Macrophage-Colony-Stimulating Factor-Induced Proliferation and Lipopolysaccharide-Dependent Activation of Macrophages Requires Raf-1 Phosphorylation to Induce Mitogen Kinase Phosphatase-1 Expression¹

Macrophages serve essential roles as regulators of immunity. A number of stimuli regulate macrophage gene expression and determine their final biological function. In response to growth factors, such as M-CSF, macrophages differentiate and proliferate (1). In contrast, activation by LPS, a component of Gram-negative bacteria, makes these phagocytic cells stop their proliferating program and acquire effector functions such as the production of NO and proinflammatory cytokines, including TNF-α, IL-1, and IL-6 (2, 3, 4).

Although macrophages either proliferate or become activated, in both cases activation of ERK-1 and -2 is required (5). These are evolutionarily conserved proline-directed serine/threonine protein kinases, also known as p44- and p42-MAPKs, responsible for phosphorylating several transcription factors such as Elk-1, Fos, Jun, and c-Myc family members (6, 7). Interestingly, subtle differences in the time course and strength of ERK activation as well as its subcellular localization appear to specify differential downstream signaling events. In particular, the duration of ERK signaling is under tight control by positive and negative regulators and changes in the activity of these kinases have been associated with specific cell fates in a number of cell types (8). In our cellular model, an early peak of ERK activation (5 min) correlated with cellular proliferation whereas a later peak (15 min) was associated with the activation program (5). Among the regulatory mechanisms that affect ERK phosphorylation, dual specificity mitogen kinase phosphatases (MKP⁴ or DUSP) dephosphorylate tyrosine and threonine residues, which results in ERK inactivation (9, 10). In vitro studies and transfection experiments showed that MKP-1 expression in response to mitogens dephosphorylates mainly ERK-1/2 and has lower activity toward other MAP kinases such as JNK and p38 (9). In bone marrow-derived macrophages, MKP-1 is a key regulator of the time course of ERK activity and inactivation of ERK-1/2 tightly follows the induction of this phosphatase (5). However, in contrast to what has been reported in other cell types (11), MKP-1 induction by M-CSF and LPS in these cells is independent of ERK activation (12, 13). Interestingly, a number of conditions that inhibit MKP-1 expression in macrophages, for example extracellular matrix proteins such as decorin or fibrinogen, or treatment with cyclosporine A or FK506, elongate ERK activity and reduce proliferation (14, 15).

One of the early players in intracellular signaling is the serine/threonine kinase Raf, which regulates a broad range of functions including cellular growth, proliferation, activation, differentiation and apoptosis (11, 16). In many cell types, activated Raf kinases phosphorylate and stimulate the mitogen-activated ERK kinase (MEK), the immediate upstream regulator of the ERK module (17). Several Raf isoforms in mammalian cells have been identified; c-Raf/Raf-1, A-Raf, and B-Raf (18). A and B isoforms are predominantly expressed in neuronal and urogenital tissues, whereas Raf-1 is expressed ubiquitously (19). Raf activity can be regulated by direct phosphorylation by other kinases, including members of the serine-threonine protein kinase C (PKC) family, Src and AKT/PKB (20, 21, 22) and by complex formation with 14-3-3 proteins (23) and Hsp90 chaperone (24).

The exact effects of Raf-1 activation on macrophages are controversial. Here we studied the role of Raf-1 in MKP-1 expression and ERK activation in primary macrophages. We show that Raf-1 activation is required for MKP-1 induction in macrophages stimulated with M-CSF or LPS. In both conditions, we observed that Raf-1 binds to and regulates PKCε activity. Distinct kinetics of Raf-1 activation were detected in response to the two stimuli. Although these time courses correlated with those of activated ERK-1/2, Raf-1 was involved in ERK activation in response to M-CSF only. During LPS signaling, an alternative pathway directed ERK activation in the absence of active Raf-1. In correlation with these effects, inhibition of Raf-1 activity compromised the progression of macrophages through the cell cycle whereas no effect was observed in the expression of proinflammatory cytokines or inducible NO synthase (iNOS) upon LPS stimulation.

LPS, actinomycin D, propidium iodide, and Abs anti-ERK^P Thr¹⁸³/Tyr¹⁸⁵, and anti-β-actin were purchased from Sigma-Aldrich. Recombinant M-CSF (1200 U/ml ≅ 10 ng/ml) and IFN-γ were obtained from R&D Systems. Sodium arsenite (SA) was obtained from Wako Pure Chemicals. The p38^P (Thr¹⁸⁰/Tyr¹⁸²) MAP kinase Ab was purchased from Cell Signaling Technology. Monoclonal anti-Raf-1^P Ser³³⁸ and polyclonal anti-Raf-1 Abs were obtained from Upstate Biotechnology and Biosource, respectively. Anti-PKCε and anti-phosphotyrosine Abs were from BD Pharmingen. Abs against ERK-1/2 or MKP-1 were purchased from Santa Cruz Biotechnology. ZM 336372, geldanamycin, and Abs anti-NOS2 and anti-MEK^P Ser²¹⁸/Ser²²² (17) were obtained from Calbiochem.

Bone marrow-derived macrophages were obtained from BALB/c mice (Harlan Ibérica) and cultured as previously described (25). Animal use was approved by the Animal Research Committee of the University of Barcelona (2523). To render cells quiescent, they were deprived of M-CSF for 18 h before stimulation (25).

Cell death was assessed by Annexin V^FITC (Bender MedSystems) using FACS analysis (Coulter Multisizer II) according to the manufacturer’s instructions. These studies were confirmed by trypan blue exclusion. Each point was performed in triplicate and the results were expressed as the mean ± SD.

The proliferative capacity of macrophages was assessed as previously described (26). A 6-h pulse of [³H]thymidine (1 μCi/ml) (Amersham Biosciences) was applied to each sample. Each point was performed in triplicate, and the results were expressed as the mean ± SD.

Total cytoplasmic extracts were obtained and Western blotting was performed as previously described (27). Detection was conducted using EZ-ECL kit (Biological Industries). β-Actin expression was measured as a control for differences in loading and transfer. Figures are representative of at least three independent experiments.

JNK activity was measured as previously described (28). Briefly, cell lysates were immunoprecipitated for JNK1. The reaction was performed with 1 μg of GST-c-jun (MBL) as JNK substrate, 20 μM ATP, and 1 μCi [γ-³²P]ATP.

PKC assay was performed as previously described (29). Briefly, total cellular lysates were obtained and immunoprecipitated for PKCε. After washing, kinase assay was performed using 1 μg of histone H1 (Santa Cruz Biotechnology) as exogenous substrate, 20 μM ATP, and 1 μCi of [γ-³²P]ATP.

Total RNA was extracted with the RNA Kit EZ-RNA (Biological Industries), separated in agarose gel and transferred to nitrocellulose membrane (Amersham Biosciences). Probes were labeled with [α-³²P]dCTP (Amersham Biosciences) using the random prime labeling system (Amersham Biosciences). 18S was used as a loading and transfer control. Results are representative of three independent experiments.

cDNA was obtained using M-MLV Reverse Transcriptase (Promega) as described (30). Primer Express software (Applied Biosystems) was used to design primer sequences. Real Time-PCR was conducted with 1× SYBR Green PCR master mix using the ABI Prism 7900 detection system (Applied Biosystems). Each sample was analyzed in triplicate. Expression levels were normalized to β-actin. Relative values from a representative experiment out of three independent experiments are represented in each graphic.

Cell cycle was analyzed as described (27). Briefly, 10⁶ treated cells were fixed and incubated in PBS, 0.2% Triton X-100, 10 μg/ml RNase A, 5 mg/ml propidium iodide, and analyzed by FACS. Cell cycle distributions were analyzed with the Multicycle program (Phoenix Flow Systems). Experiments were performed in triplicate.

Raf activity assay was performed as previously described (1). The assay is based on measurement of Raf-1-dependent phosphotransferase activity in a kinase reaction using recombinant MEK1, inactive as a Raf-1 substrate (Upstate Biotechnology). Briefly, cell lysates were immunoprecipitated with total Raf-1 Ab (Biosource). After several washes, the reaction was conducted with 1 μg of inactive MEK, 2 mCi of [γ ³²-P]ATP, and 500 μM cold ATP at 30°C for 30 min.

ERK activity was analyzed as previously described (5) using myelin basic protein (Sigma-Aldrich) as a substrate, copolymerized in the gel. Results were representative of three independent assays.

siRNA were obtained from Dharmacon and transfected by electroporation (31). Then 4 × 10⁶ cells and 1.5 μM siRNA were resuspended in 400 μl and pulsed once at 350 V, 2300 μF with a BTX ECM 600 electroporator (BTX). The siRNA sequences were AGUCAAAGAAGAGAGACCU for siRNA1 and UUCCAGAUGUUCCAGCUAA for siRNA2. siRNAu was directed against calcineurin and the sequence was AACCUCGUGUGGAUAUCUU (15).

Although resulting in two opposing cellular functions, the macrophage growth factor M-CSF and the activating bacterial compound LPS induce the expression of the phosphatase MKP-1 in macrophages. This phosphatase has gained interest as a key regulator of macrophage function by virtue of its capacity to control the duration of MAPK activity (5, 12, 13). The involvement of Raf-1 in macrophage signal transduction pathways is unclear. There is evidence that proliferating and activating stimuli also result in different Raf-1 activation requirements in macrophagic cell lines (32, 33, 34). Here we studied whether Raf-1 activity is required for the expression of MKP-1. Since primary cultures of bone marrow-derived macrophages are scarcely transfectable with plasmids (31), we were unable to use mutants or activated constructs of Raf-1 proteins. We therefore used chemical inhibitors of Raf-1, assuming that they are not always completely specific. It has been reported that fully induced Raf-1 activity requires both Ser³³⁸ and Tyr³⁴¹ phosphorylation (20, 35, 36), which can be blocked by the synthetic compound ZM 336372 (Fig. 1, A and B). Furthermore, Raf-1 activity was determined by a link MEK kinase assay measuring MEK phosphorylation after in vitro reaction (Fig. 1,A). To study the effect of Raf-1 activity on MKP-1 expression, we performed both Northern and Western blot analyses. We had previously characterized the time course of induction of this phosphatase by M-CSF and LPS; maximal mRNA induction occurred at 30 and 45 min of treatment, respectively, whereas the protein levels peaked 15 min later for each case (12, 13). ZM 336372 abolished the induction of MKP-1 mRNA and protein by M-CSF and LPS (Fig. 1, C and D). These effects were not caused by toxicity induced by ZM 336372, as shown in Fig. 5 B.

To corroborate the results obtained with the Raf-1 inhibitor, we selectively abrogated Raf-1 protein expression using siRNA technology. The efficiency of this process was tested using real time-PCR and Western blot after 48 h of electroporation. Two independent siRNAs against Raf-1 were similarly efficient at depleting Raf-1 (Fig. 1,E). As controls, we used a mock transfection and an siRNA directed to an unrelated gene, calcineurin (siRNAu) (15). To exclude toxic effects of siRNA transfection, we determined the levels of apoptosis. No increased cell death was associated with any of the conditions tested compared with actinomycin D used as a positive control (Fig. 1,F). Depletion of Raf-1 resulted in complete inhibition of MKP-1 expression even in the presence of M-CSF or LPS (Fig. 1 G). Taken together, these results indicate that Raf-1 activity is required for MKP-1 expression during M-CSF-induced proliferation and LPS activation.

We previously reported that PKCε is a key upstream regulator of MKP-1 expression in primary macrophages (12, 13). Our next interest was to examine whether cross-talk occurs between Raf-1 and PKCε during signal transduction to M-CSF or LPS. PKCε was immunoprecipitated from equivalent amounts of cellular lysates and the samples were subsequently immunoblotted with an anti-phosphoserine Raf-1 Ab, which is a good indicator of the active state of Raf-1. Phosphorylated Raf-1 coimmunoprecipitated with PKCε in response to M-CSF (Fig. 2,A). This complex was detected only from 4 to 8 min after the start of M-CSF stimulation, which indicates that the kinetics of this interaction is very tightly controlled by positive and negative regulators. The phosphorylated Raf-1-PKCε complex was also detected during the response to LPS, although its kinetics was slightly delayed; binding was detected 8 to 10 min after the start of treatment (Fig. 2 B).

To further examine whether Raf-1 regulates PKCε activity in macrophages, we performed immune complex kinase assays for PKCε. In this case, PKCε was immunoprecipitated from total cell lysates and its activity was measured in vitro using histone H1 as a substrate. Both M-CSF and LPS stimulation increased PKCε activity, which was abolished in the presence of the Raf-1 inhibitor ZM 336372 (Fig. 2, C and D). These observations indicate that Raf-1 activity is required for PKCε activation during M-CSF and LPS signaling. Moreover, in correlation with the time course of Raf-1 phosphorylation (Fig. 2, A and B), activation of PKCε by M-CSF and LPS followed distinct kinetics. Maximal levels of active PKCε were detected at 5–10 min of M-CSF stimulation and at 15–30 min of LPS treatment.

Interestingly, although Raf-1 phosphorylation was necessary for PKCε activation as assessed by ZM 336372, time course experiments indicate that Raf-1 was required only for triggering the initial activation of PKCε, since it remained active after Raf-1 dephosphorylation.

Our results in Fig. 2 indicated that M-CSF and LPS stimulation resulted in differential patterns of Raf-1 activation. We already reported subtle changes in the activation pattern of ERK-1/2 following stimulation with M-CSF or LPS (5). Here we further characterized overall changes in the Raf-1-MEK-ERK cascade in response to the proliferative or the activation signal (Fig. 3). Raf-1 was phosphorylated in both serine and tyrosine residues within 3–5 min of M-CSF stimulation (Fig. 3, A and B), which was followed by fast and transient phosphorylation of MEK-1/2 (Fig. 3,C), and subsequent activation of ERK-1/2 (Fig. 3,D and Ref. 5). As described previously, the time course of ERK-1/2 activation by LPS was more delayed than that induced by M-CSF (Fig. 3,D and Ref. 5), which correlated with the delayed kinetics of Raf-1 and MEK-1/2 phosphorylation (Fig. 3, A–C).

We next examined whether Raf-1 activity was indeed required for ERK-1/2 activation. Total cell lysates from M-CSF-stimulated macrophages were immunoblotted with an Ab that recognizes Thr¹⁸³ and Tyr¹⁸⁵ phosphorylation of ERK-1/2. Activation of ERK-1/2 by M-CSF was blocked by the Raf-1 inhibitor ZM 336372 (Fig. 4,A). Gel kinase assays using myelin basic protein as a substrate confirmed the complete inhibition of ERK-1/2 activity by ZM 336372 (Fig. 4,B). Two nonrelated inhibitors of Raf-1 function, namely SA and geldanamycin, a disruptor of Raf-1 interaction with hsp90 chaperone (37), had similar effects on ERK-1/2 phosphorylation by M-CSF (Fig. 4, C and D). Likewise, the use of two independent Raf-1 siRNAs inhibited the phosphorylation of ERK-1/2 in response to M-CSF (Fig. 4,G). Taken together, these data indicate that Raf-1 activity is required for ERK-1/2 activation by M-CSF. In contrast, inhibition of Raf-1 with ZM 336372 did not reduce ERK phosphorylation or activation (Fig. 4, E and F) after stimulation with LPS. Identical results were obtained using SA (data not shown) and siRNA technology (Fig. 4,H). These results indicate that LPS signaling includes alternative pathways that can direct ERK phosphorylation and activation in the absence of active Raf-1. Interestingly, Raf-1 inhibition in LPS-treated macrophages appeared to extend the period of ERK-1/2 activation, which can be explained by the lack of expression of MKP-1, responsible for dephosphorylating several members of the MAPK family (Fig. 1 D). Elongation of JNK and p38 MAPK activation was also observed under these conditions (data not shown).

We next determined the role of Raf-1 in macrophage proliferation by measuring [³H]thymidine incorporation as an indicator of DNA synthesis (Fig. 5,A). ZM 336372 blocked macrophage proliferation in a dose-dependent manner, with maximal inhibition at 30 μM. To exclude toxic effects of this drug, we measured the induction of apoptosis by annexin V staining. Anti-proliferative doses of ZM 336372 did not increase macrophage cell death in the presence (Fig. 5 B) or absence of M-CSF (data not shown). The lack of cellular toxicity was corroborated by trypan blue exclusion (data not shown).

In addition, SA also inhibited M-CSF-dependent proliferation (Fig. 5,C) without causing toxic side effects (data not shown). These results were corroborated using siRNA technology. Macrophages transfected with two independent siRNAs directed against Raf-1 showed decreased DNA synthesis in response to M-CSF, compared with cells transfected with a control siRNA (Fig. 5 D).

To further characterize the role of Raf-1 in the control of macrophage cell cycle, we measured the cellular DNA content by propidium iodide staining. Gated viable cells were analyzed for their DNA content and the percentage of cells in the S phase of the cycle is shown in Fig. 5,E. In the absence of growth factors, the cells were arrested at G₀-G₁ phase and very few cells progressed through the S phase (Fig. 5,E and Ref. 27). Upon M-CSF stimulation, 30% of the cells re-entered the cell cycle and were progressing through the S phase 30 h after the start of the stimulation. As a positive control of G₁-S blockage, we used IFN-γ, as previously described (27). Inhibition of Raf-1 activity with ZM 336372 decreased the numbers of cells entering the S phase compared with cells stimulated with the growth factor. No cycling activity was observed when the inhibitor was used alone. In conclusion, inhibition of Raf-1 decreased the entry of cells into the S phase of the cell cycle, which correlates with reduced [³H]thymidine incorporation (Fig. 5 A).

We also studied the effect of ZM 336372 on several regulators of the cell cycle. Progression throughout the G₁ and S phases of the cell cycle is controlled by the sequential activation of a number of cyclin D/cdk complexes (38). We thus performed real time-PCR assays to monitor the expression of various cyclins. M-CSF-stimulated cells enter the cell cycle by increasing the expression of cyclin D1, D2, and D3 (38). However, inhibition of Raf-1 by ZM 336372 did not modify the mRNA levels of any of these cyclins (Fig. 5,F; data not shown). Because cyclin-Cdk activity is also regulated by Cdk inhibitory proteins such as p21^Waf-1 and p27^Kip-1 (39, 40), we measured their mRNA levels after ZM 336372 treatment. Raf-1 inhibition increased the expression of p21^Waf-1 and p27^Kip-1 (Fig. 5, G and H) at different time points after M-CSF stimulation, indicating that these two regulators may mediate the blockage in S phase entry caused by ZM 336372.

Northern blot assays were used to measure mRNA expression of proinflammatory cytokines in response to LPS. Raf-1 inhibition did not compromise the induction of IL-1β, TNF-α, or IL-6 in macrophages treated with subsaturant concentrations of LPS. Interestingly, this inhibition enhanced their induction (Fig. 6,A), which could be related to extended MAPK activation in response to LPS (Fig. 4,F). Raf-1 had no effect on LPS-induced NOS2 expression in bone marrow-derived macrophages (Fig. 6, B and C). Taken together, these results indicate that LPS activates signaling events that are capable of bypassing Raf-1 activation and still inducing ERK-1/2 activity and proinflammatory cytokine production.

Here we report on differential requirements of Raf-1 for macrophage function. During macrophage signal transduction to M-CSF, Raf-1 plays two crucial roles that result in tight regulation of the pattern of ERK activation. First, Raf-1 activity accounts for positive regulation of the MEK-ERK module, which is later required for macrophage proliferation. Our data confirm previous reports that demonstrated that Raf-1 is critical for activation of ERK-1/2 during M-CSF stimulation (41, 42). Second, Raf-1 is involved in expression of the phosphatase MKP-1, a key negative regulator of MAPK activity (43).

Therefore, the overall duration and strength of M-CSF-induced ERK activity falls under the control of Raf-1. During the activation of bone marrow-derived macrophages by LPS, Raf-1 activity is dispensable for ERK phosphorylation, because activation of the ERK module can proceed in the absence of fully active Raf-1. Other macrophagic cells, including the RAW 264.7 cell line and primary alveolar macrophages, phosphorylate ERK during LPS stimulation without activation of Raf-1 (33, 44), which indicates that LPS generally bypasses Raf-1 activity to activate the ERK cascade in macrophages. Although MEK-1/2 are the major substrates of Raf proteins, the presence of an alternative upstream pathway that activates MEK-1/2 has been proposed and includes the serine/threonine kinase Cot/Tpl2 (45, 46), the proto-oncogene Mos kinase (47), and a 73-kDa MEKK in EGF-treated PC-12 cells (48). In correlation with these observations, our data indicate that proinflammatory cytokine production and expression of iNOS in response to LPS do not require Raf-1 activity either. Moreover, increased cytokine production was observed in the absence of MKP-1 and Raf-1 inhibition. This may be because the inhibition of MKP-1 extended MAPK activation. In fact, it has been reported that an increase of MKP-1 limits cytokine production (2, 49).

In contrast, MKP-1 induction by LPS is a Raf-1-dependent event, which indicates that Raf-1 is required for the control of the duration of MAPK activity during signal transduction to LPS. Previous reports showed that ectopic expression of v-raf in macrophages also induces MKP-1 (34). The contribution of MKP-1 phosphatase to deactivation of ERK, JNK, and p38 kinases has been widely reported (43, 52). MKP-1 was the first phosphatase identified as an in vitro ERK-specific phosphatase (53, 54, 55). Further studies reveal that MKP-1 was ubiquitously induced by a wide variety of stimuli such as mitogens, hormones, oxidative-DNA damage cellular stresses, LPS, proinflammatory cytokines, and anti-inflammatory agents (53, 55, 56, 57, 58). In addition, depending on the circumstances, MKP-1 dephosphorylates JNK and p38 (2, 49, 59). Discrepancies in the role of MKP-1 are related to the stimuli, the cell type or the kinetics of induction. In our model, we observed that MKP-1 inhibition was correlated with elongated MAPK activation, similar to that described (62). Therefore, in several cells MKP-1 is a crucial regulator of many functions, including proliferation, differentiation, activation or apoptosis (2, 63, 64, 65, 66). Furthermore, MKP-1 has been described as a novel target since it is involved in human tumors (67, 68, 69).

The mechanism used by Raf-1 to induce MKP-1 expression in macrophages probably involves the activation of PKCε. In this report, we demonstrate that Raf-1 interacts and mediates the activation of PKCε in response to M-CSF and to LPS. In our previous work, we established PKCε as a key regulator of MKP-1 expression in macrophages (12, 70). Indeed, the promoter of the MKP-1 gene contains AP-1 sites that respond to PKC-derived signals (12), and inhibition of PKC activity elongates ERK activation possibly due to inhibited induction of MKP-1. However, our observations are in agreement with previous studies that demonstrated that Ras recruits Raf-1 to the plasma membrane where it can interact with PKCε and act as a coregulator and cooperator of PKC-derived signals (36). Indeed, full Raf-1 phosphorylation is dependent upon its association with the plasma membrane (12) and PKCε is constitutively located in the plasma membrane in macrophages (5).

Although M-CSF and LPS activate a common target, the ERK pathway, distinct kinetics of ERK-1/2 activation and deactivation correlate with a range of macrophagic functions (5, 71). An early peak of ERK-1/2 activity is associated with proliferative signals, whereas macrophage-activating signals trigger a more delayed and lasting peak. Here we have demonstrated that distinct kinetics are also displayed by the upstream regulators MEK-1/2 and Raf-1 in response to M-CSF or LPS. This observation indicates that specific events that occur upstream of Raf-1 signaling may be responsible for these differences (Fig. 7). For example, Ras has been shown to be involved in M-CSF, but not in LPS, signaling toward Raf-1 activation, whereas phosphatidylcholine-phospholipase C appears to mediate Raf-1 activation in LPS-stimulated cells (71). Also, differences in receptor assembly and activation could account for the differential time courses; c-fms receptor requires noncovalent dimerization and autophosphorylation of tyrosine kinase domains (72), whereas recognition of LPS requires the assembly of TLR4 with CD14, LPS-binding protein, and MD-2 (73, 74).

The involvement of Raf-1 in mediating proliferation has also been described in other cellular types. In fibroblasts, Raf-1 activity is essential for the induced G₁-S phase transition and mitogenic effects of growth factors, as demonstrated by antisense constructs and dominant inhibitory mutants (75). Furthermore, oncogenically activated Raf-1 initiates DNA synthesis (76, 77, 78) and induces the transcription of several genes required for proliferation (77). In this context, the Raf-1 knockout has confirmed that Raf-1 is required for the viability of mice embryos and for ERK activation by mitogens in fibroblasts (41, 42). However, although oncogenic Raf-1 increases cyclin D1 protein in fibroblasts (77), we did not observe any modifications in the expression levels of any of the members of the cyclin D family. Interestingly, the Cdk inhibitors p21^Waf-1 and p27^Kip-1 were negatively regulated by Raf-1 in macrophages, which may explain the occurrence of growth arrest in the absence of active Raf-1. In agreement with our data, oncogenic activation of Raf-1 suppresses the expression of p27^Kip-1 in NIH3T3 cells (75). The data presented here provide new insights into the mechanisms used by macrophages to direct their response toward proliferation or activation.

We thank Tanya Yates for editing the manuscript.

The authors have no financial conflict of interest.