The expression of the phosphotyrosine phosphatase DEP-1/PTPeta dictates the responsivity of glioma cells to somatostatin inhibition of cell proliferation.

Here we characterize the intracellular effectors of the antiproliferative activity of somatostatin in glioma cell lines and post-surgical specimens. The responsiveness to somatostatin correlated with the expression of the phosphotyrosine phosphatase DEP-1/PTPeta, identified in C6 and U87MG cells, in which somatostatin inhibited cell growth. The expression of a dominant negative mutant of DEP-1/PTPeta in C6 cells abolished somatostatin effects, confirming the involvement of this phosphotyrosine phosphatase in such effects. Somatostatin treatment increased the activity of DEP-1/PTPeta and inhibited ERK1/2 activation. Conversely, basic fibroblast growth factor-dependent MEK phosphorylation was not affected, suggesting a direct effect on ERK1/2. In vitro experiments showed that PTPeta was able to interact and dephosphorylate ERK1/2 activated by basic fibroblast growth factor. Furthermore, by transfecting PTPeta in the somatostatin-unresponsive, DEP-1/PTPeta-deficient U373MG cells, the somatostatin-dependent control of cell proliferation was recovered. Finally we evaluated the requirement for DEP-1/PTPeta in somatostatin inhibition of cell proliferation in post-surgical specimens derived from different grade human gliomas. Although all of the glioma analyzed expressed somatostatin receptor mRNA, DEP-1/PTPeta expression was limited to 8 of 22 of the tumors. Culturing seven gliomas, a correlation between the expression of DEP-1/PTPeta and the somatostatin antiproliferative effects was identified. In conclusion we propose that the expression and activation of DEP-1/PTPeta is required for somatostatin inhibition of glioma proliferation.

Here we characterize the intracellular effectors of the antiproliferative activity of somatostatin in glioma cell lines and post-surgical specimens. The responsiveness to somatostatin correlated with the expression of the phosphotyrosine phosphatase DEP-1/PTP, identified in C6 and U87MG cells, in which somatostatin inhibited cell growth. The expression of a dominant negative mutant of DEP-1/PTP in C6 cells abolished somatostatin effects, confirming the involvement of this phosphotyrosine phosphatase in such effects. Somatostatin treatment increased the activity of DEP-1/PTP and inhibited ERK1/2 activation. Conversely, basic fibroblast growth factor-dependent MEK phosphorylation was not affected, suggesting a direct effect on ERK1/2. In vitro experiments showed that PTP was able to interact and dephosphorylate ERK1/2 activated by basic fibroblast growth factor. Furthermore, by transfecting PTP in the somatostatin-unresponsive, DEP-1/PTP-deficient U373MG cells, the somatostatin-dependent control of cell proliferation was recovered. Finally we evaluated the requirement for DEP-1/PTP in somatostatin inhibition of cell proliferation in post-surgical specimens derived from different grade human gliomas. Although all of the glioma analyzed expressed somatostatin receptor mRNA, DEP-1/PTP expression was limited to 8 of 22 of the tumors. Culturing seven gliomas, a correlation between the expression of DEP-1/PTP and the somatostatin antiproliferative effects was identified. In conclusion we propose that the expression and activation of DEP-1/PTP is required for somatostatin inhibition of glioma proliferation.
Somatostatin (SST) 1 is an endogenous regulator of proliferation in a variety of epithelial and endocrine cells (1)(2)(3)(4). The effects of SST are mediated by a family of five G proteincoupled receptors (named SSTR1-5) (5) that are variably expressed in both normal tissues and tumors (6). SST inhibition of cell proliferation involves both direct and indirect mechanisms. SST indirectly controls cell growth in vivo through its inhibitory effects on the release of growth promoting hormones (7) or through an antiangiogenic mechanism involving the regulation of proliferation and migration of endothelial cells and monocytes (8,9), whereas a direct antiproliferative activity in both normal and tumoral cells is mediated by the activation of the SSTR expressed in the cell membrane (5,7).
The modulation of phosphotyrosine phosphatase (PTP) activity is considered one of the main intracellular pathways responsible for the direct inhibition of cell growth by SST (3, 10 -13). Indeed, although protein-tyrosine kinases provide the "on switch" for both normal and tumoral cell proliferation and migration, in most cases, PTPs represent the "off switch" signal to bring back cell activity to the basal state. In the past years different endogenous compounds displaying antiproliferative activity have been reported to modulate the activity of this class of enzymes including, in addition to SST, gonadotropin hormone releasing hormone, angiotensin II, and dopamine (14 -16). However, most of the studies have been performed on PTP activation by SSTR (for review see Ref. 17). In particular, a SST-dependent increase of PTP activity has been shown to induce dephosphorylation of the epidermal growth factor receptor, inhibiting the mitogenic activity of epidermal growth factor (10,18). The SST-induced PTP activity was reported to be associated to the plasma membrane, sharing biochemical features with the SHP1 and SHP2 phosphatases (19 -23). Although SHPs are involved in the cytostatic activity of SST, more recently they were reported to play a more general role as adaptor molecules rather than specific effectors. Indeed, SHP1 and SHP2 may act both as positive or negative regulators of intracellular signaling, depending on the cell context. For instance, it was reported that SHP2 may be an activator of the signaling through epidermal growth factor receptor or Met but an inhibitor of the platelet-derived growth factor receptor-ac-tivated intracellular pathways (24 -26). Thus, it was proposed that the specific antiproliferative activity of SST may also involve other PTPs.
We recently reported that SST-dependent cytostatic effects in the PC Cl3 rat thyroid cells are mediated by the activation of a receptor-like PTP, named PTP (27), on the basis of its homology with the human DEP-1/HPTP gene (28). PTP is expressed ubiquitously, showing high levels in the brain, liver, and spleen (28). The predicted protein contains a unique intracellular catalytic domain, a short transmembrane domain, and an extracellular region containing eight fibronectin type III-like repeats (28). Experimental evidence suggests that this PTP is involved in the negative regulation of cell growth. For example, after vascular injury DEP-1 is down-regulated in migrating and proliferating endothelial cells (29), whereas the overexpression of DEP-1 in macrophages or breast cells inhibited the generation of the stable lines, likely because of the marked inhibition of the proliferation (30,31). In addition to a role as an inhibitor of proliferation, DEP-1/PTP has also been implicated in cell differentiation, because its expression increased in differentiated breast cancer cells (31) and was induced by differentiating agents (i.e. thyroidstimulating hormones) in normal thyroid cells (32). Conversely, PTP expression is down-regulated by oncogene-dependent thyroid cell transformation as well as in malignant human thyroid tumors (33,34). Reintroduction of PTP into transformed rat thyroid cells leads to a the recovery of a partially differentiated phenotype and the slowing down of the proliferation rate via the inhibition of the ERK1/2-dependent proteolysis of p27 kip1 (27,34). Importantly after SST treatment, a similar activation of SHP2 was observed in both PTP-expressing or nonexpressing transformed cells, but SST was able to induce a p27 kip1 -mediated cytostatic effect only when PTP was activated (27). These data indicate that, in this cell model, although SST activates both SHP2 and PTP, the latter enzymatic activity is required to transduce SSTR-dependent antiproliferative signals.
The aim of this paper was to establish a more general role for DEP-1/PTP as transducer of the antiproliferative activity of SST. In particular we analyzed different glioma cell lines and post-surgical specimens from human gliomas to identify a correlation between the expression of this PTP and a possible responsivity to the antitumoral effects of SST.

EXPERIMENTAL PROCEDURES
Materials-SST and PD98059 were purchased from Calbiochem (Lucerne, Switzerland), vanadate was from ICN Biochemicals Inc. (Cleveland, OH), and all other reagents were from Sigma unless otherwise specified.
Antibodies-For DEP-1/PTP detection, we used antibodies raised against the intracellular region of DEP-1 or PTP expressed as a recombinant protein fused to glutathione S-transferase (GST) and affinity-purified (34), used, in Western blot experiments, at the dilution of 1:500. The other antibodies were p44/42, phospho-p44/p42 mitogenactivated protein kinase, MEK, and phospho-MEK (New England Biolabs Inc., Beverly, MA) and c-Myc (Santa Cruz Biotechnology Inc., Santa Cruz, CA) all used at the dilution of 1:1000.
Cell Cultures, Treatments, and Transfections-C6 cell lines were grown in Ham's F-12 medium, supplemented with 10% fetal calf serum (ICN). U87 MG, U373 MG, DBTRG 05MG and CAS1 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. All of the cell lines were obtained from the Bank of Biological Material of the Interlab Cell Line Collection (National Institute for Cancer Research, Genova, Italy). Pertussis toxin (PTX) (180 ng/ml) treatment was performed in serum-free medium for 18 h before the experimental treatments, as reported (35). Transfections were performed using the FuGENE reagent (Roche Applied Science) according to the manufacturer's instructions. For the generation of stable cell lines, after the transfection with plasmids with PTP, PTP[C/S], or PTP␣ subcloned in PCDNA3, the cells were selected in G418 (500 g/ml). Single clones were isolated and kept in G418 containing medium. The expression of the transfected genes was confirmed by both RT-PCR and Western blot (WB).
Patients, Tumors, and Tissue Preparation-Post-surgical glioma samples were obtained from the Division of Neurosurgery of the Department of Neuroscience, Ophthalmology and Genetics of the University of Genova. Thirty-six patient-derived specimens (15 females and 21 males; age range; 14 -85 years; mean age, 57.3 years) were analyzed. Diagnosis and classification into histological subtypes, carried out at the Department of Pathology of San Martino Hospital (Genova, Italy), showed that one tumor was a pylocitic astrocytoma (World Health Organization grade I), five were fibrillary astrocytomas (World Health Organization grade II), nine were anaplastic astrocytomas (World Health Organization grade III) and 21 were glioblastoma multiforme (World Health Organization grade IV). After histological examination, part of each tumor was immediately frozen at Ϫ80°C until mRNA extraction, and, when possible, the remaining tissue was dispersed and grown in primary cultures, as described below.
Messenger RNA Analysis-The expression of specific mRNAs was evaluated by means of RT-PCR. Total RNA was isolated using the acidic phenol technique (36). RT-PCR was performed as previously reported (27). Briefly, 10 g of total RNA were treated for 45 min with RNasefree DNase at 37°C to remove genomic DNA contamination, phenol/ chloroform-extracted, and ethanol-precipitated. RT reaction was performed using oligo(dT) 16 primer and the avian myeloblastosis virus RT (Amersham Biosciences), for 40 min at 42°C. PCR was performed on 10 ng of cDNA as follows (final volume, 50 l): 5 min of denaturation at 94°C followed by 40 cycles of 1 min at 94°C, 1 min at 60°C, and 1 min at 72°C, followed by 7 min at 72°C, using the Taq DNA polymerase (2.5 units/reaction) (Roche Applied Science). The amplified DNA fragments were visualized by agarose gel electrophoresis.
The primers used for the amplification of both rat and human SSTR1-5 cDNAs were previously described (27,37); for the amplification of rat PTP, primers were designed according to the sequences of the nucleotides 3975-3795 and 3453-3473 of the published sequence (28); for human DEP-1, primers were designed according to the sequences of the nucleotides 3794 -3814 and 4319 -4339 of the published sequence (GenBank TM accession number U10886).
␤-Actin fragment amplification was assessed in all PCR as reaction efficiency control, using specific primers designed based on the nucleotides 2595-2682 and 2865-2988 of the published sequence (accession number J00691) (38). In addition a negative control (PCR amplification in the absence of RT reaction) was used to verify the absence of genomic DNA contamination in the cDNA samples.
Primary Cultures of Human Glioma Cells-Single cell suspensions were obtained from resected glioma pieces (patients 2, 3, 12, 15, 24, 32, and 34) after mechanical disruption under sterile conditions. The cultures were maintained in minimal essential medium containing 10% fetal calf serum and nonessential amino acids. The medium was changed every 24 h removing nonadherent cells. The cells were subcultured at least once before experiments.
[ 3 H]Thymidine Incorporation Assay-DNA synthesis activity was measured by means of the [ 3 H]thymidine uptake assay, as previously reported (16). Briefly, the cells were plated at the density of 5 ϫ 10 4 in 24-well plates. After 24 h the cells were serum-starved for 48 h and treated with the test substances for 16 h. In the last 4 h the cells were pulsed with 1 Ci/ml of [ 3 H]thymidine (Amersham Biosciences). Then the cells were trypsinized (15 min at 37°C), extracted in 10% trichloroacetic acid, and filtered under vacuum through fiber glass filters (GF/A; Whatman, Clifton, NJ). The filters were sequentially washed with 10 and 5% trichloroacetic acid and 95% ethanol. The trichloroacetic acid-insoluble fraction was counted in a scintillation counter.
PTP Assay-PTP assay was performed on immunocomplexes using the synthetic substrate para-nitrophenyl phosphate (pNPP) in a spectrophotometric assay. The samples were incubated at 30°C in a 80-l

DEP-1/PTP Expression Dictates Responsivity of Glioma Cells
volume containing 20 l of a 5ϫ reaction buffer (250 mM HEPES, pH 7.2, 50 mM dithiothreitol, 25 mM EDTA, 500 nM microcystin-leucinearginine (Alamone Laboratories, Jerusalem, Israel)), and the reaction was started by adding 20 l of 50 mM pNPP, carried out for 60 min, and stopped by adding 900 l of 0.2 N NaOH. The absorbance of the sample, directly proportional to the amount of the cleaved substrate, was measured at 410 nm (39). Immunoprecipitation-Total cell lysates (250 g) were incubated with the appropriate antibody (1 g/1 mg of protein) for 2 h at 4°C in RIPA buffer, and then protein A-Sepharose was added for an additional hour. After three washes with RIPA buffer the immunocomplexes were precipitated and analyzed in WB. For PTP immunocomplex assay, the precipitation was done with IgG-coupled magnetic beads (Dynabeads), because the protein A caused hydrolysis of pNPP (27).
Pull-down Experiments-GST-fused recombinant PTP and its fragment 993-1135 were purified by glutathione-Sepharose affinity chromatography (Amersham Biosciences) as previously described (40). Precipitation experiments were performed as reported (41). Briefly, serumstarved C6 cells were treated for 10 min with 50 ng/ml bFGF and lysed in 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 1 mM vanadate, and the Complete protease inhibitor mixture (Roche Applied Science). One mg of proteins was precleared with 20 l of slurry of glutathione-Sepharose (Amersham Biosciences), then incubated with 1 mg of recombinant PTP993-1135, and left rocking at 4°C for 1 h. ERK-PTP993-1135 complex was then precipitated by centrifugation after incubation with 20 l of slurry of glutathione-Sepharose and washed three times with the same buffer without vanadate. After the last wash the sample was resuspended in PTP assay buffer (see above) and divided in four samples into which was added buffer alone or containing different concentrations of recombinant, catalytically active PTP. The samples were incubated at 30°C for 1 h, and then the reaction was stopped by adding Laemmli buffer and boiling the samples for 1 min. The samples were then size-fractionated by SDS-PAGE and analyzed in WB for the presence of phospho-ERK1/2.
Statistical Analysis-The experiments were performed in quadruplicate and repeated at least three times. Statistical analysis was performed by means of one-way analysis of variance. A p value Ͻ0.05 was considered statistically significant.

Somatostatin Inhibition of Glioma Cell Proliferation
Correlates with the Expression of DEP-1/PTP-To establish a general role for DEP-1/PTP as an intracellular effector of the antiproliferative activity of SST, we analyzed the effects of this peptide on the proliferation of different human and rat glioma cells. We evaluated SST effects on the U87 MG, U373 MG, CAS1, and DBTRG 05MG human glioblastoma cells and the C6 rat glioma cells. All of these cell lines expressed multiple SSTR, as determined by RT-PCR experiments (Table I). In particular all of the cell lines showed SSTR2 mRNA either as single receptor (U87 MG and U373 MG) or in association with SSTR1 (DBTRG 05MG), SSTR1 and 3 (CAS1), and SSTR1, 3, and 5 (C6). However, irrespectively to the pattern of SSTR mRNA expression, only U87 MG and C6 cells showed a significant inhibition of bFGF-induced cell proliferation after SST treatment. In these cells, SST caused a dose-dependent inhibition of bFGF-stimulated DNA synthesis, with EC 50 of 1 and 10 nM, respectively (Fig. 1). Conversely, in U373 MG, DBTRG 05MG, and CAS1, SST was completely ineffective as far as cell proliferation inhibition, even when used at high concentration ( Fig.  1). Because all of the cell lines expressed SSTR subtypes able to inhibit cell proliferation, to understand the molecular determinants of the antiproliferative effects of SST, we investigated, in the different cell lines, the expression of DEP-1/PTP by means of both RT-PCR and WB. As shown in Fig. 1 (bottom right panel), we identified a precise correlation between the expression of DEP-1/PTP and the inhibitory effects of SST. Indeed, DEP-1/PTP mRNA and protein were identified only in C6 and U87 MG cells.
DEP-1/PTP Activation by Somatostatin Interferes with ERK1/2 Activity-Thus, we characterized the SST regulation of DEP-1/PTP activity and its relationship with the antiproliferative effects of SST, using the C6 cells as glioma cell model (42). The dose-dependent inhibition of cell proliferation induced by SST was completely reverted by the pretreatment with PTX (180 ng/ml, 18 h), involving the activation of G␣ i/o in SST effects, and by means of vanadate (30 M), confirming that the activation of PTPs was necessary to SST to block the proliferative stimulus of bFGF (data not shown). The direct regulation of PTP by SST was measured in a PTP assay, after immunoprecipitation of the enzyme, in control or SST-treated cells, as previously reported (27). SST treatment caused a significant increase in PTP activity starting after 15 min of treatment, and that further increased up to 60 min ( Fig. 2A). Importantly, the increase in PTP activity was completely reverted by PTX treatment (180 ng/ml, 18 h) ( Fig. 2A). To correlate the SST-dependent increase of PTP activity with the SST effects on cell proliferation, we evaluated the effects of SST on C6 cell proliferation in two selected clones stably transfected with a dominant negative mutant of PTP (PTP[C/S]) (43). In these cells SST was completely unable to inhibit bFGF-induced DNA synthesis, thus confirming that the activation of PTP is necessary for the antiproliferative effects of SST (Fig. 2B).
To delve deeper into the mechanisms regulated by PTP to control cell proliferation, we evaluated the effects of SST on ERK1/2 activity stimulated by bFGF. SST added simultaneously to the growth factor or with a short pretreatment (5 min) did not significantly modify the bFGF-dependent ERK1/2 phosphorylation. However, longer pretreatments (15-60 min), although ineffective per se in the modulation of ERK1/2 activity (data not shown), powerfully inhibited bFGF effects (Fig. 3A), showing a time course similar to that observed in the PTP activity. Also, the SST dose-response curve (measured after 60 min of pretreatment) was comparable with that obtained in the proliferation experiments (maximal inhibition at 1 M, with a down-regulation of the effect at higher concentrations) (Fig.  3B). These data suggest that the SST-dependent activation of PTP may control C6 cell growth through the inactivation of ERK1/2. To identify the possible site of action of PTP along the ERK1/2 activation cascade, we evaluated the effect of SST treatment on MEK phosphorylation. Using the same cell extracts in which a significant inhibition of ERK1/2 phosphorylation was observed, we did not detect significant changes in MEK1 phosphorylation (Fig. 3C), although a slight reduction was observed after 1 h of pretreatment with SST. These data suggest that PTP acts downstream of MEK and thus likely directly on ERK1/2. To explore this possibility, we performed in vitro dephosphorylation experiments using a recombinant, catalytically active PTP (evaluated as pNPPase activity; data not shown) after immunoprecipitation of bFGF-activated ERK. To better identify ERK and avoid interferences of the antibody with PTP activity, we transiently transfected a plasmid expressing a Myc-tagged ERK2. The cells were then treated for 10 min with 50 ng/ml bFGF and immunoprecipitated with anti-Myc antibody, incubated 60 min with or without 1 g of recombinant PTP, and evaluated in WB for the presence of phospho-ERK2. As shown in the Fig. 4A, ERK2 immunoprecipitated from bFGF-treated cells was highly phosphorylated, and the incubation with PTP significantly reduced the amount of phosphorylated Myc-ERK2 in both control and treated cells. The presence of equal amounts of immunoprecipitated ERK2 was demonstrated by WB, using anti-Myc antibody  (Fig. 4A). Moreover, to confirm these data and identify a possible site of interaction between PTP and ERK1/2, we performed pull-down experiments, as previously reported (41), using a recombinant fragment of PTP encompassing the amino acids 993-1135 (PTP993-1135), located in the intracellular juxtamembrane domain of the PTP and devoid of catalytic activity (as measured as pNPPase activity; data not shown), isolated as fusion protein with GST. After the binding, the ERK-PTP993-1135 complex was incubated with glutathione-Sepharose and precipitated by centrifugation. The isolated ERK-PTP993-1135 complex was then incubated with catalytically active recombinant PTP and evaluated for phospho-ERK1/2 by WB. As shown in Fig. 4B, we were able to precipitate active ERK1/2 via the binding with PTP993-1135, and phosphorylated ERK1/2 was directly dephosphorylated by PTP, in vitro.

DEP-1/PTP Expression Dictates Responsivity of Glioma Cells
Reversal of Somatostatin Unresponsive Phenotype by Expression of PTP in U373 Glioma Cells-To further demonstrate the role of DEP-1/PTP in the SST antiproliferative signaling, we tried to revert the SST-unresponsive phenotype of U373 MG cells by stable expression of PTP in these cells. From transfection experiments, we obtained many clones but, in agreement with previous results in different cell types (31), a high expression level of PTP caused an exaggerate inhibition of basal cell proliferation, and after few passages the cells were irreversibly growth-arrested. Thus, we repeated the transfection and, among several clones showing the same features than the previous ones, we were able to select few clones with lower PTP levels that maintained a proliferative rate high enough to allow further analysis. We studied two independent clones that showed similar results. The simple expression of PTP in these cells, as demonstrated by both RT-PCR and WB (Fig. 5A), caused the loss of some markers of transformed cells, such as the ability to proliferate in an anchorage-independent manner (colony forming in soft agar) (Fig. 5B), a significant increase in the doubling time (Fig. 5C), and the loss of an autonomous pattern of proliferation characterized by insensitivity to growth factors, such as bFGF (Fig. 5D). In particular we found that, although growing much slower than the w.t. counterpart, PTP-transfected U373 MG cells showed a significantly higher increase over their respective basal DNA synthesis, in response to bFGF (Fig. 5D). Thus, we tested whether the expression of this PTP allowed also the recovery of the sensitivity to the inhibitory action of SST. As shown in Fig. 6, whereas the U373 MG w.t. were completely insensitive to SST treatment, the transfected cells displayed a dose-dependent (Fig. 6A) inhibition of DNA synthesis, with an EC 50 of ϳ10 nM, similar to that observed in the U87 MG, SST-responsive human glioma cells. The inhibition of cell proliferation by SSTR activation in U373-PTP cells was abolished by PTX pretreatment (Fig. 6B), confirming that all of the components of the SST transduction system are present in the w.t. cells (i.e. receptors and G proteins) and that the only lacking component was the final effec-

FIG. 4. PTP directly dephosphorylates phpspho-ERK1/2, in vitro.
A, C6 cells were transiently transfected with a Myc-ERK2-containing PC-DNA3 plasmid, treated with vehicle or bFGF (50 ng/ml) for 5 min, and then immunoprecipitated with anti-Myc antibody. The Myc-ERK2 immunoprecipitated (Ipt) was then incubated with 1 g of recombinant, catalytically active, PTP and then evaluated in WB for Myc-ERK2 phosphorylation. Equal level of immunoprecipitated proteins in the different samples was assured by WB on the same samples using anti-Myc antibody. B, the direct interaction between ERK1/2 and PTP was evaluated incubating bFGF-treated (50 ng/ml) C6 cell lysate with a recombinant GST-fused PTP fragment, corresponding to the amino acids 993-1135, that does not include the catalytic domain and thus is devoid of phosphatase activity. The ERK1/2-PTP993-1135 complex was then pulled down using glutathione-Sepharose beads. The precipitated proteins were incubated for 1 h with different concentration of catalytically active PTP and evaluated in WB for both phospho-ERK1/2 and total ERK1/2. As a control an aliquot of the cell lysate was incubated with a recombinant GST (2 g), and as expected, no ERK1/2 was pulled down in that sample.

DEP-1/PTP Expression Dictates Responsivity of Glioma Cells
tor, the PTP. Moreover, in addition to a significant PTP basal activity observed in U373-PTP cells, SST treatment also caused a 3-fold induction of PTP activity in the transfected cells (Fig. 6C), although the time course of the activation was shifted to the left, as compared with the effect observed in C6 cells, likely because of the overexpression of the PTP. As expected in U373 MG w.t. cells no PTP activity was observed after PTP immunoprecipitation, because this PTP is not expressed in these cells. SST-dependent PTP activation caused also, in the transfected cell line, a significant inhibition of the bFGF-induced ERK1/2 phosphorylation with a time course compatible with the kinetics of PTP activity (a significant inhibition observed after 5 min of SST pretreatment reflects the maximal PTP activation) (Fig. 6D). As expected no significant modulation of ERK1/2 phosphorylation was observed in the U373 w.t. cells (Fig. 6D). Interestingly, the specificity of the effects observed in the U373-PTP cells was demonstrated by transfecting another receptor-like PTP, PTP␣, in the U373 MG cells (U373-PTP␣). In these cells we did not observe a different growth kinetics compared with the w.t. cells (Fig. 5C), although a significant inhibition in the anchorage independent proliferation was detected (Fig. 5B). However, the overexpression of PTP␣ did not allow neither the recovery of the SST inhibition of bFGF-induced DNA synthesis (Fig. 6A) nor the inhibition of ERK1/2 phosphorylation (Fig. 6D), demonstrating the specificity of the PTP activation by SSTR.
Role of DEP-1/PTP in Somatostatin Antiproliferative Effects in Human Glioma Primary Cultures-Finally we tried to confirm these data using post-surgical specimens from human gliomas. We analyzed 36 gliomas of different grade as detailed under "Experimental Procedures." All of the samples, tested for the expression of SSTR by means of RT-PCR, showed the presence of at least one subtype. In particular, mRNA for SSTR1 was detected in 89% of the samples, SSTR2 in 70%, SSTR3 in 53%, SSTR4 in 39%, and SSTR5 in 30% (Table II). Moreover, in 22 of these samples, we analyzed the expression of DEP-1 mRNA to verify the possibility that the activation of this PTP by SSTR was a mechanism to control human glioma proliferation. DEP-1 mRNA was identified in a subset of the tumors analyzed (8 of 22; 36%), independently from the tumor grading (Table III). Thus, we tested the possible relationship between the expression of this PTP and the capability of SST to inhibit cell proliferation. We analyzed the effect of the SST treatment on the proliferation of primary cultures of three DEP-1-deficient (patients 2, 24, and 32) and three DEP-1expressing tumors (patients 3, 12, and 34) (Fig. 7). All of the tumors expressed mRNA for SSTR subtypes with a variable pattern of expression (Fig. 7). The cells were serum-deprived for 24 h and then treated with phorbol myristate acetate (PMA) (100 nM) as mitogen, in the presence or absence of SST. PMA is a powerful activator of protein kinase C that, through the direct phosphorylation of Raf, is able to activate the mitogenactivated protein kinase cascade (44). We confirmed that the expression of DEP-1/PTP is required for the SST antiproliferative effects because a significant inhibition of the PMAinduced DNA synthesis occurred only in the glioma cultures expressing DEP-1 (Fig. 7). Moreover, we identified another tumor (P15) in which SST significantly inhibited PMA-stimulated DNA synthesis, although because of the low amount of RNA extracted, it was not possible to evaluate the expression of DEP-1 (Fig. 7). DISCUSSION SST is a powerful negative regulator of both normal and tumoral cell proliferation. However, although significant results were obtained in preclinical studies, much more controversial are the clinical data. Indeed, despite the almost constant expression, and in many cases overexpression, of SSTR in human tumors and regardless of the histological origin, a clear therapeutic efficacy of SST was not identified. Thus, we hypothesized that the antiproliferative effects of SST may require, in addition to the expression of SSTR, the presence of specific intracellular effectors and that the lack of their expression may prevent the possible therapeutic benefit of SST treat-

DEP-1/PTP Expression Dictates Responsivity of Glioma Cells
ment. In the past years our study was dedicated to assess the role of specific PTPs in the SST antiproliferative activity. In particular, in rat thyroid cells we identified a receptor-like PTP, named PTP (DEP-1 in humans), that in addition to being selectively activated upon SST stimulation, thus mediating its cytostatic effects (3), was reported to be lost in some human malignancies (34). Indeed, being frequently deleted in human cancer, this PTP was proposed as a gene candidate for colon cancer susceptibility locus Scc1 (45). Thus, we tested whether the expression of this PTP (or of a similar enzyme) may dictate the responsivity to SST of glioma cell lines, as well as of different human glioma primary cell cultures. The choice of this kind of tumors was made in consideration of the almost constant expression of SSTR reported in these malignant cells, despite controversial results regarding the SST antiprolifera-tive activity after both in vivo and in vitro treatments (46,47).
The main conclusion that we draw from our study is that DEP-1/PTP activation is required for SST to induce a cytostatic effect. We found this correlation in both established cell lines and in post-surgical specimens: SST is devoid of effects when this PTP is not expressed. Moreover, the transfection of PTP interfering mutants in responsive glioma cell lines completely reverted SST effects. Although the analysis of a larger number of human tumors will be required to confirm these data, the identification of this unique characteristic that associates the responsivity to SST in cell lines and primary cultures to the expression of DEP-1/PTP is very suggestive of being a real discriminator for SST efficacy. In fact, we did not identify any relationship between the antiproliferative effect of SST and the pattern of SSTR subtype expressed (either in primary treatment on PTP specific activity in U373 w.t. and U373-PTP cells. No activity was detected in the w.t. cells, confirming the lack of the expression of PTP in these cells, whereas after transfection, in addition to a significant basal activity, SST treatment induced a powerful activation of PTP (ϩ300%). D, time course (5-30 min) of the effect of SST (1 M) treatment on bFGF-dependent (50 ng/ml, 5 min) ERK1/2 phosphorylation in U373 w.t., U373-PTP, and U373-PTP␣ cell lines, measured with phospho-specific antibodies. SST decreased the phosphorylation of ERK1/2 in a time-dependent manner that was effective starting after 5 min of pretreatment. Equal loading of proteins was demonstrated by using a total ERK1/2 antibody in a parallel gel (data not shown). Con, control. cultures of human tumors or in cell lines) or the grading of the human tumors analyzed, but only with the expression and activation of DEP-1/PTP. If our hypothesis is confirmed in more extensive studies, the analysis of the expression of DEP-1/PTP in human tumors will be particularly relevant from a therapeutic point of view. Indeed we found that the expression of DEP-1/PTP was present in approximately one of three of the tumors analyzed, thus indicating that only a small subset of gliomas are possibly responsive to SST or its analogues and thus explaining the contradictory results previously published.
In past years a pivotal role for many PTP in the control of the proliferation and the neoplastic transformation of the cells has been frequently reported. In particular, a role in the reversal of tumor proliferation, acting either at tyrosine kinase receptor or at downstream effectors such as ERK1/2, has been reported for PTP-SL, STEP, and TC-PTP, as well as for DEP-1 (41, 48 -50). Our experiments confirm these data, showing a direct interference between the activity of this PTP, the cell proliferation machinery, and the transformation mechanisms in glioma cells. In particular in our study we reverted some of the transformation characteristics of the U373 MG human glioblastoma cells (anchorage-independent colony-forming cells) through the simple overexpression of PTP.
However, differently from previous reports showing a DEP-1 activity on tyrosine kinase receptors (49,50) or PLC␥ (51,52), in our experimental model we identified ERK1/2 as a specific DEP-1/PTP target. In fact, we did not find significant changes in bFGF-dependent MEK activation after SST-mediated PTP activation, whereas we observed a significant reduction in ERK1/2 activation. This observation is also in line with our previous study in thyroid cells where we demonstrated that the activation of PTP caused a significant inhibition of ERK1/2 phosphorylation, also in conditions of MEK constitutive activation (27). Furthermore, as previously reported for PTP-SL and STEP (41), we were able to pull-down phospho-ERK1/2 through its interaction with a juxtamembrane fragment of PTP and to completely dephosphorylate them in vitro, after incubation with a catalytically active recombinant PTP. Thus, these data further support the evidence that PTP may directly bind ERK1/2 or other proteins forming a multimeric complex with ERK1/2 and that these kinases may represent specific substrates for PTP, in C6 cells.
Moreover, we clearly demonstrate that the activity of this PTP is under the control of SSTR. Indeed, a significant increase in PTP activity was observed after SST treatment in cells that natively express PTP (C6 cells) as well as after its exogenous introduction (U373-PTP cells). Interestingly, using selective analogues we demonstrated that both SSTR1 and SSTR2, the major SSTRs expressed in brain tumors (this paper and Refs. 47, 53, and 54), were able to activate PTP in C6 cells, 2 indicating that this PTP lays on a common intracellular pathway activated by different SSTRs to elicit a cytostatic response. Thus, because we previously demonstrated the activation of this intracellular pathway in the PC Cl3 thyroid cell line (27), the identification of its role also in the antiproliferative activity in different glioma cell lines as well as in primary glioma cell cultures allows us to propose that the activation of this PTP may represent the molecular correlate for the SST inhibition of cell proliferation. Moreover, the regulation of DEP-1/PTP by SST seems to be a very specific effect because the overexpression in the U373 MG cells of FIG. 7. Correlation between the expression of DEP-1 and the SST antiproliferative effects in primary cultures from human gliomas. Left panel, SSTR, and DEP-1 mRNA expression in a subset of the human gliomas analyzed from which were established primary cell cultures. ␤-Actin amplification was used as internal control for the efficiency of the RT reaction. The lack of contaminating genomic DNA was demonstrated by the lack of amplification products in a PCR on RNA before the RT reaction (data not shown). ND, not done. Right panel,   another receptor-like PTP, PTP␣, while inhibiting the soft agar colony formation, did not allow the recovery of the SST cytostatic effects. Interestingly, the recently developed antitumoral drugs that selectively inhibit the tyrosine kinase activity of different growth factor receptors may be nicely integrated by compounds able to activate PTP. Indeed, in identified DEP-1-expressing tumors, the combined use of compounds acting on the same intracellular pathway via different mechanisms may result in a significant synergism, allowing a better clinical output at lower dosage and less side effects.
In conclusion, we show that in both primary human glioma cells as well as in established cell lines the antiproliferative activity of SST is mediated by the activation of DEP-1/PTP and that the expression of this PTP is required for SST effects. We identified ERK1/2 as a major substrate of DEP-1/PTP, and its inactivation is responsible for the inhibition of the glioma cell proliferation induced by SST.