Vitamin D Inhibits G 1 to S Progression in LNCaP Prostate Cancer Cells through p27 Kip1 Stabilization and Cdk2 Mislocalization to the Cytoplasm*

1,25-(OH) 2 vitamin D 3 (1,25-(OH) 2 D 3 ) exerts antiprolif- erative effects via cell cycle regulation in a variety of tumor cells, including prostate. We have previously shown that in the human prostate cancer cell line LNCaP, 1,25-(OH) 2 D 3 mediates an increase in cyclin- dependent kinase inhibitor p27 Kip1 levels, inhibition of cyclin-dependent kinase 2 (Cdk2) activity, hypophosphorylation of retinoblastoma protein, and accumulation of cells in G 1 . In this study, we investigated the mechanism whereby 1,25-(OH) 2 D 3 increases p27 levels. 1,25-(OH) 2 D 3 had no effect on p27 mRNA levels or on the regulation of a 3.5-kb fragment of the p27 promoter. The rate of p27 protein synthesis was not affected by 1,25-(OH) 2 D 3 as measured by luciferase activity driven by the 5 (cid:1) - and 3 (cid:1) -untranslated regions of p27 that regulate p27 protein synthesis. Pulse-chase analysis of 35 S-labeled p27 revealed an increased p27 protein half-life with 1,25-(OH) 2 D 3 treatment. Because Cdk2-mediated phosphoryl- ation of p27 at Thr 187 targets p27 for Skp2-mediated degradation, we examined the phosphorylation status of p27 in 1,25-(OH) 2 D 3 -treated cells. 1,25-(OH) 2 D 3 de- creased levels of Thr 187

Our previous studies established that the initial growth inhibition of LNCaP and its androgen-independent derivative LNCaP-104R1 by 1,25-(OH) 2 D 3 correlates with an increase in the levels of the cyclin-dependent kinase inhibitors (CKIs) p21 WAF1,CIP1 and p27 Kip1 , a profound decrease in cyclin-dependent kinase 2 (Cdk2) activity, hypophosphorylation of pRb, and accumulation of cells in the G 1 phase of the cell cycle (12,13). 1,25-(OH) 2 D 3 also decreased transcriptional activity of the E2F transcription factor family, which regulates the expression of genes necessary for S phase entry (12).
Loss of p27 expression correlates with prostate cancer recurrence, a more aggressive phenotype, and decreased patient prognosis and survival (15)(16)(17)(18)(19)(20)(21)(22). Reduced p27 protein, however, is not the result of p27 gene mutations, which are quite rare in cancers (23)(24)(25). Activation of oncogenic signaling pathways ultimately results in accelerated p27 proteolysis and decreased p27 levels (26 -29). Understanding the mechanisms whereby p27 levels are regulated may yield new targets for potential anticancer agents. Because of the emerging role of p27 in controlling prostate cancer growth, we further investigated the mechanism of 1,25-(OH) 2 D 3 regulation of this CKI in LNCaP cells.
p27 is an important regulator of the G 1 to S phase transition. p27 binds and inhibits cyclin E/Cdk2 and thereby negatively regulates S phase entry. To traverse G 1 , cellular p27 levels must decrease. A major mechanism to achieve this regulation involves ubiquitin-dependent p27 proteolysis (30,31). Two rate-limiting steps for this process include phosphorylation at Thr 187 by Cdk2 and recognition of Thr 187 -phospho-p27 by the SCF skp2 ubiquitination system (17,(32)(33)(34)(35)(36). Recent studies have revealed a novel pathway for p27 degradation at the G 0 /early G 1 phase of the cell cycle that is independent of Thr 187 phos-phorylation (37)(38)(39). Although not well understood, this pathway is thought to be activated by mitogenic signaling during early G 1 prior to Cdk2 activation. This initial phase of p27 degradation then facilitates Cdk2 activation, which results in further p27 degradation in late G 1 and ultimately entry into S phase. Although this pathway still involves the ubiquitin-proteasome system, it is independent of Thr 187 phosphorylation and may be Skp2-independent (37)(38)(39). The exact mechanisms of this novel pathway for p27 degradation remain to be resolved.
The 5Ј-untranslated region (UTR) of the p27 mRNA regulates the translation of p27. In quiescent cells, p27 translation is enhanced despite an overall decrease in the synthesis of most other proteins (31,(42)(43)(44)(45). The translation of most proteins involves recognition of the mRNA 5Ј 7-methylguanosine cap by the eukaryotic initiation factor 4E. Eukaryotic initiation factor 4E activity is regulated by mitogenic stimulation (reviewed in Ref. 44). In the absence of mitogenic signaling, this cap-dependent translation is decreased through down-regulation of eukaryotic initiation factor 4E activity. The 5Ј-UTR of p27, however, contains an internal ribosomal entry site, which allows cap-independent translation of p27. This is enhanced by binding of HuR, a member of the embryonic lethal, abnormal vision (ELAV) family of RNA-binding proteins and binding of heterogeneous nuclear ribonucleoproteins C1 and C2, factors implicated in mRNA stability and processing, to the 5Ј-UTR of p27 (42)(43)(44)(45)48).
Cell Culture-LNCaP-FGC cells (ATCC) were passaged and maintained in RPMI medium supplemented with 10% fetal bovine serum, 100 IU/ml penicillin, 100 g/ml streptomycin, and 100 g/ml L-glutamine. All of the cultures were maintained at 37°C in a humidified atmosphere of 5% CO 2 .
[ 3 H]Thymidine Uptake Assays-LNCaP cells were treated with either ethanol vehicle or 10 nM 1,25(OH) 2 D 3 . Following the 48-h treatment period, the media were replaced with [ 3 H]thymidine-containing medium, and the cells were incubated for 18 h. Acid soluble tritium was removed by a trichloroacetic acid wash; the cells were then lysed, and [ 3 H]thymidine uptake was determined by scintillation counting.
Reporter Plasmids and Luciferase Assay-The reporter plasmids p27PFLuc and control (41) were the generous gifts of Dr. Toshiyuki Sakai (Kyoto Prefectural University of Medicine, Kawaramachi-Hirokoji). The reporter plasmids 5Ј-SvL-3Ј and controls (43) were provided by Dr. Andrew Koff (Memorial Sloan-Kettering Cancer Center, New York). For luciferase assays, the cells were transfected with 5.0 g of reporter plasmid and 1.0 g of cytomegalovirus-␤-galactosidase vectors and treated with either ethanol vehicle or 10 nM 1,25-(OH) 2 D 3 or serum-starved for the appropriate time periods. Following the treatment period, the cell lysates were collected, and luciferase activity was observed. ␤-Galactosidase activities were measured to normalize for transfection efficiency.
Pulse-Chase Analysis-One day after plating, the cells were treated with either ethanol vehicle or 10 nM 1,25-(OH) 2 D 3 for 24 h. Following the treatment period, the medium was replaced with Dulbecco's modified Eagle's medium (Ϫmet) and 10% dialyzed fetal bovine serum for 1 h. Next, the cells were pulsed for 1 h with 500 Ci of [ 35 S]Met (PerkinElmer Life Sciences). Following the pulse, the cells were then chased for 0, 1, 3, and 6 h with Dulbecco's modified Eagle's medium (Ϫmet) medium containing 40 mM cold methionine and 10% dialyzed fetal bovine serum. After the chase times, the cells were lysed in sample buffer containing 50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 0.5% Nonidet P-40, 10 g/ml aprotonin, 10 g/ml leupeptin, 50 mM NaF, 0.1 mM sodium orthovanadate. Extracts (500 g protein) were precleared with 1 g of normal rabbit IgG preadsorbed with protein A-agarose beads. Precleared extracts were then incubated with 1 g of polyclonal p27 antibody and agitated overnight at 4°C. The samples were then equilibrated with lysis buffer and 25 l of packed volume of protein Aagarose beads for 4 h at 4°C with agitation. Following three washes with lysis buffer, immune complexes were eluted by boiling in 20 l of Laemmli gel loading buffer. The samples were then subjected to SDS-PAGE, transferred to nitrocellulose membrane filters, and exposed to autoradiography.
Western Blot Analysis-The cells were treated with ethanol vehicle, 10 nM 1,25-(OH) 2 D 3 , or serum-starved for appropriate times, washed, and lysed in sample buffer. The protein concentrations were determined by the Bio-Rad D c Protein Assay according to the manufacturer's instructions. 50 g of cell extract proteins were subjected to standard SDS-PAGE and transferred to nitrocellulose membrane filters. The filters were processed for Western blotting using standard procedures. Briefly, the filters were incubated overnight at 4°C in blocking solution (5% dry milk, 0.1% Tween in 1ϫ wash buffer (20 mM Tris, 50 mM NaCl, 2.5 mM EDTA)) followed by incubation with the primary antibody for 1 h. For phospho-specific antibodies, the filters were blocked overnight in 5% bovine serum albumin, 0.1% Tween in PBS. Actin antibody (Roche Applied Science) was used at 0.5 g/ml. p27 and Cdk2 antibodies were used at 1.0 g/ml. After washing, the blots were incubated with horseradish peroxidase-conjugated secondary antibody, and the proteins were visualized using the ECL system (Amersham Biosciences) following the supplier's instructions.
In Vitro Cdk2 Kinase Assay-LNCaP cells plated at 40 -50% confluency were treated with either ethanol vehicle or 10 nM 1,25-(OH) 2 D 3 . Following the treatment period, the cells were washed in PBS and solubilized in TNE buffer (50 mM Tris, pH 7.5, 140 mM NaCl, 5 mM EDTA) containing 1% Nonidet P-40, 1:100 dilution of the protease inhibitor mixture (Sigma), 50 mM NaF, and 0.1 mM sodium orthovanadate. The protein concentrations of the lysates were determined as described above. 200 g of proteins were then incubated with 3 g of rabbit anti-human Cdk2 or rabbit anti-human cyclin E antibodies for 1 h at 4°C. Following this incubation, the samples were adsorbed with 40 l of anti-rabbit IgG-agarose beads (Sigma) for 1 h at 4°C. After washing once with TNE buffer and three times with kinase buffer (50 mM Tris, pH 7.4, 10 mM MgCl 2 ), immune complexes were incubated in 30 l of kinase buffer containing 1 g of histone H1, 25 M ATP, and 10 Ci of [␥-32 P]ATP for 30 min at 30°C. The reactions were stopped by the addition of 4ϫ Laemmli buffer. After 5 min of boiling, the samples were then subjected to SDS-PAGE and transferred to a nitrocellulose membrane. Phosphorylated histone H1 was visualized by autoradiography. After the autoradiography, the membrane was subjected to Western blotting for Cdk2 and cyclin E.
Subcellular Fractionation-The cells were treated with ethanol vehicle, 10 nM 1,25-(OH) 2 D 3 or serum-starved for appropriate times. Following the treatment periods, the cells were washed in PBS and lysed in transport buffer (20 mM HEPES, pH 7.4, 110 mM potassium acetate, 2 mM MgCl 2 ) containing 15 g/ml digitonin (Calbiochem, Cambridge, MA) and 10 g/ml aprotonin, 10 g/ml leupeptin, 50 mM NaF, and 0.1 mM sodium orthovanadate. The lysates were then centrifuged for 5 min at 800 ϫ g at 4°C, and the supernatant was collected (cytosolic fraction). The nuclear fraction was obtained by sonication of the pellet in transport buffer. The protein concentrations of both fractions were determined as described above. 25 g of cytosolic and 50 g nuclear proteins were then subjected to SDS-PAGE and transferred to nitrocellulose membrane filters. The filters were processed for Western blotting of Cdk2, p27, nucleolin (for fractionation purity), and actin (loading control).
Immunofluorescence-One day after plating on coverslips, the cells were treated with ethanol vehicle, 10 nM 1,25-(OH) 2 D 3 , or serumstarved for appropriate times. Following the treatment period, the cells were fixed in 2% formaldehyde in PBS for 10 min. Next, the cells were permeablized in PBS containing 1% bovine serum albumin, 0.1% Triton X-100 for 5 min. The cells were then incubated for 1 h with rabbit anti-Cdk2 (1:50 dilution in PBS) and mouse anti-nucleoporin antibodies (mAB414), washed in PBS, and incubated for 1 h with Cy3-conjugated goat anti-rabbit (1:1000) and fluorescein isothiocyanate-conjugated donkey anti-mouse antibodies. Following three additional washes, the coverslips were incubated with 4,6-diamidino-2-phenylindole for 5 min and mounted on glass slides in anti-fade medium. The images were then collected using confocal microscopy (Zeiss).
Quantification and Statistical Analysis-Western blot data were quantified using densitometry. The data were analyzed via either a two-tailed unpaired t test or a one-way analysis of variance followed by a Bonferroni post test using GraphPad Prism version 3.00 for Windows (GraphPad Software, San Diego, CA). For confocal image analysis, fluorescence intensities were determined using ImageJ (rsb.info.nih.

FIG. 3. Vitamin D does not increase p27 protein synthesis. A,
LNCaP cells were transfected with either SvL control or 5Ј-3Ј SvL luciferase reporter constructs (luciferase activity governed by the 5Јand 3Ј-untranslated regions of p27) and treated with either vehicle or 10 nM 1,25-(OH) 2 D 3 . The cells were harvested 24 h later, and luciferase activity was assessed. As a positive control, transfected cells were also subjected to serum starvation, a known stimulator of p27 protein synthesis. B, parallel dishes of LNCaP cells were treated with either vehicle or 10 nM 1,25-(OH) 2 D 3 or subjected to serum starvation for 24 h. The cell extracts were subjected to immunoblot analysis with antibodies against p27. Actin was used as a loading control. The data in A represent two experiments performed in triplicate. Luciferase activity in serum-starved cells was significantly different from vehicle-treated cells (p Ͻ 0.05). Vit D, vitamin D.
To test whether 1,25-(OH) 2 D 3 regulates p27 at the transcriptional level, we utilized a reporter gene construct in which the luciferase reporter is driven by a 3.5-kb fragment of the p27 promoter (41). It was previously shown that 1,25-(OH) 2 D 3 activates this construct in U937 leukemic cells despite the apparent lack of vitamin D response elements in the p27 promoter (41). However, although serum starvation of LNCaP cells enhanced luciferase activity, 1,25-(OH) 2 D 3 did not increase luciferase activity driven by this fragment of the p27 promoter (Fig. 2). The time points utilized to assay p27 promoter activity following 1,25-(OH) 2 D 3 treatment were within the range in which p27 protein is up-regulated (Ref. 13 and Figs. 3B and 5). Northern blot analysis confirmed this finding because no changes in p27 mRNA levels were observed in response to 1,25-(OH) 2 D 3 (data not shown). These data show that 1,25-(OH) 2 D 3 -mediated up-regulation of p27 protein was not due to increased p27 transcription or increased p27 mRNA levels.

1,25-(OH) 2 Vitamin D 3 Stabilizes p27 Protein via Decreased Cdk2
Phosphorylation of Thr 187 -To determine whether 1,25-(OH) 2 D 3 -mediated increases in p27 protein are due to decreased p27 degradation, pulse-chase analysis was performed. As seen in Fig. 4, 1,25-(OH) 2 D 3 increased the half-life of p27 protein from ϳ4.5 h in asynchronously proliferating vehicletreated cells to greater than 6 h in 1,25-(OH) 2 D 3 -treated cells. The half-lives of p27 in both treated and untreated cells are comparable with those found in other studies for growth arrested and asynchronous cells (31,49).
For cells to enter S phase, p27 levels must decline. This process is regulated in part by phosphorylation of p27 at Thr 187 by cyclin E/Cdk2 complexes. This phosphorylation event targets p27 for degradation by allowing its recognition by the SCF skp2 ubiquitin ligase complex. To investigate further the 1,25-(OH) 2 D 3 -mediated regulation of p27 proteolysis, the levels of phospho-Thr 187 -p27 were assessed. As seen in Fig. 5, 1,25-(OH) 2 D 3 treatment at 24 and 48 h results in decreased phospho-Thr 187 -p27, whereas total p27 was increased at these two points. These results affirmed that 1,25-(OH) 2 D 3 -mediated upregulation of p27 in LNCaP cells was due to decreased targeting of p27 for proteolysis.

1,25-(OH) 2 Vitamin D 3 Treatment Results in Decreased Skp2
Levels-The SCF skp2 ubiquitin ligase complex directs Thr 187 phosphorylation-dependent p27 proteolysis. It was recently reported that the vitamin D analog EB1089 mediates up-regulation of p27 in AT-84 head and neck squamous carcinoma cells via down-regulation of Skp2 and decreased association of p27 with Skp2, which leads to p27 stabilization (50). We found that Skp2 protein levels were decreased by 1,25-(OH) 2 D 3 after 24 and 48 h of treatment (Fig. 6). Because Skp2 is destabilized as cells exit from the cell cycle (51), the down-regulation of Skp2 may be the result of growth arrest by 1,25-(OH) 2 D 3 . Thus, we examined a more proximal event of p27 degradation, namely the localization and activation of Cdk2.

1,25-(OH) 2 Vitamin D 3 Stabilization of p27
Correlates with Decreased Nuclear Localization of Cdk2-The late G 1 phase of p27 degradation that increases at the G 1 to S phase transition is dependent on Cdk2 phosphorylation of p27 at Thr 187 . For Cdk2 to be fully activated, newly synthesized Cdk2 must translocate into the nucleus to be phosphorylated by Cdk-activating kinase (CAK) and to be dephosphorylated by the phosphatase Cdc25A. CAK phosphorylates Cdk2 at Thr 160 , and Cdc25A dephosphorylates Cdk2 at Thr 14 and Tyr 15 (52,53). Regulation of Cdk2 nucleocytoplasmic trafficking is a poorly understood process. To test whether 1,25-(OH) 2 D 3 -mediated effects on p27 degradation might involve regulation of Cdk2 localization, vehicle-and 1,25-(OH) 2 D 3 -treated LNCaP cells were subjected to immunofluorescence staining of Cdk2. Examination of these cells using confocal microscopy revealed decreased Cdk2 nu-  (Fig. 7). Subcellular fractionation and Western blot analysis confirmed this shift to predominantly cytoplasmic Cdk2 localization (Fig. 8). These results suggest that 1,25-(OH) 2 D 3 -mediated growth inhibition involves decreased Cdk2 nuclear localization, which leads to the inhibition of cyclin-Cdk2 activity and the impairment of Cdk2-dependent p27 proteolysis. DISCUSSION 1,25-(OH) 2 D 3 has been shown to decrease the growth of many cancers, including prostate. The mechanism of the initial growth inhibitory events is by blocking the G 1 to S phase transition in the cell cycle (10,(12)(13)(14). 1,25-(OH) 2 D 3 -mediated growth inhibition has also been attributed to the induction of apoptosis (programmed cell death) (54). However, the dose and timing of 1,25-(OH) 2 D 3 treatment as well as the cell cycle stage and expression levels of cell cycle genes may dictate the response to hormone.
In this study, we show that 1,25-(OH) 2 D 3 -mediated induction of p27 protein levels in LNCaP prostate cancer cells occurred through a reduction in the rate of p27 degradation. 1,25-(OH) 2 D 3 -mediated stabilization of p27 correlated with a decrease in Thr 187 phosphorylation of p27, the critical phosphorylation event necessary for targeting p27 for cyclin E-Cdk2dependent proteolysis in late G 1 . We made the novel observation that 1,25-(OH) 2 D 3 decreased nuclear localization of Cdk2. Because full activation of Cdk2 by the regulatory CAK and Cdc25A requires nuclear translocation of Cdk2, cytoplasmic sequestration of Cdk2 would prevent cyclin E-Cdk2 activation and reduce the overall activity of this complex. Thus, in 1,25-(OH) 2 D 3 -treated cells, the decreased Cdk2 nuclear localization may also act to inhibit p27 degradation and may be a key event in 1,25-(OH) 2 D 3 -mediated growth inhibition of LNCaP cells.
1,25-(OH) 2 D 3 acts through its cognate receptor, the VDR, which is a ligand-activated transcription factor. Growth inhibition of prostate cancer cells by 1,25-(OH) 2 D 3 requires the VDR; however, VDR expression is not sufficient to cause this effect (9,11). p27 regulation by 1,25-(OH) 2 D 3 , however, occurs post-translationally through effects on p27 degradation. Transcriptional regulation of p27 by 1,25-(OH) 2 D 3 occurs in the myelomonocytic cell line U937 despite the p27 promoter lacking canonical vitamin D response elements (41). p27 up-regulation in these cells is a result of increased binding of Sp-1 and NF-Y factors to the p27 promoter, leading to enhanced p27 transcription. Because the expression and regulation of transcription factors vary from cell to cell, 1,25-(OH) 2 D 3 regulation of these factors may depend on cell-specific mechanisms that may be absent in LNCaP cells.
Translational regulation of p27 by 1,25-(OH) 2 D 3 has not been reported but does occur with other growth arresting agents such as lovastatin (43). The rate of p27 protein synthesis is governed by heterogeneous nuclear ribonucleoproteins C1, C2, and the embryonic lethal, abnormal vision (ELAV) family of RNA-binding proteins, which all have been implicated in the enhanced polysomal association of p27 mRNA (43)(44)(45)48). Currently, no evidence exists for 1,25-(OH) 2 D 3 -mediated regulation of heterogeneous nuclear ribonucleoproteins.
Our results showing p27 stabilization by 1,25-(OH) 2 D 3 are in accord with studies utilizing G 1 arresting agents that up-regulate p27 via post-translational mechanisms (31,50,55). p27 degradation is a complex process and involves phospho-Thr 187dependent and -independent mechanisms (30,31,(37)(38)(39). As cells traverse the G 1 to S phase transition, p27 levels decrease because of both decreased translation and mitogen-mediated Cdk2-independent p27 proteolysis in early G 1 . This initial reduction in p27 leads to cyclin E/Cdk2 activation. In late G 1 and early S phase, p27 degradation is governed by Cdk2-dependent phosphorylation of p27 at Thr 187 . Phosphorylation of p27 at Thr 187 allows p27 recognition by the ubiquitin ligase SCF skp2 complex, which ubiquitinates p27 and targets p27 for proteolysis. In LNCaP cells, 1,25-(OH) 2 D 3 appears to regulate the Cdk2-dependent phase of p27 proteolysis via sequestration of Cdk2 to the cytoplasm.
Similar to the actions of the vitamin D analog EB1089 on AT-84 head and neck squamous carcinoma cells, 1,25-(OH) 2 D 3 decreased Skp2 protein levels in LNCaP cells. Although the reduction in Skp2 levels may play a role in 1,25-(OH) 2 D 3mediated p27 stabilization in LNCaP, this effect may be a result of G 1 accumulation. Skp2 is rapidly degraded during the G 0 /G 1 phase of the cell cycle (51). Also, the lowered Skp2 levels cannot account for the 1,25-(OH) 2 D 3 -mediated decline in Thr 187 phosphorylation of p27. In contrast, decreased nuclear Cdk2 available for cyclin binding and activation would result in decreased phospho-Thr 187 -p27 targeted for degradation. Thus, decreased nuclear Cdk2 may be the more critical event in 1,25-(OH) 2 D 3 -mediated growth inhibition of LNCaP cells.
The nucleocytoplasmic trafficking of Cdk2 is not well understood. Reports suggest a possible role of the ERK signaling pathway in the regulation of Cdk2 translocation (52,53,56). The decrease in nuclear Cdk2 following 1,25-(OH) 2 D 3 treatment may occur via decreased nuclear import and/or increased nuclear export of Cdk2. Import of the cyclin E-Cdk2 complex into the nucleus occurs via the importin ␣/␤ system (57). The mechanism of Cdk2 nuclear export and its possible dependence The cells were then fractionated into nuclear and cytosolic fractions. Proteins from each respective fraction were subjected to immunoblot analysis with antibodies to Cdk2 or p27. Actin was used as a loading control. To assess the purity of the fractions, immunoblot analysis for the nucleolar protein nucleolin was performed. The data are representative of three independent experiments. B, quantification of Cdk2 levels was performed via densitometry. Statistical analysis was performed using an unpaired two-tailed t test. *, p Ͻ 0.01 versus vehicle-treated cells. Vit D, vitamin D; Nuc, nuclear; Cyt, cytosolic. on the nuclear export protein Crm1, however, is not known. Cytoplasmic Cdk2 mislocalization may be a key event in 1,25-(OH) 2 D 3 -mediated growth inhibition of LNCaP cells because this leads to decreased cyclin E/Cdk2 activity. Therefore, G 1 accumulation would be due to decreased Cdk2 availability in the nucleus for cyclin binding and decreased activation of cyclin/Cdk2 complexes by CAK and Cdc25A. In addition, mislocalization of Cdk2 would impair the Cdk2-dependent phase of p27 proteolysis. Thus, regulation of Cdk2 localization may be an important mediator of the fate of cells during the G 1 to S phase transition of the cell cycle.