Soluble E-selectin Acts in Synergy with Platelet-activating Factor to Activate Neutrophil β2-Integrins

Selectins play a critical role in neutrophil recruitment to sites of inflammation, in tethering and rolling of neutrophils on vascular endothelium, as well as triggering β2-integrin-mediated adhesion. We have previously demonstrated potential pro-inflammatory effects of soluble E-selectin upon neutrophil effector functions, using a soluble recombinant molecule (E-zz), which increased β2-integrin-mediated adhesion, decreased β2-integrin-dependent migration, and triggered reactive oxygen species generation and release. In this study, we have examined the intracellular signals following neutrophil activation by soluble E-selectin. We show that exposure of neutrophils to E-selectin and platelet-activating factor (PAF) in combination induced a synergistic effect upon β2-integrin-mediated adhesion. Although soluble E-selectin did not induce Ca2+ mobilization in neutrophils by itself, elevation of intracellular Ca2+ was specifically prolonged in response to PAF but not leukotriene B4 orN-formyl-Met-Leu-Phe. The prolonged Ca2+mobilization observed in the presence of E-selectin was dependent on Ca2+ influx from intracellular stores rather than influx of extracellular Ca2+ through SKF 96365-sensitive channels. The specific alteration of Ca2+ mobilization reported here appears not to have a role in the synergistic effects of E-selectin and PAF upon neutrophil O⨪2 release but may be involved in augmentation of β2-integrin-mediated adhesion.

Selectins play a critical role in neutrophil recruitment to sites of inflammation, in tethering and rolling of neutrophils on vascular endothelium, as well as triggering ␤ 2 -integrin-mediated adhesion. We have previously demonstrated potential pro-inflammatory effects of soluble E-selectin upon neutrophil effector functions, using a soluble recombinant molecule (E-zz), which increased ␤ 2 -integrin-mediated adhesion, decreased ␤ 2integrin-dependent migration, and triggered reactive oxygen species generation and release. In this study, we have examined the intracellular signals following neutrophil activation by soluble E-selectin. We show that exposure of neutrophils to E-selectin and platelet-activating factor (PAF) in combination induced a synergistic effect upon ␤ 2 -integrin-mediated adhesion. Although soluble E-selectin did not induce Ca 2؉ mobilization in neutrophils by itself, elevation of intracellular Ca 2؉ was specifically prolonged in response to PAF but not leukotriene B 4 or N-formyl-Met-Leu-Phe. The prolonged Ca 2؉ mobilization observed in the presence of E-selectin was dependent on Ca 2؉ influx from intracellular stores rather than influx of extracellular Ca 2؉ through SKF 96365-sensitive channels. The specific alteration of Ca 2؉ mobilization reported here appears not to have a role in the synergistic effects of E-selectin and PAF upon neutrophil O 2 . release but may be involved in augmentation of ␤ 2 -integrin-mediated adhesion.
Recruitment of inflammatory cells in response to injury or infection involves coordinated regulation of specific receptor/ counter-receptor pairs from the integrin, immunoglobulin, selectin, and mucin families (1). The control of events involved in leukocyte recruitment is critical for the development of effective anti-microbial defenses and also for efficient wound healing. Failure to recruit granulocytes is life threatening, resulting in inadequate clearance of opportunistic pathogens. However, excessive inflammatory cell recruitment or inappropriate cell activation leads to the development of chronic inflammatory conditions that may favor fibrotic repair mechanisms and, ultimately, loss of organ function (2).
It is now well established that a multistep molecular process allows leukocyte recruitment to different tissue sites. In acute inflammatory responses, neutrophil granulocyte recruitment involves L-selectin and P-selectin glycoprotein ligand-1 that engage endothelial counter-receptors including endothelial Pand E-selectin (3). Conversion of "rolling" adhesion to "firm" adhesion necessary for subsequent transmigration requires an additional co-stimulus, for example via intracellular signals from occupancy of G protein-coupled seven-transmembrane receptors for bacterial peptides, inflammatory mediators, and chemokines (4,5). Trans-endothelial migration is clearly a complex process but is thought to involve dynamic modulation of integrin-mediated adhesion events to allow formation of new sites of contact at the leading edge of the cell and selective detachment at the trailing end (6 -8). The existence of multiple levels of control for adhesion processes ensures that "inappropriate" recruitment of leukocytes does not occur. However, a breakdown in the normal adhesion regulatory mechanisms may contribute to pathogenesis of a number of diseases in which the granulocyte has been implicated (9).
Cellular recruitment is critically dependent on formation of molecular interactions between receptor/counter-receptor pairs. However, it is now clear that many adhesion receptors are also coupled to the intracellular signal transduction machinery, providing "outside-in" signals that modulate leukocyte responses. Initiation of signal transduction pathways following integrin occupancy is relatively well established (10). For example, release of degradative enzymes and toxic oxidant species required for antimicrobial defense is augmented following adhesion to matrix components. However, the role of other adhesion receptors in triggering signaling events has only recently begun to be defined. Engagement of CD31 has been demonstrated to augment the rate of integrin-dependent migration, providing a mechanism for ensuring efficient diapedesis (11). Recent evidence also suggests that engagement of receptors involved in rolling adhesion may also transduce specific signals that alter neutrophil behavior. In particular, ligation of neutrophil L-selectin and P-selectin glycoprotein ligand-1 augment ␤ 2 -integrin-mediated adhesion and increase production of damaging reactive oxygen intermediates (12,13). Selectin-dependent alterations in neutrophil function have been suggested to require co-stimulation with other agents, e.g. platelet-activating factor (PAF) 1 (14,15). The functional alter-ations induced by agents that activate neutrophils appear distinct from those caused by "priming" agents, which have little effect on generation of reactive oxygen species by themselves. Together these findings indicate that the repertoire of receptors engaged on the neutrophil surface determines the functional responses in terms of adhesion, migration, and secretion. One implication is that although it may be possible to define experimental conditions in vitro under which neutrophils are "unprimed," "primed," or "activated," the functional status of neutrophils during recruitment in vivo may approximate to a continuum of activation states.
Our previous studies have demonstrated that neutrophils treated with soluble recombinant E-selectin (E-zz) comprising the N-terminal C-type lectin domain, epidermal growth factor domain and two of the short consensus repeats with two protein A domains in tandem cause increased ␤ 2 -integrin-mediated adhesion with a concomitant decrease in ␤ 2 -integrin-dependent migration. In addition, generation and release of reactive oxygen species is increased following exposure to Eselectin (16). These effects could be blocked by F(abЈ) 2 anti-Eselectin monoclonal antibody binding to the lectin domain of E-selectin. The kinetics of the neutrophil response to E-zz raised the possibility that the effects might be indirect because of secondary release of inflammatory mediators from neutrophils, for example PAF. In this paper we examine the generation of intracellular signals following exposure of neutrophils to E-selectin. We have found that E-selectin specifically prolongs [Ca 2ϩ ] i increase in neutrophils treated with PAF but not Nformyl-Met-Leu-Phe (fMLP) or leukotriene B 4 (LTB 4 ). In addition, we have addressed whether soluble E-selectin effects were due to a second release of PAF and found that soluble Eselectin was not acting directly through PAF receptors. However, our data suggest that specific changes in [Ca 2ϩ ] i are not required for E-selectin-dependent augmentation of functional responses in neutrophils.
Neutrophil Isolation-Polymorphonuclear leukocytes were isolated from peripheral blood of healthy donors as described previously (16). Briefly, after centrifugation of citrated whole blood at 300 ϫ g for 20 min and removal of platelet-rich plasma, leukocytes were separated from erythrocytes by dextran sedimentation using 0.6% dextran T500. Polymorphonuclear leukocytes were then separated from mononuclear leukocytes using discontinuous isotonic Percoll gradients. Polymorphonuclear leukocytes were 95-98% neutrophils using morphological criteria, and viability was Ͼ99% as assessed by trypan blue exclusion.
Neutrophil Adhesion to Albumin-coated Latex Beads-␤ 2 -Integrindependent adhesion of ACLB to freshly isolated neutrophils was measured as described previously (17). Fluorescent latex beads were washed in HBSS and incubated with 10 mg/ml human serum albumin for 15 min before washes in HBSS and resuspension at 0.5% in HBSS. Neutrophils (155 l at 10 7 /ml in HBSS) in the presence or absence of E-zz (5 g/ml) were added to 25 l of ACLB and 25 l of agonist as described under "Results." For inhibition experiments, neutrophils were preincubated with TMB-8 (250 M), SKF 96365 (5 M), or UK-74505 (1 M) for 10 min before addition of ACLB. Cells and ACLB were then incubated at 37°C for 15 min in a shaking water bath at 110 beats/min. After fixation of the cells with 0.5 ml of 0.5% glutaraldehyde, nonadherent ACLB were removed by three washes. Bead binding to neutrophils was measured by flow cytometry using a FacsCalibur (Becton Dickinson, Oxford, UK).
Determination of [Ca 2ϩ ] i by Spectrofluorimetry-Freshly isolated neutrophils were incubated at 10 7 /ml in Ca 2ϩ -and Mg 2ϩ -free HBSS with 2 M Fura-2-AM for 30 min at 37°C. Cells were then washed twice and resuspended at 2 ϫ 10 6 /ml in HBSS containing Ca 2ϩ and Mg 2ϩ essentially as described (18). A SPEX Fluoromax spectrofluorimeter with excitation wavelengths at 340 and 380 nm and an emission wavelength recorded at 510 nm was used for measuring [Ca 2ϩ ] i . [Ca 2ϩ ] i was calculated as described (18,19) using the equation: where r is the measured Fura-2 fluorescence ratio between 340 and 380 nm. r max , the maximum 340/380 nm ratio, was obtained by lysing cells with 25 M digitonin. r min , the minimum 340/ 380 ratio, was obtained by addition of 10 mM EGTA. K d is the dissociation constant of the Fura-2/Ca 2ϩ complex (224 nM), and B is the ratio of fluorescence at 380 nm at 0 and saturating Ca 2ϩ concentration.
Measurement of Superoxide Anion Release-The determination of release of superoxide anions by freshly isolated neutrophils was performed as described previously (16). Briefly, neutrophils were preincubated with genistein (74 M) or Ro-318220 (1 M) for 15 min at 37°C before stimulation with soluble recombinant proteins, fMLP, or PMA in the presence of cytochrome c (1 mg/ml). After 15 min at 37°C, the reaction was stopped by placing the cells on ice, followed by a centrifugation (13,000 ϫ g, 2 min, 4°C). The superoxide dismutase (200 units)-inhibitable reduction of cytochrome c was determined for each supernatant by measuring the peak absorbance between 535 and 565 nm using a Pye-Unicam scanning spectrophotometer. Results are expressed as nanomoles of superoxide anions generated per 10 6 neutrophils.
Statistical Analysis-Using the paired Student's t test, the difference between the mean of the assays and control groups was considered significant when p Ͻ 0.05. The results are expressed as the means Ϯ S.D. of the number (n) of independent experiments each using cells isolated from separate donors.

Requirement for Ca 2ϩ in E-selectin-mediated Augmentation of ␤ 2 -Integrin
Adhesion-Our previous studies demonstrated that there was a temporal association between the extent of augmentation of neutrophil effector function and the duration of treatment with soluble E-selectin (E-zz), raising the possibility that E-zz triggered secondary release of inflammatory mediators, e.g. PAF. We therefore tested whether the PAF antagonist UK-74505 affected E-zz-induced ␤ 2 -integrin-mediated adhesion using a well characterized assay involving binding of albumin-coated latex beads (16,17). Although UK-74505 blocked PAF-induced ␤ 2 -integrin activation, it had no effect on ␤ 2 -integrin activation following treatment with either fMLP or E-zz (Fig. 1A), suggesting that E-zz effects are independent of PAF release and ligation of the PAF receptor.
Careful analysis of the effects of E-zz and PAF alone or in combination demonstrated that treatment of neutrophils with both agents induced a synergistic effect upon ␤ 2 -integrin activation (Fig. 1B). Thus, although PAF (0.1 nM) or E-zz alone resulted in a small augmentation of ␤ 2 -integrin-mediated adhesion (1.8-and 3.5-fold, respectively), the combination resulted in 8.6-fold increase in neutrophil ␤ 2 -integrin activation to a level comparable with that seen with 10 nM fMLP. Because changes in [Ca 2ϩ ] i are an important early event following ligation of the seven-transmembrane receptor family (20,21), we next assessed whether E-zz affected the PAF response by altering Ca 2ϩ fluxes within neutrophils by using different Ca 2ϩ channel inhibitors. Treatment of neutrophils with SKF 96365 (5 M), a receptor-operated channel inhibitor, did not block augmentation of ␤ 2 -integrin activity with E-zz alone or in combination with PAF and only partially blocked ␤ 2 -integrin-mediated adhesion induced by PAF at 10 nM. In contrast, treatment of neutrophils with 250 M TMB-8, an intracellular calcium antagonist, completely abolished E-zz-and PAF-mediated ␤ 2 -integrin activation (Fig. 1B). fMLP-induced ␤ 2 -integrin activation was not significantly inhibited by TMB-8 (Fig. 1B), indicating that Ca 2ϩ mobilization is not required for induction of neutrophil ␤ 2 -integrin activity and providing strong evidence that TMB-8 does not nonspecifically inhibit ␤ 2 -integrin activation.
Soluble E-selectin Prolongs PAF-induced Calcium Mobilization-The distinct specific effects of TMB-8 upon ␤ 2 -integrin activation induced by PAF and E-zz led us to consider whether E-zz might act to alter Ca 2ϩ mobilization within neutrophils in response to PAF. In contrast to the rapid and transient elevation of [Ca 2ϩ ] i in response to PAF, treatment of neutrophils with E-zz alone did not induce any changes in [Ca 2ϩ ] i ( Fig. 2A), indicating that E-zz may act through a pathway distinct from IP 3 -dependent Ca 2ϩ mobilization. Interestingly, preincubation of neutrophils with E-zz did not affect the initial increase of [Ca 2ϩ ] i in response to PAF but caused a subsequent sustained increase in [Ca 2ϩ ] i ( Fig. 2A). Preincubation of neutrophils with V-zz did not affect [Ca 2ϩ ] i in response to PAF ( Fig. 2A), indicating that the effects of E-zz upon calcium mobilization are specific. Analysis of the concentration responses revealed that E-zz was maximally effective following treatment with higher concentrations of PAF (Fig. 2B), suggesting that a threshold increase in [Ca 2ϩ ] i is required above which E-zz prolongs calcium mobilization.
PAF-induced Ca 2ϩ mobilization was ablated by TMB-8 (Fig.  3A), indicating that release of Ca 2ϩ from intracellular stores may be an essential trigger for PAF-induced binding to ACLB. The prolongation of [Ca 2ϩ ] i elevation in the presence of E-zz was independent of SKF 96365-sensitive Ca 2ϩ channels (Fig. 3,  B and C). Thus, although SKF 96365 blocked PAF-induced extracellular Ca 2ϩ influx, Ca 2ϩ mobilization in the presence of E-zz was relatively insensitive to SKF 96365. Together with data presented in Fig. 1, these results indicate that E-zz treatment may affect PAF responses via the release of Ca 2ϩ from intracellular stores rather than influx of extracellular Ca 2ϩ through SKF 96365-sensitive receptor-operated channels.
Specificity of Soluble E-selectin Effects-We next considered whether treatment of neutrophils with E-zz prolonged Ca 2ϩ mobilization in response to other agonists that act via distinct seven-transmembrane G protein-coupled receptors. As shown in Fig. 4, in contrast with the effects of PAF, changes in [Ca 2ϩ ] i following stimulation with either LTB 4 or fMLP were not affected by preincubation with E-zz. These findings raise the possibility that sustained elevation of [Ca 2ϩ ] i observed when E-zz-pretreated cells were stimulated with PAF might account for the synergistic effects on ␤ 2 -integrin-mediated adhesion. We therefore tested whether ␤ 2 -integrin activation observed following LTB 4 or fMLP treatment was affected by preincubation with E-zz. As shown in Fig. 5, E-zz treatment alone induced a small but significant increase in ␤ 2 -integrin-mediated adhesion. The ␤ 2 -integrin activation response of neutrophils to LTB 4 and fMLP was not significantly affected by E-zz pretreatment, whereas E-zz and PAF together induced a significant increase in the percentage of cells that were capable of ␤ 2 -

FIG. 2. Soluble E-selectin prolongs PAF-induced Ca 2؉ mobilization.
A, after loading neutrophils with Fura-2-AM (2 M) for 30 min at 37°C in Ca 2ϩ and Mg 2ϩ -free HBSS, cells were washed and resuspended in HBSS containing Ca 2ϩ and Mg 2ϩ and then preincubated with or without E-zz or V-zz (5 g/ml) for 15 min at 37°C. Cells were then transferred into 3-ml cuvettes for measurement of [Ca 2ϩ ] i using a SPEX Fluoromax spectrofluorimeter followed by analysis as described under "Experimental Procedures." After 30 s of recording, neutrophils were stimulated with PAF (10 nM); Ⅺ, untreated cells; ‚, cells preincubated with E-zz; E, cells preincubated with V-zz; or E-zz (5 g/ml) ϫ E-zz (5 g/ml). B, after preincubation with (f) or without (Ⅺ) E-zz (5 g/ml), neutrophils were stimulated with different concentrations of PAF as indicated. We obtained curves similar to those in A, which have been integrated to calculate area under each curve using GraphPad Prism software (GraphPad Inc.). Data presented are representative of five separate experiments that were performed. integrin-mediated adhesion.
Although integrin activation following E-zz/PAF treatment may parallel prolonged Ca 2ϩ mobilization, analysis of the effects of E-zz upon neutrophil O 2 . release provides evidence that E-zz effects were independent of Ca 2ϩ mobilization. Classical priming agents such as PAF or TNF-␣ have no effect upon neutrophil superoxide release when used alone but prime release in response to a second stimulus such as fMLP or the chemokine interleukin-8. cells), confirming that the neutrophils used in these experiments were minimally activated or unprimed (Fig. 6). Similarly, treatment with TNF-␣ alone did not induce neutrophil O 2 . release, but preincubation with TNF-␣ caused a 7-fold increase in fMLP (100 nM)-induced O 2 . release (Fig. 6). In contrast, E-zz treatment alone significantly induces O 2 . release, and preincubation of neutrophils with E-zz prior to fMLP treatment resulted in a further two-fold increase in O 2 . release to levels seen with TNF-␣ and fMLP in combination (Fig. 6). Similar effects were seen when interleukin-8 (10 ng/ml) or PAF (0.1 nM) was used in combination with E-zz (data not shown), demonstrating that distinct mechanisms underlie the effects of . release when compared with classical priming agents such as TNF-␣. These data also suggest that the effects of E-zz upon neutrophil superoxide production and release are largely independent of effects upon Ca 2ϩ mobiliza-  Figs. 2A and 6). Second, our data indicate that E-zz does not affect Ca 2ϩ mobilization in response to fMLP (Fig. 4), yet E-zz is able to convert fMLP from a nonsecretory to a secretory stimulus in terms of O 2 . release (Fig. 6). Together with our data relating to the effects of E-zz on ␤ 2 -integrin-mediated adhesion, we conclude that the specific alteration of Ca 2ϩ mobilization reported here does not have a role in the synergistic effects of E-zz and PAF upon neutrophil O 2 . release but is involved in ␤ 2 -integrin-mediated adhesion. Superoxide Anion Release Induced by Soluble E-selectin Is Tyrosine Phosphorylation-dependent-To further define the signaling pathways involved in neutrophil activation following soluble E-selectin treatment, we next tested the effect of inhibitors of protein kinases involved in intracellular signaling upon E-zz-induced neutrophil O 2 . release. E-zz-induced neutrophil O 2 . release was inhibited by genistein (74 M), a potent tyrosine kinase inhibitor, but not by Ro-318220 (1 M), a protein kinase C inhibitor (Fig. 7). In contrast, PKC-dependent O 2 . release induced by the phorbol ester (PMA) was blocked by Ro-318220 and not by genistein. These results demonstrate that Ro-318220 is indeed active in our assay, because PMA-stimulated O 2 . release is mediated predominantly via PKC activation.
In addition, our results show that genistein is not causing a global nonspecific inhibition of neutrophil responsiveness. These findings suggest that activation of the NADPH oxidase following E-zz treatment is PKC-independent but requires protein tyrosine kinase activity. We are currently investigating specific targets of protein tyrosine kinases in neutrophils, which may be associated with altered neutrophil effector function in response to E-zz treatment.

DISCUSSION
Appropriate recruitment of neutrophils to sites of infection or tissue injury is critical for the initiation and progression of the inflammatory response. A number of studies have shown the importance of selectins in the orchestration of adhesion and migration of neutrophils in the initial capture and subsequent rolling on vascular endothelial ligands (22)(23)(24). Selectins also have a more complex regulatory role in inflammatory processes, initiating intracellular signal transduction cascades that lead to neutrophil activation and altered functional responses (12, 25, 26, 28 -30). We have previously shown that soluble E-selectin (E-zz) specifically potentiates neutrophil adhesion and augments reactive oxygen species production and release (16). In this paper we further examine the nature of the signaling pathway that may underlie these regulatory effects.
The data presented here demonstrate several important findings relating to altered functional status following E-zz treatment, which are consistent with the concept of a juxtacrine activation process demonstrated for neutrophil-platelet interactions involving PAF and P-selectin (15). First, in support of our suggestion that elevated levels of soluble E-selectin may cause alterations in neutrophil behavior that are proinflammatory, E-zz is able to markedly alter neutrophil production of reactive oxygen species. Unlike classical priming agents (PAF or TNF-␣), E-zz is able to trigger O 2 . release when used alone. However, in combination with stimuli that do not normally trigger O 2 . release from unprimed carefully prepared neutrophils (for example the bacterial peptide fMLP), E-zz potentiates O 2 . production (Fig. 6). Second, we have observed that E-zz acts synergistically with PAF but not fMLP or LTB4 to augment ␤ 2 -integrin activity, providing a mechanism for alteration of the adhesion repertoire of neutrophils exposed to specific inflammatory mediators at sites of extravasation. Third, analysis of the effects of E-zz upon Ca 2ϩ mobilization within neutrophils revealed that E-zz specifically prolongs elevation of [Ca 2ϩ ] i in response to PAF but not fMLP or LTB 4 . It has been suggested that [Ca 2ϩ ] i elevation was a necessary step in the signaling between the fMLP receptor and NADPH oxidase (20). Our data show that Ca 2ϩ mobilization is not required for fMLP-induced ␤ 2 -integrin activation. Moreover, we have shown that the effect of E-zz upon PAF-induced ␤ 2 -integrin adhesion is dependent on a threshold increase in [Ca 2ϩ ] i and is blocked when neutrophils are treated with an intracellular Ca 2ϩ antagonist (Fig. 1B). The observed dissociation between the effects of E-zz on ␤ 2 -integrin activation, augmentation of O 2 . release, and Ca 2ϩ mobilization in response to PAF indicates that altered Ca 2ϩ mobilization following PAF receptor ligation is not important for the synergistic action of PAF and E-zz upon O 2 . release.
We have examined a number of parameters of neutrophil activation that might be affected by changes in PAF-induced Ca 2ϩ mobilization, including the extent and duration of neutrophil polarization and whether transient ␤ 2 -integrin mediated adhesion responses are prolonged following E-zz treatment. Although we have observed synergism between E-zz and PAF in terms of ␤ 2 -integrin-mediated adhesion, the precise relationship between these events is difficult to determine. Because E-zz induces a low level of ␤ 2 -integrin activation by itself yet fails to induce Ca 2ϩ mobilization, a distinct functional correlate of prolonged Ca 2ϩ elevation following E-zz and PAF treatment remains to be defined. One possibility is that [Ca 2ϩ ] i elevation may have important consequences for arachidonic acid metabolism. Treatment of macrophages with PAF required a sustained increase in [Ca 2ϩ ] i to induce arachidonic acid release (32). Although the precise mechanisms of E-zz action upon PAF-induced Ca 2ϩ mobilization remain to be determined, it would be interesting to investigate whether neutrophil arachidonic acid metabolism is affected by E-zz/PAF.
We are currently investigating the effects of E-zz upon neutrophil migration in response to PAF, where prolonged elevation of [Ca 2ϩ ] i may interfere with the dynamic activation/deactivation cycles of integrin ligand binding that are necessary for an efficient migration.
PAF induces an initial rapid increase in [Ca 2ϩ ] i because of IP 3 -dependent release of Ca 2ϩ from intracellular stores local-ized in the endoplasmic reticulum (33), which is blocked by TMB-8 and is essential for E-zz effect upon PAF-induced Ca 2ϩ mobilization. The second phase of PAF response is due to Ca 2ϩ influx from extracellular sources through receptor-operated channels. The effects of E-zz upon PAF-induced Ca 2ϩ mobilization was not blocked by SKF 96365, indicating that the sustained [Ca 2ϩ ] i elevation following incubation with E-zz was not due to extracellular Ca 2ϩ entry through SKF 96365-sensitive receptor-operated channels. Alternatively, because the E-zz effect requires a threshold increase in [Ca 2ϩ ] i following Ca 2ϩ release from IP 3 -dependent intracellular stores, the sustained changes in [Ca 2ϩ ] i observed in the presence of E-zz could be due to extracellular Ca 2ϩ entering the cells via store-operated channels (34), which are primed by E-zz and activated by intracellular Ca 2ϩ release from endoplasmic reticulum. One hypothesis for the mechanism of activation of store-operated channels is the release of an intracellular messenger from the endoplasmic reticulum (for example arachidonic acid generated by the activation of phospholipase A 2 ), which would bind to the store-operated channels and increase the probability of opening of the channels (35,36). Another possibility would be an effect of E-zz upon intracellular Ca 2ϩ stores, which may be insensitive to IP 3 .
Inhibition of E-zz effects upon neutrophil O 2 . release by genistein suggests that tyrosine phosphorylation represents a critical step in E-selectin-mediated signaling, possibly controlling downstream functional events, such as the oxidative burst. Although L-selectin-induced Ca 2ϩ mobilization has been reported to be tyrosine phosphorylation-dependent (28), we show for the first time that tyrosine phosphorylation is required for E-selectin induced signaling. Whether the sustained [Ca 2ϩ ] i elevation following E-zz preincubation requires tyrosine phosphorylation has been difficult to assess, partly because PAFinduced Ca 2ϩ mobilization itself is suppressed by tyrosine kinase inhibitors (data not shown). One possibility is that suppression of PAF-induced [Ca 2ϩ ] i elevation was not sufficient to trigger E-zz-induced prolongation of [Ca 2ϩ ] i elevation (Fig. 2B). Cross-linking of neutrophil L-selectin induced a rapid and transient [Ca 2ϩ ] i increase and O 2 . generation, which were both inhibited by genistein (37), indicating that activation of NADPH and Ca 2ϩ mobilization may be downstream of the early signal cascade and require protein tyrosine kinase activation.
In conclusion, we present data that soluble E-selectin can specifically alter PAF induced Ca 2ϩ mobilization in neutrophils but has little effect on Ca 2ϩ mobilization in response to other agonists. Exposure of neutrophils to soluble E-selectin also alters the capacity for production of damaging reactive oxygen species in a manner that is distinct from classical priming agents. In addition, neutrophil ␤ 2 -integrin-mediated adhesion is also altered following exposure to E-selectin, with some degree of synergy observed when E-selectin-treated neutrophils are stimulated with other agonists, e.g. PAF. It is possible that engagement of E-selectin receptors on neutrophils may act physiologically to fine tune processes involved in adhesion and emigration. In view of recent findings that ligation of other neutrophil receptors such as CD31 (11), L-selectin (27), or ICAM-3 (31) influences either ␤ 2 -integrin-mediated adhesion and/or production and release of reactive oxygen species, it seems likely that the repertoire of adhesion receptors engaged provides regulatory signals that determine neutrophil functional status.