HPV16 E6-E7 induces cancer stem-like cells phenotypes in esophageal squamous cell carcinoma through the activation of PI3K/Akt signaling pathway in vitro and in vivo

High-risk human papillomavirus (HPV), especially HPV16, correlates with cancerogenesis of human esophageal squamous cell carcinoma (ESCC) and we have reported that HPV16 related with a poor prognosis of ESCC patients in China. We aim to investigate the potential role and mechanism of HPV16 in ESCC development and progress. Our following researches demonstrated that ESCC cells which were stably transfected by HPV16 E6-E7 lentiviral vector showed a remarkable cancer stem-like cells (CSCs) phenotype, such as: migration, invasion, spherogenesis, high expression of CSCs marker in ESCC---p75NTR, chemoresistance, radioresistance, anti-apoptosis ability in vitro and cancerogenesis in vivo. HPV16 E6-E7 induced PI3K/Akt signaling pathway activation and this affect could be effectively inhibited by LY294002, a specific PI3K inhibitor. It was also indicated that the inhibition of PI3K/Akt signaling pathway by PI3K and Akt siRNA reverse the effect which induced by HPV16 E6-E7 in ESCC cells. Taken together, the present study demonstrates that HPV16 E6-E7 promotes CSCs phenotype in ESCC cells through the activation of PI3K/Akt signaling pathway. Targeting the PI3K/Akt signaling pathway in HPV16 positive tissues is an available therapeutic for ESCC patients.


INTRODUCTION
Esophageal carcinoma was the sixth and ninth lethal malignancy worldwide in men and women in 2012 [1], respectively. Although the advanced treatment such as radical surgical resection, chemotherapy and radiotherapy are extensively applied, the 5-year survival rate of esophageal squamous cell carcinoma (ESCC) patients remained less than 30%, resulting in 400,200 deaths in 2012 worldwide [2].
The etiology of ESCC has not been fully elucidated although epidemiological and experimental studies have identified several risk factors [1,3,4]. Human papillomavirus (HPV) infection in ESCC was first suggested as one of the risk factors in 1982 [5] and has been detected in ESCC in high-risk areas in China recently [6]. High-risk HPV, including HPV16 and HPV18, expresses oncogenes E6 and E7 which can be bound to the p53 and Rb tumor suppressor, respectively, leading the rapid p53 degradation and the loss of Rb products. This might suggest the etiological involvement of high-risk HPV in ESCC [7][8][9][10].
Cancer stem-like cells (CSCs) are a small subpopulation of tumor cells with the capacity of selfrenewal maintaining tumor growth and cell differentiation. CSCs also contribute to the tumorigenic potential of cancer, including spherogenesis, resistance to cytotoxic drugs and ionizing radiation [11][12][13]. The low-affinity p75 Research Paper www.impactjournals.com/oncotarget neurotrophin receptor (p75NTR) was first identified as a stem/progenitor cell marker in human normal esophageal epithelial cells by Okumura in 2003 [14]. Moreover, p75NTR is necessary for survival and maintenance of ESCC tumors and could be considered as a potential target for novel therapies [15]. The p75NTR positive cells are supposed to be capable of repopulating all known epithelial cell subtypes with a slow cycling rate and a relatively immature phenotype. It could be concluded that p75NTR could be regarded as a candidate biomarker of CSCs in ESCC cells.
Activation of PI3K/Akt signaling pathway was required for colony-formation ability in vitro and tumorigenicity in vivo in breast CSCs [16]. It was also found that targeting PI3K benefits for prostate cancer by eliminate prostate CSCs [17,18]. Up-regulation of hypoxia-inducible factor 1α inhibits proliferation and promotes survival of prostate cancer stem cells through the PI3K [19].Furthermore, PI3K/Akt inhibition is considered to be an effective strategy in the solid-tumor treatment which reduced the bulk tumor burden significantly combined with classical chemotherapy [20].
Our previous study demonstrated that HPV16 infection and highly PI3K expression in ESCC patients were independently associated with poor overall survival [21]. Therefore, a better understanding of ESCC biology with high-risk HPV infection is urgently needed to be detected. In the present study, we first detected the influence and the potential mechanism of HPV16 E6-E7 expression in ESCC cells in vitro and in vivo. LY294002, a specific PI3K inhibitor suppressing PI3K/Akt pathways which plays a vital role in carcinogenesis, was selected due to its small size and ease of use in vitro and in vivo systems. The results demonstrated that HPV16 E6-E7 expression promoted CSCs phenotypes in ESCC cells via activating PI3K/Akt signaling pathway in vitro and in vivo.

Lentivirus transfection of ESCC cells and confirmation of HPV16 E6-E7 expression
HPV16 E6-E7 protein was stably expressed in HPV16 E6-E7 lentiviral vector transfected ESCC cells, which were labeled as Eca109-psb and TE-1-psb cells, while no expression of HPV16 E6-E7 protein was observed in control vector transfected ESCC cells, which were labeled as Eca109-control and TE-1-control cells. The transfection efficiency were detected and confirmed by western blotting ( Figure 1A). The transfection outcome of HPV16 E6-E7 lentiviral vector was observed using a fluorescence microscope 72 h after experimental treatment of Eca109 and TE-1 cells ( Figure 1B). HPV16 E6-E7 promotes CSCs phenotypes, increases the ratio of CSCs marker-p75NTR, and reduces G 0 /G 1

cell cycle arrest in ESCC cells in vitro
The migration and invasion capacities of ESCC cells were conducted by the transwell assay, as results were shown ( Figure 1C and 1E) and quantified ( Figure 1D and 1F). Relative to control cells, there were dramatically (P<0.05-0.01) improvements in the migration capacity or invasion capacity of Eca109-psb and TE-1-psb cells after 24h (without matrigel) migration assay or 48h (with matrigel) invasion assay, respectively. It could be concluded that HPV16 E6-E7 increases migration and invasion potential in ESCC cells in vitro.
The studies above demonstrated that HPV16 E6-E7 behaved as a tumor-promoting factor in ESCC cells. We next examined the effect of HPV16 E6-E7 on the spherogenesis ability of ESCC cells which was considered to be an important characteristic of CSCs in vitro [22][23][24]. It was observed obviously (P<0.001) increased in the number of spheres in Eca109-psb and TE-1-psb cells compared to control cells on day 14 (Figure 2A and 2B). It shows that HPV16 E6-E7 enhances the spherogenesis of ESCC cells in vitro.
Previous studies [14,15] reported that the p75NTR could be a candidate CSCs marker in ESCC cells. In this study, flow cytometry was applied to investigate whether HPV16 E6-E7 increased p75NTR positive cells in ESCC cells, as it increased spherogenesis as demonstrated above. As shown in Figure 2C-2F, significant (P<0.05-0.001) improvement of p75NTR positive cells were detected in Eca109-psb and TE-1-psb cells than control cells. Similarly, it was also found that the p75NTR positive cells significantly (P<0.01-0.001) increased in Eca109-psb and TE-1-psb cell spheres compared to control cell spheres. This suggests that HPV16 E6-E7 increases the ratio of p75NTR positive cells in ESCC cells in vitro.
Flow cytometry was performed to investigate the effect of HPV16 E6-E7 on ESCC cells cycle progression. As shown in Figure 2G and 2H, compared to control cells, HPV16 E6-E7 significantly (P<0.05-0.01) decreased cells that were arrested at the G0/G1 transition evidenced by a less accumulated cells in the G0/G1 peak. A marked (P<0.001) accumulation of cells at the G2/M phase was also observed. This indicates that HPV16 E6-E7 leads to the accumulation of cells in G2/M and extensively decreases the percentage of cells in G0/G1 phase in ESCC cells in vitro.
The colony formation assay was performed to determine the influence of HPV16 E6-E7 on radioresistance in ESCC cells. Cells were divided and exposed to 0, 2, 4, 6, 8 Gy of ionizing radiation ( Figure  3B). The Single hit multi-target model SF=1-(1-e -D/D0 ) n [25] was applied to fitting the result and dose-survival curves of ESCC cells ( Figure 3C). Constant values of   radiation biology were calculated basing on the data of colony formation assay (Supplementary Table S1). It was found that the value of D 0 , k and SF 2 significantly (P<0.001-0.05) increased in Eca109-psb and TE-1-psb cells compared to control cells (Supplementary Table S1) while no significant differences were observed in the value of N, Dq. This indicates that HPV16 E6-E7 also increases radioresistance of ESCC cells in vitro.
Annexin V-APC/PI flow cytometry was applied to assess the effect of HPV16 E6-E7 on ESCC cells apoptosis after ionizing radiation. The fourth quadrant (Q4) represents early apoptosis cells and the second quadrant (Q2) represents late apoptotic and necrotic cells ( Figure  3D). The ratio of apoptosis cells increased with radiation dose accumulating in all the cells ( Figure 3E). Meanwhile, the ratio of apoptosis cells was significantly (P<0.05-0.001) lower in Eca109-psb and TE-1-psb cells than control cells when treated by 4, 6 and 8Gy ( Figure 3E), with no significant differences by 2Gy (P=0.065, 0.087, respectively) or untreated (P=0.118, 0.069, respectively). This means that HPV16 E6-E7 inhibits the apoptosis in ESCC cells when treated by ionizing radiation, especially accumulating and highly dose of ionizing radiation in vitro.

HPV16 E6-E7 increased p75NTR expression through PI3K/Akt signaling pathway in ESCC cells
As described above, more and more recent studies show that PI3K/Akt pathway plays a vital role in the maintenance of CSCs [26][27][28]. So, we speculate that HPV16 E6-E7 may increase the CSCs properties of ESCC cells through up-regulation of PI3K/Akt signaling.
Western blotting and immunofluorescence were applied to examine the hypothesis above. The western results were shown in Figure 4A and relative density were analyzed and quantified as histograms in Figure 4B. The expression of PI3K, phosphorylation-Akt (ser473), also named as p-Akt (ser473), and p75NTR in the Eca109psb and TE-1-psb cells significantly (P<0.001) increased compared to control cells. Whereas, no significant alterations could be observed in the expression of total Akt in Eca109-psb and TE-1-psb cells compared to control cells ( Figure 4A and 4B). Immunofluorescence of ESCC cells ( Figure 4C) demonstrated that PI3K, p-Akt (ser473) and p75NTR expressions were markedly increased in Eca109-psb and TE-1-psb cells compared to control cells. Meanwhile, immunofluorescence of spheres ( Figure 4D) also demonstrated that PI3K, p-Akt (ser473) and p75NTR expression are obviously stronger in spheres derived from Eca109-psb and TE-1-psb cells than spheres derived from control cells. It suggests that HPV16 E6-E7 activates the PI3K/Akt signaling pathway in ESCC cells and spheres derived from ESCC cells.
To investigate whether inhibiting the PI3K/ Akt signaling pathway can reverse CSCs phenotypes caused by HPV16 E6-E7. LY294002 was applied at the concentration of 10μmol/L in the cell culture medium incubating for 24h as our previously study described [29]. Consequently, there were no significant difference in the expression of PI3K, p-Akt (ser473) and p75NTR as they all decreased (P<0.001) in ESCC cells after LY294002 treatment ( Figure 4A and 4B). However, there were still no alternations in the expression of total Akt in the ESCC cells after being treated with LY294002 ( Figure 4A and 4B).
Based on the results above, LY294002 treatment reverse the expression of PI3K, p-Akt (ser473) and p75NTR in Eca109-psb and TE-1-psb cells, and this further demonstrated that the activation of PI3K/Akt signaling pathway, induced by HPV16 E6-E7, plays a vital role in maintaining CSCs phenotypes in ESCC cells [30].
Specific siRNAs were utilized to inhibit PI3K/ Akt pathway for further investigation. Negative control siRNA (NC siRNA) and GAPDH siRNA were both utilized for negative and positive control. Results were shown and relative density were analyzed and quantified as histograms in Figure 5 and Figure 6.  6B, 6D). p-Akt (ser473) and p75NTR expressions were significantly (P<0.05-0.001) higher in Eca109-psb and TE-1-psb cells, compared to control cells, in control siRNA group with no significant difference of that in the Akt siRNA group. GAPDH was used for positive control to make sure siRNA knock-down efficiency and β-actin was used as the endogenous reference.
The xenograft tumour growth markedly (P<0.05) increased in mice bearing Eca109-psb cells, compared to the Eca109-control cells, both in Control group and Radiation group ( Figure 7A, 7B). However, there was no significant difference in the tumor volumes between LY294002 group and the LY294002+radiation group. Moreover, no marked changes in body weight were found in the four groups while the growth of body weight in LY294002+radiation group is the slowest in four groups ( Figure 7D). This indicates that tumors derived from Eca109-psb cells grow markedly and significantly (P<0.05) faster than tumors derived from control cells in Control group and Radiation group. Interestingly, the significant difference could be eliminated when intraperitoneally injected with LY294002.
After tumors were surgically excised and fixed, IHC were performed to detect the mechanism involved in vivo. Consistent with the results of the western blotting in vitro, PI3K, p-Akt (ser473) and p75NTR expression were significantly increased in tumors derived from Eca109psb cells compared with that of control cells in control and radiation groups in vivo ( Figure 7E). However, in the LY294002 and LY294002+radiation groups, no significant difference could be observed in the expression of PI3K, p-Akt (ser473) and p75NTR between the tumors derived from Eca109-psb cells and tumors derived from Eca109-control cells in vivo ( Figure 7E). What's more, the expressions of PI3K, p-Akt (ser473) and p75NTR are all inhibited by LY294002 ( Figure 7E).
Taken together, it could be easily concluded that HPV16 E6-E7 promotes the tumorigenesis and radioresistance in ESCC cells in vivo. This effect could be eliminated by intraperitoneally being injected with LY294002. IHC results suggest that the activation of PI3K/Akt signaling pathway induced by HPV16 E6-E7 could be inhibited by LY294002 in vivo.

DISCUSSION
Cancer development is a highly orchestrated process that origins from complex factors, and in ESCC high-risk HPV infection is one of them [31,32]. Our previous study has proposed that HPV16 infection predicts poor prognosis in ESCC patients [21]. However, no clear mechanism has been reported. The present study firstly aims at finding the influence on cell behavior and potential mechanism after HPV16 E6-E7 infecting the ESCC cells.
It have been proved that cancer recurrence, radioresistance and chemo-resistant were mostly attributed to the presence of CSCs [33][34][35][36]. The specific ability of CSCs is to form spheres in vitro under non-adherent culture conditions [22,37,38]. Meanwhile, it is well established that p75NTR is one of the most important CSCs markers in ESCC [14,39], which mainly express in the basal layer of esophageal epithelium [40]. Our previous study also demonstrated that p75NTR positive cells significantly increased in Eca109R-50Gy cells (Eca109 cells achieved by accumulated 50 Gy ionizing radiation with high radioresistance and characteristics of CSCs), compared to Eca109 cells [41,42]. Basing on this, we sent out to investigate the role of HPV16 E6-E7 in the biological behavior of ESCC cells. Transwell assay in this study found that HPV16 E6-E7 promoted the migration and invasion ability significantly ( Figure  1C-1F). The spherogenesis assay performed in this study found that HPV16 E6-E7 also induced spherogenesis in ESCC cells (Figure 2A, 2B). Next, flow cytometry was applied to analysis of p75NTR positive cells, and the results showed that HPV16 E6-E7 induced the stemness in ESCC cells because of the increased ratio of p75NTR positive cells in Eca109-psb, TE-1-psb cells and spheres derived from them ( Figure 2C-2F). All of the results above indicate that HPV16 E6-E7 induces CSCs phenotypes in ESCC cells.
One of the important regulatory mechanisms of cell growth is the cell cycle distribution [43]. Cell cycle analysis showed that HPV16 E6-E7 caused an accumulation of cells in G2/M phase with significantly reduction in G0/G1 phase ( Figure 2G, 2H). In the cell proliferation analysis, CCK8 cell viability assay ( Figure 3A) and colony formation assay ( Figure 3B, 3C, Supplementary Table S1) suggested HPV16 E6-E7 promoted chemoresistance and radioresistance in ESCC cells, respectively. Then the cell apoptosis assay performed by flow cytometry analysis revealed that HPV16 E6-E7 increased the anti-apoptotic ability of ESCC cells when treated by ionizing radiation (Figure 3D, 3E). All in all, HPV16 E6-E7 increases the chemoresistance, radioresistance and plays an anti-apoptotic effect in ESCC cells.
As the results described above, the potential mechanism that HPV16 E6-E7 induces CSCs phenotypes in ESCC should be detected. Many signaling pathways are commonly involved in maintaining the CSCs phenotypes [44,45] and the PI3K/Akt is one of them in the ESCC [26,46]. Side population cells (SP cells) which could not be stained by Hoechst 33342 dye have been proven to be enriched in CSCs in various tumors and cells [47][48][49]. The PI3K/Akt regulates the SP cells in ESCC cells and specimens [50]. Akt is the major effector kinases in the downstream of PI3K/Akt pathway, inhibition of PI3K/ Akt activity attenuate CSCs metabolism and is critical for CSCs quiescence [19,20].
So we investigated the PI3K/Akt signaling pathway in ESCC cells with LY294002, PI3K siRNA and Akt siRNA. Western blotting indicated that HPV16 E6-E7 increased PI3K, p-Akt (ser473) and p75NTR expression in ESCC cells ( Figure 4A, 4B, Figure 5 and Figure 6). Akt siRNA decreases the p-Akt(ser473) and p75NTR expression in Eca109-psb and TE-1-control, TE-1-psb cells ( Figure 6). Immunofluorescence also demonstrated that HPV16 E6-E7 improved PI3K, p-Akt (ser473) and p75NTR expression in ESCC cells and spheres from ESCC cells (Figure 4C and 4D). All the data above show that inhibiting PI3K/Akt signaling pathway decreases the p75NTR expression in Eca109-psb and TE-1-psb cells. Consistently, Tetsuji et.al [51] also proved that p75NTR expression was applied for CSCs identification and isolation in ESCC cells. Moreover, the PI3K/Akt/c-Myc Axis was closely associated with CSC-like features is ESCC cells as well as there was significant correlation between strong c-Myc expression and a short overall survival in ESCC patients [30,50].
The xenograft study found that the proliferation effect significantly promoted by HPV16 E6-E7 in Eca109 cells could be inhibited by LY294002 in vivo ( Figure 7A,  7B). Furthermore, the IHC found the expression of PI3K, p-Akt (473) and p75NTR in the Eca109-psb cells tissues is definitely higher than control cells tissues in the Control group and Radiation group while no difference could be seen in LY294002 group and LY294002+Radiation group ( Figure 7E). This further confirms our hypothesis that HPV16 E6-E7 promotes CSCs phenotypes of ESCC cells which can be partly inhibited by the blocking of PI3K/Akt signaling pathway in vivo.
In summary, although many previous studies declare the important role of HPV16 infection in the incidence of ESCC [5,7], the present study is firstly investigate the potential role and mechanism of HPV16 E6-E7 in the development of ESCC. It is supposed that HPV16 E6-E7 function as a CSCs phenotypes promoter in ESCC cells through activation of PI3K/Akt signaling pathway with the up-regulate of p75NTR, a CSCs marker in ESCC cells. However, it is also indicated that the inhibition of PI3K/ Akt signaling pathway reverses the effect which induced by HPV16 E6-E7 in ESCC cells, in vitro and in vivo. These findings proved that HPV16 may be a potential high risk factor for the ESCC development. Targeting the CSCs in ESCC through the inhibition of PI3K/ Akt signaling pathway in HPV16 positive tissues is an available therapeutic.

Cell lines and cancer stem-like cell culture
ESCC cell lines Eca109 and TE-1 were purchased from the Chinese Academy of Sciences Cell Bank (Shanghai, China). The authenticity of cancer cell lines was tested by short tandem repeat profiling. All cells were maintained in Dulbecco's Modified Eagle Medium (DMEM, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, USA), 100μg/ mL streptomycin (Sigma-Aldrich, USA) and 100U/mL penicillin (Sigma-Aldrich, USA) at 37°C with 5% CO 2 in a humidified incubator.

Lenti-virus transfection and confirmation of the transfection
The HPV16 E6-E7 lentiviral vector was synthesized by Shanghai Sunbio Medical Biotechnology CO., Ltd. It was constructed by ligating the HPV16 E6-E7 precursor (789bp) into the EcoR I and Xba I sites of the pLVX-EGFP-T2A-Puro lentivector (8934bp). The lentivector was packaged using pCD/NL-BH*DDD Packaging Plasmid mix (Addgene) and transiently cotransfected into 293T cells to generate recombinant virus particles. After 48h of infection, lentivirus in the supernatant was transduced into esophageal squamous carcinoma cells, Eca109 and TE-1, using 5μg/ml of polybrene (Sigma-Aldrich) at the optimal MOI (multiplicity of infection) of each cell. The control lentiviral vector pLVX-EGFP-T2A-Puro which constructed by nonsense sequences stably expressed green fluorescent protein (GFP), while the HPV16 E6-E7 lentiviral vector stably expressed GFP and HPV16 E6-E7. Stable clones were maintained on 5μg/ml of puromycin (Sigma-Aldrich). Fluorescence intensity of GFP was observed by fluorescence microscope to indicate the lentivirus transfection efficiency. The expression of HPV16 E6-E7 was confirmed by Western blotting.

Gene silencing by small interfering RNA (siRNA)
Silencing PI3K and Akt expression using siRNA was performed following the manufacturer's instructions. PI3K siRNA 1178, 619 and 856; Akt siRNA 945, 652 and 567; Negative control siRNA and GAPDH siRNA were all synthesised by GenePharma. All the siRNA sequences were listed in Supplementary Table S2. Briefly, at 60%-70% confluence, ESCC cells were transfected with Select siRNA duplexes, or the Negative Control siRNA, or the GAPDH siRNA at 10nM using Lipofectamine 3000® (Invitrogen, Waltham, MA). Following 6h/37°C siRNA incubation, fresh medium was then added for another 60h/37°C incubation.

In vitro migration and invasion assays
Cells (5*10 4 cells per well) were plated on the top side of polycarbonate transwell filters (for migration assay) or plated on the top side of polycarbonate transwell filter coated with matrigel (for invasion assay) in the top chamber of the 24-Well Cell Invasion Assay (Cell Biolabs, Inc). Cells were suspended in medium without serum, and medium supplemented with 20% serum was used as a chemo-attractant in the bottom chamber. The cells were incubated at 37°C for 24 hours (migration assay) or 48 hours (invasion assay). The nonmigratory or noninvasive cells in the top chambers were removed with cotton swabs. The migrated and invaded cells on the lower membrane surface were fixed in 100% methanol for 3 minutes, airdried, then stained with crystal violet and counted under a microscope.

In vitro spherogenesis assay
Cell suspensions were then plated into ultra-low attachment 96-well plates (Corning) at a density of 2000 cells per well in SFM. To keep volume of media in the well constant, 20 μL SFM was added every 5 days. At the end of the experiment 10μL of trypan blue dye solution was added to each well to detect dead cells. Sphere growth was monitored for 14 days and the number of spheres was counted on day 14. An average of 6 wells was seeded for each radiation dose.

Cell viability assay
The exponentially growing cells were plated onto 96-well plates (Corning) in complete DMEM at a density of 6.0*10 3 cells per well and treated by cisplatin (Sigma-Aldrich, UK; dissolved in 0.9% NaCl,) at different concentration: 2, 5, 10, 20, 50 μmol/L, then incubated at 37°C, 5% CO 2 . After 24, 48 and 72h, the medium was replaced with 100μL DMEM supplemented with 10% FBS and 10μL CCK-8 (Cell counting Kit-8, Beyotime, China) solution and incubated for additional 2h at 37°C. The solution of each sample was aspirated and the absorbance was measured spectrophotometrically at 450 nm. Six parallel replicates of each sample at each time point were prepared during this cell viability assay.

Flow cytometry
Flow cytometry was performed to detect the ratio of p75NTR + cells in ESCC cells. Briefly, Cells were trypsinized and collected by centrifugation. Then cells were washed and suspended in buffer and incubated with p75NTR-PE antibody (BD Bioscience) at 4°C for 2h. The ratio of p75NTR + cells in each samples were analyzed by FASC Calibur MT flow cytometer (BD Bioscience).
Cell cycle was detected by flow cytometry using a cell cycle analysis kit (BD Bioscience). In brief, at least 1*10 6 cells were harvested and washed, then cells were fixed in ice-cold 70% ethanol for at least 2h at 4°Cwashed the cells again and stained with a solution containing 50μg/ml PI and 50 RNase at room temperature for 30 min. Cell cycle was analyzed with a FACS Calibur MT flow cytometer (BD Bioscience).
Cell apoptosis was detected by flow cytometry using an Annexin V-APC apoptosis detection kit (BD Bioscience). In brief, cells were harvested and washed, then cells were suspended in binding buffer and incubated with Annexin V-APC and propidium iodide (PI) at 4°C, the cells were double stained with Annexin V-APC and PI according to the manufacturer's instructions. Early apoptosis and the late apoptosis were determined by Annexin V+/PI-staining and Annexin V+/PI+ staining, respectively. The percentage of apoptosis cells in each sample was examined using the FASC Calibur MT flow cytometer (BD Bioscience).

Colony formation assay
The cells in the monolayer culture were irradiated with graded doses of X-ray by using a linear accelerator at the dose of 0, 2, 4, 6, 8 Gy. The cells that remained attached were removed from the culture flask by exposing to 0.05% trysin solution. Afterward, these cells were replated in triplicate onto 6-well plates in complete DMEM to determine the colony-forming ability. After 14d of incubation, the plates were stained with 0.5% crystal violet in absolute methanol, and the colonies consisting of at least 50 cells were recorded. The survival fraction (SF) was calculated by dividing the number of colonies at a certain radiation dose by the product of the number of cells seeded at this dose and the corresponding colonyforming rate. Lines are fitted using the Single hit multitarget model (SF=1-(1-e -D/D0 ) n ) with GraphPad Prism version 5.01.

Immunofluorescence
Cells seeded on coverslips were fixed with 4% paraformaldehyde for 30 min after three times PBS washing and permeabilized in PBS with 0.1% TritonX-100. Then cells were blocked with 5% BSA at room temperature for 1 h. Primary antibody incubation was performed overnight at 4°C. Fluorescent secondary antibody was incubated for 1h at room temperature after three times PBS washing and DAPI (#4083, CST, USA) was applied to counterstain nucleus. All the primary antibodies were the same with western blotting. Images were examined and captured under the microscopy (Leica, Heidelberg, Germany).

Xenograft tumor model and treatment
The animal studies were approved by the Institutional Animal Ethics Committee of The First affiliated hospital of Xi'an Jiao Tong University and experiments were performed in accordance with the Animal Ethics guidelines of The First affiliated hospital of Xi'an Jiao Tong University. Female BALB/c nude mice (four weeks old with body weight approximately 15g) were purchased from the Experimental Animal Center of School of Medicine (Xi'an Jiaotong University) and housed in it. After a week's acclimation, the tumor model was established by subcutaneously injecting 2*10 6 cells per 200μL, Eca109-control and Eca109-psb cells, in the back of each mouse. The mice were randomly distributed into four groups (N=6):Control, Radiation, LY294002 and LY294002 +Radiation group. The mice in the LY294002 group were intraperitoneally injected with LY294002 at a dose of 5 mg/kg body weight twice a week for five weeks. This dose was administered based on the result of our preliminary experiment, which is 10% of the dose with the median effective concentration of LY294002 when administered alone [29]. The mice treated by radiation alone were immobilized in a customized harness. The back was exposed and the remaining parts of the body were shielded with a thick piece of lead. The tumor area was irradiated by an X-ray linear accelerator with a dose of 2Gy twice a week for five weeks. The mice in the group treated with the combined LY294002 and radiation were irradiated after the injection of LY294002 every time. The control group received a 0.1% DMSO/sodium chloride solution without drugs. Above all of these are operated after tumor cells injection for three weeks when the tumor volume in each group could be macroscopied and measured. The cumulative radiation exposure dose in the radiation group and combined group are 20 Gy. The mice were monitored daily for body weight, behavior and water/food consumption. Tumor growth was monitored by measuring the tumor size in two orthogonal dimensions with Vernier calipers every 4 days; the tumor volume was calculated according to the following formula: length* width 2 *0.5. The mice were anesthetized and sacrificed with eight weeks feeding time. The tumors were surgically excised, weighed, fixed in 10% formalin, embedded in paraffin, and cut into thick slices (4μm) for hematoxylin and eosin staining (H&E) to confirm tumor formation. Inhibition rate was defined as: 1-tumor weight or volume of the treatment group / tumor weight or volume of the control group. The solid tumors were then surgically removed and fixed in 10% formalin neutral buffer solution.

Immunohistochemistry (IHC)
Immunohistochemical staining was performed according to a standard protocol using a Hypoxyprobe™-1 kit. All the primary antibodies were the same with western blotting and selected at the concentration of 1:100. PBS, instead of primary antibody, was used as a negative control. All of the animal procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Experimental protocols were approved by the Animal Care and Use Regulation of Xi'an Jiaotong University. Images were captured using a microscopy (Leica, Heidelberg, Germany) and five random fields of each were chosen from each specimen for quantitative analysis.

X-ray radiation treatment
Cells were irradiated with 0, 2, 4, 6, 8 Gy of X-ray radiation using the X-ray linear accelerator (Clinac 2100EX, Varian Medical Systems). The treated cells were then used for the experiments as described above.

Statistical analysis
All results are present as mean ± SD from at least three independent experiments. Statistical analysis was performed using Student's t-test and followed by ANOVA analysis in SPSS 16.0. Differences between groups were considered to be statistically significant at P<0.05 (*). Graphs were made in GraphPad Prism 5.01 (GraphPad Software, San Diego, California, USA).