, an ATP-hydrolyzing enzyme, on hypotonic stress-induced Ca 2 transients. Figure 3A shows that hypotonic stress ( 20%) failed to evoke the fast Ca 2 transient reminiscent of Ca 2 release in the presence of 30 M suramin but induced a gradual [Ca 2 ] i increase instead (n 5). Furthermore, when extracellular ATP was scavenged by 2 U/ml apyrase, hypotonic stress ( 20%) did not induce the fast Ca 2 transient but induced a gradual increase in [Ca 2 ] i (n 7, Fig. 3B). It seems therefore probable that the fast Ca 2 transient is elicited by ATP released from endothelial cells. Bath application of a low concentration of ATP elicited a Ca 2 transient in an all-or-none manner with a threshold of 10 nm (n 5, Fig. 3C). Furthermore, exogenously applied ATP also induced a fast Ca 2 transient in the presence of 30 M suramin but with a higher threshold (100 nm, n 5, Fig. 3D). The maximal rates of the [Ca 2 ] i increase during the Ca 2 transients at the threshold concentrations of ATP in the absence and presence of suramin were similar, i.e., 10.1 1.6 nm/s (n 5) in control cells and 8.6 1.6 nm/s

H634 HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM Fig. 3. Possible involvement of extracellular ATP in hypotonic stress-induced fast Ca 2 transient. A: hypotonic Ca 2 -free solution ( 20%) was applied to the cell in the presence of 30 M suramin. A gradual, but not steep, increase of [Ca 2 ] i is observed. B: in the presence of 2 U/ml apyrase, hypotonic stress ( 20%) did not induced fast Ca 2 transient either. Gradual increase in [Ca 2 ] i was observed instead. C: effect of sequential application of low concentrations of ATP on a bovine aortic endothelial cell. Ca 2 transient was elicited in an all-or-none manner with a threshold of 10 nm. The same result was observed in four other cells. D: in the presence of 30 M suramin, ATP-induced Ca 2 transient was observed with a threshold of 100 nm. The same result was observed in four other cells. (n 5) in the presence of suramin (P 0.05). Therefore, hypotonic stress-induced gradual [Ca 2 ] i increase in the presence of suramin, whose maximal rate of increase was 0.33 1.6 nm/s (n 5, P 0.05 compared with 100 nm ATP in the presence of suramin), seems not to be mediated by ATP. Measurement of [ATP] o in bovine aortic endothelial cells. To confirm that the hypotonic stress-induced fast Ca 2 transient is due to release of ATP from the endothelium, we then measured [ATP] o using the luciferinluciferase assay. By carefully replacing the culture medium with isotonic or hypotonic Krebs solutions, we generated photons through the catalytic reaction of luciferase, which indicates the release of ATP from the cell (Fig. 4A). The photon count per 10 minutes was converted into [ATP] o using photon-[atp] standard curves (Fig. 4B), and then converted into the amount of released ATP by using the volume of solution (50 M) and the cell number of each well. In isotonic Krebs solution, [ATP] o of the well was increased by 2.71 0.36 nm [change in [ATP] o ( [ATP] o )] in 10 min, and the amount of released ATP was 27.1 3.6 amol/cell per 10 min (means SE, values from 16 wells). On the other hand, the values obtained in hypotonic Krebs solutions were significantly larger than those in isotonic Krebs solutions (Fig. 4C; 20%, [ATP] o 5.60 0.71 nm and ATP release 56.0 7.1 amol/cell per 10 min, 40%, [ATP] o 9.10 0.74 nm and ATP release 91.0 7.4 amol/cell per 10 min, n 15, all P 0.01 compared with isotonic Krebs). The release of ATP from the endothelium, however, may be because of the cell damage by hypotonic cell swelling. Therefore, we then measured hypotonic stress-induced cytotoxicity, assessed by the leakage of LDH into extracellular space. We set the mean LDH leakage in isotonic Krebs solution as 0% (with a SE of 2.1%, n 15) and 0.2% Tween-20-induced mean LDH leakage as 100% (with a SE of 5.8%, n 11). With this condition, the leakage of LDH at 10, 20, and 40% hypotonic stress was calculated as 0.53 2.4 (n 9), 0.47 2.6 (n 13), and 0.16 1.9% (n 13), respectively (all P 0.05, compared with isotonic Krebs, Fig. 5A), thereby indicating that hypotonic stress does not disrupt endothelial cells.

HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM H635 Fig. 4. Measurement of extracellular ATP by luciferin bioluminescence in bovine aortic endothelial cells. A: photon imaging of ATP released from endothelium in isotonic or 40% hypotonic Krebs solutions. Images are obtained by the accumulation of the illuminated photon for 10 min. Cells were grown in 96-well culture plate, and two representative wells for each condition are shown. B: dependence of ATP concentration-photon counting relationships on extracellular solution. Photon-counting values are smaller in isotonic Krebs solution than hypotonic solution. Continuous lines represent linear fits with a correlation coefficient of 0.992 and 0.990 for isotonic and hypotonic Krebs solutions, respectively. C: ATP released in 10 min from a single endothelial cell in isotonic and hypotonic Krebs solution. Values were calculated from [ATP], which was obtained using photon counting-[atp] relations in B. Furthermore, to exclude a possibility that hypotonic stress might induce microlysis of the membrane that would allow ATP to leak out of the cell, we examined the uptake of ethidium bromide (mol wt 394.3, which is smaller than ATP) into the cell. As shown in Fig. 5B, 40% hypotonic stress did not induce an apparent increase in the intensity of ethidium bromide fluorescence. The following perfusion of 0.2% Tween-20 induced a marked increase in ethidium bromide fluorescence (Fig. 5B), thereby indicating that the fluorescence of ethidium bromide appropriately reflected the membrane leakage. Thus it seems that even 40% hypotonic stress does not increase the leakage of the plasma membrane. These results clearly indicate that hypotonic stress stimulates ATP release, which is not due to nonspecific cell damage or leakage, from bovine aortic endothelial cells. Hypotonic stress-induced gradual [Ca 2 ] i increase mediated by arachidonic acid. Hypotonic stress still induced a gradual [Ca 2 ] i increase if the ATP-induced fast Ca 2 transient was inhibited by neomycin (Fig. 2A) or suramin (Fig. 3A). As shown in Fig. 2, B and C, this gradual [Ca 2 ] i increase was inhibited by phospholipase A 2 inhibitors but not by downstream inhibitors, suggesting that this gradual [Ca 2 ] i increase is mediated by arachidonic acid. To confirm that arachidonic acid is capable of inducing Ca 2 transients in bovine aortic endothelial cells, we examined the effects of exogenously applied arachidonic acid on [Ca 2 ] i. As shown in Fig. 6A, 30 M arachidonic acid induced a gradual [Ca 2 ] i increase in the Ca 2 -free solution with a time course similar to that of the hypotonic stress-induced gradual [Ca 2 ] i increase in the presence of neomycin (Fig. 2A, n 6). This increase was not inhibited by U-73122 (not shown), suggesting that the response was independent of phospholipase C-mediated responses. The EC 50 value of the arachidonic acid-induced [Ca 2 ] i increase was 13.3 M (Fig. 6B), which was much larger than that in human umbilical vein endothelium (0.07 M) (15). Therefore, we conclude that the gradual increase in [Ca 2 ] i induced by hypotonic stress is mediated by arachidonic acid.

H636 HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM Fig. 5. Absence of cell damage by hypotonic stress. A: leakage of lactose dehydrogenase (LDH) into extracellular solution was measured as described in MATERIALS AND METHODS. The mean basal LDH leakage in isotonic Krebs and 0.2% Tween-20 solution were set as 0 and 100%, respectively. B: cellular uptake of ethidium bromide in the extracellar medium into the cell was measured as an indicator of microlysis. Left, intensity of 510 nm fluorescence excited by 490 nm wavelength was sampled every 30 s. No increase in ethidium bromide fluorescence was elicited even by 40% hypotonic solution. Broken line indicates the basal level of fluorescence, emitted from chamber glass and solution, which was obtained from cell-free area. Right, statistical analysis of net increment of ethidium bromide fluorescence. n.d., No significant difference from isotonic Krebs solution. DISCUSSION In the present experiment, we have shown that neomycin, U-73122 (but not the inactive analog U-73343), suramin, and apyrase inhibited the steep Ca 2 transient induced by hypotonic stress in bovine aortic endothelial cells. Therefore, the P 2 receptor-mediated production of Ins (1,4,5) is responsible for the hypotonic stress-induced Ca 2 transient. Because neither of the solutions used in the present experiment contained ATP or other Ins (1,4,5) P 3 -producing agonists, it can be excluded that a small contamination with agonist elicited the Ca 2 transient, though such a possibility was reported previously (3). Therefore, the most possible explanation for these results would be that ATP was released from the endothelium itself in the bovine aorta. Because 30 M suramin inhibited the hypotonic stress-induced steep Ca 2 transient, it can be speculated that the local [ATP] o after hypotonic stress would be between the thresholds of the ATP-induced Ca 2 transient in the absence and presence of 30 M suramin, i.e., between 10 and 100 nm. We confirmed this by measuring [ATP] o through the luciferin-luciferase assay and found that [ATP] o was increased from 2.71 nm in isotonic Krebs solution to 9.10 nm in 40% hypotonic solution. However, the local [ATP] o just above the cell membrane may be much larger than this value, because the standard photon- [ATP] relationships were obtained by homogenous ATP solutions (Fig. 4B). Therefore, to evaluate the hypotonic stress-induced ATP release more properly, we also calculated the amount of ATP molecules released from one cell, and the value was 91.0 amol/cell per 10 min in 40% hypotonic solution compared with 27.1 amol/cell per 10 min in isotonic Krebs solution. We suppose that the amount of ATP released in the isotonic Krebs solution in the present experiment might be an inevitable artifact (9) or a basal ATP release. Thus the net ATP release induced by hypotonic stress would be 63.9 amol/cell in 10 min. We have also shown by measuring LDH leakage and ethidium bromide uptake that cells were not damaged significantly even by 40% hypotonic stress (Fig. 5). This is not surprising, because the cell membrane is flexible and such an increment in cell volume is normal for the cell during cell division. Therefore, although it is generally considered that the intracellular concentration of ATP is of millimolar range (1, 24), we suppose that the release of ATP by hypotonic stress was not due to the artificial cell lysis or leakage. Even though ATP is essential for many cellular metabolic pathways, it is obvious that the loss of such a tiny amount of ATP would not affect cellular metabolism significantly. On the other hand,

HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM H637 Fig. 6. Effects of exogenously applied arachidonic acid on [Ca 2 ] i. A: arachidonic acid (30 M) was applied to a cell in Ca 2 -free isotonic solution. Gradual increase in [Ca 2 ] i was elicited. B: dose-response relationship of arachidonic acid-induced [Ca 2 ] i increase. Experiment was performed as in A, and the peak value of [Ca 2 ] i increase was measured. Data were taken from the number of cells indicated in parentheses. Continuous line represents fit to the logistical equation with an ED 50 of 13.3 M. such small amounts of ATP suffice to induce Ca 2 responses (as shown in the present study) and furthermore, we (11) have previously reported the significance of ATP-induced Ca 2 oscillations in the endothelium. The elevation of [Ca 2 ] i is known to be necessary for the production of nitric oxide (NO) (19), and indeed, ATP is reported to generate NO in endothelial cells (4). So we suppose that the mechanosensitive ATP release would play a significant physiological role in the vascular microenvironment. Hypotonic stress also induced a gradual increase in [Ca 2 ] i when the ATP-induced fast Ca 2 transient was inhibited by neomycin, suramin, or apyrase. This gradual [Ca 2 ] i increase was not due to incomplete inhibition of ATP-induced Ca 2 transient, because the time course of the hypotonic stress-induced gradual [Ca 2 ] i increase (Fig. 3A) was significantly slower than that induced by exogenously applied ATP in the presence of suramin (Fig. 3D). Furthermore, when phospholipase A 2 was inhibited by pretreatment of the cell with U-73122 (Fig. 2B) or neomycin together with pbpb (Fig. 6A), no gradual increase of [Ca 2 ] i was elicited, but rather a decrease of [Ca 2 ] i was observed, probably because of swelling-induced dilution of [Ca 2 ] i.sowe conclude that arachidonic acid contributes to the gradual [Ca 2 ] i increase in bovine aortic endothelial cells. We previously reported that arachidonic acid could induce a gradual release of Ca 2 from intracellular store sites in human umbilical vein endothelial cells (15). Other groups also reported that hypotonic stress activates phospholipase A 2 in Ehrlich cells (23). We also observed in this study that arachidonic acid induces Ca 2 release from intracellular store sites (Fig. 6B). However, the EC 50 value of the arachidonic acidinduced Ca 2 transient was much larger in bovine aortic endothelium (13.3 M) than in the human umbilical cord vein (0.07 M) (15). This may be one of the reasons why the arachidonic acid-induced [Ca 2 ] i transient was not pronounced in bovine aortic endothelial cells, whereas it was the predominant [Ca 2 ] i response in endothelial cells from the human umbilical cord vein. However, the present study could not clarify whether arachidonic acid releases Ca 2 through a selective pathway or in a nonspecific way such as a passive Ca 2 leak. Further study is needed to clarify the detailed mechanism of the arachidonic acid-induced Ca 2 release. We have shown that bovine aortic endothelium is potentially capable of producing Ca 2 signals to hypotonic stress by two different mechanisms, i.e., ATPand arachidonic acid-mediated mechanisms. We suppose that hypotonic stress always activates both ATP release and arachidonic acid production mechanisms and that the priority of ATP-induced Ca 2 release is simply because of its larger efficiency for Ca 2 mobilization than arachidonic acid in bovine aortic endothelial cells. This may also indicate that arachidonic acidinduced Ca 2 release does not play an important role under physiological conditions in the bovine aortic endothelium. If not, what is the significance of mechanosensitive arachidonic acid production? One possibility is that the production of arachidonic acid would lead to the generation of prostaglandins, the importance of which as endothelium-derived mediators are now widely considered (13). Furthermore, the endothelium is known to regulate vascular tonus by releasing NO and endothelium-derived hyperpolarizing factor (EDHF). Metabolites of arachidonic acid have been reported repeatedly as a candidate of EDHF (10, 20, 21), although another candidate, K, has also been proposed recently (7). Therefore, production of arachidonic acid by mechanical stress may lead to the release of EDHF from endothelial cells. Actually, it has been reported recently that EDHF is released by pulsatile flow in the porcine coronary artery (20). So this would be another possibility of the physiological significance of hypotonic stress-induced production of arachidonic acid.

H638 HYPOTONIC CA 2 TRANSIENT IN AORTIC ENDOTHELIUM In conclusion, we have shown the dual responses of hypotonic stress-induced Ca 2 mobilization in cultured bovine aortic endothelial cells: prominent ATP-induced Ca 2 oscillation and arachidonic acid-induced gradual [Ca 2 ] i increase. Each of these substances would have physiological roles, such as the generation of NO and EDHF, and warrants further investigation as to the significance of these mechanical stress-induced responses. We thank Drs. B. Nilius and G. Droogmans for the critical reading of the manuscript. This study was carried out as part of the Ground Research Announcement for the Space Utilization promoted by the National Space Development Agency of Japan and Japan Space Forum (to M.Oike) and as part of the Fund for Basic Experiments Oriented to Space Station Utilization promoted by the Institute of Space and Astronautical Science of the Ministry of Education of Japan. This study was also supported in part by a Grant-In-Aid from the Ministry of Education of Japan. REFERENCES 1. Allue I, Gandelman O, Dementieva E, Ugarova N, and Cobbold P. Evidence for rapid consumption of millimolar concentrations of cytoplasmic ATP during rigor-contracture of metabolically compromised single cardiomyocytes. Biochem J 319: 463 469, 1996. 2. Barakat AI, Leaver EV, Pappone PA, and Davies PF. A flow-activated chloride-selective membrane current in vascular endothelial cells. Circ Res 85: 820 828, 1999. 3. Bodin P and Burnstock G. ATP-stimulated release of ATP by human endothelial cells. J Cardiovasc Pharmacol 27: 872 875, 1996. 4. Bogle RG, Coade SB, Moncada S, Pearson JD, and Mann GE. Bradykinin and ATP stimulate L-arginine uptake and nitric oxide release in vascular endothelial cells. Biochem Biophys Res Commun 180: 926 932, 1991. 5. Cass WA, Larson G, Fitzpatrick FA, and Zahniser NR. Inhibitors of arachidonic acid metabolism: effects on rat striatal dopamine release and uptake. J Pharmacol Exp Ther 257: 990 996, 1991. 6. Davies PF. Flow-mediated endothelial mechanotransduction. Physiol Rev 75: 519 560, 1995. 7. Edwards G, Dora KA, Gardener MJ, Garland CJ, and Weston AH. K is an endothelium-derived hyperpolarizing factor in rat arteries. Nature 396: 269 272, 1998. 8. Folkman J and Shing Y. Angiogenesis. J Biol Chem 267: 10931 10934, 1992. 9. Grygorczyk R and Hanrahan JW. CFTR-independent ATP release from epithelial cells triggered by mechanical stimuli. Am J Physiol Cell Physiol 272: C1058 C1066, 1997. 10. Hecker M, Bara AT, Bauersachs J, and Busse R. Characterization of endothelium-derived hyperpolarizing factor as a cytochrome P450-derived arachidonic acid metabolite in mammals. J Physiol (Lond) 481: 407 414, 1994. 11. Kimura C, Oike M, and Ito Y. Acute glucose overload abolishes Ca 2 oscillation in cultured endothelial cells from bovine aorta: a possible role of superoxide anion. Circ Res 82: 677 685, 1998. 12. Knudsen HL and Frangos JA. Role of cytoskeleton in shear stress-induced endothelial nitric oxide production. Am J Physiol Heart Circ Physiol 273: H347 H355, 1997. 13. Luscher TF, Boulanger CM, Dohi Y, and Yang ZH. Endothelium-derived contracting factors. Hypertension 19: 117 130, 1992. 14. Milner P, Kirkpatrick KA, Ralevic V, Toothill V, Pearson J, and Burnstock G. Endothelial cells cultured from human umbilical vein release ATP, substance P and acetylcholine in response to increased flow. Proc R Soc Lond B Biol Sci 241: 245 248, 1990. 15. Oike M, Droogmans G, and Nilius B. Mechanosensitive Ca 2 transients in endothelial cells from human umbilical vein. Proc Natl Acad Sci USA 91: 2940 2944, 1994. 16. Oike M and Ito Y. Dynamic regulation of intracellular Ca 2 concentration in aortic endothelial cells. Eur J Pharmacol 319: 291 298, 1997. 17. Oike M, Schwarz G, Sehrer J, Jost M, Gerke V, Weber K, Droogmans G, and Nilius B. Cytoskeletal modulation of the response to mechanical stimulation in human vascular endothelial cells. Pflügers Arch 428: 569 576, 1994. 18. Olesen SP, Clapham DE, and Davies PF. Haemodynamic shear stress activates a K current in vascular endothelial cells. Nature 331: 168 170, 1988. 19. Palacios M, Knowles RG, Palmer RM, and Moncada S. Nitric oxide from L-arginine stimulates the soluble guanylate cyclase in adrenal glands. Biochem Biophys Res Commun 165: 802 809, 1989. 20. Popp R, Fleming I, and Busse R. Pulsatile stretch in coronary arteries elicits release of endothelium-derived hyperpolarizing factor: a modulator of arterial compliance. Circ Res 82: 696 703, 1998. 21. Randall MD and Kendall DA. Endocannabinoids: a new class of vasoactive substances. Trends Pharmacol Sci 19: 55 58, 1998. 22. Schwarz G, Callewaert G, Droogmans G, and Nilius B. Shear stress-induced calcium transients in endothelial cells from human umbilical cord veins. J Physiol (Lond) 458: 527 538, 1992. 23. Thoroed SM, Lauritzen L, Lambert IH, Hansen HS, and Hoffmann EK. Cell swelling activates phospholipase A 2 in Ehrlich ascites tumor cells. J Membr Biol 160: 47 58, 1997. 24. Wu ML and Vaughan-Jones RD. Effect of metabolic inhibitors and second messengers upon Na -H exchange in the sheep cardiac Purkinje fibre. J Physiol (Lond) 478: 301 313, 1994.