Cell Volume-dependent Regulation of L-selectin Shedding in Neutrophils

Neutrophil-mediated organ damage is a common feature of many disease states. We previously demonstrated that resuscitation with hypertonic salt solutions prevented the endotoxin-induced leukosequestration and consequent lung injury, and this effect was partially attributed to an altered surface expression of adhesion molecules, CD11b and L-selectin. In this study we investigated the mechanisms whereby osmotic stress evokes L-selectin shedding. The metalloprotease inhibitor RO 31-9790 prevented the osmotic down-regulation of L-selectin, indicating that this process was catalyzed by the same “sheddase” responsible for L-selectin cleavage induced by diverse inflammatory stimuli. The trigger for hypertonic shedding was cell shrinkage and not increased osmolarity, ionic strength, or intracellular pH. Volume reduction caused robust tyrosine phosphorylation and its inhibition by genistein and erbstatin abrogated shedding. Shrinkage stimulated tyrosine kinases Hck, Syk, and Pyk2, but prevention of their activation by the Src-family inhibitor PP1 failed to affect the L-selectin response. Hypertonicity elicited the Src family-independent activation of p38, and the inhibition of this kinase by SB203580 strongly reduced shedding. p38 was also essential for the N-formyl-methionyl-leucyl-phenylalanine- and lipopolysaccharide-induced shedding but not the phorbol ester-induced shedding. Thus, cell volume regulates L-selectin surface expression in a p38-mediated, metalloprotease-dependent manner. Moreover, p38 has a central role in shedding induced by many inflammatory mediators.

Although neutrophils are indispensable for efficient host defense against invading microorganisms, their uncontrolled or excessive activation is a major contributor to disease. Thus, neutrophil-mediated tissue damage is a key pathologic factor in a variety of conditions including overwhelming infection, ischemia-reperfusion injury, or post-hemorrhage adult respiratory distress syndrome (1,(22)(23)(24). It is therefore not surprising that a variety of approaches have been tried to mitigate neutrophil/ endothelium interactions and neutrophil activation. Potential strategies for intervention include the blockade of adhesion molecules with specific antibodies (10), the neutralization of the released tissue-toxic molecular species such as reactive oxygen intermediates (e.g. by antioxidants) (25), and the interference with signaling pathways or membrane traffic involved in the potentially deleterious activation and degranulation of the cells. A novel example for this latter approach is the application of hypertonic salt solutions, which brings about an efficient, safe, and reversible suppression of numerous neutrophil functions (23, 26 -28). Using an animal model, we have recently shown that an elevation of serum osmolarity protected against neutrophil-mediated post-hemorrhage lung injury (23). Importantly, hypertonicity prevented the lipopolysaccharide (LPS) 1induced up-regulation of CD11b and caused adhesion-independent shedding of L-selectin in vivo and in vitro (23). Interference with proper neutrophil adhesion appears to have contributed significantly to the protective, anti-inflammatory effect of hypertonicity.
To date, the cellular mechanisms underlying the neutrophilsuppressive effects of hyperosmolarity remained largely undefined. The primary aim of the present study was to explore the signaling pathways that might participate in the osmotic shedding of L-selectin. Because we and others have shown that hypertonicity triggers robust tyrosine phosphorylation in various cell types including neutrophils (23, 29 -32), we hypothesized that osmosensitive kinases might be involved in this process. We therefore intended to further investigate which kinases are activated by osmotic stress in neutrophils and to assess their potential contribution to L-selectin shedding. Moreover, the signaling mechanisms leading to L-selectin cleavage upon exposure of the cells to inflammatory stimuli such as formyl peptides (FMLP) or LPS are also poorly under-stood. The involvement of protein kinase C-dependent and independent, presumably tyrosine kinase-dependent, mechanisms have been documented (17), but the relevant tyrosine kinases have not been identified. Thus, in addition to the specific osmolarity-or cell volume-related regulation of L-selectin expression, the present studies were performed to gain further insight into the general mechanisms responsible for the control of the surface expression of this adhesion molecule.
Our results demonstrate that the trigger for the osmotic effect on L-selectin is cell shrinkage, which initiates a p38 mitogen-activated protein kinase-dependent, metalloproteasemediated cleavage of this adhesion molecule. In addition, p38 seem to play an important role in L-selectin shedding evoked by classical inflammatory stimuli as well.

EXPERIMENTAL PROCEDURES
Materials-Ficoll, dextran T500, the enhanced chemiluminescence detection system (ECL), and horseradish peroxidase-coupled anti-rabbit and anti-mouse antibodies were purchased from Amersham Pharmacia Biotech. FMLP, PMA, urea, sucrose, lipopolysaccharide (Escherichia coli O111: B4), bovine serum albumin, diisopropylfluorophosphate, and propidium iodide were from Sigma, and Triton X-100 was from Caledon Lab. BCECF was obtained from Molecular Probes Inc.; genistein, erbstatin analog, SB203580, DEVD-fmk, and staurosporin were from Calbiochem; the proteinase inhibitor mixture was from PharMingen; and ATF-2 was from Santa Cruz Biotechnology. RO 31-9790 was a kind gift from Dr. W. H. Johnson from Roche Products Ltd. All chemicals used were of the highest purity available.
Fluorescein isothiocyanate-conjugated monoclonal anti-L-selectin (CD62L) antibody was purchased from Immunotech, the anti-phosphotyrosine monoclonal antibody (4G10) was from Upstate Biotechnology Inc. and the polyclonal phospho-and nonphospho-specific anti-p38 were from New England Biolab. Monoclonal anti-p72 syk and polyclonal anti-Hck were both from Santa Cruz Biotechnology. Mouse monoclonal anti-Pyk2 was from Transduction Laboratories.
Neutrophil Isolation-Cells were isolated from fresh blood drawn by venipuncture as described earlier (23). Isolated neutrophils were suspended in either Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (Hyclone Lab Inc) plus penicillin/streptomycin (Life Technologies, Inc.) or bicarbonatefree RPMI 1640 (Life Technologies, Inc.) buffered to pH 7.4 with 10 mM HEPES or isotonic saline buffer containing 140 mM NaCl, 5 mM KCl, 5 mM glucose, 1 mM MgCl 2 , 1 mM CaCl 2 , 10 mM HEPES, pH 7.4. Hypertonic medium (500 mosM) was obtained by the addition of extra 100 mM NaCl or 200 mM urea or sucrose into the isotonic medium.
L-selectin Measurement-Surface L-selectin was determined by flow cytometry. Neutrophils were incubated with various inhibitors and exposed to iso-or hypertonicity as detailed under the figures. Subsequently, aliquots of cells were incubated at 1:10 dilution of fluorescein isothiocyanate-conjugated anti-L-selectin antibody for 20 min at 4°C and analyzed on a FACScan (Becton-Dickison, Palo Alto, CA) using FL1 detector (488-nm excitation wavelengths). Typically, 5,000 cells were analyzed per condition. The number of positive cells was determined by comparison of histograms with those obtained using fluorescein isothiocyanate-conjugated nonreactive antibody as described by Preece et al. (11). Soluble L-selectin was measured using an enzyme-linked immunosorbent assay kit from Bender MedSystems, according to the manufacturer's instructions.
Measurement of Cell Volume and Intracellular pH-For cell volume determination neutrophils were suspended in various media as indicated under Fig. 3, and their volumes were assessed by electronic sizing using a Coulter Counter model ZM, equipped with a Channelyzer. For intracellular pH determination, 1 ϫ 10 6 neutrophils suspended in isotonic saline were incubated with 2 M of BCECF for 15 min at 37°C, sedimented, and resuspended in isotonic saline buffer in the presence or absence of 10 M HOE694 for 1 min. The intracellular pH was monitored spectrofluorimetrically using a Perkin-Elmer fluorimeter as in Ref. 33. After obtaining a baseline reading, neutrophils were challenged with hypertonicity by the addition of 100 mM NaCl.
SDS-PAGE and Immunoblotting-Following specified treatments, the neutrophils were rapidly sedimented in a microcentrifuge at 4°C, and the pellet was resuspended in hot Laemmli sample buffer and boiled for 10 min. Samples were subjected to 10% SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane (Schleicher & Schuell), and blocked for 1 h at room temperature in Tris-buffered saline containing 5% bovine serum albumin. The mem-branes were then incubated with the corresponding antibodies at room temperature for 1 h. The dilution was 1:3000 for anti-phosphotyrosine (4G10) and 1:1000 for anti-p38, anti-Syk, anti-Pyk2, and anti-Hck. After washing, the membranes were incubated with either peroxidaseconjugated anti-mouse or anti-rabbit secondary antibody (1:4000) and visualized using ECL.
Immunoprecipitation-To minimize proteolysis, neutrophils were pretreated with 1 mM diisopropylfluorophosphate for 30 min, washed, and resuspended at a concentration of 5 ϫ 10 6 cells/ml in HEPESbuffered RPMI. After specific treatments, the same volume of ice-cold buffer of the corresponding osmolarity was added, and the cells were rapidly sedimented in a microcentrifuge. The cells were dissolved in ice-cold lysis buffer containing 100 mM NaCl, 30 mM HEPES, 20 mM NaF, 1 mM EGTA, 1% Triton X-100, 1 mM sodium vanadate, 20 l/ml proteinase inhibitor mixture. The lysate was kept in ice for 5 min and centrifuged for 10 min at 4°C, and the Triton-insoluble pellet was discarded. The lysate was precleared and incubated with specific antibody for 2 h. Subsequently protein G-agarose beads (40 l) were added, and the lysate was incubated at 4°C for 1 h. Following centrifugation, the immune complexes were sedimented and washed three times with washing buffer (25 mM Tris, pH 7.4, 1 mM sodium vanadate, 150 mM NaCl) and then the precipitate was boiled in LSB (10% glycerol, 5% mercaptoethanol, 2% SDS, 0.025% bromphenol blue, 62.5 mM Tris, pH 6.8) for 10 min.
p38 Kinase Assay-Following immunoprecipitation, the immune complexes were resuspended in 50 l of kinase reaction mix containing 20 mM HEPES, pH 7.6, 200 M MgCl 2 , 20 Ci of [␥-32 P]ATP, 2 mM dithiothreitol, 100 M sodium vanadate, 25 mM ␤-glycerolphosphate, pH 7.2, and 5 g of recombinant fragment of ATF-2 for 15 min at 37°C. The reaction was terminated with hot LSB, and the samples were boiled for 10 min. Proteins were separated by 15% SDS-PAGE, and the phosphorylated ATF-2 was visualized by autoradiography.
Apoptosis Analysis by Flow Cytometry-Neutrophils were incubated with specific inhibitors prior to challenge with iso-or hypertonic medium for 18 h at 37°C. The cells were then centrifuged and resuspended in propidium iodide (500 ng/ml) and stored at 4°C for 30 min. DNA fragmentation analysis was carried out by flow cytometry. Apoptosis was scored by the appearance of a sub-G 0 peak as described by Frasch et al. (34).
Statistical Analysis-Data are presented as mean Ϯ S.E. for n experiments as indicated with duplicate readings within each experiment. Blots are representative of at least three independent studies. Significance was assessed using one-way analysis of variance with posthoc testing using the Student-Newman-Keuls Multiple Comparisons Test. A probability of p Ͻ 0.05 was considered significant.

RESULTS
The Effect of Hypertonicity on Surface Expression of L-selectin-To study the effect of hyperosmolarity on L-selectin expression, the tonicity of the medium containing 1 ϫ 10 6 neutrophils was increased from 290 to 500 mosM by the addition of extra NaCl. After various times of hypertonic exposure, the cells were resuspended in isotonic medium and stained for L-selectin analysis by flow cytometry. Hypertonicity caused a significant, time-dependent reduction in L-selectin expression on the neutrophil surface ( Fig. 1). In contrast to the rapid shedding of L-selectin, which occurs within minutes following receptor ligation or chemokine activation (2,18,19,35), the hypertonicity-elicited response was a slower process, which was clearly detectable in 30 -60 min and progressed over time, reaching almost 100% loss of L-selectin from the surface by 4 h.
Upon stimulation, L-selectin is proteolytically cleaved yielding a soluble fragment (11,12). To confirm that the reduction in L-selectin expression was due to shedding, we determined soluble L-selectin in the supernatant of hypertonically challenged cells, using enzyme-linked immunosorbent assay. Hypertonicity (for 4 h) caused essentially the same increase in soluble L-selectin as a 20-min treatment with phorbol ester, a condition known to cause an almost complete shedding of L-selectin from neutrophils (11) (not shown). Taken together, these data show that hypertonicity induces a gradual shedding of L-selectin that results in decreased cell surface expression.
The Effect of the Metalloprotease Inhibitor RO 31-9790 on the Hypertonicity-induced Shedding of L-selectin-It has been proposed that a membrane-bound protease, referred to as L-selectin sheddase, is the common final pathway catalyzing the cleavage of L-selectin after exposure of the cells to a variety of inflammatory stimuli (11)(12)(13)(14)(15)(16)36). To determine whether hypertonicity-induced shedding of L-selectin is also mediated by this enzyme, a newly developed zinc-chelating hydroxamic acid derivative (RO 31-9790), reported to be a potent sheddase inhibitor was used (11). Consistent with its known activity, preincubation of cells with RO 31-9790 completely abolished the shedding of L-selectin induced by LPS, FMLP or the PKC stimulator PMA. Importantly, the drug was equally effective in preventing the hypertonicity-triggered L-selectin shedding. Specifically, the number of L-selectin-positive neutrophils remained at the level observed in the isotonically treated cells even after prolonged hypertonic incubation (Fig. 2). Because RO 31-9790 has been shown to exert inhibitory effects on matrix proteases (11,15,37), one possible mechanism of its effect on shedding might have been through inhibition of soluble proteolytic factor(s) released in response to hypertonic exposure. To test this possibility, 1.66 ϫ 10 6 neutrophils were incubated in 1 ml of hypertonic solution for 4 h and pelleted, and their supernatant was collected. After the supernatant was diluted 1.66-fold with water to reestablish isotonicity, it was then added to 1 ϫ 10 6 fresh neutrophils for 4 h. The supernatant of hypertonically treated cells caused no shedding of L-selectin from the test neutrophils, arguing against the possibility that hypertonicity induced the release of a soluble proteolytic factor that cleaved L-selectin. To further demonstrate that shedding was not mediated by a matrix proteinase, we added EDTA to the medium, an intervention that selectively inhibits these enzymes without affecting L-selectin sheddase (13,38). Incubation of the cells with 5 mM EDTA had no effect on the hypertonicity-induced shedding (not shown). Together, these findings strongly support the notion that the L-selectin shedding caused by osmotic shrinkage is mediated by the same membrane-bound protease as that responsible for shedding after exposure to other stimuli.
The Role of Cell Shrinkage in the Hypertonic Shedding of L-selectin-In addition to causing cell shrinkage, exposure to a hypertonic environment results in changes of intra-and extracellular osmotic concentration, ionic strength, and intracellular pH. The following experiments were designed to establish which of these variables is critical to trigger shedding. To separate the effect of cell shrinkage from that of medium os-molarity, we used incubation conditions that allowed manipulation of these parameters (30,31). Initially, the medium osmolarity was increased while maintaining the cell volume constant. This was achieved by the addition of 200 mM urea to the medium. Urea is a freely cell-permeant compound and thus rapidly equilibrates across the cell membrane, causing no cell shrinkage but increasing both intra-and extracellular osmolarity. In contrast to hypertonic NaCl, incubation of cells in urea did not alter neutrophil volume (Fig. 3B). Under the hyperosmolar but isovolemic conditions caused by urea, there was no shedding of L-selectin (Fig. 3A). To discern whether a rise in Na ϩ or Cl Ϫ concentration or ionic strength of the medium triggered L-selectin shedding, neutrophil shrinkage was induced by the addition of a nonionic osmolyte in the form of 200 mM sucrose. As shown in Fig. 3B, hypertonicity caused by the addition of 200 mM sucrose resulted in cell shrinkage of a magnitude similar to that observed with hypertonic NaCl (500 mosM) and also provoked a comparable degree of L-selectin shedding (Fig. 3A). These observations indicate that L-selectin shedding is related to a reduction in cell volume and not an increase in extracellular NaCl concentration or ionic strength.
Shrinkage is known to stimulate Na ϩ /H ϩ exchanger 1, the isoform of the Na ϩ /H ϩ exchanger expressed in neutrophils leading to a rise in the intracellular pH of the shrunken cells (39). To evaluate whether Na ϩ /H ϩ exchanger 1-mediated cytosolic alkalization contributes to L-selectin shedding, we compared the effect of hypertonicity in the presence and absence of HOE694, a potent and highly selective inhibitor of the antiporter. As shown on Fig. 4A, hypertonicity caused a sizable cytosolic alkalization that was completely abolished by HOE694. However, HOE694 did not influence the hypertonicity-induced shedding of L-selectin (Fig. 4B), suggesting that the latter process is not caused by the shrinkage-induced alkalization. The combined results of these experiments demonstrate that volume reduction (shrinkage) is the critical factor triggering the osmotic shedding of L-selectin, regardless of changes in osmolarity, ionic strength, or the intracellular pH.
The Role of Tyrosine Phosphorylation in L-selectin Shedding-Protein tyrosine phosphorylation in response to various inflammatory stimuli has been shown to be important in the regulation of surface adhesion molecule expression (40 -43). Recent studies have reported the ability of hypertonicity to induce extensive tyrosine phosphorylation in different cell types, including neutrophils (29 -32). We therefore hypothesized that tyrosine phosphorylation might be involved in shrinkage-induced L-selectin shedding. To address this possibility, we initially used pharmacological means. Specifically, the effect of two broad-spectrum tyrosine kinase inhibitors, genistein and erbstatin, with different mechanisms of action was studied. Fig. 5A shows that hypertonicity causes a large increase in accumulation of phosphotyrosine residues on many proteins, and this effect was substantially inhibited by both inhibitors. In addition, genistein and erbstatin analog significantly reduced the hypertonicity-stimulated shedding of L-selectin (Fig. 5B). This effect was not due to a direct inhibition of the sheddase because the drugs did not affect the PKC-mediated L-selectin shedding (not shown). These findings therefore suggest that tyrosine kinase activation is involved in the shedding of L-selectin following osmotic shrinkage. Further investigation addressed the identity of the tyrosine kinases involved in this process.
Cell shrinkage has been shown to induce tyrosine phosphorylation of important signal transducing molecules in various cell types. These include certain nonreceptor tyrosine kinases (e.g. Hck and Syk) (30,44,45) and members of the focal adhesion kinase family (Pyk2) (46) and the mitogen-activated protein kinase superfamily (extracellular signal-regulated kinase, c-Jun N-terminal kinase, and p38) (34,47). Studies were performed to determine whether specific members of these groups might be involved in shrinkage-induced shedding of L-selectin in human neutrophils.
We first investigated Hck, an abundant Src family kinase expressed in neutrophils. Hck was immunoprecipitated from lysates obtained from iso-and hypertonically treated cells and probed with anti-phosphotyrosine antibody to confirm its phosphorylation following cell shrinkage. As shown on Fig. 6A, shrinkage induced a marked increase in the tyrosine phospho- rylation of Hck. This effect was prevented by genistein and completely abolished by PP1, a newly developed pyrazolo pyrimidine-type compound, shown to be a Src family-specific inhibitor (48). Reprobing the same blots with anti-Hck antibodies confirmed that similar amount of Hck was precipitated from the iso-and hypertonic samples (Fig. 6A, lower panel). Thus, Hck is activated by shrinkage and its phosphorylation is due to either autophosphorylation or phosphorylation by other Srclike kinase(s). Interestingly, however, PP1 failed to inhibit the hypertonic shedding (Fig. 6D). The differential effect of genistein and PP1 was used to investigate the identity of the tyrosine kinases involved in the shrinkage-induced shedding of L-selectin.
The possible role of Syk was then examined. This nonreceptor tyrosine kinase has been shown to be rapidly phosphorylated and activated in B-cells by both osmotic and oxidative stress (44,45), although its osmotic sensitivity has not been examined in neutrophils. As shown in Fig. 6B, Syk was strongly phosphorylated upon cell shrinkage. This increase was impaired by PP1, suggesting that hypertonicity-activated Syk through heterophosphorylation catalyzed by Src-type kinases, because PP1 does not to inhibit Syk itself (48). These findings suggest that Syk is similarly not involved in L-selectin shedding. Further, piceatannol, a Syk-selective kinase inhibitor, also failed to inhibit L-selectin shedding by hypertonicity (data not shown).
One further candidate kinase that might contribute to Lselectin shedding is Pyk2, a calcium-dependent member of the FAK family, which has been shown to be activated by osmotic stress in rat liver epithelial cells (46) and reported to be involved in the integrin-related signaling of hematopoetic cells (49,50). As shown on Fig. 6C, Pyk2 was constitutively phosphorylated to a modest extent in control isotonic neutrophils, and shrinkage induced further accumulation of phosphotyrosine in this protein. PP1 completely abolished not only the hypertonicity-induced increase but also the basal tyrosine phosphorylation of this kinase.
To rule out that the lack of effect of PP1 on shedding was not simply due to its metabolism or spontaneous decomposition, we tested whether PP1 could inhibit phosphorylation even after 2 h of hypertonic treatment (i.e. when the shedding was measured). As shown in Fig. 6 (A-C, 2 h panels), hypertonicity induced sustained tyrosine phosphorylation in all three kinases. This response was, however, completely prevented in the presence of PP1, confirming that the drug exerted a longlasting effect.
Therefore, each of the three kinases evaluated proved to be osmo-sensitive in neutrophils, and their phosphorylation was strongly inhibited by PP1, a finding suggesting that that the phosphorylation was catalyzed by the Src family. However, despite their activation, these kinases do not seem to be responsible for the hypertonicity-induced L-selectin shedding.
The Role of Mitogen-activated Protein Kinase p38 -Three distinct families of mitogen-activated protein kinases have been identified in mammalian cells: p42/44 extracellular signal-regulated kinases, c-Jun N-terminal kinase/stress-activated protein kinases, and p38. Although it has been established that in neutrophils neither p42/44 extracellular signalregulated kinases nor c-Jun N-terminal kinase/stress-activated protein kinase (except in extreme hypertonicity) are activated by osmotic stress, the studies concerning osmotic activation of p38, have reported conflicting results (30,32,34).
To investigate whether shrinkage triggers phosphorylation of p38, the cells were subjected to hypertonic stress, lysed, and immunoblotted with a specific anti-phospho-p38 antibody. As shown in Fig. 7A, hypertonic exposure of the cells for 10 min caused a modest but readily detectable phosphorylation of p38. Because the effect of hypertonicity on L-selectin was a slow process, studies were performed to discern whether more prolonged hypertonic treatment might induce a persistent and stronger response. Fig. 7A shows that by 2 h of hypertonic exposure, there was a marked increase in p38 phosphorylation, a time course comparable with that of L-selectin shedding. This observation supports the involvement of this kinase in the shrinkage-induced shedding of L-selectin (Fig. 7, A and B). To verify whether the increased phosphorylation was associated with an increased kinase activity even after 2 h of treatment, p38 activity was determined in immunocomplex kinase assays using ATF-2 as exogenous substrate. Fig. 7B shows that the kinase activity of p38 obtained from hypertonically treated cells was significantly increased compared with the isotonic controls. In addition, treatment of the cells with genistein completely prevented p38 phosphorylation, whereas PP1 had absolutely no effect on this process (Fig. 7, A and B). This inhibitory profile correlated with the differential ability of genistein and PP1 to affect the hypertonicity-induced shedding of L-selectin (Figs. 5B and 6D, respectively).
To determine whether p38 activation is in fact involved in L-selectin shedding, neutrophils were incubated with the inhibitor SB203580 before exposure to hypertonicity. SB203580 is a highly specific inhibitor of the p38 kinase activity with no effect on p42/44 extracellular signal-regulated kinases and c- Jun N-terminal kinase/stress-activated protein kinases (51). Because in suspension of human neutrophils substantial inhibition of p38 was attained at 10 -30 M of SB203580 (52), we routinely applied 20 M. As demonstrated in Fig. 7C, SB203580 inhibited the hypertonic shedding of L-selectin, suggesting that p38 activation is required for this process. This novel finding led us to investigate whether p38 might play an important signaling role in the shedding of L-selectin induced by other stimuli such as LPS, FMLP, and PMA. LPS and FMLP are strong stimulators of p38 in neutrophils (53,54), whereas PMA acts through the activation of PKC and has little effect on p38 (54). Inhibition of p38 by SB203580 abrogated the shedding caused by LPS and FMLP but not by PMA (Fig. 7C). Conversely, the PKC inhibitor staurosporin prevented the PMAinduced shedding but had no effect on the LPS-or hypertonicity-induced response (data not shown). These data suggest that p38 plays a major role in the hypertonicity-, LPS-, and FMLPmediated shedding of L-selectin, whereas the direct stimulation of PKC induces shedding in a p38-independent manner.
The Relationship between Apoptosis and the Hypertonic Shedding of L-selectin-Aging neutrophils undergo apoptosis spontaneously, a process that is accompanied by shedding of L-selectin (55-57). Frasch et al. (34) recently reported that p38 is involved in the signaling of the stress-activated neutrophil apoptosis. It was therefore conceivable that L-selectin shedding might be a consequence of the hypertonicity-induced apoptosis or vice versa. To investigate the relationship between these events, we examined whether the inhibition of apoptosis interferes with shedding, and conversely, whether the inhibition of shedding interferes with apoptosis.
To abrogate apoptosis we used DEVD-fmk, a potent caspase-3 inhibitor. Neutrophils suspended overnight in isotonic Dulbecco's modified Eagle's medium/fetal calf serum underwent spontaneous apoptosis that was significantly in-creased by hyperosmolar stress (Fig. 8A). Treatment with DEVD-fmk protected the cells from the hypertonicity-induced apoptosis but did not inhibit the hypertonicity-induced shedding of L-selectin (Fig. 8). Further, treatment with the sheddase inhibitor RO 31-9790 prevented the hypertonic shedding ( Fig. 8B) but had no effect on the hypertonicity-induced increase in apoptosis (Fig. 8). These data suggest that the hypertonicity-induced shedding of L-selectin is not a consequence of apoptosis. Although p38 is involved in the upstream signaling of both processes, shedding and apoptosis are not necessarily coupled to each other. DISCUSSION The present study provides evidence that hypertonicity induces shedding of L-selectin from the surface of neutrophils. This process is triggered by a decrease of cell volume, is mediated by a RO 31-9790-sensitive protease (sheddase), and involves tyrosine kinase activation. Although several candidate tyrosine kinases appear to be osmo-sensitive, our studies determined that p38 stress kinase plays an important role in L-selectin shedding in response to hypertonicity.
Other pro-inflammatory stimuli including bacterial peptides (53,58), LPS (54), and cytokines (59,60) have been previously reported to activate p38 in neutrophils, although the role of this kinase in mediating the functional responses to these stimuli has not been fully elucidated. Activation of p38 appears to participate in the regulation of phospholipase A 2 (52), in the FMLP-or tumor necrosis factor-␣-induced superoxide production (59,61), in ␤ 2 -integrin-dependent adhesion and oxidative burst (62), and in stress-induced apoptosis (34). To our knowledge, the present studies are the first to implicate p38 as an essential mediator for the shedding of L-selectin. The central role of this stress kinase is underlined by our finding that its inhibition completely abolished the shedding triggered by both for either a short term (1 min for Hck, 5 min for Syk, and 2 min for Pyk2) or long term (2 h). Using specific monoclonal antibodies, the kinases were immunoprecipitated as described under "Experimental Procedures," separated by SDS-PAGE, and probed with anti-phosphotyrosine (4G10) mAb. The same membranes were stripped and reprobed with the corresponding antibodies to verify equal loading. One representative experiment of three similar ones is shown for each group. genist, genistein; PY, tyrosine phosphorylation. D, the lack of effect of PP1 on hypertonicity-induced shedding of L-selectin. Neutrophils were treated with PP1 (10 M) for 20 min prior to and during exposure to hypertonicity (100 mM extra NaCl) for 2 h. The data represent the means Ϯ S.E., n ϭ 10 separate studies. *, p Ͻ 0.001 versus control; HT, hypertonic treatment. by LPS and by FMLP and strongly abrogated the process induced by hypertonic stress. Further, activation of p38 (e.g. by LPS or hypertonicity) can trigger L-selectin shedding in the presence of PKC inhibitors, suggesting independence from the PKC signaling pathway. In support of this notion, L-selectin shedding elicited via the direct stimulation of PKC by phorbol esters was insensitive to the inhibition of p38. PKC-independent but tyrosine kinase-dependent shedding of L-selectin has been recently reported in lymphocytes after ligation of the Leu-13 surface antigen (17). Considered together, these results indicate that the activation of PKC and p38 represent two distinct signaling pathways contributing to the regulation of L-selectin shedding.
The molecular mechanisms by which p38 activation leads to cleavage of L-selectin remain to be elucidated. It is conceivable that p38 is involved in the activation or membrane insertion (exocytosis) of the sheddase, or it may be necessary for rendering L-selectin susceptible for proteolytic cleavage. In this regard, the cytosolic tail of L-selectin has also been shown to bind to various cytoskeletal molecules (63). Phosphorylation of these interacting proteins may result in a conformational change that exposes the membrane proximal ecto-domain of L-selectin to a constitutively active sheddase (2,11). Alternatively, Lselectin could be a direct substrate of p38. Direct phosphorylation of L-selectin on serine residues has been shown to occur after activation of PKC in rat basophilic leukemia cells (64). However, this reaction seemed to be independent of the subsequent shedding, because no reduction was observed in the cleavage of L-selectin variants in which target serines were exchanged for alanines.
Hypertonicity has been shown to stimulate p38 in many different cell types (31,32,65). Although an earlier study reported that p38 may not be affected by hypertonicity in neutrophils (30), recent data, including those described herein, unambiguously show that this kinase is both phosphorylated and stimulated by osmotic shock in neutrophils (32,34). However, there are striking differences in the kinetics of p38 activation by the various stimuli, a finding that provides a plausi- hypertonicity (100 mM extra NaCl) for 10 min or 2 h. Cells were lysed, subjected to SDS-PAGE, and immunoblotted using polyclonal anti-phospho-p38. The same blot was then stripped and reprobed with an anti-p38 antibody. B, following the same treatment as in A and exposure to hypertonicity for 2 h, p38 was immunoprecipitated with a specific anti-p38 antibody. In vitro kinase assay was then performed using ATF-2 as a substrate, and the results were visualized with autoradiography. C, neutrophils were treated in the presence or absence of the p38 inhibitor SB203580 (20 M) for 20 min. The medium was then supplemented with an extra 100 mM NaCl for 2 h (HT), LPS (1 g/ml) for 1 h, FMLP (100 nM), or PMA (50 nM) both for 20 min. L-selectin surface expression was measured using a mAb and flow cytometry. The data represent the means Ϯ S.E., n ϭ 6 separate studies. *, p Ͻ 0.001 versus control (CON); **, p Ͻ 0.05 versus control; ***, p Ͻ 0.01 versus control; Genist., genistein; HT, hypertonic medium. ble explanation for the apparent controversy. Thus, FMLP provokes a robust activation within a minute, whereas LPS requires approximately 20 min for maximal action. In contrast, the effect of hypertonicity, although detectable after 10 min, continues to increase for as long as 2-3 h. For each stimulus, the time dependence of p38 activation correlates well with the kinetics of L-selectin shedding, further confirming the causal relationship between the two phenomena. An interesting and possibly functionally important feature of this slow kinetics is that the slight p38 activation induced by hypertonicity at early time points may render the kinase refractory to the much stronger stimulatory effects of FMLP and LPS. For example, Junger et al. (32) demonstrated that hypertonicity impaired FMLP-induced exocytosis, whereas our group recently reported that osmotic shrinkage inhibited LPS-stimulated CD11b upregulation (66). In both studies, exposure to hypertonicity inhibited ligand-induced activation of p38 stress kinase, without altering the interaction of the stimulus with the surface ligand. Further studies are warranted to elucidate the mechanisms contributing to the refractoriness.
The present studies showed that two important tyrosine kinases, Syk and Pyk2, are osmotically activated in neutrophils and are probably downstream from the Src family. Rather surprisingly, however, these kinases do not seem to be responsible either for the osmotic activation of p38 or for the osmotic shedding of L-selectin. This finding does not exclude the possibility that the activation of Src family members may play a role in the shedding induced by other stimuli. In fact, the LPSand to a lesser extent the FMLP-triggered L-selectin release was partially sensitive to PP1, suggesting that the Src family may be involved in p38 stimulation induced by these agonists (not shown).
Hypertonicity has been previously shown to interfere with several neutrophil functions including superoxide production (26), phagocytosis (27), adhesion, and transmigration (23,28). In addition to these inhibitory effects on neutrophil function, the present studies show that shrinkage actively promotes the premature loss of an important surface adhesion molecule. This loss may impair not only the initial endothelium-neutrophil adhesive interactions but also the L-selectin-mediated intracellular signaling, which helps prime cells for augmented superoxide production and attachment induced by subsequent stimuli (19). Further, because hydroxamate-sensitive metalloprotease(s) are involved in the shedding of other surface molecules including tumor necrosis factor and Fc receptors (67), it is conceivable that these proteins may also be lost from the surface of shrunken cells, rendering them unresponsive to their corresponding ligands. When considered together, these findings suggest that hypertonic solutions may exert a net antiinflammatory effect, in addition to their resuscitative effects. The therapeutic application of hyperosmotic fluids may represent a safe, reversible, and inexpensive means of lessening organ dysfunction in disease processes characterized neutrophil-mediated tissue injury such as ischemia/reperfusion and adult respiratory distress syndrome.
In summary, we show that cell volume regulates the expression of L-selectin on the surface of neutrophils. This observation led us to the recognition that p38 stress kinase plays an important general role in the control of L-selectin shedding and thereby in the regulation of adhesive interactions of neutrophils with their environment.