If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by National Science Council of the Republic of China, Taiwan, Grants NSC 92-2320-B006-080 (to S.-J. T.) and NSC 92-2320-B006-038 (to L.-Y. C. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Fibroblast growth factor-9 (FGF9) is a potent mitogen that stimulates normal and cancer cell proliferation though the signaling mechanism is not fully understood. In this study, we aimed to unravel the signaling cascades mediate FGF9 actions in human uterine endometrial stromal cell. Our results demonstrate that the mitogenic effect of FGF9 is transduced via two parallel but additive signaling pathways involving mammalian target of rapamycin (mTOR) and extracellular signal-regulated kinase. Activation of mTOR by FGF9 induces p70 ribosomal S6 kinase (S6K1) phosphorylation, cyclin expression, and cell proliferation, which are independent of phosphatidylinositol 3-kinase and Akt. Coimmunoprecipitation analysis demonstrates that mTOR physically associates with S6K1 upon FGF9 treatment, whereas ablation of mTOR activity using RNA interference or pharmacological inhibitor blocks S6K1 phosphorylation and cell proliferation induced by FGF9. Further study demonstrates that activation of mTOR is regulated by a phospholipase Cγ-controlled calcium signaling pathway. These studies provide evidence to demonstrate, for the first time, that a novel signaling cascade involving phospholipase Cγ, calcium, mTOR, and S6K1 is activated by FGF9 in a receptor-specific manner.
). Despite its well studied biological functions, the signaling mechanism of FGF9 remains to be identified.
The signaling of FGF is mediated via complex interactions between specific members of the FGF family and one or more FGF receptor isoforms. Receptors for FGF (FGFR1 to -4) are tyrosine kinase receptors consisting of two intracellular tyrosine kinase domains, a single transmembrane domain, and an extracellular portion that contains three immunoglobulin-like (Ig) domains. The third Ig domain (Ig III), which exerts the highest impact on FGF receptor binding specificity and tissue-specific expression patterns, is the region in which alternative splicing occurs. Three different splice variants (designated as IIIa, IIIb, and IIIc) have been identified for FGFR1 and FGFR2, whereas only the IIIb and IIIc variants have been detected for FGFR3 (
). As of yet, no splice variant for FGFR4 has been identified. The splicing variant “IIIa” of FGFR is a secreted protein, whereas IIIb and IIIc are both membrane-bound receptors containing mutually exclusive Ig III domains. It is generally believed that the IIIb isoform of FGFRs is expressed in epithelial lineages, whereas the IIIc variant is restricted to mesenchymal origin (
Activation of FGFR triggers several intracellular signaling cascades. These include phosphorylation of Src and phospholipase Cγ (PLCγ), leading finally to activation of PKC as well as activation of Crk and Shc (
). The adaptor protein, FRS2, serves as an alternative link of FGFR to the activation of PKC and, in addition, activates the Ras signaling cascade. FRS2 contains multiple tyrosine phosphorylation sites in the C-terminal tail that serve as binding sites for the Grb2 and for the Src homology 2 domain-containing protein tyrosine phosphatase, Shp2 (
). Furthermore, phosphatidylinositol 3-kinase (PI3K)/Akt/p70 ribosomal S6 kinase (S6K1) cascade also play important roles in mediating FGFR signaling. Activation of PI3K by FGF-FGFRs has been demonstrated in studies using biochemical, pharmacological, and genetic approaches (
Although the signaling pathways of FGFR have been investigated extensively, most studies failed to distinguish individual signaling being activated by specific isoforms of FGFR or particular members of the FGF family. For example, most studies on FGF receptor-mediated signal transduction have been carried out using FGFR1 as the prototypical FGFR, which is bound by many members of FGF family. Others using an overexpression system to characterize the signaling pathways of a given splice variant have had trouble in interpreting the significance of their data, since it is not obtained under physiological conditions (
). We previously reported that endometrial stroma expresses predominantly the FGFR2IIIc splice variant by quantitative reverse transcription-PCR analysis and immunohistostaining and that FGF9 potently stimulates endometrial stromal cell proliferation (
). This unique feature of the endometrial stromal cell makes it the ideal model for investigation of the signaling pathway of FGF9 devoid of unnecessary interference from other FGFRs or forced overexpression of FGFR in a cell that normally does not express such receptor. Results from this study should provide novel information in dissecting the complex FGF-FGFR signaling network and bring new insight in understanding physiological and pathological functions of FGF9.
Materials—Primary antibodies (anti-Akt, anti-phospho-Akt (Ser473), anti-phospho-Akt (Thr308), anti-Erk1/2, anti-phospho-Erk1/2 (Thr202/Tyr204), anti-S6K1, anti-phospho-S6K1 (Thr389), anti-phospho-S6K1 (Thr421/Ser424), anti-mTOR, anti-phospho-mTOR (Ser2448), anti-phospho-mTOR (Ser2481), anti-TSC2, anti-phospho-TSC2 (Thr1462), anti-PKCα, and anti-PKCδ) were from Cell Signaling Technologies (Beverly, MA); anti-β-actin was from Oncogene Research Products (Cambridge, MA); anti-proliferating cell nuclear antigen was from Zymed Laboratories (South San Francisco, CA); and anti-cyclin A, anti-cyclin D1, anti-PI3K 110α, anti-total PKC, and anti-PDK1 were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The small interference RNA (siRNA) of mTOR, SignalSilence™ mTOR siRNA, was purchased from Cell Signaling. Cell culture materials (phenol red-free Dulbecco's modified Eagle's medium/F-12, fetal bovine serum, and antibiotics) were purchased from Invitrogen.
Tissue Collection and Stromal Cell Purification—Eutopic endometrial tissues from patients of reproductive age undergoing hysterectomy for leiomyoma or ovarian pathology (n = 25) were collected as previously described (
). Human ethics approval was obtained from the Clinical Research Ethics Committee at The National Cheng Kung University Medical Center, and informed consents were obtained from the patients. Tissues were immersed in Hanks' solution supplemented with HEPES and antibiotics and transported to the laboratory for further processing. The biopsies were minced and subjected for isolation of stromal cells as previously described (
). The stromal cell population was free of epithelial cell contamination.
Cell Cultures—Stromal cells were cultured in culture medium consisting of Dulbecco's modified Eagle's medium/F-12, 10% FCS, penicillin (100 μg/ml), streptomycin (100 units/ml), and fungizone (50 μg/ml) in a humidified atmosphere with 5% CO2 at 37 °C. The medium was changed every other day. When the cells reached confluence, they were subcultured in phenol red-free Dulbecco's modified Eagle's medium/F-12 medium supplemented with 10% FCS, penicillin (100 μg/ml), streptomycin (100 units/ml), and fungizone (50 μg/ml) until 70% confluence was reached. After serum starvation for 36 h, the cells were treated with vehicle, FGF9 (50 ng/ml) in the presence or absence of different inhibitors under serum-free, phenol red-free medium conditions. For the siRNA experiment, cells were cultured in a 12-well plate and transfected with SignalSilence™ mTOR siRNA or SignalSilence™ control siRNA according to procedures recommended by the manufacturer (Cell Signaling). Two days after transfection, cells were treated with FGF9 for 15 min. Cells were harvested under Tris-sucrose-EDTA buffer (10 mm Tris, 250 mm sucrose, and 0.1 mm EDTA, pH 7.4) containing protease and phosphatase inhibitors (1 mm phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml pepstatin A, 1 mm NaVO3, and 1 mm NaF) and centrifuged at 600 × g for 30 min at 4 °C to remove debris. Protein concentrations were determined by the Lowry method (
[3H]Thymidine Incorporation Assay—Subcultured stromal cells were deprived of serum for 36 h, treated with recombinant human FGF9 (50 ng/ml; R&D Research, San Diego, CA) for 18 h in the presence or absence of different inhibitors (as described in the figure legends), and subjected to [3H]thymidine incorporation assay as previously described with minor modification (
). In brief, cells were incubated with [3H]thymidine (1 μCi/ml) for 4 h and then washed twice with phosphate-buffered saline. Cells were incubated with 10% ice-cold trichloroacetic acid for 20 min and then washed with phosphate-buffered saline. The acid-insoluble fractions were dissolved by 1 n NaOH and then neutralized with an equal volume of 1 n HCl. Five hundred-microliter aliquots were transferred to a scintillation vial containing 3.5 ml of counting fluid (Ready Safe; Beckman). The radioactivity was measured by a liquid scintillation counter.
Immunoprecipitation and Western Blot Analysis—Whole cell lysate (500 μg of protein) was precleaned using Sepharose-conjugated protein A (Amersham Biosciences) at 4 °C for 1 h, and then anti-S6K1 antibody (4 μg/ml) was incubated with the lysate for 16–18 h at 4 °C with rotation. Protein A-Sepharose was then added to the lysate to capture the S6K1-antibody complex. After extensive washing with ice-cold washing buffer (0.5% sodium deoxycholate, 1% Nonidet P-40, 0.1% SDS, and 0.2 mm phenylmethylsulfonyl fluoride in 1× phosphate-buffered saline), immune complex was recovered by the protein A-Sepharose beads. They were boiled in 2× SDS sample buffer (125 mm Tris-HCl, 10% 2-mercaptoethanol, 4% SDS, 20% glycerol, 0.01% bromphenol blue) and subjected to SDS-PAGE separation. Proteins were transferred onto a polyvinylidene difluoride membrane (Millipore Corp., Bedford, MA), which was blocked in 5% skimmed milk at 4 °C overnight. Membrane was then incubated with specific primary antibodies for 1 h at 37 °C. After washing with TBST (10 mm Tris, pH 8.0, 150 mm NaCl, 0.05% Tween 20) for 3 × 10 min, membrane was further incubated with horseradish peroxidase-conjugated second antibodies for 1 h at room temperature. Membrane was washed for 6 × 10 min with PBST and detected by ECL (Amersham Biosciences). Membranes were then stripped with striping buffer (100 mm 2-mercaptoethanol, 2% SDS, and 62.5 mm Tris-HCl, pH 6.7) and redetected as described above using different set of primary and secondary antibodies. For statistical analysis, the intensity of individual band in a given Western blot was quantified using the AlphaImager analyzing software (Alpha Innotech Corp, San Leandro, CA).
PI3K Kinase Activity Assay—Serum-starved stromal cells were stimulated with 50 ng/ml FGF-9 for 5 min at 37 °C. The samples were lysed for 20 min on ice in 200 μl of lysis buffer (50 mm Tris, pH 7.4, 150 mm NaCl, 0.1% (w/v) SDS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 5 mm dithiothreitol, 1 mm phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml pepstatin, 10 μg/ml aprotinin, 1 mm Na3VO4, 2.5 mm sodium pyrophosphate). Immunoprecipitation was performed using anti-p85α, anti-phosphotyrosine, or an irrelevant mouse IgG at 4 °C overnight. Protein A-Sepharose was then added, and the incubation was continued for 1 h at 4 °C. The immunoprecipitates were washed three times in lysis buffer and once in kinase activity buffer (50 mm Tris-HCl, pH 7.4, 10 mm MgCl2). PI3K activity was initiated in 50 μl of final volume by the addition of 45 μl of kinase activity buffer, 5 μl of 1 ng/ml phosphoinositides, 10 μCi of [γ-32P]ATP (3000 Ci/mmol; Amersham Biosciences) and incubated for 10 min at room temperature. Phosphoinositides were sonicated for 15 min in 50 mm Tris-HCl, pH 7.4, and added to the samples (final concentration, 0.1 ng/ml). The reactions were stopped by the addition of 100 μl of 1 n HCl and 160 μl of CHCl3/MeOH (1:1) (v/v). The organic phase was collected, re-extracted with 80 μl of MeOH, 1 n HCl, 2 mm EDTA (1:1) (v/v). The samples were spotted onto Silica Gel 60 TLC (MERCK) plates pretreated with 1% (w/v) potassium oxalate, 2 mm EDTA, 40% MeOH (v/v) for 30 min and air-dried at room temperature. The plates were developed in CHCl3/MeOH/H2O/NH4OH (45:35:8.5:1.5) (v/v), and the radiolabeled spots were detected by autoradiography.
Calcium Measurement—Stromal cells were serum-starved for 36 h and then labeled for 40 min at room temperature with 5 μm Fura-2/AM, the fluorescent Ca2+ indicator. Extracellular Fura-2/AM was washed away, and the stromal cells were incubated for an additional 20 min at 37 °C in Fura-2-free buffer. The Petri dish was mounted on a computer-connected inverted microscope with epifluorescence attachments (Diaphot 300; Nikon, Tokyo, Japan). The excitation light from a xenon lamp was filtered to provide wavelengths of 340 and 380 nm using a high speed rotating filter wheel (Lambda 10-2, Sutter, Novato, CA). Cells were carefully focused and then sequentially treated with different concentrations of FGF9 (50, 100, and 200 ng/ml) at a 7-min interval. In a separated experiment, cells were pretreated with LY294002 (0.1, 1, 10 μm) or wortmannin (100 nm) 30 min prior to calcium measurement. Results from five randomly picked cells were used to calculate single cell calcium levels. At the end of each experiment, the calcium concentration was calibrated by applying ionomycin (5 μm) in the presence of 5 μm EGTA, followed by 10 μm CaCl2. All signals were corrected for autofluorescence determined by exposing the tissue to 5 mm manganese to quench cytosol Fura-2 fluorescence. Intracellular free calcium was estimated after subtracting background and autofluorescence according to the established formula.
Statistical Analysis—The data were expressed as means ± S.E. Data were analyzed using GraphPad Prism 4.02 (GraphPad Software, San Diego, CA). Duncan's procedure was used to test the difference between individual treatment groups, whereas Dunnet's test was applied to compare treatment groups versus control once significance was found by F test. p < 0.05 was considered statistically significant.
FGF9-induced Stromal Cell Proliferation Is Partially Mediated by the Ras-Erk Signaling Pathway—To investigate the mitogenic effect of FGF9, endometrial stromal cells were cultured in serum-free, phenol red-free Dulbecco's modified Eagle's medium/F-12 for 36 h and treated with FGF9 to induce cell proliferation. The addition of FGF9 (50 ng/ml) induced a greater than 10-fold increase in [3H]thymidine incorporation over basal level (Fig. 1A). Concordantly, expression of cell cycle regulators such as cyclin D1 (marker of G1 to S phase) and cyclin A (marker of S to G2 phase) was time-dependently up-regulated by treatment with FGF9. Elevation of cyclin D1 induced by FGF9 was first evident after treatment with FGF9 for 12 h and sustained for at least 48 h (Fig. 1B). Up-regulation of cyclin A was induced between 12 and 24 h after FGF9 treatment and remained elevated until 48 h (Fig. 1B). Together, this demonstrates that FGF9 can promote cell cycle progression, and the process is unremitting.
To investigate the signaling pathway that governs FGF9-induced stromal cell proliferation, cells were pretreated with selective inhibitors for Ras and MEK. Cell proliferation induced by FGF9 was inhibited by pretreatment with MEK inhibitor, PD98059, or Ras inhibitor, FPT, in a dose-dependent manner (Fig. 2, A and B). Interestingly, disruption of Ras-Erk signaling pathway using FPT or PD98059 cannot completely inhibit FGF9-induced stromal cell proliferation without causing cell death (Fig. 2, A and B, and data not shown).
To verify that incomplete inhibition of FGF9-induced stromal cell proliferation by Ras and MEK inhibitors was not due to incomplete blockage of Ras-MEK signaling, phosphorylation and nuclear translocation of MEK substrate, Erk1/2, was investigated. The addition of FGF9 to cultured stromal cells induced rapid and transient phosphorylation of Erk1/2 and nuclear translocation (Fig. 2D). In contrast, p38 mitogen-activated protein kinase and c-Jun N-terminal kinase were not phosphorylated by treatment with FGF9 (data not shown). Pretreatment with Ras inhibitor, FPT (10 μm), or MEK inhibitor, PD98059 (10 μm), inhibited Erk1/2 phosphorylation induced by FGF9 (Fig. 2E). Interestingly, expression of cyclin D1 was only partially inhibited by pretreatment with FPT and PD98059 (Fig. 2F). Thus, it is likely that the Ras-Erk cascade is only responsible for a portion of FGF9 actions, and one or more pathway(s) may exist to transduce FGF9 signaling.
FGF9 Action Is Independent of PI3K and Akt Activation—It is generally believed that PI3K-S6K1 is an important downstream effector system of peptide growth factors. To test whether PI3K-S6K1 signaling is involved in FGF9-induced cell proliferation, we first used PI3K inhibitor, LY294002, to block FGF9-induced stromal cell proliferation. Treatment with LY294002 dose-dependently inhibited FGF9-induced stromal cell proliferation (Fig. 3A). We next examined whether the downstream effector, S6K1, was activated by FGF9. Treatment of endometrial stromal cell with FGF9 induced phosphorylation of S6K1 at Thr389 and Thr421/Ser424 within 5 min and peaked at 15 min (Fig. 3B). The phosphorylation of S6K1 disappeared after 30 min of treatment (Fig. 3B). Phosphorylation of S6K1 was accompanied with nuclear translocation (Fig. 3B), although the significance remains unknown. Pretreatment with LY294002 (10 μm) abrogated FGF9-induced S6K1 phosphorylation and nuclear translocation induced by FGF9 (Fig. 3C).
Phosphorylation of Akt is usually accompanied by the activation of PI3K. Therefore, we tested whether Akt was phosphorylated by FGF9 treatment and served as an upstream effector for S6K1. To our surprise, Akt was not phosphorylated after the addition of FGF9 even after 60 min (Fig. 4A and data not shown). To confirm whether the human endometrial stromal cell was able to utilize Akt as a signaling molecule, EGF (10 ng/ml), IGF-1 (10 ng/ml), insulin (10 μg/ml), or FCS (10%) was added to the cell. A side-by-side experiment showed that EGF, IGF-1, insulin, and FCS all induced Akt phosphorylation on Ser473 and Thr308 by 5 min, whereas FGF9 was not able to exert such an effect (Fig. 4B).
The phosphorylation status of Akt might not entirely correlate with its enzyme activity; therefore, we directly evaluated the ability of Akt to phosphorylate its downstream substrates, forkhead transcription factor (FOXO) and tuberin (TSC2). Direct phosphorylation of FOXO by Akt resulted in cytoplasmic retention and inactivation. We therefore examined the subcellular localization of FOXO3a as an indicator of Akt activation. Treatment with serum (10% FCS) or IGF-1 (10 ng/ml) caused FOXO3a phosphorylation and cytosolic retention (Fig. 4C). In contrast, FOXO3a was accumulated in the nucleus under serum-free conditions or treated with 50 ng/ml FGF9 (Fig. 4C). In a parallel experiment, phosphorylation of TSC2 by Akt was evaluated. It has been shown that phosphorylation of TSC2 relieves inhibition of mTOR, which is a direct indicator for Akt-mediated mTOR activation (
). As shown in Fig. 4D, treatment with IGF-1 (10 ng/ml) markedly induced TSC2 phosphorylation, which can be blocked by LY294002 and wortmannin. In contrast, FGF9 failed to induce TSC2 phosphorylation (Fig. 4D). Taken together, these results provided further evidence to support that Akt was not activated by FGF9 in human endometrial stromal cells.
It is obvious that FGF9 does not activate the PI3K-dependent Akt, but whether PI3K is activated remains unanswered despite the effectiveness of LY294002 in blocking S6K1 phosphorylation and DNA synthesis. We therefore decided to determine whether PI3K was phosphorylated by FGF9 treatment. The PI3K protein was immunoprecipitated by anti-phosphotyrosine antibody and then blotted with anti-p85 subunit of PI3K. The result demonstrated that IGF-1 induced robust PI3K phosphorylation, whereas FGF9 failed to induce any PI3K phosphorylation (Fig. 4E). Again, both EGF and serum could induce PI3K phosphorylation, but the extent was much less than that of IGF-1 (Fig. 4E). Finally, we directly measured the kinase activity of PI3K. In concordance with other data, treatment with FGF9 did not induce any measurable kinase activity of PI3K (Fig. 4F). In contrast, PI3K kinase activity was significantly elevated by IGF-1 treatment and by the addition of 10% FCS to a lesser extent (Fig. 4F). Taken together, these results provided direct evidence to support that PI3K is not involved in FGF9-induced stromal cell proliferation.
Phosphorylation of S6K1 Induced by FGF9 Is Mediated by mTOR—To address what is the upstream effector leading to S6K1 phosphorylation, we immunoprecipitated stromal cell lysates with anti-S6K1 antibody and blotted with antibodies against several potential candidates. The result demonstrated that mTOR directly associated with S6K1 in cells treated with FGF9 (Fig. 5A). The physical association may be with phosphorylated form S6K1, since the level of total S6K1 was not different between FGF9-treated and control cells (Fig. 5A). Western blot analysis demonstrated that mTOR was phosphorylated on Ser2481 and Ser2448, which are critical sites for mTOR kinase activity, within 5 min after FGF9 treatment and remained phosphorylated even up to 60 min (Fig. 5B). To confirm that activation of mTOR resulted in S6K1 phosphorylation, stromal cells were pretreated with rapamycin, a selective inhibitor for mTOR at 30 min before FGF9 treatment. Pretreatment with rapamycin (10 nm) specifically inhibited S6K1 phosphorylation (Fig. 5C), providing evidence to support that mTOR is an upstream effector of S6K1.
An alternative approach using RNA interference technique was conducted to evaluate effect of mTOR on S6K1 phosphorylation. Transfection with mTOR-specific siRNA knocked down about 80% of mTOR, whereas transfection with nonrelated siRNA did not affect expression of mTOR (Fig. 5D). The expression of eIF4B was not affected by either mTOR siRNA or control siRNA (Fig. 5D). Although mTOR siRNA did not completely ablate mTOR expression, it was sufficient to blunt FGF9-induced mTOR phosphorylation (Fig. 5D). Concordantly, mTOR siRNA blocked FGF9-induced S6K1 phosphorylation without altering the expression level of S6K1 (Fig. 5E). Taken together, these results provided evidence showing that mTOR is an upstream effector of S6K1 in FGF9-treated stromal cells.
Finally, we aimed to determine whether blockage of mTOR activity is sufficient to inhibit FGF9-induced cell proliferation. Treatment with rapamycin dose-dependently (from 0.1 nm to 1 μm) inhibited stromal cell proliferation (Fig. 5F). Nevertheless, rapamycin still cannot completely inhibit FGF9-induced stromal cell proliferation without causing cell death (Fig. 5F).
Activation of mTOR by FGF9 Is Mediated through PLCγ—So far, we have demonstrated that mTOR is activated by FGF9 treatment and is able to induce phosphorylation of S6K1, but the upstream effector of mTOR remains uncharacterized. Activation of PLCγ is another signaling pathway that mediates FGF actions. To determine whether PLCγ is involved in mTOR activation, we first examined whether PLCγ could be phosphorylated by treatment with FGF9. Fig. 6A shows that phosphorylation of PLCγ can be detected by 2.5 min after treatment with FGF9. FGF9-induced PLCγ phosphorylation was still evident at 5 min after FGF9 treatment but disappeared after 15 min. The role of PLCγ on FGF9 action was further supported by evidence that pretreatment with a selective PLCγ inhibitor, U73122 (10 μm), effectively inhibited FGF9-induced mTOR phosphorylation and cell proliferation (Fig. 6, B and C).
Activation of PLCγ results in conversion of phosphatidylinositol bisphosphate to diacylglycerol and inositol 1,4,5-trisphosphate, which causes PKC activation and calcium influx, respectively. To determine whether PKC is involved in FGF9-induced mTOR phosphorylation, we first examined translocation of PKC to the plasma membrane, a sign of PKC activation. FGF9 failed to induce PKC translocation from cytosol to plasma membrane, whereas phorbol 12-myristate 13-acetate did cause PKC translocation (Fig. 6D). Prolonged treatment of stromal cells with FGF9 up to 24 h also failed to induce PKC translocation (data not shown). Another sign of PKC activation is phosphorylation of PKC. Five isoforms of PKC, including α, β, δ, θ, and λ, were detected in endometrial stromal cells, whereas others were undetected under our conditions. By using different anti-phosphorylated PKC antibodies, no FGF9-induced PKC phosphorylation was observed (Fig. 6E). Finally, pretreatment with the general PKC inhibitor (GF109203X; 1 or 5 μm) failed to affect FGF9-induced [3H]thymidine incorporation further substantiates that PKC is not involved in FGF9 signaling (Fig. 6F).
Calcium/Calmodulin-dependent Protein Kinase (CaMK) Mediates FGF9-induced mTOR Activation—Since PKC is not involved in FGF9-induced mTOR activation, we next aimed to test the involvement of calcium signaling. Treatment with FGF9 caused a significant increase in free intracellular calcium concentration in time- and dose-dependent manners (Fig. 7, A and B). Increased intracellular free calcium concentration is known to activate CaMK. We thus used several inhibitors to block the calcium-CaMK pathway and evaluated the phosphorylation of mTOR and its downstream substrate, S6K1. Results from three independent experiments using different batches of cells demonstrated that pretreatment with W7 (calmodulin inhibitor, 10 μm), KN93 (CaMK inhibitor, 10 μm), or EDTA (calcium chelator, 10 mm) attenuated FGF9-induced phosphorylation of mTOR on Ser2481 (p < 0.05, Fig. 7C) but not on Ser2448 (p > 0.05, data not shown). Phosphorylation of S6K1 after FGF9 treatment was also reduced by pretreatment with inhibitors that selectively block PLCγ, calmodulin, or CaMK, or by depletion of calcium (Fig. 7D). Together, these evidences demonstrate that PLCγ-mediated calcium signaling may play a role in FGF9-induced mTOR and subsequently S6K1 activation.
It has been reported that LY294002 can block growth factor-induced calcium influx (
). We thus decided to determine whether the effect of LY294002 on inhibiting FGF9-induced cell proliferation is mediated via blocking calcium signaling. Stromal cells were pretreated with LY294002 or wortmannin for 30 min and then subjected to calcium measurement. Pretreatment with LY294002 dose-dependently inhibited FGF9-induced calcium influx, whereas pretreatment with wortmannin had no such effect (Fig. 7E). As a positive control, IGF-1 also induced significant increase in intracellular free calcium concentrations. Again, LY294002 effectively blocked IGF-1-induced calcium influx whereas wortmannin was not effective (Fig. 7F). The effective dose of LY294002 (0.1 μm) was much lower than that on inhibiting PI3K activity (10 μm), suggesting that its effect on blocking calcium influx was not mediated via suppression of PI3K activity.
FGF9-induced Ras-Erk and mTOR-S6K1 Pathways Are Parallel to Each Other—We have thus far identified that both Ras-Erk and PLCγ-calcium-mTOR-S6K1 were activated by FGF9. It has been reported that PI3K-S6K1 and Erk pathways can cross-talk to each other. Since mTOR is a PI3K-like kinase, we sought to investigate whether the mTOR and Ras-Erk can mutually affect each other. Fig. 8A shows that inhibition of Ras-Erk signaling pathway by treatment with FPT or PD98059 had no substantial effect on FGF9-induced S6K1 phosphorylation (Fig. 8A). Again, pretreatment with rapamycin blocked both FGF9-induced and basal level of S6K1 phosphorylation (Fig. 8A). On the other hand, inhibition of mTOR activity by treatment with rapamycin (Fig. 8B) or ablation of mTOR by RNAi (Fig. 8C) failed to prevent Erk1/2 phosphorylation, whereas PD98059 and FPT significantly blocked Erk1/2 phosphorylation (Fig. 8B). In addition, inhibition of PLCγ activity or blockage of calcium signaling pathways has no substantial effect on FGF9-induced Erk phosphorylation (Fig. 8B). These results suggest that the PLCγ-calcium-mTOR and Ras-Erk pathways may function independently in FGF9-treated endometrial stromal cells.
Previous results showed that blockage of Erk signaling pathway or mTOR signal cascade resulted in partial inhibition of [3H]thymidine incorporation (Figs. 2A and 5F). To confirm that Erk and mTOR signaling pathways additively regulate FGF9-induced cell proliferation, PD98059 and rapamycin were used either alone or in combination to block FGF9 actions. Pretreatment with either inhibitor alone only resulted in partial inhibition of FGF9-induced cyclin D1 expression (Fig. 8D). The expression level of cyclins were nicely mirrored by [3H]thymidine incorporation data with U73122, rapamycin, and PD98059 alone only exerting partial inhibitory effect on FGF9-induced cell proliferation (Fig. 8E). Combined treatment with U73122 and PD98059 or rapamycin and PD98059 completely abrogated FGF9-induced stromal cell proliferation (Fig. 8E). Taken together, these data demonstrated that Ras-Erk and mTOR-S6K1 signaling pathways independently and additively control FGF9 action in human endometrial stromal cells.
Fibroblast growth factor-9 plays critical roles in many physiological and pathological processes including sex determination, embryonic development, neuron degeneration, cancer formation, and endometriosis. The mitogenic effect of FGF9 is mediated via binding to its high affinity receptors, although the cellular and molecular mechanisms remain to be determined. This study shows that FGF9 stimulates cell proliferation via two parallel but additive pathways, the Ras-Erk and mTOR-S6K1 cascades in human uterine endometrial stromal cells (Fig. 9). We demonstrate that FGF9-induced cell proliferation is independent of the well established PI3K-Akt signaling pathway utilized by most peptide growth factors such as IGF-1, EGF, and other FGFs (
). Instead, our data provide evidence to show that activation of S6K1 is directly mediated by mTOR. Importantly, we further demonstrate that PLCγ and calcium-dependent kinase are involved in FGF9-induced mTOR phosphorylation. To our knowledge, this is the first report to characterize the signal pathways responsible for FGF9 function under a physiological condition and to demonstrate a novel signaling cascade involving PLCγ, calcium, mTOR, and S6K1 in a ligand receptor-specific manner.
It has been established that the Ras-mitogen-activated protein kinase pathway plays an important role in signaling via FGF receptors (
). The adapter protein Grb2 links FGFRs with the Ras signaling pathway by binding to the guanine nucleotide-releasing factor Sos through its Src homology 3 domains and to tyrosine-phosphorylated receptors or docking molecules via its Src homology 2 domain (
). Our current results agree with this notion by showing that Ras-Erk cascade is activated by treatment with FGF9. Nevertheless, we found that activation of Ras-Erk signaling only accounts for partial FGF9-induced stromal cell proliferation.
Activation of PI3K-Akt-S6K1 signaling cascade by growth factor not only mediates cell survival but also controls DNA synthesis (
). Surprisingly, we found no sign of PI3K activation and Akt phosphorylation upon FGF9 treatment in human endometrial stromal cells indicating that S6K1 was phosphorylated at Thr389, Thr421, and Ser424. The inability of FGF9 to utilize the PI3K-Akt pathway is not due to a lack of either signal molecule in the endometrial stromal cell but rather to specific ligand-receptor interaction. Four lines of evidence show that the PI3K-Akt signaling cascade is intact and functional in endometrial stromal cells. First, treatment of endometrial stromal cells with IGF-1 or serum significantly increases PI3K phosphorylation and kinase activity. Second, serum or other peptide growth factors such as EGF, IGF-1, and insulin effectively cause Akt phosphorylation. Third, phosphorylation of TSC2, a hallmark of Akt activity, is markedly elevated after IGF-1 treatment. Fourth, cytosolic retention of FOXO3a caused by Akt-dependent phosphorylation (
) is evident after IGF-1 or serum treatment, suggesting that the enzymatic activity of Akt is intact in the endometrial stromal cells. Most notably, the cytosolic retention of FOXO3a after IGF-1 treatment was conducted in primary cultured cells without forced overexpression of Akt. These results clearly show that FGF9, via binding to its high affinity receptor (likely to be FGFR2IIIc), stimulates endometrial stromal cell proliferation independently of the PI3K-Akt signaling cascade.
At first glance, the findings that FGF9-induced S6K1 phosphorylation and DNA synthesis can be inhibited by LY294002, a putatively selective PI3K inhibitor, are contradictory to the conclusion that PI3K is not activated by FGF9 in endometrial stromal cell. This disparity leads us to ponder that the target of LY294002 may not be PI3K per se. Our hypothesis is supported by reports that LY294002 and wortmannin can inhibit activities of PI3K and mTOR (
). Our current results demonstrate that LY294002 is a much more effective inhibitor at blocking calcium influx than at inhibiting PI3K activity, whereas wortmannin is only effective in suppressing PI3K activity. These results provide another possibility that the effect of LY294002 on FGF9-induced cell proliferation may be mediated via inhibiting the calcium signaling pathway. Nevertheless, the molecular target of LY294002 remains undetermined and merits further investigation.
mTOR (also known as FRAP, RAFT, and RAPT) is a member of phosphatidylinositol kinase-related kinase family that plays a central role in controlling cellular growth and DNA synthesis (
). To test whether mTOR is the direct activator of S6K1 in FGF9-treated stromal cells, we used anti-S6K1 antibody to bring down the immunocomplex and blotted with anti-mTOR antibody. Our results show that mTOR physically associates with S6K1 upon FGF9 treatment. Furthermore, ablation of mTOR by siRNA completely blocks FGF9-induced S6K1 phosphorylation. Results from the current study and previous reports (
) provide clear evidence to show that mTOR is the direct upstream activator of S6K1 in response to FGF9 signaling. Concordantly, rapamycin inhibited FGF9-induced cyclin A, D1, and B1 expression and [3H]thymidine incorporation similar to, if not more potent than, that inhibited by LY294002. All of this evidence implicates mTOR as the alternative signaling molecule that mediates FGF9-induced S6K1 phosphorylation and thus DNA replication.
The upstream effectors that activate mTOR have been extensively investigated in the past couple years (
). However, none of the above mentioned molecules seems to be a likely candidate to activate mTOR in FGF9-treated endometrial stromal cells. The activation of PLCγ by FGF9 provides a new direction to test whether its downstream effectors, PKC and calcium, can activate mTOR. mTOR undergoes multiple phosphorylations. The phosphorylation of Ser2448 and Ser2481 under various stimuli is used to reflect the activity of mTOR (
). Our results show that inhibitor for PLCγ attenuates FGF9-induced mTOR phosphorylation on Ser2481 and its ability to induce S6K1 phosphorylation. Further study suggests that calcium but not the PKC pathway is involved in phosphorylation of Ser2481, which is concordant with a recent report demonstrating that activation of α1A-adrenergic receptor by phenylephrine induces mTOR phosphorylation on Ser2481 is calcium-dependent and PKC-independent (
). Another mTOR residue, Ser2448, was also phosphorylated after FGF9 treatment but the phosphorylation appears to be independent of calcium signaling. The underlying mechanism is not clear but it has been reported that amino acid induces mTOR phosphorylation is not mediated by Akt, Erk, or calcium signaling pathways in primary culture rat adipocytes (
). Furthermore, though Ser2448 is a consensus phosphorylation site for Akt, recent studies reveal that Ser2448 can be phosphorylated even when PI3K/Akt is not activated and that phosphorylation of this site might not be sufficient for mTOR kinase activity (
). Taken together, these lines of evidence implicate that phosphorylation of mTOR induced by FGF9 might also be mediated by other signaling molecules. Further investigation is warranted to identify the involvement of other possible signaling molecules in FGF9 signaling and unravel the mechanism responsible for mTOR phosphorylation.
It has been reported that Erk is one of the downstream effectors of PI3K/Akt pathway and that Erk may regulate the activation of S6K1 (
). However, our data demonstrated that these two signaling pathways are parallel but additive in mediating FGF9-induced stromal cell proliferation by four lines of evidence. First, blockage of either pathway only results in partial inhibition of FGF9-induced [3H]thymidine incorporation, whereas simultaneous disruption of both pathways completely inhibits FGF9 action. Second, phosphorylation of Erk is not affected by treatment with LY294002, U73122, rapamycin, W7, or KN93. Third, ablation of mTOR by siRNA does not influence Erk phosphorylation. Fourth, the addition of Ras or MEK inhibitors fails to inhibit S6K1 phosphorylation.
The complexity of FGF-FGFR signaling has hampered researchers in dissecting functions of an individual member of FGF family. For example, FGF2 can bind to FGFR1, FGFR2IIIc, FGFR3IIIc, and FGFR4 with high affinity and FGFR1 can sever as common receptor for FGF1, -2, -4, -5, and -6 and more (
), showing distinct binding affinity and/or specificity of FGFs to FGFRs, it is reasonable that FGF9 may utilize a distinct signaling pathway, since it does not bind FGFR1. The human uterine endometrial stromal cell is an excellent system to characterize the signaling pathway of FGF, because the predominantly receptor isoform in this cell type is FGFR2IIIc (
). Therefore, it provides a clean system to study the signaling pathway activated by the ligand, such as FGF9 and an exclusive isotype of the receptor devoid of interference by other variants. Although it may be argued that the signaling pathways from different FGFRs are similar, giving a high degree of homology at the amino acid level, some caveats still exist due to the biochemical nature of the ligand-receptor interaction. Thus, different members of FGF family may utilize distinct signal pathways to exert their specific actions. For example, a recent report by Portnoy et al. shows that FGF7 stimulates alveolar cell proliferation through Erk and PI3K/Akt pathways by binding to FGFR2IIIb (
). In the current study, our results show that FGF9 follows the PLCγ-calcium-mTOR-S6K1 cascade, clearly demonstrating that this is different from the conventional PI3K-Akt signaling pathway. In sum, the results presented here provide compiling evidence to demonstrate specific ligand-mediated FGF receptor signaling, which is important for understanding signaling pathways by different FGF-FGFRs and the consequence of their physiological and pathological functions.
We are grateful for the excellent technical assistance from Dr. Yang-Kao Wang in the PI3K kinase activity assay.