The human gastric cancer cell lines, MGC-803 and SGC-7901 were purchased from China Academia Sinica (Shanghai, China). All cells were cultured in DMEM and RPMI-1640 medium supplemented with 10% FBS (HyClone, Logan, UT, USA) and were maintained in a humidified atmosphere with 5% CO₂ at 37˚C.

Lentiviruses containing the HOTAIR overexpression sequence (Lenti-HOTAIR), a HOTAIR inhibitor sequence (Lenti-HOTAIR si) and a negative control sequence (Lenti-NC and NC) were obtained from Shanghai GenePharma (Shanghai, China). For the transient transfection of cells, the tumor cells were transfected with negative control siRNA oligonucleotides against EZH2 and SUZ12 using X-treme GENE siRNA transfection reagent (Roche, Shanghai, China) according to the manufacturer’s instructions. The siRNA sequences were as follows: si-EZH2, 5′GUCGCAACGGACCAGUUAA-3′; and si-SUZ12, 5′-GCTTACGTTTACTGGTTTCTT-3′. Subsequent experiments were then carried out 48 h following transfection.

RIPA lysis buffer was used to extract the total protein, and the Nuclear and Cytoplasmic Protein Extraction kit (Beyotime Biotechnology, Shanghai, China) was used to extract the nuclear protein, according to the manufacturer’s instructions. Equal amounts of protein per lane (30 µg) were separated by 10 and 15% SDS-polyacrylamide gel and transferred onto PVDF membranes. The membranes were blocked in 5% skim milk at 37˚C for 1 h and then incubated with a specific antibody at 4˚C overnight. The antibodies used in this study were as follows: H3K4me3 (ab8580), H3K4ac (ab240195), H3K27me2 (ab24684), H3K27me3 (ab6002), HEK27ac (ab45173) and β-catenin (ab16051) antibodies (Abcam, Cambridge, UK; 1 µg/ml); EZH2 (5246), SUZ12 (3737), CBP (7389), E-cadherin (14472), N-cadherin (13116), snail (3879), slug (9585), twist (46702), H3K4me1 (5326), H3K4me2 (9725) and H3K27me1 (84932) (Cell Signaling Technology, Danvers, MA, USA; 1 µg/ml); H3 (AF0009; Beyotime Biotechnology; 0.2 µg/ml); GAPDH (TA-08; Zhongshan Bio Corporation, Beijing, China; 0.2 µg/ml). They were then incubated with their respective secondary antibodies (ZB-2305, ZB2306; Zhongshan Bio Corporation; 1:2,000 dilution). The membranes were stripped, and the bands were visualized using the SuperSignal protein detection kit (Pierce/Thermo Fisher Scientific, Waltham, MA, USA) following normalization with the density of GAPDH and H3. ImageJ 1.41 software (National Institutes of Health, Bethesda, MD, USA) was employed to quantify protein expression.

TRIzol reagent (Invitrogen/Thermo Fisher Scientific) was used to extract the RNA from the gastric cancer cells. Equal amounts of RNA were converted into cDNA using HOTAIR and CDH1 primers by using a PrimeScript RT Reagent kit (Takara Bio, Inc., Otsu, Japan). Quantitative PCR (qPCR) was performed using a Reverse Transcription System (Promega, Madison, WI, USA) on a CFX96 Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) according to the instructions of the manufacturer. Each sample was analyzed in triplicate. The primers used were as follows: HOTAIR forward, 5′-AAATATGGCGGCGTCTACACGGA-3′ and reverse, 5′-TCCAGAACCCTCTGACATTTGCCT-3′; CDH1 forward, 5′-ACGCCGAGAGCTACACGTTC-3′ and reverse, 5′-AGGCTGTCCTTTGTCGACCG-3′; GAPDH forward, 5′-CCGGGAAACTGTGGCGTGATGG-3′ and reverse, 5′-AGGTGGAGGAGTGGGTGTCGCTGTT-3′. Relative expression was calculated using the 2^−ΔΔCq method (25). The PCR conditions were as follows: 35 cycles of 95˚C for 30 sec, 56˚C for 30 sec, 72˚C for 90 sec, and a final extension at 72˚C for 5 min.

The MGC-803 cells were transfected with the Lenti-HOTAIR and Lenti-HOTAIR si, NC and Lenti-NC as a control to detect the amounts of H3K27me3/ac in the nuclei, and images were captured using a microscope (Olympus IX81; Olympus Corporation). The antibodies used were as follows: H3K27me3 (ab6002; 5 µg/ml), H3K27ac (ab45173; 5 µg/ml) (both from Abcam) and E-cadherin (14472; Cell Signaling Technology; 5 µg/ml). DAPI was used to dye the nuclei into blue.

The transfected and control cells were plated in 96-well plates at a cell density of 10⁴ cells/well with 6 replicate wells. The cells were incubated at 37°C for an additional 4 h by adding 10 µl of CCK-8 (5 g/l; Sigma, St. Louis, MO, USA) into each well. The optical density (OD) was measured at 450 nm at 0, 24, 48, 72, 96 and 120 h, and the data were expressed as a percentage of the control, which are derived from triplicate samples.

Transwell invasion assays were performed using Matrigel-coated Transwell membranes (BD Biosciences, San Jose, CA, USA). The Matrigel solidified after a 30-min incubation at 37°C and post-transfection, the cells (10⁵) in 100 µl of serum-free RPMI-1640 medium were added to the upper compartment of the chamber. As a chemoattractant, 20% FBS in RPMI-1640 medium was added to the lower chambers of the wells. The cells were then incubated at 37°C in 5% CO₂ for 48 h, and the lower cells were stained with crystal violet (Zhongshan Bio Corporation) for 5 min at room temperature. The average number of cells per field was calculated by counting cells in 6 random fields per well under the microscope at ×200 magnification (Olympus Corporation).

To determine whether H3K27me3/ac binds to the E-cadherin promoter, ChIP was performed using the EZ-ChIP kit (Millipore, Billerica, MA, USA) according to the manufacturer’s instructions. Cross-linked chromatin was extracted from the gastric cancer cells by adding ChIP lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl pH 8.0) and sonicated into 200-1,000 fragments. The chromatin was immunoprecipitated by the addition of anti-H3K27ac (ab45173, 1 µg/ml) and anti-H3K27me3 (ab6002, 1 µg/ml) antibodies, and a negative control IgG (ab171870, 1 µg/ml) (both from Abcam), which was incubated overnight at 4°C with Protein G agarose. The Protein G agarose-antibody/chromatin complex was resuspended in wash buffer and centrifuged (70 × g, 3 min) to collect the protein/DNA complex, which were reversed to obtain free DNA. qPCR was conducted according to the method described above. Input sample represents the amount of chromatin used in the ChIP, the results of each antibody sample represent the percent of Input. The primers used were as follows: CDH1a forward, 5′-CCCTATGACTGCACCCCTTG-3′ and reverse, 5′-GCAGATAACAGACCCCGCAT-3′; CDH1b forward, 5′-CAAATGGCCCTGGGAGGTTC-3′ and reverse, 5′AGCCCTCTTGGCCCAATACA-3′.

The ENCODE Histone Modification Tracks embedded in the UCSC Genome Browser (University of California Santa Cruz, Santa Cruz, CA, USA) was used to find the promoter.

A total of 12 BALB/c-A nude female mice at 6 weeks of age and weighing 8 g were purchased from the Animal Center of the Cancer Institute, Chinese Academy of Medical Science and bred at Tianjin Institute of Biological Engineering. The maintenance conditions for the mice were as follows: Temperature, 26-28°C; humidity, 40-60%; the food and water were bacteria-free without any special requirements. The mice were randomly assigned to 2 groups (6 mice a group), and were separately subcutaneously injected with 10⁷ MGC-803 cells transfected with Lenti-NC and Lenti-HOTAIR si in 100 µl of PBS. The tumor volumes and the weight of the mice were measured every 2 days, and the tumor weight was surveyed at the endpoint of the study. At 25 days post-tumor implantation, the xenograft samples were collected and subjected to immunohistochemistry (IHC). Animal experiments were approved by the Ethics Committee of Tianjin Medical University General Hospital.

For staining, 5-µm-thick sections were cut and dried, deparaffinized, and immunostaining was performed on the sections using the avidin-biotin-complex method. Primary antibodies specific for H3K27me3 (ab6002), H3K27ac (ab45173) (Abcam; 1:100 dilution) and E-cadherin (14472) (Cell Signaling Technology; 1:50 dilution) were diluted in PBS with 0.1% Tween-20 and incubated overnight at 4°C. A secondary antibody (PV9003; Zhongshan Bio Corp.; 1:100 dilution) was added to the slides at 37°C for 1 h, followed by avidin-biotin complex solution for an additional 1 h. Protein expression was detected by coloration with diaminobenzidine (DAB) buffer for 5 min, and the sections were counterstained with hematoxylin for 5 min at room temperature. All images were captured using a microscope (Olympus Corporation).

All statistical analyses were performed using SPSS version 16.0 (SPSS, Inc., Chicago, IL, USA) or GraphPad 5.0 (GraphPad Software, Inc., La Jolla, CA, USA) software. Data are expressed as the means ± standard deviation; the experiments were repeated 3 times. The statistical significance of differences between groups was determined using a Student’s t-test, and one-way ANOVA was used to determine the differences among at least 2 groups. P-values <0.05 were considered to indicate statistically significant differences.

Based previous results regarding HOTAIR (23), in this study, we examined the effect of HOTAIR on the malignant properties of GC cells. HOTAIR expression was significantly knocked down or overexpressed following lentiviral infection detected by qPCR (Fig. 1A). The results of Transwell assays revealed that the inhibition of HOTAIR decreased MGC-803 and SGC-7901 cell invasion (Fig. 1B). In addition, the CCK-8 assay was used to observe the proliferation of MGC-803 and SGC-7901 cells following HOTAIR knockdown in vitro. We estimated the proliferation rate of the GC cells at 0, 24, 48, 72, 96 and 120 h and compared this with that of the negative control. The results indicated that HOTAIR knockdown suppressed the growth of the GC cells (Fig. 1C). Taken together, these results indicate that HOTAIR is associated with the malignant properties of GC cells.

Since the malignant properties of cancer cells are associated with EMT (2), we examined the effects of HOTAIR on EMT processes in GC cells. Previous research revealed that HOTAIR is involved in GC cell ECM function (16). Consequently, we detected the of mRNA and protein expression levels of EMT-related TFs following HOTAIR knockdown or overexpression. The expression levels of the epithelial marker, E-cadherin (also known as CDH1), in cells in which HOTAIR was knocked down were increased compared with the control cells. By contrast, the expression levels of mesenchymal markers, including N-cadherin, Snail, Slug, Twist and β-catenin, were decreased in the GC cells in which HOTAIR was knocked down. However, in the HOTAIR-overexpressing cells, the expression levels of EMT markers exhibited the opposite effects (Fig. 2A and B). We then performed immunofluorescence staining to evaluate E-cadherin-related morphological changes induced by HOTAIR knockdown or overexpression. As shown in Fig. 2C, the Lenti-HOTAIR si-infected cells exhibited an epithelial morphology, which was characterized by the increased expression of E-cadherin. By contrast, the cells infected with Lenti-HOTAIR exhibited a mesenchymal morphology, as indicated by the decreased expression of E-cadherin in the cytoplasm. The E-cadherin protein is a hallmark of EMT and is crucial for maintaining an epithelial phenotype (26). These results suggested that HOTAIR may promote GC cell metastasis by negatively regulating E-cadherin expression at the mRNA and protein level.

Previous research has indicated that long intergenic non-coding RNA HOTAIR serves as a scaffold for linking a histone methylase and a demethylase, which separately affects histone H3 lysine 27 methylation and lysine 4 demethylation (12). It has also been reported that the loss of H3K27me3 results in increased levels of H3K27ac, which are specifically dependent on PRC2 activity (27). It is believed that the methylation of H3K27 mediates transcriptional repression. However, the methylation and acetylation of H3K4, and the acetylation of H3K27 are regarded as transcriptionally active markers (28). Therefore, in this study, we performed western blot analysis to explore the molecular mechanisms through which HOTAIR regulates transcription in SGC-7901 and MGC-803 GC cells. The expression levels of H3K4me1/2/3 and H3K27ac were increased, whereas those of H3K4ac and H3K27me1/2/3 were decreased following infection with Lenti-HOTAIR si. By contrast, infection with Lenti-HOTAIR increased the H3K4ac and H3K27me1/2/3 expression levels, and the decreased H3K4me1/2/3 and H3K27ac expression levels (Fig. 3A). The results of western blot analysis indicated that HOTAIR regulated the methylation and acetylation of H3K4/27 and mediated the transcriptional repression or activation. Furthermore, we analyzed the gray value of H3K27me3/ac. The results indicated that HOTAIR may regulate histone H3K27 methylation and acetylation using an antagonistic approach (Fig. 3B).

The above-mentioned findings demonstrated that H3K27 methylation and acetylation were regulated by HOTAIR in opposing ways. Subsequently, we aimed to identify the histone-modifying enzymes that may be involved in HOTAIR-associated H3K27 methylation and acetylation. It is known that HOTAIR targets PRC2, which consists of H3K27 methyltransferase EZH2, SUZ12 and EED (29). In this study, the GC cell lines, SGC-7901 and MGC-803, were infected with Lenti-HOTAIR si or Lenti-HOTAIR. Western blot analysis was then used to examine the expression levels of EZH2 and SUZ12 and the results revealed that these were not markedly altered by HOTAIR knockdown or overexpression. However, the expression levels of acetyltransferase CBP increased with Lenti-HOTAIR si infection and decreased with the overexpression of HOTAIR. Additionally, the expression levels of H3K27me3 were decreased, whereas the expression levels of H3K27ac were increased by Lenti-HOTAIR si infection. The opposite effects were observed as regards the expression levels of these factors with HOTAIR overexpression (Fig. 4A). To further explore the association between the PRC2 complex and CBP, the expression levels of EZH2, SUZ12, CBP, H3K27me3 and H3K27ac were detected following the inhibition of EZH2 and SUZ12 expression. The results were similar to those observed with Lenti-HOTAIR si infection (Fig. 4B). Following this, immunofluorescence assays were performed to further verify the HOTAIR-mediated antagonistic switch between histone H3K27 methylation and acetylation in GC cells (Fig. 4C and D). All the results mentioned above indicated that the antagonistic switch of methylation and acetylation of H3K27 were specifically dependent on PRC2 activity, which was induced by HOTAIR. Conversely, acetyltransferase CBP was considered to play an important role in histone H3K27ac without the existence of the PRC2 complex.

To examine the underlying mechanisms through which E-cadherin transcription is suppressed by HOTAIR-mediated H3K27 methylation and acetylation, UCSC Genome Browser was used to identify H3K27me3 and H3K27ac enrichment peaks in the E-cadherin promoter region (Fig. 5A). ChIP was performed to obtain specific DNA, including the DNA of the CDH1 promoter binding to H3K27me3 and H3K27ac using agarose gel electrophoresis (Fig. 5B). The results suggested that H3K27me3 and H3K27ac were enriched in the promoter region of the CDH1 gene. In addition, qPCR analysis revealed that HOTAIR knockdown decreased the binding of H3K27me3 to the E-cadherin promoter in GC cells (Fig. 5C); however, HOTAIR knockdown induced the recruitment of H3K27ac to the E-cadherin binding site at the promoter of E-cadherin (Fig. 5D). Therefore, the knockdown of HOTAIR appears to relieve the repressive histone marker, H3K27me3, and induces a switch to the active marker, acetylated H3K27, at the promoter of E-cadherin. Taken together, these data provide evidence of a HOTAIR-mediated competition between H3K27me3 and H3K27ac in regulating CDH1 expression.

To further verify the role of HOTAIR in GC cells and to determine the therapeutic potential of HOTAIR knockdown, we established a subcutaneous tumor model of GC using MGC-803 GC cells. Nude mice were subcutaneously injected with MGC-80 cells that were infected with Lenti-NC or Lenti-HOTAIR si. From 10 days post-injection, the weight of the mice and tumor volume were measured every 2 days for a total of 25 days. The tumor weight was measured after the mice were sacrificed (Fig. 6A and B). Compared with the nude mice injected with Lenti-NC-infected MGC-803 cells, the mice bearing tumors derived from MGC-803 cells in which HOTAIR was knocked down exhibited suppressed GC tumor growth. In addition, IHC was performed to evaluate the pathological changes in each of the GC orthotopic tumors (Fig. 6C). Compared with the Lenti-NC tumors, treatment with Lenti-HOTAIR si significantly elevated E-cadherin expression. Furthermore, the methylation of H3K27 was decreased and the acetylation of H3K27 was increased following infection with Lenti-HOTAIR si, which was in line with our results in vitro.

Taken together, the findings of this study indicate that HOTAIR promotes the EMT of gastric cancer by inhibiting the expression of E-cadherin through an antagonistic switch of histone H3K27 acetylation to methylation at the E-cadherin promoter (Fig. 7).