INTRODUCTION

Many types of inflammatory cells, in particular granulocytes, possess secretory granules that contain microbicidal and proinflammatory substances. Exocytosis of these granules plays an important role not only in the host defense, but also in the development of inflammatory tissue damage. Thus, understanding the regulation of granule release from these cells is crucial for the understanding of the inflammatory response. However, while research over the last years has greatly advanced our knowledge about the mechanisms of exocytosis in neuronal cells, the mechanisms of exocytosis in inflammatory cells are still poorly understood. Exocytosis in neurons is intimately linked to the activation of voltage-operated Ca²⁺ channels (1). A low affinity Ca²⁺-sensitive step in the exocytotic machinery of neurons (EC₅₀ up to 200 µM in the synapse) has been demonstrated. It is thought that only Ca²⁺ influx through Ca²⁺ channels is able to achieve submembraneous [Ca²⁺] which are sufficiently high to activate this low affinity step (2, 3). In contrast to excitable cells, inflammatory cells secrete in response to receptor activation, not membrane depolarization. Voltage-activated Ca²⁺ channels are absent in inflammatory cells. Previous reports suggested that the cytosolic free Ca²⁺ concentration ([Ca²⁺]_c)¹ induces exocytosis in inflammatory cells with a high affinity (EC₅₀ 0.5-3 µM in human neutrophils) (4). Thus, at this point there is no compelling evidence that Ca²⁺ influx through ion channels plays the same role in exocytosis by inflammatory cells as it does in excitable cells.

Inflammatory cells, however, do have plasma membrane Ca²⁺ channels, and Ca²⁺ influx occurs in response to stimulation with receptor agonists (5). The predominant type of Ca²⁺ influx in inflammatory cells is the so-called store-operated Ca²⁺ influx (or capacitative Ca²⁺ entry) (6), and the underlying channels are referred to as store-operated Ca²⁺ channels (or I_CRAC channels) (7). As opposed to voltage-operated Ca²⁺ channels, store-operated Ca²⁺ channels are found in virtually all cell types, and their most obvious function is the refilling of intracellular Ca²⁺ stores after stimulated Ca²⁺ release. In neuronal cells, it has been suggested that store-operated Ca²⁺ channels are important for refilling of Ca²⁺ stores but do not play a relevant role in shaping the [Ca²⁺]_c transients during cell activation. In inflammatory cells however, store-operated Ca²⁺ influx is likely to play an important role for cell activation, with the best documented example being lymphocyte proliferation (8). The role of store-operated Ca²⁺ influx in the exocytotic secretion by inflammatory cells has hitherto received little attention. It has previously been noted that the Ca²⁺ sensitivity of exocytosis in inflammatory cells is modulated by other agonist-generated signals (4). However, the identity of these signals and their relationship to store-operated Ca²⁺ influx have not been revealed.

In this study we demonstrate that store-operated Ca²⁺ influx plays a premier role in the regulation of receptor-stimulated exocytosis in HL-60 granulocytes. However, as opposed to the situation in neuronal cells and as opposed to previous results obtained with Ca²⁺ ionophores in granulocytes, even maximal activation of store-operated Ca²⁺ influx is not sufficient to induce exocytosis in HL-60 granulocytes by itself, and synergistic signals generated through receptor activation are required for exocytosis to occur. The synergistic signaling cascade involves pertussis toxin sensitive G proteins and a phosphatidylinositol 3-kinase (PI 3-kinase).

EXPERIMENTAL PROCEDURES

Cell culture media were obtained from Life Technologies, Inc. (Paisley, Scotland), ³H-platelet-activating factor from Amersham (Little Chalfont, UK), silica plates from Merck (Darmstadt, Germany), erbstatin analog (methyl 2,5-dihydroxycinnamate) and LY294,002 (2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one) from Alexis (Läufelfingen, Switzerland), U73122 and U73343 from Calbiochem, and fura-2/AM from Molecular Probes (Eugene, OR). SK&F96365 (1-(-[3-(4-methoxyphenyl)propoxy]-4-methoxyphenethyl)-1H-imidazole hydrochloride) was provided by Dr. J. Merritt, SmithKline Beecham (Welwyn Hertfordshire, UK). All other chemicals were purchased from Sigma or Fluka. The "Ca²⁺-free medium" contained 143 mM NaCl, 6 mM KCl, 1 mM MgSO₄, 5.6 mM glucose (=0.1%), 20 mM HEPES, pH 7.4. The free Ca²⁺ concentration of this medium was about 5 µM as determined with a Ca²⁺-sensitive electrode (9). The "Ca²⁺ medium" consisted of Ca²⁺-free medium supplemented with 1 mM CaCl₂.

HL-60 cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum, penicillin (50 units/ml), streptomycin (50 µg/ml), and L-glutamine (2 mM) (9). The cells were cultured at 37 °C in a humidified atmosphere of 5% CO₂, 95% air. The culture was passaged twice weekly, and granulocytic differentiation was initiated by addition of dimethyl sulfoxide (final concentration 1.3% for 3 days, then 0.65% for 1 or 2 days).

2.5 × 10⁶ cells were suspended in 500 µl of medium containing 2.5 µg of cytochalasin B and warmed to 37 °C for 5 min before addition of stimuli at the indicated concentration and time. Incubation was terminated by rapid cooling with an equal volume of ice-cold buffer and further cooling in ice before centrifugation (800 × g for 10 min). N-Acetyl--glucosaminidase in the supernatant was measured fluorimetrically with 4-methylumbelliferyl-glucosamine (10). The results were calculated as percentage of initial total cellular content. Background exocytosis (5-10% of total cellular content) was subtracted. This background was independent of any preincubation. The total enzyme content was determined from an aliquot of the same cell suspension treated with 0.025% Triton X-100 for 5 min at 37 °C.

HL-60 granulocytes (2 × 10⁷/ml) were loaded with 2 µM fura-2/AM in Ca²⁺ medium plus 0.1% bovine serum albumin for 45 min at 37 °C, then diluted to 10⁷ cells/ml and kept on ice. Just before use, 0.5 ml of this cell suspension was centrifuged and resuspended in 2.4 ml of the indicated medium including 5 µg/ml cytochalasin B. Fura-2 fluorescence (F) was measured in a thermostated cuvette (37 °C) (LS3 fluorimeter, Perkin-Elmer Corp.) at 340-nm excitation and 505-nm emission wavelength. Calibration was performed for each cuvette by sequential addition of 1 mM Ca²⁺ (if not already present), 1 µM ionomycin to measure Ca²⁺-saturated fura-2 (F_max), then 8 mM EGTA plus 30 mM Tris (pH 9.3) and 0.1% Triton X-100 to measure Ca²⁺-free fura-2 (F_min). Free Ca²⁺ concentrations were calculated from the fura-2 fluorescence according to the formula of Grynkiewicz et al. (11): [Ca²⁺]_c = K_d × (F  F_min)/(F_max  F) with K_d = 227 nM.

The absolute values for [Ca²⁺]_c beyond 1 µM (see Figs. 3 and 5) must be interpreted with caution, since the Ca²⁺ indicator fura-2 reaches saturation in the micromolar range. However, the relative differences obtained with the different experimental conditions were highly reproducible as indicated by the error bars. Genistein caused a quench of fura-2 fluorescence but allowed reliable [Ca²⁺]_c measurements. Erbstatin analog and LY294,002 have multiple effects on fura-2 fluorescence and interfere with the calibration; their effects on [Ca²⁺]_c could therefore not be determined.

[View Larger Version of this Image (16K GIF file)]

Effect of fMLP, GTPS, and pertussis toxin on Ca²⁺-induced exocytosis.

[View Larger Version of this Image (37K GIF file)]

Respiratory burst activity was measured by chemiluminescence originating from horseradish peroxidase-catalyzed luminol oxidation on Luminometer 1250 Wallac LKB (12, 13). Cells (2 × 10⁶/ml in Ca²⁺-free medium supplemented with 1 unit/ml horseradish peroxidase, 50 µM of luminol, and 90 µM of NaN₃) were preincubated for 5 min at 37 °C with 5 µg/ml cytochalasin B before stimulation with 50 nM thapsigargin or 0.5 µM ionomycin and 0-2 mM Ca²⁺ 3 min later. Chemiluminescence was measured continuously at 37 °C.

GTPS Loading by Hypo-osmotic Shock

HL-60 granulocytes were loaded with GTPS by hypo-osmotic shock as described elsewhere (14). Briefly, 40 × 10⁶ cells were incubated in 250 µl of a buffer containing 143 mM NaCl, 6 mM KCl, 1 mM MgSO₄, 20 mM HEPES, pH 7.4, 0.1% glucose, 375 mM sucrose, 7.5% polyethylene glycol 1000, 7.5% fetal calf serum, in the presence or absence of 50 mM GTPS. After 15 min at 25 °C to allow fluid-phase endocytosis, hypo-osmotic lysis of the endosomes was induced by addition of 4 ml of H₂O. After 60 s, iso-osmolarity was restored by addition of 3.5 ml of 1.8% NaCl. The intracellular concentration of GTPS was estimated to be 50 µM, homogeneously distributed in over 90% of the cells (14). Cell viability was >90%.

Phospholipase D activity was determined by measuring the formation of phosphatidylethanol, a specific product of this enzyme, produced in the presence of ethanol. The cellular pools of phosphatidylcholine were labeled with 1-O-[³H]octadecyl-platelet-activating factor, as described by Pai et al. (15). Cells were suspended 3 × 10⁷/ml in Ca²⁺-free medium supplemented with fatty acid-free bovine serum albumin (1 mg/ml) and incubated in the presence of 1-O-[³H]octadecyl-platelet-activating factor (5 mCi/ml) for 75 min at 37 °C. The labeled cells were then washed and resuspended in Ca²⁺-free medium. Fifteen seconds before stimulation, ethanol was added (0.5%). Each sample contained 8 × 10⁶ cells in a final volume of 0.4 ml. Stimulation was terminated by adding 1.6 ml of chloroform/methanol/acetic acid (100/200/4 by volume). After 30 min, the phases were separated by adding 533 µl of chloroform and 560 µl of H₂O (13). The lower phase (chloroform containing the lipids) was recovered, dried under N₂ and spotted onto Silica Gel 60 thin layer chromatography plates. The plates had been previously run in chloroform/methanol/water (65/35/8 by volume). The separations were performed in ethyl acetate/iso-octane/acetic acid/water (130/20/30/100 by volume). The plates were then dried and autoradiographed. Spots were identified, scraped off, and quantified by liquid scintillation counting.

All experiments were performed at least three times with different cell batches unless specified in the text. Background controls with solvent instead of agonists were run and subtracted from the values obtained with agonist. Data are presented as mean ± standard error of mean (S.E.), n = 3 or 4, unless stated otherwise in the text.

RESULTS

We first studied the effect of extracellular Ca²⁺ on the [Ca²⁺]_c signal and on primary granule release in response to the chemotactic peptide fMLP (for the definition of granule subpopulations in neutrophil granulocytes, see Borregaard et al. (16)). In the absence of extracellular Ca²⁺, the [Ca²⁺]_c signal showed a marked decrease in duration, but only a slight decrease in amplitude (Fig. 1, A and C). The initial rapid rise from basal [Ca²⁺]_c  100 nM to over 1 µM, was the result of Ca²⁺ release from intracellular stores. In Ca²⁺-containing medium, the initial peak was followed by a prolonged influx of Ca²⁺ from the extracellular medium. In the absence of extracellular Ca²⁺, only the release from intracellular stores occurred. Peak [Ca²⁺]_c was reached about 15 s after stimulation in both conditions. fMLP stimulated exocytosis was completed in 30 s. The amplitude of exocytosis was four times higher in the presence of extracellular Ca²⁺ (Fig. 1B), whereas peak [Ca²⁺]_c was only 25% higher (Fig. 1C). Thus Ca²⁺ influx was necessary for efficient triggering of exocytosis. To further study the importance of Ca²⁺ influx for fMLP-induced exocytosis, we analyzed the effect of inhibitors of Ca²⁺ influx in granulocytes (6, 17). Two classes of inhibitors, trivalent metal ions and imidazole derivatives, reduced exocytosis to levels that were obtained in the absence of extracellular Ca²⁺ (Fig. 1D).

[View Larger Version of this Image (20K GIF file)]

Thus, Ca²⁺ influx is a necessary signal for fMLP-stimulated exocytosis in HL-60 granulocytes. However, is it also sufficient to induce exocytosis? To separate the effect of fMLP on [Ca²⁺]_c from its effect on other intracellular signals, we designed protocols to control [Ca²⁺]_c independently of receptor stimulation. We used thapsigargin, an inhibitor of the Ca²⁺ store Ca²⁺-ATPase that provokes release of Ca²⁺ from intracellular stores by inhibiting the refilling of the stores (18). The empty stores generate a yet unknown signal that activates store-operated Ca²⁺ channels in the plasma membrane (19). These channels are the physiological pathway for Ca²⁺ entry after fMLP stimulation (6). Without extracellular Ca²⁺, thapsigargin induced a slow, transient elevation of [Ca²⁺]_c, which returned to baseline levels within 5 min (Fig. 2A). Because thapsigargin acts irreversibly, the stores can not be refilled, and the plasma membrane Ca²⁺ channels remain permanently activated. Thus, subsequent addition of Ca²⁺ to the extracellular medium caused a sustained Ca²⁺ influx, and [Ca²⁺]_c reached high plateau levels (Fig. 2, A and B). Despite its essential role in fMLP-stimulated exocytosis (see above), store-operated Ca²⁺ influx did not efficiently trigger exocytosis by itself (Fig. 2B, right side, compared with Fig. 1B). Addition of fMLP did not alter [Ca²⁺]_c in this protocol (Fig. 2, C and D). 5 min after thapsigargin treatment, the stores were almost empty, and fMLP induced only minimal Ca²⁺ release. Under these conditions, fMLP was unable to stimulate exocytosis underlining the essential role of Ca²⁺ in fMLP-stimulated exocytosis (Fig. 2D, center). However, together with store-operated Ca²⁺ influx, fMLP markedly stimulated exocytosis (Fig. 2, D, right side, compared with B, right side). Thus, fMLP activates additional signaling pathways that synergize with [Ca²⁺]_c elevations.

[View Larger Version of this Image (28K GIF file)]

The inefficiency of thapsigargin-induced Ca²⁺ influx to stimulate exocytosis contrasts with the well known stimulation of neutrophil exocytosis by the Ca²⁺ ionophore ionomycin (see, for example, Lew et al. (10)). To investigate this discrepancy, we stimulated HL-60 granulocytes with either ionomycin or thapsigargin in the presence of increasing extracellular Ca²⁺ concentrations and measured [Ca²⁺]_c elevations (Fig. 3A), exocytosis (Fig. 3B), and superoxide production (Fig. 3C). The results show that the Ca²⁺ ionophore raised average [Ca²⁺]_c to levels substantially higher than thapsigargin-induced store-operated Ca²⁺ influx. They also show that the Ca²⁺ ionophore induced substantial exocytosis, while thapsigargin, even at the highest extracellular Ca²⁺ concentration tested, did not trigger exocytosis. However, the lack of thapsigargin-induced exocytosis cannot solely be explained by the lower average [Ca²⁺]_c levels. Indeed, when average [Ca²⁺]_c values of ~500 nM were obtained with ionomycin, significant exocytosis occurred, while store-operated Ca²⁺ influx did not induce exocytosis even when it raised average [Ca²⁺]_c levels above 1 µM. The absence of exocytosis in thapsigargin-stimulated cells was not due to an inhibition of the exocytotic machinery through thapsigargin, as costimulation of cells with thapsigargin did not diminish fMLP-induced exocytosis (data not shown). To better understand the discrepancy between the thapsigargin effect and ionomycin effect, we studied O₂ generation under the same experimental conditions (Fig. 3C). Both thapsigargin and ionomycin were able to stimulate O₂ generation in the presence of extracellular Ca²⁺. However, the extracellular Ca²⁺ concentrations necessary to induce maximal Ca²⁺-dependent superoxide generation were much lower for ionomycin than for thapsigargin. This was particularly striking with 0.2 mM extracellular [Ca²⁺], where ionomycin induced maximal O₂ generation, while thapsigargin had a very small effect. Thus, even under conditions where the fura-2 measurements did not reveal differences in thapsigargin- and ionomycin-stimulated average [Ca²⁺]_c elevations, the measurements of Ca²⁺-dependent O₂ generation clearly suggests that ionomycin leads to higher [Ca²⁺]_c at the site of Ca²⁺ action. This discrepancy most likely reflects a different spatial organization of the thapsigargin- and the ionomycin-induced Ca²⁺ signal (see "Discussion"). Note also the inhibition of the ionomycin-induced O₂ generation by increasing Ca²⁺ concentrations. This probably is an in vivo reflection of the previously described inhibition of the assembled NADPH oxidase activity in membrane preparations by high Ca²⁺ concentrations (20).

It appears that fMLP is able to activate additional signals at resting [Ca²⁺]_c (Fig. 2) and that these signals have a synergistic action when store-operated Ca²⁺ influx is allowed to occur. To investigate the kinetics of these synergistic signals, we used the Ca²⁺ store depletion protocol as outlined above (Fig. 2), varied the delay between fMLP and Ca²⁺ readdition from 0 to 15 min, and compared exocytosis and Ca²⁺ influx (Fig. 4A). The level of [Ca²⁺]_c that was reached under these conditions was similar at all time points of Ca²⁺-addition, albeit a small transient decrease during the first minutes after fMLP was observed. Exocytosis was highest, when Ca²⁺ and fMLP were given at the same time (0 min). When Ca²⁺ was added 30 s to 2 min after fMLP, exocytosis was reduced by 50%. With longer delays between fMLP stimulation and Ca²⁺ addition, exocytosis reached almost the same level as at time 0. This suggests at least three possible explanations: (i) the synergistic signaling pathway has complex kinetics, including a transient inactivating component; (ii) there are several distinct synergistic signals with different kinetics; and (iii) the small transient decrease in average [Ca²⁺]_c reflects a large decrease in submembraneous [Ca²⁺], which accounts for the large decrease in exocytosis. Given the possibility that the submembraneous [Ca²⁺] rather than the average cellular [Ca²⁺] is crucial for the induction of exocytosis, we favor the last possibility.

[View Larger Version of this Image (20K GIF file)]

The fMLP receptor is known to be rapidly inactivated upon ligand binding (21). It was therefore surprising to see that, under our experimental conditions, the synergistic fMLP-stimulated signals were maintained for as long as 15 min. To understand whether this long lasting effect was due to a long lasting receptor activity, or rather the persistence of activated downstream signals, we studied the effect of the fMLP antagonist N-tert-butoxycarbonyl-methionyl-leucyl-phenylalanine, boc-MLP (Fig. 4B). Experiments were performed as shown in Fig. 2; however, where indicated, boc-MLP was added to the medium 3 min after the addition of fMLP. As shown in Fig. 4B, exocytosis was almost completely inhibited by boc-MLP. Thus, the long lasting fMLP effect to synergize with store-operated Ca²⁺ influx was due to a long lasting receptor/ligand interaction and not to the persistence of activated downstream signals.

We then investigated the nature of the synergistic signal by generating steady-state [Ca²⁺]_c between 100 nM and 10 µM with low levels of ionomycin (0.5 µM) in the presence of 0-2 mM extracellular Ca²⁺. For any given average [Ca²⁺]_c value from 100 nM to 10 µM, fMLP enhanced exocytosis (Fig. 5A). A fit of the data in Fig. 5A with a logistic equation suggested that fMLP enhanced both maximal exocytosis (Y_max: 18.2 ± 0.9 to 24.2 ± 0.5% release in the absence and presence of fMLP, respectively) and its apparent affinity for [Ca²⁺]_c (EC₅₀ for average [Ca²⁺]_c values: 1000 ± 150 nM and 670 ± 50 nM in the absence and presence of fMLP, respectively). This suggests that the role of the synergistic signals generated by fMLP is not only an enhancement of exocytosis, but also a shift of its Ca²⁺ sensitivity to [Ca²⁺]_c values that can be achieved by physiological Ca²⁺ influx. However, while a shift of the Ca²⁺ sensitivity can be reliably detected with the fura-2 method, the discrepancy between the submembraneous [Ca²⁺]_c values (seen by the exocytotic machinery) and the average [Ca²⁺]_c values (measured by the fluorescent dye) precludes calculations of the absolute values of the Ca²⁺ sensitivity of the exocytotic process (see also Fig. 3). Loading of cells with GTPS (for details, see Jaconi et al. (14)) induced an increase in the amplitude and the Ca²⁺ sensitivity of the ionomycin-induced exocytosis similar as seen for fMLP (Fig. 5, A and B). This suggests that GTP-dependent processes are involved in the second signal pathway. To understand whether these GTP-dependent processes act along the same synergistic signaling pathway as fMLP, we investigated whether fMLP stimulation of GTPS-loaded cells leads to a further enhancement of exocytosis. As shown as in Fig. 5C, no additivity between fMLP and GTPS was detectable compatible with the concept that both act along the same pathway. To date most, if not all, fMLP-activated signaling cascades were found to be mediated by pertussis toxin-sensitive G proteins, presumably by members of the G_i family of heterotrimeric G proteins. To analyze whether such proteins might be involved in the activation of the synergistic signal by fMLP, we investigated the effect of pertussis toxin on the synergistic signal generated by fMLP. A complete suppression of the synergistic signal was observed (Fig. 5D). Taken together, these results suggest that fMLP stimulates a synergistic signal via a pertussis toxin-sensitive G protein to increase the amplitude and the affinity of Ca²⁺-activated exocytosis.

Phospholipase D (PLD) has been implicated in the regulation of exocytosis in neutrophils (22). To investigate, whether PLD activation could account for the synergistic signaling pathway, we investigated activation of the enzyme under our experimental conditions (Fig. 6). As described above, cells were similarly suspended in Ca²⁺-free medium and store-depleted by the addition of thapsigargin (0 min), followed by addition of fMLP or Me₂SO (5 min). Under these conditions, fMLP did not stimulate PLD activity, demonstrating that there is an absolute requirement for [Ca²⁺]_c elevations in the fMLP stimulation of PLD activity. When store-operated Ca²⁺ influx was allowed to occur (1 mM CaCl₂; 10 min), a significant increase of PLD activity was observed, both in the absence (6-fold) or in the presence (12-fold) of fMLP (Fig. 6A). Thus, store-operated Ca²⁺ influx by itself is a good activator of PLD. Ethanol is an inhibitor of PLD-mediated signaling. Pretreatment with ethanol diminished both fMLP- and ionomycin-induced exocytosis in a dose-dependent manner (Fig. 6C). Thus, PLD activity may have a role downstream from store-operated Ca²⁺ influx. A major role of PLD in the synergistic signals elicited by fMLP as shown in Fig. 2 appears less likely but cannot be excluded. There might be a high threshold (>6 phosphatidylethanol per total lipid, Fig. 6A) above which PLD activity could synergize with Ca²⁺ to induce exocytosis.

[View Larger Version of this Image (28K GIF file)]

A phosphatidylinositol bisphosphate-specific phospholipase C (PLC) is activated by fMLP stimulation of granulocytes via pertussis toxin-sensitive G proteins. The PLC reaction generates inositol 1,4,5-trisphosphate, which is the mediator of the Ca²⁺ signal. In addition, PLC also generates the protein kinase C activator diacylglycerol and diminishes the phosphatidylinositol bisphosphate content of the plasma membrane (although the latter effect is relatively small and short lived) (23). Thus, PLC could also be involved in the synergistic signaling pathway. We therefore investigated the effect of the phospholipase C inhibitor U73122 (24). As expected, U73122 inhibited fMLP-induced [Ca²⁺]_c increase with an IC₅₀ of 250 nM, while the thapsigargin-induced [Ca²⁺]_c increase was not affected (Fig. 7B). Next, we investigated whether U73122 had an additional effect on fMLP-induced exocytosis, independent of the inhibition of [Ca²⁺]_c increase. To obtain the obligatory rise in [Ca²⁺]_c even in the presence of PLC inhibitor, the cells were pretreated with thapsigargin in the presence of 1 mM Ca²⁺ and then stimulated with fMLP (Fig. 7A). Under these conditions, U73122 did inhibit exocytosis. However, as (i) exocytosis was markedly less sensitive to U73122 than was the [Ca²⁺]_c increase (IC₅₀ 754 ± 104 nM versus 256 ± 11 nM, p < 0.01, for exocytosis and [Ca²⁺]_c, respectively), and (ii) ionomycin-induced exocytosis was also affected, the inhibitory effect of U73122 on exocytosis is probably unrelated to PLC inhibition.

[View Larger Version of this Image (16K GIF file)]

The tyrosine kinase inhibitor genistein moderately reduced fMLP- but not ionomycin-induced exocytosis at higher concentrations (Fig. 8A). However, at these concentrations genistein also reduced fMLP-mediated Ca²⁺ release and Ca²⁺ influx (Fig. 8B; inhibition of Ca²⁺ influx by genistein has been reported previously (25)). Another tyrosine kinase inhibitor, the erbstatin analog methyl 2,5-dihydroxycinnamate, did not inhibit fMLP-induced exocytosis (with 1, 3, and 10 µM erbstatin analog exocytosis was at 95.3 ± 3.0%, 89.5 ± 7.8%, and 93.9 ± 5.0% of the control level, mean ± S.E. of four determinations from two independent experiments).

[View Larger Version of this Image (14K GIF file)]

Wortmannin is a potent inhibitor of PI 3-kinases and of granulocyte superoxide generation (26). Fig. 9A shows that wortmannin inhibited fMLP- but not ionomycin-induced exocytosis. The inhibition of fMLP-induced exocytosis was incomplete (maximal inhibition 62 ± 4%); however, its affinity was relatively high (IC₅₀ 70 nM), compatible with an effect on a PI-3-kinase. No relevant effects of wortmannin on Ca²⁺ release and Ca²⁺ influx were observed. These results would be compatible with an involvement of PI-3 kinase in the synergistic signaling pathway, as wortmannin inhibited fMLP-induced exocytosis, but neither ionomycin-induced exocytosis nor fMLP-induced [Ca²⁺]_c elevations. To obtain additional evidence, we tested a second PI 3-kinase inhibitor, namely LY294,002, which inhibits the enzyme with 250-fold lower potency (27). LY294,002 also inhibited fMLP-induced exocytosis (10, 30, and 100 µM LY294,002 reduced exocytosis to 76.5 ± 5.2%, 58.3 ± 6.9%, and 46.4 ± 15.4% of control, mean ± S.E. of four determinations from two independent experiments). Thus, taken together, our results suggest that PI 3-kinase activation is a part of the synergistic signaling pathway toward granulocyte exocytosis. Ptasznik et al. (28) recently reported that tyrosine kinases were involved in the stimulation of PI 3-kinases by fMLP in neutrophils. The inhibition of exocytosis at high genistein concentrations (see Fig. 8), might therefore involve an effect on tyrosine kinase-mediated PI 3-kinase activation.

[View Larger Version of this Image (16K GIF file)]

DISCUSSION

Our results demonstrate that store-operated Ca²⁺ influx plays a key role in the regulation of exocytosis of primary granules in HL-60 granulocytes. Thus, as seen in neuronal cells, there is a similar tight correlation between Ca²⁺ channel activation and activation of exocytosis.

Unlike the situation in neuronal cells, Ca²⁺ influx through Ca²⁺ channels is by itself not sufficient to induce exocytosis in HL-60 granulocytes. Nevertheless, exocytosis of primary granules has been activated by Ca²⁺ alone in ionomycin-stimulated intact granulocytes (29), in permeabilized granulocytes (30, 31), as well as in whole cell patch-clamped granulocytes.² A high [Ca²⁺]_c affinity of primary granule release (EC₅₀ ~2.6 µM) has been determined using the "ionophore clamp" technique in combination with fluorescence measurements of average [Ca²⁺]_c values (4). Our ionophore-clamp experiments (Fig. 5A) yield similar results; however, the data obtained with thapsigargin-induced Ca²⁺ influx are not compatible with such a high affinity for Ca²⁺. In digitonin-permeabilized and in electropermeabilized cells, a much higher [Ca²⁺] (between 10 and 100 µM) was needed to achieve primary granule release (31, 32), and in patch-clamp studies pipette Ca²⁺ concentrations up to 4 µM were unable to induce substantial exocytosis (33). Thus, most likely, the Ca²⁺ affinity of primary granule exocytosis in granulocytes is underestimated by the ionophore experiments. In addition, [Ca²⁺]_c elevations at the site of exocytosis achieved through store-operated Ca²⁺ influx in granulocytes are probably lower than those achieved through voltage-operated Ca²⁺ influx in neurons. When normalized for the cell surface, peak Ca²⁺ currents through store-operated channels in inflammatory cells are at least 10 times smaller than currents through voltage-operated Ca²⁺ channels in excitable cells (current density ~1-3 pA/pF in inflammatory cells (8, 34), versus 30-90 pA/pF in excitable cells (35, 36)). The requirement for additional signals that activate exocytosis in synergy with store-operated Ca²⁺ influx may be a direct consequence of the relatively low amplitude of submembraneous [Ca²⁺] increases generated through store-operated Ca²⁺ influx.

Both thapsigargin and ionomycin stimulate store operated Ca²⁺ influx by emptying intracellular Ca²⁺ stores. In addition, ionomycin induces Ca²⁺ influx through ionophore pores in the plasma membrane. This difference is witnessed by the saturation of the steady-state [Ca²⁺]_c levels with increasing extracellular [Ca²⁺] in thapsigargin-treated, but not ionomycin-treated, cells (Fig. 3A).

Ionomycin, but not thapsigargin, stimulated exocytosis, and ionomycin activated a maximal respiratory burst response at a 10 times lower extracellular Ca²⁺ concentration than did thapsigargin. Thus ionomycin appears to induce much higher Ca²⁺ concentrations at the site of Ca²⁺ action (i.e. the submembraneous space) than does thapsigargin. These local Ca²⁺ elevations are poorly reflected by the average [Ca²⁺]_c measurements probably because (i) the submembraneous space represents a minor portion of the cellular volume and (ii) fura-2 is already saturated below the [Ca²⁺]_c values that can be reached in the submembraneous space.

At least two mechanisms might contribute to the generation of much higher submembraneous Ca²⁺ concentrations by ionomycin compared with thapsigargin: (a) ionomycin insertion into the plasma membrane and (b) Ca²⁺ buffering by a ionomycin-sensitive, thapsigargin-resistant Ca²⁺ store. High capacity low affinity Ca²⁺ stores that might buffer [Ca²⁺]_c elevations in a thapsigargin-resistant fashion include mitochondria (37), and recently described endoplasmic reticulum subcompartments (38).

The synergistic signals enhance exocytosis and promote a left-shift of the [Ca²⁺]_c sensitivity of exocytosis. Such a shift through receptor agonists or various pharmacological activators has been observed in intact cells (Fig. 5C) (4), as well as in permeabilized cells (39, 40).

Our results do not favor a role of phospholipase C in the synergistic signaling pathway. Phospholipase D could have a role when its activated above a high threshold. Previous studies have already argued against a role of protein kinase C enzymes in this context, as primary granules release in response to fMLP is not inhibited by various protein kinase C inhibitors and activation of protein kinase C by phorbol ester does not induce primary granule release (41, 42).

fMLP receptor-dependent signaling has been shown to occur in many instances through pertussis toxin-sensitive G proteins. However, biochemical analysis has also suggested that the fMLP receptor may interact with low molecular weight (i.e. pertussis-toxin insensitive) GTP-binding proteins (43). Small GTP-binding proteins are thought to be involved in the regulation of exocytosis in various cellular systems (44). Our results (Fig. 5) show that the activation of pertussis toxin-sensitive G proteins (presumably G_i2 or G_i3) (22) is an obligatory step in the activation of the second signaling pathway. These results do not exclude the downstream involvement of other G proteins in the exocytotic process.

PI 3-kinases are likely to be involved in the regulation of exocytosis (26). This was supported by the recent observation that synaptotagmin, a potential Ca²⁺ sensor in neuronal exocytosis, is able to bind phosphatidylinositol 3,4,5-trisphosphate (i.e. a product of PI 3-kinase activity) in a Ca²⁺-dependent fashion (45). In HL-60 granulocytes, we found an inhibition of fMLP-induced exocytosis by the PI 3-kinase inhibitors wortmannin and LY294,002. The IC₅₀ for the inhibition of exocytosis (70 nM for wortmannin, 15 µM for LY294,002) was in the range of PI 3-kinase inhibition. Thus, our results suggest that PI 3-kinase is involved in the synergistic signaling pathway. However, maximal inhibition obtained through PI 3-kinase inhibitors did not exceed 60%. Therefore, activation of PI 3-kinases alone cannot fully account for the synergistic signal pathway and additional mechanisms are likely to be involved.

To our knowledge, this study is the first in-depth analysis of the role of store-operated Ca²⁺ influx in the regulation of exocytosis in granulocytes. In previous studies, the consequence of omission of extracellular Ca²⁺ during chemoattractant-stimulated exocytosis in granulocytes has varied from a minor reduction in exocytosis (10, 46), to a profound inhibition (29, 47, 48). However, in all studies, chemoattractant stimulation of exocytosis was Ca²⁺-dependent. Thus, it appears that Ca²⁺ release from intracellular stores can under certain circumstances substitute for Ca²⁺ influx. Two explanations may be considered. (i) Synergistic receptor-generated signals (e.g. priming) might lead to a more pronounced left-shift of the exocytosis/[Ca²⁺]_c curve, and Ca²⁺ release from intracellular stores might then suffice to trigger exocytosis. (ii) Positioning of Ca²⁺ stores close the exocytotic machinery similar to their accumulation around phagosomes (49) might generate localized high [Ca²⁺] by Ca²⁺ release.

In summary, store-operated Ca²⁺ influx is necessary for efficient stimulation of HL-60 exocytosis by fMLP, but it is not a sufficient stimulus by itself and requires the concomitant activation of synergistic signals to activate the exocytotic machinery. Our results suggest that the synergistic signaling pathway is initiated by the interaction of the fMLP receptor with a pertussis toxin-sensitive G protein and it involves PI 3-kinases. The need for synergistic signals in the regulation of exocytosis in granulocytes might stem from the relatively low amplitude of submembraneous [Ca²⁺]_c changes that can be achieved through store-operated Ca²⁺ channels. However, the fact that maximal Ca²⁺ influx through physiological pathways does not induce exocytosis may serve as a safety mechanism. An inappropriate exocytosis of primary granules is most harmful to the organism. It may be vital for the granulocyte to allow the occurrence of Ca²⁺ influx without invariably inducing exocytosis.