Mitogen-activated protein kinase in neutrophils and enucleate neutrophil cytoplasts: evidence for regulation of cell-cell adhesion.

We employed neutrophils and enucleate neutrophil cytoplasts to study the activation of the mitogen-activated protein kinases (MAPKs) p44 and p42 in neutrophils by inflammatory agonists that engage G protein-linked receptors. Formyl-methionyl-leucylphenylalanine (FMLP) rapidly and transiently activated MAPK in neutrophils and cytoplasts, consistent with a role in signaling for neutrophil functions. FMLP stimulated p21activation in neutrophils and Raf-1 translocation from cytosol to plasma membrane in cytoplasts, with kinetics consistent with events upstream of MAPK activation. Insulin, a protein tyrosine kinase receptor (PTKR) agonist, stimulated neutrophil MAPK activation, demonstrating an intact system of PTKR signaling in these post-mitotic cells. FMLP- and insulin-stimulated MAPK activation in cytoplasts were inhibited by BtcAMP, consistent with signaling through Raf-1 and suggesting a mechanism for cAMP inhibition of neutrophil function. However, BtcAMP had no effect on FMLP-stimulated MAPK activation in neutrophils. The extent of MAPK activation by various chemoattractants correlated with their capacity to stimulate neutrophil and cytoplast homotypic aggregation. Consistent with its effects on MAPK, BtcAMP inhibited FMLP-stimulated aggregation in cytoplasts but not neutrophils. Insulin had no independent effect but primed neutrophils for aggregation in response to FMLP. Our studies support a p21-, Raf-1-dependent pathway for MAPK activation in neutrophils and suggest that neutrophil adhesion may be regulated, in part, by MAPK.

Circulating neutrophils are terminally differentiated, postmitotic phagocytes that constitute the first line of host defense against microorganisms. In contrast to dividing cells that respond slowly to mitogens, neutrophils respond rapidly to inflammatory stimuli. One class of neutrophil agonists, the chemoattractants, engage seven transmembrane-spanning domain receptors that activate G i proteins. The only well-documented effector downstream of neutrophil G i is phospholipase C ␤ , a regulatory enzyme not directly linked to the MAPK pathway. Nevertheless, the chemoattractant N-formyl-methionylleucyl-phenylalanine (FMLP) has been shown to activate MAPK (19) and to stimulate MAPK autophosphorylation in neutrophils in a pertussis toxin-sensitive fashion (20). MAPKs thus represent candidate effectors for the signaling pathway(s) leading from G protein activation to rapid neutrophil responses. Because currently understood mechanisms of neutrophil activation fail to explain the observation that agents that elevate intracellular cAMP inhibit some chemoattractant-stimulated neutrophil responses, the possibility that chemoattractants activate MAPKs in a p21 ras /Raf-1-dependent fashion is an attractive hypothesis.
In the present study we utilize intact neutrophils and neutrophil cytoplasts (enucleate, granule-poor, metabolically active cell fragments) to demonstrate that MAPK activation by FMLP is associated with activation of p21 ras and translocation of Raf-1 to the PM and that cAMP acts via PKA to inhibit FMLP-stimulated MAPK activation in cytoplasts but not neutrophils. We also show that insulin, known to activate MAPK via p21 ras and Raf-1 in mitotic cells, activates MAPK in neutrophils and cytoplasts. Finally, we observed a strong correla-tion between MAPK activation and cell-cell adhesion in neutrophils and cytoplasts suggesting a new regulatory role for MAPK in a process critical for inflammation.
Neutrophil Isolation, Cytoplast Preparation, and Subcellular Fractionation-Neutrophils were prepared by the method of Boyum (21). Cytoplasts were prepared by the method of Roos et al. (22).
Gel Renaturation MAPK Activity Assay-Neutrophils (5 ϫ 10 7 /ml) or cytoplasts (5 ϫ 10 8 /ml) were suspended in cell buffer and incubated in the absence or presence of agonists and/or inhibitors. Neutrophil lysates were prepared by the method of Torres et al. (19). Cytoplast incubations were stopped by addition of SDS sample buffer. Neutrophil lysate (10 6 cell eq) or cytoplast lysate (10 7 cell eq) was analyzed using the gel renaturation method of Kameshita and Fujisawa (23) except that prepared gels contained 0.25 mg/ml myelin basic protein (MBP) or bovine serum albumin (BSA), and phosphorylation buffer contained 2.5 Ci/ml [ 32 P]ATP. MAPK activity was analyzed by phosphorimaging and quantitated as MBP phosphorylation of the p44 erk1 /p42 erk2 region of each gel.
MBP Peptide Substrate (MBPp) Kinase Activity Assay-Neutrophils (2 ϫ 10 8 /ml) or cytoplasts (10 9 /ml) were incubated in the absence or presence of Bt 2 cAMP for 10 min at 37°C, followed by incubation in the absence or presence of 100 nM FMLP for 1 min. Reactions were stopped by addition of lysis buffer (20 mM Tris, pH 7.4, 1 mM NaEGTA, 2 mM sodium vanadate, 25 mM sodium fluoride, 0.5% Triton X, 2 mM PMSF, 10 trypsin inhibitor units/ml aprotinin, and 10 g/ml each of chymostatin, antipain, and pepstatin). Neutrophil lysates were centrifuged (14,000 ϫ g for 10 min at 4°C) to remove nuclei. Lysates were kept on ice for 15 min, followed by incubation for 15 min at 37°C in a buffer (25 mM Tris, pH 7.4, 12.5 mM MgCl 2 , 125 mM NaEGTA, 1.25 mM sodium fluoride, 2 mM dithiothreitol, 220 M ATP, and 25 Ci/ml [ 32 P]ATP) containing 500 M MBPp or a control peptide in which valine was substituted for threonine. Reactions were stopped by the addition of 15% formic acid. The lysates were spotted onto phosphocellulose papers that were washed thoroughly with distilled water and quantitated by scintillation counting. Duplicate assays in the absence of MBP peptide were performed to determine non-MAPK background kinase activities.
Membrane Translocation of Raf-1, p44 erk1 /p42 erk2 , and SOS-Assays for FMLP-stimulated translocation of proteins from cytosol (CS) to PM in cytoplasts were performed as described previously (24). Localization of proteins was determined by SDS-polyacrylamide gel electrophoresis and immunoblotting of CS and PM fractions with appropriate antisera.

FMLP Stimulates MAPK Activation in Neutrophils and
Enucleate Neutrophil Cytoplasts-MAPK activity was detected in lysates of intact neutrophils as discrete MBP kinase activities using a gel renaturation assay (Fig. 1A). FMLP-stimulated MBP kinase activity was detected only in the 42-44-kDa region, consistent with the molecular weight of the MAPKs, p44 erk1 and p42 erk2 . Because the presence of nucleic acids and granular proteases in SDS lysates of intact neutrophils im- paired the resolving capacity of MBP-impregnated gels, we also analyzed MAPK activation by this method in enucleate, granule-depleted cytoplasts. In cytoplasts as in neutrophils, FMLP stimulated MBP kinase activity in the 42-44-kDa region of the gel (Fig. 1B). In some but not all gels resolution of cytoplast proteins was sufficient to allow identification of two discrete FMLP-sensitive MBP kinase activities, consistent with the molecular weight of p44 erk1 and p42 erk2 . Duplicate gels in which BSA was substituted for MBP revealed no phosphoproteins, excluding autophosphorylation and confirming MBP kinase activity for each band. Thus, the only MBP kinases stimulated by FMLP had molecular weights consistent with the MAPKs p44 erk1 and p42 erk2 .
To establish that the FMLP-sensitive MBP kinase activity observed in the gel renaturation assay corresponded to MAPKtype phosphorylation of the MBP molecule, which has numerous non-MAPK phosphorylation sites, we tested the ability of lysates of FMLP-stimulated cells or cytoplasts to phosphorylate MBPp, a synthetic peptide containing only the MBP amino acid sequence specifically phosphorylated on threonine by Erk (PRTP) (26). Neutrophil and cytoplast lysates both contained MBPp kinase activity that was markedly stimulated by FMLP ( Fig. 1, C and D). A peptide in which valine was substituted for threonine gave only background counts. The fold-increase stimulated by FMLP was greater in neutrophils than cytoplasts, consistent with the results obtained in the gel renaturation assay.
We confirmed the identity of cytoplast FMLP-sensitive MBP kinases as p44 erk1 and p42 erk2 by immunoprecipitating p44 erk1 and p42 erk2 from unstimulated or FMLP-stimulated cytoplasts and analyzing the precipitates by immunoblot and gel renaturation MBP kinase assays. Coimmunoprecipitation of p44 erk1 and p42 erk2 followed by immunoblot using a third antiserum recognizing both Erks revealed two polypeptides of expected molecular weight that were unaffected by FMLP stimulation ( Fig. 2A). Neither protein was precipitated by a control antiserum. In contrast, p44 erk1 /p42 erk2 antisera precipitated MBP kinase activity in the 42-44-kDa region of the gel only from lysates of cytoplasts that had been stimulated with FMLP (Fig.  2B). The resolving power of the MBP-impregnated gels in these experiments was inadequate to distinguish p44 erk1 from p42 erk2 kinase activity. However, when p44 erk1 and p42 erk2 were immunoprecipitated separately from FMLP-stimulated lysates each precipitate contained 42-44-kDa MBP kinase activity, although somewhat less than the coprecipitate (Fig. 2B). FMLP-stimulated lysates immunoprecipitated with control antisera contained no 42-44-kDa MBP kinase activity, although higher molecular weight activities were pulled down nonspecifically. Thus, the measurement of radioactivity in immunoprecipitates in the absence of simultaneous assessment of the molecular weight of the kinase is inadequate for measuring MAPK activity specifically. Accordingly, these data are the clearest demonstration to date that FMLP, acting through a G protein-linked receptor, activates p44 erk1 and p42 erk2 in neutrophils. Moreover, they demonstrate that the signaling pathway for formyl peptide receptor-stimulated MAPK activation does not depend on nuclear or granular elements since it is retained in neutrophil cytoplasts.
Because neutrophils required post-stimulation processing that made precise kinetic measurements difficult, we studied the kinetics of FMLP-stimulated MAPK activation in cytoplasts that could be rapidly lysed in SDS sample buffer without releasing nucleic acids and granular proteases (Figs. 3A and 4). FMLP-stimulated activation of p44 erk1 and p42 erk2 was transient, peaking at 1 to 2 min and returning to base line by 10 min. Because MAPK activity is associated with tyrosine phosphorylation of p44 erk1 and p42 erk2 , we compared MBP kinase activities with tyrosine phosphorylation of cytoplast proteins following stimulation with FMLP (Fig. 3B). Cytoplast lysates contained a prominent 42-kDa tyrosine phosphoprotein whose phosphorylation was stimulated by FMLP with kinetics identical to those of the p42 erk2 kinase activity. When the blot was stripped and reprobed with an antiserum directed to p44 erk1 / p42 erk2 (Fig. 3C), the 42-kDa phosphoprotein aligned precisely with p42 erk2 . The amount of p44 erk1 and p42 erk2 detected by immunoblot did not change with FMLP stimulation, confirming that the antiphosphotyrosine blot detected changes in phosphotyrosine content, not amount of protein. Since the anti-Erk antiserum used has greater affinity for p44 erk1 than p42 erk2 , the immunoblot results revealing a darker p42 erk2 band suggest that more p42 erk2 than p44 erk1 is expressed in human neutrophils. This could explain why the anti-phosphotyrosine antibody was apparently only sensitive enough to detect phosphorylated p42 erk2 , a result consistent with earlier studies (27). These data confirm that enucleate neutrophil cytoplasts retain the signaling molecules necessary to respond to FMLP stimulation by phosphorylating and then dephosphorylating Erk on tyrosine and transiently activating MAPK activity.
Pathways of MAPK Activation in Neutrophils and Cytoplasts-G protein-stimulated MAPK activation has been Cytoplasts stimulated with 100 nM FMLP at 37°C for the times indicated were analyzed for MAP kinase activity by gel renaturation assay (A) and immunoblotting for phosphotyrosine-containing proteins (B) as described under "Experimental Procedures." C, the nitrocellulose from B was stripped (19) and reprobed with an anti-p44 erk1 /p42 erk2 antiserum. Results shown are representative of two experiments.
shown to proceed via both p21 ras /Raf-1-dependent and independent pathways. We therefore studied whether MAPK activation by FMLP in neutrophils and cytoplasts was preceded by p21 ras and Raf-1 activation. We first determined by immunoblot analysis that terminally differentiated neutrophils retain all of the elements of the classical p21 ras /Raf-1-dependent MAPK pathway, including Grb2, SOS, Raf-1, MAPK or Erk kinase, and p44 erk1 and p42 erk2 in CS and p21 ras in PM (not shown). To confirm that neutrophils retain the capacity to respond to stimuli that activate p21 ras /Raf-1-dependent MAPK pathways, we examined the effect of PTKR agonists on neutrophils. Epidermal growth factor had no effect on neutrophil or cytoplast MAPK activity, suggesting a lack of expression of EGF receptor on myeloid cells. In contrast, insulin, at concentrations (20 -200 nM) required to maximally activate MAPK in mitotic cells expressing insulin receptors (28), activated MAPK in neutrophils (312 Ϯ 113% control) and cytoplasts (148 Ϯ 21% control) with kinetics distinct from those of FMLP stimulation (Fig. 4). Thus at least one PTKR known to activate MAPK in a p21 ras /Raf-1-dependent fashion is functionally expressed in human neutrophils suggesting that p21 ras /Raf-1 signaling is intact.
To study the rapid FMLP-stimulated p21 ras activation predicted by the kinetics of MAPK activation, we studied FMLPstimulated guanine nucleotide exchange on p21 ras . Because isolated neutrophils had a bench life (Ͻ6 h by lactate dehydrogenase release assay) inadequate to ensure equilibrium labeling of nucleotide pools by metabolic labeling with [ 32 P]orthophosphate, we permeabilized neutrophils by electroporation in the presence of [␣-32 P]GTP to rapidly label intracellular pools of GTP. In this system p21 ras activation is measured as total [␣-32 P]guanine nucleotide loading (guanine nucleotide exchange with or without GTPase activation). Electroporated neutrophils were stimulated with FMLP (100 nM), p21 ras was immunoprecipitated from cell lysates, and the amount of [␣-32 P]guanine nucleotide associated with p21 ras was determined by thin layer chromatography (Table I). The amount of [␣-32 P]guanine nucleotide associated with p21 ras following FMLP stimulation was 164 Ϯ 20% control at 30 s and remained stable for as long as 5 min, suggesting that FMLPstimulated guanine nucleotide exchange on p21 ras in neutrophils peaks no later than 30 s after stimulation. Since hydrolysis of [␣-32 P]GTP on p21 ras results in p21 ras ⅐[␣-32 P]GDP, total [␣-32 P]guanine nucleotide associated with p21 ras cannot distinguish transient from persistent p21 ras activation. However, the percentage of [␣-32 P]guanine nucleotide associated with p21 ras as [␣-32 P]GTP declined after 1 min, suggesting a GTPaseactivating protein activity limiting activation.
Membrane association of Raf-1 appears to be required for its activation by PTKR (7-9). Since Raf-1 is recruited to the PM of cells transfected with oncogenic, activated p21 ras (7,9), we tested whether FMLP, in addition to activating p21 ras , can stimulate Raf-1 translocation from the CS to PM of cytoplasts. When unstimulated cytoplasts were sonicated and separated by centrifugation into soluble and insoluble fractions, 14 Ϯ 3% of total immunodetected Raf-1 (supernatant ϩ pellet) was associated with the pellet. This analysis is likely to overestimate the true membrane-associated pool since cytoplast disruption may not have been complete and vesiculated cytoplast membrane is likely to sequester CS. FMLP stimulated a 2.2-fold increase in the amount of Raf-1 associated with cytoplast membranes. Thus, despite the potential overestimation of basal membrane-associated Raf-1, our system was sensitive enough to detect membrane translocation of this molecule. Kinetic analysis (Fig. 5) revealed that Raf-1 translocation to the membrane peaked at 30 s to 1 min and remained stable for at least 10 min following FMLP stimulation. In contrast to Raf-1, we observed no FMLP-stimulated translocation of p44 erk1 or p42 erk2 to PM in cytoplasts. Although neutrophil CS contained an abundant supply of SOS, this molecule also did not translocate from cytoplast CS to PM in response to FMLP. Thus, both FMLP-stimulated p21 ras activation and Raf-1 translocation preceded MAPK activation, consistent with a role for both events upstream of Erk activation in the FMLP-stimulated pathway.
PTKR activation of MAPK via p21 ras and Raf-1 may be down-regulated by cAMP-dependent, PKA-mediated phosphorylation of Raf-1, resulting in impaired interactions between p21 ras and Raf-1 (12,13). We therefore tested whether FMLP-stimulated MAPK activity in neutrophil cytoplasts can be similarly inhibited by cAMP. The membrane-permeable, phosphodiesterase-resistant cAMP analog dibutyryl cAMP (Bt 2 cAMP) (1 mM) completely inhibited insulin-stimulated cytoplast MAPK activity and inhibited FMLP-stimulated cytoplast MAPK activity by 46.4 Ϯ 11.7% (Fig. 6A). Agents that raise intracellular cAMP by indirect mechanisms, including isobutrylmethylxanthine (50 M), forskolin (50 M), isoproterenol (10 M), and the adenosine A 2 receptor agonist NECA (10 M) also inhibited FMLP-stimulated MAPK activity by approximately 50% (Fig. 6B). To confirm that cAMP inhibition of FMLP-stimulated MAPK in cytoplasts is PKA-dependent, we tested the effect of KT5720 that, at concentrations below 2 M, is a specific inhibitor of PKA (29). KT5720 (1 M) reversed the inhibitory effect of Bt 2 cAMP on FMLP-stimulated MAPK activation (Fig. 6C). These data  To explore this discrepancy we compared the effect of Bt 2 cAMP on FMLP-stimulated MAPK activity in neutrophils and cytoplasts (Fig. 6D). As measured by the MBPp kinase activity assay, 1 mM Bt 2 cAMP inhibited FMLP-stimulated cytoplast MAPK activity by 53 Ϯ 13% but had no effect on FMLP-stimulated MAPK activity in neutrophils. Thus, a regulatory role for cAMP in MAPK activation can be observed in vitro in lysates from cytoplasts but not from intact neutrophils suggesting that a factor(s) derived from the nucleus and/or cytoplasmic granules masks the effect.

Role of MAPK in Neutrophil Homotypic
Aggregation-To determine a potential role for MAPK activation in neutrophil function, we compared MAPK activation with neutrophil adhesiveness as measured by homotypic aggregation in neutrophils and cytoplasts. Although each of the best-studied neutrophil chemoattractants (FMLP, LTB 4 , C5a, and Il-8) acts through a similar seven transmembrane-spanning receptor linked to a similar or identical G protein (likely G i2 ), their efficacies at stimulating neutrophil functions vary dramatically. We therefore compared the ability of each chemoattractant to stimulate MAPK activity with its effect on aggregation. Like FMLP, saturating concentrations of LTB 4 , C5a, and Il-8 stimulated MAPK activation in intact neutrophils by 1 min (Table II); however, LTB 4 , C5a, and Il-8 were significantly less efficacious than FMLP in activating MAPK. The extent of aggregation stimulated by each agonist correlated well (R 2 ϭ 0.92) with the degree of MAPK activation (Fig. 7, A and C). LPA, which acts via a G protein-linked receptor and p21 ras to stimulate MAPK in cells in culture (14), had no effect on neutrophil MAPK activity or homotypic aggregation (not shown). In cytoplasts, whereas FMLP stimulated MAPK activity (318 Ϯ 34% control), the other chemoattractants demonstrated little or no effect on MAPK activity (Table II). As in intact neutrophils, the ability of the chemoattractants to stimulate cytoplast aggregation correlated well (R 2 ϭ 0.99) with their ability to stimulate cytoplast MAPK activity (Fig. 7, B and D). These data confirm divergent signaling from distinct G protein-linked receptors and argue strongly for an association between MAPK activation and adhesion. The ability of Bt 2 cAMP to inhibit FMLP-stimulated MAPK activity in cytoplasts but not neutrophils suggested that similar effects might be observed on homotypic aggregation. Indeed, Bt 2 cAMP (1 mM) had no effect on FMLP-stimulated neutrophil aggregation but inhibited FMLP-stimulated cytoplast aggregation by 33 Ϯ 2% (Fig. 8). This observation further supports a role for MAPK activation in neutrophil adhesion.
In contrast to the chemoattractants, insulin activated neutrophil and cytoplast MAPK but failed to stimulate neutrophil aggregation, suggesting that MAPK activation may be necessary but not sufficient to support aggregation. We therefore tested whether insulin could prime neutrophils for chemoat-  tractant-stimulated functions. Preincubation of neutrophils with insulin had little or no effect on homotypic aggregation stimulated by concentrations of chemoattractants inducing maximal aggregation responses. However, preincubation with insulin for 10 min primed neutrophils for aggregation in response to concentrations of FMLP and LTB 4 that induced submaximal aggregation responses (Fig. 9). Shorter incubations (i.e. times at which insulin failed to stimulate MAPK activity) had no effect on aggregation. The priming effect of insulin on FMLP-and LTB 4 -stimulated aggregation was dose-dependent, peaking at 200 nM. A trend toward insulin priming of neutrophils for C5a-stimulated aggregation was observed but did not achieve statistical significance. Insulin had no effect on Il-8stimulated aggregation. Thus, the effect of 200 nM insulin on neutrophil homotypic aggregation by submaximal concentrations of chemoattractants was proportional to the ability of these chemoattractants to stimulate MAPK activity in neutrophils (FMLPϾLTB 4 ϾC5aϾIl-8). Insulin potentiation of FMLPand LTB 4 -stimulated homotypic aggregation was glucose-independent. In contrast, insulin had no direct effect on FMLPinduced O 2 . generation and ␤-glucuronidase release and potentiated these responses only in the presence of glucose, presumably by increasing glucose transport and affecting metabolism.

DISCUSSION
Although well established, the link between chemoattractant-stimulated G protein signaling pathways and the MAPK cascade is poorly elucidated. Neutrophils are a good system in which to study G protein-mediated signaling because the cellular responses are rapid and easily quantitated. Enucleate, granule-depleted neutrophil cytoplasts retain the capacity to respond to chemoattractants (30) and thus represent a simplified system useful in studying chemoattractant signaling through G proteins. We employed neutrophils and cytoplasts to study the kinetics of chemoattractant-stimulated activation of p21 ras , Raf-1, and MAPK and observed an association between MAPK activation and cell-cell adhesion.
The analysis of MAPK activity in cytoplast lysates by an MBP kinase gel renaturation assay offered distinct advantages over similar studies of lysates of intact neutrophils, including resolution of two MAPKs in cytoplast lysates, identified by immunoprecipitation as p44 erk1 and p42 erk2 . Moreover, the ability to terminate stimulation by direct addition of SDS sample buffer permitted more accurate kinetic analysis of cytoplasts than of neutrophils. FMLP-stimulated cytoplast MAPK activation was rapid and transient, consistent with a role for MAPK in signaling pathways for neutrophil functions such as O 2 . generation, degranulation, and cell-cell adhesion but slower than the previously reported kinetics of neutrophil MAPK activation (19,27,31). The greater precision afforded by kinetic analysis of MAPK in cytoplasts thus allowed comparison with the kinetics of activation of other putative elements in the FMLP-stimulated MAPK cascade, such as p21 ras and Raf-1.
The use of cytoplasts also permitted observation of MAPK signaling in the absence of nuclear or granular elements. Thus, phosphorylation and dephosphorylation on tyrosine residues of p42 erk2 in cytoplasts, with kinetics paralleling those of MAPK activation, indicate that the molecular machinery required for regulating MAPK activity by Erk kinase and phosphatase activities is retained in cytoplasts and so independent of any nuclear factors that may regulate MAPK. This observation may distinguish neutrophils from proliferating cells, in which activated MAPK translocates to the nucleus where it is downregulated by dual phosphothreonine/phosphotyrosine phosphatases such as PAC-1 (32).
Although a wide variety of G protein-linked and non-G protein-linked receptors have been demonstrated on neutrophils, none of the classical PTKRs have been reported. Insulin, however, has been shown to bind to human neutrophils (33), stimulate chemokinesis (34), and prime for chemotaxis to FMLP (35), indicating that PTKRs for insulin are expressed on these cells. Our observation that insulin activated MAPK in neutrophils and cytoplasts suggests that a p21 ras /Raf-1 pathway is functionally intact and can be engaged by at least one PTKR. Thus, the neutrophil formyl peptide receptor may also activate MAPK through the p21 ras /Raf-1 pathway. However, the longer latency for insulin-than for FMLP-stimulated MAPK activation indicates that the pathways to p21 ras activation may be distinct.
Our observation that FMLP activated p21 ras in neutrophils supports p21 ras /Raf-1 signaling. Although a previous study came to the same conclusion using neutrophils metabolically labeled with [ 32 P]orthophosphate (36), we found the bench life of neutrophils insufficient to label nucleotide pools to equilibrium, an absolute requirement for interpreting GTP/GDP ratios of GTPase-bound nucleotide as an indicator of p21 ras activation. We therefore analyzed total labeled guanine nucleotide associated with immunoprecipitated p21 ras from lysates of cells electroporated in the presence of [␣-32 P]GTP and found maximal increase after 30 s of exposure to FMLP (i.e. preceding peak MAPK activity). Concordant with a prior report (36), the proportion of [␣-32 P]GTP associated with p21 ras declined by 5 min, suggesting sequential guanine nucleotide exchange factor and GTPase-activating protein activities following FMLP stimulation.
Raf-1 has been shown to translocate from CS to PM in cultured cells exposed to serum (7). However, Raf-1 translocation in response to ligation of neither a specific PTKR nor a G protein-linked receptor has been demonstrated in any cell type. We have shown that cytoplasts are an ideal system with which to assay PM translocation of cytosolic proteins (24). Using this system, we now report FMLP-stimulated translocation of Raf-1 to the PM. These data complement those of Worthen et al. (36) who reported FMLP-stimulated Raf-1 kinase activity. Like p21 ras activation, FMLP-stimulated Raf-1 translocation preceded MAPK activation. Our failure to observe SOS translocation in response to FMLP suggests that SOS may not participate in G protein activation of p21 ras . Alternatively, translocation of SOS may not be necessary for its activity, or the kinetics of SOS translocation may be too rapid to have been appreciated in our assay.
Our observation that MAPK activity in cytoplasts was inhibited by agents that raise intracellular cAMP and that a PKA antagonist reversed this inhibition is also consistent with p21 ras /Raf-1-dependent signaling since cAMP has been shown to down-regulate MAPK activation by PKA-dependent phosphorylation of Raf-1, inhibiting p21 ras /Raf-1 interactions (12,13). Bt 2 cAMP has been shown to inhibit neutrophil Raf-1 kinase activity (36), supporting an effect of cAMP in neutrophils at the level of Raf-1. However, Bt 2 cAMP inhibition of FMLPstimulated MAPK activation has not previously been demonstrated. Indeed, Yu et al. (37) reported that Bt 2 cAMP does not inhibit FMLP-stimulated MAPK in cytochalasin B-treated neutrophils. Our data confirm this observation in intact neutro-phils but show that cytoplasts express a cAMP-sensitive pathway. The exposure of a cAMP-sensitive pathway in cytoplasts may be explained by increased phosphodiesterase or Raf-1 phosphatase activities in detergent lysates of granule-replete, nucleated neutrophils. Alternatively, neutrophils may possess both cAMP-sensitive and -insensitive G protein-linked pathways of MAPK activation, the latter preferentially inactivated during cytoplast preparation. Indeed, Faure and Bourne (38) have recently shown that cell lines in which stimulation of MAPK activity by LPA is cAMP-insensitive nevertheless demonstrate cAMP inhibition of Raf, suggesting a Raf-independent pathway of MAPK activation.
Several groups have proposed a role for MAPK in neutrophil O 2 . generation (39 (37,40) and its failure to inhibit MAPK activation in FMLP-stimulated neutrophils (37), in our studies Bt 2 cAMP inhibited neither FMLP-stimulated MAPK activation nor FMLP-stimulated homotypic aggregation in neutrophils but significantly inhibited both of these responses in cytoplasts. Thus the effect of Bt 2 cAMP on FMLP-stimulated MAPK activity in both neutrophils and cytoplasts correlated with its effect on cell-cell adhesion. Our discovery that insulin both activated MAPK in human neutrophils and primed these cells for homotypic aggregation in response to chemoattractants demonstrates a further correlation between MAPK activation and cell-cell adhesion. The inability of insulin to directly stimulate aggregation suggests that MAPK may be necessary but not sufficient to directly or indirectly regulate adhesion molecules. The failure of insulin to stimulate or prime neutrophils for O 2 . generation or degranulation in the absence of extracellular glucose supports the hypothesis that MAPK regulates some but not all neutrophil functions. Thus, although O 2 . generation can be dissociated from MAPK activation, our studies of neutrophils and cytoplasts support a role for MAPK activation in cell-cell adhesion. Neutrophil homotypic aggregation is mediated by activation of the ␤ 2 integrin CD11b/CD18 (25). The activation states of integrins appear to be regulated by interactions of the cytoplasmic domains of these heterodimeric transmembrane glycoproteins with the actin cytoskeleton through focal adhesion plaques (41). Thus, MAPK might regulate cell-cell adhesion through phosphorylation of molecules regulating focal adhesion plaques. In addition to a hypothetical role in regulating the actin cytoskeleton, MAPKs have a well-established role in regulating the microtubule cytoskeleton by associating with and phosphorylating microtubule-associated proteins (42). The relationship between MAPK activation and leukocyte adhesion suggests new targets for anti-inflammatory drugs since leukocyte adhesion to vascular endothelium is the first committed step in the inflammatory response. Furthermore, insulin stimulation of neutrophil MAPK and priming for chemoattractantstimulated adhesion suggest a molecular mechanism for impaired neutrophil function in type I diabetes, a state of insulin deficiency associated with increased susceptibility to bacterial infection.
Since neutrophils are terminally differentiated, non-mitotic cells, the effects of MAPK on transcription factors related to growth and differentiation are unlikely to be relevant. Our studies with enucleate cytoplasts support this view. Marshall (43) has recently proposed that the outcome of MAPK signaling is dependent largely on its duration of activation. If so, rapid MAPK activation in neutrophils may represent a distinct category of signaling. Alternatively, differentiated cells might also be distinguished by their complement of MAPK substrates. The only well-defined MAPK substrate also implicated in neutrophil activation is cytoplasmic phospholipase A 2 (44). Further studies are likely to identify other MAPK substrates involved in rapid neutrophil responses.