Mitochondrial reactive oxygen species accelerate gastric cancer cell invasion.

Tumor invasion is the most important factor to decide patient's prognosis. The relation between reactive oxygen species and tumor invasion is mainly reported that nicotinamide adenine dinucleotide phosphate oxidase in the cell membrane is a reactive oxygen species producer for formulating an invadopodia. On the other hand, mitochondrion was known as one of the most important reactive oxygen species-producer in the cell via an energy transfer system. However, the relation between mitochondrial reactive oxygen species and the tumor invasion was not well clarified. In this study, we evaluated the relation between mitochondrial reactive oxygen species and tumor invasion using a normal gastric mucosal cell-line (RGM-1) and a cancerous mutant RGM-1 cell-line (RGK-1). Manganese superoxide dismutase-expressing RGK-1 cell-lines were used for a scavenging mitochondrial reactive oxygen species. The cells have been evaluated their movement ability as follows; cellular ruffling frequencies, wound healing assay to evaluate horizontal cellular migration, and invasion assay using matrigel to analyze vertical cellular migration. All cellular movement abilities were inhibited by scavenging mitochondrial reactive oxygen species with manganese superoxide dismutase. Therefore mitochondrial reactive oxygen species was one of factors enhancing the tumor invasion in gastric cancer.

Introduction

Cancer cellular reactive oxygen species (ROS) play an important role in their proliferation, invasion and metastasis.( The ROS regulates tumor invasion through actin remodeling proteins and focal adhesion proteins.( Tumor invasion ability is the most important factor to decide patient’s prognosis. Tumors without invading abilities are usually diagnosed as benign lesions, and treatments decreasing these abilities possibly improve the prognosis. NADPH oxidase (NOX) was mainly reported to generate ROS involving tumor invasion. NOX is an enzyme located in cellular membrane to produce superoxide anion (O2•) which accelerates invadopodia formation.(

Mitochondrion is a main source of ROS, because it used to show complex I dysfunction to generate O2• in cancer cells. This dysfunction is induced by cancerous gene mutation.( Mitochondrial ROS derived cancer cellular progression and metastasis.( However, the relation between mitochondrial ROS and tumor invasion is not well investigated.

Manganese superoxide dismutase (MnSOD) is particularly localized at mitochondria to scavenge their O2•.( Overexpression of MnSOD after its gene transfection suppressed tumor proliferation and metastasis.( Mitochondrial ROS is thus likely to important for organizing tumor properties.

Recently we established two cell lines; a rat normal gastric mucosa cell-lines (RGM-1) and a cancerous RGM-1 cell-lines (RGK-1).( These cell lines are useful to compare cellular properties between normal cells and cancerous cells. In addition, we also established MnSOD-expressing cell lines, because these cell-lines used to be regarded as good experimental materials to clarify whether intracellular ROS were derived from mitochondria or not.( Using them, we have reported the roles of mitochondrial ROS to involve gastric epithelial cellular injuries after treatments with NSAIDs, bisphosphonate and exposure to acidic environments.(

In this study, we elucidated whether mitochondrial ROS involved tumor cellular invasion or not. For this aim, we investigated living cellular ROS spectra of electron paramagnetic resonance (EPR) using RGM-1, RGK-1 and MnSOD-overexpressing stable clone RGK-1 cell lines (RGK-1 MnSOD). We also evaluated cellular ruffling frequencies, wound healing assay and invasion assay as described in “Materials and Methods”.

Materials and Methods

Materials

Aminophenyl Fluorescein (APF) (SEKISUI MEDICAL CO., LTD., Tokyo, Japan), and 2-[5,5-Dimethyl-2-oxo-2λ5-(1,3,2)dioxaphosphinan-2-yl]-2-methyl-3,4-dihydro-2H-pyrrole 1-oxide (CYPMPO) (Radical Research Inc., Tokyo, Japan) were purchased. Can get signal (PVDF blocking reagent, Solution 1, Solution 2) and WB Stripping Solution Strong (Nacalai tesque Inc., Kyoto, Japan) were used for western blotting analysis.

Cell culture

RGM-1 and RGK-1 were cultured in DMEM/F12 (Life Technologies Co., CA) and Dulbecco’s modified EAGLE’S medium nutrient mixture F-12 HAM (Life Technologies Co.), respectively. These culture mediums contains 10% inactivated FBS and 1% penicillin/streptomycin. Culture medium for RGK-1 MnSOD also contained antibiotic G418 sulfate for being stable MnSOD expression. Gene vector transfected RGK-1 cell line was used as the control for a gene transfect. The different number in RGK-1 MnSOD means different clone like RGM-1 MnSOD-overexpression cell lines.( Each cells showed differ MnSOD-expression. All cells were cultured in 5% CO2 cell culture incubator at 37°C.

ROS measurement by EPR

EPR measurement was performed according to previous report.( Briefly, cells were cultured on cover slide glasses (49 × 5 × 0.2 mm). Cells were immersed in culture medium with 5 mM respiratory substrates (succinate, glutamate, and malate), 5 mM nicotinamide adenine dinucleotide (NADH) and 10 mM CYPMPO. Immersed-slide glasses were put on the tissue glass, thereafter the couple of glasses were measured by EPR apparatus.

Standard signals of H2O2 by Fenton reaction and O2• by xanthine/xanthine oxidase reaction were studied according to the report.( Briefly, same volume of 10 mM CYPMPO, 0.1 mM H2O2, 50 mM PBS (pH 7.4) and 3 mM FeSO4 were mixed for Fenton reaction, and the mixture was analyzed by EPR. For xanthine/xanthine oxidase reaction, 50 mM CYPMPO, 2 mM hypoxanthine, 0.4 U/mL xanthine oxidase and 50 mM PBS (pH 7.4) were mixed, and the mixture was used as O2• standard.

All EPR spectra were recorded by using a JEOL-TE X-band spectrometer (JEOL Ltd., Tokyo, Japan). All EPR spectra were obtained under the following conditions: 10 mW incident microwave power, 9.42 GHz frequency, 0.1 mT field modulation amplitude, and 15 mT scan range. The hyperfine splitting constants (hfsc) and spectral computer simulation were analyzed using a Win-Rad Radical Analyzer System (Radical Research Inc.). All EPR spectra were representative of at least 3 independent experiments.

Measurements of oxidative stress by APF

Free radicals (hydroxyl radical and peroxynitrite) were detected by APF as follows; APF was diluted with PBS and it exposed to the cells at the concentration of 1 µM for 30 min. After incubation, cells were washed using a cold PBS twice. The intensities of APF-fluorescent were measured by Varioskan (Thermo Fisher Scientific K.K., Kanagawa, Japan) at Ex. 490 nm and Em. 515 nm.

Western blotting analysis

Western blotting analysis was studied according to previous reports.( Briefly, the protein from cell-lysate was obtained according to RIPA buffer manufacture’s protocol. The amount of total protein in sample solution was determined using Protein Quantification kit - Rapid (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). The electrophoresis and the blotting procedure were studied as follows; 15 µl (10 µg) protein samples were applied in the well of NuPAGEgel (Life technologies). After the samples were applied in the lane, it was electrophoresed under 200 V for 30 min on ice. And then, proteins were blotted to PVDF membrane under 30 V for 10 min.

The definition of protein using sandwich immune assay were performed according to manufacture protocol of SNAP i.d. (Millipore Co., MA). Rabbit polyclonal MnSOD antibody (Enzo Life Sciences, Inc., NY) and β-actin rabbit antibody (Cell Signaling Technology, K.K., Tokyo, Japan) were used as first anti-body, and anti-rabbit IgG HRP-linked antibody (Cell Signaling Technology) was used as second anti-body. After Luminata forte western HRP substrate (Millipore) added, luminescence was detected by LAS4000 (GE Health care Japan Co., Ltd., Tokyo, Japan).

Evaluation of cell membrane ruffling

Cell membrane ruffling was determined with a method of Borm et al.( Briefly, the movie was created by 75 pictures (1 pictures/min). The movie was changed a black-and-white image (8 bit, tone; 256, Black; less than tone 80, white; more than tone 80), and the region of interest (ROI: 5,000 pixels/each movies) was set around the cells. Ruffling frequencies were calculated based on the sum of dark area of the movies using Image J software ver. 1.43 (National Institute of Health, MA).

Wound healing assay

Wound healing assay was studied according to previous report.( Briefly, the free-cell area was created using yellow-chip on the confluent cell. After incubated for 0, 6, 12 and 24 h, The free-cell area (µm2) was measured on the microscopic picture by the analysis software (ZEN software, Carl Zeiss Co., Ltd., Jena, Germany).

Tumor invasion assay

Matrigel was created in the well of 96-well plate. 50 µl Matrigel solutions were poured to the well at 4°C, and then the plate was incubated at 37°C for 30 min. After polymerization of matrigel, the cells (RGM-1, RGK-1, RGK-1 MnSOD2, RGK-1 MnSOD3 and RGK-1 MnSOD9; 500 cells/well) were dispersed on matrigel and it was incubated for 24 h. After cells were attached on the matrigel, the invasion assay was started (this time was 0 h in Fig. 6). The invaded-length was taken a pictured using the z-stack function of confocal microscopy after 0, 24 and 48 h. The pictures were obtained 6.02 µm/slice through z-axis movement. The measurement of invaded-length was calculated based on 3D matrigel pictures, those were created by 2D pictures. This experiment was studied twice independently.

Fig. 6

Invasion assay using matrigel. The cells (500 cells/100 µl) were harvested on matrigel and incubated overnight. After cells were attached, it was took pictures using the z-stack function of confocal microscopy (6.02 µm/slices) for 0, 24 and 48 h. 3D structure picture was created based on obtained-2D-pictures on the axio-vision software. The invaded-length was measured on the 3D picture. This assay was experimented twice, and the tendency of invasion was same.

Statistical analysis

Statistical analysis was calculated on the Origin software. Significant static value (p value) was calculated using ANOVA followed by Scheffe’s F-test.

Results

Cellular ROS concentration in normal cells and cancerous mutant cells

Generally, cancerous cells used to show higher ROS concentration than that of normal cells. We have evaluated the intracellular ROS concentration of two cell-lines. One is a normal RGM-1, and another is a RGK-1.( The intracellular ROS concentration was determined by EPR measurement with a spin trap reagent CYPMPO. Standard spectra of H2O2 by Fenton reaction and O2· by xanthine/xanthine oxidase reaction showed different EPR spectra from that of living cells (Fig. 1). RGK-1 spectrum of EPR was stronger than that of RGM-1 (Fig. 1). Additionally, we have measured cellular ROS with a fluorescent dye APF, of which intensity can be a representative of the amounts of hydroxyl radicals and peroxynitrite.( The intensity of RGK-1 was about twice higher than that of RGM-1 (Fig. 2).

Fig. 1

Cellular ROS concentration in living cells by EPR. CYPMPO was used as a spin trapping reagent.

Fig. 2

Cellular ROS concentration of RGM-1, RGK-1 control, RGK-1 MnSOD2, RGK-1 MnSOD3, and RGK-1 MnSOD9. The cellular ROS concentration was detected by APF. The MnSOD expression decreased the cellular ROS concentration significantly. The data are expressed as mean ± SD (n = 6). *p<0.05, **p<0.01.

MnSOD-expression and subcellular location in RGK-1 MnSOD

The amount of MnSOD expressed in each cell-line was determined by Western blotting analysis and digitalized calculations of MnSOD expression with immunohistochemically stained specimens, respectively (Fig. 3). The transfected MnSOD vector has the mitochondrial targeting sequences, the expressing-MnSOD can thereby scavenge mitochondrial ROS particularly.( After the gene vector transfection, the amount of MnSOD in RGK-MnSOD was larger than that of non-gene transfected RGK-1 (Fig. 3a), and these MnSOD was colocalized with mitochondria (Fig. 3b). Fig. 1 and Fig. 2 show intracellular ROS concentration of RGK-1 MnSOD. The EPR spectra of these RGK-1 MnSOD cells were weaker signals than that of RGK-1 control. Together, these results indicated that mitochondrial ROS were scavenged by MnSOD. The ROS concentration in RGK-1 was significantly higher than in normal cells, and MnSOD overexpression significantly decreased this cancer specific higher ROS concentration.

Fig. 3

Determination of the amount of MnSOD by Western blotting analysis and the colocarization of MnSOD and mitochondria. (a) Western blotting analysis of MnSOD. (b) MnSOD was immunostained with FITC labeled second antibody and mitochondria was stained with Mitotracker Red. The change of their fluorescent intensity showed almost same tendency.

Scavenging mitochondrial ROS by MnSOD decreased cellular ruffling frequencies

Supporting information Supplemental Table 1 shows the analytical results of cellular movement assays in this research. The cellular ruffling activity is one of essential movement relating cellular migration. The expression of MnSOD decreased cellular ruffling significantly (Fig. 4). These data of ruffling frequencies showed same tendency of the intracellular ROS concentration.

Fig. 4

Determination of cellular ruffling activity. These data shows the average ruffling area in 5000 pixels of pictures. MnSOD reduced the cellular ruffling activity. The data are expressed as mean ± SD. (n = 6), *p<0.05.

Scavenging mitochondrial ROS by MnSOD suppressed horizontal cellular migration and cell invasion to matrigel

The relation between horizontal cellular migration and mitochondrial ROS was evaluated. The wound healing assay was studied as the evaluation of horizontal cellular migration. The results of wound healing assay might be affected by the cell growth. Therefore we have evaluated at first the cell growth speed (Supplemental Fig. 1). There were no significant differences in each RGK-1 cell lines. The results of wound healing assays indicated that the fastest cell-line of horizontal cellular migration ability was RGK-1, RGK-1 MnSOD was the second one and RGM-1 was the latest one (Fig. 5). These results indicated that mitochondrial ROS played an important role in cellular migration.

Fig. 5

Wound healing assay. (a) The pictures of free-cell area. The free-cell area was surrounded with yellow line. (b) The percentage of affected area compared as the pore size incubated after 0 h. These results indicated that the migration is enhanced by tumorigenesis and it was decreased reducing cellular ROS-concentration by MnSOD. The data are expressed as mean ± SD (n = 6).

Scavenging mitochondrial ROS by MnSOD inhibited the cellular invasion to matrigel

To elucidate the relation between cellular invading abilities and ROS, we measured length of each cellular invasion into matrigel layer, an artificial model of extracellular matrix.( The length of vertical cellular migration from the top of matrigel to the deepest cells with a confocal microscopy was measured. These results showed the same tendency of horizontal cellular migration: the RGK-1 showed the highest performance of invasion ability, RGK-1 MnSOD showed the second one and RGM-1 showed the lowest (Fig. 6).

Discussion

The intracellular ROS concentration in RGK-1 was about twice higher than that in RGM-1 according to the results of EPR measurement using living cells. RGK-MnSOD significantly decreased intracellular ROS concentration. We thus concluded that cancerous mutation involved mitochondrial ROS production. The eliminated-mitochondrial ROS should be superoxide anion because MnSOD reduced the intensity of ESR spectra in RGK-1.

The tumor cellular invasion has suppressed by scavenging mitochondrial ROS. Cellular invasion needs alteration of cellular morphology including invadopodia to move. The data suggested the participation of Tropomyosin 1 and NOX-derived superoxide anion for forming invadopodia, which show an obvious amount in RGK-1 but few in RGM-1.( These factors should be impaired by expressed MnSOD. We confirmed the activation of cytoskeletal-associated protein due to clarify the mechanism of cellular movements through a mitochondrial ROS. However the amount of Rac1/cdc42 did not reduce by the expression of MnSOD (data not shown). The mitochondrial ROS might have a different mechanism for the tumor invasions from the NOX-ROS pathway. Cellular invasion also needs the ability of degrading collagens to create the road in extracellular matrix. Mitochondrial ROS controls matrix metalloproteinase signaling (MMP) by activating MMP gene expression or proenzymes.( MMP enhances tumor invasion ability by degrading collagens. MMP signals may play an important role for tumor invasion. We now under-investigate it.

Conclusions

We demonstrated for the first time that mitochondrial ROS involved cancer cellular invasion. The intensity of EPR spectra from RGK-1 cells was twice higher than that of RGM-1 cells, and this EPR signals was decreased by scavenging mitochondrial ROS. The cellular membrane ruffling, cellular migration and cell invasion were suppressed by scavenging mitochondrial ROS.