Activation of Mitogen-Activated Protein Kinase Pathways by Mycoplasma fermentans Membrane Lipoproteins in Murine Macrophages: Involvement in Cytokine Synthesis

Mycoplasmas, the smallest self-replicating bacteria, are characterized by a wall-less envelop and an unusually small genome size. They are closely related to Gram-positive bacteria having a low guanine/cytosine (G/C) content (1). The Mycoplasma genus belongs to the Mollicutes class, which consists of eight genera. Although mycoplasmas are generally commensal parasites in human, some species are real pathogens capable of causing a wide variety of diseases (2). One features of mycoplasmas is their ability to interact with and stimulate cells of the immune system, including B and T cells and monocyte/macrophages. Mycoplasma fermentans is suspected of being a pathogen of several human diseases, mainly it is suspected to play a role in rheumatoid arthritis and proposed to be involved in immune defect (for review, see 3 .

Stimulation of monocytes and resident macrophages by mycoplasmas induces the production of numerous cytokines (i.e., IL-1β, TNF-α, IL-6, IL-10, IFN-γ, and granulocyte/mcrophage-CSF) (4, 5, 6). This immunomodulatory effect seems to reside in the lipid-associated membrane protein (LAMP)² fraction of mycoplasmas (4). We have previously reported that protein tyrosine phosphorylation is an early event in macrophage activation by LAMP and that blocking of protein tyrosine kinases (PTKs) inhibits downstream pathways leading to cytokine production in response to LAMP (4). However, the biochemical events involved in the mycoplasma-induced cytokine synthesis by macrophages are poorly understood.

Mitogen-activated protein kinases (MAPKs) are a group of serine/threonine-specific, proline-directed protein kinases that are activated by a wide spectrum of extracellular stimuli. They are important mediators involved in the intracellular network of interacting proteins that transduce extracellular signals to intracellular responses. To date, several distinct MAPKs expressed in vertebrates have been identified, including extracellular signal-regulated kinase 1 and 2 (ERK1/2), c-Jun NH2-terminal kinase (JNK)/stress-activated protein kinases (SAPK), and p38/RK/Mpk2 (7). Each of these effectors is regulated by other upstream kinases. MAPK/ERK kinase 1 (MEK-1) is responsible of ERK1/2 activation, and Raf1 is the kinase upstream of MEK-1. Likewise, SEK1 (stress-activated protein kinase/ERK kinase) is responsible of SAPK/JNK phosphorylation, and MEKK1 (mitogen-activated protein kinase/ERK kinase kinase) is the upstream effector of SEK1. MKK3 (mitogen-activated protein kinase kinase homologue) activates p38, which in turn phosphorylates a MAPKAPK2 (MAPK-activated protein kinases 2) (7).

Bacterial Gram-negative LPS, a very potent inducer of cytokine synthesis by monocytic cells, has recently been demonstrated to activate multiple MAPK-related pathways (8). To examine further aspects of signaling pathways involved in macrophage activation by mycoplasmas, we tested the ability of M. fermentans-derived LAMP (LAMPf) to induce the activation of the different MAPKs and the involvement of these pathways in cytokine synthesis. The present study demonstrates that treatment of murine macrophage cells with LAMPf results in the activation of ERK, JNK, and p38 kinases. Furthermore, our results show that MAPK pathways are involved in LAMPf-induced IL-1β, TNF-α, and IL-6 production by murine macrophages. Unlike LPS, LAMPf do not require serum proteins or cell surface CD14 to exert these effects.

LPS from Escherichia coli O55:B5, myelin basic protein (MBP), and polymyxin B were purchased from Sigma (Saint Quentin, Fallavier, France). Genistein, PD-98059, SB203580, and GST-c-Jun(1–79) were from Biomol Research Laboratories (Philadelphia, PA). Triton X-114 (TX-114) was obtained from Merck (Nogent sur Marne, France). Anti-JNK1 (C17), anti-ERK (K23), and anti-p38 (C20) polyclonal Abs were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phosphotyrosine and agarose-coupled anti-phosphotyrosine mAbs (anti-PY 4G10 Ab), anti-rat MAPK R2 (anti-ERK1-CT), GST-MAPKAPK2, and p38/RK/Mpk2 assay kits were obtained through Upstate Biotechnology (Lake Placid, NY). Peroxidase-coupled anti-phosphotyrosine mAb (anti-PY RC20 Ab) was from Transduction Laboratories (Lexington, KY). Enhanced chemiluminescence kit (ECL), p42/44^MAPK enzyme assay, and [γ-³²P]ATP (3000 Ci/mmol) were commercially available from Amersham (Les Ulis, France). MY4, an anti-human CD14 mAb, was purchased from Coulter Diagnostics (Hialeah, FL).

Mycoplasma fermentans (PG18 strain) and Mycoplasma pneumoniae (FH LIV strain) were cultivated in medium containing 20% horse serum (Life Technologies, Cergy Pontoise, France), 10% freshly prepared yeast extract, 1% glucose and 1000 U/ml penicillin G. Mycoplasma cultures were incubated at 37°C and 5% CO₂, then quantified as described by Rodwell and Whitcomb (9) and expressed as CCU (color changing unit) per milliliter. For heat inactivation, mycoplasmas were isolated by centrifugation (15,000 × g at 4°C for 30 min), washed, and resuspended in Hayflick medium at 10⁷ CCU/ml, followed by incubation at 60°C for 30 min. Heat-inactivated mycoplasmas (HIM) were stored at −70°C until needed. Complete heat inactivation was verified by inoculation of 10⁶ CCU into 10 ml of culture medium to monitor the growth of mycoplasmas; no growth could be observed over a 2-wk period. Mycoplasma membrane lipoproteins were prepared by hydrophilic/hydrophobic fractionation using the TX-114 partitioning method as described previously (10). Protein concentrations were determined by microBCA assay (Pierce, Rockford, IL). The endotoxin level of both HIM and LAMP preparations was <60 pg/ml, as determined by Limilus amebocyte lysate assay (Haemachem, St. Louis, MO).

The murine macrophage cell line RAW 264.7, HeLa cell line, and 3T6 cell line (from American Type Culture Collection, Rockville, MD) were cultured (37°C, 5% CO₂) in DMEM culture medium (Life Technologies) containing 10% FCS (Life Technologies), 2 mM l-glutamine, and antibiotics. The human monocytic cell line THP-1 was cultured (37°C, 5% CO₂) in RPMI culture medium (Life Technologies) containing 10% FCS, 2 mM l-glutamine, and antibiotics. Cell lines were tested every 2 wk by a PCR-based detection assay for mycoplasma contamination (11). For stimulation experiments, cells were seeded at 10⁶ cells/ml density and then cultivated overnight. Cells were stimulated with either HIM (10⁶ CCU/ml) or LAMP (1 μg/ml) for appropriate time intervals. LPS was used in control experiments at 1 μg/ml to stimulate RAW 264.7 cells and at 100 ng/ml to stimulate THP-1 cells. For phosphotransferase assays, cell lysates were prepared as described elsewhere (8). Briefly, cells were washed twice with ice-cold PBS containing 1 mM Na₃VO₄. For each 10⁶ cells that were initially seeded, 100 μl of the following lysis buffer was added: 20 mM MOPS, pH 7.2, 5 mM EDTA, 1% (w/v) Nonidet P-40, 1 mM DTT, 75 mM β-glycerol phosphate, 1 mM Na₃VO₄, and protease inhibitor mixture (Boehringer Gmbh, Mannheim, Germany). Lysis was performed at 4°C for 20 min with continuous shaking. Cell lysates were centrifuged (10,000 × g for 10 min at 4°C), and supernatant was aliquoted and stored at −80°C. Protein concentration in cell lysates was determined by microBCA assay (Pierce).

In experiments in which chemical inhibitors or anti-CD14 Ab were used, cells were pretreated for 1 h with various concentrations of each item before stimulation.

Monocytes were prepared from PBMC by Ficoll/Hypaque density gradient centrifugation (Pharmacia LKB, Uppsala, Sweden). 12 × 10⁶ PBMC/well were allowed to adhere to six-well tissue culture plates (Costar, Cambridge, MA) for 1 h at 37°C, 10% CO₂, in DMEM containing 1% human serum (CNTS, Saint Antoine, France), 2 mM l-glutamine, and antibiotics. To remove nonadherent cells, wells were washed twice with prewarmed culture medium. Adherent cells (approximately 1.2 × 10⁶ cells/well) were cultured (37°C, 10% CO₂) in 2 ml of culture medium overnight. Human monocytes were stimulated with LAMPf (1 μg/ml) for appropriate time intervals, and cell lysates were prepared for phosphotransferase assays as described above.

RAW 267.4 cells and THP-1 cells were cultivated and stimulated in 24-well tissue culture plates (Costar) as described above. After 24 h of stimulation, cells were lysed by two consecutive cycles of freezing/thawing. Thus, the samples represented the total amount of cytokines produced (both intracellular and those that have been released into the supernatant). The cytokine concentration from murine macrophages was measured using murine IL-1β, IL-6, and TNF-α ELISA from Genzyme (Boston, MA). For cytokine measurement from the human monocytic cell line THP-1, cytokine ELISA kits from R&D Systems (Abingdon, U.K.) were used. The assays were performed according to the manufacturer’s instructions.

For total tyrosine-phosphorylated protein immunoprecipitation, 500 μg of proteins from cell lysates were incubated with 10 μg of agarose-coupled anti-PY 4G10 Ab overnight at 4°C with continuous rotation. The mixtures were then centrifuged (7000 × g for 2 min at room temperature), and agarose beads were washed three times with PBS buffer and resuspended in 20 μl of SDS sample buffer. Following electrophoresis, immunoblot was performed using a second peroxidase-coupled ant-PY RC20 Ab.

ERK1/2 and JNK were immunoprecipitated by incubating 500 μg of cell lysates with 2 μg of anti-JNK1 Ab or 5 μg of anti-ERK1-CT Ab, respectively, at 4°C for 4 h with continuous rotation. Then, 30 μl of protein A-Sepharose was added, and the incubation was extended for 2 more hours. The mixtures were then centrifuged (7000 × g for 2 min at room temperature) and protein A-Sepharose beads were washed three times with buffer B (12.5 mM MOPS, pH 7.2, 0.5 mM EGTA, 12.5 mM β-glycerol phosphate, 7.5 mM MgCl₂, 1 mM DTT, 1% Nonidet P-40) containing 250 mM NaCl. The beads were resuspended either in 50 μl of buffer B containing 10 mM MgCl₂ and 1 mM MnCl₂ for phosphotransferase assays or in 20 μl of SDS sample buffer for electrophoresis and immunoblotting using anti-PY 4G10 Ab. For p38 immunoprecipitation, the protocol was slightly modified. Anti-p38 Ab (10 μg) was first coupled to A/G-Sepharose beads (Santa Cruz Biotechnology) for 2 h at 4°C, then washed with buffer B, and then added to 500 μg of cell lysates. Immunoprecipitation was allowed overnight at 4°C with continuous rotation and immunoprecipitates were analyzed as described above.

SDS-PAGE was performed on 12% separating gels. Samples were boiled for 10 min in the presence of SDS sample buffer. After electrophoresis, proteins were transferred onto a polyvinylidene difluoride membrane (Immobilon-P; Millipore, Molshein, France). Membranes were blocked overnight at 4°C with PBS containing 3% BSA, 1% goat serum, and 0.01% Tween-20. The membranes were washed twice with PBS containing 150 mM NaCl and 0.1% Tween-20, then incubated for 1 h at room temperature with the anti-PY 4G10 Ab or peroxidase-coupled anti-PY RC20 Ab at 1/1500 dilution in PBS containing 0.3% BSA and 0.1% Tween-20. Membranes were washed five times and directly subjected to ECL detection when peroxidase-coupled anti-PY RC20 Ab was used. Otherwise, membranes were incubated with a second Ab, a goat anti-rabbit coupled to peroxidase, washed five times, and then processed for ECL detection as described in the manufacturer’s instructions.

10 μg of MBP, 2 μg of GST-c-Jun, or 200 ng of GST-MAPKAP K2 in the presence 50 μM [γ-³²P]ATP were added to ERK1/2, p38, or SAPK/JNK immunoprecipitates, respectively. The reactions were conducted at 30°C for 30 min duration, then terminated by adding SDS sample buffer to 1× final concentration. Samples were analyzed by SDS-PAGE using 12% gels for GST-MAPKAP K2 and GST-c-Jun and 16% gels for MBP. Gels were fixed in 10% acetic acid and 50% methanol, then embedded in cellophane sheets and dried. Dried gels were first autoradiographed for 3 to 16 h to qualitatively determine ERK1/2, JNK, and p38/Mpk2 activity by visualization of phosphorylated MBP, GST-c-Jun, or GST-MAPKAPK2, respectively, and quantitatively evaluated using a PhosphorImager and ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

ERK and p38 activation was additionally determined by measuring radioactively their respective phosphotransferase activities toward peptide substrate using p42/44^MAPK or p38/MpK detection kits. Assays were performed according to the manufacturer’s instructions. Activity was expressed as [³²P]ATP cpm.

Constructs expressing JNK1 dominant negative mutant (dnJNK-APF) and JNK1 wild-type (JNK-WT) were kindly provided by Dr. B. Dérijard (CNRS, Nice, France) and Dr. R. J. Davis (University of Massachusetts, Boston, MA). RAW 264.7 cells were grown up to 80% confluence, then transfected with the indicated plasmids by the electroporation method described by Stacey et al. (12). After transfection, cells were cultivated for 36 h before subsequent treatment. Cells were stimulated with LAMPf, and cytokine production was measured as described above.

Mycoplasmas, and more precisely their lipid-associated membrane proteins (or LAMPs), have been demonstrated to interact with monocytes and induce proinflammatory cytokine synthesis by these cells. We selected the murine macrophage cell line RAW 264.7 to analyze the biochemical events triggered by mycoplasmas in macrophages, because this cell line have been extensively used to characterize signaling pathways in response to extracellular stimuli such as bacterial LPS (8). First, we assessed the capability of heat-inactivated M. fermentans (HIMf) as well as LAMP extracted from this species (LAMPf) to induce cytokine production by RAW 264.7 cells. Cells treated with either HIMf (10⁶ CCU/ml) or LAMPf (1 μg/ml) for an 18-h period produced a significant amount of IL-1β, IL-6, and TNF-α (Fig. 1,A). As with other monocytic cell lines, IL-1β was found to be almost exclusively intracellular in RAW 264.7 (4, 13). M. pneumoniae-derived LAMP preparations (LAMPp) (Fig. 1,A) as well as heat-inactivated M. pneumoniae (data not shown) failed to induce cytokine production by murine macrophages, thus indicating that this biologic activity is not a general rule for any mycoplasma species. Cytokine production by LAMPf-treated murine macrophages was found to be dose dependent in a concentration range from 10 ng/ml to 1 μg/ml of lipoproteins (Fig. 1 B). At concentrations higher than 1 μg/ml, the stimulation effect declined. Similar results were obtained with the human monocytic cell line THP-1 (data not shown; (14)).

Contamination with endotoxin is not responsible for the cytokine induction described above. Apart from the fact that our preparations were systematically tested for the absence of endotoxin (see Material and Methods), the pretreatment of LAMPf preparations with polymyxin B does not affect cytokine-induced production, while it abolished LPS stimulation (see Fig. 2,A for TNF-α, and data not shown). Additionally, the anti-CD14 mAb MY4, which inhibited LPS-induced cytokine production by human monocytes, had no effect on cytokine production by LAMPf-stimulated human monocytic THP-1 cells (Fig. 2 B).

PTK activation is one of the earliest responses of monocytes to LAMPf (4); we therefore examined phosphotyrosine-containing proteins in RAW 264.7-LAMPf-stimulated cells. Cell lysates were immunoprecipitated with an agarose-coupled anti-PY 4G10 mAb, electrophoresed on SDS-PAGE, and immunoblotted using a second anti-PY RC20 mAb. As shown in Figure 3,A, LAMPf enhanced tyrosine phosphorylation of several proteins, with the maximal effect occurring at 30 min after challenging the cells. PTK activation plays a crucial role in cytokine induction by mycoplasmas because genistein, a well-known inhibitor of PTKs, completely blocked cytokine synthesis induced in RAW 264.7 cells by LAMPf (Fig. 3 B).

It has recently been demonstrated that LPS stimulation of murine macrophages leads to increased tyrosine phosphorylation and activation of the three subgroups of MAPK identified in mammals: ERK, JNK, and p38 (8). Since both LAMPf and LPS have been demonstrated to induce the synthesis of proinflammatory cytokines by monocytic cells, we examined whether LAMPf was capable of activating the MAPK pathways. RAW 264.7 cells were treated with LAMPf for different time intervals and cell lysates tested for ERK1/2, SAPK/JNK, and p38 activation by measuring their respective phosphotransferase activities toward different substrates. Activation of ERK1/2 was first determined by measuring the phosphorylation of MBP after the immunoprecipitation of lysates with an anti-ERK Ab. This assay allowed the detection of a significant phosphotransferase activity in lysates from LAMPf (1 μg/ml)-treated macrophages (Fig. 4,A). Similar results were obtained when an ERK1/2-specific peptide was used as a substrate to determine the phosphotransferase activity in total cell lysates (Fig. 4,B). In both assays maximal activity was detected at 30 min of stimulation. At 60 min, this activity declined to 60% of the maximal value. Substrate phosphorylation paralleled an increase in both ERK1 and ERK2 tyrosine phosphorylation. As shown in Figure 4 C, immunoprecipitation with an ERK1/2 polyclonal Ab and Western blotting with an anti-PY 4G10 mAb allowed the detection of two bands at a molecular mass of 44 and 42 kDa.

p38 phosphotransferase activity was measured after immunoprecipitation with an anti-p38 Ab using a kinase assay with either GST-MAKAPK2 (Fig. 5,A) or a p38-specific peptide as substrate (Fig. 5,B). In both assays, significant p38 phosphotransferase activity was found in cell lysates from LAMPf-treated RAW 264.7 macrophages compared with untreated cells. p38 activation in response to LAMPf peaked at 30 min, but the kinetics differed from ERK1/2 activation: 1 h after LAMPf challenge, the p38 phosphotransferase activity dramatically dropped to 10% of maximal activity. p38 was demonstrated to undergo tyrosine phosphorylation by immunoprecipitation with a specific p38 Ab followed by immunoblotting with the anti-PY 4G10 mAb (Fig. 5 C).

Activity of immunoprecipitated JNK was assayed using its physiologic substrate, c-Jun. As shown in Figure 6, stimulation of RAW 264.7 cells with LAMPf strongly induced JNK activity. Once again, the maximal activation occurred at 30 min after stimulation.

We previously demonstrated that LAMPf were capable of inducing PTK activation in human monocytes/macrophages and in the THP-1 cell line (4). Thus, it was particularly interesting to determine whether ERK, p38, and JNK activation could be observed in human monocytes. We therefore challenged human monocytes from healthy donors with LAMPf for different time intervals and then measured the phosphotransferase activity of the herein-studied MAPK pathways. Results displayed in Figure 7 show that M. fermentans lipoproteins induce the activation of ERK (Fig. 7,A), p38 (Fig. 7,B), and JNK (Fig. 7 C) in these cells. The time course activation of MAPK pathways observed in human monocytes was comparable with that observed using the murine macrophage cell line RAW 264.7. The same results were obtained using elutriated human monocytes from healthy donors (data not shown).

These results show that membrane lipoproteins from M. fermentans are capable of activating the three major MAPK pathways described in mammals. As shown above, M. pneumoniae preparations are not capable of stimulating cytokine production in murine macrophages (Fig. 1 A). Interestingly, LAMPp also did not induce any MAPK activation in these cells (data not shown). We did not observe any differences in MAPK activation when LAMPf preparations were preincubated with polymyxin B (1000 U/ml) for 1 h before cell stimulation (data not shown), thus indicating that the observed effect does not account for endotoxin contamination.

A serum protein named LBP (i.e., LPS-binding protein) has been identified as a major mediator of LPS effects (15, 16). We therefore investigated the requirement of serum to mediate the effect of LAMPf on murine macrophages. While the absence of serum completely abolished the activation of MAPK pathways in LPS-treated RAW 264.7 cells, it did not significantly affect the MAPK activation induced by LAMPf treatment (Fig. 8,A). Interestingly, the presence of serum is not absolutely required for the induction of cytokines by mycoplasmas. As depicted in Figure 8 B, 18-h serum starvation slightly reduced the cytokine production by mycoplasma-stimulated macrophages, whereas it prevented cytokine induction by LPS.

We additionally tested whether MAPK activation by LAMPf could be triggered in cell types other than monocytes. The treatment of HeLa cells, a human epithelioid cervix carcinoma, with LAMPf did not induce any activation of MAPK pathways (Fig. 9). Control treatment with PMA (50 ng/ml) showed an increased MAPK and p38 phosphotransferase activity in a time interval as short as 10 min after challenge. The same results were obtained with the murine fibroblast cell line 3T6 (data not shown). This suggests that M. fermentans triggers MAPK activation in a cell-specific manner.

To examine the involvement of MAPKs in mycoplasma-mediated cytokine production by murine macrophages, we specifically blocked each of the three MAPK pathways and monitored the cytokine production when RAW 264.7 cells were challenged with LAMPf.

PD-98059 is a synthetic compound that specifically inhibits the ERK-activating MAPK kinase MEK-1 (17, 18). This compound selectively inhibited the LAMPf-mediated activation of ERK1/2 in murine macrophages without significantly affecting the stimulation of p38 or JNK (data not shown). As shown in Figure 10 A, PD-98059 decreased TNF-α and IL-1β production by LAMPf-stimulated RAW 264.7 cells in a dose-dependent manner. TNF-α production was completely blocked at 10 μM of PD-98059, and IL-1β production was inhibited up to 80% at 30 μM. On the contrary, only 20% IL-6 inhibition could be observed at the highest dose used (30 μM).

SB203580 is a bicyclic imidazole compound able to specifically inhibit p38 (19). This compound selectively inhibited the LAMPf-mediated activation of p38 in RAW 264.7 cells without significantly affecting the stimulation of ERK1/2 or JNK (data not shown). As depicted in Figure 10,B, pretreatment of cells with 1 μM of SB203580 completely abrogates mycoplasma-mediated IL-1β and TNF-α production. A higher dose of this compound was necessary to completely block the IL-6 production (Fig. 10 B).

To investigate the involvement of the JNK pathway, RAW 264.7 cells were transiently transfected with a dominant negative mutant of JNK1, dnJNK-APF, in which the phosphorylation residues Thr¹⁸³ and Tyr¹⁸⁵ were changed to Ala and Phe, respectively (20). The expression of dnJNK-APF has been shown to efficiently block JNK1 phosphorylation in response to extracellular stimuli such as UV irradiation (20, 21). In our hands, dnJNK-APF selectively affected the LAMPf-mediated activation of JNK without changing ERK1/2 or p38 activities (data not shown). The transfection of RAW 264.7 cells with JNK1-APF significantly reduced LAMPf-induced IL-6 secretion as compared with cells transfected with the wild-type JNK1 DNA (JNK-WT), whereas no changes were detected in the IL-1β or TNF-α levels (Fig. 11).

These results demonstrate that MAPK pathways are involved in M. fermentans-mediated cytokine induction. p38 plays an essential role in the LAMPf-triggered signaling leading to the synthesis of the three studied cytokines. Our data strongly suggest that the ERK pathway is involved in IL-1β and TNF-α production induced by LAMPf, while the JNK pathway is most likely implicated in IL-6 production only.

Although the molecular basis of mycoplasma pathogenicity remains unclear, host immune reaction appears to play a major role in the development of mycoplasma infections. Actually, mycoplasmas have been shown to trigger a wide range of immunomodulatory effects, and they exert a profound stimulatory action on cells of the immune system. One of the best documented effect of mycoplasmas is the induction of numerous cytokines by monocytic cells. We have previously demonstrated that the ability of mycoplasmas to stimulate monocytes/macrophages to secrete cytokines resides in the membrane lipoprotein fraction and that PTK activation is absolutely required for this biologic effect (4, 5). In the present study, we demonstrate that lipid-associated membrane proteins derived from the human M. fermentans, or LAMPf, are capable of activating the MAPK family members ERK1/2, p38 and JNK in the murine macrophage cell line RAW 264.7 and in human monocytes/macrophages.

This capacity of M. fermentans to stimulate MAPK pathways is not shared by any species of mycoplasma because our findings indicate that M. pneumoniae-derived LAMP failed to induce the activation of MAPKs. Interestingly, this species is not capable of inducing the production of cytokines by monocytic cells (Fig. 1 A; (4)). However, M. pneumoniae has been shown to stimulate cytokines production by total blood cells (22), suggesting that this pathogen might induce the cytokine production by stimulating a cell population other than one of monocyte/macrophage lineage. Recent findings indicate that M. pneumoniae interacts with CD4 T cells to induce proinflammatory cytokines (23), supporting the foregoing hypothesis. It could be speculated that the failure of M. pneumoniae to stimulate macrophages is part of a survival strategy that this pathogen employs.

LPS is a potent stimulator of the MAPK pathways in monocytes and macrophages and a very efficient inducer of proinflammatory cytokines in these cells (8). Classical Gram negative bacterial LPS is absent from mycoplasmas, but membranes of these microorganisms are rich in lipoproteins; therefore, it is interesting to compare the effects of LPS and LAMPs on monocytic cell lines. The first clear difference between LAMPf and LPS is the quite dissimilar kinetics of MAPK activation. MAPKs peak activity induced in RAW 264.7 cells in response to LPS have been observed between 10 and 15 min after stimulation (data not shown; (8)), while LAMPf induced MAPK’s maximal activity 30 min after challenge.

A second major difference consists in the lack of serum requirement for LAMPf-mediated MAPK activation. Although the absence of serum did not affect ERK1/2, p38, or JNK activation, it slightly reduced LAMPf-induced cytokine production, suggesting that pathways other than MAPKs could contribute to the signaling. The possible involvement of a Ca²⁺-dependent biochemical pathway, but not PKC, in mycoplasma-induced TNF-α by THP-1 cells has been pointed out (24), suggesting that other secondary messengers, such as phospholipase C, might also contribute to this effect. This issue should be addressed in further studies. LPS-induced cell activation involves the membrane receptor CD14, which recognizes LPS complexed to the serum protein LBP (LPS-binding protein). This could explain the inability of LPS to activate MAPK pathways and stimulate cytokine synthesis after a 12-h serum starvation. Interestingly, an anti-CD14 human mAb, MY4, able to block LPS activation, did not effectively reduce cytokine levels produced by LAMPf-stimulated human monocytic cell line THP-1. From our data, it can be hypothesized that M. fermentans membrane lipoproteins may recognize specific membrane receptor(s) present in the monocyte/macrophage lineage and most probably absent in epithelial cells and fibroblasts, since this type of cell does not respond to mycoplasma challenge by activating the MAPK pathways. The putative receptor(s) might be coupled, directly or indirectly, to PTKs that subsequently trigger the MAPKs activation.

Presently, it is not clearly demonstrated whether lipid fraction of LAMPf or protein is involved in monocytes/macrophages activation. Based on proteinase K digestions, several reports have suggested that lipid fractions is the active component (4, 14), but the possibility that activity resides in small lipopeptides or small peptides resulting from this treatment cannot be ruled out. Hall et al. (25) have recently characterized a new lipoprotein gene (p48) from M. fermentans. When p48 gene was expressed as a MBP-P48 fusion protein in E. coli, the fusion protein was shown able to activate the HL-60 cell line, although it is not anchored to lipids. Furthermore, Muhlradt et al. (26) have also reported that the N-terminal peptide (the first 14 amino acids) of P48 coupled to fatty acid is also capable of stimulating monocytes. Lipid-uncoupled N-terminal peptide from P48 has not been tested for its ability to interact with monocytes/macrophages. We are currently investigating this issue.

The importance of the different intracellular signaling routes for cytokine production has not been addressed extensively. In response to extracellular stimuli, activated MAPKs can activate a number of substrates, including transcription factors, and control the synthesis of cytokines. In this way, the LPS-mediated induction of IL-1β and TNF-α has been demonstrated to be controlled by p38 (19). Recently, it has been shown that the activation of ERK1 is necessary for FcγR cross-linking-induced TNF-α synthesis (27, 28). Our study points out the importance of MAPK pathways in mycoplasma-mediated cytokine production by monocytic cells. By using specific inhibitors of MEK-1 (the kinase upstream ERK1/2) and p38 or a JNK dominant negative construct, we could delineate the contribution of each of these signaling pathways in the cytokine induction by LAMPf. Results obtained with the specific inhibitor of p38 SB203580 clearly show that the p38 pathway is crucial for IL-1β, TNF-α, and IL-6 production by LAMPf-stimulated RAW 264.7 cells. MEK-1 inhibitor PD-98059 blocked LAMPf-induced IL-1β and TNF-α but not IL-6, while the JNK dominant negative mutant inhibited only IL-6. Our results suggest that at least two MAPK-derived signals are required for the synthesis of a given cytokine in response to mycoplasmas. Whereas p38 appears to trigger a signal common to the three studied proinflammatory cytokines, ERK and JNK differentially contribute to the synthesis of IL-1β, TNF-α, and IL-6.

In summary, we have demonstrated that M. fermentans membrane lipoproteins activate in macrophages the three MAPK identified in mammals, and we shown the importance of these pathways in the induction of proinflammatory cytokines by these microorganisms. Further studies to characterize molecules involved in the interaction between mycoplasmas and monocytic cells should improve the understanding of the immunomodulatory activity and pathogenicity of these infectious agents.

The authors thank Dr. B. Dérijard (CNRS, Nice, France) and Dr. R. Davis (University of Massachusetts Medical School, Boston, MA) for providing pCDNA3-JNK1 and pCDNA3-JNK1(APF) constructs, respectively. The assistance of Dr. C. Marie (Pasteur Institute, Paris, France) is gratefully acknowledged. We are indebted to A. Dujeancourt and C. Prevost for helpful technical assistance.