Epstein-Barr Virus LMP1-Activated mTORC1 and mTORC2 Coordinately Promote Nasopharyngeal Cancer Stem Cell Properties

Abstract

Epstein-Barr virus (EBV) is associated with several malignant diseases, including Burkitt's lymphoma, nasopharyngeal carcinoma (NPC), certain types of lymphomas, and a portion of gastric cancers. The virus-encoded oncoprotein, LMP1, induces the epithelial-to-mesenchymal transition (EMT), leading to cancer stem cell formation. In the current study, we investigated how LMP1 contributes to cancer stem cell development in NPC. We found that LMP1 plays an essential role in acquiring cancer stem cell (CSC) characteristics, including tumor initiation, metastasis, and therapeutic resistance by activating the PI3K/mTOR/Akt signaling pathway. We dissected the functions of distinct signaling (mTORC1 and mTORC2) in the acquisition of different CSC characteristics. Side population (SP) formation, which represents the chemotherapy resistance feature of CSC, requires mTORC1 signaling. Tumor initiation capability is mainly attributed to mTORC2, which confers on NPC the capabilities of proliferation and survival by activating mTORC2 downstream genes c-Myc. Both mTORC1 and mTORC2 enhance cell migration and invasion of NPC cells, suggesting that mTORC1/2 coregulate metastasis of NPC. The revelation of the roles of the mTOR signaling pathways in distinct tumorigenic features provides a guideline for designing efficient therapies by choosing specific mTOR inhibitors targeting mTORC1, mTORC2, or both to achieve durable remission of NPC in patients. IMPORTANCE LMP1 endows NPC to gain cancer stem cell characteristics through activating mTORC1 and mTORC2 pathways. The different mTOR pathways are responsible for distinct tumorigenic features. Rapamycin-insensitive mTORC1 is essential for CSC drug resistance. NPC tumor initiation capacity is mainly attributed to mTORC2 signaling. mTORC1 and mTORC2 coregulate NPC cell migration and invasion. The revelation of the roles of mTOR signaling in NPC CSC establishment has implications for novel therapeutic strategies to treat relapsed and metastatic NPC and achieve durable remission.

Keywords: Epstein-Barr virus (EBV); LMP1; cancer stem cell; mTORC1; mTORC2; nasopharyngeal carcinoma (NPC).

FIG 1

Effect of LMP1 expression on mTOR signaling. (A) E/M and xM subpopulations sorted from S26-LMP1/2A cell line, and the E subpopulation sorted from S26 cells were analyzed by Western blot for the expression and phosphorylation of mTOR components and substrates. (B) CNE-2 S26 and S18 cells were transfected with various doses of pcDNA3.1-LMP1. Forty-eight hours posttransfection, cell lysates were analyzed by Western blot for the expression and phosphorylation of mTOR components and substrates. The expression or phosphorylation of proteins was quantitated by densitometry and plotted. (C) S26 and S18 cells that stably express LMP1 were examined by Western blot for activation of mTOR signaling. β-actin served as the loading control.

FIG 2

LMP1-mediated activation of mTORC1 confers on NPC cells cancer stemness properties. (A) CNE-2 S18 cells were transduced with shRNA lentiviruses against mTOR, Raptor, and Rictor genes. The knockdown efficiencies and the functional consequences (phosphorylation of AKT at S473 and p70S6K at T389) were analyzed by Western blots. (B and C) The effect of silencing mTOR signaling in S18 cells on Vimentin expression was analyzed by Western (B) and immunofluorescent assays (630×) (C). (D) CNE-2 S26 and S18 cells were transfected with pcDNA3.1-LMP1 or a combination of 0.3 μg/mL pcDNA3.1-LMP1 and 1 μM BEZ235. Forty-eight hours posttransfection, cells were subjected to flow cytometry (Aldefluor) for ALDH^br cells. ALDH^br gating is established using the cells treated with ALDH inhibitor DEAB. (E) S18 and S26 cells were transduced with shRaptor, shRictor, and shmTOR lentiviruses, selected with puromycin for 7 days and then analyzed by Aldefluor assay. (F) S18 cells were treated with mTORC1/2 dual inhibitor BEZ235 and mTORC1 inhibitor rapamycin for 24 h. The effects of these treatments were determined by monitoring ALDH^br populations using Aldefluor assay.

FIG 3

Rapamycin-insensitive mTORC1 is essential for NPC drug resistance. (A) Flow cytometry analysis of side population (SP) in CNE-2 S26 cells and S18 cells stained with Hoechst 33342. (B) S26 and S18 cells were transfected with pcDNA3.1-LMP1 or treated with combination 1 μM BEZ235 and 0.2 μg/mL pcDNA3.1-LMP1. Forty-eight hours posttransfection, cells were subjected to flow cytometry analysis for SP cells. (C) S18 cells were transduced by specific shRNA lentiviruses to silence the expression of Raptor (shRaptor), Rictor (shRictor), and mTOR (shmTOR), respectively. The knockdown efficiency of each shRNA was verified by Western blot as well as the functional consequences (phosphorylation of 4E-BP1 [T37/46] and 4E-BP1). (D) Effects of shRNA-mediated knockdown of mTOR components were analyzed for SP by flow cytometry. (E) S18 cells and EBV-positive CNE2 cells (CNE2+) were treated mTORC1/2 dual inhibitor BEZ235 and mTORC1 inhibitor rapamycin in various doses for 24 h. The effects of these treatments on SP were analyzed by flow cytometry after staining with Hoechst 33342. (F) The effects of the treatment on phosphorylation of 4E-BP1 at T37/46, p70S6K at T389, and AKT at S473 in S18 cells were examined. (G) Side population cells (SP) and non-side population cells (non-SP) were sorted from CNE2-S18 cells by FCS assay. The expression and phosphorylation of mTOR components and substrates in SP and non-SP cells were analyzed by Western blot. (H) The expression of ABCG2 mRNA relative to GAPDH in SP, non-SP, S18, and S26 cells was determined by RT-qPCR (mean ± SD of three biological replicates).

FIG 4

Tumor initiation ability of nasopharyngeal cancer stem cells is mainly attributed to mTORC2. (A) CNE2-S26 cells expressing LMP1 were treated with 1 μM BEZ235 or 100 nM rapamycin and the effects of mTOR inhibitors on tumor initiation ability were analyzed by the tumorsphere-forming assay. Representative images are shown (50×) (mean ± SD, n = 3). (B) CNE2-S18 cells and EBV-positive TW03 cells (TW03+) were treated with 1 μM BEZ235 or 100 nM rapamycin and subjected to a tumorsphere-forming assay. Representative images are shown (50×) (mean ± SD, n = 3 per group). (C and D) S18 cells and S26 cells were transduced with shRNA lentiviruses targeting Raptor (shRaptor), Rictor (shRictor), or mTOR (shmTOR), respectively. The effects of silencing mTOR components on tumor initiation ability were analyzed by tumorsphere-forming assay and representative images are shown (50×). The tumorsphere number of each sample was quantitated (mean ± SD, n = 3). (E) The knockdown efficiency of each shRNA and functional consequences (the phosphorylation of AKT and FoxO and the expression of pluripotent transcription factors Bmi-1, c-Myc, Sox2, KLF4, and Nanog) in CNE2-S18 cells were examined by Western blot. (F) S18 cells and mTOR component knockdown cells were analyzed for apoptosis by annexin V flow cytometry assay. (G) S18 and these knockdown cells were subjected to CFSE dye dilution assays for cell proliferation ability. (H) Cell proliferation upon the shRNA knockdown of mTOR components was compared between S18 and S26 cells by CFSE dye dilution assays. (I) Does-response curves of S18 and S26 cells to the treatment with cisplatin, rapamycin, and BEZ235 were compared (mean ± SD, n = 3).

FIG 5

Effects of silencing mTORC1 and mTORC2 on NPC cell migration and invasion capabilities. (A) CNE2-S26 LMP1-expressing cells were treated with 1 μM BEZ235 or 100 nM rapamycin and effects of the treatments on cell migration and invasion abilities were assayed using Transwell migration and invasion assays. Cells migrated to the lower chamber were fixed, stained with crystal violet, and counted (200×, mean ± SD, n = 3). (B) CNE2-S18 cells were treated with 1 μM BEZ235 or 100 nM rapamycin and analyzed for invasion and migration. The S18 cells that invaded into the lower chamber were fixed, stained, and counted (200×, mean ±SD, n = 3). (C and D) S18 cells and cells expressing shRNAs targeting Raptor (shRaptor), Rictor (shRictor), and mTOR (shmTOR), respectively, were subjected to a wound-healing assay for their migration ability. The knockdown efficiencies of shRNAs on each target gene were determined by Western blot after 7 days of puromycin selection. (E) S18 cells and cells expressing shRaptor, shRictor, and shmTOR were assayed for their migration and invasion abilities using Transwell migration and invasion assay (200×; mean ± SD, n = 3).

FIG 6

Inhibition of mTOR signaling delayed tumor growth in vivo. (A) CNE-2 S18 cells (5 × 10⁵ cells, with Matrigel) and cells in that Raptor, Rictor, and mTOR expression had been silenced by specific shRNAs were transplanted subcutaneously into the right flanks of BALB/c-nu/nu mice. After 30 days, tumors were stripped, and tumor mass was weighted. Data are represented as mean ± SD and P value of one-tailed unpaired t test. (B) Tumor sections were examined by H&E staining, IHC for Ki67, and TUNEL staining. (C) The effects of mTOR dual inhibitor BEZ235 on NPC tumor formation were examined in the NPC xenograft model and mice were treated with BEZ235 in a dose of 45 mg/kg or 25 mg/kg body weight daily by intragastric administration. Images of NPC tumors treated with BEZ235 or vehicles are shown. (D to F) The effect of BEZ235 on tumor volume (D), tumor weight (E), and body weight of mice (F) were illustrated.

FIG 7

Schematical illustration of a model for mTOR signaling regulating NPC cancer stem cell characteristics.