EBV-miR-BART7-3p promotes the EMT and metastasis of nasopharyngeal carcinoma cells by suppressing the tumor suppressor PTEN

Oncogene (2015) 34, 2156 2166 2015 Macmillan Publishers Limited All rights reserved 0950-9232/15 www.nature.com/onc ORIGINAL ARTICLE EBV-miR-BART7-3p promotes the EMT and metastasis of nasopharyngeal carcinoma cells by suppressing the tumor suppressor PTEN L-M Cai 1,8, X-M Lyu 1,2,8, W-R Luo 1,8, X-F Cui 3,8, Y-F Ye 1, C-C Yuan 1, Q-X Peng 4, D-H Wu 5, T-F Liu 6, E Wang 7, F-M Marincola 7, K-T Yao 1, W-Y Fang 1, H-B Cai 4 and X Li 1 The epithelial-mesenchymal transition (EMT) is crucial to cancer progression and metastasis. Although multiple cellular mirnas have been identified to regulate the EMT and metastasis in cancers, the role of viral mirnas in cancer progression remains largely unknown. Nasopharyngeal carcinoma (NPC) is an Epstein-Barr virus (EBV)-associated malignancy typically characterized by its early metastasis. In the present study, we have discovered the involvement of a viral mirna, EBV-miR-BART7-3p, in the EMT and metastasis of NPC cells. Initially, we observed that EBV-miR-BART7-3p was highly expressed in NPC and positively correlated with lymph node metastasis and clinical stage of NPC. Subsequently, we demonstrated that EBV-miR-BART7-3p enhanced cell migration/ invasion in vitro, cancer metastasis in vivo, and particularly the EMT characterized by loss of epithelial markers and gain of mesenchymal features in NPC cells. Furthermore, mechanistic studies disclosed that EBV-miR-BART7-3p targeted a major human tumor suppressor PTEN, modulating PI3K/Akt/GSK-3β signaling and eventually leading to the high expression and nuclear accumulation of Snail and β-catenin, which favor EMT. Knockdown of PTEN could phenocopy the effect of EBV-miR-BART7-3p, whereas re-expression of PTEN resulted in a phenotypic reversion. Moreover, these findings were supported by an observation of an EBV-positive cell model in which silencing of endogenous EBV-miR-BART7-3p partially attenuated cell migration/invasion and altered EMT protein expression pattern via reverting PI3K/Akt, Snail and β-catenin expression. Thus, this study suggests a novel mechanism by which EBV-miR-BART7-3p modulates the EMT and metastasis of NPC cells, and a clinical implication of EBV-miR- BART7-3p as a potential biomarker or therapeutic target. Oncogene (2015) 34, 2156 2166; doi:10.1038/onc.2014.341; published online 27 October 2014 INTRODUCTION Metastasis is majorly responsible for the recurrence, poor prognosis and death of cancer. The epithelial-mesenchymal transition (EMT) is crucial to the acquisition of metastatic potential in cancer cells. The initial stage of metastatic progression is essentially dependent on EMT, which is characterized by the loss of epithelial markers and the gain of mesenchymal markers as well as a fundamental change in cellular morphology and phenotype with increased ability to migrate. 1,2 Epithelial cell cell junctions are deconstructed and some components of adherens junctions such as β-catenin are degraded and/or re-localized in the initiation of EMT. 3 Several transcription factors including zinc finger proteins, such as Snail, provide important crosstalk between the relevant signaling pathways 4 or directly regulate EMT-associated genes and trigger EMT. 5,6 As understanding of the molecular processes that govern EMT increases, the ability to target EMT multiple stages for therapy will provide a promise in the treatment of epithelial malignancy and metastasis. MiRNAs exert a variety of important regulatory functions related to multiple processes. The aberrant expression of many mirnas has been linked to human diseases including cancers. 7 To date, multiple cellular mirnas, such as mir-148a, 8 mir-155, 9 mir-377, 10 mir-92a, 11 mir-200c, 12 mir-23a, 13 mir-130b, 14 mir-30a, 15 mir-149 16 and mir-93 17 have been reported to modulate the EMT and metastasis in various cancers, but only three of them are implicated in the EMT in nasopharyngeal carcinoma (NPC), a malignancy typically featured by its early metastasis and invasion. 17 19 Therefore, it is necessary to identify other unknown mirnas that regulate the EMT and metastasis of NPC cells. It is estimated that viral infections contribute to 15 20% of all human cancers. 20,21 Oncogenic viruses are involved in the different stages of carcinogenesis, and the relationship of a virus with a given cancer can be anywhere from 15 100%. 22 Epstein- Barr virus (EBV) infection has been long believed to be a major etiologic factor of NPC. This virus expresses more than 40 mature mirnas in latently infected cells. Of them, some mirnas encoded by BamH I-A rightward transcripts (BARTs) of EBV are disclosed to be highly expressed in some cancers such as gastric cancer, 23 lymphomas 24 and NPC, 25 27 contributing to viral latency, 28 30 host cell survival, 30,31 immune escape, 32 cell proliferation, 33 cell 1 Cancer Research Institute, Southern Medical University, Guangzhou, China; 2 Department of Laboratory Medicine, the Third Affiliated Hospital, Southern Medical University, Guangzhou, China; 3 Department of ENT, 463 Hospital of the Chinese PLA, Shenyang, China; 4 School of Chinese Traditional Medicine, Southern Medical University, Guangzhou, China; 5 Department of Radiation Oncology, Nanfang Hospital, Southern Medical University, Guangzhou, China; 6 Department of Pathology, Southern Medical University, Guangzhou, China and 7 Infectious Diseases and Immunogenetics Section, DTM, Clinical Center, the National Institutes of Health, Bethesda, MD, USA. Correspondence: Professor X Li or Professor W-Y Fang, Cancer research institute, Southern medical university, Guangzhou 510515, China or Associate Professor H-B Cai, School of Chinese Traditional Medicine, Southern medical university, Guangzhou 510515, China. E-mail: xinli268@gmail.com or fangweiyi1975@yahoo.com.cn or chbing2008@126.com 8 These authors contributed equally to this work. Received 5 February 2014; revised 1 July 2014; accepted 12 September 2014; published online 27 October 2014

apoptosis 31,34 and cancer metabolism. 35 However, there have been few reports of the relationship between viral mirnas and EMT in cancer to date. In the present study, we have observed that the high expression of EBV-miR-BART7-3p is positively related to the degree of spread to regional lymph nodes (N stage) and clinical stage of NPC, and disclosed the effects of EBV-miR-BART7-3p on the EMT and metastasis of NPC cells in vitro and in vivo. Most importantly, we have discovered that EBV-miR-BART7-3p can specifically target a major human tumor suppressor PTEN, consequently regulating PI3K/Akt/GSK-3β signaling and ultimately leading to the EMT and metastasis of NPC cells. This study reports a novel mechanism by which EBV-miR-BART7-3p modulates the EMT and metastasis of NPC cells. RESULTS Highly expressed EBV-miR-BART7-3p enhances the metastasis of NPC cells In the present study, we initially conducted a mirna expression profiling analysis in 16 NPC tissue samples and 20 non-cancerous 2157 nasopharyngeal tissue (NP) samples (Supplementary Table S1). Along with several EBV-miR-BARTs, EBV-miR-BART7-3p was found to be highly expressed in NPC samples relative to NP samples (Figure 1a, Supplementary Table S2). This was consistent with a previous study by Wong et al. 36 Furthermore, in additional 42 NPC patients, we observed a positive correlation of EBV-miR-BART7-3p expression with the degree of spread to regional lymph nodes (N stage) and clinical stage of NPC (Figure 1b), hinting the relevance of EBV-miR-BART7-3p to NPC progression and metastasis. Next, using lentiviral particles carrying EBV-miR-BART7-3p precursor, we generated two EBV-negative NPC cell lines stably expressing EBV-miR-BART7-3p (CNE1-BART7-3p and 5-8F-BART7-3p cells). The expression levels of EBV-miR-BART7-3p in these two stable cells were within a similar physiological range to pooled NPC tissue samples (Supplementary Figure S1). Notably, in vitro migration assays revealed that the exogenic expression of EBV-miR-BART7-3p significantly enhanced the migration of these two cell lines (Figure 1c), consistent with the results of the invasion assays using a matrigel invasion chamber system (Supplementary Figure S2a). Conversely, silencing of EBV-miR- BART7-3p attenuated the migration and invasion of NPC cells Figure 1. Highly expressed EBV-miR-BART7-3p enhances the metastasis of NPC cells. (a) Microarray analysis showing EBV-miR-BARTs differentially expressed between NPC and NP samples. (b) EBV-miR-BART7-3p expression normalized to U6 snrna was detected by qpcr in NPC samples with clinical stage and N stage. The data were shown as the mean ± s.e.m. (*Po0.05; ***Po0.001). (c) 5-8 F-BART7-3p, CNE1- BART7-3p and their corresponding NC cells were applied to in vitro migration assay. Moreover, after transfecting anti-mir or anti-c into 5-8 F-BART7-3p and CNE1-BART7-3p cells, respectively, cells were applied to in vitro migration assay. The data were shown as the mean ± s.e.m. (**Po0.01, ***Po0.001). (d) 5-8F-BART7-3p cells or 5-8F-NC cells (control cells) were inoculated under the liver capsules of nude mice. The fluorescent images were displayed for two representative mouse models. Xenograft tumor and lymph node metastases were marked by the red arrows and white arrows, respectively. The number of lymph node metastases in two mouse groups were shown in the bar chart (n = 6 per group. Mean ± s.e.m.s, 0.6667 ± 0.6667 and 5.667 ± 1.687 for 5-8F-NC and 5-8F-BART7-3p group, respectively. Po0.05). 2015 Macmillan Publishers Limited Oncogene (2015) 2156 2166

2158 stably expressing EBV-miR-BART7-3p (Figure 1c, Supplementary Figure S2b). To further validate the role of EBV-miR-BART7-3p in NPC metastasis, we established the mouse models of NPC metastasis. 5-8F cells stably expressing EBV-miR-BART7-3p (5-8F-BART7-3p) and its control cells (5-8F-NC) were transplanted under the liver capsule of mice. All mice were killed in 3 weeks. Notably, the fluorescent image detection confirmed that lymph node metastases dominantly appeared in 5-8F-BART7-3p cell-injected mouse models compared with the control mice, which was consistent with the high lymphatic metastases in clinical NPC patients. 37 Lymph node metastases were present in 83% (5/6) of 5-8F-BART7-3p-injected mice and 17% (1/6) of the control mice. The average number of metastatic lymph nodes was significantly bigger in 5-8F-BART7-3p cell-injected mice (5.667 ± 1.687) than that of control mice (0.6667 ± 0.6667) (Po0.05) (Figure 1d, Supplementary Figures S3a, b). Thus, the above results collectively indicate that EBV-miR- BART7-3p contributes to the metastasis of NPC cells. EBV-miR-BART7-3p promotes the EMT in NPC cells As EMT has been regarded as an important mechanism that augments tumor cell motility and leads to metastasis, we wonder whether EBV-miR-BART7-3p is involved in the EMT of NPC cells. Using in vitro gain-of-function analysis, we firstly observed that the introduction of EBV-miR-BART7-3p led to the deceased expression of epithelial marker (E-cadherin) and increased expression of mesenchymal markers (vimentin and fibronectin) in two NPC cell lines (Figure 2a) along with the EMT-like morphological changes identified by relatively losing cell-to-cell adhesions and acquiring spindle-like phenotype (Figure 2b). Relatively high mrna expression levels of some other EMT-associated genes were also observed upon exgenic EBV-miR-BART7-3p expression (Supplementary Figure S4). Furthermore, silencing of EBV-miR- BART7-3p reversed the changes in cellular shape and EMT protein expression pattern of two NPC cell lines stably expressing EBVmiR-BART7-3p (Figures 2a and b). Consistently, using immunohistochemistry assay to detect the expression of EMT markers in tumor tissues derived from the mouse models of NPC metastasis, we observed the loss of E-cadherin and the increased expression of vimentin and fibronectin in EBV-miR-BART7-3p-induced tumor tissues relative to control tumor tissues. The cell contact loss and spindle-like phenotype were more easily observed in EBV-miR-BART7-3pinduced tumor tissues (Figure 2c). Taken together, these data suggest that EBV-miR-BART7-3p promotes a transition from epithelial to mesenchymal phenotype and may contribute to cancer metastasis by inducing EMT. EBV-miR-BART7-3p directly targets human PTEN gene We further explored how EBV-miR-BART7-3p could promote EMT and cancer metastasis. Predicted target genes of EBV-miR-BART7-3p were firstly retrieved from TargetScan and RNAhybird. Phosphatase and tensin homolog (PTEN) was one of the potential candidates, which gained our attention because it is one of the most important tumor suppressors associated with the EMT and metastasis in various cancers including NPC. Bioinformatics Figure 2. EBV-miR-BART7-3p promotes the EMT of NPC cells. (a) The EMT proteins were detected by western blot in indicated cells passaged three times after the introduction of EBV-miR-BART7-3p or NC. Moreover, EMT proteins were examined in indicated cells transfected with anti-mir or anti-c. β-actin served as the internal control. Densitometry quantifications were performed with Gel-Pro Analyzer software. (b) Morphology of indicated cells was shown under phase contrast microscopy. Magnification, 100. (c) The expressions of EMT markers were evaluated by immunohistochemistry in tumor tissues derived from the mouse models of NPC metastasis and the control models. Magnification, 400. Oncogene (2015) 2156 2166 2015 Macmillan Publishers Limited

analysis showed that the 3 UTR of PTEN was well matched by the seed sequence of EBV-miR-BART7-3p (Figure 3a). To determine whether this viral mirna could directly target PTEN gene, we performed luciferase reporter assays. The target sequence of PTEN 3 UTR (wt 3 UTR) or the mutant sequence (mt 3 UTR) was cloned into a luciferase reporter vector. 293T cells were co-transfected with wt or mt 3 UTR vector and EBV-miR-BART7-3p mimics or anti-mir. Interestingly, we observed that EBV-miR- BART7-3p significantly attenuated the luciferase activity of reporter vector with the wt 3 UTR of PTEN, whereas this effect was abolished when the 3 UTR-binding site was mutated (Figure 3b). Moreover, anti-mir increased the luciferase activity of reporter vector with wt PTEN 3 UTR instead of mt 3 UTR (Figure 3b). To further confirm the relationship between EBV-miR-BART7-3p and PTEN, we detected PTEN mrna expression in 42 NPC and 16 NP tissue samples. As shown in Figure 3c, PTEN expression was significantly lower in NPC samples than in NP samples. It was negatively correlated with EBV-miR-BART7-3p expression as well 2159 (Figures 1b and 3c). Consistently, in vitro experiment showed that exogenic EBV-miR-BART7-3p expression obviously attenuated the endogenic protein expression level of PTEN in two NPC cell lines (Figure 3d). The immunohistochemistry assay displayed a markedly reduction of PTEN expression in tumor tissues derived from 5-8F-BART7-3p cell-injected mouse models (Figure 3e). Moreover, we observed that downregulated PTEN expression was correlated with the high degree of spread to regional lymph nodes and advanced clinical stage of NPC (Figure 3f). Collectively, these data suggest that EBV-miR-BART7-3p has its roles in NPC through directly suppressing PTEN expression. EBV-miR-BART7-3p regulates the PI3K/Akt/GSK-3β signaling inducing the expression and relocalization of Snail and β-catenin It is well known that PTEN can negatively regulate the PI3K/Akt/ GSK-3β signaling that has emerged as a central feature of EMT. As a result, we next elucidated whether EBV-miR-BART7-3p could regulate this signaling pathway by suppressing PTEN. As shown in Figure 3. EBV-miR-BART7-3p directly regulates the tumor suppressor PTEN. (a) EBV-miR-BART7-3p and its putative binding sequences in the 3 UTR of PTEN. Mutation was generated in the complementary site that binds to the seed region of EBV-miR-BART7-3p. (b) 293T cells were cotransfected with EBV-miR-BART7-3p mimic or NC and luciferase reporters carrying either the predicted mirna target site in PTEN 3 UTR (wt) or its corresponding mutant (mut). Inhibition of EBV-miR-BART7-3p by anti-mir or anti-c was also performed. The data are shown as the mean ± s.e.m. (***Po0.001). (c) The levels of EBV-miR-BART7-3p and PTEN were detected by qpcr in NPC and NP tissue specimens, normalized to the U6 snrna and GAPDH, respectively. The correlation between EBV-miR-BART7-3p and PTEN expression levels was calculated. The data were shown as the mean ± s.e.m. (***Po0.001). (d) Western blotting analysis of endogenous PTEN protein expression levels in 5-8F and CNE1 cells treated with EBV-miR-BART7-3p mimic or NC. β-actin served as the internal control. Densitometry quantifications were performed. (e) PTEN expression was evaluated using immunohistochemistry assay in tumor tissues derived from NPC metastasis mouse models and the control models. Magnification, 400. (f) PTEN mrna expression was detected by qpcr in NPC samples (with clinical stage and N stage) and NP samples. The data were shown as the mean ± s.e.m. (*Po0.05; ***Po0.001). 2015 Macmillan Publishers Limited Oncogene (2015) 2156 2166

2160 Figures 4a and b, the introduction of EBV-miR-BART7-3p could enhance the activated phosphorylation of Akt-ser473 and the inhibitory phosphorylation of GSK-3β-ser9, whereas the silencing of EBV-miR-BART7-3p attenuated p-akt-ser473 and p-gsk-3β-ser9 levels. Impressively, silencing of PTEN expression by sirna could phenocopy the effect of EBV-miR-BART7-3p on NPC cells; it also activated PI3K/Akt and GSK-3β-ser9 (Figure 4c), leading to the EMT and metastasis of NPC cells (Figures 4c and d, Supplementary Figure S5a). Snail and β-catenin are two important EMT-associated molecules lying downstream of PI3K/Akt/GSK-3β signaling. By increasing GSK-3β-ser9 phosphorylation, activated Akt can stabilize Snail and β-catenin which translocate to the nucleus to promote a gene expression programme that favors EMT. 3,38 Notably, western blots and confocal analysis showed that Snail and β-catenin were highly expressed and accumulated in the nuclei of NPC cells upon exogenic expression of EBV-miR-BART7-3p (Figures 5a and b), which were also confirmed by the immunohistochemistry analysis of tumor tissues derived from the mouse models of NPC metastasis (Figure 5c). To further validate PTEN functional effects on EBV-miR-BART7-3p-mediated EMT and metastatic phenotypes, experiments were performed wherein PTEN (lacking the 3 UTR) was re-expressed in 5-8F-BART7-3p and CNE1-BRT7-3p cells. We observed that PTEN re-expression could attenuate the ability of EBV-miR-BART7-3p to activate PI3K/Akt, p-gsk-3β-ser9, Snail and β-catenin (Figure 6a), thereby reverting EBV-miR-BART7-3p-induced cell migration/ invasion (Figure 6b, Supplementary Figure S5b) and EMT-like phenotype (Figures 6c and d). Therefore, these data suggest that EBV-miR-BART7-3p promotes the EMT and metastasis of NPC cells through regulating PTEN/ PI3k/Akt, GSK-3β, Snail and β-catenin. Silencing of endogenous EBV-miR-BART7-3p attenuates the EMT and cell migration in EBV-positive NPC cells EBV-miR-BARTs were preferentially expressed in EBV-infected epithelial cells, 25,39,40 so we replicated our experiments in the EBV-positive NPC cell line to validate our findings. NPC EBVpositive cell lines (C666-1-EBV, HONE1-EBV and HK1-EBV) were initially confirmed to express EBV-miR-BART7-3p at a similar level to CNE1-BART7-3P cells, 5-8F-BART7-3p cells and NPC tissues (Supplementary Figure S6a). HONE1-EBV cell line was selected as a representative cellular model because it grew better than other EBV-positive NPC cell lines and was suitable for in vitro experiments. Notably, we found that silencing of endogenous EBV-miR- BART7-3p (Figure 7a) could partially recover PTEN protein expression and suppress PI3k/Akt, Snail and β-catenin in HONE1- EBV cells, subsequently altering the expression of EMT markers and decreasing the migratory and invasive ability of HONE1-EBV cells (Figure 7b, Supplementary Figure S6b). Of the EBV latent genes, LMP1 and LMP2A have been reported to be able to influence PI3K/Akt pathway, 41 43 so we detected their expression levels in HONE1-EBV cells after transfected with anti-mir or anti-c. Notably, LMP2A protein was not expressed in Figure 4. EBV-miR-BART7-3p regulates PI3K/Akt/GSK-3β signaling by suppressing PTEN. (a) PTEN, AKT, p-akt and p-gsk-3β were detected by western blot in 5-8F and CNE1 cells transfected with EBV-miR-BART7-3p mimic or NC. β-actin served as the internal control. (b) 5-8 F-BART7-3p and CNE1-BART7-3p cells were transfected with anti-mir or anti-c, and then analyzed by western blot. β-actin served as the internal control. (c) PTEN, p-akt and GSK-3β were determined by western blot in 5-8F and CNE1 cells treated with sirna against PTEN or si-nc. β-actin was internal control. (d) The expression of EMT markers in indicated cells was detected by western blot. β-actin was internal control. Densitometry quantifications were performed for all above western blots. (e) si-pten or si-nc-treated NPC cells were applied to in vitro migration assays. The data were shown as the mean ± s.e.m. (*Po0.05; **Po0.01). Oncogene (2015) 2156 2166 2015 Macmillan Publishers Limited

2161 Figure 5. EBV-miR-BART7-3p enhances the expression and nuclear accumulation of Snail and β-catenin. (a) The expression of Snail and betacatenin was detected by western blot in indicated cells treated with lentiviral particles carrying EBV-miR-BART7-3p, anti-mir and si-pten as well as their corresponding NCs, respectively. β-actin served as the internal control. Densitometry quantification was performed for all western blots. (b) Confocal analysis was used to compare the expression and nuclear accumulation of snail and β-catenin between NC cells and BART7-3p NPC cells. Magnification, 1200. (c) The expression and nuclear accumulation of Snail and β-catenin were evaluated by immunohistochemistry in tumor tissues derived from NPC metastasis mouse models. Magnification, 400. (d) Graphical representation illustrating the role of EBV-miR-BART7-3p-mediated pathway in the EMT of NPC. HONE1-EBV cells (data not shown), which was possibly due to the degradation through an ubiquitin-dependent mechanism. 44,45 LMP1 expression remained unchanged after silencing of EBVmiR-BART7-3p (Figure 7c). These data imply that EBV-miR-BART7-3p may regulate PTEN/PI3K/Akt and its downstream signals independent of LMP1 and LMP2A in HONE1-EBV cells though these two latent genes are involved in the regulation of PI3K/Akt. Next, PTEN (lacking the 3 UTR) was also re-expressed in HONE1- EBV cells. Our results revealed that re-expressed PTEN could considerably reverse the relative expression of p-akt and E-cadherin, and reduce the cell migration and invasion in HONE1-EBV cells (Figure 7d, Supplementary Figure S6c). Thus, these results validate that EBV-miR-BART7-3p provides a portion of the regulatory influence of the EMT and metastasis in EBV-positive NPC cells by regulating PTEN/PI3K/Akt, GSK-3β, Snail and β-catenin. DISCUSSION Cancer metastasis occurs in multiple sequential steps underlain by genetic and epigenetic changes within probably a fraction of malignant cells. NPC is typically characterized by a higher propensity to invade adjacent regions and metastasize to regional lymph nodes or distant organs, leading to higher recurrence and worse prognosis. 37,46 Investigation of NPC metastasis may be helpful for fully understanding the molecular mechanism of cancer metastasis and developing new clinical management strategies. Here, we discovered an involvement of a viral mirna, EBV-miR-BART7-3p, in the EMT and metastasis of NPC cells and further demonstrated that this viral mirna played its roles by mainly regulating PTEN/PI3K/Akt, GSK-3β, Snail and β-catenin. This reveals a mechanistic connection between a viral mirna and cancer progression, and implicates EBV-miR-BART7-3p as a potential biomarker or therapeutic target. EBV is the first discovered human tumor virus 47 that has been linked to the development of cancers and serious conditions including Burkitt's lymphoma, Hodgkin lymphoma, gastric cancer and NPC. The role of EBV-encoded mirnas has recently emerged as an important aspect in EBV-associated cancers. 39 Two clusters of BART mirnas, encoded by the BamHI-A region of EBV genome, are the most abundant transcripts; approximately 15.57% of the total mirna content in NPC cells is attributed to BART mirnas, suggesting that they may have a considerable role in regulating both viral and cellular genes in NPC cells. 25,48 However, to date, only a few have been implicated in cancer progression. EBV-miR- BART5 targets PUMA gene to suppress p53-induced apoptosis. 31 EBV-MiR-BART3* overcomes the growth suppressive activity of DICE1 and stimulates cell proliferation. 26 EBV-miR-BART1 is probably involved in the regulation of metabolism genes. 35 EBVmiR-BART7 has been recently reported to be detectable in NPC patient plasma samples and enhance the in vitro proliferation and migration of NPC cells. 33 In the present study, we not only observed that EBV-miR-BART7-3p was highly expressed in NPC relative to NP, clinically related to lymph node metastasis and clinical stage of NPC, and significantly promoted the metastatic potential of NPC cells in vitro and in vivo, but also discovered that this viral mirna led to decreased expression of epithelial markers and increased the expression of mesenchymal markers in NPC 2015 Macmillan Publishers Limited Oncogene (2015) 2156 2166

2162 Figure 6. Functional validation of PTEN effects on EBV-miR-BART7-3p-mediated EMT and metastatic phenotypes. (a) The expression of PTEN, p-akt, p-gsk-3β, Snail and β-catenin in 5-8F-BART7-3p and CNE1-BART7-3p cells was measured by western blot in the presence and absence of PTEN (lacking the 3 UTR) construct. β-actin served as the internal control. Densitometry quantification was performed. (b) 5-8F-BART7-3p and CNE1- BART7-3p cells were applied to in vitro migration assays in the presence and absence of PTEN (lacking the 3 UTR) construct. The data were shown as the mean ± s.e.m. (**Po0.01). (c) Phase-contrast microscopy images of 5-8F-BART7-3p and CNE1-BART7-3p cells in the presence and absence of PTEN (lacking the 3 UTR) reconstitution. Magnification, 100. (d) EMT-associated proteins were measured by western blot. β-actin served as the internal control. Densitometry quantification was performed. cells along with EMT-like morphological changes. Silencing of endogenous EBV-miR-BART7-3p could partially attenuate the EMT and cell migration of EBV-positive NPC cells. This is the first study to indicate that EBV-miR-BART7-3p promotes a transition from epithelial to mesenchymal phenotype that contributes to cancer metastasis. Interestingly, a recent study has reported the effect of EBV-miR-BART9 on EMT by suppressing E-cadherin expression. 36 Combining the research data with ours, it can be supposed that EBV BART clusters may contain several mirnas targeting host genes to induce EMT; as a result, they deserve to be fully explored in the future. PTEN is an important tumor suppressor with diphosphohydrolase activity, affecting multiple cellular processes including cell growth, proliferation and cell migration. Up to now, several cellular mirnas, such as mir-205, 49 mir-141 50 and mir-144, 51 have been identified to regulate PTEN in various cancers 52,53 including NPC, but no viral mirnas are reported to directly repress it. The present study first reported that a viral mirna, EBV-miR- BART7-3p, could directly target this major human tumor suppressor, based on several lines of evidence. First, bioinformatics prediction and luciferase reporter assay showed that EBVmiR-BART7-3p was able to directly suppress PTEN by binding to its 3 UTR. Second, there was an inverse relationship between EBVmiR-BART7-3p and PTEN expression in clinical samples. Third, exogenic EBV-miR-BART7-3p expression could attenuate endogenic PTEN expression in vitro and in vivo. Fourth, the reexpression of PTEN could reverse EBV-miR-BART7-3p-induced phenotypes. Fifth, a previous study also theoretically predicted that PTEN is a potential target of EBV-miR-BART7. 54 It is believed that PTEN is a key negative regulator of the PI3K/ Akt signaling that activates the EMT programme. 55 GSK-3β can be inactivated by PI3K/AKT pathway via increasing inhibitory GSK-3β phosphorylation, 56,57 which stabilizes Snail and β-catenin or facilitates the nuclear location of them, and leads to the enhanced invasiveness and EMT of cancer cells. 8,38,58,59 In the present study, we observed that EBV-miR-BART7-3p downregulated PTEN expression, which could also increase the activated phosphorylation of Akt-ser473 and inhibitory phosphorylation of GSK-3β-ser9, consequently enhancing the expression and nuclear location of Snail and β-catenin in NPC cells. The silencing of EBV-miR-BART7-3p suppressed PI3k/Akt, Snail and β-catenin accordingly. These observations support that EBV-miR-BART7-3p induces EMT and cancer metastasis through regulating PI3K/Akt, GSK-3β, Snail and β-catenin. Interestingly, several previous reports showed that the inactive form of GSK-3β could activate both Snail and β-catenin during gastrin-induced migration, 58 Fas-induced EMT, 38 and Endothelin-1 mediated-emt and cell invasion 60 in cancers. Therefore, our results also suggest that exogenic viral mirna can, like host factors, induce GSK-3β-mediated activation of Snail and β- catenin in EMT and cancer metastasis through regulating its associated signaling pathways. It is known that Snail and β-catenin are two of the most important EMT-associated proteins. In the initiation of EMT, with the destabilization of adherens junctions and the degradation of E-cadherin, β-catenin can no longer interact with E-cadherin; so it is either degraded or re-located into the nucleus to engage some transcription factors and promote a gene expression programme that favors EMT. 3 Snail is a key transcriptional factor that represses Oncogene (2015) 2156 2166 2015 Macmillan Publishers Limited

2163 Figure 7. Silencing of endogenous EBV-miR-BART7-3p attenuates the EMT and cell migration in EBV-positive NPC cells. (a) qpcr analysis showing EBV-miR-BART7-3p expression in HONE1-EBV cells transfected with anti-mir or anti-c. The data were shown as the mean ± s.e.m. (***Po0.001). (b, c) After transfected with anti-mir or anti-c, HONE1-EBV cells were detected by western blot for the expression levels of PTEN, p-akt, β-catenin, Snail, EMT-associated proteins and LMP1, and then applied to in vitro migration assay (*Po0.05, **Po0.01). (d) After transfection with PTEN construct or NC, HONE1-EBV cells were detected by western blot for PTEN, p-akt and E-cadherin expression, and then applied to in vitro migration assay (**Po0.05, ***Po0.001). For all above western blots, β-actin served as the internal control and densitometry quantification was performed. the epithelial marker genes and activates genes associated with the mesenchymal phenotype. It represses E-cadherin expression by recruiting the Polycomb repressive complex 2 (PRC2) to the E-box sequences of E-cadherin proximal promoter region. 46 It activates MMP expression by cooperating with other transcriptional regulators such as ETS1. 61 High ZEB expression often follows the activation of Snail expression, consistent with Snail directly targeting the ZEB1 gene. Twist often cooperates with Snail in the induction of ZEB1 expression. 62 Of note, our qpcr data also showed a general increase in the expression of some EMTassociated proteins including ZEB1, ZEB2, Twist, MMP2 and MMP9 in NPC cells overexpressing EBV-miR-BART7-3p, supporting that this viral mirna does enhance a gene expression programme that facilitates EMT and cancer metastasis through regulating an important tumor suppressor and its key downstream signals. EBV-miR-BART7-3p may be not the only factor regulating PTEN/ PI3K/AKT pathway and some other regulators should be involved. Our results have reflected this point; although we confirmed that EBV-miR-BART7-3p could regulate PI3K/AKT pathway by inhibiting PTEN, it was in EBV-positive NPC cells that silencing of endogenous EBV-miR-BART7-3p could not fully recover PTEN expression and its downstream signals as well as reduce cell migration and invasion, suggesting the existence of other factors, which need to be further investigated. LMP1 and LMP2A, as two major EBV latent genes in NPC, have been reported to influence PI3K/AKT pathway 41 43 as well, but we did not observe the obvious expression alteration of LMP1 and LMP2A upon the silencing of endogenous EBV-miR-BART7-3p in HONE1-EBV cells, indicating that EBV-miR-BART7-3p modulates PTEN/PI3K/Akt and its downstream signals independent of LMP1 and LMP2A. In summary, our study shows, for what we believe is the first time, that EBV-miR-BART7-3p has important roles in the EMT and cancer metastasis by directly suppressing the tumor suppressor PTEN, which provides a molecular basis for the regulation of PI3K/ Akt, GSK-3β, Snail and β-catenin though there may be more EBVmiR-BART7-3p targets that likely contribute to the EMT and metastasis of NPC cells. Uncovering this viral mirna-mediated mechanism underlying the EMT and cancer metastasis provides a novel insight into understanding virus-mediated tumor progression and metastasis. Moreover, our study implicates the significance of our findings in clinical application; EBV-miR-BART7-3p may be a potential biomarker for the clinical prognosis or diagnosis of NPC. Silencing of oncogenic virus-encoded mirnas may be a promising therapeutic approach for targeting therapy of virus-related cancers including NPC in the future. MATERIALS AND METHODS Ethics statement All subjects involved in this study signed informed consent. The research was approved by the Ethics Committee of Southern Medical University, Guangzhou, and Zhongshan People s Hospital, Guangdong, China. Animal experimentation was approved by the Ethical Committee of Animal Research at Southern Medical University. The experimental protocol was established according to the associated national guidelines from Ministry of Science and Technology of China. Tissue specimens All NPC (not pretreated with radiotherapy or chemotherapy) and NP specimens were collected and confirmed pathologically in Zhongshan People s Hospital, Guangdong, China. Sixteen NPC and 20 NP specimens were used for microarray analysis (Supplementary Table S1) and additional 42 NPC specimens with TNM staging and 16 NP specimens were collected for qpcr and clinical data analysis (Supplementary Table S3). Staging was 2015 Macmillan Publishers Limited Oncogene (2015) 2156 2166

2164 performed according to the 1992 Fuzhou NPC staging system of China. 63 N staging indicates the degree of spread to regional lymph nodes, ranging from N0 to N3. Clinical staging is determined by combining the T, N and M classifications. The hematoxylin-and-eosin-stained frozen sections matched for each NPC tissue were examined to ensure that each tissue should contain more than 80% of homogeneous cancer cells in its cross-sectional area. mirna microarray analysis Total RNA was extracted from each NPC and NP specimen using Trizol Reagent (Invitrogen, Carlsbad, CA, USA). mirna microarrays (CCDTMmiRNA850-V4p1.4) were printed at Infectious Disease and Immunogenetics section, Department of Transfusion Medicine, Clinical Center, National Institutes of Health, United States. Briefly, 6 μg of total RNA was labeled by mircury LNA mirna Power Labeling Kit (Exiqon Life Sciences, Vedbaek, Denmark) according to the manufacturer s instruction. The test sample and the reference were labeled with Hy5 and Hy3, respectively. After labeling, the sample and the reference were co-hybridized to the mirna array. The washed slides were scanned by LuxScanTM 10K microarray Scanner (Capital Bio. Corporation, Beijing, China). The resulting gene expression data were uploaded to madb database and analyzed using BRBarrayTools (NCI, Bethesda, MD, USA) further. Clustering and visualization of expression profiles were preformed with Cluster (University of California, Berkeley, CA, USA) and Treeview software (Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA, USA). Cell line and cell culture Two EBV-negative NPC cell lines (CNE1 and 5-8F) and 293T cells were obtained from Cancer Research Institute of Southern Medical University, Guangzhou, China. Three EBV-positive NPC cell lines (C666-1-EBV, HONE1- EBV, and HK1-EBV) were kindly offered by Prof. George S. W. Tsao from the University of Hong Kong. All cell lines were cultured in RPMI-1640 (HyClone, Logan, UT, USA) with 10% calf serum (Gibco, Grand Island, NY, USA) at 37 C and 5% CO 2. Lentivirus production and infection Lentiviral (GV209, H1-MCS-CMV-EGFP) particles carrying EBV-miR-BART7-3p precursor (BART7-3p for short) and its franking control sequence (NC for short) were constructed by GeneChem, Shanghai, China. CNE1 and 5-8F cells were infected with recombinant lentiviral transducing units plus 8 mg/ml Polybrene (Sigma-Aldrich, St Louis, MO, USA) for 2 days according to the protocol from GeneChem. The EBV-miR- BART7-stably-expressed NPC cells and their corresponding NC cells (GFP + ) were sorted with BD FACS Aria cell sorter (Becton and Dickinson Company, Franklin Lakes, NJ, USA) for the following experiments. EBV-miR-BART7-3p expression was confirmed by qpcr. Two stable cell lines were indicated as 5-8F-BART7-3p and CNE1-BART7-3p, respectively. RNA oligoribonucleotides, PTEN overexpression vector and cell transfection EBV-miR-BART7-3p mimic, inhibitor (anti-mir, 2 -O-methyl modification) and their corresponding NCs were synthesized by Genepharma (Shanghai, China) (Supplementary Table S5). PTEN sirna (h2) (Santa Cruz, sc-44272, Santa Cruz, CA, USA) and its control sirna-a (Santa Cruz, sc-37007) were indicated as si-pten and si-nc, respectively. h-pten lentiviral vector (ABM Inc. Cat. LV276502, San Jose, CA, USA; lacking the 3 UTR) and blank control Lentivector (ABM Inc., Cat. LVP591) were kindly offered by Professor Xiao Dong from cancer research institute of Southern Medical University, Guangzhou, China. Before transfection, the medium was changed to RPMI-1640 (HyClone) with 10% fetal bovine serum (Gibco). All cells were maintained in a humidified atmosphere of 95% air and 5% CO 2 at 37 C, and seeded 24 h prior to transfection. EBV-miR-BART7-3p mimic, anti-mir, si-pten, PTEN lentiviral vector and their NCs were transfected into cells at a final concentration of 50 nmol/l using Lipofectamine 2000 (Invitrogen, 11668-019) in serum-free conditions. Six hours later, the medium was changed to fresh RPMI-1640 (HyClone) with 10% fetal bovine serum (Gibco). RNA extraction and qpcr Total RNA was extracted with TRIzol reagent (Invitrogen). cdna was synthesized with the PrimeScript RT reagent Kit (TaKaRa, Dalian, China). PCR analyses were performed with SYBR Premix Ex Taq (TaKaRa). The primers used are shown in Supplementary Table S6. Data were normalized to GAPDH expression and further normalized to the negative control unless otherwise indicated. Quantification of EBV-miR-BART7-3p was performed by TaqMan mirna assays (Applied Biosystems, Foster City, CA, USA) per the manufacturer s instructions with 70 ng of total RNA. Data were normalized to small nuclear RNA RNU6B (U6 snrna) expression. All reactions were run in triplicate and repeated in three independent experiments. The fold changes were calculated through relative quantification (2 ΔΔCt ). In vitro migration and invasion assays Transwell migration was performed to assess cell migration. Transfected or infected NPC cells were cultured for 48 h. Subsequently, dissociated 1 10 5 cells were resuspended in serum-free media and placed in inserts containing 8 um pores (Corning Co, Corning, NY, USA). These inserts were put in wells with 10% FBS serum-containing media. After 24 h of incubation, the cells on the upper surface of the membrane were removed. The cells that had migrated to the lower surface of the membrane were fixed with 100% methanol and stained with 1% toluidine, followed by counting in six random optical fields under a microscope. Cell invasion assay (Boyden assay) was similar to cell migration assay except transwell membrane was pre-coated with 24 mg/ml matrigel. In vivo metastasis assay All mice that were 4 5 weeks old with 18 20 g in weight were provided by the Central Animal Facility of Southern Medical University. To establish tumor metastasis models, we firstly made a small cut on the abdominal region of each mouse, softly pushed its liver out of abdominal cavity, injected 50 ul of 5-8F cells (5 10 6 ) stably expressing GFP/EBV-miR- BART7-3p or an equal number of control cells under the liver capsule of each mouse (six mice for each group), and then carefully pushed its liver back into the abdominal cavity after cleaning and lightly pressing the pinhole with alcohol cotton balls for 2 min. All mice were killed in 3 weeks. Their whole bodies and resected internal organs including lymph nodes (aligned on culture plates) were subjected to fluorescent image detection using LT-9MACIMSYSPLUS whole-body imaging system (Lighttools Research, Encintas, CA, USA). Western blot analysis Western blotting analyses were performed with standard methods. Briefly, cell pallets were lysed in the radio-immunoprecipitation assay buffer containing protease inhibitors (Sigma-Aldrich) and phosphatase inhibitors (Keygen, Nanjing, China). Proteins were separated by 10% SDS PAGE gels, and blotted onto polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA). The membrane was probed with the specific antibodies (Supplementary Table S7), and then with peroxidase-conjugated secondary antibodies. Beta-actin was used as a protein loading control. The bands were visualized by eecl Western Blot Kit (CWBIO Technology, Beijing, China). The images were captured with ChemiDocTM CRS+ Molecular Imager (Bio-Rad, Hercules, CA, USA). LMP2A and LMP1 antibodies were kindly offered by Prof. Zeng Mu-sheng and Zhu Xiao-feng from Sun Yatsen university cancer center, Guangzhou, China (Supplementary Table S7). Immunohistochemical staining The paraffin sections prepared from in vivo experiments were applied to immunohistochemistry assays for detecting protein expression levels of PTEN, Snail and EMT proteins. The indirect streptavidin-peroxidase method was used as the manufacturer s introduction. The immunohistochemically stained tissue sections were reviewed separately by two pathologists. The information of antibodies was shown in Supplementary Table S7. Dual luciferase assay 293T cells (1 10 4 ) were cultured in 24-well plates and co-transfected with 20 nm EBV-miR-BART7-3p mimic or NC, 5 ng of prl-cmv Renilla luciferase reporter and 30 ng of luciferase reporter that contained the wild-type or mutant 3 UTR of PTEN. For antagonism experiments, cells were also cotransfected with 20 nm anti-mir or anti-c. Transfections were performed in duplicate and repeated in three independent experiments. Forty-eight hours after transfection, the luciferase activities were analyzed with a Dual- Luciferase Reporter Assay System (Promega, Madison, WI, USA). Oncogene (2015) 2156 2166 2015 Macmillan Publishers Limited

Immunofluorescent staining Cell lines were cultured on coverslips, rinsed with phosphate-buffered saline after 24 h and fixed with 4% paraformaldehyde for 5 min at 20 C. The cells were next blocked for 30 min in 0.3% Txiton X-100, and then incubated with primary monoclonal antibodies (Snail and β-catenin) in phosphate-buffered saline for 2 h at room temperature. After three washes in phosphate-buffered saline, the coverslips were incubated for 1 h in the dark with secondary antibodies (Bioworld Technology, Inc, Nanjing, China). After further washing for three times, the slides were stained with 4-6- diamidino-2-phenylindole for 5 min to visualize the nuclei, and viewed by confocal microscope (Olympus FV1000, Tokyo, Japan). Statistical analysis The data are expressed as the mean ± s.e.m. from at least three independent experiments. Comparisons between two groups were performed using Student s t-test, unless otherwise indicated. The association between EBV-miR-BART7-3p and PTEN gene was analyzed using Spearman s correlation coefficient. Statistical analyses were performed with the SPSS 13.0 statistical software package (SPSS Inc. Chicago, IL, USA). All statistical tests were two-sided, and Po0.05 was considered to be statistically significant. CONFLICT OF INTEREST The authors declare no conflict of interest. ACKNOWLEDGEMENTS This study was financially supported by grants from National Natural Science Foundation of China (No. 81372895, 81172586), Research Fund for the Doctoral Program of Higher Education of China (No. 20134433110013), Natural Science Foundation of Guangdong Province (No. S2011010004157) and Guangzhou Science and Technology research project (No. 2014J4100149). REFERENCES 1 Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymal transitions in development and disease. Cell 2009; 139: 871 890. 2 Brabletz T. EMT and MET in metastasis: where are the cancer stem cells? Cancer Cell 2012; 22: 699 701. 3 Lamouille S, Xu J, Derynck R. Molecular mechanisms of epithelial-mesenchymal transition. Nat Rev Mol Cell Biol 2014; 15: 178 196. 4 Talbot LJ, Bhattacharya SD, Kuo PC. Epithelial-mesenchymal transition, the tumor microenvironment, and metastatic behavior of epithelial malignancies. Int J Biochem Mol Biol 2012; 3: 117 136. 5 Villarejo A, Cortes-Cabrera A, Molina-Ortiz P, Portillo F, Cano A. Differential role of Snail1 and Snail2 zinc fingers in E-cadherin repression and epithelial to mesenchymal transition. J Biol Chem 2013; 289: 930 941. 6 Lim SO, Gu JM, Kim MS, Kim HS, Park YN, Park CK et al. Epigenetic changes induced by reactive oxygen species in hepatocellular carcinoma: methylation of the E-cadherin promoter. Gastroenterology 2008; 135: 2128 2140, 2140 2141. 7 Zhang J, Ma L. MicroRNA control of epithelial-mesenchymal transition and metastasis. Cancer Metastasis Rev 2012; 31: 653 662. 8 Zhang JP, Zeng C, Xu L, Gong J, Fang JH, Zhuang SM. MicroRNA-148a suppresses the epithelial-mesenchymal transition and metastasis of hepatoma cells by targeting Met/Snail signaling. Oncogene 2013; 33: 4069 4076. 9 Johansson J, Berg T, Kurzejamska E, Pang MF, Tabor V, Jansson M et al. MiR-155- mediated loss of C/EBPbeta shifts the TGF-beta response from growth inhibition to epithelial-mesenchymal transition, invasion and metastasis in breast cancer. Oncogene 2013; 32: 5614 5624. 10 Formosa A, Markert EK, Lena AM, Italiano D, Finazzi-Agro' E, Levine AJ et al. MicroRNAs, mir-154, mir-299-5p, mir-376a, mir-376c, mir-377, mir-381, mir- -487b, mir-485-3p, mir-495 and mir-654-3p, mapped to the 14q32.31 locus, regulate proliferation, apoptosis, migration and invasion in metastatic prostate cancer cells. Oncogene 2014; 33: 5173 5182. 11 Zhang G, Zhou H, Xiao H, Liu Z, Tian H, Zhou T. MicroRNA-92a functions as an oncogene in colorectal cancer by targeting PTEN. Dig Dis Sci 2014; 59: 98 107. 12 Chen D, Zhang Y, Wang J, Chen J, Yang C, Cai K et al. MicroRNA-200c overexpression inhibits tumorigenicity and metastasis of CD117+CD44+ ovarian cancer stem cells by regulating epithelial-mesenchymal transition. J Ovarian Res 2013; 6: 50. 13 Cao M, Seike M, Soeno C, Mizutani H, Kitamura K, Minegishi Y et al. MiR-23a regulates TGF-beta-induced epithelial-mesenchymal transition by targeting E-cadherin in lung cancer cells. Int J Oncol 2012; 41: 869 875. 14 Colangelo T, Fucci A, Votino C, Sabatino L, Pancione M, Laudanna C et al. MicroRNA-130b promotes tumor development and is associated with poor prognosis in colorectal cancer. Neoplasia 2013; 15: 1218 1231. 15 Kumarswamy R, Mudduluru G, Ceppi P, Muppala S, Kozlowski M, Niklinski J et al. MicroRNA-30a inhibits epithelial-to-mesenchymal transition by targeting Snai1 and is downregulated in non-small cell lung cancer. Int J Cancer 2012; 130: 2044 2053. 16 Ke Y, Zhao W, Xiong J, Cao R. mir-149 inhibits non-small-cell lung cancer cells EMT by targeting FOXM1. Biochem Res Int 2013; 2013: 506731. 17 Lyu X, Fang W, Cai L, Zheng H, Ye Y, Zhang L et al. TGFbetaR2 is a major target of mir-93 in nasopharyngeal carcinoma aggressiveness. Mol Cancer 2014; 13: 51. 18 Luo Z, Zhang L, Li Z, Jiang C, Dai Y, Liu X et al. mir-149 promotes epithelialmesenchymal transition and invasion in nasopharyngeal carcinoma cells. Zhong Nan Da Xue Xue Bao Yi Xue Ban 2011; 36: 604 609. 19 Luo Z, Dai Y, Zhang L, Jiang C, Li Z, Yang J et al. mir-18a promotes malignant progression by impairing microrna biogenesis in nasopharyngeal carcinoma. Carcinogenesis 2013; 34: 415 425. 20 Zur HH. Viruses in human cancers. Science 1991; 254: 1167 1173. 21 McLaughlin-Drubin ME, Munger K. Viruses associated with human cancer. Biochim Biophys Acta 2008; 1782: 127 150. 22 Parkin DM. The global health burden of infection-associated cancers in the year 2002. Int J Cancer 2006; 118: 3030 3044. 23 Kim DN, Chae HS, Oh ST, Kang JH, Park CH, Park WS et al. Expression of viral micrornas in Epstein-Barr virus-associated gastric carcinoma. J Virol 2007; 81: 1033 1036. 24 Ramakrishnan R, Donahue H, Garcia D, Tan J, Shimizu N, Rice AP et al. Epstein-Barr virus BART9 mirna modulates LMP1 levels and affects growth rate of nasal NK T cell lymphomas. PLoS One 2011; 6: e27271. 25 Yang HJ, Huang TJ, Yang CF, Peng LX, Liu RY, Yang GD et al. Comprehensive profiling of Epstein-Barr virus-encoded mirna species associated with specific latency types in tumor cells. Virol J 2013; 10: 314. 26 Lei T, Yuen KS, Xu R, Tsao SW, Chen H, Li M et al. Targeting of DICE1 tumor suppressor by Epstein-Barr virus-encoded mir-bart3* microrna in nasopharyngeal carcinoma. Int J Cancer 2013; 133: 79 87. 27 Lo AK, Dawson CW, Jin DY, Lo KW. The pathological roles of BART mirnas in nasopharyngeal carcinoma. J Pathol 2012; 227: 392 403. 28 Barth S, Pfuhl T, Mamiani A, Ehses C, Roemer K, Kremmer E et al. Epstein-Barr virus-encoded microrna mir-bart2 down-regulates the viral DNA polymerase BALF5. Nucleic Acids Res 2008; 36: 666 675. 29 Lo AK, To KF, Lo KW, Lung RW, Hui JW, Liao G et al. Modulation of LMP1 protein expression by EBV-encoded micrornas. Proc Natl Acad Sci USA 2007; 104: 16164 16169. 30 Lung RW, Tong JH, Sung YM, Leung PS, Ng DC, Chau SL et al. Modulation of LMP2A expression by a newly identified Epstein-Barr virus-encoded microrna mir-bart22. Neoplasia 2009; 11: 1174 1184. 31 Choy EY, Siu KL, Kok KH, Lung RW, Tsang CM, To KF et al. An Epstein-Barr virusencoded microrna targets PUMA to promote host cell survival. J Exp Med 2008; 205: 2551 2560. 32 Nachmani D, Stern-Ginossar N, Sarid R, Mandelboim O. Diverse herpesvirus micrornas target the stress-induced immune ligand MICB to escape recognition by natural killer cells. Cell Host Microbe 2009; 5: 376 385. 33 Chan JY, Gao W, Ho WK, Wei WI, Wong TS. Overexpression of Epstein-Barr virusencoded microrna-bart7 in undifferentiated nasopharyngeal carcinoma. Anticancer Res 2012; 32: 3201 3210. 34 ChoiH,LeeH,KimSR,GhoYS,LeeSK.Epstein-Barrvirus-encodedmicroRNABART15-3p promotes cell apoptosis partially by targeting BRUCE. JVirol2013; 87: 8135 8144. 35 Ye Y, Zhou Y, Zhang L, Chen Y, Lyu X, Cai L et al. EBV-miR-BART1 is involved in regulating metabolism-associated genes in nasopharyngeal carcinoma. Biochem Biophys Res Commun 2013; 436: 19 24. 36 Wong AM, Kong KL, Tsang JW, Kwong DL, Guan XY. Profiling of Epstein-Barr virusencoded micrornas in nasopharyngeal carcinoma reveals potential biomarkers and oncomirs. Cancer 2012; 118: 698 710. 37 Zheng H, Li W, Wang Y, Liu Z, Cai Y, Xie T et al. Glycogen synthase kinase-3 beta regulates Snail and beta-catenin expression during Fas-induced epithelialmesenchymal transition in gastrointestinal cancer. Eur J Cancer 2013; 49:2734 2746. 38 Qiu J, Cosmopoulos K, Pegtel M, Hopmans E, Murray P, Middeldorp J et al. A novel persistence associated EBV mirna expression profile is disrupted in neoplasia. PLoS Pathog 2011; 7: e1002193. 39 Cai X, Schafer A, Lu S, Bilello JP, Desrosiers RC, Edwards R et al. Epstein-Barr virus micrornas are evolutionarily conserved and differentially expressed. PLoS Pathog 2006; 2: e23. 40 Dawson CW, Tramountanis G, Eliopoulos AG, Young LS. Epstein-Barr virus latent membrane protein 1 (LMP1) activates the phosphatidylinositol 3-kinase/Akt 2165 2015 Macmillan Publishers Limited Oncogene (2015) 2156 2166