Abstract
Maternally expressed gene 3 (MEG3), a long non-coding RNA (lncRNA), has tumor-suppressor properties and its expression is lost in several human tumors. However, its biological role in esophageal squamous cell carcinoma (ESCC) tumorigenesis is poorly defined. The present study determined the role and methylation status of MEG3 in esophageal cancer cells and ESCC clinical specimens, and further observed the competing endogenous RNA (ceRNA) activity of MEG3 in the pathogenesis and development of ESCC. Significant downregulation of MEG3 was detected in esophageal cancer cells and ESCC tissues and the expression level of MEG3 was significantly increased in cancer cells after treated with the DNA methyltransferase inhibitor 5-Aza-dC. Upregulation of MEG3 led to the inhibition of proliferation and invasiveness of the cancer cells. The aberrant promoter hypermethylation of MEG3 indicates silencing of its expression. Furthermore, MEG3 acts as a ceRNA to regulate the expression of E-cadherin and FOXO1 by binding hsa-miR-9. Upregulation of miR-9 was detected in esophageal cancer cell lines and ESCC tissues, and miR-9 promoted esophageal cancer cell proliferation and invasion. Finally, downregulation and hypermethylation of MEG3 was associated with ESCC patients' survival.
Implications: MEG3 functions as a tumor-suppressive lncRNA and aberrant promoter hypermethylation is critical for MEG3 gene silencing in ESCC. In addition, MEG3 acts as a ceRNA to regulate expression of E-cadherin and FOXO1 by competitively binding miR-9 and may be used as a potential biomarker in predicting ESCC patients' progression and prognosis. Mol Cancer Res; 15(7); 800–10. ©2017 AACR.
This article is featured in Highlights of This Issue, p. 787
Introduction
Esophageal cancer ranks the eighth of the incidence of malignant tumors and the sixth cause of cancer-related death worldwide, and carries a high morbidity and mortality in the high-incidence regions (1, 2). Esophageal squamous cell carcinoma (ESCC) is the predominant type of esophageal cancer in China, and in spite of the continuous improvement in available therapies, the overall survival rate of ESCC patients remains unsatisfactory (3). ESCC has the characteristics of familial aggregation and carries high incidence in some areas, especially in some counties in northern China bordering Shanxi, Henan, and Hebei Provinces. Although some tumor suppressors and oncogenes have been identified to play crucial roles in the tumorigenesis and development of ESCC (4); however, the exact molecular mechanisms for ESCC carcinogenesis and progression have not been well elucidated.
Long non-coding RNAs (lncRNA), within the length range from 200 nt to 100 kb, are emerging as new factors in many biological processes (5). Genome-wide transcriptomic analyses have revealed highly aberrant lncRNA expression in tumor cell lines and primary tumor tissues, and certain differentially expressed lncRNAs may function as indicator and new regulators in cancer metastasis or prognosis (6, 7). As an lncRNA, maternally expressed gene 3 (MEG3) is located at 14q32, a region in which chromosomal abnormalities are associated with the pathogenesis and progression of multiple tumors (8–12). Growing evidence suggests that MEG3 has tumor-suppressor properties and is involved in tumorigenesis (13).
MEG3 is highly expressed in several types of normal tissues but down-regulated in malignant tissues (14). Over expression of MEG3 may inhibit proliferation, migration, and invasion of tumor cells by interacting with several molecules such as p53, MDM2, and cyclic AMP (15, 16). DNA methylation plays an essential role in silencing MEG3 in several tumor types (17–22). The competing endogenous RNA (ceRNA) hypothesis postulates that RNA transcripts sharing multiple microRNA (miRNA) response elements (MREs) in their 3′ UTRs communicate to each other and regulate the expression levels by competing for a limited pool of miRNAs (23). Accumulating studies have revealed that lncRNAs are the targets of miRNAs and may act as ceRNAs in regulating the expression of mRNAs (24, 25). Bioinformatics analysis predicted two hsa-miR-9 binding sites within the MEG3 transcripts, indicating its potential ceRNA role. However, to the best of our knowledge, the effects and methylation status of MEG3, and its roles as ceRNA in ESCC tumorigenesis have not been investigated and clarified so far. In the current study, we detected the role and methylation status of MEG3 in esophageal cancer cell lines and ESCC tissues, and further detected whether MEG3 could regulate ESCC carcinogenesis as a ceRNA and served as a potential target for ESCC metastasis and prognosis.
Materials and Methods
Patients and specimens
One hundred and forty-three pairs of surgical primary ESCC tissues and corresponding adjacent normal tissues were obtained directly after surgical resection between the years of 2007 and 2010 at the Fourth Affiliated Hospital, Hebei Medical University. Tissue samples were divided into two parallel parts, one part was formalin-fixed and paraffin-embedded, the other parts were frozen and stored at −80°C to extract genomic DNA and RNA. All study subjects were ethnically homogeneous Han nationality and residents of Hebei Province and its surrounding regions and written informed consent was obtained from each case before surgery. The patients consisted of 106 males and 37 females with a median age of 58 years (ranged from 37 to 75 years). Individuals with at least one first-degree relative or at least two second-degree relatives having esophageal/cardia/gastric cancer were defined as having family history of upper gastrointestinal cancers (UGIC). Information on clinicopathologic characteristics was available from hospital recordings and pathological diagnosis. Recurrence and survival data were ascertained through the tumor registry and hospital chart review (Supplementary Table S1). The study was approved by the Ethics Committee of the Fourth Affiliated Hospital, Hebei Medical University.
Cell culture and treatment
Human esophageal cancer cell lines TE1, TE13, Eca109, YES2, and T.Tn, and human normal esophageal epithelial cell line HEEpiC were purchased from the ATCC. Cells were detected and identified as mycoplasma and bacteria free according to the ATCC's instructions during the past 3 months. Cells were seeded at a low density and incubated for 24 hours before treatment with DNA methyltransferase inhibitor 5-aza-2′-deoxycytidine (5-Aza-dC). All five esophageal cancer cells and normal esophageal epithelial cells (2 × 105/mL) were treated with 5 μmol/L 5-Aza-dC (Sigma) for 72 hours and medium containing 5-Aza-dC was changed every 24 hours. Control cells received no drug treatment.
RNA isolation and quantitative real-time RT-PCR assay
Tumor cells were separated from frozen tissue sections by laser capture microdissected (LCM) technique and total RNA was extracted from laser capture microdissected tumor cells and corresponding normal tissues, and cell lines by standard methods using TRizol reagent (Invitrogen). For MEG3, E-cadherin, and FOXO1 expression, 2 μg RNA was used to synthesize single-stranded cDNA using the advantage RT-for-PCR Kit (Clontech Laboratories) and the cDNA from each sample was used as quantitative real-time RT-PCR template. Power SYBR Green PCR Master Mix (Life Technologies) was used as amplification reaction mixture. The primers and reaction conditions for MEG3, E-cadherin, and FOXO1 were listed in Supplementary Table S2. For miR-9-5p expression, miRcute miRNA First-strand cDNA Synthesis Kit (Tiangen Biotech, China) was used to synthesize the first strand of cDNA, and miRcute miRNA qPCR Detection Kit (SYBR Green; Tiangen Biotech, China) was used to detect miR-9-5p expression. Human GAPDH and U6 snRNA were used for MEG3/ E-cadherin/FOXO1 and miR-9-5p normalization, respectively. The fold change for the target genes were calculated using the 2–ΔΔCT method (26). All the samples were run in triplicate.
Northern blot assay
Ten μg of the indicated RNA was subjected to formaldehyde gel electrophoresis and transferred to a Biodyne Nylon membrane. Then, a biotin-16-dUTP (Roche)–labeled MEG3 cDNA probe was prepared and after 60 minutes of prehybridization in ULTRAhybTM-Oligo buffer (Ambion), the membrane was hybridized at 68°C for 12 hours in ULTRAhybTM-Oligo buffer containing the denatured probe. Washes were performed as described in the Ambion NorthernMax Kit.
DNA extraction, bisulfate modification, and methylation analysis of MEG3
Total DNA was isolated from 5-Aza-dC–treated and –untreated cells using DNAzol (Invitrogen) according to the manufacturer's recommendation. Genomic DNA from tumor and corresponding normal sections was isolated from flash-frozen tissues using a simplified Proteinase K (Merck) digestion method. To examine the DNA methylation patterns, 1 μg of genomic DNA was bisulfate modified using Epitect Fast Bisulfite Conversion Kits (Qiagen) according to the manufacturer's instructions. The methylation status of MEG3 was then determined by methylation-specific polymerase chain reaction (MSP) method as described previously (17). The primers and reaction conditions were listed in Supplementary Table S2. Genomic DNA, methylated in vitro by CpG methyltransferase (Sss I) following the manufacturer's directions (New England BioLabs), was used as a positive control and water blank was used as a negative control. MSP products were analyzed on 2% agarose gel with ethidium bromide staining, and were determined to have methylation if a visible band was observed in the methylation reaction. Reactions were performed in duplicate with each of the samples.
Cell transfection
For overexpression of MEG3, the sequence of MEG3 was synthesized and subcloned into pcDNA3.1 (Invitrogen). The Eca109 and YES2 cell lines were transfected with MEG3 expression plasmid (pcDNA3.1-MEG3) or the empty vector (pcDNA3.1-EV) as control at a final concentration of 2 μg/uL using FuGENE HD Transfection Reagent (Promega) in accordance with the manufacturer's instructions. For inhibition of MEG3, HEEpiC cells were transfected with MEG3 specific shRNA plasmid using FuGENE HD Transfection Reagent, and disrupted order of shRNA was used as a negative control. For inhibition of miR-9 function, YES2 cells were transfected with a specific microRNA inhibitor for miR-9 (miR-9 inhibitor) or its negative control RNA (inhibitor control) at a final concentration of 25 nmol/L (Ambion). For overexpression of miR-9, Eca109 cells were transfected with miR-9 mimic or mimic control (Ambion) at a final concentration of 25 nmol/L.
Cell proliferation assay
The proliferation of MEG3 transfected or knocked down or miR-9 inhibitor treated cells was measured by cell-counting kit-8 (CCK-8) assay. Before proliferation detected, 10 μL of CCK8 (Dojindo Molecular Technologies, Inc., Japan) was added to the 100 μL cultured cells, and after incubated for 2 hours in a humidified incubator containing 5% CO2 at 37°C, the absorbance of each well was detected at a wavelength of 450 nm. Proliferation rates were determined at 0, 24, 48, 72, 96 hours after transfection. All experiments were performed in triplicate.
Cell invasion assay
The invasiveness of MEG3 transfected or knocked down or miR-9 inhibitor treated cells was evaluated in 24-well Transwell chambers (Corning). The number of cells invaded through the membrane to the lower surface was counted in five microscopic fields (at ×100 magnification) per filter. The experiments were repeated in triplicate.
Subcellular fractionation
To determine the cellular localization of MEG3, cytosolic and nuclear fractions were collected from HEEpiC cells using Nuclear/Cytosol Fractionation Kit (BioVision) according to the manufacturer's protocol.
Anti-AGO2 RIP assay
For anti-AGO2 RIP, Eca109 cells were transfected with miR-9 mimic or mimic control. After 48 hours, cells were used to perform RNA immunoprecipitation (RIP) experiments using an AGO2 antibody (Millipore) and the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) according to the manufacturer's instructions.
Luciferase reporter assay
The Eca109 cells were transfected with either miR-9 mimic (25 nmol/L) or mimic control (25 nmol/L) at 24 hours after transfection with the pGL3 luciferase expression vector containing wild type MEG3 (pGL3-MEG3) or diverse mutant MEG3 using FuGENE HD Transfection Reagent (Promega). Renilla luciferase vector (pRL-TK) was cotransfected into Eca109 cells as a normalizing transfection control. After 14 hours, the reporter luciferase activity was measured with the Dual-Luciferase Reporter assay system (Promega) according to the manufacturer's instructions. All transfection assays were carried out in triplicate.
Western blot analysis
Western blot analysis of E-cadherin and FOXO1 expression was carried out using standard method. Cells were homogenized and centrifuged. Proteins were separated by 10 % SDS-PAGE, and were transferred electrophoretically onto PVDF membranes (Millipore). Antihuman mouse monoclonal antibody for E-cadherin (1: 100 dilution; Santa Cruz Biotechnology) and FOXO1 (1: 100 dilution; Santa Cruz Biotechnology) was used as the primary antibody. The membranes were stained using an ECL kit according to the manufacturer's instructions. An anti-GAPDH antibody was used as a control.
Statistical analysis
Statistical analysis was performed with SPSS19.0 software package (SPSS Company). The real-time RT-PCR results were expressed as the mean ± SD. The Student t test was used to compare the expression means between different groups. The status of gene methylation between different groups was analyzed using the Pearson's χ2 test. Survival curves were made by using the Kaplan–Meier method and the Log-rank or the Breslow tests were used as needed for the univariate comparison of MEG3 expression and methylation categories. The Cox's multivariate test applied in a stepwise forward method was used to adjust for potentially confounding variables and to evaluate the role of MEG3 as an independent predictor of patients' prognosis. All statistical tests were two sided; and a P value of <0.05 was considered to be statistically significant.
Results
Frequent silencing of MEG3 and upregulation of the gene by 5-Aza-dC treatment in esophageal cancer cell lines
As shown in Fig. 1A, the expression of MEG3 was evidently decreased or silenced in five esophageal cancer cell lines. However, the expression level of MEG3 was significantly increased in the cell lines after treated with DNA methyltransferase inhibitor 5-Aza-dC, suggesting the important role of abnormal methylation in the inactivation of MEG3 in esophageal cancer cell lines.
The expression and methylation status of MEG3 in human esophageal cancer cell lines and tumor tissues. A, Expression of MEG3 in five esophageal cancer cell lines and human normal esophageal epithelial cells treated or untreated with 5-Aza-dC by qRT-PCR. *, Compared with untreated cells, P < 0.05. B, Schematic structure of MEG3 CpG islands. MSP regions analyzed are indicated. C, The methylation status of MEG3 detected by MSP analysis in various cancer cell lines with or without 5-Aza-dC treatment. M, methylated; U, unmethylated. D, Relative expression of MEG3 in normal tissues and corresponding ESCC tumor tissues, as determined by qRT-PCR. *P < 0.05. E, Northern blot analysis of MEG3 in normal tissues and corresponding ESCC tumor tissues. N, normal tissues; T, tumor tissues. F, Relative expression of MEG3 in different subgroups. *, P < 0.05. G, The methylation status of MEG3 in ESCC tumor tissues and corresponding normal tissues determined by MSP analysis. m: methylated; u: unmethylated. H, Relative expression of MEG3 in the tumor tissues with and without methylation of the gene. *, P < 0.05.
The aberrant methylation of MEG3 induces silencing of its expression
MEG3 promoter region is rich in CpG islands predicted by MethPrimer (Fig. 1B). The MSP method was then used to detect the methylation status of MEG3 in esophageal cancer cell lines and the methylation primers were located in the enhancer region of MEG3 as previously reported (17–19). As shown in Fig. 1C, fully or partially methylation of MEG3 was detected in esophageal cancer cell lines. However, the aberrant methylation status of the cells was reversed after treated with 5-Aza-dC, which suggested that hypermethylation of MEG3 may be one of the mechanisms in silencing its expression.
Decreased expression of MEG3 in clinical specimens
We then further detected the expression level of MEG3 in ESCC tissues. Compared with corresponding normal tissues, the expression level of MEG3 in ESCC tumor tissues was significantly decreased (P < 0.01; Fig. 1D), and the results were verified by Northern blot assay (Fig. 1E). Furthermore, the expression levels of MEG3 in the tumor tissues of 117 patients (81.8%) were less than 50% of that in corresponding normal tissues. When stratified for clinicopathologic characteristics, the expression level of MEG3 in ESCC tumor tissues was associated with TNM stage, depth of invasion, lymph node metastasis, and distant metastasis or recurrence (P < 0.05; Fig. 1F).
Aberrant methylation of MEG3 in clinical specimens
The methylation analysis was successfully performed in all tissue specimens by the MSP method (Fig. 1G). Fully methylation, partially methylation was respectively observed in 82(57.3%), 27(18.9%) primary tumor tissues, the methylation frequency of MEG3 in tumor tissues was significantly higher than that in corresponding normal tissues (Supplementary Table S3). When stratified for clinicopathologic characteristics, the methylation status of MEG3 was associated with TNM stage, depth of invasion, lymph node metastasis, and distant metastasis or recurrence (P < 0.05; Table 1). The association between MEG3 expression and methylation was further detected in ESCC tissues. The expression level of MEG3 in ESCC tissues with unmethylation of the gene was significantly higher than that with hypermethylation of the gene (Fig. 1H; P < 0.05).
Methylation status of MEG3 in ESCC tissues
MEG3 inhibits esophageal cancer cell proliferation and invasion
The function of MEG3 was then investigated in esophageal cancer cell lines. The construct containing MEG3 transcripts (pcDNA3.1-MEG3) was transfected into Eca109 and YES2 cells. As shown in Fig. 2A, the expression level of MEG3 was significantly upregulated in transfected Eca109 and YES2 cells. Transfection of MEG3 led to a significant inhibition of Eca109 and YES2 cells proliferation and invasiveness (Fig. 2B and C). However, knockdown of MEG3 in HEEpiC cells led to increased proliferation and invasiveness of the cells (Fig. 2D–F).
The functional analysis of MEG3 in human esophageal cancer cell lines and normal esophageal epithelial cell line. A, Upregulation of MEG3 was detected by qRT-PCR in MEG3-transfected Eca109 and YES2 cells compared with empty vector transfected cells. *, P < 0.05. B, Overexpression of MEG3 inhibited Eca109 and YES2 cells proliferation detected by CCK-8 assay. *, Compared with empty vector, P < 0.05. C, Overexpression of MEG3 inhibited Eca109 and YES2 cells invasiveness detected by Transwell invasion assay. The results were measured by determined cell counts that penetrated through Matrigel-coated Transwell chambers (8-μm pore size). *, Compared with empty vector, P < 0.05. D, Downregulation of MEG3 was detected by qRT-PCR in shRNA transfected HEEpiC cells. *, P < 0.05. E, Knockdown of MEG3 increased HEEpiC cells proliferation detected by CCK-8 assay. *, Compared with shRNA control, P < 0.05. F, Knockdown of MEG3 increased HEEpiC cells invasiveness detected by Transwell invasion assay. For all quantitative results, the data are presented as the mean ± SD. The experiments are representative of three independent experiments with similar results.
MEG3 is physically associated with miR-9
Many RNA transcripts have been reported to function as ceRNAs by competitively binding common miRNAs. To investigate the possibility of MEG3 as ceRNA, we first detected the level of the gene in nuclear and cytoplasmic fractions of HEEpiC cells, which expressed high levels of endogenous MEG3. As shown in Fig. 3A, qRT-PCR analysis revealed that MEG3 was predominantly expressed in cytoplasm, which suggested that it may act as a ceRNA. Then, the webserver lnCeDB (http://gyanxet-beta.com/lncedb/index.php) and DIANA TOOLS (http://diana.imis.athena-innovation.gr/DianaTools) were adopted to predict potential lncRNA–miRNA interactions. Among the results, two miR-9-5p–binding sites scattering the MEG3 transcripts were found (Fig. 3B). The miRNAs are known to bind their targets and cause RNA degradation and/or translational repression in an AGO2-dependent manner. Endogenous MEG3 pull down by AGO2 was specifically enriched in miR-9 transfected Eca109 cells (Fig. 3C), suggesting that miR-9 are bona fide MEG3-targeting miRNAs.
The association between MEG3 and miR-9. A, The levels of MEG3 in nuclear and cytoplasmic fractions of HEEpiC cells detected by qRT-PCR. GAPDH: cytoplasmic control; U6: nuclear control. Data are mean ± SD. Error bars are representative of three independent experiments. B, Schematic diagram of potential miR-9-5p target sites in MEG3. C, Anti-AGO2 RIP was performed in Eca109 cells transiently overexpressing miR-9, followed by qRT-PCR to detect MEG3 associated with AGO2. *, P < 0.05. D, Expression of miR-9 in a pool of 6 nontumorous tissues (N), human normal esophageal epithelial cells, and five esophageal cancer cell lines. U6 was set as an endogenous control. *, Compared with N or human normal esophageal epithelial cells, P < 0.05. E, Relative expression of miR-9 in normal tissues and corresponding ESCC tumor tissues, as determined by qRT-PCR. *, P < 0.05. F, Relative expression of miR-9 in different subgroups. *, P < 0.05. G, The expression correlation between MEG3 and miR-9 in ESCC tissues. H, Downregulation of miR-9 was detected in YES2 cells after treated with miR-9 inhibitor (25 nmol/L) compared with inhibitor control (25 nmol/L) by qRT-PCR. *, P < 0.05. I, Cell proliferation was inhibited after YES2 cells was treated with miR-9 inhibitor compared with inhibitor control, as determined by CCK-8 assay. *, P < 0.05. J, Cell invasiveness was inhibited after YES2 cells was treated with miR-9 inhibitor compared with inhibitor control, as determined by Transwell invasion assay. *, P < 0.05. Results are expressed as mean ±SD of three independent experiments.
Upregulation of miR-9 in esophageal cancer cell lines and clinical specimens
According to the ceRNA theory, the expression level and function of miR-9 may be reversely associated with MEG3. As shown in Fig. 3D, upregulation of miR-9 was detected in YES2 and T.Tn cell lines. The average miR-9 expression level was significantly higher in tumor tissues than that in corresponding normal tissues (Fig. 3E; P < 0.05). When stratified for clinicopathologic characteristics, the expression level of miR-9 in ESCC tumor tissues was associated with TNM stage, depth of invasion, lymph node metastasis, and distant metastasis or recurrence (P < 0.05; Fig. 3F). The expression level of miR-9 in ESCC tissues was inversely correlated with MEG3 (Fig. 3G). Furthermore, functional analyses of miR-9 demonstrated that transfected YES2 cells with miR-9 inhibitor showing decreased proliferation and invasiveness (Fig. 3H–J). These results suggested the oncogenic roles of miR-9 in ESCC cells, which was contrary to the roles of MEG3.
MEG3 acts as a sponge for miR-9
To verify whether MEG3 is associated with miR-9, a series of luciferase reporters were constructed that contained the wild-type MEG3 (pGL3-MEG3), or mutant MEG3 with mutations of two predicted miR-9–binding sites (pGL3-MEG3-M1-2). We found that transfection with miR-9 mimic significantly reduced the luciferase activities of the pGL3-MEG3 reporter vector but not empty vector or mutant vector pGL3-MEG3-M1-2 in Eca109 cell line, suggesting the binding of miR-9 to these sites (Fig. 4A). For further analysis, we cloned the each putative miR-9-target site or its point mutant in sequences corresponded to “seed sequence” of miR-9 into a pGL3 vector (pGL3-MEG3-Site1, pGL3-MEG3-Site M1, pGL3-MEG3-Site2, pGL3-MEG3-Site M2) and performed reporter assays. As shown in Fig. 4B and C, when miR-9 mimic was transfected into Eca109 cells, luciferase activity was significantly repressed in each site, whereas the repressions were completely abolished when mutation was introduced in each site. Ectopically expressed MEG3, but not the mutant, reduced the level of miR-9, suggesting the sponge role of MEG3 for miR-9 (Fig. 4D).
The interaction of MEG3 with miR-9, E-cadherin, and FOXO1. A–C, The Eca109 cells were transfected with either miR-9 mimic (25 nmol/L) or mimic control (25 nmol/L) at 24 hours after transfection with the pGL3 luciferase expression vector containing wild-type MEG3 or diverse mutant MEG3 as indicated. After 14 hours, reporter luciferase activity was evaluated. Data are presented as the relative ratio of firefly luciferase activity to renilla luciferase activity. The values are shown relative to the value obtained with mimic control (n = 3; *, P < 0.05). D, The miR-9 expression levels after the transfection of indicated plasmids into Eca109 cells; *, P < 0.05. E, Upregulation of E-cadherin and FOXO1 in YES2 cells after treated with miR-9 inhibitor detected by qRT-PCR. *, P < 0.05. F, Western blot analysis of E-cadherin and FOXO1 protein in YES2 cells after treated with miR-9 inhibitor as described in E. G, The expression of E-cadherin and FOXO1 in YES2 cells transfected with empty vector, wild type MEG3, mutant MEG3, or MEG3 and miR-9 mimic as indicated detected by qRT-PCR. *, P < 0.05. H, Western blot analysis of E-cadherin and FOXO1 protein in YES2 cells as described in G.
MEG3 regulates E-cadherin and FOXO1 expression as ceRNA
We further detected whether MEG3 regulates the expression of other mRNA by competitive binding to miR-9. Studies have reported that miR-9 could promote the progression of tumor metastasis through regulating E-cadherin and FOXO1 expression directly and promoting epithelial–mesenchymal transition (EMT; refs. 27, 28). As epithelial marker, E-cadherin was downregulated in the course of EMT, and FOXO1 was also involved in tumor metastases. Because of the evident effects of MEG3 on ESCC progression and metastasis, we detected the ceRNA role of MEG3 in regulating the expression of E-cadherin and FOXO1. Upregulation of E-cadherin and FOXO1 was detected in YES2 cells after transfected with miR-9 inhibitor by qRT-PCR and Western blot analysis (Fig. 4E and F). As shown in Fig. 4G and H, the expression level of E-cadherin and FOXO1 in YES2 cells with transfected wild type MEG3 was significantly higher than that in the other transfected groups, suggesting the important role of MEG3 in modulating E-cadherin and FOXO1 by competitively binding miR-9.
Survival analysis of MEG3 in ESCC
Kaplan–Meier analysis indicated that downregulation (the expression levels of MEG3 in the tumor tissues were less than 50% of that in corresponding normal tissues) or hypermethylation of MEG3 was significantly associated with poorer ESCC patients' overall survival (log-rank test, P < 0.05; Fig. 5A and B). Compared with the ESCC cases with both high expression and unmethylation of MEG3, ESCC cases with both low expression and hypermethylation of MEG3 demonstrated worse survival rates (Fig. 5C). ESCC patients in stage III and IV, with low expression or hypermethylation of MEG3 also demonstrated worse survival rates and were most prone to develop metastatic disease (Fig. 5D and E). As shown in Fig. 5F, overexpression of miR-9 (the expression levels of miR-9 in the tumor tissues were equal to or more than 2-fold than those in corresponding normal tissues) was significantly associated with poorer ESCC patients' survival. Cox multivariate analysis revealed that MEG3 methylation may be an independent prognostic factor for poor survival of patients with ESCC (Table 2).
Kaplan–Meier univariate survival analysis of MEG3 expression and methylation, miR-9 expression in ESCC cases. A, Showing a direct correlation between low MEG3 expression and poor patient survival. B, Showing a direct correlation between MEG3 methylation and poor patient survival. C, ESCC cases with simultaneous low expression and methylation of MEG3 showing poor patient survival. D, ESCC cases in stage III and IV and with low expression of MEG3 showing poor patient survival. E, ESCC cases in stage III and IV and with methylation of MEG3 showing poor patient survival. F, Showing a direct correlation between high miR-9 expression and poor patient survival.
Multivariate analysis of survival in ESCC cases (Cox's test)
Discussion
MEG3 has been reported to be highly expressed in the adrenal gland, brain, pituitary, ovary, placenta, testes, pancreas, mammary gland, spleen, and liver (29). However, downregulation of MEG3 has been detected in many types of tumor cell lines and primary tumors (17–22, 29–34). In the current study, silenced or decreased expression of MEG3 was found in esophageal cancer cell lines. However, the expression level of MEG3 was significantly increased in esophageal cancer cell lines after treated with 5-Aza-dC. It has been reported that MEG3 gene expression is tightly controlled by at least two differentially methylated regions (DMR): the MEG3-DMR and the intergenic DMR (IG-DMR). The two DMRs are located upstream of the MEG3 gene. Zhao and colleagues (21) found that two 5′-flanking regions, immediately in front of and approximately 1.6–2.1 kb upstream of the first exon (overlaps with the MEG3-DMR), respectively, were hypermethylated in pituitary tumors without MEG3 expression compared with the normal pituitary. Promoter hypermethylation of MEG3 was also observed in epithelial ovarian cancer (17), myeloma (18), acute myelogenous leukemia, and myelodysplastic syndromes (19), meningioma (20), and hepatocellular cancer (30). Downregulation of MEG3 was also found to be associated with abnormal hypermethylation of the MEG3 promoter in these types of carcinomas. Consistent with these studies, we also found promoter hypermethylation of MEG3 in esophageal cancer cell lines and ESCC tissues, suggesting that inactivation of MEG3 in these tumors partially result from promoter hypermethylation of the gene.
Except for methylation analysis, miRNA-dependent regulation of MEG3 expression was also studied in recent years. Deregulated miR-29 could represent a mechanism for the silencing of MEG3 in hepatocellular cancer through modulation of DNMTs (31), miR-26a may target the DNMT3B transcript and its inhibition may result in the upregulation of MEG3 in tongue squamous cell carcinoma (TSCC) tissues (34). Peng et al found that MEG3 inhibited gastric cancer cell proliferation, migration and invasion by operating as a ceRNA for the miR-181 family, and Bcl-2 was validated as a downstream target of MEG3 ceRNA function (35). In the present study, we found that MEG3 may act as a ceRNA to regulate expression of E-cadherin and FOXO1 by competitively binding miR-9 in ESCC. These results indicate that MEG3 may regulate ESCC carcinogenesis as a ceRNA and may be used as a potential target for antitumor therapy.
To highlight the impact of dysregulated expression and methylation of MEG3, we further detected the critical role of MEG3 in the prognosis of ESCC. It has been reported that non–small cell lung cancer (NSCLC) patients with lower expression levels of MEG3 had a relatively poor prognosis (33); TSCC patients with combined low expression levels of both MEG3 and miR-26a demonstrated poor clinical outcome (34). We found that downregulation or hypermethylation of MEG3, over expression of miR-9 was significantly associated with poorer ESCC patients' overall survival. Thus, MEG3 expression and hypermethylation may be useful markers for ESCC tumor progression and prognosis; however, further studies with larger sample size are needed to verify the results.
In conclusion, MEG3 may act as a tumor-suppressor gene and aberrant promoter hypermethylation is critical for MEG3 gene silencing in ESCC. In addition, MEG3 may act as a ceRNA to regulate expression of E-cadherin and FOXO1 by competitively binding miR-9. Furthermore, MEG3 may be used as a potential biomarker in predicting ESCC patients' progression and prognosis, and may be an effective target for antitumor therapies.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: W. Guo
Development of methodology: F. Xu, S. Shen, W. Guo
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): F. Lu, G. Zhang, W. Guo
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Liu, Y. Guo, W. Guo
Writing, review, and/or revision of the manuscript: Z. Dong, A. Zhang, W. Guo
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Zhang, Y. Shi, J. Liang, W. Guo
Study supervision: W. Guo
Grant Support
This study was supported by Grants from the National Natural Science Foundation (no. 81472335 and no. 81572441), Natural Science Foundation of Hebei Province (no. H2015206196 and no. H2015206420).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Footnotes
Note: Supplementary data for this article are available at Molecular Cancer Research Online (http://mcr.aacrjournals.org/).
- Received November 1, 2016.
- Revision received January 23, 2017.
- Accepted February 16, 2017.
- ©2017 American Association for Cancer Research.
References
Volume 15, Issue 7
