Signal transduction pathways activated by RET oncoproteins in PC12 pheochromocytoma cells.

Gene alterations in the ret proto-oncogene, which encodes a receptor tyrosine kinase, have been found to associate with several human diseases. In this study, we showed that induction of the vgf promoter activity is a good molecular indicator for RET activation in PC12 cells, a rat pheochromocytoma cell line. We demonstrated that all forms of RET oncoprotein, including RET chimeric oncoproteins found in human papillary thyroid carcinomas (RET/PTC) as well as RET oncoproteins found in patients with multiple endocrine neoplasia type 2A and 2B (2A/RET and 2B/RET) can induce vgf promoter activity in PC12 cells. In contrast, a RET mutant found in a patient with Hirschsprung's disease, as well as a RET/PTC1 mutant with deletion of the dimerization domain, failed to induce vgf promoter activity in PC12 cells. We further determined that the signaling events mediated by phosphorylated Tyr294 and phosphorylated Tyr451 binding sites are essential for RET/PTC1 to induce vgf promoter activity in PC12 cells. We also showed that RET/PTC1, 2A/RET, and 2B/RET induce ELK-, cAMP-responsive element binding protein (CREB), or JUN-mediated gene expression in PC12 cells, and these three signaling events are mediated by phosphorylated Tyr294 and phosphorylated Tyr451 binding sites in RET/PTC1.

The ret proto-oncogene (c-ret) encodes a receptor tyrosine kinase with an extracellular domain, a transmembrane domain, and a cytoplasmic tyrosine kinase domain. The ligands for c-RET have been recently identified as glial cell line-derived neurotrophic factor (1-3) and neurturin (4,5). Gene alterations in c-ret have been found to associate with five different diseases, including papillary thyroid carcinoma (PC), 1 multiple endocrine neoplasia type 2A (MEN2A), MEN2B, familial medullary thyroid carcinoma, and a congenital developmental defect, Hirschsprung's disease (HSCR). The ret/PTC oncogenes encoding chimeric oncoproteins with the RET tyrosine kinase domain fused to the N terminus of other genes are frequently detected in human papillary thyroid carcinomas (see review, see Ref. 6). Missense ret mutations found in patients with MEN2 syndromes are clustered in specific codons, whereas ret mutations found in patients with HSCR are scattered along the gene without any hot spots (for review, see Ref. 7). It was reported that both ret/PTC and ret with MEN2A or MEN2B mutations (2A/ret or 2B/ret) function as dominant oncogenes with transforming activity in NIH/3T3 cells (8 -12), whereas ret with HSCR mutations (HSCR/ret) encodes nonfunctional proteins (13). It has been shown that mutations found in MEN2A lead to formation of disulfide-linked 2A/RET homodimers (10 -12), whereas the mutation found in MEN2B changes the substrate specificity of 2B/RET from a receptor tyrosine kinase to a cytoplasmic tyrosine kinase (14). We recently demonstrated that RET/PTC1 oncoprotein, the major form of RET/PTC found in human papillary thyroid carcinoma, forms constitutive dimers through the leucine zipper motif in the N terminus of H4 (15,16).
The intracellular domain of c-RET consists of 14 tyrosine residues, and the long isoform of c-RET has two extra tyrosine residues at the C terminus (see review, see Ref. 17). Among those tyrosine residues, Tyr 905 (P), Tyr 1015 (P), and Tyr 1062 (P) of c-RET (corresponding to Tyr 294 (P), Tyr 404 (P), and Tyr 451 (P) of RET/PTC1 or Tyr 429 (P), Tyr 539 (P), and Tyr 586 (P) of RET/PTC2) were identified as the docking sites for Grb7/Grb10, PLC␥, and Shc/Enigma, respectively (18 -22). The roles of these phosphotyrosine residues in RET-induced mitogenic activity and transforming activity have been studied vigorously in mouse fibroblasts. The docking site for PLC␥ was reported to be essential for full oncogenic activation of RET/PTC2 in NIH/3T3 cells (19). The docking site for Grb7/Grb10 was reported to be essential for the transforming activity of 2A/RET in NIH/3T3 cells (23). However, both phosphotyrosine residues are not essential for mitogenic signaling of RET/PTC2 in mouse fibroblasts (22). The docking site for Shc/Enigma was reported to be essential not only for the mitogenic activity of RET/PTC2 in mouse fibroblast cells (22) but also for the transforming activity of 2A/RET and 2B/RET in NIH/3T3 cells (24).
RET-induced signaling pathways appear to be cell type-specific. Santoro et al. (25) reported that RET activation induces Ras activation but is not able to activate the MAPK signaling pathway or to induce phosphatidylinositol 3-kinase activity in NIH/3T3 transfectants. However, other investigators have reported that RET activation not only induces the activation of the Ras-MAPK pathway but also induces phosphatidylinositol 3-kinase activity in human neuroblastoma cells that express either exogenous or endogenous c-RET (26 -28). In this paper, we report our studies of RET-induced signaling pathways in a physiologically relevant cell line, PC12 cells. For patients with MEN2A or MEN2B, 10 -50% develop pheochromocytomas. PC12 cells derived from a rat pheochromocytoma are able to differentiate into sympathetic neuron-like cells with neurite outgrowth by stimulation with NGF (29). It has been shown that expression of RET/PTC1, 2A/RET, or 2B/RET in PC12 cells can also induce the expression of various neuronal markers and lead to morphological changes with neurite outgrowth (30 -32). Among these changes, the induction of vgf promoter activity in PC12 cells appears to be specific to NGF stimulation and RET activation, as both epidermal growth factor and bovine fibroblast growth factor are unable to stimulate the vgf promoter activity.

EXPERIMENTAL PROCEDURES
Cell Culture-PC12-N21 cell line (33), kindly provided by Dr. Richard Burry at the Ohio State University, is a subclone of PC12 cells. The cells were grown in DMEM (Life Technologies, Inc.) supplemented with 5% horse serum and 5% fetal bovine serum. COS-7 cells (ATCC 1651) were grown in DMEM supplemented with 10% fetal bovine serum. LA-N-5 cells, a neuroblastoma cell line that was kindly provided by Dr. Jack Dixon at University of Michigan, were cultured in DMEM medium supplemented with 15% fetal bovine serum.
Expression Plasmids-The DNA fragment containing the vgf promoter and the 5Ј noncoding region upstream of the first methionine (from Ϫ178 to ϩ710) (34) was PCR-amplified from the genomic DNA isolated from PC12 cells. The DNA fragment was then inserted into pGL3Basic vector (Promega) to generate the pvgf-luc plasmid. Mutations changing Tyr 294 , Tyr 404 , or Tyr 451 to Phe were introduced into PTC1 cDNA by site-directed mutagenesis (Muta-gene M13 in vitro mutagenesis kit, Version 2, Bio-Rad). These mutated DNA fragments were then subcloned into the pRcCMV expression vector. 2A/RET, 2B/RET, and HSCR/RET carry mutations Cys-634 3 Arg, Met-918 3 Thr, and Arg-897 3 Glu, respectively (12). RET/PTC1⌬N (49 kDa) had the N terminus of H4 (amino acid residues 3-52) deleted but retained the leucine zipper motif. RET/PTC1⌬Zip (46 kDa) had the leucine zipper motif of H4 (amino acid residues 56 -102) deleted (16). The PathDetect in vivo signal transduction pathway reporting systems containing fusion activator plasmids and a reporter plasmid were purchased from Stratagene. pCH110 (Pharmacia Biotech Inc.) is a plasmid encoding ␤-galactosidase.
Transfection Experiments-For the COS-7 transient transfection assay, cells (1 ϫ 10 6 ) were plated in 100-mm Petri dishes 24 h before transfection. Each plasmid (10 g) was transfected into COS-7 cells using the calcium phosphate transfection method (Life Technologies, Inc.). Twenty-four hours after transfection, fresh growth medium was added, and the cells were grown for an additional 48 h before harvest.
For the vgf promoter assay, PC12 cells (3 ϫ 10 5 cells) were plated in 35-mm Petri dishes 24 h before transfection. Various RET plasmids (500 ng), 1 g of pvgf-luc, and 0.8 g of pCH110 (Pharmacia) were incubated with 40 l of Superfect transfection reagent (Qiagen) in 400 l of DMEM. The total DNA amount used for transfection was 8 g, which was obtained by adding various amounts of salmon sperm DNA (Life Technologies, Inc.). After incubation for 10 min at room temperature, the transfection mixture was then divided into four aliquots and added onto four Petri dishes with plated cells. Three h after transfection, the medium was aspirated, and cells were washed with phosphatebuffered saline once before regular PC12 growth medium was added. Twenty-four hours after transfection, NGF (100 ng/ml) (Boehringer Mannheim) was added to two of the four parallel cultures for 48 h, and the total proteins were then isolated for luciferase assay. All experiments were repeated three times.
To investigate the induction of ELK-, CREB-, or JUN-mediated gene expression by RET in PC12 cells and LA-N-5 cells, PC12 cells or LA-N-5 cells (5 ϫ 10 5 ) were plated in 35-mm Petri dishes 24 h before transfection. Various RET plasmids or pcDNA3 vector (150 ng), 3 g of pFR-luc, 150 ng of chimeric transactivator plasmid, and 360 ng of pCH110 were mixed in 120 l DMEM and incubated with 20 l of Superfect transfection reagent for 10 min at room temperature. The transfection mixture was then divided into three aliquots and added onto three Petri dishes with plated cells. Three h after transfection, the medium was aspirated, and cells were washed with phosphate-buffered saline once before regular growth medium was added. Twenty-four h after transfection, the medium was replaced with DMEM medium containing 0.5% fetal bovine serum. Cells were grown for an additional 24 h before the total proteins were isolated for luciferase assay. The experiment was repeated three times.
Immunoprecipitation and Western Blot Analysis-To ensure the proper expression of these RET proteins from the corresponding DNA constructs, various DNA constructs were transfected into COS-7 cells, and total proteins were extracted for Western blot analysis using anti-body against the cytoplasmic domain of c-RET. COS-7 cells were lysed in TN1 buffer (150 mM NaCl, 25 mM Tris-HCl, pH 8.0, 1% Triton X-100, 20 mM EDTA, 20 mM sodium pyrophospate, 20 mM NaF, 3 mM Na 3 VO 4 , 2 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin). Immunoprecipitation and Western blot analysis were carried out as described previously (12). The antibody against RET used in immunoprecipitation was first conjugated to protein-A beads before lysates were immunoprecipitated by antibody (Ab) against the C terminus of c-RET, and RET expression was detected by Western blot analysis using antibody against the tyrosine kinase domain of c-RET. Tyrosine phosphorylation (PTyr) of these RET proteins was detected by Western blot analysis using antibody against phosphotyrosine (4G10) on the same blot. IP, immunoprecipitation; WB, Western blot analysis. RET/PTC1-HSCR is derived from the long isoform of RET with C-terminal 51 amino acids, whereas other RET/PTC1 DNA constructs are derived from the short isoform of RET with C-terminal 9 amino acids. It has been shown previously that C-9 or C-51 does not affect the tyrosine phosphorylation levels of RET oncoproteins (32). The RET proteins detected are indicated by arrowheads.
use. For Western blot analysis, a monoclonal antibody against c-RET was used. Anti-phosphotyrosine antibody 4G10 is a commercially available monoclonal antibody (Upstate Biotechnology Inc.).

Various RET Proteins Encoded by Wild-type or Mutant RET DNA Constructs Can Be Expressed in COS-7 Cells-Various
forms of RET used in this study are schematically shown in Fig.  1a. All of these RET proteins can be expressed at comparable levels in COS-7 cells (Fig. 1, b and c). RET/PTC1, RET/PTC2, and RET/PTC3 represent the three forms of RET/PTC oncoproteins found in papillary thyroid carcinoma. We demonstrated that mutations changing Tyr 294 , Tyr 404 , or Tyr 451 to  PTC1dN,PTC1⌬N; PTC1dZip, PTC1⌬Zip. c, mutation changing the phosphotyrosine binding site, Tyr 294 (P) or Tyr 451 (P), but not Tyr 404 (P) significantly impairs RET/PTC1 to induce vgf promoter activity. For the above experiments, PC12 cells were transfected with pvgf-luc and pCH110 alone or together with pcDNA3 vector or together with the indicated RET cDNA construct. Twenty-four h after transfection, NGF (100 ng/ml) was added to two of the four parallel transfection plates, and the cells were incubated for an additional 48 h before the total proteins were isolated to determine the vgf promoter activity by luciferase assay (see "Experimental Procedures"). Luciferase activity, normalized by dividing with the corresponding ␤-galactosidase activity, is reported as -fold increases above the basal activity of the pvgf-luc reporter gene in PC12 cells that were transfected in the absence of RET cDNA constructs. The results represent the average -fold activation of at least four transfections. The error bars represent the S.D. from the mean.

The vgf Promoter Activity Can Be Induced by the Expression of RET/PTC, 2A/RET, and 2B/RET in PC12 Cells-The vgf
gene is specifically expressed in neuroendocrine cells upon NGF stimulation (34). It is known that c-RET is also predominantly expressed in cells or tissues of neuroendocrine origin (35)(36)(37). In this study, we showed that ret/PTC1, ret/PTC2, ret/PTC3, 2A/ret, and 2B/ret oncogenes all can induce vgf promoter activity at a rate greater than six times that of the basal activity of pvgf-luc in PC12 cells (Fig. 2a). As expected, pcDNA3 vector and HSCR/ret cannot induce vgf promoter activity. This result suggests that 2A/RET, 2B/RET, and RET/PTC activate common signaling pathways in PC12 cells, leading to the induction of vgf expression regardless of the nature of mutations as well as the cellular localization of these oncoproteins. Therefore, the vgf promoter activity is a good molecular marker to monitor RET activation in PC12 cells.
Consistent with a previous report (30), NGF treatment can only induce vgf promoter activity about three times higher than the basal activity of pvgf-luc in PC12 cells. However, in PC12 cells that express RET oncoproteins, NGF treatment did not significantly further induce vgf promoter activity. This result is consistent with the previous finding that vgf promoter activity induced by RET/PTC1 can not be further induced by NGF treatment (30). This finding suggests that RET activation and NGF stimulation utilize overlapping signaling pathways in PC12 cells. A small increase in the vgf promoter activity was observed in the NGF-treated PC12 cells expressing RET oncoproteins compared with those without NGF stimulation. A possible explanation could be that a subpopulation of cells, which were not transfected by RET, remained fully responsive to NGF treatment. Taken together, these results indicated that vgf promoter activity induced by NGF or RET was not additive nor synergistic in PC12 cells.
The Leucine Zipper Motif Is Essential for RET/PTC1 to Induce vgf Promoter Activity in PC12 Cells-To test whether the leucine zipper motif is important for the induction of vgf promoter activity by RET/PTC1, we co-transfected ret/ PTC1⌬Zip with pvgf-luc into PC12 cells. Unlike RET/PTC1, RET/PTC1⌬Zip can not induce vgf promoter activity (Fig. 2b). As a control, RET/PTC1⌬N, with deletion of the extreme Nterminal amino acid residues but retention of the leucine zipper motif, was able to induce the vgf promoter activity at a similar rate to RET/PTC1. This result indicates that leucine zipper-mediated dimerization is essential for RET/PTC1 to induce vgf promoter activity in PC12 cells.
Phosphotyrosine Docking Sites Tyr 294 (P) and Tyr 451 (P), but Not Tyr 404 (P), Are Important for RET/PTC1 to Induce vgf Promoter Activity in PC12 Cells-We introduced mutations changing Tyr 294 , Tyr 404 , or Tyr 451 to Phe into RET/PTC1 and investigated the effect of these mutations on RET/PTC1-induced vgf promoter activity (Fig. 2c). RET/PTC1 with the mutation changing Tyr 294 or Tyr 451 to Phe was no longer able to induce vgf promoter activity. These results indicate that signaling events mediated by the phosphotyrosine binding sites Tyr 294 (P) and Tyr 451 (P) are essential for RET/PTC1-induced vgf promoter activity in PC12 cells. The mutation changing Tyr 404 to Phe had little effect on the ability of RET/PTC1 to induce vgf promoter activity. This finding is consistent with previous studies showing that PLC␥ is not involved in the neuronal differentiation process in PC12 cells (38).
The importance of these three phosphotyrosine binding sites on the biological effects of RET activation has been studied in other cellular systems. The docking site for Shc or Enigma Tyr 451 (P) appears not only essential for the transforming activity of 2A/RET and 2B/RET in NIH/3T3 cells and the mitogenic activity of RET/PTC2 in mouse fibroblasts but is also essential for RET/PTC1-induced vgf promoter activity in PC12 cells. However, the roles of the docking site for Grb10 or Grb7 Tyr 294 (P) and the docking site for PLC␥ Tyr 404 (P) on the various biological effects induced by RET activation appear to be inconsistent among different assay systems in vitro. To investigate the importance of each phosphotyrosine binding sites in RET/PTC1 oncogenicity in vivo, we are currently establishing transgenic mice lines with thyroid-targeted expression of the RET/PTC1 carrying mutations changing Tyr 404 , Tyr 294 , or Tyr 451 to Phe.
ELK-, CREB-, or JUN-mediated Gene Expression Is Acti-vated by RET in PC12 Cells-A CREB binding site, a JUN binding site, and an SRE element were identified within the vgf promoter region by GenBank TM sequence analysis. It has been shown that CREB and JUN are phosphorylated and activated by protein kinase A and c-Jun NH 2 -terminal kinase, respectively, and the SRE element is the binding site for ELK, which is phosphorylated and activated by MAPK kinase 1 (39,40). This leads us to hypothesize that RET may activate ELK-, CREB-, or JUN-mediated gene expression in PC12 cells. The PathDetect signal transduction reporting system was used to test this hypothesis. In this assay system, the gene to be characterized is co-transfected with a reporter plasmid and a plasmid encoding a pathway-specific transactivator (ELK for MAPK pathway, JUN for c-Jun NH 2 -terminal kinase pathway, and CREB for protein kinase A pathway) fused to GAL4 binding domain. If expression of the gene to be characterized results in direct or indirect phosphorylation of the ELK, JUN, or CREB fusion protein, luciferase expression will be activated above the background level. As shown in Fig. 3, ELK-mediated gene expression was activated by RET/PTC1, 2A/RET, and 2B/RET more than 50 times higher than that activated by pcDNA3 vector. As expected, HSCR/RET could not activate ELK-mediated gene expression. This result is consistent with a previous report showing that c-RET activation upon glial cell line-derived neurotrophic factor stimulation activates the MAPK pathway in a neuroblastoma cell line (28). MAPK kinase 1, which directly phosphorylates ELK, could induce ELK-mediated gene expression about 1.6 times higher than RET/PTC1 did. RET/ PTC1 carrying a mutation abolishing the phosphotyrosine binding site Tyr 451 (P) or Tyr 294 (P) did not significantly activate ELK-mediated gene expression. This result indicates that Tyr 451 (P)-and Tyr 294 (P)-mediated signaling cascades involve MAPK activation.
As shown in Fig. 4, RET/PTC1 as well as 2A/RET and 2B/ RET, but not HSCR/RET, could induce gene expression mediated by CREB. This is the first report showing that RET activation can induce CREB-mediated gene expression. It has been reported that treatment of PC12 cells with NGF, epidermal growth factor, or 12-O-tetradecanoylphorbol-13-acetate induces CREB phosphorylation at Ser 133 through the activation of a CREB kinase, RSK2, via Ras-MAPK pathway (41). Therefore, we speculate that RET may activate CREB-mediated gene expression through a similar mechanism. CREB-mediated gene expression induced by the positive control protein kinase A, which directly phosphorylates CREB, was 1.6 times higher In this study, LA-N-5 cells were transfected with pFR-luc reporter, pCH110, together with pcDNA3 vector or the indicated cDNA constructs. In addition, pFA-ELK, pFA-CREB, or pFA-JUN was co-transfected in the corresponding ELK, CREB, or JUN system. Luciferase activity was normalized by dividing with the corresponding ␤-galactosidase activity. For clearer presentation, the promoter activity was reported as 1,000-fold values of the normalized luciferase activity. Each transfection was done in triplicate. The error bars represent the S.D. from the mean. MEK1, MAPK kinase 1; MEKK, MAPK kinase kinase. than that of RET/PTC1. Similar to ELK-mediated gene expression, CREB-mediated gene expression could not be activated by  RET/PTC1 carrying a mutation abolishing phosphotyrosine  binding site Tyr 451 or Tyr 294 . Finally, we also demonstrated that RET/PTC1 as well as 2A/RET and 2B/RET could induce JUN-mediated gene expression in PC12 cells through the signaling events mediated by phosphotyrosine binding sites Tyr 451 (P) and Tyr 294 (P) in RET/ PTC1 (Fig. 5). This is in contrast to the study in a neuroblastoma cell line demonstrating that c-Jun NH 2 -terminal kinase pathway was not able to be induced by the activation of c-RET through glial cell line-derived neurotrophic factor stimulation (28). It is possible that the promoter analysis assay system we used in this study is more sensitive than the biochemical characterization assay system used in the study of neuroblastoma cells. Indeed, in PC12 cells, JUN-mediated gene expression induced by RET was much lower than that induced by MAPK kinase kinase, which directly phosphorylates and activates c-Jun NH 2 -terminal kinase. Alternatively, this discrepancy might suggest that the activation of c-Jun NH 2 -terminal kinase pathway by RET is restricted to PC12 cells.
ELK-, CREB-, or JUN-mediated Gene Expression Is Activated by RET in PC12 Cells-As a comparison to PC12 cells, we also transfected RET/PTC1 cDNA constructs into a neuroblastoma cell line, LA-N-5 cells. Consistent with the published report (28), RET/PTC1 could significantly induce ELK-mediated gene expression in LA-N-5 cells (Fig. 6). However, in contrast to the study carried out in PC12 cells, neither CREBnor JUN-mediated gene expression could be induced by RET/ PTC1 in LA-N-5 cells (Fig. 6). Interestingly, our study showed that protein kinase A could not induce CREB-mediated gene expression in LA-N-5 cells. These results indicate that the stimulation effect of RET on ELK-, CREB-, or JUN-mediated gene expression could be different in different cells.
Concluding Remarks-Activation of the MAPK pathway by RET in neuronal cells has been demonstrated by both promoter analysis and detailed biochemical characterization. However, our study is the first report showing that RET activation can induce CREB-mediated gene expression by promoter analysis. Furthermore, we showed that RET activation in PC12 cells could activate JUN-mediated gene expression, despite the fact that our results and those of others showed that JUN-mediated gene expression was not induced by RET activation in LA-N-5 cells. To fully understand how RET activates CREB-or JUNmediated gene expression in PC12 cells, detailed biochemical characterization of the RET signaling pathways that lead to activation of CREB or JUN in PC12 cells is warranted. The information obtained could provide significant insights into the molecular mechanisms underlying the predisposition of RET activation in the development of MEN2 diseases.