Reentry into the cell cycle of contact-inhibited vascular endothelial cells by a phosphatase inhibitor. Possible involvement of extracellular signal-regulated kinase and phosphatidylinositol 3-kinase.

Vascular endothelial cells are unique in that they exit from the cell cycle when they come into contact with each other. Although the phenomenon is called "contact inhibition," little is known about the cellular mechanisms involved. Here we show that the phosphatase inhibitor sodium orthovanadate (SOV) induced the reentry of contact-inhibited human umbilical vascular endothelial cells (HUVECs) into the cell cycle and that reentry was associated with activation of the extracellular signal-regulated kinase (ERK) and phosphatidylinositol 3-kinase (PI 3-K)/Akt pathways. SOV stimulated [(3)H]thymidine uptake of contact-inhibited HUVECs in a time- and dose-dependent manner. SOV-induced increase in [(3)H]thymidine uptake was significantly inhibited by the mitogen-activated protein kinase kinase inhibitor PD98059 and by the PI 3-K inhibitor LY294002. SOV also stimulated the expression of cyclin D1, cyclin E, and cyclin A, and the activity of CDK2 kinase, whereas it decreased the expression of p27(kip1). In marked contrast, growth media alone did not induce these changes. Furthermore, these SOV-induced changes were abolished by pretreatment with PD98059 and LY294002. SOV stimulated phosphorylation of ERK and Akt in contact-inhibited HUVECs, while growth media alone did not. This phosphorylation was associated with inhibition of phosphatase activity in the cells. Finally, overexpression of high cell density-enhanced protein tyrosine phosphatase 1 inhibited c-fos and cyclin A promoter activity. Taken together, our results suggest that in contact-inhibited HUVECs, increased phosphatase activity suppressed the ERK and PI 3-K/Akt pathways, resulting in exit from the cell cycle by down-regulation of cyclin D1, cyclin E, and cyclin A and by up-regulation of p27(kip1).

Vascular endothelial cells (ECs) 1 play a variety of pathophysiological roles such as provision of a barrier through which substances are transported into vessel walls, maintenance of vascular tone by releasing vasoactive substances including nitric oxide and endothelin, and oxidization of lipoproteins (1). ECs are unique because they grow as a strict monolayer, and they exit from the cell cycle once they come into contact with each other. Although the phenomenon is well known and called "contact inhibition," little is known about the molecular mechanisms involved.
Several lines of evidence suggest that up-regulation of phosphatase activity may be implicated in density-dependent growth arrest. NRK-1 cells were transformed by treatment with the phosphatase inhibitor sodium orthovanadate (SOV), and transformation was accompanied by increases in protein phosphorylation in the cells (2). Protein-tyrosine phosphatase (PTPase) activity was increased in human umbilical vein endothelial cells (HUVECs) harvested at high density (3). The increase in PTPase activity was also observed in Swiss 3T3 fibroblasts whose growth was arrested at high density (4). Recently, a novel class of receptor-like PTPases was isolated and named high cell density-enhanced PTPase 1 (DEP-1). DEP-1 has an extracellular domain that contains eight fibronectin type III motifs, a single transmembrane domain, and a single intracellular PTPase domain. The expression level and the PTPase activity of DEP-1 were increased in WI-38 and AG1518 cells harvested at high density (5). Although these findings suggest that increased PTPase activity in the cells might counteract with protein phosphorylation, which leads to cell proliferation, little is known as to which intracellular signaling pathways are affected by the increase in PTPase activity and how the increased PTPase activity finally affects the cell cycle regulatory machinery.
Cell cycle progression is regulated by serine/threonine kinases termed cyclin-dependent kinases (CDKs), the activities of which oscillate during the cell cycle. CDKs are associated with the positive coactivators cyclins and the negative regulators, CDK inhibitors (6,7). In mammalian cells, cyclin D-CDK4/CDK6, cyclin E-CDK2, cyclin A-CDK2, and cyclin B-Cdc2 are the main cyclin-CDK complexes that regulate the progression of G 1 , G 1 /S, S, and G 2 /M phases, respectively. CDK inhibitors comprise two families, the Ink4 and Cip/Kip families. CDK inhibitors of the Cip/Kip family are of particular interest in that they inhibit the activity of a broader spectrum of CDKs including CDK2, CDK4, and CDK6 (8). The Cip/Kip family is composed of p21 waf1/cip1 , p27 kip1 , and p57 kip2 (9 -12). The expression level of p27 kip1 is reportedly increased in growth-arrested cells by contact inhibition or by stimulation with transforming growth factor-␤ (13)(14)(15)(16)(17).
Recent evidence suggests that several intracellular signaling pathways are linked to the cell cycle regulatory machinery. Among them, the p21 ras (RAS)/mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK) pathway is the most characterized pathway. It is reported that constitutively active MEK is sufficient to transform cells or induce differentiation (18,19), suggesting that the MEK/ERK pathway is implicated in the cell cycle regulation. Several studies have indicated, using a dominant-negative Ras mutant, that the Ras signaling pathway is involved in the induction of cyclin D1 protein, CDK4 kinase activity, and CDK2 kinase activity and in the down-regulation of p27 kip1 (20,21). Furthermore, the ERK pathway is reportedly implicated in vascular endothelial cell growth factor-induced EC proliferation by stimulating cyclin D1 synthesis and CDK4 kinase activity (22). Protein kinase C (PKC) stimulates cell proliferation or induces cell cycle arrest in vascular ECs, which appears to depend upon the timing of PKC activation during the cell cycle and PKC isozymes expressed in the cells (23)(24)(25)(26). The PI 3-Kmediated pathways also seem to be implicated in cell cycle progression, because a retrovirus-encoded PI 3-K could transform fibroblasts (27). However, it remains unclear as to how these intracellular signaling pathways are regulated in contact-inhibited vascular ECs.
To investigate the cellular mechanisms for density-dependent growth arrest and to apply those mechanisms to the regulation of the growth of other cell types, such as vascular smooth muscle cells, we examined in the present study whether treatment with the phosphatase inhibitor SOV induced reentry of contact-inhibited vascular ECs into the cell cycle by examining [ 3 H]thymidine incorporation, expression levels of cyclins, CDKs and CDK inhibitors, and CDK2 kinase activity. We also studied the effects of inhibition of the ERK-, PKC-, and PI 3-K-mediated pathways on SOV-induced cell cycle reentry. Finally, we examined the effect of overexpression of DEP-1 on c-fos and cyclin A promoter activity.
Cell Culture-HUVECs were maintained in medium 199 containing 20% fetal bovine serum, 100 g/ml EC growth supplement, and 50 units/ml heparin. BAECs were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. After reaching confluence, medium was replaced with fresh growth medium, and the cells were further incubated for 48 h. Cells were then stimulated with fresh growth medium in the presence and absence of SOV.
Preparation of Protein Extracts-For Western blot analyses and the CDK2 kinase assay, we used Nonidet P-40 cell lysis buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% Nonidet P-40) containing 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, and 2 g/ml aprotinin. Cells were washed twice with ice-cold phosphate-buffered saline (PBS) and collected by centrifugation. Cells were then lysed in Nonidet P-40 cell lysis buffer for 30 min on ice. After centrifugation, the supernatant was stored at Ϫ80°C. For the ERK1 and Akt kinase assay, we used Triton X-100 cell lysis buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1% Triton X-100, 10% glycerol) containing 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, and 2 g/ml aprotinin. Cells were lysed in the buffer for 30 min on ice and centrifuged for 10 min at 4°C. The cleared supernatant was used for the ERK1 and Akt kinase assay. Protein concentration was measured according to Bradford's method (Bio-Rad).
Western Blot Analysis-Protein extracts were separated on 10% SDS-polyacrylamide gels and transferred onto nylon membranes (Millipore Corp., Bedford, MA) using a semidry blotting system (Amersham Pharmacia Biotech, Uppsala, Sweden). After blocking in 1ϫ PBS, 5% nonfat dry milk, 0.2% Tween 20 at 4°C overnight, the membranes were incubated with the primary antibodies in blocking buffer (1ϫ PBS, 2% nonfat dry milk, 0.2% Tween 20) for 1 h at room temperature. Antibodies were used at a dilution of 1:100, except for phosphospecific anti-ERK1/2 antibody and phosphospecific anti-Akt antibody, which were diluted at 1:500. The membranes were washed three times with the blocking buffer and then incubated with secondary antibodies, which were conjugated with horseradish peroxidase (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom) at a final dilution of 1:7,000. After final washes with 1ϫ PBS, 0.2% Tween 20, the signals were detected using ECL chemiluminescence reagents (Amersham Pharmacia Biotech).
In Vitro Kinase Assays-For the CDK2 kinase assay, 75 g of each protein extract was precleaned with protein A-agarose beads (Roche Molecular Biochemicals) for 1 h at 4°C in the Nonidet P-40 cell lysis buffer. The extracts were then incubated with 1 g of anti-CDK2 antibody for 1 h at 4°C and with protein A-agarose beads for another 1 h at 4°C with constant rocking. After centrifugation, the pellets were washed twice with the Nonidet P-40 cell lysis buffer and then three times with a kinase buffer (50 mM Tris-HCl (pH 8.0), 10 mM MgCl 2 , 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, and 2 g/ml aprotinin). The pellets were then incubated in 30 l of the kinase buffer containing 3 g of histone H1, 10 M ATP, and 10 Ci of [␥-32 P]ATP for 30 min at room temperature. The reactions were terminated by adding 30 l of 2ϫ SDS loading buffer. After boiling for 5 min, half of the samples were separated on a 12% SDS-polyacrylamide gel. The ERK1 kinase assay was also performed in the same way. Fifty g of each protein extract was precleaned in the Triton X-100 cell lysis buffer in the presence of protein A-agarose beads, and the extract was incubated with 1 g of anti-ERK1 antibody. After incubation with protein A-agarose beads, the beads were washed twice with Triton X-100 cell lysis buffer and three times with the kinase buffer. The pellets were incubated in 20 l of the kinase buffer containing 1 g of myelin basic protein, 10 M ATP, and 10 Ci of [␥-32 P]ATP, for 30 min at room temperature. Half of the samples were separated by electrophoresis using a 12% SDS-polyacrylamide gel. For the Akt kinase assay, 100 g of each protein extract was precleaned in the Triton X-100 cell lysis buffer, and the extract was incubated with 1 g of anti-Akt antibody. After incubation with protein A/G PLUS agarose beads (Santa Cruz Biotechnology), the beads were washed twice with Triton X-100 cell lysis buffer and three times with the kinase buffer. The pellets were incubated in 20 l of the kinase buffer containing 2 g of histone H2B, 10 M ATP, and 10 Ci of [␥-32 P]ATP for 30 min at room temperature. Half of the samples were separated by electrophoresis using a 15% SDS-polyacrylamide gel. To confirm that almost equal amounts of CDK2, ERK1, or Akt protein were used for the kinase assays, the same amounts of the protein extracts were immunoprecipitated with anti-CDK2, -ERK1, or -Akt antibody and immunoblotted with the same antibody as the internal controls. The signals were visualized using an image analyzer (BAS 2000, Fuji Film, Tokyo) and the intensities of the signals were analyzed by densitometry.

Measurement of [ 3 H]Thymidine Incorporation-Confluent
HUVECs were maintained in growth medium for 48 h and restimulated with fresh growth medium for 8, 16, and 24 h in the presence and absence of 50 M SOV. [ 3 H]Thymidine (2 Ci/ml; Amersham Pharmacia Biotech) was added to each well 2 h before the end of the incubation period. Cells were washed twice with ice-cold 1ϫ PBS and incubated with ice-cold 10% trichloroacetic acid for 30 min. After washing them twice with distilled water, the cells were lysed with 0.2 N NaOH, neutralized with 0.2 N HCl, and subjected to liquid scintillation counting.
Measurement of Phosphatase Activity-Phosphatase activity was measured as previously reported (28). In brief, protein extracts were incubated in 200 l of buffer containing 50 mM imidazole (pH 7.5), 0.1% ␤-mercaptoethanol, and 10 mM p-nitrophenyl phosphate, for 10 min at room temperature. The reaction was stopped by adding 800 l of 0.25 N NaOH. Absorbance at 410 nm was measured to estimate the amount of hydrolyzed p-nitrophenyl phosphate.
Plasmids-The 5Ј-flanking region of the mouse c-fos gene was amplified by polymerase chain reaction. One g of mouse genomic DNA (Promega, Madison, WI) was subjected to polymerase chain reaction using LA Taq DNA polymerase (Takara shuzo, Osaka, Japan). The polymerase chain reaction conditions were 1 min at 95°C, 1 min at 63°C, and 1 min at 72°C for 35 cycles, with final extension for 10 min at 72°C. The polymerase chain reaction-amplified product was digested with SacI and XhoI and subcloned in the pGL2 vector (pGL2-c-fos). Primer sequences used for the reaction were as follows: sense primer, 5Ј-GAGCTCTTGCTTCTCCTAATACCAGAGACT-3Ј; antisense primer: 5Ј-CTCGAGTGCAGTCGCGGTTGGAGTAGTA-3Ј.
The nucleotide sequence of the construct was confirmed by cycle sequencing using an ABI PRISM 310 Genetic Analyzer (Perkin-Elmer). pGL2-cyclin A/Ϫ3200/ϩ245 encodes approximately 3.2 kilobase pairs of the promoter region of the human cyclin A gene upstream of the luciferase gene (29). pME18S-mouse DEP-1 encodes the full length of mouse DEP-1 cDNA in the expression vector pME18S (30). pcDNA3-HA-mouse RasS17N encodes the amino-terminally HA-tagged mouse Ras cDNA in which Ser 17 was replaced with Asn, and pcDNA3-HA-mouse RasG12V encodes the amino-terminally HA-tagged mouse Ras cDNA in which Gly 12 was replaced with Val, as described previously (31). pRL-TK, which encodes the sea pansy luciferase gene, was purchased from Toyo Ink (Tokyo, Japan) and used as the internal control for the luciferase assays.
Statistical Analyses-The values are the mean Ϯ S.E. The effects of SOV on cyclins/CDKs/CDK inhibitors, kinase activities, and [ 3 H]thymidine uptake were assessed using analysis of variance followed by the Student-Neumann-Keul test. Differences with a p value of Ͻ0.05 were considered statistically significant.  Fig. 2A, left panels). SOV did not affect the expression of CDK2 and CDK4. On the contrary, growth medium without SOV did not affect the expression of those factors ( Fig. 2A, right panels). We next examined CDK2 kinase activity using histone H1 as the substrate. The basal CDK2 kinase activity was negligible. SOV stimulated CDK2 kinase activity in a time-dependent manner. CDK2 kinase activity was substantially increased 16 h (3.95 Ϯ 1.15-fold increase versus 0 h, p Ͻ 0.05, n ϭ 4) and 24 h (5.66 Ϯ 1.65-fold increase versus 0 h, p Ͻ 0.05, n ϭ 4) poststimulation (Fig. 2B, left panels). In contrast, growth medium alone did not induce CDK2 kinase activity (Fig. 2B, right panels). Collectively, the results described above suggest that growth medium containing 50 M SOV induced reentry of contact-inhibited HUVECs into the cell cycle, whereas growth medium alone did not.

Effects of MEK and PI 3-K Inhibition on SOV-induced Cell Cycle Reentry in Contact-inhibited HUVECs-
The results described above suggested that inhibition of phosphatases by SOV might result in stimulation of phosphorylation of a variety of proteins that are involved in transmitting mitogenic signals. We therefore examined the intracellular signaling pathways that might be implicated in reentry of contact-inhibited HU-VECs into the cell cycle induced by SOV. We studied the effects of the MEK1/2 inhibitor PD98059, the PKC inhibitor calphostin C, and the PI 3-K inhibitor LY294002 on dose-dependent fashion (Fig. 1B). We also used another PI 3-K inhibitor, wortmannin, and obtained basically the same results (data not shown). Calphostin C, at a concentration of 100 nM, which seems to be sufficient to inhibit PKC (32), tended to inhibit [ 3 H]thymidine uptake; however, the difference was not statistically significant. Calphostin C, at a concentration of 200 nM, significantly inhibited [ 3 H]thymidine uptake (data not shown). However, some cells were rounded up and detached in 48 h, probably because of its cytotoxic effects at that concentration. We also used GF109203X, another PKC inhibitor. GF109203X, at concentrations varying from 0 to 3.5 M, did not significantly inhibit [ 3 H]thymidine uptake (data not shown). Overall, PKC inhibitors used in the experiments did not significantly inhibit SOV-induced increase in [ 3 H]thymidine uptake in contact-inhibited HUVECs. We therefore focused on the effects of MEK and PI 3-K inhibition on SOV-induced reentry of contact-inhibited HUVECs into the cell cycle.
We next examined the effects of MEK and PI 3-K inhibition on the protein expression levels of cyclin D1, cyclin E, cyclin A, and p27 kip1 , because their expression was significantly affected by SOV treatment (Fig. 2A). The basal expression of these factors was not affected by pretreatment with PD98059 or LY294002. SOV-induced increase of cyclin D1, cyclin E, and cyclin A was significantly inhibited by pretreatment with 50 M PD98059 (Fig. 3A) or 50 M LY294002 (Fig. 3B), while SOVinduced decrease of p27 kip1 was restored by these inhibitors. We also examined the effects of MEK and PI 3-K inhibition on SOV-induced activation of CDK2 kinase. The basal CDK2 kinase activity was not affected by pretreatment with PD98059 or LY294002. SOV-induced increase of CDK2 kinase activity was significantly inhibited to 26.5 Ϯ 12.6% (24 h poststimulation, p Ͻ 0.01, n ϭ 3) and 19.8 Ϯ 15.3% (24 h poststimulation, p Ͻ 0.01, n ϭ 3), respectively, by pretreatment with 50 M PD98059 (Fig. 4A) and 50 M LY294002 (Fig. 4B).
Activation of ERK and Akt by SOV Treatment in Contactinhibited HUVECs-The results described above suggest that the ERK-and PI 3-K-mediated pathways were activated by SOV in contact-inhibited HUVECs, resulting in the activation of the cell cycle regulatory machinery. To confirm this hypothesis, we examined whether ERK and Akt, a major downstream target of PI 3-K, were activated by SOV in contact-inhibited HUVECs. Neither ERK nor Akt was significantly phosphorylated under basal conditions. SOV, at a concentration of 50 M, stimulated phosphorylation of ERK1/2 and Akt in a time-dependent manner. Phosphorylation of ERK1/2 and Akt peaked around 15 min poststimulation (Fig. 5A, left panels). In marked contrast, neither ERK nor Akt was significantly phosphorylated in the absence of SOV in contact-inhibited HUVECs, although fresh growth medium was added (Fig. 5A, right panels), suggesting that both pathways were reversibly shut down in contact-inhibited HUVECs. We also examined the kinase activities of ERK1 and Akt using myelin basic protein and histone H2B, respectively, as the substrate. SOV, at a concentration of 50 M, stimulated the kinase activities of ERK and Akt in a time-dependent fashion (Fig. 5B, left panels). The kinase activities of ERK1 and Akt peaked around 15 min poststimu-

FIG. 2. SOV-induced changes of cell cycle regulatory factors and CDK2 kinase activity in contact-inhibited HUVECs.
A, time course of SOV-induced changes in the expression levels of cell cycle regulatory factors. Confluent HUVECs were stimulated with growth medium in the presence (SOV) and the absence (GM) of 50 M SOV for the indicated periods. Seventy-five g of each protein extract was immunoblotted with the indicated antibodies. The same protein extracts were used for all of the immunoblots. Note that some photographs in the right panels were exposed longer to show the faint bands. B, time course of SOV-induced changes in CDK2 kinase activity. Experiments were performed in the same way as in A. Seventy-five g of each protein extract was immunoprecipitated with anti-CDK2 antibody, and CDK2 kinase activity was measured using histone H1 as the substrate (upper panels). The same amounts of each protein extract were immunoprecipitated and immunoblotted with anti-CDK2 antibody as the internal control (lower panels). lation. In contrast, the kinase activities of neither ERK1 nor Akt were significantly increased in the absence of SOV (Fig.  5B, right panels). Because the time course of their kinase activities was basically paralleled with that of their phosphorylation, we examined the phosphorylation of ERK and Akt to show their activities in subsequent experiments. We next examined whether PD98059 and LY294002 indeed inhibited the ERK-and PI 3-K-mediated pathways, respectively. SOV-induced phosphorylation of ERK and Akt was completely inhibited by pretreatment with PD98059 (Fig. 6A, left panels) and LY294002 (Fig. 6B, right panels), respectively. We also asked whether PD98059 inhibited Akt phosphorylation or LY294002 inhibited ERK phosphorylation, because both compounds inhibited SOV-induced cell cycle reentry to similar extents. However, PD98059 did not inhibit Akt phosphorylation (Fig. 6B, left  panels), nor did LY294002 inhibit ERK phosphorylation (Fig.  6A, right panels). We also examined whether SOV indeed inhibited phosphatase activity in confluent HUVECs. Phosphatase activity was assessed by measuring hydrolysis of p-nitrophenyl phosphate. Absorbance at 410 nm changed linearly with the amount of protein input, which varied from 0 to 200 g (Fig.  7A). We measured phosphatase activity in the presence and absence of 200 M SOV to calculate SOV-sensitive phosphatase activity. SOV-sensitive phosphatase activity was calculated by subtracting SOV-resistant phosphatase activity (phosphatase activity in the presence of SOV) from total phosphatase activity (phosphatase activity in the absence of SOV). SOV-resistant phosphatase activity was estimated to be approximately 20% of total phosphatase activity. As shown in Fig. 7B, SOV-sensitive phosphatase activity was suppressed to 58% (p Ͻ 0.01) of the control level 15 min after SOV treatment. The down-regulation of phosphatase activity lasted at least for 2 h.

FIG. 3. Effects of PD98059 (PD) and LY294002 (LY) on SOVinduced changes in the expression of cell cycle regulatory factors in contact-inhibited HUVECs.
Overexpression of DEP-1 Inhibits c-fos and Cyclin A Promoter Activity in BAECs-To examine whether increased phos-phatase activities could affect the intracellular signaling pathway and inhibit cell cycle progression in vascular ECs, vascular ECs were transiently co-transfected with the luciferase gene, the expression of which was driven by the c-fos or cyclin A promoter, and mouse DEP-1 cDNA. In this case, we used BAECs instead of HUVECs, because transfection efficiency was poor in HUVECs. We used c-fos promoter activity as an indicator for the activity of the Ras/MEK/ERK-dependent pathway. It is also reported that c-fos promoter activity was increased via the PI 3-K-dependent pathway (33). As shown in Fig. 8A, c-fos promoter activity was significantly suppressed to 57% of the control level by co-transfection with DEP-1 (p Ͻ 0.0001). The activity was restored by co-transfection with the constitutively active Ras mutant RasG12V. We also examined the effect of a dominant-negative Ras mutant on c-fos promoter activity. The RasS17N mutant inhibited c-fos promoter activity to 32% of the control level (p Ͻ 0.0001). We next examined the effect of DEP-1 overexpression on cyclin A promoter activity. We used this promoter activity as an indicator for S phase progression, because a previous report showed that cyclin A promoter activity was strikingly down-regulated in contactinhibited BAECs (29). As shown in Fig. 8B, cyclin A promoter activity was significantly inhibited to 42% of the control level by co-transfection with DEP-1 (p Ͻ 0.001). The suppression was overcome by co-transfection with RasG12V. The RasS17N mutant inhibited cyclin A promoter activity to 5% of the control

FIG. 5. SOV-induced changes in the phosphorylation levels and kinase activities of ERK and Akt in contact-inhibited HU-VECs.
A, SOV-induced phosphorylation of ERK and Akt in contactinhibited HUVECs. Confluent HUVECs were stimulated with fresh growth medium in the presence (SOV) and absence (GM) of 50 M SOV for the indicated periods. One hundred g of each protein extract was immunoblotted with anti-phosphospecific ERK antibody (P*-ERK) or anti-phosphospecific Akt antibody (P*-Akt). Fifty g of each protein extract was immunoblotted with anti-ERK antibody (ERK) and anti-Akt antibody (Akt), which recognize total ERK and Akt, respectively, regardless of whether or not they are phosphorylated. Shown are the results of a representative experiment among three independent experiments in which the same results were obtained. B, SOV-induced ERK1 and Akt kinase activities in contact-inhibited HUVECs. Confluent HU-VECs were stimulated with fresh growth medium in the presence (SOV) and absence (GM) of 50 M SOV for the indicated periods. Protein extracts (50 g each for the ERK1 kinase assay and 100 g each for the Akt kinase assay) were immunoprecipitated with anti-ERK1 antibody or anti-Akt antibody. ERK1 kinase assay and Akt kinase assay were performed using myelin basic protein (MBP) and histone H2B (HH2B), respectively, as the substrates. The same amounts of each protein extract were immunoprecipitated and immunoblotted with anti-ERK1 antibody or anti-Akt antibody as the internal control. The positions of immunoglobulin heavy chain (Ig) are indicated. level (p Ͻ 0.001). The result suggested that increased phosphatase activity could potentially inhibit cell cycle progression in vascular ECs, although the effect of DEP-1 appeared to be weaker than that of the dominant negative Ras mutant. DISCUSSION In the present study, we have shown that contact-inhibited HUVECs reentered the cell cycle in the presence of SOV by examining [ 3 H]thymidine incorporation, protein expression levels of cyclins, CDKs, and CDK inhibitors, and CDK2 kinase activity. We have also demonstrated that SOV-induced cell cycle reentry of contact-inhibited HUVECs was associated with activation of ERK and Akt and that SOV-induced cell cycle reentry was inhibited by pretreatment with the MEK1/2 inhibitor PD98059 and the PI 3-K inhibitor LY294002. Furthermore, we have shown that overexpression of DEP-1 inhibited c-fos and cyclin A promoter activity in BAECs, which was restored by co-expression of the constitutively active Ras mutant RasG12V. A previous report has suggested that at an appropriate concentration, SOV could transform cells and that transformation was associated with increased protein phosphorylation in the cells (2). However, little is known as to which intracellular signaling pathways are activated by SOV and how treatment with SOV is finally linked to the cell cycle regulatory machinery. Our results showed down-regulation of cyclin D1, cyclin E, cyclin A, and CDK2 kinase activity and up-regulation of p27 kip1 in contact-inhibited HUVECs. Our results also demonstrated that treatment of contact-inhibited HUVECs with SOV finally induced an increased expression of cyclin D1, cyclin E, and cyclin A, an increased CDK2 kinase activity, and down-regulation of p27 kip1 , all of which could potentially stimulate cells to reenter the cell cycle. Recent reports suggested that up-regulation of p27 kip1 was associated with density-dependent growth arrest or growth arrest induced by transforming growth factor-␤ (13)(14)(15)(16)(17). However, it is also reported that p27 kip1 (Ϫ,Ϫ) cells were growth-arrested by contact inhibition (34), suggesting that up-regulation of p27 kip1 was not the sole factor that induced growth arrest. In this regard, it is of inter-est to note that the transcript level and promoter activity of cyclin A were reduced in contact-inhibited BAECs (29). Thus, it is possible that the down-regulation of cyclin D1, cyclin E, and cyclin A contributes to the growth arrest by contact inhibition in HUVECs. Although p21 cip1 expression level was also increased by SOV, this was transient. It has been reported that p21 cip1 expression was transiently increased by mitogens (35). Our results are compatible with those of that report.
Our results indicated that inhibition of the MEK-and PI 3-K-mediated pathways resulted in suppression of SOV-induced up-regulation of cyclin D1, cyclin E, cyclin A, and CDK2 kinase activity and restoration of SOV-induced down-regulation of p27 kip1 . Several lines of evidence have suggested that the Ras/MEK/ERK-and PI 3-K-mediated pathways are linked to the cell cycle regulatory machinery. A constitutively active MEK could transform cells or induce differentiation, which depended upon cell types (18,19). In several studies, it was demonstrated, using a dominant negative Ras mutant, that Ras signaling pathways were involved in the up-regulation of cyclin D1, cyclin A, and CDK2 kinase activity and in downregulation of p27 kip1 (20,21). Involvement of MEK/ERK in cyclin D1 up-regulation was also suggested in a report in which PD98059 was used to inhibit platelet-derived growth factorinduced ERK activation (36). PI 3-K also appears to be implicated in cell cycle progression. A retrovirus-encoded PI 3-K could transform fibroblasts (27). PI 3-K may induce cell cycle progression via activation of p70 S6K , which seems to play roles in the initiation of protein synthesis (37,38). A recent report showed a direct link between Akt activation and stabilization of cyclin D1. In that report, glycogen synthase kinase-3␤ phosphorylated cyclin D1 on Thr 286 , which stimulated the degradation of cyclin D1. PI 3-K/Akt phosphorylated and inactivated glycogen synthase kinase-3␤, resulting in the stabilization of cyclin D1 (39). Furthermore, it was suggested that the PI 3-K-mediated pathway was implicated in p27 kip1 down-regulation by mitogens, because LY294002 and wortmannin restored p27 kip1 expression (21,40). Thus, our results were basically compatible with those of previous reports. Although PKC inhibition did not significantly affect the SOV-induced cell cycle reentry in our system, the involvement of PKC could not be excluded, because we did not examine which classes of PKC were indeed inhibited by calphostin C or GF109203X.
Our results suggested that the MEK/ERK and PI 3-K/Akt pathways were shut down in contact-inhibited HUVECs and that these pathways were reactivated in the presence of SOV. Furthermore, the reactivation of the MEK/ERK and PI 3-K/Akt pathways by SOV treatment correlated with the suppression of SOV-sensitive phosphatases, suggesting that SOV-sensitive phosphatases might be involved in the down-regulation of those pathways, although it was not clear whether the SOVsensitive phosphatases inhibited phosphorylation of molecules such as Shc, which leads to the activation of the MEK/ERK and PI 3-K/Akt pathways, or stimulated dephosphorylation of ERK and Akt or both. It should be noted that the phosphorylation of ERK was reportedly decreased in contact-inhibited vascular endothelial cells, while activation of Ras and MEK was not impaired (41). The results indicate that dephosphorylation of ERK was stimulated and that the pathways located upstream of ERK were intact in contact-inhibited vascular endothelial cells. Recent reports have suggested that PKC-, a PKC isozyme that belongs to a class of atypical PKC, directly phosphorylates and activates MEK, resulting in activation of ERK (42,43). It has also been reported that PKC-is a downstream target of PI 3-K (44,45). We therefore examined whether LY294002 inhibited ERK phosphorylation via inhibition of PI 3-K/PKC-. Our results, however, suggested that ERK phosphorylation occurred in the presence of LY294002, indicating that the involvement of PKC-was not the major pathway of activation of the ERK pathway by SOV. Our results also showed that PD98059 did not inhibit Akt phosphorylation, suggesting that activation of either the ERK pathway alone or the PI 3-K/Akt pathway alone with the inhibition of the other pathway was not sufficient to induce cell cycle reentry in contact-inhibited HUVECs.
To study the role of a specific phosphatase in vascular ECs, we examined the effects of DEP-1 overexpression on c-fos and cyclin A promoter activity in vascular ECs. Our results suggest that overexpression of DEP-1 could potentially inhibit cell cycle progression in vascular ECs. These results were compatible with those of a previous study in which growth of breast cancer cells was inhibited by overexpressed DEP-1 (46). Although we did not exclude the possibility that RasG12V inhibited DEP-1 activity, it is possible that the inhibitory effect of DEP-1 on cell cycle progression was mediated by, at least in part, suppression of the Ras-dependent pathways such as ERK and PI 3-K, because RasG12V restored the activity of the cyclin A promoter. Our results showed that the inhibitory effect of DEP-1 on cyclin A promoter activity was weaker than that of RasS17N. However, this did not mean that the role of phosphatases in densitydependent growth arrest was small, because DEP-1 was not the only phosphatase whose activity was increased in densitydependent growth arrest (3,4).
Taken together, our results and those of other authors (3)(4)(5)41) suggest the following scenario. The increased phosphatase activities in contact-inhibited vascular ECs cause down-regulation of the MEK/ERK and PI 3-K/Akt pathways, which in turn leads to the exit from the cell cycle due to down-regulation of cyclin D1, cyclin E, and cyclin A expression, and CDK2 kinase activity, and up-regulation of p27 kip1 expression. However, these changes are reversible, and once the phosphatases are inhibited, contact-inhibited vascular ECs reenter the cell cycle. Thus, phosphatases, especially those whose activities are increased in density-dependent growth arrest, can be used to control the growth of other cell types, such as vascular smooth muscle cells. Further studies are required to identify other specific phosphatases that are implicated in density-dependent growth arrest and to elucidate the mechanisms by which the MEK/ERK and PI 3-K/Akt pathways are down-regulated by phosphatases.

FIG. 8. DEP-1 inhibits c-fos and cyclin A promoter activity in BAECs.
A, effects of DEP-1 on c-fos promoter activity. BAECs were transiently transfected with 1.5 g of pGL2-c-fos and 0.5 g of pRL-TK, along with 1 g of pME18S (Vector), 1 g of pcDNA3-HA-mouse RasS17N (RasS17N), 1 g of pME18S-mouse DEP-1 (DEP-1) or 1 g of DEP-1 and 1 g of pcDNA3-HA-mouse RasG12V (RasG12V) using LipofectAMINE. Total amounts of DNA used for the transfection were adjusted to 4 g/well by adding pME18S. After 48 h, subconfluent cells were harvested for the luciferase assay. The vertical axis shows the ratio of Photinus pyralis luciferase activity to sea pansy luciferase activity. *, p Ͻ 0.001 versus vector (n ϭ 6). B, effects of DEP-1 on cyclin A promoter activity. The experiments were performed basically in the same way as in A except that pGL2-cyclin A/Ϫ3200/ϩ245 was used instead of pGL2-c-fos. *, p Ͻ 0.001 versus vector (n ϭ 6).