LY294002, epidermal growth factor (EGF) and 5-FU were purchased from Sigma (St. Louis, MO, USA).

Human GC cell lines SGC7901 and MKN45 were provided by the Shanghai Institute of Cell Biology, CAS (Shanghai, China). Cells were maintained in RPMI-1640 culture medium supplemented with 100 g/ml streptomycin, 100 U/ml penicillin and 10% fetal bovine serum (both from HyClone, USA) at 37°C in a humidified atmosphere containing 5% carbon dioxide.

The cells were subcultured every 2–3 days. The third to fifth subcultures were harvested, and CD133⁺ GC cells were isolated utilizing a CD133 immunomagnetic cell sorting kit (Miltenyi Biotec, Bergisch Gladbach, Germany). The CD133⁺ cells were cultured in serum-free 1640 medium at 37°C in a humidified atmosphere containing 5% carbon dioxide (23,24).

CD133-specific siRNA fragments were designed and synthesized from the CD133 gene sequence (Shanghai GenePharma, Shanghai, China); sense strand, 5′-GUCCUUCCUAUAGAACAAUTT-3′ and antisense strand, 5′-AUUGUUCUAUAGGAAGGACTT-3′. A non-specific siRNA sequence was synthesized as a negative control; sense strand, 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense strand, 5′-ACGUGACACGUUCGGAGAATT-3′. Unsorted SGC7901 cell concentration was adjusted to 1.5×10⁵ cells/ml and spread on three 6-well plates (2 ml/well) (control, negative control group; and CD133, siRNA group), and cultured overnight. Transfection solution A was prepared as follows: the siRNAs were dissolved in deionized water at 20 μmol/l and mixed with RPMI-1640 at a ratio of 10:250 μl/well. Transfection solution B was prepared by mixing Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) and RPMI-1640 at a ratio of 5:250 μl/well. After 5 min, the two transfection mixtures were mixed together and allowed to stand for 20 min. This transfection solution of CD133 siRNA was added to the corresponding wells. RPMI-1640 (500 μl/well) was added to the uninterfered group as a control. After 24-h of transfection, the solution was exchanged with serum-containing RPMI-1640.

Cells were cultured up to a 60–80% confluence state. The CD133 complementary deoxyribonucleic acid (cDNA) plasmid was extracted with plasmid extraction pail (Qiagen, Düsseldorf, Germany) and transfected using Lipofectamine^® LTX reagent (Invitrogen, Tokyo, Japan) in accordance with the manufacturer’s protocol. Following transfection, cells were cultured for 72 h and intermediate samples were collected at 24 and 48 h for further analysis.

Quantified protein lysates were resolved on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred onto polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA), and immunoblotted with mouse anti-human CD133/1 (1:100; Miltenyi Biotec), P-gp (1:500; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), rabbit anti-human P-Akt (Ser473, 1:1,000), Akt (1:1,000), Bcl-2 (1:1,000), Bax (1:1,000) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1:2,000) (Cell Signaling Technology Inc., Boston, MA, USA). After immunoblotting, lysates were incubated with horseradish peroxidase-labeled goat or mouse anti-rabbit immunoglobulin G secondary antibody (1:2,000; Jackson, Mukilteo, WA, USA) at room temperature. Blots were visualized using enhanced chemiluminescence (Amersham Biosciences Inc., Piscataway, NJ, USA).

Total ribonucleic acid (RNA) was extracted using TRIzol reagent (Invitrogen). The extracted RNA (500 ng) was reverse-transcribed into cDNA using a commercial kit (Takara Bio Inc., Otsu, Shiga, Japan) under the following conditions: 42°C for 30 min, 99°C for 5 min and 5°C for 5 min. The cDNA was used as a template for PCR amplification with forward and reverse primers (Shanghai Sangon, Shanghai, China), as shown in Table I. The primer annealing temperatures were: CD133 at 57°C, multidrug resistance protein 1 at 72°C, Bcl-2 at 72°C and Bax at 72°C. GAPDH was used as an internal control for the PCR reaction (Shanghai Sangon) and the annealing temperature was 55°C. The final products of RT-PCR amplification were checked by agarose gel electrophoresis (Bio-Rad, Hercules, CA, USA). The results were photographed on a gel imaging system (Bio-Rad) and the relative gray value of each DNA band was estimated. Each measurement was repeated thrice.

The Cell Counting Kit-8 (CCK-8) assay was used to assess the cell viability. Cells were seeded onto 96-well plates at a concentration of 8×10³/well and incubated overnight under the usual culture conditions. Cells were exposed to each of 5-FU at various concentrations (0, 0.1, 1, 10, 100 and 1,000 μM; dissolved in dimethyl sulfoxide). After 10 μl of CCK-8 solution was added in each well, the plates were incubated for 1 h at 37°C. The absorbance of individual wells was read at 490 nm (test wavelength) and 530 nm (reference wavelength) using a microplate reader (Bio-Rad Laboratories Inc.). The sensitivity of tumor cells to 5-FU (5-FU treatment alone or in combination with LY294002) was determined by estimating the IC₅₀ values (doses that induce 50% growth inhibition) for 5-FU from the dose-response curves. Cell growth was examined using a CCK-8 kit (Cayman, Ann Arbor, Michigan, USA). Cell growth inhibition (%) was calculated as follows: [1-optical density (OD) values of 5-FU+/OD values of 5-FU-] × 100.

A staining solution of Hoechst 33258 was prepared immediately before use. After drug treatment, cells were incubated with Hoechst 33258 on 6-well plates (1 ml/well) for 20–30 min and washed thrice with phosphate-buffered saline. Cells were then assessed for Hoechst fluorescence using Nikon Intensilight C-HGF1 fluorescent microscope (Nikon, Tokyo, Japan) (magnification, ×200).

Statistical analyses were performed using SPSS version 13.0 software (Chicago, IL, USA). The results are expressed as means and standard deviations (± SD). Comparisons between groups were performed using one-way analysis of variance (ANOVA). Values of P<0.05 were considered to indicate a statistically significant difference.

Investigation of the correlation between CD133 and chemoresistance to 5-FU in GC cells was performed. First, CD133⁺ and CD133⁻ GC cell lines (SGC7901 and MKN45 GC cell lines) were treated with 5-FU (0–1,000 μM). CD133⁺ GC cells were found to be significantly resistant to 5-FU compared to autologous unsorted and CD133⁻ GC cells (Fig. 1A). The IC₅₀ values of CD133⁺, unsorted and CD133⁻ cells were: 37.74±1.32 to 26.39±2.04, 15.80±0.14 to 11.25±1.37 and 5.46±0.98 to 3.05±0.32 μM, respectively. Then, the presence of apoptosis was confirmed by Hoechst 33258 staining (Fig. 1B and C), which showed less frequent peripheral chromatin condensation and nuclear fragmentation in CD133⁺ GC cells than unsorted and CD133⁻ GC cells. Western blotting (Fig. 1D) and RT-PCR (Fig. 1E) were performed to detect the expression of P-gp, Bcl-2 and Bax in GC cells. The results showed that CD133⁺ GC cells had a tendency to express much higher amounts of P-gp and Bcl-2 than unsorted and CD133⁻ cells, but lower expression of Bax. These data suggest that CD133 may contribute to the observed resistance to apoptosis of CD133⁺ GC cells. Notably, overexpression of P-gp and Bcl-2 may protect CD133⁺ GC cells from apoptosis induced by 5-FU.

To further investigate the importance of CD133, gene silencing by RNA interference was performed. As shown in Fig. 2A, the expression of CD133 in GC cells was successfully knocked down by transfecting with CD133 siRNA. In addition, P-gp and Bcl-2 were decreased in CD133 siRNA-expressing cells compared to control siRNA-expressing cells, while the expression of Bax was increased (Fig. 2A and B). Furthermore, it was found that CD133 silencing increased the cytotoxicity of 5-FU in GC cells compared to cells without CD133 silencing (Fig. 2C). The IC₅₀ values of control, negative control and siRNA CD133 were: 16.65±4.54, 16.94±4.55 and 9.66±2.01 μM, respectively. The presence of apoptosis was confirmed by Hoechst 33258 staining (Fig. 2D and E), which showed more frequent peripheral chromatin condensation and nuclear fragmentation in CD133 siRNA-expressing cells than in control siRNA cells. Taken together, these results indicate that downregulation of CD133 may enhance the cytotoxicity of 5-FU in GC cells by regulating the expression of P-gp and Bcl-2 family.

To confirm the role of CD133 in 5-FU resistance in GC cells, CD133 was activated by transfecting the CD133 gene into SGC7901 cancer cells (Fig. 3A). Western blotting (Fig. 3B and C) showed that CD133 expression was stably increased in CD133-expressing cells compared with vector control cells. Moreover, P-gp and Bcl-2 were increased in CD133-expressing cells compared to vector control cells, while the expression of Bax was decreased (Fig. 3D). In addition, the IC₅₀ values of control, vector control and Flag-CD133 were: 8.69±1.03, 8.87±1.20 and 26.03±3.18 μM, respectively after treatment with 5-FU for 48 h (Fig. 3E), which indicated that CD133 overexpression increased resistance to 5-FU. The Hoechst 33258 staining (Fig. 3F and G) showed less frequent peripheral chromatin condensation and nuclear fragmentation in CD133-expressing cells than vector control cells. These data demonstrate that CD133 protects GC cells from 5-FU-induced cell death.

Recently, it was reported that Akt overexpression decreases the chemosensitivity of GC cells to 5-FU in vitro (26). Although CD133 is a downstream substrate of Akt, CD133 was shown to enhance Akt phosphorylation in several types of cancer cells (16,22,23,27). Thus, the present study investigated whether CD133-induced 5-FU resistance is attributed to Akt/p70S6K activation in GC cells. As a result, it was found that although levels of total Akt proteins were similar between CD133⁺ and CD133⁻ tumor cells, the phosphorylation of Akt on S473 and p70S6K was markedly upregulated in CD133⁺ GC cells compared with matched CD133⁻ GC cells (Fig. 4A and B). Thus, these findings suggest that elevated Akt activity may be a distinctive feature of CD133⁺ GC cells. Furthermore, western blotting (Fig. 4C and D) showed that CD133 activation and silencing directly increased and decreased the expressions of p-Akt and p-p70S6K, the active form of Akt and p70S6K, respectively.

To confirm whether the positive correlation between CD133 and 5-FU resistance is mediated by Akt, the present study assessed the effects of LY294002 and EGF treatment on 5-FU-induced cytotoxicity. Western blotting (Fig. 5A–C) showed that LY294002 treatment for 48 h effectively blocked p-Akt and p-p70S6K expression in CD133-expressing cells, while EGF-treatment for 48 h reversed the effect of CD133 siRNA on these two phosphorylated proteins. Moreover, the changes of P-gp and Bcl-2 were in accordance with the phosphorylated levels of Akt and p70S6K. In contrast, the expression of Bax changed in the opposite direction (Fig. 5A). In addition, CD133-expressing cells treated with LY294002 showed higher 5-FU cytotoxicity, while EGF reversed the effect of CD133 siRNA on 5-FU cytotoxicity (Fig. 5D). In addition, the IC₅₀ values of groups 1–4 were: 27.3±3.18, 18.01±0.18, 8.63±1.22 and 24.30±8.08 μM, respectively. Notably, Hoechst 33258 staining (Fig. 5E and F) illustrated that treatment with LY294002 or transfection with CD133-siRNA caused more apoptosis in 5-FU-treated CD133-expressing cells, while EGF partly reversed this change. Collectively, the above results indicate that CD133 enhances 5-FU resistance through regulation of P-gp and Bcl-2 family mediated by the PI3K/Akt/p70S6K pathway.