Gastric cancer stem cells survive in stress environments via their autophagy system

Cancer stem cells (CSCs) play an important role in the progression of carcinoma and have a high potential for survival in stress environments. However, the mechanisms of survival potential of CSCs have been unclear. The aim of this study was to clarify the significance of autophagy systems of CSCs under stress environments. Four gastric cancer cell line were used. Side population (SP) cells were sorted from the parent cells, as CSC rich cells. The expression of stem cell markers was examined by RT-PCR. The viability of cancer cells under starvation and hypoxia was evaluated. The expression level of the autophagy molecule LC3B-II was examined by western blot. The numbers of autophagosomes and autolysosomes were counted by electron microscope. SP cells of OCUM-12 showed a higher expression of stem cell markers and higher viability in starvation and hypoxia. Western blot and electron microscope examinations indicated that the autophagy was more induced in SP cells than in parent cells. The autophagy inhibitor significantly decreased the viability under the stress environments. These findings suggested that Cancer stem cells of gastric cancer might maintain their viability via the autophagy system. Autophagy inhibitors might be a promising therapeutic agent for gastric cancer.


Effect of starvation and hypoxic stress on the proliferation of OCUM-12 cells.
We examined the effect of starvation and hypoxic stress on the proliferation activity of the parent OCUM-12 cells and the OCUM-12/SP cells. The proliferation of parent cells and SP cells was quantified using an MTT colorimetric assay and cell counting assay. Both assays showed the proliferative activity of the OCUM-12/SP cells was significantly higher than that of the parent OCUM-12 cells in amino-acid-free medium, FBS-free medium, and under a 1%O 2 condition, but not in a low-glucose medium (Fig. 3).

Effect of starvation and hypoxic stress on the SP fraction of OCUM-12 cells.
The effect of starvation and hypoxic stress on the SP fraction was analyzed by flow cytometry. Starvation stress and hypoxia significantly increased the SP fraction (Fig. 4). The SP fraction observed when using the low-glucose medium, aminoacid-free medium, FBS-free medium, and 1%O 2 condition was significantly higher than that of the control.  and NANOG staining was performed to determine the CSCs of OCUM-12. LC3B and p62 staining was performed to evaluate the autophagic activity. Both positive of SOX2 and NANOG was evaluated as CSC-positive, and both positive of LC3B and p62 was evaluated as autophagy-positive. CSC-positive was significantly (p = 0.016) associated with autophagy-positive (Fig. 5).
Effect of starvation and hypoxic stress on autophagy. The LC3B-II expression in the parent OCUM-12 cells was greatest when using the amino-acid-free medium, while that in the OCUM-12/SP cells was greatest with the FBS-free medium. With the control medium and low-glucose medium, the LC3B-II expression was slight in both types of cells. The LC3B-II expression when using the amino-acid-free medium, FBS-free medium, and 1%O 2 condition was higher in the OCUM-12/SP cells than in the parent OCUM-12 cells, and the difference in LC3B-II expression was greatest with the FBS-free medium (Fig. 6). In addition, the numbers of autophagosomes and autolysosomes were counted using an electron microscope (Fig. 7A). A significant increase in the numbers of autophagosomes and autolysosomes in OCUM-12/SP cells was observed when using the aminoacid-free medium (p < 0.001), FBS-free medium (p = 0.011), 1%O 2 condition (p = 0.001), while no significant difference was found in the control medium and low-glucose medium (Fig. 7B).
The effect of an autophagy inhibitor, CQ, on the proliferation of OCUM-12 cells. The proliferation activity of OCUM-12/SP cells in the amino-acid-free medium, FBS-free medium, and under the 1%O 2 condition was significantly decreased following the addition of the autophagy inhibitor, CQ, in comparison to the parent OCUM-12 cells (Fig. 8A).
The effect of CQ on the proportion of SP cells. CQ reduced the SP fraction. CQ combined with amino-acid-free medium, FBS-free medium, and under the 1%O 2 condition greatly reduced the SP fraction compared to the control medium ( Fig. 8B,C). The proliferative activity of the OCUM-12/SP cells was significantly higher than that of the parent OCUM-12 cells in amino-acid-free medium, FBS-free medium, and under a 1%O 2 condition, but not in a low-glucose medium. *p < 0.05, **p < 0.01, *** p < 0.005.

Discussion
We aimed to clarify the significance of the autophagy systems of CSCs under stress environments such as starvation and hypoxia. It has been reported that SP cells of gastric cancer possess cancer stem-like properties 8,9 . We   www.nature.com/scientificreports/ associated with autophagy-positive, both positive of LC3B and p62. The interpretation of p62 positivity in immunohistochemical staining remains to be controversial. It has reported that the accumulation of p62 may indicate the inhibition of autophagy 16 . On the other hand, it has also reported that the accumulation of p62 may indicate autophagy-induced autophagosomes 17,18 . In this study, we determined that the both positive expression of LC3B and p62 of the xenografted tumor by OCUM-12 cells may indicate as the autophagy positive, because the CSC-rich OCUM-12/SP cells significantly increased the number of autophagosomes in compared to the parent OCUM-12 cells by electron microscopic examination. LC3B-I is conjugated and forms LC3B-II, which is recruited to autophagosomal membranes. Autophagosomes fuse with lysosomes to form autolysosomes and are degraded by lysosomal hydrolases. The autophagosomal marker LC3B-II reflects starvation-induced autophagic activity 19 . Western blot of LC3B-II indicated that the autophagy system was induced in both the parent cells and SP cells in FBS starvation, amino acid starvation, and hypoxia (1%O 2 condition), but that of SP cells was greater than that of parent cells. The number of autophagosomes and autolysosomes observed by electron microscopy was also increased in SP cells under FBS starvation, amino acid starvation and hypoxia (1%O 2 condition), which supported the results of Western blot. The growth activity of SP cells was greater than that of parent cells under these stress environments. These findings might suggest that the survival potential of CSCs was greater than parent cells inducing autophagy system.
The autophagy inhibitor, CQ, significantly decreased the growth activity of the OCUM-12/SP cells under FBS starvation, amino acid starvation, and hypoxia (1%O 2 condition), in compared to the parent OCUM-12 cells. The heterogenicity of starvation and hypoxic lesions in the tumor microenvironment is often associated with malignant transformation of solid tumors 20 . CSCs might survive under these stress environments via autophagy systems, and the CQ might decrease the survival activity of the CSCs of gastric cancer under stress environments. There are only two reports about the effect of CQ on CSCs in ovarian cancer 21 and hepatocellular carcinoma 22 . This is the first report of the inhibitory effect of the autophagy inhibitor, CQ, on the proliferation of gastrointestinal CSCs. Gastrointestinal CSCs have been proposed to play an important role in the progression of carcinomas including the metastasis and chemoresistance of cancer cells. These findings suggested that the autophagy inhibitor might be useful for advanced stage patients with metastatic carcinoma or chemo-resistant gastric carcinoma.
In conclusion, CSCs of gastric cancer might maintain their viability under the stress environments of starvation and hypoxia via the autophagy system. Autophagy inhibitors might be promising therapeutic agents for gastric cancer.  To determine fold changes in each gene, RT-PCR was performed on the ABI Prism 7500 (Applied Biosystems, Foster City, CA, USA), with commercially available gene expression assays (Applied Biosystems) for, CD44, CD133, SOX2, NANOG, and ATP-binding cassette, sub-family G, member 2 (ABCG2). ACTβ was used as an internal standard to normalize mRNA levels. The primer sequences were listed in Supplementary Table 1. Effect of starvation or hypoxia and chloroquine on SP fraction. Each cell was incubated 12 h in each medium or hypoxia, either alone or in the presence of 10 μM chloroquine (CQ; SIGMA). Cells were suspended at 5 × 10 5 cells/ mL in PBS. SP fraction was analyzed by flow cytometry.
Cell growth assay and cytotoxic assay. The proliferation of parent cells and SP cells and cytotoxicity of CQ were quantified using an MTT colorimetric assay (Dojindo, Kumamoto, Japan) and cell counting assay. Each cell was seeded at a cell density of 1.5 × 10 3 /well in 96-well plates for MTT colorimetric assay, and seeded at a cell density of 1.0 × 10 4 /well in 12-well plates for cell counting assay in each medium and hypoxia, and incubated 96 h at 37 °C for proliferation assay. For cytotoxicity assay, each cell was seeded at a cell density of 3.0 × 10 3 /well in 96-well plates in each medium and hypoxia with or without 10 μM CQ, and incubated 48 h at 37 °C. MTT solution was added to each well at 500 μg/ ml. The cells were incubated for 2 h and lysed in dimethyl sulfoxide and the absorbance value was analyzed using a Varioskan LUX (Thermo Fisher scientific, Waltham, MA, USA) at a wavelength of 535 nm.
Western blot analysis. Each cell was seeded at a cell density of 5.0 × 10 4 /well in 6-well plates under each medium or hypoxia in absence (− CQ) or presence (+ CQ) of 10 μM CQ and incubated 6 h at 37 °C. Cell lysates were made by standard methods. The protein concentration of each sample was measured using a Bio-Rad protein assay kit II (Bio-Rad Laboratory, Richmond, CA, USA). For SDS-PAGE, 2.5 μg of proteins from each sample www.nature.com/scientificreports/ were subjected to electrophoresis on 10-15% polyacrylamide gels. Proteins were electrophoretically transferred to polyvinylidene difluoride membranes with a tank transfer systems (Bio-Rad Laboratory), then blocked with buffer containing 3.0% skim-milk and 0.1% Tween-20 in Tris-buffered saline (TBST) at room temperature for 1 h. The blotting membranes were cut prior to hybridization with primary antibodies. Primary antibody of a microtubule-associated protein-light chain 3B (LC3B; L7543, SIGMA) was used at 1:1000 dilution, GAPDH (sc-47724, Santa Cruz, Dallas, TX) was used at 1:5000 dilution in TBST containing 3.0% skim-milk. The membranes were incubated with primary antibody overnight at 4 °C, washed (3 × 5 min) with TBST, followed by incubation with a horseradish peroxidase-conjugated secondary antibodies (NA931V and NA934V, GE Healthcare, Chicago, IL, USA) in TBST for 1 h at room temperature, then washed (3 × 10 min) with washing buffer. Detection of chemiluminescence was performed with immunoStar LD (WAKO) following the manufacturer's instructions. Band intensity were estimated using ImageQuant TL (GE).
Counting the autophagy flux by electron microscope. Each cell was incubated 12 h in each medium and hypoxia, and was trypsinized and pelleted. The cells were then suspended and fixed 30 min at 4 °C in 2.5% glutaraldehyde with 2% paraformaldehyde in 0.1% M Phosphate buffer. The cells were then rinsed 4 times in 0.1% M Phosphate buffer and spun down into 3% agarose at 55 °C, and cooled to form blocks. The agarose blocks were incubated in 1% osmium tetroxide in 0.1 MPB for 2 h at room temperature. The agarose blocks were rinsed 4 times in 0.1% MPB and dehydrated in graded steps of alcohol and embedded in propylene oxide. Following polymerization overnight at 60 °C, 80-nm sections were cut on a UltracutUCT (Leica, Wien, Austria) and picked up on copper grids. The grids were post-stained in uranyl acetate and Reynolds solution. The sections were observed in a Talos F200C (FEI, Hillsboro, OR, USA). The numbers of autophagosomes and autolysosomes in 10 cells was counted under each condition.
Statistical analysis. The statistical analysis was done using the Student's t-test. Statistical significance was set at ≤ 0.05.