INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

ATP IS AN IMPORTANT MEDIATOR involved in signaling between cardiovascular cells. Extracellular levels of ATP are normally maintained at extremely low levels due to ubiquitous ectonucleotidases, which rapidly hydrolyze nucleotides. However, in the vasculature significant amounts of extracellular ATP may locally accumulate at the site of thrombus formation (6) due to a release from activated platelets or in the hypoxic myocardium (4) when ATP is released together with degradation products such as adenosine from energy-depleting myocardial cells. A systemic increase in ATP and its immediate hydrolytic products up to micromolar concentrations is found in blood plasma under conditions of traumatic shock (9).

ATP may act on endothelial cells through multiple receptors and second messengers. The majority of known effects elicited by nonhydrolyzed ATP are mediated via P_2y and P_2u receptors (5). These receptors are both coupled to phospholipase C (PLC), but via distinct G proteins (21). Signal transduction events stimulated by ATP include the inositol lipid-Ca²⁺ signaling cascade, protein kinase C (PKC), and, directly or indirectly, activation of soluble guanylate cyclase or adenylate cyclase (6). ATP may also act on endothelial cells after degradation to adenosine via adenosine receptors (25).

The information on ATP effects on endothelial barrier function is scarce and inconsistent. Depending on the endothelial cell population, ATP can modulate endothelial barrier function in one direction or the other. In the special case of microvascular endothelium contained in venular microvessels from frogs (13-16), ATP was found to increase paracellular permeability. Similar to effects of inflammatory mediators, this rise of permeability in response to ATP coincided with a rise in cytosolic Ca²⁺ concentration ([Ca²⁺]_i) within the cells. In other populations, ATP was found to decrease paracellular permeability. Examples are bovine aortic endothelial cells (12) or, from preliminary experiments of the present study, cultured endothelial cells from other macrovessels such as the porcine aorta, porcine pulmonary artery, bovine aorta, or human umbilical vein. In this second type of reaction, the role of [Ca²⁺]_i is unclear. The present study was undertaken to analyze the role of [Ca²⁺]_i changes in endothelial cell preparations in which ATP reduces permeability. The effects of ATP were contrasted with those of ionomycin, a Ca²⁺ ionophore, which causes a rise in [Ca²⁺]_i in a receptor-independent manner. Endothelial barrier function was studied by determining the passage of albumin through confluent monolayers of porcine aortic endothelial cells. Variations of albumin passage in this model are to be attributed to changes in paracellular permeability.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell cultures. Endothelial cells from the bovine aorta, porcine aorta, and pulmonary artery were isolated and cultured as previously described (26). Human endothelial cells from umbilical cords were isolated and cultured according to van Hinsberg et al. (30). Confluent cultures of primary endothelial cell were trypsinized in phosphate-buffered saline [PBS; composed of (in mM) 137 NaCl, 2.7 KCl, 1.5 KH₂PO₄, and 8.0 Na₂HPO₄, pH 7.4, supplemented with 0.05% (wt/vol) trypsin and 0.02% (wt/vol) EDTA] and seeded at a density of 7 × 10⁴ cells/cm² on 24-mm round polycarbonate filters (pore size 0.4 µm), 5 × 20-mm glass coverslips, or 30-mm culture dishes for determination of albumin permeability, [Ca²⁺]_i, or cyclic nucleotide contents, respectively. Experiments were performed with confluent monolayers 4 days after seeding.

Macromolecule permeability. The permeability of the endothelial cell monolayer was studied in a system of two compartments separated by a filter membrane (23, 24). Both compartments contained as basal medium modified Tyrode solution [composition in mM: 150 NaCl, 2.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 1.0 CaCl₂, and 30.0 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES); pH 7.4, 37°C] supplemented with 5% (vol/vol) heat-inactivated newborn calf serum (NCS; 10 min at 60°C). There was no hydrostatic pressure gradient between the compartments. The "luminal" compartment containing the monolayer had a volume of 2.5 ml, and the "abluminal" compartment had a volume of 13 ml. The fluid in the abluminal compartment was constantly stirred. Trypan blue-labeled albumin (60 µM) was added to the luminal compartment. The appearance of labeled albumin in the abluminal compartment was continuously monitored by pumping the liquid through a spectrophotometer (Specord 10; Zeiss, Jena, Germany). Increases in the concentration of labeled albumin were detected with a time delay of <15 s. The concentration of labeled albumin in the luminal compartment was determined every 10 min during incubation. It did not change significantly in the time frame of the experiments.

Experimental protocols. The basal medium used in incubations was modified Tyrode solution (composition as described in Macromolecule permeability). Macromolecule permeability of the endothelial monolayer, transferred to the incubation chamber, was determined after an initial equilibration period of 20 min. The basal albumin permeability of each monolayer-filter system was then determined during another 20 min of incubation. Agents were added as indicated in RESULTS, and the response of albumin permeability was recorded for another 40-80 min. For the incubations under Ca²⁺-free extracellular conditions, a Ca²⁺-free basal medium (composition in mM: 150 NaCl, 2.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, 0.5 ethylene glycol-bis(-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), and 30.0 HEPES; pH 7.4, 37°C) supplemented with heat-inactivated 5% (vol/vol) NCS was used. Stock solutions of bisindolylmaleimide I (BIM), Gö-6976, ionomycin, 1H-[1,2,4]oxadioazolo[4,3-a]quinoxalin-1-one (ODQ), SQ-22536, U-37122, U-37343, and thapsigargin were prepared with dimethyl sulfoxide (DMSO). A stock solution of 1-O-octadecyl-2-O-methyl-rac-glycero-3-phosphorylcholine (ET-18-OCH₃) was prepared with absolute ethanol. Appropriate volumes of these solutions were added to the cells, yielding final solvent concentrations 0.1% (vol/vol). The same final concentrations of DMSO or ethanol were also included in all respective control experiments. Stock solutions of all other substances were prepared in basal medium (composition as described in Macromolecule permeability). Appropriate volumes of these solutions were added to the cells. Identical additions of basal medium were included in all respective control experiments.

Cytosolic Ca²⁺. Free [Ca²⁺]_i were determined using the fluorescent Ca²⁺ indicator fura 2. Confluent endothelial monolayers cultured on 5 × 20-mm glass coverslips were incubated in medium 199 supplemented with 5% (vol/vol) heat-inactivated NCS and the addition of 5 µM fura 2-AM (acetoxymethyl ester of fura 2) at 20°C in the dark. After a 50-min incubation period, the extracellular fura 2-AM was removed by medium change. This was followed by a 20-min incubation period in the same medium before measurements were started. The coverslips were then aligned in a quartz cuvette into the beam of a fluorescence spectrophotometer (LS 50B; Perkin-Elmer, Überlingen, Germany). During incubations, the excitation wavelength was alternated between 340 and 380 nm (bandwidth 5 nm). Emitted light was detected at 510 nm (bandwidth 2 nm). Fura 2 fluorescence was calibrated according to the method described by Grynkiewicz et al. (10). For this purpose, the cells were exposed to 5 µM ionomycin in modified Tyrode solution containing either 3 mM Ca²⁺ or 5 mM EGTA to obtain the maximum (R_max) and minimum (R_min) of the ratio of fluorescence (R), respectively. [Ca²⁺]_i was calculated according to the equation where K_d is the dissociation constant of fura 2 (10) and  is the ratio of the 380-nm excitation signals of ionomycin-treated cells at 5 mM EGTA and 3 mM Ca²⁺.

Loading of BAPTA-AM. Confluent endothelial monolayers cultured on either filter membranes or 5 × 20-mm glass coverslips were incubated in medium 199 supplemented with 5% (vol/vol) heat-inactivated NCS and the addition of 10 µM 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid-AM (acetoxymethyl ester of BAPTA) at 37°C. After a 30-min incubation period, the extracellular BAPTA-AM was removed by medium change and the experiments were started.

Determination of PKC activity. The activity of PKC was determined in the membrane fraction of endothelial cells using the method of Chakravarthy et al. (3) that allows measurement of PKC activity in its native membrane-associated state. Cultured endothelial monolayers were incubated for the time indicated in RESULTS, and the cultures were then rinsed twice with ice-cold PBS and subsequently covered with ice-cold hypotonic lysis buffer (composition in mM: 1 NaHCO₃, 5 MgCl₂, and 0.1 PMSF; pH 7.4). The cells were scraped off the dish and vigorously mixed at room temperature for 2 min. The lysates were centrifuged at 1,000 g for 5 min at 4°C to sediment nonlysed cells and nuclei. The postnuclear supernatants were centrifuged at 4°C for 10 min at 425,000 g. The sedimented endothelial membranous fractions were resuspended in 200 µl of assay buffer (composition in mM: 0.002 CaCl₂, 10 MgCl₂, 0.2 PMSF, 2 NaF, 0.2 Na₃P₂O₇, 0.2 Na₃VO₄, and 50 Tris · HCl buffer; pH 7.4).

1

1

Cellular contents of cyclic nucleotides. Cultured endothelial monolayers were incubated for the time indicated in RESULTS. At the end of the incubations, the incubation medium was aspirated, ice-cold ethanol was added, and the culture dishes were stored at 80°C. To determine the intracellular cyclic nucleotide contents, the ethanol was evaporated at 60°C, and the samples were suspended in double-distilled water, transferred into Eppendorf reaction tubes, and centrifuged for 5 min at 14,000 g. Cyclic nucleotide concentrations of the supernatants were determined by using radioimmunoassays (Amersham, Braunschweig, Germany). The protein contents of the samples were determined according to Bradford (1) using bovine serum albumin as the standard.

Materials. ODQ was from Alexis Biochemicals (Grünberg, Germany); Falcon plastic tissue culture dishes were from Becton Dickinson (Heidelberg, Germany); acrylodan-labeled MARCKS peptide (as PKC fluorescence substrate), ATP, adenosine 5'-O-(3-thiotriphosphate) (ATPS), ADP, AMP, and UTP were from Boehringer Mannheim (Mannheim, Germany); BIM, ET-18-OCH₃, Gö-6976, ionomycin, SQ-22536, U-37122, U-37343, and thapsigargin were from Calbiochem (Bad Soden, Germany); Transwell polycarbonate filter inserts (24-mm diameter, 0.4-µm pore size) were from Costar (Bodenheim, Germany); NCS, medium 199, penicillin-streptomycin, and trypsin-EDTA were from GIBCO Life Technologies (Eggenstein, Germany); and 5'-(N-ethylcarboxamido)adenosine (NECA) was from Sigma (Deisenhofen, Germany). All other chemicals were of the best available quality, usually analytic grade.

Statistical analysis. Data are given as means ± SD of n = 6 experiments using independent cell preparations. Statistical analysis of data was performed according to Student's unpaired t-test. Probability (P) values of <0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Macromolecule permeability. Exposure of aortic endothelial monolayers to 10 µM ATP reduced their albumin permeability from a control level of 5.8 ± 0.7 to 3.5 ± 0.9 × 10⁶ cm/s after 20 min (Fig. 1). Albumin permeability remained that low for a further 60 min. This observation was not restricted to porcine aortic endothelial monolayers but was also made in endothelial monolayers derived from the porcine pulmonary artery, bovine aorta, and human umbilical vein (Table 1). For the remainder of this study, endothelial cells from the porcine aorta were used. In contrast to the effect on permeability induced by ATP, the addition of ionomycin (1 µM) increased permeability rapidly to a peak value of 8.9 ± 0.5 × 10⁶ cm/s within 10 min. Afterward, permeability decreased but remained significantly elevated during the time of observation. The effects of ATP and ionomycin on permeability were concentration dependent. Both agents were tested in a concentration range between 10⁸ and 10⁵ µM (Fig. 2). The permeability experiments were performed in the presence of 5% heat-inactivated NCS. This small amount of serum was included in the incubation medium because basal permeability of endothelial monolayers was then stable up to 2 h. Under serum-free conditions, permeability was also reduced by 10 µM ATP (from a control value of 6.7 ± 0.5 to 4.1 ± 0.5 × 10⁶ cm/s after 20 min) or increased by 1 µM ionomycin (from a control value of 6.7 ± 0.5 to 9.7 ± 0.8 × 10⁶ cm/s after 10 min).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effects of ionomycin (1 µM; ) and ATP (10 µM; ) on albumin permeability of aortic endothelial monolayers. , Control. Data are means ± SD of n = 6 separate experiments of independent cell preparations.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Concentration-dependent effects of ionomycin () and ATP () on albumin permeability. , Control (C). Data are means ± SD, after 20 min of incubation, of n = 6 separate experiments of independent cell preparations. * P < 0.05 vs. control.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent effects of ATP (), adenosine 5'-O-(3-thiotriphosphate) (ATPS; ), ADP (), and UTP () on albumin permeability. , Control (C). Data are means ± SD, after 20 min of incubation, of n = 6 separate experiments of independent cell preparations. At nucleotide concentrations 106 M, albumin permeabilities were significantly different from control (P < 0.05).

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Effect of 8-phenyltheophylline (8-PT; 10 µM) on reduction of albumin permeability induced by ATP (10 µM), AMP (10 µM), and 5'-(N-ethylcarboxamido)adenosine (NECA; 100 nM). Data are percentages of control, where 100% corresponds to an albumin permeability of 5.8 ± 0.6 × 106 cm/s of barrier formed by endothelial monolayer and filter support. Data are means ± SD of n = 6 separate experiments of independent cell preparations. * P < 0.05 vs. control. # P < 0.05 vs. AMP or NECA alone.

ATP-induced increase in [Ca²⁺]_i and permeability. Exposure of endothelial cells to 10 µM ATP, a concentration that caused a marked reduction of albumin permeability, elicited a transient rise of [Ca²⁺]_i with a maximum of 503 ± 47 nM (Fig. 5). The effect of ATP on [Ca²⁺]_i was comparable to that induced by 1 µM ionomycin. As shown in Table 2, ATPS, ADP, and UTP also transiently increased [Ca²⁺]_i. Significant differences in peak [Ca²⁺]_i among ATP, ATPS, ADP, or UTP were not found. AMP and NECA did not increase [Ca²⁺]_i.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Effects of ionomycin (1 µM) and ATP (10 µM) on cytosolic Ca2+ concentration ([Ca2+]i) of aortic endothelial monolayers. Representative records are shown.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Influence of internal Ca2+ chelation on ATP-induced permeability (A) and [Ca2+]i (B). Control cells () were neither exposed to ATP nor loaded with 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA). Other cells were loaded with BAPTA (10 µM; ) for 30 min before additional administration of ATP (10 µM; ). Results were compared with effects of ATP alone (10 µM; ). Data are means ± SD of n = 6 separate experiments of independent cell preparations.

6

6

View larger version (38K): [in this window] [in a new window]   Fig. 7.   Influence of external and internal Ca2+ deprivation on ATP-induced permeability (A) and [Ca2+]i (B). Control cells () were incubated in Ca2+-free incubation medium and were treated with neither ATP nor thapsigargin. Cells were treated with thapsigargin (300 nM; ) to empty endogenous Ca2+ stores (external and internal Ca2+ deprivation) and after addition of ATP (10 µM; ). Results were compared with effects of ATP alone (10 µM; ) under extracellular Ca2+-free conditions of endothelial cells not treated with thapsigargin. Data are means ± SD of n = 6 separate experiments of independent cell preparations.

ATP-induced effects on PKC, PLC, and permeability. To analyze whether ATP causes activation of PKC, we isolated the membrane fraction of ATP-stimulated endothelial cells and determined the PKC activity in the native membrane-associated state. Exposure of endothelial cells to 10 µM ATP for 10 min increased PKC activity in the membrane fraction by 90% above the control level (Table 3). The increase in PKC activity in the membrane fraction was virtually abolished when cells had been pretreated for 20 min with 1 µM BIM, a panspecific inhibitor of PKC, or 100 nM Gö-6976, a selective inhibitor of Ca²⁺-dependent PKC isoenzymes (Table 3). Exposure of endothelial cells to 1 µM BIM or 100 nM Gö-6976 alone had no effect on basal PKC activity in the membrane fraction.

View larger version (40K): [in this window] [in a new window]   Fig. 8.   Influence of U-73122 (U2), an inhibitor of phospholipase C, and its inactive analog U-73343 (U3) on ATP effects on permeability (A) and [Ca2+]i (B). Control cells () were exposed to neither ATP nor phospholipase C inhibitor. Other cells were pretreated with U-73122 (1 µM; ) or U-73343 (1 µM; ) for 10 min before additional administration of ATP (10 µM;  or , respectively). Results were compared with effects of ATP alone (10 µM; ). Data are means ± SD of n = 6 separate experiments of independent cell preparations.

ATP-induced effects on cyclic nucleotides and permeability. We analyzed whether ATP stimulates cyclic nucleotide synthesis in endothelial cells and whether this synthesis is related to the ATP-induced reduction of macromolecule permeability.

View larger version (40K): [in this window] [in a new window]   Fig. 9.   Influence of 1H-[1,2,4]oxadioazolo[4,3-a]quinoxalin-1-one (ODQ), an inhibitor of guanylate cyclase, on ATP effects on permeability (A) and cellular cGMP content (B). Control cells () were exposed to neither ATP nor guanylate cyclase inhibitor. Other cells were exposed to either ATP (10 µM; ) or ODQ (20 µM; ) alone. In 1 set of experiments, cells were pretreated with ODQ (20 µM) for 10 min before additional administration of ATP (10 µM; ). Data are means ± SD of n = 6 separate experiments of independent cell preparations.

View larger version (40K): [in this window] [in a new window]   Fig. 10.   Influence of SQ-22536 (SQ), an inhibitor of adenylate cyclase, on ATP effects on permeability (A) and cellular cAMP content (B). Control cells () were exposed to neither ATP nor adenylate cyclase inhibitor. Other cells were exposed to either ATP (10 µM; ) or SQ-22536 (50 µM; ) alone. In 1 set of experiments, cells were pretreated with SQ-22536 (50 µM) for 10 min before additional administration of ATP (10 µM; ). Data are means ± SD of n = 6 separate experiments of independent cell preparations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The most prominent findings of the present study are that ATP induced a reduction of albumin permeability even though it provoked transient rises of [Ca²⁺]_i, cGMP, cAMP, and translocation of PKC in porcine aortic endothelial monolayers. The ATP-induced rises of [Ca²⁺]_i, cGMP, cAMP, and translocation of PKC could be prevented without abolishing the effect of ATP on permeability. The effect of ATP on permeability, however, was sensitive to inhibitors of PLC. These data show that ATP reduces macromolecule permeability via a PLC-mediated signaling pathway that is independent of concomitant effects on [Ca²⁺]_i, cyclic nucleotides, or PKC.

Exposure of endothelial cells from the porcine aorta to ATP caused a sustained reduction of albumin permeability. Similar effects were observed in endothelial monolayers derived from the porcine pulmonary artery, bovine aorta, and human umbilical vein. A detailed analysis was carried out on endothelial cells from the porcine aorta. The effect of ATP on barrier function was concentration dependent. ATPS, which is much less hydrolyzable than ATP, reduced permeability in the same manner as ATP, indicating that ATP, but not its degradation products, caused the reduction of permeability. It has been reported (5, 21) that ATP can stimulate endothelial cells via purinergic receptors, in particular P_2y and P_2u. The latter of these receptors has a high affinity for UTP. In endothelial monolayers UTP also caused a reduction of permeability in a concentration-dependent manner, but this reduction was less potent than that induced by ATP. The fact that the action of ATP can be mimicked by other nucleotides suggests that ATP exerts its effect on permeability by acting on purinergic receptors. The tested nucleotides were in the following rank order of potency: ATP = ATPS > ADP = UTP. The order of potency is not identical to that known for either P_2y or P_2u. It may therefore correspond to that of yet another one of the multitude of purinergic receptors identified in cloning experiments.

To validate the conclusion that ATP and not a derivative is responsible for its effect on permeability, the actions of ATP, AMP, and the stable adenosine analog NECA were compared. We reported in a previous study (31) that NECA reduces endothelial albumin permeability via stimulation of endothelial adenosine receptors in the same model used in the present study. It has now been confirmed that NECA reduces endothelial permeability and that AMP has a comparable effect. Both effects could be fully antagonized by 8-PT, an adenosine-receptor antagonist. In contrast, 8-PT failed to antagonize the reduction of permeability obtained with ATP. These data therefore show independently that the effect of ATP on permeability is not due to the actions of adenosine.

In a further set of experiments, we tested the question of whether the reduction of permeability in the presence of ATP depends on the concomitant rise in [Ca²⁺]_i. Two different experimental maneuvers were used. In the first maneuver, [Ca²⁺]_i was kept at low intracellular levels during ATP stimulation by loading cells with the chelator BAPTA. In the second maneuver, endothelial cells were incubated in Ca²⁺-free medium and endogenous Ca²⁺ stores were preemptied by thapsigargin. It has been shown previously (24) that, with the latter protocol, the possibility of a rise in [Ca²⁺]_i via extra- or intracellular causes is abolished. In the present study it was found that the loading of endothelial cells with BAPTA as well as extra- and intracellular Ca²⁺ deprivation completely precluded the ATP-evoked rise in [Ca²⁺]_i, but these maneuvers did not affect the ATP-induced reduction of permeability. The experiments with the Ca²⁺ ionophore ionomycin show, from a different point of view, the dissociation between [Ca²⁺]_i rise and reduction of permeability evoked by ATP. At the applied concentration, ionomycin causes a transient increase in [Ca²⁺]_i that is comparable to the [Ca²⁺]_i rise in the presence of ATP. The action of ionomycin circumvents the signal transduction activated by ATP that causes [Ca²⁺]_i rise and a multitude of other effects (6). Experiments with ionomycin show that a sole rise in [Ca²⁺]_i, without the other mechanisms also activated by ATP, causes a spontaneous increase in endothelial permeability. This is in accordance with many other observations (14-16, 18, 24). The contrasting effect of ATP must therefore be due to an intracellular mechanism activated in endothelial cells in response to ATP that overrides the Ca²⁺-activated increase in permeability.

It has been described previously (7, 27) that stimulation of endothelial cells with ATP causes activation of PLC. Here, we studied whether the activation of PLC is involved in the effects of ATP on permeability by using an inhibitor approach. The specific inhibitor of PLC, U-73122, was applied and compared with U-73343, an inactive analog of U-73122. The PLC inhibitor U-73122, but not its inactive analog, abolished the ATP-induced reduction of permeability. This indicates that the effect of ATP on permeability is mediated through activation of PLC. This conclusion is supported by the observation that ET-18-OCH₃, a PLC inhibitor chemically distinct from U-73122, can also attenuate the ATP-induced reduction of permeability. Apart from its effect on permeability, U-73122 also abolished the increase in [Ca²⁺]_i. This shows that the ATP-induced increase in [Ca²⁺]_i, even though unrelated to permeability changes, is another PLC-dependent effect. Taken together, the experiments using the PLC inhibitors indicate that the permeability-reducing mechanism observed in the presence of ATP is activated on a level downstream of PLC but upstream of PLC-mediated [Ca²⁺]_i rise.

PLC activation can lead to activation of PKC. A parameter of PKC activation is translocation of PKC activity into cell membranes. In the presence of ATP such an increase in membranous PKC activity was indeed observed. This translocation was found to be abolished by pretreatment of cells with the PKC inhibitor U-73122, indicating that it occurs secondarily to PLC activation. PKC translocation was also inhibited in the presence of the nonselective PKC inhibitor BIM or in the presence of Gö-6976, an inhibitor of the Ca²⁺-dependent isoforms of PKC. This suggests that only Ca²⁺-dependent isoforms are translocated in response to stimulation with ATP. Endothelial cells are known to express the Ca²⁺-dependent PKC isoforms  and  (2).

In contrast to their effects on PKC translocation into the membrane fraction, the PKC inhibitors did not affect the ATP-induced rise in monolayer permeability. This indicates that the latter is independent of an activation of classic PKC isoforms that would all be inhibited by BIM. It leaves open the possibility that nonclassic isoforms of PKC are involved.

We found that the presence of ATP causes a transient rise in cGMP and cAMP in the cells. Processes that stimulate cGMP or cAMP synthesis may cause a reduction of macromolecule permeability of endothelial monolayers (17, 18). The changes in cyclic nucleotides, observed in the presence of ATP, could be blocked in the presence of specific inhibitors, ODQ for soluble guanylate cyclase and SQ-22536 for adenylate cyclase. The effects of the inhibitors indicate that the rises in cyclic nucleotides are due to activation of synthesis. We did not analyze this signaling of ATP stimulation toward cyclic nucleotide synthesis. Instead, we analyzed whether the changes in cAMP or cGMP contents affected the ATP-lowering effect on permeability. This was not the case. The ATP effect on permeability is thus independent of the changes in cyclic nucleotides.

In conclusion, ATP can stimulate both a rise in [Ca²⁺]_i and a reduction of macromolecule permeability in the types of endothelial monolayers tested here. ATP seems to be the first mediator described that can exert such apparently contrasting effects. In general, rapid effects on endothelial permeability, as investigated here, seem to be due to two kinds of mechanisms: 1) a modulation of tension of the endothelial contractile machinery and 2) changes in cell-cell or cell-matrix adhesion. A transient rise of [Ca²⁺]_i seems to increase endothelial permeability primarily by activation of the second type of mechanism (8). One may speculate that ATP overrides these [Ca²⁺]_i-mediated effects on cell adhesion by the activation of specific Ca²⁺-independent signal transduction pathways. These have yet to be identified.