The samples were obtained from the surgical tissues of 69 patients with hepatic carcinoma at the Lianyungang First People's Hospital. Samples >3 cm from the edge of the cancerous tissue were used as the adjacent-cancerous tissues. All patients signed an informed consent form approved by the Lianyungang No. 1 People's Hospital ethics committee (permit no. LW-20140322001), which complied with medical ethics regulations. Of the enrolled patients, 39 were male and 30 were female. The patients' age ranged from 20–71 years, with a median age of 52 years. None of the patients received adjuvant treatments such as radiotherapy, chemotherapy or radiofrequency ablation, and the postoperative specimens were confirmed to be hepatocellular carcinoma by pathological examination. The specimens of the control group were adjacent tissues located at least 3 cm from the surgical margin, and no cancer cells were found on postoperative pathological examination. All specimens were quickly frozen in liquid nitrogen after being obtained and stored in a refrigerator at −80°C.

Hepatoma cell line (SUN449), hepatoblastoma cell line (Huh-6), liver cancer cell line (HepG2) and transformed human liver epithelial-2 cells (normal hepatocytes; THLE-2) were purchased from the American Type Culture Collection. The short tandem repeat method was used for authentication. The cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) containing 10% fetal calf serum (both Invitrogen; Thermo Fisher Scientific, Inc.) and placed in an incubator at 37°C containing 5% CO₂. After 0.25% trypsin digestion, the cells were passaged, and the cells in the logarithmic growth phase were used for the experiments.

A total of 3 short hairpin RNA (sh)-lncRNA XIST sequences, overexpression vector (oe)-lncRNA XIST, miR-320a mimic, miR-320a inhibitor, PIK3CA inhibitor, and their corresponding controls were designed and synthesized by IBSBIO Co. Ltd. Cells were co-transfected with shlncRNA XIST (2 µg; Ibsbio Co., Ltd.) and the viral packaging plasmid (2 µg) (Nanjing Cobioer Gene Technology Co., Ltd.) The expression was detected after incubation at 37°C for 72–96 h. After 72 h, G418 was added for screening. After 48 h of transfection, G418 was added to each group to achieve a final concentration of 0, 200, 400, 600, 800 and 1,000 mg·l⁻¹, respectively. After continuous observation for 4–7 days, the G418 concentration, which can lead to complete cell death in the untransfected group, was selected as the optimal working concentration for the subsequent screening of transfected cells. According to the selected optimal G418 concentration, the transfected cells were maintained in the culture until a stable viable cell line was obtained. Lipofectamine^® 2000 reagent (Thermo Fisher Scientific, Inc.) was also used to transfect oe-lncRNAXIST. The corresponding empty vector was used as a control. Similarly, the miR-320a mimic, miR-320a inhibitor, PIK3CA inhibitor, and their corresponding controls were also transfected by Lipofectamine 2000 at 37°C, as aforementioned. Non-targeting sequence was used as shNC, inhibitor NC, siNC and NC for negative control experiments. A non-targeting sequence was the negative control for transfections of miR-320a mimic. After 24 or 48 h, the following experiments were conducted.

The TRIzol^® method was used to extract the total RNA from hepatic carcinoma tissues and cells according to the manufacturer's instructions (Invitrogen; Thermo Fisher Scientific, Inc.). The primers were designed using Primer 5 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) and synthesized by the Shanghai Yingwei Jieji Company. For cDNA synthesis, a 20 µl reaction solution was prepared according to the reverse transcription kit (Roche Diagnostics). The amplification conditions of the thermal cycler were as follows: 25°C for 10 min; 55°C for 30 min; and 85°C for 5 sec. RT-qPCR was used to quantitatively detect the expression of each gene. A total of 2 µl cDNA of each clinical specimen was obtained; then, 10 µl SYBR qPCR mix (Toyobo Life Science) and 2 µl primer mixture were added, and dH₂O was added to the total volume of 20 µl. The LightCycler^®480 was used to detect relative expression. Gene expression was quantified using the 2^−ΔΔCq method (18). The primer sequences were as follows: lncRNA-XIST forward, 5′-ACGCTGCATGTGTCCTTAG-3′; lncRNA-XIST reverse, 5′-GAGCCTCTTATAAGCTGTTTG-3′; miR-320a forward, 5′-CAACAGAAGGCTCGAGGGAAGTCTGCGTGGCAGG-3′; miR-320a reverse, 5′-ATTCTGATCAGGATCCGAGGCGAATCCTCACATTG-3′; PIK3CA forward, 5′-GGTGACTGTGTGGGACTTATTG-3′; PIK3CA reverse, 5′-TGATGTAGTGTGTGGCTGTTG-3′; GAPDH forward, 5′-CATCACCATCTTCCAGGAGCG-3′; GAPDH reverse, 5′-TGACCTTGCCCACAGCCTTG-3′; U6 forward, 5′-CTCGCTTCGGCAGCACA-3′; U6 reverse, 5′-AACGCTTCACGAATTTGCGT-3′.

Following treatment, the cells in each group were trypsinized and collected, centrifuged at 300 × g (4°C) for 5 min and the supernatant was removed, and the complete medium was added to adjust the cell concentration to 5×10³/ml. The cells were seeded in a 96-well plate at 100 µl cells per well, with three replicate wells. After culturing for 12, 24, 48 and 72 h, a total of 10 µl Cell Counting Kit (CCK)-8 reagent (Beyotime Institute of Biotechnology) was added to each well, and incubated at 37°C for 1.5 h in the incubator. A microplate reader was used to determine the absorbance value of each well; the wavelength was 450 nm, and the proliferation curve was drawn. The cell count assay is an indirect form that can be used to measure cell proliferation.

According to the instructions of the apoptosis kit (Dojindo Molecular Technologies, Inc.), the cells in each group were trypsinized and collected and washed twice with phosphate buffered saline (PBS). The cells were centrifuged at 300 × g for 5 min (4°C) and then 100 µl Annexin V buffer was added to suspend the cells. Finally, PI and Annexin V staining solutions were added separately and mixed by pipetting. After 15 min of staining at room temperature and in the dark, 400 µl Annexin V buffer was added to the sample, and cell apoptosis was detected by flow cytometry.

After the cells from each group were digested with trypsin, the cell concentration was adjusted to 1×10⁸/l. Matrigel on the Transwell membrane was used to detect cell invasion ability, but Matrigel was not required to measure cell migration ability. A total of 100 µl DMEM without FBS was added to the upper chamber of each Transwell, while 500 µl of DMEM was added to the lower chamber. The Transwell chamber was incubated in an incubator with 5% CO₂ at 37°C for 24 h. The cells were fixed with 4% paraformaldehyde for 15 min at room temperature, and washed thrice with PBS. Then, the cells were stained with 1% crystal violet at room temperature for 10 min. After washing away the crystal violet staining solution, cell migration and invasion were observed under a light microscope (Leica Microsystems GmbH). Ten fields were selected to calculate the average number of cells that migrated and invaded with a magnification of ×400.

The binding regions of lncRNA XIST and PIK3CA with miR-320a were predicted based on the StarBase (http://starbase.sysu.edu.cn/) and TargetScan databases (http://www.targetscan.org/vert_72/). Wild-type lncRNA XIST (lncRNA XIST-wt), wild-type PIK3CA non-coding region (PIK3CA-WT), mutant lncRNA XIST (lncRNA XIST-Mut) sequence, and mutant PIK3CA (PIK3CA-Mut) missing the miR-320a binding region, as well as mutations missing the miR-320a binding region were designed and the luciferase reporter vector was constructed by Nanjing Cobioer Gene Technology Co., Ltd. The lncRNA XIST-wt, lncRNA XIST-Mut, PIK3CA-WT and PIK3CA-Mut luciferase reporter vectors were transfected into the control and experimental cells at 37°C for 48 h (using Lipofectamine^® 2000), and the Promega dual-luciferase reporter gene kit (Promega Corporation) was used to test the luciferase activity after transfection. Fluorescent enzyme activity was recorded as the ratio of firefly luciferase to Renilla luciferase activity.

The RNA pull-down experiment was performed according to the instructions of the RNA pull-down kit (Thermo Fisher Scientific, Inc.). The brief steps were as follows: The labeled target RNA was combined with streptavidin magnetic beads, and 30 Ult of streptavidin magnetic beads were added to a 1.5-ml centrifuge tube. The magnetic beads were collected, and an equal volume of RNA capture buffer was used for resuspension and gentle pipetting. A total of 50 pmol biotin-labeled RNA was added, mixed well, and incubated for 30 min at room temperature. The magnetic beads were collected and 100 µl RNA binding buffer was added and mixed. Then, the liquid was discarded and 100 µl of the mixed RNA binding reaction solution was combined with the magnetic beads. Elution buffers (50 µl) were used to elute the complex. The eluate was collected, RNA was extracted with TRIzol^® and stored at −80°C, and the amount of enriched miR-320a was detected.

Western blotting was performed in accordance with conventional methods. Protein extraction buffer (Shanghai Kanglang Biotechnology Co., Ltd.) was used to extract the total protein of each group of hepatocellular carcinoma cells. The concentration of each histone was detected by the bicinchoninic acid (BCA) protein quantitative method, and all proteins were boiled and denatured at 100°C. A total of 50 µg of each protein group was subjected to 10% acrylamide gel electrophoresis, and transferred to a polyvinylidene fluoride membrane at a constant current. The membrane was sealed in 5% skim milk at room temperature for 1 h, washed with PBS, and incubated with the primary antibody PIK3CA (1:1,000; cat. no. ab40776; Abcam) at 4°C overnight. After washing with TPBS, the membrane was incubated with Alexa Fluor^® 488 goat anti-rabbit IgG (H+L) (ab150077; 1:2,000; cat. no. ab150077; Abcam) at room temperature for 1 h. Finally, the protein bands were illuminated, developed and fixed by enhanced chemiluminescence (Pierce; Thermo Fisher Scientific Inc.). Signals were quantified using ImageJ 1.63 software (National Institutes of Health).

For any quantified parameter, each group had three parallels, and the experiment was repeated three times. SPSS 21.0 (IBM Corp.) was the statistical software used for the data analysis. The data are represented by the mean ± standard deviation. Two-way ANOVA (two-way analysis of variance) with post hoc test was used. Kaplan-Meier survival analysis was applied to examine the association between gene expression and survival rate, and statistical significance was assessed using the Mantel-Cox log rank test. Pearson's correlation was used to test the gene expression correlation. P<0.05 was used to indicate a statistically significant difference.

RT-qPCR was used to detect the expression level of lncRNA XIST in hepatic carcinoma tissues and cells. Compared with normal tissues, lncRNA XIST was highly expressed in hepatic carcinoma tissues (P<0.01; Fig. 1A). Moreover, with increasing tumor stage, the expression of lncRNA XIST also increased (Fig. 1B). Based on the median value of lncRNA XIST expression (cut-off value, 1.815), the 69 patients were divided into the highly expressed lncRNA XIST group (n=37) and low-expressed lncRNA XIST group (n=32). As shown in Fig. 1C, the survival rate was significantly lower in the highly expressed lncRNA XIST group. Moreover, tumor size, vascular invasion and classification were significantly different between the two groups (Table I). The lncRNA XIST was highly expressed in all hepatic carcinoma cells compared with the THLE-2 cell line (Fig. 1D). Among the hepatic carcinoma cell lines, lncRNA XIST was expressed the highest in the HepG2 and SUN449 cell lines. Thus, the HepG2 and SUN449 cell lines were used in the following experiments.

After 3 shlncRNA XIST transfection, the transfection efficiency was detected. In terms of the results, the relative lncRNA XIST level was significantly decreased, and shlncRNA XIST-1 was associated with the highest transfection efficiency (Fig. 2A). For the following experiments, shlncRNA XIST-1 was used and defined as shlncRNA XIST. Following shlncRNA XIST transfection for 72 h, the proliferation of both HepG2 and SUN449 cell lines significantly decreased (Fig. 2B). Moreover, the apoptosis rate was significantly higher in the shlncRNA XIST group in both cell lines (Fig. 2C). The invasion and migration of hepatic carcinoma cells were also inhibited by shlncRNA XIST (Fig. 2D and E). Thus, shlncRNA XIST inhibited cell proliferation, invasion and migration, while promoting the apoptosis of hepatic carcinoma cells.

Starbase was applied to predict the targeting association between lncRNA XIST and miR-320a. The binding site is shown in Fig. 3A. In the lncRNA XIST-WT group, the miR-320a mimic resulted in significantly decreased luciferase activity, while the miR-320a inhibitor increased its activity. However, both the miR-320a mimic and miR-320a inhibitor had no effect on luciferase activity in the lncRNA XIST-MUT group (Fig. 3B). Compared with the empty vectors and mutants, biotin-labeled lncRNA XIST was used to pull down miR-320a by affinity in vitro, and the results confirmed the endogenous binding between miR-320a and lncRNA XIST (Fig. 3C). In addition, the expression level of miR-320a was also detected in hepatocellular carcinoma cells and tissues. As shown in Fig. 3D and E, the miR-320a demonstrated a lower expression level in hepatocellular carcinoma cells and tissues. Pearson's correlation analysis also confirmed that miR-320a and lncRNA XIST demonstrated a negative association in patients with hepatocellular carcinoma (Fig. 3F). Notably, the relative miR-320a expression level was significantly increased following shlncRNA XIST transfection (Fig. 3G).

The Starbase online tool was used to predict the targeting association between PIK3CA and miR-320a. The binding site of PIK3CA and miR-320a is shown in Fig. 4A. miR-320a mimic, miR-320a inhibitor and miR-NC were transfected in cells; the transfections were successful (Fig. 4B). The luciferase reporter gene experiment was undertaken to verify whether binding occurred. In the PIK3CA-WT group, the miR-320a mimic significantly decreased the luciferase activity, while the miR-320a inhibitor increased the activity. However, both the miR-320a mimic and miR-320a inhibitor had no effect on luciferase activity in the PIK3CA-MUT group (Fig. 4C). The expression of PIK3CA in hepatocellular carcinoma cells was detected by RT-qPCR. As shown in Fig. 4D, the level of PIK3CA was significantly higher in hepatocellular carcinoma cells compared with the THLE-2 cell line. In hepatocellular carcinoma and normal tissues, the expression of PIK3CA exhibited a similar trend (Fig. 4E). Following transfection with miR-320 mimic, the expression level of PIK3CA significantly decreased (Fig. 4F). Pearson's correlation analysis was used to confirm the correlation between PIK3CA and lncRNA XIST or miR-320a expression. The results showed that PIK3CA was negatively correlated with miR-320a, while PIK3CA was positively correlated with lncRNA XIST (Fig. 4G and H).

The rescue assay was performed in SUN449 cell lines. Small interfering RNA (si)PIK3CA significantly decreased PIK3CA expression, indicating the transfection was successful (Fig. 5A). Based on transfection, the cells were divided into MOCK, oelncRNA XIST, oelncRNA XIST+miR-320a mimic and oelncRNA XIST+PIK3CA inhibitor groups. After transfection, oelncRNA XIST significantly increased lncRNA XIST expression level (Fig. 5B). OelncRNA XIST significantly increased PIK3CA, while miR-320a mimic and PIK3CA inhibitor decreased the expression of PIK3CA at the gene and protein level (Fig. 5C). With respect to the effects of miR-320a, PIK3CA and lncRNA XIST levels on hepatocellular carcinoma development, CCK-8, flow cytometry and the Transwell assay were processed. When examining the results, oelncRNA XIST increased the cell number, invasion and migration, while it decreased the apoptosis of SUN449 cells. The cell number is an indirect measure of proliferation. More importantly, the miR-320a mimic and PIK3CA inhibitor demonstrated the recovery effects of oelncRNA XIST in the aforementioned processes (Fig. 5D-G). The aforementioned findings indicate that lncRNA XIST accelerates hepatic carcinoma progression by targeting the miR-320a/PIK3CA axis.