Ras/ERK1/2-mediated STAT3 Ser727 Phosphorylation by Familial Medullary Thyroid Carcinoma-associated RET Mutants Induces Full Activation of STAT3 and Is Required for c-fos Promoter Activation, Cell Mitogenicity, and Transformation*

The precise role of STAT3 Ser727 phosphorylation in RET-mediated cell transformation and oncogenesis is not well understood. In this study, we have shown that familial medullary thyroid carcinoma (FMTC) mutants RETY791F and RETS891A induced, in addition to Tyr705 phosphorylation, constitutive STAT3 Ser727 phosphorylation. Using inhibitors and dominant negative constructs, we have demonstrated that RETY791F and RETS891A induce STAT3 Ser727 phosphorylation via a canonical Ras/ERK1/2 pathway and that integration of the Ras/ERK1/2/ELK-1 and STAT3 pathways was required for up-regulation of the c-fos promoter by FMTC-RET. Moreover, inhibition of ERK1/2 had a more severe effect on cell proliferation and cell phenotype in HEK293 cells expressing RETS891A compared with control and RETWT-transfected cells. The transforming activity of RETY791F and RETS891A in NIH-3T3 cells was also inhibited by U0126, indicating a role of the ERK1/2 pathway in RET-mediated transformation. To investigate the biological significance of Ras/ERK1/2-induced STAT3 Ser727 phosphorylation for cell proliferation and transformation, N-Ras-transformed NIH-3T3 cells were employed. These cells displayed elevated levels of activated ERK1/2 and Ser727-phosphorylated STAT3, which were inhibited by treatment with U0126. Importantly, overexpression of STAT3, in which the Ser727 was mutated into Ala (STAT3S727A), rescued the transformed phenotype of N-Ras-transformed cells. Immunohistochemistry in tumor samples from FMTC patients showed strong nuclear staining of phosphorylated ERK1/2 and Ser727 STAT3. These data show that FMTC-RET mutants activate a Ras/ERK1/2/STAT3 Ser727 pathway, which plays an important role in cell mitogenicity and transformation.

The precise role of STAT3 Ser 727 phosphorylation in RET-mediated cell transformation and oncogenesis is not well understood. In this study, we have shown that familial medullary thyroid carcinoma (FMTC) mutants RET Y791F and RET S891A induced, in addition to Tyr 705 phosphorylation, constitutive STAT3 Ser 727 phosphorylation. Using inhibitors and dominant negative constructs, we have demonstrated that RET Y791F and RET S891A induce STAT3 Ser 727 phosphorylation via a canonical Ras/ERK1/2 pathway and that integration of the Ras/ERK1/2/ELK-1 and STAT3 pathways was required for up-regulation of the c-fos promoter by FMTC-RET. Moreover, inhibition of ERK1/2 had a more severe effect on cell proliferation and cell phenotype in HEK293 cells expressing RET S891A compared with control and RET WT -transfected cells. The transforming activity of RET Y791F and RET S891A in NIH-3T3 cells was also inhibited by U0126, indicating a role of the ERK1/2 pathway in RET-mediated transformation. To investigate the biological significance of Ras/ERK1/2-induced STAT3 Ser 727 phosphorylation for cell proliferation and transformation, N-Ras-transformed NIH-3T3 cells were employed. These cells displayed elevated levels of activated ERK1/2 and Ser 727phosphorylated STAT3, which were inhibited by treatment with U0126. Importantly, overexpression of STAT3, in which the Ser 727 was mutated into Ala (STAT3 S727A ), rescued the transformed phenotype of N-Ras-transformed cells. Immunohistochemistry in tumor samples from FMTC patients showed strong nuclear staining of phosphorylated ERK1/2 and Ser 727 STAT3. These data show that FMTC-RET mutants activate a Ras/ERK1/2/STAT3 Ser 727 pathway, which plays an important role in cell mitogenicity and transformation.
Different activating missense mutations in the RET protooncogene cause multiple endocrine neoplasia type 2 (MEN2), a dominant inherited cancer syndrome affecting several neuroendocrine tissues (3). Three different clinical subtypes, MEN2A, MEN2B, and familial medullary thyroid carcinoma (FMTC) (3), can be recognized depending on the affected tissues and mutations found. However, despite a clear phenotypegenotype correlation, the molecular mechanisms connecting the mutated receptors with the different clinical subtypes are largely unknown (4).
In MEN2, aberrant activation of STAT3 through Tyr 705 phosphorylation by mutated RET receptors has been reported (5)(6)(7). STAT3 is a latent transcription factor implicated in several types of cancer when aberrantly activated (8,9). Activation of STAT3 is triggered by phosphorylation of Tyr 705 in its Src homology 2 domain, resulting in dimerization, nuclear translocation, and transcriptional activation of target genes. Additionally, phosphorylation of STAT3 on Ser 727 , situated in the C-terminal transactivation domain (Fig. 1), results in enhanced transcriptional activation and DNA binding capacity (10). Various kinases have been shown to phosphorylate STAT3 on Ser 727 , depending on the cytokines and growth factors involved and the cellular context (10 -14). However, the precise role of STAT3 Ser 727 phosphorylation in cell transformation and in RET-mediated oncogenesis was unknown.
Here we have investigated the different signaling pathways activated by FMTC-RET at the level of STAT3 Ser 727 and whether the deregulation of such pathways by oncogenic RET plays a crucial role in cell mitogenicity and transformation. In this study, we have shown that RET Y791F and RET S891A activate a Ras/ERK1/2/STAT3 Ser 727 pathway required for cell transformation and tumor cell proliferation, and we further extend our findings to tumor samples from patients carrying a germ line-activating RET S891A mutation.

EXPERIMENTAL PROCEDURES
Cell Lines and Cell Culture Reagents-Human embryonic kidney 293 (HEK293) cells were grown as described previously (7). NIH-3T3 cells and the N-Ras-transformed NIH-3T3 cell line 149169 were provided by Prof. C. Marshall (The Institute of Cancer Research, London, UK) and grown in Dulbecco's modified Eagle's medium, supplemented with 5% donor calf serum (Invitrogen), 100 IU/ml penicillin, 1 mg/ml streptomycin, and 2 mM L-glutamine (Invitrogen). To generate stably transfected cell lines, 1 ϫ 10 6 HEK293, NIH-3T3, or 149169 cells were plated in 90-mm dishes and transfected with 0.5 g of expression plasmid by the calcium phosphate method. 24 h after transfection, the cells were rinsed with phosphate-buffered saline and cultured in fresh medium containing 500 g/ml G418 (Sigma-Aldrich). Clones were picked after 2 weeks and screened by Western blot analysis. GDNF (Peprotech, Rocky Hill, NJ), STI571 (Novartis, Basel, Switzerland), PP2, and U0126 (Promega, Madison, WI) were used as indicated.
Transformation Foci and Soft Agar Assays-Transformed foci assays were carried out as previously described (18). Briefly, NIH-3T3 and 149169 cells were transfected with 0.5 g of the indicated plasmid using calcium phosphate. The next day, cells were washed with phosphate-buffered saline, and fresh medium containing 500 g/ml G418 was added. After 3 weeks, the cells were fixed in 4% paraformaldehyde for 1 h and then washed and stained for 1 h with 0.1% crystal violet.
For soft agar colony formation assays, 1% agar (DNA grade; Invitrogen) was boiled and mixed 1:1 with 2ϫ Dulbecco's modified Eagle's medium/F-12 with standard additives and 5 ml added to 90-mm dishes to form the base agar. For the top agar, 2 ϫ 10 4 cells were mixed with 5 ml of a 1:1 mix of 0.7% agarose and 2ϫ Dulbecco's modified Eagle's medium/F-12 and poured on the top of the base agar. Plates were incubated at 37°C in a humidified incubator for 10 -14 days prior to counting of colonies.
Luciferase Reporter Assays-Reporter assays were undertaken as previously described (5,7,16). Briefly, 24 h after transfection, cells were washed and treated with ligand or inhibitors as indicated. The following day, cells were harvested in lysis buffer (Promega, Madison, WI), and luciferase activity was determined using the SteadyLite HTS kit (PerkinElmer Life Sciences). In all transfections, a ␤-galactosidase expression plasmid (pDM2LacZ) was included to normalize luciferase activities. ␤-Galactosidase activity was determined in 100 mM Na 2 HPO 4 /NaH 2 PO 4 , 1 mM MgCl 2 , 100 mM 2-mercaptoethanol, and 0.67 mg/ml O-nitrophenylgalactopyranoside. Data represent three independent experiments performed in triplicate.
Statistical Analysis-Statistical analysis was performed using Prism software. Cell and colony number assays were compared using one-way analysis of variance for repeating measures and the Dunnett test. Differences were considered statistically significant when p Ͻ 0.05.

RESULTS
RET Y791F and RET S891A , two mutations targeting the tyrosine kinase domain of RET and associated with the FMTC phenotype, signal independently of GDNF and induce STAT3 Tyr 705 phosphorylation through a Src-and Janus kinase-dependent pathway (7). To determine whether RET Y791F and RET S891A contribute to the aberrant transcriptional activity of STAT3, first we investigated whether these mutants also induce STAT3 Ser 727 phosphorylation. Western blot analyses of HEK293 cells expressing RET WT , RET Y791F , and RET S891A indicated that both mutant receptors are activated in the absence of GDNF, as shown by RET Tyr 1062 phosphorylation (Fig. 1A). Furthermore, these mutant receptors induced STAT3 phosphorylation on Tyr 705 and Ser 727 (Fig. 1A). Next, the effect of RET S891A -induced STAT3 Ser 727 phosphorylation on the transcriptional activity of STAT3 was investigated. STAT3␣, STAT3␤, and a STAT3 point mutant, STAT3 S727A (Fig. 1B), were expressed in HEK293 cells in combination with RET S891A and an interleukin-6 response element (IRE) luciferase reporter. Activation of the IRE-reporter by RET S891A was increased by co-expression of STAT3␣ but not by co-expression of either STAT3␤ or STAT3 S727A (Fig. 1B). The same reporter experiments were performed using the promoter regions of the Cyclin D1, ICAM1 and Bcl-x L genes fused to a luciferase reporter gene. In all promoters tested, STAT3␣ potentiated reporter activation by RET S891A , whereas co-expression of STAT3 S727A did not result in increased transcriptional activation (data not shown). These data indicate that STAT3 Ser 727 phosphorylation is important for RET-induced transcriptional activation of STAT3 target genes and may be required for RET-mediated oncogenesis.
STAT3 Ser 727 is located in a consensus sequence Pro-Xaa-(Ser/ Thr)-Pro that can be phosphorylated by ERK1/2 (20). As ERK1/2 can be activated by wild-type RET (21), a number of approaches were taken to investigate whether FMTC-RET mutants aberrantly activate the ERK1/2-MAPK pathway in the absence of GDNF. First, HEK293 cells were transfected with RET WT , RET Y791F , or RET S891A in combination with an SRE-luciferase reporter, as a readout for downstream activation of ERK1/2 (17). In the absence of GDNF, RET Y791F and RET S891A induced a 3-and 5-fold activation of the SRE reporter, respectively, whereas no induction was observed with RET WT (Fig. 2A). To confirm these results, Western blot analysis of HEK293 cells transiently transfected with RET WT , RET Y791F , or RET S891A showed increased levels of ERK1/2 phosphorylation by both mutants (Fig. 2B).
The SRE element is known to be activated by ELK-1, an ERK1/2 target. Consequently, HEK293 cells were transfected with the RET constructs and a GAL4/ELK-1 reporter system in which ELK-1 is fused to the DNA binding domain of GAL4 in combination with an upstream activating sequence (UAS)-Luc reporter (Fig. 2C). Robust activation of the UAS-reporter by the RET Y791F and RET S891A mutants (200-and 300-fold, respectively) was observed compared with a lower activation observed with RET WT (70-fold). To further investigate the activation of ERK1/2 by mutant FMTC-RET receptors, the mRNA levels of c-fos and Egr-1 genes in established HEK293-expressing RET Y791F and RET S891A were analyzed using reverse transcription-PCR. Up-regulation of both of these genes was observed in the mutant cell lines when compared with control HEK293 cells (Fig. 2D). Taking these results together, we conclude that both RET Y791F and RET S891A activate the ERK1/2 pathway independent of GDNF.
To determine whether the STAT3 Ser 727 phosphorylation induced by FMTC-RET was mediated through a Ras/ERK1/2 pathway, established HEK293 cells expressing RET S891A were treated during 24 h with the tyrosine kinase receptor inhibitor STI571 (22). A reduction of RET Tyr 1062 phosphorylation accompanied by a reduction of RET expression (22) was observed at 20 M STI571. Treatment with this inhibitor also resulted in a reduction in phosphorylation of STAT3 Ser 727 and ERK1/2 (Fig. 3A). These results suggest that RET signaling is required for STAT3 Ser 727 and ERK1/2 phosphorylation. To delineate the pathway from RET to ERK1/2 activation, GAL4-ELK-1/UAS-Luc reporter assays were used. Co-expression of either dn-Ras (RasN17) or dn-Raf (a kinase-dead Raf construct) resulted in a complete loss of UAS-reporter activation, indicating that Ras and Raf are required for ELK-1 activation by RET Y791F and RET S891A (Fig. 3B). To integrate the Ras/ERK1/2 pathway with STAT3 signaling, the IRE-Luc reporter was used in combination with a series of dominant negative constructs (Fig. 3B). dn-Ras and dn-Raf partially inhibited reporter activation by RET S891A , whereas expression of a dn-SEK1 (15) had no effect on RET S891A -induced IRE-Luc activity, indicating that Ras and Raf (but not SEK1) are involved in ERK1/2-mediated STAT3 activation. To test whether MEK1/2 (a MAPK kinase upstream of ERK1/2) was activating ERK1/2 in response to RET S891A , the effect of the MEK1/2 inhibitor U0126 on RET S891A -induced UAS-Luc and IRE-Luc activation was assessed. U0126 (10 and 40 M) reduced both UAS-Luc and IRE-Luc activation by RET S891A (Fig. 3C). To further confirm these results, Western analysis demonstrated that U0126 was able to inhibit RET S891A -induced ERK1/2 and STAT3 Ser 727 phosphorylation (Fig. 3D). Together these data demonstrated that RET S891A -induced STAT3 Ser 727 phosphorylation is mediated by a canonical Ras/Raf/MEK1-2/ERK1-2 pathway. To elucidate whether FMTC-associated RET mutant signaling differs from GDNF-stimulated wild-type RET, Western analyses were performed (Fig. 4). RET WT , when stimulated with GDNF in the presence of GFR␣-1, was able to induced ligand-dependent receptor activation followed by ERK1/2 and slight STAT3 Ser 727 phosphorylation, which was independent of STAT3 Tyr 705 phosphorylation (Fig. 4). In contrast, RET S891A induced complete activation of the STAT3 pathway as promoted by both Tyr 705 and Ser 727 STAT3 phosphorylation and also ERK1/2 activation (Fig. 4) in the absence of the ligand. When established RET WT cells were treated with the Src and RET kinase inhibitor PP2, a decrease in receptor phosphorylation was accompanied by a decrease in ERK1/2 and STAT3 Ser 727 phosphorylation (Fig. 4).
FMTC-RET mutants induced c-fos up-regulation (Fig. 2). The c-fos promoter contains various transcription factor binding sites; among them there are STATs (SIE) and ELK-1 (TCF) binding elements (17). To determine a possible cooperativity between STAT3 and ERK1/2-ELK-1 pathways in c-fos promoter activation by FMTC-RET, luciferase reporter assays were performed in established cell lines expressing vector alone (control), RET WT , and RET WT transiently transfected with GFR␣-1 and stimulated with GDNF and RET S891A (Fig. 5). Activation of the wild-type c-fos promoter was seen by RET S891A and GDNF-stimulated RET WT but not by RET WT or vector (alone)-expressing cells (Fig. 5). A partial reduction in reporter activation by RET S891A was observed with the p18 mutant (in which the TCF binding site required for ELK-1 binding is mutated) and the SIE mutant (in which the STAT binding site is mutated) c-fos promoter constructs. However, when both transcriptional binding sites were mutated in the c-fos promoter (p18/SIE mutant), reporter activation by RET S891A was completely abolished (Fig. 5). GDNF-stimulated RET WT induced a ligand-dependent activation of the c-fos promoter but to a lesser extent compared with RET S891A . When the p18 mutant c-fos promoter was used, a 2-fold decrease in promoter activation was seen, whereas a weak effect was observed with the SIE mutant (Fig. 5). This supports the data presented in Fig. 4; hence we concluded that activation of the Ras/ERK1/2/ELK-1 and STAT3 pathways is required for ligand-independent c-fos promoter upregulation by FMTC-RET.
To further investigate the biological significance of the functional interaction between ERK1/2 and STAT3, we determined the effect of the MEK1/2 inhibitor U0126 on the morphology and proliferation of established HEK293 cells. First, we examined the phenotype of established HEK293 cells transfected with vector alone (control), RET WT , or RET S891A (Fig. 6A). HEK293-RET WT cells displayed epithelial morphological features, a low number of cell extensions, and thin distal cell processes. They were indistinguishable from vector (alone)-transfected HEK293 cells. In contrast, HEK293-RET S891A cells showed a less differentiated phenotype, being more scattered and with a higher number of neural-like extensions and thin distal cell processes (Fig. 6A, arrows). When HEK293-RET S891A cells were cultured for 5 days in the presence of U0126 (10 M), a lower number of neural-like extensions and thin distal cell processes was observed, as well as a lower degree of cell scattering (Fig. 6A). Next, we inhibited STAT3 activation by RET S891A with PP2 (7). Treatment of cells with PP2 (3 M) reverted the phenotype of the mutant cell lines as well (Fig. 6A). No effect of either inhibitor was seen in HEK293 control and HEK293-RET WT (data not shown). In addition to the cell morphology analysis, proliferation assays demonstrated that HEK293-RET S891A cells grew faster than HEK293 control and HEK293-RET WT cells. Furthermore, HEK293-RET S891A cells had a higher sensitivity for the MEK1/2 inhibitor, as the proliferation rate was significantly more impaired by U0126 compared with HEK293 control and HEK293-RET WT cells (Fig. 6B).
To investigate the importance of the Ras-ERK1/2 pathway in the transforming activity of RET Y791F and RET S891A , NIH-3T3 cell transformation foci assays (18) were performed in the presence or absence of the inhibitors U0126 or PP2 (Fig. 7). RET Y791F and RET S891A were able to induce transformed foci when compared with cells transfected with RET WT . As a positive control, RET C634R was used. When cells were cultured in the presence of U0126 or PP2, the transforming capacity of both FMTC mutants was significantly reduced (Fig. 7). These data, in combination with the previous results, indicate that ERK1/2 activation is required for STAT3 Ser 727 phosphorylation and for the proliferation and transforming activity of FMTC-RET mutants and it further points toward the ERK1/2 and STAT3 pathways as targets for therapeutic intervention in MTCs (see "Discussion").
To demonstrate that integration of the ERK1/2 and STAT3 pathways is required for cell transformation and that STAT3 Ser 727 phosphorylation plays a crucial role in this process, experiments were undertaken using an NIH-3T3 cell line expressing the N-Ras oncogene (149169 cell line). First, Western blot analysis of cell lysates from NIH-3T3 and 149169 cells demonstrated that, in the N-Ras-transformed cells, high levels of phosphorylated ERK1/2 correlated with high levels of Ser 727 phosphorylated STAT3 (Fig. 8A). When cells were treated with the MEK1/2 inhibitor U0126, a reduction in phosphorylated ERK1/2 and STAT3 on Ser 727 was observed (Fig. 8A). Second, the transforming activity of 149169 cells transfected with STAT3␣, STAT3␤, or STAT3 S727A was scored by transformed foci. Cells transfected with STAT3␣ resulted in a significantly increased number of transformed foci when compared with 149169 cells transfected with STAT3␤ or STAT3 S727A mutant (Fig. 8B). Next, the transformed phenotype of established 149169 cells expressing the distinct STAT3 molecules was assessed. Vector (alone)-transfected 149169 cells showed typical transformed features, such as long neural-type elongations, poorly differentiated shape, and high polarization with a small cytoplasm and nucleus (Fig. 8C, arrows). 149169 cells expressing STAT3␣ displayed large filaments and a poorly differentiated phenotype (see arrows) with a small cellular body, and they showed reduction in contact inhibition compared with vector (alone)-transfected 149169 cells (Fig. 8C). Cells expressing STAT3 S727A showed different morphological features, such as less pronounced elongations and thin distal cell processes, a more differentiated mesenchymal phenotype, less polarization with an evident cytoplasm and nucleus, and a fibroblast-type shape typical of parental NIH-3T3 cells, suggesting a reversal in the transformed features (Fig. 8C, asterisks).
Transformed cells are characterized by their capacity to proliferate faster and growth in an anchorage-independent manner. To assess these oncogenic features, proliferation and soft agar assays were performed. As expected, proliferation of N-Ras-transformed 149169 cells was higher than the parental NIH-3T3 (Fig. 8D). In 149169 cells expressing STAT3␣, the proliferation rate was slightly enhanced, whereas for 149169 cells expressing STAT3 S727A , the proliferation rate was significantly reduced compared with 149169 cells expressing STAT3␣ (Fig. 8D). In the same line of evidence, N-Ras-transformed cells were able to grow in an anchorage-independent manner when compared with the parental NIH-3T3, whereas 149169 cells expressing STAT3 S727A showed no increase in colony formation in soft agar. In contrast, the 149169 cells overexpressing STAT3␣   showed a significantly increased anchorage-independent growth when compared with 149169 cells expressing STAT3 S727A (Fig. 8D). Similar results were obtained with two independent clones for each cell line (data not shown). Taking these results together, we conclude that the Ras/ ERK1/2/STAT3 Ser 727 pathway plays an important role in cell transformation and tumor cell proliferation.
To investigate the in vivo relevance of these findings, tumor samples from five patients carrying a RET S891A germ line mutation were analyzed. Data from one patient are shown in Fig. 9. Similar results were obtained with the samples of the other four patients (data not shown). All tumor samples displayed high levels of RET expression at the plasma membrane and in the cytoplasm (Fig. 9, A and B). A section containing both normal (N) and tumor (T) tissue was analyzed as control (Fig. 9C). In the area containing normal tissue, highly organized thyroid follicles with low levels of RET expression were observed (Fig. 9C,  arrows). In the area containing the tumor cells, disorganized tissue with loss of the follicular pattern and high levels of RET expression were detected (Fig. 9C). Strong nuclear staining for Ser 727 -phosphorylated STAT3 was seen in all MTC tumors (Fig. 9, D and E). As a control, a biopsy section combining normal and tumor tissue was shown (Fig. 9F). Weak staining was observed for Ser 727 -phosphorylated STAT3 in areas lacking tumor tissue, as seen in the follicular cells (see arrows), whereas strong staining was detected in the area where the hyperplasia of the C cell localizes (Fig. 9F). Finally, strong nuclear and cytoplasmic phospho-ERK1/2 staining was observed in the tumor samples (Fig. 9, G and H). As a control, a section showing normal tissue is depicted (Fig. 9I). In areas where normal follicular epithelium was seen, low levels of phosphorylated ERK were detected (Fig. 9I, arrows). Controls performed with depletion of the primary antibody showed no staining in any of the tumors tested (data not shown).

DISCUSSION
RET signaling and the molecular genetics of RET-associated diseases are among the best examples of phenotypic heterogeneity and cross-talk of signaling pathways. However, despite a clear phenotype-genotype correlation, the molecular mechanisms connecting the different mutant receptors with their associated clinical phenotypes are not well understood (3). FMTC is characterized only by tumors arising from the C-cells of the thyroid, and hence it is considered the less aggressive clinical subtype of the cancer syndrome MEN2 (3). However, the understanding of the signaling pathways activated by intracellular point mutations targeting the tyrosine kinase domain of RET, in particular those mutations associated with the FMTC phenotype, is very limited. Iwashita and co-workers (18) show that RET S891A displays constitutive tyrosine phosphorylation and was able to induce transformed foci in NIH-3T3 cells. In another study, Pasini et al. (23) show similar results with RET E768D and RET V804L (two mutations associated with FMTC). Recently, a work by Mise et al. (24) has shown that RET Y791F increased the proliferative properties, as indicated by cells in S-phase, colony formation in soft agar, and tumor growth in nude mice. RET Y791F also increased apoptotic resistance, which was accompanied by enhanced levels of active AKT and BCL2 expression. The downstream signaling of FMTC-RET mutants was also investigated by Plaza-Menacho et al. (7). We showed that RET Y791F and RET S891A signaled independently of GDNF as monomeric oncoproteins and that RET Y791F and RET S891A induced STAT3 Tyr 705 phosphorylation via Src-and Janus kinase 1/2dependent mechanism (7). This study has addressed whether STAT3 Ser 727 phosphorylation was induced by RET Y791F and RET S891A , the pathway connecting these FMTC-RET receptors with STAT3 Ser 727 phosphorylation, and whether phosphorylation of this residue contributes to aberrant transactivation and cell mitogenicity and transformation.
First, RET Y791F and RET S891A are able to induce STAT3 Ser 727 phosphorylation in the absence of GDNF. To prove the importance of this phosphorylation event, STAT3 S727A reduced the levels of RET-mediated STAT3 transcriptional activity. These results are comparable with those obtained by Huang et al. (25) in which the activation of a STAT3-responsive reporter by the papillary thyroid carcinoma-RET variant RET/ PTC1 was reduced by STAT3 S727A . However, despite activating STAT3 on Tyr 705 , RET/PTC1 did not induce phosphorylation of STAT3 on Ser 727 (25).
Second, we demonstrated that FMTC-RET was also able to constitutively activate a Ras/ERK1/2/ELK-1 pathway (Fig. 2). Levels of ERK1/2 phosphorylation and SRE and ELK-1 reporter activation displayed by RET S891A were higher than RET Y791F . These results correlated with the levels of receptor activation and the degree of STAT3 activation previously observed (Fig.  1A). These data suggest that the RET S891A mutant has a higher capacity to trigger proliferative signals. Transformation foci assays performed in NIH-3T3 cells indeed showed increased transforming capacity of RET S891A than RET Y791F . Interest-ingly, a study by Jimenez et al. (26) shows that patients with the germ line RET S891A mutation develop both MTC and pheochromocytoma, indicating that this mutation could be considered as MEN2 and hence associated with a stronger phenotype. The transforming activity of both FMTC-RET mutants was affected by U0126, suggesting an important role of the Ras/ERK1/2 MAPK pathway in RET-mediated transformation (Fig. 7). Recent studies support this conclusion; for example, a report by Sawai et al. (27) demonstrates that the polymorphism RET G691S increases the levels of ERK1/2 activation and induces pancreatic cell invasion. Another recent study by Melillo et al. (28) shows that activation of the RET/ PTC-Ras/Raf/ERK1/2 pathway induces cell proliferation and invasion of thyroid follicular cells via up-regulation of the CXCL1 and CXCL10 chemokines. Third, we delineated a Ras/ERK1/ 2/STAT3 Ser 727 pathway triggered by RET S891A (Fig. 3). To find specific signaling profiles between wild-type RET and oncogenic RET, we showed that GDNF-stimulated wild-type RET signaled preferentially via the Ras/ERK1/2 pathway rather than the STAT3 route, in contrast to oncogenic FMTC-RET, which was able to induce full activation of the STAT3 pathway as well constitutive activation of ERK1/2 (Fig. 4). Furthermore, we have shown that the integration of both STAT3 and Ras/ERK1/2/ELK-1 pathways was required for c-fos promoter activation by RET S891A (Fig.  5). These data suggest that c-fos could be one of the key early genes implicated in the pathogenesis of FMTC. Indeed, cDNA microarray studies using various established cell lines expressing FMTC-mutated RET support this hypothesis. 3 The biological importance of the Ras/ERK1/2/STAT3 Ser 727 pathway in cell mitogenicity and transformation was analyzed in a model based on an N-Ras NIH-3T3-transformed cell line (149169). These cells displayed higher levels 3 I. Plaza-Menacho and R. M. W. Hofstra, manuscript in preparation. of ERK1/2 and STAT3 Ser 727 phosphorylation than the parental NIH-3T3 cell line (Fig. 8A). Inhibition of MEK1/2 by U0126 resulted in a decrease in ERK1/2 and STAT3 Ser 727 phosphorylation (Fig. 8A). Furthermore, stable expression of a STAT3 S727A mutant in 149169 cells showed a reversion of the transformed phenotype (Fig. 8C). The cells showed less transforming activity, less anchorage-independent growth, and less proliferation rate compared with 149169 cells or 149169-overexpressing STAT3␣ (Fig. 8D). These results suggest that (i) STAT3, in particular when phosphorylated on Ser 727 , is one of the transcription factors required by Ras for cell transformation; (ii) these two independent pathways (Ras/ERK1/2 and STAT3) can cooperate during signaling and, when deregulated, play an important role in cell mitogenicity and transformation; (iii) targeting both pathways with small inhibitor molecules should be considered as a therapeutic strategy for RET-related tumors.
Immunohistochemical analyses of tumor samples from patients carrying the germ line RET S891A mutation supported our in vitro data, as high levels of RET expression at the plasma membrane were observed in combination with strong nuclear staining of both phospho-Ser 727 STAT3 and phospho-ERK1/2 (Fig. 9). These data, in combination with previous studies (7), suggest that activation of both the ERK1/2 and STAT3 pathways occurs in vivo in MTCs.
In conclusion, our data give new insights into the signaling networks implicated in FMTC. We show that aberrant activation of STAT3 by FMTC-RET not only involves the constitutive activation of STAT3 by Tyr 705 phosphorylation but that these mutants further enhance the transcriptional activity of STAT3 by constitutive phosphorylation of Ser 727 via a Ras/ERK1/2 pathway. These data show that deregulation of the Ras/ERK1/2/STAT3 Ser 727 pathway plays an important role in cell transformation and in the development of MTC and points toward the inhibition of this pathway as a potential therapeutic strategy for patients with RET-related neuroendocrine tumors.