A total of 22 tumor samples and 10 normal adjacent tissues were collected from intrahepatic cholangiocarcinoma (ICC) patients who were admitted to Srinagarind Hospital, Faculty of Medicine, Khon Kaen University, Thailand. The project was approved by the Khon Kaen University Ethics Committee in human research (HE571044). All patients gave written informed consent. Fresh tumor tissues were tested for Histoculture Drug Response Assay. Paralleled tissues were kept under liquid nitrogen until used for protein extraction.

Tumor tissues were classified as 5-FU sensitive and 5-FU resistant based on results obtained from an ex vivo histoculture drug response assay (HDRA), using the median of inhibition index (% II) as previously described (28). In brief, fresh tumor tissues were washed and a 3-mm diameter punch was used aseptically to take samples that were placed on collagen gel sponges. Each was cultured at 37ºC in a 6-well plate containing RPMI medium supplemented with 2.5% v/v penicillin-streptomycin-fungizone (PSF; Invitrogen, Carlsbad, CA, USA) and 5-FU at 200 μM. For a control sample, no 5-FU was added in the culture medium. After 4 days of culture, the viability of tumor cells in the cultured tissues was examined using TUNEL staining. TUNEL-positive cells were identified as dead cells. The efficacy of 5-FU was calculated and expressed as the % II using the following formula: % II = (1 − % viable tumor cells in 5-FU-treated tumor tissue/% living tumor cells in control tissue) × 100.

Two CCA cell lines, KKU-M139 and KKU-M214 were established from primary tumors of human intrahepatic CCA at the Liver Fluke and Cholangiocarcinoma Research Center, Khon Kaen University Thailand (16,29,30). Both were cultured at 37ºC in RPMI medium (Gibco BRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS), 1% v/v penicillin-streptomycin solution (Gibco, BRL) under 5% CO₂ atmosphere. These CCA cells were named as KKU-M139P and KKU-M214P, to identify the parental cell lines at the beginning of drug resistance induction.

5-FU-resistant cell lines were generated from the parental cell lines KKU-M139P and KKU-M214P by stepwise increases of the concentration of 5-FU (Boryung Pharm, Korea) as described previously (16). In brief, 1×10⁵ cells were seeded in 25-cm² flasks and cultured without 5-FU for 24 h. Subsequently, cells were exposed with 5-FU at 6 μM for KKU-M139 (1X IC₅₀ value of KKU-M139P) and 4 μM for KKU-M214 (1X IC₅₀ value of KKU-M214P) for 72 h. Cells were then cultured in a drug-free medium until they reached 70% confluence. These cells were continuously maintained in 1X IC₅₀ for several passages until these cells were stable before being subjected into 2X IC₅₀. The 5-FU-resistant clones were finally obtained after continuous selection by several passages for 18 months. The IC₅₀ of the resistant clones was checked by the SRB assay (31). The resistant clones were then passaged into a drug-free medium for 2 weeks before being stored as a stock of the resistant cell lines (KKU-M139R and KKU-M214R) at −80ºC. KKU-M139R and KKU-M214R were cultured in 5-FU-free medium for ≥2 weeks to eliminate potential long-term effects of 5-FU unrelated to drug resistance prior to being performed in all experiments.

Expression of Δ133p53 in KKU-M139R and KKU-M214R cells was suppressed using a siRNA technique. The sequences of two specific siRNA targeting human Δ133p53 (Δ133p53a and Δ133p53b) as described previously (18), were purchased from Ambion (Austin, TX, USA). The cells (2×10⁶ cells/well) were seeded in a 6-well plate and cultured overnight before being transfected separately with 100 pM of siΔ133p53a and siΔ133p53b, while siGFP (Green fluorescence protein, Applied Biosystems/Ambion, Carlsbad, CA, USA) was used as a siRNA control. Transfection was carried out using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. After 24 h of transfection, culture medium was added and the plates were incubated at 37ºC for a further 48 h. At 72 h-post transfection, total proteins were extracted using TRIzol reagent (Invitrogen). The level of Δ133p53 protein was determined by western blot analysis using β-actin as a loading control.

The parental CCA cells (KKU-M139P and KKU-M214P) and the 5-FU-resistant cells (KKU-M139R and KKU-M214R) were seeded at 1×10⁴ cells/well in triplicate into a 96-well culture plate and incubated at 37ºC for 24 h. All cell lines were then treated with various concentrations of 5-FU ranging from 2–128 μM in triplicate for 72 h, while 0.9% saline was used as a negative control. The cytotoxicity was performed using a sulforhodamine B (SRB) assay as previously described (31). The cells were fixed using a 10% cold trichloroacetic acid (TCA), washed and air-dried at room temperature. SRB (Sigma-Aldrich, MO, USA) solution (100 μl/well) was added and followed by three quick rinses with 1% acetic acid to remove unbound dye. SRB was solubilized in a 10 mM Tris base solution and the absorbance at 490 nm was measured using a microplate reader (Tecan Ltd., Reading, UK). Percentage of cell viability was calculated [(mean OD_sample − mean OD_day0/mean OD_{negative control} − mean OD_day0) × 100] and used to generate the curve by which IC₅₀ was calculated. Resistance index was defined as a ratio of the IC₅₀ value of drug resistant cells to parental cells.

To assess cell growth, the population doubling time (PDT) of the cells was assessed in triplicate. Cells (2×10⁵) were cultured in a drug free medium supplemented with 10% FBS at 37ºC in a humidified 5% CO₂ atmosphere. When reaching 70% confluence, the cells were trypsinized, stained with trypan blue and counted on a hemocytometer. PDT was calculated using the following formula as described previously (32): (T-T₀) log2/logN-logN₀, where N₀ is the initial cell number, N is the final cell number, T is the time interval between N₀ and N, and T₀ is the initial time.

KKU-M139R and KKU-M214R cells were pre-treated with either siΔ133p53 or siRNA control. The cells were seeded at 200 cells/well in a 6-well plate containing 2 ml RPMI culture medium supplemented with 10% v/v FBS and cultured at 37ºC in a humidified 5% CO₂ atmosphere for 5 days. The cells were washed twice with PBS, stained with H&E and the colonies were counted.

The analysis of Annexin V binding was carried out with the Annexin V-FITC Detection Kit I (eBioscience, San Diego, CA, USA) according to the manufacturer's instructions. Briefly, cells were incubated with or without siΔ133p53 or a control scramble RNA for 48 h. Cells were collected, washed twice with cold PBS, centrifuged at 1,800 rpm for 3 min, and resuspended in 1X binding buffer at a concentration of 10⁶ cells/ml. Then 100 μl of the solution (10⁵ cells) was transferred to a 5-ml culture tube; 5 μl of Annexin V-FITC and 5 μl of PI were added. Cells were incubated for 15 min at room temperature in the dark. Furthermore, 200 μl of 1X binding buffer were added to each tube, and samples were analyzed by FACScan flow cytometry (BD FACSCanto II; BD, USA). For each sample, 10,000 ungated events were acquired. Annexin V⁺/PI⁻ cells represented the early apoptotic populations, and Annexin V⁺/PI⁺ cells the late apoptotic populations.

KKU-M139R and KKU-M214R cells (10⁵ cells/ml) were incubated with or without siΔ133p53 or a siRNA control for 48 h. The cells were collected, washed with cold PBS, fixed in cold 100% ethanol, treated with DNase-free RNase, and stained with 40 μg/ml of propidium iodide (PI). The distribution of the cells between phases of the cell cycle was deduced from the DNA content on a FACScan flow cytometer (BD FACSCanto II; BD, USA). For each sample, 10,000 gated events were acquired.

Protein was extracted from CCA tissues and cell lines using TRIzol (Invitrogen) and 40 μg aliquots were fractionated on 15% polyacrylamide gel electrophoresis. Primary antibodies; CM-1 (1:100, Signet, Emeryville, CA, USA), p21 (1:400), p27 (1:400), Bcl-2 (1:200), Bax (1:200), and p73 (ab-4) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) as well as a loading control, β-actin (1:4,000, Sigma Chemical Co.) were used followed by the secondary antibody peroxidase-labeled anti-rabbit (1:10,000, Abcam, UK) The proteins were detected by chemiluminescence using the ECL Plus system (GE Healthcare, UK). Protein band intensity was calculated by Scion image program, and normalized to β-actin.

The paraffin-embedded sections (5 μm) of CCA cell pellets fixed with 10% formalin solution were deparaffinized. Antigen retrieval was performed in boiling 0.01 M citrate buffer (pH 6.0) as described previously (26). Endogenous peroxidase was inactivated with 100 μl of 3% H₂O₂. Non-specific binding was further treated with a blocking buffer containing phosphate-buffered saline with Tween-20 (PBS-T), 30% casein and 5% FBS. For p53 protein detection, mutant p53 was detected with the primary antibody clone DO-7, 1:100 (Dako, Glostrup, Denmark), which recognizes an epitope between amino acids 1–45 of human p53. The slides were incubated overnight at room temperature with primary antibody. Proteins were detected using the EnVision system (Dako) for 1 h at room temperature. Color was developed with DAB solution (Dako) and nuclei were counterstained with hematoxylin. Positive staining was observed as brown color in blue/gray nuclei. Positive with DO-7 antibody indicates overexpression of the mutated p53 due to its stability (26,33–35).

Statistical analyses were performed using SPSS for Windows version 15 (SPSS, Inc., IL, USA). The Mann-Whitney U test was used for comparison of two groups. Data are expressed as mean ± SD from three independent experiments. Statistically significant differences are indicated in the figures as ^*P<0.05, ^**P<0.01, ^***P<0.001.

5-FU sensitivity in 22 CCA patients was classified based on the results obtained from histoculture drug response assays (HDRA) and treatment outcome data. For the HDRA-based classification, we used the median of the % inhibition index (% II) which was 36.5% and this identified 11 patients as 5-FU-sensitive and 11 as 5-FU-resistant. Only 12 of 22 CCA patients underwent complete course of chemotherapy. The treatment outcome was further followed up every 6 months for ≥12 months. Poor response (n=7) and favorable response (n=5) were defined in terms of having tumor progression before and after 6 months after treatment, respectively. All clinical data are summarized in Table I.

Western blot analysis of 22 tumors and 10 normal paired tissues using CM-1 antibody revealed the presence of various p53 isoforms (Fig. 1). At least 3 isoforms; TAp53, Δ40p53 and Δ133p53 were observed in tumor tissues, in which two different sizes of Δ133p53 at 35 kDa corresponding to Δ133p53α and 28 kDa to Δ133p53β, γ were defined (Fig. 1A). Interestingly, no Δ133p53 isoform was observed in the normal tissues (Fig. 1A). The box-plot analysis of Δ133p53 protein expression normalized to β-actin and 5-FU sensitivity classified by HDRA showed significantly increased Δ133p53 in 5-FU-resistant cases compared to sensitive ones (P=0.019) (Fig. 1B). It seemed that Δ133p53 protein was highly expressed in CCA patients with poor outcome compared to CCA cases with favorable outcome but was not statistically significant (P=0.264) (Fig. 1C).

Two resistant CCA cell lines; KKU-M139R and KKU-M214R were successfully induced from the parental cells; KKU-M139P and KKU-M214P. Drug toxicity is presented in Fig. 2. The IC₅₀ values of 5-FU in KKU-M139P, KKU-M139R, KKU-M214P and KKU-M214R cells were 6.2±2.09, 38.8 ±7.11, 3.9±2.04 and 39.5±3.45 μM, respectively. The resistance index calculated from the ratio of IC₅₀ of resistant to parental CCA cell lines was 6.26 and 10.12 for KKU-M139 and KKU-M214, respectively. Furthermore, population doubling times of both resistant cell lines were shorter than those of parental cell lines as summarized in Table II.

The expression of p53 isoforms was assessed in both parental and resistant CCA cell lines in response to 5-FU concentration covering their IC₅₀ values. When challenged with 5-FU with 5 and 10 μM, TAp53, Δ40p53 and Δ133p53 protein isoforms were markedly increased in both parental cell lines (KKU-M139P and KKU-M214P) in a dose-dependent manner (Fig. 3). This finding indicates the increased usage of both P1 and P2 promoters under 5-FU stress. In contrast to parental cells, only Δ133p53 protein was upregulated in both resistant cell lines (KKU-M139R and KKU-M214R) induced by 5-FU (Fig. 3) suggesting the enhancement of P2 promoter usage. Upon 5-FU challenge with 20 and 50 μM, KKU-M214R showed rapid response to 5-FU in which the upregulation of Δ133p53 was detected at a lower dose (20 μM of 5-FU) compared to KKU-M139R (50 μM of 5-FU).

Silencing of Δ133p53 in both KKU-M214R and KKU-M139R was used to investigate the role of Δ133p53 in 5-FU-resistant CCA cells. The expression of Δ133p53 was successfully suppressed with both siΔ133p53a and siΔ133p53b compared to the siRNA control (Fig. 4A and B) with normal apparent morphology (Fig. 4C). The effect of silenced Δ133p53 on apoptotic and cell cycle markers in KKU-M139R and KKU-M214R was investigated. The suppression of Δ133p53 protein resulted in significant upregulation of Bax expression in both KKU-M214R and KKU-M139R (P<0.01) as well as downregulation of Bcl-2 in KKU-M214R (P<0.001) and KKU-M139R (P<0.01) (Fig. 4D). Accordingly, Annexin V/PI staining showed significantly increased cell apoptosis in the silenced siΔ133p53 of both KKU-M139R and KKU-M214R (siΔ133p53a at P<0.01 and siΔ133p53b at P<0.001, respectively) compared with control (Fig. 5).

Additionally p21 and p27 proteins were significantly upregulated with P<0.01 and P<0.001 in M139R and KKU-M214R compared to the siRNA control (Fig. 4A and 4D). Moreover, cells were significantly arrested at G2 in both siΔ133p53a and siΔ133p53b treated KKU-M139R and KKU-M214R compared to the siRNA controls (Fig. 6). Interestingly, significant upregulation of p73 was observed in both types of silenced CCA cells (Fig. 4A and D). Moreover, no mutated p53 was revealed in siΔ133p53 treated (KKU-M139R and KKU-M214R) or siRNA control compared to their parental cells (KKU-M139P and KKU-M214P) using immunostaining with DO-7 (Fig. 7). These results imply that the induced p73 protein might help mediating apoptosis upon the absence of wild-type TAp53 (Fig. 3A) and mutated p53 (Fig. 7). Moreover, the induced cell cycle arrest by Δ133p53 silencing leads to growth retardation, as shown by the prolonged PDT (Fig. 8A) and the inhibition of colony forming capability (Fig. 8B). Hence, suppression of Δ133p53 expression affects certain tumor characteristics of these CCA resistant cells.

The chemosensitivity of KKU-M139R and KKU-M214R cells transfected with siΔ133p53a and siΔ133p53b was re-assessed using an SRB assay of cell numbers (Table III). Strikingly, the IC₅₀ of KKU-M139R-siΔ133p53a and KKU-M139R-Δ133p53b cells to 5-FU was decreased 12- and 45-fold, as compared with the IC₅₀ of the siRNA control cells. Similar effects were also observed in KKU-M214R with 11- and 20-fold decreases, respectively. The IC₅₀ of silenced CCA cells was lower than those of parental cells suggesting that silencing of Δ133p53 can re-sensitize 5-FU resistance in CCA cell lines.