Epstein-Barr virus nuclear antigen 1 (EBNA1) protein induction of epithelial-mesenchymal transition in nasopharyngeal carcinoma cells

Authors

  • Lu Wang MD,

    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    Search for more papers by this author
    • The first 2 authors contributed equally to this work.

  • Wen-Dong Tian MD,

    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    Search for more papers by this author
    • The first 2 authors contributed equally to this work.

  • Xia Xu MD,

    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    Search for more papers by this author
  • Biao Nie MD,

    1. Department of Gastroenterology, Nanfang Hospital, Southern Medical University, Guangzhou, China
    Search for more papers by this author
  • Juan Lu MD,

    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    Search for more papers by this author
  • Xiong Liu MD,

    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    Search for more papers by this author
  • Bao Zhang PhD,

    1. Institute of Molecular Biology, Southern Medical University, Guangzhou, China
    Search for more papers by this author
  • Qi Dong PhD,

    1. Stanley Ho Center for Emerging Infectious Diseases, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong, China
    Search for more papers by this author
  • John B. Sunwoo MD,

    1. Department of Otolaryngology-Head and Neck Surgery, School of Medicine, Stanford University, Palo Alto, California
    Search for more papers by this author
  • Gang Li MD,

    Corresponding author
    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    2. Department of Otolaryngology-Head and Neck Surgery, School of Medicine, Stanford University, Palo Alto, California
    • Corresponding authors: Xiang-Ping Li, MD, Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, 1838 Guangzhou Avenue North, Guangzhou 510515, China; Fax: (011) 86-20-616-42001; lp133402@aliyun.com; Gang Li, MD, Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, 1838 Guangzhou Avenue North, Guangzhou 510515, China; Fax: (011) 86-20-616-42001; climber1999g@gmail.com

    Search for more papers by this author
  • Xiang-Ping Li MD

    Corresponding author
    1. Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China
    • Corresponding authors: Xiang-Ping Li, MD, Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, 1838 Guangzhou Avenue North, Guangzhou 510515, China; Fax: (011) 86-20-616-42001; lp133402@aliyun.com; Gang Li, MD, Department of Otolaryngology-Head and Neck Surgery, Nanfang Hospital, Southern Medical University, 1838 Guangzhou Avenue North, Guangzhou 510515, China; Fax: (011) 86-20-616-42001; climber1999g@gmail.com

    Search for more papers by this author

  • We thank Dr. Guang-Hui Xiao for kindly providing the plasmids (plvthm, psPAX2, and pMD2.G) and the core facility of Nanfang Hospital (Guangzhou, China) for their support.

Abstract

BACKGROUND

The Epstein-Barr virus (EBV)-encoded EB nuclear antigen 1 (EBNA1) protein is required for maintenance and transmission of the viral episome in EBV-infected cells. The objective of this study was to investigate the role of EBNA1 protein in nasopharyngeal carcinoma (NPC).

METHODS

Tissue samples from 48 patients with NPC and 12 patients with chronic nasopharyngitis were subjected to immunohistochemical analysis of EBNA1 expression. EBNA1 combinational DNA was used to overexpress EBNA1 protein in NPC cell lines to assess tumor cell epithelial-mesenchymal transition (EMT), colony formation, migration and invasion, and gene expression.

RESULTS

EBNA1 protein was highly expressed in NPC tissue specimens, and its expression was associated with NPC lymph node metastasis. EBNA1 expression affected NPC cell morphology and the expression of EMT markers in vitro. Furthermore, overexpression of EBNA1 inhibited the expression of microRNA 200a (miR-200a) and miR-200b and, in turn, up-regulated expression of their target genes, zinc finger E-box binding homeobox 1 ( ZEB1) and ZEB2, which are well known mediators of EMT. In addition, EBNA1-regulated miR-200a and miR-200b expression was mediated by transforming growth factor-β1.

CONCLUSIONS

The current findings provided novel insight into the vital role of EBNA1 in manipulating a molecular switch of EMT in EBV-positive NPC cells. Cancer 2014;120:363–372. © 2013 American Cancer Society.

INTRODUCTION

Nasopharyngeal carcinoma (NPC) is the most common endemic cancer in southern China and Southeast Asia. Early metastasis is the distinctive feature of NPC, and diagnosis of NPC is often made by lymph node biopsy. Most patients who have NPC present with stage III or IV disease at diagnosis, which contributes to the poor prognosis of NPC.[1, 2] Epstein-Barr virus (EBV) is the most significant risk factor for NPC, especially in endemic areas like southern China and Southeast Asia.[3] EBV infection identified in NPC is mainly the type II latent infection in which EBV only expresses EB nuclear antigen 1 (EBNA1), latent membrane protein-1 (LMP-1), LMP-2, EBV-encoded RNAs (EBERs), and BamHI-A rightward transcript (BART) microRNAs (miRNAs).[3, 4] Emerging evidence has demonstrated that virus-encoded genes play a crucial role in the regulation of epithelial-mesenchymal transition (EMT).[5] Recent studies have suggested that LMP1 and LMP2A induce EMT in NPC through Snail, Twist, and the transforming growth factor beta (TGF-β) signaling pathway.[6-8] EMT triggers tumor cell invasion and dissemination and, thus, is considered the initiating step in cancer metastasis.[9, 10] EBNA1 is a multifunctional and dimeric viral protein that is detected in latent infections and in all patients who have NPC with EBV infection.[11, 12] Moreover, EBNA1 is indispensable for viral replication, genome maintenance, and viral gene expression. Although EBNA1 is a well characterized protein, its role in carcinogenesis—and especially in EMT—is less well defined.[13, 14] Thus, in this study, we explored the role of EBNA1 in NPC cell phenotypes and gene expression levels.

MATERIALS AND METHODS

Tissue Specimens and Cell Lines

This study was approved by the ethics committee of Nanfang Hospital. Specimens from 48 patients with NPC and 12 patients with chronic nasopharyngitis were collected from the NPC tissue bank of Nanfang Hospital. Written informed consent was obtained from each patient. The tissue specimens were routinely fixed in 10% formalin, embedded in paraffin, and sectioned according to routine procedures.

The human NPC CNE1, CNE2, and 5-8F cell lines were used in our previous study.[15] In brief, the CNE1 cell line was derived from squamous cell carcinoma and is EBV-negative; the CNE2 cell line was derived from undifferentiated carcinoma tissue; and the 5-8F cell line was produced subsequently from SUNE1 cells, which were derived from undifferentiated carcinoma tissue. The early passages of both CNE2 cells and 5-8F cells were positive for EBV; however, after long-term in vitro culture, they lost EBV expression.[16] Thus, all 3 cell lines were EBV-negative.

Immunohistochemistry

Tissue sections were deparaffinized and rehydrated. Antigen retrieval was performed by using 0.05 M citrate buffer, pH 6.0, at 100°C for 20 minutes. Next, the sections were stained with an antibody against human EBNA1, cytokeratin 18 (CK18), or vimentin. Staining was repeated at least twice in sequential sections from the same tissue blocks, and the sections were reviewed by 2 pathologists and classified using the World Health Organization criteria for NPC. The ratio of positive cells per specimen was evaluated quantitatively and scored as follows: 0 for staining in ≤1% of cells, 1 for staining in 2% to 25% of cells, 2 for staining in 26% to 50% of cells, 3 for staining in 51% to 75% of cells, and 4 for staining in ≥75% of cells. Staining intensity was scored as 0, no staining; 1, weak staining; 2, moderate staining; and 3, strong staining. A total score from 0 to 12 was finally calculated and graded as negative (−; score, 0-1), weak (+; score, 2-4), moderate (++; score, 5-8), and strong (+++; score 9-12).

EBNA1 Overexpression

Plasmid pCEP4 containing EBV replication initiation Orip and EBNA1 cDNA was purchased from Invitrogen (Carlsbad, Calif). Lipofectamine 2000 (Invitrogen) was used for pCEP4 transfection. Hygromycin B was added to the cell culture medium (150 μg/mL) to select cells with stable EBNA1 expression for the cell population used in our experiments.

Lentiviral Production and Tumor Cell Infection

To produce a lentivirus that carried EBNA1 short hairpin RNA (shRNA), a double-stranded DNA encoding EBNA1 shRNA EBNA1 (shEBNA1) was synthesized and cloned into the lentiviral vector pLVTHM (pLVTHM/shEBNA1), which contained green fluorescent protein (GFP) cDNA. Human embryonic kidney 293t cells were cotransfected with pLVTHM or pLVTHM/shRNA EBNA1 plus helper vectors (psPAX2 and pMD2.G) using Lipofectamine 2000 according to the manufacturer's instructions. Next, lentivirus carrying control shRNA (control) or shEBNA1 was harvested from 293t cells, and the multiplicity of infection was tested before it was used for tumor cell infection.

These lentiviral particles were then used to infect the CNE1, CNE2, and 5-8F NPC cell lines and generated EBNA1-overexpressed cells (CNE1/EBNA1, CNE2/EBNA1, and 5-8F/EBNA1, respectively) using flow cytometry to sort GFP. After 72 infections, the cells were grown to obtain stable cell sublines, which were named LV-control, EBNA1/LV-control, and EBNA1/LV-shEBNA1, respectively.

Quantitative Reverse-Transcriptase Polymerase Chain Reaction Analysis

To detect gene expression, quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) was performed using an ABI 7500 PCR system (Applied Biosystems, Foster City, Calif) with the FastStart DNA Master SYBR Green I Kit (Roche Applied Science, Indianapolis, Ind). Messenger RNA (mRNA) expression levels were then normalized to the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or U6 and were compared with expression levels in the control NPC cells. The experiments from each sample were performed in triplicate and were repeated 2 or 3 times.

Protein Extraction and Western Blot Analysis

Total cellular protein was extracted from cells using a radioimmunoprecipitation assay buffer plus leupeptin and aprotinin. After quantification, 40 μg of the cell lysates were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred onto polyvinylidene difluoride membranes (Millipore, Bedford, Mass), and incubated with 5% dry milk in Tris-buffered saline plus Tween 20 for 1 hour at room temperature. The target protein levels were measured by immunoblotting with the corresponding antibodies for 2 hours. The secondary antibodies were used at 1:1000 dilutions and incubated for 1 hour at room temperature. A Pierce enhanced chemiluminescence plus Western blotting substrate kit (Thermo Scientific, Rockford, Ill) was used to observe the positive band.

Tumor Cell Wound-Healing and Boyden Chamber Invasion Assays

Cells were seeded into a 6-well cell culture plate and cultured to 100% confluence. The confluent cells were then scraped with a 1-mL pipette tip to create a wound. The wound width was monitored at 12 hours and 24 hours after scratching.

We assessed tumor cell invasion capacity using a BD BioCoat Matrigel Invasion Chamber (Becton Dickinson, Mountain View, Calif) according to the manufacturer's instructions. In brief, cells (1 × 104 per well for EBNA1-overexpressed cells and 5 × 104 per well for EBNA1 knocked down cells) were resuspended in 100 μL serum-free medium plus 0.2% bovine serum albumin and were added to the top chamber, and 200 μL of normal growth medium containing 20% fetal bovine serum were added to the bottom chamber. After 24 to 48 hours, the cells on the top surface of the membrane were swabbed away using a cotton swab, and the cells under the bottom membrane were stained with crystal violet.

Tumor Cell Clone-Formation Assay

Cells (500 cells per well for EBNA1-overexpressed cells and 1 × 103 per well for EBNA1 knocked down cells) were seeded into each well of a 6-well culture plate in triplicate and cultured at 37°C for 14 days, and the medium was refreshed every 3 days. Cells were then washed with phosphate-buffered saline and stained with Giemsa stain. The number of colonies that contained >50 cells was counted under a microscope and was summarized using the following formula: plate clone formation efficiency = (number of colonies / number of cells inoculated) × 100%.

Immunofluorescence Assay

Cells were cultured on chamber slides for 48 hours, fixed with methanol for 15 minutes, permeabilized with 0.2% Triton X-100 for 10 to 15 minutes, then incubated with different primary antibodies (see Protein Extraction and Western Blot Analysis, above), and incubated further with an antirabbit immunoglobulin G (IgG)/DyLight594 or antimouse IgG/DyLight594 antibody (Zhongshan Goldenbridege Biotechnology Inc., Beijing, China). Positive staining was analyzed under a confocal microscope (TCS SP2 AOBS; Leica Microsystems UK, Milton Keynes, United Kingdom).

Transfection of MicroRNA Precursors

Stably infected CNE1/EBNA1-positive cell lines were seeded at 2 × 105 cells per well in 6-well plates and transfected with synthetic miRNA precursors (pre-miR) (miRNA Precursor Molecules; Ambion, Austin, Tex) using siPORT NeoFX Transfection Agent (Ambion). The final concentrations of each pre-miR were 20 nM for pre-miR-200a or pre-miR-200b and 40 nM (20 nM pre-miR-200a and 20 nM pre-miR-200b). Cells transfected with 20 nM pre-miR miRNA Precursor Negative Control No. 1 from Ambion were used as a negative control.

Statistical Analysis

Immunohistochemical data were analyzed using the Mann-Whitney U test and the Kruskal-Wallis test. Expression levels of EBNA1, CK18, and vimentin in NPC were compared using Spearman correlation analysis. A 2-way independent samples t test was performed to compare qRT-PCR, invasion, colony formation, Western blot, and immunofluorescence data. Wound-healing data were analyzed with a 2-way, factorial design analysis of variance. All statistical analyses were performed using the SPSS 13.0 statistical software package (SPSS Inc., Chicago, Ill). P values < .05 were considered statistically significant.

RESULTS

Association of EBNA1 Expression With Nasopharyngeal Carcinoma Lymph Node Metastasis and Epithelial-Mesenchymal Transition Markers

EBNA1 protein was expressed in 42 of 48 (87.5%) NPC tissues but was not detectable in chronic nasopharyngitis samples (0 of 12 tissues). EBNA1 expression was associated with NPC clinical stage (P = .000), pathologic stage (P = .018), and lymph node metastasis (P = .003) (Table 1). Furthermore, EBNA1 expression was inversely associated with CK18 expression (r = −0.331; P = .021) but was associated with vimentin expression (r = 0.311; P = .031) (Fig. 1). Expression of CK18 and vimentin proteins was inversely associated in these NPC tissues (r = −0.329; P = .022).

Table 1. Association of Epstein-Barr Virus Nuclear Antigen 1 Expression and Clinicopathologic Features in Patients With Nasopharyngeal Carcinoma and Nasopharyngitis
 EBNA1 Expression: No. of Patientsa 
Clinicopathologic Feature++++++P
  1. Abbreviations: EBNA1, Epstein-Barr virus nuclear antigen 1; NPC, nasopharyngeal carcinoma

  2. a

    Immunohistochemical staining was scored as negative (−) (score, 0–1), weak (+) (score, 2–4), moderate (++) (score, 5–8), and strong (+++) (score, 9–12).

Nasopharyngitis (n = 12)12000.000
NPC (n = 48)63273 
Pathologic tumor classification     
T1 (n = 6)3111.018
T2 (n = 17)21500 
T3 (n = 11)1901 
T4 (n = 14)0662 
Pathologic lymph node classification     
N0 (n = 11)4700.003
N1 (n = 15)01320 
N2 (n = 14)2831 
N3 (n = 8)0323 
Pathologic disease stage     
I–II (n = 14)31010.000
III (n = 16)31300 
IV (n = 18)0864 
Figure 1.

Immunohistochemical analyses of Epstein-Barr virus nuclear antigen 1 (EBNA1), cytokeratin 18 (CK18), and vimentin in nasopharyngeal carcinoma (NPC) tissues are illustrated. (A) Photomicrographs illustrate immunohistochemical analyses of (a,d,g) EBNA1, (b,e,h), CK18, and (c,f,i) vimentin in tissue samples from 3 patients with NPC (original magnification ×400). (B) Correlations between EBNA1 and CK18 expression and between EBNA1 and vimentin expression are illustrated from an analysis of 48 NPC tissue specimens.

EBNA1 Expression Induced Morphologic Transformation of Nasopharyngeal Cells

Western blot analysis confirmed that EBNA1 proteins were expressed in CNE1/EBNA1, CNE2/EBNA1, and 5-8F/EBNA1 cells; whereas no EBNA1 proteins were detected in CNE1, CNE2, or 5-8F cells (Fig. 2A); in shRNA experiments, the expression of EBNA1 was decreased by 56.47%, 45.40%, and 49.53%, respectively (Fig. 2B). Ectopic expression of EBNA1 resulted in morphologic changes in all 3 NPC cell lines. Particularly, the original round or polygonal 5-8F and CNE1 cells became fibroblast-like, with a narrow spindle shape and a long lamellipodium, whereas CNE2 cells exhibited the appearance of polygonal cells with an elongated filopod, similar to neurons with an axon (Fig. 2C). However, knockdown of EBNA1 expression reversed the morphologic alterations in these EBNA1-transfected NPC cells (Fig. 2D).

Figure 2.

Changes in cell morphology are illustrated after Epstein-Barr virus nuclear antigen 1 (EBNA1) transfection in nasopharyngeal carcinoma (NPC) cells. (A) Western blot analysis revealed EBNA1 expression in NPC parental cell lines (CNE1, CNE2, and 5-8F) and in EBNA1-overexpressed cells. (B) The expression of EBNA1 after short hairpin RNA (shRNA) knockdown was detected using Western blot analysis. Lane 1, lentivirus (LV)-control; lane 2, EBNA1/LV-control; lane 3, EBNA1/LV-shEBNA1. The results shown are representative of 3 experiments with similar results. (C) Changes in cell morphology are illustrated after overexpression of EBNA1 in CNE1, CNE2, and 5-8F cells. Cells were grown and transfected with EBNA1 combinational DNA, treated with antihygromycin B, and the altered cell morphology was then recorded under an inverted microscope (original magnification ×200). (D) The effect of EBNA1 shRNA knockdown on the regulation of tumor cell morphology is illustrated. GFP indicates green fluorescent protein (original magnification ×200.

EBNA1 Induction of Epithelial-Mesenchymal Transition in Nasopharyngeal Carcinoma Cells

EBNA1 expression reduced the transcription and expression of epithelial cell markers and increased the expression of mesenchymal cell markers (Figs. 3, 4). Specifically, levels of E-cadherin and CK18 mRNA in the CNE1, CNE2, and 5-8F cell lines were decreased by 67.70% and 85.24%, by 62.74%, and 40.42%, and by 81.81% and 96.94%, respectively; whereas the level of vimentin mRNA was increased 1.65-fold, 2.08-fold, and 5.40-fold in these cell lines, respectively. Similarly, levels of β-catenin and N-cadherin mRNA also increased by 1.64-fold and 2.69-fold, by 1.67-fold and 5.28-fold, and by 2.71-fold and 3.18-fold in the respective cell lines (Fig. 3A-C). Figure 3D indicates that expression of E-cadherin and CK18 was decreased, whereas N-cadherin and vimentin expression was increased in overexpressing EBNA1 cells. Immunofluorescence and confocal images (Fig. 4A) were consistent with Western blot data. Furthermore, expression of β-catenin protein was localized on the cell membrane of NPC cells that expressed low levels of EBNA1, whereas β-catenin protein was localized in the nuclei of NPC cells that expressed high levels of EBNA1 (Fig. 4B), supporting the notion that EBNA1 expression promotes EMT conversion.

Figure 3.

Epstein-Barr virus nuclear antigen 1 (EBNA1)-induced expression of epithelial-mesenchymal transition regulatory proteins in nasopharyngeal carcinoma cells is illustrated. The expression of epithelial and mesenchymal markers was determined by quantitative reverse-transcriptase polymerase chain reaction and Western blot analyses in (A) CNE1 cells, (B) CNE2 cells, and (C) 5-8F cells with and without EBNA1 transfection. CK18 indicates cytokeratin 18. A single asterisk indicates P < .05; double asterisks, P < .01. (D) Western blot analysis was used to identify epithelial and mesenchymal cell markers in all 3 cancer cell lines. The results shown are representative of 3 experiments.

Figure 4.

Epstein-Barr virus nuclear antigen 1 (EBNA1)-induced expression of epithelial-mesenchymal transition regulatory proteins was investigated in 3 nasopharyngeal carcinoma cell lines (CNE1, CNE2, and 5-8F) using immunofluorescence. (A) Micrographs reveal the expression of epithelial and mesenchymal cell markers (original magnification ×60). CK18 indicates cytokeratin 18. (B) The nuclear accumulation and distribution of β-catenin protein in NPC cells are observed after cells were transfected with a vector carrying EBNA1-transfected cells or control cells (original magnification ×60).

EBNA1 Induction of Tumor Cell Migration and Invasion

Overexpression of EBNA1 protein enhanced CNE1 cell migration by 2.75-fold at 24 hours. Similarly, migration of CNE2/EBNA1 and 5-8F/EBNA1 cells was increased by 1.5-fold and 2.5-fold, respectively (Fig. 5A). Conversely, silencing of EBNA1 expression decreased such capabilities by 50% to 60% in these cells (Fig. 5B).

Figure 5.

Epstein-Barr virus nuclear antigen 1 (EBNA1) promotion of tumor cell migration and invasion capacity in nasopharyngeal carcinoma (NPC) cells is illustrated. (A) Quantitative measurement of wound gaps using Image J (public domain software; National Institutes of Health, Bethesda, Md) revealed increased cellular motility in EBNA1-overexpressed cells (the human NPC cell lines CNE1, CNE2, and 5-8F) compared with control cells. (B) The results from a wound-healing assay are illustrated after knockdown of EBNA1 expression by short hairpin RNA (shRNA) lentivirus infection. (C) The results from a Matrigel invasion assay are illustrated after EBNA1 overexpression in CNE1, CNE2, and 5-8F cells. (D) The results from a Matrigel invasion assay are illustrated after EBNA1 knockdown by shRNA lentivirus (LV) infection (LV-shEBNA1). (E) Statistical analysis of the colony-formation assay is illustrated. Colonies were evaluated between EBNA1-negative and EBNA1-positive CNE1 cells, CNE2 cells, and 5-8F cells. (F) These are results from the colony-formation assay after NPC cells were infected with the EBNA1 shRNA virus. All experiments were replicated 3 times. A single asterisk indicates P < .05; double asterisks, P < .01.

Furthermore, CNE2/EBNA1 and 5-8F/EBNA1 cells also exhibited a dramatic increase in tumor invasion capacity compared with control cells. Knockdown of EBNA1 expression using LV-shEBNA1 decreased tumor cell invasion capacity by 66% to 81%. Parental CNE1 cells are nonmetastatic, and the overexpression of EBNA1 protein induced CNE1 cells to acquire a more aggressive phenotype by invading through the Matrigel (Fig. 5C,D).

EBNA1 Expression Increased Tumor Cell Colony Formation

We also assessed the effects of EBNA1 expression on the regulation of NPC cell colony formation. Colony numbers of CNE1, CNE2, and 5-8F cells after EBNA1 overexpression were increased by 2.44-fold, 2.00-fold, and 2.10-fold, respectively (Fig. 5E). In contrast, knockdown of EBNA1 expression by shEBNA1 decreased cell colony formation by 53.4%, 48.7%, and 64.2% in CNE1/EBNA1, CNE2/EBNA1, and 5-8F/EBNA1 cells, respectively (Fig. 5F).

EBNA1 Induction of Nasopharyngeal Cell Epithelial-Mesenchymal Transition by Regulation of the TGF-β1/miR-200/ZEB Pathway

To explore the underlying molecular mechanisms responsible for EBNA1-induced changes in gene expression and EMT, we examined expression of the zinc finger E-box binding homeobox (ZEB)/miR-200 genes, because this feedback loop has been regarded as 1 of the crucial signaling pathways in the molecular cascade of EMT. qRT-PCR data demonstrated that expression levels of miR-200a and miR-200b were significantly reduced after EBNA1 transfection, whereas the target genes of miR-200a and miR-200b were up-regulated by EBNA1 transfection (Fig. 6A).

Figure 6.

Epstein-Barr virus (EBV) nuclear antigen 1 (EBNA1)-regulated expression of epithelial-mesenchymal transition (EMT)-related proteins by activation of the transforming growth factor β1 (TGF-β1)/microRNA-200 (miR-200)/zinc finger E-box binding homeobox (ZEB) pathway is illustrated. (A) The expression of TGF-β1, miR-200, and ZEB is illustrated in human CNE1 nasopharyngeal carcinoma cells with or without EBNA1 as measured by quantitative reverse-transcriptase polymerase chain reaction analysis. A single asterisk indicates P < .05; double asterisks, P < .01. (B) In Western blot analysis, the expression of ZEB1 and ZEB2 proteins was reduced significantly after EBNA1-overexpressed cells (EBNA1+) were transfected with miR-200a and miR-200b. Neg indicates negative. (C) In immunofluorescence experiments, the cells from B were stained for E-cadherin and N-cadherin expression. (D) CNE1 cells were transfected with a control or TGF-β1 small interfering RNA (siRNA) in the presence or absence of EBNA1 and then subjected to quantitative reverse-transcriptase polymerase chain reaction analysis of miR-200a and miR-200b levels. (E) The study hypothesis is illustrated, indicating that the EBNA1-regulated TGF-β1/ZEB/miR-200 signaling pathway promotes EMT in NPC cells. A single asterisk indicates P < .05; double asterisks, P < .01.

Moreover, we performed experiments to regain the balance of the ZEB/miR-200 feedback loop by overexpression of miR-200a and miR-200b in these cells. After transient transfection of miRNA mimics, levels of miR-200a and miR-200b were increased by 1064-fold and 221-fold, respectively, compared with controls. Meanwhile, ZEB1 and ZEB2 proteins in CNE1/EBNA1 cells were dramatically down-regulated (Fig. 6B). Similarly, expression levels of E-cadherin and N-cadherin in these EBNA1-positive cell lines also were changed, which suggested that miR-200 promoted a reverse EMT in these EBNA1-positive cell lines (Fig. 6C).

We also assessed the role of other EMT regulatory proteins by knockdown of several EMT-inducing transcription factors, such as Twist, AP-1, Bim1, and TGF-β1, in these tumor cells. Our data indicated that only TGF-β1 silencing resulted in increased levels of miR-200a and miR-200b expression in EBNA1-positive NPC cells (Fig. 6D), suggesting that EBNA1 induced EMT at least in part through a TGF-β1–mediated miR-200/ZEB feedback loop (Fig. 6E).

DISCUSSION

Previous studies have demonstrated that the transcriptional factor-like functions of EBNA1 can promote NPC oncogenesis though different pathways, including the oxidative stress response pathway[17] and the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) pathway,[18] or though disruption of promyelocytic leukemia (PML) nuclear bodies (NBs).[19] But these were not enough evidence of its role and underlying mechanism in EMT. Our current data indicated that EBNA1 overexpression in 3 EBV-negative NPC cell lines led to distinct transformations in cell morphology, migration, and invasion. Reversal of epithelial and mesenchymal markers, including E-cadherin, CK18, vimentin, N-cadherin, and β-catenin, clearly demonstrated initiation of the EMT process. β-Catenin relocalization upon EBNA1 overexpression also was observed in NPC cells. Highly expressed β-catenin protein in NPC tissues has been associated with tumor metastasis, and β-catenin relocalization from the cell membrane to the nucleus indeed induces EMT.[20, 21] These results indicate that EBNA1 is able to induce EMT in NPC cell lines in vitro.

Our current data also demonstrated that EBNA1 can induce EMT through the TGF-β/ZEB/miR-200 pathway. It has been demonstrated that the zinc finger E-box binding homeobox (ZEB) family (ZEB1 and ZEB2) is involved in tumor invasion and metastasis and promotes cancer cell invasion by inhibition of miR-200.[22-25] Conversely, miR-200 family members (miR-200a, miR-200b, miR-200c, miR-141, and miR-429) can repress ZEB expression in a feedback manner.[26, 27] Thus, it has been postulated that the ZEB/miR-200 double-negative feedback loop controls the balance between epithelial and mesenchymal states of tumor cells.[28, 29] Furthermore, EMT not only confers cellular motility but also induces stem cell-like properties as well as preventing apoptosis and senescence. These processes provide the basis for cellular plasticity in development and adult tissue homeostasis.[30] Xia et al observed that overexpression of miR-200a not only promotes the transition of mesenchyme-like EBV-positive C666-1 cells to the epithelial state, but it also causes a significant reduction of stem-like cell features.[31] Moreover, those authors suggested that the effect and mechanism of miR-200a on shifting NPC cells occurs through a reversible process, and they coined this process as epithelial-mesenchymal to stem-like transition (EMST).

In the current study, we observed that EBNA1 expression may direct EBV-positive NPC cells toward a more mesenchymal and undifferentiated state. In addition, the induction and maintenance of a stable mesenchymal phenotype requires continuous EBV latent infection.[32] The ZEB/miR-200 feedback loop also plays key roles in regulating the latent-lytic switch of EBV. ZEB1/ZEB2 can repress the EBV lytic-initiate gene BZLF1 in transient transfection assays by directly binding its promoter, which, in turn, maintains EBV latency and EBNA1 expression.[33] Thus, we hypothesize that in NPC pathogenesis, EBV latency, EMT, and cell stemness are closely related to each other through the ZEB/miR-200 molecular motor, thereby forming a vicious cycle. EBNA1 may keep this vicious cycle running by driving the ZEB/miR-200 molecular motor and contribute to the early metastasis, high recurrence rate, and therapy resistance of NPC.

However, several questions remain to be answered. TGF-β is a multifunctional cytokine that plays critical roles in EMT[33] though induction of the δEF1/ZEB1, SIP1/ZEB2, and Snail/SNAI1 signaling pathways.[34] The induction and maintenance of a stable mesenchymal phenotype requires the establishment of autocrine TGF-β signaling to drive sustained ZEB expression, and manipulation of the ZEB/miR-200 balance is able to repeatedly switch cells between epithelial and mesenchymal states.[32]

A previous study reported that EBNA1 was able to suppress TGF-β1 in both C666-1 NPC and Hodgkin lymphoma cells.[35, 36] However, our current study demonstrated that overexpression of EBNA1 protein in these 3 NPC cell lines up-regulated the expression of TGF-β1 protein and decreased miR-200a and miR-200b expression, but it significantly increased ZEB1 and ZEB2 expression; whereas silencing of EBNA1 in NPC cells partially reversed this gene pathway by the up-regulation of TGF-β1 expression. The reason for this discrepancy is unclear; however, as a multifunction nuclear protein, EBNA1 could either increase or decrease TGF-β in different systems. Despite of its role in episome maintenance, EBNA1 also induces EBV reactivation by inducing the loss of PML NBs immediately after the lytic cycle is initiated.[19] It is possible that, under certain circumstances, the existence of EBNA1 may induce EMT through TGF-β but counter the EBV lytic reactivation function of TGF-β at the same time.

Apparently, miR-200/ZEB1 is not the only signaling pathway that mediates the metastasis-promotion function of EBNA1 in NPC cells. EBNA1 up-regulated 3 proteins that affect metastatic potential (stathmin 1, maspin, and Nm23-H1) and several proteins in the oxidative stress response pathway, including the antioxidants superoxide dismutase 1 (SOD1) and peroxiredoxin 1 (Prx1).[17] Moreover, through a cross-talking of multiple signaling pathways, including Src, Ras, integrins, Wnt/-catenin, and Notch, EBNA1 continues to drive the reverse differentiation, which, in turn, initiates chain reactions in metastasis cascades.[37]

In this study, we demonstrated for the first time that EBNA1 expression can initiate the transition of epithelium-like NPC cells to a more mesenchymal phenotype and promote NPC cell migration and invasion. Furthermore, the expression of EBNA1 protein was associated with NPC lymph node metastasis and with the expression of EMT regulatory genes in ex vivo experiments. These observations extend the understanding of EBV infection in NPC carcinogenesis and/or progression and also implicate miR-200 as a therapeutic target in NPC.

FUNDING SUPPORT

This study was supported in part by grants from the National Natural Science Foundation of China to Xiang-Ping Li (no. 30772469 and no. 81172053) and Gang Li (no. 30801381) and from the Natural Science Foundation of Guangdong Province, China to Wen-Dong Tian (no. 10151051501000068).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary