Cell culture and reagents
The human cholangiocarcinoma cell lines KKU-213 and KKU-100 were established from CCA patients of Srinagarind Hospital, Khon Kaen University and purchased from the Japanese Collection of Research Bioresources (JCRB) Cell Bank, Osaka, Japan. KKU-213 is a mixed (papillary and non-papillary) cholangiocarcinoma which was established from a 58-year-old male patient, whereas KKU-100 is a poorly differentiated cholangiocarcinoma established from a 65-year-old female patient. All cell lines were grown in DMEM medium (Gibco Life Technology, Carlsbad, CA, USA), supplemented with NaHCO3, 100 units/ml penicillin, 100 mg/ml streptomycin and 10% fetal bovine serum at 37°C containing 5% CO2 in a humidified incubator. Three reagents were used: SSZ, an inhibitor of xCT (Sigma-Aldrich, MO, USA), cisplatin purchased from Boryung Pharmaceutical Co., Ltd. (Gyeonggi-do, South Korea), and gemcitabine purchased from the Eli Lilly Corporation (Indianapolis, IN, USA).
Liver tissues were fixed in 10% buffered formaldehyde, embedded in paraffin blocks and then sectioned at a thickness of 4 μm. Sections were deparaffinized in xylene and rehydrated in an ethanol series. Immunohistochemical staining was performed for CD44 variants 9 (CD44v9), and Ki-67 (proliferative marker) according to standard methods as previously described [32, 33]. Anti-CD44v9 was purchased from Cosmo Bio (1:50 dilution, Cosmo Bio, Tokyo, Japan) and Ki-67 was purchased from Abcam (1:300 dilution, Abcam, Cambridge, United Kingdom). The sections were observed under a light microscope at ×200 and ×400 magnifications (Axioscope A1, Carl Zeiss, Jena, Germany). The scoring system of IHC was performed as previously described .
The tissue sections were processed for immunohistochemical staining and retrieved by heating in 0.01 M sodium citrate containing 0.05% Tween 20 (pH 6.0) for 10 min at 110°C. The samples were then exposed to 3% bovine serum albumin before being incubated at 4°C overnight with primary antibodies: CD44v9 (1:50 dilution, Cosmo Bio, Tokyo, Japan) and CK-19 (1:300 dilution, Abcam, Cambridge, United Kingdom). After washing with PBS, the samples were incubated with Alexa Fluor 488 or 555-conjugated secondary antibodies (Invitrogen, Waltham, MA, USA) and mounted in Hoechst 33342 (Invitrogen, Waltham, MA, USA). The fluorescence signals were detected under a Zeiss LSM800 confocal laser scanning microscope (Carl-Zeiss, Oberkochen, Germany).
The tumor tissue sections were deparaffinized and rehydrated using xylene, and 100%, 90%, 80%, 70% alcohol twice each for 5 min, then incubated with proteinase K (20 ug/ml in 10 mM Tris-HCl pH 7.4) for 30 min in a 37 C° incubator. The samples were then exposed to hydrogen peroxide before being incubated at 37°C for 60 min with TUNEL reaction mix (Enzyme solution 50 uL+ nucleotide mixture solution 450 uL). After washing with PBS, the samples were incubated with convertor-POD at 37°C for 30 min. The signals were detected using the DAB system and observed under a light microscope (Axioscope A1, Carl Zeiss, Jena, Germany).
Flow cytometry analysis for the apoptosis assay
KKU-213 cells were seeded at 5x105 cells/well. After the cells had adhered, they were treated with the drug for 48 h. Both dead and living cells were collected then washed with 1xPBS. The cells were incubated with annexin V and propidium iodide (Sigma-Aldrich, MO, USA) for 15 min. The signal of positive cells was detected with Flow cytometry analysis using FACSCanto II (BD Bioscience, San Jose, CA, USA).
Cell proliferation and cell cytotoxicity
The number of viable cells was evaluated with a Cell Titer-Glo luminescence cell viability kit (Promega, Madison, WI, USA). Briefly, CCA cells (2x103 cells per well) were plated into 96-black well plates for 24 h. Cells were then treated with SSZ (0, 200, 400, 600, 800, 1000 μM) and CIS (0, 10, 20, 40, 60, 80, 100 μM) for 48 h, and GEM (0, 10, 20, 40, 60, 80, 100 μM) for 72 h. In addition, 300 μM of SSZ was used in combination with CIS or GEM. The luminescence signal was detected on a SpectraMaxL microplate reader. The experiments were done in triplicate.
The Animal Ethics Committee of Khon Kaen University (AEKKU 6/2560) approved the study protocol. Female nude mice (4 to 6-week-old purchased from Nomura Siam International CO., Ltd., Japan) were used to establish subcutaneous xenograft mice. They were injected subcutaneously with 2 ´ 106 KKU-213 CCA cells on both flank sides. This cell line previously demonstrated highly expressed CD44v9 protein. A day after tumor injection, the mice were orally administered either a vehicle or sulfasalazine (250 mg/kg body weight) daily for 15 days. In the case of cisplatin treatment, the mice were injected intravascularly with cisplatin (2 mg/kg body weight) twice a week. The body weight of each animal was observed and tumor masses were removed and weighed 22 days after inoculation.
1H NMR analysis of tissue extraction
To extract tumor tissues for detection of metabolomics, we weighted 100 mg the tumor tissue and washed this with 1xPBS pH 7.4. The samples were then extracted with methanol and chloroform, the supernatant was separated into a polar phase and a lipophilic phase after centrifugation at 1000 g for 15 min. The solvents were removed using a speed vacuum concentrator (Labconco, Kansas City, MO, USA). The tissues were re-suspended with 560 ml of 100 mM sodium phosphate buffer, pH 7.4 in D2O containing 0.1 mM 3-trimethysilypropionic acid (TSP) (Cambridge Isotype Laboratories, Tewksbury, MA, USA) as a chemical shift reference (d=0 p.p.m.) and optionally 0.2% NaN3. Next, the extracted tissue samples were transferred to a NMR tube and after being vortex centrifuged at 12,000 g for 5 min. Proton NMR spectra were acquired using a 400 MHz NMR spectrometer (Bruker, Ettlingen, Germany). All samples were detected using a standard 1-dimensional pulse sequence (recycle delay-90°-t1-90°-tm-°-acquisition) with t1 to 3 ms, tm to 10 ms, abd 90° pulse to 10 µs in 32 scans.
MATLAB (R2015a) was used for multivariate analysis, both unsupervised and supervised multivariate statistical methods, and in-house developed scripts were employed. The unsupervised analysis involved principal component analysis (PCA) which generates a model showing the intrinsic similarities or differences without prior class information and reduces the complexity and intricacy of the data. Orthogonal projection to latent structures-discriminant analysis (OPLS-DA) was the supervised multivariate statistical method. This analyses the statistical models of class membership data which is then used to optimize the separation between the different classes. The fitness and predictability of the OPLS-DA models were determined by R2 and Q2 values. The assignments of discriminatory metabolites were confirmed using ChenomxNMR Suite software analysis and statistical total correlation spectroscopy (STOCSY) on 1-dimensional NMR spectra. SPSS software version 17.0 (IBM Corporation, Armonk, NY, USA) was used for statistical analysis. The differences among each group of samples were analyzed using a t-test. The data were expressed as a graph of mean ± S.D. using Graph Pad prism 5. All analyses were two-tailed and P-values < 0.05 were considered statistically significant.