Interferon-inducible protein 10 and macrophage inflammatory protein-1 alpha inhibit growth factor stimulation of Raf-1 kinase activity and protein synthesis in a human growth factor-dependent hematopoietic cell line.

Stimulatory cytokines, including granulocyte-macrophage colony-stimulating factor (GM-CSF) and steel factor (SLF), act in a synergistic manner to stimulate the growth of hematopoietic progenitor cells, an effect also demonstrated for the growth factor-dependent human hematopoietic cell line MO7e. While little is known about the mechanisms responsible for mediating synergistic interactions of cytokines, Raf-1, a component of the MAP kinase signaling pathway, is thought to play a role in the stimulatory response evoked by several cytokines, including SLF and GM-CSF. Interferon-inducible protein-10 (IP-10) and macrophage inflammatory protein-1α (MIP-1α) are members of the chemokine family of suppressive cytokines. Prior exposure of hematopoietic cells to chemokines, including IP-10 and MIP-1α, inhibits the synergistic action of growth factors on stimulating cell proliferation. We report that treatment of MO7e cells with the combination of GM-CSF and SLF directly stimulates statistically significant synergistic increases in the phosphorylation and activation of Raf-1 kinase, and in cellular protein synthesis levels. Pretreatment of MO7e cells with IP-10 or MIP-1α blocked synergistic growth factor action, resulting in statistically significant suppression of cell proliferation, protein synthesis, and Raf-1 phosphorylation and activation. IP-10 and MIP-1α treatment also evoked significant increases in intracellular cAMP levels. Pretreatment of cells with agents which serve to raise intracellular cAMP levels, or with cAMP analogs inhibited the synergistic actions of GM-CSF and SLF in a manner similar to IP-10 and MIP-1α. In addition, treatment of cells with a potent inhibitor of cAMP-dependent protein kinase A blocked the suppressive action of MIP-1α and IP-10 on Raf-1 kinase activity and on MO7e cell proliferation. The ability of IP-10 and MIP-1α to antagonize the synergistic action of GM-CSF and SLF appears to involve inactivation of Raf-1 and the down-regulation of protein synthesis. Our findings suggest that both MIP-1α and IP-10 mediate their suppressive effects in MO7e cells by stimulating increases in cellular cAMP levels and activating protein kinase A, a mechanism we believe to be unique to these chemokines and not one applied to all growth suppressive members of the chemokine superfamily (for example, interleukin 8 and platelet factor 4).

Stimulatory cytokines, including granulocyte-macrophage colony-stimulating factor (GM-CSF) and steel factor (SLF), act in a synergistic manner to stimulate the growth of hematopoietic progenitor cells, an effect also demonstrated for the growth factor-dependent human hematopoietic cell line MO7e. While little is known about the mechanisms responsible for mediating synergistic interactions of cytokines, Raf-1, a component of the MAP kinase signaling pathway, is thought to play a role in the stimulatory response evoked by several cytokines, including SLF and GM-CSF. Interferon-inducible protein-10 (IP-10) and macrophage inflammatory protein-la (MlP-1α) are members of the chemokine family of suppressive cytokines. Prior exposure of hematopoietic cells to chemokines, including IP-10 and MlP-1α, inhibits the synergistic action of growth factors on stimulating cell proliferation. We report that treatment of MO7e cells with the combination of GM-CSF and SLF directly stimulates statistically significant synergistic increases in the phosphorylation and activation of Raf-1 kinase, and in cellular protein synthesis levels. Pretreatment of MO7e cells with IP-10 or MlP-1α blocked synergistic growth factor action, resulting in statistically significant suppression of cell proliferation, protein synthesis, and Raf-1 phosphorylation and activation. IP-10 and MlP-1α treatment also evoked significant increases in intracellular cAMP levels. Pretreatment of cells with agents which serve to raise intracellular cAMP levels, or with cAMP analogs inhibited the synergistic actions of GM-CSF and SLF in a manner similar to IP-10 and MIP-1α. In addition, treatment of cells with a potent inhibitor of cAMP-dependent protein kinase A blocked the suppressive action of MIP-1α and IP-10 on Raf-1 kinase activity and on MO7e cell proliferation. The ability of IP-10 and MIP-1α to antagonize the synergistic action of GM-CSF and SLF appears to involve inactiva-tion of Raf-1 and the down-regulation of protein synthesis. Our findings suggest that both MIP-1α and IP-10 mediate their suppressive effects in MO7e cells by stimulating increases in cellular cAMP levels and activating protein kinase A, a mechanism we believe to be unique to these chemokines and not one applied to all growth suppressive members of the chemokine superfamily (for example, interleukin 8 and platelet factor 4).
Growth of hematopoietic progenitor cells is coordinated by a number of stimulatory and inhibitory cytokines. Several stimulatory cytokines, including granulocyte-macrophage colonystimulating factor (GM-CSF)1 and steel factor (SLF), promote the growth of hematopoietic progenitor cells in a synergistic manner when administered in combination (1). Upon ligand binding and activation, cytokine receptors set into motion a cascade of phosphorylation and dephosphorylation reactions designed to transmit information from the cell membrane to other portions of the target cell (2). Activation of some cytokine receptors leads to phosphorylation and activation of receptorassociated proteins, including mSOS and GRB-2 (3), and the subsequent activation of ras (4). Activated Ras is thought to target inactive Raf-1 proteins to the cell membrane where they are phosphorylated and become active kinases (5). Both Ras and Raf-1 are components of the MAP kinase signaling pathway, a major stimulatory pathway within cell systems. While the mechanisms responsible for growth regulation within hematopoietic cells are not completely known and even less is known regarding synergistically induced cell proliferation, it appears likely that cross-talk between proteins associated with and activated by separate cytokine receptors mediate changes within target cells which are necessary for synergistic stimulation of cell proliferation.
The chemokine family of cytokines includes macrophage inflammatory protein 1-α (MIP-1α) and interferon inducible protein-10 (IP-10) (6). These molecules suppress the synergistic action of combinations of stimulatory cytokines on hematopoietic progenitor cell growth (7)(8)(9)(10). Other suppressive chemo- kines include interleukin 8 (IL-8), platelet factor 4 (PF4), MIP-2α, and macrophage chemotactic and activating factor (MCAF; also designated MCP-1) (8)(9)(10). Members of the chemokine family which do not suppress progenitor cell growth include MIP-1β and MIP-2β, GRO-α, and RANTES (7)(8)(9). While the suppressive effects of chemokines have been characterized, the cellular mechanisms through which growth inhibition is carried out have not been elucidated. Part of the difficulty is that chemokine suppression of growth factor action generally occurs during synergistic stimulation of cell proliferation. Therefore, studying the mechanism of action of specific chemokines is often limited to hematopoietic systems which not only display synergistic growth effects in response to growth factors but can be used readily for various biochemical analyses. Due to the rarity of hematopoietic stem and progenitor cells and the difficulty of isolating enough purified cells of this type for biochemical analyses, growth factor-dependent cell lines have been used (11). Growth arrest resulting from serum deprivation or growth factor deprivation is often associated with profound declines in protein synthesis rates for many cell systems (12). Since growth suppression mediated by chemokines may likely trigger responses similar to those evoked by factor deprivation, we set out to determine whether cytokine or chemokine treatment could alter protein synthesis rates in MO7e cells.
We and others have shown previously that treatment of MO7e cells with either GM-CSF or SLF results in increased phosphorylation and Raf-1 kinase activity (13,14). More recently, we have shown that treatment of MO7e cells with SLF results in the physical association between Ras and Raf-1 (15), an event now considered necessary for activation of Raf-1. Since SLF is known to synergize with a number of cytokines in promoting cell growth, we investigated whether Raf-1 kinase activity could be influenced by exposure of MO7e cells to growth factors in the presence or absence of various chemokines. Given that recent evidence has shown that Raf-1 can be inactivated through phosphorylation of Ser43, and a second serine residue in the kinase domain, by cAMP-dependent protein kinase A (16), we also investigated whether chemokines or growth factors could alter cAMP levels in MO7e cells.
Cells -The human factor-dependent cell line M07e was obtained from Genetics Institute (Boston, MA). Biological characteristics and culture conditions for the MO7e cell line have been described (17). MO7e cells were maintained in RPMI 1640 culture medium supplemented with 20% FBS and 100 units/ml rhuGM-CSF. Prior to growth factor or chemokine treatment, MO7e cells were washed with RPMI 1640 and "factor starved" in serum-free RPMI 1640 supplemented with 0.5% bovine serum albumin for 16-18 h at 37 °C.
Proliferation of MO7e Colony Forming Cells (CFC) -The percent MO7e CFC in S-phase were estimated by the high specific activity tritiated thymidine ([3H]Tdr) kill assay. Factor starved M07e cells were pretreated at 37 °C for 1 h with control diluent or 50 ng/ml of a specific chemokine. Cells were washed two times and treated with either control medium or 50 µCi of high specific activity [3H]Tdr (20 Ci/mmol; DuPont NEN) at 37 °C for 30 min prior to washing twice. Treated cells were then plated at 1.25 X 103 cells/ml in 0.3% agar culture medium with 10% FBS and in the presence of 100 units/ml rhuGM-CSF and 50 ng/ml rhuSLF. Colonies were scored after 7-8 days of incubation at 5% CO2 and lowered (5%) O2, conditions conducive for detection of the suppressive effects of MIP-1α on MO7e cell proliferation (18).
Iodination of Chemokines and Ligand Binding Assays -The specific ligand binding affinity and binding capacity of several chemokines were determined for MO7e cells. Carrier-free rhuMIP-1α and rhuMIP-1β were radiolabeled with 125I using the Bolton-Hunter (diiodo-) reagent method, as described previously (19). 125I-IL-8 was purchased from Amersham Corp. IP-10 was not examined for ligand binding since the process of labeling the molecule with radioactive iodine has been shown to render the molecule biologically inactive.2 Factor-starved MO7e cells were resuspended at 5 X 106 cells/ml in PBS containing 0.5% bovine serum albumin. For each chemokine tested, M07e cells were incubated on ice with various concentrations of radiolabeled ligand (ranging from 10 to 0.1 nM) in the presence or absence of a 100 X molar excess of cold, unlabeled ligand for 2 h at 4 °C. Free radiolabeled ligand was separated from ligand bound to cells at the end of the incubation period by passing each cell suspension through 200 µl of PBS containing 25% sucrose. The amount of radioactivity, as a measure of radiolabeled ligand present, was determined for supernantants and cell pellets using a gamma counter. Values were plotted as the amount of ligand bound to cells versus free ligand and then transformed into Scatchard plots for determination of the dissociation constant of ligand binding (Kd) and binding capacity (sites/cell) using the non-linear kinetics curve-fitting computer program Enzfitter (Elsevier Biosoft, Cambridge, U.K.).
[3H]Leucine Incorporation -Factor-starved MO7e cells were resuspended at 106/ml in leucine-free RPMI supplemented with 5 µCi/ml [3H]leucine and 0.5% FBS. Following stimulation with growth factors, chemokines, or other agents, cells were harvested into microcentrifuge tubes maintained on ice and washed two times with ice-cold PBS. Cells were resuspended in lysis buffer (50 mM Tris-HCl, pH 7.4, 10 mM EDTA, 10% glycerol, 140 mM NaCl, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, aprotinin, and 10 µg/ml leupeptin) and sonicated on ice for 30 s. Aliquots were removed and analyzed for total protein content using a BCA protein assay kit (Pierce). Incorporation of [3H]leucine into protein was determined for each sample by precipitating equivalent amounts of total protein (150 µg) onto glass fiber filter discs using ice-cold trichloroacetic acid, as described (20). Total counts of [3H]leucine present within precipitated protein and in the total protein of whole cell lysate aliquots were determined by liquid scintillation counting. Treatment of cells with the protein synthesis inhibitor cycloheximide (21) served as a means to determine the percentage of counts which represented ongoing protein synthesis in control, untreated cells.
cAMP Assay -Factor-starved MO7e cells were plated at a density of 5 X 105 cells/well in 24-well tissue culture plates. Following treatment with various agents, cells were harvested at different time points and collected by centrifugation at 500 X g. Cell pellets were resuspended in 150 µl of cold extraction buffer (50 mM Tris-HCl, pH 7.5, 4 mM EDTA) and homogenized on ice using a Dounce homogenizer. Extracts remained on ice for 15 min, with vigorous mixing every 5 min. Aliquots were removed and analyzed for protein content. Extract samples were boiled for 10 min. Cell debris was removed by centrifugation, and supernatants were transferred to fresh microcentrifuge tubes maintained on ice. A commercially available [3H]cAMP assay kit (Amersham Corp.) was used to measure cAMP content of samples, as per kit instructions. Incubations were carried out at 4 °C for 4 h, and charcoaldextran-treated samples were analyzed by liquid scintillation counting.
Immunoprecipitation -Factor-starved MO7e cells were washed, resuspended in phosphate-free RPMI 1640 medium containing 0.5% bovine serum albumin, and incubated for 1 h at 37 °C. Cells were then resuspended at 3 x 106 cells/ml in phosphate-free medium containing carrier-free [32P]orthophosphate at 1.0 mCi/ml for 2 h. Radiolabeled cells were treated with chemokines and/or growth factors and then placed directly into lysis buffer. Lysates were centrifuged to remove insoluble particles, and aliquots were normalized for protein content prior to immunoprecipitation. Immunoprecipitations were conducted by combining 150 µg of each sample with 5 µg of rabbit anti-MAP kinase (ERK1) antibody and incubating on ice for 1 h. Protein G-coated Sepharose beads were added for an additional 30 min and then pelleted at 500 x g. This preclearing step served to remove active MAP kinase from the assay sample, since MAP kinase is known to phosphorylate Mek1 (22). Precleared samples were then combined with 5 µg of rabbit anti-Raf-1 antibody and incubated on ice for 1.5 h. Raf-1-antibody complexes were collected by protein G-coated Sepharose beads. Immunoprecipitates were washed twice with 1 ml of high LiCl buffer (0.5 M LiCl, 100 mM Tris-HCl, pH 7.6), once with low LiCl buffer (0.1 M LiCl, 100 mM Tris-HCl, pH 7.6), and twice with lysis buffer. 32P-Labeled immunoprecipitates were then used directly in Raf-1 kinase activity assays or analyzed by SDS-PAGE and subsequent autoradiography and immunoblotting for Raf-1 content.
Immunoblotting -Raf-1 immunoprecipitates, or reaction mixtures from Raf-1 kinase activity assays, were combined with SDS protein sample buffer containing β-mercaptoethanol, boiled for 5 min, and the proteins separated by SDS-PAGE. Following electrophoresis, proteins were transferred to PVDF membrane (Millipore, Bedford, MA) using a Bio-Rad transblot apparatus (Hercules, CA) and then exposed to film. Following autoradiography, membranes were probed for specific proteins using a horseradish peroxidase-based detection system. All incubations were carried out at room temperature. Briefly, residual PVDFbinding sites were blocked by incubation of blots with a 5% milk solution in TBS (10 mM Tris-HCl, pH 8.0, 150 mM NaCl) for 1.5 h. Blots were then incubated with primary antibody (anti-Raf-1 Ab, 1:5000; anti-Mek1 Ab, 1:2500) in TBST (TBS plus 0.05% Tween 20) for 1.5 h. Following two wash steps with TBST, blots were incubated with horseradish peroxidase-linked anti-rabbit IgG secondary antibody (Amersham Corp.) in TBST for 1 h. Blots were washed twice with TBST and placed in a solution containing a 1:1 mixture of detection reagents from an ECL Western blotting detection kit (Amersham Corp.) for 1 min. Blots were drained of excess liquid, wrapped in thin plastic wrap, and exposed to film. Protein content was determined by densitometric analysis of exposed films.
GST-Mekl Fusion Protein -A GST-Mekl fusion protein constructed by Dr. Zhi-jun Luo (Harvard University Medical School) was kindly provided to us for use as a substrate in our Raf-1 kinase assay. Briefly, a cDNA encoding human Mek1 (23) and six consecutive histidines at its carboxyl terminus was constructed in the vector pGEX-KG. The recombinant fusion protein was purified from the Escherichia coli strain JM109 harboring the plasmid using Ni2+-NTA agarose (In Vitrogen) followed by GSH-agarose (Sigma) affinity purification. The purified fusion protein was dialyzed against a buffer containing 20 mM Tris-HCl, pH 7.9, 5 mM MgCl2, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 50% glycerol and stored at -20 °C.
Raf-1 Kinase Activity Assay -Raf-1 immunoprecipitates were combined with the GST-Mek1 fusion protein (1 µg/sample) and 0.1 mM [32P]ATP in 50 µl of kinase assay buffer (50 mM β-glycerophosphate, 0.03% BRIJ-35, pH 7.3, 10 mM MgCl2). The kinase reaction was conducted at 30 °C for 30 min and then terminated upon addition of protein sample buffer. Reaction mixtures were boiled and then loaded onto 12% SDS-PAGE gels. After electrophoresis, proteins were transferred to PVDF membranes and exposed to film. Autoradiograms were subjected to densitometric analysis in order to assess differences in phosphorylation intensity between treatment groups. Following autoradiography, the same membranes were immunoblotted for both Raf-1 and Mek1 in order to assess the degree of protein content present in relation to phosphorylation state for both proteins.
Statistical Analysis of Raf-1 Phosphorylation and Kinase Assay Data -Phosphorylation data obtained from Raf-1 phosphorylation studies and Raf-1 kinase assays were subjected to statistical analyses. Densitometric values were analyzed as percentages (the area under the peak for each band expressed as a percentage of total area under the curve analyzed). In order to satisfy the normality assumptions (Gaussian) of the statistical analysis to be used, the data were first transformed by means of the angular transformation (24). This transformation is particularly suited to transform percentage data to normality. As a check, χ2 tests for goodness-of-fit (25,26) were also conducted. Raf-1 and GST-Mek1 phosphorylation data were analyzed using analysis of variance, with experiments as blocks and GM-CSF, SLF, GM-CSF+SLF, and Control as effects. The means of GM-CSF, SLF, and GM-CSF+SLF groups were compared to the Control mean using Dunnett's test (27). The MIP-1α+GM-CSF+SLF and IP-10+GM-CSF+SLF means were compared to the GM-CSF+SLF mean also using the Dunnett test. To test for synergistic interaction between GM-CSF and SLF, an appropriate contrast was specified. All tests were two sided. The SAS (28) and Stata (29) software were used. χ2 tests indicated that the angularly transformed data were normally distributed.

Statistical Analysis of Protein Synthesis and cAMP Data -
The statistical significance between treatment groups for the protein synthesis and cAMP assay data were determined by Student's t test.

Influence of Chemokines on Cycling of MOTe CFC -We had
previously demonstrated in preliminary experiments that rmuMIP-1α resuspended in acetonitrile (ACN)-based buffer could decrease the percentage of MO7e CFC in cycles that were responsive to the stimulatory effects of the combination of GM-CSF plus SLF, but not those MO7e CFC responsive to only GM-CSF (18). This mimicked the effects of chemokines on subsets of normal myeloid progenitor cells which were responsive to stimulation by the combination of SLF with a colonystimulating factor such as GM-CSF (8)(9)(10). In the present study, the non-ACN-treated chemokines rhuIP-10, rhu and rmuMIP-1α, natural huPF4, rhuIL-8 and rhuMIP-1β, each at 50 ng/ml, were assessed for their effects on GM-CSF plus SLF-responsive MO7e CFC (Table I). Pulse exposure of MO7e cells for 1 h at 37 °C in vitro to IP-10, MIP-1α, PF4, and IL-8, chemokines which have suppressive effects on colony formation by normal myeloid progenitor cells (8-10), significantly decreased by approximately 50% the percentage of MO7e CFC in S-phase of the cell cycle (p < 0.01). The rhu and rmu preparations of MlP-1α were equally suppressive, and the results for these were combined. There was no significant difference between the suppressive effects of IP-10, MIP-1α, PF4, or IL-8. This pulse exposure or leaving the chemokines in with MO7e cells for the duration of the 7-8-day semisolid medium culture period did not significantly influence the number of colonies formed (5 ± 4% change from control numbers, p > 0.05); however, addition of the suppressive chemokines once a day for 4 days resulted in 35-45% significant (p < 0.01) inhibition of total colony formation, suggesting that the suppressive effects on MO7e CFC cycling were reversible. Pulse exposure of MO7e cells to MIP-1β or GRO-α, which do not have suppressive activity on normal myeloid progenitor cells (8,9), did not influence the percentage of MO7e CFC in S-phase (Table I)  For cycloheximide treatment, cells were exposed to cycloheximide (50 µM) for 1 h, washed with PBS, and then received control vehicle only for the remainder of the treatment duration. Whole cell lysates were prepared and analyzed for protein content.
[3H]Leucine incorporation was determined for lysate aliquots (150 µg/sample) by trichloroacetic precipitating labeled proteins onto glass fiber filters and counting the amount of [3H]leucine present in dried filters by liquid scintillation counting. Each point represents the mean of three separate determinations. Incorporation levels were significantly lower than control values (p < 0.05) for all groups treated with MIP-1α, IP-10, or PF4 at concentrations of 20 ng/ml or greater. In contrast, treatment with MIP-1β at any concentration did not significantly effect basal levels of leucine incorporation. Similar results were obtained in each of three separate experiments.
ious Chemokines -Factor-starved MO7e cells were treated with various chemokines in the presence of [3H]leucine. Proteins were trichloroacetic acid precipitated from whole cell lysates and analyzed for the amount of [3H]leucine incorporated into protein. Treatment of cells with IP-10, MIP-1α, or PF4 modestly yet significantly decreased (p < 0.05) the basal level of leucine incorporation 15-20% below control levels within 12 h (Fig. 1). In contrast, MIP-1β or GRO-α (not shown) did not alter the level of leucine incorporation. Cycloheximide (50 µM) blocked 95% of [3H]leucine incorporation detected for control cells (Fig. 1), suggesting that the level of leucine incorporated into MO7e cell protein related directly to the level of active, ongoing protein synthesis.
Growth Factor-stimulated Increases in Protein Synthesis Are Suppressed by IP-10, MlP-1α, and Cholera Toxin -Treatment of MO7e cells with the growth factors GM-CSF (100 units/ml) and SLF (50 ng/ml) resulted in significant 40-65% increases (p < 0.05) in [3H]leucine incorporation by 12 h (Fig. 2, A-C). Treatment with either GM-CSF or SLF alone stimulated increases in protein synthesis 15-20% above control levels (not shown). To determine whether chemokine treatment could block the stimulatory effect of growth factors, MO7e cells were exposed to IP-10 (50 ng/ml) or MIP-1α (50 ng/ml) for 1 h at 37 °C prior to treatment with the combination of GM-CSF and SLF. Pretreatment with either chemokine was sufficient to significantly (p < 0.05) block the increase in protein synthesis stimulated by GM-CSF and SLF down to control, unstimulated levels within 12 h (Fig. 2A). Pretreatment of cells for 1 h with cholera toxin (1 µg/ml), which increases cAMP in MO7e cells (30), blocked the stimulatory action of GM-CSF and SLF on protein synthesis (Fig. 2B). Cholera toxin exposure decreased the basal level of protein synthesis in MO7e cells to a level similar to that evoked by chemokine pretreatment (Fig. 2, A  and B). Pretreatment with the chemokines IL-8 (50 ng/ml) or PF4 (50 ng/ml) also served to block the stimulatory action of GM-CSF and SLF (Fig. 2C), although to a lesser degree in comparison to MIP-1α and IP-10 ( Fig. 2A). In contrast, pretreatment of cells with 50 ng/ml GRO-α ( Fig. 2A) or 50 ng/ml MIP-1β (Fig. 2C) failed to block the stimulatory action of GM-CSF plus SLF on protein synthesis, nor did exposure of cells to GRO-α alter basal protein synthesis levels ( Fig. 2A).
Chemokines Bind to MO7e Cells with Similar Affinity -While the ability of specific chemokines to suppress the growth of hematopoietic progenitor cells and MO7e cells has been well documented (8-10, 18, 30), studies relating relative binding affinity to suppressive activity for each chemokine have not been reported for MO7e cells. In order to determine whether the ability of a chemokine to suppress cell growth or protein synthesis in MO7e cells might be related to ligand binding, we set out to determine the ligand binding affinity and capacity of binding sites on the surface of MO7e cells for several chemokines. rhuMIP-1α bound specifically to MO7e cells with a relatively high affinity, with a calculated dissociation constant of 1.2 nM and a capacity of 2,266 binding sites/cell (Table II). This observation is consistent with the results of our previous study in which the dissociation constant for MIP-1α resuspended in an acetonitrile-based buffer was determined to be 1.0 nM (31). We found that IL-8 also bound specifically to the surface of MO7e cells, with an affinity and binding capacity similar to MlP-1α (Table II). While the number of binding sites determined for MIP-1β in MO7e cells was consistent with capacities determined for the other chemokines, we calculated the dissociation constant for MIP-1β to be 3.9 nM, an affinity value 2-3-fold lower than those calculated for MlP-1α and IL-8 (Table II).
Inhibitory Action of IP-10 Requires a Minimal Pretreatment Duration -Prior exposure of MO7e cells to MIP-1α antagonizes the proliferative effects of growth factors on cell proliferation (18). In order to determine if the suppressive activity of IP-10 on protein synthesis also required a minimal pretreatment duration, MO7e cells were treated with IP-10 at the same time as, or for various times prior to, treatment with GM-CSF plus SLF in the presence of [3H]leucine. Coadministration of GM-CSF and SLF stimulated steadily increasing levels of leucine incorporation above control levels throughout a 12-h treatment period. Cotreatment with IP-10 at the same time appeared to have no effect on the ability of GM-CSF and SLF to increase protein synthesis levels. Pretreatment with IP-10 for 15 min prior to stimulation with GM-CSF plus SLF blocked the increases in protein synthesis 70-80% (Fig. 3). Administration of IP-10 30 min prior to growth factor treatment suppressed the level of [3H]leucine incorporation down to control levels (Fig. 3).
IP-10 and MIP-1 Stimulate Significant Increases in cAMP-Since cholera toxin, which acts by increasing cellular levels of cAMP, could mimic the action of IP-1O and MIP-1α in antagonizing the stimulatory effects of GM-CSF plus SLF on protein synthesis levels, we determined whether these or other chemokines could be acting by altering intracellular cAMP levels. Whole cell lysates were prepared from factor-starved MO7e cells treated for different times with various chemokines or cholera toxin, and then assayed directly for cAMP content. Treatment of MO7e cells with 50 ng/ml IP-10 or MIP-1α evoked significant (p < 0.05), 3-4-fold increases, respectively, in cAMP levels within 2 h (Fig. 4). This observation is consistent with results of our recent study in which ACN-treated rhuMIP-1α was shown to significantly increase cAMP in MO7e cells, while rhuRANTES was not (31). Treatment with cholera toxin (1 µg/ml) stimulated increases in cAMP 5-6-fold higher than the maximal level stimulated by either chemokine (not shown). In contrast to these results, treatment with other chemokines, Cell lysates were analyzed for protein content and [3H]leucine incorporation, as described in the legend to Fig. 1. Each point represents the mean of three separate determinations. Incorporation levels for the GM-CSF plus SLF treatment group were significantly higher than controls at 12, 18, and 24 h for experiments shown in panels A-C (p < 0.05). Incorporation levels for MIP-1α, IP-10, or cholera toxin pretreatment groups were significantly less than those for the GM-CSF plus SLF group at 12, 18, and 24 h (p < 0.05).

TABLE II Chemokine binding affinity and capacity in MO7e cells
Factor-starved M07e cells resuspended in PBS containing 0.5% BSA were incubated in the presence of various concentrations of 125I-labeled chemokines (range 10-0.1 nM) +/-a 100X molar excess of unlabeled chemokine for each concentration. Microfuge tubes were maintained on ice at 4 °C for 2 h. Specific ligand binding, as measured by radioactivity, was determined for each chemokine as described under "Experimental Procedures." Each ligand dissociation constant and binding capacity were determined from Scatchard plot analysis of each individual experiment using the non-linear kinetics curve-fitting program Enzfitter. Each value represents the mean +/-S. including PF4, IL-8, MIP-1β, and GRO-α, failed to stimulate any changes in cellular cAMP levels (Fig. 4).
Increased Raf-1 Phosphorylation Stimulated by Growth Factors Is Antagonized by Chemokine Pretreatment -Since GM-CSF and SLF each alone are believed to exert at least part of their effects within target cells through an activation of the Ras/Raf-1/MAP kinase cascade (13,15,32,33), we set out to determine whether treatment with the growth factors GM-CSF and SLF, alone or in combination with various chemokines, could alter the phosphorylation state of the Raf-1 kinase protein. Factor-starved MO7e cells maintained in phosphate-free medium were incubated with [32P]ATP in the presence of various growth factors and/or chemokines. 32P-Labeled proteins were immunoprecipitated from whole cell lysates using anti-Raf-1 antibodies, separated by SDS-PAGE, transferred to PVDF membranes, and exposed to film. Raf-1 appears as a single band at an approximate molecular weight of 74 kDa, as indicated by the arrow (Fig. 5, A-C). Treatment of factorstarved MO7e cells with the combination of GM-CSF (100 units/ml) and SLF (50 ng/ml) synergistically increased the phosphorylation of Raf-1 (Fig. 5A, lanes 1-4). Pretreatment with IP-10 (50 ng/ml) or MIP-1α (50 ng/ml) for 1 h prior to growth factor treatment greatly reduced the increase in Raf-1 phosphorylation (Fig. 5A, lanes 5 and 6). Pretreatment with IP-10 for 15 min was less effective than the 1-h pretreatment duration at blocking Raf-1 phosphorylation (Fig. 5A, lane 7).

FIG. 4. IP-10 and MIP-1α significantly increase cAMP levels in MO7e cells.
Factor-starved MO7e cells were treated for the indicated times with IP-10 (50 or 100 ng/ml), MIP-1α (50 or 100 ng/ml), PF4 (100 ng/ml), IL-8 (50 ng/ml), MIP-1β (50 ng/ml), or 1 µg/ml cholera toxin (not shown). Cell lysates were analyzed directly for protein content and for cAMP content, using a commercially available [ 3 H]cAMP assay kit (Amersham). Each sample was assayed in duplicate. Each bar represents the mean ± S.E. of separate determinations obtained from three separate experiments. cAMP levels evoked by treatment with either concentrations of IP-10 or MIP-1α were significantly higher than control levels at 2 and 4 h (p < 0.05). cAMP levels of cells treated with PF4, IL-8, or MIP-1β did not differ significantly from control levels.
Pretreatment with cholera toxin (1 µg/ml) or forskolin (50 µM) for 1 h also reduced the increases in phosphorylation of Raf-1 stimulated by GM-CSF and SLF (Fig. 5A, lanes 8 and 9). In contrast to these results, pretreatment of cells with 50 ng/ml GRO-µ, MIP-1β, or IL-8 failed to block phosphorylation of Raf-1 stimulated by GM-CSF plus SLF (Fig. 5B). Immunoblot anal-  1, panels A and B) received vehicle only. Raf-1 proteins were immunoprecipitated from whole cell lysates by anti-Raf-1 antibodies, separated by 12% SDS-PAGE, transferred to PVDF membrane, and the intensity of 32 P-labeling visualized by autoradiography. At left is indicated the position of the molecular weight markers. Raf-1 appears as a single band at approximately 74 kDa, as indicated by the position of the arrow. C, Raf-1 protein content was determined by immunoblotting PVDF membranes used for 32 P analysis with anti-Raf-1 antibodies and horseradish peroxidase-linked protein G. Raf-1 proteins were visualized upon exposure of ECL-treated membranes to film, as described under "Experimental Procedures." Treatment groups are the same as in panel A. ysis of PVDF membranes for Raf-1 protein using anti-Raf-1 antibodies demonstrated equivalent protein loading between lanes (Fig. 5C). Results of densitometric analysis of phospho- P]orthophosphate were treated either with growth factors for 10 min, chemokines for 1 h, or pretreated for 1 h with 50 ng/ml MIP-1α or IP-10 prior to a 10-min treatment with GM-CSF plus SLF. Raf-1 proteins were immunoprecipitated from whole cell lysates, separated by SDS-PAGE and blotted to polyvinylidene difluoride membranes, as described under "Experimental Procedures." Intensity of 32 P-labeled proteins were visualized by autoradiography of membranes and differences between treatment groups determined by densitometry of autoradiogram films. Raf-1 protein content was determined by immunoblotting polyvinylidene difluoride membranes with anti-Raf-1 antibodies and horseradish peroxidase-linked protein G. ECL-treated membranes were exposed to film, and autoradiograms analyzed by densitometry. rylated bands are presented in the summary statistics of Table III.
Growth Factor Stimulation of Raf-1 Kinase Activity Is Suppressed by Chemokine Pretreatment -We determined whether the actions of IP-10 and MIP-1α on Raf-1 phosphorylation state were related to changes in kinase activity. Raf-1 proteins, immunoprecipitated from MO7e whole cell lysates by anti-Raf-1 antibodies and protein G-Sepharose, were combined with [ 32 P]ATP and a GST-Mek1 fusion protein. Mek1, also known as MAP kinase kinase, is phosphorylated by activated Raf-1 (33). Following incubation, assay proteins were separated by SDS-PAGE, transferred to PVDF membrane, and subjected to autoradiography. GST-Mek1 appeared as a single band at approximately 69-70 kDa (Fig. 6, A and B, lower arrow at right). The fainter band appearing immediately above GST-Mek1 migrated at approximately 74 kDa and corresponded to autophosphorylated Raf-1 (Fig. 6, A and B, upper arrow at right). Treatment of cells with either GM-CSF or SLF was sufficient to activate Raf-1 kinase and increase Mek1 phosphorylation (Fig.  6A, lanes 2 and 3). Cotreatment with GM-CSF plus SLF stimulated synergistic increases in Raf-1 kinase activity, as evidenced by higher levels of phosphorylated Mek1 substrate (Fig.  6A, lane 4). Treatment of cells with either MIP-1α or IP-10 1 h prior to growth factor stimulation resulted in the suppression of GM-CSF plus SLF synergistically stimulated Raf-1 kinase activity (Fig. 6A, lanes 5 and 6). Pretreatment of cells with cholera toxin or forskolin (Fig. 6A, lanes 7 and 8), or the cAMP analog 8-bromo-cAMP (Fig. 6B, lane 6) also served to block growth factor stimulation of Raf-1 kinase activity. In contrast to these results, pretreatment of cells with either GRO-α, MIP-1β, PF4, or IL-8 failed to block Raf-1 activation and phosphorylation of GST-Mek1 stimulated by GM-CSF and SLF (Fig.  6B). The appearance and intensity of radiolabeled protein bands detected below the GST-Mek1 band (Fig. 6A) and above Raf-1 in Fig. 6B varied between lanes within the same experiment and from one experiment to another. These phosphorylated protein bands represent artifacts from the isolation of GST-Mek1 or immunoprecipitation of Raf-1 and do not appear to be related to overall protein levels or the ability of raf-1 to phosphorylate Mek1. Immunoblot analysis of PVDF membranes with anti-Mek1 antibodies indicated the relative amount of Mek1 protein present in each lane (Fig. GC). Results of densitometric analysis of phosphorylated protein bands are presented in the summary statistics of Tables III and IV. GM-CSF and SLF Synergistically Stimulate Raf-1 Kinase-Summary statistics representing densitometric analysis of Raf-1 phosphorylation and kinase activity assay data are presented in Tables III and IV, respectively. Densitometric values were normalized to Raf-1 or Mek1 protein content prior to statistical analysis. For both Raf-1 phosphorylation (Table III) and GST-Mek1 phosphorylation (Table IV), the means of GM-CSF, SLF and GM-CSF + SLF were significantly higher than the Control mean (Dunnett, p < 0.05). Statistical analysis showed that a significant synergistic interaction existed between GM-CSF and SLF (F test, p < 0.001) for both Raf-1 phosphorylation and Raf-1 kinase (GST-Mek1 phosphorylation) activity. In addition, the mean MIP-1α+GM-CSF+SLF and mean IP-10+GM-CSF+SLF values were significantly lower than the mean of GM-CSF + SLF (Dunnett, p < 0.05) for both data sets.
Suppressive Activity of MIP-la and IP-10 Are Blocked by a Potent Inhibitor of cAMP-dependent Protein Kinase A -Since the results of our studies suggested a correlation between changes in cellular cAMP levels and the suppressive activities of MIP-1α, we set out to determine whether changes in cAMP levels could be mediating the inhibitory effects of this chemokine. Since Raf-1 is known to be inhibited directly by the action of cAMP-dependent protein kinase A, we made use of the specific protein kinase A inhibitor PKI (34) to block the action of protein kinase A within MO7e cells. As expected, treatment of MO7e cells with GM-CSF in combination with SLF stimulated an increase in Raf-1 kinase activity, as shown by the increased phosphorylation of the GST-Mek1 substrate in a Raf-1 kinase assay (Fig. 7, lower band, lane 2). Pretreatment with PKI (10 µg/ml) had no effect upon the ability of GM-CSF plus SLF to activate Raf-1 (lane 3). Pretreatment of MO7e cells with MIP-1α prior to growth factor treatment resulted in a decrease in the phosphorylation and activity of Raf-1 (lane 5). Pretreatment of MO7e cells with MIP-1α in combination with PKI, however, resulted in a block to the suppressive action of MIP-1α and allowed for the activation of Raf-1 by GM-CSF and SLF, as evidenced by increased Mek1 phosphorylation (Fig. 7,  lane 6).
In addition to examining the ability of PKI to block the inhibitory effects of MIP-1α on Raf-1 activation, we set out to determine whether PKI could also block the growth suppres-  (Table V). The suppressive activity of MlP-1α and IP-10 were blocked completely upon coincubation of cells with either chemokine and 10 µg/ml PKI (Table V). In contrast, coincubation of cells with PKI and either IL-8 or PF4 failed to block the suppressive activity of these two chemokines (Table V). PKI did not effect the percentage of MO7e in S-phase of the cell cycle treated with Control medium alone (Table V). DISCUSSION Raf-1 kinase plays a key role in the MAP kinase signaling pathway which links membrane-associated events with other metabolic processes occurring within target cells (35). We present here our findings showing that the phosphorylation state and kinase activity of Raf-1 in MO7e cells can be increased synergistically in response to treatment with a combination of GM-CSF and SLF. We believe that this is the first time that direct, statistically significant synergistic activation of Raf-1 by combined cytokine treatment has been reported in a complete study. A preliminary study conducted by others and reported in abstract form observed that treatment of murine factor-dependent cells with physiological doses of SLF and either IL-3 or erythropoietin synergistically phosphorylated and activated Raf-1 (36). Our present results demonstrating GM-CSF and SLF synergistic activation of Raf-1 differ from those of our previous study (13) and that of others (14). Differences in the immunoprecipitation buffer conditions and the amount of radionucleotide used to label target cells may account for these differences. More importantly, however, our assay system utilizes a GST-Mekl fusion protein as a Raf-1 kinase substrate. Since Mekl is a biological substrate for activated Raf-1, phosphorylation of GST-Mekl provides a more sensitive and biologically relevant assay system to study Raf-1 kinase activity than previous studies which employed more broad spectrum substrates, such as histone H1 (13).
Phosphorylation of Raf-1 at serine 43 and at a second, currently unidentified serine residue in the kinase domain, by protein kinase A serves to inactivate Raf-1 kinase and prevent its association with activated Ras (37). We report here that MlP-1α and IP-10 pretreatment can inhibit Raf-1 phosphorylation and decrease Raf-1 kinase activity stimulated by growth factor treatment, while simultaneously increasing cAMP levels in MO7e cells. The suppressive action of these chemokines can be mimicked by treatment of cells with agents which serve to raise intracellular levels of cAMP, such as cholera toxin and forskolin, and by the cAMP analog 8-bromo-cAMP. These results are consistent with those we reported in a recent study in which cAMP and cAMP analogs were shown to inhibit the number of MO7e CFC in cycle, in a manner similar to the suppressive action of ACN-treated MIP-1α (31). Taken together with results of Raf-1 kinase assays and MO7e CFC thymidine kill assays in which we demonstrated that the suppressive action of MIP-la and IP-10 are blocked by the protein kinase A inhibitor PKI, our observations strongly suggest that part of the inhibitory mechanism employed by MIP-la and IP-10 relate directly to alterations in intracellular cAMP levels. However, our observations that rhuIL-8 and PF4 fail to alter cAMP levels, fail to block activation of Raf-1 stimulated by growth factors, and do not appear to be sensitive to inhibition of protein kinase A in MO7e cells indicate that inactivation of Raf-1 by cAMP may not be applied as a general inhibitory mechanism for all suppressive chemokines, but as one which is limited to a few members of the chemokine superfamily.
A receptor which binds MIP-1α with high affinity has recently been cloned (38). Based upon sequence information, this receptor is thought to be comprised of seven membrane-spanning domains, a protein structure which is consistent with membrane-bound receptors coupled through G-proteins to adenylyl cyclase and other effector molecules (38). While this observation is of interest, neither direct nor indirect activation of adenylyl cyclase has been reported for any of the chemokine TABLE IV Densitometric analysis of Raf-1 kinase assay data Factor-starved MO7e cells were treated either with growth factors for 10 min, chemokines for 1 h, or pretreated for 1 h with 50 ng/ml MIP-la or IP-10 prior to a 10-min treatment with GM-CSF plus SLF. Raf-1 proteins were immunoprecipitated from whole cell lysates and combined with GST-Mekl in the presence of [ 32 P]ATP for 30 min at 30 °C, as described under "Experimental Procedures." Proteins were separated by SDS-PAGE, transferred to polyvinylidene difluoride membranes and the intensity of 32 P incorporation into Mekl and Raf-1 was visualized by autoradiography. Mek1 content was determined by immunoblotting polyvinylidene difluoride membranes with anti-Mek1 antibodies and horseradish peroxidaselinked protein G. ECL-treated membranes were exposed to film and developed films were analyzed by densitometry.  receptors. Our observation that both growth suppressive chemokines, such as MIP-1α and IL-8, and non-suppressive chemokines, such as MIP-1β, can bind with moderately high affinity to MO7e cells indicates that the degree of suppressive activity may not be related directly to the number or specificity of chemokine-binding sites, but rather to the effector pathways activated in response to ligand binding. However, the 2-3-fold difference between binding affinities for MIP-1α and MIP-1β may account for the ability of excess MIP-1β to block the suppressive action of MIP-1α, since we have shown that greater concentrations of MIP-1β are required in order to block MIP-1α suppressive action (18).
We have demonstrated that GM-CSF and SLF can stimulate increases in protein synthesis in MO7e cells. This provides an additional model system through which the growth promoting effects of cytokines can be examined. By making use of protein synthesis regulation as an assay system, we have shown that pretreatment of MO7e cells with the chemokines IP-10 and MIP-1α, and IL-8 and PF4 to a lesser extent, can block the stimulatory effects of GM-CSF and SLF. Since declines in protein synthesis are often associated with quiescent or growtharrested cells, the ability of these chemokines to alter protein synthesis levels may be related to altering the cell cycle progression of target cells. Our observations are consistent with results of the thymidine kill assays in which MO7e cells were treated with various agents and then exposed to high specific activity [ 3 H]thymidine. In this assay system, cells in S-phase incorporate [ 3 H]thymidine and are reproductively sterilized as a result. MO7e cells pretreated with specific chemokines, including MIP-1α and IP-10, tended to survive the exposure to the thymidine, suggesting that chemokines protect these cells by altering their progression into cycle and/or by slowing down their growth rate (18).
Down-regulation of both protein synthesis and Raf-1 activity evoked by IP-10 and MIP-1α within the same cell type suggests that these two events may be linked to growth suppression. Since the ability of these chemokines to block the stimulatory action of growth factors appears to require a minimal pretreatment duration, it is likely that stimulatory and inhibitory pathways utilized by cytokines and chemokines may share several key components which, depending upon how they are modified, may ultimately lead to growth activation or suppression. Although Raf-1 activation of MAP kinase is thought to play a role in the growth promoting effects of GM-CSF and SLF on hematopoietic cells (13)(14)(15), the connections to factors down stream of MAP kinase which are ultimately responsible for activation of cell growth have not been completely identified. Since active cell growth and division are associated with periods of increased protein synthesis, it is likely that a key component of growth regulation by cytokines may reside in the activation of the protein synthesis machinery within target cells. Activation of the initiation factor eIF-4E, as an example, is thought to represent at least one of the rate-limiting steps in the stimulation of protein synthesis in eukaryotic systems (39). What is of particular interest regarding eIF-4E is that it can be phosphorylated and activated through an unknown mechanism directly upon exposure to active Ras and, in a manner similar to Raf-1, can be activated in response to phosphorylation by protein kinase C (40). Since exposure of hematopoietic cells and MO7e cells to cAMP has been shown to be inhibitory for cell growth (30,31,41), it may be through inactivation of eIF-4E or related proteins that protein synthesis is shut down and cell growth is halted. Conversely, activation of protein synthesis in 32b 32b a CFC values represent the mean +/-S.E. of the number of responsive colonies observed for three separate determinations. b Significant difference from control medium, p < 0.01. some cell systems can be achieved through inactivation of inhibitory initiation factors. Phosphorylation of the initiation factor eIF-2α by double-stranded RNA-dependent protein kinase has recently been shown to result in the inhibition of protein synthesis in murine H7 cells (42). Subsequent studies conducted by the same group have shown that treatment with IL-3 leads to the dephosphorylation and inactivation of both double-stranded RNA-dependent protein kinase and eIF-2α, and the subsequent activation of protein synthesis in IL-3 deprived murine hematopoietic cells (43). Similar to our present findings, results of studies involving IL-3 and eIF-2α demonstrate a clear connection between the stimulatory action of growth factors and regulation of protein synthesis (42,43), providing evidence for one possible mechanism by which this regulation is accomplished. By exploring these and other links which may exist between the regulation of protein synthesis and interactions among signaling pathways, including Raf-1 activation and alterations in cellular cAMP levels, we may be able to determine the sequence of events through which hematopoietic cell growth is regulated by multiple factors.