|Home | About | Journals | Submit | Contact Us | Français|
Fibroblast growth factor receptors (FGFRs) are cell surface tyrosine kinases that function in cell proliferation and differentiation. Aberrant FGFR signaling occurs in diverse cancers due to gene amplification, but the associated oncogenic mechanisms are poorly understood. Using a proteomics approach, we identified STAT3 as a receptor binding partner that is mediated by Tyr677 phosphorylation on FGFR. Binding to activated FGFR was essential for subsequent tyrosine phosphorylation and nuclear translocation of STAT3, along with activation of its downstream target genes. Tyrosine phosphorylation of STAT3 was also dependent upon concomitant FGFR-dependent activity of SRC and JAK kinases. Lastly, tyrosine (but not serine) phosphorylation of STAT3 required amplified FGFR protein expression, generated either by enforced overexpression or as associated with gene amplification in cancer cells. Our findings show that amplified FGFR expression engages the STAT3 pathway, and they suggest therapeutic strategies to attack FGFR-overexpressing cancers.
Fibroblast growth factor receptors (FGFRs) of the receptor tyrosine kinase family mediate a diversity of biological processes, including regulation of embryogenesis, proliferation, differentiation, migration, cell survival and angiogenesis (1). These functions are executed by transmitting signals that initiate gene expression. Thus, upon ligand binding FGFRs dimerize, triggering tyrosine kinase activity, leading to autophosphorylation of the intracellular domains of both receptors and receptor associated adaptors and subsequent recruitment of signaling partners. Formation of these multiprotein complexes results in activation of several well characterized downstream pathways, including MAPK and PI3K/Akt (1).
Abnormal expression of FGFRs is often linked with development and progression of a variety of human cancers. FGFR gene amplifications and protein expression have been described in several tumor types, including breast (2-4), gastric (5, 6) and prostate (7). As a result of this, and other oncogenic manifestations of FGFR function, the FGFR pathway has elicited significant interest as a target for the development of therapeutic interventions (8). However, despite rapidly advancing knowledge of the prevalence of FGFR amplicons in tumors, the exact impact of enhanced FGFR expression on downstream signaling processes is unknown. Here we implicate the transcription factor STAT3 as a mediator of amplified FGFR signaling.
The Signal Transducers and Activators of Transcription (STATs) were identified as transcription factors mediating signaling via the cytokine family of signaling receptors (9). Activation of STATs mediates cell migration, differentiation, proliferation, apoptosis and wound healing and is classically induced via the cytokine receptor-associated JAK family of non-receptor tyrosine kinases. JAKs phosphorylate tyrosine residues within the intracellular domain of cytokine receptors, providing SH2 domain mediated docking sites for STAT3. STAT3 is then phosphorylated by JAKs on Tyr705 within the C-terminus yielding STAT dimerisation (10). Dimeric STAT3 translocates to the nucleus where it induces downstream transcriptional activation of specific target genes (9). The dynamic status of the STAT3/JAK pathway is regulated by two key determinants: STAT3 de-phosphorylation and nuclear export (11) as STAT3 constitutive shuttles between the cytoplasm and nucleus and tyrosine phosphorylation inactivates nuclear export (12).
STAT3 is often expressed at a high level and constitutively activated in many human tumor specimens (13-15) and its role in cancer development is firmly established (16). The first observation of oncogenic STAT3 activation was made in cells transformed by v-Src (17). In addition, a constitutively active form of STAT3 was shown to be able to transform fibroblasts and induce tumorigenicity (18).
In this study, we demonstrate a connection between over-expression of FGFRs and STAT3 activation. We report that STAT3 is a binding partner for phosphorylated Tyr677 of FGFR1 and this tyrosine is shown to be critical for STAT3-FGFR1 association. Importantly, tyrosine activation of STAT3 requires over-expression of FGFR1 or FGFR2: either in tumor cells, SUM-52PE, expressing high level of endogenous FGFR2 or experimentally by transient transfection of FGFR1. Moreover, our results imply that JAK2 and Src also form a complex with active FGFR and TyrSTAT3 phosphorylation by FGF is JAK2 and Src-dependent. These findings suggest that oncogenic FGFR amplification, and hyper-expression of receptor, results in ectopic activation of the STAT3 transcriptional response.
HEK293T, MCF7 and HeLa cells were cultured in DMEM supplemented with 10% FBS, 0.2 U/ml penicillin, 0.1 mg/ml streptomycin and 2 mM L-glutamine. SUM-52PE cells were cultured in RPMI1640 supplemented with 10% FBS, 0.2 U/ml penicillin, 0.1 mg/ml streptomycin and 2 mM L-glutamine.
Antibodies obtained from Santa Cruz Biotechnology Inc: rabbit anti-Bek, rabbit anti-Flg, rabbit anti-ERK1, mouse anti-phospho-ERK1, rabbit anti-JunB, rabbit c-fos. Antibodies obtained from Cell Signaling Technology: rabbit anti-phosphoJak2 (Tyr1007/1008), rabbit anti-Jak2, mouse anti-Src, rabbit anti-phosphoSrc family (Tyr416), mouse anti-STAT3, rabbit anti-phosphoSTAT3 (Tyr705), rabbit anti-phosphoSTAT3 (Ser727), rabbit anti-phosphoSTAT5 (Tyr694), mouse anti-phosphoFGFR (Tyr653/654), rabbit anti-c-myc, mouse myc-tag. Mouse anti-α-tubulin (Sigma), goat anti-human IgG-Fc-HRP conjugate (Pierce), goat anti-mouse IRDye-conjugated, goat anti-rabbit IRDye-conjugated (Li-cor), goat anti-mouse IgG Texas Red, goat anti-rabbit Alexa Fluor 594 and goat anti-rabbit IgG FITC (Molecular Probes). Inhibitors: SU5402, SU6656, Jak Inhibitor I, SP600125, SB203580, Calphostin C, Cucurbitacin I (Calbiochem), U0126 (Cell Signaling Technology), PD173074 (Sigma). LDH toxicology kit and MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) (Sigma) were used according to manufacturer’s instructions.
Dynabeads MyOne Streptavidin T1 (Invitrogen) were washed and re-suspended in PBST. 0.1μM of desthiobiotinylated FGFR1 peptides was added to Dynabeads, incubated for 1 hour and washed with PBST before addition of HEK293T whole cell lysate. After 1 hour whole cell lysate was removed and beads were extensively washed with PBST. Following addition of 2x reduced SDS sample buffer, proteins were resolved by SDS-PAGE and analyzed by Western blotting.
FcFGFR1 pEF-BOS constructs were previously described (19). Human full length FGFR1WT pcDNA3.1 was provided by Pamela Maher (Department of Cell Biology, The Scripps Research Institute, La Jolla, California). Mouse STAT3 pZeoSV2 was provided by Valeria Poli (Molecular Biotechnology Center, University of Turin, Italy). Point mutations in FGFR1, Y677F, and STAT3, R609L, constructs were obtained by QuickChange site-directed mutagenesis (Stratagene) and were confirmed by DNA sequencing. N-terminal myc-tagged-JAK2 construct was made by Gateway cloning (Invitrogen) according to manufacturer’s instructions. siRNA ON-TARGET oligonucleotides against FGFR2 were purchased from Thermo Fishers Scientific. C-Src and Jak2 siRNA were purchased from Santa Cruz Biotechnology Inc. Non-targeting control siRNA were purchased from Qiagen. Gene Juice reagent (Novagen, MERCK Bioscience) and DharmaFECT3 (Dharmacon) were used according to the manufacturer’s instructions.
For all experiments cells were serum-starved, stimulated with 20 ng/ml FGF1 and 10 μg/ml heparin for 20 minutes (or as indicated). Cell lysis, immunoprecipitation were essentially performed as described previously (20). Proteins were separated on NuPAGE 4-12% Bis-Tris gels (Invitrogen) and transferred to PVDF membrane (Millipore). Primary antibodies were incubated with the membrane overnight at 4°C. Membranes were washed and probed with InfraRedDye-conjugated secondary antibody, and then proteins were visualized using fluorescence detection on the Odyssey Infrared Imaging System (Li-cor).
Cells growing on coverslips were fixed with 4% paraformaldehyde, permeabilized for 5 min with methanol at −20°C and incubated with 4% BSA-PBS blocking buffer for 1 hour. Cells were incubated with primary antibody followed by incubation with fluorescence-conjugated secondary antibody. Coverslips were mounted with Mowiol solution on the slide and observed under a laser scanning confocal microscope (Leica TCS SP2). All further image processing was carried out by Adobe Photoshop 7.0.
In order to identify phosphorylation-dependent protein partners for FGFR1, we performed peptide pull-down experiments from HEK293T cells using phosphorylated and non-phosphorylated pairs of FGFR1 peptides. Using mass spectrometry STAT3 was identified as a novel binding partner for the FGFR1 peptide containing phospho-Tyr677 (Supplementary Table 1). Tyrosine 677 is part of a pYxxQ motif on FGFR1, which is a consensus binding motif for the STAT3 SH2 domain (21) and conserved in human FGFRs 2, 3 and 4 and other vertebrate FGFRs. To confirm the interaction, whole cell lysates from HEK293T cells expressing either STAT3WT (wild type) or STAT3 R609L, a mutation that specifically disrupts the SH2 binding pocket of STAT3, were incubated with Tyr677 FGFR1 peptides. Figure 1A shows an increased association of STAT3WT with phospho-Tyr677 FGFR1 peptide compared to the non-phosphorylated peptide. R609L mutation in STAT3 abolished its ability to bind the FGFR1 peptide, demonstrating the importance of the STAT3 SH2 domain (Fig. 1A). These results reveal the interaction between the phospho-Tyr677 FGFR1 peptide and STAT3, and show that the interaction is dependent on the SH2 domain of STAT3.
The STAT3-FGFR1 interaction was further verified by co-immunoprecipitation. An FGFR1 construct was created where tyrosine 677 was substituted with phenylalanine. FGFR1WT and FGFR1 Y677F were expressed in HEK293T cells (Fig. 1B and C). Over-expression of FGFR1WT caused constitutive activation of the receptor, so to compare conditions where FGFR1 was phosphorylated and de-phosphorylated, we used SU5402, FGFR inhibitor. The interaction between activated FGFR1WT and STAT3 was confirmed by independent immunoprecipitation experiments where over-expressed FGFR1 or endogenous STAT3 were pulled-down (Fig. 1B and C). In both experiments association between FGFR1 Y677F and STAT3 was decreased, confirming the importance of Tyr677 of FGFR1 in this interaction (Fig. 1B and C). Additionally, tyrosine phosphorylation of STAT3 was significantly enhanced when FGFR1WT was over-expressed and phosphorylated (Fig. 1B). Increased STAT3 tyrosine phosphorylation was not observed when FGFR1 Y677F was expressed, however, the Y677F point mutation in FGFR1 did not affect serine phosphorylation of STAT3 (Fig. 1B). Serine phosphorylation of STAT3 was also detected in cells expressing endogenous FGFR1, in the presence of FGF1, suggesting that SerSTAT3 was activated as a downstream effect of FGFR1 and was independent of Tyr677. Furthermore, STAT3 recruitment to phosphorylated FGFR1 was shown to be independent of FRS2 (Supplementary Fig. 1). Taken together, our results indicate that STAT3 and FGFR1 specifically interact with each other in a phospho-dependent-manner and Tyr677 of FGFR1 plays a critical role in STAT3 recruitment and TyrSTAT3 phosphorylation. Given the crucial roles of the receptor YxxQ motif and the STAT3 SH2 domain, it is most likely that this interaction is direct. Nevertheless, the formal possibility that it is mediated via another protein cannot be excluded.
In the cytokine receptor pathway, activated STAT3 is accumulated inside the nucleus where it exerts its biological effects. Immunofluorescence studies on wild type HeLa cells showed no STAT3 nuclear accumulation after FGF1 stimulation (data not shown), therefore, we examined localization of STAT3 in HeLa cells expressing high levels of FGFR1. Nuclear accumulation of STAT3 was observed in HeLa cells over-expressing FGFR1WT (Fig. 1D). Moreover, treatment with SU5402 suppressed STAT3 nuclear accumulation proving that the observed effect was due to FGFR1WT activation (Fig. 1D). Transfection with control vector did not induce any effect on STAT3 location (Fig. 1D). Additionally, immunofluorescence analysis for phospho-TyrSTAT3 was performed. HeLa cells transfected with FGFR1WT showed nuclear accumulation of tyrosine-phosphorylated STAT3 which was effectively decreased with SU5402 (Supplementary Fig. 2). These results confirm that TyrSTAT3 is phosphorylated only in cells expressing FGFR1WT to higher levels than those required to elicit Tyr677–independent serine phosphorylation of STAT3.
We examined the effect of FGF1 stimulation on tyrosine and serine phosphorylation of STAT3 in several FGF responsive cell lines to probe this point further. No tyrosine phosphorylation of STAT3 was observed upon FGF1 stimulation in MCF7, HEK293T, HeLa and NIH3T3 cells although STAT3 protein was clearly detectable (data not shown). This suggested that stimulation of endogenous FGFR1 in these cells had no detectable effect on STAT3 tyrosine activation. On the other hand, FGF-induced SerSTAT3 phosphorylation did not rely on high level of FGFRs (Fig. 1B). FGF-induced STAT3 serine activation in MCF7 cells was also demonstrated (Fig. 2). In order to examine which Ser/Thr kinases phosphorylate SerSTAT3 we used specific inhibitors. Decreased SerSTAT3 phosphorylation with U0126 (MEK inhibitor) and SP600125 (JNK inhibitor), but not Jak Inhibitor I, SB203580 (p38 inhibitor) and calphostin C (PKC inhibitor), was observed (Fig. 2). These data suggest that SerSTAT3 is phosphorylated downstream of JNK and Ras/MAPK pathways which are both activated upon FGF1 stimulation. When FGFR1 and Src were inhibited by treatment with SU5402 and SU6656 respectively, the level of phospho-SerSTAT3 was also significantly reduced (Fig. 2), presumably due to lack of JNK and MEK activation. Thus, JNK and MEK kinases are responsible for phosphorylation of SerSTAT3 which is a downstream effect of FGF1-mediated signaling.
As phosphorylation of TyrSTAT3 required over-expression of FGFR1, we hypothesized that this pathway functions only in cells that express high levels of FGFRs. Therefore, we examined several tumor cells in which the FGFR gene had been amplified. Surprisingly, from this set only SUM-52PE exhibited elevated FGFR protein expression. SUM-52PE is a breast cancer cell line reported to have 12-fold amplification of the FGFR2 gene and 40-fold over-expression of FGFR2 isoform IIIb (2). The expression level of FGFR2 in two cell lines, MCF7 and SUM-52PE, was compared. Figure 3A shows that FGFR2 in MCF7 cells is barely detectable, whereas SUM-52PE cells express the receptor at high levels. Importantly, the phosphorylation of TyrSTAT3 was only observed in SUM-52PE cells, but not in MCF7, which supports the notion that TyrSTAT3 activation by FGF requires high levels of FGFR expression.
To specifically activate FGFR2 isoform IIIb we used FGF7 (1). FGF7 stimulation induced STAT3 tyrosine and serine phosphorylation (Fig. 3B). Moreover, tyrosine STAT5 phosphorylation was also induced (Fig. 3B) and to lesser extent TyrSTAT1 (data not shown) which implicates the possibility of interaction between FGFRs and other members of STAT family. To further verify that this effect was due to FGFR2 activity we used increasing concentrations of two FGFR inhibitors, SU5402 and PD173074, following stimulation with FGF7 (Fig. 3C). Phosphorylation of TyrSTAT3 gradually decreased together with the inhibition of FGFR2 (Fig. 3C) suggesting that receptor activity was essential.
In order to validate the data showing that elevated level of FGFR2 caused tyrosine STAT3 activation, the effect of receptor depletion was studied. Even though the knock-down of FGFR2 was not complete, a reduction of tyrosine STAT3 phosphorylation was observed (Fig. 3D). To exclude the possibility of an off-target effect, two siRNA oligonucleotides against FGFR2 were used and the same effect was detected (Fig. 3D). This reveals a significant role for FGFR2 kinase activity in FGF-induced tyrosine STAT3 phosphorylation in SUM-52PE cells.
We further verified whether STAT3 interacts with FGFR2. An immunoprecipitation experiments were performed in which endogenous FGFR2 or endogenous STAT3 were pulled-down from whole cell lysates (Fig. 4A and B). As expected, the association between STAT3 and FGFR2 was enhanced by FGF1 stimulation, whereas, treatment with SU5402 efficiently reduced the interaction, confirming that STAT3 binds preferentially to activated receptor (Fig. 4A and B).
STAT3 is a transcription factor that, when tyrosine phosphorylated, translocates to the nucleus where it regulates transcription of gene targets. The increased expression of c-fos, c-myc and JunB was observed after FGF1 stimulation in SUM-52PE cells (Fig. 5A). Because expression of these genes could also be induced by other pathways, including Ras/MAPK cascade, SUM-52PE cells were treated with Cucurbitacin I to confirm that the observed effect was due to STAT3 activity. STAT3 inhibition significantly reduced the expression of c-fos, c-myc and JunB (Fig. 5A) indicating that STAT3 contributes to regulation of the expression of early response transcription factors via its activation by FGFR2. Furthermore, tyrosine STAT3 phosphorylation and its subsequent nuclear accumulation in SUM-52PE cells were confirmed by immunofluorescence (Supplementary Fig. 3). Phosphorylation of TyrSTAT3 was effectively reduced with SU5402 treatment indicating importance of FGFR2 activity (Supplementary Fig. 3).
We next examined the effect of blocking FGFR and STAT3 activity on proliferation and viability of SUM-52PE cells. PD173074 treatment led to growth inhibition and cell death as assessed by either cell number-dependent catabolism of the dye MTT or LDH leakage to media (Fig. 5B and C). Cucurbitacin I treatment resulted in suppression of cell proliferation but did not induce cell death (Fig. 5B and C). Combined effect of both inhibitors gave similar results to PD173074 alone. These data suggest that while STAT3-independent FGFR2 activity in SUM-52PE cells is essential for their survival, FGF-activated STAT3 plays a role in regulation of cell proliferation. Similar dependence of receptor activity in breast and lung cancer cells over-expressing FGFR has been shown before (4, 22, 23). However, we cannot eliminate the formal possibility that STAT3 activation in SUM-52PE cells occurs by an FGR kinase dependent mechanism that does not involve Tyr677 and is not active in HEK293T cells.
To identify the role of Src and JAK in TyrSTAT3 activation, inhibitors for Src and JAK family members were pre-incubated with SUM-52PE cells prior to FGF1 stimulation (Fig. 6A). Both inhibitors significantly reduced TyrSTAT3 activation indicating that Src and JAK contribute to TyrSTAT3 phosphorylation in FGF1-stimulated SUM-52PE cells (Fig. 6A). Furthermore, the same effect was observed using other pharmacological inhibitors, Dasatinib and AG490 (Supplementary Fig. 4A). Additionally, we verified the observed effect by Src and JAK2 depletion using siRNA oligonucleotides. Knock-down experiment confirmed that both kinases, Src and JAK2, take part in FGF-induced TyrSTAT3 phosphorylation (Fig. 6B).
To further investigate the involvement of Src and JAK2 in STAT3-FGFR pathway we performed an immunoprecipitation experiment to establish if they associate with the receptor. Src kinase was identified as a binding partner for Tyr730 of FGFR1 in our initial peptide pull-down (Supplementary Table 1) and the interaction between Src and active FGFR1 was confirmed by immunoprecipitation (Fig. 6C). We transiently transfected four FcFGFR1 constructs into HEK293T cells: the FcFGFR1 kinase active, the kinase dead construct, the truncated version lacking the C-terminal intracellular domain and the FcFGFR1 VT- construct that is kinase active but unable to bind FRS2 (19). We found that Src interacted with kinase active and VT- FGFR1 (Fig. 6C). Increased phosphorylation of Src was demonstrated with kinase active and VT- FGFR1 expression (Fig. 6C). Also, the interaction between FGFR1 and myc-JAK2 was demonstrated. Similar to Src, JAK2-FGFR1 association was phospho-dependent and FRS2-independent (Fig. 6D). Phosphorylation of JAK2 was detected with kinase active FGFR1 and VT- suggesting that JAK2 acted downstream of FGFR1 and FGFR1 kinase activity was necessary for its activation (Fig. 6D). The reverse immunoprecipitation experiments for Src and JAK2 are presented in Supplementary Fig. 4B. Altogether, our results show that FGFR kinase activity is crucial for formation of a complex with Src and JAK2 which are activated downstream of FGFR. The results of pharmacological inhibition studies, as well as Src and JAK2 knock-down, reveal that both non-receptor kinases are required for tyrosine phosphorylation of STAT3.
In this study, a connection between over-expression of FGFRs and STAT3 phosphorylation in cancer cells is demonstrated. We show that FGF-induced STAT3 activation via tyrosine phosphorylation requires high expression level of FGFRs and is Src and JAK2-dependent. This was demonstrated in breast cancer cell line SUM-52PE which has amplified FGFR2 gene and elevated levels of receptor expression. STAT3 phosphorylation induced by FGF stimulation in these cells leads to nuclear accumulation and regulation of gene expression. Moreover, we show a phospho-dependent interaction between FGFRs and STAT3. The association between FGFR1WT and STAT3 is mediated by tyrosine 677 of FGFR1 that is part of a highly conserved YxxQ SH2-domain binding motif for STAT3 (21). We demonstrate that mutation of this tyrosine residue significantly reduced the binding of STAT3 to FGFR1WT. These data are with agreement with our initial peptide pull-down experiment followed by mass spectrometry where STAT3 was identified as a binding partner for Tyr677 of FGFR1. Furthermore, a mutation in the SH2 domain of STAT3 abolished the association with a Tyr677 FGFR1 peptide suggesting that SH2 domain is necessary for the interaction with the receptor. Thus, the physical interaction between phospho-Tyr677 of FGFR1 and SH2 domain of STAT3 is mandatory for STAT3 recruitment and its subsequent phosphorylation.
The interaction between STATs and growth factor receptors and subsequent STATs activation was previously demonstrated (24-28). FGF-mediated phosphorylation of STAT3 was described in cells harboring FGFR mutations however the mechanism of the interaction between them remained elusive. The ability to activate STAT1 and STAT3 seems to be restricted to FGFR3 with K650E mutation, which leads to constitutive activation of the receptor (29-32). The same mutation generated in other FGFR isoforms results in a similar activation of both STATs (33). STAT1 was also described as a downstream substrate for FGF signaling that negatively regulates proliferation of chondrocytes (34). However, Krejci et al. were unable to detect STAT1 tyrosine activation by FGF in RCS chondrocytes (35). Recently, FGF-induced activation of STAT5 via JAK2 and Src was described as necessary for angiogenesis (36). Here, we tested activation of STAT3 following FGF1 stimulation: SerSTAT3 phosphorylation was observed at endogenous level of FGFRs in several cell lines, whereas TyrSTAT3 phosphorylation required over-expression of FGFRs. Transient transfection of FGFR1WT in HEK293T or HeLa cells induced TyrSTAT3 phosphorylation and nuclear accumulation in contrast to cells with endogenous level of FGFRs. Additionally, FGF stimulation of SUM-52PE resulted in enhanced TyrSTAT3 phosphorylation which was significantly decreased by de-activation of the receptor using chemical inhibitors, as well as by FGFR2 knock-down. Thus, it is possible that oncogenic over-expression of FGFRs induces STAT3 activation as an additional signaling pathway amplified in cancer cells. Moreover, FGFR auto-phosphorylation is an ordered and sequential event that when disturbed, for example by an activating mutations, might contribute to aberrant assembly of FGFR binding partners (37). Thus, possible changes in the order of FGFR phosphorylation, including Tyr677, due to over-expression could lead to modification of adaptor proteins recruitment.
Serine STAT3 phosphorylation was believed to enhance STAT3 transcription activity (38) however, recently other aspects of SerSTAT3 activity are emerging (39) and the role of SerSTAT3 in FGFR signaling remains unclear. Udayakumar et al. demonstrated the induction of promatrilysin expression in prostate carcinoma cell line by FGF1-induced SerSTAT3 (40). Here, we show that STAT3 serine phosphorylation is induced by ERK and JNK as a downstream effect of FGFR signaling and it does not require over-expression of FGFRs. However, the physiological function of SerSTAT3 activation remains to be determined since the involvement of serine STAT3 in oncogenesis was suggested and its activity in cancer cells might be as important as tyrosine phosphorylated STAT3 (41).
We also investigated the potential role of the non-receptor tyrosine kinases, Src and JAK2. Both kinases act downstream of FGFR and their activity is also essential for TyrSTAT3 phosphorylation. Src has been described as an important kinase in STATs phosphorylation in many cell types under various conditions (24, 25, 28, 42) but its activity is usually dependent on upstream receptor tyrosine kinase activation (15). Src activation upon FGF stimulation was previously described (43) and both direct and indirect interaction between Src family members and FGFR was presented (20, 44). We show that complex formation between active FGFR1 and Src is FRS2-independent. We also demonstrate that Src potentially plays a direct role in the FGF-induced STAT3 phosphorylation. On the other hand, Src can be indirectly involved in STAT3 activation by maintaining activity and dynamics of FGFR as activation of Src and FGFR is inter-dependent (43).
STAT3 activation by various RTKs might be JAK-dependent or JAK-independent and both models of STAT3 phosphorylation are presented in the literature (25-28, 36, 45). In our study, we found that JAK2 was necessary for TyrSTAT3 phosphorylation in SUM-52PE cells but its activity depended on the concurrent tyrosine kinase activity of FGFR. Furthermore, immunoprecipitation results showed association between JAK2 and kinase active FGFR1. Phospho-dependent JAK2 association with FGFR1 suggests that JAK2 is recruited only to the activated receptor. The exact mechanism for JAK recruitment to FGFR remains to be further determined.
Finally, we report that STAT3 can be a mediator of amplified FGFR signaling in cancer cells, SUM-52PE, which is Src and JAK2-dependent process. We propose a model where over-expression of FGFR leads to recruitment of STAT3 to the receptor. At the same time Src and JAK2 form a complex with active receptor and both kinases are subsequently phosphorylated. Then, Src and JAK2 facilitate tyrosine STAT3 phosphorylation, which leads to STAT3 dimerization, translocation in the nucleus and regulates cell proliferation. Emerging evidence indicates that aberrant FGFR signaling leads to cancer development (8), however, the mechanism is so far unclear. STAT3, meanwhile, has been found to be active in various cancer cells and its role in cancer development is firmly established (16). This study supports a potential contribution of STAT3 in the process of tumorigenesis in cells over-expressing FGFRs, in which case the FGFR-STAT3 pathway might be an attractive therapeutic target.
We thank Nicholas Turner for SUM-52PE cells; Valeria Poli and Pamela Maher for constructs. We are grateful to Sue Brewer for technical support and to members of Heath’s group for useful discussions. The work was funded by the FP6 Endotrack grant number REAZ 12101 and Cancer Research UK.