The Cyclin-dependent Kinase Inhibitor p21CIP/WAF Is a Positive Regulator of Insulin-like Growth Factor I-induced Cell Proliferation in MCF-7 Human Breast Cancer Cells*

To study the role of IGF-I receptor signaling on cell cycle events we utilized MCF-7 breast cancer cells. IGF-I at physiological concentrations increased the level of p21CIP/WAF mRNA after4has well as protein after8hby 10- and 6-fold, respectively, in MCF-7 cells. This IGF-1 effect was reduced by 50% in MCF-7-derived cells (SX13), which exhibit a 50% reduction in IGF-1R expression, demonstrating that IGF-1 receptor activation was involved in this process. Preincubation with the ERK1/2 inhibitor U0126 significantly reduced the IGF-1 effect on the amount of p21CIP/WAF protein in MCF-7 cells. These results were confirmed by the expression of a dominant negative construct for MEK-1 suggesting that the increase of the abundance of p21CIP/WAF in response to IGF-1 occurs via the ERK1/2 mitogen-activated protein kinase pathway. Using an antisense strategy, we demonstrated that abolition of p21CIP/WAF expression decreased by 2-fold the IGF-1 effect on cell proliferation in MCF-7. This latter result is explained by a delay in G1 to S cell cycle progression due partly to a reduction in the activation of some components of cell cycle including the induction of cyclin D1 expression in response to IGF-1. MCF-7 cells transiently overexpressing p21 showed increased basal and IGF-I-induced thymidine incorporation. Taken together, these results define p21CIP/WAF as a positive regulator in the cell proliferation induced by IGF-1 in MCF-7 cells.

Insulin-like growth factors are potent mitogens for the estrogen receptor-positive breast cancer cells including the MCF-7 human breast cancer cell line (1,2). The proliferative effect is mediated mainly by the insulin-like growth factor-I receptor (IGF-1R), 1 which can be inhibited by using a blocking antibody, ␣IR3 (3), or an antisense RNA (4) to this receptor. In vivo, some clinical and experimental data showed that the IGF-1R is involved in tumorigenesis of breast tissue (5). Furthermore, the tumorigenic potential of IGF-1R is usually dependent on the hyperactivation of IGF-1 signaling pathways.
The IGF-1R is a tyrosine kinase receptor, which is activated following ligand binding. Once autophosphorylated, IGF-1R binds and phosphorylates on tyrosine various substrates such as the insulin receptor substrates (IRS-1 to 4) (6 -9) and Shc (10). These substrates serve as docking molecules for other proteins containing SH2 domains including the p85 regulatory subunit of PI3K and Grb2 that lead to the activation of the two main signaling pathways, the PI3K/Akt (11) and the MAPK pathways (12). In the ER-positive cell lines, IRS-1 is the main substrate activated by the IGF-1R in the growth stimulatory effects of IGF-1 (13,14). The Shc pathway supports growth but also plays a role in the processes of cell motility and cell-cell aggregation (15). In MCF-7, the PI3K pathway but not the MAPK is crucial in the cell cycle progression induced by IGF-1 (1).
The mechanisms by which IGF-1 induces cellular proliferation involve a regulation of the cell cycle machinery. Progression through the cell cycle is regulated by sequential activation and subsequent inactivation of a series of cyclin-dependent kinases (CDKs) at different phases of the cell cycle (16,17). In normal cells, the transition from G 1 to S phase requires the activity of two classes of CDKs, CDK4/6, and CDK2. As cells emerge from quiescence in response to mitogenic stimuli, Dtype cyclins are synthesized and associate with CDK4/6. This cyclin D-CDK4/6 complex hyperphosphorylates the retinoblastoma protein (Rb), leading to its release from the transcription factor E2F (18,19). The free transcription factor E2F then activates the genes responsible for cellular proliferation including cyclin E. Cyclin E binds then to CDK2 contributing to kinase activation and G 1 -to-S phase progression. CDK activity is regulated by multiple mechanisms, including phosphorylation (20 -23) and association with both positive and negative regulatory proteins. CDKs activity can be inhibited by two different families of cyclin-dependent kinase inhibitors (CD-KIs). Members of the INK4 family, including p15 INK4b , p16 INK4a , p18 INK4c , and p19 INK4d bind specifically to monomeric CDK4/6 and prevent its association with a D-cyclin (24,25). Members of the WAF1/Cip1 family, including p21 WAF1/Cip1 , p27 kip1 , and p57 kip2 bind to G 1 cyclin-Cdk complexes and not to monomeric cyclins or CDKs (18,27). IGF-1 has been shown to regulate both the expression and the activity of various molecules involved in the cell cycle progression. In Rat L6E9 myoblasts, IGF-1 increases cyclin D1, and CDK4 gene expression and Rb phosphorylation (28). In MCF-7 cells, IGF-1 induces cyclin D1 expression and hyperphosphorylation of Rb through the PI3K and not the MAPK pathway (1). IGF-1 has also been shown to regulate the CDKIs. In melanoma cells, IGF-1 is involved in the redistribution of p27, which is a mechanism for growth arrest (29). Surprisingly, in a previous study from our laboratory and in a good agreement with another study, we showed that p21 CIP/WAF expression is increased by IGF-1 in MCF-7 cells (1,2). Here, we investigated the regulation of p21 CIP1/WAF by IGF-1 and the role of p21 CIP/WAF in the mitogenic effects of IGF-1 in MCF-7 cells. Based on the signal transduction pathways activated in MCF-7 cells and the effects of inhibitory agents as well as those of the expression of dominant negative constructs, we showed that the ERK1/2 MAP kinase is important in the up-regulation of both mRNA and protein of p21 CIP/WAF after IGF-1 stimulation. Using an antisense of p21 CIP/WAF to inhibit p21 CIP/WAF expression, we show that cell proliferation induced by IGF-1 is reduced by at least 2-fold following p21 CIP/WAF down-regulation.
Cell Culture and Stable Transfection-MCF-7 and T47D cells from ATCC were cultured in IMEM (MCF-7) or DMEM (T47D) supplemented with 10% fetal bovine serum, glutamine (2 mM), penicillin (100 IU/ml), and streptomycin (100 g/ml). MCF-7 cells stably transfected with an Serum-starved and G 0 synchronized MCF-7 cells were incubated with IGF-1 (10 Ϫ8 M) for the indicated times. A, total RNA was extracted from the cells, resolved by agarose gel electrophoresis, and transferred to a nylon membrane. Membranes were hybridized to a cDNA probe for p21 CIP/WAF , and results were quantified by densitometry using a PhosphoImager apparatus (Fujifilm). 18 S ribosomal RNA was used as a loading control. The corrected densities are expressed as a percent of the value at time 0 (starved cells). The error bars represent the mean Ϯ S.D. from three separate experiments. B, cells were lysed and proteins were separated on SDS-PAGE and immunoblotted with an antibody to p21 CIP1/WAF . Samples contained equal levels of protein, as confirmed by reprobing the membrane with an anti-actin antibody. Immunoreactivity was quantified by scanning densitometry and expressed as percentage of that for unstimulated cells. The results are representative of three independent experiments.
antisense IGF-1R cDNA (SX13) (4) and corresponding control cell lines transfected with the empty vector (NEO) were maintained in the same medium supplemented with 800 g/ml G418 (Geneticin, Invitrogen). The full-length p21 cDNA was synthesized by RT-PCR using total RNA from MCF-7 cells. The forward and reverse primers were 5Ј-GGAAT-TCCATGTCAGAACCGGCTG-3Ј and 5Ј-GGAATTCCTGTGGCGGAT-TAG-3Ј, respectively (30). The mammalian expression vector pBK-CMV, driven by the cytomegalovirus immediate-early promoter and containing the neomycin resistance gene, was purchased from Stratagene Cloning Systems (La Jolla, CA.). The full-length p21 CIP/WAF cDNA fragment (521 bp) was subcloned in an inverse orientation into the EcoRI and Xba sites of pBK-CMV. This antisense construct was designated pBK-CMV-AS-p21. The antisense orientation of pBK-CMV-AS-p21 was verified by PCR and DNA sequencing. The p21 CIP/WAF cDNA was also subcloned in a sense orientation into pBK-CMV, and the construct was designated pBK-CMV-S-p21. For stable transfection experiments, MCF-7 cells were cultured in phenol-red free IMEM medium supplemented with 10% fetal bovine serum. The transfection method using Effectene (Qiagen, Valencia, CA), was carried out according to the manufacturer instructions. Subconfluent MCF-7 cells were transfected with the empty vector pBK-CMV (as a negative control) or pBK-CMV-AS-p21. After 48 h, the cells were switched into the same medium supplemented with 500 g of Geneticin per ml. After 2 weeks, the surviving colonies were counted and cloned using limiting dilution of cells and several replications.
Cell Proliferation Assays-For growth studies, cells were seeded in 96-well plates (3-5,000 cells per well) in IMEM phenol red-free medium containing 5% charcoal-stripped fetal bovine serum. One day later, the medium was changed to that containing the serum-free medium plus the anti-estrogen ICI 182, 780, (10 nM) for 48 h to arrest cells in the G 0 /G 1 phase. The medium was then changed to that containing phenol red-free medium without serum and the various stimuli as indicated in the figure legends. Growth was analyzed using the 3-[4,5-dimethyltiazol 2-yl]2,5-diphenyltetraolium bromide (MTT) assay as previously described (31).
For [ 3 H]thymidine uptake, MCF-7 and T47D cells transiently transfected or not with p21 were treated for 24 h with IGF-1 (10 nM). DNA synthesis was assessed following the addition of 1 Ci/ml of [ 3 H]thymidine for a period of 4 -6 h prior to the end of the treatment protocol. Cells were then washed twice with cold PBS, and cold 5% trichloroacetic acid was added for 30 min to precipitate DNA. The precipitates were washed twice with cold PBS and resuspended in 0.5 M NaOH. Aliquots were counted in a scintillation counter.
Flow Cytometry and Cell Cycle Analysis-After stimulation with IGF-1 (10 nM), cells were trypsinized, pelleted and washed twice in PBS. Cell pellets were resuspended in the citrate buffer (250 mM sucrose, 40 mM trisodium citrate, pH 7.6, 5% Me 2 SO) and stored at Ϫ70°C. Nuclei were obtained by incubation of the cells in 300 l of solution A (3.4 mM trisodium citrate, pH 7.6, 1 mM Tris, 3 mM spermine tetrahydrochloride, 0.2% Nonidet P40, 100 g/ml trypsin) for 5 min at room temperature and in 300 l of solution B (3.4 mM trisodium citrate, pH 7.6, 1 mM Tris, 3 mM spermine tetrahydrochloride, 0.2% Nonidet P40, 100 g/ml ribonuclease A, and 500 g/ml trypsin inhibitor) for 5 min at room temperature. Finally, the nuclei were stained with 30 l of propidium iodide solution (1 mg/ml), and DNA staining was analyzed by flow cytometry on a FACSCalibur using CellQuest software. Cell cycle analysis was performed using ModfitLT TM software Version 2 (Verify Software House, Inc). Immunoprecipitation and Immunoblotting-Cells were prepared in lysis buffer A (10 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.5% Nonidet P40) containing various protease inhibitors (2 mM phenylmethylsulfonyl fluoride, 10 mg/ml leupeptin, 10 mg/ml aprotinin) and phosphatase inhibitors (100 mM sodium fluoride, 10 mM sodium pyrophosphate, 2 mM sodium orthovanadate). Lysates were centrifuged at 12,000 ϫ g for 20 min at 4°C and the supernatants were aliquoted and protein concentrations determined using the BCA protein assay. Cell lysates (250 g) were immunoprecipitated with 5 g of appropriated antibodies at 4°C overnight. The immunocomplexes were precipitated with 40 l of protein A-agarose for 1 h at 4°C. After two sequential washes using buffer A at a 1:2 dilution, the resulting pellets were boiled for 4 min in reducing Laemmli buffer containing 80 mM dithiothreitol. Proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes. Blots were blocked and probed with the different antibodies as indicated in the figure legends. After extensive washings, immune complexes were detected with the horseradish peroxidase conjugated with specific secondary antibodies followed by enhanced chemiluminescence reaction. Densitometry was performed by scanning the radiographs and then analyzing the bands with the software Mac Bas V2.52 (Fuji Photo Film). In some experiments, Western blotting was performed on whole cell lysates, using 50 g of protein.
RNA Isolation and Northern Blot-Total RNA from MCF-7 and MCF-7-derived cells was extracted at the indicated times using Trizol reagent according to the manufacturer's instructions. Twenty micrograms of total RNA was separated by denaturing formaldehyde electrophoresis, then transferred by capillary blot to positively charged nylon membrane overnight, and immobilized by exposure to UV light. Blots were prehybridized for 2 h at 42°C in a buffer containing formamide 50%, Denhart 5ϫ, SDS 1%, SSC 5ϫ, and salmon sperm 100 g/ml and hybridized overnight at the same temperature with 2 MCPM/ml of [ 32 P]dCTP-labeled DNA probe in a buffer containing formamide 50%, Denhart 2.5ϫ, SDS 1%, SSC 5ϫ, dextran sulfate 10ϫ, and salmon sperm 100 g/ml. The full-length p21 cDNA probe was labeled using the rediprime labeling kit (Amersham Biosciences). Finally, blots were washed to high stringency and hybridized radioactivity was measured using a PhosphoImager apparatus (FujiFilm, Stamfordn CT). Autoradiography was also carried out at Ϫ70°C. The integrity and the quantification of p21 CIP/WAF transcript were assessed using the human RNA 18 S probe from Ambion, Inc (Austin, TX).
Transient Transfection in MCF-7 and T47D Cells-MCF-7 or T47D cells were seeded the day before the transfection. Cells were then transiently transfected using the Effectene transfection reagent (Qiagen, CA) according to the manufacturer's instruction with 1 g of a MEK-1 (MEK-1 Ϫ/Ϫ ) dominant negative construct, the pBK-CMV-S-p21 construct or an empty vector (empty vector). 24 h after transfection with the MEK-1 construct, MCF-7 cells were switched to serum-free medium for 16 h, followed by the addition of IGF-1 (10 nM, 10 min) at 37°C. Protein extracts were prepared as described previously, and Western blot analyses were performed with antiphospho-ERK1/2 antibodies to detect ERK1/2 phosphorylation induced by IGF-1 stimulation as well as with the ERK-1 antibodies to detect ERK1/2 protein. 24 h after transfection with the pBK-CMV-S-p21 construct, MCF-7 and T47D cells were switched to serum-free medium for 16 h, followed by the addition of IGF-1 (10 nM, 12 h) at 37°C, and Western blot analyses were performed with p21 antibodies and actin as a loading control.
Statistical Analysis-All reported values are the means Ϯ S.D. Statistical comparisons were made by two-sided Students's t test. Statistical significance was assumed if a null hypothesis could be rejected at the p Ͻ 0.05.

IGF-1 Induced p21 CIP/WAF Expression in MCF-7 Cells-
MCF-7 cells were partially synchronized in the G 0 phase of the cell cycle by serum deprivation and using the anti-estrogen ICI 182, 780 (10 nM) for 48 h. Under these conditions, more than 75% of cells were synchronized in G 0 (data not shown). Next, cells were treated with IGF-1 for various times. IGF-1 (10 nM) increased p21 CIP/WAF mRNA level maximally after 4 h of stimulation. This effect was maintained for 12 h and abolished after 24 h of stimulation (Fig. 1A). p21 CIP/WAF protein increased 7-fold after 12 h of IGF-1 stimulation (Fig. 1B). This increase was observed for up to 24 h of IGF-1 treatment. However, after 48 h of stimulation, the effect of IGF-1 was abolished. Thus, the time course for p21 CIP/WAF protein expression correlates well with the expression of p21 CIP/WAF mRNA induced by IGF-1. In order to confirm that IGF-1 increases p21 CIP/WAF expression through the IGF-1R, MCF-7-derived cells (SX13), which exhibit a 50% reduction in IGF-1R expression (2,4) were stimulated with IGF-1 (10 nM) for 4 h. As shown in Fig. 2, the increase in p21 CIP/WAF mRNA level in response to IGF-1 was significantly reduced in SX13 cells as compared with the MCF-7/NEO cells. Taken together, these data demonstrate that IGF-1R activation is necessary to increase p21 CIP/WAF expression in MCF-7 cells.
MAP Kinase ERK1/2 Is Involved in p21 CIP/WAF Expression Induced by IGF-1-In order to study the signaling pathways involved in the increase of p21 CIP1/WAF protein level by IGF-1 in MCF-7 cells, inhibitors specific for PI3K, p38, JNK1/2, and ERK1/2 were used. First, we determined the active concentration of each inhibitor. MCF-7 cells were treated for 1 h with the inhibitors prior to the IGF-1 stimulation (10 nM, 10 min). Western-blot analysis revealed the dose-dependent inhibition by LY294002 (PI3K pathway), SB202190 (p38 pathway), SP600125 (JNK1/2 pathway), and U0126 (ERK1/2 pathway), respectively. 2 M U0126 was capable of inhibiting ERK1/2 phosphorylation induced by IGF-1 (data not shown). In the same manner, 50 M SB202190, 50 M SP600125, and 50 M LY294002 inhibited p38, JNK1/2, and Akt phosphorylation (data not shown). Western blot analyses, performed to detect all forms of ERK1/2, p38, JNK1, and Akt kinases, revealed that protein levels were similar among the samples. Fig. 3 indicates that the JNK1/2 pathway inhibitor SP600125, the p38 pathway inhibitor SB202190 and the PI3K pathway inhibitor LY294002 did not affect the increase in the p21 CIP1/WAF protein level induced by IGF-1. However, the addition of U0126 reduced by 80% the increase in the p21 CIP1/WAF expression induced by IGF-1 stimulation. Thus, in MCF-7 cells, the increase in the p21 CIP1/WAF protein level by IGF-1R activation is apparently mediated through the MAPK ERK1/2 pathway.
In order to confirm these data, MCF-7 cells were transiently transfected with a dominant negative MEK-1 construct (MEK-1 Ϫ/Ϫ ). Upon IGF-1 stimulation, there was a 70% reduction in ERK1/2 phosphorylation after transfection of the MEK-1 dominant negative construct (Fig. 4A). Western-blot analysis was performed on MCF-7 transfected with MEK Ϫ/Ϫ and with the empty vector as a control (Fig. 4B). Upon IGF-1 stimulation, in cells transfected with the empty vector, p21 CIP1/WAF protein level was increased by 7-fold as compared with unstimulated cells. However, when MCF-7 cells were transfected with a dominant negative MEK-1 the increase in expression of p21 CIP1/WAF by IGF-1 was reduced by about 80%. Thus, the inhibition of the ERK1/2 pathway, either by specific inhibitors or by the expression of a dominant negative MEK-1 in MCF-7 results in inhibition of the increase of p21 CIP/WAF protein expression by IGF-1.
Growth Inhibition of MCF-7 Cells in Response to IGF-1 by Stable Expression of p21 CIP/WAF Antisense Ribonucleic Acid-In order to understand the role of p21 CIP/WAF in the mitogenic effect of IGF-1, we generated MCF-7/p21 CIP/WAF AS (antisense) clones. The MCF-7/p21 CIP/WAF clones were developed by stable transfection of MCF-7 cells with the pBK-CMV-AS-p21 CIP/WAF expression vector as described under "Experimental Procedures." Thirty G418 resistant clones were analyzed by immunoblotting, using an anti-p21 CIP/WAF antibody. Five clones were selected, designated as C1, C2, C3, C4, and C5, and analyzed. In parallel, MCF-7 cells were stably transfected with the pBK-CMV vector, and two clones of control cells (Neo 1 and Neo 2) were generated. p21 CIP/WAF (21 kDa) immunoreactivity was detected in clone Neo 1 and Neo 2, but this was strongly reduced in the C4 and C5 clones, and totally absent in C1, C2, and C3 clones (Fig. 5A). The cellular proliferation rates of cells expressing p21 CIP/WAF antisense RNA were then compared with those in MCF-7 cells transfected with vector alone (Neo controls) in response to IGF-1 (10 nM) or estradiol (1 nM). Neo and p21 CIP/WAF -AS cell lines were plated at the same density (3,000 -5,000 cells/well) and incubated in the presence or absence of IGF-1 (10 nM) and estradiol (1 nM) separately or combined for 24 or 48 h. Results from such experiments, using Neo 1 and C1, are shown in Fig. 5B. Similar results were obtained with the Neo 2, C3, and C4 clones. Cell growth in response to IGF-1 was significantly reduced (p Ͻ 0.05) by at least 2-fold in the C1 clone, which overexpressed p21 CIP/WAF -AS, as compared with Neo 1 cells. However, cell growth in response to estradiol was similar in C1 and Neo 1 cells. Thus, a reduction in the p21 CIP/WAF expression leads to a decrease of IGF-1-induced cell proliferation in MCF-7 cells. In order to show whether p21 affects IGF-1-induced cell proliferation, we transiently transfected the p21 sense construct in MCF-7 cells to levels observed after treatment of MCF-7 with IGF-1. As shown in the Fig. 6A, in the basal state (no IGF-1 treatment), MCF-7 cells transiently transfected with the pBK-CMV-S-p21 construct (p21 ϩ/ϩ cells) incorporated [ 3 H]thymidine to a greater degree than MCF-7 cells (p Ͻ 0.05) transfected with the empty vector (control cells). Moreover, IGF-1 treatment further improved thymidine incorporation (p Ͻ 0.05). We have also transiently transfected the T47D cells, another estrogen receptor positive breast cancer cell line with p21 CIP/WAF . As shown in Fig. 6B similar results are observed in T47D as with MCF-7 cells. Indeed, thymidine incorporation is higher in T47D p21 ϩ/ϩ cells than T47D control cells in the basal state (p Ͻ 0.05) and once again IGF-1 increased cell proliferation (p Ͻ 0.05). G 1 Growth Arrest in Response to IGF-1 in MCF-7 p21 CIP/WAF Antisense Cells-FACS scan analysis was used to determine the effects of IGF-1 on the distribution of cells through the phases of the cell cycle in the MCF-7/p21 CIP/WAF antisense clones, C1, and C2 and their controls, Neo 1 or Neo2. Neo and p21 CIP/WAF antisense cells were synchronized for 48 h using the anti-estrogen ICI 182,780 10 nM (time 0) and stimulated with IGF-1 10 nM for the indicated times. Representative results for C1 and Neo 1 are indicated in Table I. After 15 h of IGF-1 stimulation, 3% of C1 cells and 22% of Neo 1 cells progressed through S phase. Similar results were obtained with Neo 2 and C2 cells. Thus, the decrease in proliferation rate in response to IGF-1 in MCF-7/p21 CIP/WAF antisense cells correlates with an increase in the proportion of cells in G 1 suggesting a delay in G 1 to S progression.
Expression of Cyclin D1 and Its Association with CDK4 and CDK2 and the Level of Rb Phosphorylation in Response to IGF-1 Is Reduced in MCF-7 p21 CIP/WAF Antisense Cells-In order to explain further the G 1 growth arrest in MCF-7 p21 CIP/WAF antisense cells in response to IGF-1, we studied some cell cycle components. The elements of the cell cycle we examined included cyclin D1 expression and its association with CDK4 or CDK6 (Fig. 7, A and B) and the level of Rb phosphorylation and cyclin E expression (Fig. 8). Neo and p21 CIP/WAF antisense cells were synchronized for 48 h by serum deprivation and using the anti-estrogen ICI 182,780 10 nM and then stimulated with IGF-1 10 nM for the indicated times. As shown in the time course experiment of Fig. 7A, cyclin D1 immunoreactivity in response to IGF-1 is reduced by about 2-fold in MCF-7 p21 CIP/WAF antisense as compared the Neo cells. Consequently, we showed a marked reduction in the association of cyclin D1 with CDK4 or CDK6 in MCF-7 p21 CIP/WAF antisense as compared with the Neo cells (Fig. 7B). These results were observed with a decrease in cyclin D1 protein levels but without any changes in CDK4/CDK6 protein levels (data not shown) in p21 CIP/WAF antisense compared with the Neo cells. Finally, we determined the phosphorylation state of Rb and the expression level of cyclin E. As shown in the time course experiment of Fig.  8, similar to the cyclin D1 expression, phosphorylation of Rb (Fig. 8A) and protein level of cyclin E (Fig. 8B) were significantly reduced in MCF7/p21 CIP/WAF antisense cells as compared with Neo cells. Thus, the G 1 growth arrest in response to IGF-1 in MCF-7/p21 CIP/WAF antisense cells is due at least to some extent to the reduction of the activation of several components of the cell cycle including the cyclin D1 expression.

DISCUSSION
The cyclin-dependent kinase inhibitor, p21 CIP1/WAF is induced in various biological situations including cell cycle arrest (32), differentiation (33,34), and apoptosis (35,36). Several studies described p21 CIP/WAF as a potent inhibitor of cell proliferation in various cell culture models (37). For instance, in A431 cells, EGF inhibited cell growth by increasing p21 CIP/WAF protein levels and, more specifically, the half-life of p21 CIP/WAF mRNA (38). Similarly, treatment with fibroblast growth factor (FGF) induced an increase in p21 CIP/WAF mRNA and protein levels and resulted in inhibition of cellular proliferation in  Neo1  74  16  10  P21AS-C1  79  13  8  15  Neo1  55  38  7  P21AS-C1  75  16  9 FIG. 6. Overexpression of p21 increases cell proliferation in MCF-7 and T47D cells. Upper panel, MCF-7 (A) or T47D (B) cells were transiently transfected with a p21 CIP/WAF sense construct (p21 ϩ/ϩ ) or the pBK-CMV empty vector (control cells) as described under "Experimental Procedures." 50 g of protein from whole cell lysates was subjected to SDS-PAGE followed by immunoblotting with a monoclonal p21 CIP/WAF antibody or actin as loading control. Lower panel, p21 ϩ/ϩ or control MCF-7 (A) and T47D (B) cells were treated for 24 h with IGF-1. Cell proliferation was determined after thymidine incorporation as described under "Experimental Procedures." The error bars represent the mean Ϯ S.D. from three separate experiments. . p21 inhibits CDKs that function in the G1 and S phases. A recent study shows that p21 CIP/WAF is an important physiological target of antiestrogen action that inhibits cdk4 in addition to cdk2 in human breast cancer cells (40). Interestingly, in a previous study, we have observed that IGF-1 increases the protein levels of p21 CIP/WAF (2). The protein levels of p21 CIP/WAF are known to be regulated by post-translational degradation and also by transcriptional activation, although the exact mechanism of protein degradation is unclear (41). Here, we studied the regulation of p21 CIP/WAF mRNA and protein levels by IGF-1 and found that the MAPK ERK1/2 signaling pathway is involved in the activation of p21 CIP/WAF gene expression by IGF-1 stimulation. In PC12 and Hela cells, some studies have suggested that the p21 CIP/WAF gene expression is regulated at the level of transcription, primarily through the MAP kinase pathway in response to growth factors (42). In the differentiating chondrogenic cell line MCT as well as in primary mouse chondrocytes, the Raf-1/MEK/ERK pathway is the major regulator of p21 CIP/WAF gene transcription (43). The MAPK pathway is also involved in p21 CIP/WAF induction by TGF␤ in the HaCaT cells (44). However, p21 CIP/WAF was also reported to be regulated by a MAP kinase independent pathway (45). Two studies demonstrated that the PI3 kinase pathway, which is regulated by growth factors and which up-regulates AKT kinase activity, was involved with the modulation of p21 CIP/WAF levels (46,47). Moreover, a recent study showed that the stability of p21 CIP/WAF is regulated by AKT -mediated phosphorylation at residue Ser-146 (48). In MCF-7 cells, Dufourny et al. (1) showed that some stimulatory effects of IGF-1 on cell cycle progression such as cyclin D1 synthesis and pRb hyperphosphorylation were blocked by the specific PI3K inhibitory LY294002. From these results, they suggested that the PI3K activity but not MAPK activity was required for transduction of the mitogenic IGF-1 signal in these cells (1). In the present study, we showed that the IGF-1 induced p21CIP/WAF expression through the MAPK ERK1/2 is involved in the mitogenic IGF-1 effect. Thus, both PI3K and MAPK are important signaling pathways involved in the IGF-1 effect on the cell cycle progression in MCF-7 cells.

Time of stimulation Cells in cell cycle
The induction of p21 CIP/WAF gene expression by IGF-1 appears to be contradictory to the previously described role of p21 CIP/WAF as a negative regulator of cell growth via its ability FIG. 7. p21 CIP/WAF antisense overexpression reduces cyclin D1 expression and its association with CDK4 and CDK6. Neo1 and p21AS C1 were synchronized in G 0 phase by starvation and using the anti-estrogen ICI 182, 780 (10 nM) for 48 h. A, cells were stimulated with IGF-1 (10 nM) for various times (1-6 h), and then the protein expression of cyclin D1 was determined by Western blotting. The results obtained for Neo1 and p21AS C1 are represented in odd numbered lanes (1,3,5,7,9,11,13) and even numbered lanes (2,4,6,8,10,12,14), respectively. Samples contained equal levels of protein, as confirmed by reprobing each membrane with an anti-actin antibody. These results are representative of two independent experiments. B, cells were stimulated with IGF-1 (10 nM) for the indicated times prior to be lysated. Cyclin D1 association was determined with CDK4 (panel 1) and CDK6 (panel 2) in both cell lines. Cyclin D1 was immunoprecipitated from whole cell lysates. Samples were then subjected to Western blotting with antibodies against either CDK4 or CDK6. Membranes were stripped and reprobed to determine the cyclin D1 total protein in each lane. Results shown (mean Ϯ S.D.) are from three experiments.
to act as a CDK inhibitor. However, a number of studies described a positive function of p21 CIP/WAF in the cell cycle progression. In addition, several studies show a marked increase in the levels of p21 CIP/WAF in the rapidly proliferating tumors, a result that is counterintuitive in the light of negative role of p21 CIP/WAF in the cell cycle progression (49,50). Some studies demonstrated that although p21 inhibits the cyclin E-CDK2 complexes it also promotes the assembly and activation of cyclin D1-CDK4/6 complexes, which function as sensors of growth factors at the G 1 /S phase (51)(52)(53)(54). Another study showed that p21 is active as a CDK inhibitor in MCF-7 but the elevated level of cyclin D3 titrates p21 away from cyclin D1-CDK4/6 complexes and CDK2 complexes resulting in increased CDK kinase activities and cell cycle progression (55). Here, we demonstrate for the first time that an abrogation of p21 CIP/WAF expression by overexpression of p21 CIP/WAF cDNA antisense in MCF-7 results in a 50% decrease in the mitogenic effects of IGF-1. Moreover, we show that this reduction in mitogenic activity is positively correlated with a delay in G 1 /S progression. This indicates that, at least in MCF-7 cells, p21 CIP/WAF has a positive role in growth factor-induced mitogenesis. Mantel et al. (56) using p21 CIP/WAF -deficient mice, suggested also that p21 CIP/WAF plays a positive role in the mitogenic response of myeloid cells to the steel factor or the granulocyte-macrophage colony-stimulating factor especially in the synergistic proliferative response to stimulation with the combination of these cytokines. To confirm that p21 is involved in cell proliferation, we have transiently transfected p21 and measured thymidine incorporation in response to IGF-1 in both MCF-7 and T47D cells. In both cell lines, in basal state (no IGF-1 treatment) p21 overexpression leads to an increase in the cell proliferation. This indicates that in MCF-7 and T47D cells, p21 has a positive role in the cell proliferation. Moreover, in both cell lines tranfected with p21 CIP/WAF , IGF-1 increased cell proliferation. Overexpression of the human p21 CIP/WAF gene promoted further proliferation in mouse myeloid progenitor cells (57). Another study indicate that tumors arising in MMTVras/p21 CIP/WAFϪ/Ϫ mice displayed higher S-phase fractions and correspondingly increased cyclin D1 and E/CDK activity than MMTV-ras tumors. In contrast, MMTV-myc/p21 CIP/WAFϪ/Ϫ tumors had lower S-phase fractions and levels of cyclin D1 and E/CDK activity than MMTV-myc tumors (58). Thus, p21 CIP/WAF might play a role in promoting either growth arrest or proliferation, depending on the specific cellular context.
In our study, we showed for the first time that a reduction of p21 CIP/WAF expression in MCF-7 cells is associated with a reduction in the IGF-1-induced cyclin D1 expression. Consequently, the levels of CDK4/cyclin D1 and CDK6/cyclin D1 complexes as well as the Rb phosphorylation and the level of cyclin E are reduced in response to IGF-1. These data probably explain the delay in cell cycle progression and the 2-fold reduction in the IGF-1 induced cell proliferation. Cyclin D1 plays an important role in the cell cycle control of the mammary gland and clinical studies of human breast cancer confirm its importance. For example, cyclin D1 is overexpressed in the majority of human breast cancers (59). Moreover, p21 CIP/WAF does correlate with cyclin D1 overexpression; 93% of breast tumors that overexpressed D-type cyclins also overexpressed p21 CIP/WAF, suggesting that a relationship exists between p21 CIP/WAF and cyclin D1 levels in some human malignancies (60). Our results suggest that in addition to regulating the activity of cyclin/CDK complexes through direct interaction, p21 CIP/WAF can also either directly or indirectly lead to changes in cyclin D 1 protein levels. A study analyzing the effects of loss of both p21 CIP/WAF and p27 in mouse embryo fibroblasts demonstrated not only a loss of cyclin D/CDK activity but also markedly reduced cyclin D protein levels (61). Cyclin D1 has been also reported to be less stable in cells lacking p21 CIP/WAF and p27 kip1 , suggesting that the CIP/KIP protein levels were required to maintain cyclin D1 protein levels (53). Although it is unclear whether the effect of p21 CIP/WAF on cyclin D1 is direct or indirect, there is some evidence for p21 CIP/WAF -induced transcription of cyclin D1 (59) possibly through pRb activation. A recent study has also demonstrated that in Ras-transformed cells, cyclin D1 is required to protect p21 CIP/WAF from proteasomal degradation (26).
In summary, we have elucidated the molecular mechanisms involved in the increase of p21 CIP/WAF expression in response to IGF-1 in MCF-7 cells. Using antisense strategy, we also demonstrated that p21 CIP/WAF is a positive key regulator in the IGF-1 induced cell proliferation. Moreover, we showed for the first time that a reduction of p21 CIP/WAF protein is associated to a decrease in the cyclin D1 expression in MCF-7 cells. Further studies will enable us to understand the interaction between IGF-1R, p21 CIP/WAF , and cyclin D1.
FIG. 8. p21 CIP/WAF antisense overexpression reduces Rb phosphorylation and cyclin E expression. Neo1 and p21AS C1 cells were arrested by starvation and using the anti-estrogen ICI 182, 780, (10 nM) for 48 h. Cells were stimulated with IGF-1 (10 nM) for various times (0 -28 h) and then the phosphorylation of Rb and cyclin E expression was determined by Western blotting. The results obtained for Neo1 and p21AS C1 are represented in odd numbered lanes and even numbered lanes, respectively. Samples contained equal levels of protein, as confirmed by reprobing each membrane with either an Rb antibody or an anti-actin antibody.