A Novel Activating Mutation in the RET Tyrosine Kinase Domain ...

0888-8809/06/$15.00/0 Printed in U.S.A.

Molecular Endocrinology 20(7):1633–1643 Copyright © 2006 by The Endocrine Society doi: 10.1210/me.2004-0447

A Novel Activating Mutation in the RET Tyrosine Kinase Domain Mediates Neoplastic Transformation Aaron Cranston, Cristiana Carniti, Sam Martin, Piera Mondellini, Yvette Hooks, Jean Leyland, Shirley Hodgson, Sue Clarke, Marco Pierotti, Bruce A. J. Ponder, and Italia Bongarzone Cambridge Institute for Medical Research and Cancer Research UK Department of Oncology (A.C., S.M., Y.H., J.L., B.A.J.P.), University of Cambridge, Cambridge CB2 2XY, United Kingdom; Department of Experimental Oncology (C.C., P.M., M.P., I.B.), Istituto Nazionale per lo Studio e la Cura dei Tumori, 20133 Milan, Italy; and Division of Medical and Molecular Genetics (S.H.) and Department of Nuclear Medicine (S.C.), Guys’ Hospital, London SE1 9RT, United Kingdom We report the finding of a novel missense mutation at codon 833 in the tyrosine kinase of the RET proto-oncogene in a patient with a carcinoma of the thyroid. In vitro experiments demonstrate that the R833C mutation induces transformed foci only when present in the long 3ⴕ splice isoform and, in keeping with a model in which the receptor has to dimerize to be completely activated, glial cell linederived neurotrophic factor stimulation leads the RETR833C receptor to a higher level of activation. Tyrosine kinase assays show that the RETR833C long isoform has weak intrinsic kinase activity and phosphorylation of an exogenous substrate is not elevated even in the presence of glial cell linederived neurotrophic factor. Furthermore, the

R833C mutation is capable of sustaining the transformed phenotype in vivo but does not confer upon the transformed cells the ability to degrade the basement membrane in a manner analogous to metastasis. Our functional characterization of the R833C substitution suggests that, like the V804M and S891A mutations, this tyrosine kinase mutation confers a weak activating potential upon RET. This is the first report demonstrating that the introduction of an intracellular cysteine can activate RET. However, this does not occur via dimerization in a manner analogous to the extracellular cysteine mutants. (Molecular Endocrinology 20: 1633–1643, 2006)

T

cause MEN 2A (11, 12). A small number of cases of MEN 2A and familial MTC have causative mutations in the intracellular TK domain (12). MEN 2B is almost exclusively caused by a single missense mutation at codon 918 (M918T) (13–15), with a mutation at codon 883 (A883F) being responsible for the remainder of cases (15, 16). Approximately 75% of MTC cases are sporadic, defined as having no family history with a peak age of onset of 50–60 yr of age. However, population studies have demonstrated that 20–40% of tumors have somatic RET mutations, whereas between 1.5 and 10% actually represent occult MEN 2 cases with germline mutations in RET (17–20). Nonetheless, this leaves a significant percentage of true sporadic MTC cases in which somatic RET mutations have been identified (5). The MEN 2B-causative M918T mutation is the most common being present in 20–40% of MTCs (4, 20). Somatic mutations involving the cysteine codons, characteristic of MEN 2A and familial MTC are rare in sporadic MTCs (20–25) as are somatic mutations in the RET intracellular domain (26, 27). Interestingly, the less frequent allele of a polymorphism at codon 836 has been reported to be overrepresented in sporadic MTC patients that have the somatic M918T mutation (28), suggesting low penetrance variants that might contribute to the etiology of MTC.

HE RET PROTO-ONCOGENE encodes a receptor tyrosine kinase that, in combination with coreceptors, is normally regulated by soluble ligands belonging to the glial cell line-derived neurotrophic factor (GDNF) family of growth factors (reviewed in Ref. 1). RET is mainly expressed in the nervous system and in the urogenital system and is involved in the development of the neural crest (2, 3). RET activation in these locations promotes neuronal cell survival and differentiation. Gain-of-function mutations in RET have been identified in patients with the inherited cancer syndrome, multiple endocrine neoplasia (MEN) type 2, and in some apparently sporadic cases of medullary thyroid carcinoma (MTC) and pheochromocytoma (reviewed in Refs. 4–10). Typically, amino acid substitutions affecting cysteine residues in the RET extracellular domain that result in ligand-independent covalent dimerization and constitutive activation of the tyrosine kinase (TK) First Published Online February 9, 2006 Abbreviations: DAPI, 4⬘,6-Diamidino-2-phenylindole; GDNF, glial cell line-derived neurotrophic factor; MBP, myelin basic protein; MEN, multiple endocrine neoplasia; MTC, medullary thyroid carcinoma; TK, tyrosine kinase. Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.

1633

1634 Mol Endocrinol, July 2006, 20(7):1633–1643

RESULTS Clinical History The patient presented at 59 yr of age with a mass in the neck. She underwent partial thyroidectomy; the pathology is reported as medullary carcinoma of the thyroid. She subsequently underwent surgery for the remaining thyroid lobe. Investigation for pheochromocytoma (urine catecholamines) and parathyroid disease (serum calcium) showed no abnormalities. Regarding the thyroid, she remains well 20 yr later with no clinical features to suggest MEN 2 syndromes. To date, family members have not been screened for mutations in the RET gene, so the possibility that this case of MTC has a familial origin cannot be entirely excluded. The patient was recruited as part of a systematic study in which the mutation status of RET was investigated in apparently sporadic cases of MTC. Sequencing of genomic DNA extracted from a blood sample from the patient revealed the presence of a missense mutation, CGC3TGC, at codon 833 in exon 14 of the RET proto-oncogene (data not shown). This is predicted to change the arginine to a cysteine. The presence of the mutation was confirmed by sequencing in both forward and reverse directions. Arg833 is located within the kinase insert region that splits the intracellular tyrosine kinase domain of RET in two. According to the primary sequence outlined in van der Geer et al. (29), the residue is located between ␣-helices D and E of the intracellular tyrosine kinase catalytic domain. The RETR833C Mutation Induces Transformation in Vitro The transforming activity of the RETR833C mutation in both long and short 3⬘ splice isoforms was assayed by transfecting the constructs into NIH3T3 cells in parallel with other known activating mutations (C634R, V804M, S891A, and M918T). The cells transfected with the RETR833C expression vectors displayed a similar morphology to those transfected with the MEN 2A mutation (C634R) and different from that of the parental NIH3T3 line, suggesting that the R833C mutation is capable of inducing the transformed phenotype in NIH3T3 cells (Fig. 1A). We then analyzed biochemically the expression of each RET mutation in two independent foci of transformation and all foci demonstrated expression of the RET protein as shown by Western blotting (Fig. 1B). To quantify the transforming potential of both 3⬘ splice isoform RETR833C mutants relative to the other known activating mutations, the number of focus forming units per microgram DNA transfected was assayed either in the presence or absence of chronic GDNF stimulation (Fig. 1C). We observed a transforming activity of the long isoform of RET when carrying the R833C, S891A, or V804M mutation in the absence of GDNF. However, the transforming potential of these TK mutants was about 5- to 15-fold lower, depending

Cranston et al. • A Novel Oncogenic Mutation in RET

on mutation type, than that of the RETM918T mutation, also on the long-form. This is in agreement with a previous study by Iwashita et al. (30) on kinase domain mutants. With the exception of the RETM918T shortform receptors, we did not observe transforming activity in the 3⬘ splice short isoforms of RETR833C or RETS891A in the absence of GDNF; however, each of these mutant isoforms became maximally activated when GDNF was administered. The transforming activity of the RETV804M short isoform could not be determined. Complementation Studies The above experiments were conducted in NIH3T3 fibroblasts, which do not express RET, and therefore may not accurately reflect adrenal medullary or thyroid cells that have undergone transformation and where both mutant and wild-type RET are expressed. To address this limitation, we carried out a further series of experiments in which we performed dual transfections to obtain RETWT and RETR833C coexpression in NIH3T3 cells in the presence of GDNF chronic stimulation. As controls, we also transfected RETR833Ciso9, RETR833Ciso51, RETM918Tiso51, and the combination RETWTiso9 plus RETM918Tiso51. Morphologically transformed cell clones from WTiso9 and R833Ciso51 or WTiso51 and R833Ciso9 transfections were examined by immunofluorescence microscopy to reveal the expression of the short (iso9; red) and long (iso51; green) isoforms; the expressing cells were assessed by 4⬘,6-diamidino-2phenylindole (DAPI) nuclear staining; the yellow staining in the merged photomicrographs indicates coexpression of both the RET isoforms (Fig. 2A). Transient transfections of HEK293T cells were performed with RETWTiso9 or RETWTiso51 alone or in combination with the complementary isoform of the RETR833C mutant or RETM918Tiso9 mutant for further biochemical studies. Expression of the correct RET isoform in both singly and doubly transfected cells was also verified by Western blotting using C terminus-specific antibodies (Fig. 2B, upper panels). We used the kinase activity of the mutant RET as a readout of activation. Kinase activity was evaluated directly by examining the antiphosphotyrosine profile obtained after Western blotting (Fig. 2B, bottom panel). Consistent with the results of our NIH3T3 transformation assays, this series of experiments confirmed that R833C mutation was indeed activating. RETR833C Sustains Tumor Formation in Vivo The same transformed foci that were characterized in vitro were tested for the ability to induce and sustain tumor formation in vivo. However, only mutations on the long isoform of RET were examined as these displayed transforming activity without the need for GDNF stimulation. After sc injection in immunocompromised mice, every mutation was found to be capable of causing tumor formation in vivo. The time taken in days for each mutation to form tumors of a


Mol Endocrinol, July 2006, 20(7):1633–1643 1635

Fig. 1. The RETR833C Mutation Induces Transformation In Vitro A, NIH3T3 cells transfected with the RET mutations display a transformed phenotype when grown on plastic. Image shows the morphology of untransfected NIH3T3 cells (A), NIH3T3 cells transfected with the MEN 2A mutation (C634R) (B), and NIH3T3 cells transfected with the R833C mutation on the long isoform of RET, displaying the transformed phenotype (C). B, Transfectant foci express mutant RET protein. RET protein was immunoprecipitated (IPC) from transfected cells and Western immunoblotted (WB) using anti-RET antibodies. For each mutation transfected, two foci were analyzed (lanes 1 and 2); NIH3T3 cells which do not express RET were used as a negative control (⫺ve); cells transfected with the MEN 2A mutation (C634R) and known to express the mutant protein were used as positive controls (⫹ve). C, Transformation potential of the two RET isoforms (iso9 and iso51) carrying activating mutations in the presence or absence of GDNF stimulation (10 ng/␮l). SDs from the mean are shown.

similar size was plotted as the latency period (Fig. 3). As expected, NIH3T3 cells, used as a negative control, did not induce tumor formation within the study period of 49 d. The MEN 2A mutation (C634R) gave rise to tumors with a mean latency of 16.6 d (range, 13–21 d) and was used as a comparison for the other mutations. The MEN 2B mutation (M918T) was not assayed in vivo for technical reasons. Generally, the mean time to tumor formation was found to be similar across the mutations. The data were analyzed statistically using a one-way ANOVA test (MINITAB), and no statistical significant difference was found between mutation types in relation to latency (P ⫽ 0.684). This suggests that each mutation was capable of inducing tumors in vivo at similar rates. Overall, these results confirm our in vitro transformation results and demonstrate for the first time that

the RET R833C mutation is functional and capable of elucidating tumor formation in vivo. RET R833C Does Not Confer the Potential to Invade We assessed the ability of cells harboring the various RET mutations on the long isoform to degrade matrigel, representing the basement membrane, and invade in a manner analogous to metastasis, using an in vitro assay. Only those cells with the ability to degrade the matrigel can migrate through an underlying porous membrane and are labeled. The human fibrosarcoma cell line HT-1080, which is commonly used as a highly invasive cell line, served as a positive control. NIH3T3 cells, which are considered noninvasive, were included to establish a baseline for the experiment. The mutants were as-



Fig. 2. Coexpression Studies A, NIH3T3 cell clones expressing WTiso9 and R833Ciso51 or WTiso51 and R833Ciso9 viewed by immunofluorescence microscopy to reveal the expression of the short (iso9; red staining) and long (iso51; green staining) RET carboxyl isoforms. The yellow staining in the merged images indicates coexpression of both RET isoforms. The expressing cells were assessed using DAPI nuclear staining. Final magnification 400-fold. B, To obtain RETWT and RETR833C coexpression and evaluate their activity, 293T cells were transiently transfected with RETWTiso9 or RETWTiso51 alone or in combination with the complementary isoform of RETR833C mutant. As controls, we also transfected RETR833Ciso9, RETR833Ciso51, RETM918Tiso51, and the combination RETWTiso9 plus RETM918Tiso51. The expression of the correct RET isoform was verified by using C terminus-specific antibodies in immunoblotting on both RET singly and doubly transfected cells (upper panels). The activity of the RET sequences were evaluated by examining the antiphosphotyrosine patterns obtained in immunoblotting experiments using an antibody specific for phosphorylated tyrosines (lower panel). IB, Immunoblot.



Fig. 3. In Vivo Tumor Latency Scatter Plot for Each Transfected Mutation against Time in Days Cells transfected with the MEN 2A and C634R mutations, and NIH3T3 cells were used as positive and negative controls, respectively. Each closed circle data point represents an individual animal that was killed when the tumor reached 1 cm in diameter; open triangles, mean tumor latency; square boxes, no tumor formation by end of study (49 d).

sayed in duplicate, and experiments were replicated four times. As shown in Fig. 4, the C634R and M918T mutations confer invasion potential to NIH3T3 cells; in both cases, this invasion potential is statistically significant (t test, P ⫽ 0.02). None of the other mutations assayed (V804M, R833C, S891A) confers the potential to invade (t test, P ⫽ 0.36, 0.2, and 0.69, respectively). The results are in line with our transformation and kinase data, indicating that the R833C mutation behaves in a manner similar to that of the 804 and 891 mutations rather than the classical 2A and 2B mutations. This reinforces the evidence that the R833C mutant belongs to a class of mutations that confer low activating potential to RET. RETR833C Exhibits Weak Kinase Activity and Does Not Induce Receptor Dimerization To assess the effects of the Arg3Cys at codon 833 on RET kinase activity in vitro, RET tyrosine-phosphorylated proteins were analyzed by Western blotting with antiphosphotyrosine antibodies. The intensity of the bands corresponding to autophosphorylated RET and transphosphorylated myelin basic protein (MBP) substrate was quantified and expressed as the fold increase relative to unstimulated wild-type RET (Fig. 5A). Control reactions, in which RET immunoprecipitates were omitted, showed no MBP phosphorylation (data not shown). It is interesting to note that, of the mutants assayed, only the M918T mutation possesses strong autophosphorylation and transphosphorylation potential, whereas the potential of the R833C mutant is about 6-fold less, which is similar to the V804M and S891A mutants (Fig. 5B). Introduction of cysteine residues may determine the formation of new intermolecular or intramolecular disulfide bond formation, thus altering conformation and activity. We postulated that introduction of a cysteine

Fig. 4. Invasion Potential of Cells through Matrigel FluoroBlok Assay Wells Only those cells that degrade the matrigel analogous to the basement membrane can migrate through and are labeled. Values are given as mean fluorescence units, and SDs from the mean are plotted. Noninvasive NIH3T3 cells were used as a negative control and invasive HT-1080 cells as positive controls in the assay. Each experiment was performed in duplicate and replicated four times (n ⫽ 8). For reference, the threshold for invasion, as indicated by that of NIH3T3 cells, is shown as a dotted line. Unless otherwise stated, mutations were introduced on the RET long isoform (iso51). Only the positive control cell line HT-1080 and the MEN 2 mutations, C634R and M918T, were observed to be statistically significantly different from the noninvasive NIH3T3 cell line (t test, P ⫽ 0.02).

residue in this region of the receptor could introduce a structural modification consistent with disulfide bond formation. To test whether cysteine 833 may be involved in the formation of an intermonomeric disulphide bond, thereby promoting dimer formation, we performed nonreducing SDS-PAGE analysis of S35MetCys metabolically labeled RETR833C-expressing cells. We did not observe dimerization of the RETR833C mutant receptors in either reducing or nonreducing Western blot experiments or in pulse-chase experiments, although, as expected, we were able to identify dimers in the C634R lysate under nonreducing conditions (Fig. 5C). These results suggest that the transforming potential of cysteine 833 is not mediated by disulfide bond formation.

DISCUSSION This is the first report of a mutation at codon 833 in the RET proto-oncogene. The R833C mutation involves a nonconserved arginine residue contained within the 27-amino acid insert that splits the tyrosine kinase domain of RET in two. According to the primary



Fig. 5. In Vitro Immunokinase and Dimerization Assays of the Mutant RET Proteins The long isoform of RETK758R [RET kinase dead (KD)], RETM918T, wild-type RET (WT), RETS891A, RETV804M, and RETR833C were transiently expressed in HEK293T cells, immunoprecipitated with anti-RET antiserum, and subjected to autokinase assay and immunokinase assay against the MBP. The kinase-deficient RETK758R (RET KD) protein was used as a negative control. Equal amounts of the immunoprecipitates were subjected to Western blotting using anti-RET antibodies. A, The anti-RET immunoprecipitates subjected to autokinase assay and immunokinase assay against the MBP were eluted and separated in 6% (for RET) or 15% (for MBP) SDS-PAGE. For each RET mutant, 32P-labeled bands corresponding to RET (upper panel) and to MBP (lower panel) were visualized by autoradiography. B shows the phosphorylation activity measured as the fold increase relative to wild-type RET, quantified by ImageQuant software (Media Cybernetics, Silver Spring, MD). SDs from the mean are shown. C, Dimerization analysis of S35MetCys metabolically labeled RETR833C-expressing cells. NIH3T3 and NIH3T3 cells expressing RETC634R and RETM918T mutant proteins were included in the analysis for comparison. The C634R mutation induces RET dimerization and was used as a positive control, whereas the M918T mutation, which does not induce RET activation by dimerization, was used as a negative control. Cell extracts were immunoprecipitated with anti-RET antiserum, and the complexes were separated by SDS-PAGE under reducing and nonreducing conditions. RET immunoprecipitates revealed the presence of RET dimers in the cell line expressing the C634R mutant. Stable dimer formation was not detected in cell lines expressing RETM918T and RETR833C, nor under reducing conditions.


sequence outlined in van der Geer et al. (29), the residue is located between ␣-helices D and E of the intracellular tyrosine kinase catalytic domain. It is also the first example of an activating mutation that involves introduction of a cysteine, which leads to activation of the receptor tyrosine kinase. RET was originally identified by virtue of its ability to transform NIH3T3 fibroblasts, and transformation of these cells resulting in focus formation remains the classical way in which to assess the transforming potential of new mutations. Our earlier studies demonstrated that both the short and long isoforms of the wild-type RET receptor induce cell transformation with the same efficiency but only in the presence of GDNF chronic stimulation (31) (our unpublished results). However, the transforming activity of the short isoform of RET carrying the MEN 2B mutation is strongly increased from a baseline transforming activity by GDNF stimulation, whereas the long isoform showed complete GDNF-independent activation. We have demonstrated in NIH3T3 transformation assays that the R833C mutation (along with S891A and V804M) confers a GDNF-independent activation to the long isoform of RET but at a level that is lower than that of the long isoform when activated by the MEN 2B mutation (M918T). The above experiments were conducted in NIH3T3 fibroblasts, which do not express RET, and therefore may not necessarily reflect the situation in cells, which having undergone transformation, now express both mutant and wild-type alleles of RET. To address this, we assessed activation of the mutant RET in the context of wild-type RET. Initially, we performed complementation experiments in which we cotransfected 3T3 cells with mutant and wild-type RET. Then, as an extension, and to confirm the results of these studies, we introduced the R833C mutant allele into the human embryonic kidney cell line 293T, which endogenously expresses normal, wild-type RET. Together, these experiments enabled us to express mutant RET in the context of wild-type RET, a situation analogous to expression of the mutated allele in the heterozygous state in transformed cells. In both cases, kinase assays performed on these cells as a readout of activation indicate that the R833C mutation is indeed activating in the heterozygous state and confirm the results of our initial transformation assays performed in NIH3T3 cells. Thus, results from all three series of experiments were similarly confirmatory and indicate that RETR833C is activated even in the presence of wild-type RET, a situation analogous to the heterozygous expression of the mutated allele in adrenal medullary and thyroid cells where both mutant and normal RET are expressed. By analogy to other receptor tyrosine kinases, the RET receptor should require ligand binding for dimerization and subsequent activation, leading to autophosphorylation and phosphorylation of specific downstream targets (32). In the absence of ligand, we did not observe activation of RETR833C


short-form receptors. However, when present in the long isoform, the R833C mutation activated the RET receptor to low levels in the absence of exogenous growth factors. This transforming activity could be further increased to almost maximal levels by the addition of GDNF. This observed difference in oncogenicity of the two 3⬘ carboxyl splice isoforms of RETR833C may be explained by the downstream interactions of RET. Alternative splicing of the 3⬘ exons results in three RET protein isoforms with either nine (short form), 43 (intermediate), or 51 (long form) unique amino acids. These C-terminal isoforms differ in the relative binding of adaptor molecules, such as SHC, to Tyr1062, which lies in different amino acid sequence contexts in each of the three isoforms (33). Furthermore, the adaptor protein, GRB2, binds to Tyr1096, which is only present in the RET long isoform. Consistent with these differences, our results demonstrate that the short and long isoforms harboring the R833C mutation display different transformation efficiencies. Alternatively, as the R833C mutation is responsible for a basal oncogenicity of the mutated long-form receptor even in the absence of ligand, it is tempting to speculate on the role of the specific C-terminal domain in removing some of the inhibitory constraints acting on the kinase domain. Specifically, this may relate to the length of the C terminus and/or the sequence context of the C-terminal domain. Nonetheless, and in accordance with a model in which the receptor has to dimerize to become completely activated, we have demonstrated that ligand stimulation leads both RETR833C short- and long-form receptors to an increased level of activation. Our in vitro biological and biochemical analyses bear out the fact that the V804M mutation is weakly transforming when compared with isoforms with the C634R or M918T mutations. This is in agreement with a previous study (34) and also studies investigating the transforming activity of a similar mutation, V804L (30, 35). Furthermore, mutation of codon 804 has been associated with later onset and less aggressive disease in sporadic MTC (36). Similar weak transforming activities have been reported for mutations of codons 768 and 891 (Ref. 30; this study), suggesting that mutations in exons 13, 14, and 15 may represent lower penetrance mutations with a less aggressive phenotype. Certainly, our biochemical data support this conclusion: all three mutations (V804M, R833C, and S891A) demonstrated a weak ability to undergo both autophosphorylation and to transphosphorylate an exogenous substrate, suggesting that these mutations might not be able to confer sufficient release from autoinhibitory signals to be strongly transforming. When considered together, the characterization of the R833C substitution indicates that, like the V804M and S891A mutations, this TK mutation can be placed into the conceptual class of conferring a reduced activating potential upon RET (8).

1640

Mol Endocrinol, July 2006, 20(7):1633–1643

Our in vivo tumorigenicity results provide functional evidence that the R833C mutation is tumorigenic and, in parallel, demonstrate that the V804M and S891A mutations are tumorigenic. However, we were unable to demonstrate a statistically significant difference in the onset of tumor induction in vivo between the mutation types. Although this is the first example of an activating RET mutation that involves the introduction of a cysteine residue, it is not entirely clear how this Arg3Cys substitution in the insert region of the intracellular tyrosine kinase domain induces oncogenic activation of the RET receptor. We considered the possibility that the substitution of cysteine for arginine at codon 833 in RET might induce dimerization of mutant tyrosine kinase domains through covalent intermolecular disulfide bridges in a manner similar to that of the MEN 2A extracellular cysteines. Corroboration for this possibility is provided by two studies in which a reduction in dimer formation and kinase activity was observed both in RET and a fusion product when an intracellular residue was mutated from cysteine to alanine, raising the possibility that the cysteine residue at this position might have been involved in disulfide bond-mediated dimerization of intracellular domains (37, 38). However, in the present study, we did not observe dimerization of tyrosine kinase domains using pulse-chase and nonreducing Western blot assays. This suggests that the R833C mutation does not, in fact, lead to activation of the kinase via disulfide bond-mediated dimerization of monomeric RET receptors. Consequently, dimerization, through ligand binding, is still necessary for RETR833C to be elevated to a further level of activity. Understanding further the specificity with which RET mutations are associated with specific disease states is of fundamental clinical importance. The ability to identify further new low penetrance activating mutations and/or modifiers may allow us to begin to elucidate the basis of specific genotype-phenotype correlations, upon which we may be able to begin to specify genotype-based treatment regimens. Further understanding of the downstream pathways in tissues that have undergone neoplastic transformation should allow us to develop novel therapeutic approaches with which to interfere in the disease state: currently, these are focused on inhibiting tyrosine kinase domain activity (39–43).

MATERIALS AND METHODS Patients Patients were recruited to the UK MEN 2 Register held in the University of Cambridge/Cancer Research UK Department of Oncology under the supervision of Prof. Bruce Ponder. The register has ethical approval from the Eastern Multi-Centre Research Ethics Committee. The patient, MEN0929, was recruited as part of a systematic study of RET mutations in sporadic cases of MTC.


Mutation Detection Genomic DNA was extracted from peripheral blood and exons 13, 14, and 15 of the RET gene individually amplified by PCR and then sequenced using the DyeDeoxy terminator cycle sequencing kit (Applied Biosystems, Foster City, CA). Sequence variations were confirmed both in the forward and reverse direction. Site-Directed Mutagenesis The cDNAs of the human long and short RET isoforms were cloned into expression vector pRc/CMV (Invitrogen, San Diego, CA) and mutations introduced using the GeneEditor site-directed mutagenesis kit (Promega, Madison, WI) and PAGE-purified mutagenic primers (Sigma-Aldrich, St. Louis, MO). The cDNAs were sequenced to ensure that only the desired change(s) were introduced. Cell Lines and In Vitro Transformation Assays NIH3T3 and HEK293T cell lines were maintained in DMEM supplemented with 10% Colorado calf serum (Celbio, Milan, Italy) and grown at 37 C and 10% CO2. For transformation assays, the expression constructs were transfected into the mouse fibroblast NIH3T3 cell line, and their ability to transform NIH3T3 cells was assayed in triplicate as previously described (44). Where indicated, growth factor stimulation was given by supplying GDNF (10 ng/␮l). HEK293T cells were transiently transfected by calcium phosphate precipitation. Xenograft Studies Twelve-week-old SCID mice (Charles River, Margate, Kent, UK) were randomly assigned to groups and the cell lines coded so that the study was conducted “blind.” A total of 0.2 ⫻ 106 cells was injected sc into the flank of each mouse. Only mutations in the RET long isoform were analyzed. Mice were weighed and checked thrice weekly for signs of tumor development. Mice were killed when the tumor reached 1 cm in diameter. The study was repeated twice. Animals were housed in individually ventilated cages in specific pathogenfree conditions and handled accordingly. All animals were treated in strict accordance with best practice, the United Kingdom Coordinating Committee on Cancer Research guidelines, and within the remit of the appropriate Home Office license. Expression Assays After transfection, cells were harvested to confirm expression from the construct. Expression of RET mutants was confirmed by immunoprecipitation and Western blotting (45). Expression of the correct RET isoform was verified using C terminus-specific antibodies (C19 and C20; Santa Cruz Biotechnology, Santa Cruz, CA). The kinase activity of the RET mutants was evaluated in Western blotting using an antiphosphotyrosine antibody (4G10; Upstate Biotechnology, Lake Placid, NY). For immunofluorescence microscopy studies, cells were plated on 35-mm-diameter tissue culture dishes coated with filtered 0.01% poly-L-lysine. Cells were fixed in 4% paraformaldehyde and 2% sucrose and permeabilized with 0.5% Triton X-100 in PBS. Cells were incubated at 4 C with anti-RET C19 antibody (Santa Cruz Biotechnology) and anti-RET C20 antibody (Santa Cruz Biotechnology), which recognize the short and long isoforms, respectively. Immunostaining with primary antibodies was followed by incubation with Alexa Fluor 546, goat antirabbit IgG Red (Molecular Probes, Eugene, OR) or fluorescein isothiocyanate-


conjugated AffiniPure Donkey Anti-Goat IgG (Jackson ImmunoResearch, West Grove, PA). The cells were assessed using nuclear staining with DAPI and images obtained using a Nikon Eclipse E1000 microscope. Final magnification in photomicrographs is 400-fold.


ing conditions and pulse-chase experiments as described previously (46).

Acknowledgments We thank Dr. Michael Festing for help with statistics.

In Vitro Invasion Assays The ability of cells to degrade an in vitro matrix analogous to the basement membrane was assessed using the BioCoat HTS FluoroBlok Invasion System (BD Biosciences, Mountain View, CA). NIH3T3 cells were included as a noninvasive control and the commonly used invasive human fibrosarcoma HT-1080 cell line was used as an invasive control. Matrigel inserts were hydrated with PBS for 2 h at 37 C. Cells grown to confluence were trypsinized and insert wells were seeded with 2.5 ⫻ 104 cells in 0.5 ml serum-free DMEM. Growth medium (0.75 ml) was added to the bottom well (the serum acts as a chemoattractant) and the plate was incubated for 22 h at 37 C in 10% CO2. Only cells that degraded the matrix and invaded through the 8 ␮m pore membrane were labeled with a fluorescent dye, Calcein AM (Molecular Probes) (0.5 ml/well of 4 ␮g/ml Calcein AM in HBS) for 90 min at 37 C and counted on an Applied Biosystems CytoFluor 4000 plate reader with excitation/emission wavelengths of 485/530 nm at a gain of 55 and 4 reads per well. A FluoroBlok insert system was used to establish background polyethylene terepthalate membrane autofluorescence readings. Each experiment was performed in duplicate in the plate and replicated four times. Statistically significant differences were determined using a two-tailed, unpaired (type 2) t test for pairwise comparisons. Kinase Assays The long isoform of RETK758R [RET kinase dead (KD)], RETM918T, RET wild type (WT), RETS891A, RETV804M, and RETR833C were transiently expressed in HEK293T cells, and the kinase activity of the mutants was assessed in vitro by the ability of cells to autophosphorylate and transphosphorylate MBP (Sigma-Aldrich). Transfected cells were lysed in Nonidet P-40 buffer and protein extracts were immunoprecipitated with anti-RET antibodies, washed twice with lysis buffer and once with incubation buffer [50 mM HEPES (pH 7.2), 20 mM MnCl2, 5 mM phenylmethylsulfonyl fluoride], and incubated for 15 min at 4 C in 20 ␮l of the same buffer containing 0.5 mM dithiothreitol, 4 ␮Ci [␥-32P]ATP diluted with unlabeled ATP to the final concentration of 26 pmol ATP per sample, and 50 ␮M MBP. The reactions were stopped by an equal volume of 2⫻ reducing Laemmli buffer, and the products were heated at 100 C for 5 min. Proteins were eluted and separated in 6 or 15% SDS-PAGE. 32P-Labeled bands were visualized by autoradiography. The intensity of the bands corresponding to phosphorylated RET and to phosphorylated MBP was quantified by PhosphorImager (Molecular Dynamics, Sunnyvale, CA) and expressed as the fold increase relative to unstimulated wild-type RET. Immunoprecipitation of equal amounts of RET was verified by anti-RET Western blotting, and the quantity was measured using 125I-labeled protein A. The intensity of the bands corresponding to RET was quantified by PhosphorImager (Molecular Dynamics) analysis. Experiments were repeated in triplicate. Dimerization Assays Lysates from each of the mutant cell lines were subjected to one-dimensional SDS-PAGE under reducing and nonreduc-

Received November 5, 2004. Accepted January 31, 2006. Address all correspondence and requests for reprints to: Aaron Cranston, Cambridge Institute for Medical Research and Cancer Research UK Department of Oncology, University of Cambridge, Hills Road, Cambridge CB2 2XY, United Kingdom. E-mail: [email protected]. This work was supported by grants from Cancer Research UK, the Italian Association for Cancer Research, and the Te´lethon Foundation. B.A.J.P. is a Gibb Fellow of Cancer Research UK.

REFERENCES 1. Takahashi M 2001 The GDNF/RET signaling pathway and human diseases. Cytokine Growth Factor Rev 12: 361–373 2. Pachnis V, Mankoo B, Costantini F 1993 Expression of the c-ret proto-oncogene during mouse embryogenesis. Development 119:1005–1017 3. Edery P, Eng C, Munnich A, Lyonnet S 1997 RET in human development and oncogenesis. Bioessays 19: 389–395 4. Ponder BA, Smith D 1996 The MEN II syndromes and the role of the ret proto-oncogene. Adv Cancer Res 70: 179–222 5. Eng C, Mulligan LM 1997 Mutations of the RET protooncogene in the multiple endocrine neoplasia type 2 syndromes, related sporadic tumours, and hirschsprung disease. Hum Mutat 9:97–109 6. Eng C 1999 RET proto-oncogene in the development of human cancer. J Clin Oncol 17:380–393 7. Ponder BA 1999 The phenotypes associated with ret mutations in the multiple endocrine neoplasia type 2 syndrome. Cancer Res 59(Suppl):1736–1741 8. Santoro M, Carlomagno F, Melillo RM, Billaud M, Vecchio G, Fusco AJ 1999 Molecular mechanisms of RET activation in human neoplasia. Endocrinol Invest 22:811–819 9. Hansford JR, Mulligan LM 2000 Multiple endocrine neoplasia type 2 and RET: from neoplasia to neurogenesis. J Med Genet 37:817–827 10. Jhiang SM 2000 The RET proto-oncogene in human cancers. Oncogene 19:5590–5597 11. Asai N, Iwashita T, Matsuyama M, Takahashi M 1995 Mechanism of activation of the ret proto-oncogene by multiple endocrine neoplasia 2A mutations. Mol Cell Biol 15:1613–1619 12. Santoro M, Carlomagno F, Romano A, Bottaro DP, Dathan NA, Grieco M, Fusco A, Vecchio G, Matoskova B, Kraus MH, Di Fiore PP 1995 Activation of RET as a dominant transforming gene by germline mutations of MEN2A and MEN2B. Science 267:381–383 13. Carlson KM, Dou S, Chi D, Scavarda N, Toshima K, Jackson CE, Wells Jr SA, Goodfellow PJ, Donis-Keller H 1994 Single missense mutation in the tyrosine kinase catalytic domain of the RET protooncogene is associated with multiple endocrine neoplasia type 2B. Proc Natl Acad Sci USA 91:1579–1583 14. Hofstra RM, Landsvater RM, Ceccherini I, Stulp RP, Stelwagen T, Luo Y, Pasini B, Hoppener JW, van Amstel HK, Romeo G, Lips CJM, Buys CHCM 1994 A mutation in the RET proto-oncogene associated with multiple endocrine neoplasia type 2B and sporadic medullary thyroid carcinoma. Nature 367:375–376


15. Smith DP, Houghton C, Ponder BA 1997 Germline mutation of RET codon 883 in two cases of de novo MEN 2B. Oncogene 15:1213–1217 16. Gimm O, Marsh DJ, Andrew SD, Frilling A, Dahia PL, Mulligan LM, Zajac JD, Robinson BG, Eng C 1997 Germline dinucleotide mutation in codon 883 of the RET protooncogene in multiple endocrine neoplasia type 2B without codon 918 mutation. J Clin Endocrinol Metab 82: 3902–3904 17. Eng C, Mulligan LM, Smith DP, Healey CS, Frilling A, Raue F, Neumann HP, Ponder MA, Ponder BA 1995 Low frequency of germline mutations in the RET proto-oncogene in patients with apparently sporadic medullary thyroid carcinoma. Clin Endocrinol 43:123–127 18. Wohllk N, Cote GJ, Bugalho MM, Ordonez N, Evans DB, Goepfert H, Khorana S, Schultz P, Richards CS, Gagel RF 1996 Relevance of RET proto-oncogene mutations in sporadic medullary thyroid carcinoma. J Clin Endocrinol Metab 81:3740–3745 19. Schuffenecker I, Virally-Monod M, Brohet R, Goldgar D, Conte-Devolx B, Leclerc L, Chabre O, Boneu A, Caron J, Houdent C, Modigliani E, Rohmer V, Schlumberger M, Eng C, Guillausseau PJ, Lenoir GM 1998 Risk and penetrance of primary hyperparathyroidism in multiple endocrine neoplasia type 2A families with mutations at codon 634 of the RET proto-oncogene. J Clin Endocrinol Metab 83:487–491 20. Scurini C, Quadro L, Fattoruso O, Verga U, Libroia A, Lupoli G, Cascone E, Marzano L, Paracchi S, Busnardo B, Girelli ME, Bellastella A, Colantuoni V 1998 Germline and somatic mutations of the RET proto-oncogene in apparently sporadic medullary thyroid carcinomas. Mol Cell Endocrinol 137:51–57 21. Donis-Keller H, Dou S, Chi D, Carlson KM, Toshima K, Lairmore TC, Howe JR, Moley JF, Goodfellow P, Wells Jr SA 1993 Mutations in the RET proto-oncogene are associated with MEN 2A and FMTC. Hum Mol Genet 2:851–856 22. Zedenius J, Wallin G, Hamberger B, Nordenskjold M, Weber G, Larsson C 1994 Somatic and MEN 2A de novo mutations identified in the RET proto-oncogene by screening of sporadic MTCs. Hum Mol Genet 3:1259–1262 23. Eng C, Mulligan LM, Smith DP, Healey CS, Frilling A, Raue F, Neumann HP, Pfragner R, Behmel A, Lorenzo MJ, Stonehouse TJ, Ponder MA, Ponder BAJP 1995 Mutation of the RET protooncogene in sporadic medullary thyroid carcinoma. Genes Chromosomes Cancer 12: 209–212 24. Jhiang SM, Fithian L, Weghorst CM, Clark OH, Falko JM, O’Dorisio TM, Mazzaferri EL 1996 RET mutation screening in MEN2 patients and discovery of a novel mutation in a sporadic medullary thyroid carcinoma. Thyroid 6:115–121 25. Marsh DJ, Andrew SD, Learoyd DL, Pojer R, Eng C, Robinson BG 1998 Deletion-insertion mutation encompassing RET codon 634 is associated with medullary thyroid carcinoma. Human Mutation 1(Suppl):3–4 26. Eng C, Smith DP, Mulligan LM, Healey CS, Zvelebil MJ, Stonehouse TJ, Ponder MA, Jackson CE, Waterfield MD, Ponder BA 1995 A novel point mutation in the tyrosine kinase domain of the RET proto-oncogene in sporadic medullary thyroid carcinoma and in a family with FMTC. Oncogene 10:509–513 27. Marsh DJ, McDowall D, Hyland VJ, Andrew SD, Schnitzler M, Gaskin EL, Nevell DF, Diamond T, Delbridge L, Clifton-Bligh P, Robinson BG 1996 The identification of false positive responses to the pentagastrin stimulation test in RET mutation negative members of MEN 2A families. Clin Endocrinol 44:213–220 28. Gimm O, Neuberg DS, Marsh DJ, Dahia PL, Hoang-Vu C, Raue F, Hinze R, Dralle H, Eng C 1999 Over-representation of a germline RET sequence variant in patients with


29. 30.

31.

32. 33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

sporadic medullary thyroid carcinoma and somatic RET codon 918 mutation. Oncogene 18:1369–1373 van der Geer P, Hunter T, Lindberg RA 1994 Receptor protein-tyrosine kinases and their signal transduction pathways. Annu Rev Cell Biol 10:251–337 Iwashita T, Kato M, Murakami H, Asai N, Ishiguro Y, Ito S, Iwata Y, Kawai K, Asai M, Kurokawa K, Kajita H, Takahashi M 1999 Biological and biochemical properties of Ret with kinase domain mutations identified in multiple endocrine neoplasia type 2B and familial medullary thyroid carcinoma. Oncogene 18:3919–3922 Bongarzone I, Vigano E, Alberti L, Borrello MG, Pasini B, Greco A, Mondellini P, Smith DP, Ponder BA, Romeo G, Pierotti MA 1998 Full activation of MEN2B mutant RET by an additional MEN2A mutation or by ligand GDNF stimulation. Oncogene 16:2295–2301 Heldin CH 1995 Dimerization of cell surface receptors in signal transduction Cell 80:213–223 Ishiguro Y, Iwashita T, Murakami H, Asai N, Iida K, Goto H, Hayakawa T, Takahashi M 1999 The role of amino acids surrounding tyrosine 1062 in ret in specific binding of the shc phosphotyrosine-binding domain. Endocrinology 140:3992–3998 Iwashita T, Murakami H, Kurokawa K, Kawai K, Miyauchi A, Futami H, Qiao S, Ichihara M, Takahashi M 2000 A two-hit model for development of multiple endocrine neoplasia type 2B by RET mutations. Biochem Biophys Res Commun 268:804–808 Pasini A, Geneste O, Legrand P, Schlumberger M, Rossel M, Fournier L, Rudkin BB, Schuffenecker I, Lenoir GM, Billaud M 1997 Oncogenic activation of RET by two distinct FMTC mutations affecting the tyrosine kinase domain. Oncogene 15:393–402 Fink M, Weinhusel A, Niederle B, Haas OA 1996 Distinction between sporadic and hereditary medullary thyroid carcinoma (MTC) by mutation analysis of the RET protooncogene Study Group Multiple Endocrine Neoplasia Austria (SMENA). Int J Can 69:312–316 Kato M, Iwashita T, Akhand AA, Liu W, Takeda K, Takeuchi K, Yoshihara M, Hossain K, Wu J, Du J, Oh C, Kawamoto Y, Suzuki H, Takahashi M, Nakashima I 2000 Molecular mechanism of activation and superactivation of Ret tyrosine kinases by ultraviolet light irradiation. Antioxid Redox Signal 2:841–849 Takeda K, Kato M, Wu J, Iwashita T, Suzuki H, Takahashi M, Nakashima I 2001 Osmotic stress-mediated activation of RET kinases involves intracellular disulfidebonded dimer formation. Antioxid Redox Signal 3:473–482 Carlomagno F, Vitagliano D, Guida T, Napolitano M, Vecchio G, Fusco A, Gazit A, Levitzki A, Santoro M 2002 The kinase inhibitor PP1 blocks tumorigenesis induced by RET oncogenes. Cancer Res 62:1077–1082 Carlomagno F, Vitagliano D, Guida T, Ciardiello F, Tortora G, Vecchio G, Ryan AJ, Fontanini G, Fusco A, Santoro M 2002 ZD6474, an orally available inhibitor of KDR tyrosine kinase activity, efficiently blocks oncogenic RET kinases. Cancer Res 62:7284–7290 Cohen MS, Hussain HB, Moley JF 2002 Inhibition of medullary thyroid carcinoma cell proliferation and RET phosphorylation by tyrosine kinase inhibitors. Surgery 132:960–966 Sawyer TK, Bohacek RS, Dalgarno DC, Eyermann CJ, Kawahata N, Metcalf III CA, Shakespeare WC, Sundaramoorthi R, Wang Y, Yang MG 2002 SRC homology-2 inhibitors: peptidomimetic and nonpeptide. Mini Rev Med Chem 2:475–488 Carniti C, Perego C, Mondellini P, Pierotti MA, Bongarzone I 2003 PP1 inhibitor induces degradation of RETMEN2A and RETMEN2B oncoproteins through proteosomal targeting. Cancer Res 63:2234–2243 Bongarzone I, Monzini N, Borrello MG, Carcano C, Ferraresi G, Arighi E, Mondellini P, Della Porta G,


Pierotti MA 1993 Molecular characterization of a thyroid tumor-specific transforming sequence formed by the fusion of ret tyrosine kinase and the regulatory subunit RI ␣ of cyclic AMP-dependent protein kinase A. Mol Cell Biol 13:358–366 45. Borrello MG, Pelicci G, Arighi E, De Filippis L, Greco A, Bongarzone I, Rizzetti M, Pelicci PG, Pierotti MA 1994 The oncogenic versions of the Ret and Trk tyrosine ki-


nases bind Shc and Grb2 adaptor proteins. Oncogene 9:1661–1668 46. Bongarzone I, Vigano E, Alberti L, Mondellini P, Uggeri M, Pasini B, Borrello, MG, Pierotti MA 1999 The Glu632-Leu633 deletion in cysteine rich domain of Ret induces constitutive dimerization and alters the processing of the receptor protein. Oncogene 18: 4833–4838

Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.

American Society for Bone and Mineral Research (ASBMR) 28th Annual Meeting September 15-19, 2006 Pennsylvania Convention Center Philadelphia, Pennsylvania, USA For more information, call (202) 367-1161, e-mail [email protected], or visit our web site www.asbmr.org Contemporary Diagnosis and Treatment of Vitamin D-Related Disorders A Scientific Meeting Sponsored by the American Society for Bone and Mineral Research December 4-5, 2006 Crystal Gateway Marriott Arlington, Virginia, USA Abstract Submission Deadline: September 1, 2006 For more information, call (202) 367-1161, e-mail [email protected], or visit our web site www.asbmr.org American Society for Bone and Mineral Research (ASBMR) 29th Annual Meeting September 16-20, 2007 Honolulu Convention Center Honolulu, Hawaii, USA For more information, call (202) 367-1161, e-mail [email protected], or visit our web site www.asbmr.org