Human NPC cell lines namely HONE1 and HK1, were derived from poorly and well differentiated NPC, respectively (13,14). HONE1 and HK1 cells were maintained in RPMI-1640 medium additionally supplemented with 10% fetal bovine serum, 200 U/ml penicillin G sodium, 200 µg/ml streptomycin sulfate, and 0.5 µg/ml amphotericin B in a 37°C humidified incubator with 5% CO₂. Cell irradiation was performed by Gammacell^® 3000 Elan system (Best Theratronics, Ottawa, ON, Canada).

HONE1 and HK1 cells were transiently transfected with miR-138 precursor and negative control (Applied Biosystems, Carlsbad, CA, USA) by HiPerFect reagent (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. After 72 h of transfection, cells were collected and the efficiency of miR-138 precursor transfection was determined.

Total RNA was extracted by TRIzol (Life Technologies, Grand Island, NY, USA) according to the protocol of the manufacturer. High-Capacity cDNA Reverse Transcription kit (Applied Biosystems) was used for reverse transcription. Transcript levels of miR-138-5p and U6 control snRNA were measured by TaqMan Gene Expression assays (Applied Biosystems). The PCR primer-probe pairs for EIF4EBP1 were as follows: forward, 5′-AGCCCTTCCAGTGATGAGC-3′; reverse, 5′-TGTCCATCTCAAACTGTGACTCTT-3′; probe no. 21 of Universal ProbeLibrary (Roche Applied Science, Indianapolis, IN, USA). Primers were synthesized by Integrated DNA Technologies (Coralville, IA, USA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an endogenous control and quantified by Universal ProbeLibrary Human GAPD Gene assay (Roche Applied Science). The fold changes were calculated using 2^−∆∆Ct method.

Colony formation assay was carried out in 6-well cell culture plate with 600 cells per well. Cells were irradiated at a single dose of 2, 4, or 6 Gy and allowed to grow for two weeks. Then, colonies were stained with 0.5% crystal violet after fixation with 70% ethanol and colonies with >50 cells were counted.

Proliferation assay was performed by using RTCA DP instrument of xCELLigence Real-Time Cell Analyzer (Roche Applied Science). Cells were inoculated into E-Plate 16 and were irradiated by Gammacell 3000 Elan system (Best Theratronics). The proliferation rates of HONE1 and HK1 cells were examined constantly.

HONE1 and HK1 cells were plated on chamber slides and irradiated at a single dose of 2, 4, or 6 Gy. Then, cells were washed, fixed and incubated with rabbit polyclonal anti-γH2AX antibodies (Abcam, Cambridge, UK) and CF™488A Secondary Antibody Conjugates (Biotium, Hayward, CA, USA). Nucleus and F-actin were counterstained with DAPI (Life Technologies) and Alexa Fluor^® 635 phalloidin (Life Technologies) respectively. Images were taken under a fluorescence microscope (Nikon, Tokyo, Japan) and the number of γH2AX foci in the cells was counted.

AO (Sigma-Aldrich, St. Louis, MO, USA) staining was performed to detect the formation of acidic vesicular organelles (AVOs) during autophagy process. Cells were stained with AO at the concentration of 4 µg/ml for 15 min at 37°C. Stained cells were observed by a fluorescent microscope (Nikon).

Affymetrix HG-U133 Plus 2 array (Affymetrix, Santa Clara, CA, USA) was used to detect global gene expression profiling. The quality of total RNA was evaluated using Agilent 2100 Bioanalyzer (Agilent Technologies). Data analysis was performed using GeneSpring version 13 (Agilent Technologies). Microarray was conducted by the Centre for Genomic Sciences, the University of Hong Kong.

The sense and antisense strands of wild-type and mutant 3′-UTR of EIF4EBP1 were synthesized. The oligonucleotides were as follows: wild-type sense strand: 5′-CCAGGGCACCTGCCCCCTCCTCTTCGTGAACACCAGCAGATACCTCCTTGTGA-3′; wild-type antisense strand: 5′-AGCTTCACAAGGAGGTATCTGCTGGTGTTCACGAAGAGGAGGGGGCAGGTGCCCTGGAGCT-3′; mutant sense strand: 5′-CCAGGGCACCTGCCCCCTCCTCTTCGTGAAAGTACTAAGATACCTCCTTGTGA-3′; mutant antisense strand: 5′-AGCTTCACAAGGAGGTATCTTAGTACTTTCACGAAGAGGAGGGGGCAGGTGCCCTGGAGCT-3′. The annealed sense and antisense strands of wild-type or mutant 3′-UTR of EIF4EBP1 were cloned into the SacI and HindIII sites of pMIR-Report Luciferase vector (Applied Biosystems) to produce Luc-wild-type or Luc-mutant construct, respectively. HONE1 cells were co-transfected with 200 ng Luc-wild-type vector or Luc-mutant vector, 50 nM miR-138 precursor or negative control (Applied Biosystems), together with 200 ng pMIR-Report β-galactosidase control vector (Applied Biosystems) using Lipofectamine 2000 (Life Technologies). After 48 h, firefly luciferase and β-galactosidase activities were determined using Dual-Light luminescent reporter gene assay kit (Applied Biosystems) on an LB 96V microplate luminometer (EG&G Berthold, Bad Wildbad, Germany).

EIF4EBP1 siRNA-1, EIF4EBP1 siRNA-2 and negative control siRNA (Qiagen) were transfected into HONE1 and HK1 cells using HiPerFect transfection reagent (Qiagen) in accordance with the protocol from the manufacturer. After 72 h, cells were collected and the efficiency of EIF4EBP1 silencing was determined by QPCR.

Student's t-test was employed to compare quantitative variables between two groups. All the tests were 2-sided. Results were only considered statistically significant if p<0.05.

To evaluate the effect of miR-138-5p on radiosensitivity of NPC cells, we overexpressed miR-138-5p by transfection of miR-138 precursor. A dose-dependent increase in miR-138-5p expression was observed (Fig. 1A). Colony formation assay was performed to assess the sensitivity to radiation treatment. Cells transfected with 10 nM miR-138 precursor or negative control were exposed to radiation at the dose of 2, 4 or 6 Gy and the survival fraction was counted after 14 days. An apparent decrease in colony number was observed in cells transfected with miR-138 precursor in comparison with negative control under radiation treatment (Fig. 1B and C).

The proliferation rate after radiation treatment was used to assess the radiosensitivity. Cells transfected with 10 nM miR-138 precursor were exposed to radiation at 24 hours after transfection and cell proliferation was continuously monitored. Although miR-138 precursor itself did not affect the proliferation rate of cells without radiation exposure, cells transfected with miR-138 precursor exhibited a significant reduction in proliferation rate in a time-dependent manner under radiation treatment (Fig. 2A and B). Radiation killed cells by inducing DNA double-strand breaks (DSBs). DSBs were recognized by varied protein complexes including γH2AX. Staining of cells with anti-γH2AX antibody generated foci corresponding to DSBs in nuclei (15). Overexpression of miR-138-5p resulted in an obvious increase in the number of γH2AX foci after radiation treatment (Fig. 2C and D), indicating a higher degree of radiation-induced DNA damage. Radiation could induce autophagy in cancer cells. Formation of AVOs was a characteristic feature of autophagy. In AO-stained cells, cytoplasm and nucleolus displayed green fluorescence. In contrast, AVOs fluoresced bright red (16). An increase in red signal was observed in NPC cells transfected with miR-138 precursor compared to negative control after exposure to radiation (Fig. 2E), implicating a higher level of radiation-induced autophagy.

First, we used prediction databases to predict the targets of miR-138-5p. Considering that individual prediction database probably generated false positive targets, we employed Bioinformatics resource manager v2.3 (BRM) software (http://www.sysbio.org/dataresources/brm.stm) for prediction (17). BRM utilizes three databases including microCosm/miRBase, TargetScan and miRNA.org to predict targets. Those targets that are conserved across all three databases were considered as high-confidence ones. By the prediction function of BRM, 225 genes were predicted to be targets of miR-138-5p by all three databases. Given that miRNA promoted the degradation of target mRNA, we carried out microarray analysis to identify transcripts downregulated by miR-138-5p, which were its putative targets. The global gene expression profiling of HONE1 cells transfected with miR-138 precursor was examined. With a fold-change cut-off of 3 and p<0.05, 186 genes were found to be downregulated by miR-138-5p. Then, the 225 high-confidence targets were integrated with the 186 transcripts downregulated by miR-138-5p. By this novel strategy, we identified 5 potential targets including DMKN, EIF4EBP1, NINJ1, PER1 and ST6GALNAC4 (Fig. 3A).

The binding sites between miR-138-5p and 3′-UTR of EIF4EBP1 are shown in Fig. 3B. Results from QPCR further confirmed the downregulation of EIF4EBP1 by miR-138-5p (Fig. 3C). To explore whether miR-138-5p modulated the expression of EIF4EBP1 by binding to the 3′-UTR, luciferase reporter assay was performed in HONE1 cells transfected with luciferase vector containing 3′-UTR of EIF4EBP1 together with miR-138 precursor or negative control. Overexpression of miR-138-5p significantly reduced the luciferase activity of luciferase vector harboring wild-type 3′-UTR of EIF4EBP1 (Fig. 3D). In contrast, mutation of the binding sites within 3′-UTR of EIF4EBP1 abrogated the suppressing effect (Fig. 3D).

To gain insights into the clinical implications of EIF4EBP1 in NPC, the expression of EIF4EBP1 was analyzed in publicly available microarray dataset GSE12452 in the Oncomine database (www.oncomine.org) (18). GSE12452 contained expression data of 31 NPC tissues and 10 normal controls. A 1.516-fold increase in EIF4EBP1 expression was observed in NPC tissues in comparison with normal controls (Fig. 3E).

To investigate the role of EIF4EBP1 in radiosensitivity of NPC cells, the expression of EIF4EBP1 was silenced by 2 specific siRNA. Transfection of siRNA at the concentration of 2.5 nM successfully inhibited the expression of EIF4EBP1 in HONE1 cells, while siRNA at the concentration of 10 nM was required to suppress the expression in HK1 cells (Fig. 4A). NPC cells transfected with EIF4EBP1 siRNA were exposed to radiation treatment at the dose of 2, 4 or 6 Gy and the ability to form colony was measured after 14 days. NPC cells with silenced expression of EIF4EBP1 displayed a significant decrease in the number of colonies after radiation treatment compared to negative control siRNA (Fig. 4B and C). Results from AO staining showed that silence of EIF4EBP1 enhanced radiation-induced autophagy as evidenced by an increase in red signal under radiation treatment (Fig. 4D).