A total of 16 GC cell lines, namely SNU001, SNU005, SNU016, SNU216, SNU484, SNU520, SNU601, SNU620, SNU638, SNU668, SNU719, AGS, KATOIII, MKN01, MKN45 and MKN74, were obtained from the Korean Cell Line Bank (cellbank.snu.ac.kr/main/index.html) and cultured in RPMI-1640 medium (Welgene, Inc.) supplemented with 10% fetal bovine serum and 1% antibiotic-antimycotic solution (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C in a humidified 5% CO₂ incubator. All experiments were performed with mycoplasma-free cells and all GC cell lines were authenticated using short tandem repeat profiling within the past 3 years. A total of 145 frozen tumor and paired adjacent non-tumor tissue samples were obtained from the Chungnam National University Hospital (CNUH; Daejeon, South Korea), a member of the Korea Biobank Network. The samples included 31 stage I, 30 stage II, 46 stage III and 38 stage IV tumors from 50 females and 95 males, aged 23–83 years (average age, 59 years). All samples were obtained with informed consent and their use was approved by the Internal Review Board of CNUH (approval no. CNUH 2018-01-056).

Bisulfite sequencing was performed on two different regions proximal to the STK31 promoter. Two sets of PCR primer were designed using Methprimer (urogene.org/cgi-bin/methprimer/methprimer.cgi): Region 1, forward, 5′-TCTAGAAAATGCAAACTAATATATGGTGAC-3′ and reverse, 5′-GAAAGGAACTGGCCCTAGACCCCAC-3′ to yield a 581-bp product containing 9 CpG sites and Region 2, forward, 5′-GAAAGGAATTGGTTTTAGATTTTAT-3′ and reverse, 5′-TCTAACACCCCTCTAAAATAAC-3′ to yield a 468-bp product containing 36 CpG sites. Genomic DNA (2 µg) from GC cells or tissue was modified by sodium bisulfite using an EZ DNA methylation kit (Zymo Research Corp.). Bisulfite-modified DNA (1 µl) was amplified in a 20 µl volume using 2X Dye Mix polymerase (Enzynomics Co., Ltd.) and the aforementioned primers. Samples were heated to 95°C for 10 min and amplified for 40 cycles of 95°C for 45 sec, 60°C for 45 sec and 72°C for 60 sec, then incubated at 72°C for 10 min and cooled to 4°C. The PCR products were visualized on a 1% agarose gel by ethidium bromide staining, purified from the gel using a Qiagen Gel Extraction kit (Qiagen, Inc.) and cloned using the pGEM-T Easy Vector (Promega Corporation). A total of ten clones were randomly chosen for sequencing. Complete bisulfite conversion was assured when <0.01% of cytosines in non-CG dinucleotides in the final sequence had not converted.

Total RNA was extracted from GC cells or tissue using the RNeasy plus mini kit (Qiagen), treated with DNase I (Promega Corporation), and reverse-transcribed with Superscript II reverse transcriptase (Invitrogen; Thermo Fisher Scientific, Inc.). RT-qPCR for STK31 was performed with a CFX96 Real-Time PCR Detection system (Bio-Rad Laboratories, Inc.) using the following thermocycling conditions: 94°C for 5 min, followed by 35 cycles of 94°C for 30 sec, 64°C for 30 sec and 72°C for 30 sec, with a final step of 72°C for 7 min. β-actin served as the PCR control. The PCR products were analyzed on 1.5% agarose gels stained with ethidium bromide. The primer sequences for RT-qPCR were as follows: STK31 forward, 5′-AACCTGCTTCTCCAGGTTCA-3′ and reverse, 5′-ATTGCTCCTTTGGCATCAAG-3′ to yield a 228-bp product and β-actin forward, 5′-CAAGAGATGGCCACGGCTGCT-3′ and reverse, 5′-TCCTTCTGCATCCTGTCGGCA-3′ to yield a 275-bp product. RT-qPCR for STK31 was performed using a C1000 Thermal Cycler (Bio-Rad Laboratories, Inc.). cDNA (100 ng) was amplified as aforementioned for 45 cycles with 2X SYBR-Green Supermix (Bio-Rad Laboratories, Inc.). β-actin was used as a control. The relative amount of target mRNA was quantified using comparative threshold cycle (Cq) method (16).

Two CpG sites, namely CpG#23 and #24, in Region 2 were selected for quantification of the extent of methylation. Bisulfite-modified DNA (100 ng) was amplified by PCR in a 20 µl reaction using 2X Dye Mix polymerase (Enzynomics Co., Ltd.) to yield a 221-bp product using the following primers: Forward, 5′-TGTTTGGGGGTAGGTAGTAGTTAG-3 and reverse, 5′-CCCTAAACCCACATACTAAACTTTC-3′. PCR was performed using an initial melting step of 94°C for 5 min, followed by 40 cycles of 94°C for 30 sec, 59°C for 30 sec and 72°C for 30 sec, with a final incubation at 72°C for 7 min. Pyrosequencing was performed as previously described (17) using a sequencing primer (5′-AGGAGTAGTGTGGGGTTT-3′) and PSQ HS 96A system (Biotage AB).

SNU638, SNU719 and MKN74 cells were maintained as aforementioned. Cells were seeded in 100-mm dishes at a density of 1×10⁶ cells per dish, then treated with 10 µM DNA methylation inhibitor 5-aza-dC (Sigma-Aldrich; Merck KGaA) every 24 h for 3 days and harvested. Another group of these cells at the same cell density was treated with 0.5 µM histone deacetylase inhibitor TSA (Sigma-Aldrich; Merck KGaA) for 3 days and harvested. In order to test the combined effect of 5-aza-dC and TSA, cells were treated with 10 µM 5-aza-dC every 24 h for 3 days and then with 0.5 µM TSA for 1 day. After 2–5 days, cells were washed with phosphate-buffered saline, and total RNA was extracted using an RNeasy Mini kit (Qiagen, Inc.). All experiments were performed at 37°C. RT-qPCR for STK31 was performed as aforementioned. A total of three independent experiments was performed.

STK31-knockdown (KD) cells were established using TRCN0000368917 and TRCN0000003276 (STK31_sh#1 and STK31_sh#4, respectively; Sigma-Aldrich; Merck KGaA) targeting STK31 mRNA; pLKO.1-puro (Sigma-Aldrich; Merck KGaA) was used as a control. For lentivirus construction, 293T cells were obtained from Koram Biotech Corp. and co-transfected with 2 µg MISSION Lentiviral Packaging Mix and 2 µg control or STK31 short hairpin (sh) RNA using a 2nd Generation lentiviral system and Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.). In order to establish STK31-expressing cell lines, 2 µg full-length STK31 cDNA was cloned into vector pCDH-CMV-MCS-EF1-Puro (System Biosciences, LLC). For lentivirus construction, 293T cells were co-transfected using a 2nd generation lentiviral system with 2 µg MISSION Lentiviral Packaging Mix plus empty vector or STK31-expressing vector using Lipofectamine 2000. After 48 and 72 h incubation at 37°C, 5% CO₂, supernatant containing the lentivirus was collected from 293T cells and centrifuged at 250 × g for 2 min at room temperature, then filtered and applied to target cells. For lentiviral infection, cells (3×10⁵) for KD experiments or ectopic expression were seeded onto 6-well culture plates before addition of viral supernatant. After 72 h, the medium was changed to RPMI-1640 medium containing 1 µg/ml puromycin (Sigma-Aldrich; Merck KGaA). After 2 weeks of puromycin selection, the change in expression levels was confirmed by RT-qPCR (as aforementioned) and western blotting.

For protein extraction from GC cell lines, RIPA lysis buffer (Invitrogen; Thermo Fisher Scientific, Inc.) with protease inhibitor cocktail (Sigma-Aldrich; Merck KGaA) was used and then concentration was quantified via Bradford Protein Assay (Bio-Rad Laboratories, Inc.) Protein samples (50 µg/lane) were loaded onto 10% acrylamide gel. Electrophoresis was performed using a Bio-Rad Western Blotting system. (Bio-Rad Laboratories, Inc.) Proteins were transferred to a polyvinylidene fluoride membrane (Sigma-Aldrich; Merck KGaA) and blocked in 5% skimmed milk in Tris-buffered saline (0.1% Tween-20) for 30 min at room temperature. The membranes were incubated with primary antibodies (1:1,000) at 4°C overnight. The antibodies were as follows: Anti-STK31 (cat. no. ab155172; Abcam), anti-Caspase3 (cat. no. 9662; Cell Signaling Technology, Inc.), anti-PARP (cat. no. 9542; Cell Signaling Technology, Inc.) and anti-Tubulin (cat. no. T5168; Sigma-Aldrich; Merck KGaA), then Probed with mouse anti-rabbit IgG conjugated with horseradish peroxidase (1:5,000; cat. no. sc-2357; Santa Cruz Biotechnology, Inc.). Immunopositive bands were visualized using an enhanced luminescence image analyzer LAS-4000 (FUJIFILM Wako Pure Chemical Corporation) and the intensity for each band was estimated by Image J version 1 software (National Institutes of Health).

For cell proliferation assays, 1×10³ cells were plated in a 96-well plate and proliferation was measured with the EZ-Cytox Cell Viability Assay kit (Itsbio) using a microplate reader (Molecular Devices, LLC) at 450 nm. For colony-forming assays, 1×10³ cells were plated in a 6-well plate. RPMI-1640 Media (Welgene Inc.) were replaced every 3 days then cells were incubated at 37°C. After 2 weeks, colonies were stained with crystal violet solution (0.5 crystal violet, 3.7 formaldehyde, 30.0% ethanol) for 2 h at room temperature, and the number of viable cells was manually counted in each well. All assays were performed in triplicate.

Transwell migration assays were performed in a 24-well Transwell chamber (Corning, Inc.) fitted with a polycarbonate membrane (pore size, 8 mm). Cells were suspended in 100 µl serum-free RPMI-1640 medium (Welgene, Inc.), and 2×10⁴ cells were seeded in the upper chamber. The lower chamber was filled with RPMI-1640 medium containing 10% fetal bovine serum. After 16–24 h incubated at 37°C, migrated cells were stained for 2 h at room temperature with 0.5% crystal violet solution. A total of three independent fields of view were observed using a fluorescence microscope (magnification, ×10) for each membrane, and migrated cells were manually counted in each field.

MKN1 cells were treated 0.05% trypsin for 3 min at 37°C and then harvested at a density of 2×10⁶ per ml. Cells were washed with ice-cold PBS and fixed with 70% ethanol for 24 h at 4°C. Prior to analysis, cells were stained with 50 µg/ml propidium iodide solution for 2 min at room temperature. Cell cycle analysis was performed using a flow cytometer (FacsCalibur; BD Biosciences) with blue laser 488 nm and FlowJo 10.7.1 and Cell Quest Software (BD BioSciences).

Mobility of STK31-expressing MKN74 cells was measured using gap closure assay (Ibidi Gmbh). Cell suspension at a density of 1×10⁶ per ml in 70 µl volume were seeded in Cultured-Inserts 2-well (diameter, 35 mm). Following incubation at 37°C for 24 h, the Culture-Insert was detached from the well using forceps and filled with serum-free RPMI-1640 medium. Cell images were captured after 24, 48, 72 and 96 h under a fluorescence microscope (magnification, ×10). The area of gap closure in three fields of view was calculated with ImageJ software (v.1.8.0; National Institutes of Health).

A total of 10 female BALB/C nude mice (age, 4 weeks; weight, 20±2 g) were purchased from Central Animal Laboratory (Shizuoka, Japan) and maintained with regular mouse chow and water at a constant temperature of 22±1°C and 50–60% humidity under a 12-h light/dark cycle and specific pathogen-free conditions for the xenograft assay. Mice were anesthetized before injection using isoflurane (3–4% oxygen) to minimize pain. After 1 week, parental or STK31-expressing MKN74 cells (5×10⁶ cells per mouse) were subcutaneously injected into nude mice. Mice were monitored daily to check sickness and feed daily and to determine weight loss >20% twice a week. Once palpable tumors devloped, tumor size was measured using Vernier calipers and tumor volume was calculated according to the following formula: Volume (mm³)=width² × length/2. All mice were euthanized using 30–70% volume/min CO₂ in chamber at day 53. The humane endpoints were when the largest tumor size exceeded 20 mm or feed intake or drinking water were affected due to necrosis, infection or ulcer. None of the mice died before endpoints of the study. Experiments using mouse were conducted under the Institutional Animal Care and Use Committee-approved protocols at KRIBB in accordance with institutional guidelines (approval no. KRIBB-AEC-15102).

The 450K HumanMethylation BeadChip data (accession number GSE103186) for 39 gastric mucosae and 76 IM tissue samples from GC-free patients (18) were downloaded to compare methylation status in the STK31 promoter. Gene expression and 450K HumanMethylation BeadChip data for 230 primary GCs and 450K HumanMethylation BeadChip data for two normal tissue samples (19) were downloaded from The Cancer Genome Atlas (TCGA) portal (portal.gdc.cancer.gov/) and the National Center for Biotechnology Information Gene Expression Omnibus (ncbi.nlm.nih.gov/geo/). Because methylation data for normal tissue are limited in the TCGA database, additional data for 10 gastric mucosa samples were obtained public data [accession nos. GSE50192 (n=4) and GSE31848 (n=6)] (20,21) produced by the same platform. In order to assess the activation status of the STK31 promoter in primary GCs, public data (accession no. GSE51776) for histone modifications produced by nano-scale chromatin immunoprecipitation-sequencing (Nano-ChIP-seq) of paired GC and non-tumor tissue was downloaded (22). The downloaded sequence reads were quality controlled using cut-adapt (v1.1) on a public site (github.com/marcelm/cutadapt/). Quality-control parameters were read for quality >30 (Phred score), read length >20 bp and replicated level <40%. The sequence reads were mapped with bowtie2 (v2.2.2) using default parameters (m=1). Duplicate reads were removed using the picard (v.2.18.14) markduplicates function. Unique reads were identified by peak calling using macs2 (v.2.0.5) with default parameters. For predicting patient outcomes, gene expression data (accession no. GSE26253) for GC (n=432) in the Samsung Medical Center cohort (SMC) (23) was downloaded from the Gene Expression Omnibus database. The samples included 68 stage I, 167 stage II, 130 stage III and 66 stage IV tumors and one non-information tumor, taken from 152 females and 280 males (average age, 52 years). All samples were obtained with written informed consent, and their use was approved by the Internal Review Board of the SMC, Seoul, South Korea (approval no. SMC 2010-10-025).

A paired t-test was used to examine differences in mRNA levels and methylation between paired gastric tumor and adjacent non-tumor tissues. Values are expressed as the mean ± SD of ≥3 independent repeats.. Correlations between SKT31 expression levels and CpG methylation was determined using Pearson's correlation coefficient. For the comparison of multiple groups of GC type from TCGA data, pairwise P-values were calculated using the Wilcoxon rank-sum the Student's t-test, respectively, and were corrected for multiple comparison by Bonferroni method (n=3). For comparison of of STK31 mRNA levels in GC cells following drug treatment, pairwise P-values were calculated using paired Student's t-test and was corrected for multiple comparison by Bonferroni method (n=3). For in vitro assays of STK31 expression levels in STK31-KD cells, pairwise P-values were calculated using paired Student's t-test and were corrected for multiple comparison by Bonferroni method (n=2). Kaplan-Meier survival analysis with log-rank test was used to estimate the difference in overall survival between STK31 high and low expression groups in combined cohorts of CNUH and SMC. All statistical analysis was performed using the R statistical programming language (Version 4.0.2; r-project.org/).

Our previous study (8) identified 174 hypomethylated promoters in GC cells (91 GC-specific and 83 early-onset) via methylome analysis with laser-capture microdissected cells of a single patient with intestinal-type GC (IGC). Early-onset hypomethylation was defined as a methylation difference >2-fold in IM and GC compared with gastric mucosa cells. STK31 was an early-onset hypomethylated targets (Fig. 1B). RRBS data from the UCSC Genome Browser (hg19) revealed that CpG methylation signatures (purple vertical lines) at the STK31 promoter were predominant in GM cells but mostly absent in IM and completely absent in GC (Fig. 1B).

RT-qPCR analysis was performed with four paired gastric tumor and adjacent non-tumor tissues, revealing that STK31 was silenced in non-tumors but expressed in all tumor samples tested (Fig. 2A). Bisulfite sequencing analysis with three of four paired clinical tissues revealed 51.4–69.4% methylation in Region 1 of non-tumors but 16.7–30.3% in paired tumors (Fig. 2B). For Region 2, the difference was also significant (non-tumor, 92.7–96.4%; tumor, 64.8–86.1%; Fig. 2B), although not as large as for Region 1. These data suggested that STK31 expression levels in tumors may be a consequence of CpG hypomethylation at the promoter comprising Regions 1 and 2.

STK31 expression levels and promoter methylation were analyzed using 145 paired clinical tissue samples of the CNUH cohort. RT-qPCR analysis of STK31 revealed significant upregulation in tumor (16.70±3.60%) compared with non-tumor samples (13.29±5.49%; Fig. 2C; P=0.02). A a significant increase in expression was defined as >2-fold with respect to the values for tumors compared with paired non-tumors; increased STK31 expression was apparent in approximately half (72/145) of tumors. Methylation at two CpG sites, namely CpG#23 and #24, within Region 2 was quantified by pyrosequencing (Fig. 1C) of 120 paired clinical tissues and compared with the corresponding RT-qPCR data. Pyrosequencing of these two sites revealed 92.1±6.4% methylation in non-tumors and 79.7±13.0% in tumors, the difference for which was significant (Fig. 2D; P<2.2×10⁻¹⁶). Finally, there was a negative correlation between methylation at the two CpG sites and STK31 expression in 120 matched tumors (Fig. 2E; r=−0.34; P=1.1×10⁻⁴).

From the public data for 450K HumanMethylation BeadChip of the TCGA, methylation status between GC and non-tumors at CpG sites proximal to the STK31 promoter was compared; the promoter and upstream region were hypomethylated in primary GC, whereas there was no difference in CpG methylation on the gene body (Fig. 3A). For example, methylation at cg05000488 within the STK31 promoter was significantly decreased in GC, especially in IGC (Fig. 3B), whereas STK31 mRNA expression levels were significantly increased in GC (Fig. 3C), revealing a negative correlation between CpG methylation and STK31 mRNA expression levels (Fig. 3D). In addition, methylation status was compared at the same CpG sites proximal to the STK31 promoter from the public data (GSE103186) for 450K BeadChip of gastric mucosae and IM tissue from GC-free patients (17). Methylation at cg05000488 within the STK31 promoter was significantly decreased in IM compared with that in gastric mucosae (Fig. 3E).

In order to investigate the association between STK31 expression and methylation of its promoter in GC cell lines, RT-qPCR and bisulfite sequencing analysis of GC cell lines were performed. Based on the RT-qPCR results, cell lines were divided into two groups according to the median relative STK31 expression level. Fig. 4A and B); the STK31-expressing group (+; >median value) included SNU484, SNU520, KATOIII and MKN1 and the weakly/non-expressing group (−; <median value) included SNU001, SNU638, SNU719 and MKN74. For Region 1, bisulfite sequencing revealed that members of the STK31 (+) group, except for SNU484, exhibited low mean methylation (12.1–69.1%), whereas the STK31 (−) group had relatively high methylation (81.2–92.6%; Fig. 4B). In Region 2, however, members of both groups were highly methylated (86.8–99.1%; Fig. 4B). The association between methylation status at two CpG sites (CpG#23 and #24) and STK31 expression was examined in both groups; methylation status tended to be decreased in STK31 (+) group compared with the STK31 (−) group and STK31 expression levels tended to be increased in STK31 (+) group compared with the STK31 (−) group (Fig. 4C) in both bisulfite sequencing and pyrosequencing analyses. In order to determine whether STK31 expression is controlled epigenetically, cells were treated with 5-aza-dC and/or TSA. Treatment with 5-aza-dC significantly restored STK31 expression levels in SNU-638, SNU-719, and MKN74 cells (Fig. 4D). Treatment of GC cells with both 5-aza-dC and TSA also restored STK31 expression levels (Fig. 4D), suggesting that its expression in GC cells may be regulated epigenetically.

From the public data (GSE51776), unique reads were identified by peak calling using macs2 with default parameters. This yielded peak regions for H3K4me3 (chromatin mark for active promoters), H3K4me1 (chromatin mark for active enhancers and promoters) and H3K27ac (chromatin mark for active regulatory elements), for which five paired normal and gastric tumor tissue samples were merged. Then, signatures for promoter activity were examined near the STK31 promoter. Gain of STK31 promoter activity (increased H3K4me3 and H3K27ac) was evident in primary GC but not in normal mucosae (Fig. 5).

It was next investigated whether STK31 expression in SNU484 or MKN1 cells, in which STK31 was highly expressed (Fig. 6A), could be knocked down by two shRNAs. RT-qPCR and western blotting analysis confirmed that STK31 expression levels significantly decreased in STK31-KD SNU484 and MKN01 cells (Fig. 6A). Each shRNA significantly decreased colony formation (Fig. 6B), proliferation (Fig. 6C) and migration (Fig. 6D) of both STK31-KD cell lines compared with cells treated with control shRNA. In order to investigate the underlying mechanisms for the cell proliferation of STK31 suppression, the effect of STK31 silencing on cell cycle progression in MKN1 cells was assessed by flow cytometry. STK31-KD MKN1 cells was increased in the G₀/G₁ phase but decreased in the G₂/M phase compared with the control (Fig. 6E). Next, the expression levels of apoptosis regulatory proteins such as Caspase3 and PARP, were assessed. Western blotting revealed that Cleaved caspase-3 and Cleaved PARP were increased in STK31-KD cells compared with the control (Fig. 6F). These results suggest that shRNAs targeting STK31 mRNA modulated the oncogenic potential of STK31 by inducing G₁ arrest and apoptosis in GC cells.

The effect of ectopic STK31 expression was assessed in MKN74 or AGS cells, in which STK31 was repressed or weakly expressed (Fig. 7A). RT-qPCR and western blotting confirmed that STK31 mRNA was stably expressed in STK31-transfected MKN74 (STK31-MKN74) and AGS (STK31-AGS) cells (Fig. 7A). Moreover, ectopic STK31 expression significantly induced colony formation in each of STK31-MKN74 and STK31-AGS cells (Fig. 7B) compared with control cells. Ectopic STK31 expression also significantly induced the proliferation (Fig. 7C) and cell migration (Fig. 7D-F) in both STK31-transfected cell lines. Finally, in vivo experiments revealed that tumors in mice engrafted with STK31-MKN74 cells were significantly larger than those of control cells (Fig. 7G and H), demonstrating that tumorigenicity was promoted by ectopic STK31 expression in a xenograft mouse model.

In order to assess the prognostic value of STK31, clinical data from the CNUH (n=145) and SMC (n=432) cohorts were combined to improve predictability. Using the mean cut-off risk score, each cohort was divided into two groups based on the mean expression value of STK31, and the upper and lower groups were combined and analyzed. Kaplan-Meier survival analysis revealed a significant difference in survival rate between the two groups in the combined cohort (Fig. 8; log-rank test; P=0.02), indicating that patient outcome was significantly poorer in the STK31 high expression group compared with the STK31 low expression group.