Inhibitory Role of the Somatostatin Receptor SST2 on the Intracrine-regulated Cell Proliferation Induced by the 210-Amino Acid Fibroblast Growth Factor-2 Isoform

The fibroblast growth factor (FGF)-2 isoform of 210 amino acids (HMW FGF-2) contains a nuclear localization sequence (NLS) and is targeted to the nucleus. This FGF-2 isoform allows cells to grow in low serum concentrations through still unknown mechanisms called intracrine regulations. Different peptide hormones and cytokines have been found to be nuclearized through NLS and to induce cell proliferation. The existence of molecules acting as negative regulators of the intra-crine-induced cell growth has not been explored. Pancreatic cells AR4-2J were stably transfected to express selectively the HMW FGF-2. We demonstrated that activation of the somatostatin receptor subtype SST2 by the somatostatin analogue RC-160 in serum-deprived medium inhibits the mitogenic effect of the HMW FGF-2, without affecting growth of control cells. The signaling pathway implicates G (cid:1) i /JAK2/SHP-1. The G (cid:1) i inhibitor pertussis toxin and the JAK2 inhibitor AG490 abrogate the inhibitory effect of RC-160 on HMW FGF-2-induced cell growth. Co-immunoprecipitation

The fibroblast growth factor (FGF)-2 isoform of 210 amino acids (HMW FGF-2) contains a nuclear localization sequence (NLS) and is targeted to the nucleus. This FGF-2 isoform allows cells to grow in low serum concentrations through still unknown mechanisms called intracrine regulations. Different peptide hormones and cytokines have been found to be nuclearized through NLS and to induce cell proliferation. The existence of molecules acting as negative regulators of the intracrine-induced cell growth has not been explored. Pancreatic cells AR4-2J were stably transfected to express selectively the HMW FGF-2. We demonstrated that activation of the somatostatin receptor subtype SST2 by the somatostatin analogue RC-160 in serum-deprived medium inhibits the mitogenic effect of the HMW FGF-2, without affecting growth of control cells. The signaling pathway implicates G␣ i /JAK2/SHP-1. The G␣ i inhibitor pertussis toxin and the JAK2 inhibitor AG490 abrogate the inhibitory effect of RC-160 on HMW FGF-2-induced cell growth. Co-immunoprecipitation studies demonstrate the constitutive association of JAK2 and SHP-1, and RC-160 induces a rapid activation of both proteins followed by the dissociation of the complex. AG490 prevents the RC-160 induced SHP-1 activation indicating the implication of JAK2 in this process. JAK2 and SHP-1 are immunoprecipitated with SST2 in basal conditions indicating the existence of a functional signaling complex at the receptor level. In summary, these data provide the following evidence: 1) the intracrine-induced proliferation can be reversed by extracellular acting polypeptides; 2) SST2 inhibitory signaling may involve the JAK2/SHP-1 pathway.
Basic FGF 1 (FGF-2), a protein belonging to the family of heparin-binding growth factors, is responsible for the growth promotion, blood supply, and invasiveness in cancer (for review, see Ref. 1). These pleiotropic effects have been related previously to auto-paracrine regulations involving extracellular FGF-2 and FGF receptors of high and low affinities (2). However, unlike normal adult cells, tumor cells express FGF-2 as different molecular weight forms with a predominance of high molecular weights. FGF-2 peptides are synthesized from a single mRNA at AUG and CUG start codons (3)(4)(5) and under the control of internal ribosomal entry sequences (6). Translation of the high molecular forms of 21-34 kDa is initiated at CUG codons located upstream from the AUG codon. The latter codon initiates synthesis of the low molecular weight peptide of 155 amino acids (LMW FGF-2 of 18 kDa). The high molecular weight forms only differ from the LMW peptide by the presence of N-terminal extensions of different lengths containing nuclear localization sequences (NLS) (5,7). Confocal and immunohistochemical analyses localized the HMW FGF-2 in the nucleus (1, 8 -10). Furthermore, chimeric cytoplasmic proteins containing the NLS sequence of the large FGF-2 isoforms were found in the nucleus, confirming the translocation efficiency of the NLS (9). Because all FGF-2 forms lack the secretory signal peptide, the LMW peptide is highly concentrated in the cytoplasmic compartment, whereas the other isoforms are nuclear. The LMW form can be secreted by a novel energy-dependent secretory mechanism involving Na,K-ATPase (11,12). After an endocytotic process, small amounts of the LMW peptide are targeted to the nucleus, and in contrast to the HMW, the LMW FGF-2 is specifically localized in the nucleolar compartment (13). The puzzling feature of the HMW FGF-2 is now supported by the more recent discovery of nuclear migration of various growth factors, peptide hormones, and cytokines, many of which possess NLS sequences (for review see Refs. 14 and 15). Localization of these peptides in different intracellular compartments has been extensively analyzed by electron microscopy and confocal microscopy. Nuclear translocation of several transmembrane receptors including the FGF receptor-1 has also been observed (16,17). One conceivable advantage of direct action of cytokines and/or their receptors in the nucleus is a more specific response than activation of overlapping signaling pathways (14). The implication of these nuclearized peptides or receptors in specific regulatory functions, for instance in the control of gene transcription, has also been reported (15)(16)(17). In some cases the known biological functions related to nuclear localizations concern proliferation and/or antiapoptotic responses (among the factors involved are FGF-2, insulin-like growth factor-1, and parathyroid hormone-related protein), and in other cases they have been related to angiogenic processes (among the factors involved are FGFs, vascular endothelial growth factors, and platelet-derived growth factors) (14,15,18). However, some of the biological functions of the nuclearized peptides still remain to be established, and the intracellular regulatory mechanisms involved, called "intracrine regulations," are still poorly understood (16). Because of the importance of FGF-2 in several physiological and pathological situations, its biological effects have been more extensively studied. The results obtained after the FGF-2 selective expression either of the secreted form of 155 amino acids (LMW FGF-2) or of the nuclear translocated high molecular weight form of 210 amino acids (HMW FGF-2) indicate the existence of some differences in their biological functions. For instance, expression of the LMW FGF-2 enhances cell migration (10) and increases integrin (19), tissue plasminogen activator, and PKC⑀ expressions (20,21). By contrast, the HMW FGF-2 specifically promotes cell growth at low serum concentration (10,22,23), reduces cell migration (10,20), expression of laminin B1 (8), tissue plasminogen activator (20), plasminogen activator inhibitor-1 (20), PKC⑀ (21), adenylate cyclase activity (24), and inhibits interleukin-6 promoter activity (25). In these reports, which also include the cell line AR4-2J-A3 used in the present study, addition of neutralizing anti-FGF-2 antibodies, which recognize the common sequence of all FGF-2 forms, or expression of a dominant negative FGF receptor-1 were found to counteract the biological effects of the extracellular FGF-2 but not those of the HMW FGF-2 (10, 20, 26 -28). Thus, the LMW peptide regulates cell functions predominantly through auto-paracrine mechanisms, whereas the larger one acts via intracrine regulations independently of the activation of cell surface FGF receptors (1,18,29). Human pancreatic cancers overproduce the HMW FGF-2 of 210 aa, and this peptide facilitates tumor growth (30). Inhibition of tumor growth is a crucial point in cancer therapy; therefore, inhibition of the mitogenic effect of the HMW FGF-2 might represent a critical point for new therapeutic approaches. However, at the present time, experimental data demonstrating the possible inhibition of the intracrine FGF-2-mediated cell proliferation are not available.
Somatostatin is a widely distributed inhibitory neuropeptide that directly antagonizes the mitogenic action of growth factors in many cell types (for review see Ref. 31). This antimitogenic effect is only observed when cells are induced to proliferate. SST2 belongs to the G-protein-coupled somatostatin receptor family, which consists of five subtypes (SST1 to SST5). In the pancreatic cancer cell line AR4-2J, the SST2-mediated antiproliferative effect involves the activation of the tyrosine phosphatase SHP-1 (32,33). SHP-1 dephosphorylates and consequently inhibits several signal transducers involved in the cell cycle progression (34). The purpose of the present study was to analyze whether somatostatin was able to inhibit the intracrine-induced growth of pancreatic cancer cells AR4-2J expressing the HMW FGF-2 of 210 aa. Our data show that somatostatin via its receptor SST2 inhibits the HMW FGF-2induced proliferation by a mechanism involving JAK2 kinase activation. This is the first demonstration of inhibition of an intracrine-regulated cell growth.

EXPERIMENTAL PROCEDURES
Cell Culture-A3 cells expressing the HMW FGF-2 of 210 amino acids and control CAT cells were routinely grown in DMEM containing 4.5 g/liter glucose (Invitrogen) supplemented with 10% fetal calf serum (Invitrogen). Trypsin (0.05%)/EDTA (0.02%) (Invitrogen) was used to dissociate the cells for successive passages, and medium was replaced every 2 days. Cultures were regularly checked for the absence of contamination by mycoplasma.
Experiments were carried out in serum-free media. Cells were plated in DMEM containing 10% fetal calf serum (FCS), and after an overnight attachment media were replaced by DMEM minus serum, buffered at pH 7.3 with 20 mM Hepes. Growth experiments were performed either by counting the cells (Coulter Counter Z1) or by adding to cell cultures 18 h after drug addition 1 Ci/ml of [ 3 H]thymidine (Amersham Biosciences, 44 Ci/mmol, final concentration 7 Ci/mmol), and the incorporated radioactivity was measured 6 h later.
The PKC inhibitor GF109203X, the Src inhibitor PP2, the JAK2 inhibitor AG490 (tyrphostin B42), the MEK inhibitor PD 98059, the p38 inhibitor SB 203580, the phosphatidylinositol 3-kinase inhibitor wortmannin, and the G␣ i inhibitor pertussis toxin (PTX) were from Calbiochem. Most of these drugs were solubilized in Me 2 SO and used at the final concentration of 0.1%. They were added to serum-free cultures for 24 h, and control cells were grown in the presence of Me 2 SO.
Plasmids and Transfections-The retroviral transfection of AR4-2J cells was described previously (22). Briefly, the cell line AR4-2J-A3 (called A3 cells) was obtained by transfecting the pancreatic cancer cell line AR4-2J with an FGF-2 cDNA encoding only the FGF-2 of 210 amino acids. The cell line AR4-2J-CAT (called CAT cells) was obtained by transfecting the pancreatic cancer cell line AR4-2J with the same vector in which the FGF-2 cDNA was replaced by the gene encoding the chloramphenicol acetyltransferase. The geneticin-resistant cells were shown to express low levels of FGF-2 (22). The transfected cell lines used for the present experiments were not cloned.
For experiments on SHP-1 functions, A3 cells were transiently transfected with the inactive mutant SHP-1 (SHP-1-C453S) vector (33). Transfections were realized overnight with FuGENE 6, and the total amount of the transfected vector was normalized to 0.2 g/well, with the void vector. Experiments were carried the 3rd day after transfection and 48 h after serum deprivation.
Analysis of SST1 and SST2 Receptors Expression by Real Time PCR-Total RNA from control CAT and A3 cells was prepared from 24-h starved cells by using the TriPure isolation reagent (Roche Applied Science). The RNase-free DNase (Promega) (2.5 IU) was then added to the purified RNA and incubated for 30 min at 37°C to hydrolyze genomic DNA. The DNase was then inactivated at 94°C for 5 min. Reverse transcription was then performed with 1 g of total RNA in RT buffer containing 0.5 mM each of dNTP, 20 units of RNasin (Promega, Charbonnières, France), 200 units of Superscript II reverse transcriptase (Invitrogen), and 10 mM dithiothreitol. The real time PCR was performed on a GeneAmp 5700 (Applied Biosystems), using the Master Mix SYBR Green I (Roche Applied Science). The thermal profiles were 2 min at 50°C, followed by 10 min at 95°C and 1 min at 60°C. Standard curves for expression of each gene were generated by serial dilutions of RT templates. Relative quantifications of signals were done by normalizing the signals of SST1 and -2 genes with those of ␤-actin. Primers were synthesized by MWG-Biotec (Courtaboeuf, France). The following primers were used (35): SST1, forward primer 5Ј-GCAACAT-GCTCATGCC-3Ј and reverse primer 5Ј-GCGTTATCCATCCAGC-3Ј; SST2, forward primer 5Ј-GGGCGAATCCGGGGCA-3Ј and reverse primer 5Ј-GTTTGGAGGTCTCCATTG-3Ј. Primers specific for rat ␤-actin were obtained from Clontech (BD Biosciences).
Western Blots-Cells were grown for 24 h in a serum-free medium buffered at pH 7.3 with 20 mM Hepes. After washings with cold phosphate-buffered saline, cells were harvested in lysis buffer (phosphatebuffered saline, 1% Nonidet P-40, 150 mM NaCl, 1 mM EGTA, pH 7.4) containing the protease inhibitor Pefabloc (Roche Applied Science) and phosphatase inhibitors (25 mM NaF and 10 mM sodium orthovanadate). Lysates were clarified by centrifugation, 10 min at 10,000 ϫ g at 4°C. Protein was determined on the supernatant using the BCA assay (Pierce). SHP-1, JAK2, SST1, and SST2 were analyzed on a 7.5% SDS-PAGE gel and then electroblotted onto a polyvinylidene difluoride membrane (Schleicher & Schü ll). For co-immunoprecipitation studies, equal amounts of proteins (1 mg) from unstimulated or RC-160-stimulated cells were immunoprecipitated with antibodies directed against SHP-1, JAK2, SST1, or SST2, overnight at 4°C in the presence of protein A-agarose with gentle rocking. The immune complexes were recovered by centrifugation at 2,000 ϫ g, washed 3 times with the lysis buffer containing the different inhibitors, solubilized in the sample buffer, and then boiled 5 min before SDS-PAGE separation. The anti-SHP-1 antibody was from Transduction Laboratories, and lysates from Jurkat cells were used as control of migration. The anti-JAK2 and the antiphosphotyrosine 4G10 antibody were from Upstate Biotechnology, Inc. (Euromedex, Mundolsheim, France). The anti-SST2 and anti-SST1 receptor antibodies were described previously (36,37). Antigen-anti-body complexes were detected using enhanced chemiluminescence (ECL assay kit, Amersham Biosciences) and signals quantified with a laser densitometer (Amersham Biosciences).
SHP-1 and JAK2 Activities-SHP-1 activity was measured using 33 P-labeled poly(Glu-Tyr) as substrate (Sigma) as described previously (38). Immunoprecipitated proteins with anti-SHP-1 antibodies as described above were resuspended in a Tris-HCl buffer (50 mM, pH 7.0) containing 1 mg/ml bovine serum albumin, 0.5 mg/ml bacitracin, 5 mM dithiothreitol, 10 M genistein, and 10 M PP2. The poly(Glu-Tyr) was phosphorylated with [␥-33 P]ATP (38). The reaction was started by adding the phosphorylated substrate and allowed to proceed for 10 min at 30°C and then stopped with trichloroacetic acid. One unit of SHP-1 activity was defined as the amount of phosphatase that released 1 pmol of phosphate/min/mg solubilized protein.
JAK2 kinase activity was measured by autophosphorylation, on the immunoprecipitates obtained with the anti-JAK2 antibody in the presence of protease and tyrosine phosphatase inhibitors as described above. Immunocomplexes were washed twice with lysis buffer and three times with Tris-buffered saline at 4°C and then resuspended with cold kinase buffer containing 10 mM Tris-HCl, 5 mM MgCl 2 , 10 mM MnCl 2 , 0.1 mM sodium orthovanadate, 3.75 M ATP, and 0.5 Ci/l [␥-32 P]ATP at pH 7.4. Kinase reactions were allowed to proceed at 30°C for 10 min (39); proteins were resolved by SDS-PAGE, and 32 P-labeled proteins were detected by autoradiography. Quantifications were performed on a PhosphorImager (Amersham Biosciences). The same blot was also probed with the anti-JAK2 antibody, and the antigen-antibody complex was revealed by ECL prior to quantification with the laser densitometer.
Statistics-Statistical significance was determined according to the unpaired Student's t test.

RC-160 Inhibits Proliferation of AR4-2J Cells
Expressing the HMW FGF-2 of 210 aa-The pancreatic cell line AR4-2J was chosen for its interesting properties; it express FGF-2 receptors but does not synthesize detectable amounts of FGF-2 (22). In transfected AR4-2J-A3 cells (A3 cells), the expression of the HMW FGF-2 of 210 amino acids was under the control of the ␤-actin promoter. A weak promoter rather than a strong viral promoter was chosen to obtain low expression levels of HMW FGF-2 (22). In the presence of low serum concentrations, A3 cells grew faster and reached higher saturation densities than control cells transfected with the same vector in which the HMW FGF-2 cDNA was replaced by CAT (CAT cells) (Fig. 1), as already reported for other cell types (1). Somatostatin has been shown to inhibit cell growth induced by exogenous growth factors or serum (31). In order to investigate whether soma-tostatin was able to inhibit the overgrowth induced by the expression of the HMW FGF-2, A3 and control CAT cells were grown in the absence of FCS, with or without vapreotide (RC-160), a stable analogue of somatostatin. After 48 h of culture in serum-deprived medium, overgrowth of A3 versus CAT cells reached 26.2 Ϯ 3.1% (n ϭ 10; p Ͻ 0.01). RC-160 did not affect proliferation of CAT cells in serum-free medium. By contrast, 100 nM RC-160 abrogated CAT cell growth induced by 1% FCS (result not shown). RC 160 inhibited in a concentration-dependent manner proliferation of A3 cells grown in serum-free conditions, with a maximal effect of about 85% at 100 nM ( Fig.  2A) and a half-maximal effect at 1 Ϯ 0.9 nM. Thus, 100 nM RC-160 treatment decreased A3 cell proliferation to levels of control CAT cells. These data indicate that expression of HMW FGF-2 results in an increase of cell growth that can be suppressed by addition of RC-160.
Only SST1 and -2 receptor subtype mRNAs have been reported in AR4-2J cells (38), and the IC 50 value for RC-160 observed above corresponds to SST2. We investigated the effect of the SST1-specific analogue CH275 on CAT and A3 cell growth in order to exclude the involvement of this SST receptor subtype. CH275 appeared unable to modify cell proliferation either of CAT or A3 cells up to concentrations of 100 nM (Fig.  2B). These data indicate that the growth advantage of A3 cells resulting from the expression of the HMW FGF-2 in these cells can be reversed by somatostatin and further suggest that activation of the somatostatin receptor subtype SST2 is responsible for the somatostatin-induced growth inhibition.
The Inhibitory Effect of RC-160 on A3 Cells Is Unrelated to an Increased Expression of SST Receptors-In order to explain the inhibition of A3 cell growth induced by RC-160, we first investigated whether the HMW FGF-2 affected the expression of SST receptors. The real time PCR analysis indicated that the HMW FGF-2 did not modify SST1 and SST2 mRNA levels in A3 versus CAT cells (data not shown). Furthermore, Western blot analysis also indicated that the SST2 protein was expressed at comparable levels in both cell lines (Fig. 3). By contrast, we were unable to detect SST1 protein in CAT and A3 cells (data not shown), suggesting that the SST1 mRNA is not translated in these cell lines as already reported for AR4-2J cells (40). These data indicate that the effect of RC-160 in HMW FGF-2-producing cells was unrelated to modification in SST1 and SST2 receptor levels or to altered SST1/SST2 ratios.
G␣ i and Tyrosine Phosphatase SHP-1 Are Involved in the RC-160-induced Growth Inhibition-Because the pertussis toxin-sensitive G␣ i protein is involved in SST2 signaling cascade (33), we asked whether G␣ i protein was implicated in the RC-160 effect on A3 cells. To test this hypothesis A3 cells were grown for 24 h in the absence of serum, in the presence or not of RC-160. The G␣ i inhibitor PTX (100 ng/ml) was added to culture media 30 min before RC-160 addition. Whereas the G␣ i inhibitor had no effect alone, it abolished the inhibitory effect of 10 nM RC-160 on [ 3 H]thymidine incorporation (Fig. 4). This result indicates that a G␣ i protein is implicated in the antimitogenic pathway activated by the somatostatin analogue.
Activation of the tyrosine phosphatase SHP-1 was shown previously (33,41) to play a key function in the growth inhibition elicited by the SST2 receptor occupancy. Western blot analysis indicated that SHP-1 levels were unmodified by HMW FGF-2 expression (result not shown). In order to test whether the activation of SHP-1 was involved in growth inhibitory effect mediated by RC-160, A3 cells were treated with 10 nM RC-160 for 10 min, and SHP-1 was immunoprecipitated with the anti-SHP-1 antibody. The SHP-1 phosphatase activity was then measured in the immunoprecipitates (Fig. 5A). A time course study showed a 1.8-fold increase in SHP-1 activity as early as 30 s after RC-160 addition and was sustained during the stimulation period. This somatostatin-induced activation of SHP-1 suggests that in A3 cells SHP-1 is involved early in the SST2 pathway.
To validate the implication of SHP-1 in the growth inhibition elicited by RC-160, A3 cells were transiently transfected with increasing concentrations of a dominant negative mutant of SHP-1 (SHP-1-C453S) vector and then grown for 24 h in serum-free medium with or without 10 nM RC-160. The results reported in Fig. 5B show that transfection of 0.15 g/dish of the SHP-1-C453S vector abolished the RC-160-induced growth inhibition. Whereas transfection of the mock vector in the absence of RC-160 did not affect cell growth up to 0.2 g/dish (not shown). Control experiments were performed to check the transfection efficiency. Western blot analysis with anti-SHP-1 antibodies (Fig. 5C) showed an overexpression of the tyrosine phosphatase in cells transfected with the vector containing the mutated SHP-1 (0.2 g/well), thus confirming the transfection efficiency. RC-160-stimulated cells presented a slight and not significant increase in SHP-1 expression in mock and transfected cells (Fig. 5C). Such an increase did not abolish the effect of the dominant negative SHP-1-C453S (see Fig. 5B; 0.2 g/ well). These results indicate the critical role for the tyrosine phosphatase SHP-1 in the somatostatin antimitogenic effect.
JAK2 Implication in RC-160-mediated Growth Inhibition-To characterize the signal transduction pathway responsible for the antimitogenic effect of RC-160, specific inhibitors of several pathways were tested, including the Src inhibitor PP2, the phosphatidylinositol 3-kinase inhibitor wortmannin, the p38 kinase inhibitor SB 203580, the PKC inhibitor GF109203X, and the MEK inhibitor PD 98059 (Fig. 6A). None of these compounds affected the antimitogenic action of RC-160 on A3 cells. In contrast, the JAK2 inhibitor AG490 reversed the effect of RC-160 in a concentration-dependent manner (Fig. 6B). AG490 alone did not modify A3 cell growth, whereas it already abrogated the effect of the somatostatin analogue at concentrations as low as 10 M (Fig. 6B). This finding suggests that JAK2 plays an essential role in the signaling pathway activated by RC-160.
We then tested whether RC-160 regulated JAK2 activity in A3 cells. For that purpose, serum-starved A3 cells were incubated with or without 10 nM RC-160, and JAK2 activity was measured in anti-JAK2 immunoprecipitates. RC-160 induced a time-dependent activation of JAK2 that was already evident at 30 s and was maximal at 3 min (Fig. 7). This finding indicates that the ligand-SST2 receptor interaction stimulates the kinase activity of JAK2. Interaction between SHP-1 and JAK2-The direct association of JAK2 with SHP-1 has been described in hematopoietic cells during erythropoietin (Epo) stimulations (42). The physical interaction between the two signaling proteins has been reported to cause activation of SHP-1 by JAK2 (43). It was then tempting to suggest that in A3 cells occupancy of SST2 might induce the activation of JAK2, and consequently the activated JAK2 might be responsible for SHP-1 activation. In order to test this hypothesis, the interaction between SHP-1 and JAK2 was first analyzed. Lysates obtained from A3 cells treated or not with RC-160 for 10 min were immunoprecipitated with the anti-JAK2. Immunoprecipitated proteins were immunoblotted with the anti-SHP-1 antibody. Our data showed that a JAK2-SHP-1 complex was already present in resting cells (Fig. 8A). Immunoprecipitations carried out with the anti-SHP-1 antibody co-immunoprecipitated SHP-1 and JAK2 (data not shown), which confirmed the constitutive association between both proteins. Fig. 8A also shows that RC-160 induced a timedependent dissociation of the protein complex SHP-1-JAK2, which was observed as early as 30 s.
In order to test whether JAK2 was responsible for SHP-1 activation, we measured SHP-1 activity in the presence of the JAK2 inactivator AG490. For that purpose, A3 cells were preincubated for 30 min with AG490 (10 M) before addition of RC-160. Cell lysates were immunoprecipitated with the anti-SHP-1 antibody, and the phosphatase activity was then assessed. As shown in Fig. 8B, AG490 abrogated RC-160-induced SHP-1 activation for any time of RC-160 treatment tested. Control experiments confirmed that 10 M AG490 effectively inactivated JAK2, as measured in a JAK2 kinase assay (not shown). These results show the critical role played by JAK2 in the SHP-1 activation induced by RC-160.
JAK2 and SHP-1 Are Associated with SST2-We then investigated whether JAK2 and SHP-1 were associated with the SST2 receptor subtype in A3 cells. For that purpose, lysates prepared from RC-160-treated cells were immunoprecipitated with the anti-SST2 antibody. Immunoblots were then probed concomitantly with anti-JAK2, anti-SHP-1, and SST-2 antibodies, taking into account the different molecular weights of these proteins, and were then reprobed with the antiphosphotyrosine antibody. In resting cells, JAK2 and SHP-1 were found associated with SST2 (Fig. 9A). Addition of RC-160 did not significantly affect the pool of JAK2 associated to SST2, whereas it decreased by about 60% the receptor-associated SHP-1 within 3 min of treatment (Fig. 9B, upper panel). These data suggest the presence in resting cells of a trimeric SST2-JAK2-SHP-1 complex and that RC-160 treatment resulted in SHP-1 dissociation from this complex. Considering the level of the SST2-coupled JAK2 during RC-160 stimulation (Fig.  9B, upper panel), the calculated P-JAK2/JAK2 ratio in the SST2 complex was maximal at 30 s (230% increase) and remained at a high level (Fig. 9B, lower panel). In contrast, a similar analysis for the SST2-associated SHP-1 indicates that RC-160 up-regulated the P-SHP-1/SHP1 ratio up to 3 min (360% increase), which was followed by a decrease at 10 min (Fig. 9B, lower panel).
Altogether these data indicate the existence at the receptor level of a SST2-JAK2-SHP-1 complex. Ligand binding to SST2 induces both the phosphorylation of the receptor-associated JAK2 and SHP-1 and the dissociation of SHP-1.

DISCUSSION
In the present study we have demonstrated that the somatostatin analogue RC-160 inhibits the proliferation induced by the expression of the HMW FGF-2 of 210 amino acids in pancreatic cancer cells. This growth-inhibitory effect was observed in A3 cells producing the HMW FGF-2 but not in control CAT cells. Somatostatin has been shown to inhibit the proliferation induced by growth factors or serum (31). Because control CAT cells were growth-inhibited by RC-160 after serum stimulation, the present data show that in the A3 cells there is a relationship between HMW FGF-2 expression and RC-160-induced growth inhibition.
Because several studies (1,18,20,22,23,29) have reported the mitogenic effect of the HMW FGF-2 via intracrine regulations in various cell types including A3 cells, the present finding demonstrates that RC-160 activates signaling pathways counteracting the intracrine HMW FGF-2-induced cell growth. To our knowledge this is the first report of a receptor-mediated inhibition of an intracrine-mediated cell proliferation.
Although mRNAs encoding the somatostatin receptor subtypes SST1 and SST2 are present in AR4-2J cells (38) and in the derived cell lines CAT and A3, only SST2 protein was found either in AR4-2J (44) or in CAT and A3 cells. Comparison of SST2 mRNA and protein levels did not unravel significant modifications between CAT and A3 cells. We then hypothesized that the difference in CAT and A3 cell sensitivity to RC-160mediated inhibition of cell growth depends on a direct inhibitory effect of RC-160 on HMW FGF-2 mitogenic signals. Experiments performed with the somatostatin analogues that present different affinities for SST1 and SST2 receptors confirmed that only RC-160 was effective, indicating the specific involvement of SST2 in somatostatin action.
Signaling events involved in inhibition by RC-160 of the HMW FGF-2-induced mitogenic effects were then explored. It is well established that SST2 couples to G␣ i and that PTX uncouples G␣ i from SST2 (33). The observation that PTX blocks the antiproliferative action of RC-160 on A3 cells indicates that the SST2-mediated antiproliferative response is dependent on G␣ i proteins (33).
In CHO cells co-expressing SST2 and SHP-1 as well as in AR4-2J cells, co-immunoprecipitation studies showed that SHP-1 formed a complex with SST2 (33), and previous studies (32) on AR4-2J cells demonstrated that SHP-1 mediated the growth inhibition induced by the ligand-stimulated SST2 receptor. SHP-1 is a cytoplasmic tyrosine phosphatase containing two Src homology 2 domains at the amino region and a phosphatase catalytic domain at the C terminus (45). SHP-1 involvement in the antimitogenic effect of RC-160 on A3 cells was first suggested by the finding that SHP-1 was activated during RC-160 stimulation. The implication of this phosphatase was then confirmed by using the dominant negative SHP-1 C453S. Indeed, cell transfection with increasing amounts of SHP-1 C453S cDNA induced a concentration-dependent reduction of RC-160 antiproliferative effect.
At the present time, the signaling requirements for SHP-1 activation during somatostatin stimulation remain to be established. In A3 cells expressing the HMW FGF-2, we demonstrated that AG490, a JAK2 inactivator, induced a concentration-dependent reduction of the antimitogenic effect of RC-160. The JAK tyrosine kinases bind to members of the cytokine receptor family and are activated by trans-phosphorylation after ligand binding to these receptors (46). A number of studies demonstrated an interaction between JAK2 and SHP-1 (43) and analyzed its role in Epo signaling (42). In these studies SHP-1 and JAK2 were found constitutively associated, and interestingly, the SHP-1/JAK2 association was demonstrated to involve the N-terminal region of JAK2 independently of its Src homology 2 domain (42). Cytokine stimulation has been reported to induce JAK2 activation, which rapidly activates SHP-1. Consequently, the active SHP-1 plays the role of negative regulator of JAK2 signaling by de-phosphorylating JAK2. However, the question as to whether SHP-1 is only implicated in JAK2 de-phosphorylation or also plays a role in receptor regulations remains to be clarified (42,43). Only a few reports point out JAK2 association with G-protein-coupled receptors (GPCR), including angiotensin II (47) and RANTES chemokine receptors (48). However, in contrast to somatostatin, the GPCR-binding molecules angiotensin II and RANTES are mitogenic. Furthermore, with regard to angiotensin II-AT1 receptor interactions, the active SHP-1 plays the role of down-regulator of JAK2, as observed for the cytokine receptors (47). In the case of the chemokine RANTES, which activates a G␣ i -coupled receptor, the involvement of SHP-1 has not been reported (48). The implication of JAK2 in the signaling cascade initiated by the SST receptors has not been described. In A3 cells, coimmunoprecipitation of JAK2 and SHP-1 in unstimulated cells suggests the existence of a constitutive interaction between these two proteins. Furthermore, because SHP-1 is not activated by RC-160 in the presence of the JAK2 inhibitor AG490, we hypothesize that active JAK2 is required for the somatostatin-mediated SHP-1 activation. These data suggest that RC-160 activates the JAK2/SHP-1 pathway as already described for Epo signaling (42,43).
Immunoprecipitation studies also show that JAK2 and SHP-1 are associated with SST2, indicating the presence in resting cells of a ternary complex SST2-JAK2-SHP-1. Such an association of JAK2 with the inhibitory GPCR SST2 is a novel information. During RC-160 stimulation, we observed a rapid tyrosine phosphorylation of the SST2-coupled JAK2, in agreement with the effect of RC-160 on JAK2 activation. In addition, RC-160 induces SHP-1 tyrosine phosphorylation, as reported during cytokine stimulation (42,43). In the course of RC-160 treatment, the amount of the SST2-associated SHP-1 protein decreases, indicating the dissociation of SHP-1 from the ternary complex. Since during RC-160 stimulation SHP-1 was found to remain in the active state (Fig. 5A), the dissociated phosphatase must exert biological functions. Dissociation of SHP-1 from SST2 was already observed in CHO-SST2 cells during somatostatin treatment (33). The released SHP-1 might be translocated to other cellular compartments in order to trigger antimitogenic signals. A number of studies indicate that the JAK2 signaling pathway is involved in mitogenic cell responses (46). The data presented here strongly suggest that JAK2 may also play antimitogenic effects when functionally associated with a receptor that negatively regulates cell growth.
Altogether the present findings indicate that G␣ i , JAK2, and SHP-1 sustain the SST2-induced antimitogenic effect in A3 cells. The physical interactions among these proteins and the detailed steps of the signaling cascade arising from the ligand activation of SST2 remain to be clarified and are the subject of our ongoing investigations.
The somatostatin-activated SHP-1 pathway has been shown to lead to G 1 cell cycle arrest in CHO-SST2 cells (34). This pathway involves the induction of the cyclin-dependent kinase inhibitor p27 Kip , the inhibition of cyclin E-Cdk2 kinase activity, and the accumulation of hypophosphorylated retinoblastoma (34,49). The same type of regulation of the cell cycle machinery might also happen in A3 cells, or SHP-1 might directly inactivate some of the signaling proteins mediating the growth effect of the HMW FGF-2 (21).
The results of the present study clearly indicate that the mitogenic effect of the HMW FGF-2 on cancer cells can be down-regulated. Our findings might have predictable interests in physiological and pathological situations in which cell proliferation occurs through intracrine regulations (14,15,50). The mitogenic signal transduction pathways involved in the intracrine regulation activated by growth factors or nuclear translocated growth factor receptors still remain to be clarified. It has been reported that the nuclear translocated epidermal growth factor receptors (16) and FGF receptor-1 (17) up-regulated the cyclin D1 expression. Because SST2 receptors activate the signaling proteins that negatively regulate the cell cycle (34), it will then be of a great interest to check the effect of somatostatin in the down-regulation of the cell growth induced by other intracrine-acting factors.