Recombinant mutant forms of human Gal-9 lacking the linker peptides were expressed and purified as previously described (19). Cell Counting Kit-8 (CCK-8) was purchased from Dojindo Laboratories (Kumamoto, Japan), and all other chemicals were obtained from Sigma Chemical (Tokyo, Japan).

Following antibodies were used: anti-β-actin monoclonal antibody (A5441, used at a dilution of 1:3,000; Sigma-Aldrich, St. Louis, MO, USA); cyclin D1 (RB-9041, used at 1:1,000); cyclin E (used at 1:1,000) (both from Thermo Fisher Scientific, Waltham, MA, USA); Cdk6 (sc-177, used at 1:1,000); Cdk4 (sc-749, used at 1:1,000); Cdk2 (sc-163, used at 1:2,000) (all from Santa Cruz Biotechnology, Santa Cruz, CA, USA); phosphorylated retinoblastoma protein (Rb; no. 558385, used at 1:1,000; BD Pharmingen); Rb (#9309, used at 1:1,000; Cell Signaling Technology); and secondary horseradish peroxidase (HRP)-linked anti-mouse and anti-rabbit IgG antibodies (used at 1:2,000; GE Healthcare, UK).

The two human cholangiocarcinoma TFK-1 and HuH-28 cell lines were obtained from the Japanese Cancer Research Resources Bank (Osaka, Japan) and passaged at our laboratory for <6 months. The cell lines were authenticated by the Cell Bank using short tandem repeat PCR. TFK-1 was grown in RPMI-1640 and HuH-28 was grown in MEM, with both media types supplemented with 10% fetal bovine serum (FBS) and 100 mg/l of penicillin-streptomycin in a humidified atmosphere with 5% of CO₂ at 37°C.

Cell proliferation assays were conducted in TFK-1 and HuH-28 cells. Each cell line (5×10³) was seeded into 96-well plates and cultured in 100 μl of culture medium supplemented with 10% FBS. After 24 h, the cells were treated with 0.03, 0.1 or 0.3 μM Gal-9 and then cultured for an additional 48 h. CCK-8 reagent (10 μl) was added to each well, and the plates were incubated at 37°C for 3 h. The absorbance of each well was measured at 450 nm using an auto-microplate reader.

M30 Apoptosense enzyme-linked immunosorbent assay (ELISA) kits obtained from PEVIVA AB (Bromma, Sweden) was used to evaluate caspase-cleaved keratin 18 (CCK18) levels (20). Cell line cells (5×10³ cells) were seeded into 96-well plates and cultured in 100 μl of culture medium for 24 h. The cells were then treated with 0.3 μM Gal-9, and the assays were performed according to the manufacturer's instructions. The amount of antigen in controls and samples were calculated by interpolation to a standard curve.

To evaluate the mechanism of growth inhibition by Gal-9, cell cycle profiles were analyzed after treatment with Gal-9. TFK-1 cells (1.0×10⁶ cells in a 100-mm dish) were treated with or without 0.3 μM Gal-9 for 24 h. The cell cycle distribution was analyzed by measuring the amount of propidium iodide (PI)-labeled DNA in ethanol-fixed cells. Fixed cells were washed with phosphate-buffered saline (PBS) and stored at −20°C until analysis by flow cytometry. On the day of analysis, the cells were washed with cold PBS, suspended in 100 μl of PBS plus 10 μl of RNase A (250 μg/ml) and incubated for 30 min. To each suspension was added 110 μl of PI stain (100 μg/ml), and the cells were incubated at 4°C for at least 30 min prior to analysis. Flow cytometry was performed using a Cytomics FC500 flow cytometer (Beckman Coulter) with an argon laser (488 nm). The percentages of cells in different phases of the cell cycle were analyzed by FlowJo software (TreeStar, Ashland, OR, USA). All experiments were performed in triplicate to assess consistency of response.

TFK-1 cells (1.0×10⁶/dish) were seeded in 100-mm culture dishes and cultured for 24 h; Gal-9 was added, and the cells were further cultured for 48 h. The cells were washed twice in PBS and lysed using a protease inhibitor cocktail (ProPrep, 'Complete' protease inhibitor mixture; iNtRON Biotechnology, Sungnam, Korea). Protein concentration was quantified using the NanoDrop 2000 fluorospectrometer (Thermo Scientific Corp., USA). Whole-cell lysates (1–10 μg) were resolved by SDS-polyacrylamide gel electrophoresis on 10% Tris-glycine gradient gels by SDS-PAGE (21), and the proteins were transferred to nitrocellulose membranes. After blocking, the membranes were incubated with primary antibodies and then incubated with HRP-conjugated secondary antibodies (22). Immunoreactive proteins were visualized with an enhanced chemiluminescence detection system (Perkin-Elmer Co., Waltham, MA, USA) on X-ray film.

Animal experiments were performed according to the guidelines of the Committee on Experimental Animals of Kagawa University, Kagawa, Japan.

Eighteen female athymic mice (BALB/c-nu/nu; 6-weeks old; 20–25 g) were purchased from Japan SLC Inc. (Shizuoka, Japan). The animals were maintained under specific pathogen-free conditions using a laminar airflow rack. The mice had continuous free access to sterilized (γ-irradiated) food (CL-2; CLEA Japan Inc.) and autoclaved water.

Each mouse was subcutaneously inoculated with TFK-1 cells (5×10⁶ cells/animal) in the flank. Two weeks later, the xenografts were identifiable as masses of maximal diameter >6 mm. We randomly assigned the animals to 3 groups. These groups were treated with 30 μg of Gal-9 (n=6), 90 μg of Gal-9 (n=6) or control (PBS only; n=6).

The Gal-9-treated groups were injected subcutaneously with Gal-9 (30 or 90 μg) 3 times a week. The control group was administered PBS alone. Tumor growth was monitored daily by the same investigators (K. Kobayashi and T. Masaki), and tumor size was measured weekly by measuring the 2 largest perpendicular dimensions. Tumor volume was calculated as follows: tumor volume (mm³) = [tumor length (mm) x tumor width (mm)²]/2 (23). All animals were sacrificed 7 weeks after treatment, with all animals remaining alive during this period.

The Human apoptosis antibody array kit (R&D Systems, Minneapolis, MN, USA) was used according to the manufacturer's instructions.

Human p-RTKs were assayed using Human p-RTK array kits (R&D Systems) according to the manufacturer's instructions. Briefly, p-RTK array membranes were blocked with 5% BSA/TBS (0.01 M Tris-HCl, pH 7.6) for 1 h and incubated with 2 ml of lysate prepared from cell lines after normalization so that the amounts of protein were equal. After 3 washes for 10 min each with TBS plus 0.1% v/v Tween-20 and 2 washes for 10 min with TBS alone to remove unbound materials, the membranes were incubated with anti-phospho-tyrosine-HRP antibody for 2 h at room temperature. The unbound HRP antibody was washed out with TBS plus 0.1% Tween-20. Finally, each array membrane was exposed to X-ray film using a chemiluminescence detection system (Perkin-Elmer Co.).

The RayBio Human Angiogenesis Antibody Array (RayBiotech Inc.) was used according to the manufacturer's protocol. This method is a dot-based assay enabling detection and comparison of 20 angiogenesis-specific cytokines. Each array membrane was exposed to X-ray film using a chemiluminescence detection system.

Total RNA was extracted from tumor samples using RNeasy Mini kits (Qiagen) according to the manufacturer's instructions. RNA samples typically showed A_260/280 ratios between 1.9 and 2.1, using an Agilent 2100 Bioanalyzer (Agilent Technologies).

After RNA measurement with an RNA 6000 Nano kit (Agilent Technologies), the samples were labeled using a miRCURYHy3/Hy5 Power Labeling kit and were hybridized to a human miRNA Oligo chip (v.19.0; Toray Industries). The chips were scanned with a 3D-Gene Scanner 3000, and the results were analyzed by 3D-Gene extraction version 1.2 software (both from Toray Industries). Differences in miRNA expression between Gal-9-treated and control samples were assessed by analyzing the raw data using GeneSpring GX v10.0 (Agilent Technologies). The samples were first normalized relative to 28S RNA and baseline corrected to the median of all samples.

Replicate data were consolidated into 2 groups (those from Gal-9-treated cells and those from control cells) and were organized using the hierarchical clustering and ANOVA functions in GeneSpring software. Hierarchical clustering was performed using the clustering function (condition tree) and Euclidean correlation as a distance metric. Two-way ANOVA analysis and asymptotic P-value computation without any error correction on the samples were performed to determine the miRNAs varying most prominently across the groups. The P-value cut-off was set to 0.05. Only changes >50% for at least one of the time points for each sample were considered significant. All analyzed data were scaled by global normalization. The significance of differentially expressed miRNAs was analyzed by the Student's t-test.

All statistical analyses were performed using computer-assisted JMP 9.0 (SAS Institute, Cary, NC, USA). Paired analysis between groups used t-tests. A P-value <0.05 was considered significant.

To evaluate the effect of the growth activity of Gal-9 on human cholangiocarcinoma cells in vitro, we examined the effect of Gal-9 on proliferation in two cholangiocarcinoma cell lines, TFK-1 and HuH-28. The cells were grown in 10% FBS and treated with 0, 0.03, 0.1 or 0.3 μM Gal-9 for 48 h. Gal-9 led to a dose-dependent and strong inhibition of cell proliferation in the two cholangiocarcinoma cell lines (Fig. 1).

To clarify the mechanism of the growth inhibitory effect of Gal-9, we first examined the induction of apoptosis by Gal-9.

CCK18 expression was determined by ELISA to confirm whether apoptosis is involved in Gal-9-induced cell death. Gal-9 increased the levels of CCK18 in the two CCA cell lines (Fig. 2).

It became intriguing to clarify whether effects of Gal-9 on cell cycle are simultaneously involved in the suppression of cholangiocarcinoma cells. The effects of Gal-9 on the expression of various cell cycle-related molecules in TFK-1 cells were evaluated by western blotting. Cells were treated 0 or with 0.3 μM Gal-9 for 24–48 h. Assays of the expression of other proteins associated with the G₀ to G₁ transition showed that the Cdk6, catalytic subunit of cyclin D1, Cdk2 and Cdk4 were not changed at 24 and 48 h after the addition of 0.3 μM Gal-9 (Fig. 3A).

Next, we performed flow cytometric analysis of the cell cycle to evaluate the contribution of cell cycle arrest to the suppression of CCA cell lines by Gal-9. TFK-1 cells treated with 0.3 μM Gal-9 showed no change in the cell cycle profile (Fig. 3B). These results suggested that Gal-9 suppresses cholangiocarcinoma cell growth through the tumor cell apoptosis but not through cell cycle arrest.

Next experiments were performed to clarify whether Gal-9 also exhibits tumor growth suppressive activity in vivo. Nude mice were injected subcutaneously with TFK-1 cells, followed by the subcutaneous injection of Gal-9. Gal-9 significantly suppressed the tumor growth of TFK-1 cells compared with untreated control mice (Fig. 4A and B). Throughout the in vivo study, Gal-9 had no apparent effects on these mice and did not affect their weights. All animals survived throughout the experiment.

We used an apoptosis array system to identify the apoptosis-associated proteins associated with the antitumor effect of Gal-9. Using an antibody array enabled the expression of 35 apoptosis-associated proteins to be screened in TFK-1 cells in the presence and absence of Gal-9 (Fig. 5A). Gal-9 increased the levels of expression of cytochrome c (Fig. 5B). Densitometry analysis showed that the ratio of cytochrome c spots of Gal-9-treated to untreated cells was 249.9% (Fig. 5C). Expression levels of the apoptosis-associated proteins other than cytochrome c were not affected by Gal-9.

We used a p-RTK array system to identify the key RTKs associated with the antitumor effect of Gal-9. Using an antibody array enabled the expression of 49 activated RTKs to be screened in TFK-1 cells in the presence and absence of Gal-9 (Fig. 6A). Gal-9 reduced the levels of expression of phosphorylated epidermal growth factor receptor (p-EGFR) and phosphorylated insulin-like growth factor-1 receptor (p-IGF-1R), as well as reduced the expression of hepatocyte growth factor receptor (HGFR) and fibroblast growth factor receptor 3 (FGFR3; Fig. 6B). Densitometry analysis showed that the ratio of p-EGFR, p-HGFR, p-IGF-1R and p-FGF R3 spots of Gal-9-treated to untreated cells was 40.1, 38.0, 16.1 and 18.3%, respectively (Fig. 6C).

We used an angiogenesis array system to identify the key angiogenesis-related molecules associated with the antitumor activity of Gal-9 on TFK-1 cells (Fig. 7A). The 20 angiogenesis molecules screened showed no change by treatment with Gal-9 (Fig. 7B).

Using a custom microarray platform, we analyzed the level of 2555 miRNA probe in tumorous tissues in vivo that were treated with and without Gal-9.

In a tumor xenograft model, in the Gal-9 group, there were 22 upregulated and 27 downregulated miRNAs of the 985 miRNAs (GEO, accession no. GSE30289; Table I).

Unsupervised hierarchical clustering analysis using Pearson's correlation showed that tumorous tissues in vivo treated with Gal-9 clustered together, and separately from untreated cell lines (Fig. 8). We found that miR-198 was upregulated in tumor tissues treated with Gal-9.