To examine the metabolic activity of CYP24A1 in the event of CYP24A1 manipulation, normal phase HPLC analysis was performed using ³H-labelled 1α,25(OH)₂D₃ on an Alliance 2695 separation module, equipped with a 996 photodiode array detector (Waters, Milford, MA, USA), as described elsewhere in detail (11).

The breast cancer cell lines MCF7 and MDA-MB-231 were obtained from a local distributor (Summit Pharmaceuticals International, Tokyo, Japan) of the American Type Culture Collection (Manassas, VA, USA). MCF7 cells are estrogen receptor (ER)-positive cells, and MDA-MB-231 cells are ER-negative cells. All cells were maintained in Dulbecco's modified Eagle's medium (Sigma, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, Carlsbad, CA, USA), 100 U/ml penicillin, and 100 µg/ml streptomycin (Sigma).

Cells were transfected with different types of CYP24A1-specific small-hairpin RNA (shRNA)-expressing lentivirus plasmids (Sigma) using FuGENE6 (Roche, Basel, Switzerland) to generate stable transfectants. Transfected clones were selected in 0.6 µg/ml puromycin (Sigma). Drug-resistant clones were picked after more than 14 days of selection and screened for CYP24A1 expression. We obtained the following MDA-MB-231 cell transfectants: CYP24A1 shRNA #1296 (harboring shRNA clone NM_000782.2-1296s1c1) and CYP24A1 shRNA #1016 and CYP24A1 shRNA #1016S (harboring shRNA clone NM_000782.2-1016s1c1). CYP24A1-shRNA #1016S is a transfectant originating from a single clone that showed efficient suppression of constitutive CYP24A1. A negative-control shRNA was used as a control. Cells were also transfected with CYP24A1-specific small-interfering RNAs (siRNAs; Santa Cruz Biotechnology, Santa Cruz, CA, USA), or a plasmid containing the full-length CYP24A1 cDNA (OriGene, Rockville, MD, USA).

Total RNA was extracted using the TRIzol reagent (Invitrogen), and subsequent RT-PCR was performed using the Superscript II Reverse Transcriptase kit (Invitrogen). Samples were incubated at 42°C for 50 min, after which the reactions were terminated by incubation at 70°C for 15 min. Then, the cDNA was incubated with 0.5 U of Taq DNA polymerase (Takara, Shiga, Japan) and appropriate primers to amplify the genes of interest. The cycling conditions were as follows: 20–40 cycles of 30 sec at 96°C, 30 sec at 58°C, and 1 min at 72°C, followed by a final elongation time of 7 min at 72°C. The sequences of the PCR primers used are available upon request. The expression of each gene of interest was analyzed using cycling parameters that were optimized previously for detection linearity, allowing for semiquantitative analysis of signal intensities. PCR experiments were repeated in 3 independent experiments to ensure that the quantified expression was reproducible. Densitometric analyses of gel bands were performed using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Gene-expression changes caused by CYP24A1 suppression were examined by microarray analysis. Total RNA was analyzed using the Agilent Whole Human Genome Oligo Microarray (4×44K; Agilent Technologies, Santa Clara, CA, USA). Gene-expression analysis was outsourced to Bio Matrix Research (Chiba, Japan). Briefly, total RNA (100 ng) was converted into cDNA, and complimentary RNA (cRNA) was synthetized by in vitro transcription and subsequently labeled with cyanine 3-CTP and cyanine 5-CTP, using the Low Input Quick Amp Labeling kit. Labeled cRNA was hybridized with a microarray chip for 17 h using the Gene Expression Hybridization kit, and slides were scanned using an Agilent Microarray Scanner and Feature Extraction Software 9.5, with default settings. Raw data were normalized using the Quantile algorithm of the Gene Spring Software package, version 11.0 (all from Agilent).

Aliquots of whole cell lysates (20 µg) were separated on 12% sodium dodecyl sulfate-polyacrylamide gels and electroblotted onto nitrocellulose membranes. Membranes were then immunoblotted with antibodies against CYP24A1 (sc-32166; 1:50, Santa Cruz Biotechnology), the cleaved form of caspase-3 (sc-22171-R; 1:75, Santa Cruz Biotechnology), Pdlim2 (SAB1407872; 1:100, Sigma), and β-actin (A5316; 1:1,000, Sigma). The membranes were incubated with appropriate peroxidase-labeled secondary antibodies (1:2,000, Dako, Glostrup, Denmark), and bands were visualized using enhanced chemiluminescence (GE Healthcare, Buckingham, UK).

Oxidative stress-induced apoptosis was stimulated by incubating cells in 0-100 µM H₂O₂ for 24 h. Anoikis was induced by adding MDA-MB-231 cells to agarose-coated dishes to avoid cell attachment in the presence or absence of 100 nM 1α,25(OH)₂D₃ (Sigma).

Apoptosis was assessed in cells cultured on collagen-coated glass coverslips by performing TUNEL assays. In addition, tumor tissues developed in mice were fixed in 70% methanol and studied in TUNEL assays. Apoptotic cells were visualized using an In Situ Cell Death Detection kit (11684817910; M7240, Roche). The procedure was also performed without terminal deoxynucleotidyl transferase, as a negative control. To examine the cell-proliferation rate, cells were manually counted every 24 h up through day 7 after an equal number of cells had been plated. In addition, cellular DNA synthesis was assessed by immunohistochemical labeling of Ki-67 (MIB-1 clone; 1:100, Dako). The cells of interest were counted under a light microscope under low magnification (×100) in 10 randomly selected fields in each section. The results were confirmed in triplicate independent analyses.

Soft agar assays were performed in 6-cm dishes to assess colony formation in 3 dimensions. Cells (2.5×10³) were uniformly suspended in 6 ml of 0.33% agarose gel with growth medium supplemented with 5% FBS and then overlaid onto the base layer of 1% agarose gel. Plates were incubated for at least 3 weeks, and cell clusters approximately >50 µm in diameter were defined as positive. Colonies developing on cell culture dishes in soft agar suspensions were counted using phase-contrast microscopy under low magnification (×100) in 10 separate random fields in each plate. For quantitative analysis, the number of colonies formed by control cells was defined as 100%.

We injected cells of parental MDA-MB-231 and its transfectants (2×10⁵ cells in 50 µl/mouse) into the mammary fat pads of 6 or 7-week-old female athymic nude mice (Charles River Japan, Yokohama, Japan). Animals were euthanized when skin ulceration of the primary tumor occurred. Tumor volumes were calculated in 2 dimensions in a time-dependent manner, after the inoculation of cells in independent duplicate experiments. The volume (V) of the primary tumors was calculated by the following equation: V = (π/6) × (L × W²), where L is the length representing the largest tumor diameter and W is the perpendicular width of tumors. We examined primary tumors macroscopically for the formation of tumors, and then the tumors were subjected to histological evaluations. The maintenance and handling of animals were conducted using protocols approved by the Animal Care Committee of Kochi University School of Medicine.

Unless otherwise specified, all data are expressed as means ± standard deviations from at least 3 independent experiments, each performed in triplicate wells. Statistical differences were analyzed using Student's t-test. A P-value of <0.05 was considered statistically significant.

We transfected MDA-MB-231 cells with siRNAs against CYP24A1 to examine the suppressive effect of CYP24A1 silencing on vitamin D metabolism by HPLC analysis (Fig. 1A). We also transfected different types of CYP24A1-specific shRNAs into MDA-MB-231 cells to establish several different stable transfectants with various CYP24A1 expression levels. Although our preliminary observations revealed that the phenotypes of at least 3 independent clones were similar, the cell populations were mixed to establish a stably transfected cell line to avoid possible clonal variation within the clonal transfectants. Finally, we obtained pools of transfectants, designated as CYP24A1 shRNA #1296 and CYP24A1 shRNA #1016. To select for clones with more specific phenotypes, we generated a clonal transfectant designated as CYP24A1 shRNA #1016S, which was isolated following independent transfections and subsequent screening for low CYP24A1 expression. CYP24A1 expression was lower in the shRNA transfectants, compared to the constitutive CYP24A1 expression observed in control cells (Fig. 1B). CYP24A1 expression was significantly lower in CYP24A1 shRNA #1016 cells than in CYP24A1 shRNA #1296 cells. A striking effect of CYP24A1-specific shRNA was observed in CYP24A1 shRNA #1016S cells, which showed no detectable CYP24A1 protein expression. In addition, the CYP24A1 expression level was closely correlated with the metabolism of 1α,25(OH)₂D₃ in each transfectant (Fig. 1C).

We examined alterations in the gene-expression profile following CYP24A1 suppression, using microarray analysis to test the possibility that mRNAs induced or suppressed by CYP24A1 silencing were responsible for the tumor pathology (Fig. 2). Breast tumors expressing hormone receptors, such as ER, generally display a less aggressive phenotype and are associated with prolonged disease-free and overall survival. To exclude the possibility that the hormone receptor status might affect the tumor pathology, we examined gene-expression alterations following shRNA-mediated stable CYP24A1 silencing, both in ER-positive MCF7 cells and ER-negative MDA-MB-231 cells, although these cell lines show different oncogenic behaviors. We found that the changes in gene-expression patterns between these cell types were different, but at least shared some similarities (Fig. 2A). CYP24A1 silencing affected the expression of an extensive range of oncogenic pathway members, including oncogenes, tumor-suppressor genes, and components of the apoptosis machinery (Fig. 2B and C). We also observed altered signaling in pathways mediating cell growth and differentiation, as well as with a variety of signal-transduction molecules. As evidenced by the altered expression of cell-adhesion molecules, CYP24A1 suppression may affect the tumor microenvironment and tumor cell-interstitial cell interactions. Signaling regulating angiogenesis was also modulated by silencing CYP24A1 expression. Further, our data provided evidence of the possible association of CYP24A1 downregulation with gene families involved in tumor cell invasion and metastasis, such as different types of matrix metalloproteinases. PDLIM2 expression was also associated with CYP24A1 expression (Fig. 2D), which is consistent with previous data showing that PDLIM2 expression is driven by vitamin D (12). The alteration of gene expression observed suggested that many of the genes involved in cancer-related pathways are modulated to favor cell survival when CYP24A1 is constitutively expressed.

To investigate the effect of CYP24A1 signaling on cell proliferation, cell numbers were analyzed in a time-dependent manner. Manual cell counting revealed an appreciable, but not significant reduction in cell growth, regardless of the presence of 1α,25(OH)₂D₃. However, CYP24A1-silenced cells that were exposed to 1α,25(OH)₂D₃ exhibited significantly decreased cell growth (Fig. 3A) and DNA-synthesis rates (Fig. 3B).

We also examined the effect of altered CYP24A1 expression on apoptosis because 1α,25(OH)₂D₃ is known to exert pro-apoptotic effects in cells (6). When cells were incubated under oxidative stress caused by H₂O₂, TUNEL assay results consistently revealed that 1α,25(OH)₂D₃ treatment enhanced the sensitivity of cells to H₂O₂-induced cell death, and CYP24A1 suppression conferred a significant increase in sensitivity to apoptosis (Fig. 3C). The CYP24A1 expression level inversely correlated with the apoptosis-sensitizing effect caused by 1α,25(OH)₂D₃, suggesting that the catabolic activity of the enzyme was important for apoptosis sensitivity. In addition, cells with downregulated CYP24A1 expression gained increased sensitivity to anoikis, and 1α,25(OH)₂D₃ enhanced this effect (Fig. 3D).

Although we observed that the basal levels of cell growth and apoptosis differed between the control- and CYP24A1-shRNA transfectants, there was evidence that normal FBS contained active concentrations of various growth and differentiation factors, including 1α,25(OH)₂D₃, which can modulate the transcriptional machinery driven by these biologically relevant factors (13). This finding was also explained by the observation that basal levels of cell growth and apoptosis were unchanged following CYP24A1 suppression in media supplemented with charcoal-dextran-treated FBS (data not shown). These data indicated that the expression level of CYP24A1 contributes to determine the sensitivity of cells to apoptosis.

Consistent with the apoptosis-sensitizing effect of CYP24A1 inhibition, CYP24A1 downregulation significantly suppressed colony formation in soft agar, as evidenced by a decreased number and size of colonies (Fig. 4A and B). Treatment with 1α,25(OH)₂D₃ inhibited anchorage-independent growth, and this effect was more evident following CYP24A1 suppression (Fig. 4A). Our observations suggested that the expression level of CYP24A1 was significantly associated with the colony-formation capability of cells, supporting that CYP24A1 expression had a stimulatory oncogenic effect on breast carcinoma cells.

To address whether CYP24A1 affects tumorigenicity in vivo, we employed an orthotopic-transplantation mouse model. CYP24A1 suppression inhibited tumorigenicity in vivo, showing a markedly decelerated rate of tumor growth (Fig. 5A and B). In addition, the cells with lower metabolic activity showed greater retardation of tumor formation in mammary fat pads. CYP24A1-expressing tumors showed high-grade malignant characteristics, such as invasive growth into the surrounding tissue (Fig. 5C). However, CYP24A1-suppressed tumors did not show invasiveness, and the tumor tissues underwent massive necrosis, which was accompanied by increased apoptosis and reduced mitosis (Fig. 5D). These findings are partially explained by the ability of CYP24A1 suppression to cause upregulation of the anti-invasion marker PDLIM2 (Fig. 2D). Histological examinations also revealed a marked decrease of CD34-positive microvascular components in tumors developed from CYP24A1-silenced cells (Fig. 5D). Consistently, cancer cells with CYP24A1 downregulation showed significantly lower expression levels of pro-angiogenic genes, such as angiopoietin (Fig. 2B). This observation was in good agreement with evidence that vitamin D can inhibit angiogenesis (6).