LNCaP and 293T cells were purchased from the Korean Cell Line Bank (KCLB; no. 21573), Korean Cell Line Research Foundation, Seoul, Korea, in 2017. The cells were maintained in RPMI-1640 (Welgene, Daegu, Korea) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 0.1% antibiotic/antimycotic solution (Welgene) at 37°C in a 5% CO₂ humidified chamber.

HT29 colon cancer cells were used as a source of CD133. Cloning of CD133 was performed as previously described (15). Plasmid pcDNA3.1/N-terminal green fluorescent protein (NT-GFP) lacking the CD133 insert was used as a transfection control. 293T cells were used as a positive control to evaluate the translated fusion product.

Stable cell lines overexpressing the CD133 protein were obtained through co-transfection of confluent LNCaP-luciferase cells in 100-mm plates with 20 µg pcDNA3.1/NT-GFP:CD133 or pcDNA3.1/NT-GFP plasmid using the FuGENE HD transfection reagent, as previously described (15).

After reaching approximately 80% confluency, the medium was removed, and the cells were washed twice with phosphate-buffered saline (PBS, pH 7.4). Cell lysates were prepared in 200 µl cold lysis buffer [1% NP-40, 50 mM Tris-HCl, pH 7.5, 150 mM sodium chloride (NaCl), 0.02% sodium azide, 150 mg/ml of phenylmethylsulfonyl fluoride (PMSF), 2 mg/ml of aprotinin, 20 mg/ml of leupeptin, and 1 mg/ml of pepstatin A]. Approximately 30 mg of tissue lysate was separated via 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene difluoride (PVDF) membrane (Amersham, Piscataway, NJ, USA). Each membrane was blocked for 30 min with a blocking solution containing 5% skim milk in Tris-buffered saline containing Tween-20 (TBST, 2.42 g/l Tris-HCl, 8 g/l NaCl, 0.1% Tween-20, pH 7.6) and rinsed with TBST. The membrane was incubated overnight at 4°C with appropriate primary antibodies, including anti-CD133 (1:1,000; cat. no. MBS850595; MyBioSource, San Diego, CA, USA), anti-octamer-binding transcription factor 4 (Oct-4, 1:1,000; cat. no. 2750, Cell Signaling Technology, Beverly, MA, USA), anti-NANOG (1:1,000; cat. no. 3580; Cell Signaling Technology), anti-(sex determining region Y)-box 2 (SOX2, 1:1,000; cat. no. 2748; Cell Signaling Technology), anti-E-cadherin (1:1,000; cat. no. sc-8426; Santa-Cruz Biotechnology Inc., Dallas, TX, USA), anti-transcription factor 4 (TCF-4, 1:1,000; cat. no. 2953; Cell Signaling Technology), anti-vimentin (1:1,000; cat. no. sc-6260; Santa-Cruz Biotechnology Inc.), anti-β-catenin (1:1,000; cat. no. 9562; Cell Signaling Technology), and anti-macrophage migration inhibitory factor (MIF, 1:1,000; cat. no. 88186; Cell Signaling Technology). A mouse monoclonal immunoglobulin G (IgG) specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 1:1,000; cat. no. sc-47724; Santa-Cruz Biotechnology Inc.) was used as a control. The membrane was rinsed with TBST, and protein immunoreactivity was detected using an enhanced chemiluminescence detection kit (ECL, Amersham).

Immunolabeled cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) provided in ProLong Gold antifade mounting medium (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) to visualize nuclear morphology. Digital images were acquired at the Korea Basic Science Institute Gwangju Center using a TCS SP5 AOBS laser-scanning confocal microscope (Leica Microsystems, Heidelberg, Germany).

Cells were seeded at a density of 1,000 cells/well in non-adherent 24-well culture plates coated with a 10% poly (2-hydroxyethyl methacrylate) (polyHEMA) solution (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) in absolute ethanol and then dried overnight. After seeding, cells were incubated in serum-free Dulbecco's modified Eagle's medium (DMEM) supplemented with 200 ng/ml epidermal growth factor (Sigma-Aldrich; Merck KGaA), 20 ng/ml of basic fibroblast growth factor (Sigma-Aldrich; Merck KGaA), and B-27 supplement (Invitrogen; Thermo Fisher Scientific, Inc.). After 5 days of incubation, the number of spheroids in each well was counted using a light microscope (Zeiss, Zena, Germany).

Cells were seeded at 1×10⁵ cells/well in 6-well plates and cultured to approximately 90% confluency in 1 ml of DMEM supplemented with 10% FBS. The media were removed from the wells, and a straight transverse line was drawn through the adherent cells using a ruler and sterile 200-µl plastic micropipette tip to produce a uniform gap. Serum-free DMEM was added, and the distance between the gaps was measured immediately and 24 h later, following image capture of six random microscopic fields. The distance measured immediately was defined as 100%.

Cell invasion was assessed using a chemotaxis chamber (Neuro Probe, Gaithersburg, MD, USA). The cells (5×10⁴ cells in 0.35-ml serum-free DMEM per well) were seeded into the top chamber of a 10-well invasion chamber assay plate. DMEM supplemented with 10% FBS was placed in the lower chamber, and a matrigel-coated membrane was inserted between the two chambers. After incubation for 24 h at 37°C, the membranes were fixed and stained with a Hemacolor rapid staining kit (Merck KGaA). Cells from five random microscopic fields (each 0.5 mm²) were counted using a hematocytometer under a light microscope (Zeiss).

To confirm morphological changes, cells were washed thrice with PBS, fixed in 4% paraformaldehyde (PFA) for 10 min at room temperature, and permeabilized with PBS containing 0.25% Triton X-100 (PBST) for 10–15 min at room temperature. After three washes with PBS, cells were blocked with 1% bovine serum albumin for 30 min. Samples were incubated overnight at 4°C with the relevant antibodies anti-MIF (1:500; Cell Signaling Technology) and anti-F actin (1:500; Santa-Cruz Biotechnology Inc.) for 2 h at room temperature. After three washes with PBS, immune-labeled cells were counterstained with 50 µl of DAPI at 37°C for 10 min. Cells were analyzed using a laser scanning confocal microscope (Leica Microsystems).

Immunoprecipitation analysis was carried out as previously described (16). Briefly, cells were lysed using a protein lysis buffer [1% (v/v) Triton X-100, 150 mM NaCl, 50 mM Tris-HCl, pH 8.0, 5 mM ethylenediaminetetraacetic acid (EDTA), and protease inhibitors], and cell lysates were incubated overnight with primary anti-β-catenin and protein A or G beads (GE Healthcare, Uppsala, Sweden) at 4°C on a rotator. The beads were washed, boiled in Laemmli buffer, and processed for western blotting.

Five-week-old male athymic nude mice (BL-6/Nu, 20–25 g; Orient Bio Co., Ltd., Seoul, Korea) were housed under controlled light conditions and provided water ad libitum. All experimental procedures involving animals were performed in compliance with institutional and governmental requirements and approved by the Institutional Animal Care and Use Committee (approval no. CIACUC2015-A0032), Chosun University, Gwangju, Korea.

Twenty mice were subjected to intracardiac injections of LNCaP^Vec or LNCaP^CD133+ cells and 10 mice subjected PBS as mock mice to examine the ability of PC cells to metastasize to bone, respectively. Before injection, the cells were transfected with the pGL4.5 vector plasmid encoding luciferase (Promega Corporation, Madison, WI, USA). The stably transfected cells were grown in Hank's balanced salt solution (HBSS; Welgene) supplemented with hygromycin (2 mg/ml; Sigma-Aldrich; Merck KGaA) and, luciferase activity was measured using the Luciferase Assay System (Promega Corporation) and detected by luminometer GENios Plus (Tecan Group Ltd., Salzburg, Austria). The severe combined immunodeficient mice were anesthetized using isoflurane gas. Subsequently, LNCaP^Vec+ or LNCaP^CD133+ cells (1×10⁵ cells per mouse) were injected into the left cardiac ventricle, as per the technique first described by Arguello et al (17) with some modifications. After 4 weeks, images of tumor-bearing tissues excised from mice during necropsy were obtained.

After intracardiac injection, mice were weekly imaged via bioluminescence for 4–6 weeks using the IVIS^® imaging system (PerkinElmer, Waltham, MA, USA) at the Korea Basic Science Institute (Gwangju, Korea). Images were captured and processed using the Living Image^® v.4.2 software. Anesthesia was induced using inhaled isoflurane and maintained with 2% isoflurane mixed with oxygen/nitrogen via nose cone delivery. d-Luciferin (3 mg dissolved in water) was administered at 150 mg/kg in Dulbecco's phosphate-buffered saline (DPBS) via intraperitoneal injection. Optical imaging was acquired approximately 10–15 min later.

Tumor-bearing tissues were fixed in cold 4% PFA. Bone tissue was first decalcified using a sodium citrate solution before processing onto histological slides. Decalcified bones were cut at the midpoint and embedded in paraffin blocks. Fluorescence from serial paraffin sections was monitored via fluorescence microscopy (Leica Microsystems). Tissues were stained with hematoxylin and eosin (H&E) stain, and images were acquired using a microscope slide scanner (3D-HISTECH Ltd., Budapest, Hungary).

Supernatants from LNCaP^Vec and LNCaP^CD133+ cells and sera from tumor-bearing mice were collected and assayed using a cytokine array kit (R&D Systems, Minneapolis, MN, USA), according to the manufacturer's instructions. The cytokines examined using this technique are listed in Table I. Cytokines were detected using the ECL detection kit (Amersham) and quantified via densitometric analysis using ImageJ software (Scion Corp, MD, USA).

Total RNA was extracted from cells using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.). cDNA was synthesized from 2 µg total RNA using the Super-Script II First-Strand Synthesis System (Invitrogen; Thermo Fisher Scientific, Inc.). mRNA levels were measured via qPCR, and the Gapdh gene was used as an endogenous control. Sequences of the primers used to target various genes are listed in Table II.

Paraffin sections were deparaffinized in three solutions of xylene and rehydrated in a graded series of ethanol solutions. For antigen retrieval, slides were placed in 0.01 M citrate buffer at pH 6.0 and heated in a steamer for 30 min. Endogenous peroxidases were quenched by incubating the samples with 3% hydrogen peroxide for 20 min at room temperature. Sections were incubated overnight at 4°C using a 1:50 dilution of the primary antibody against MIF (Santa-Cruz Biotechnology Inc.). Sections were then incubated for 30 min with a biotinylated secondary antibody (LSAB system HRP kit; DakoCytomation, Glostrup, Denmark), rinsed in PBS, and incubated for 30 min with a streptavidin-peroxidase conjugate (LSAB; DakoCytomation). The reaction was developed for 5 min using 3, 30-diaminobenzidine tetrahydrochloride (Sigma-Aldrich; Merck KGaA). Slides were counterstained in hematoxylin, dehydrated, and covered with a cover slip. Negative and positive controls were simultaneously analyzed. Positive controls comprised mammary tissues. The slides were captured using a microscope slide scanner (3D-HISTECH Ltd., Budapest, Hungary).

One-way analysis of variance (ANOVA) followed by Sidak's multiple comparison test (unless specifically mentioned otherwise) were used for statistical analyses. P<0.05 was considered to indicate a statistically significant difference. Data are expressed as the mean ± standard deviation (SD) unless specified otherwise. Data were analyzed using the SPSS v.20.0 software program for Windows (SPSS, Inc., Chicago, IL, USA). GraphPad Prism v.6.00 software program for Windows (GraphPad, La Jolla, CA, USA) was used to analyze data from in vitro and in vivo experiments.

To evaluate the effect of CD133 overexpression in vitro, we established a stable cell line overexpressing CD133 under the control of a constitutive promoter. We transfected LNCaP cells with either CD133 or an empty vector tagged with GFP. A representative CD133 clone was selected from LNCaP^CD133+ cells and compared against control cells transfected with the empty vector (LNCaP^Vec). Transiently transfected 293T^Vec or 293T^CD133+ cells were used as a positive control. Basal expression of CD133 protein in LNCaP^Vec cells was very low; however, the basal expression of CD133 was high in stable CD133-transfected cells (LNCaP^CD133+) and transiently transfected 293T cells (Fig. 1A). Green fluorescence was observed in the cytosol and membrane of LNCaP^CD133+ cells (Fig. 1B).

To explore the function of CD133 in cancer stemness, we evaluated the acquisition of CSC properties such as increased ability to form tumor spheres and expression of stem cell markers such as Oct-4, SOX2, and NANOG. The expression levels of Oct-4 and NANOG were significantly elevated in LNCaP^CD133+ cells, consistent with the increase in stemness properties (Fig. 1C). Stemness was also confirmed by a colony-formation assay. Colony-forming ability was significantly elevated in LNCaP^CD133+ cells, in line with the elevated expression of stemness factors (Fig. 1D). These results suggest that the increase in colony-forming ability and Oct-4 and NANOG expression conferred stem cell-like properties during ectopic overexpression of CD133.

CD133 overexpression in LNCaP cells, as determined by wound healing and cell invasion assays, induced a significant increase in cell migration and invasion. After 24 h, the artificial wound gaps became significantly narrower in LNCaP^CD133+ cells than in LNCaP^Vec cells (Fig. 2A). Furthermore, the invasive abilities of LNCaP^CD133+ cells significantly increased (Fig. 2B). To monitor morphological changes in LNCaP cells, immunofluorescence analysis for F-actin was performed using confocal microscopy. The number of spindle and discontiguous cells observed in the LNCaP^CD133+ group was higher than that observed in the LNCaP^Vec group (Fig. 2C).

To identify factors associated with altered invasiveness and migration caused by ectopic overexpression of CD133, we monitored levels of epithelial and mesenchymal markers in PC cells. Protein expression analyses revealed a slightly lower expression of epithelial marker E-cadherin in LNCaP^CD133+ cells than in LNCaP^Vec cells (Fig. 2D). In contrast, expression of the mesenchymal marker vimentin was significantly upregulated in LNCaP^CD133+ cells. Further, analysis of β-catenin-associated signal transduction was carried out via immunoprecipitation in LnCaP^CD133+ cells (Fig. 2E). CD133 overexpression led to a significant increase in binding between β-catenin and TCF-4, an EMT-inducing transcriptional factor in cancer cells in compared with LnCaP^vec cells as a negative control. These data strongly support the idea that CD133 is essential for increased and decreased expression of vimentin and E-cadherin, respectively, leading to subsequent EMT and metastasis in PC cells.

Both of LNCaP^Vec and LNCaP^CD133+ cells were labeled with luciferase, and the cells showed a similar luciferase activities (Fig. 3A). An then we injected LNCaP^Vec and LNCaP^CD133+ cells into the left cardiac ventricle of male nude mice at 0.5×10⁶ cells/mouse. The mice were imaged weekly, starting from week 4 after intracardiac injection. The results of PBS (mock) injection and injection of LNCaP^Vec and LNCaP^CD133+ cells are shown in the whole-mouse image in Fig. 3B. A large hot-spot and a small spot of bioluminescence were observed in the whole body following LNCaP^CD133+ cell injection. Only 20% of animals inoculated with LNCaP^Vec cells developed skeletal tumors (n = 10) as compared with 80% of those inoculated with LNCaP^CD133+ cells (Table III).

We performed H&E staining to examine the histological features of skeletal tumors formed by LNCaP^Vec/LNCaP^CD133+ cells inoculated into mice via intracardiac injection. As shown in Fig. 3C, gross examination of excised spines in mice injected with LNCaP^Vec or LNCaP^CD133+ cells showed that the tumor mass occupied the primary spongiosum (trabecular epiphysis) and displaced the bone marrow cells. In particular, an apparent margin between the spinal cord and metastatic tumors was observed in LNCaP^Vec-inoculated mice. However, metastatic tumors invading the spinal cord and discursive osteolytic features of vertebrae were observed in LNCaP^CD133+ cell-inoculated mice. We aimed to identify the secreted molecules that played a critical role in LNCaP^CD133+ cell-induced bone metastasis through a cytokine screening using a proteome array cytokine kit (Fig. 3D). CD133 overexpression in LNCaP cells led to a marked increase in the release of interleukin 8 and MIF from mouse serum. MIF expression increased in the supernatants of LNCaP^CD133+ cells. These data suggest that ectopic overexpression of CD133 might lead to a significantly increased risk of bone metastatic cancer. Further, increased MIF expression may contribute to PC metastasis in the bone.

To investigate the role of MIF in bone metastasis of LNCaP cells, its mRNA and protein expression levels were analyzed. MIF protein (Fig. 4A) and mRNA (Fig. 4B) levels were higher in LNCaP^CD133+ cells than in LNCaP^Vec cells. To monitor the cellular distribution of MIF in LNCaP cells, MIF immunofluorescence was performed. MIF expression in the membrane and cytosol was higher in LNCaP^CD133+ cells than in LNCaP^Vec cells (Fig. 4C).

In the immunohistochemical study, MIF expression increased around the tumor mass and trabecular epiphysis region of bones inoculated with LNCaP^CD133+ cells. These observations suggest that ectopic overexpression of CD133 leads to increased MIF expression in CD133+ cells and activation of bones adjacent to the tumor, particularly in LNCaP cell-injected bones, and that CD133 expression is associated with PC metastasis to the bone.