A Role for Protein Phosphatase-2A in p38 Mitogen-activated Protein Kinase-mediated Regulation of the c-Jun NH2-terminal Kinase Pathway in Human Neutrophils*

Human neutrophil accumulation in inflammatory foci is essential for the effective control of microbial infections. Although exposure of neutrophils to cytokines such as tumor necrosis factor-α (TNFα), generated at sites of inflammation, leads to activation of MAPK pathways, mechanisms responsible for the fine regulation of specific MAPK modules remain unknown. We have previously demonstrated activation of a TNFα-mediated JNK pathway module, leading to apoptosis in adherent human neutrophils (Avdi, N. J., Nick, J. A., Whitlock, B. B., Billstrom, M. A., Henson, P. M., Johnson, G. L., and Worthen, G. S. (2001) J. Biol. Chem. 276, 2189–2199). Herein, evidence is presented linking regulation of the JNK pathway to p38 MAPK and the Ser/Thr protein phosphatase-2A (PP2A). Inhibition of p38 MAPK by SB 203580 and M 39 resulted in significant augmentation of TNFα-induced JNK and MKK4 (but not MKK7 or MEKK1) activation, whereas prior exposure to a p38-activating agent (platelet-activating factor) diminished the TNFα-induced JNK response. TNFα-induced apoptosis was also greatly enhanced upon p38 inhibition. Studies with a reconstituted cell-free system indicated the absence of a direct inhibitory effect of p38 MAPK on the JNK module. Neutrophil exposure to the Ser/Thr phosphatase inhibitors okadaic acid and calyculin A induced JNK activation. Increased phosphatase activity following TNFα stimulation was shown to be PP2A-associated and p38-dependent. Furthermore, PP2A-induced dephosphorylation of MKK4 resulted in its inactivation. Thus, in neutrophils, p38 MAPK, through a PP2A-mediated mechanism, regulates the JNK pathway, thus determining the extent and nature of subsequent responses such as apoptosis.

The responses of eukaryotic cells to external stimuli are regulated in part by the activation of three major MAPK 1 signaling pathways, p42/44 ERK, p38 MAPK, and JNK, which mediate many of the cellular processes associated with growth, survival, and death. Activation of the JNK pathway occurs in response to environmental and physiological stresses, including genotoxins, protein synthesis inhibitors, oxidative agents, and pro-inflammatory cytokines such as TNF␣ (1)(2)(3)(4), and has been linked to many important cellular functions and processes, including cell survival, death, gene expression, and embryogenesis. Our previous studies on adherent human neutrophils have focused on characterization of the TNF␣-induced activation of JNK, defining both upstream signaling components and a downstream functional response of this pathway, apoptosis (26). However, mechanisms that regulate the JNK pathway remain poorly defined.
In the yeast Saccharomyces cerevisiae, MAPK pathways were initially conceptualized as linear hierarchies of signaling molecules that regulate the expression of specific phenotypic responses (5,6). Recent evidence suggests, however, that the yeast homolog of mammalian p38 MAPK, Hog1, limits the activation of other pathways, thus altering the repertoire of MAPKs induced by osmotic stress (7).
In a variety of eukaryotic cells, there is mounting evidence that the p38 MAPK pathway negatively regulates both the JNK and ERK signaling pathways in response to diverse stimuli. In papilloma-producing mouse keratinocytes, inhibition of p38 MAPK augments both native and okadaic acid-induced JNK and ERK activation, increases AP-1 binding to the 12-Otetradecanoylphorbol-13-acetate response element, and enhances synthesis of JunD and FosB (8). Similarly, JNK activity in osmotically stressed Madin-Darby canine kidney cells (9) and in 1,25-dihydroxyvitamin-differentiated HL-60 cells (10) is greatly increased in the presence of p38 MAPK inhibitors. Our group and others (11) define cross-talk or cross-signaling as the regulation of one signaling pathway by another. We hypothesized that p38 MAPK may regulate the TNF␣-induced JNK pathway in human neutrophils via such a mechanism.
However, regulation of MAPK pathways also involves the dynamic interplay between kinases and phosphatases. Because kinase-mediated protein phosphorylation is utilized in every aspect of cellular function, the requirement that the extent and amplitude of this action be tightly controlled is essential. Recent studies have demonstrated the importance of phosphatase participation in such essential regulatory processes (12). Activation of ERK, JNK, and p38 MAPKs involves a reversible/dual phosphorylation of Tyr and Thr residues, which are located in specific motifs, TEY, TPY, and TGY, respectively (13), whereas both the MAPK kinases and MAPK kinase kinases are phosphorylated at Ser/Thr residues (14). Conversely, inactivation is facilitated by the MAPK phosphatases, which include the single-specificity tyrosine phosphatases; the dual-specificity phos-phatases, which dephosphorylate both Tyr and Ser/Thr residues; and the serine/threonine phosphatases (12,(15)(16)(17), which comprise four major classes: PP1, PP2A (including PP4, PP5, and PP7), PP2B (calcineurin), and PP2C.
The highly conserved and extremely abundant PP2A family (18) contributes extensively to the serine/threonine phosphatase activity of cells (19). The PP2A holoenzyme comprises the core enzyme (PP2A D ), consisting of a 36-kDa catalytic subunit (PP2A c ) strongly bound to the 65-kDa regulatory scaffold-like subunit A (PR65), while a third B regulatory subunit associates with the core. Knockout studies in yeast (20) and mice (21) have reported embryonic lethality for genes encoding the catalytic subunit of PP2A, demonstrating an absolute requirement for this enzyme (which is involved in almost every cellular process) (22).
Herein, we present studies detailing a novel mechanism in which p38 MAPK regulation of the JNK pathway in human neutrophils is mediated via PP2A. These results demonstrate that p38 MAPK activation has a profound inhibitory effect on the JNK pathway. We have defined the effect of p38 both on each member of the putative JNK module (JNK, MKK4/7, and MEKK1) and also on apoptosis, an important functional consequence of JNK pathway activation. Furthermore, our data support a role for PP2A downstream of p38 MAPK in the dephosphorylation of MKK4 and attenuation of JNK activation. Together, these findings advance our understanding of the regulation of the neutrophil response to pro-inflammatory stimuli and the potential effects of inhibiting the p38 MAPK pathway. This study has important implications for the clinical setting where the use of p38 inhibitors has already been proposed as an adjunct for treatment of patients in the resolution of certain inflammatory responses.
Recombinant Proteins-Active MEKK, unactive and active GST-MKK4/SEK1, active p38, and wild-type JNK were from Upstate Biotechnology, Inc. c-Jun, a gift from Dr. Gary L. Johnson, was expressed in Escherichia coli and purified as previously described (25).
JNK, MKK4/7, and MEKK Immunoprecipitation and in Vitro Kinase Assays-Neutrophils were preincubated as described previously (26) to promote adherence-like conditions. Stimulation with TNF␣, lysis of cells, immunoprecipitation, and assays were also carried out as previously described (26).
PP2A Dephosphorylation of MKK4 -Purified PP2A was protein A-Sepharose-immunoprecipitated with anti-PP2A antibody in JNK lysis buffer (26) for 2 h. The beads were washed once with JNK lysis buffer and twice with dephosphorylation assay buffer containing 0.1 mM EDTA, 1 mg/ml bovine serum albumin, 20 mM imidazole HCl (pH 7.63), and 0.1% ␤-mercaptoethanol. PP2A-bound beads were incubated with active MKK4 (150 ng) in 30 l of dephosphorylation mixture (20 l of dephosphorylation assay buffer and 10 l of solution buffer containing 50 mM Tris-HCl (pH 7.0), 0.1 mM EDTA, 15 mM caffeine, and 1% ␤-mercaptoethanol) for 1 h at 30°C. The supernatant (25 l) containing MKK4 was transferred to a clean tube containing 35 l of a kinase mixture (500 ng of c-Jun, 1 g of wild-type JNK, and 15 Ci of [ 32 P]ATP in cell-free assay buffer) and incubated for 25 min at 30°C. Reactions were terminated with 5ϫ Laemmli sample buffer at 100°C; proteins were separated by SDS-PAGE and transferred to nitrocellulose; and phosphorylation changes were visualized by autoradiography and phosphorimaging.
Phosphatase Assays-Two different phosphatase assays were employed. The first involved release of 32 P i from 32 P-labeled phosphorylase a, the preferred substrate for PP1 and PP2A. Neutrophils were stimulated with TNF␣ as described above, and cell pellets were lysed with phosphatase lysis buffer containing 50 mM Tris (pH 7.14), 10% glycerol, 0.1% IGEPAL, 0.1 mM EGTA, 0.1% ␤-mercaptoethanol, 5 g/ml leupeptin, and 5 g/ml aprotinin. Lysates were passed five times through a 20-gauge needle and subjected to swift freeze-thawing, and phosphatase activity was determined according to the manufacturer's instructions (Protein Phosphatase Assay Systems, Invitrogen). The second method involved phosphate release from a phosphopeptide (250 M), which was quantified using a malachite green reagent (Upstate Biotechnology, Inc., Ser/Thr Phosphatase Assay Kit I) according to the manufacturer's instructions. PP2A from neutrophils, lysed in JNK lysis buffer (26), was immunoprecipitated, and PP2A-bound beads were washed with JNK lysis buffer and then with Ser/Thr assay buffer (50 mM Tris-HCl, pH 7.0, 100 nM CaCl 2 ). 250 M phosphopeptide in Ser/Thr assay buffer was added for 5 min at 30°C. Microcentrifuge tubes were centrifuged at 5000 rpm for 15 s at room temperature; 25 l was transferred to an assay plate; and 100 l of malachite green reagent was added according to the manufacturer's instructions for 15 min 30°C. Changes in absorbance were measured at 630 nm. Phosphatase activity was determined as percent unstimulated control.
Apoptosis Assessment-Neutrophils were incubated as described previously (26) and then stimulated with TNF␣ for 1.0, 1.5, 2.0, and 2.5 h (all tubes were incubated for the same time period so that all tube reactions were terminated together). Neutrophils were cytocentrifuged onto glass slides and stained with a Hema 3 staining kit (Fisher), and apoptosis was assessed from changes in cell morphology, including nuclear and cytoplasmic condensation (27).

Inhibition of p38 MAPK Enhances TNF␣-stimulated JNK
Activity-Because it is particularly difficult in the human neutrophil to gain access to interacting points between kinase pathways, as these primary cells cannot be transfected, the delineation of such events relies at present on the use of specific pharmacological inhibitors (32,42). Small cell-permeant inhibitors such as SB 203580, which readily enters cells, provide certain advantages because they may be utilized to determine not only the roles and functional consequences of endogenous kinases, but also the effects of these kinases on other pathways. Previously, we have shown that TNF␣ induces JNK activation in human neutrophils in an adherence-dependent manner (26).
To study the effect of p38 MAPK on TNF␣-induced JNK signaling, human neutrophils were pretreated with SB 203580, a pyridinylimidazole that inhibits both p38␣ n and p38␤ 2 MAPK (28,29), and TNF␣-induced JNK activation was determined from in vitro phosphorylation of the transcription factor c-Jun. JNK was activated in response to TNF␣, as we have previously reported (26); and this activity was significantly augmented in TNF␣-stimulated neutrophils pretreated with SB 203580 (Fig.  1A). Increased JNK activity in unstimulated neutrophils following SB 203580 pretreatment was also observed, but to a lesser extent than in TNF␣-stimulated neutrophils. Western blots probed with a phospho-specific antibody directed against the active form of JNK demonstrated increased JNK phosphorylation in unstimulated and TNF␣-stimulated neutrophils following inhibition of p38 MAPK (Fig. 1B), which was consistent with the observed increase in JNK activity (Fig. 1A). Reprobing these same immunoblots with an antibody directed against JNK demonstrated that equal amounts of JNK protein had been immunoprecipitated under all conditions (Fig. 1B), ruling out the possibility that the increase in JNK activity was due to the presence of greater amounts of JNK protein in the p38 inhibitor-pretreated samples.
To validate the in vivo specificity of the enhanced TNF␣induced JNK response induced by SB 203580 and because of the reported effects of SB 203580 on other kinases at higher concentrations (24, 30 -32), a structurally distinct and more highly specific p38 inhibitor, M 39, was used in parallel. Pretreatment of neutrophils with M 39 similarly increased TNF␣induced JNK phosphorylation compared with non-M 39-pretreated cells (Fig. 1B). In contrast, TNF␣-induced JNK phosphorylation in neutrophils that had been pretreated with SB 202474, a non-active analog of SB 203580, was not augmented.
UV Irradiation-and Anisomycin-induced Activation of JNK Is Augmented by Inhibition of p38 MAPK-JNK activation can be induced by DNA-damaging agents such as UV irradiation and by protein synthesis inhibitors such as anisomycin, in addition to the pro-inflammatory cytokine TNF␣. Therefore, we questioned whether p38 inhibition would similarly augment JNK pathway activation induced by UV irradiation and anisomycin. Pretreatment with SB 203580 and M 39 resulted in augmentation of UV-induced JNK activation (Fig. 2, A and B). Although anisomycin only weakly induces JNK activation in neutrophils, pretreatment with M 39 also enhanced the anisomycin-induced JNK activation (Fig. 2B). Thus, increased JNK activation following p38 inhibitor pretreatment occurs when the JNK pathway is activated by three mechanistically different neutrophil stimuli.
Activation of p38 MAPK Inhibits JNK-The augmentation of TNF␣-stimulated JNK activation observed following p38 MAPK inhibition suggests that either p38 or a downstream effector of the p38 pathway may regulate JNK activity. We hypothesized that if p38 MAPK inhibition resulted in an enhanced response to TNF␣, then pre-activation of p38 MAPK in the neutrophil should cause a concomitant decreased JNK response to TNF␣. To further investigate this, TNF␣-induced JNK activation was studied in neutrophils in which p38 had been activated by prior exposure to PAF, a lipid mediator that we have previously shown to induce rapid activation of p38 MAPK in neutrophils (33). Under the given conditions, neither JNK (Fig. 3B) nor ERK (33) was activated by PAF. PAF pretreatment for 5 min significantly decreased the TNF␣-induced JNK response by 47% in neutrophils (Fig. 3A). This reduction could be reversed upon exposure of neutrophils to a p38 inhibitor prior to PAF stimulation. To verify that p38 was indeed active at this time, we exposed neutrophils to PAF over a period of 5 min and determined p38 activity in immunoprecipitates using an in vitro kinase assay. Additionally, immunoblots from PAF-stimulated lysates were probed with an antibody directed against phosphorylated p38 MAPKs. Both of these methods of assessment indicated that p38, but not JNK, was rapidly activated by PAF and that this activation was sustained over a 5-min period (Fig. 3B). p38 immunoblots demonstrated that equal amounts of p38 were present under all conditions, thus normalizing the p38 activity.
Inhibition of p38 MAPK Enhances TNF␣-stimulated Activation of MKK4, but Not MKK7-To determine whether interac- FIG. 2. p38 inhibition augments UV irradiation-and anisomycin-induced JNK activation. A, SB 203580 (ϩ)-or Me 2 SO (control (Ϫ))-treated neutrophils were treated with buffer (B) or 10 ng/ml TNF␣ (T) or irradiated with ultraviolet light (U; 254 nm) from a UV transilluminator for 10 min and then maintained for an additional 20 min prior to lysis. JNK1 immunoprecipitates were exposed to in vitro kinase assay using c-Jun as substrate, and JNK activity was quantified as described previously (26). B, M 39 pretreatment enhanced both UV-and anisomycin-stimulated JNK activation. Neutrophils were exposed to 1 M M 39 (ϩ) or Me 2 SO (control (Ϫ)) for 1 h at 37°C prior to stimulation in the presence of buffer (B), 10 ng/ml TNF␣ (T), or 50 ng/ml anisomycin (A) for 15 min or irradiated with ultraviolet light (U) for 10 min and then allowed to rest for an additional 20 min as described for A prior to lysis. JNK1 immunoprecipitates were subjected to in vitro kinase assay, SDS-PAGE, and transfer as described in the legend to Fig. 1. c-Jun phosphorylation was determined by PhosphorImager analysis.
tion between the p38 MAPK and JNK pathways occurs at the level of JNK itself or an upstream JNK pathway member, we studied the effects of p38 MAPK inhibition on the JNK-activating kinases MKK4 and MKK7. We have shown previously that both MKK4 and MKK7 are activated in TNF␣-stimulated neutrophils (26). Neutrophils were pretreated in the absence and presence of both SB 203580 and M 39, and MKK4 activity was determined in MKK4 immunoprecipitates from TNF␣-stimu-lated neutrophils. MKK4 activity was greatly augmented under conditions in which p38 activity was inhibited, both for SB 203580 (Fig. 4A) and for M 39 (data not shown). SB 202474, an inactive analog of SB 203580, failed to induce any augmentation of TNF␣-induced MKK4 pathway activity (Fig. 4A). Western blots of these samples demonstrated that equal amounts of MKK4 protein had been immunoprecipitated under all conditions (data not shown). In contrast, p38 MAPK inhibition did not result in a similar augmentation of TNF␣-induced MKK7 activity (Fig. 4B). Analysis of these samples by Western blotting again confirmed that equal amounts of MKK7 protein had been immunoprecipitated (data not shown).
p38 Inhibition Does Not Enhance MEKK Activity-Although a variety of kinases appear to lie upstream of MKK4/7, we have previously shown that MEKK1 is activated in TNF␣-stimulated human neutrophils (26), and others have shown that MEKK1 is involved in the JNK pathway (34 -39). To determine whether inhibition of p38 MAPK acts upstream of MKK4 on MEKK1, neutrophils were pretreated in the absence and presence of SB 203580, and MEKK1 activity was determined in vitro with a coupled kinase assay. Pretreatment of neutrophils with SB 203580 did not enhance the TNF␣-induced MEKK1 activity compared with non-inhibitor-pretreated cells (Fig. 4C). This observation could be interpreted to indicate that the p38 pathway may act at the level of MKK4, but does not preclude the possibility that other MEKKs may be upstream of MKK4.
p38 Does Not Directly Inhibit MKK4 -A cell-free system composed of recombinant components, including p38 MAPK and members of the JNK pathway module (MEKK1/MKK4/ JNK/c-Jun), was developed to determine whether p38 MAPK modulates the MEKK1-mediated activation of MKK4. In vitro phosphorylation of recombinant proteins in the presence of 32 P-labeled ATP was measured. Active recombinant MEKK1 greatly enhanced activation of a JNK module consisting of

p38 MAPK Regulates JNK via PP2A
MKK4/JNK and c-Jun; and in its absence, MKK4 had only moderate ability to activate JNK and c-Jun (Fig. 5), thus confirming previous studies showing that in vitro reconstitution behaves robustly, requires all components and results in amplification of the preceding signal (67). In the presence of active p38 MAPK, these characteristics did not appear to change. We were unable to detect any inhibitory effects of p38 MAPK activation on the fully reconstituted system or on any of the downstream components. The inclusion of active p38 MAPK neither inhibited MEKK1-induced activation of MKK4 and the sequential substrates JNK and c-Jun nor reduced the phosphorylation of MKK4 once activated. In fact, in the presence of active p38 MAPK, a slight increase in MKK4 phosphorylation was detected, as reported by others (40).
Inhibitors of the Ser/Thr Phosphatases PP1 and PP2A Activate JNK-Recent evidence for the involvement of the p38regulated protein phosphatase PP2A in MEK1 and MEK2 activity (41) raised the possibility that p38 MAPK may regulate TNF␣-induced JNK pathway activation via its action on a phosphatase. Therefore, we investigated the effects of the PP1/ PP2A inhibitors calyculin A and okadaic acid on TNF␣-induced JNK activation. Pretreatment of neutrophils with calyculin A induced activation of JNK in the absence of TNF␣ (Fig. 6A), but did not augment the TNF␣-induced response. However, if neutrophils were pretreated with M 39 prior to calyculin A exposure, a further augmentation in the JNK response was observed both in the presence and absence of TNF␣. Pretreatment of neutrophils with okadaic acid resulted in enhanced JNK phosphorylation activation both in unstimulated and TNF␣-stimulated neutrophils (Fig. 6B), although the okadaic acid effect was more obvious in the unstimulated cells.
TNF␣ Induces Ser/Thr Phosphatase Activity in Neutrophils-Upon establishing that inhibition of the PP1/PP2A phosphatases induced JNK activation and that this activation could be further augmented by prior treatment with a p38 inhibitor, we proceeded to study phosphatase activity in TNF␣stimulated neutrophils. Neutrophil whole cell lysates demonstrated increased phosphatase activity at 2 min following TNF␣ stimulation (Fig. 7A). Phosphatase activity appeared to peak at 5 min following TNF␣ exposure and was still present at 10 min.
The actions of PP1 and PP2A may be differentiated by the use of okadaic acid and calyculin A because the IC 50 of PP2A activity lies between 0.05 and 0.5 nM for okadaic acid and 0.3 and 7 nM for calyculin A, whereas 100-fold greater concentrations (IC 50 ϭ 10 -15 nM) will inhibit PP1. Thus, the relative inhibition of phosphatase activity with increasing concentrations of these phosphatase inhibitors could be used to infer which phosphatase species were involved. Unstimulated and TNF␣-stimulated neutrophils were lysed, and phosphatase activity was determined in the absence and presence of increasing concentrations of okadaic acid and calyculin A. At the lowest concentrations of okadaic acid and calyculin A, which primarily affect PP2A, the TNF␣-induced phosphatase activity was decreased to unstimulated levels (Fig. 7, B and C), indicating PP2A activity. However, substantial basal activity remained at this point. Higher concentrations of okadaic acid and calyculin A reduced the basal phosphatase activity in a dose-dependent fashion, pointing to the potential involvement of PP1 in tonic regulation of the unstimulated cells (Fig.  7, B and C).
TNF␣ Induces p38-regulated Phosphatase Activity in PP2A Immunoprecipitates-To further clarify the role of PP2A in TNF␣-stimulated neutrophils, phosphatase activity was determined in PP2A immunoprecipitates. Increased phosphatase activity was observed in PP2A immunoprecipitates from TNF␣-treated neutrophils compared with untreated cells (Fig.  8A). To establish whether p38 played a role in the modulation of this phosphatase activity in PP2A immunoprecipitates, neutrophils were pretreated with SB 203580, and phosphatase activity was determined. Pretreatment with SB 203580 prior to stimulation with TNF␣ reduced PP2A activity to unstimulated levels. Western blot studies of PP2A immunoprecipitates confirmed the presence of PP2A (Fig. 8B).
PP2A Directly Dephosphorylates MKK4 -Having demonstrated a p38-regulated PP2A phosphatase activity in TNF␣stimulated neutrophils, a cell-free system was employed to determine the nature of the interaction between active purified PP2A and active recombinant MKK4. In the presence of ATP, active MKK4 phosphorylated and activated JNK, which could then activate c-Jun (Fig. 8C). However, following incubation with active PP2A, MKK4 autophosphorylation was blocked, and MKK4 activity was down-regulated. JNK phosphorylation, FIG. 5. p38 does not directly inhibit MKK4 activation. The autoradiograph shows the effect of activated p38 on the phosphorylation of JNK cascade components in a cell-free system using an in vitro kinase assay. Recombinant proteins (active MEKK1 (MEKK-1*), MKK4, JNK, and the transcription factor c-Jun) were incubated (as described under "Experimental Procedures") in the indicated combinations in the absence (Ϫ) and presence (ϩ) of active p38 (p38*) and [ 32 P]ATP. Reactions were terminated by the addition of Laemmli sample buffer. Samples were subjected to SDS-PAGE, and separated proteins were transferred to nitrocellulose. Phosphorylation of proteins was visualized by phosphorimaging.
FIG. 6. Exposure of neutrophils to calyculin A and okadaic acid activates JNK. A, effect of calyculin A on JNK activity. Neutrophils pretreated in the presence (ϩ) and absence (Ϫ) of M 39 (1 M, 45 min) were exposed to 50 M calyculin A (Ca), where indicated, for 15 min at 37°C and then stimulated with buffer (B) or 10 ng/ml TNF␣ (T) for an additional 15 min. JNK activity was determined in JNK immunoprecipitates as described previously (26). B, immunoblot depicting the effect of okadaic acid on JNK phosphorylation. Neutrophils were pretreated in the absence (Ϫ) and presence (ϩ) of okadaic acid (Oka; 1 M) for 1 h prior to stimulation in the absence (buffer (B)) and presence of TNF␣ (T; 10 ng/ml) for 15 min. Whole cell lysates were subjected to SDS-PAGE; separated proteins were transferred to nitrocellulose; and membranes were probed with anti-phospho-JNK antibody.

p38 MAPK Regulates JNK via PP2A
while still present, was diminished, and phosphorylation of c-Jun was undetectable. Dephosphorylation of MKK4 following PP2A treatment was confirmed by probing immunoblots with anti-phospho-MKK4 antibody, which recognizes only the active form of MKK4 (Fig. 8C).
Inhibition of p38 MAPK Augments TNF␣-induced Apoptosis-We have previously shown that, under adherent condi-tions, TNF␣ induces apoptosis in neutrophils, which is associated with activation of the JNK pathway (26). To determine whether p38 MAPK inhibition and the resulting enhanced JNK FIG. 7. TNF␣ stimulation induces Ser/Thr phosphatase activity in neutrophils. A, time course of phosphatase activity. Neutrophils were stimulated in the absence (Ⅺ) and presence (f) of TNF␣ (10 ng/ml) for 0, 2.5, 5, and 10 min. Samples were lysed and processed as described under "Experimental Procedures." Lysates were incubated in the presence of 32 P-labeled phosphorylase a and trichloroacetic acid-precipitated, and the released 32 P i was directly quantified by scintillation counting. TNF␣ induced a significant increase in phosphatase activity at 5 min. *, p Ͻ 0.05 (n ϭ 5; Student's t test). B, dose-dependent okadaic acid inhibition of phosphatase activity. Neutrophils were stimulated with TNF␣ for 5 min, lysed, and processed as described for A. Lysates were incubated in the absence or presence of increasing concentrations of okadaic acid (0.3, 3, and 30 M) during the phosphatase assay. Released 32 P i was quantified as described for A. C, dose-dependent calyculin A inhibition of phosphatase activity. Lysates from neutrophils stimulated and processed as described for B were exposed to 0.5 and 5 M calyculin A during the phosphatase assay. The amount of 32 P i released was determined as described for A. 8. PP2A activity is mediated by p38. A, phosphatase activity in PP2A immunoprecipitates. Neutrophils pretreated in the absence (Ϫ) and presence (ϩ) of SB 203580 were stimulated in the absence (Ϫ) and presence (ϩ) of TNF␣ for 5 min and lysed, and PP2A was immunoprecipitated. PP2A-bound beads were washed once with JNK lysis buffer and twice with Ser/Thr assay buffer. The PP2A-bound beads were then incubated with 250 M specific phosphopeptide for 5 min at 30°C and centrifuged briefly, and the supernatant containing released P i was exposed to malachite green reagent for 15 min at 30°C. P i release was quantified from changes in absorbance, and activity was normalized to unstimulated controls. Statistical analysis was performed by two-way analysis of variance. Both the TNF␣-induced phosphatase activity (TNF␣ versus buffer; *) and the SB 203580-induced inhibition of TNF␣-induced phosphatase activity (TNF␣ versus SB 203580/TNF␣; **) were found to be statistically significant (p Ͻ 0.05; n ϭ 4). B, PP2A immunoblot of PP2A immunoprecipitates. Neutrophils pretreated with SB 203580 and stimulated with TNF␣ as described for A were lysed, and PP2A was immunoprecipitated. Samples containing PP2A eluted from beads with Laemmli sample buffer were subjected to SDS-PAGE; separated proteins were transferred to nitrocellulose; and membranes were probed with an antibody directed against the PP2A catalytic subunit, PP2A c . C, PP2A dephosphorylation of active MKK4. p38 MAPK Regulates JNK via PP2A activation would exert functional consequences on TNF␣-induced apoptosis, neutrophils were pretreated with SB 203580 and exposed to TNF␣ over a period of 2.5 h, and the degree of apoptosis was assessed from characteristic changes in neutrophil morphology (Fig. 9). Both the rate and extent of TNF␣induced apoptosis were significantly greater in cells that had been pretreated with the p38 inhibitor, suggesting that p38 regulation of the JNK pathway activation similarly affected a functional outcome. The p38 inhibitor pretreatment did not change the rate or extent of spontaneous apoptosis in the absence of TNF␣ (data not shown). DISCUSSION We have previously reported that, in adherent human neutrophils, exposure to the cytokine TNF␣ results in the activation of a JNK pathway module composed of MEKK1 and MKK4/7 and in the induction of apoptosis (26). The studies described here demonstrate functional cross-talk between the p38 and JNK pathways in TNF␣-stimulated neutrophils, whereby p38 MAPK acts to limit the activation of JNK by PP2A-mediated inhibition of MKK4 (Fig. 10). We propose that the interaction between p38 and PP2A provides a novel mechanism for the regulation of JNK activation and functional responses (such as apoptosis) in response to TNF␣.

FIG.
It has recently been shown in yeast and transfectable cell lines that signaling pathways do not work linearly in isolation, but instead exhibit "cross-talk," whereby they interact with and regulate one another, creating a complex and often intricate web of MAPK pathway events superimposed upon interactions between the MAPK and other signaling pathways. Although representing only a subset of potential regulatory interactions necessary for maintaining the specificity and temporal precision of each cellular event, cross-talk nonetheless provides the cell with an important means to modify the repertoire of MAPK module responses for a given stimulus.
Several different experimental systems indicated an inhibitory role of p38 MAPK in JNK activation. The use of three different inhibitors specific for p38 MAPK, SB 203580, M 39 and the initial prototype SK&F 86002 (data not shown), at least two of which are structurally distinct, permitted the detection of comparable inhibitory effects that were not observed with the structurally similar negative control analog SB 202474, thus decreasing the likelihood that the effect of the inhibitor was artifactual. In addition to TNF␣, ultraviolet irradiation and anisomycin, two mechanistically different stress pathway (JNK/p38) inducers, similarly demonstrated increased JNK activation following p38 MAPK inhibition, indicating that this response probably represents a general phenomenon in neutrophils. Furthermore, pre-activation of the p38 pathway through PAF (a lipid mediator) resulted in significantly attenuated TNF␣-induced JNK activation. The fact that this attenuation was partial may be attributed to additional or other unknown effects induced following PAF-induced p38 MAPK activation.
Studies on upstream members of the JNK cascade undertaken to elucidate the mechanism by which p38 regulates JNK activity demonstrated that the TNF␣-induced activation of MKK4, but not MKK7 or MEKK1, is similarly enhanced upon neutrophil pretreatment with a p38 inhibitor (Fig. 4). Although MKK4 and MKK7 are dual-specificity kinases, studies have demonstrated that MKK4 preferentially phosphorylates JNK at Tyr 185 , whereas MKK7 prefers Thr 183 , and that a combination of both will invoke a synergistic response (46). Furthermore, MKK4 and MKK7 are both required for maximal activation of JNK (43-45). Hence, it is possible that inhibition of p38 bypasses the synergistic effects of both MKK4 and MKK7 on JNK, but it remains to be determined whether synergism be- FIG. 9. p38 inhibition augments TNF␣-induced apoptosis. Neutrophils were stimulated with TNF␣ (10 ng/ml) for 1, 1.5, 2, and 2.5 h following pretreatment in the absence (f) and presence (q) of the p38 inhibitor SB 203580. Unstimulated neutrophils were also pretreated in the absence (Ⅺ) and presence (data not shown) of SB 203580. Apoptosisassociated changes in neutrophil morphology were examined, and percent apoptosis was determined for each time point. Statistical analysis using two-way analysis of variance demonstrated that apoptosis was significantly increased in TNF␣-stimulated neutrophils compared with unstimulated cells at 2 and 2.5 h post-TNF␣ stimulation (p Ͻ 0.05), whereas SB 203580 (*) significantly enhanced the percent apoptosis in neutrophils exposed to TNF␣ for 2 h (p Ͻ 0.05). p38 MAPK Regulates JNK via PP2A tween MKK4 and MKK7 is still required for full JNK activation.
Utilization of a cell-free system to determine whether p38 MAPK interacts directly with MKK4 and whether this interaction has a modulating effect on events farther downstream from JNK represents, perhaps, an oversimplification of events occurring within a live cell, but a clear advantage lies in its ability to demonstrate direct interactions between the various components. Data from such experiments did not support evidence for a direct interaction between p38 and JNK pathway components, but rather suggested that a p38-mediated intermediary may regulate MKK4.
Previous studies have implicated the serine/threonine phosphatases PP1 and PP2A in TNF␣ signaling (47)(48)(49), whereas others have shown that p42/44 ERK/MAPK-induced expression of collagen in fibroblasts is regulated at the level of MEK1 through an interaction between the p38 pathway and PP2A (41). Hence, we hypothesized that the Ser/Thr phosphatases PP2A and PP1 are involved in p38-modulated regulation of JNK neutrophil activity.
Confirmation of the role of PP2A in the regulation of JNK was based on several lines of evidence. A time-dependent increase in TNF␣-induced Ser/Thr phosphatase activity was demonstrated and attributed principally to PP2A through studies that exploited the differential inhibition of PP1 and PP2A by calyculin A and okadaic acid. Furthermore, the presence of a TNF␣-stimulated PP2A activity in PP2A immunoprecipitates from neutrophils not only confirmed these data, but was shown to be p38-dependent. PP2A-mediated dephosphorylation of the active form of MKK4, determined by both in vitro kinase assay and immunochemical studies with an antibody that specifically recognizes active MKK4, subsequently provided a possible link between p38 and MKK4 activity. Interestingly, the negative regulatory activities of both PP2C in TNF␣-stimulated HeLa cells (50) and PP2A in lipopolysaccharide-stimulated THP-1 cells (51) have been linked to JNK activation, probably via MKK4.
We have shown previously that adherent human neutrophils exposed to TNF␣ undergo apoptosis that may be linked to activation of the JNK pathway (26). Our studies here demonstrate that p38 MAPK inhibition augments and accelerates apoptosis following TNF␣ stimulation. During an inflammatory response, activation of p38 MAPK by a variety of receptorligand interactions (lipopolysaccharide and G-proteins such as PAF) may act to limit TNF␣-induced apoptosis, prolonging the life span of the neutrophil and thus enabling it to carry out functional responses required in the inflammatory process. However, it is important to emphasize that, although the effect of p38 inhibition on apoptosis mirrors its effect on JNK activation, p38-mediated regulation of MKK4 via PP2A may not be the sole mechanism. p38 inhibition may modulate one of the many steps between c-Jun phosphorylation and apoptosis. Furthermore, additional pathways are also associated with apoptosis, and the inhibition of p38 may influence apoptosis via these signaling events.
The importance of p38 MAPK cross-regulation is not necessarily limited to MAPK signaling modules and apoptosis. p38 MAPK-induced inhibitory modulation of other cellular activi-ties, including gene expression (62-64), 2 STAT3 (signal transducer and activator of transcription-3) activity (65), and cell differentiation (66), has been described. This study is novel in that it describes a distinct target of p38 inhibition, MKK4. p38 negatively regulates the TNF␣-induced activation of MKK4 via PP2A, a serine/threonine phosphatase, and subsequently modulates JNK activation. These observations suggest a mechanism to regulate apoptosis, among other functional responses consequent to JNK activation, and hence modify the outcome of an inflammatory event.