sst2 Somatostatin Receptor Mediates Cell Cycle Arrest and Induction of p27 Kip1

Activation of the somatostatin receptor sst2 inhibits cell proliferation by a mechanism involving the stimulation of the protein-tyrosine phosphatase SHP-1. The cell cycle regulatory events leading to sst2-mediated growth arrest are not known. Here, we report that treatment of Chinese hamster ovary cells expressing sst2 with the somatostatin analogue, RC-160, led to G1cell cycle arrest and inhibition of insulin-induced S-phase entry through induction of the cyclin-dependent kinase inhibitor p27 Kip1 . Consequently, a decrease of p27 Kip1 -cdk2 association, an inhibition of insulin-induced cyclin E-cdk2 kinase activity, and an accumulation of hypophosphorylated retinoblastoma gene product (Rb) were observed. However, RC-160 had no effect on the p21 Waf1/Cip1 . When sst2 was coexpressed with a catalytically inactive mutant SHP-1 in Chinese hamster ovary cells, mutant SHP-1 induced entry into cell cycle and down-regulation of p27 Kip1 and prevented modulation by insulin and RC-160 of p27 Kip1 expression, p27 Kip1 -cdk2 association, cyclin E-cdk2 kinase activity, and the phosphorylation state of Rb. In mouse pancreatic acini, RC-160 reverted down-regulation of p27 Kip1 induced by a mitogen, and this effect did not occur in acini from viable motheaten (me v /me v) mice expressing a mutant SHP-1 with markedly deficient enzymes. These findings provide the first evidence that sst2 induces cell cycle arrest through the up-regulation of p27 Kip1 and demonstrate that SHP-1 is required for maintaining high inhibitory levels of p27 Kip1 and is a critical target of the insulin, and somatostatin signaling cascade, leading to the modulation of p27 Kip1 .

Somatostatin is a widely distributed inhibitory hormone that plays an important role in several biological processes including neurotransmission, inhibition of exocrine and endocrine secretions, and cell proliferation. The diverse biological effects of somatostatin are mediated through a family of five somatostatin receptors (sst1-sst5) that belong to the family of Gprotein-coupled receptors and that regulate diverse signal transduction pathways including adenylate cyclase, phospholipase C-␤, phospholipase A 2 , guanylate cyclase, ionic conduct-ance channels, and tyrosine phosphatase (1,2).
The ability of somatostatin and its stable analogues to promote inhibition of normal and tumor cell growth has been demonstrated in various cell types including mammary, prostatic, gastric, pancreatic, colorectal, and small cell lung cancer cells (3,4). However, the mechanisms of cell growth arrest by somatostatin are still poorly understood. Somatostatin analogues induce a G 0 /G 1 cell cycle arrest and thus prevent DNA synthesis in GH3 rat pituitary tumor cells, whereas they induce a transient G 2 /M blockade as well as apoptosis in MCF7 human mammary tumor cells (5,6). These tumor cells express multiple somatostatin receptors and the question of whether different somatostatin receptor(s) may be involved in eliciting these effects still remains to be clarified. A specific role for sst3 in transducing apoptosis through an induction of p53 and Bax has been reported (7). Control of the cell cycle machinery by other receptors is an important problem that remains to be addressed.
Our studies on the expression of somatostatin receptor subtypes in heterologous systems led us to demonstrate that sst2 selectively mediates the antiproliferative effect of somatostatin analogues on serum-or insulin-induced cell growth through the stimulation of a protein-tyrosine phosphatase (8), which was recently identified as SHP-1 (9,10). SHP-1, a protein-tyrosine phosphatase with two SH2 domains, plays a role in terminating growth factor and cytokine signals by dephosphorylating critical molecules (reviewed in Ref. 11). We reported that SHP-1 is activated by somatostatin and participates in the negative regulation of mitogenic insulin signaling as a result of its association with and dephosphorylation of insulin receptor as well as associated molecules (9,10). However, the effect of sst2 as well as the role of SHP-1 on cell cycle parameters remain unknown.
Cell cycle progression is dependent on the coordinated interaction, posttranslational modification, and degradation of cyclins and their catalytic partners, cyclin-dependent kinases (cdks). 1 Cyclins are expressed at particular stages of the cell cycle and associate with specific cdks to form active complexes that phosphorylate multiple proteins and promote cell cycle progression. In mammalian cells, progression through early to middle G 1 phase of the cell cycle is dependent on cdk4/and/or cdk6, which are activated by D-type cyclins. Transition through middle G 1 to S phase is regulated by activation of cdk2 by cyclin E, cdk2, and cyclin A is required for late G 1 to S-phase progres-sion and throughout S phase. One of the critical targets of cyclin-cdk complexes is the retinoblastoma gene product (Rb). Rb acts as a transcriptional repressor. In its hypophosphorylated form, it binds to the E2F family of cell cycle transcription factors during G 1 phase and inhibits E2F activity. Rb is inactivated by cdk phosphorylation in mid to late G 1 phase of the cell cycle and dissociates from E2F, leading to activation of genes containing E2F sites and a progression from G 1 to S phase (reviewed in Ref. 12).
Another level of regulation of cdk activity results from the action of cdk inhibitors that bind cyclin-cdk complexes and either inhibit their kinase activities or prevent their activation by cdk-activating kinase (reviewed in Refs. 13 and 14). In mammalian cells, cdk inhibitors comprise two classes of proteins, the Ink4 family including p16 Ink4a , p15 Ink4b , p18 Ink4c , and p19 Ink4d , which specifically inhibit cyclin D-dependent kinases, cdk4 and cdk6 (12), and the p21 family including p21 Cip1/Waf1 , p27 Kip1 , and p57 Kip2 (15)(16)(17)(18)(19), which can interact with many different cyclin-cdk complexes. Among them, p27 Kip1 is a widely distributed cdk inhibitor that has an important role regulating entry into and exit from the cell cycle. p27 Kip1 is abundantly expressed in normal quiescent cells and is down-regulated by mitogens. The decrease in p27 Kip1 expression occurs through protein degradation via the ubiquitin-proteasome pathway after p27 Kip1 phosphorylation by cyclin E-cdk2 complexes (20 -22). Increased levels of p27 Kip1 induced by transforming growth factor ␤, contact inhibition, serum deprivation, rapamycin, or staurosporine have been associated with a G 1 arrest (18,23,24). In contrast, an overexpression of p27 Kip1 antisense cDNA results in mitogen-independent G 1 progression, demonstrating the importance of p27 Kip1 in controlling cell cycle exit (25,26). The involvement of p27 Kip1 in the negative regulation of cell proliferation is related to its binding and subsequent inhibition of the kinase activity of cdk2-cyclin complexes (17,18,23,27).
In this study, we investigated the potential effects of sst2 on cell cycle progression and expression of cell cycle regulatory proteins in CHO cells expressing sst2 (CHO/sst2). Activation of sst2 caused a G 1 cell cycle arrest in-phase accompanied by an increased expression of cdk inhibitor p27 Kip1 , which resulted in an increase of its association with cdk2 and a decrease in cdk2 activity and led to dephosphorylation of protein Rb. The role of SHP-1 in sst2-mediated regulatory mechanisms was investigated in CHO cells coexpressing sst2 and a negative SHP-1 (C453S) mutant as well as in acini isolated from viable motheaten mice expressing a mutant SHP-1 with markedly deficient enzyme activity. Our results provide evidence that SHP-1 is a critical regulator of p27 Kip1 . This enzyme is required for the maintenance of cell quiescence and is involved in the sst2mediated up-regulation of p27 Kip1 leading to cell cycle arrest.

EXPERIMENTAL PROCEDURES
Materials-Monoclonal anti-p27 Kip1 and anti-Rb antibodies, known to react with murine and human proteins, were purchased from Transduction Laboratories and Pharmingen, respectively. Monoclonal antihuman p21 Waf1/Cip1 antibodies were from Transduction Laboratories. Polyclonal anti-mouse p21 Waf1/Cip1 and anti-cdk2 antibodies that react with human and murine proteins were from Santa Cruz Biotechnology.
Monoclonal anticyclin E antibodies react with human and murine proteins and were from Calbiochem. RC-160 was synthesized as described previously (28). [␥ 33 P]ATP (3,000 Ci/mmol) was purchased from Isotopchim (France), histone H1 was from Sigma, and the enhanced chemiluminescence (ECL) immunodetection system was from Amersham Pharmacia Biotech.
DNA Transfection-The 1.2-kilobase XbaI fragment of mouse sst2A cDNA subcloned into pCMV6c vector was stably co-transfected in CHO (DG44 variant) cells using Lipofectin reagent with pSV2neo as described (kindly donated by Dr. G. I. Bell, Howard Hughes Medical Institute, University of Chicago and Dr. T. Reisine, University of Penn-sylvania, School of Medicine, Philadelphia) (8). Stable transfectants were selected in ␣MEM (minimal essential medium) containing geneticin at 600 g/ml. Geneticin-resistant clones expressing sst2 (CHO/ sst2) were screened for somatostatin binding using [ 125 I-Tyr 11 ] somatostatin as tracer as described (8). The 2.1-kilobase HindIII/NotI fragment of human SHP-1 cDNA (a gift of Dr. M. L. Thomas, Howard Hughes Medical Institute, Washington University, St. Louis, MO) was subcloned into the expression vector pcDNA I neo vector (Invitrogen). The SHP-1 (C453S) mutant (a gift of Dr. C. Nahmias, ICGM, Paris) was constructed as described (10). The mouse sst2 gene in the pCMV6c vector was stably co-transfected in CHO cells using Lipofectin reagent with the SHP-1 (C453S) mutant in pcDNA I neo. Stable colonies obtained by selection with G418 (600 g/ml) were screened for somatostatin binding and the presence of SHP-1 as described (9).
Pancreatic Acini from Viable Motheaten (me v /me v ) Mice-C57BL6me v /me v mice were obtained by mating heterozygous C57BL6-me v /ϩ mice (Jackson Laboratories, Bar Harbor, ME) breeding pairs. Homozygous me v /me v were screened by reverse transcription-polymerase chain reaction as described (29). Me v /me v were identifiable by 10 -15 days of age, because of the motheaten appearance of the skin. Me v /me v or their unaffected littermates were sacrificed at 3 weeks of age, and pancreases were removed. Pancreatic acini were prepared using enzymatic digestion of the pancreas with 1.5 mg/ml collagenase in an oxygenated Krebs-Ringer buffer containing 0.2% bovine serum albumin and 0.01% soybean trypsin inhibitor as described previously (buffer A) (30). After washing with buffer A, acinar cells were incubated in the same buffer in the presence or absence of peptides at 25°C for indicated times. Acini were then transferred to 0.3 M sucrose, sedimented at 800 ϫ g for 5 min at 4°C, and solubilized in lysis buffer (50 mM HEPES, 150 mM NaCl, 10 mM EDTA, 10 mM Na 4 P 2 O 7 , 100 mM NaF, 2 mM sodium orthovanadate (pH 7.4)) (buffer B) containing 0.1% Triton X-100, 1 mM benzamidine, and 0.01% soybean trypsin inhibitor at 4°C for 30 min. Lysates were collected and centrifuged at 13,000 ϫ g for 10 min at 4°C and used for immunoprecipitation or immunoblotting.
Flow Cytometric Analysis-Cells were harvested by trypsin (0.5 mg/ ml) and EDTA (0.02 mg/ml), washed twice with phosphate-buffered saline (pH 7.4), and centrifuged at 800 ϫ g for 5 min at 4°C. Cells were then incubated in the presence of 200 l of trypsin (30 mg/ml) for 10 min. 200 l of RNase (0.1 mg/ml) and trypsin inhibitors (0.5 mg/ml) were added for 10 min, and cells were stained with 250 l of propidium iodide solution (125 g/ml) for at least 15 min at 4°C. Fluorescence of labeled cell nuclei was measured by flow cytometry using a FACScan (Beckton Dickinson) with a minimum of 10,000 events performed for each sample. Data were analyzed using LYSIS II software.
Immunoprecipitation and Immunoblotting-Cells were washed with phosphate-buffered saline and then with buffer B. Cells were lysed in 500 l of buffer B containing 1% Triton X-100, 1 mM phenylmethylsulfonylfluoride, 20 g/ml aprotinin, 20 M leupeptin. After a 15-min incubation at 4°C, the lysate was collected and centrifuged at 13,000 ϫ g for 10 min at 4°C. Soluble proteins (300 -500 g) were incubated for 3 h at 4°C with specific antibodies or preimmune serum prebound to Sepharose-protein A beads prewashed in buffer B. The beads were then washed twice with buffer B and resuspended in sample buffer for immunoblotting.
For immunoblotting, solubilized proteins or immunoprecipitated proteins (see above) were resolved on 7.5% or 12% SDS-polyacrylamide gels, transferred to a nitrocellulose membrane, and immunoblotted with specific antibodies as described previously (9). Immunoreactive proteins were visualized by the ECL immunodetection system and quantified by image analysis using a Biocom apparatus (Biocom, Paris, France).
Kinase Assay-Immunoprecipitated proteins with anti-cdk2 or anticyclin E antibodies were collected by centrifugation and washed four times with buffer B and twice with 50 mM HEPES buffer containing 1 mM dithiotreitol (pH 7.4). The beads were suspended in 40 l of kinase buffer containing 50 mM HEPES (pH 7.5), 10 mM MgCl 2 , 1 mM dithiothreitol, 4 g of histone H1, 2.5 mM EGTA, 0.1 mM sodium orthovanadate, 1 mM NaF, 50 M ATP, and 10 Ci of [␥-33 P]ATP (1000 -3000 Ci/mmol) and incubated for 30 min at 25°C. Each sample was mixed with 20 l of 2ϫ SDS sample buffer, heated for 5 min at 100°C, and subjected to SDS-PAGE. The gel was fixed in 40% methanol, 10% acetic acid and exposed to hyperfilm ECL.

Somatostatin Promotes G 1 Cell Cycle Arrest and Blocks
Induction of the S Phase-We previously reported that in CHO cells expressing sst2, the addition of the somatostatin analogue, RC-160, for 24 h to the culture medium led to inhibition of the mitogenic effect of insulin (10). To analyze whether RC-160-mediated inhibition of cell proliferation reflects a stage-specific arrest of the cell cycle, cells were rendered quiescent by serum deprivation and incubated with 100 nM insulin in the presence or absence of 1 nM RC-160 and then analyzed by flow cytometry. Cells grown in the absence of fetal calf serum were taken as control values. The treatment of cells with insulin increased the percentage of cells in the S phase, which reached 27% at 6 h and increased up to 50% at 24 h of treatment (data not shown). The simultaneous treatment of cells with insulin and RC-160 for 6 h prevented cells from entering into the S phase, RC-160 causing a decrease in the percentage of cells in the S phase (Ϫ43%) and an accumulation of cells in the G 1 phase, which increased from 57% in the absence of RC-160 to 72% (Fig. 1). For longer treatment, RC-160 had no significant effect on the G 1 /S transition (data not shown). We concluded that activation of sst2 by ligand induces a G 1 cell cycle arrest in CHO/sst2 cells. Furthermore, treatment of cells with 1 M orthovanadate suppressed the RC 160-induced decrease of number of cells in the S phase as well as the increase of cells in the G 1 phase, indicating that a tyrosine phosphatase was required in the RC-160 effects (data not shown). G 1 progression depends on an orderly and coordinated expression of cyclins that bind to and activate cdks, the activity of which is negatively regulated by their association with a family of cdk inhibitory proteins. Therefore we investigated whether sst2-mediated cell cycle arrest is associated with a change in the expression of the cdk inhibitors.
Somatostatin Analogue Induces a Rapid Accumulation of p27 Kip1 -We first examined the expression of the Kip/Cip family cdk inhibitor p27 Kip1 , which has been demonstrated to be involved in the regulation of cell cycle progression induced by various antiproliferative stimuli that cause G 1 -phase arrest (23)(24)(25)(26). CHO/sst2 cells were treated for various times with insulin in the presence or not of RC-160, and the level of p27 Kip1 was investigated by Western blot analysis. As observed in Fig. 2, p27 Kip1 was expressed at high level in growth-arrested control cells, and after 3 h of insulin treatment, its expression decreased by 45% (p Ͻ 0.05), consistent with previous results reported for mitogenic signals (26,31,32). The decrease of p27 Kip1 level was transient, the level of this cdk inhibitor being not significantly different from that in control cells by 24 h. The addition of RC-160 resulted in a 4-fold increase (p Ͻ 0.02) in the level of p27 Kip1 during the first 3 h. Elevated levels of p27 Kip1 were found to return to control levels at 24 h of treatment with RC-160. Treatment of cells with 1 M orthovanadate for 3 h suppressed the RC 160-induced increase of p27 Kip1 , indicating that this effect is dependent on a tyrosine phosphatase (Fig. 2).
The expression of the other member of the Kip/Cip family cdk inhibitors, p21 Waf1/Cip1 , was also examined in CHO/sst2 cells. In contrast to p27 Kip1 , p21 Waf1/Cip1 was found to be barely detectable in control cells (Fig. 3), as observed by others in quiescent cells (33). As reported for mitogens in other cell  systems (32,34), insulin induced an increase of its expression up to 24 h, suggesting that elevated p21 Waf1/Cip1 is not related to insulin-mediated G 1 /S transition. However, the addition of RC-160 did not significantly modify the insulin-induced expression of p21 Waf1/Cip1 irrespective of the time of treatment, suggesting that this inhibitor is not involved in the somatostatinmediated growth arrest.
Somatostatin Analogue Induces Inhibition of cdk2 Kinase Activity-It has been shown in many cell types that among the G 1 cyclin-cdk complexes negatively regulated by p27 Kip1 , the up-regulation of p27 Kip1 in response to growth inhibitory factors favors its association with cyclin E-cdk2, resulting in kinase inhibition and contributing to cell growth arrest (34). Therefore, we first tested whether somatostatin analogue-mediated increases in the level of p27 Kip1 should be reflected in a change of the kinase activity of cdk2-associated complexes, as measured by an in vitro assay on cdk2 immunoprecipitates using histone H1 protein as a substrate. In comparison with the control activity detected in resting CHO/sst2 cells, cdk2 kinase activity was increased by about 2-fold after 3 h of treatment with insulin, as revealed by the heavily phosphorylated histone H1 level. When RC-160 was added to the culture, the cdk2-dependent kinase activity was inhibited by 80% during the first 3 h of culture (Fig. 4A). Similarly, RC-160 induced a decrease of about 80% insulin-induced increase of cyclin E-cdk2 associated kinase activity (Fig. 4B). In addition, the amount of p27 Kip1 associated with cdk2 was decreased by 70% after treatment of cells for 3 h with insulin, whereas the addition of RC-160 increased the level of the complexes by 87% (Fig. 4C). These results indicate that insulin and RC-160 could contrarily modulate the level of cdk2-cyclin E complexes, free from the constraining influence of p27 Kip1 inhibitor and the resulting kinase activity of the complexes.
We then analyzed the steady-state level of expression of cyclin E and cdk proteins by immunoblotting. As shown in Fig.  4D, the amount of cdk2 detected in control cells and in insulintreated cells was not significantly modified by RC-160 treatment. Anticyclin E antibodies revealed that cyclin E was barely detectable in control cells, and treatment of cells for 3 h with insulin induced an increase of the level of cyclin E as observed for agents that promote DNA synthesis (31). The addition of RC-160 resulted in a 50% inhibition of cyclin E expression after 3 h of treatment.
Somatostatin Analogue Prevents pRb Phosphorylation-One of the targets of cdk includes pRb protein, as its hyperphospho-rylated form was proven to be a critical check point involved in regulating progression through late G 1 and into S phase. The extent of pRb phosphorylation was first analyzed in soluble extracts of CHO/sst2 cells incubated in the presence of 100 nM insulin with or without 1 nM RC-160 for various times. A specific antibody able to detect the active hypophosphorylated form of pRb (Fig. 5, lower band) as well as the inactive slower migrating hyperphosphorylated form of the protein (Fig. 5, upper band) was used. In serum-starved cells, pRb was found in its active hypophosphorylated form in CHO/sst2 cells, and the inactive form of pRb became apparent after 3 h of insulin stimulation and remained present 6 and 24 h after insulin treatment. During those periods, the cells expressed fully hyperphosphorylated inactive pRb, in agreement with the reported relation of the appearance of the slowly migrating hyperphosphorylated form of pRb with the ability of mitogens to subsequently induce DNA synthesis (31). Treatment of cells with RC-160 produced an accumulation of the fast-migrating hypophosphorylated form of pRb, which was evident at 3 h. Increased levels of the hypophosphorylated form of pRb were no longer detectable by 6 -24 h after the addition of RC-160, indicating that the effect was transient.
SHP1 Is Required for Somatostatin Analogue-mediated Inhibition of S-phase Entry-SHP-1 has been previously demonstrated to play a role in the negative feedback of growth factor signaling and to be required for early events in somatostatinactivated sst2 signaling (9 -11). However the role of SHP-1 in the regulation of cell cycle machinery has not been delineated. To investigate whether SHP-1 is critical for cell cycle arrest and p27 Kip1 regulation, a mutated SHP-1 cDNA in which the active cysteine at position 453 was mutated to serine was generated and resulted in a catalytically inactive enzyme. We stably co-transfected the cDNA coding for the SHP-1 mutant and sst2 in CHO cells and selected the clones (CHO/sst2-SHP-1(C453S)) that expressed sst2 receptors at a level similar with that observed in CHO/sst2 cells (10). These clones overexpressed the SHP-1 mutant protein approximately 4-fold as observed by Western blotting (not shown).
Analysis of CHO/sst2-SHP-1(C453S) cells by flow cytometry revealed that in the absence of serum, cells did not undergo G 1 arrest as observed with CHO/sst2 cells, and 22% of cells remained in S phase. Furthermore, insulin and RC-160 did not modify the S phase, indicating that expression of SHP-1 mutant promoted G 1 progression of cells and nullified the modulatory effect of insulin and RC-160 on cell cycle progression

FIG. 3. Effect of insulin and RC-160 on p21 Waf1/Cip1 expression in CHO/ sst2 cells.
A, Serum-starved CHO/sst2 cells were incubated at 37°C for indicated times with 0.1 M insulin and with (InsϩRC) or without (Ins) 1 nM RC-160 or were not treated (control (Cont)) and solubilized as described under "Experimental Procedures." A, soluble proteins were subjected to SDS-PAGE and immunoblotted (Blot) with anti-p21 Waf1/Cip1 antibodies. The arrow indicates the position of p21 Waf1/Cip1 . B, immunoblots were analyzed densitometrically, and the data were plotted as the percentage of control values obtained from cells incubated in serum-free ␣MEM at time 3, 6, and 24 h. Data from three separate experiments are presented as mean ϮS.E. (Fig. 6). These results are in agreement with the previously observed increase in basal proliferation and abrogation of regulatory effect of insulin as well as RC-160 on cell proliferation in CHO/sst2-SHP-1(C453S) cells (10) and argue in favor of a role for SHP-1 in maintenance of cell quiescence.

SHP-1 Is Required for Somatostatin Analogue-mediated
Induction of p27 Kip1 , Inhibition of cdk2 Activity, and pRb Hypophosphorylation-We then examined whether SHP-1 mutant affected p27 Kip1 protein levels. Western blotting analysis demonstrated that in cells expressing mutant SHP-1, the basal level of p27 Kip1 decreased significantly as compared with control CHO/sst2 cells (57 Ϯ 5.6% of control) (Fig. 7A). Furthermore, the dominant negative mutant SHP-1 was found to prevent insulin-mediated down-regulation as well as RC-160induced up-regulation of p27 Kip1 (Fig. 7B). These results strongly suggest that SHP-1 might dephosphorylate some key substrate(s) in the insulin-and sst2-mediated signaling pathway in order for p27 Kip1 regulation to occur. We previously demonstrated that the phosphotyrosine insulin receptor is an early substrate of somatostatin-activated SHP-1 (11), suggesting that SHP-1 may exert at least part of its effects on p27 Kip1 expression by dephosphorylating the insulin receptor.
These results prompted us to examine the effect of mutant SHP-1 on cdk2 and cyclin E-associated kinase activities. Expression of mutant SHP-1 prevented the effect of insulin and RC-160 on cdk2 as well as cyclin E-associated kinase activity (Fig. 8A). In addition, the decrease of p27 Kip1 basal level observed in cells expressing mutant SHP-1 was paralleled by a decrease in the association of p27 Kip1 with cdk2 (Fig. 8B). The amount of p27 Kip1 associated with cdk2 was decreased by 45% under basal conditions, and insulin and RC-160 no longer had an effect on the association p27 Kip1 -cdk2. These results provide evidence that SHP-1 is involved in the retargeting of p27 Kip1 to cyclin E-cdk2 complexes and, in turn, in inhibition of the associated kinase activity.
The effect of mutant SHP-1 on p27 Kip1 and cdk2-associated kinase activity suggests that mutant SHP-1 may affect the extent of pRb phosphorylation. Immunoblotting of pRb showed that only the hyperphosphorylated inactive form of pRb was detected in CHO/sst2-SHP-1(C453S) grown in serum-starved conditions and that mutant SHP-1 prevented the mobility shift from hypo-to hyperphosphorylated pRb as well as from hyperto hypophosphorylated after growth in the presence of insulin and RC-160, respectively (Fig. 8C).
To further examine the role of SHP-1 in p27 Kip1 regulation, experiments were performed using isolated pancreatic acini from viable motheaten (me v ) mice. These animals express negligible SHP-1 catalytic activity consequent to loss-of-function mutation in the gene encoding SHP-1 (35). Pancreatic acini from normal as well as me v mice expressed a high level of sst2 receptors as revealed by anti-sst2 immunoblotting. 2 However the p27 Kip1 protein level was decreased by about 50% in the me v pancreatic acinar cells as compared with normal acini (Fig.  9). We observed that incubation of normal acini for 3 h with 1 nM epidermal growth factor (EGF) at 25°C induced a downregulation of p27 Kip1 . This effect was reversed by the addition of 1 nM RC-160, in agreement with the mitogenic effect of EGF and the antiproliferative effect of somatostatin on pancreatic acini (36). In contrast, when acini were isolated from me v mice, EGF down-regulated p27 Kip1 as observed in control acini, but RC-160 had no more significant effect on the level of p27 Kip1 (Fig. 9). breast cancer cells and AtT-20 mouse pituitary tumor cells (5,37). All these cell types express multiple somatostatin receptor subtypes, and the involvement of each receptor subtype in somatostatin response remains to be clarified. We have previously demonstrated the role of sst2 in the somatostatin-mediated inhibition of cell growth and the involvement of SHP-1 in the transduction of the inhibitory growth signal (8 -10, 38). The present investigation was undertaken to further delineate the basis of sst2-mediated control of cell cycle machinery. We have demonstrated that in CHO cells expressing sst2, the inhibition of proliferation in response to somatostatin analogue results from the suppression of cell cycle progression and the arrest of cells in the G 0 /G 1 phase, which correlates with an increase in expression of p27 Kip1 but not p21 Waf1/Cip1 . This is accompanied by an increase of association of p27 Kip1 with cdk2, a concomitant inhibition of cyclin E-cdk2 activity and a consequent decrease in the phosphorylation of pRb that precedes the inhibition of entry into S phase. On the other hand, the data presented provide strong support for the involvement of the tyrosine phosphatase SHP-1 in maintaining cell quiescence as well as sst2-induced cell cycle arrest. The expression of dominant negative SHP-1 is sufficient to induce G 1 /S transition, allowing mitogen-independent cell proliferation and abrogating the inhibitory effect of somatostatin. Consistent with its effect on cell cycle progression, expression of dominant negative SHP-1 down-regulates p27 Kip1 , which results in a decrease of association of p27 Kip1 , an increase of cyclin E/cdk2 kinase activity, and a subsequent inactivation of the growth-suppressive function of pRb protein, thus linking SHP-1 to control of p27 Kip1 expression.
The pivotal role of p27 Kip1 in controlling cdk function and, thus, cell cycle progression is well established. p27 Kip1 mediates cell cycle arrest in response to various antimitogenic signals, including transforming growth factor ␤, rapamycin, cAMP, cell-cell contact, and anti-epidermal growth factor antibody (18, 39 -41). Overexpression of p27 Kip1 leads to cell cycle arrest (17) and inhibition of both normal and transformed human mammary epithelial cell growth (42). Antisense inhibition of p27 Kip1 expression can prevent quiescence upon withdrawal of growth factor (25,26). In this study, insulin treatment of CHO/sst2 cells reduces the level of p27 Kip1 , as expected for mitogenic signaling pathways. In the presence of soma-  tostatin analogue, p27 Kip1 is rapidly up-regulated, and this induction precedes somatostatin-induced G 1 cell cycle arrest. Such a G 1 blockade has not yet been observed in cells expressing sst3, because it has been reported that somatostatin induces a decrease of G 0 /G 1 and an increase in S phase via sst3 (7). These results argue in favor of a receptor subtype selectivity for somatostatin-mediated regulation of cell cycle progression. In contrast to the ability of interferon ␥ (43) and transforming growth factor-␤ (34) to up-regulate both p27 Kip1 and p21 Waf1/Cip1 cdk inhibitors, p21 Waf1/Cip1 does not appear to play a role in the sst2-mediated antimitogenic effect of somatostatin. However, we cannot rule out the possibility that other cdk inhibitors may be involved in this effect.
p27 Kip1 associates with G 1 -specific cyclin-cdk complexes and inhibits their catalytic activities. Among them, cyclin E-cdk2 p27 Kip1 governs cdk2 activity, which is required for G 1 progression and contributes to the phosphorylation and inactivation of the pRb protein. As reported for mitogenic factors, our results demonstrate that after insulin treatment, the down-regulation of p27 Kip1 protein level is accompanied by an increase of cyclin E-and cdk2-associated kinase activity, which is related to a decrease of the amount of p27 Kip1 associated with cdk2, thereby providing conditions favorable for entry into S phase (21,31). In contrast, somatostatin analogue decreases the amount of p27 Kip1 associated with cyclin E-cdk2 complexes, inactivates cyclin E-cdk2 complexes, and induces the disappearance of the hyperphosphorylated form of pRb protein. It has been reported that p27 Kip1 inactivates cyclin E-cdk2-associated kinases (23), and conversely, inactivation of cyclin E-cdk2-associated kinases can lead to accumulation of p27 Kip1 , because cyclin E-cdk2-associated kinases phosphorylate p27 Kip1 and induce its destruction by the ubiquitin pathway (21). Somatostatin also decreases cyclin E protein level. However, cyclin E overexpression does not prevent cell cycle exit (44), and there is no significant relationship between cyclin E expression and cyclin E-cdk2 activity (45). It is likely that the negative regulation by somatostatin of the cdk2 activity could be one of the consequences of the rise of p27 Kip1 , leading to an increased interaction of p27 Kip1 with cyclin E-cdk2 complexes and, thus, resulting in an inhibition of their activity.
A critical finding of the present study is that unlike CHO cells expressing sst2, which accumulate in G 0 /G 1 , CHO cells coexpressing sst2 and a dominant negative mutant SHP-1 remain distributed through the cell cycle after serum withdrawal, with a high portion of cells being in the S phase. These cells express a constitutively low level of p27 Kip1 ; this in turn leads to an activation of cyclin E-cdk2 complexes and accumulation of hyperphosphorylated pRb. Furthermore, the expression of dominant negative SHP-1 circumvents the requirement for insulin in G 1 cell cycle progression and is associated with a resistance to the antiproliferative effect of somatostatin. Ours results provide strong support for the hypothesis that SHP-1 is required to revert mitogen-induced down-regulation of p27 Kip1 and clearly demonstrate the importance of SHP-1 in sst2 signaling for blockade of p27 Kip1 down-regulation. This hypothesis is strengthened by our results, obtained with acinar cells isolated from me v mice that express a high level of sst2 and a defective SHP-1 (35). 2 As observed in CHO/sst2 cells expressing a mutant SHP-1, acini from me v mice express a low level of p27 Kip1 that can be no longer up-regulated by somatostatin, demonstrating the importance of SHP-1 in the retention of high levels of p27 Kip1 and its central role as the downstream target of the sst2 signaling pathway leading to up-regulation of p27 Kip1 . The functional role of SHP-1 in pancreatic cells is not FIG. 8. Effect of insulin and RC-160 on cell cycle proteins, cyclin E-and cdk2-associated kinase activity, and p27 Kip1 -cdk2 association in CHO/sst2-SHP1 (C453S) cells. Serum-starved CHO/ sst2-SHP1 (C453S) cells were incubated at 37°C for 3 h with 0.1 M insulin and with or without 1 nM RC-160 or were not treated (control) and solubilized. A, soluble proteins were subjected to immunoprecipitation (Ip) with anti-cdk2 or anticyclin E antibodies. cdk2 kinase activity was assayed in cdk2 or cyclin E immunoprecipitates using histone H1 as substrate, followed by SDS-PAGE and autoradiography. B, to detect the amount of p27 Kip1 associated with cdk2, cells were subjected to immunoprecipitation with anti-p27 Kip1 antibodies. Immunoprecipitates were resolved by SDS-PAGE and analyzed by immunoblotting (Blot) with anti-cdk2 antibodies. The arrow indicates the position of cdk2. Immunoblots were analyzed densitometrically, and the data were plotted as percentages of control values obtained from cells incubated in serum-free ␣MEM. Data from three separate experiments are presented as mean ϮS.E. C, soluble proteins were subjected to SDS-PAGE and immunoblotted with anti-pRb antibodies to detect pRb. The upper band represents the hyperphosphorylated (pRb phos ), whereas the lower band (pRb) represents the hypophosphorylated form of protein pRb. known, but the demonstration that SHP-1 is necessary for regulation of p27 Kip1 suggests that SHP-1 may be important for the pancreatic cell development. It is notable that EGF downregulates p27 Kip1 in acini from me v mice, suggesting that SHP-1 is not the only negative regulatory protein-tyrosine phosphatase in growth factor signaling or that another proteintyrosine phosphatase can substitute for SHP-1 when it is not functional. Most of the previous studies have focused on the role of SHP-1 in response of quiescent cells to mitogenic stimulation. We and others identified SHP-1 as a critical negative regulator of cytokine as well as growth factor signaling; the recruitment of this enzyme to activated membrane receptors causes dephosphorylation of the receptors or/and of downstream signaling molecules (10,46). The results presented here extend these observations by demonstrating that SHP-1 is necessary for negative regulation of cell cycle progression in the G 1 phase and is a key mediator of sst2 signaling pathway in controlling high inhibitory levels of p27 Kip1 .
Regulation of p27 Kip1 occurs by different mechanisms, including transcriptional, post-transcriptional, or post-translational mechanisms. Indeed, post-transcriptional mechanisms have been implicated in the up-regulation of p27 Kip1 protein induced by antimitogen-transforming growth factor-␤ (47) and interferon ␥ (48), whereas post-translational mechanisms involving desequestration of p27 Kip1 as a consequence of downregulation of the N-myc gene (49) or degradation of the protein (50) have been proposed for the retinoic acid-and lovastatininduced increases in p27 Kip1 , respectively. It has been reported that the mitogen-induced decrease in p27 Kip1 expression occurs through posttranslationally regulated protein degradation via the ubiquitin-proteasome pathway (20). Recent data demonstrated the key role for Ras signaling pathway in the downregulation of p27 Kip1 and the involvement of RhoA in ubiquitinmediated p27 Kip1 degradation (51). However, a regulation at a transcriptional level can also occur, as shown in v-Src oncoprotein-transformed cells, v-Src reducing the level of p27 Kip1 mRNA and preventing cellular quiescence (52). The mechanisms involved in the SHP-1-induced up-regulation of p27 Kip1 remain to be elucidated. We previously demonstrated that SHP-1 is associated with activated insulin receptor and is involved in down-regulation of insulin signaling. In addition, upon sst2 stimulation, somatostatin negatively regulates insulin signal transduction by controlling first the recruitment of SHP-1 to insulin receptor and its activation and then causing a dephosphorylation and an inactivation of insulin receptor and its substrates, thus leading to an inhibition of the insulin downstream signaling (12). In agreement with these results, activated SHP-1 may regulate the level of p27 Kip1 as a consequence of SHP-1-induced dephosphorylation of insulin receptors and blockade of the insulin-induced catalytic cascade leading to down-regulation of p27 Kip1 . However, other downstream effectors of somatostatin-activated SHP-1, different from growth factor receptors and not yet identified, could also be involved in the SHP-1-induced p27 Kip1 regulation.
In conclusion, this investigation shows that activation of sst2 promotes cell growth arrest through the ability of somatostatin to maintain high levels of p27 Kip1 and inactivate cyclin E-cdk2 complexes, thus leading to hypophosphorylation of pRb. Our findings provide evidence that SHP-1 may be required for accumulation of p27 Kip1 and inhibition of cell cycle progression and indicate that SHP-1 is a key mediator of sst2-induced p27 Kip1 up-regulation and subsequent cell cycle arrest.