We first measured the expression level of lincRNA-p21 by qRT-PCR in 40 pairs of gastric cancer tissues and adjacent non-cancerous tissues. We first expressed this paired group data as fold changes in the GC tissues relative to the paired normal tissues and then separated the paired tissues into low (n=21, ratio≤0.11) and high (n=19, ratio>0.11) expression groups according to the median ratio (Figure 1A). Our statistical analysis indicated that lincRNA-p21 was significantly down-regulated in tumor tissue samples compared to paired normal tissue (p<0.05) (Figure 1C). Further analysis indicated a significantly lower level of lincRNA-p21 in GC tissues with distant metastasis (p<0.001) (Figure 1D). However, the reduced lincRNA-p21 expression was not significantly correlated with lymph node metastasis (p>0.05) (Figure 1E). The comparison of the clinicopathological characteristics between relative low expression group and high expression group demonstrated that the lower expression level of lincRNA-p21 was correlated with higher invasion depth grade (p=0.024), distant metastasis (p=0.009) and more advanced TNM stage (p=0.011) (Table 1). Next, we performed qRT-PCR to examine the expression of lincRNA-p21 in gastric cancer cell lines. The results showed that lincRNA-p21 was significantly decreased in multiple gastric cancer cell lines compared with the normal gastric mucosal cell line GES-1 (Figure 1B).

In order to study the biological function of lincRNA-p21 in gastric cancers, we performed knockdown and overexpression experiments. The expression level of lincRNA-p21 was significantly down-regulated in the si-lincRNA-p21-transfected cells (p<0.05, Figure 2A). Meanwhile, the expression level of lincRNA-p21 was significantly up-regulated in the cells transfected with a pcDNA 3.1-lincRNA-p21 overexpresssion vector (p<0.01, Figure 2B). CCK-8 assays demonstrated that the knockdown of lincRNA-p21 promoted cell growth in both MGC-803 and MKN-45 cells (Figure 2C) while lincRNA-p21 overexpression showed the opposite result (Figure 2D). Additionally, the Edu assay also showed enhanced cell proliferation ability in the MGC-803 (Figure 2E and 2F) and MKN-45 cells (Figure 2G and 2H) transfected with si-lincRNA-p21. In contrast, the cell proliferation ability of MGC-803 was impaired during lincRNA-p21 overexpression (Figure 2I and 2J).

We next evaluated the effect of down-regulated lincRNA-p21 on the EMT process. We found that the morphology of both MGC-803 and MKN-45 dramatically changed at 48h post-transfection of si-lincRNA-p21. They lost their cell to cell contacts and acquired spindle-like appearance (Figure 3A). Our western blots showed that vimentin and N-cadherin were significantly elevated in si-lincRNA-p21 group (Figure 3B) while lincRNA-p21 overexpression showed an opposite result (Figure 3C). In wound healing assays, we showed that the wound area became narrower in si-lincRNA-p21 groups compared with control groups in both MGC-803 (Figure 3D) and MKN-45 (Figure 3E) cells. Quantitative analysis revealed that the decrease of the wound area was statistically significant at both 24h and 48h post-transfection time point in MGC-803 (Figure 3F) and MKN-45 (Figure 3G) cells. Transwell assays were also performed to test the cell migration and invasion ability. We observed that the knockdown of lincRNA-p21 lead to more cells passing through the chambers with or without the coated matrigel in MGC-803 and MKN-45 cells (Figure 4A and 4C). Further quantitative analysis showed a significantly increased number of cells passing through the chamber in MGC-803 and MKN-45 experiments during knockdown of lincRNA-p21 (Figure 4B and 4D). In lincRNA-p21 overexpression experiments, fewer cells passed through the chambers compare to control groups (Figure 4E) and quantitative analysis supported this finding (Figure 4F).

We tested the expression level of several vital proteins in the GC cells transfected with si-lincRNA-p21. Our result showed up-regulated YAP, β-catenin and NF-κB in the si-lincRNA-p21 group, which is consistent with several recent reports [22, 23], while the expression of P-ERK was somehow decreased (Figure 5A). We found up-regulated YAP and P-YAP level in the lincRNA-p21 knockdown cells (Figure 5B). However, the expression of LATS1, the main upstream activator of YAP cytoplasmic accumulation in Hippo pathway, remained almost unchanged (Figure 5B). Next qRT-PCR assays showed that the down-regulated lincRNA-p21 expression correlated with elevated levels of YAP mRNA, connective tissue growth factor (CTGF) and cysteine-rich angiogenic inducer 61 (CYR-61), in MGC-803 cells (Figure 5C) and MKN-45 cells (Figure 5D). CTGF and CYR-61 are two targets of YAP activation [24]. Moreover, we also found lower levels of YAP and P-YAP expression in the MGC-803 cells transfected with pcDNA 3.1-lincRNA-p21 (Figure 5E). Furthermore, qRT-PCR assays also showed decreased mRNA levels of YAP, CTGF and CYR-61 in MGC-803 cells transfected with pcDNA 3.1-lincRNA-p21 (Figure 5F).

Besides the elevated mRNA of YAP, we wondered whether the elevated YAP protein level was correlated with increased YAP nuclear translocation. We first extracted nuclear and cytoplasmic proteins separately and analyzed them by western blot. We found increased YAP protein levels in both the nucleus and cytoplasm in two GC cell lines transfected with si-lincRNA-p21 (Figure 6A and 6B). Overexpression of linRNA-p21 showed an opposite result (Figure 6C). We also performed immunofluorescence staining experiments to visualize YAP intracellular localization. Staining revealed an obvious accumulation of YAP in the nucleus in si-lincRNA-p21 treated-cells compared to control siRNA-treated cells, in which YAP was mainly cytoplasmic (Figure 6D and 6E). We conclude that lincRNA-p21 may regulate YAP nuclear translocation.

Since YAP plays an important role in tumorigenesis and triggering the EMT process [16, 17], we wondered whether the involvement of lincRNA-p21 in the GC development was related to the YAP expression. The knockdown efficiency of si-YAP was tested by both qRT-PCR (Figure 7A) and western blot (Figure 7B). Western blots indicated that cells transfected with siRNAs targeting both lincRNA-p21 and YAP counteracted the induction of YAP, N-cadherin, Vimentin and C-myc observed in single si-lincRNA-p21 transfection group in both MGC-803 (Figure 7C) and MKN-45 (Figure 7D) cells. Overexpression experiments validated these observations (Figure 7E). In this case, we came to the conclusion that the down-regulation of lincRNA-p21 contributes to GC development, possibly through the upregulation of YAP.

40 pairs of fresh GC tissues and adjacent non-tumor tissues along with the patients’ information were acquired from the tissue bank of West China Hospital, Sichuan University, China. The tissues were stored in liquid nitrogen for further RNA extraction. This study was approved by the Research Ethics Committee of West China Hospital. The patients’ baseline information was listed in Table 1.

Five gastric cancer cell lines (MGC-803, MKN-45, BGC-823, MKN-28, SGC-7901) and a normal gastric mucosal cell line (GES-1) were obtained from the cell depository of our laboratory. These cell lines were cultured in Dulbecco's modified Eagle (DMEM) (GIBCO, USA) supplemented with 10% fetal bovine serum (FBS) (GIBCO, USA) in a humid atmosphere of 5% CO2 at 37%.

The total RNA of patients’ tissue specimen and cells were both extracted using the NucleoZOL reagent (MACHEREY-NAGEL, Germany) in accordance to the protocol. The concentration and quality of RNA was detected by NanoDrop. The extracted RNA was converted to cDNA by using PrimeScript™ RT reagent Kit with gDNA Eraser (Takara, Japan) following the manufacturer's instruction. The cDNA template was mixed with gene specific primers as well as SYBR Green 2X mixture (Bio-Rad, CA, USA). Quantitative real time-polymerase chain reaction (qRT-PCR) was carried out on Bio-Rad CFX96 Touch (Bio-Rad, USA). The primer sequences were listed in Supplementary Table 1. The expression level of target mRNA was calculated by 2^−ΔΔCT values normalized to GAPDH.

The small interfering (si) RNA targeted linc-RNA-p21 was designed as same as the known sequence in a former report [38]. Both the si-lincRNA-p21 and a non-specific si-NC as a negative control were both purchased from GenePharma (GenePharma, China). The sequence of the siRNA was listed in Supplementary Table 1. Plasmid used for overexpression, pcDNA 3.1-lincRNA-p21, was constructed and sequencing confirmed by Sangon Company (Sangong, China). MGC-803 and MKN-45 were seeded in a 6-well plate 24h before transfection at approximately 1×10⁵ cells/per well. Then either si-lincRNA-p21 or si-NC was transfected using Lipofectamine 2000 (Invitrogen, USA) under the guidance of the protocol. The newly constructed plasmid was extracted by EndoFree Plasmid Mini Kit (CWBIO, China) follow the instruction. We used pcDNA 3.1 as the internal control which was directly obtained from the stock of the lab. 48h later, the transfected cells were harvested for further analysis.

Cell proliferation was tested by Cell Counting Kit-8 (CCK-8) (Dojindo, China) following the manufacturer's instructions. MGC-803 and MKN-45 were seeded in a 96-well plate 24h before the transfection of si-lincRNA-p21 or si-NC. 100μl of serum free DMEM contains 10% CCK-8 reagent was added into each well at 0h and 48h post transfection respectively and then cultured for 1h. The absorbance was assessed at a wavelength of 450nm by a microplate reader (Bio-Rad, USA). Experiments were repeated at least three times.

Cell proliferation ability was also tested by EdU Apollo DNA in vitro kit (RIBOBIO, China) according to instructions. MGC-803 and MKN-45 were seeded in a 96-well plate 24h before the transfection of either si-lincRNA-p21 or si-NC. At the 48h time point, 100 μl of 50 μM EdU was added in each well and incubated for 2 h at 37%. After that, the cells were fixed in 100ul of 4% paraformaldehyde (prepared in PBS) for 30 minutes at room temperature. The cells were incubated in 50 μl of 2 mg/ml glycine for 5 minutes followed by PBS washing. Afterwards, the cells were permeabilized with 1% TritonX and reacted with 1X Apollo solution for 30 minutes at the room temperature in dark subsequently. Finally, 40μl of Hoechst solution was added in each well and incubated for 15 minutes at room temperature in dark with PBS washing followed. In the end, the cells were visualized and photographed under a fluorescence microscopy. Experiments were repeated at least three times.

MGC-803 and MKN-45 were seeded in a 6-well plate 24h before the transfection of either si-lincRNA-p21 or si-NC. At the 48h time point, the transfected cells were harvested and starved in serum free DMEM and then placed in the upper chambers with or without matrigel coated in duplicate. The lower chambers were filled with DMEM containing 20% FBS. After wiping off the cells on the upper surface by wet cotton swab, the cells migrated to the lower surface of the chamber were fixed in 4% paraformaldehyde and stained with 0.5% crystal violet solution and then photographed under the microscope. Experiments were repeated at least three times.

MGC-803 and MKN-45 were seeded in a 6-well plate 24h before the transfection of either si-lincRNA-p21 or si-NC. Right after the transfection, a scratch was made through the center of each well using the 200ul sterile pipette tip. The scratch was observed and pictured at 0h, 24h, 48h post-transfection. Experiments were repeated at least three times.

Cellular protein extracts were obtained from cultured cells with RIPA lysis buffer (Bioteke, Beijing, China). The Nuclear and Cytoplasmic Extraction Kit (KeyGEN, Jiangsu, China) was used to prepare cytoplasmic and nuclear extracts separately according to the manufacture's instruction. The protein extracts were separated by 8% SDS-polyacrylamide gels and transferred electrophoretically onto a PVDF membrane using the Bio-Rad semidry transfer system. Membranes were blocked with 5% milk overnight and incubated with primary antibodies (diluted in 1:1000) for 2 hours at room temperature. Second antibodies (diluted in 1:100) were added after washing and incubated for 1 hour in dark. After washing again, the membranes were exposed in a dark room using machine. All of the primary antibodies and secondary antibodies were purchased from the Abcam campany.

MGC-803 and MKN-45 cells were transfected by either si-lincRNA-p21 or si-NC 48hrs before plating into 24-well dishes containing 12 mm glass coverslips. After 24hrs, cells were fixed in 4% paraformaldehyde and then incubated 1 h in blocking buffer (0.5% triton X-100 and 1% BSA in PBS) prior to primary antibody staining. Both primary and secondary antibodies were diluted in blocking buffer. Coverslips were incubated with primary antibodies for 1 h at room temperature followed by 1h second antibodies incubation. In the end, diluted DAPI was added on the coverslips. 10 minutes later, coverslips were mounted on the glass slide. Images were captured under Nikon Eclipse 80i (Nikon, Japan) microscope with NIS-Elements software (version 4.30.01).

The analysis between clinicopathological characteristics and lincRNA-p21 expression was performed using SPSS 22.0 software (IBM, USA). The rest data analysis was performed using Graphpad Prism 5.0. Student t test was used for the comparison between two groups and Chi-square test or Fisher's exact test for analyzing the correlation between lincRNA-p21 expression level and patient's clinicopathological characteristics. Quantitative analysis from pictures was performed by ImageJ v1.04 software (NIH, Bethesda, MD). All of the p values were two-sided and p values less than 0.05 were considered to be statistically significant.