Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

During infection, bacterial endotoxins potently activate macrophages functions. In these cells, LPS and/or zymosan, two natural products, stimulate the production of arachidonic acid metabolites (Huwiler and Pfeilschifter, 1993) and the secretion of proinflammatory cytokines such as TNF (Vassalli, 1992). After stimulation, bacterial endotoxins rapidly induce tyrosine phosphorylation of multiple proteins, which mediates some of the M responses (Weinstein et al., 1991). In particular, LPS activates cJun N-terminal kinase, p38 kinase and Erk1/Erk2, which all belong to a proline-directed kinase family named MAP-kinase (Sanghera et al., 1996). MAP-kinase represents a heterogeneous group of serine/threonine kinases characterized by the Thr-X-Tyr regulatory motif. It was first reported that LPS stimulates the phosphorylation of the Erk1 and Erk2 isoforms of MAP-kinases in M (Weinstein et al., 1992), and the phosphorylation cascade leading to activation of these kinases is now well defined. After LPS stimulation, Erk1/Erk2 are phosphorylated on both threonine and tyrosine residues by the MEK1,2, which is itself activated by the serine/threonine kinase raf1 (Reimann et al., 1994). In M, this cascade can be selectively triggered by phorbol esters, although it can also be activated by PKC-independent pathways (Qiu and Leslie, 1994). Erk1/Erk2 are involved in the control of many cell functions and play a pivotal role in growth-signaling pathways (Bokemeyer et al., 1996). In inflammation, they trigger cytosolic release of arachidonic acid by direct activation of PLA₂ (Qiu and Leslie, 1994) and participate in the regulation of TNF production (Dong et al., 1993).

TNF is a major mediator of inflammation (Vassalli, 1992) and a therapeutic approach to controlling its production has been the focus of considerable research. A number of publications have recently reported that TNF synthesis could be inhibited, for instance, by tyrphostins, i.e., nonselective inhibitors of tyrosine kinases (Novogrodsky et al., 1994), by selective cyclic nucleotide phosphodiesterase type IV inhibitors (Kambayashi et al., 1995), by selective p38 kinase inhibitors (Lee et al., 1994) or by new compounds of the benzamide family (Lang et al., 1995). Indeed, we have demonstrated that some benzamide derivatives, which showed anti-inflammatory activity in vivo (Robert et al., 1992; Robert et al., 1995), also potently inhibited the release of TNF from endotoxin-stimulated M. Moreover we have shown that these compounds interacted with a PKC-dependent pathway to block cytokine production. In this report, we further analyzed the mode of action of our compounds by testing their activity on PKC itself and on the phosphorylation status of Erk1/Erk2. One of our pyridinyl-benzamide derivatives (JM34), inhibited the phosphorylation of Erk2 without directly inhibiting PKC activity. In parallel, a comparable inhibition of cytosolic release of arachidonic acid was observed in resident M stimulated by zymosan, which suggests the inhibition of PLA₂ activation. In contrast, we present evidence that inhibition of TNF production by JM34 cannot be fully explained by its effect on PKC-dependent pathways.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. The following animals, drugs and chemicals were kindly provided by or obtained from the sources indicated: Sprague-Dawley rats (200 g) and male Swiss CF mice (20-25 g) (CREJ, France), (RPMI) and FCS (GIBCO, Les Ulis, France), [-³³P]ATP (3000 Ci/mmol) and [³H]arachidonic acid (100 Ci/mmol) (ISOTOPCHIM, Ganagobie, France), myelin basic protein synthetic pseudopeptide (MBP_4-14) and BIM GF109203X (CALBIOCHEM, Meudon, France), peroxydase-labeled goat anti-rabbit IgG and (polyvinylidene difluoride) membranes (AMERSHAM, Les Ulis, France), anti-ERK2 antibody (a polyclonal rabbit antiserum partially coreactive with Erk1) (Dr. Moolenaar, Division of Cellular Biochemistry, The Netherlands Cancer Institute, Amsterdam), LPS (from E. coli 055:B5), zymosan (from S. cerevisiae), phorbol ester di-butyrate (PdBu), A23187 and others products (SIGMA Saint-Quentin Fallavier, France). The benzamide derivative JM34 [N-(4,6-dimethylpyridin-2-yl)-furane-2-carboxamide)] was synthesized in the Therapeutical Chemistry Department, University of Nantes.

Isolation of M. Thioglycollate-elicited mouse M (TPM) and resident mouse M were isolated as previously described (Lang et al., 1995). Briefly, after cervical dislocation, M were obtained by peritoneal lavage with Ca-Mg-free PBS. The cell suspension (1-2 × 10⁶ cells/well) was incubated in 24-well culture plates in RPMI 10% FCS for 2 h at 37°C, 5% CO₂. Each well was then washed twice with Ca-Mg-free PBS, and adherent cells were incubated in RPMI 2% FCS for the indicated times.

Preparation of rat brain PKC and TPM PKC. A PKC mixture was prepared from rat brain as previously described by Kitano et al. (1986) with minor modifications. Sprague-Dawley rats were decapitated. Their brains were quickly removed and homogenized in a polytron homogenizer (Polytron, Luzern, Switzerland) with ice-cold extraction buffer (20 mM Tris-HCl, pH 7.5, 0.5 mM EDTA, 0.5 mM EGTA, 10 mM 2-mercaptoethanol, 0.5% Triton X-100, 25 µg/ml leupeptine, 0.25 M sucrose). The homogenate was then centrifuged at 4°C for 60 min at 100,000 × g. The supernatant was loaded on a DEAE-cellulose column, and kinase activity was eluted with 0.2 M NaCl. PKC from TPM was obtained as follows: cell suspension was incubated (20 × 10⁶ cells), in culture plates 150 mm in diameter, in RPMI 10% FCS for 2 h. Culture plates were then washed twice with Ca-Mg-free PBS, and TPM were scraped into 1 ml of ice-cold extraction buffer at 4°C. Cells were lysed for 3 × 10 s with a Branson B15 sonifier and centrifuged for 60 min at 100,000 × g. PKC from TPM was then partially purified as described above for rat brain PKC.

PKC assay. PKC activity was determined using the method of Yasuda et al. (1990), with minor modifications. Partially purified PKC was incubated for 10 min at 30°C in a reaction medium containing 20 mM Tris-HCl, pH 7.5, 1.4 mM EDTA, 1.4 mM EGTA, 20 mM MgCl₂, 10 µg/ml phosphatidylserine, 1 µg/ml diolein, 1 mM CaCl₂, 25 µM PKC substrate MBP_4-14 and 10 µM [-³³P]ATP (5 µCi/ml) in the absence or presence of the indicated concentrations of drugs. JM34 was solubilized extemporally as a maleate salt (JM34). Samples were spotted on phosphocellulose WHATMAN P81, washed twice with 1% H₃PO₄, and assayed by scintillation counting. In some tubes, we omitted phosphatidylserine and diolein in order to determine the phospholipid-independent phosphorylation of the PKC substrate. PKC activity was expressed as total cpm minus cpm of phospholipid-independent phosphorylation of the PKC substrate. Percentage of inhibition of PKC activity was expressed as [1  (PKC activity in the presence of drug / PKC activity in the absence of drug)] × 100.

M stimulation and lysis. Before stimulation, M were preincubated for 1 h with the concentrations of compounds to be tested. M were then stimulated by the addition of appropriate concentrations of LPS, zymosan, PdBu or A23187. After stimulation, cells were washed twice with ice-cold Ca-Mg-free PBS and lysed for 20 min at 4°C in lysis buffer containing 20 mM Tris-HCl, pH 7.5, 140 mM NaCl, 2 mM EDTA, 25 mM -glycerophosphate, 1% (w/v) Igepal, 1 mM orthvanadate, 0.5 mM phenylmethyl sulfonyl fluoride, 1 mM NaF and 20 µg/ml leupeptine. The detergent-insoluble material was pelleted by centrifugation (16,000 × g for 15 min at 4°C), and the soluble material was stored at 20°C. Protein concentration of cell lysates was determined using PIERCE bicinchoninic acid assay (Interchim, Montluçon, France).

Electrophoretic mobility shift assay. M lysates were subjected to 10% SDS-PAGE (acrylamide/bis-acrylamide, 99/1, w/w). After dilution with 2× Laemmli's sample buffer and boiling for 3 min, 30 µg of M lysate protein/lane was loaded onto the gel and run for 18 h at 200 V. After separation, the gel was washed and equilibrated in transfer buffer (25 mM Tris-HCl, pH 8.3, 192 mM glycine and 20% methanol) for 10 min and was electrotransferred to PVDF membrane for 24 h at 150 mA in Trans-Blot Cell (Bio-Rad, Ivry-sur-seine, France). The PVDF membrane was then blocked with 5% skim milk and 0.1% Tween-20 in TBS (TTBS) for 1 h at room temperature. The membrane was washed with TTBS and was incubated with primary antibody (anti-Erk2 in TTBS, 1/5000 dilution) for 90 min at room temperature. The membrane was then washed three times with TTBS before incubation with the second antibody (peroxydase-labeled goat anti-rabbit IgG in TTBS, 1/2500 dilution) for 90 min at room temperature. The membrane was washed three times with TTBS, and immunoreactive proteins were detected using a nonisotopic chemoluminescent system (ECL, Amersham, Les Ulis, France). The phosphorylation of Erk2 was quantified by imaging densitometer (Bio-Rad, Ivry-sur-Seine, France), and its percentage was expressed as [densitometry of phosphorylated Erk2 band / densitometry of total Erk2 bands (phosphorylated + nonphosphorylated)] × 100. Percentage of inhibition of Erk2 phosphorylation in drug-treated cells was expressed as [1  (% of Erk2 phosphorylation in drug-treated cells minus % of background phosphorylation in unstimulated cells / % of maximal Erk2 phosphorylation in untreated cells minus % of background phosphorylation)] × 100.

Determination of [³H]arachidonic acid release. Adherent resident M were radiolabeled for 18 h with [³H]arachidonic acid (0.1 µCi/ml) in RPMI containing 10% FCS. Thereafter, the medium was removed, and the cells were washed twice with PBS containing 0.2% bovine serum albumin to remove all nonincorporated [³H]arachidonic acid. Labeled cells were incubated for 60 min in RPMI containing 0.1% bovine serum albumin, in the absence or presence of the indicated concentrations of drugs. Resident M were then stimulated for 90 min with opsonized zymosan (100 µg/ml) or A23187 (0.5 µg/ml). At the end of the incubation, medium was removed, and cells were lysed with 0.5% triton-X 100. Radioactivity in supernatant and cell lysate was determined by scintillation counting. Released radioactivity was expressed as a percentage of total incorporated radioactivity.

TNF assay. After a 4-h stimulation, TPM supernatants were collected and assayed for TNF content using the WEHI164 clone 13 cytotoxic assay (Espevik and Nissen-Meyer, 1986) as previously described (Lang et al., 1995). Briefly, 50 µl of supernatant (1/200 dilution) was added to 50 µl of actinomycin D-treated (2 µg/ml) WEHI cells (6 × 10⁵ cells/ml) in flat-bottom 96-well plates and incubated for 18 h at 37°C, 5% CO₂. In each experiment, a reference curve was obtained by using serial dilutions of mouse recombinant TNF- (Boehringer Mannheim, Meylan, France), starting at 100 pg/ml and proceeding down to 0.02 pg/ml. After incubation, 50 µl of tetrazolium salts (2.5 mg/ml in PBS), were added to each well and incubated for 3 h. Formazan cristals were solubilized with 100 µl of TNF lysis buffer (1V N-N dimethyl formamide, 2V 30% SDS, adjusted to pH 4.7 with acetic acid), and optical density was read at 570 nm with an ELISA plate reader (Molecular Devices Corporation, Menlo Park, CA).

Statistical analysis. Results are expressed as mean ± S.E.M. Arachidonic acid release data were analyzed using Student's t test.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of JM34 on PKC activity in vitro. We previously demonstrated that our compounds could interact with a PKC-dependent pathway to alter cytokine production (Lang et al., 1995). In order to analyze further the molecular mechanism of these compounds, we first examined the effect of the benzamide derivative JM34 on PKC activity. The JM34 compound was selected because it potently inhibits TNF release and because it was easier to solubilize in aqueous solutions. Its activity on PKC extracted from rat brain or TPM was compared with the activity of a chlorhydrate salt of BIM, a selective PKC inhibitor (Toulec et al., 1991). We first tested the effect of increasing concentrations of BIM on rat brain PKC, which contains the four major calcium-dependent isoforms (Sekiguchi et al., 1987). We found an IC₅₀ value of about 25 nM (fig. 1A) which is very similar to the value previously described for BIM on bovine PKC (Toulec et al., 1991). As shown in figure 1B, JM34 did not significantly modify the activity of rat brain PKC, even at the highest concentrations tested in three distinct experiments. Similar results (data not shown) revealed no inhibition by JM34 of PKC from TPM, whereas BIM inhibited 34% of kinase activity at 25 nM. Therefore, JM34 does not interact with PKC pathways by directly blocking PKC activity.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Effect of drugs on PKC activity. Rat brain PKC activity was assayed in the absence or presence of increasing concentrations of BIM (panel A, n = 2) or JM34 (panel B, n = 3) in a vesicle assay using MBP4-14 as a selective PKC substrate (see "Materials and Methods"). BIM was used at 25 nM in panel B.

Effect of JM34 on the phosphorylation status of Erk2. We next examined the effect of JM34 on a molecular target located downstream of PKC. We decided to investigate the phosphorylation status of the MAP-kinase Erk1/Erk2, because growing evidence suggests a pivotal role of these kinases in PKC-dependent signaling pathways. We first used PdBu (50 nM) in combination with calcium ionophore A23187 (0.5 µg/ml) in order to focus on the PKC-dependent phosphorylation of Erk2 in TPM. Kinetic experiments revealed that optimal Erk2 phosphorylation was observed after a 10-min stimulation (data not shown). We chose this time of stimulation to compare the effects of JM34 and BIM, and in three distinct experiments, we found that JM34 potently and dose-dependently inhibited PdBu-induced Erk2 phosphorylation. Figure 2 shows a typical experiment, in which JM34 inhibited 85% of PdBu-induced Erk2 phosphorylation at 300 µM. BIM at 3 µM totally blocked PdBu-induced phosphorylation of this MAP-kinase isoform, a result consistent with the previously reported inhibition of Erk1,2 activity by BIM (Qiu and Leslie 1994). In those cells, A23187 (0.5 µg/ml) alone did not stimulate Erk2 phosphorylation (fig. 2).

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Effect of JM34 on PdBu-induced Erk2 phosphorylation. TPM were stimulated by PdBu (50 nM) in combination with A23187 (0.5 µg/ml) for 10 min before cell lysis. Whole-cell lysates were normalized for protein content, separated by 10% SDS-PAGE, transferred to PVDF membrane and probed with anti-Erk2 antibody. Immunoreactive proteins were detected using an ECL system. The data shown are typical of three independent experiments. JM34 dose-dependently inhibited Erk2 phosphorylation from 33% at 75 µM to 85% at 300 µM. A23187 alone did not stimulate Erk2 phosphorylation.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Effect of JM34 on LPS-induced Erk2 phosphorylation. TPM were stimulated with LPS (0.5 µg/ml) for 10 or 15 min before cell lysis. Erk2 was separated and detected as described in figure 2. The data shown are typical of four independent experiments. After a 15-min stimulation, BIM (3 µM) and JM34 (300 µM) inhibited 24% and 27% of Erk2 phosphorylation, respectively.

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Effect of JM34 on zymosan-induced Erk2 phosphorylation. Resident M were stimulated with zymosan (50 µg/ml) for 10 or 15 min before cell lysis. Erk2 was separated and detected as described in figure 2. The experiment shown is representative of three independent experiments. After a 15-min stimulation, BIM (3 µM) and JM34 (300 µM) inhibited 32% and 23% of Erk2 phosphorylation, respectively.

Effect of JM34 on [³H]arachidonic acid release. It has been recently reported that Erk1/Erk2 stimulate cytosolic release of arachidonic acid in various cell types through direct phosphorylation and activation of PLA₂ (Mukherjee et al., 1994). Therefore, we tested whether the effect of JM34 on zymosan-induced phosphorylation of Erk2 was associated with an inhibition of [³H]arachidonic acid release in zymosan-stimulated resident M. As shown in figure 5, JM34 dose-dependently inhibited the release of radioactive arachidonate metabolites in M supernatants. The mean inhibition achieved with 200 µM JM34 was 26.8 ± 1.1% (n = 4, P < .001) and the inhibition decreased to 9.1 ± 3.7% (P < .05) with 25 µM JM34. BIM (3 µM) inhibited 34.1 ± 4.5% of the response of zymosan-stimulated resident M, in agreement with previously published results (Qiu and Leslie, 1994). In contrast, JM34 did not inhibit arachidonic acid release from A23187-stimulated M, although only a weak stimulation was obtained with the calcium ionophore alone (data not shown).

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Effects of JM34 and BIM on the release of arachidonate metabolites. [3H]arachidonic acid was incorporated into resident M for 18 h. Cells were then preincubated with BIM (3 µM) or with increasing concentrations of JM34 for 1 h before stimulation with zymosan (50 µg/ml) for 90 min. Supernatants were collected and assessed for radioactivity content (n = 4). P < .001 for JM34 at 200 µM vs. control and P < .05 for JM34 at 25 µM vs. control.

Effect of BIM and JM34 on TNF production by LPS-stimulated TPM. A number of previous reports, including our own, have demonstrated that PKC was implicated in TNF production in LPS-stimulated M, using various PKC inhibitors (i.e., staurosporine, H7, sphingosine) that turned out not to be strictly selective for PKC. This might have led to an overestimation of PKC involvement in this process. Accordingly, we wanted to reassess the importance of PKC-dependent pathways in TNF production by using the more selective PKC inhibitor BIM and to compare its effect with that of JM34. As shown in figure 6A, the maximal inhibition of TNF release obtained with BIM was 38 ± 7.1%. In contrast, JM34 dose-dependently inhibited TNF production with a maximal inhibition of 90 ± 3.3% at 300 µM, which suggests that the additional inhibition observed, as compared with the effect of BIM, was due to the inhibition of an additional pathway independent of PKC (fig. 6B).

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Effects of BIM and JM34 on TNF release. TPM were preincubated with increasing concentrations of BIM (panel A, n = 3) or JM34 (panel B, n = 4) before stimulation with LPS for 4 h. Supernatants were assessed for TNF content using a biological assay (see "Materials and Methods").

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We previously demonstrated that new anti-inflammatory benzamide derivatives inhibited the production of TNF from endotoxin-activated M, probably through interaction with a PKC-dependent pathway (Lang et al., 1995). We have further investigated the mode of action of our compounds, and we now report that the derivative JM34 inhibited the PKC-dependent phosphorylation of the MAP-kinase isoform Erk2 without inhibiting PKC activity itself.

Endotoxins rapidly trigger PKC pathways by increasing PKC activity in monocytes (Shapira et al., 1994) and M (Qiu and Leslie, 1994). Our previous results strongly suggested that our benzamide compounds acted mainly on these pathways, because they lost nearly all their inhibitory effects on TNF production once PKC pathways had been inactivated either by PKC inhibitors or by long-term treatment with phorbol esters (Lang et al., 1995). Therefore, in this study, we decided to test first whether JM34 could block PKC activity itself. PKC was first characterized on the basis of its activation in vitro by calcium, phospholipids and DAG (Nishizuka, 1984). However, it is now clear that PKC represents a large family of isoforms that differ in their requirements for activation: the classic calcium- and DAG-dependent isoforms, the novel calcium-independent but DAG-dependent isoforms and the atypical calcium- and DAG-independent isoforms (Hug and Sarre, 1993). Because we had already shown that our compounds inhibited TNF production after activation by phorbol esters with or without calcium ionophore, we decided to focus on the DAG-dependent isoforms of PKC. Activation of those PKCs is inhibited mainly by drugs that interact with the binding site of DAG or ATP, located on their regulatory and catalytic domains, respectively (Basu, 1993). We chose a previously described vesicle assay (Yasuda et al., 1990) to investigate the effects of our compound on rat brain PKCs. In the original assay, rat brain PKCs were activated by the combination of calcium (100 µM CaCl₂), phosphatidylserine and DAG, but we found that such a high concentration of calcium decreased their dependence on DAG for activation. Indeed, in our hands, this combination of calcium and phosphatidylserine was sufficient to induce major kinase activity. To examine specifically the potential interaction of JM34 with either DAG or ATP binding sites, we lowered the final calcium concentration (see "Materials and Methods") to unmask the DAG-dependent activation of PKC. We have previously reported that this combination of low calcium concentration and phosphatidylserine was not sufficient to activate rat brain PKC, whereas major activity was recovered by adding DAG (Vernhet et al., 1996). In those experimental conditions, we found that JM34 inhibited neither rat brain PKC activity nor TPM PKC activity, in contrast with BIM, a selective PKC inhibitor interacting with the ATP binding site. These results led us to hypothesize that JM34 interacted with a biochemical event located downstream of PKC activation.

In M, phorbol esters, which directly activate PKC in a manner very similar to DAG, selectively trigger the phosphorylation of Erk1 and Erk2, with no or weak effects on phosphorylation of cJun N-terminal kinase and p38 kinase, two other members of the MAP-kinase family (Sanghera et al., 1996; Dziarski et al., 1996). In this study, we present evidence that JM34 potently and dose-dependently inhibited PdBu-induced Erk2 phosphorylation in TPM, a result that confirms that our benzamide derivative directly interacted with a PKC-dependent pathway. Our results also showed that JM34 and BIM partially blocked phosphorylation of Erk2 at comparable levels, in LPS-stimulated TPM and zymosan-stimulated resident M. This confirms that LPS and zymosan activated PKC-dependent pathway(s) responsible in part for the phosphorylation of Erk1/Erk2 and suggests that JM34 selectively blocked these pathway(s). Our data are consistent with previous studies that have reported the inhibitory effects of BIM on Erk1/Erk2 activity in zymosan-stimulated M (Qiu and Leslie, 1994). Using an analog of staurosporine as a PKC inhibitor, some investigators showed that the LPS-induced phosphorylation (Weinstein et al., 1992) and activation (Sanghera et al., 1996) of Erk1/Erk2 could also be mediated through PKC-independent pathways in M cell lines. Indeed, it is now established that both PKC-dependent and PKC-independent pathways can stimulate the Raf1/MEK1,2/Erk1,2 cascade. Raf1 is associated with the chaperones hsp90 and p50 to form a multi-subunit complex (Wartmann and Davis, 1994) that needs to be translocated to the membrane for activation (Leevers et al., 1994). Once translocated, the Raf1 complex can be either directly phosphorylated and activated by PKC (Kolch et al., 1993) or activated by a distinct but yet unknown mechanism that involves Src tyrosine kinases (Gupta et al., 1995). Further studies will be needed to determine whether JM34 can interact with the Raf1 complex to inhibit its activation by PKC and thus inhibit Erk2 phosphorylation.

In inflammation, Erk1/Erk2 regulate some M responses, such as eicosanoid formation (Mukherjee et al., 1994). The production of these lipid mediators is largely dependent on the availability of free cytosolic arachidonate that is released from membrane phospholipids after PLA₂ activation. In many cell types, including M, agonist-induced arachidonic acid release involves activation of the 85-Kda cytosolic PLA₂ (cPLA₂) (Qiu et al., 1993), activation that is regulated mainly by intracellular calcium level and by phosphorylation. It has been reported that Erk1/Erk2 could increase the cytosolic release of arachidonic acid by direct phosphorylation of cPLA₂ on ser-505 (Lin et al., 1993). Our present data show that in JM34-treated resident M, the partial inhibition of Erk2 phosphorylation was associated with a similar inhibition of arachidonic acid release after zymosan stimulation. In addition, we found that the percentages of inhibition induced by JM34 and BIM were very similar and that JM34 did not block the response induced by A23187 alone, which does not induce Erk1/Erk2 activation in resident M (Qiu and Leslie, 1994). These results suggest that in resident M, JM34 decreased arachidonic acid release mainly through its inhibition of PKC-dependent Erk2 phosphorylation.

PKC-dependent pathways have been clearly implicated in the production of TNF- by endotoxin-stimulated cells (Kovacs et al., 1988; Geng et al., 1993; Shapira et al., 1994; Tschaikowsky, 1994, Lang et al., 1995). Some of these reports have shown that PKC could regulate TNF- gene expression in human monocytes (Geng et al., 1993; Shapira et al., 1994) or murine M (Kovacs et al., 1988), because PKC inhibitors, such as staurosporine and H7, potently blocked LPS-induced TNF- mRNA accumulation. However, the importance of PKC-dependent pathways in TNF- production remains unclear, because 1) the inhibitors used in previous studies were not strictly selective for PKC and 2) phorbol esters, which strongly activate PKC and Erk2 phosphorylation, only weakly stimulate TNF- production by murine M as compared with LPS (Hambleton et al., 1995; Lang et al., 1995). We confirmed a role for a PKC-dependent pathway in LPS activation of TNF production, because BIM, a very selective PKC inhibitor, inhibited 38% of TNF production in LPS-stimulated M. Nevertheless, the inhibition obtained with BIM was lower than the inhibition we previously obtained with staurosporin or sphingosine (83% and 48%, respectively). This further supports the hypothesis that these latter PKC inhibitors interacted with other enzymes than PKC to inhibit TNF production in activated M. Our results showed that JM34 was a much more potent inhibitor of TNF production than BIM (fig. 6), whereas both products inhibited Erk2 phosphorylation at similar levels. In addition, it was recently reported that selective activation of the raf1/Erk1,2 cascade only partially mimicked LPS-induced signaling events in murine M. In particular, TNF production after selective Erk1,2 cascade activation was 20-fold less than that obtained after stimulation with LPS (Hambleton et al., 1995). Taken together, these results strongly suggest that JM34 interacted with additional pathway(s) distinct from the Erk1,2 cascade to block LPS-induced TNF production. It has been recently reported that new bicyclic pyridinyl-imidazol compounds, which inhibit TNF production by selectively blocking p38 kinase activity (Lee et al., 1994), exhibit potent anti-inflammatory activity on endotoxin shock and experimental arthritis in murine models (Badger et al., 1996). It was further suggested that these compounds acted mainly at a post-transcriptional level of TNF production. Our own preliminary results indicate that JM34 also acts mainly on post-transcriptional events (L. Vernhet, unpublished data). This prompts us to investigate the effect of JM34 on the p38 pathway.

In conclusion, we have shown that our benzamide derivative, JM34, inhibited the release of two major proinflammatory molecules from activated M: arachidonate metabolites and TNF. Its inhibitory effect on PKC-dependent phosphorylation of Erk2 may fully account for its PLA₂ inhibition and subsequent arachidonic acid release, but its inhibition of TNF production probably involves an additional signaling pathway(s) that remains to be determined.