Distinct regions of the human granulocyte-colony-stimulating factor receptor cytoplasmic domain are required for proliferation and gene induction.

Using two different cell systems, we show that the cytoplasmic domain of the granulocyte-colony-stimulating factor receptor (G-CSFR) may be composed of at least two functional regions. The first, within the membrane-proximal 57 amino acids, is absolutely required to deliver a proliferative signal. This region contains two sequence motifs conserved between members of the hematopoietin receptor family. The second functional region resides between amino acids 57 and 96. This region is required for the induction of acute-phase plasma protein gene expression when the G-CSFR is transfected into human hepatoma cell lines. The G-CSFR-transfected hepatoma cells respond to G-CSF by increasing the production of the same set of plasma proteins as stimulated by interleukin-6, suggesting that the two cytokines share a common signal transduction pathway.

Molecular cloning and characterization of a novel receptor protein tyrosine kinase from human placenta.

Using a polymerase chain reaction-based approach we have isolated and characterized a cDNA (HPK-6) from human placental RNA encoding a novel receptor protein tyrosine kinase. This receptor tyrosine kinase has a unique extracellular domain, with an immunoglobulin-like domain at the amino terminus followed by three EGF-like cysteine repeats and three fibronectin type III repeats, giving the HPK-6 gene extracellular domain a novel combination of structural motifs. A comparison of the HPK-6 sequence with other receptor tyrosine kinases shows that the HPK-6 gene is the human homolog of the murine tek gene and very closely related to the recently described receptor tyrosine kinase tie. The HPK-6 gene is expressed predominantly in placenta and lung, with a lower level in umbilical vein endothelial cells, brain and kidney. The HPK-6 cDNA, when transfected into COS-7 cells, encodes a 140-kDa protein with in vitro kinase activity.

Murine interleukin-1 receptor: differences in binding properties between fibroblastic and thymoma cells and evidence for a two-chain receptor model.

The concentration of porcine interleukin-1 beta (pIL1 beta) required to elicit half-maximal IL2 production from NOB-1, a subline of murine thymoma EL4, was 100-fold greater than for p1L alpha. In contrast, similar doses of each type of IL1 stimulated increased lactate production by Balb/C 3T3 fibroblasts. Receptor-bound 125I-IL 1 alpha was displaced with equal efficiency by both unlabelled forms from 3T3 cells, but a 20-fold lower affinity for p1L1 beta was observed using NOB-1. Crosslinking experiments suggested that the IL1 receptors on each line consisted of two polypeptides of 80 and 100 kDa. The results provide the first evidence for a multiple-component IL1 receptor within which IL1 alpha and IL1 beta may bind at different loci, and suggest the receptors may have evolved differently in the two lines.

The interleukin-1-stimulated protein kinase that phosphorylates heat shock protein hsp27 is activated by MAP kinase.

In KB cells, interleukin-1 (IL-1), epidermal growth factor and phorbol ester transiently activated both MAP kinase and a serine kinase which phosphorylated the heat shock protein hsp27. Extracts made from IL-1-stimulated KB cells phosphorylated recombinant hsp27, in vitro, on serine residues 78 and 82 which are contained within Arg-X-X-Ser motifs similar to those phosphorylated by the ribosomal protein S6 kinases. Upon size exclusion chromatography, however, hsp27 kinase eluted as a single peak of activity at 50-60 kDa, clearly separated from ribosomal protein S6 kinases. Treatment of partially purified hsp27 kinase with protein phosphatase-2a reduced its activity by 80%. De-phosphorylated hsp27 kinase could be approximately 50% reactivated by a factor present in IL-1-treated cell extracts in the presence of ATP. This factor co-eluted with MAP kinase after partial purification by DEAE-cellulose, phenyl Sepharose, and size exclusion chromatography. Purified sea star p44mpk and recombinant ERK2 MAP kinases were also capable of re-activating hsp27 kinase to a similar extent. These data suggest that hsp27 kinase is downstream from, and probably a direct target of MAP kinase.

Binding of interleukin-18 to the interleukin-1 receptor homologous receptor IL-1Rrp1 leads to activation of signaling pathways similar to those used by interleukin-1.

Interleukin-18 (IL-18) is an inflammatory cytokine that has been shown to enhance a variety of Th1 type T cell responses. Because IL-18 is homologous to IL-1, we tested binding of IL-18 to the known IL-1R family members. We could show binding of IL-18 to the orphan receptor IL-1Rrp1 but not to other IL-1R homologous proteins. IL-1Rrp1 and IL-1RI share highly conserved domains within their cytoplasmic regions. Comparison of the IL-1 and IL-18 signaling mechanisms showed that they activate identical cytoplasmic messengers. IL-18, like IL-1, induced association of its receptor with IRAK and subsequent recruitment of TRAF6. IL-18 activated p38 MAP kinase, jun kinase, and beta casein kinase (TIP kinase), an apparently novel kinase previously thought to be specifically activated by IL-1 and tumor necrosis factor (TNF). IL-18 activated NF-kappaB in EL4/6.1 thymoma cells but not in COS-7 cells, even though the latter presumably contain all components required for the IL-1 signaling pathway. From our binding and signaling studies, we conclude that the IL-18 receptor complex consists of IL-18, the IL-1Rrp1, and another thus far unidentified receptor molecule.

Identification of a common class of high affinity receptors for both types of porcine interleukin-1 on connective tissue cells.

Interleukin-1 (IL-1) is the name given to the polypeptides produced by activated mononuclear phagocytes which were originally defined as lymphocyte activating factors (LAF). Administration of IL-1 in vivo causes fever and synthesis of acute phase proteins. In vitro they have been shown to cause cartilage and bone resorption, and to stimulate fibroblasts and chondrocytes to make prostaglandins and latent collagenase. IL-1 has therefore been proposed to be an important inflammatory mediator and may be involved in the destruction of cartilage and bone that is a feature of rheumatoid arthritis and other inflammatory diseases of joints. We therefore looked for IL-1 receptors on connective tissue cells which might be targets for therapeutic intervention. Here we report the iodination, to high specific activity and with retention of full biological potency, of the two types of natural porcine IL-1. These ligands have been used to demonstrate high affinity dissociation constant (approximately 10(-10) M) specific binding sites on pig chondrocytes and synovial fibroblasts, human dermal fibroblasts and murine osteoblasts (3,000-5,000 sites per cell). Most interestingly, the two different Il-1 proteins show a similar affinity for a common class of receptors.

The effect of a vitamin B-6 antagonist, 4-deoxypyridoxine, on the cross-linking of collagen in the developing chick embryo.

The vitamin B-6 antimetabolite 4-deoxypyridoxine, when injected into 13-day chick embryos, has the effect of increasing the amount of collagen solubilized from the leg bones by buffered saline solutions, 24 h after the injection. This effect is similar to, but less marked than, that produced by the administration of the lathyrogen beta-amino-propionitrile. Since that fraction of the total collagen which is solubilized by saline represents the least-cross-linked pool, it is concluded that 4-deoxypyridoxine is a lathyrogen, decreasing the cross-linking in the developing embryo, and confirming the importance of vitamin B-6 in that process. Lysyl oxidase, the cross-linking enzyme, was measured in extracts made from the epiphysial cartilages of embryos 24 h after the injection of either 4-deoxypyridoxine or beta-aminopropionitrile. The injection of 5 mg of beta-aminopropionitrile causes the lysyl oxidase activity to fall to 61% of that of saline-injected controls; after treatment with 4-deoxypyridoxine, the activity is 74% of the control value. In the latter case, full activity is not restored to the extracts by preincubation with pyridoxal phosphate. The results are discussed in relation to the early development of the connective tissues.

Studies on the fate of receptor-bound 125I-interleukin 1 beta in porcine synovial fibroblasts.

Binding of porcine interleukin 1, radiolabeled with Bolton-Hunter reagent (125I IL 1), to monolayers of porcine synovial fibroblasts (PSF) was found to be a temperature-dependent process. The rate of uptake and the amount of cell-associated ligand was higher at 37 degrees C than at 4 degrees C or 19 degrees C, and exceeded the apparent equilibrium binding capacity. The amount of bound 125I IL 1 that was removed by brief treatment with acidic buffers decreased from 80% at 4 degrees C to 35% for PSF incubated at 37 degrees C; this procedure was used to distinguish surface-bound from internalized ligand. In untreated PSF, surface binding was maximal at 1 hr and was maintained for at least 5 hr during which time the internal pool continued to increase. The lysosomotropic agent methylamine (20 mM) decreased surface binding by 50%; monensin (20 microM) decreased the rate and extent of internalization. Cycloheximide (10 micrograms/ml) did not affect ligand uptake, hence, continual expression of surface receptors could not be ascribed to their de novo synthesis. 40% of the radioactivity taken up by PSF during incubation at 37 degrees C subsequently appeared in the culture medium upon prolonged postincubation (5 hr) in the absence of added 125I IL 1: 60% of this fraction was trichloroacetic acid-soluble in untreated cultures, but the extent of degradation was halved by treatment with methylamine or monensin. Direct measurement of the rate of internalization of prebound 125I IL 1 was obtained by monitoring the formation of covalently cross-linked ligand-receptor complexes after warming PSF monolayers to 37 degrees C. By using gel electrophoresis we observed a decrease (t1/2 = 9 to 11 min) in labeling of the major cross-linkable species.

Down-modulation of epidermal growth factor receptor affinity in fibroblasts treated with interleukin 1 or tumor necrosis factor is associated with phosphorylation at a site other than threonine 654.

Interleukin 1 or tumor necrosis factor alpha can cause a transient down-modulation of epidermal growth factor (EGF) binding to quiescent fibroblast monolayers; the effect results from a reduction in EGF receptor (EGF-R) affinity and appears to be mediated by a protein kinase C (PKC)-independent mechanism. Here we show transient increases in EGF-R serine/threonine phosphorylation which are temporally coordinated with the effects on EGF binding; we also demonstrate that the cytokine-mediated phosphorylations, unlike those caused by PKC activators, have little discernible effect upon intrinsic EGF-R-associated tyrosine kinase activity. Cytokine-mediated EGF-R phosphorylation is resistant to staurosporine, an extremely potent inhibitor of PKC. Analysis of tryptic 32P-phosphopeptides reveals that Thr654, the unique site of PKC-mediated phosphorylation, is not phosphorylated in cytokine-treated cells, but a different, relatively acidic, peptide containing phosphoserine can be detected instead.

IL-1 and TNF transmodulate epidermal growth factor receptors by a protein kinase C-independent mechanism.

The effect of the human rIL-1 alpha and rTNF-alpha on the binding of 125I-labeled epidermal growth factor ([125I]EGF) to its receptor (EGF-R) has been studied in human gingival fibroblasts (HuGi). Incubation of these cells with recombinant cytokines at 37 degrees C caused a rapid, dose-dependent decrease in their ability to subsequently bind subsaturating levels of [125I]EGF at 4 degrees C. Inhibition was evident at 5 min after addition of cytokines, reached a maximal level (60-70% reduction) after 15 to 30 min, and declined thereafter. Normal EGF binding was attained by 2 h. Half-maximal inhibition of EGF binding occurred at 10 pM IL-1 and 50 pM TNF. The two cytokines were not additive in their effect. Competition experiments at 4 degrees C showed that the cytokines did not interact directly with EGF-R; Scatchard analysis of binding of [125I]EGF to HuGi after treatment with IL-1 and TNF revealed an increase in EGF-R Kd from 0.75 nM to 2.9 nM with no change in receptor number. The effect of IL-1 and TNF on EGF-R was compared with that of the tumor-promotor PMA which is known to "transmodulate" EGF-R affinity by activating protein kinase C which then phosphorylates EGF-R. PMA caused a greater inhibition of EGF binding to HuGi (80 to 85% inhibition; ED50 = 500 pM), and recovery of binding was much slower. Importantly, in HuGi made deficient in protein kinase C by prolonged incubation with PMA, addition of fresh PMA no longer affected EGF binding, while the response to IL-1 and TNF was intact. Cytokine- but not PMA-mediated EGF-R transmodulation was partially reversed by treatment of the cells with millimolar concentrations of the kinase inhibitor amiloride. HuGi were incubated with H3 32PO4, stimulated with PMA or cytokines, and EGF-R were immunoprecipitated; IL-1 and TNF, like PMA, caused a 2- to 5-fold increase in receptor phosphorylation. We conclude that occupation of IL-1 and TNF-R activates a protein kinase, distinct from kinase C, for which EGF-R is a substrate.

Mechanism for the decreased biosynthesis of cartilage proteoglycan in the scorbutic guinea pig.

Our previous work showed that vitamin C deficiency caused about a 70-80% decrease in the incorporation of [35S]sulfate into proteoglycan of guinea pig costal cartilage, coordinately with a decrease in collagen synthesis (Bird, T. A., Spanheimer, R. G., and Peterkofsky, B. (1986) Arch. Biochem. Biophys. 246, 42-51). We examined the mechanism for decreased proteoglycan synthesis by labeling normal and scorbutic cartilage in vitro with radioactive precursors. Proteoglycan monomers from scorbutic tissue were of a slightly smaller average hydrodynamic size than normal but there was no difference in the size of the glycosaminoglycan chains isolated after papain digestion. The type of glycosaminoglycans synthesized and the degree of sulfation were unaffected as determined by chondroitinase ABC digestion and duel labeling with [35S]sulfate and [3H]glucosamine. Conversion of [3H]glucosamine to [3H]galactosamine also was unimpaired. There was about a 40% decrease in core protein synthesis, measured by [14C]serine incorporation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Nevertheless, decreased incorporation of [35S]sulfate into scorbutic tissue persisted in the presence of p-nitrophenyl-beta-D-xyloside and cycloheximide, which indicated that the site of the scorbutic defect was beyond core protein synthesis and xylosylation. Galactosyltransferase activity in scorbutic cartilage decreased to about one-third the levels in control samples in parallel with the decreases in proteoglycan and collagen synthesis. Our results suggest that the step catalyzed by this enzyme activity, the addition of galactose to xylose prior to chondroitin sulfate chain elongation, is the major site of the scorbutic defect in proteoglycan synthesis. Decreased enzyme activity may be related to increased cortisol levels in scorbutic serum.

Regulation of collagen synthesis and mRNA levels in articular cartilage of scorbutic guinea pigs.

Previous studies suggested that decreased type I collagen synthesis in calvaria of ascorbate-deficient guinea pigs was correlated with weight loss rather than defective proline hydroxylation. The generality of this correlation was examined in articular cartilage, which synthesizes mainly type II collagen, by measuring collagen synthesis and proline hydroxylation in vitro in tissue from ascorbate-supplemented and scorbutic guinea pigs. Ascorbate concentrations in tissues were almost completely depleted after 1 week of deficiency, but proline hydroxylation remained normal until after approximately 3 weeks, when it had decreased only by 10%. At that point collagen synthesis had decreased to about 50% of the control value. There was little additional effect on proline hydroxylation but collagen synthesis decreased further to 20% of normal. Procollagen mRNA levels in cartilage, as measured by dot-blot hybridization with a type II-specific cDNA probe, were unchanged after 2 weeks of scurvy, which correlated with the lack of effect on collagen synthesis during that period. Thereafter, during the period when collagen synthesis decreased, procollagen mRNA levels decreased to 20% of control values. Refeeding ascorbate to acutely scorbutic animals led to reversal of defective proline hydroxylation within 24 h with a slower increase in collagen synthesis and mRNA levels. Collagen synthesis returned to the normal level after 4 days with no further increase, while mRNA levels continued to increase to 2.7 times the control values after 7 days. Thus the major mechanism for regulation of collagen synthesis in articular cartilage during scurvy and ascorbate repletion occurs independently of the effect on proline hydroxylation and is associated with changes in mRNA levels. The lack of precise coordination between collagen synthesis and mRNA levels during repletion, however, suggests that there may be additional regulation through post-transcriptional mechanisms.

Coordinate regulation of collagen and proteoglycan synthesis in costal cartilage of scorbutic and acutely fasted, vitamin C-supplemented guinea pigs.

The effects of ascorbic acid deficiency and acute fasting (with ascorbate supplementation) on the synthesis of collagen and proteoglycan in costal cartilages from young guinea pigs was determined by in vitro labeling of these components with radioactive proline and sulfate, respectively. Both parameters were coordinately decreased by the second week on a vitamin C-free diet, with a continued decline to 20-30% of control values by the fourth week. These effects were quite specific, since incorporation of proline into noncollagenous protein was reduced by only 30% after 4 weeks on the deficient diet. The time course of the decrease in collagen and proteoglycan synthesis paralleled the loss of body weight induced by ascorbate deficiency. Hydroxylation of proline in collagen synthesized by scorbutic costal cartilage was reduced to about 60% of normal relatively early, and remained at that level thereafter. Neither collagen nor proteoglycan synthesis was returned to normal by the addition of ascorbate (0.2 mM) to cartilage in vitro. Administration of a single dose of ascorbate to scorbutic guinea pigs increased liver ascorbate and restored proline hydroxylation to normal levels by 24 h, but failed to increase the synthesis of collagen or proteoglycan. Synthesis of both extracellular matrix components was restored to control levels after four daily doses of ascorbate. A 96-h total fast, with ascorbate supplementation, produced rates of weight loss and decreases in the synthesis of these two components similar to those produced by acute scurvy. There was a linear correlation between changes in collagen and proteoglycan synthesis and changes in body weight during acute fasting, scurvy, and its reversal. These results suggest that it is the fasting state induced by ascorbate deficiency, rather than a direct action of the vitamin in either of these two biosynthetic pathways, which is the primary regulatory factor.

Studies on the murine interleukin-1 receptor.

The multiple biological actions of interleukin-1 (IL-1) on its diverse range of target tissues is consequent upon interaction of the cytokine with specific, high affinity cell surface receptors. In this report we describe the covalent crosslinking of natural porcine IL-1 and recombinant human IL-1 to intact cells of the murine EL-4 6.1 [NOB-1] thymoma subline, and the solubilization of functional receptors from these cells. Crosslinking studies revealed the existence of two polypeptides, of 100 Kda and 80 Kda, which are involved in IL-1 recognition. Chromatographic studies and ligand-blotting of the soluble receptor demonstrated that the smaller of these two polypeptides, which appears to be a glycoprotein, is capable of interaction with both forms of IL-1.

Murine interleukin 1 receptor. Direct identification by ligand blotting and purification to homogeneity of an interleukin 1-binding glycoprotein.

Functional receptors (IL1-R) for the proinflammatory cytokine interleukin 1 (IL1) were solubilized from plasma membranes of the NOB-1 subclone of murine EL4 6.1 thymoma cells using the zwitterionic detergent 3[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS). Membrane extracts were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to nitrocellulose membranes, and "ligand blotted" with 125I-labeled recombinant human IL1 alpha in order to reveal proteins capable of specifically binding IL1. A single polydisperse polypeptide of Mr approximately equal to 80,000 was identified in this way, which bound IL1 alpha and IL1 beta with the same affinity as the IL1-R on intact NOB-1 cells (approximately equal to 10(-10) M). The IL1-binding polypeptide was only seen in membranes from IL1-R-bearing cells and did not react with interleukin 2, tumor necrosis factor alpha, or interferon. IL1-R was purified to apparent homogeneity from solubilized NOB-1 membranes by affinity chromatography on wheat germ agglutinin-Sepharose and IL1 alpha-Sepharose. Gel electrophoresis and silver staining of purified preparations revealed a single protein of Mr approximately equal to 80,000 which reacted positively in the ligand-blotting procedure and which we identify as the ligand-binding moiety of the murine IL1-R. Purified IL1-R exhibited the same affinity and specificity as the receptor on intact cells. The relationship of this protein to proteins identified by covalent cross-linking studies is discussed.

Decreased extracellular matrix production in scurvy involves a humoral factor other than ascorbate.

Our recent studies suggested that decreased collagen synthesis in bone and cartilage of scorbutic guinea pigs was not related to ascorbate-dependent proline hydroxylation. The decrease paralleled scurvy-induced weight loss and reduced proteoglycan synthesis. Those results led us to propose that the effects of ascorbate deficiency on extracellular matrix synthesis were caused by changes in humoral factors similar to those that occur in fasting. Here we present evidence for this proposal. Exposure of chick embryo chondrocytes to scorbutic guinea pig serum, in the presence of ascorbate, led to effects on extracellular matrix synthesis similar to those seen in scorbutic animals. The rates of collagen and proteoglycan synthesis were reduced to approximately 30-50% of the levels in cells cultured in normal guinea pig serum plus ascorbate, but proline hydroxylation and procollagen secretion were unaffected. Similar results were obtained with serum from fasted guinea pigs supplemented in vivo with ascorbate. The growth rate of the chondrocytes was not significantly affected by scorbutic guinea pig serum.

Similar hormonal changes in sera from scorbutic and fasted (vitamin C-supplemented) guinea pigs, including decreased IGF-I and appearance of an IGF-I reversible mitogenic inhibitor.

We previously proposed that the decreased rates of synthesis of collagen and proteoglycans in vitamin C-deficient guinea pigs were unrelated to the role of ascorbate in proline hydroxylation but might result from modulation of hormones known to change during fasting. In the present studies, we found that sera from guinea pigs on an ascorbate-free diet for 24-28 days or from those fasted for 4 days, with vitamin C supplementation, showed similar changes in the concentrations of several hormones. EGF and IGF-II concentrations were unchanged, but cortisol was increased 3-5 times and growth hormone was increased to approximately twice normal levels. Thyroxine and IGF-I concentrations were decreased to 40% and 25-33% of normal levels, respectively. The decrease in serum IGF-I must occur by a growth hormone-independent pathway. The extent of changes in hormone concentrations in sera from ascorbate-deficient guinea pigs was correlated with the extent of weight loss. Sera from scorbutic and fasted guinea pigs failed to stimulate DNA synthesis in quiescent BALB 3T3 cells in the presence of saturating concentrations of EGF and PDGF. Addition of experimental sera to normal serum showed that lack of mitogenic activity was due to the presence of an inhibitor. Inhibition was not related to IGF-I concentrations in the sera, although it was reversed by the addition of IGF-I to sera from scorbutic or fasted animals. These results support our proposed model and suggest that IGF-I, as well as an inhibitor of its activity, plays a role in the regulation of growth by vitamin C and other nutrients.

The inhibition of lysyl oxidase in vivo by isoniazid and its reversal by pyridoxal. Effect on collagen cross-linking in the chick embryo.

Isonicotinic acid hydrazide (isoniazid) causes a large increase in the salt-solubility of collagen when injected into chick embryos; this change is accompanied by the inactivation of lysyl oxidase (EC 1.4.3.13), the enzyme responsible for initiating cross-link formation in collagen and elastin. In addition, isoniazid markedly decreases the liver content of pyridoxal phosphate. The depletion of pyridoxal phosphate takes approx. 6 h, whereas the inhibition of lysyl oxidase and the increase in collagen solubility occur more slowly. A reversal of these effects of isoniazid can be produced by the subsequent injection of a stoichiometric amount of pyridoxal, supporting the role of pyridoxal as a cofactor for lysyl oxidase. Treatment of chick embryos with beta-aminopropionitrile, an irreversible inhibitor of lysyl oxidase, causes an inhibition of the enzyme, which begins to recover within 24 h but which is not affected by the administration of pyridoxal; with isoniazid inhibition, however, lysyl oxidase activity does not show any sign of recovery by 48 h. It is proposed that isoniazid may cause the inhibition of lysyl oxidase by competing for its obligatory cofactor, pyridoxal phosphate. The potential clinical implications in the therapeutic control of fibrosis are briefly discussed.

Phorbol ester induces phosphorylation of the 80 kilodalton murine interleukin 1 receptor at a single threonine residue.

The cytoplasmic domains of some cell surface receptors become phosphorylated in cells treated with phorbol esters. The present study was undertaken in order to determine whether this is also true of the 80 kDa interleukin 1 receptor (IL1R). Recombinant murine IL1R, transfected into chinese hamster ovary (CHO) cells or murine fibroblasts, was immunoprecipitated from [32P]orthophosphate-labelled cells. IL1R phosphorylation was only detected in cells pretreated with phorbol 12-myristate 13-acetate (PMA) and occurred solely on phosphothreonine. In contrast to a previous report, little or no IL1R phosphorylation occurred in response to IL1. By using a truncated receptor and receptors in which threonine residues were changed to alanines, we established that Thr537, near the carboxy-terminus, is the major site of PMA-induced phosphorylation. The human IL1R has a different sequence at this locus, and is apparently not phosphorylated. Binding studies showed that PMA-induced phosphorylation had no discernible effect on ligand binding or internalization.

Oncostatin M and leukemia inhibitory factor trigger overlapping and different signals through partially shared receptor complexes.

Leukemia inhibitory factor (LIF) and oncostatin M (OSM) both bind to the same receptor with high affinity and thus mediate an overlapping spectrum of biological activities, the signal transduction mechanisms for which are unclear. We show that mitogen-activated protein kinases are involved in both the LIF and OSM signal transduction pathways. However, we found that OSM is a much more potent inducer of both mitogen-activated protein kinase activity and biological response, both of which correlate with the expression of a second OSM receptor that does not bind LIF. In addition, different patterns of tyrosine-phosphorylated proteins were stimulated by OSM and LIF. We therefore suggest that the two receptors for OSM can be coupled to different signal transduction events.