Inhibition of Tumor Necrosis Factor-induced p42/p44 Mitogen-Activated Protein Kinase Activation by Sodium Salicylate*

Tumor necrosis factor (TNF) activates both p42 and p44 mitogen-activated protein kinases (MAPK) in human FS-4 fibroblasts, cells for which TNF is mitogenic. We now show that TNF activates p42 MAPK in two cell lines whose growth is inhibited by TNF. A mutant TNF that binds only to the p55 TNF receptor (TNFR) produced a similar degree of activation as wild-type TNF in FS-4 fibroblasts, indicating that the p55 TNFR is sufficient to mediate p42/p44 MAPK activation. The up- stream intracellular signals that couple the TNFR to MAPK activation are still poorly defined. We now show that neither phorbol ester-sensitive protein kinase C nor G i (cid:97) link TNF to p42/p44 MAPK activation, because pretreatment of FS-4 cells with phorbol ester to down-regulate protein kinase C or pretreatment with pertus- sis toxin to block G i (cid:97) does not inhibit p42/p44 MAPK activation by TNF. To further analyze MAPK activation in FS-4 cells, we compared p42/p44 MAPK activation by TNF and epidermal growth factor (EGF). While tyrosine phosphorylation of p42/p44 MAPK was detected almost immediately (30 s) after stimulating cells with EGF, TNF-induced tyrosine phosphorylation was detected only after a more prolonged time interval (initially de- tected at 5 min and peaking at 15–30 min). In addition, the anti-inflammatory drug sodium salicylate, previ- ously demonstrated to inhibit activation


Tumor necrosis factor (TNF) activates both p42 and p44 mitogen-activated protein kinases (MAPK) in human FS-4 fibroblasts, cells for which TNF is mitogenic.
We now show that TNF activates p42 MAPK in two cell lines whose growth is inhibited by TNF. A mutant TNF that binds only to the p55 TNF receptor (TNFR) produced a similar degree of activation as wild-type TNF in FS-4 fibroblasts, indicating that the p55 TNFR is sufficient to mediate p42/p44 MAPK activation. The upstream intracellular signals that couple the TNFR to MAPK activation are still poorly defined. We now show that neither phorbol ester-sensitive protein kinase C nor G i␣ link TNF to p42/p44 MAPK activation, because pretreatment of FS-4 cells with phorbol ester to downregulate protein kinase C or pretreatment with pertussis toxin to block G i␣ does not inhibit p42/p44 MAPK activation by TNF. To further analyze MAPK activation in FS-4 cells, we compared p42/p44 MAPK activation by TNF and epidermal growth factor (EGF). While tyrosine phosphorylation of p42/p44 MAPK was detected almost immediately (30 s) after stimulating cells with EGF, TNF-induced tyrosine phosphorylation was detected only after a more prolonged time interval (initially detected at 5 min and peaking at 15-30 min). In addition, the anti-inflammatory drug sodium salicylate, previously demonstrated to inhibit NF-B activation by TNF, blocked the activation of p42/p44 MAPK in response to TNF but not in response to EGF. These findings demonstrate that the TNF and EGF receptors utilize distinct signaling molecules to couple to MAPK activation. Elucidation of the mechanism whereby sodium salicylate blocks TNF-induced p42/p44 MAPK activation may help to clarify TNF-activated signaling pathways.
Tumor necrosis factor (TNF), 1 a cytokine originally described as a mediator of endotoxin-induced hemorrhagic necrosis of tumors, possesses potent immunomodulatory capacities and is believed to play key roles in inflammation, septic shock, and cachexia (1,2). Two TNF receptors (TNFRs) of 55 kDa (p55) and 75 kDa (p75) are expressed on many types of cells and transduce the TNF signal (3,4). The extracellular portions of the receptors possess structural features also present in the extracellular domain of the nerve growth factor receptor and other receptors comprising the TNFR superfamily (3). Although the intracellular domains of the p55 and the p75 TNFRs do not display significant homology to each other, the cytoplasmic portion of the p55 TNFR contains a death domain also found in the Fas antigen, necessary for signaling cell death (5). Recent studies have identified several proteins associated with the cytoplasmic domains of the p55 (6 -8) and p75 (9) TNFRs.
The initial event in TNF signal transduction involves association of a trimeric TNF molecule with its receptor and subsequent receptor oligomerization. There is evidence for the involvement of several receptor-distal elements in the propagation of the TNF signal, including GTP-binding proteins (10,11), protein kinase A (12), and protein kinase C (PKC) (13,14). In addition, TNF has been shown to activate phosphatidylcholine-specific phospholipase C (15) and cytosolic phospholipase A 2 (16) and to cause an increase in arachidonic acid metabolism (16 -18). There is also evidence for sphingomyelinase activation by TNF, leading to the generation of ceramide, a potent second messenger (19). Finally, TNF has been shown to activate several members of a large and growing family of mitogen-activated protein kinases (MAPKs) (20 -24).
MAPKs constitute a family of related and evolutionarily conserved serine/threonine kinases that are important in cell growth and differentiation (25)(26)(27). They become activated by phosphorylation on threonine and tyrosine in response to many external stimuli. MAPKs preferentially recognize a minimal substrate consensus sequence, Ser/Thr-Pro (28), present in many proteins. Accordingly, MAPKs appear to have many possible substrates, including transcription factors such as STAT1␣ (29) and the serum response factor accessory protein Elk-1 (30), as well as cytosolic proteins such as cytosolic phospholipase A 2 (31). Consistent with the role of MAPKs in modulating transcription factor function is the observation that MAPKs frequently translocate from the cytoplasm into the nucleus following growth factor stimulation (32). Three MAPK subfamilies have been identified, consisting of the p42 and p44 MAPKs (also termed extracellular signal-regulated kinases (ERKs)), the c-Jun N-terminal kinases (JNK)/stress-activated protein kinases (SAPK), and the p38 MAPKs (20,(33)(34)(35) (see  Ref. 36 for a review). Although each of the subfamilies appears to have its own set of activators, there is a certain amount of crossover in the pathways. For example, while MAPK-ERK kinase kinase (MEKK) preferentially functions in the JNK/ SAPK pathway, it can also activate MAPK-ERK kinase (MEK), located in the p42/p44 MAPK pathway (37,38). TNF can activate all three MAPK subfamilies (20 -22). In particular, TNF has been shown to activate both MEK (39,40) and MEKK (40). The activation pathway for p42/p44 MAPK by receptor tyrosine kinases typically involves Ras, either Raf or MEKK, and MEK (41)(42)(43)(44). While there is evidence for both MEK (39,40) and MEKK (40) involvement in TNF actions in some cells, it is not clear what other components may participate in the TNFmediated activation of p42/p44 MAPK.
In earlier work, we showed that TNF activated p42/p44 MAPK in normal human FS-4 fibroblasts (21). The goal of this study was to further characterize the activation of p42/p44 MAPK by TNF and to compare this activation with that mediated by other agents, especially EGF. Our findings indicate that in addition to activating p42/p44 MAPK with different kinetics, TNF and EGF-activated pathways differ in their susceptibility to inhibition by sodium salicylate. Whereas TNFinduced p42/p44 MAPK activation was strongly inhibited, sodium salicylate showed no significant effect on MAPK activation by EGF. It is possible that this newly demonstrated inhibition of TNF signaling contributes to the anti-inflammatory action of sodium salicylate. Elucidation of the mechanism of the inhibitory effect of sodium salicylate may help to clarify the signaling pathways activated by TNF.

EXPERIMENTAL PROCEDURES
Cell Culture-All tissue culture media were obtained from Life Technologies Inc. Normal human diploid FS-4 fibroblasts (45) were used at passage 15. FS-4 cells were cultured in Eagle's minimal essential medium supplemented with 6 mM Hepes, 3 mM Tricine, 100 units/ml penicillin, 100 g/ml streptomycin, and 5% heat-inactivated fetal bovine serum (FBS; Gemini Bioproducts, Inc.). Before use in the experiments, cells were washed twice with phosphate-buffered saline (136 mM NaCl, 3 mM KCl, 8 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , pH 7.4) and then serum-starved for 2-3 days in minimal essential medium with 0.25% FBS. HeLa B human carcinoma cells, obtained from Dr. Ed Ziff (NYU Medical Center), were cultured in minimal essential medium with 10% FBS and serum-starved overnight in minimal essential medium with 0.5% FBS. HT-29 human adenocarcinoma cells were cultured in Mc-Coy's 5a medium supplemented with 10% FBS and serum-starved overnight in McCoy's 5a medium with 0.5% FBS.
Detection of MAPK Activity Using MBP-containing SDS-Polyacrylamide Gels-This assay was performed as described previously (21). Briefly, whole cell lysates were generated using a buffer consisting of 1% Nonidet P-40, 50 mM Hepes, pH 7.5, 100 mM NaCl, 2 mM EDTA, 1 mM pyrophosphate, 10 mM sodium orthovanadate, 3 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, and 100 mM sodium fluoride. Lysates were subsequently subjected to 10% SDS-polyacrylamide gel elec-trophoresis in a gel copolymerized with 0.1 mg/ml MBP. Following electrophoresis, the proteins in the gel were successively denatured, renatured, and subjected to an in situ kinase reaction with 125 Ci of [␥-32 P] ATP. The gel was then dried and autoradiographed.
Immunoblotting-This was performed essentially as described previously (21), with some modifications. Equal amounts of cell lysate (generated as described above) were subjected to 10% SDS-polyacrylamide gel electrophoresis and then transferred to Immobilon-P membrane using transfer buffer (25 mM Tris, 192 mM glycine, 20% (v/v) methanol). The membrane was first incubated in TBS (10 mM Tris, pH 7.4, 150 mM NaCl) and then blocked overnight at room temperature in TBS-5% bovine serum albumin before being probed with either anti-p42 MAPK antibody 1B3B9 diluted 1:10,000 in TBS-5% bovine serum albumin, or anti-Tyr(P) diluted 1:200 in TBS-5% bovine serum albumin. For Fig. 1A, antibody-antigen complexes were detected with biotinylated goat anti-mouse IgG, a streptavidin/alkaline phosphatase conjugate, and the nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate development reagents (Life Technologies, Inc.). For all other immunoblots, the antibody-antigen complexes were detected with the aid of horseradish peroxidase-conjugated Protein A and a chemiluminescent development kit (Kirkegaard and Perry Laboratories, Inc.).

Activation of p42 MAPK Does Not Correlate with Mitogenic
Activity of TNF-We have previously shown that TNF activates both p42 and p44 MAPK in normal human FS-4 fibroblasts (21). This activation is manifested by increased tyrosine phosphorylation of p42 and p44 MAPK, concomitant with a mobility shift detectable by Western blotting with anti-MAPK antibodies, and an increased ability to phosphorylate a substrate, MBP. Since TNF is mitogenic for FS-4 cells (45), it was of interest to determine whether TNF would induce MAPK activation in cells for which TNF treatment is growth-inhibitory, such as the HeLa and the HT-29 adenocarcinoma cell lines (46). Western blot analysis with an anti-p42 MAPK antibody showed that in HeLa and HT-29 cells, as well as in FS-4 cells, TNF treatment caused the appearance of a band with decreased mobility, characteristic of phosphorylation and activation of p42 MAPK (Fig. 1A). In addition, treatment of FS-4 cells with IL-1, previously demonstrated to activate MAPK in other cell types (47), caused an identical mobility shift for p42 MAPK. In contrast, double-stranded RNA poly(I)⅐poly(C), which induces signaling pathways that resemble those of some cytokines (48), did not cause a mobility shift for p42 MAPK in FS-4 cells. As a corollary to the immunoblotting analysis, the lysates described above were subjected to an in-gel kinase assay with MBP co-polymerized in the gel as a substrate. Cells stimulated with either TNF or IL-1 showed increased kinase activity of a band consistent in size with p42 MAPK (Fig. 1B). The relative intensities of the other bands in this gel were not significantly altered by any of the stimuli. These results indicate that the ability of TNF to activate p42 MAPK is not limited to cell lines for which TNF is mitogenic.
Activation of p42/p44 MAPK via the p55 TNF Receptor-Human TNF binds with a high affinity to both the p55 TNFR and the p75 TNFR (4). In order to determine the relative contribution of the p55 TNFR to MAPK activation, FS-4 cells were stimulated with a mutant TNF (Trp 32 /Thr 86 ) that can bind the p55 TNFR but has lost its ability to interact with the p75 TNFR (49). Lysates were generated and immunoblotted with anti-Tyr(P). We have shown earlier (21) that, with this antibody as a probe, p42 and p44 MAPK are the only proteins in this molecular weight range whose tyrosine phosphorylation is visibly increased upon TNF stimulation of FS-4 cells. Both wild-type and mutant TNF proteins caused an increase in the tyrosine phosphorylation and hence presumed activation of p42 and p44 MAPK (Fig. 1C). A faint band migrating under the pp42 MAPK, visible after TNF or mutant TNF treatment, has not been positively identified but may represent p38 MAPK, known to be inducible by TNF (22). Treatment with TNF or mutant TNF also caused a mobility shift for p42 MAPK (data not shown). These data indicate that triggering of the p55 TNFR alone is capable of mediating p42/p44 MAPK activation by TNF.
Kinetics of TNF-and EGF-induced MAPK Activation-As a first step toward the examination of pathways mediating p42/ p44 MAPK activation by TNF, we compared the kinetics of p42/p44 MAPK activation produced by TNF with those mediated by EGF, a well characterized activator of MAPK. In agreement with previous studies in other cell lines (50,51), treatment of FS-4 cells with EGF caused a rapid and persistent activation of p42/p44 MAPK (Fig. 2). EGF produced an increase in p42 MAPK tyrosine phosphorylation after 30 s, and a marked increase in tyrosine phosphorylation of both p42 and p44 MAPK occurred by 2 min of EGF stimulation ( Fig. 2A, top  panel). This increase correlated with a mobility shift for p42 MAPK (Fig. 2A, bottom panel). Increased tyrosine phosphorylation persisted for at least 1 h of EGF treatment (Fig. 2B). In contrast, TNF activated p42 and p44 MAPK only after 5-10 min of stimulation ( Fig. 2A), and this activation was relatively transient, peaking at 15 min and ending after 40 min of TNF treatment (Fig. 2B). No second wave of activation was seen when the cells were maintained in the continuous presence of TNF for up to 24 h (data not shown). It is noteworthy that EGF induces the tyrosine phosphorylation of a number of proteins in FS-4 cells, including the EGF receptor (top band in top panel of Fig. 2A) and that this phosphorylation peaks before that of p42/p44 MAPK.
Effect of Pertussis Toxin Pretreatment on MAPK Activation-Several studies have suggested roles for G proteins in TNF signal transduction (10,11). In order to determine whether a G i␣ -protein is involved in the TNF-mediated activation of p42/ p44 MAPK, FS-4 cells were treated for 18 h with 500 ng/ml PT before their stimulation with LPA, TNF, or EGF. LPA is known to activate p42 MAPK through a PT-sensitive G protein-coupled receptor pathway (52). PT pretreatment of FS-4 cells abolished both the increase in tyrosine phosphorylation and the partial mobility shift of p42 MAPK induced by LPA (Fig. 3). In contrast, PT pretreatment had no effect on the tyrosine phosphorylation of p42/p44 MAPK or on the mobility shift of p42 MAPK induced by either TNF or EGF. These results suggest that the activation of p42/p44 MAPK by TNF (and EGF) occurs via a G i␣ -independent pathway. It should be noted, however, that these data do not preclude the possibility that other types of G proteins, not inhibited by PT, may mediate the activation of p42/p44 MAPK by TNF.
Effect of High Dose Phorbol Ester Pretreatment on MAPK Activation-A number of studies have examined possible roles for PKC in TNF signal transduction (13,14). We examined the contribution of the TPA-sensitive PKC isotypes to the TNFmediated activation of p42/p44 MAPK. A brief stimulation of FS-4 cells with a low dose of TPA (20 ng/ml) caused MAPK activation, as judged by both an increase in phosphotyrosine content for p42 and p44 MAPK and a mobility shift for p42 MAPK (Fig. 4). Following a 24-h pretreatment of FS-4 cells with a high dose of TPA (1 g/ml) in order to deplete TPAsensitive PKC, the ability of TPA to activate p42/p44 MAPK was severely diminished. However, under the same conditions, the ability of TNF and EGF to activate p42/p44 MAPK was not affected. Therefore, TPA-sensitive PKC isoforms are not involved in the activation of p42/p44 MAPK by TNF (or EGF). This result does not exclude the possibility that TNF may activate MAPK via a PKC-dependent pathway because of the existence of several PKC isoforms, some of which are TPAinsensitive. One such isoform is PKC, shown to be necessary for mitogenic activation in fibroblasts (53), to bind Ras and to function downstream of Ras following growth factor stimulation (54), and to associate with and activate MEK-MAPK complexes in vitro (13). Since TNF has been shown to activate PKC (13,55), it is plausible that one pathway of p42/p44 MAPK activation by TNF involves PKC.
Effect of Sodium Salicylate on MAPK Activation by TNF-Given the central role that TNF plays in many inflammatory processes (1,2), it was of interest to determine the effect of an anti-inflammatory drug on p42/p44 MAPK activation by TNF. Sodium salicylate has previously been demonstrated to inhibit the activation of the transcription factor NF-B in response to TNF and other agents (56). Pretreatment of FS-4 cells with 20 mM sodium salicylate inhibited the TNF-mediated increase in tyrosine phosphorylation of p42/p44 MAPK, with this inhibition becoming significant after 30 min of sodium salicylate pretreament and more prominent after 60 min of pretreatment (Fig. 5A). Immunoblotting the lysates with an anti-p42 MAPK antibody demonstrated that this inhibition of TNF-induced MAPK tyrosine phosphorylation correlated with an inhibition of the TNF-induced p42 MAPK mobility shift and was not simply due to down-regulation of MAPK protein levels by sodium salicylate treatment (data not shown). Sodium salicylate also inhibited the tyrosine phosphorylation of p42/p44 MAPK induced specifically via the p55 TNFR by the Trp 32 /Thr 86 mutant TNF (Fig. 5A). The effect of sodium salicylate on p42/p44 MAPK activation appears to be specific for TNF, since sodium salicylate did not block p42/p44 MAPK tyrosine phosphorylation in FS-4 cells in response to EGF stimulation (Fig. 5B) or in response to platelet-derived growth factor (data not shown). To rule out the possibility that the inhibitory effect was due to toxicity produced by the combination of sodium salicylate and TNF, cells pretreated with sodium salicylate were stimulated with a mixture of TNF and EGF. Under these conditions, sodium salicylate failed to inhibit the increase in EGF-induced p42/p44 MAPK tyrosine phosphorylation (Fig. 5B), indicating a lack of toxicity. A dose-response experiment demonstrated that only the relatively high dose of 20 mM sodium salicylate caused a prominent inhibition of the TNF-mediated tyrosine phosphorylation of p42/p44 MAPK after a 30-min pretreatment (data not shown). DISCUSSION Previous studies have demonstrated the TNF-mediated activation of p42/p44 MAPK in several different types of cells (20 -24). However, the signaling pathways linking the TNFRs to p42/p44 MAPK activation remain largely unknown. The main goal of the present study was to compare p42/p44 MAPK activation by TNF and by EGF in the human diploid FS-4 fibroblasts and to determine whether the pathways of activation by these agents can be distinguished. A clear difference was seen in the kinetic patterns of activation by the two agents, as judged by the increases in phosphotyrosine contents of p42 and p44 MAPK, such that EGF produced both a more rapid and a more sustained activation than TNF. In view of an earlier observation that TNF-induced NF-B activation can be inhibited by acetylsalicylic acid or sodium salicylate (56), we examined the effect of the latter agent on p42/p44 MAPK activation in our system. Sodium salicylate produced a marked inhibition of TNF-induced p42 and p44 MAPK tyrosine phosphorylation but failed to suppress EGF-induced activation. This different response to sodium salicylate indicates that TNF and EGF produce p42/p44 MAPK activation via different routes and suggests an important role for phospholipid metabolism in the TNF-activated pathway. Our demonstration of the inhibitory action of sodium salicylate on TNF-induced p42/p44 MAPK activation may help to design new approaches to the analysis of TNF signaling.
Our finding that the p55 TNFR is sufficient for activation of p42/p44 MAPK confirms similar findings by others in macrophages (57), and complements earlier observations that the p55 receptor independently can mediate TNF's antiviral activity (58) and also stimulate PKC, NF-B, phospholipase A 2 , and sphingomyelinase activities (59). The cytoplasmic portion of the p55 receptor has been shown to possess a so-called "death domain" responsible for signaling cytotoxicity, which is homol-  ogous to the intracellular portion of the Fas antigen (5). Since Fas has also been shown to activate MAPK (60), it will be of interest to determine whether it is the death domain that mediates the activation of p42/p44 MAPK by the p55 receptor. Pagès et al. (25) have demonstrated that p42/p44 MAPK is required for the proliferation of fibroblasts in response to growth factor stimulation. We showed that TNF activates p42/ p44 MAPK in FS-4 fibroblasts, cells for which it is mitogenic (45). However, we now show that TNF also activates p42 MAPK in HeLa and HT-29 cells, in which TNF is cytostatic (46). These findings indicate that while p42 MAPK activation may be required for growth factor-induced mitogenesis, it probably serves a more general role in TNF signaling.
The kinetics of p42/p44 MAPK tyrosine phosphorylation in FS-4 fibroblasts in response to TNF treatment, with increased phosphorylation beginning at 5 min and ending after approximately 40 min of stimulation, are in agreement with the findings of other investigators in different cell types (24,57). This relatively transient TNF-induced tyrosine phosphorylation of p42/p44 MAPK is also consistent with the findings demonstrating that while TNF is a potent activator of the SAPK/JNK pathway, it is a relatively weak activator of p42/p44 MAPK (33,61). The relatively brief period of p42/p44 MAPK activation by TNF and the apparent existence of redundancy in target selection by members of the MAPK subfamilies raise questions as to the possible functions of p42/p44 MAPK in TNF signaling. For example, the Elk-1 transcription factor, known to be phosphorylated by p42/p44 MAPK in response to TPA treatment, appears to be principally phosphorylated by JNK in response to IL-1 (and most probably TNF) stimulation (62). Further studies will be required to determine the specific downstream targets of TNF-activated p42/p44 MAPK.
Our finding that sodium salicylate blocks the TNF-mediated tyrosine phosphorylation of p42/p44 MAPK in FS-4 fibroblasts complements earlier observations demonstrating the sodium salicylate-mediated inhibition of NF-B activation in response to TNF and other agents (56). In addition to its effects on NF-B, sodium salicylate has been shown to induce DNA binding of the heat shock transcription factor in HeLa cells (63) and to function as a signaling molecule in the response of plants to infection with various pathogens (64,65). Recent work has also demonstrated that sodium salicylate, at suprapharmacological concentrations, can inhibit inducible nitric-oxide synthase expression and nitrite production in murine macrophages (66). To our knowledge, the data described in our present paper represent the first demonstration of sodium salicylate's interference in a cytokine-induced kinase cascade. Along with the above mentioned actions of sodium salicylate, this interference with the TNF/MAPK pathway may contribute to the well documented anti-inflammatory effect of sodium salicylate in vivo. It is likely that sodium salicylate may also block the pathways for MAPK activation by other inflammatory cytokines and/or that sodium salicylate may block other signaling pathways employed by TNF and other inflammatory agents.
The sodium salicylate-mediated inhibition of TNF-induced p42/p44 MAPK activation was clearly demonstrable only at the relatively high sodium salicylate dose of 20 mM. This same dose has also been used by others to demonstrate several other actions of sodium salicylate (63,66). In view of the pH dependence of sodium salicylate action in at least one system, in which decreased local pH increased the effectiveness of sodium salicylate (63), it is possible that lower concentrations of sodium salicylate will inhibit the TNF-induced p42/p44 MAPK pathway under some conditions. It will be important to determine the specific step at which sodium salicylate interferes with the activation of p42/p44 MAPK by TNF. In view of its inhibitory action on NF-B activation by TNF (56), it is possible that sodium salicylate exerts a global inhibitory effect on TNF signaling, and acts at a far upstream, TNF receptor-proximal site. In this regard, our preliminary experiments in FS-4 cells indicate that the TNF-mediated activation of JNKs is also significantly decreased by sodium salicylate pretreatment. A principal mechanism of action of nonsteroidal anti-inflammatory drugs is thought to be interference with pathways of arachidonic acid metabolism (67)(68)(69). Because TNF has been demonstrated to cause an increase in arachidonic acid metabolism (17,18) and because arachidonic acid itself can cause MAPK activation in at least one cell type (70), sodium salicylate may be blocking the TNF-mediated induction of p42/p44 MAPK by interfering with arachidonic acid metabolism. Alternatively, because sodium salicylate has been shown to inhibit phospholipase C activity (71) and because TNF is known to activate  3 and 7), or co-stimulated for 15 min with both TNF and EGF at the above concentrations (lanes 4 and 8). Lysates were then generated, blotted, and probed with anti-Tyr(P). Arrows in both panels A and B denote positions of activated (pp42 and pp44) MAPKs. phospholipase C-dependent pathways (15), sodium salicylate may be blocking the TNF-mediated induction of p42/p44 MAPK by interfering with phospholipase C metabolism. In any case, our present work broadens the spectrum of potential targets for the anti-inflammatory actions of sodium salicylate and could serve as a tool to further dissect TNF signaling pathways.