Phosphatidylinositol 3′-Kinase-independent p70 S6 Kinase Activation by Fibroblast Growth Factor Receptor-1 Is Important for Proliferation but Not Differentiation of Endothelial Cells*

p70s6k has a role in cell cycle progression in response to specific extracellular stimuli. The signal transduction pathway leading to activation of p70s6k by fibroblast growth factor receptor-1 (FGFR-1) was examined in FGF-2-treated rat L6 myoblasts. p70s6k was activated in a biphasic and rapamycin-sensitive manner. Although phosphatidylinositol 3′-kinase was not activated in the FGF-2 treated cells, as judged fromin vitro and in vivo analyses, wortmannin and LY294002 treatment inhibited p70s6k activation. Inhibition of protein kinase C (PKC), by bisindolylmaleimide or by chronic phorbol ester treatment of the FGFR-1 cells, suppressed but did not block p70s6k activation. In cells expressing a point-mutated FGFR-1, Y766F, unable to mediate PKC activation, p70s6k was still activated, in a bisindolylmaleimide- and phorbol ester-resistant manner. The involvement of S6 kinase in FGFR-1-dependent biological responses was examined in murine brain endothelial cells. In response to FGF-2, these cells differentiate to form tube-like structures in collagen gel cultures and proliferate when cultured on fibronectin. p70s6k was not activated in endothelial cells on collagen, whereas activation was observed during proliferation on fibronectin. In agreement with this finding, rapamycin inhibited the proliferative but not the differentiation response. Our results indicate that FGFR-1 mediates p70s6k activation by a phosphatidylinositol 3′-kinase-independent mechanism that does not require PKC activation and, furthermore, proliferation, but not differentiation of endothelial cells in response to FGF-2, is associated with p70s6k activation.

The p70 S6 kinase (p70 s6k ) 1 is a serine/threonine kinase that phosphorylates 40 S ribosomal protein S6, in response to a number of extracellular stimuli (1,2). The two isoforms of p70 s6k , the 70-kDa s6k ␣II (cytosolic form) and the 85-kDa s6k ␣I (nuclear form), are derived from alternatively spliced products (3,4) from a single gene (3). Extracellular stimuli induce acute phosphorylation on multiple serine and threonine residues within p70 s6k , which are associated with its activation. Four of these residues are located in the carboxyl terminus of p70 s6k (5). These sites are potential mitogen-activated protein kinase targets. However, mitogen-activated protein kinase fails to activate p70 s6k after phosphorylation of these sites in vitro (6), and p70 s6k activation lies on a Ras-independent pathway (7,8). Protein kinase C (PKC), phosphatidylinositol 3-kinase (PI3-kinase), and protein kinase B have been implicated as upstream signaling molecules of p70 s6k activation in insulin-, platelet-derived growth factor (PDGF)-, epidermal growth factor (EGF)-and interleukin-2-treated cells (9). Recently, p70 s6k was moreover shown to complex with and to be activated by the GTP-binding Rho family proteins Rac1 and Cdc42 (10). However, the relative contribution of these pathways to activation of p70 S6K is unclear (8).
Rapamycin is a potent and specific inhibitor of p70 s6k , preventing phosphorylation and activation of p70 s6k by all known external stimuli (11)(12)(13)(14)(15). After binding of rapamycin to its cellular receptor, the FK506-binding protein-12 (FKBP-12), this complex targets TOR kinases in Saccharomyces cerevisiae or the related protein FKBP-12-rapamycin-associated protein/ rapamycin-FKBP target 1/mammalian TOR in mammalian cells (16 -18). Inactivation of p70 s6k by rapamycin is associated with selective dephosphorylation of a unique set of serine and threonine sites, flanked by large aromatic residues, in p70 s6k (19). Rapamycin is known to inhibit growth of many types of cells; causing G 1 arrest in T lymphocytes and delaying entry into S phase in fibroblasts (11,12). Microinjection of a neutralizing antibody against p70 s6k or p85 s6k has also been shown to block the entry into S phase of injected cells (20,21). These results indicate that p70 s6k activation is important for cell cycle progression.
Fibroblast growth factors (FGF) are heparin-binding polypeptide growth factors, which form a family of nine members (22). Extracellular signaling by FGFs is transduced via specific receptor tyrosine kinases, denoted FGF receptor-1 to -4 (23,24). Heparin and heparan sulfate proteoglycans are known to modulate ligand binding to the receptor tyrosine kinase. Binding of FGFs to the receptor tyrosine kinase leads to receptor dimerization and activation of the kinase domain, followed by autophosphorylation of the receptor and association with downstream signaling components. Thus far, only one Src homology 2 (SH2) domain-containing protein, phospholipase C-␥ (PLC-␥), has been shown to bind directly to FGF receptor-1, via a carboxyl-terminal autophosphorylation site at Tyr-766 in the receptor (25). The FGF receptors are known to mediate a variety of cellular responses, such as cell proliferation, migration, and differentiation (23,24). We have previously shown that murine brain capillary endothelial cells respond to FGF-1 and FGF-2 treatment either by proliferation or by differentiation, the latter visualized in vitro as tube formation of cells cultured in collagen gels (26). It is likely that several distinct signal transduction pathways, coupling directly or indirectly to the receptor, contribute to establish these responses. p70 s6k has been shown to be activated in FGF-treated cells (27). In this paper, we have used different inhibitors of signal transduction pathways, known to contribute to p70 s6k activation, as well as a mutant FGF receptor unable to bind PLC-␥, to characterize FGF-induced p70 s6k activation biochemically and to investigate its function in cellular responses to FGF.

MATERIALS AND METHODS
Cell Culture-Rat L6 myoblasts expressing wild-type FGF receptor-1 and Y766F point-mutated FGF receptor-1 (28) were cultured in Dulbecco's modified Eagle's medium (DMEM; Life Technologies, Inc., London, UK) supplemented with 10% fetal bovine serum, at 37°C. The IBE cell line is a capillary endothelial cell line established from H-2K b -tsA58 SV40 large T transgenic mouse (Immortomouse) brain (26). The parental IBE cells were cultured routinely in Ham's F12 medium containing 5 g/ml insulin, 10 ng/ml epidermal growth factor (EGF), 20 units/ml mouse interferon-␥ (IFN-␥), 75 g/ml endothelial growth supplement, and 20% heat-inactivated fetal bovine serum. Dependent on the assay, as outlined below, IBE cells were cultured on dishes coated with 20 g/ml human plasma fibronectin (Sigma), on dishes covered with type I collagen gels (denoted collagen-coated dishes), or, alternatively, in the middle of two layers of type I collagen gels (tube formation assay). To prepare collagen-coated dishes, 4 volumes of type I collagen solution (Vitrogen 100; Celtrix Pharmaceuticals, Inc. Santa Clara, CA), 4 volumes of 0.012 M HCl, 1 volume of 10 ϫ concentrated Ham's F-12, and 1 volume of concentrated buffer (260 mM NaHCO 3 , 200 mM HEPES, 50 mM NaOH) were mixed at 4°C, and the mixture was poured into 6-cm dishes (1.5 ml per dish), which were incubated at 33°C.
Signal Transduction Pathway Inhibitors-Rapamycin (Biomol, Plymouth Meeting, PA) was dissolved in ethanol at a concentration of 20 mg/ml as a stock solution and kept at Ϫ20°C. This solution was diluted using culture medium and added to the cells at indicated concentrations 30 min before FGF-2 stimulation. Wortmannin (Sigma) was dissolved in Me 2 SO at a concentration of 1 mM. After further dilution in Me 2 SO, 100 nM (final concentration) was added to the cells 30 min before FGF-2 stimulation. Bisindolylmaleimide (Calbiochem) was dissolved in Me 2 SO at a concentration of 2.4 mM and kept at Ϫ20°C. Prior to use, this solution was diluted 20-fold in Me 2 SO, and 120 nM (final concentration) was added to the cells 30 min before FGF-2 treatment. Phorbol 12-myristate 13-acetate (PMA; Calbiochem) was dissolved in Me 2 SO at a concentration of 5 mM and kept at Ϫ70°C. For downregulation of PKC, 5 M (final concentration) was added to the cells and the culture was continued for 24 h, at which point FGF-2 stimulation was performed.
Analysis of Receptor Tyrosine Kinase Activity-Receptor tyrosine kinase activity was measured as autophosphorylation of receptor protein in vitro as described previously (26). Briefly, cells were cultured in 10-cm dishes, and the culture medium was replaced by DMEM containing 0.1% FBS and cultured overnight. Cells were either unstimulated or stimulated by 100 ng/ml FGF-2 for 8 min, and cell lysate was prepared in 20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 0.1 mM sodium orthovanadate, 100 units/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride (PMSF), 2.5 mM EDTA, and 1 mM dithiothreitol. After incubation of lysate with anti-FGF receptor-1 antiserum (29), receptor protein was precipitated followed by incubation with Protein A-Sepharose beads. After washing, beads were incubated with [␥-32 P]ATP on ice for 10 min, and proteins were separated by SDSpolyacrylamide gel electrophoresis (SDS-PAGE). The gel was fixed and treated with 1 M KOH at 55°C for 35 min and then dried and exposed on x-ray film (Fuji).
p70 s6 Kinase Activity Assay-The p70 s6k activity assay was performed as described previously (3) with some modifications. L6 cells cultured on 6-cm dishes were either unstimulated or stimulated with 100 ng/ml ligand for 10 or 60 min. IBE cells were inoculated on either fibronectin-or collagen-coated 6-cm dishes with Ham's F-12 medium containing 0.25% BSA, cultured for 4 h at 33°C, and then stimulated with 100 ng/ml FGF-2, for indicated times. Cells were washed with extraction buffer (EB), which is composed of 50 mM Tris-HCl, pH 8.0, 120 mM NaCl, 20 mM NaF, 1 mM benzamidine, 1 mM EDTA, 6 mM EGTA, 15 mM Na 4 P 2 O 7 , 30 mM 4-nitrophenyl phosphate, 0.1 mM PMSF. After washing, cells were lysed in EB containing 1% Nonidet P-40 on ice. Collected cell lysate was frozen in liquid nitrogen and kept at Ϫ80°C until assay. After thawing, the lysate was centrifuged at 10,000 ϫ g for 30 min and then 100 g of lysed protein was incubated with the anti-p70 s6k antiserum M5 (3) at 4°C for 2 h, followed by precipitation with Protein A-Sepharose beads. Beads were washed 3 times with EB containing Nonidet P-40, once with dilution buffer (DB; 50 mM MOPS, pH 7.2, 5 mM MgCl 2 , 0.2% Triton X-100, 1 mM dithiothreitol) and resuspended in 5 l of DB. The kinase reaction was initiated by addition of 5 l of DB containing 200 M ATP, 10 mM 4-nitrophenyl phosphate, 10 g of 40 S ribosomal protein, and 3 Ci of [␥-32 P]ATP to the beads. The reaction proceeded at 37°C for 30 min and was stopped by addition of SDS sample buffer, followed by electrophoresis in 15% SDS-polyacrylamide gels. After electrophoresis, the gel was fixed and dried and then exposed on an Image Analyzer screen (Fuji).
PI3-Kinase in Vivo Assay-L6 cells expressing FGFR-1 were washed with phosphate-free DMEM containing 0.1% fatty acid-free BSA, 0.0375% sodium bicarbonate, and 20 mM HEPES, pH 7.4, and labeled for 90 min in medium containing 300 Ci/ml of [ 32 P]P i . Washed cells were stimulated with FGF-2 for 5 min, and the lipids were extracted into chloroform/methanol, deacylated using monomethylamine, and deglycerated using periodate as described (30). The generated inositol phosphates (which corresponded to the inositol lipids) were separated by anion exchange high performance liquid chromatography on a 25-cm partisphere 5SAX column, eluted with a linear gradient of ammonium dihydrogen phosphate (0.5 M, pH 3.8) at 1 ml/min over 110 min. Fractions were collected every 0.5 min and [ 32 P] determined by scintillation counting; peak retention times were compared with authentic 3 H-labeled standards which were run every 5th injection.
Diacylglycerol (DAG) Kinase-linked Assay-Changes in DAG mass were determined by the DAG kinase-linked assay as described (31). Briefly, cells were stimulated as above; the medium was aspirated, and incubations were terminated by the addition of ice-cold methanol. The lipids were extracted into chloroform/methanol and dried in vacuo. The solubilized lipids were then incubated with Escherichia coli DAG kinase and [␥ 32 P]ATP; the generated [ 32 P]phosphatidate was separated by thin layer chromatography and analyzed using a PhosphorImager, and the mass of DAG was determined by comparison to a standard curve generated in parallel.
Cell Proliferation Assay-Mouse brain capillary endothelial cells (IBE cells) were inoculated into 24-well culture plates coated with human plasma fibronectin at a density of 1.5 ϫ 10 4 cells/cm 2 (3 ϫ 10 4 cells/well) in Ham's F-12 medium containing 5% FBS and cultured at 33°C. The next day, medium was changed to Ham's F-12 medium containing 2% FBS, or 0.2% FBS, with either vehicle (ethanol) or rapamycin. Twenty minutes later, 1 ng/ml FGF-2 was added as indicated, and the culture was continued for 3 days. Cells were detached from the well by trypsinization, and the cell number was counted by a hemocytometer.
Tube Formation Assay-Tube formation assays were performed as described previously (26). In brief, IBE cells were inoculated on the first layer of collagen gels made in wells of 12-well plates at a density of 8 ϫ 10 5 cells/well in Ham's F-12 medium containing 0.25% BSA with or without rapamycin. Thirty minutes later, 5 ng/ml FGF-2 was added to indicated cells. Culture was continued for additional 3.5 h, at which point the medium was removed and the second layer of collagen was added onto the cells. After gelation of the second layer of collagen, Ham's F-12 medium containing 0.25% BSA and rapamycin or vehicle with or without FGF-2 was added and cells were cultured overnight. Photographs were taken under phase-contrast microscopic examination.

RESULTS
FGF Receptor-1 Transduces Signals for Activation of p70 s6k -p70 s6k is known to be activated in response to various mitogens, including insulin, PDGF, EGF, and serum (9). Agonist-stimulated activation of p70 s6k is biphasic, with a rapid transient peak, followed by a sustained plateau (27,32). The relevance of the first peak is not clear, but the sustained phase is thought to represent biological activity (32). Several distinct signal transduction pathways have been implicated in p70 s6k activation. We used transfected L6 myoblasts, expressing either the wild-type FGF receptor-1, or a point-mutated FGF receptor-1, Y766F, to examine signal transduction pathways leading to p70 s6k activation in response to FGF. Fig. 1A shows that wild-type and Y766F FGF receptor-1transfected L6 cells expressed similar levels of receptors and that the receptors responded to ligand stimulation with induction of tyrosine kinase activity. In contrast, the parental cells lack detectable expression of FGF receptors, and FGF-2 stimulation of the untransfected parental cells failed to induce p70 s6k activation (data not shown). p70 s6k activation was examined by immunoprecipitation of p70 s6k , from unstimulated and FGF-2-stimulated cells, followed by incubation in kinase buffer and [␥-32 P]ATP, in the presence of 40 S ribosomes, serving as a substrate for the immunoprecipitated p70 s6k . After SDS-PAGE of the samples, 32 P-labeled 40 S ribosomal protein s6 was quantified using a PhosphorImage Analyzer. p70 s6k activation was analyzed after 10 and 60 min of FGF-2 treatment, to measure early and sustained phases, respectively. Fig.  1B shows that p70 s6k was activated in a sustained manner in FGF-2-stimulated L6 cells expressing the wild-type FGF receptor-1. In the mutant Y766F receptor expressing cells (Fig. 1B), the level of p70 s6k activity was comparatively lower at 10 min treatment. However, at 60 min, similar induction (250 -300%) of p70 s6k activity was seen in the wild-type and mutant FGF receptor-1 expressing cells. These results indicate that signals for p70 s6k activation can be transduced via FGF receptor-1 and that the major autophosphorylation site, Tyr-766, which is required for binding and activation of PLC-␥, is not obligatory for FGF receptor-1-mediated p70 s6k activation.
Effects of Signal Transduction Pathway Inhibitors on p70 s6k Activation-Different signal transduction pathways have been implicated in p70 s6k activation, and inhibition of the functions of PI3-kinase and PKC have been used to demonstrate roles for pathways involving these enzymes. We analyzed the effects of rapamycin, wortmannin (PI3-kinase inhibitor), and bisindolylmaleimide (PKC inhibitor) in FGF-2-induced p70 s6k activation. In addition, chronic treatment of cells with PMA was used to down-regulate PKC. Fig. 2 shows that treatment of the wildtype FGFR-1 expressing L6 cells with either of these four different drugs attenuated FGF-2-induced p70 s6k activation, both at 10 and 60 min. Rapamycin failed to bring the activity of p70 down to basal, even at the relatively high dose of 100 ng/ml. Treatment of cells with another PI3-kinase inhibitor, LY294002, brought p70 s6k activity down to the basal level when used at a concentration of 30 M (data not shown).
We have previously shown that activation of FGF receptor-1, expressed in porcine aortic endothelial cells, does not lead to activation of PI3-kinase in vitro (33). In agreement, we failed to detect activation in vitro of PI3-kinase in L6 myoblasts expressing FGF receptor-1 (Fig. 3). The experiment was performed using L6 myoblasts expressing FGF receptor-1, which were stimulated for 5 or 30 min with FGF-2. As a positive control, PDGF-BB was used to stimulate PDGF receptors, endog- FIG. 1. A, wild-type and Y766F point-mutated FGF receptor-1-transfected L6 cells express similar levels of kinase active receptors. Lysates from stimulated (ϩ) or unstimulated (Ϫ) L6 cells transfected with either wild-type or Y766F mutant FGF receptor-1 were immunoprecipitated with anti-FGF receptor-1 antiserum and incubated in the presence of kinase buffer and [␥-32 P]ATP, and the samples were separated by SDS-PAGE. The gel was fixed, KOH-treated, dried, and exposed. wt, L6 cells expressing wild-type FGF receptor-1; Y766F, L6 cells expressing Y766F point-mutated FGF receptor-1. B, both wild-type and Y766F point-mutated FGF receptor-1 transduce signals for activation of p70 s6k . Lysates from stimulated (10 or 60 min) or unstimulated L6 cells transfected with either wild-type or Y766F point-mutated FGF receptor-1 were immunoprecipitated with anti-p70 s6k antiserum, M5. The precipitates were incubated in the kinase buffer and [␥-32 P]ATP in the presence of 40 S ribosomal protein, and the samples were separated by SDS-PAGE. The gel was fixed, dried, and exposed on an Image Analyzer screen. Incorporation of 32 P into the 40 S protein was quantified by Image Analyzer software. enously expressed in the L6 cells. Cells were lysed and immunoprecipitated with phosphotyrosine antibodies, followed by analysis for phosphorylation of phosphatidylinositol by thin layer chromatography. As expected, the PDGF-stimulated cells contained active PI3-kinase, indicated by the 5-fold increase in [ 32 P]PIP relative to basal. In the FGF-2-treated cells, the levels of [ 32 P]PIP did not change as compared with the control, neither at 5 nor 30 min of treatment. Similar results were obtained using antibodies against the p85 regulatory subunit of PI3-kinase for immunoprecipitation (data not shown).
Furthermore, analysis of in vivo 32 P-labeled L6 cells expressing FGFR-1 showed that FGF-2 also did not stimulate PI3kinase activity as determined by changes in 3Ј-phosphorylated lipids ( Table I). The table shows that FGF-2 stimulated significant turnover of inositol phospholipids with a greater than 2-fold increase in the radioactivity incorporated into PIP within 5 min and an approximate 50% increase in radioactivity associated with PI(4,5)P 2 . There was a small increase in radioactivity associated with PIP 3 ; however, the level of PIP 3 is extremely low and the apparent change is probably due to an increase in basal PI3-kinase activity acting upon the elevated [ 32 P]PI(4,5)P 2 ; indeed, the radioactivity associated with PIP 3 compared with PIP 2 remains constant at approximately 0.03%. These data strongly imply that inhibition by wortmannin of p70 s6k activation in the FGF-2-stimulated cells did not involve the classical PI3-kinase.
Role of PKC in p70 s6k Activation-The Y766F FGF receptor-1 mutant lacks the ability to mediate phosphorylation and activation of PLC-␥. Active PLC-␥ hydrolyzes PIP 2 to inositol 1,4,5-trisphosphate and diacylglycerol (DAG), leading to intracellular Ca 2ϩ fluxes and activation of PKC, respectively. In the Y766F mutant FGF receptor-1 expressing cells, PKC could potentially still be activated in response to FGF-2, via PLC-␥independent DAG formation. We therefore analyzed formation of DAG, by use of a DAG kinase-linked assay, on FGF-2stimulated wild-type or Y766F FGFR-1 expressing L6 cells (Table II). In the wild-type receptor expressing L6 cells DAG formation was increased approximately 2-fold, in response to FGF-2, at 10 and 60 min of stimulation. A similar level of DAG formation was induced by vasopressin treatment (data not shown). In the mutant Y776F receptor expressing L6 cells, DAG formation did not increase in response to FGF-2 stimulation (Table II).
As shown in Fig. 1B, the extent of p70 s6k activity at the mitogenically relevant, sustained phase was very similar in wild-type FGF receptor-1 and Y766F mutant receptor expressing cells. Moreover, Fig. 4 shows that the level of p70 s6k activity in the Y766F mutant receptor expressing cells was not affected by treatment with bisindolylmaleimide, to inhibit PKC, or chronic exposure to PMA, to deplete cellular PKC levels. In these cells, rapamycin (Fig. 4) and wortmannin (data not shown) still had strong inhibitory effect on p70 s6k activation. These data indicate that in the absence of PKC activation,  -20). Precipitated proteins were incubated with phosphatidylinositol (PI) and [␥-32 P]ATP, followed by extraction of lipids and separation by thin layer chromatography. The TLC plate was exposed on an Image Analyzer screen, and the extent of 32 P incorporation was measured. The plate was then reexposed on an x-ray film. PIP, phosphatidylinositol phosphate. p70 S6K Activation via FGF Receptor-1 rapamycin-and wortmannin-sensitive pathways were modulated to compensate for the absence of the PKC-dependent pathway for p70 s6k activation.
Role of p70 s6k in Biological Responses-It has been reported that p70 s6k is involved in G 1 -S transition in the cell cycle (11,20) and therefore crucial for cell proliferation. We examined the effect of inhibiting FGF-2-stimulated p70 s6k in a murine brain capillary endothelial cell line undergoing proliferation or differentiation. The cell line was established from H-2K b -tsA58 SV40 large T transgenic mice (26). We have shown that, dependent on the growth conditions, the cells will either differentiate (i.e. form tube-like structures when grown on a collagen-coated dish or in three-dimensional collagen gels) or proliferate (when grown on fibronectin-coated dishes) in response to FGF-2. Cells cultured on fibronectin-coated dishes never showed a differentiated phenotype. On the other hand, cells grown on a collagen-coated dishes failed to grow in response to FGF-2 treatment as judged from lack of increase in labeling index. 2 We analyzed the parental and chimeric receptor expressing endothelial cells for their level of activated p70 s6k in response to FGF-2 treatment. As seen in Fig. 5, prominent activation of p70 s6k was seen in cells grown on fibronectin. A considerably weaker activation of p70 s6k was seen in cells grown on collagen gels (in cells stimulated for 30 min the extent of p70 activation was less than 30% of that demonstrated in cells cultured on fibronectin).
Next, we treated the endothelial cells, grown on fibronectin or in collagen gels, with rapamycin to block p70 s6k activation in response to FGF-2. The effects of rapamycin on proliferation was analyzed by assessing the increase in cell number of cells in fibronectin-coated wells. As seen in Fig. 6, rapamycin treatment inhibited proliferation of cells treated with FGF-2 in 0.2% serum (data not shown) or 2% serum (Fig. 6) in a dose-dependent manner. However, tube formation of endothelial cells in collagen gels in response to FGF-2 was not affected even by a high dose of rapamycin (Fig. 7), in agreement with the inefficient p70 s6k activation, in these cultures. DISCUSSION Here we show that FGF receptor-1 mediates activation of p70 s6k in different cell types and that treatment of the cells with different drugs, including rapamycin, wortmannin, and bisindolylmaleimide, inhibits p70 s6k activation. These drugs are known to attenuate signal transduction pathways involved in regulation of the serine/threonine kinase. However, the relationship between these pathways remains unclear. The immunosuppressant drug rapamycin forms a complex with FKBP-12, a peptidyl-propyl cis-trans-isomerase that may be involved in protein folding, and the complex binds and inhibits the function of yeast and mammalian TOR (9). The TOR proteins, which recently have been identified, all have a structural domain similar to PI3-kinase and PI4-kinase; however, phylogenetic analysis shows that the TOR proteins constitute a distinct family of lipid kinases (16). According to recent reports, mammalian TOR, derived from rat brain, possesses PI4-kinase activity (34,35). Whether this activity is intrinsic to TOR has been disputed, based on the finding that kinase-inactive TOR possessed intact level of lipid kinase activity, as compared with wild-type TOR (18). In addition, the PI-4 kinase activity was not affected by rapamycin (see review, Ref. 36). In any case, TOR appears to be obligatory for S6 kinase activation, since rapamycin inhibits S6 kinase activation independent of cell The medium was changed to Ham's F-12 medium supplemented with 2% FBS with vehicle or rapamycin and after 20 min incubation; 1 ng/ml FGF-2 was added to indicated wells, and the culture was continued for 3 days. Cells were trypsinized and counted. Data are expressed as the means Ϯ S.D. for triplicate wells. Similar results were obtained when analyzing cells starved in 0.2% FBS. type and stimulus (11,12). The rapamycin-FKBP-12 complex does not appear to inhibit the kinase activity of TOR proteins, at least in yeast. Instead, inhibition might be exerted on the level of TOR protein binding to or phosphorylation of G 1 effectors (16). How TOR couples to upstream elements, such as receptor tyrosine kinases, has not been elucidated. It also remains to be shown whether TOR directly regulates S6 kinase activity. Indeed, recent data show that an amino-terminally truncated mutant of S6 kinase, which no longer is inhibited by rapamycin, still becomes phosphorylated at the critical regulatory Thr-389 site in response to serum, indicating that TOR is not involved in this phosphorylation (37).
The level of activation of p70 s6k by FGF was about 3-fold as compared with basal, in different cell types; this is about half the level of activation of p70 s6k seen in PDGF-stimulated cells (data not shown). This difference in efficiency between FGFand PDGF-stimulated p70 s6k activation might be due to PDGF initiating multiple pathways leading to p70 s6k activation. Thus, previous studies using the fungal metabolite wortmannin, to study the regulation of p70 s6k activation in PDGF stimulated cells, has implicated PI3-kinase as an upstream regulator of p70 s6k (38). In agreement, expression of constitutively active PI3-kinase resulted in substantial activation of p70 (39). Rapamycin and wortmannin inhibit phosphorylation at the same set of serine sites in p70 s6k , indicating that these drugs target the same pathway (15). However, since susceptibility to wortmannin and rapamycin has been mapped to different domains of p70 s6k (40) the input signals that these drugs modulate appear to be different. The serine/threonine kinase c-Akt/ protein kinase B has been postulated to bridge between PI3kinase and p70 s6k , possibly with several intermediates (41,42). In this work, wortmannin was used to inhibit FGF receptor-1 mediated p70 s6k activation, in the absence of any detectable activation, in vitro or in vivo, of PI3-kinase. Despite the failure to activate PI3-kinase, FGFR-1 expressed in L6 cells mediated activation of Akt/protein kinase B in response to FGF-2 (data not shown). The ability of wortmannin to inhibit S6 kinase activity independent of PI3-kinase was also shown by Hara et al. (43). In cells expressing ⌬p85, a dominant negative form of the regulatory subunit of PI3-kinase, S6 kinase was still activated in response to insulin. In contrast, wortmannin treatment attenuated insulin-induced S6 kinase activity. p70 s6k itself appears not to inhibited by wortmannin (35); different reports show lack of effect on TOR/FKBP-12-rapamycin-associated protein (18) or inhibition of TOR autokinase activity (44) by wortmannin. Wortmannin inhibits PI3-kinase at nM concentrations (45) but also affects other enzymes, such as phospholipase A 2 (30). The PI3-kinase inhibitor Ly294002 also efficiently inhibits S6 kinase activation in FGF-2-stimulated L6 cells expressing FGFR-1 (data not shown). Although LY294002 has been suggested to be a more specific PI3-kinase inhibitor, the range of targets for this drug as well as wortmannin is still not clear (46,47). Thus, it is possible that the true target of wortmannin in the S6 kinase pathway is not PI3-kinase but a so far unidentified lipid kinase, or another type of enzyme, upstream of S6 kinase.
PKC is activated in response to FGF receptor-1 stimulation through PLC-␥-catalyzed hydrolysis of PIP 2 to DAG and phosphatidylinositol 3,4,5-trisphosphate (48). DAG is also formed as a result of hydrolysis of phosphatidylcholine; the kinetics of phosphatidylcholine-generated DAG formation are different, with a later and sustained peak than for PIP 2 -derived DAG. One of the enzymes responsible for generation of phosphatidylcholine-derived DAG is phospholipase D, which is activated by FGF-2. In cells expressing an FGF receptor-1 mutant lacking the major autophosphorylation site, Tyr-766, neither PLC-␥ (49, 50) nor phospholipase D 3 is activated in response to FGF-2. In consequence, DAG formation and PKC activation are suppressed. S6 kinase p70 activation in FGF-2-stimulated L6 cells expressing the Y766F mutant was only moderately affected (Fig. 1B). Treatment of cells expressing either the wild-type or mutant receptors with bisindolylmaleimide or PMA inhibited p70 S6k in the wild-type receptor cells but had no effect in cells expressing the mutant receptor. We infer from these data that other pathways, such as the rapamycin-and wortmannin-sensitive pathways, for p70 s6k activation were up-regulated in response to FGF-2 stimulation of the Y766F mutant, as compared with the wild-type FGFR-1. It has been shown that p70 s6k activity is induced in response to insulin, even though PKC is not activated by this mitogen. Furthermore, inhibition of PKC attenuates EGF-induced but not PDGF-induced p70 s6k activation (51). Thus, PKC appears to play a modulatory role in p70 s6k activation but is not absolutely required.
The p70 s6k -mediated phosphorylation of the 40 S ribosomal subunit S6 has been implicated in the selective up-regulation of a family of essential gene products. Inhibition of S6 phosphorylation, using neutralizing antibodies, or treatment with rapamycin, leads to a block in cell cycle progression (2,9). However, although rapamycin blocks p70 s6k activation in most cell types, it impedes cell cycle progression preferentially in hematopoietic cells (11). This has been suggested to be due to the redundancy of signal transduction pathways involved in the proliferative response. FGF-induced proliferation of an endothelial cell line was efficiently blocked by rapamycin. Dependent on the culture conditions, these cells form tubes in response to FGF-2 treatment. This response was not affected by rapamycin, which is in agreement that tube formation is independent of DNA synthesis. 4 A critical task for the future is to identify the point in G 1 progression, which is targeted by S6 kinase. FIG. 7. Rapamycin has no effect on differentiation (tube formation) of murine brain capillary endothelial cells cultured in collagen gels. Cells were cultured between two layers of collagen gels in the presence or absence of 5 ng/ml FGF-2 and/or rapamycin treatment for 18 h.