MiRNA-203 Reduces Nasopharyngeal Carcinoma Radioresistance by Targeting IL8/AKT Signaling Free

Nasopharyngeal carcinoma (NPC) is the most frequent head and neck tumor in southern China and Southeast Asia, which poses one of the most serious public health problems in these areas (1). Radiotherapy is the major therapeutic modality used to treat NPC. Although NPC is sensitive to radiotherapy, a major impediment to achieve long-term survival is radioresistance that has been linked to an increased likelihood of recurrence and a distant metastasis (2, 3).

MiRNAs (miR) are endogenous small non-coding RNAs that act as crucial gene regulators at posttranscriptional level by binding to the 3′ untranslated region (UTR) of target mRNAs. MiRs are believed to have fundamental roles in the human cancers, and have great potential in the diagnosis and treatment of cancers (4). Regulation of tumor radiosensitivity via miRs-associated mechanisms has attracted much attention in the recent years (5–8). Over the past few years, several miRs involving in tumor radioresistance such as miR-21 (9), miR-95 (10), miR-23b (11), let7 (12), miR-205 (13), miR-210 (14), miR-181a (15), miR-125b (16), and miR-324-3p (17) have been identified.

We previously used microarrays to compare the differences of miR and mRNA expression profiles in the NPC cell lines with different radiosensitivity, and found that miR-203 is downregulated, but IL8 is upregulated in the radioresistant NPC cells (18). Bioinformatics analysis identified IL8 as a putative target gene of miR-203 (18). MiR-203 is frequently downregulated and functions as a potential tumor suppressor in the various types of cancers, including NPC (19–21). Recent studies showed that the miR-203 expression level positively correlates with chemosensitivity in colon (22) and prostate (23) cancer cells. However, the function and mechanism of miR-203 in tumor radioresistance have not been characterized.

IL8 is a proinflammatory cysteine-X-cysteine (CXC) chemokine (24). Previous studies have shown that IL8 promotes tumor angiogenesis, growth, and metastasis in the various types of cancers, including NPC (25–31). Increased serum and tissue IL8 levels are associated with worse prognosis of NPC patients, and can serve as an independent prognostic factor for overall patient survival (31, 32). However, it is undetermined whether high IL8 expression in NPC contributes to radioresistance, leading to worse patient prognosis. IL8 executes its biologic functions by activating cellular signaling pathways. AKT signaling plays a role in the tumorigenesis and progression of NPC (33). It has been reported that irradiation of NPC cells can activate AKT (34), and activated AKT is associated with NPC radioresistance (13, 35). Activation of AKT by IL8 signaling, which promotes tumor metastasis, has been shown in NPC (31). However, it is unknown whether activation of AKT by IL8 increases NPC radioresistance.

In this study, we found that miR-203 expression was frequently downregulated in the radioresistant NPC tissues, restoration of miR-203 expression decreased NPC radioresistance both in vitro and in vivo, and reduced miR-203 increased NPC radioresistance by targeting IL8/AKT signaling. Our data strongly suggest that the miR-203 expression level could serve as a potential marker for predicting NPC radioresponse and patient prognosis, and NPC patients with reduced miR-203 may benefit from specific targeted therapies directed at miR-203/IL8/AKT signaling.

Radioresistant human NPC cell lines CNE2-IR and CNE1-IR cells as well as their corresponding radiosensitive cell lines CNE2 and CNE1 cells were previously established by us (36, 37), and cultured with RPMI-1640 medium containing 10% FBS (Invitrogen). Radioresistant CNE2-IR and CNE1-IR cells were derived from parental CNE2 and CNE1 cells, respectively, by treating the cells with four rounds of sublethal ionizing radiation (36, 37). Radiosensitive CNE2 and CNE1 cells, used as a control, were treated with the same procedure except sham irradiated. Experiments were performed with the CNE2-IR and CNE1-IR cells within 4 to 10 passages after the termination of irradiation, and their radioresistance was tested by a clonogenic survival assay before use.

One hundred and eleven NPC patients without distant metastasis (M0 stage) at the time of diagnosis who were treated by radical radiotherapy alone in the Affiliated Cancer Hospital of Central South University, China between January 2006 and December 2008 were recruited in this study. The radiotherapy was administered for a total dose of 60–70 Gy (2 Gy/fraction, 5 days a week). The neck received 60 Gy for lymph node–negative cases and 70 Gy for lymph node–positive cases. NPC tissue biopsies were obtained at the time of diagnosis before any therapy, fixed in 4% formalin and embedded in paraffin. We also acquired 16 cases of formalin-fixed and paraffin-embedded normal nasopharyngeal mucosa in the same period. On the basis of the 1978 WHO classification (38), all tumors were histopathologically diagnosed as poorly differentiated squamous cell carcinomas (WHO type III). The clinical stage of the patients was classified according to the 2008 NPC staging system of China (39).

The radiotherapy response was evaluated clinically for primary lesions based on nasopharyngeal fiberscope and MRI one month after the initiation of radiotherapy according to the following criteria. Radioresistant NPC patients were defined as ones with persistent disease (incomplete regression of primary tumor and/or neck lymphonodes) at >3 months or with local recurrent disease at the nasopharynx and/or neck lymphonodes at ≤12 months after completion of radiotherapy. Radiosensitive NPC patients were defined as ones without the local residual lesions (complete regression) at >3 months and without local recurrent disease at >12 months after completion of radiotherapy. On the basis of the above criteria, 111 NPC patients comprised 53 radioresistant and 58 radiosensitive ones.

The patients were followed up, and the follow-up period at the time of analysis was more than 72 months (average, 77.5 ± 11.8 months). Disease-free survival (DFS) was calculated as the time from the completion of primary radiotherapy to the date of pathologic diagnosis or clinical evidence of local failure and/or distant metastasis. Overall survival (OS) was defined as the time from the initiation of primary radiotherapy to the date of cancer-related death or when censured at the latest date if patients were still alive. The clinicopathologic parameters of the patients used in the present study are shown in Supplementary Table S1.

Total RNA was extracted from NPC cells with TRizol reagent (Invitrogen), or from paraffin-embedded NPC tissues with the RecoverAll Total Nucleic Acid Isolation Kit (Ambion) according to the manufacturer's instructions. qRT-PCR detection of miR-203 and IL8 expression in NPC cells and tissues was performed as previously described (18). The products were quantitated using the 2^−ΔΔCt method against GAPDH or U6 for normalization. The primer sequences are listed in Supplementary Table S2.

A dual luciferase reporter plasmid with IL8 3′UTR (HmiT009678-MT01; GeneCopoeia), without IL8 3′UTR of (CmiT000001-MT01; GeneCopoeia), or with mutated IL8 3′UTR in the predicted miR-203–binding sites constructed by GeneCopoeia, and miR-203 or control mimic (RiboBio) were cotransfected into NPC cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Cells were harvested 48 hours after transfection, both firefly luciferase and renilla luciferase activities were measured using the dual-luciferase reporter assay system (Promega) according to the manufacturer's instruction, and luciferase activity was estimated using a luminometer (Promega).

Cells were cultured with RPMI-1640 medium containing 10% FBS overnight, and then 50 μmol/L miR-203 mimic, miR-203 inhibitor, and their respective negative control (Ribobio) were transfected into cells using the riboFect CP Transfection Kit (Ribobio) according to the manufacturer's instruction, respectively. Twenty-four hours after transfection, cells were subjected to further analysis.

Psi-LVRU6GP-IL8 shRNA and psi-LVRU6GP-scramble non-target shRNA vectors, which were established by GeneCopoeia and confirmed by sequencing, were transfected into CNE2-IR cells using Lipofectamine 2000, respectively. Cells were selected using puromycin for 2 weeks, and stable knockdown of IL8 CNE2-IR cell lines and control cell lines were obtained. The targets for human IL8 shRNAs are shown in Supplementary Fig. S1.

A clonogenic survival assay was performed as previously described by us (36). Briefly, cells were exposed to a range of radiation doses (1–8 Gy) and 12 days after irradiation surviving colonies were stained with 0.5% crystal violet and counted. The survival fraction was calculated as the numbers of colonies divided by the numbers of cells seeded times plating efficiency. Radiation dose–response curves were created by fitting the data to the linear quadratic equation S = e^{−αD−βD^2} using GraphPad Prism 5.0, where S is the surviving fraction, α and β are inactivation constants, and D is the dose in Gy. The AUC that represent the mean inactivation dose (MID) was calculated using GraphPad Prism 5.0. The radiation protection factor (RPF) was calculated by dividing the MID of the test cells by the MID of control cells.

Nude male mice that were 4-weeks-old were obtained from the Laboratory Animal Center of Central South University (Changsha, China). A total of 5 × 10⁶ CNE2-IR cells were injected s.c. into the right flanks of 5-week-old nude mice. When the xenograft volumes reached approximately 50 mm³, the transplanted mice were randomly divided into 2 groups (n = 5 mice each), 5 nmol control or miR-203 agomir (RiboBio) in 25 μL saline buffer was intratumorally injected into the tumor mass at multiple sites per mouse, and next day an 8-Gy dose of ionizing radiation was delivered to the tumor. Three days after irradiation, 5 nmol control or miR-203 agomir was intratumorally injected into the tumor mass at multiple sites per mouse. Five weeks after irradiation, the mice were killed by cervical dislocation, and their tumors were excised, weighted, and cut in half, with one half fixed embedded in paraffin for TUNEL and immunohistochemical staining, and the remaining half flash-frozen in liquid nitrogen until use. Tumor volume (in mm³) was measured by caliper measurements performed every 3 to 4 days and calculated by using the modified ellipse formula (volume = length × width²/2).

Flow-cytometry analysis of cell cycle in response to irradiation was performed as previously described by us (36).

Hoechst 33258 staining was performed to detect apoptotic cells after irradiation as previously described by us (37).

Proteins were exacted from cells and tissues. An equal amount of protein in each sample was subjected to SDS-PAGE separation, followed by blotting onto a polyvinylidene difluoride membrane. After blocking, blots were incubated with anti-IL8 (ab18672; Abcom), phospho-AKT (Thr308; #4056; Cell Signaling Technology), or AKT antibody (#4691; Cell Signaling Technology) overnight at 4°C, followed by incubation with horseradish peroxidase–conjugated secondary antibody for 1 hour at room temperature. The signal was visualized with an enhanced chemiluminescence detection reagent (Pierce). β-Actin was detected as a loading control.

Terminal deoxynucleotidyl transferase(TdT)–mediated dUTP nick end labeling (TUNEL) was performed to detect apoptotic cells of formalin-fixed and paraffin-embedded tissue sections of xenograft tumors after irradiation with the In Situ Cell Death Detection Kit (Roche) according to the manufacturer's instruction. Quantitative evaluation of apoptotic cells was done by examining the sections in 10 random microscopic fields and counting the number of TUNEL-positive cancer cells among 1,000 carcinoma cells under the light microscope. The apoptotic index was expressed as positive cells per 100 cancer cells.

Immunohistochemical staining of IL8, phospho-AKT (Thr308), and γH2AX (phospho-S139) was performed on formalin-fixed and paraffin-embedded tissue sections. Briefly, after antigen retrieval tissue sections were incubated with anti-IL8 antibody (ab18672; Abcom), phospho-AKT1 (Thr308; ab5626; Abcom), or γH2AX antibody (ab2893, Abcom) overnight at 4°C, and then were incubated with biotinylated secondary antibody followed by avidin–biotin peroxidase complex (DAKO). Finally, tissue sections were incubated with 3′, 3′-diaminobenzidine (Sigma) and counterstained with hematoxylin. In negative controls, primary antibodies were omitted. The immunoreactions were evaluated independently by two pathologists as described previously. Staining intensity was categorized: Absent staining as 0, weak as 1, moderate as 2, and strong as 3. The percentage of stained cells was categorized as no staining = 0, <30% of stained cells = 1, 30% to 60% = 2, and >60% = 3. The staining score (ranging from 0 to 6) for each tissue was calculated by adding the area score and the intensity score. A combined staining score of ≤3 was considered to be low expression; and a score of >3 was considered to be high expression. Quantitative evaluation of DNA-damaged cells was done by examining the sections in 10 random microscopic fields and counting the number of γH2AX-positive nuclei among 1,000 carcinoma cells under the light microscope. The rate of DNA-damaged cells was expressed as positive cells per 100 cancer cells.

Cells were cultured with RPMI-1640 medium containing 10% FBS for 24 hours, and then incubated in a serum-free medium for an additional 12 hours. Conditioned medium was collected, followed by filtration to remove the debris, and IL8 concentration in the conditioned medium was measured using the Human IL8 Sandwich ELISA Kit (Excell) according to the manufacturer's instruction.

Cells were cultured with RPMI-1640 medium containing 2% FCS and 2.5 μg/mL mouse anti-human IL8 antibody (ab18672; Abcom) or 2.5 μg/mL mouse control IgG1 (ab188776; Abcom) for 24 hours, and then cells were subjected to further analysis.

Cells were cultured with RPMI-1640 medium containing 10% FBS for 24 hours, and incubated in a serum-free medium for an additional 12 hours, and then recombinant human IL8 (Life Technologies) were added into the medium. Twelve hours after stimulation, cells were subjected to further analysis.

Cells were cultured with RPMI-1640 medium containing 10% FBS one day before transfection. The transfection of 2 μmol/L DNAzyme-targeting AKT1 (the sequence: 5′-TGGTCCACAGGCTAGCTACAACGACCTGCGGCC-3′) was performed with oligofectamine reagent (Life technologies) as previously described (40). Forty-eight hours after transfection, cells were subjected to further analysis.

All experiments were carried out at least three times. Data were presented as the mean ± SD. Statistical analysis was conducted using SPSS 20.0 software. For comparisons between two groups, a Student t test or χ² test was used. Survival curves were obtained by using the Kaplan–Meier method, and comparisons were made by using the log-rank test. Univariate and multivariate survival analyses were conducted on all parameters by using the Cox proportional hazards regression model. The Spearman rank correlation coefficient was used to determine the correlation between two parameters. P values less than 0.05 were considered to be statistically significant.

This study was approved by the ethics committee of Xiangya School of Medicine, Central South University, China. Written informed consent was obtained from all participants in the study. All animal experiments were undertaken in accordance with the Guide for the Care and Use of Laboratory Animals of Central South University, with the approval of the Scientific Investigation Board of Central South University.

Our previous integrated analysis of differential miR and mRNA expression profiles in the radioresistant CNE2-IR and radiosensitive CNE2 cells identified 11 differential miRs anti-correlated with mRNA expression (Supplementary Table S3; ref. 18), including miR-23a and let-7e, whose family members have been demonstrated to be associated with tumor radioresponse (11, 12, 41). We were interested in reduced miR-203 in the CNE2-IR cells, because the function and mechanism of which in tumor radioresistance have not been characterized.

Using qRT-PCR, we confirmed that miR-203 expression was significantly decreased in the CNE2-IR cells compared with CNE2 cells, and in the additional radioresistant NPC CNE1-IR cells compared with radiosensitive CNE1 cells (Fig. 1A). We proceeded to detect miR-203 levels in a cohort of NPC tissues. Compared with the radiosensitive NPCs, miR-203 expression was significantly decreased in the radioresistant NPCs (Fig. 1A), and negatively correlated with NPC radioresistance (r = −0.70, P < 0.001). The cutoff value of miR-203, which could be used to differentiate between the radioresistant and radiosensitive NPCs, was 1.05. It was determined by receiver-operating characteristic analysis (Supplementary Fig. S2) and used to separate the patients into low and high miR-203 level groups. Survival analyses revealed that low miR-203 level in NPCs correlated with the markedly reduced DSF and OS of the patients (Fig. 1B). A univariate Cox regression analysis showed that the miR-203 expression level and clinical TNM stage significantly affected the DFS and OS of NPC patients (Table 1). A multivariate Cox regression analysis confirmed that low miR-203 expression was an independent predictor for the reduced DFS and OS of NPC patients (Table 1). These results indicated the importance of the miR-203 expression level in the NPC radioresistance and patient prognosis.

To determine the effect of reduced miR-203 on NPC cell radioresistance in vitro, CNE2-IR cells were transiently transfected with control or miR-203 mimic, and then cell radiosensitivity was determined. A clonogenic survival assay showed that transfection of miR-203 significantly decreased CNE2-IR cell radioresistance compared with transfection of control mimic [AUC, 1.14 (miR-203 mimc) vs. 1.86 (control mimic); P < 0.05; RPF = 0.61; Fig. 2A]. Irradiation primarily leads to double-strand DNA breaks (DSB), and unrepaired or misrepaired DSBs in the DNA lead to cell apoptosis. The apoptosis resulting from irradiation is, to a considerable degree, understood as radiosensitivity (41). Therefore, we also analyzed the effect of miR-203 mimic on the irradiation-induced apoptosis of CNE2-IR cells. Hoechst 33258 staining showed that transfection of miR-203 mimic significantly increased irradiation-induced apoptosis of CNE2-IR cells compared with transfection of control mimic (Fig. 2B). Our previous study showed that compared with radiosensitive CNE-2 cells, more CNE-2-IR cells were found detained in S phase with less cells in G₂–M phase after 6 Gy irradiation (36), which is consistent with the typical radioresistant phenotype (42). Accordingly, the difference in response to radiation between control or miR-203 mimic–transfected CNE2-IR cells was further studied by cell-cycle analysis using flow cytometry. As showed in Fig. 2C, no difference was induced by irradiation in G₀–G₁ phase at 24 hours after 6 Gy irradiation, whereas compared with control mimic–transfected cells, less miR-203 mimic–transfected cells were found detained in S phase with more miR-203 mimic–transfected cells in G₂–M phase, which is consistent with the typical radiosensitive phenotype (42). Moreover, transfection of miR-203 mimic into the additional radioresistant NPC CNE1-IR cells was also able to reduce cell radioresistance [AUC, 1.97 (miR-203 mimic) vs. 2.56 (control mimic); P < 0.05; RPF = 0.77; Fig. 2A]. Taken together, these results demonstrated that restoration of miR-203 expression in the radioresistant NPC cells significantly decreased in vitro cell radioresistance, i.e. that miR-203 could sensitize NPC cells to irradiation.

To determine the effect of reduced miR-203 on NPC radioresponse in vivo, we generated subcutaneous tumors in nude mice using CNE2-IR cells. Control or miR-203 agomir was injected into the tumors before and after 8 Gy ionizing radiation, and then tumor radioresponse was assessed. As shown in Fig. 3A, radioresistance of miR-203 agomir–injected tumors was significantly lower than that of control agomir–injected tumors as demonstrated by tumor growth and weight. Hematoxylin and eosin (H&E) staining of tumor tissue sections showed that more necrosis was noted in miR-203 agomir–injected tumors compared with control agomir–injected tumors [29.90% ± 4.16% (miR-203 agomir) vs. 13.32% ± 3.53% (control agomir); P < 0.01; Fig. 3B]. TUNEL assay showed that more apoptotic cells were present in the miR-203 agomir–injected tumors relative to control agomir–injected tumors [27.3% ± 4.21% (miR-203 agomir) vs. 8.92% ± 1.60% (control agomir); P < 0.01; Fig. 3C]. Immunohistochemical staining indicated that more positive cells of γH2AX, that is, more cells with DNA damage, were present in the miR-203 agomir–injected tumors compared with control agomir–injected tumors [18.35% ± 4.28% (miR-203 agomir) vs. 7.34% ± 2.27% (control agomir); P < 0.01; Fig. 3D]. Taken together, these results demonstrated that restoration of miR-203 expression obviously decreased in vivo NPC radioresistance, suggesting that in vivo administration of miR-203 agomir had a considerable potential for NPC radiosensitization.

To confirm IL8 as a direct target of miR-203, we cotransfected a dual luciferase reporter plasmid with wild-type (wt) IL8 3′UTR into CNE2-IR cells with control or miR-203 mimic. The results revealed a significant reduction in luciferase activity in miR-203 mimic–transfected cells compared with control mimic–transfected cells, whereas miR-203 mimic had no obvious effects on the luciferase activity of a dual luciferase reporter plasmid without IL8 3′UTR or with mutated IL8 3′UTR in the two miR-203–binding sites (Fig. 4A). Furthermore, cellular and secretory IL8 levels were significantly decreased in the miR-203 mimic–transfected CNE2-IR cells, whereas significantly increased in the miR-203 inhibitor–transfected CNE2 cells compared with their corresponding controls (Fig. 4B). Taken together, these results proved that IL8 was a direct target of miR-203 in NPC cells. Notably, IL8 level was significantly downregulated in the miR-203 agomir–injected xenografts compared with control agomir–injected xenografts (Fig. 4C), supporting our in vitro results.

Next, we investigated whether IL8 mediates miR-203–regulated NPC radioresponse. CNE2-IR cell lines with stable knockdown of IL8 and control cell lines were established (Fig. 5A and Supplementary Fig. S1). A clone survival assay showed that IL8 knockdown decreased cell radioresistance [AUC, 1.03 (IL8 shRNA 1) vs.1.72 (scramble shRNA), P < 0.05; RPF = 0.58; AUC, 0.98 (IL8 shRNA 2) vs.1.72 (scramble shRNA), P < 0.05; RPF = 0.55; Fig. 5B], phenocopying that seen in the miR-203 mimic–transfected CNE2-IR cells. Moreover, neutralization of secretory IL8 using anti-IL8 antibody decreased CNE2-IR cell radioresistance compared with control IgG [AUC, 1.42 (IL8 antibody) vs.1.74 (control IgG), P < 0.05; RPF = 0.81; Fig. 5C]. We also investigated whether exogenous IL8 increases NPC cell radioresistance, and found that exogenous IL8 stimulation increased radioresistance of radiosensitive CNE-2 cells as demonstrated by a clone survival assay [AUC, 1.28 (1.5 ng/mL IL8) vs.1.01(control), P < 0.05; RPF = 1.27; AUC, 1.48 (4.5 ng/mL IL8) vs.1.01(control), P < 0.05; RPF = 1.46; Fig. 5D]. Importantly, exogenous IL8 stimulation markedly abolished the radiosensitizing effect of miR-203 mimic in the radioresistant CNE2-IR cells [AUC, 1.59 (miR-203 mimic plus IL8) vs. 1.12 (miR-203 mimic), P < 0.05; RPF = 1.42; Fig. 5E], and IL8 knockdown markedly abolished radioresistance induced by transfection of miR-203 inhibitor in the radiosensitive CNE2 cells [AUC, 1.04 (miR-203 inhibitor plus IL8 knockdown) vs. 1.43 (miR-203 inhibitor plus scramble shRNA), P < 0.05; RPF = 0.73; Fig. 5F]. Taken together, our results demonstrated that miR-203 regulated NPC cell radioresponse through targeting IL8, and reduced miR-203 increased NPC radioresistance by upregulating IL8.

Previous studies indicated that activated AKT plays a critical role in NPC radioresistance (13, 34, 35), and IL8 can activate AKT in NPC (31). Therefore, we investigated whether IL8/AKT signaling mediates miR-203–regulated NPC cell radioresponse. Western blotting showed that either IL8 stimulation or transfection of miR-203 inhibitor significantly enhanced phospho-AKT level in the radiosensitive CNE2 cells, whereas either IL8 knockdown or transfection of miR-203 mimic significantly reduced phospho-AKT level in the radioresistant CNE2-IR cells (Fig. 6A). Importantly, IL8 knockdown abrogated increment of phospho-AKT level induced by transfection of miR-203 inhibitor in the radiosensitive CNE2 cells, and IL8 stimulation restored phospho-AKT level decreased by transfection of miR-203 mimic in the radioresistant CNE2-IR cells (Fig. 6A). Moreover, a clone survival assay showed that either transfection of AKT DNAzyme or PI3 kinase inhibitor LY294002 treatment decreased CNE2-IR cell radioresistance [AUC, 1.32 (DNAzyme) vs. 1.82 (vehicle), P < 0.05; RPF = 0.73; AUC, 1.31 (LY294002) vs. 1.83 (vehicle), P < 0.05; RPF = 0.72; Fig. 6B]. Importantly, both transfection of AKT DNAzyme and LY294002 treatment significantly abolished radioresistance induced by IL8 stimulation in the radiosensitive CNE2 cells [AUC, 1.11(IL8 plus LY294002) vs.1.51 (IL8), P < 0.05; RPF = 0.73; AUC, 1.17 (IL8 plus DNAzyme) vs. 1.51 (IL8), P < 0.05; RPF = 0.77; Fig. 6C].

In the cohort of NPC tissues, the levels of IL8 and phospho-AKT were significantly higher in the radioresistant NPCs than those in the radiosensitive NPCs (Fig. 6D; Table 2). Correlation analyses revealed that IL8 level was positively associated with phospho-AKT level (r = 0.57, P < 0.001), whereas negatively associated with miR-203 level (r = −0.70, P < 0.001), and phospho-AKT level was negatively associated with miR-203 level (r = −0.61, P < 0.001; Supplementary Fig. S3). Moreover, phospho-AKT expression was significantly reduced in the miR-203 agomir–injected xenografts relative to control agomir–injected xenografts (Fig. 4C). Taken together, these results suggested that IL8/AKT signaling mediated miR-203–regulated NPC cell radioresponse, and reduced miR-203 increased NPC radioresistance by activating IL8/AKT signaling.

Radioresistance is the main obstacle in the clinical management of NPC (2, 3). Investigating the role of miRs in radioresistance is a promising avenue, given their ability to regulate multiple oncogenic processes, including response to therapy (8). In this study, we focused on miR-203, one of downregulated miRs in the radioresistant NPC cells (18), because the roles of miR-203 in tumor radioresistance are unknown.

To gain insight into miR-203 function in NPC radioresistance, we performed in vitro and in vivo radioresponse assays, and found that upregulation of miR-203 decreased radioresistance in NPC cells and xenograft tumors, demonstrating that reduced miR-203 enhances NPC radioresistance. These findings are clinically relevant, given our discovery that miR-203 was frequently downregulated in the radioresistant NPC tissues, and its decrement correlated with NPC radioresistance and poor patient survival, outlining a potential marker for predicting the radioresponse and prognosis of NPC patients. To our knowledge, it is for the first time reported that miR-203 modulates tumor radioresponse.

Our previous bioinformatics analysis identified IL8 as one of miR-203 potential target genes (18). In this study, we proved that IL8 is a direct target of miR-203 in NPC cells, which is for first time reported IL8 as a miR-203 target in tumor cells. Next, we investigated whether IL8 mediates miR-203–regulated NPC radioresponse. We observed that IL8 knockdown or antibody neutralization of secretory IL8 decreased NPC cell radioresistance, whereas IL8 stimulation increased NPC cell radioresistance. Furthermore, IL8 stimulation abolished the radiosensitizing effect of miR-203 mimic, and IL8 knockdown abolished radioresistance induced by transfection of miR-203 inhibitor in NPC cells. In the clinical NPC samples, IL8 level was significantly higher in the radioresistant NPCs than that in the radiosensitive NPCs, and negatively associated with miR-203 level. These results demonstrate that reduced miR-203 enhances NPC radioresistance through targeting IL8.

miRs modulate tumor radiosensitivity by targeting signal pathways (8). Previous studies indicated that activated AKT is associated with NPC radioresistance (13, 34, 35) and IL8 can activate AKT in NPC (31). Therefore, we investigated whether IL8 mediates miR-203–regulated NPC cell radioresponse by activating AKT. We observed that transfection of miR-203 inhibitor or IL8 stimulation enhanced, whereas transfection of miR-203 mimic or IL8 knockdown reduced phospho-AKT level in NPC cells. Importantly, IL8 knockdown abrogated increment of phospho-AKT level induced by transfection of miR-203 inhibitor, and IL8 stimulation restored phospho-AKT level decreased by transfection of miR-203 mimic in NPC cells. miR-203 agomir reduced the levels of IL8 and phospho-AKT in the NPC cell xenografts. Moreover, AKT DNAzyme or LY294002 not only decreased NPC cell radioresistance, but also markedly abrogated NPC cell radioresistance induced by IL8 stimulation. In the clinical NPC samples, phospho-AKT level was significantly higher in the radioresistance NPCs than that in the radiosensitive NPCs, and positively associated with IL8 level whereas negatively associated with miR-203 level. Taken together, our results demonstrate that activation of AKT by IL8 enhances NPC radioresistance, and suggest that reduced miR-203 increases NPC radioresistance by activating IL8/AKT signaling.

Methods for radiosensitization of NPC attract much attention (43–45). In this study, we confirmed that restoration of miR-203 expression by using miR-203 agomir enhanced NPC radiosensitivity in NPC xenografts, suggesting its considerable potential in radiosensitizing NPC. Nucleic-acid drugs, such as miRs, can be directly synthesized and modified to be more lipophilic that improves penetration. Such modification includes cholesterylation. Our delivery of miR-203 agomir, cholesterylated miR mimic, successfully increased NPC radiosensitivity in the intratumoral injection model, suggesting that miR-203 agomir has a potential for further drug development. Our results also highlight the possibility of radiosensitizing NPC with reduced miR-203 by inhibiting IL8/AKT signaling, given that increased level of IL8 and phospho-AKT was seen in the radioresistant NPC tissues, and AKT DNAzyme or LY294002 significantly abrogated NPC cell radioresistance.

Although IL8/AKT signaling seems to largely account for the radioresistant phenotype of NPC cells induced by reduced miR-203, indeed a single miR has been thought to target multiple mRNAs to regulate gene expression (46). Therefore, there may be other molecules or signaling pathways that are also targeted by miR-203(19, 22, 47–50), and some of them may be still unknown in NPC. This presumption may raise interesting future work to reveal the entire functions of miR-203 in NPC radioresistance. Moreover, our results indicated that miR-203 is a potential marker for predicting the radioresponse and prognosis of NPC patients, but this study was a retrospective study that is limited to only 111 patients from a single institution. Therefore, a prospective study involving larger populations from multiple centers will be required to further validate the usefulness of miR-203.

In summary, our data demonstrate that (i) miR-203 is frequently downregulated in the radioresistant NPC tissues, and its decrement significantly correlates with NPC radioresistance, and is an independent predictor for the poor survival of NPC patients; (ii) upregulation of miR-203 decreases NPC cell radioresistance in vitro and in vivo; (iii) reduced miR-203 increases NPC radioresistance through activating IL8/AKT signaling. Our study demonstrates that miR-203 is a critical determinant of NPC radioresponse, and suggests that targeting the miR-203/IL8/AKT signaling axis is a promising approach for enhancing NPC sensitivity to radiotherapy.

No potential conflicts of interest were disclosed.

Conception and design: J.-Q. Qu, H.-M. Yi, Z.-Q. Xiao

Development of methodology: T. Xiao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.-Q. Qu, H.-M. Yi, X. Ye, J.-F. Zhu, H. Yi, L.-N. Li, T. Xiao, L. Yuan, J.-Y. Li, Y.-Y. Wang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.-Q. Qu, H.-M. Yi, J. Feng, Q.-Y. He, S.-S. Lu

Writing, review, and/or revision of the manuscript: J.-Q. Qu, Z.-Q. Xiao

This work was supported by Major Program of National Natural Science Foundation of China (81230053; to Z.Q. Xiao), National Key Technology R&D Program of China (2013BAI01B07; to Z.Q. Xiao), National Basic Research Program of China (2013CB910502; to Z.Q. Xiao), National Natural Science Foundation of China (81172559; to Z.Q. Xiao), and Collaborative Innovation Center for Chemistry and Molecular Medicine of Hunan Province, China (to Z.Q. Xiao).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.