Antagonism Increase

of Platelet-Activating in Cytosolic in Human

Free

Factor-Induced

Calcium

Endothelial

Concentration Cells

Masahiko Hirafuji and Hisashi Shinoda Department of Pharmacology, Tohoku University School of Dentistry, 4-1 Seiryo-machi, Aoba-ku, Sendai 980, Japan Received

June

29, 1991

Accepted

December

9, 1991

ABSTRACT-The effects of platelet-activating factor (PAF) antagonists on the agonist-induced increase in cytosolic free calcium concentration, [Ca2+]i, in human vascular endothelial cells grown in monolayer were investigated by a continuous su perfusion technique using a calcium fluorescent probe, fura-2. PAF caused a small but dose-dependent increase in [Ca2+];. Seven structurally dissimilar PAF antagonists dose-dependently suppressed the peak response, among which WEB 2086 was the most potent, followed by WEB 2170 > FR 900452 ONO 6240 > BN 52021 kad surenone CV 3988. These antagonists except for CV 3988 were specific for PAF, since they had no effects on calcium mobilization induced by thrombin or histamine, while CV 3988 had a non-specific effect. PAF in the same range of concentration in creased prostacyclin release from human endothelial cells. WEB 2086 also inhibited the PAF-induced prostacyclin release, while it had no effect on the release induced by histamine and thrombin. These results demonstrate the specificity and dose-response characteristics of PAF antagonists in cultured human endothelial cells and suggest that a PAF antagonist could be a valuable therapeutic agent in certain human diseases where PAF activation of endothelial cells may have a critical role.

Platelet-activating factor (PAF) is an ether phospholipid with diverse and potent biologi cal activities suggesting an important role in a variety of pathophysiological conditions such as acute inflammation, endotoxic shock and acute allergic diseases (1). The interaction of this lipid mediator with vascular endothelium could be important in vascular diseases, since the endothelium plays an important role in hemostasis and in the modulation of vascular tone and permeability (2). Several reports in dicate a direct agonistic action of PAF on vascular endothelial cells in culture. PAF in duces a dose-dependent synthesis of prostacy clin (3), and it alters the molecular organiza

tion of cytoskeletal proteins that control en dothelial permeability (4) in cultured human endothelial cells. In bovine pulmonary artery endothelial cells, PAF induces significant al teration in morphology as well as metabolic functions such as phosphoinositide turnover, -adrenergic receptor adenyl ate cyclase sys tem, and prostacyclin and thromboxane re lease (5). PAF also induces alterations in cal cium fluxes in cultured endothelial cells (6, 7). Since the initial description of CV 3988 as a selective PAF receptor antagonist (8), a large number of compounds have been reported as selective PAF antagonists (1). These com pounds have been evaluated primarily on the

basis of inhibition of PAF-induced platelet aggregation or [3H]-PAF binding to platelets. However, despite the multiple functional re sponses evoked by PAF in endothelial cells as described above and possible therapeutic advantage of the antagonists, few studies have been made to evaluate the potency of PAF antagonists in endothelial cells. In a previous report, we showed that PAF increased cytoso lic free calcium concentration, [Ca21]i, in hu man endothelial cells cultured in monolayer by a continuous superfusion technique using a calcium fluorescence dye, fura-2 (9). An in crease in [Ca21]i is an essential initial event coupled to cell activation. Therefore, in the present study, the potencies of seven structur ally dissimilar PAF antagonists were evaluated by measuring PAF-induced increase in [Ca2+]i to determine if PAF antagonists primarily re ported and characterized with respect to platelet aggregation are also effective and spe cific in human endothelial cells. MATERIALS AND METHODS Culture of human endothelial cells Human endothelial cells were isolated from umbilical cord veins and cultured essentially as described previously (10). Briefly, cells were collected by 0.1% collagenase treatment, seeded onto coverglasses (8 X 16 mm) and grown until confluence (3 5 days) in Ham's F-12 culture medium supplemented with 15% heat-inactivated fetal bovine serum, 1% Ultrosel G, a serum substitute, penicillin (50 U/ml) and streptomycin (50 ,ug/ml). Measurement of cytosolic free calcium concen tration The cytosolic free calcium concentration, [Ca2+]i, of human endothelial cells grown in monolayer was measured by a continuous su perfusion technique as described previously (9, 11), using a fluorescent calcium indicator, fura-2/AM (12). During the measurement procedure, endothelial cells loaded with fura-2 were continuously superfused with Hanks' bal anced salt solution (pH 7.4, HBSS) containing

10 mM HEPES and 0.1% bovine serum albu min (BSA) at the rate of 1 ml/min. When PAF antagonists were used, cells were con tacted with the antagonist for 3 min prior to and during the PAF stimulation. The fura-2 fluorescence was monitored by a fluorescence spectrophotometer (Hitachi F-4000, Japan) coupled to a microcomputer (NEC PC-9801, Japan), which was programmed to shuttle be tween the two excitation wavelengths of 340 and 380 nm while keeping the emission wavelength of 505 rim. Rm;n, the minimum fluorescence ratio, was measured by adding 10 ,uM ionomycin followed by perfusion with calcium-free HBSS containing 1 mM EGTA to chelate free calcium. Rmax, the maximal ratio, was obtained by exposing the cells to 0.05% Triton X-100 with a temporary stop in the perfusion, thus permeabilizing the cells to cal cium and ensuring that essentially all fura-2 was complexed with calcium. The cells were finally superfused with 0.05% Triton X-100 to obtain the autofluorescence. After the subtrac tion of autofluorescence, [Ca21]i was calcu lated as described by Grynkiewicz et al. (12) based on the equation: [Ca2+]; = Kd(R

Rmin)/(Rmax R)(Sf2/Sb2)

and assuming the Kd of the Ca2+-fura-2 in teraction to be 225 nM in the cytosolic en vironment (12). Measurement of prostacyclin Endothelial cells grown to confluence in 24 well culture plates (Primaria, Falcon, U.S.A.) were washed 3 times with 0.5 ml of the Hanks' solution. Cells were then incubated for 10 min at 37°C without (control) or with different concentrations of PAF in a total volume of 0.4 ml of the buffer. When a PAF antagonist was used, cells were contacted with the antagonist for 3 min prior to and during the PAF stimula tion. At the end of the experiment, cell num bers in the selected wells were counted in a haemocytometer after trypsin digestion. Pros tacyclin released into the incubation medium was determined by radioimmunoassay as its stable metabolite 6-keto-PGFla (13).

Aggregation of washed rabbit platelets Blood was collected from the central ear artery of unanaesthetized adult male rabbits and immediately mixed with EDTA (5 mM, final concentration). Platelet-rich plasma (PRP) was obtained by centrifugation (20 min at 500 X g) at room temperature and further centrifuged for 15 min at 1,500 X g to obtain the platelet pellet. Platelets were resuspended in calcium-free Tyrode buffer (137 mM NaCI, 2.6 mM KCI, 0.4 mM NaH2PO4, LO MM MgCl2, 5.6 mM glucose, pH 6.5) containing 0.25% BSA, 10 mM HEPES and 0.2 mM EGTA, washed twice with the same buffer, and counted using a Celltac Counter (Nihon Kohden MEK-4150, Japan). Platelets were finally resuspended at a concentration of 5 X 108platelets/ml in the Tyrode buffer (pH 7.4) containing calcium (1.3 mM) and no EGTA and kept at room temperature. The aggrega tion study was carried out at 37°C using an aggregometer (Chronolog C-440, U.S.A.). Aggregation responses were allowed to de velop until a maximum change in light trans mission occurred. Antagonists were added to the cuvette 1 min before the addition of PAF. Materials PAF (1-O-hexadecyl-2-acetyl-sn-glycero-3 phosphocho line), bovine serum albumin (BSA, fatty acid-free) and bovine thrombin were obtained from Sigma, U.S.A.; ionomy cin, from Calbiochem, U.S.A.; HEPES and fura-2 acetoxymethyl ester (fura-2/AM), from Dojin, Japan; collagenase (Clostridium histoly ticum) and histamine dihydrochloride, from Wako, Japan; Hanks' balanced salt solution, Ham's F-12 culture medium, fetal bovine serum, penicillin and streptomycin, from Gib co, U.S.A.; Ultroser G, from IBF, France. Kadsurenone (2-(3,4-dimethoxyphenyl)-2,3-di hydro-3a-methoxy-3-methyl-5-(allyl)-6-2H-oxo benzofuran) was a kind gift from Dr. T.Y. Shen, Merck, Sharp & Dohme, U.S.A.; BN 52021 (3-tert-butyl-hexahydro-4,7b,11-trihydro xy -8-methyl-9H-1,7a -epoxymethano-1H,6aH cyclopenta[c]furo[2,3-b]furo[3',2':3,4]cyclopen ta[1,2-d]furan-5,9,12(4H)trione), from Dr.

J.M. Mencia-Huerta, Institut Henri Beaufour, France; CV 3988 (rac-3-(N-n-octadecylcarbam oyloxy)-2-methoxypropyl 2-thiazolioethyl phos phate), from Dr. K. Nishikawa, Takeda Chem ical Indust. Japan; ONO 6240 (1-O-hexa decyl 2RS -O -ethyl-3-O -(7 thiazolinoheptyl) glycerol methanesulfonate), from Ono Phar maceut. Co., Ltd., Japan; WEB 2086 (3-[4-(2 chlorophenyl)-9-methyl-6H-thieno[3,2 f ] [1,2,4] triazolo [4,3 a] [1,4]-diazepine-2 yl]-1-(4-mor pholinyl)-1-propanone) and WEB 2170 (6-(2 chlorophenyl) 8,9 -dihydro-l-methyl -8-(4-mor pholinyl -carbonyl) 4H,7H -cyclopenta[4,5]thi eno[3,2 j][1,2,4]-triazolo-[4,3-a][1,4]daizepine), from Dr. H. Heuer, Boehringer Ingelheim KG, FRG; FR 900452 (1-methyl-3-(1-(5-me thylthiomethyl-6-oxo-3-(2-oxo-3-cyclopenten-l ylidene)-2-piperazinyl)ethyl)-2-indolinone) eth azino[1,2-a]indole-1,6-dione), from Dr. M. Okamoto, Fujisawa Pharmaceut. Co., Ltd., Japan. The structural formulas of these PAF antagonists are shown in Fig. 1. [3H]-6-Keto PGFIa was purchased from NEN, U.S.A. The standard compound 6-keto-PGFI0 and its anti serum were kindly provided by Ono Phar maceut. Co., Ltd., Japan. PAF was dissolved in physiological saline containing 0.25% BSA at a concentration of 10 mM and stored at -20°C. Kadsurenone and BN 52021 were dissolved in dimethylsulfo xide (DMSO) at a concentration of 20 mM and stored at -20°C. FR 900452 was dissolved in ethanol freshly on the day of experiment. Final concentration of DMSO and ethanol never exceeded 0.3%, which had no effects on [Ca2+]i change. Other PAF antagonists were dissolved in physiological saline. These stock solutions were diluted to the desired concen trations with the incubation buffer before use. Statistical analyses Results were expressed as means ± S.E.M. Student's t-test was used to determine the sig nificance of difference between means, and P values less than 0.05 were considered to indi cate a significant difference.

Fig.

1.

Structural

formulas

of PAF

antagonists

studied.

RESULTS Changes in [Ca2+] of human endothelial cells in response to PAF, thrombin and histamine As illustrated in Fig. 2, PAF induced a rapid and transient elevation in [Ca2+]i in hu man umbilical cord vein endothelial cells, which was maximal by 1 min after the stimula tion and returned progressively to the basal level within the next 3 4 min. This calcium mobilizing effect was dose-dependent, and the threshold concentration of PAF was 0.1-1 nM, depending on the batch of cells. The maximal response was observed with 100 nM PAF, which caused a significant increase in [Ca2+]i from the basal level of 31.9 ± 3.5 to the peak of 85.4 ± 10.2 nM (mean ± S.E.M., n=8). Thrombin (0.001-1 U/ml) and histamine (0.01-10 ,uM) also produced a rapid and concentration-dependent increase in [Ca2+];. However, the peak [Ca2+]i elevations in re sponse to these agonists were of much higher intensity as compared to those induced by PAF. In addition, no return of [Ca2+]i level to the prechallenge value was observed with higher concentrations of thrombin and hista mine (Fig. 2).

Fig. 2. Effects of PAF, thrombin and histamine on [Ca 2+]i in human endothelial cells cultured in mono layers. The trace is a change in [Ca2+]i value calcu lated from fluorescence changes of fura-2. The result for each agonist was obtained from each independent experiment using a different batch of the cells, but is a representative of at least three experiments.

Suppression of [Ca2+], increase by PAF antago nists The antagonistic properties of the seven structurally dissimilar PAF antagonists shown in Fig. 1 were evaluated by comparing the peak level in [Ca2+]i induced by 100 nM PAF in the presence or the absence of the antago

nist. As illustrated in Fig. 3, a PAF antagonist, WEB 2086 suppressed the maximal PAF-in duced increase in [Ca2+]i level of human en dothelial cells in a concentration-dependent manner. Figure 4 summarizes the mean per centage inhibition produced by the seven PAF antagonists studied. Among these antagonists,

Fig. 3. Effect of the PAF antagonist WEB 2086 on PAF-induced increase in [Ca 2+]; of human endothelial cells. Cells were superfused without (control) or with the antagonist at the indicated concentrations for 3 min prior to and during the stimulation with 100 nM PAF (arrows). The result is a representative of four experiments.

Fig. 4. Effects of several PAF antagonists on PAF-induced increase in [Ca 2+]i of human endothelial cells. Results were expressed as % of the control increase induced by 100 nM PAF in each experiment. Each point represents the mean of 3-4 determinations performed in four experiments; vertical lines show S.E.M. Symbols: WEB 2086 (L), WEB 2170 (0), FR 900452 (A), ONO 6240 (40), BN 52021 (*), CV 3988 (o), kadsurenone (•).

WEB 2086 was the most potent with an IC50 value of 23.1 ± 2.9 nM (n = 4) followed by WEB 2170 > FR 900452 ONO 6240 > BN 52021 _ kadsurenone CV 3988. IC50 values for the seven antagonists are presented in Table 1. Six of the tested antagonists induced no change in the resting [Ca2+]i level at the con centration used, and only FR 900452 caused a decrease in basal [Ca2+]i value at concentra tions higher than 3,uM. The effect of FR 900452 seemed to be a physico-chemical effect because this compound caused a quenching of the fluorescence of fura-2 excited at a wave length of 340 nm. Therefore, the effect of this antagonist higher than 3 ,uM was not evalu ated. Also, none of the compounds: WEB 2086, WEB 2170 and FR 900452 at a concentration of 1 pM, ONO 6240 at 10 ,uM, BN 52021 and kadsurenone at 100 pM, and CV 3988 at 30 ,uM suppressed the peak responses of [Ca 2+]i induced by 0.01 U/ml thrombin and 0.1,uM histamine, which significantly increased the [Ca2+]i level from the basal levels of 35.8 ± 3.9 nM and 38.2 ± 4.5 nM to the peak levels of 204.7 ± 30.3 nM and 189.6 ± 20.5 nM (n = 4) , respectively. However, as illustrated in Fig. 5, CV 3988 (30,uM) caused a disappear ance of the sustained phase following the

rapid peak response in [Ca2+], change observed with a higher concentration of hista mine (1 and 10 pM) without the effect on the rapid peak response, while the other antago nists did not. PAF-induced prostacyclin release PAF stimulated prostacyclin release from human vascular endothelial cells in a concentration-dependent manner. The stimula tory effect was observed in the same concen tration range as that inducing an increase in [Ca2+];. However, the basal release and the maximal response of prostacyclin induced by 100 nM PAF were variable depending on the batch of cells (0.04 ± 0.02 and 0.50 ± 0.12 ng/105 cells, n = 4 in triplicate, respectively). Figures 6(A) and 6(B) illustrate representative results on the concentration and time-de pendent effect of PAF on prostacyclin release from endothelial cells, respectively. As demonstrated in Fig. 7, the PAF an tagonist WEB 2086 showed a dose-dependent inhibitory effect on prostacyclin release in duced by 100 nM PAF with an IC50 value of 44.2 ± 8.1 nM (n = 4, in triplicate). Hista mine at a concentration of 0.1 u M and throm bin at a concentration of 0.01 U/ml signifi cantly stimulated prostacyclin release from 0.05 ± 0.01 to 0.66 ± 0.05 and 1.89±0.11

Table 1. Antagonism of PAF-induced increase in [Ca 2+]; of human endo thelial cells and aggregation of washed rabbit platelets by several PAF an tagonists

Fig. 5. Effect of CV 3988 on histamine-induced calcium mobilization in human endothelial cells. Cells were stimulated with 10,uM histamine (arrows), and 30 pM CV 3988 (closed circle) was added 3 min prior to the stimulation. Note that CV 3988 caused a disappearance of the sustained phase following the rapid peak response. [Ca2+]i levels at 5 min after histamine challenge were 273.2 and 69.3 nM for non-treated and CV 3988-treated cells, respectively. The result is a representative of three experiments, and the other ex periments using different batches of cells showed a similar result.

Fig. 6. Effect of PAF on prostacyclin release from human endothelial cells. Endothelial cells grown to confluence in a 24-well culture plate were stimulated for 10 min without (0) or with (0) PAF at the indi cated concentrations (A) or were stimulated without (0) or with (0) 100 nM PAF for the indicated time (B). Prostacyclin released into the incubation medium was determined by radioimmunoassay. Each result represents the mean of 3 determinations of a repre sentative of three experiments; vertical lines show S.E.M.

Fig. 7. Effect of WEB 2086 on PAF-induced pros tacyclin release from human endothelial cells. Endo thelial cells were preincubated for 3 min without (con trol) or with WEB 2086 at the indicated concentra tions, and then stimulated with 100 nM PAF. Results were expressed as a % of the control increase induced by PAF in each experiment. Each point represents the mean of 4 experiments performed in triplicate; vertical lines show S.E.M.

ng/105 cells (mean ± S.E.M. of a representa tive experiment measured in triplicate), re spectively. WEB 2086 at a concentration of I ,uM, which completely inhibited PAF-induced prostacyclin release, caused no significant in hibition of prostacyclin release induced by these vasoactive agents (data not shown). Effects of PAF antagonists on platelet aggrega tion PAF caused a concentration-dependent aggregation of washed rabbit platelets, with the mean maximum increase in light transmis sion of 80.1 ± 3.5% (n = 8) induced by 10 nM PAF. The effect of antagonists on PAF-in duced platelet aggregation was evaluated using submaximal 1 nM PAF, which caused a 73.8 ± 1.1% (n = 8) increase in light transmission. All PAF antagonists used in the present study inhibited PAF-induced aggregation in a concentration-dependent manner, and the IC50 values obtained are presented in Table 1. The rank order of potency was found to be WEB 2086 > WEB 2170 > ONO 6240 > FR 900452 > kadsurenone > BN 52021 > CV 3988. DISCUSSION In a previous report, using the calcium fluorescence dye fura-2 (12), we showed that PAF increased cytosolic free calcium concen tration in human endothelial cells cultured in monolayer mainly by discharging calcium from its intracellular stores (9). In the present study, the potency of PAF antagonists on the agonist-induced endothelial cell activation was assessed by evaluating PAF-induced calcium mobilization, an early event coupled to cell activation. The PAF antagonists used in this study are as follows: PAF analogues, CV 3988 (8) and ONO 6240 (14); terpene, BN 52021 (15); lignan, kadsurenone (16); gliotoxin, FR 900452 (17); triazolobenzodiazepines, WEB 2086 (18) and WEB 2170 (19). All these com pounds have been described to be specific PAF antagonists primarily with respect to platelet activation or [3H]-PAF binding to platelets. In our hands, indeed, these com

pounds antagonized PAF-induced platelet aggregation of washed rabbit platelets with the following order of potency: WEB 2086 > WEB 2170 > ONO 6240 > FR 900452 > kad surenone > BN 52021 > CV 3988. The pres ent study demonstrated that all of these an tagonists suppressed the PAF-induced increase in [Ca2+]i in endothelial cells in a dose-depend ent manner. Among these antagonists, WEB 2086 was the most potent and the order of potency was found to be WEB 2086 > WEB 2170 > FR 900452 ONO 6240 > BN 52021 kadsurenone = CV 3988. The existence of subtypes and species differ ence of PAF receptor has been recently pro posed based on the differences in the potencies of the antagonists (20 22). Here, while the rank order of potencies of the antagonists was similar for the two cell types when analyzed by linear regression analysis (r = 0.95, P < 0.001), the relative potencies of the antago nists between these cell types were much dif ferent. This might suggest a possibility that some antagonists would be more potent against endothelial cells and some more potent against platelets. However, this particular problem may require more precise study obtaining pA2 values, because the apparent potencies of antagonists as measured by IC50 values are highly dependent on the agonist concentration. In the present study, PAF has been used at 100 nM in endothelial cells but at 1 nM in platelets; in addition, other different conditions have been employed such as BSA concentration, antagonist preincubation, an tagonist superfusion, etc. As emphasized by O'Donnell and Barnett (23), the valid com parison of antagonist potencies should be directly made with the pA2 values. In the pres ent study, no attempt was made to establish the pA2 values because of the insufficient number of primary human endothelial cells. Histamine and thrombin induced a dose-de pendent and marked increase in [Ca2+]i of hu man endothelial cells cultured in monolayer. The specificity of PAF antagonists were evalu ated against the responses to histamine and thrombin at the supra-threshold concentrations

producing changes in [Ca2+]i. All antagonists were found to be specific PAF antagonists in human endothelial cells, since these com pounds did not inhibit calcium mobilization in duced by histamine or thrombin. However, CV 3988 seemed to have a nonspecific effect on calcium mobilization, since it was observed that this compound (30,uM) caused a dis appearance of the sustained phase following the maximal change in [Ca 2+]i level induced by high concentration of histamine, while it had no effect on the rapid peak response. A previous study showed that lyso PAF, a deacetylated metabolite of PAF, was unable to induce calcium mobilization, and that PAF caused a rapid and homologous desensitization of the PAF response (9). Together with these results, the present study support the idea that PAF-induced calcium mobilization is mediated via specific membrane receptors on human endothelial cells; and furthermore, the present results indicate that the mechanism of action of PAF antagonists is due to the ability of these compounds to block PAF binding and hence suppress PAF-induced calcium mobiliza tion. The presence of specific PAF receptors on human endothelial cells is further suggested by a binding study demonstrating the presence of specific binding sites for [3H]-PAF (24). The present study further demonstrated that PAF in the same concentration range inducing the intracellular calcium mobilization also stimulated prostacyclin release. The maximal [Ca 2+]i response induced by PAF in human endothelial cells was observed 1 min after the stimulation, while prostacyclin release reached the plateau by 10 min, indicating that the cal cium mobilization preceded the mediator re lease. Therefore, the PAF-induced increase in [Ca21]i may represent an early event coupled to the mediator release. The potent PAF an tagonist, WEB 2086 dose-dependently inhib ited prostacyclin release induced by PAF but not by thrombin or histamine, suggesting the specificity of the antagonist and the involve ment of PAF specific receptors in the func tional response. Consistent with our result, D'Humieres et

al. (3) have shown that PAF stimulates the re lease of prostacyclin from human endothelial cells, while others have reported that PAF has no effect (6, 25). The conflicting results may be due to the culture conditions, in particular, with or without the subcultivation of the cells. Indeed, in our four experiments in which cells after the first passage were used, PAF did not stimulate prostacyclin release (data not shown), while it reproducibly stimulated the release when the primary endothelial cells were used. Furthermore, in agreement with Hallam et al. (25), PAF induced only a 2-4 fold increase over the basal level of [Ca2+]i even in primary endothelial cells, and this re sponse was much smaller than that of hista mine or thrombin. However, as potential roles of endothelial cells as target cells of PAF in a diverse range of human pathologies have been suggested (1, 26), the responsiveness of hu man endothelial cells to PAF may be rapidly diminished by their culture. Finally, it should be mentioned that the [Ca2+]i in resting human endothelial cells cali brated in the present study (34.5 ± 3.9 nM, n = 16) is less than half of the level reported by other workers (around 100 nM) in these cells loaded with fura-2 or quin-2 (25, 27-29). However, it should be noted that the value of [Ca 2+]i derived from the equation of Grynk iewicz et al. (12) is not absolute, because it has been reported that the Kd value is variable according to the ionic strength, pH, magne sium concentration, and the concentration of cytosolic proteins (30, 31). Moreover the vis cosity affects the fluorescent intensity (31, 32). All these factors in the cytosolic environment might not be constant but vary as a function of [Ca 2+1i. In addition, while a homogeneous and selective distribution of fura-2 dye in the cytosol is a prerequisite for accurate measure ment of [Ca2+]i, fura-2 fluorescence has been shown to be localized to the mitochondria in endothelial cells (33), a fact which might affect Rmax and Rmin values. In the present study, we calibrated [Ca2+]i using Rmaxand Rmin de termined for each cell preparation in situ. When, however, Rmaxand Rmin values deter

mined in fura-2 dye standard solution contain ing saturating Ca2-1 (1.3 mM) and Ca 21-free with EGTA, respectively, were used, the rest ing [Ca2+]i was calculated to be 84.8 ± 4.3 nM (n = 8). Although several explanations might be possible for this difference, it would not affect the conclusions of the present study. Most importantly, despite the low resting [Ca2+]i value, the changes of [Ca2+]i caused by vasoactive mediators and the PAF antago nists were dose-dependent and reproducible. In conclusion, the present study demon strates the specificity and dose-response charac teristics of seven structurally dissimilar PAF antagonists in cultured human endothelial cells, and it suggests that PAF antagonists could be a valuable therapeutic agent in cer tain human diseases where PAF activation of endothelial cells may have a critical role. These data should assist their use for evaluat ing the potential role of PAF in vascular pathophysiology and the potential of the an tagonists as therapeutic agents. Acknowledgments The authors are grateful to the individuals and com panies indicated in the manuscript for their generous donations of PAF antagonists. The authors are also grateful to Dr. T. Adachi and the staff at Adachi Maternity Hospital (Sendai) for providing the umbilical cords; Drs. K. Maeyama and M. Sasaki, Department of Pharmacology I, Tohoku University School of Medi cine, for helpful discussion and suggestions for calcium measurement; and Mr. T. Kikuchi, for technical assist ance. This work was supported in part by a Grant-in Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan. REFERENCES 1 Vargaftig, B.B., Pretolani, M., Coeffier, E. and Chignard, M.: Platelet activating factor: biology, receptors and antagonists. In Handbook of In flammation, Vol. 6, Mediators of the Inflammatory Process, Edited by Henson, P.M. and Murphy, R.C., p. 113-146, Elsevier, Amsterdam (1989) 2 Vane, J.R., Anggard, E.E. and Botting, R.M.: Regulatory functions of the vascular endothelium. N. Engl. J. Med. 323, 27-36 (1990) 3 D'Humieres, S., Russo-Marie, F. and Vargaftig,

4

5

6

B.B.: PAF-acether induced synthesis of prostacy clin by human endothelial cells. Eur. J. Pharma col. 131, 13-19 (1986) Bussolino, F., Camussi, G., Aglietta, M., Bra quet, P., Bosia, A., Pescarmona, G. et al.: Hu man endothelial cells are target for platelet activat ing factor. I. Platelet-activating factor induces changes in cytoskeleton structures. J. Immunol. 139, 2439-2446 (1987) Grigorian, G.Y. and Ryan, U.S.: Platelet-activat ing factor effects on bovine pulmonary artery en dothelial cells. Circ. Res. 61, 389 395 (1987) Bussolino, F., Aglietta, M., Sanavio, F., Stacchini, A., Lauri, D. and Camussi, G.: Alkyl-ether phos

phoglycerides influence calcium fluxes into human endothelial cells. J. Immunol. 135, 2748-2753 (1985) 7 Brock, T.A. and Gimbrone, M.A., Jr.: Platelet activating factor alters calcium homeostasis in cul tured vascular endothelial cells. Am. J. Physiol. 250, H1086-H1092 (1986) 8 Terashita, Z., Tsushima, S., Yoshioka, Y., Momura, H., Inada, Y. and Nishikawa, K.: CV 3988-A specific antagonist of platelet activating factor (PAF). Life Sci. 32, 1975-1982 (1983) 9 Hirafuji, M., Maeyama, K., Watanabe, T. and Ogura, Y.: Transient increase of cytosolic free cal cium in cultured human vascular endothelial cells by platelet-activating factor. Biochem. Biophys. Res. Commun. 154, 910-917 (1988) 10 Hirafuji, M., Mencia-Huerta, J.M. and Benve niste, J.: Regulation of PAF-acether (platelet-acti vating factor) biosynthesis in cultured human vascular endothelial cells stimulated with throm bin. Biochim. Biophys. Acta 930, 359-369 (1987) 11 Maeyama, K., Sasaki, M. and Watanabe, T.: Simultaneous determination of intracellular cal cium concentration and histamine secretion in rat basophilic leukemia cells (RBL-2H3). Anal. Biochem. 194, 316-325 (1991) 12 Grynkiewicz, G., Poenie, M. and Tsien, R.Y.: A new generation of calcium indicators with greatly improved fluorescence properties. J. Biol. Chem. 260, 3440-3450 (1985) 13 Hirafuji, M., Akiyama, Y. and Ogura, Y.: Receptor-mediated stimulation of aortic prostacy clin release by 5-hydroxytryptamine. Eur. J. Phar macol. 143, 259-265 (1987) 14 Miyamoto, T., Ohno, H., Yano, T., Okada, T., Hamanaka, N. and Kawasaki, A.: ONO-6240: a new potent antagonist of platelet-activating factor. Adv. Prostaglandin, Thromboxane Leukotriene Res. 15, 719-720 (1985) 15 Braquet, P., Spinnewyn, B., Braquet, M., Bour

gain, R.H., Taylor, J.E., Etienne, A. et al.: BN 52021 and related compounds: a new series of highly specific PAF-acether receptor antago nists isolated from Ginkgo biloba. Blood Vessels 16, 559-572 (1985) 16 Shen, T.Y., Hwang, S.-B., Chang, M.N., Doeb ber, T.W., Lam, M.H., Wu, M.S. et al.: Charac terization of a platelet-activating factor receptor antagonist isolated from haifenteng (Piper futokad sura): specific inhibition of in vitro and in vivo platelet-activating factor-induced effects. Proc. Natl. Acad. Sci. U.S.A. 82, 672-676 (1985) 17 Okamoto, M., Yoshida, K., Nishikawa, M., Ando, T., Iwami, M., Kohsaka, M. et al.: FR 900452, a specific antagonist of platelet activating factor (PAF) produced by Streptomyces phaeofa ciens. J. Antibiot. (Tokyo) 39, 198-204 (1986) 18 Casals-Stenzel, J., Muacevic, G. and Weber, K.H.: Pharmacological actions of WEB 2086, a new specific antagonist of platelet activating fac tor. J. Pharmacol. Exp. Ther. 241, 974-981 (1987) 19 Heuer, H.O.: Action of the new hetrazepinoic PAF-antagonist WEB 2170 on shock states induced by PAF, antigen or endotoxin/TNF. Arch. Phar macol. 339, Supp. R78 (1989) 20 Voelkel, N.F., Chang, S.-W., Pfeffer, K.D., Worthen, S.G., McMurtry, I.F. and Henson, P.M.: PAF antagonists: different effects on platelets, neutrophils, guinea-pig ileum and PAF induced vasodilatation in isolated rat lung. Pros taglandins 32, 359-372 (1986) 21 Stewart, A.G. and Dusting, G.L.: Characteriza tion of receptors for platelet-activating factor on

22

23

24

platelets, polymorphonuclear leukocytes and mac rophages. Br. J. Pharmacol. 94, 1225-1233 (1988) Hwang, S.-B. and Lam, M.-H.: Species difference in the specific receptors of platelet activating fac tor. Biochem. Pharmacol. 35, 4511-4518 (1986) O'Donnell, S.R. and Barnett, C.J.K.: pA2 values for antagonists of platelet activating factor on aggregation of rabbit platelets. Br. J. Pharmacol. 94, 437-442 (1988) Korth, R., Hirafuji, M., Bidault, J., Canton, B., Russo-Marie, F. and Benveniste J.: Comparison between paf-acether receptors on intact washed human platelets and human endothelial cells in

25

26

27

28

29

30

31

32

33

culture. In Vascular Endothelium; Receptors and Transduction Mechanisms, Edited by Catravas, J.D., Gillis, N. and Ryan, U.S., p.89-98, Ple num, New York (1989) Hallam, T.J., Pearson, J.D. and Needham, L.A.: Thrombin-stimulated elevation of human endo thelial-cell cytoplasmic free calcium concentration causes prostacyclin production. Biochem. J. 251, 243 249 (1988) Braquet, P., Paubert-Braquet, M., Bourgain, R.H., Bussolino, F. and Hosford, D.: PAF/ cytokine auto-generated feedback networks in microvascular immune injury: consequences in shock, ischemia and graft rejection. J. Lipid Mediat. 1, 75 -112 (1989) Rotrosen, D. and Gallin, J.I.: Histamine type 1 receptor occupancy increases endothelial cytosolic calcium, reduces F-actin, and promotes albumin diffusion across cultured endothelial monolayers. J. Cell Biol. 103, 2379-2387 (1986) Jaffe, E.A., Grulich, J., Weksler, B.B., Hampel, G. and Watanabe, K.: Correlation between thrombin-induced prostacyclin production and inositol triphosphate and cytosolic free calcium levels in cultured human endothelial cells. J. Biol. Chem. 262, 8557-8565 (1987) Wickham, N.W.R., Vercellotti, G.M., Moldow, C.F., Visser, M. and Jacob, H.S.: Measurement of intracellular calcium concentration in intact mono layers of human endothelial cells. J. Lab. Clin. Med. 112, 157-176 (1988) Konishi, M., Olson, A., Hollingworth, S. and Baylor, S.M.: Myoplasmic binding of Fura-2 in vestigated by steady-state fluorescence and adsorb ance measurements. Biophys. J. 54, 1089 -1104 (1988) Uto, A., Arai, H. and Ogawa, Y.: Reassessment of Fura-2 and the ratio method for determination of intracellular Ca2+ concentrations. Cell Calcium 12, 29-37 (1991) Poenie, M.: Alteration of intracellular Fura-2 fluorescence by viscosity: Asimple correction. Cell Calcium 11, 85-91 (1990) Steinberg, S.F., Bilezikian, J.P. and Al-Awqati, Q.: Fura-2 fluorescence is localized to mitochon dria in endothelial cells. Am. J. Physiol. 253, C744 C747 (1987)