A total of 60 paired samples of fresh frozen GC tissues and corresponding adjacent non-tumor tissue samples (>5 cm away from tumor) were obtained from The Third Affiliated Hospital of Harbin Medical University (Harbin, China) between March 2006 and March 2008. All cases were reviewed by pathologists and histologically confirmed as GC. Patients were not subjected to local or systemic treatment prior to the procedure. The age range of the 60 patients (46 male and 14 female) was 40–76 years, with a median age of 57 years. The study was approved by the Ethics Committee of Southeast University affiliated to Zhongda Hospital (Nanjing, China).

The normal gastric epithelial GES-1 cell line and the GC AGS, MKN-45 and MKN-74 cell lines were purchased from The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences. The cells were cultured in RPMI-1640 medium containing 10% FBS (both Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin/streptomycin at 37°C in a humidified atmosphere with 5% CO₂. The culture medium was changed every 1–2 days and subcultured when the cell confluence reached 80–90%.

Total RNA was extracted from the fresh GC tissues and cell lines with TRIzol^® reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. cDNA was synthesized using PrimeScript@ RT Reagent kit (Takara Bio, Inc.) under standard conditions according to the manufacturer's instructions. RT-qPCR was performed to determine the expression levels of specific genes using a SYBR Premix Ex Taq kit (Takara Bio, Inc.), and β-actin was used as the internal control for normalization of the data. The thermocycling conditions for amplification were as follows: 95°C for 5 min, followed by 40 cycles at 95°C for 30 sec, 60°C for 30 sec and 72°C for 30 sec, and a final step at 72°C for 10 min. All the experiments were performed on a StepOne Plus system (Applied Biosystems; Thermo Fisher Scientific, Inc.), and the sequences of the primers for RT-qPCR are shown in Table SI. Relative gene expression compared with its control was assessed using the 2^−ΔΔCq method (31). Each sample was analyzed in triplicate.

For in vitro assays, the small interfering (si)RNA sequences and non-targeting negative controls (NCs) were synthesized by Shanghai GenePharma Co., Ltd., and were as follows: si-LINC00659-1, 5′-GGGACUUGGAUGCUUAACATT-3′; si-LINC00659-2, 5′-GCCGUGCUCUGGAUAUAUATT-3′; and si-NC, 5′-UUCUCCGAACGUGUCACGUTT-3′. Short hairpin (sh)RNA sequences were designed according to the aforementioned siRNA sequences. Human embryonic kidney 293T cells were used for lentiviral packing (second-generation packaging system). 293T cells were provided by Shanghai GeneChem Co., Ltd., and originally purchased from the American Type Culture Collection. The cells were cultured in 10-cm dishes for 2–3 days until they reached 90–95% confluency. The recombinant virus plasmid GV112 (20 µg), which encoded sh-LINC00659, and the control, together with packaging plasmids pHelper 1.0 (15 µg) and pHelper 2.0 (10 µg), were co-transfected into 293T cells using Lipofectamine^® 2000 (all from Shanghai GeneChem Co., Ltd.). After 48 h of transduction, the lentiviral particles contained in the supernatant of 293T cells were harvested and then concentrated by passing through a 0.45-µm filter.

GC cells (1×10⁵ cells/well) were seeded on a 24-well plate, and cell transfection was performed when the cell number reached ~2×10⁵ cells/well. Fresh medium containing 6 µg/ml polybrene was added for transfection at a multiplicity of infection of 10 with control or sh-LINC00659 lentiviruses (Shanghai GeneChem Co., Ltd.) at 37°C after the cells had been washed 3 times with PBS. After 8 h of incubation at 37°C, fresh medium was added to the cells, and puromycin (2 µg/ml) was added to screen the transfected cells on the 3rd day of transfection. After 7 days, the cells were used for subsequent experimentation.

The infected cells were suspended and seeded in a 96-well plate at a density of 2,000 cell/well. After 0, 24, 48, 72 and 96 h of incubation at 37°C, the culture medium was replaced with 100 µl CCK-8 reagent (Dojindo Molecular Technologies, Inc.). After 4 h of incubation, the cell proliferative ability was determined by measuring the optical density at 450 nm.

The human GC AGS and MKN-74 cell lines, which were stably transfected with sh-LINC00659, were seeded in 6-well plates and reached a density of ~90% after one day. A 200-µl pipette tip was used to scratch the cell culture surface. The cells were washed three times with PBS and were then cultured in serum-free medium. The cells were cultured at 37°C in a humidified atmosphere with 5% CO₂. Wound healing was recorded at 0 and 24 h using an inverted light microscope (Olympus Corporation; magnification, ×200).

The cell invasive and migratory potential was measured using Transwell chambers (EMD Millipore) containing a polycarbonate membrane with 8.0-µm pores. The human GC AGS and MKN-74 cell lines were stably transfected with sh-LINC00659 using a lentivirus system, as aforementioned. After 48 h of transfection, 1.5×10⁴ cells in serum-free medium were placed into the upper chamber of an insert for migration, and the cells were allowed to migrate for 18 h at 37°C. For the invasion assay, 2×10⁴ cells in serum-free medium were seeded in Matrigel-coated (Sigma-Aldrich; Merck KGaA) inserts and allowed to invade for 36 h at 37°C. In both assays, the lower Transwell chamber was filled with medium supplemented with 10% FBS. After incubation, the cells that had migrated or invaded through the membrane were stained with methanol and 0.1% crystal violet for 30 min at room temperature, imaged and counted using an IX71 inverted light microscope (Olympus Corporation; magnification, ×200).

Processed TCGA expression data including tumor (n=408) and normal samples (n=36) of patients with GC were downloaded from TCGA database (http://cancergenome.nih.gov/), and the upper quantile normalized fragments per kilobase per million values were collected and converted into log2 data form. The data were used to analyze the survival of patients with GC based on lncRNA expression with Kaplan-Meier survival curves. The samples were first divided into two groups (high and low LINC00659 groups) according to the median expression levels (FPKM value, 37) of LINC00659. The 5-year survival analysis was performed using the survival package (https://cran.r-project.org/web/packages/survival) in the R software.

To explore the potential downstream pathways of LINC00659, LINC00659 expression was categorized into two groups of high and low expression, according to the median value (FPKM value, 37) of LINC00659 expression. Gene set enrichment analysis (GSEA) 2-2.2.3 (JAVA version) was downloaded from the GSEA website (http://www.gsea-msigdb.org/gsea/downloads.jsp). Subsequently, the downloaded dataset was imported using the GSEA software. Gene sets associated with biological signal transduction were identified on the Molecular Signatures Database (MSigDB; http://software.broadinstitute.org/gsea/msigdb). The simulation was repeated 1,000 times for each analysis, according to the default weighted enrichment statistical method. Gene sets with a false discovery rate <0.25 and P<0.05 were selected.

The correlation between the co-expression of LINC00659 and all protein coding genes (PCGs) was determined by calculating the Pearson correlation coefficients and z test. The PCGs positively or negatively correlated with LINC00659 were considered as LINC00659-associated PCGs (|Pearson correlation coefficient|>0.4 and P<0.01).

All experiments were performed independently at least three times and data are presented as the mean ± SD. Statistical analyses were performed using SPSS 18.0 software (SPSS, Inc.). Comparisons between two groups were made using the independent sample t-test; unpaired Student's t-test was used to analyze the differences between two groups, and paired Student's t-test was used to compare the paired tissue samples. Comparisons among >2 groups were made using ANOVA followed by Fisher's LSD post-hoc test (for 3 groups) or Dunnett's post-hoc test (for >3 groups). Fisher's exact test or χ² test was used to analyze the association between LINC00659 expression and clinicopathological parameters. Survival curves were plotted using the Kaplan-Meier method, and the difference in OS survival rates was assessed using the log-rank test. P<0.05 was considered to indicate a statistically significant difference.

Alteration of chromosomal duplication and gene amplification are crucial for functional gain and overexpression of oncogenes. Based on the CGH analysis, several chromosomal aberrations, especially recurrent gain and amplification of the long arm of chromosome 20 (20q13.33), have been observed in numerous types of cancer, including GC (12–17). In order to explore the contribution of lncRNAs located on amplified 20q13.33 (Fig. 1A), TCGA database was screened, and the expression levels of lncRNAs in GC were analyzed. The expression levels of four lncRNAs, namely RTEL1-TNFRSF6B, SLCO4A1-AS1, ZBTB46-AS1 and LINC00659, were found to be significantly upregulated in gastric tumor tissues compared with in non-tumor tissues (Fig. 1B). Subsequently, Kaplan-Meier analysis using a log-rank test was performed to investigate whether the expression levels of these four lncRNAs were associated with the survival of patients with GC. The results demonstrated that patients with a high level of LINC00659 had a significantly poorer OS rate (P=0.04) compared with patients with a low level of LINC00659 expression (Fig. 1C). However, the other three lncRNAs, RTEL-TNFRSF6B, SLCO4A1-AS1 and ZBTB46-AS1, were not associated with OS rate or tumor stage (Figs. S1 and S2). Additionally, the expression levels of LINC00659 were significantly increased in patients with advanced GC (stage IV) compared with in those with stages I–III (Fig. 1D). These data indicated that patients with GC with increased levels of LINC00659 expression may be prone to progress to a more advanced stage of the disease.

The relative expression levels of LINC00659 were detected in 60 paired GC and adjacent non-cancerous tissues via RT-qPCR. LINC00659 expression was significantly upregulated in the cancerous tissues compared with in their non-cancerous counterparts (Fig. 2A). Furthermore, the association of LINC00659 expression with clinical features of GC was examined, revealing that higher expression levels of LINC00659 were associated with differentiation and vessel invasion; however, no association was observed with sex, age, tumor size, lymph node metastasis and Helicobacter pylori infection (Table SII). Additionally, the expression levels of LINC00659 were higher in cancer tissues than in adjacent non-cancerous tissues in patients with lymph node metastasis (Fig. 2B). Subsequently, LINC00659 expression was detected in three GC cell lines (AGS, MKN74 and MKN-45) and the normal gastric epithelial GES-1 cell line via RT-qPCR. As presented in Fig. 2C, all three GC cell lines exhibited significantly higher expression levels of LINC00659 compared with the normal gastric epithelial cell line. Furthermore, LINC00659 was more highly expressed in the copy number-amplified group than in the normal copy number group (P<0.01; Fig. 2D). This finding indicated that the number of amplifications of LINC00659 may be closely associated with its expression. Overall, these results indicated that upregulation of LINC00659 expression may have important roles in GC development and progression.

To explore the biological function of LINC00659, GSEA analysis based on the mRNA expression levels of LINC00659 in GC was performed using TCGA data. Enrichment plots of gene expression signatures illustrated that epithelial cell migration and epithelial-mesenchymal transition pathways were associated with the expression levels of LINC00659 (Fig. 3A and B). With further analysis, the potential function of LINC00659 was aggregated at a few functional clusters, including calcium signaling pathway, phosphorylation pathway and adhesion-related pathway (Fig. 3C). These data implied that the function of LINC00659 may be associated with cell migration and invasion involved in GC metastasis.

To assess the effects of LINC00659 expression on GC cells, sh-LINC00659 or sh-NC were used to infect AGS and MKN-74 cells, which exhibited higher LINC00659 expression. The efficiency of infection was confirmed by RT-qPCR, and a significant decrease in LINC00659 expression was observed following transfection with both sh-LINC00659 molecules (Fig. 4A), with sh-LINC00659-2 being chosen for subsequent experiments. The CKK-8 assay revealed that LINC00659 inhibition by sh-LINC00659-2 did not affect the proliferation of AGS and MKN-74 cells (Fig. S3).

In order to clarify the function of LINC00659 in terms of cell migration and invasion, the effects of LINC00659-knockdown on cell migration and invasion were assessed. The results of the wound-healing assays revealed that the migratory ability of AGS and MKN-74 cells was decreased in the LINC00659-knockdown group compared with in the NC group (Fig. 4B). Transwell migration assays displayed that LINC00659-knockdown significantly suppressed the migratory capabilities of both AGS and MKN-74 cancer cells compared with the control group (Fig. 4C and E). Moreover, GC cells with LINC00659-knockdown exhibited significantly weaker abilities to invade through Matrigel compared with the control group (Fig. 4D and F). These results indicated that knocking down LINC00659 led to a clear suppression of the migratory and invasive abilities of GC cells.

To investigate the underlying mechanism of LINC00659 in cell migration and invasion, the correlation between the expression levels of LINC00659 and all PCGs was examined using two-sided Pearson correlation coefficients and z-test. The PCGs positively or negatively correlated with LINC00659 were considered as LINC00659-associated PCGs (|Pearson correlation coefficient|>0.4 and P<0.01). The expression levels of eight genes, namely IQGAP3, MMP15, integrin subunit α6 (ITGA6), G protein-coupled receptor class C group 5 member A (GPRC5A), hematological and neurological expressed 1 (HN1), keratin 8 (KRT8), engulfment and cell motility 3 (ELMO3), and TNF receptor associated factor 4 (TRAF4), were found to be positively correlated with LINC00659 expression (Fig. 5A-H). To determine whether these eight genes were downstream targets of LINC00659, their expression patterns were investigated in GC cells where LINC00659 expression had been knocked down by sh-LINC00659-2. As shown in Fig. 6A and B, the expression levels of IQGAP3 and MMP15 were significantly downregulated in LINC00659-knockdown GC cells, and ITGA6 expression was significantly downregulated only in AGS cells. GPRC5A, HN1, KRT8, ELMO3 and TRAF4 exhibited no change in expression when LINC00659 was knocked down in AGS and MKN-74 cells (Fig. 6C). These results suggested that LINC00659 may be associated with metastasis via IQGAP3 and MMP15 as its downstream targets.