Subjects and tumor cell lines
A total of 45 liver tumor specimens were obtained from pediatric patients undergoing surgical resection in our clinic. Normal liver matching was available from seven patients (N146, N175, N198, N227, N253, N254, and N629). Written informed consent was obtained from each patient, and the study protocol was approved by the Committee of Ethics of the Ludwig-Maximilians-University of Munich. We used the HB cell lines HUH6 (Japanese Collection of Research Bioresources, JCRB, Osaka, Japan), HepT1 , HepT3 , and HepG2 , as well as the hepatocellular carcinoma (HCC) cell line HUH7 (JCRB). All cell lines were maintained as the suppliers recommended.
Real-time reverse transcription-PCR (RT-PCR)
The total RNA was extracted from macroscopically dissected frozen tumor tissue (at least 80% tumor cells), frozen normal liver tissue and HB cell lines, depleted from residual DNA, and reverse transcribed as previously described . PCR amplifications were carried out with 40 ng of cDNA, 500 nM forward and reverse primers and iTaq SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA) on a Mastercycler Realplex2 cycler (Eppendorf, Hamburg, Germany) with 40 cycles consisting of a 15 sec denaturation at 95°C, primer annealing for 15 sec at 55°C, and extension for 30 sec at 72°C. We used the following primer pairs (5'- > 3' orientation): IGF2, CCTCCGACCGTGCTTCC, GGTGGACTGCTTCCAGGTGT; PLAG1, ACAAGTGCATACAACAAGACTGCA, CAGGAGAATGAGTAGCCATGTGC; IGFBP3, GTCCAAGCGGGAGACAGAATAT, CCTGGGACTCAGCACATTGA; TIA1, TTAGCCAGATTGGACCTTGTAAAAA, CGATGCTCATGAAACTCCACA; TBP, GCCCGAAACGCCGAATAT, CCGTGGTTCGTGGCTCTCT. Amplification of the house-keeping gene TATA-Box-binding-Protein (TBP) was performed to standardize the amount of sample RNA according to a previous study . PCR efficiencies for all assays were determined, with slopes ranging from -3.34 to -3.69. The relative quantization of gene expression was performed using the ΔΔct method as previously described .
Genomic DNA from frozen tumor and normal liver samples (see above) and cell lines was extracted with phenol and chloroform, precipitated with ethanol and dissolved in TE buffer following standard procedures. Genomic DNA (10 μg) from a healthy person was methylated in vitro using 40 U CpG methyltransferase (SssI), S-adenosylmethionine, and NEBuffer2 (New England Biolabs, Frankfurt, Germany) at 37°C for 4 h, precipitated with ethanol, dissolved in TE buffer, and used as a positive control for methylated alleles. Genomic DNA (2 μg) was bisulfite-treated using the EpiTect® Bisulfite Kit (Qiagen, Hilden, Germany) and amplified using primers not specific for methylation status (IGFBP3-BS-F, GGTGTTGAGTTGGTTAGGAGT; IGFBP3-BS-R, AAACAACACCAACAAAATCAA). We cloned the PCR products into the pCR2.1 TOPO vector (Invitrogen, Karlsruhe, Germany) and sequenced six independent clones per sample (MWG Biotech, Ebersberg, Germany).
In addition, the methylation status of the promoter region of IGFBP3 (from -314 to -147 bp upstream of transcriptional start site) gene was analyzed by methylation-specific-PCR (MSP) using the following primer sets 5'- > 3' orientation): methylated (IGFBP3-M-F, TGATTCGGGTTTCGGGCGTGC; IGFBP3-M-R, GCCGACCGCTATATAAAAACCG) and unmethylated (IGFBP3-U-F, GGTGATTTGGGTTTTGGGTGTGTGTAT; IGFBP3-U-R, AAACACACCAACCACTATATAAAAACCAAA). MSP primer design and PCR conditions were performed according to . For DNA demethylation experiments, we used 0.5 μM 5-aza-2'-deoxycytidine (5-Aza-dC; Sigma-Aldrich, Seelze, Germany) for HUH6 and HepT3 cells and 1.25 μM for HepT1, HepG2 and HUH7 cells; 5-the Aza-dC was applied for 5 days and changed daily. Alternatively, Trichostatin A (Sigma-Aldrich) was applied for 24 h in a concentration of 0.1 μM (HUH6 and HepT3) and 0.25 μM (HepT1, HepG2 and HUH7).
HepT1 cells (5 × 105 cells/6-well plate) were transfected with 1 μg DNA of the pIRES-IGFBP3 expression vector containing full-length IGFBP3 cDNA  or the empty vector control using the FuGene 6 transfection reagent (Roche Diagnostics, Mannheim, Germany). After 24 h of transfection, the cells were changed to media containing 1 μg/ml puromycin (Sigma-Aldrich). After 2 weeks of selection, puromycin-resistant colonies were selected and cultured as stable transfected HepT1 clones. Western blot analysis was performed using rabbit polyclonal anti-IGFBP3 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and rabbit anti-human β-actin (Cell Signaling, Technology, Danvers, MA, USA) antibodies, as previously described .
Cell viability assay
For the proliferation assay, 5 × 103 cells were seeded into 96-well plates, and the viability was assessed at the time points indicated using the Cell Proliferation Kit I (Roche Diagnostics) according to the manufacue's protocol. The optical density was measured at a wavelength of 595 nm after the addition of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) labeling reagent on the GENios microplate reader (Tecan, Männedorf, Switzerland).
Colony formation assay
HepT1 cells (5 × 105 cells/well) were transfected in a 6-well plate format with 1 μg of the pIRES-IGFBP3 expression vector or control vector using the FuGene 6 transfection reagent (Roche Diagnostics). They were subsequently cultured in selection media containing 1 μg/ml puromycin (Sigma-Aldrich) for 2 weeks. Colonies were fixed with 100% methanol, stained with 0.1% crystal violet and counted.
For annexin V-based apoptosis analysis, cells were trypsinized, washed with PBS, and suspended in 500 μl of calcium-containing binding buffer. Cy5-conjugated annexin V (1:100; BioVision, Mountain View, CA, USA) and 5 μM calcein (Invitrogen) were added to the cell suspension. Early apoptotic cells (annexin V and calcein positive) were detected using cell fluorescence assays with an Agilent 2100 Bioanalyzer. The cleavage of poly(ADP-ribose) polymerase was detected as previously described  using antibodies for human poly(ADP-ribose) polymerase and human β-actin (both from Cell Signaling Technology, Danvers, MA, USA).
Cell migration assay
HB cells were seeded into 24-well plates and grown to confluency. A wound of approximately 1 mm was inflicted to cell monolayers with a pipette-tip. The wells were washed twice with PBS to remove detached cells and incubated at 37°C with medium in the presence or absence of 1 μg/ml recombinant human IGFBP3 (R&D systems, Wiesbaden, Germany) for 72 h. Images were taken at 0, 24, and 48 h after scratching, and the wound widths were measured and quantified.
Transwell permeable supports (8 μm pore polycarbonate inserts; Corning Incorporated, Corning, NY, USA) were coated either with collagen or 10% Matrigel® (BD Biosciences, Heidelberg, Germany) in DMEM and subsequently added to 24-wells containing DMEM (negative control) or DMEM/10% FCS/50 ng/ml recombinant human HGF (NatuTec, Frankfurt, Germany) as a chemoattractant. Cells (1 × 105) were seeded in DMEM in the inside compartments and allowed to migrate for 16 h (HUH7) or 72 h (HepG2) in the presence or absence of 1 μg/ml recombinant human IGFBP3 (R&D systems). Afterwards, the inserts were stained with crystal violet solution. Cells from the upper side of the insert were removed by using cotton swabs. Cells attached to the bottom side of the insert were photographed using a Zeiss Axiovert 25 microscope and a Canon 450D camera. For each sample, eight pictures were taken, and the number of cells was calculated by using ImageJ (National Institute of Health, Bethesda, MD, USA) and the Particle Counter plugin. Data for three independent experiments were analyzed using GraphPad Prism Version 3.0 (GraphPad Software, La Jolla, CA, USA).
The data are presented either as dot plots or bar graphs, indicating the mean ± SEM. The statistical analyses and Kaplan Meier calculations were performed with GraphPad Prism Version 3.0 using Student's unpaired t-test, Mantel-Cox test, Mann-Whitney-U test, Spearman's rank correlation, one-way ANOVA and Dunnett's test. P < 0.05 was considered to be significant.