Cross-Regulation between Oncogenic BRAFV600E Kinase and the MST1 Pathway in Papillary Thyroid Carcinoma

Background The BRAFV600E mutation leading to constitutive signaling of MEK-ERK pathways causes papillary thyroid cancer (PTC). Ras association domain family 1A (RASSF1A), which is an important regulator of MST1 tumor suppressor pathways, is inactivated by hypermethylation of its promoter region in 20 to 32% of PTC. However, in PTC without RASSF1A methylation, the regulatory mechanisms of RASSF1A-MST1 pathways remain to be elucidated, and the functional cooperation or cross regulation between BRAFV600E and MST1,which activates Foxo3,has not been investigated. Methodology/Principal Findings The negative regulators of the cell cycle, p21 and p27, are strongly induced by transcriptional activation of FoxO3 in BRAFV600E positive thyroid cancer cells. The FoxO3 transactivation is augmented by RASSF1A and the MST1 signaling pathway. Interestingly, introduction of BRAFV600Emarkedly abolished FoxO3 transactivation and resulted in the suppression of p21 and p27 expression. The suppression of FoxO3 transactivation by BRAFV600Eis strongly increased by coexpression of MST1 but it is not observed in the cells in which MST1, but not MST2,is silenced. Mechanistically, BRAFV600Ewas able to bind to the C-terminal region of MST1 and resulted in the suppression of MST1 kinase activities. The induction of the G1-checkpoint CDK inhibitors, p21 and p27,by the RASSF1A-MST1-FoxO3 pathway facilitates cellular apoptosis, whereasaddition of BRAFV600E inhibits the apoptotic processes through the inactivation of MST1. Transgenic induction of BRAFV600Ein the thyroid gland results in cancers resembling human papillary thyroid cancers. The development of BRAFV600Etransgenic mice with the MST1 knockout background showed that these mice had abundant foci of poorly differentiated carcinomas and large areas without follicular architecture or colloid formation. Conclusions/Significance The results of this study revealed that the oncogenic effect of BRAFV600E is associated with the inhibition of MST1 tumor suppressor pathways, and that the activity of RASSF1A-MST1-FoxO3 pathways determines the phenotypes of BRAFV600E tumors.


Introduction
Activating mutations in the BRAF gene are found at high frequency in various human cancers, and BRAF V600E is the most common of these activating mutations, especially in papillary thyroid cancer, where it is found at a frequency of 40-70% [1,2,3]. In BRAF V600E -positive thyroid cancer cell lines and BRAF V600E transgenic mice, this mutation is responsible for tumor initiation, transformation, growth, proliferation and dedifferentiation [4,5,6]. Research into the molecular mechanisms of BRAF V600E -positive tumors has revealed that the missense valine to glutamic acid mutation increases kinase activity, promoting the constitutive activation of MEK-ERK signaling [7,8,9,10,11] and enhancing ERK-dependent transcriptional output [12,13]. However, other signaling pathways except MEK-ERK [14,15]regulated in BRAF V600E tumors are not fully characterized [16]. Moreover, tumor suppressor systems which may be controlled by BRAF V600E in thyroid cancer remain to be identified.
The tumor suppressor gene RASSF1A (Ras association domain family 1A) is epigenetically inactivated through de novo promoter methylation in the early stages of thyroid tumorigenesis [17,18]. Interestingly, RASSF1A has recently been described as an important activator of MST1, which in turn phosphorylates and promotes the nuclear translocation of the forkhead transcription factor FKHRL1 (FoxO3), inducing cell death [19,20,21]. This suggests that FoxO3 transactivation could be induced by the RASSF1A-MST1 pathway and function as a tumor suppressor system in response to specific oncogenic signals, such as BRAF V600E . However, promoter hypermethylation of RASSF1A could only be detected in a relatively small percentage of PTC (20 to 32%) [18,22]. These observations predict that novel RASSF1A-MST1-FoxO3 pathways regulated by BRAF V600E might function during the development of PTC whichdoes nothave RASSF1Apromoter methylation.
FoxO3 transactivation is effectively inhibited by RET/PTC (rearranged in transformation/papillary thyroid carcinomas) kinase [23], the gene rearrangement of which is the most common rearrangement in papillary thyroid cancer. The inactivation of FoxO3 could therefore be a signature molecular event that should also occur in BRAF V600E thyroid tumors. Several molecular mechanisms which are possibly regulated by BRAF V600E may control FoxO3 activity in thyroid cancer. First, as RET/PTC kinase inhibits FoxO3 transactivation through an Akt/PKB dependent pathway, BRAF V600E might also activate Akt/PKB signaling pathway [24]. Second, the constitutive activation of ERK by BRAF V600E could inhibit FoxO3 activation via the ubiquitinproteasome pathway [25]. Finally, BRAF V600E could act through crosstalk with the RASSF1A-MST1-FoxO3 pathway.
Based on these hypotheses, we decided to investigate the regulation of the MST1-FoxO3 pathway by BRAF V600E . This resulted in the identification of novel cross-talksignaling between BRAF V600E and MST1, thereby demonstrating the functional activity of the RASSF1A-MST1-FoxO3 tumor suppressor system. Furthermore, in vivo experiments showed that MST1 knockout mice exhibited more aggressive BRAF V600E tumor phenotypes.

Cell culture and transfection
FRO cells and 293T cells were cultured in RPMI and DMEM supplemented with 10% fetal bovine serum [20,30]. Cells (2610 5 / 35 mm dish) were transfected with siRNA at a concentration of 50 nM using the HiPerfect Transfection Reagent (Qiagen, Valencia, CA) and/or with the indicated vectors using Fugen6 Transfection Reagent (Roche Applied Science, Indianapolis, IN). The MEK inhibitor (U0126) and PI3K inhibitors (Ly294002, Wortmannin) were purchased from Cell Signaling Technology (Beverley, MA).

Luciferase-based reporter assay
To assay the activity of FoxO3, 293T cells were transfected with 3X-IRS-Luc, or p27-Luc and pRL-SV40, which encodes Renilla luciferase (Promega, Madison, WI) using Lipofectamine in OptiMEM (Invitrogen). Luciferase levels were determined using the Dual Luciferase Assay Kit (Promega). Firefly and Renilla luciferase activities were measured sequentially from the same sample using a Fusion Alpha-Microplate Analyzer (Perkin Elmer, Waltham, MA). For each sample, firefly luciferase activity was normalized to Renilla luciferase activity.

Immunoprecipitation and Immunoblot analysis
All immunoprecipitation and immunoblotting procedures were performed at 4uC. Cells were washed twice with PBS and lysed in lysis buffer containing protease inhibitors. Lysates were centrifuged at 14,000 g for 15 min at 4uC. For the IP analysis, lysates were precleared with protein A/G beads (Santa Cruz Biotechnology) for 30 min, and supernatants were incubated with the indicated primary antibody for 3 h with agitation. Protein A/G beads were then added to the mixture and incubated for at least 2 h at 4uC. Beads were washed with lysis buffer and the immunoprecipitates were subjected to immunoblotting. Cell lysates were separated by SDS-PAGE, and transferred to nitrocellulose membranes. Membranes were blocked with TBS containing 5% milk and 0.1% Tween20 for 1 h and then incubated overnight with primary antibody at 4uC. Membranes were washed in TBS/T and incubated with secondary antibodies conjugated to horseradish peroxidase (Cell Signaling) for 1 h at RT. Blots were developed using the LumiGLOChemiluminescent substrate (Cell Signaling).

Immunofluorescence staining
293T cells were grown on 6-well plates and transfected with the indicated plasmids using LipofectAmine (Invitrogen). Twenty-four hours after transfection, the cells were washed three times with PBS and fixed in 3.7% formaldehyde for 20 min. Cells were permeabilized with PBS containing 0.1% Triton X-100 and 0.1 M glycine at RT, incubated for 15 min, washed three times with PBS, and blocked with 3% BSA in PBS for 10 min at RT. Cells were incubated with primary anti-Myc and anti-Flag for 1 h at 37uC, washed three times with PBS, and incubated for 1 h at 37uC with Cy5-conjugated anti-Rabbit and TRITC-conjugated anti-Mouse for Myc-BRAF V600E and Flag-MST1, respectively (Jackson ImmunoResearch Laboratories, West Grove, PA). Cells were observed with a CellomicsArrayScanV HCS reader (Thermo Scientific, Pittsburgh, PA).
In vitro kinase assay of MST1 MST1 was immunoprecipitated from cell lysates and the washed precipitates were incubated for 30 minutes at 30uC with 1 mg histone H2B (Roche) in 25 mL kinase assay buffer comprised of 40 mM HEPES-NaOH (pH 7.4), 20 mM MgCl 2 , 1 mM DTT, phosphatase inhibitor mixture, 10 mM unlabeled ATP, and 1 mCi [c-32P]ATP. The reaction was terminated by the addition of Laemmli sample buffer and phosphorylated proteins were detected by SDS-PAGE followed by autoradiography.

Apoptosis Assay
Image based apoptosis assays were carried out using the APO-BrdU TUNEL Assay Kit (Invitrogen). Cells was transfected with FoxO3, MST1, and/or BRAFV 600E , as indicated, and then stained with DAPI and terminal deoxynucleotidyltransferasemediated dUTP nick end labeling (TUNEL) reagents.The proportion of apoptotic cells was determined asthe percentage of TUNEL-positive cells among all DAPI-stainedcells.

Mouse Experiments and Histology
The strategy and methods used for the generation of thyroid specific transgenic BRAF V600E FVB/N mice with a bovine thyroglobulin promoter have been described previously [4]. A detailed explanation of how MST1 null mice are generated is described in a recent paper from our group [31]. The MST1-null mice appeared to have normal reproductive capacity when compared with wild-type mice. Tg-BRAF V600E mice were crossed with MST1-null mice to obtain Tg-BRAF V600E /MST1-null mice and Tg-BRAF V600E littermates. These mice were sacrificed at 16 weeks and their thyroids were removed after genotyping. Thyroid tissues were fixed in 10% formalin and embedded in paraffin. Five mM-thick sections were prepared and stained with H&E. Immunohistochemical analysis of paraffin embedded thyroid tissue was also performed using anti-cyclin D1 rabbit monoclonal antibody (92G2, Cell Signaling). ImageJ (http://rsbweb.nih.gov/ ij/) for microscopy was used to estimate areas without follicular architecture or colloid formation in thyroid tissues. Tg-BRAF V600E /MST1 null mice (n = 10) and Tg-BRAF V600E littermates (n = 10) were collected and three sections of each mouse were used for immunohistochemical and imageJ analysis. Differences in the number of undifferentiated foci and areas without follicular architecture or colloid formation between both groups were assessed by the Mann-Whitney U-test. All animal experiments were approved by the Institutional Animal Care and Use Committee of the Chungnam National University School of Medicine (Approval ID; 2009-3-25).

Statistical Analysis
Results were expressed as the means 6 standard deviations. Statistical differences between two groups were analyzed by the Mann-Whitney U test. Differences were considered to be significant at P,0.05.All statistical analyses were performed using SPSS Version 16.0 for Windows (SPSS Inc., Chicago, IL).

BRAF V600E suppresses p21 and p27 induction by RASSF1A in FRO cells
To confirm that RASSF1A induces FoxO3 transactivation, RASSF1A was transiently transfected into FRO cells harboring the BRAF mutated alleles and then p21 and p27, the known transcriptional targets of FoxO3, were assessed by immunoblotting [30]. As shown in Fig 1A, endogenous RASSF1A was not detected, which is compatible with previous reports showing hypermethylation of RASSF1A in undifferentiated thyroid cancer cell lines [17,32]. Fig. 1A shows the RASSF1A-induced dose dependent increase in p21 and p27, which suggests that RASSF1A might be linked with FoxO3 transactivation. To investigate the possible crosstalk between RASSF1A and BRAF V600E , the effect of BRAF V600E on RASSF1A-mediated p21 and p27 induction was evaluated. Interestingly, BRAF WT did not affect the induction of p21 and p27 by RASSF1A, whereas BRAF V600E effectively decreased p21 and p27 (Fig. 1B), suggesting that oncogenic BRAF V600E can regulate the RASSF1A-FoxO3 signaling pathway. As BRAF V600E can inhibit FoxO3 transactivation via the ERK or Akt/PKB signaling pathways, a MEK inhibitor (UO126) and PI3K inhibitors (Wortmannin, LY294002) were used to block these pathways and observe the effect of BRAF V600E on RASSF1A-induced FoxO3 transactivation. As shown in Fig. 1C, the BRAF V600E -mediated suppression of p21 and p27 induction by RASSF1A was not reversed by the MEK inhibitor U0126 or the PI3K inhibitors Wortmannin and LY294002, showing that BRAF V600E inhibits the RASSF1A-induced p21 and p27 upregulation independently from the ERK and PI3K-Akt/PKB signaling pathways. Recently, RASSF1A was shown to interact with MST1 via its SARAH domain and increase MST1 kinase activity [20], suggesting that RASSF1A-MST1-FoxO3 might be operational in response to oncogenic signals. To verify this hypothesis, MST1 expression was silenced using siRNA, and the RASSF1A induction of p21 and p27 was assessed. Fig. 1D shows that RASSF1A did not induce p21 and p27 expression in siMST1-treated cells, indicating that MST1 plays an important role in the p21 and p27 induction by RASSF1A.
Because the transcriptional activity of FoxO3 depends on its nuclear translocation, a FoxO3-GFP vector was constructed to confirm the data obtained with the 3XIRS reporter by immunofluorescence imaging. As shown in Fig. 2E, GFPfluorescence was predominantly observed in the cytoplasm in control samples; introduction of MST1 caused a shift in GFPfluorescence detection to the nucleus. When BRAF V600E was cotransfected with MST1, nuclear GFP fluorescence was barely detected. Interestingly, BRAF V600E was exclusively localized in the cytoplasm (blue), MST1 showed dominant cytoplasmic localization (red), and merged images revealed a combined color (lilac) that suggests the co-localization of BRAF V600E and MST1. To quantify FoxO3-fluorescence, the nuclear GFP-fluorescence signal intensity was analyzed using the BioApplication of CellomicsAr-rayScan, following the manufacturer's protocol. In agreement with the immunofluorescence images, MST1 increased intensity of the nuclear GFP-fluorescence signal, and BRAF V600E inhibited the nuclear detection of GFP-fluorescence (Fig. 2F). Furthermore, we performed fractionation assay to define the subcellular localization of FoxO3-GFP and observed that MST1 increased nuclear FoxO3-GFP and BRAF V600E could reverse this MST1-induced FoxO3 nuclear localization (Fig. S2).Taken together, these results indicate that BRAF V600E inhibits nuclear translocation of FoxO3 and suggest that this inhibition might be mediated by direct interaction between BRAF V600E and MST1. BRAF V600E represses FoxO3 transactivation via MST1 kinase-dependent pathway RAF-1 is an homologue of BRAF that interacts with MST2, suppressing MST2-mediated apoptosis [37,38], which may be important for the regulation of certain biological processes [39]. In our study, however, FoxO3 inhibition by BRAF V600E was not altered by RAF-1 silencing (Fig. S3). Furthermore, this inhibitory effect of BRAF V600E was not reversed by the inhibition of MEK or PI3-kinase (Fig. S4), confirming that BRAF V600E is able to suppress FoxO3 transactivation through a mechanism independent from MEK/ERK and PI3-kinase signaling.
The inhibition of RASSF1A or MST1-induced FoxO3 transactivation and nuclear translocation by BRAF V600E led to the hypothesis that BRAF V600E might regulate FoxO3 transactivation via a MST1-dependent pathway. To verify this hypothesis, reporter experiments were carried out in siMST1-treated cells. Remarkably, following knockdown of MST1, FoxO3 could not induce full activation of the 3XIRS reporter and BRAF V600E no longer had an effect on FoxO3 induced activation of 3XIRS reporter (Fig. 3A). These observations suggest that MST1 plays a role in FoxO3 transactivation and that BRAF V600E inhibits FoxO3 transactivation by suppressing MST1 activity. The reporter assay was also used to test whether there is signal-crosstalk between MST1 and BRAF V600E on FoxO3 transactivation. As expected, BRAF V600E inhibited MST1-induced 3XIRS reporter activation in a dose-dependent manner and MST1 abolished the inhibitory effect of BRAF V600E on FoxO3 transactivation (Fig. 3B & 3C). These reporter assays indicated that increased levels of MST1 could prevent BRAF V600E -induced FoxO3 inactivation. Taken together, these data suggest that the cross-regulation between BRAF V600E and MST1 determines the degree of FoxO3 transactivation.
BRAF V600E shows mutation-specific binding with the Cterminal of MST1 and inhibits MST1 kinase activity in vitro Because immunofluorescence images and reporter assays suggested that BRAF V600E and MST1 might be colocalized and have competitive effects on FoxO3 transactivation, immunoprecipitation (IP) assays were carried out to define the direct interaction of BRAF V600E and MST1. As shown in Fig. 4A, BRAF V600E was co-precipitated with Flag-tagged MST1 in lysates from 293T cells transfected with BRAF V600E and Flag-tagged MST1 and vice versa. To verify the region of MST1 important for the interaction with BRAF V600E , IP assays using MST1 deletion constructs, such as MST1 N-terminal (residues 1-326) and Cterminal (327-487) constructs (Fig. 4B), were carried out. These IP assays consistently demonstrated that BRAF V600E binds to the Cterminal region of MST1 (Fig. 4C), which contains of regulatory binding domain with the Sav/RASSF1/Hpo (SARAH) domain of RASSF1A [40,41,42]. To verify theBRAF V600E mutation specific binding with MST1, we performed additional IP assays using BRAF WT and BRAF V600E constructs. Furthermore, as shown in Fig. 4D, BRAF V600E was consistently co-precipitated with Flagtagged MST1 whereas wild-type BRAF did not. Taken together, these results show that BRAF V600E can bind to the C-terminal domain of MST1 via mutation-specific protein-protein interaction.
Because RASSF1A also binds to the C-terminal domain of MST1 and increases MST1 kinase activity, the interaction between BRAF V600E and MST1 might affect MST1 kinase activity by interrupting the heterotypic interaction of MST1 with RASSF1A. To verify this, the effect of BRAF V600E on the kinase activity of MST1 was examined in vitro using 293T cells, which were transfected with Flag-MST1 and BRAF V600E as indicated. The immunoprecipitates obtained with an antibody to Flag were used to assay for MST1 kinase activity using histone H2B as a substrate. As seen in Fig. 4E, the expression of BRAF V600E completely abolished MST1 kinase activity. These data suggest that BRAF V600E can negatively regulate MST1 kinase through direct interaction with its C-terminus.

BRAF V600E generates an anti-apoptotic effect via repression of p21 and p27 induction
FoxO3 acts as a cell-cycle regulator by inducing the transcription of genes such as p21 and p27 [43], the promoters of which contain IRS elements. The results of the assessment of the regulatory effect of BRAF V600E on the p27 reporter are shown in Fig. 5A. FoxO3 increased p27 reporter activity and MST1 potentiated this FoxO3-induced p27 activation. In addition, MST1-FoxO3 mediated p27 reporter activation was abrogated by co-transfection with BRAF V600E . To confirm that BRAF V600E can affect p27 reporter activity, the inhibitory effect of BRAF V600E on p21 and p27 induction was investigated by Western blot analysis (Fig. 5B). The results confirmed the MST1-FoxO3 induction of p21 and p27, which was almost completely reversed by BRAF V600E co-transfection. Notably, these two G1-checkpoint CDK inhibitors have been associated with the initiation and progression of human tumors because p21 and p27 bind and inhibit cyclin-CDK complexes, inducing cell cycle arrest and apoptosis. Therefore, the TUNEL assay was used to determine whether MST1-FoxO3 promotes apoptosis and if BRAF V600E reverses the MST1-FoxO3 effect (Fig. 5C). Transient coexpression of FoxO3 and MST1 increased the number of TUNEL positive 293T cells, but this number was decreased by the introduction of BRAF V600E (see also right panel of Fig. 5C for statistical analysis). These results imply that the MST1-FoxO3 pathway induction of G1-checkpoint CDK inhibitors promotes apoptosis, and oncogenic BRAF V600E inhibits the activity of these tumor suppressor systems through the regulation of MST1.

Loss of MST1 aggravates PTC phenotypes in thyroidspecific transgenic BRAF V600E mice
To validate the pathophysiological consequences of the interaction between BRAF V600E and MST1 in vivo, animal experiments using thyroid-specific transgenic BRAF V600E mice (Tg-BRAF V600E mice) and MST1 null mice were designed. Tg-BRAF V600E mice were crossed with MST1 null mice and the phenotypic differences between Tg-BRAF V600E /MST1 null mice and their Tg-BRAF V600E littermates were analyzed at 16 weeks [4]. The thyroid glands of MST1 null mice with wild-type BRAF were normal (data not shown). However, Tg-BRAF V600E /MST1 null mice showed aggressive PTC tumor phenotypes compared with their Tg-BRAF V600E littermates. Although the presence of lymph node metastasis was not observed in either of the two groups, Tg-BRAF V600E /MST1 null mice showed abundant foci of undifferentiated carcinomas and large areas without follicular architecture or colloid formation ( Figure 6A-H). For the statistical analysis of these results, immunohistochemical staining of cyclin D1 was performed, which showed an intensive cyclin D1 staining area as undifferentiated foci (Fig. 6B&C, F-G). As shown in Figure 6D, the number of undifferentiated foci in female Tg-BRAF V600E /MST1 null mice was 3.360.86, but that of their Tg-BRAF V600E littermates was 1.0560.51, reflecting a statistical difference between the two groups (P,0.001). Similarly, male Tg-BRAF V600E /MST1 null mice showed a higher number of undifferentiated foci than their Tg-BRAF V600E littermates (3.560.83and 1.3560.67, respectively; P,0.001). The total area composed of sheets of spindle cells without follicular architecture or colloid formation was estimated and ImageJ (http://rsbweb. nih.gov/ij/) was used to quantify these areas and analyze the differences between Tg-BRAF V600E /MST1 null mice and their Tg-BRAF V600E littermates ( Figure 6H). In female mice, Tg-BRAF V600E /MST1 null mice had a significantly larger area without follicular architecture or colloid formation than their Tg-BRAF V600E littermates (75.4763.83% and 52.6967.93%, respectively; P = 0.034). In male mice, Tg-BRAF V600E /MST1 null mice also showed a larger area without follicular architecture or colloid formation compared to their Tg-BRAF V600E littermates (70.0263.68% and47.3167.93%, respectively; P = 0.021). This in vivo study suggests that the direct interaction between BRAF V600E and MST1 could affect the behavior of papillary thyroid cancer. Another piece of evidence demonstrating the aggressive tumor phenotype in these Tg-BRAF V600E /MST1 null mice was the frequent presence of muscle invasion ( Figure 6I-L), which was not observed in their Tg-BRAF V600E littermates.

Discussion
BRAF V600E constitutively activates the MEK-ERK signaling pathway, promoting ERK-dependent transcriptional activity, which regulates cell growth, cell proliferation, apoptosis evasion, replicative potential, angiogenesis, and tissue invasion [44,45]. RASSF1A primarily interacts with small GTPases such as Ras to facilitate cell cycle arrest and apoptosis. Transcriptional silencing of RASSF1A by promoter methylation has been reported in many human cancers including thyroid cancer [46,47]. However, epigenetic alteration has been regarded as a slow and time-consuming process for gene silencing, implying that tumor suppressors such as RASSF1A are inhibited to allow the initiation and maintenance of oncogenic processes necessary for the establishment of clinically significant thyroid cancers [48]. The results of the present study suggest that BRAF V600E tumors might  /well), and Flag-MST1-C (1 mg/well) was performed as indicated. Twenty-four hours after transfection, cells were lysed and the lysates were used for immunoprecipitation analyses with the indicated primary antibodies. The immunoprecipitates were separated by SDS-PAGE and subjected to immunoblotting with anti-Flag or anti-BRAF antibodies. (e) Co-transfection with BRAF V600E (1 mg/well) and Flag-MST1 (indicated amount/well) was performed. Twenty-four hours after transfection, immunoprecipitation of MST1 from cell lysates was performed and the washed precipitates were incubated for 30 minutes at 30uC with 1 mg histone H2B in 25 mL kinase assay buffer. The reaction was terminated by the addition of Laemmli sample buffer and phosphorylated proteins were detected by SDS-PAGE followed by autoradiography. doi:10.1371/journal.pone.0016180.g004 regulate RASSF1A function via a novel mechanism. RASSF1A transfection induced p21 and p27 in a MST1-dependent manner, suggesting that the RASSF1A-MST1-FoxO3 pathway can induce cell cycle arrest. In addition, BRAF V600E suppressed the RASSF1A-MST1-mediated signal, inhibiting p21 and p27 induction, supporting the hypothesis of the present work. It has been suggested that BRAF V600E -activated ERK inhibits FoxO3 by murine double minute 2 (MDM2)-mediated degradation [25].
However, BRAF V600E did not alter HA-FoxO3 protein levels, as shown by Western blot analysis, and, interestingly, treatment with a MEK inhibitor did not abolish the BRAF V600E -mediated inactivation of FoxO3. In the present study, BRAF V600E did not induce phosphorylation of Akt/PKB and, furthermore, the addition of PI3 kinase inhibitors did not affect the inhibition of FoxO3 by BRAF V600E . Therefore, BRAF V600E did not affect FoxO3 transactivation by a PI3-Akt/PKB-dependent signaling pathway, whereas RET/PTC inactivates FoxO3 by Akt/PKB activation [23]. In addition, the silencing of RAF-1, which has been reported to bind with MST2, did not alter the inhibition of FoxO3 transactivation by BRAF V600E .Taken together,these data suggest that FoxO3 inactivation might be a common process in the carcinogenesis of PTC and that BRAF V600E inhibits FoxO3 transactivation through a MEK-ERK-, Akt/PKB-, or RAF-1independent pathway. The results of the present study suggest the existence of a novel FoxO3 inhibitory mechanism that might operate in BRAF V600E -induced PTC.
MST1 kinase had a counter-regulatory effect on FoxO3 activation and localization compared with Akt/PKB [49,50]. In contrast to Akt/PKB, MST1 phosphorylation of FoxO3 promotes its nuclear translocation, thereby inducing cell death [19]. A novel inhibitory mechanism of FoxO3 transactivation might be opera-tional in BRAF V600E -induced PTC, which led to the evaluation of BRAF V600E regulation of MST1-FoxO3 signaling pathways. A reporter assay revealed that MST1-induced FoxO3 transactivation was completely inhibited by co-transfection with BRAF V600E . In addition, over-expression of MST1 prevented BRAF V600E -induced FoxO3 inactivation. These data suggested that cross-regulation between MST1 and BRAF V600E might play a critical role in regulating FoxO3 transactivation. Supporting this idea, immunofluorescence studies revealed not only the BRAF V600E -mediated inhibition of MST1-induced FoxO3 nuclear translocation, but also a direct interaction between MST1 and BRAF V600E , which was confirmed through IP assays that demonstrated that BRAF V600E binds to the C-terminal dimerization domain of MST1.
In fact, the dimerization domain of MST1/2 mediates MST1 kinase activity by interacting with binding partners such as RASSF1A and hWW45 [20,51]. Because RASSF1A promotes MST1 activation and apoptosis via this hetero-dimerization, we postulated that the direct interaction of MST1 with BRAF V600E might affect MST1 kinase activity. Actually, our in vitro kinase assay clearly showed that MST1 kinase activity on histone H2B was markedly decreased by BRAF V600E . Another possible mechanism of MST1 regulation may be the inhibitory phosphorylation of MST1 by BRAF V600E . Prior research has shown that Akt/PKB can phosphorylate the Thr 387 residue of MST1 and prevent its proteolytic activity, providing a mechanism for the Akt/PKB-mediated inactivation of MST1-FoxO3 pathways [50]. Although the BRAF V600E -induced MST1 phosphorylation was not assessed, it remains possible that BRAF V600E may phosphorylate and inactivate MST1 through direct interaction with the Cterminal domain of MST1.
Few mouse models exist for the study of tumor suppressor systems in PTC. In the present study, a new in vivo model was generated to investigate the cross-regulation between BRAF V600E and MST1, and the results obtained with transgenic mice were consistent with the in vitro generated data. Although MST1 knockout mice did not show any thyroid gland pathological changes compared to wild-type mice, Tg-BRAF V600E /MST1 null mice showed an aggressive phenotype, supporting the hypothesis that MST1 might have a modifying effect on BRAF V600E -induced carcinogenesis. Although the lack of an appropriate antibody for the specific immunohistochemical detection of serine 207 residue in FOXO3 phosphorylated by MST1 kinaseprevented the detection of FoxO3 activity modified by MST1, these results, coupled with the results of in vitro assays, suggest that BRAF V600E may alleviate the tumor suppressor function of RASSF1A-MST1 and vice versa.
In conclusion, the present study suggests that a direct interaction between oncogenic BRAF V600E and MST1 kinase plays a crucial role in determining tumor behavior in PTC. This novel interaction might provide advanced insight into the pathogenesis of the disease, and form the basis for novel drug development for the treatment of BRAF V600E -induced PTC. In light of this, it will be important to delineate the precise regulatory mechanism of the RASSF1A-MST1 pathway and verify the action of BRAF V600E on this signal pathway in human PTC. Figure S1 MST1 significantly increases 3XIRS reporter activities. (A) 293T cells were cultured in 12 well dishes until they reached 80% confluence, and co-transfected with 3XIRS Luc (100 ng/well), FoxO3 (0.5 mg/well), and MST1 (0.5 mg/well) as indicated for 24 hours. To verify transfection efficacy, total lysates were immunoblotted with anti-HA, anti-Flag, and anti-Actin antibodies. (B) Cells were co-transfected with 3XIRS Luc (100 ng/ well), FoxO3 (0.5 mg/well), and MST1 (indicated amount/well). Total lysates were immunoblotted with anti-HA, anti-Flag, and anti-Actin antibodies. For each sample, firefly luciferase activity was normalized to Renilla luciferase activity and expressed as relative-fold change compared to basal luciferase activity. All data are presented as mean6SD: (*) P,0.01 between two groups. (TIF) Figure S2 BRAF V600E inhibits MST1-induced nuclear translocation of FoxO1. 293T cells were cultured in a six-well dish until they reached 80% confluence and co-transfected with FoxO3-GFP (0.5 mg/well), Flag-MST1 (0.5 mg/well), and Myc-BRAF V600E (0.5 mg/well) as indicated. Twenty-four hours after transfection, the cells were prepared for subcellular fractionation using the Nuclear/Cytosol Fractionation kit (BioVision, Inc. CA). The markers, origin recognition complex subunit 1 (ORC1) and b-tubulin, were used to verify the identity and purity of the nuclear and cytosolic fractions, respectively. Based on these markers, a good overall yield was obtained without mixing of the fractions. (TIF) Figure S3 BRAF V600E mediated FoxO3 inhibition was not altered by RAF-1. 293T cells were cultured in 12 well dishes until they reached 80% confluence and co-transfected with 3XIRS Luc (100 ng/well), FoxO3 (0.5 mg/well), BRAF V600E (0.5 mg/well), and SiRAF-1 (20 mM/well Stealth TM RNA) for 24 h as indicated. Total lysates were immunoblotted with anti-HA, anti-BRAF, anti-RAF-1, and anti-Actin antibodies. For each sample, firefly luciferase activity was normalized to Renilla luciferase activity and expressed as relative-fold change compared to basal luciferase activity. All data are presented as mean6SD: (*) P,0.01 between two groups. (TIF) Figure S4 BRAF V600E suppresses FoxO3 transactivation via a MEK/ERK-, PI3 kinase-independent pathway. 293T cells were cultured in 12 well dishes until they reached 80% confluence, and co-transfected with 3XIRS Luc (100 ng/well), FoxO3 (0.5 mg/ well), and BRAF V600E (0.5 mg/well) as indicated for 24 h. MEK inhibitor (lane 3, U0126 20 mM/well) and PI3 kinase inhibitors (lane 4, Wortmannin 200 nM/well, and lane 5, LY294002 20 mM/well) were added. Total lysates were immunoblotted with anti-HA, anti-BRAF, anti-pERK, anti-ERK, anti-pAkt/PKB, anti-Akt/PKB, and anti-Actin antibodies. For each sample, firefly luciferase activity was normalized to Renilla luciferase activity and expressed as relative fold change compared to basal luciferase activity. All data are presented as mean6SD. Abbreviations: U0, U0126; WT, Wortmannin; LY, LY294002; and Con, control. (TIF)