SNU601, SNU638 and SNU216 human GC cell lines obtained from the Korea Cell Line Bank were grown in RPMI-1640 medium (Invitrogen) supplemented with 10% (v/v) fetal bovine serum and 1% antibiotics at 37°C in 5% CO₂.

Treated cells were incubated with 1 μg/ml Hoechst 33342 at 37°C, 5% CO₂ for 15 min in the dark, and both floating and attached cells were collected by centrifugation. The pooled cell pellets were washed with ice-cold phosphate-buffered saline (PBS) and fixed in 3.7% formaldehyde on ice, then washed again with PBS, re-suspended and a fraction of the suspension was centrifuged in a cytospinner (Hanil, Korea). The slides were air dried, mounted in an anti-fade solution, and examined using a DM5000 fluorescence microscope (Leica, Germany) at respective excitation/emission wavelengths of 340/425 nm. A total of 500 cells from randomly chosen fields were counted and the number of apoptotic cells was expressed as a percentage of the total number of cells counted. For staining of autophagic vacuoles, treated cells were incubated with MDC for 30 min and washed with PBS and observed under a fluorescence microscope at 340/525 nm.

Cells were incubated with MTT solution (0.5 mg/ml) for 4 h and solubilized using dimethylsulfoxide, and the solubilized formazan product was quantified using an enzyme-linked immunosorbant assay (ELISA) plate reader at 595 nm; absorbance of untreated cells was designated as 100% and cell survival was expressed as a percentage of this value.

Cells treated with cholesterol for 12 h were trypsinized, washed and re-plated (2,000 cells/60-mm dish). After incubation for 14 days in a 37°C/5% CO₂ incubator, colonies were stained using crystal violet and scored (>1 mm).

Equal amounts of protein were electrophoretically separated using SDS-PAGE and transferred to a nitrocellulose membrane using a standard technique. Antibodies were used to probe for active P-ERK1/2, P-MEK1, MEK1, P-JNK1/2, P-p38, p38, caspase-3 (Cell Signaling Technology), PARP, Fas, caveolin-1, ERK2, JNK1/2, cytochrome c (Santa Cruz), DR4, DR5 (Pro Sci), ATG5 and LC3II (MBL International). Anti-α-tubulin (BioGenex) was used as a loading control. Acquisition of signals was performed using an image analyzer (Image Station 4000MM, Kodak).

The caspase-8 activity assay was performed using a FADD-like IL-1β-converting enzyme (FLICE) colorimetric assay kit (BioVision), according to the manufacturer’s protocol. Briefly, 200 μg protein lysates in a 50-μl volume was mixed with reaction buffer, mixed with IETD-pNA substrate, incubated for 90 min, and the absorbance was measured at 405 nm. Fold increase in FLICE activity was determined by comparing the results of treated samples with the level of the untreated control.

The constitutively active mutant of MEK1 (CA-MEK1) was designed by substitution of the regulatory phosphorylation sites, Ser²¹⁸ and Ser²²², with aspartic acid (S218D/S222D mutant), as described previously (18), and cloned into the pCMV vector. Then, cells were transfected transiently for 48 h and the protein amounts of MEK1 and p-ERKs were confirmed by western blot analysis. For the RNAi, siRNA were obtained from Bioneer and 5–8 μg/ml RNA were used for transfection by use of AMAXA nucleofector kit (Amaxa Biosystems GmbH) according to the manufacturer’s protocol. Sequences of siRNA are as follows: control sense, 5′-CCUACGCCACCAAUUUCGU(dTdT)-3′; DR5 sense, 5′-CAGACUUGGUGCCCUUUGA(dTdT)-3′.

All numerical data are reported as mean ± SE. All data represent the results of at least three independent experiments. Groups were compared using Student’s t-test.

To assess the effect of cholesterol on the viability of GC cells, a number of different human stomach cancer cell lines, SNU601, SNU638 and SNU216, were used in the study. Incubation of these cells in a culture medium with addition of different concentrations of water-soluble cholesterol resulted in a dose-dependent decrease in cell viability in all three cell lines, as determined by the MTT assay. In addition, treatment with cholesterol also resulted in a marked, dose-dependent decrease in colony forming ability of these cells (Fig. 1). Thus, cholesterol may exert antitumor activity in GC cells.

To exclude the specific effect of increased water solubility of cholesterol, we tested the effect of hydrophobic cholesterol solubilized in acetic acid in these cells; water-insoluble hydrophobic cholesterol also reduced cell viability when compared with the control group, the same as water-soluble cholesterol treatment (data not shown). Then, in order to explore the cause of cholesterol-induced viability reduction, we attempted to determine whether cholesterol induces apoptosis in GC cells. Cells exposed to cholesterol were stained by Hoechst 33342 and apoptotic nuclei were then counted. Upon exposure to cholesterol, apoptotic body formation showed a clear increase in SNU601 cells, however, only a slight increase was observed in SNU638 and SNU216 cells (Fig. 2A). Cholesterol also strongly induced cleavage of caspase-3, degradation of PARP and activation of caspase-8 in SNU601 cells, and weakly in SNU638 and SNU216 cells, as demonstrated by immunoblot analysis and caspase activity assay (Fig. 2B). Although viability of all three cell lines was decreased by cholesterol, active apoptotic induction was observed only in SNU601 cells. Thus, we also attempted to determine whether other types of cell death are involved in cholesterol-induced reduction of cell viability. Because cleavage of LC3I to LC3II is a sign of autophagy, we examined the question of whether autophagy occurred in this event by detection of LC3II protein level. Treatment with cholesterol resulted in a mild increase in LC3II level in SNU601 cells, with a stronger increase in SNU638 and SNU216 cells (Fig. 2C). Furthermore, essential autophagic factor, ATG5 was also accumulated in SNU638 and SNU216 cells. However, necrotic features, PI inclusion or LDH release were not observed in these cell lines by 48 h (laboratory observation). Therefore, cholesterol seemed to trigger cell killing by apoptosis and autophagy in GC cells.

Autophagy may function as a protective mechanism and as a cell death mechanism. If this cholesterol-induced autophagy exerts a protective role, prevention of autophagy will further enhance cell death upon exposure to cholesterol. Nevertheless, inhibition of autophagy by 3-methyladenine (3-MA) did not aggravate cholesterol-mediated cell toxicity (Fig. 3A). This implies that cholesterol-induced autophagy is not a protective autophagy, but a process of a cell death. However, inhibition of autophagic death did not restore viability of cholesterol-treated cells either. Autophagy inhibitor, 3-MA resulted in a decrease in the number of autophagic vacuoles, as detected by MDC staining. However, instead of a decrease of autophagic death, 3-MA led to an increase in apoptosis, as detected by elevated apoptotic body formation, cleavage of procaspase-3 and PARP, and activation of caspase-3 and caspase-8 (Fig. 3B–D). Therefore, it appears that the fate of the cell has already been determined as death before decision of cell death mode between autophagic death and apoptosis; thus, blockade of the autophagic pathway may take a bypass to the apoptotic pathway instead of cell survival.

Mitogen-activated protein kinase (MAPK) family proteins play critical roles in regulation of cell survival and death; therefore, in order to determine the signaling mechanism involved in cholesterol-induced GC cell death, we first examined the effect of cholesterol on activities of MAPK family members. Exposure to cholesterol triggered an abrupt and transient decrease of ERK1/2 phosphorylation not affecting JNK or p38MAPK in both SNU601 and SNU638 cell lines. MEK1, upstream kinase of ERK1/2, was also transiently dephosphorylated by cholesterol (Fig. 4A). To examine the role of the ERK1/2 pathway in cholesterol-treated cells, we observed the effect of cholesterol in CA-MEK1 overexpressed cells. In comparison with control cells, CA-MEK1 expressed cells demonstrated significant rescue of cell viability (Fig. 4B). Thus, reduction of ERK activity appears to be responsible for cholesterol-mediated cell death. Nevertheless, because phospho-status of ERK was recovered soon after dephosphorylation upon treatment with cholesterol, we investigated the role of restored ERK activity. MEK inhibitor, U0126, was added to medium 3 h after treatment with cholesterol in order to inhibit reactivation of ERK1/2. Combination of U0126 resulted in strongly increased apoptosis compared to cholesterol treatment alone, as demonstrated by enhancement of apoptotic body formation and cleavage of procaspase-3 and PARP (Fig. 4C and D). However, the level of autophagic marker, LC3II, was not altered by U0126 in either cell type (Fig. 4D), indicating that inhibition of the late ERK pathway resulted in activation of strong apoptotic signaling. These results suggest that cholesterol-induced early downregulation of ERK signaling evokes loss of cell viability, and subsequent restoration of ERK activity may be responsible for inhibition of apoptotic pathway.

Cholesterol is an essential component of the cell membrane and cholesterol-enriched membrane lipid rafts can affect membrane receptor-linked signal regulation. Furthermore, in this study, treatment with cholesterol resulted in activation of caspase-8, a representative death receptor-mediated apoptotic enzyme. Thus, we attempted to determine whether the death receptor pathway is involved in this cholesterol-mediated cell death event. In immunoblot analysis to assess expression of death receptors; TRAIL-R1/DR4, TRAIL-R2/DR5 and Fas/CD95, increase of TRAIL-R2/DR5 protein was detected after treatment with cholesterol in both cell lines, however, protein levels of TRAIL-R1/DR4 and Fas/CD95 were unchanged by cholesterol or were too low to be observed in these cells (Fig. 5A). In order to understand the role of TRAIL-R2/DR5 overexpression in cholesterol-mediated apoptosis, we performed an interference assay using siRNA of TRAIL-R2/DR5. Compared to scrambled control RNA (CTL RNAi) transfection, TRAIL-R2/DR5 siRNA transfection resulted in a decrease in cholesterol-induced apoptosis in both GC cell lines (Fig. 5B). In addition, knockdown of TRAIL-R2/DR5 reduced not only cleavage of caspase-3 and PARP but also elevation of LC3II level (Fig. 5C). Thus, TRAIL-R2/DR5 appeared to be linked to apoptotic and autophagic death pathways in response to cholesterol.