The phosphotyrosine phosphatase inhibitor vanadyl hydroperoxide induces morphological alterations, cytoskeletal rearrangements and increased adhesiveness in rat neutrophil leucocytes

Neutrophils are actively motile cells whose primary role in the acute inflammatory response is to emigrate from the bloodstream into the tissues. The emigration phase can be broadly divided into three categories: (i) adhesion to acti-vated endothelium; (ii) transmigration between endothelial cells; and (iii) migration through the interstitial tissues towards an inflammatory focus. Central to this process is the regulation and stimulation of cell motility, including the recognition of a gradient of chemoattractant (Zigmond, 1974).

Neutrophil movement has been the subject of intense investigation, and a number of cytoskeletal proteins have been implicated (Howard et al., 1990; Niggli and Jenni, 1989); however, although increases in filamentous actin have been documented, the mechanism underlying the regulation of motility has not yet been defined and in particular the signalling pathway leading to enhanced motility has yet to be elucidated (Downey et al., 1990; Southwick and Young, 1990; Bengtsson et al., 1990; Howard et al., 1990). In vitro, treatment of neutrophils with chemotactic factors such as formylmethionyl-leucyl-phenylalanine (fMLP) results in a variety of responses, including enhanced motility, increased phagocytic activity, and production of superoxide due to activation of an NADPH-dependent oxidase. Binding of fMLP also results in the production of several intracellular signals. Phospholipase C is rapidly activated by a G-protein-mediated process, leading to the hydrolysis of phosphatidylinositol 4,5-bisphosphate (PI-4,5-P₂) to generate the second messengers inositol 1,4,5-trisphosphate (IP₃) and diacyl-glycerol as well as other products (Stephens et al., 1991) that result in increases in intracellular calcium and activation of protein kinase C (PKC), respectively. Whilst there have been several reports documenting the activation of PKC during PMN stimulation, the role of PKC-mediated phosphorylation in cell locomotion is complex, since several inhibitors of PKC have been shown to either stimulate or inhibit locomotion, depending on the conditions (Gaudry et al., 1988; Keller et al., 1990; Niggli and Keller, 1991). At present it is not clear how these signalling events relate to changes in cell motility. Recently, it has been shown that when neutrophils are stimulated by chemotactic factors there is a rapid increase in tyrosine phosphorylation. Evidence that this plays a functional role is derived from reports that tyrosine kinase inhibitors such as erbstatin or ST638 will inhibit several chemotactic factor-induced responses (Gomez-Cambronero et al., 1989; Naccache et al., 1990; McColl et al., 1991; Fuortes et al., 1993). However, the putative kinases that are activated to give rise to the enhanced tyrosine phosphorylation have not been identified. In cells tyrosine phospho-rylation is regulated by a balance between tyrosine kinase activity and tyrosine phosphatase activity; thus, it is possible that in addition to differential kinase activity, there may also be changes in phosphatase activity when cells are stimulated. In support of this, the potent tyrosine phosphatase inhibitor vanadate stimulates oxygen consumption by per-meabilised neutrophils, which is accompanied by increased protein tyrosine phosphorylation (Grinstein et al., 1990). Such studies are difficult in whole cells, however, because the plasma membrane is relatively impermeable to sodium orthovanadate. Therefore, in order to investigate the role of tyrosine phosphorylation in neutrophil locomotion in unpermeabilised cells we have used a membrane-permeant derivative, vanadyl hydroperoxide (pervanadate) to study the functional consequences of inhibiting phosphotyrosine phosphatase activity in neutrophils. We report here that, in addition to causing the rapid accumulation of phosphoty-rosyl proteins, pervanadate induces a motile morphology concurrent with a reorganisation and transient changes in the polymerisation state of cellular actin, and increased cell spreading.

All general chemicals were Analar grade or better, from BDH. Sodium orthovanadate, phenylmethylsulfonyl fluoride (PMSF), hydrogen peroxide (30% stock), 3′,3′-diaminobenzidine, NBD-, rhodamine- and FITC-labelled phallicidin were purchased from Sigma. Erbstatin analogue and tyrphostin 47 were purchased from Calbiochem. Monoclonal anti-phosphotyrosine 4G10 was obtained from Upstate Biotechnology Inc. All other secondary antibodies were obtained from Bio-Rad.

Rat neutrophils were prepared by intraperitoneal injection of 0.5% oyster glycogen (Sigma) in sterile saline into Wistar rats; 18 hours later the animals were killed and the cells were harvested by peritoneal lavage. The cells were washed three times in balanced salt solution (BSS: 8 g/l NaCl, 0.4 g/l KCl, 0.2 g/l MgCl₂6H₂0, 0.14 g/l CaCl₂, 1 g/l glucose, 2.388 g/l Hepes, pH 7.4) and resuspended at 5×10⁷ cells per ml for use. Differential staining revealed that this suspension was >85% neutrophil polymorphonuclear leucocytes.

Pervanadate was prepared by mixing equimolar concentrations of hydrogen peroxide and sodium orthovanadate. Equal volumes of 10 mM of each were mixed to give 5 mM vanadyl hydroperoxide (pervanadate). This stock solution was diluted to give various concentrations of pervanadate (1 μM to 1 mM).

Purified rat neutrophils in BSS were incubated at 37°C for 15 minutes with various concentrations of fMLP, H₂O₂ and pervanadate (1 μM to 1 mM). After the incubation period, the neutrophils were fixed by the addition of one-tenth volume 2.5% glutaraldehyde for 10 minutes at room temperature. The cells were then examined under a Zeiss Axioskop microscope and scored for shape-changed (polarised) cells. At least 100 cells were counted for each mea-surement point.

A total of 5×10⁵ cells were plated onto 8-well Lab-tek glass slide chambers (Nunc) in BSS + 1% bovine serum albumin (BSA). After allowing the cells to settle for 5 minutes, 500 μM (final con-centration) pervanadate or hydrogen peroxide was added. At the indicated times formaldehyde was added to a final concentration of 3% for 10 minutes, and the slides were processed for F-actin staining as described below. In some experiments cells were pre-incubated with tyrosine kinase inhibitors for 15 minutes at 37°C prior to plating. Spread cells were quantified by counting the number of cells exhibiting flattened lobed nuclei and extensive cytoplasm with a diameter >150% of that of unstimulated cells using phase-contrast optics and F-actin staining.

Cells were suspended at 2×10⁷/ml, stimulated with 10⁻⁹ M fMLP or 200 μM pervanadate for 15 minutes and fixed by the addition of an equal volume of 7.4% formaldehyde in phosphate-buffered saline (PBS). A 100 μl portion of 0.2% Triton X-100 was added to 100 μl of fixed neutrophils, for 10 minutes to permeabilise the cells. The cells were stained for F-actin by the addition of rho-damine-labelled phalloidin (Sigma) to a final concentration of 330 nM for 1 hour at room temperature, mixing every 15 minutes to resuspend the cells. The cells were washed three times, resus-pended in Citifluor (Citifluor Ltd, London) and viewed using a Zeiss Axioscop microscope equipped with epifluorescence. Double fluorescence staining was performed by a modification of the procedure described previously (Horvath et al., 1990), stain-ing the permeabilised cells with monoclonal anti-phosphotyrosine at 2 μg/ml (4G10, Upstate Biotechnology), followed by biotinylated swine anti-mouse Ig Fab fragments (Dako), each for 1 hour with three washes. At this stage fluorescein-labelled streptavidin and rhodamine-labelled phalloidin were added simultaneously and the cells washed as described above.

Rat PMNs were suspended at 10⁸ cells/ml and stimulated with pervanadate or fMLP as described above. The cells were lysed by the addition of an equal volume of twice strength reducing SDS sample buffer (Laemmli, 1970) and immediately boiled for 5 minutes. Insoluble debris were removed by centrifugation and 15 μl aliquots run on 10% SDS-polyacrylamide minigels (Bio-Rad) under reducing conditions. The proteins were transferred to nitro-cellulose using a semidry transfer apparatus (Bio-Rad) and the nitrocellulose was immersed in 5% non-fat milk powder for 1 hour to block nonspecific binding sites. The nitrocellulose was then washed on a shaker 3 times for 10 minutes in PBS containing 0.05% Tween-20. Anti-phosphotyrosine monoclonal antibody 4G10 at 1 μg/ml was then added to the nitrocellulose and incubated at room temperature for 1 hour with shaking. The nitrocellulose was washed 3 times for 10 minutes each with PBS/Tween, and a 1:3000 dilution of peroxidase-conjugated goat anti-mouse Ig (Bio-Rad) in PBS/Tween was added to the nitrocellulose for 1 hour. This was washed 3 times for 10 minutes in PBS/Tween, twice for 10 minutes with PBS, rinsed in distilled water, developed with 50 mg 3′,3′-diaminobenzidine (Sigma) in 100 ml PBS and 100 μl 30% H₂O₂, and finally dried on blotting paper. In some experiments 10 mM free phosphotyrosine was added to the pri-mary antibody as a specificity control.

After activation 2×10⁷ neutrophils were extracted with 200 μl of extraction buffer (BSS containing 1% Triton X-100 plus 1 mM PMSF) for 10 minutes on ice. The insoluble residue was pelleted in a microcentrifuge for 2 minutes at 10,000 g and washed with 200 μl of extraction buffer. The pellet (cytoskeleton) was solubilised in 450 μl SDS sample buffer, and 40 μl of 10× SDS sample buffer was added to the 400 μl of soluble material. Equal volumes containing equal amounts of cell equivalents were loaded into each well for western blotting.

F-actin was quantified using a phalloidin binding assay followed by flow cytometry (Howard et al., 1990). Briefly 2×10⁵ cells in 100 μl were activated and an equal volume of PBS was added, containing 6.4% paraformaldehyde, 200 μg/ml lysophosphatidyl-choline and 0.6 μM NBD-phallicidin (Sigma), for 30 minutes. The cells were washed twice in PBS containing 0.1% BSA and resus-pended in 1 ml of the wash buffer. The cells were then analysed by flow cytometry using a FACScan (Becton Dickinson); 10⁵ cells were analysed for each time point and the mean fluorescence per cell was expressed in arbitrary units.

Neutrophils are naturally motile cells and their velocity can be stimulated by the addition of chemotactic factors. How-ever, common measurements of motility rely on migration across 2- or 3-dimensional surfaces, which can also be influenced by changes in substratum adhesion. Change of cell shape from spherical to non-spherical has been shown to be a close correlate of stimulated motility and since this is performed in suspension there is no interference from changes in adhesiveness (Keller 1983). We therefore used this assay to examine the induction of the motile pheno-type in rat peritoneal neutrophils. Washed peritoneal exudate neutrophils exhibited a spherical morphology (Fig. 1A); however, after treatment with 10⁻⁹ M fMLP > 80% of the cells became non-spherical, extending pseudopodia and membrane processes (Fig. 1B). After treatment with 200 μM pervanadate alone the cells exhibited a morphology similar to that of fMLP-stimulated cells with >75% of the population showing a polarised morphology with mem-brane processes or pseudopodia (Fig. 1D). Control cells treated with 200 μM H₂O₂ showed no shape change and were indistinguishable from untreated cells (Fig. 1C). The concentrations of pervanadate necessary to induce the motile morphology were investigated. Dose-response studies showed that 50% activation of shape change occurred at 40-50 μM pervanadate, it reached a plateau at about 200 μM and was then maintained with >80% of the cells showing shape changes (Fig. 2A). Control experiments showed that Na₃VO₄ or H₂O₂ at the same concentrations had little effect on morphology (Fig. 2A). To confirm that the morphological changes were due to the vanadium ion concentration and not the peroxide concentration, cells were treated with various concentrations of vanadate whilst maintaining the H₂O₂ concentration at 100 μM. The dose-response for various vanadate concentrations showed a curve similar to that of pervanadate concentrations (Fig. 2B). Thus the shape changes were induced by the vanadate moiety of the pervanadate solution.

The time course of shape change induced by pervanadate was investigated. The development of shape change in pervanadate-treated cells was rapid, with more than 50% of the cell showing morphological changes within 1.5 minutes. This was slower than with fMLP; however, shape change was still rapid, with >70% of the cells changed within 3 minutes of stimulation, and this motile morphology was maintained for at least 20 minutes (Fig. 3).

Stimulation of neutrophils commonly results in changes in the localisation and configuration of actin. We examined the effects of pervanadate on both actin localisation and polymerisation state using fluorescent phalloidin. In resting cells F-actin staining was diffuse throughout the cell (Fig. 4A,B). Treatment with fMLP resulted in a reorganisation and polarisation of F-actin to restricted, subcortical regions (Fig. 4C,D). A similar reorganisation of F-actin was found when the cells were treated with 200 μM pervanadate (Fig. 4G,H); however, control experiments showed that 200 μM H₂O₂ had no effect on F-actin localisation (Fig. 4E,F). In both fMLP- and pervanadate-stimulated cells the F-actin was commonly concentrated in uropod-like regions (Fig. 4C,G).

The effect of pervanadate on the polymerisation state of actin was also investigated using FITC-phalloidin staining followed by flow cytometry. fMLP induced a 3-to 4-fold increase in F-actin content, which peaked at about 20 sec-onds and then gradually declined (Fig. 5). Pervanadate also induced an increase in cellular F-actin; however, this was slower than that induced by fMLP, peaking at about 4 min-utes, with an initial slight delay (Fig. 5). Although the peak was smaller than that induced by fMLP, because the basal level was slightly depressed there was still a 3-fold increase in F-actin content induced by pervanadate, compared with unactivated cells. After pervanadate treatment the amount of F-actin remained elevated, and declined at a much slower rate than that induced by fMLP (initial t_G of fMLP-induced actin decline = 10 minutes; initial t_G of pervanadate-induced F-actin decline = 30 minutes).

Treatment of adherent PMNs with pervanadate resulted initially in enhanced membrane activity including ruffles and blebs (data not shown), followed by spreading on the sub-stratum (Fig. 6). The time course of pervanadate-induced spreading indicated that maximal spreading occurred after 30 minutes; however, there was evidence of spreading as early as 5 minutes after pervanadate addition, with many of the cells fully spread after 15 minutes (Fig. 6A,B,C). Cells treated with 500 μM H₂O₂ for 30 minutes showed no induction of spreading (Fig. 6D) and unstimulated cells only partially spread, with some cells adopting a polarised mor-phlogy (Fig. 6E). At higher power the extent of spreading could be seen more clearly as evidenced by the flattened, lobed nucleus seen under phase-contrast optics (Fig. 6F). Spreading was accompanied by a redistribution of F-actin to a peripheral localisation near the plasma membrane, particularly in membrane ruffles, as the cells spread (Fig. 6C, G, arrow).

To ascertain that pervanadate inhibited phosphotyrosyl phosphatases, pervanadate-treated PMN lysates were west-ern-blotted using a monoclonal anti-phosphotyrosine. Dose-response studies of pervanadate treatment revealed that changes in phosphotyrosyl proteins were detectable with 20 μM pervanadate and increased in intensity up to 50 μM (Fig. 7A). The time course of blotting showed that there was an increase in phosphotyrosyl proteins, detectable within 1 minute and peaking at 5 minutes (Fig. 7B, lanes 3-9). No increases in phosphotyrosyl proteins were found when cells were treated with H₂O₂ alone (Fig. 7B, lane 2). The major phosphotyrosyl bands were found at 20 kDa, 40 kDa, 60 kDa, 70 kDa, 75 kDa and 105 kDa. Free phos-photyrosine blocked the detection of phosphotyrosyl protein on the nitrocellulose, demonstrating the specificity of the antibody for tyrosine-phosphorylated proteins (data not shown).

Although we found a close correlation between activation of cells by pervanadate and increased protein tyrosine phosphorylation, we wished to determine more directly whether pervanadate activation was mediated by increases in protein phosphotyrosine resulting from inhibition of phosphatases. We therefore examined the effects of pre-incubation of cells with the tyrosine kinase inhibitors tyrphostin and erbstatin.

Pre-incubation of cells with 90 μM of the tyrosine kinase inhibitor tyrphostin 47 substantially inhibited the pervanadate-induced increases in protein tyrosine phosphorylation. This concentration of tyrphostin also inhibited fMLP- and pervanadate-induced shape change by 80-90%, although the DMSO vehicle also had some inhibitory activity (Table 1). However, 20 μM tyrphostin inhibited both pervanadate and fMLP-induced shape change by 40-50% whereas the equivalent concentration of vehicle had no effect (Table 1). In addition, 20 μM erb-statin completely blocked fMLP- and pervanadate-induced shape change (Table 1) and actin polarisation (data not shown). Similar results were obtained when the effect of tyrosine kinase inhibitors on pervanadate-induced cell spreading was examined. 20 μM tyrphostin partially inhibited cell spreading, and 90 μM tyrphostin completely inhibited spreading, as did 20 μM erbstatin (Table 2). Tyr-phostin at 90 μM also significantly inhibited actin poly-merisation (Fig. 8). Thus inhibition of constitutive tyro-sine kinase activity resulted in lack of activation of cells by pervanadate or fMLP. This strongly implies that the action of pervanadate on cell function is through its effects on tyrosine phosphorylation.

Since pervanadate treatment resulted in cytoskeletal changes concurrent with increases in phosphotyrosyl pro-teins the localisation of tyrosine-phosphorylated proteins with respect to F-actin was investigated using immunoflu-orescence. In untreated cells staining for phosphotyrosyl proteins was faint and diffuse (Fig. 9B); however, after fMLP treatment phosphotyrosine staining was stronger and the phosphotyrosyl proteins co-localised with F-actin (Fig. 9C,D; arrows). Similarly, H₂O₂-treated cells showed faint phosphotyrosine staining (Fig. 9F), but pervanadate treat-ment resulted in increased phosphotyrosine staining and co-localisation of phosphotyrosyl proteins with F-actin to a restricted cellular location commonly at one pole of the cell and in pseudopods (Fig. 9G,H, arrows). Western blotting of PMNs after fractionation with 1% Triton X-100 into detergent-resistant (cytoskeleton) and detergent-soluble fractions demonstrated that a substantial number of tyro-sine-phosphorylated proteins remained associated with the PMN cytoskeleton after activation with pervanadate (Fig. 10).

Chemotactic factors such as fMLP or interleukin-8 rapidly stimulate neutrophils into enhanced motility. The profound effects of pervanadate on neutrophil morphology, adhesiveness and cytoskeletal reorganisation strongly imply that these events are regulated by tyrosine phosphorylation. Whilst increases in tyrosine phosphorylation have been observed after chemotactic factor-mediated activation of neutrophils, our results provide more direct evidence that increased tyrosine phosphorylation activates the motile apparatus of neutrophils. Experiments directly measuring PMN motility using a computer-assisted time-lapse video micrography system showed no apparent increase in motil-ity in response to pervanadate (data not shown). This was due to the increased adhesiveness that anchored the cells to the substratum, thereby inhibiting movement, and this occurred at all pervanadate concentrations. This does not occur with chemotactic (10⁻⁹ M) concentrations of fMLP, but only at much higher (>10⁻⁶ M) fMLP concentrations. Treatment of neutrophils with GM-CSF results in tyrosine phosphorylation of a number of cellular proteins; however, pretreatment with the tyrosine kinase inhibitor erbstatin inhibits GM-CSF-mediated tyrosine phosphorylation in addition to inhibiting intracellular alkalinisation, downreg-ulation of LTB₄ receptors, priming of calcium fluxes and c-fos expression (McColl et al., 1991). Erbstatin at 25 μM also blocks fMLP-induced tyrosine phosphorylation, alkalinisation and superoxide production, but not calcium fluxes, actin polymerisation or secretion of granule com-ponents (Naccache et al., 1990) and, similarly, treatment with another tyrosine kinase inhibitor ST638 blocks fMLP-induced tyrosine phosphorylation and superoxide production, but not actin polymerisation (Gomez-Cambronero et al., 1989). In contrast to these results, we have found that tyrphostin and erbstatin inhibit pervanadate-induced shape change and actin polymerisation. However, in agreement with other groups we found that tyrphostin did not inhibit fMLP-induced actin polymerisation (data not shown) although it did inhibit shape change, suggesting that these different events may be mediated by different pathways. Our results showing inhibition of pervanadate-mediated responses at tyrphostin concentrations of 20-90 μM are consistent with inhibition of tyrosine kinases, since the IC₅₀ of tyrphostin 47 for pp60^v-src in vitro is about 8 μM (O’Dell et al., 1991). These results also demonstrate that there is no strict relationship between the transient net actin polymerisation and maintained cell shape change after agonist stimulation. Activation of PKC promotes cytoskeletal association of NADPH oxidase components, and this is inhibitable by staurosporine (Nauseef et al., 1991). The role of PKC in cytoskeletal organisation appears to more be complex, since the PKC inhibitor CGP41251 had no effect on fMLP-induced F-actin increases, whereas staurosporine was inhibitory, suggesting the involvement of a PKC-like pro-tein (Niggli and Keller, 1991).

The coupling of tyrosine phosphorylation to other signalling molecules has been investigated by several groups. Uings et al. (1992) found that tyrosine kinase inhibitors inhibited fMLP-induced phospholipase D (PLD) activity, but not IP₃ generation, and that pervanadate stimulated PLD activity, suggesting that tyrosine phosphorylation is coupled to PLD but not to PI-4,5-P₂-specific phospholipase C in neutrophils. There are at least two fMLP receptors that are members of the seven transmembrane-spanning, G protein-coupled superfamily and receptor occupancy leads to cytoskeletal rearrangements, exocytosis and the respiratory burst. In addition to enhanced tyrosine phosphorylation fMLP induces a number of signals, including G-protein-dependent activation of phospholipase C (Cockroft and Stutchfield, 1989), leading to IP₃, diacylglycerol generation, protein kinase C activation (Traynor-Kaplan et al., 1989; Stephens et al., 1991), transient increases in intracellular calcium (Sha’afi and Molski, 1988) and activation of phos-phatidylinositol-3-kinase (Vlahos and Matter, 1992). The functional roles of tyrosine phosphorylated proteins are largely undefined; however, recently it has been shown that extracellularly regulated kinase 1 is phosphorylated on tyro-sine in neutrophils in response to fMLP, raising the possibility of tyrosine and serine/threonine kinase cascades during activation (Grinstein and Furuya, 1992). The relationships between tyrosine kinases and G-protein-mediated events have not been completely elucidated in any system; however, recently the guanine nucleotide exchanger Grb2 and the guanine nucleotide releasing factor Sos have been reported to couple receptor tyrosine kinases to ras signalling (Li et al., 1993; Rozakis-Adcock et al., 1993; Gale et al., 1993). It is possible that a similar mechanism is operative in PMNs, since the ras-related protein rac regulates actin configuration, membrane ruffling and also the activation of the leukocyte NADPH oxidase (Ridley et al., 1992; Abo et al., 1991). Many of the proteins involved in ras activation contain src homology 2 (SH2) domains and the inter-molecular associations of these proteins are regulated by tyrosine phosphorylation, thus pervanadate may activate neutrophils by maintaining tyrosine phosphorylation of SH2 domain-binding proteins, thus stabilising functional com-plexes within the cell.

The tyrosine kinase activated after PMN stimulation has not been identified, and the results presented here demon-strate that inactivation of tyrosine-specific phosphatases is sufficient to induce activation of the motile response in neutrophils. The inhibition of pervanadate-induced stimulation by tyrosine kinase inhibitors indicates that these responses are mediated by constitutive activity of cellular tyrosine kinases. Several members of the src family of nonreceptor tyrosine kinases are present in neutrophils: hck, fyn, lyn and fgr. c-fgr is the most abundant kinase in neutrophils and its expression is restricted to circulating monocytes, granulocytes, some tissue macrophages and natural killer cells (Perlmutter et al., 1988; Biondi et al., 1991; Ley et al., 1989; Inoue et al., 1990). Furthermore, differentiation of HL60 or U937 cells to monocyte/macrophages or granulocytes is accompanied by increases in c-fgr message and protein levels (Katagiri et al., 1991; Willman et al., 1991), thus expression of c-fgr is associated with cellular functions such as chemotaxis, phagocytosis and respiratory burst. The translocation of c-fgr from secondary granules to the plasma membrane during PMN activation has been reported (Gutkind and Robbins, 1989). The association of phospho-tyrosyl proteins with the neutrophil cytoskeleton is reminiscent of the situation in v-src-transformed cells, where the majority of tyrosine-phosphorylated proteins associate with the cytoskeleton (Hamaguchi and Hanafusa, 1989) and specific cytoskeletal proteins, such as vinculin, talin and integrins, are phosphorylated (Kellie et al., 1986; DeClue and Martin, 1987; Horvath et al., 1990), although the functional relevance of this is still questionable (Kellie et al., 1991). Nevertheless, association of the v-src kinase with the cytoskeleton is essential for morphological transforma-tion of fibroblasts (Hamaguchi and Hanafusa, 1987) and by analogy the association of c-src-related kinases with the neutrophil cytoskeleton may play a role in the activation of these cells as has been postulated for platelets (Horvath et al., 1992). Vanadate ions induce actin polymerisation in permeabilised HL60 cell models (Trudel et al., 1990), sug-gesting a role for tyrosine phosphorylation in actin regulation. Pervanadate induces tyrosine phosphorylation in MDCK cells with a time course similar to that found in neutrophils, and also induces changes in cell morphology, including cell rounding and deterioration of F-actin-containing bundles concomitant with accumulation of phos-photyrosine-containing proteins in cytoskeleton-associated adherens junctions, similar to our finding of tyrosine-phos-phorylated proteins co-localising with F-actin in neutrophils (Volberg et al., 1991,1992).

The neutrophil tyrosine phosphatases that are inhibited by pervanadate have not been identified. Neutrophils express CD45, a tyrosine phosphatase that is an important regulator of T cell activation (Shiroo et al., 1992). Several antibodies against CD45 have been shown to inhibit fMLP-induced neutrophil chemokinesis; however, there are no data to show whether the phosphatase activity of the CD45 is altered by antibody binding, and CD45 does not seem to be involved in the processing of spatial information required for chemotaxis (Harvath et al., 1991; Gatewood and Zigmond, 1992). Other phosphatases have also been implicated in neutrophil activation, such as type 1 and type 2 phosphatases (Ding and Badwey, 1992), and calcineurin, a type 2B phosphatase (Hendey et al., 1992).

Pervanadate has been shown to activate signalling events in other cell types. Heffetz et al. (1990, 1992) have shown that pervanadate is an insulinomimetic, activating the insulin receptor kinase. Pervanadate induces secretion, actin polymerisation and shape change in platelets (Pumiglia et al., 1992); will enhance the zymosan-induced respiratory burst in bone marrow macrophages (Green et al., 1992) and will activate peripheral blood T cells (O’Shea et al., 1992). Our results demonstrate additional effects of pervanadate on the neutrophil cytoskeleton, and show that inhibition of cellular tyrosine phosphatases is sufficient to activate the motile responses of neutrophils. This raises the possibility that the increased tyrosine phosphorylation and stimulation of cells by agonists might be mediated by inactivation of phosphatases rather than activation of kinases.

We are grateful to Dr J. Lackie for helpful discussions and Christine Moorhouse for her assistance in the preparation of this manuscript.