Abrogation of nerve growth factor-induced terminal differentiation by ret oncogene involves perturbation of nuclear translocation of ERK.

Oncogenic variants of the receptor tyrosine kinase, Ret, cause formation of tumors of neuroendocrine derivation in the multiple endocrine neoplasia type 2 and, thus, likely interfere with antiproliferative and/or differentiative extracellular signals. Here we took advantage of two rat pheochromocytoma-derived cell lines (PC12/MEN2A and PC12/MEN2B) to investigate whether Ret-induced nerve growth factor (NGF) unresponsiveness might involve impairment of ERK signaling. In fact, these cells, stably transfected with distinct forms of the active ret oncogene, fail to block proliferation, even upon NGF stimulation. In these cells we show the presence of both chronic ERKs activity and high expression levels of MKP-3, an ERK-specific phosphatase. Despite the presence of MKP-3, ERK activity can be further stimulated by NGF, but it fails to translocate into the nucleus and consequently to induce immediate-early gene transcription. Because of the presence of MKP-3, our results suggest the existence of a negative regulatory feedback acting on ERKs as a mechanism responsible for the abrogation of NGF-induced terminal differentiation. Indeed, MKP-3 seems to be implicated in the persistence of ERKs in cell cytoplasm. This interpretation is further supported by the observation that in ret-transfected cells, forced expression of an active form of MEK-1 may overcome this block; it restores transcription from the c-fos promoter, induces translocation of ERKs into the nucleus, and inhibits cell proliferation.

Ret is a receptor tyrosine kinase whose expression is restricted to neuronal cells of the central and peripheral nervous system. Recently four ligands have been described for this receptor as follows: glial cell-derived neurotrophic factor, neurturin, artemin, and persephin (1).
Germ line mutations of the receptor tyrosine kinase, Ret, are responsible for the multiple endocrine neoplasia (MEN) 1 type 2A and 2B syndrome and for the familial medullary thyroid carcinoma (2)(3)(4)(5). MEN-2A and MEN-2B are distinct hereditary neoplastic syndromes both characterized by the presence of medullary thyroid carcinomas and pheochromocytomas (6,7). Missense mutations at one of five cysteine residues (Cys-609, -611, -618, -620, and -634) clustered in the extra cytoplasmic domain of ret are the most frequent causative genetic events of familial medullary thyroid carcinoma and the MEN-2A syndrome (8). A single point mutation, which results in a Thr for Met substitution at codon 918 within the Ret catalytic domain, is responsible for the MEN-2B syndrome (8). These mutations convert ret into a dominant transforming gene and cause constitutive activation of its intrinsic tyrosine kinase activity, although their mechanism of activation differs (8 -10).
In MEN-2 syndromes, the molecular mechanisms by which the mutated Ret contribute to the development of neuroendocrine neoplasms remain largely unknown. Indeed, the inheritance of mutated ret alleles implicates them in the pathogenesis of a generalized hyperplasia of the entire population of thyroid C-cells and of the adrenal medulla chromaffin cells (11,12). In this study we investigate whether a biological mechanism by which Ret mutants participate in neoplastic progression in MEN-2 syndromes might involve the unresponsiveness of neuroendocrine cells to extracellular growth inhibitory signals.
Many polypeptide growth factors activate sequentially the components of the signal transduction cascade, which includes activation of Ras, Raf, the mitogen-activated protein kinase (MAPK) kinase (MEK), and the extracellular signal-regulated kinases, ERK-1 and ERK-2 (also called p42/p44MAPK) (13). MAP kinase activation is a crucial step of several cellular processes, including proliferation, differentiation, and long term potentiation (14 -17). In turn, the activity of MAP kinases is modulated by several dual specificity MAP kinase phosphatases (DSP). Among them, the MKP-3, which is mainly found in cell cytoplasm, is involved in modulating nuclear translocation of the kinase (18). In PC12 cells terminal differentiation is mediated by the tyrosine phosphorylation of the NGF receptor and leads to the persistent activation and subsequent nuclear translocation of ERKs (Refs. 19 and 20 and for a review see Ref. 21). Although many aspects of the ret signaling pathway have been elucidated, there is no direct evidence that ERK activation is involved in the transforming mechanism of Ret. Moreover, acute stimulation of ret in neuroblastoma cell lines and primary cultures from neuroendocrine tumors, but not in NIH-3T3 fibroblasts, results in ERK phosphorylation (22)(23)(24)(25)(26).
To investigate further the mechanisms by which the ret oncogene causes uncontrolled cell proliferation in neuroendocrine tumors, we addressed the question of whether the retinduced ERK cascade might be involved in abrogating terminal differentiation. We used two recently established stable isolates of the PC12 cells that express the active Ret variants, Ret C634Y (PC12/MEN2A) and Ret M918T (PC12/MEN2B). Expression of the ret-active mutants induces the PC12 cells to become partially differentiated but incapable of undergoing terminal differentiation, even upon NGF stimulation (27). Thus, because a major consequence of NGF-induced differentiation is inhibition of cell proliferation, we investigated whether abrogation of NGF responsiveness in ret-transfected PC12 cell lines might involve an impairment in the nuclear signal transmission through ERK kinases.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection Experiments-PC12 cells were grown in RPMI 1640 (Life Technologies, Inc.) supplemented with 2 mM Lglutamine, 10% horse serum, and 5% fetal calf serum. PC12/MEN2A and PC12/MEN2B, which are PC12-derived cell lines that express the human RET9 isoform with the C634Y and the M918T mutation, respectively, were grown as described previously (27). For transient transfection assays, cells were plated at 3 ϫ 10 5 cells in a 60-mm diameter tissue culture dish 24 -36 h before transfection. Transfection experiments were performed using the Lipofectin reagent according to the manufacturer's instructions (Life Technologies, Inc.), as previously reported (28). Transient transfections were carried out with 2 g of reporter plasmid, pfos-CAT (Ϫ356 to ϩ109) (29), together with increasing amounts (0, 2.5, 5, and 7.5 g) of activated MEK-1 (N3-S218E-S222D) mutant (30). The same DNA concentration was reached by adding various amounts of the control vector. Nerve growth factor 2.5 S (Upstate Biotechnology Inc.) (100 ng/ml) was added to the culture medium as indicated.
Chloramphenicol Acetyltransferase Assays-Cell extracts were prepared 72 h after transfection, and CAT activity was analyzed by thin layer chromatography with 95% chloroform, 5% methanol, as described previously (28). Each experimental point was cut from the thin layer chromatography plate and counted. For each independent experiment, the percentage of conversion to acetylated [ 14 C]chloramphenicol was calculated and normalized for the transfection efficiency. Values from three independent experiments, each made in duplicate, were used to calculate standard deviation and were plotted on an arbitrary scale as relative promoter induction. By staining cells with anti-HA antibodies, in parallel transfection experiments with HA-MEK-1, we assessed that the number of Ha ϩ cells was similar in PC12/MEN2A and PC12/MEN2B. Northern Blot Analysis-RNA was prepared from cultured cells using a guanidine thiocyanate-based reagent (Ultraspec RNA Isolation System, Biotecx) as recommended by the manufacturer. 20 g of total RNA per lane from each sample was size-fractionated with 1% agarose formaldehyde denaturing gel electrophoresis and blotted onto nylon filters (Hybond-N, Amersham Pharmacia Biotech). To obtain the Krox-24, MKP-3, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probes, short specific fragments were synthesized by reverse transcriptase-polymerase chain reaction. The vgf probe was excised from the pV2-2 plasmid (31).The random oligonucleotide primer kit (Amersham Pharmacia Biotech) was used for 32 P labeling of the vgf Krox24, MKP-3, and GAPDH probes. Hybridization and washing were carried out under stringent conditions as follows: 0.1ϫ SSC, 0.1% SDS, 60°C. Autoradiography was performed using Kodak X-AR films at Ϫ70°C for 1-7 days with intensifying screens.
Kinase Assay-The untreated and treated PC12 cell lines were lysed with 1% Nonidet P-40 lysis buffer containing 10 mM Tris, pH 8, 150 mM NaCl, 0.4 mM EDTA, 0.1 mg/ml PMSF, 2 g/ml leupeptin, 2 mM NaOV, 10 mM NaF, 10 mM sodium pyrophosphate, and 2 g/ml aprotinin. Protein concentrations were estimated using a modified Bradford assay (Bio-Rad). 400 g of proteins from each sample were immunoprecipitated with anti-ERK 1 rabbit polyclonal antibodies (C-16, Santa Cruz Biotechnology) at 4°C for 2 h. The immunoprecipitates were washed once with lysis buffer and twice with assay buffer and were then assayed for kinase activity by incubating with 8 g of myelin basic protein (MBP) and 8 Ci of [␥-32 P]ATP in 30 l of assay buffer containing 20 mM Hepes, pH 7.5, 2 mM NaOV, 10 mM magnesium acetate, 0.1 mg/ml PMSF, 2 g/ml aprotinin, 2 g/ml leupeptin, 100 M ATP for 30 min at 30°C. Reactions were terminated by the addition of 30 l of SDS buffer, and proteins were separated by 14% SDS-polyacrylamide gel electrophoresis. 32 P-Labeled bands were revealed by autoradiography of the dried gel and quantified on a PhosphorImager (Molecular Dynamics).
Immunofluorescence-Cells were seeded at low confluence on glass coverslips coated with poly-L-lysine (15 g/ml) (Sigma), serum-starved for 16 h, and then stimulated with NGF for 1 h. Cells were washed twice with PBS containing 0.9 mM calcium and 0.5 mM magnesium chloride (PBS/CM), fixed for 10 min in PBS containing 3% paraformaldehyde (Serva) and 2% sucrose (Sigma), and then treated with cold (Ϫ20°C) mixture 1:1 methanol acetone for 1 min, washed three times with PBS/CM, and incubated in humidified atmosphere at room temperature with polyclonal antibody anti-ERK-1 (Santa Cruz C-16) diluted 1:100 in PBS/CM containing 0.2% gelatin (Sigma) and 0.075% saponin (Sigma) (PBS/CM/GS) for 1 h at room temperature. Cells were then washed three times with PBS/CM and incubated in humidified atmosphere at room temperature with secondary goat anti-rabbit rhodamine-conjugated antibody (Jackson ImmunoResearch) diluted 1:30 in PBS/CM/GS. As a negative control, cells were also stained with goat serum alone (data not shown). The cells were then washed three times with PBS/CM and for nuclear staining (2Ј-[4-hydroxyphenyl]-5-[4-methyl-1-piperazinyl]2, 5Ј-bi-1H-benzimidazole) Hoechst-33258 (Sigma) 0.5 g/ml dissolved in PBS was used. The cells were washed three times with PBS/CM and finally mounted in 50% glycerol in PBS and analyzed with a Zeiss Axiophot epifluorescence microscope.
Cell Fractionation-PC12, PC12/MEN2A, and PC12/MEN2B cells were grown overnight in serum-free medium, treated with NGF (100 ng/ml) for 50 min at 37°C, washed twice with ice-cold PBS buffer, scraped in Nonidet P-40 lysis buffer (10 mM Hepes, pH 7.9, 1 mM EDTA, 60 mM KCl, 0.2% Nonidet P-40, 1 mM dithiothreitol, 1 mM PMSF, 1 mM NaOV, 10 g each of aprotinin and leupeptin per ml), and left on ice for 15 min. The cells were then centrifuged for 5 min at 2,500 rpm at 4°C. The supernatant consisting of the cytosol and most of the plasma membrane was carefully removed, transferred to a fresh tube, and centrifuged for 15 min at 14,000 rpm at 4°C. The nuclei-containing fraction was washed in 1 ml of lysis buffer except Nonidet P-40, collected by centrifugation, and resuspended in 300 l of the same buffer. The nuclei suspension was then added carefully on 300 l of 30% sucrose cushion (1:1 solution of 60% sucrose/lysis buffer 2ϫ without Nonidet P-40) and centrifuged at 6000 rpm for 10 min at 4°C. The supernatant was removed, and nuclei were lysed in nuclear resuspension buffer (250 mM Tris-HCl, pH 7.8, 60 mM KCl, 1 mM dithiothreitol, 1 mM PMSF, 1 mM NaOV, 10 g each of aprotinin and leupeptin per ml) with 3 cycles of freeze/thawing and then centrifuged at 9500 rpm for 15 min at 4°C. Protein concentrations were estimated by a modified Bradford assay (Bio-Rad). 20 g of each fraction were run on SDS-7.5% polyacrylamide gel under reducing conditions before transfer to polyvinylidene difluoride filter. The protein blot was probed overnight at 4°C with anti-pERK antibody (New England Biolabs) diluted 1:2000 in TBS, 0.05% Triton, and 0.5% Non-fat Dry Milk (NFDM). Normalization for cytoplasmic protein was with the anti-␤-tubulin antibody (Sigma) diluted 1:1000 in TBS, 0.05% Tween 20, and 5% NFDM and for nuclear proteins with the anti-CREB antibody (Upstate Biotechnology Inc.) diluted 1:1000 in TBS, 3% NFDM.

Microinjection of Cells and 5-Bromo-2-deoxyuridine (BrdUrd)
Incorporation-PC12/MEN2A cells were seeded at low confluence on glass coverslips coated with poly-L-lysine (15 g/ml). The plasmid, HA-tagged activated MEK-1 (N3-S218E-S222), was injected into cell nuclei at a concentration of 50 ng/l using a microinjection system (Zeiss). 48 h later, cells were washed twice with PBS and fixed for 10 min in PBS containing 10% paraformaldehyde at room temperature and then permeabilized in 100% methanol for 10 min at Ϫ20°C. Following a blocking step of 1% bovine serum albumin in PBS, cells were incubated for 1 h in humidified atmosphere with 2.5 g/ml of PY-204-ERK antibodies (Promega), washed three times with 1 mg/ml bovine serum albumin and 0.05% Nonidet P-40 in PBS, and incubated with secondary goat antirabbit rhodamine-conjugated antibody. After several washes in PBS, coverslips were incubated with monoclonal antibody anti-HA (clone 12CA5, Roche Molecular Biochemicals) whose working dilution was 1:80, for 1 h, washed twice with PBS, and incubated with secondary anti-mouse fluorescein conjugate antibody. Finally, nuclei were stained with Hoechst-33258, mounted, and analyzed with a Zeiss Axiophot epifluorescence microscope. BrdUrd (100 nM) was added to the culture medium for 3 h before cells were fixed; anti-BrdUrd mouse monoclonal antibody was used to detect the fraction of cells in S phase.

In the ret-transfected Cells NGF Poorly Stimulates Neuronal
Gene Expression-NGF induction of PC12 cell differentiation leads to persistent activation of ERKs and involves the expression of a complex pattern of genes, including immediate-early (fos and Krox-24) and delayed response genes (vgf, SCG10, and peripherin) (21). To assess NGF signaling in PC12/MEN2A and PC12/MEN2B, we stimulated the ret-transfected cells with NGF for different times. In agreement with our previous study (27), basal levels of vgf transcript were frankly higher in PC12/ MEN2A and PC12/MEN2B than in the parental cells. We now demonstrate that vgf transcript levels remained almost unchanged after NGF stimulation (Fig. 1). Moreover, the transcript of the immediate-early gene Krox-24 was present in unstimulated ret-transfected cells, only barely induced by NGF in PC12/MEN2A cells, and uninduced in PC12/MEN2B cells ( Fig. 1, middle panel). The unresponsiveness to NGF stimulation of the immediate-early gene transcription was confirmed by similar Northern blot experiments in which a c-fos-specific probe was used (see below and Fig. 5A).

Expression of Active ret Mutants Induces Constitutive Activation of ERK-
We thus determined whether NGF stimulation induced ERK activity in ret-transfected cells. First, we assessed ERK activity in the absence of extracellular stimulation. We used MBP as an exogenous substrate to measure MAP kinase activity in PC12/MEN2A and PC12/MEN2B cells. NGF stimulation of PC12 cells results in rapid activation of the endogenous ERKs (16 -20). Consistent with this, in the parental cells basal levels of ERKs (ERK-1 and ERK-2, see also the legend to In each experiment we verified that the same amounts of ERK were used in the kinase assay. Proteins from each of the immunoprecipitates were separated on 11% SDS-PAGE and immunoblotted with the same anti-ERK-1 antibody. Immunoblots confirmed that the same amount of ERK was used in each kinase assay (not shown). Under our experimental conditions the ERK immunoprecipitates consisted mainly (approximately 80%) of ERK-1/ p44 and to a lesser extent of ERK-2/p42, as verified by immunoblot with the same anti-ERK-1 antibody (not shown). We next determined whether NGF was capable of further stimulating ERK activity in ret-transfected cell lines. As shown in Fig. 2B, upon stimulation, the kinase activity rapidly increased in both PC12/MEN2A and PC12/MEN2B cell lines and reached a peak at 5 min. This maximal stimulation (nearly 70%) was lower than that found in the parental PC12 cells under the same conditions. To assess whether failure of the ret-transfected cells to respond to NGF might be caused by rapid inactivation of the enzyme, we measured ERK activity after 30 min and 2 h of stimulation. Consistent with previous reports (16), in the PC12 parental cells the kinase activity was persistent and still high after 30 min, and 2 h (86%), upon NGF stimulation. In both ret-transfected cell lines, either carrying a Ret-2A or a Ret-2B mutation, ERK activity decreased more rapidly than parental cells, reaching nearly steady levels within 2 h of stimulation (49 and 35%, respectively) (Fig. 2B).
NGF Did Not Trigger ERK Translocation into the Nucleus-Because, in PC12 cells, NGF-induced terminal differentiation has been correlated to translocation of ERKs from the cytoplasm into the nucleus (16 -20), we investigated the subcellular re-localization of ERK upon NGF stimulation in PC12/MEN2A and PC12/MEN2B cells, and we compared the results with findings obtained in the parental PC12 cells. By using immunofluorescence staining, we examined the intracellular distribution of ERK before stimulation and after 10, 30, and 60 min of NGF stimulation. In the parental cells, immunoreactivity was mainly restricted to the cytoplasm (Fig. 3A). After NGF treatment, part of the immunoreactivity slowly translocated into the nuclear compartment. The entire cell body was homogeneously stained at 30 min (not shown), and staining reached a maximum at 60 min (Fig. 3D). Similarly, in the ret-transfected cells, ERK immunoreactivity was mainly localized in the cytoplasm (Fig. 3, B and C). The addition of NGF did not change the distribution of immunoreactivity at 10, 30 (not shown), and 60 min (Fig. 3, E and F), indicating that NGFinduced ERK translocation was impaired in these cells.
On the other hand, since in the ret-transfected cells the immediate-early Krox-24 and the vgf genes were constitutively expressed (see Fig. 1 and Ref. 27), we asked whether a fraction of active ERK molecules was located in the cell nuclei, even in absence of extracellular NGF simulation. Thus, we determined the distribution of the phosphorylated ERKs in the soluble and nuclear fractions. The ret-transfected and parental PC12 cells were treated with NGF for 60 min, and the nuclei were fractionated on sucrose cushion. Nuclear proteins were analyzed by immunoblot with an antibody that specifically recognizes the tyrosine 204-phosphorylated form of ERK1/2 but not the ERKs that are unphosphorylated at this site (anti-ERK PY-204). As shown in Fig. 3G, the nuclear fraction of PC12 cells contains little detectable PY-204-ERKs, which accumulates in the nuclei upon NGF stimulation. Conversely, the nuclear fractions from either ret-transfected cell lines were highly immunoreactive with anti-PY-204-ERK antibodies. Moreover, in good agreement with the results from immunofluorescence experiments, upon NGF stimulation no further induction of immunoreactivity was observed in the nuclear fraction.
These results, taken together, indicate that mutant Ret proteins are able to induce the persistent accumulation of active ERK molecules in nuclei but prevent further NGF-induced ERK translocation.
Despite the fact that a portion of P-ERK molecules is localized in cell nuclei, the induction of immediate-early genes expression was prevented ( Fig. 1 and Fig. 6A). This indicates that other transducing mechanisms, necessary for immediate-early expression, may be affected in PC12/MEN2A and PC12/ MEN2B. On the other hand, we cannot exclude that some transcription takes place from the endogenous c-fos promoter as suggested by the results of the CAT experiment (Fig. 6B). In this case the absence of c-fos transcript could be explained by the short half-life of this immediate-early gene mRNA.
In PC12 cells the CREB kinase activity depends on the Ras/ERK pathway and contributes to immediate-early gene expression. As shown in Fig. 4 (upper panel) by immunoblot-ting with antibodies specific for P-Ser-133-CREB, we investigated the ability of NGF to stimulate the phosphorylation of CREB. In parental as well as in ret-transfected PC12 cells, 50 M forskolin induced comparable levels of phospho-CREB (Fig.  4, lower panel). However, NGF was unable to stimulate CREB phosphorylation in ret-transfected cells. These results confirm that signaling through Ras/ERK pathway is highly impaired in ret-transfected cells, whereas protein kinase A-mediated signaling seems not to be affected.
Expression of MAPK Phosphatase 3 Is Up-regulated in rettransfected Cells-A recently identified member of the dual specificity MAPK phosphatase family (DSP), MKP-3, is involved in preventing translocation of the kinase into the nucleus (18,(32)(33)(34). MKP-3 is highly specific for ERK1/2, and, unlike the other members of the DSP family, is mainly found in the cell cytoplasm. We wondered whether in ret-transfected cells the impairment of NGF-induced nuclear translocation of ERK and terminal differentiation might be caused by deregulation of MKP-3 expression. Consistent with previous results (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33), the MKP-3 transcripts were undetectable in parental PC12 cells and reached maximum levels of induction after 3 h of NGF stimulation (Fig. 5A, first 3 lanes). Instead, in the rettransfected cells, the steady-state levels of MKP-3 transcripts and protein were clearly up-regulated to an extent even higher than that reached in the PC12 cells upon NGF stimulation (Fig.  5A, compare 3rd lane with 4th and 7th lanes, and Fig. 5B).
On the other hand, if the presence of chronic MKP-3 activity is involved in the regulation of the steady levels of ERKs, inhibition of MKP-3 should result in more pronounced ERK activity. Thus, we treated all cell lines with (0.5 mM) sodium orthovanadate, a potent inhibitor of many tyrosine phosphatases, including MAP kinase phosphatases (35), and we measured ERK activity at different times. As shown in Fig. 5C, 10 min of sodium orthovanadate treatment was sufficient to strongly induce ERK activity in both PC12/MEN2A/MEN2B cells. As stated before, inhibition of phosphatase activity by orthovanadate is not specific for the ERK-specific phosphatase, MKP-3. Thus, these results, even though consistent with the involvement of MKP-3 in regulating ERK levels in ret-transfected cells, does not exclude that other tyrosine phosphatases may contribute in determining ERK levels of activation. Even though a likely interpretation for the presence of high constitutive levels of MKP-3 relies on the existence of a regulatory feedback (see Fig. 8). In the ret-transfected cells (Fig. 8, B-D) the chronic activation of ERKs is responsible for the expression of ERK phosphatases, which in turn, by dephosphorylating ERK molecules, maintain their kinase activity at intermediate levels. If this is the case, inhibiting ERK activity should result in depression of MKP-3 transcription. Consistently, upon overexpression of Ras (Asn-17), which depressed ERK activity, MKP-3 transcripts rapidly fall to low levels (Fig. 5D). We do not know whether the low basal levels of MKP-3 transcripts, in Ras p21 Asn-17 cells, reflect a clonal difference in the cell lines or the presence of low, non-induced levels of dominant negative Ras (Califano et al. (47)).
These results indicate that the ERK steady-state levels, in ret-transfected cells, might result from the balance between the chronic activation of the Ras-MAPK pathway (see Fig. 2) and the presence of enhanced levels of MKP-3 and, eventually, of other MAP kinase phosphatases (DSP).
Rescue of c-fos Promoter Induction after MEK-1 Transfection-We thus addressed the question of whether the steady presence of active ERKs and MKP-3 is affecting the ability of ERKs to transmit fully NGF signaling or, alternatively, whether other unidentified biochemical events are responsible for impairment of the ERK cascade. Next we asked if a more pronounced stimulation of the kinase activity could overcome this block and so transmit signaling into the nucleus.
An early consequence of activation of the ERK pathway in NGF-stimulated PC12 cells is the expression of the immediateearly genes, including c-fos (21). The c-fos transcript levels were undetectable in both ret-transfected cell lines and were insensitive to NGF stimulation. In fact, stimulation did not cause induction (no more than 1.5-fold over basal levels) in ret transfectants compared with parental cells (Fig. 6A, compare lanes 2  to lanes 4 and 6). Because of the low basal transcript levels of the c-fos in ret-transfected PC12 cells, we decided to use the fos gene as a tool to monitor nuclear transmission of the MAPK cascade.
Therefore, we attempted to rescue the MAPK-dependent induction of c-fos transcription by inducing high levels of MAPK activity in PC12/MEN2A and PC12/MEN2B. To do so, we used a construct in which the chloramphenicol acetyltransferase (CAT) gene was under the transcriptional control of the c-fos promoter, the fosCAT (Ϫ356 to ϩ109) reporter plasmid (29). The fosCAT construct was transfected either alone or with increasing amounts of a constitutively active form of MEK-1 (N3-S218E-S222D) (30). In the parental PC12 cells, either NGF stimulation or the expression of MEK-1 (N3-S218E-S222D) was sufficient to induce fosCAT activity. On the other hand, in PC12/MEN2A and PC12/MEN2B cells, the expression of MEK-1 (N3-S218E-S222D) stimulated the fos promoter (Fig.  6B), whereas NGF did not (data not shown), as already shown for the endogenous gene (Fig. 6A). It is noteworthy that the extent of stimulation in PC12/MEN2A cells was clearly less than in PC12/MEN2B cells (approximately 60% of conversion in the case of PC12/MEN2B and less than 20% in the case of PC12/MEN2A) (Fig. 6B). The presence in the ret-transfected cells, but not in parental PC12 cells, of low, but meaningful, basal levels of fosCAT activity, is probably due to the constitutive activation of ERK present in these cells (Fig. 2). Furthermore, we also show that the expression of MEK-1 induced translocation of ERK molecules, as assessed by using immunofluorescence staining with anti-ERK antibodies (data not shown).

MEK-1 Overcomes the Ret-induced Block and Induces
Terminal Differentiation of PC12/MEN2A-As shown above, the expression of an active MEK-1 in PC12/MEN2A cells induced the fosCAT promoter but to a lesser extent than in PC12/MEN2B cells (Fig. 6B). Thus, we determined whether this reduced ability to stimulate fosCAT expression may reflect a more general lack of responsiveness of the PC12/MEN2A cells to MEK-1.
We first assessed whether expression of the active form of MEK-1 was capable of inducing phosphorylation of ERKs. By microinjection, we introduced the MEK-1 (N3-S218E-S222D) expression plasmid in PC12/MEN2A cells, and 48 h later, by immunofluorescence staining, we examined immunoreactivity for the PY-204-ERKs. Although all cells showed a basal immunoreactivity, the cells microinjected with the active MEK-1 were easily distinguishable because of the clearly higher levels of immunoreactivity for PY-204-ERK antibodies (Fig. 7A and data not shown). As shown in Fig. 7B, almost all cells expressing the tagged MEK-1 (HA) were highly immunoreactive for PY-204-ERK antibodies (column HA ϩ ), and conversely, the large majority of cells highly immunoreactive for anti-PY-204-  1 and 2). PC12, PC12/MEN2A, and PC12/ MEN2B cells were either stimulated with NaOV (0.5 mM) for 10 min or left untreated. Cells lysates from each cell line were prepared, and ERK ERK antibodies expressed the tagged MEK-1 (not shown). On the other hand, a minority of cells not microinjected (approximately 6.4%) was highly immunoreactive for PY-204-ERK antibodies (column HA Ϫ). We next determined whether forced expression of MEK-1 (N3-S218E-S222D) in PC12/MEN2A was able to induce these cells to terminally differentiate. We proceeded to measure the incorporation of bromodeoxyuridine in replicating cells either in presence or in absence of the active MEK-1. Since nearly all cells highly immunoreactive with the PY-204-ERK antibodies were also immunoreactive with the HA tag of the MEK-1 construct (Fig. 7A), we assumed that the strong positivity for PY-ERK reflects the expression of the exogenous MEK-1, and thus, the antibodies against PY-204-ERK were used to localize cells that express the exogenous active MEK-1. As shown in Fig. 7B (column pERK ϩ), only a minority of cells highly immunoreactive with the PY-204-ERK (thus expressing the microinjected MEK-1 construct) efficiently incorporated bro-modeoxyuridine (approximately 5% of PY-204-ERK-positive cells) as compared with the control population (approximately 50% of cells). These results, taken together, suggest that the inability of ret-transfected cells to undergo growth arrest even upon NGF stimulation might depend on MAP kinase activation.

DISCUSSION
The fact that Ret active mutants cause uncontrolled cell proliferation in neuroendocrine tumors, associated with MEN-2 syndromes, is difficult to reconcile with the differentiating effects observed when the same mutants are overexpressed either in PC12 or in human neuroblastoma cells (25, 26, 36 -38). In this study, we report data that implicate chronic activation of the ERK cascade as a mechanism for Ret-induced uncontrolled proliferation of neuroendocrine cells. We took advantage of stable cell lines that reproduce in vitro some of the biological events caused by Ret in human tumors, including unlimited growth. In these cell lines, obtained by expressing the ret oncogene in the PC12 cells, the resulting phenotype is dependent on the ret oncogene and is associated with active proliferation (27,and Califano et al. (47)). Indeed, expression of the ret-active variants, driven by a strong promoter, may cause terminal differentiation of PC12 cells (as assessed by long neurite outgrowth and growth arrest). On the other hand, expressing ret mutants at physiological levels by less efficient promoters cause PC12 cells to change morphology and to express neuronal genes. However, differentiation is not terminal (because of absence of long neurites and of sustained cell proliferation) (27). Proliferation enables us to establish stable cell lines at high frequency. The present results indicate that in these cell lines the presence of high expression levels of the ERK-specific phosphatase, MKP-3, might be implicated in the lack of terminal differentiation. Moreover, even though NGF caused additional stimulation of ERK activity, it elicited neither further ERK translocation into the nucleus nor the expression of immediate-early genes. In PC12 cells, neuronal terminal differentiation is preceded by the coordinate expression of a complex pattern of genes, including Krox-24 and vgf transcripts, whose expression is dependent on the Ras pathway (21,39). In good agreement with previous reports of activation of this cascade by acute stimulation of Ret in neuroblastoma and in motor neuron cells (22)(23)(24)40), here we show that ERKs are constitutively active in the PC12/MEN2A and PC12/MEN2B cells with steady levels ranging between three and four times over the background. Moreover, we show that a portion of ERKs is phosphorylated in tyrosine residues, and even though mainly located in the cytosol, is also present in the nuclear compartment. This is well supported by the presence of both Krox-24 and vgf transcripts as indicative of chronically active Ras-MAPK cascade (Ref. 27,and Califano et al. (47)). However, although in ret-transfected cells the activation of this pathway elicits neuronal gene expression, it is not sufficient for further progression toward terminal differentiation, as assessed by active cell proliferation and absence of long neuritic processes (27).
Recently, a number of findings implicate NGF as an antiproliferative factor for tumor or cell lines of neuroendocrine origin (41). Indeed, loss of NGF production is associated with neoplastic progression and poor prognosis (42). Here we show that in ret-transfected cells, NGF stimulation induced a rapid increase of ERK activity, which reached a peak value after 5 min and rapidly declined. However, while NGF stimulation of the PC12 parental cells resulted in induction of immediate-early gene transcription (c-fos and Krox-24) and in nuclear translocation of the enzyme, high levels of ERK activity, induced by NGF, in the ret transfectants induced neither expression of c-fos nor further nuclear translocation of the enzyme.
Expression of the immediate-early gene c-fos is rapidly induced in PC12 cells by a variety of extracellular stimuli, including NGF (21,43). The NGF signal is, in fact, transmitted into the nucleus and stimulates c-fos transcription by means of at least two Ras-dependent protein kinases as follows: ERK, which regulates the activity of transcription factor TCF/Elk-1, and a CREB kinase, which regulates the activity of CREB (29). Indeed, the fact that, in the ret transfectants, NGF induced neither CREB phosphorylation nor c-fos expression indicates that in these cells the NGF-induced activation of ERK may be not sufficient to transmit fully NGF signaling into the nucleus. Whether other ERK-independent mechanisms contribute to the lack of immediate-early gene expression remains to be determined. Moreover, the failure of NGF to induce further ERKs translocation indicates that the lack of its nuclear relocalization might be implicated in NGF unresponsiveness. ERK1/2 are activated by phosphorylation on both threonine and tyrosine residues by dual-specific kinases (MEK-1 and -2) (44). Conversely, inactivation of ERKs is achieved by a number of DSPs that also appear to be critical regulators of ERKs activity. Unlike other DSPs, MKP-3 (also called Pyst1 and rVH6) is exclusively localized in the cell cytoplasm, where it physically associates with and selectively inactivates ERKs. Expression of MKP-3 has been implicated in controlling the cytoplasmic retention of ERK (18,32,33,45). Forced expression of a mutant of MKP-3, inactive in its catalytic domain, in NIH 3T3 cells, is able to prevent ERK translocation from the cytosol into the nucleus, showing that the binding of MKP-3 to ERK is sufficient to prevent ERK translocation (18). Here we show that, in contrast to parental PC12 cells, in PC12/MEN2A and PC12/MEN2B cells, MKP-3 is steadily expressed at levels that are even higher than those found in the NGF-stimulated parental cells. Consistently treating cells with the tyrosine phosphatase inhibitor, orthovanadate, results in further enhancement of the ERK activity, thus suggesting that the ERK levels of activation observed in these cells likely result from the balance between chronic activation induced by the active Ret oncogenes and inactivation of ERK phosphatases. Even though the presence of high levels of MKP-3 implicates this ERKspecific phosphatase in balancing Ret-induced activation, this phosphatase will be not the only candidate. Indeed, down-regulation of ERK activity is a still poorly understood mechanism that likely implicates the convergent action of distinct enzymatic activities, including the dual specificity phosphatases.
Taken together, the data presented are consistent with the existence in ret-transfected cells of a negative regulatory feedback caused by Ret-induced chronic stimulation of the  figure) induce a chronic activation of the Ras/ERK pathway. The levels of ERKs activity result from the balance between the chronic activation of ERKs and the high levels of the MKP-3 and, eventually, of other ERK phosphatases (DSP). The resultant levels of ERKs activity are sufficient to induce neuronal gene expression but not terminal differentiation. C, in PC12/MEN2A and PC12/ MEN2B cells, NGF stimulation is able to activate further ERKs but is not sufficient to overcome the negative feedback by inducing nuclear translocation of ERKs and thus terminal differentiation of the transfected cells. D, in PC12/MEN2A cells forced expression of an active variant of MEK-1 is able to induce high levels of ERKs phosphorylation, its translocation into the nucleus, and terminal differentiation. levels of ERKs which in turn cause a number of events, including high steady-state levels of MKP-3 transcripts, whose final consequences would be lack of terminal differentiation and unresponsiveness to NGF (see Fig. 8, A-C).
Moreover, since expression of active mutant variants of MEK-1 is sufficient for terminal differentiation in the PC12 (15), as a necessary corollary to this scenario we expect that a strong stimulation of ERKs by a similar MEK-1 mutant should overcome such a block and induce terminal differentiation.
Thus, the fact that stimulating ERK phosphorylation with a constitutive active mutant of MEK-1 (N3-S218E-S222D) is sufficient to rescue both c-fos transcription and inhibition of cell proliferation well supports the hypothesis that Ret-induced unresponsiveness to NGF cells may be the consequence of down-regulatory signals acting on either MAPK or MEK-1. Indeed, it is conceivable that the constitutive active MEK-1 might be insensitive to such signals, thus causing a strong stimulation of ERKs and overcoming the negative regulation acting on ERKs in ret-transfected cells (Fig. 8D).
Even though the only known substrates for MEK-1 are ERK-1 and ERK-2, we cannot rule out that induction, by the active ret oncogene, of other enzymatic activities also participate in the abrogation of terminal differentiation.
The presence of a negative feedback acting on ERKs as a mechanism involved in the resistance to the antiproliferative effects induced by NGF seems to be a mechanism triggered by Ret mutants (either Ret C634Y or Ret M918T ), which is substantially indifferent to the kind of activating mutations of these oncogenes. The more pronounced difference between PC12/ MEN2A and PC12/MEN2B cells concerns the ability to induce the fosCAT promoter. In fact, MEK-1 expression induces transcription from the exogenous c-fos promoter in both cell lines, but the levels of induction were distinctly higher in the PC12/ MEN2B as compared with the PC12/MEN2A cells. A likely explanation for such differences might be that the M918T mutation causes a shift in Ret substrate specificity (46). Indeed, even though we currently have no further experimental data supporting this possibility, a possible interpretation is that involvement of distinct enzymatic activities (triggered by the Ret-2A mutant) converging on ERK signaling may partially interfere with expression driven by the c-fos promoter. On the other hand, despite the fact that in PC12/MEN2A cells MEK-1 induces at low efficiency the exogenous c-fos promoter, its signaling was nonetheless sufficient to block proliferation, as assessed by BrdUrd incorporation. Taken together, these data support the idea that a similar mechanism is responsible for the lack of terminal differentiation in ret-2A-as in ret-2B-transfected PC12 cells, although the use of different ret mutants reveals possible signaling differences triggered by each of these mutants.
In conclusion, we propose the chronic activation of ERK by Ret mutants to be implicated in the unresponsiveness of rettransfected cells to antiproliferative extracellular signals stimulated by NGF. Moreover, since the utilization of the MAPK pathway is a branch point common to many extracellular signals, this is probably a general biochemical mechanism by which Ret mutants contribute to the hyperplasia of the chromaffin cells in MEN-2 syndromes.