SNU-601 and AGS human gastric cancer (GC) cells were obtained from the Korean Cell Line Bank and the American Type Culture Collection, respectively. Cells were cultured in RPMI-1640 medium (Invitrogen; Thermo Fisher Scientific, Inc.) supplemented with 10% (v/v) fetal bovine serum and 1% PS at 37°C in an atmosphere containing 5% CO₂. For preparation of acidic pH medium, pH 7.4 RPMI-1640 medium containing 1% FBS medium was prepared, and then adjusted to pH 6.5 or 6.2 by adding 5M HCl solution. Sulindac and NS398 were purchased from Calbiochem; EMD Millipore. For Cox inhibition, cells were exposed to pH 7.4, 6.5 and 6.2 in the presence of 20 µM NS398 or 100 µM sulindac for 72 h, and the treated cells were harvested for western blot analysis or real-time PCR assay.

To assay cell invasiveness, Matrigel-coated Transwell chambers (Costar; Corning Incorporated) were used. Cells in the logarithmic phase were incubated in RPMI-1640 medium containing 1% FBS, adjusted to pH 7.4, 6.5 or 6.2, for 72 h, after which the number of viable cells that corresponded to the equal absorbance detected in a standard curve for a viability assay were obtained from each pH condition and suspended in 1% FBS/RPMI-1640 medium adjusted to each pH condition (pH 7.4, 6.5 or 6.2). Then ~200 µl of the cell suspension was added to the upper portion of the insert, and medium containing 5% FBS was added to the lower portion of the inset. After 8 h (for AGS) and 18 h (for SNU601) of incubation at 37°C in 5% CO₂, noninvasive cells were removed from the upper surface of the Transwell membrane with a cotton swab, and the invaded cells on the lower layer surface were fixed by 4% formaldehyde for 15 at RT and stained with 0.5% crystal violet solution for 10 min RT. The numbers of invaded cells were counted under high-power magnification (×200) or the images were obtained by image analysis software (analysis FIVE; Olympus Corporation).

Before the invasion assay, a standard curve for viability in each pH medium was prepared since exposure to acidic pH medium decreased the number of viable cells. The number of viable cells was determined as follows: Cells cultured for 3 days at each pH were trypsinized, resuspended in normal pH medium and further incubated for the time required for invasion measurement, and a standard curve for viability was prepared using an MTT assay. Following MTT incubation, the purple formazan crystals were dissolved using dimethyl sulfoxide and viability was subsequently analyzed at a wavelength of 595 nm. Based on these survival curves, the number of cells corresponding to the same absorbance was calculated and used for this assay.

Treated cells were lysed in either a whole-cell lysis buffer (50 mM Hepes, 150 mM NaCl, 1% Triton X-100, 5 mM EGTA, protease inhibitor cocktail). Protein determination was performed by BCA method. Equal amounts of protein extracts (50 µg of protein per lane) were electrophoretically separated using 10–12% SDS-PAGE and transferred to a nitrocellulose membrane using a standard technique. Blocking was performed by incubation in 3% skim milk/0.1%Tween Tris-buffered saline for 30 min at 4°C. Antibodies were used to probe for KLK7 (dilution 1:500; cat. no. sc-514447; Santa Cruz Biotechnology, Inc.), cyclooxygenase (COX)2 (dilution 1:200; cat. no. 160112; Cayman Chemical), KLK8 (dilution 1:1,000; cat. no. ab150395; Abcam), COX1 (dilution 1:500; cat. no. ab695; Abcam) and α-tubulin (dilution 1:2,000; clone no. B-5-1-2; cat. no. 32-2500; Invitrogen; Thermo Fisher Scientific, Inc.). The secondary antibodies used were: Anti-rabbit HRP (dilution 1:2,500; cat. no. sc-2030; Santa Cruz Biotechnology, Inc.), anti-mouse HRP (dilution 1:2,500; cat. no. sc-2031; Santa Cruz Biotechnology, Inc.) were used. Anti-a-tubulin was used as a loading control. Protein bands were visualized with the Super Signal West Pico chemiluminescence kit (Pierce Biotechnology; Thermo Fisher Scientific, Inc.). Protein blots were quantified using Alpha Ease FC software version 4.0 (Alpha Innotech). Signals were acquired using an Image Station 4000MM image analyzer (Kodak).

For the RNAi experiment, siRNAs of KLK7, KLK8, COX1 and COX2, and a scrambled siRNA control were purchased from Bioneer Corporation. Cells were individually transfected with siRNA oligonucleotides using an Amaxa™ Transfection System (Lonza Bioscience) and grown for 24 h prior to exposure to acidic pH medium.

Real-time PCR was performed with the Light Cycler 2.0 (Roche Diagnostics) using the Fast Start DNA Master SYBR-Green I Kit (Roche Diagnostics). For verification of the correct amplification product, PCR products were analyzed on a 2% agarose gel stained with ethidium bromide. The sequences of the primers were designed as follows: For β-actin, 5′-GACTATGACTTAGTTGCGTTA-3′ and 5′-GCCTTCATACATCTCAAGTTG-3′, for COX1, 5′-GGCGGGTACATTTCTCCATC-3′ and 5′-CCTCATGTTTGCCTTCTTTGC-3′, for COX2, 5′-AACACAACAGAGTATGCGA-3′ and 5′-GTGTTAAATTCAGCAGCAATACG-3′. Primers of KLK7 (product no. P106933), KLK8 (product no. P178536), matrix metalloproteinase 7 (MMP7) (product no. P310408) and matrix metalloproteinase 9 (MMP9) (product no. P323207) were purchased from Bioneer Corporation. PCR was conducted at 95°C for 10 min, followed by 45 cycles of 95°C for 15 sec, 60°C for 5 sec, and 72°C for 7 sec. Melt curve analysis was performed to confirm that a single product was present. Negative controls without template were included in each run. Data were analyzed using Light Cycler software version 4.0 (Roche Diagnostics). The 2^−ΔΔCq method was used for analysis of relative gene expression (18).

Cells were incubated in RPMI-1640 medium containing 1% FBS, adjusted to pH 7.4, 6.5 or 6.2, for 72 h, after which culture medium was obtained and release of KLK7 and KLK8 from cells was assessed using a specific sandwich-type ELISA assay according to established procedures with some modifications (19). Briefly, 96-well polystyrene plates were coated with 500 ng/well of a specific capture antibody. After overnight incubation, the plates were washed and 50 µl of culture supernatant or standard and an equal volume of assay buffer were added and incubated for 2 h. Plates were washed and biotinylated antibodies were subsequently added. After incubation and washing, alkaline phosphatase-conjugated streptavidin was added. Then, diflunisal phosphate stock solution and terbium-based detection solution were added, and fluorescence was assessed with a fluorometer (Perkin-Elmer). Calibration and data reduction were performed automatically.

All numerical data are presented as the mean ± SE of at least three independent experiments. Student's t-test was used for simple comparisons, and one-way ANOVA with Tukey's post hoc test was applied for multiple comparisons. P<0.05 was considered to indicate a statistically significant difference.

Previous research has revealed that extracellular acidity induces tumor growth, invasion, and metastasis. It was observed that incubation of gastric cancer cell lines AGS and SNU601 in an acidic pH medium triggered increased invasion of the cells compared to those cultured in normal-pH medium. To detect invasiveness, cells were cultured in culture medium adjusted to pH 7.4, 6.5, and 6.2 for 72 h, then the same number of viable cells were obtained and added onto the upper portion of Transwell plates which were coated with Matrigel. After 8 h for AGS and 24 h for SNU601 cells, the number of cells that invaded to the lower chamber was counted. The results revealed that the number of cells invading to the lower portion of the Transwell plates increased in the acidic environment (Fig. 1A-D). In addition, the mRNA levels of proteolytic enzymes MMP7 and MMP9, which are known to be involved in cancer cell invasion, were increased, as assessed by real-time PCR (Fig. 1E-H).

Proteases play crucial roles in cancer invasion and metastasis due to their proteolytic activity to cleave extracellular matrix and adhesion molecules. Although MMPs are known to be essential proteases mediating cell invasion, a possible involvement of other proteases in cancer cell invasion under an acidic environment was explored. Notably, it was revealed that the levels of KLK7 and KLK8 were significantly increased in acidic culture conditions. Acidic pH culture conditions significantly increased the mRNA levels of KLK7 and KLK8 in both gastric cancer cell lines (Fig. 2A-H). To assess whether increased mRNAs of KLK7 and KLK8 led to increased protein levels, immunoblot assays were performed using cell lysates. When the cells were incubated in acidic medium, the protein levels of KLK8 increased but those of KLK7 decreased (Fig. 2I and J). Since KLKs are proteases mainly secreted into the extracellular space, the protein levels of KLK7 and KLK8 released into the medium were also examined using an ELISA assay. As the pH decreased, the amounts of KLK7 and KLK8 proteins released into the medium significantly increased (Fig. 2K and L). Thus, acidity induced the expression and secretion of KLK7 and KLK8.

Since KLK family members have been revealed to be involved in malignant phenotypes of certain cancers, it was hypothesized that increased KLK7 and KLK8 in acidic environments would play a role in the acidity-mediated invasion of these cancer cells. To examine this, cells treated with siRNAs specifically silencing KLK7 or KLK8 were exposed to acidic medium. Cells in which KLK7 or KLK8 were silenced exhibited low invasiveness after incubation in acidic (pH 6.5) culture medium (Fig. 3A and B). The silencing efficiency of the KLK7 and KLK8 siRNAs was confirmed by decreased protein levels (Fig. 3C and D). Thus, increased expression of KLK7 and KLK8 appears to be linked to the enhanced invasiveness of gastric cancer cells.

To understand the regulatory mechanisms involved in acidity-mediated KLK7 and KLK8 expression, potential upstream regulators were examined. Previously, acidity was revealed to stimulate inflammatory responses via the cyclooxygenase (COX) pathway. Consistent with a previous study, acidic culture conditions elevated mRNA and protein levels of COX2 in both GC cell lines (20). In addition, the mRNA and protein expression of COX1, a constitutive isoform, increased (Fig. 4). Thus, the role of COXs in KLK induction was examined by assessing the effect of COX inhibitors on the expression of KLK7 and KLK8. Upon exposure to sulindac, a broad COX inhibitor, acidity-induced mRNA expression of KLK7 and KLK8 was reduced. In contrast, exposure to selective COX2 inhibitor NS398 had little effect on the level of these peptidases (Fig. 5A-D). Likewise, the effect on cellular invasion at acidic pH was partially reduced by sulindac, but not by NS398 (Fig. 5E and F). This indicated that the COX pathway is involved in acidity-induced KLK7 and KLK8 expression, and that COX1 plays a more important role than COX2 in the expression of KLK7 and KLK8. To confirm this, cells transfected with siRNAs silencing COX1 or COX2 were exposed to acidic culture medium, and the silencing efficiency was confirmed by decreased protein levels (Fig. 6C). Knockdown of COX1, but not COX2 partially reduced the mRNA expression of KLK7 and KLK8 (Fig. 6A and B).