A total of 33 pairs of NPC tissues and adjacent normal tissues were obtained from NPC patients receiving radical surgery at Xingtai People's Hospital, Hebei Province. Among these patients, 15 were male and 18 were female. Lymph node metastasis was encountered in 19 cases, but not in 14 cases. Twenty cases were stage I or II, and 13 cases were stage III or IV in terms of TNM staging. In addition, the average age was 46.14±0.58 years. All tissues were collected and stored at −80℃ before experimentation. For survival assay, we divided the 33 patients into two groups: low NEAT1 expression group (n=14, ≥medium value) and high NEAT1 expression group (n=19, <medium value) according to the medium value of NEAT1 expression in NPC tissues. Prior written informed consent was provided by each patient, and the study was approved by the Ethics Committee of Xingtai People's Hospital, Hebei Province.

Five NPC cell lines (CNE1, CNE2, SUNE1, SUNE2, and 5-8F) and the human nasopharyngeal epithelial cell line NP69 were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). All cells were grown in RPMI-1640 medium (Gibco, Rockville, MD, USA) containing 10% FBS (Gemini Bio-Products, West Sacramento, CA, USA) and 1% penicillin/streptomycin (Gibco) in a humidified 37℃ incubator supplemented with 5% CO2.

MiRNA mimics for miR-34a-5p (miR-34a-5p mimics) and negative control miR-NC, miR-34a-5p inhibitor (in-miR-34a-5p), negative control in-miR-NC, NEAT1-special small interfering RNA (si-NEAT1), and negative control si-NC were obtained from GenePhama (Shanghai, China). The NEAT1 overexpression vector (pcDNA-NEAT1) was commercially constructed by Sigma-Aldrich (St. Louis, MO, USA), and empty pcDNA vector was used as the negative control. When cells reached 60–80% confluence, transient transfection was performed using Lipofectamine 2000 transfection reagent (Life Technologies, Carlsbad, CA, USA) referring to manufacturer's instructions.

Total RNA was freshly extracted from tissues and cells using a miRNeasy FFPE kit (Qiagen, Valencia, CA, USA). Quality of the isolated RNA was assessed by a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). For NEAT1 detection, RNA samples were reversely transcribed into cDNA with Superscript III Reverse Transcriptase (Invitrogen, Waltham, MA, USA). Then, quantitative real-time PCR (qRT-PCR) assay was performed using a SuperScript Platinum SYBR™ Green One-Step qRT-PCR Kit (Invitrogen) on a 7900HT Fast Real-Time PCR System (Thermo Fisher Scientific), with normalization to the expression of GAPDH. Fold changes were calculated using the 2^−ΔΔCt method. For miR-34a-5p detection, cDNA was synthesized using miScript II RT Kit (Qiagen), and qRT-PCR was performed by miScript SYBR Green PCR Kit (Qiagen) and miScript Primer Assay specific for miR-34a-5p (Qiagen). U6 was used as an endogenous control.

Online software Lncbase v.2 (http://carolina.imis.athena-innovation.gr/diana_tools/web/index.php?r=site%2Ftools) was used to predict miRNAs directly interacting with NEAT1. Fragments of the NEAT1 sequence containing potential miR-34a-5p complementary sites or those with mutational miR-34a-5p target sites were inserted into psiCHECK-2 vector (Promega, Madison, WI, USA) to construct NEAT1 wild-type (WT) or mutant-type (MUT) reporter vector. The constructs were transiently cotransfected into 5-8F cells with miR-34a-5p mimics or in-miR-34a-5p. After 48 h transfection, the signals of firefly and Renilla luciferase in cell lysates were measured using a dual-luciferase reporter assay system (Promega), and relative luciferase activities were calculated as the ratio of Renilla/firefly luciferase activity.

RIP assay was performed to explore whether RNA induced silencing complex (RISC) contained NEAT1 and miR-34a-5p in 5-8F cells using an Imprint RNA Immunoprecipitation (RIP) kit (Millipore, Billerica, MA, USA) referring to the directions of the manufacturers. In brief, 5-8F cells transfected with miR-34a-5p mimics were lysed in RIP lysis buffer (Takara, Beijing, China) and were incubated with protein A/G magnetic beads and Argonaute-2 antibody (anti-Ago2, Abcam, Cambridge, UK) or IgG antibody (anti-IgG, Abcam). Then, cell lysates in the miR-34a-5p mimics group were centrifuged. The supernatants were marked as output, and immunoprecipitation complexes were named as IP. Western blot analysis was performed to validate successful RIP with anti-Ago2. Lastly, qRT-PCR assay was used to detect NEAT1 enrichment in the RIP RNA fraction.

Cell proliferation capacity was measured with the Cell Counting kit-8 (CCK-8, Dojindo Laboratories, Kumamoto, Japan) referring to the manufacturer's protocol. Briefly, cells were seeded in 96-well plates and transfected with si-NEAT1 or si-NEAT1+in-miR-34a-5p. After 0, 24, 48, 72, and 96 hours transfection, 10 µL of CCK-8 solution was added into each well, followed by determination of absorbance at 450 nm using a MRX II microplate reader (Dynex Technologies, Chantilly, VA, USA).

Cell migration assay was carried out using a 24-Transwell plate with 8-µm pores (Corning Incorporated-Life Sciences, Tewksbury, MA, USA). To be brief, 5-8F cells were transfected with si-NEAT1 alone or together with in-miR-34a-5p for 48 h. Then, transfected cells (1.0×10⁵) were seeded in the upper chamber, and 500 µL of growth medium containing 10% FBS was added in the lower chamber. After incubation for 48 h at 37℃, traversed cells on the lower side were fixed in 4% phosphate-buffered formaldehyde (PFA) and stained with 0.1% crystal violet (Sigma-Aldrich, St. Louis, MO, USA). At last, images were obtained, and migrated cells were counted using a microscope (Leica, Wetzlar, Germany) with Leica Qwin IM500 software.

Cell invasion ability was assessed using Transwell filters precoated with 50 ng/mL of Matrigel (BD Bioscience, San Jose, CA, USA). Transfected cells (1.0×10⁵) were seeded into the upper chamber, and the bottom chamber was filled with 500 µL of growth medium containing 10% FBS. After incubation for 48 h at 37℃, invaded cells were fixed in 4% PFA and stained with 0.1% crystal violet.

Total protein was extracted in chilled lysis buffer comprising 50 mM Tris-HCl (pH=7.4), 1 mM EDTA, 1% Triton X-100, 150 mM NaCl, 0.1% SDS, 0.1% deoxycholic acid, and 1% Protease Inhibitor Cocktail (Sigma-Aldrich). Protein extract concentrations were measured using BCA Protein Assay Kits (Pierce, Rockford, IL, USA). Ten micrograms of protein extract were separated by 10% SDS/polyacrylamide gel electrophoresis and then electroblotted onto a PVDF membranae (Du Pont, Boston, MA, USA). After blockage with 5% nonfat milk in TBS containing 0.1% Tween 20 (TBST), membranes were hybridized with antibodies against E-cadherin (1:1000, Cell Signaling Technology, Danvers, MA, USA), N-cadherin (1:1000, Cell Signaling Technology), Vimentin (1:1000, Cell Signaling Technology), β-catenin (1:5000, Abcam), cyclin D1 (1:10000, Abcam), c-myc (1:1000, Abcam), Ago2 (1:1000, Abcam), and β-actin (1:1000, Cell Signaling Technology) overnight at 4℃. After washing three times with TBST, the membranes were further incubated by horseradish-peroxidase-linked immunoglobulin G (Cell Signaling Technology) and visualized by a chemiluminescence detection kit (New England Biolabs, Beverly, MA, USA) on the Enhanced Chemiluminescence Western blot system (GE Healthcare Bioscience, Tokyo, Japan) with ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Lentiviruses with or without a NEAT1 knockout sequence (sh-NEAT1 or sh-NC) were constructed by Sigma-Aldrich, and were used to infect 5-8F cells. Puromycin was used to screened the stable sh-NEAT1-infected cells.

Female 6-week old SCID mice were purchased from Beijing Vital River Laboratory Animal Technology Corporation (Beijing, China) and housed in a specific-pathogen-free environment. Approximately 1.0×10⁷ sh-NEAT1/sh-NC-infected cells were subcutaneously inoculated into the nude mice to form xenograft mice (n=10). Tumor volume was measured with a caliper every one week. At 5 weeks after injection, all mice were euthanized and tumor tissues were excised for weight evaluation, qRT-PCR, and western blot assays. All experimental processes were approved by the Animal Care and Use Committee of Xingtai People's Hospital, Hebei Province.

All analyses were carried out using SPSS 18.0 software (SPSS Inc., Chicago, IL, USA). Paired Student's t-test, Mann-Whitney U test, and one-way analysis of variance were used in the study to evaluate statistical differences. p<0.05 was considered statistically significant.

To validate the function of NEAT1 in NPC progression, NEAT1 expression was first measured by qRT-PCR assay in NPC tissues and adjacent non-cancerous nasopharynx tissues. The data revealed that NEAT1 expression is highly increased in NPC tissues, compared with corresponding controls (Fig. 1A and B). Subsequently, we assessed the expression of NEAT1 in NPC cell lines. Consistently, NEAT1 levels were significantly higher in NPC cell lines than those in NP69 cells (Fig. 1C).

We then sought to determine whether NEAT1 expression is associated with the prognosis of NPC patients. According to the medium value of NEAT1 expression, 33 patients were divided into two groups: low NEAT1 expression group (n=14) and high NEAT1 expression group (n=19). Kaplan-Meier survival analysis revealed that low NEAT1 expression afforded longer overall survival time that high NEAT1 expression (Fig. 1D) (log-rank p<0.05). All of these results supported that NEAT1 played an oncogenic role in the progression of NPC.

Then, we further explored correlations between NEAT1 and clinical characteristics. As shown in Table 1, NEAT1 expression was highly associated with TNM staging and lymph node metastasis. No other clinical characteristics were found to be significantly associated with NEAT1 expression.

To further explore the underlying mechanism of NEAT1 as an oncogene in NPC progression, Lncbase v.2 was explored to search for target miRNAs binding to NEAT1. The predicted data indicated that NEAT1 harbors a putative binding region with miR-34a-5p (Fig. 2A). Thus, dual-luciferase reporter assay and RIP assay were employed to verify whether NEAT1 directly binds to miR-34a-5p in NPC cells. NEAT1-WT or NEAT1-MUT constructs were transfected into 5-8F cells with miR-34a-5p mimics or in-miR-34a-5p, followed by detection of relative luciferase activity. In 5-8F cells, the relative luciferase activity of NEAT1-WT was significantly decreased by cotransfection with miR-34a-5p mimics, while it was highly increased when introduced with in-miR-34a-5p (Fig. 2B and C). However, the relative luciferase activity of NEAT1-MUT was not affected in response to alteration of miR-34a-5p levels (Fig. 2B and C). Ago2, a component of the RISC, plays a pivotal role in the mature process of miRNAs. Hence, RIP experiments were performed using anti-Ago2 in 5-8F cells transfected with miR-34a-5p mimics to validate the endogenous interaction between NEAT1 and miR-34a-5p. These results revealed that relative enrichment of NEAT1 was higher in miR-34a-5p transfection group than in the control group (Fig. 2D). Additionally, we discovered that Ago2 primarily exists in IP samples in the RIP-Ago2 group, but in output samples in the RIP-IgG group, indicating that RIP assay was successfully performed.

Then, NEAT1 overexpression vector (pcDNA-NEAT1) or NEAT1-special siRNA (si-NEAT1) was transfected into 5-8F cells to observe the effect of NEAT1 on miR-34a-5p expression. As shown in Fig. 2E, NEAT1 levels were markedly elevated by transfection with pcDNA-NEAT1, while they were inhibited by si-NEAT1 transfection. Subsequent experiments revealed that miR-34a-5p expression was significantly and negatively correlated with NEAT1 levels (Fig. 2F). Further, we detected the expression of miR-34a-5p in NPC tissues and cell lines. These data showed that, in comparison to their counterparts, miR-34a-5p levels were strikingly attenuated in NPC tissues and cell lines (Fig. 2G and H). Taken together, these results hinted that NEAT1 might suppress miR-34a-5p expression by directly binding to miR-34a-5p.

Next, loss-of-function experiments were performed by transfection of si-NEAT1 into 5-8F cells. Transwell and CCK-8 assays revealed that NEAT1 knockdown remarkably weakened cell migration (Fig. 3B), invasion (Fig. 3C), and proliferation (Fig. 3D), compared with respective controls. In parallel, cell EMT ability was significantly repressed by NEAT1 depletion, presented by an increase of E-cadherin level, as well as decreases in N-cadherin and Vimentin levels (Fig. 3E-H). These results indicated that NEAT1 knockdown suppresses cell proliferation, migration, invasion, and EMT in NPC in vitro

We then sought to explore whether the above inhibition effects exerted by NEAT1 knockdown are mediated by miR-34a-5p. Hence, by cotransfecting si-NEAT1 and in-miR-34a-5p, functional experiments were conducted in 5-8F cells. As shown in Fig. 3A, the si-NEAT1-mediated repression effect on miR-34a-5p expression was strikingly abrogated by cotransfection with in-miR-34a-5p. Subsequent functional experiments demonstrated that si-NEAT1-mediated suppression effects on cell proliferation, migration, invasion, and EMT were significantly reversed by miR-34a-5p expression restoration in 5-8F cells (Fig. 3B-H). These results indicated that NEAT1 knockdown inhibits the proliferation, migration, invasion, and EMT by miR-34a-5p in NPC cell lines.

Constitutive activation of Wnt/β-catenin signaling pathway has generally been accepted as triggering tumorigenesis and metastasis.19 Thus, we further investigated whether NEAT1/miR-34a-5p axis regulates Wnt/β-catenin signaling in NPC. 5-8F cells were transfected with si-NEAT1 alone or together with in-miR-34a-5p, followed by determination of β-catenin, cyclin D1, and c-myc levels. These data demonstrated that, compared with negative control, NEAT1 knockdown leads to a decrease in β-catenin, cyclin D1, and c-myc expression levels (Fig. 4), indicating its role as a probable inhibitor of Wnt/β-catenin signaling. Furthermore, the inhibitory effect of NEAT1 knockdown was significantly abated by the restoration of miR-34a-5p expression in 5-8F cells. All these results implied that NEAT1 knockdown might blockade Wnt/β-catenin signalling by miR-34a-5p in NPC.

Lastly, we examined the therapeutic potential of NEAT1 in animal models. These results revealed that sh-NEAT1 infection significantly represses tumor growth, represented by decreases in tumor volume (Fig. 5A) and tumor weight (Fig. 5B). As presented in Fig. 5C and D, sh-NEAT1 introduction markedly weakened NEAT1 expression, but elevated miR-34a-5p levels in xenograft tumors. Moreover, NEAT1 depletion dramatically inhibited EMT in vivo, presented as an increase in E-cadherin level and decreases in N-cadherin and Vimentin levels (Fig. 5E). Additionally, the Wnt/β-catenin pathway was inactivated by NEAT1 depletion in tumors derived from sh-NEAT1-infected cells (Fig. 5F). Together, these findings suggested that NEAT1 knockdown might repress tumor growth and EMT by blockading Wnt/β-catenin signaling.