Human ovarian cancer cell lines (SKOV-3 and OVCAR-3) and human embryonic kidney 293T (HEK293T) cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were grown according to the manufacturer's instructions. Wortmannin and LY294002 chemicals were obtained from Sigma (St. Louis, MO, USA). The primary antibodies used in this study were as follows: anti-BMI-1, anti-cyclin D1, anti-CDK4, anti-Mcl-1, anti-p53, anti-phospho-PI3K (Tyr508), anti-PI3K, anti-phospho-Akt (Ser473), anti-Akt (Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-sMEK1 (Abcam, Cambridge, UK), anti-p21, anti-p27, anti-NF-κB (Ab-1; Oncogene, Cambridge, MA, USA), anti-Bax, anti-Bcl-xL, anti-Bcl-2, anti-phospho-PDK1, anti-PDK1, anti-phospho-mTOR, anti-mTOR, anti-phospho-4E-BP1, anti-4E-BP1 (Cell Signaling, Beverly, MA, USA), and β-actin (Sigma).

For whole protein extraction, cultured ovarian cancer cells were harvested, rinsed with PBS, centrifuged, and disrupted by adding cell lysis buffer containing protease inhibitor cocktail (50 mM Tris, pH 7.2, 150 mM NaCl, 1% Triton X-100, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 2 µg/ml aprotinin, and 200 µg/ml phenylmethylsulfonyl fluoride) at 4°C for 1 h. Proteins were subsequently separated by 8–12% SDS-PAGE and transferred to Immobilon P membranes (Millipore Corp., Billerica, MA, USA). After blocking, the membranes were incubated with the indicated primary specific antibodies at 4°C overnight. The membranes were rinsed three times in TBST washing buffer and incubated with horseradish peroxidase-conjugated secondary antibodies. Protein bands were developed using the ECL detection system (GE Healthcare, Little Chalfont, Buckinghamshire, UK).

For co-immunoprecipitation, cells were rinsed in phosphate-buffered saline (PBS) and dissolved in cell lysis buffer including 50 mM Tris-HCl, pH 7.2, 150 mM NaCl, 1% Triton X-100 and a protease inhibitor cocktail (Sigma). Lysates were then incubated with anti-Flag antibody (Santa Cruz) and precipitated using protein A-agarose (Invitrogen, Carlsbad, CA, USA). Precipitated proteins were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transferred to an Immobilon-P membrane (GE Healthcare, Piscataway, NJ, USA), and subjected to immunoblot analysis using either anti-BMI-1 or anti-sMEK1 antibodies. Enhanced chemiluminescent (ECL) western blotting detection reagent (Pierce, Rockford, IL, USA) was used to visualize the gels. β-actin served as an internal control.

Oligonucleotides containing the BMI-1 siRNA sequence 5′-ATAT GAAGAGAAGAAGGGATT-3′ were synthesized using an RNAi construction kit (Ambion, Austin, TX, USA) and transfected into cells using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions to silence endogenous BMI-1 expression. Cells were seeded for 24 h followed by transfection with 150 nM siBMI-1. At 24 h after transfection, cells were prepared for MTT assays and immunoblot analysis.

Cells were seeded on 6-well plates at a density of 1×10⁵ per well and transfected. Next, cells were incubated with FITC-labeled Annexin V and propidium iodide (PI) for 15 min according to the manufacturer's instructions (BD Pharmingen, Mississauga, ON, Canada) and analyzed using a fluorescence activated cell sorting (FACS) Vantage BD FACSCalibur flow cytometer. To determine cell viability, the relative rate of cell viability was evaluated via MTT assay. Cells were grown at a density of 4.5×10³ per well in 96-well plates. Three days after transfection, fresh medium including 10% fetal bovine serum (FBS) and 20 µl of 3-(4,5-dimethylthiazol-2-yl)-2.5-diphenyl-²H-tetrazolium bromide (MTT) solution (Sigma, 5 µg/ml) was added to each well and incubated for an additional 4 h at 37°C. After centrifugation at 500 g for 10 min, the supernatant was removed from the wells, and the formazan product was dissolved in dimethyl sulfoxide (DMSO). The amount of MTT-formazan added was determined by the absorbance, calculated using a microplate reader at 540–550 nm.

BMI-1 luciferase activity in vitro was carried out as reported previously (38). In brief, cells at 85% confluency were transfected using a BMI-1 reporter plasmid vector. After lysis using RIPA buffer, lysates were cleared by centrifugation at 14,000 rpm for 15 min, and cell extracts were incubated with a luciferase substrate reagent at room temperature for 30 min according to the manufacturer's instructions. Five-microliter aliquots of each sample were measured using a MicroLumat Plus LB96V luminometer (Berthold Technologies, Bad Wildbad, Germany).

All results are presented as the means ± standard deviations (SD) from at least three independent experiments. Statistical comparisons between the different groups were analyzed using the Student's t-test. Significance was set at P<0.05.

To explore the cellular biological and physiological relevance of BMI-1 on the effects of tumor metastasis, we performed yeast two-hybrid and co-immunoprecipitation assays to identify potential BMI-1-interacting partners. The human cDNA library fused to the transcriptional activator pJG4-5/B42 gene was introduced into yeast cells containing the pGilda/LexA-BMI-1 as bait. Approximately 6.0×10⁶ independent transformants were pooled. Five positive colonies were obtained after respreading on selection media (Ura-, His-, Trp-, and Leu-). Of the five sequenced clones, all encoded sMEK1 (accession no. NM_001284280), indicating that BMI-1 interacted with sMEK1. To confirm these results, positive interaction was monitored by both cell growth on leucine-deficient plates and ONPG β-galactosidase activity. An empty plasmid (vector only) served as a negative control. As shown in Fig. 1A, β-galactosidase activity was fully activated, indicating an interaction between BMI-1 and sMEK1 (84.91±0.91), but such was not observed with the empty plasmid (vector only, 1.86±0.63). To further confirm a direct interaction between BMI-1 and sMEK1 as revealed in the yeast two-hybrid assay, co-immunoprecipitation experiments were used. Gene constructs of sMEK1 (pcDNA3.1/Flag-sMEK1) and BMI-1 (pcDNA3.1/BMI-1) or sMEK1 (pcDNA3.1/Flag-sMEK1) and vector only (pcDNA3.1), were co-transfected into ovarian carcinoma cells. Next, immunoprecipitation was performed using an anti-Flag antibody in lysates from both transfected cell lines. After immunoprecipitation, precipitated proteins were subjected to immunoblotting using anti-BMI-1 or anti-sMEK1 antibodies. As shown in Fig. 1B, pcDNA3.1-BMI-1 co-immunoprecipitated with pcDNA3.1/Flag-sMEK1 (lane 2, upper panel), but not with pcDNA3.1 (vector only) (lane 1, upper panel). Subsequently, cellular interaction between these two proteins was also demonstrated by co-immunoprecipitation of endogenous BMI-1 and sMEK1. As shown in Fig. 1C, endogenous BMI-1 co-immunoprecipitated directly with sMEK1. We next evaluated BMI-1-induced cell growth by determining the expression of the apoptosis-related proteins Bax, Bcl-xL, Mcl-1 and Bcl-2. These proteins are well-known pivotal regulators of apoptotic cell death and growth. As presented in Fig. 1D, Bax expression was decreased, while Bcl-xL, Mcl-1 and Bcl-2 were increased by BMI-1, compared with the non-transfectants. Consistent with these results, cells transfected with a specific BMI-1 RNAi (siBMI-1) showed enhanced Bax expression, whereas Bcl-xL, Mcl-1 and Bcl-2 were decreased remarkably by siBMI-1. Collectively, these results strongly indicate that BMI-1 directly interacts with sMEK1 under biologic conditions.

Cell cycle control is a fundamental process in cellular homeostasis and involves DNA repair, DNA replication and cell division. This regulation is generally mediated by cyclin and cyclin-dependent kinases (CDKs). In particular, CDKs are serine/threonine kinases that play an important role in the complex feedback regulation of cell cycle progression. Therefore, we examined expression levels of cell cycle- and apoptotic cell death-associated proteins via western blot analysis. As shown in Fig. 2A, overexpression of BMI-1 increased the expression of cyclin D1 and CDK4 remarkably, but decreased the expression of p21 and p27, acting as a CDK inhibitor. These results demonstrate that overexpression of BMI-1 strikingly interrupts sMEK1-induced apoptosis (Fig. 2B). Nuclear factor kappa B (NF-κB) and Bcl-2 family genes, as well as the p53 tumor suppressor, are major modulators of cell growth and apoptosis. To explore the functional effects of BMI-1 on sMEK1-induced apoptotic cell death, cells were transfected with either a BMI-1 expression plasmid or siBMI-1, and promoter activity was then measured using a dual luciferase reporter gene assay. As presented in Fig. 2B, the promoter activities of p21 and p53 were significantly reduced, whereas Bcl-2 and NF-κB activity were noticeably enhanced by BMI-1 overexpression. Taken together, these results demonstrate clearly that BMI-1 controls apoptosis modulator protein levels through downregulation of p53 and upregulation of NF-κB activity during sMEK1-induced apoptosis in tumor cells.

To explore the biological function of BMI-1 during sMEK1-induced apoptosis, ovarian carcinoma cells were transfected with BMI-1, sMEK1, or sMEK1 plus BMI-1. Control transfectant contained the expression vector only. According to the reduced number of cells, which is indicative of cell viability, sMEK1-transfected cells were suppressed to ~60% compared with control cells. In contrast, enhanced viability of the BMI-1 and sMEK1 plus BMI-1 transfectants was observed (Fig. 3A, left panel). Furthermore, after cotransfection with BMI-1 (0.5 µg) and sMEK1 (0–0.5 µg), apoptotic cell death increased gradually in a dose-dependent manner (Fig. 3A, right panel). Subsequently, flow cytometric analysis confirmed that the loss of viability was in fact due to cell death. Cells were transfected with a control vector (expression plasmid vector only), BMI-1, sMEK1, or a combination of sMEK1 and BMI-1. As shown in Fig. 3B, cell growth of ovarian cancer cells was inhibited by sMEK1 transfection compared with that of control transfectants. Interestingly, cell viability was recovered in sMEK1-expressing BMI-1 transfectants (sMEK1 plus BMI-1) compared with sMEK1-expressing only transfectants, whereas sMEK1-expressing siBMI-1 transfectants (sMEK1 plus siBMI-1) exhibited suppression of cell viability by 20% compared with siBMI-1-expressing cells (siBMI-1 plasmid only; data not shown). These results strongly indicate that BMI-1 has an antagonistic effect on sMEK1-mediated apoptotic cell death.

To explore the underlying molecular mechanism by which BMI-1 promotes PI3K activation, we examined the involvement of PDK1, Akt, mTOR, and 4E-BP1, all of which are downstream signaling cascade components of the PI3K pathway. Cell lysates from sMEK1-expressing cells (control) and BMI-1 transfected cells were examined by western blot analysis. BMI-1-induced phosphorylation of PI3K, Akt and mTOR plays a major role in BMI-1-stimulated tumorigenesis, migration, invasion, and tumor metastasis. As presented in Fig. 4, sMEK1-induced PI3K and Akt dephosphorylation was significantly restored by BMI-1. Ectopic BMI-1 expression also enhanced Akt phosphorylation in a dose-dependent manner (Fig. 4B). These data were comparable to those of wortmannin and LY294002, well-known PI3K inhibitors. Wortmannin and LY294002 were shown to additively decrease (by 40–45%) the PI3 kinase activity induced by sMEK1 (Fig. 4A). We next investigated the phosphorylation of essential components of the mTOR/4E-BP1 signaling pathways. As presented in Fig. 4C, sMEK1 significantly reduced PDK1 phosphorylation, as well as the phosphorylation of mTOR and 4E-BP1. Collectively, our results indicate that BMI-1 recovers sMEK1-stimulated apoptotic cell death through activation of the PI3K/mTOR/4E-BP1 signaling pathway in ovarian tumor cells.