The RET receptor tyrosine kinase has essential roles in cell survival, differentiation, and proliferation. Oncogenic activation of RET causes the cancer syndrome multiple endocrine neoplasia type 2 (MEN 2) and is a frequent event in sporadic thyroid carcinomas. However, the molecular mechanisms underlying RET's potent transforming and mitogenic signals are still not clear. Here, we show that nuclear localization of β-catenin is frequent in both thyroid tumors and their metastases from MEN 2 patients, suggesting a novel mechanism of RET-mediated function through the β-catenin signaling pathway. We show that RET binds to, and tyrosine phosphorylates, β-catenin and show that the interaction between RET and β-catenin can be direct and independent of cytoplasmic kinases, such as SRC. As a result of RET-mediated tyrosine phosphorylation, β-catenin escapes cytosolic down-regulation by the adenomatous polyposis coli/Axin/glycogen synthase kinase-3 complex and accumulates in the nucleus, where it can stimulate β-catenin–specific transcriptional programs in a RET-dependent fashion. We show that down-regulation of β-catenin activity decreases RET-mediated cell proliferation, colony formation, and tumor growth in nude mice. Together, our data show that a β-catenin–RET kinase pathway is a critical contributor to the development and metastasis of human thyroid carcinoma. [Cancer Res 2008;68(5):1338–46]
- tumor metastasis
- thyroid tumors
- MEN 2
The RET receptor is required for development of urogenital and neural crest–derived cell types ( 1). Under normal cellular conditions, RET is activated by binding of both glial cell line–derived neurotrophic factor (GDNF) ligands and cell surface bound coreceptors of the GDNF family receptor α (GFRα) proteins ( 2). However, oncogenic activation of RET, by germline point mutations, leads to ligand-independent constitutive kinase activity, giving rise to the inherited cancer syndrome multiple endocrine neoplasia type 2 (MEN 2). MEN 2 is characterized by medullary thyroid carcinoma (MTC), a tumor of thyroid C cells, and the adrenal tumor pheochromocytoma, as well as other less common tumor and developmental phenotypes (reviewed in ref. 3). MTC is the predominant disease feature, with early onset tumors and metastases to lymph nodes and distant organs ( 4). RET activation contributes to stimulation of RAS-ERK, c-Jun-NH2-kinase, phosphoinositide 3-kinase, p38 mitogen-activated protein kinase (MAPK), SRC, STAT and ERK5 signaling cascades (reviewed in ref. 1). However, the identity of the critical secondary oncogenic signals involved in the broad and early metastatic pattern of RET-mediated MTC is still largely unknown.
β-Catenin is an ubiquitously expressed multifunctional protein that plays important roles in cell adhesion and signal transduction ( 5). At the plasma membrane, β-catenin associates with E-cadherin and α-catenin in linking the cytoskeleton and adherens junctions, whereas in the nucleus, it acts as a mediator of transcription through other DNA-binding proteins, such as TCF/LEF family members (reviewed in ref. 6). Cytosolic-free β-catenin interacts with the adenomatous polyposis coli (APC) and axin proteins to form a complex, which in turn recruits glycogen synthase kinase-3 (GSK3) and casein kinase, to form a destruction complex that serine/threonine phosphorylates β-catenin and targets it to the proteosome (reviewed in ref. 7).
Abnormal expression or localization of β-catenin has been reported in many tumor types ( 5, 8), and β-catenin–mediated loss of cell-cell adhesion has been implicated in anchorage-independent cell growth and cancer metastasis ( 8). The best-characterized mechanism leading to β-catenin–mediated signal transduction is through activation of the WNT pathway by binding of WNT proteins to frizzled or LRP family cell surface receptors ( 9). However, β-catenin signaling can also be induced in response to overexpression or activation of tyrosine kinases in a WNT-independent fashion, and both these pathways converge to target β-catenin to the nucleus and induce expression of a similar set of β-catenin–specific target genes ( 10– 12). β-Catenin tyrosine phosphorylation causes its dissociation from membrane-associated E-cadherin, leading to accumulation of a pool of free cytoplasmic β-catenin ( 13). This can, in turn, increase the amount of β-catenin reaching the nucleus, where it acts as a transcription factor, up-regulating expression of genes involved in cell migration, growth, differentiation, and survival ( 8). Tyrosine phosphorylation of β-catenin, followed by functional down-regulation of E-cadherin–mediated cell-cell contact, is potentially critical in initiating cell migration in both normal physiologic processes and in tumor metastasis ( 13).
Loss of membrane-associated β-catenin, often with an accompanying relative increase in cytosolic or nuclear expression, has been noted in anaplastic and poorly differentiated thyroid carcinomas and in thyroid papillary microcarcinoma ( 14– 16). However, β-catenin had not been previously investigated in RET-mediated tumor development and metastasis. Here, we report that RET interacts with and activates β-catenin and that a RET–β-catenin signaling pathway plays roles in RET-mediated tumor growth, invasion, and metastasis.
Materials and Methods
Expression constructs. Expression constructs for full-length human RET, GFRα1, and mutant RET constructs have been previously described ( 17– 19). Intracellular RET (icRET) expression constructs, generated by fusing cDNA encoding a myristylation signal, two dimerization domains, and the intracellular portion (amino acid 658 to the COOH terminus) of RET, have been previously described ( 18). Axin and wild-type (WT) β-catenin expression constructs were a gift from Dr. J. Woodgett (Ontario Cancer Research Institute). Glutathione S-transferase (GST)–tagged expression constructs for WT and mutant β-catenin have been reported ( 20, 21).
Cell culture and transfection. The human neuroblastoma cell line SH-SY-5Y was maintained in RPMI 1640 supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich). All other cell lines were grown in DMEM (Invitrogen), supplemented with 10% FBS. Medium for HEK293-TET-ON cells, used to induce icRET, was further supplemented with 1 μg/mL doxycycline. HEK293 cells were transiently transfected with the indicated expression constructs using Lipofectamine 2000 (Invitrogen), according to the manufacturer's instructions. RET activation was induced by addition of 100 ng/mL of GDNF (Promega) for full-length RET or with 1 μmol/L AP20187 dimerizer (ARIAD Pharmaceuticals) for icRET for the time periods indicated.
Immunoprecipitation, Western, and Far-Western blotting. Whole-cell lysates were harvested 48 h after transfection and suspended in lysing buffer [20 mmol/L Tris-HCl (pH 7.8), 150 mmol/L NaCl, 1 mmol/L sodium orthovanadate, 1% Igepal, 2 mmol/L EDTA, 1 mmol/L phenylmethylsulfonyl fluoride (PMSF), 10 μg/mL aprotonin, and 10 μg/mL leupeptin; ref. 19]. Protein concentration was determined by bicinchoninic acid protein assay kit according to the manufacturer's instructions (Pierce Biotechnology). For immunoprecipitations, lysates were incubated with a 1:50 dilution of the appropriate primary antibody for 2 h at 4°C with agitation. Complexes were collected on protein AG beads (Santa Cruz Biotechnology) by centrifugation at 13,000 rpm, washed thrice with lysing buffer, and resuspended in Laemmli buffer. Protein samples were denatured at 99°C for 5 min, separated on 10% SDS-PAGE gels, and transferred to nitrocellulose membranes (Bio-Rad), as previously described ( 17, 19). Antibodies used included anti-RET (c19), anti-ubiquitin (N19), and anti–myc-tag (NE10) antibodies (Santa Cruz Biotechnology), anti–phosphorylated RET (pY905) antibody (Cell Signaling), anti–β-catenin antibody (BD Biosciences), and anti-V5 tag (axin; Invitrogen). An anti–phosphorylated tyrosine antibody (pY99; Santa Cruz) was used to detect tyrosine phosphorylation of β-catenin. For Far-Western analyses, protein lysates were immunoprecipitated for RET, resolved, and Western blotted, as above. Membranes were incubated for 2 h at 4°C in a 0.1% Tween 20/TBS solution containing a probe of 1 μg/mL GST–β-catenin, GST alone, or no probe. After three washes, bound proteins were detected with anti-GST antibody (Santa Cruz Biotechnology) or immunoblotted for RET.
Preparation of cytosolic extracts. Cells were harvested by gentle centrifugation, washed twice with PBS, and suspended in ice-cold hypotonic buffer [20 mmol/L HEPES-KOH (pH 7.0), 10 mmol/L KCl, 1.5 mmol/L MgCl2, 1 mmol/L sodium EDTA, 1 mmol/L sodium EGTA, 1 mmol/L DTT, 250 mmol/L sucrose, and protease inhibitors; ref. 22]. After incubation on ice for 15 min, cells were disrupted by dounce homogenization. Nuclei were pelleted by centrifugation, and cytosolic fractions were isolated by collection of the supernatant.
GST pull-down assays. GST fusion proteins expressed in Escherichia coli were eluted with 100 mmol/L glutathione elution buffer using a polyprep column (Bio-Rad), as described previously ( 17). For GST pull-down assays, 5 μg of GST-fusion protein and GST-sepharose-beads (GE Healthcare/Amersham Biosciences) were incubated with whole-cell lysates at 4°C for 3 h with agitation. Bound proteins were eluted by boiling in Laemmli sample buffer containing 2-mercaptoethanol and resolved by SDS-PAGE and Western blotting, as described above.
Reporter assay. For dual luciferase reporter assays, TOPFLASH or FOPFLASH vectors (Upstate Biotechnology) and pRL-TK control were cotransfected into HEK293 cells stably expressing icRET or empty vector. Luciferase activity was measured with a dual luciferase reporter kit (Promega).
Short hairpin RNA production. Four different β-catenin short hairpin RNAs (shRNA) in pLKO.1 lentiviral vectors were obtained from Open Biosystems. Lentiviral particles containing the different β-catenin shRNA constructs were grown in 293T packaging cells by transfecting a three-plasmid packaging system ( 23) according to the manufacturer's instructions. Supernatants were collected 48 and 72 h after transfection, filtered, and pooled. NIH 3T3 stably expressing the oncogenic 2A-RET (C634R) were infected with lentiviral constructs or a heterologous lentiviral control, and polyclonal stable cell lines were generated.
Cell proliferation assay. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) cell viability assays were performed as described ( 24). Briefly, HEK293 cells transiently expressing RET or empty vector, seeded in six-well plates, were either infected with lentiviral β-catenin shRNA or untreated. Cells were grown in medium supplemented with 100 ng/mL GDNF. After 3 days, MTT was added to a final concentration of 250 μg/mL for 2 h at 37°C. Culture medium was removed, and formazan crystals were dissolved in DMSO. Reduced MTT was measured spectrophotometrically at 570 nm. Statistical significance was calculated by one-way ANOVA.
Apoptosis assays. NIH 3T3 cells stably expressing 2A-RET were infected with lentiviral constructs for β-catenin shRNA or control vector and cultured for 48 h. Cells were harvested, fixed in absolute ethanol, and treated with RNaseA at 37°C overnight. Fixed cells were incubated in 5 mg/mL propidium iodide for 15 min at room temperature, and cell cycle analysis was performed using an EPICS ALTRA HSS flow cytometer (Beckman Coulter). Statistical significance was calculated by one-way ANOVA.
Soft agar colony formation assays. Soft agar colony formation assays were performed as described previously ( 17). Briefly, NIH 3T3 cells expressing 2A-RET or K758M constructs were infected with lentiviral vectors for β-catenin shRNA or control. Approximately, 5 × 104 cells were resuspended in 0.2% top agar in culture medium and plated on 0.4% bottom agar in medium. Culture medium was supplemented every 2 to 3 days. Colonies were counted after 14 days, and statistical significance was confirmed by one-way ANOVA.
Confocal microscopy. SH-SY-5Y neuroblastoma cells were cultured on glass coverslips coated with 0.2% gelatin. Twenty-four hours before fixation, 10 nmol/L retinoic acid was added to the culture medium. Cells were serum starved for 3 h, then fixed in 3% paraformaldehyde for 40 min at room temperature. Cells were then washed in PBS, permeablized with 0.15% Triton-X-100, and blocked for 30 min in 3% bovine serum albumin. Cells were double immunostained with primary antibodies specific for β-catenin (BD Biosciences) and RET (c19, Santa Cruz Biotechnology) and species-matched secondary antibodies labeled with Alexa 594 or 488, respectively. Coverslips were mounted on glass slides in Mowiol mounting medium and observed using a Leica TCS-SP2 confocal microscope. Individual channels were overlaid using Image J software, and colocalization was determined using the Image J RG2B colocalization plug-in.
Tumorigenicity in nude mice. All in vivo experiments were performed using 6-week-old to 8-week-old athymic nude mice (NIH). Experiments were performed in duplicate using a minimum of five animals per treatment group. Mice were maintained in laminar flow rooms with constant temperature and humidity. Experimental protocols were approved by the Ethics Committee for Animal Care (Queen's University). NIH 3T3 parental cells, cells stably expressing 2A-RET constructs, or polyclonal cell lines expressing 2A-RET and β-catenin shRNA (described above) were inoculated s.c. into the right flank of the mice. Cells (2 × 106 in suspension) were injected on day 0, and tumor growth was followed every 2 to 3 days by tumor diameter measurements using vernier calipers. Tumor volumes (V) were calculated using the formula: V = AB2/2 (A, axial diameter; B, rotational diameter). Mice were sacrificed at day 14, and tumor tissues were excised for protein extraction and histology. Tumor tissue was homogenized in 10 volumes of homogenization buffer [20 mmol/L Tris-HCl (pH 7.8), 150 mmol/L NaCl, 1 mmol/L sodium orthovanadate, 2 mmol/L EDTA, 1 mmol/L PMSF, 10 μg/mL aprotonin, and 10 μg/mL leupeptin]. Tissue debris was removed by centrifugation, the supernatant was collected, and cells were lysed at a final concentration of 1% Triton X-100. Samples were centrifuged to remove insoluble material. Tissue for histology was fixed in neutral buffered formalin and processed by routine methods. Paraffin-embedded sections of 5 μm were stained with H&E for histologic examination.
β-Catenin immunohistochemistry. Immunohistochemistry was performed on formalin-fixed paraffin-embedded tissues. Tumor tissue from the CALC-MEN2BRET mice was excised, fixed, and paraffin-embedded by routine methods. Human MTC samples were obtained from the Department of Pathology of the University Medical Center Utrecht. Paraffin sections (6 μm) were blocked with 0.5% casein blocking reagent (PerkinElmer Life Sciences) in 0.1% Triton-X100 in PBS and treated with 1.5% hydrogen peroxide to inhibit endogenous peroxidase activity. Sections were incubated with a mouse monoclonal anti–β-catenin antibody (BD Biosciences), 1:50 dilution, for 1 h at room temperature. Slides were incubated with peroxidase-labeled rabbit anti-mouse secondary antibody (Dako), 1:100 dilution, for 30 min at room temperature and subsequently with peroxidase-labeled swine anti-rabbit antibody (Dako), 1:100 dilution, for 30 min at room temperature. Finally, sections were incubated with 3,3′-diaminobenzidine and counterstained with Mayer's hematoxylin. Nuclear staining was considered positive when one or more positive nuclei were observed in each microscopic field (40×; ref. 25). Negative control experiments were performed by omitting the primary antibody.
Relative quantification by real-time reverse transcription–PCR. Relative differences between gene transcript levels were confirmed using the QuantiTect SYBR green reverse transcription–PCR (RT-PCR) kit (Qiagen) and a SmartCycler II (Cepheid). Primer and PCR product information are found in Supplementary Table S1. The quantitative RT-PCR (qRT-PCR) assays were repeated at least thrice. Crossing threshold (Ct) values were taken at the same threshold for each experiment; fold expressions were averaged, and mean fold change, relative to the empty vector control, was calculated.
Results and Discussion
β-Catenin nuclear localization in RET-mediated human thyroid tumors. In preliminary immunohistochemical experiments, using MTCs from CALC-MEN2BRET transgenic mice, which express a constitutively active, oncogenic RET mutant (2B-RET; ref. 26), we found that six of seven tumors had nuclear localization of β-catenin (Supplementary Fig. S1), suggesting a role for β-catenin signaling in these tumors and a relationship between RET activation and β-catenin nuclear localization. Furthermore, in human MTC samples from 20 MEN 2 patients with known oncogenic RET mutations (Supplementary Table S2), eight had nuclear β-catenin expression in a subset of cells, heterogeneously spread throughout the tumors ( Fig. 1A ). In addition, although some MTCs showed β-catenin expression at the cellular membrane, it was less prominent, particularly in tumors with strong nuclear β-catenin expression ( Fig. 1A). Association of nuclear staining with loss of membranous staining has also been reported in other cancers, such as colorectal carcinomas ( 25). Nuclear localization of β-catenin was not detected in normal or hyperplastic C cells in mouse or human tissues. Interestingly, nuclear localization was more prevalent in metastases (five of seven cases) than in primary MTCs (3 of 13 cases; Fig. 1B and Supplementary Table S2), suggesting an association of more aggressive or advanced MTC disease stage and activation of β-catenin. Conversely, in the absence of oncogenic activation of RET, we showed that the endogenous RET and β-catenin proteins colocalized at the plasma membrane in the neuroblastoma cell line SH-SY-5Y ( Fig. 1C). Together, these observations indicated that β-catenin signaling may play an important role in progression of RET-mediated tumors and suggested that a novel RET–β-catenin signaling mechanism could be taking place.
RET associates with and tyrosine phosphorylates β-catenin. As tyrosine phosphorylation of β-catenin by several kinases has been correlated with tumorigenesis and metastasis ( 11– 13), we investigated the association of RET with β-catenin. In TT cells, a line derived from a human MTC expressing endogenous RET with an activating mutation (2A-RET), we found that RET induced phosphorylation of endogenous β-catenin. Similarly, in cells cotransfected with RET and β-catenin expression constructs, we showed that treatment with the RET ligand GDNF also induced phosphorylation of β-catenin ( Fig. 2A ). Immunoprecipitation of β-catenin and immunoblotting with appropriate antibodies showed that both endogenous and transiently expressed β-catenin and RET associate in complexes ( Fig. 2A). These complexes could also be detected by immunoprecipitation of RET and immunoblotting for β-catenin (not shown). Our data show that both the ligand-activated WT RET and the oncogenic mutants 2A-RET and 2B-RET induce tyrosine-phosphorylation of β-catenin ( Fig. 2A). In the absence of GDNF, the constitutively active 2B-RET protein, but not WT RET, was able to induce significant β-catenin tyrosine phosphorylation ( Fig. 2A). Furthermore, a catalytically compromised RET mutant, K758M ( 17), was unable to induce β-catenin tyrosine phosphorylation, even in the presence of GDNF stimulation, suggesting a RET kinase–dependent mechanism of β-catenin phosphorylation ( Fig. 2A). Interestingly, however, K758M RET was still able to associate with β-catenin in a phosphorylation-independent fashion, suggesting a constitutive association between RET and β-catenin.
Although β-catenin may associate with receptor kinases and can be tyrosine phosphorylated in cells stimulated by their ligands, epidermal growth factor receptor (EGFR) is the only receptor known to directly phosphorylate β-catenin ( 27). Other receptors may induce activation of SRC, which in turn can phosphorylate β-catenin, as well as other proteins of the adherens junctions ( 13). To determine whether β-catenin tyrosine phosphorylation could result from a direct interaction with RET, we performed in vitro kinase assays using purified RET kinase ( 17) and purified recombinant GST-tagged β-catenin ( 20). We found that the purified RET kinase can be activated in the presence of ATP and can directly phosphorylate β-catenin in vitro (Supplementary Fig. S2). As endogenous WT RET and β-catenin proteins colocalize at the plasma membrane ( Fig. 1C), we confirmed that they could directly interact by Far-Western assays. We showed that purified recombinant GST-β-catenin could interact with immunoprecipitated WT RET, whereas a purified GST control could not in direct Far-Western assays ( Fig. 2B). In vivo, we further showed that the RET–β-catenin interaction can occur independent of SRC, using coimmunoprecipitations in SYF−/− cells, which lack the SRC family kinases SRC, YES, and FYN ( Fig. 2C). Our data show that RET can associate with and directly phosphorylate β-catenin and that the tyrosine phosphorylation of β-catenin is not dependent on SRC activation.
The pattern of tyrosine residue phosphorylation of β-catenin has been shown to be kinase specific. Activation of receptor tyrosine kinases FYN, FER, and MET leads to phosphorylation at Y142, whereas Y654 can be phosphorylated directly by EGFR or SRC ( 11, 12, 20, 27, 28). In pull-down assays using GST-tagged β-catenin tyrosine mutants for either Y142 (Y142F) or Y654 (Y654F; ref. 21), we showed that WT β-catenin and, to a lesser extent, the Y142F mutant were phosphorylated by WT RET but that Y654F β-catenin was not phosphorylated ( Fig. 2D), suggesting that Y654 was the major site for RET-mediated β-catenin phosphorylation. Tyrosine 654 lies within the putative E-cadherin binding region of β-catenin, and the bulkier phosphorylated residue has been shown to reduce β-catenin binding affinity for E-cadherin, leading to dissociation of β-catenin from the membrane ( 13, 28).
RET activation induces cytosolic translocation of β-catenin and its escape from the axin regulatory complex. Accumulation of cytosolic-free β-catenin is tightly regulated by the APC/axin/GSK3 complex, which binds β-catenin and leads to its serine/threonine phosphorylation, targeting it for ubiquitination, and proteosomal degradation ( 29). We have shown that RET and β-catenin colocalize at the plasma membrane ( Fig. 1C), but that activation of RET leads to a relative increase in β-catenin in the cytosol ( Fig. 3A ). Notably, the cytosolic level was not further enhanced by the more active 2B-RET kinase, perhaps suggesting that increased RET activity may enhance nuclear β-catenin levels preferentially. Interestingly, in the presence of active RET, we found a relative decrease in β-catenin associated with the APC/axin/GSK3 complex in immunoprecipitations of axin, which interacts directly with β-catenin in this complex (Supplementary Fig. S3A). Furthermore, there was an accompanying relative decrease in ubiquitinated β-catenin (Supplementary Fig. S3B), suggesting that RET-mediated tyrosine phosphorylation of β-catenin protects it from degradation. This would be consistent with data on RON and MET, which also suggest that tyrosine phosphorylation of β-catenin inhibits its interactions with the APC/axin/GSK3 complex ( 11).
RET-induced β-catenin tyrosine phosphorylation is associated with increased TCF transcriptional activity. β-Catenin also has important signaling roles in the nucleus, mediated through its interactions with the TCF family of transcription factors and other transcriptional regulators and the subsequent activation of target genes, such as cyclin D1 ( 8, 11, 12, 27). Initially, to determine whether RET activation could induce a β-catenin transcriptional program, we used the well-characterized TOPFLASH (TCF-binding site) and FOPFLASH (mutated TCF-binding site) luciferase reporters ( 11, 30) for detection of β-catenin/TCF-mediated transcription. In cells stimulated with GDNF, TCF reporter activity was significantly increased (relative to FOPFLASH) in the presence of WT RET but not in the presence of the kinase-dead mutant ( Fig. 3B), suggesting that a RET-dependent induction of a β-catenin/TCF transcriptional program may occur.
β-Catenin is required for RET-induced cell proliferation and transformation. The role of RET activation in stimulation of β-catenin–mediated transcription prompted us to investigate whether β-catenin was required for known RET-mediated processes using shRNA-knock-down of β-catenin. We evaluated four lentivirus-produced β-catenin shRNA constructs, pooled those producing the most significant knockdown of β-catenin and its target, cyclin D1 (Supplementary Fig. S4), and used these to generate polyclonal β-catenin–deficient cell lines. We used real-time qRT-PCR to evaluate the effect of loss of β-catenin on expression of a panel of known RET target genes, previously identified in gene expression analyses ( 19, 31, 32). RET activation has been shown to increase expression of cyclin D1, and our data show that both RET and β-catenin are required to induce this in NIH 3T3 cells ( Fig. 4A ; Supplementary Fig. S4). Knock-down of β-catenin expression had no significant effect on RET expression or on expression of control transcripts (e.g., axin) not known to be modulated by either RET or β-catenin ( Fig. 4A). However, it did block, or significantly reduce, RET-induced up-regulation of a subset of RET targets, including cyclin D1, JunB, and EGR1, in cells expressing both activated RET and shRNA for β-catenin ( Fig. 4A). Other RET targets were unaffected (e.g., GDF15, VGF), consistent with a β-catenin–independent mechanism of RET-mediated stimulation. Interestingly, a RET–β-catenin signaling pathway seemed to be important to induction of cell proliferative targets, including immediate early genes cyclin D1, EGR1, and JunB, but had less effect on targets associated with more differentiated cell functions, such as GDF15 and VGF (VGF nerve growth factor induced). Together, our data show that RET activation may stimulate a β-catenin transcriptional program but that not all RET targets require β-catenin for expression.
Knockdown of β-catenin expression reduces RET-mediated tumor growth and invasiveness in nude mice. As tyrosine phosphorylation of β-catenin by other kinases has been shown to divert its function from cell adhesion to increased signaling roles ( 13) and as we found a correlation of β-catenin nuclear expression and metastatasis of MTC, our gene expression data led us to postulate that RET-induced activation of the β-catenin pathway may contribute to cell proliferation and tumorigenesis. Using the MTT cell proliferation assay, we showed that GDNF-treatment of HEK293 cells significantly increased cell proliferation in the presence of WT RET but did not affect cells expressing the kinase dead K758M RET mutant or an empty vector ( Fig. 4B). Similarly, GDNF did not stimulate RET-mediated proliferation in the presence of β-catenin shRNA ( Fig. 4B). A significant decrease in colony formation induced by the oncogenic mutant 2A-RET was also detected in soft agar anchorage-independent growth assays in the presence of β-catenin shRNA ( Fig. 4C). Whereas β-catenin shRNA expression did increase apoptosis, the level of apoptotic cell death remained quite low in RET-expressing cells in the presence of β-catenin shRNA ( Fig. 4D), suggesting that a RET–β-catenin signaling pathway has roles in cell proliferation but is not as critical for cell survival.
We next used our β-catenin–shRNA knockdown model to investigate the role of β-catenin in RET-mediated oncogenesis. NIH 3T3 cells stably expressing the 2A-RET oncogenic receptor, with or without β-catenin shRNA, were injected s.c. into athymic mice and the ability of cells to form tumor outgrowths was monitored ( Fig. 5A ). NIH 3T3 cells expressing 2A-RET grew rapidly in this model, producing measurable tumors by day 9 ( Fig. 5A and B), whereas parental NIH 3T3 cells produced no detectable lesions ( Fig. 5B). In the presence of shRNA directed against β-catenin, tumor growth was slower with a notable lag in exponential growth phase. Tumors derived from cells expressing 2A-RET and control vector were nodular and adherent to the abdominal wall with frequent regional invasion and, in some cases, spreading to visceral organs, such as kidney ( Fig. 5C; Supplementary Table S3). In contrast, tumors derived from cells expressing 2A-RET and β-catenin shRNA were encapsulated, with no adhesion or invasion to surrounding tissues. The average volume of tumors was >2-fold less (P < 0.001) in the animals receiving cells containing both 2A-RET and β-catenin shRNA compared with 2A-RET and control vector, and this effect was correlated with a reduction in β-catenin protein in the corresponding primary tumor tissue ( Fig. 5D). Together, these data suggest that the highly tumorigenic potential of oncogenic RET mutants is, in part, mediated through a β-catenin signaling cascade.
These observations establish a novel signaling pathway linking the RET receptor tyrosine kinase to β-catenin. We show that RET and β-catenin can interact directly, leading to RET-mediated β-catenin tyrosine phosphorylation, nuclear translocation, and induction of a RET–β-catenin transcriptional program ( Fig. 6 ). In cell-based models, our data confirm that β-catenin is an important contributor to RET-induced cell proliferation and colony formation. In vivo, our data showing nuclear β-catenin localization in primary human MTC samples are consistent with activation of a RET–β-catenin pathway in these tumors. Nuclear localization was significantly more common in MTC metastases (P < 0.001) than in primary tumors from MEN 2 patients. Furthermore, β-catenin nuclear localization was observed in two of two cases of MTC metastasis bearing the 2B-RET mutation (M918T), the most transforming RET mutant, associated with the most aggressive clinical form of MEN 2, MEN 2B ( 17), as well as in six of seven MTCs from the CALC-MEN 2BRET mice, which bear the corresponding RET mutation (Supplementary Fig. S1). Our mouse transplantation model was consistent with a role for β-catenin in tumor outgrowth and invasiveness, suggesting that activation of the β-catenin signaling pathway by an active oncogenic RET mutant can increase the relative rate of tumor growth and permit infiltration across tissue layers. Together, these data suggest that β-catenin nuclear localization may be an important marker of tumor progression or advanced disease in human MTC. Interestingly, in familial adenomatous polyposis patients, inactivating mutations of APC that lead to increased nuclear β-catenin are also associated with increased risk for papillary thyroid carcinoma, tumors which are also associated with activating RET mutations ( 3, 33), suggesting that there could be a role for a RET-mediated up-regulation of β-catenin in these tumors as well. Targeting β-catenin pathways in thyroid cancer may in future be another avenue for therapeutic intervention in these diseases. Thus, understanding the cellular mechanisms by which RET activates the β-catenin pathway may provide potentially broad insight into the pathogenesis of multiple thyroid tumor types.
In summary, we have identified a novel RET–β-catenin signaling pathway, which is a critical contributor to enhanced cell proliferation and tumor progression in thyroid cancer. Our studies show that RET induces β-catenin–mediated transcription, cell proliferation and transformation in vitro and that β-catenin nuclear localization and the resultant RET-mediated β-catenin signaling is a key secondary event in tumor growth and spreading in vivo. This novel interaction suggests a mechanism that may underlie the broad and early metastatic potential of MTC. Our data suggest a previously unrecognized role for β-catenin signaling that may have implications for tyrosine kinase mediated tumorigenesis in multiple tumor types.
Grant support: Cancer Research Society (L.M. Mulligan) and Dutch Cancer Society (W. van Veelen, D.S. Acton, and J.W.M. Höppener). T.S. Gujral is the recipient of a Canadian Institutes for Health Research Traineeship in Transdisciplinary Cancer Research and a CIHR doctoral award, and D.S. Richardson is the recipient of a Terry Fox Foundation doctoral award.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. M.R. Canninga (Department of Pathology, University Medical Center Utrecht) for providing the human MTC material, Drs. J. Woodgett, P. Greer, and J. MacLeod for providing us with constructs, and Drs. S. Peng, P. Greer, and D. LeBrun and N. Kaur for helpful discussion.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received October 31, 2007.
- Revision received December 13, 2007.
- Accepted December 20, 2007.
- ©2008 American Association for Cancer Research.