Transcription factor ATF3 mediates the radioresistance of breast cancer

Abstract This study was designed to research the influence of activating transcription factor 3 (ATF3) on the radioresistance of breast cancer. ATF3 expression was measured by qRT‐PCR and immunohistochemistry. Cancerous cell lines were cultured in vitro, and the expression of ATF3 was gauged by both qRT‐PCR and Western blot before and after the radiation therapy. Cellular cycle and apoptosis were analysed by flow cytometry. Changes in the expression of corresponding proteins in the downstream pathways were identified by Western blot. Tumour xenograft was used to evaluate the effect of ATF3 in vivo. ATF3 was observed stronger in breast cancer tissues and cells. After radiation therapy, the expression of ATF3 in breast cancer cells was up‐regulated. Silencing ATF3 could promote G2/M phase block, facilitate cell apoptosis and decrease clonogenic survival rate. The overexpression of ATF3 could curb G2/M‐phase block, cell apoptosis and increase clonogenic survival rate. Overexpression ATF3 could increase radioresistance by up‐regulating the level of phosphorylation of Akt in the PI3K/Akt signalling pathway. Radioresistance of breast cancer cells could be alleviated by inhibiting the PI3K/Akt signalling pathway. ATF3 could also promote radioresistance in vivo. ATF3 gene was able to promote radioresistance of breast cancer cells.

Akt. 10 The PI3K/Akt signal pathway mediates cell survival by promoting aerobic glycolysis. 11 Most of the cancer cells produce abundant lactate to supply energy, but it is inefficient to generate ATP. This phenomenon is regarded as aerobic glycolysis. 12 Akt can mediate various steps of glycolysis by post-transcriptional mechanisms which contain promoting hexokinase activity and up-regulating expression of glucose transporter Glut1. 13 Recent report showed that increased expressions of glucose transporter Glut1 and lactate were examined in acquired radioresistant cells. 14 Shimura et al. unearthed that inhibition of glycolysis could control required tumour cell radioresistance. In this study, we would like to investigate the effect of ATF3 in breast cell radioresistance by controlling the production of pAkt and ATF3.
Radiosensitivity of breast cancer cells may be altered by the reversible PI3K inhibitor LY294002, which inhibits certain mammalian PI3Ks by covalent or non-covalent modification of a critical lysine residue in their phosphotransferase domains. 15 On account of the presence of the COOH-terminal sequence homology among the PI3K, we can draw a conclusion that the PI3K/Akt signalling pathway may also be sensitive to the inhibition of LY294002. 16 In a recent study of non-small cell lung cancer, it was found out that high levels of PI3K/Akt activity increased the radioresistance of these cells and suppressed the radiation-induced cell apoptosis; but once the cells were treated with LY294002, sensitivity to radiation therapy was restored. 17 The results of these studies all suggested that modulation of PI3K/Akt activity in cancer cells may alter the sensitivity of the cells to conventional radiation therapy.
In accordance with all the reports above, we have decided to disclose the relationship between the radioresistance of breast cancer cells and the expression of ATF3 in the PI3K/Akt signalling pathway.

| Patients and tissue specimens
Sixty specimens of breast cancer patients who had gone through radiotherapy and been confirmed pathologically were collected from Shengjing Hospital Affiliated China Medical University (from June 2015 to May 2016). All the breast cancer tissues and paracancerous tissues of patients were placed immediately in liquid nitrogen and kept for long-time preservation in À80°C to be measured. All participates involved in this study have signed the consent informs. Clinicopathological features of breast cancer patients were listed in Table S1.
The data of T47D cells before or after irradiation were analysed by R project (https://www.r-project.org/) and log 2 (Fold Change) >2 and P < .05 were identified as our filtration criteria. Breast cancer cell lines T47D, ZR751, MCF7, HBL10, SUM159 and   normal human mammary epithelial cell line MCF10A were  Transfection efficiency was observed at the 24th hour and the 48th hour after cell culture. Plasmid LZRS-IRES-zeo-ATF3 was built by connecting carrier LZRS-IRES-zeo and ATF3 cDNA (ATF3) and blank control group LZRS-IRES-zeo (NC) was set up.

| Immunohistochemistry
The paraffin-embedded pathological sections were waxed, dehydrated by gradient alcohol. The endogenous peroxidase was eliminated and the antigen was repaired. 100 lL antigen of ATF3 (rabbit anti-human, 1:800, Abcam, USA) was injected to be incubated overnight. HRP-labelled goat anti-rabbit IgG (ZSGB-BIO, Beijing, China) was incubated, taken out 40 minutes later and was washed with phosphate buffered saline (PBS) thrice. DAB chromogenic agent was used for colour development for about 20 seconds. After haematoxylin restraining for 1 minute, it was dehydrated with gradient alcohol, dried and mounted. The expression level of ATF3 was evaluated by semi-quantitative analysis.

| qRT-PCR
RNA primers were designed and synthesized (Table S3 for

| Western blot
In order to extract the total protein, RIPA lysis was employed to lyse the cells. The total protein concentration of each group was detected by bicinchoninic acid (BCA) method. SDS-PAGE gel was prepared for sample electrophoresis. The polyvinylidene fluoride film (PVDF) was removed and placed in a sealing fluid containing 3% of albumin from bovine serum (BSA). The primary antibody (ATF3 rabbit anti-human, 1:5000, ab207434 Abcam; b-catenin, 1:5000, ab8227, Abcam; pAkt, 1:5000, ab38449 Abcam; Akt, 1:5000, ab81283, Abcam; caspase3, 1.5 lg/mL, ab13585, Abcam) was diluted by the Tris-buffered saline solution with tween (TBST), then blocked overnight in a shaker at 4°C. PVDF film was taken out and washed three times with TBST.
Goat anti-rabbit HRP secondary antibodies (ab6741, Abcam, USA) were diluted 1:5000. Both were blocked inconstant temperature shaker for 4 hours and washed three times with TBST. After darkroom exposure, development, cleaning and fixing, stripe accumulative optical density was analysed by ImageJ software (Version 1.48u, Bethesda, USA) semi-quantitative analysis and relative expressions of target proteins were calculated, respectively.

| Clonogenic survival analysis
The cells in the logarithmic growth phase were extracted, digested into individual cells with trypsin solution, and then added RPMI1640 medium. 1000 cells were spread evenly with the appropriate concentration of the cell suspension in a petri dish at 37°C, 5% CO 2 .
When the cells were visible to the human eye, the culture was terminated. The cells were radiated with Caesium-137 Mark-I irradiator (JL Shepherd, CA, USA) of 2 Gy, 4 Gy, 6 Gy or 8 Gy for 24 hours.
Media was removed and replaced 24 hours after irradiation. The cells were allowed to grow into colonies over a period of 10-14 days or until the control plates grew visible colonies. Cells were then washed with PBS, followed by fixed with 10% methanol and 10% acetic acid for 10 minutes. The cells were dyed for 10 minutes with 0.4% crystal violet solution and dried in the air after the removal of the stain. Clones were calculated using ColCounte colony counter (Oxford Optronix Ltd, Abingdon, UK), the sensitivity was confirmed when the cell number exceeded 50. The following formula was used to calculate the Plating Efficiency (PE): PE = clone number/inoculation number. The comparison was standardized to the negative control group. The survival fraction (SF) was calculated using the following formula: SF = clone number/inoculation number 9 PE.

| Caspase 3 activity detection
Pentose nucleic acid (PNA) was diluted with the standard diluent of protein was semi-quantitatively measured by Western blot.

| Cell cycle analysis
The test is implemented following the instructions of Cell Cycle and Apoptosis Analysis Kit (Beyotime, Jiangsu, China). The radiation therapy was carried out 36 hours after the cell transfection. The cells were digested by trypsin, collected by centrifuge (1000 rpm) for 5 minutes and the culture medium was discarded. Then the cells were rinsed with cold PBS and centrifuged (1000 rpm) for 5 minutes. 1 mL of cells was added in the cooled 70% alcohol and placed at 4°C overnight. After that, the cells were washed with PBS twice and added with 100 lL of RnaseA and incubated at 37°C water for 30 minutes. 50 lL of PI was then added to the cell mixture and the solution was kept from light for 1 hour. Finally, the cell cycle was detected by flow cytometry method (FCM).

| Tumour xenograft
All animal experiments were performed according to the National Institutes of Health guide for the care and use of Laboratory animals.

| Statistical analysis
GraphPad Prism 6.0 software (Version 6, CA, USA) was used for statistical analysis. The differences between the two samples were measured by the T-test, and the differences among multiple samples were analysed using the one-way analysis of variance. When P was <.05, the difference was considered statistically significant.

| ATF3 was highly expressed in breast cancer after radiotherapy
According to the microarray analysis, 44 genes were revealed significant up-regulation in radiated breast cancer cell samples under the condition that the false positive rate was less than 0.01 and fold change >2 ( Figure 1A,B). Finally, ATF3 which was strongly expressed in multiple breast cancer cell samples were chosen for next experiments.

| ATF3 was up-regulated both in breast cancer tissues and cells
To observe the expression of ATF3 in breast cancer tissues, 60 clinical samples collected from Shengjing Hospital Affiliated China Medical University were examined by qRT-PCR. It was found out that the expression level of ATF3 in breast cancer tissues which had received radiotherapy was strongly higher than that in adjacent tissues (P < .01, Figure 2A, B). After the analysis of ATF3 expression level in different cell lines, it was found out that the mRNA expression level of ATF3 was also remarkably higher in breast cancer cell lines T47D, ZR751, MCF7, HBL100 and SUM159 than that in normal human mammary epithelial cell line MCF10A (all P < 0.05, Figure 2C). The protein expression of ATF3 was assessed by Western blot and the protein expression of ATF3 was also markedly up-regulated in breast cancer cell lines (all P < .05, Figure 2D).

| ATF3 expression increased in breast cancer cells after the radiation therapy
To simulate the clinical situation of radiation resistance, T47D, ZR751, MCF7, HBL100 and MCF7 cell lines were treated with radio ionizing radiation. The result showed that both mRNA and protein expressions of ATF3 in all cell lines were increased (P < 0.05, Figure 3A,B). As the most outstanding differences observed in the MCF7 and SUM159 cell lines, radioresistant MCF7 was used to carry out the knockout experiments whereas radiosensitive cell line SUM159 was used to carry out the overexpression experiments in the subsequent experiments.

SUM159 cells
After the MCF7 was transfected with siATF3, qRT-PCR and Western blot were utilized to measure the expression of ATF3.
Depending on the detection result, compared with the negative control group siNC, the expression level of ATF3 in the siATF3 group was significantly reduced (all P < .05, Figure S1A-S1B), which proved that the interference was effective. After the SUM159 cell line was transfected with the LZRS-IRES-zeo-ATF3, the ATF3 mRNA and protein expressions of the LZRS-IRES-zeo-ATF3 both significantly rose (P < .05, Figure S1C-S1D) in contrast to the negative control group LZRS-IRES-zeo, which proved that the overexpression was effective.

| ATF3 could promote the radioresistance of breast cancer cell
After the MCF7 cell line was transfected with siATF3, the survival rate of the cell clone in siATF3 group was significantly cut down (all P < .05, Figure S2A) compared with the negative control group.
After the SUM159 cell was transfected with the expression vector LZRS-IRES-zeo-ATF3, the survival rate of the LZRS-IRES-zeo-ATF3 cell clone was notably improved compared with the negative control (all P < .05, Figure S2B). It could be possible to conclude that the low expression of ATF3 could alleviate the radioresistance of breast cancer cells.

| Overexpression ATF3 could lessen apoptosis rate induced by radiation
The results of flow cytometry showed that, compared with the negative control group, the frequency of apoptosis rate significantly rose after the MCF7 cell line was transfected with siATF3 (P < .05, Figure 4A). And compared with the negative control group, the apoptosis rate was significantly reduced after SUM159 cell line was transfected with LZRS-IRES-zeo-ATF3 (all P < .05, Figure 4B). As shown in Figure 4C A   ACTA2  CSTA  TNFRSF10C  SLC1A1  HIST1H4H  S100A7  FAS  FGF18  MIR22  MIR612  ATF3  GADD45A  UGT2B15  IFIT1  GDF15  BTG2  OAS1  TIMP3  HIST1H2BC  NCOA2  CENPI  TRMT2B E2F8 comparison with the negative control group (P < .05, Figure 5A).
And compared with the negative control group, the G2/M phase cell rate of the LZRS-IRES-zeo-ATF3 group was significantly lessened whereas cells in the S phase notably increased (P < .05, Figure 5B).
Therefore, it could indicate that the low expression of ATF3 contributed to G2/M phase block in the cell cycle.

| ATF3 facilitated the radioresistance of breast cancer cells through the PI3K/Akt signalling pathway
Results of Western blot demonstrated that after the radiation therapy, the protein expression level of radioresistance-related key protein pAkt in siNC group and siATF3 group of MCF7 cell line both increased significantly. But pAkt protein expression was suppressed after silencing the ATF3 expression (P < .05, Figure 6A). On the contrary, the overexpression of ATF3 led to an increased level of pAkt protein and the increase was even higher after the radiation therapy (P < .05, Figure 6B). The results suggested that ATF3 could augment the radioresistance of breast cancer cells by affecting the related proteins expressions in the PI3K/Akt signalling pathway. As Akt is one of the PI3K pathway downstream regulating elements, LY294002 was utilized to interfere with activation of the pathway in the SUM159 cell line. The effects of PI3K pathway inhibitor LY294002 on radiation-induced Akt phosphorylation level in cell lines with ATF3 overexpression were then explored using the clonogenic survival assay and Western blot. The results exhibited that the radiation-induced phosphorylation was significantly inhibited by LY294002 in the SUM159 cell line (all P < .05, Figure 6C). It demonstrates that the effects of ATF3 expression and of radiation therapy on Akt phosphorylation depend on the activity of upstream PI3K pathway. The clonogenic survival assay showed that the survival rate of cells was significantly reduced after adding the inhibitor LY294002 (all P < .05, Figure 6D) compared with the ATF3 overexpression group, suggesting that the breast cancer radioresistance established by the ATF3 overexpression can be alleviated by inhibiting the PI3K/Akt signalling pathway.

| ATF3 enhanced the radioresistance of breast cancer in vivo
To further explore the effect of ATF3 in vivo, tumour xenograft experiments were performed. After 30 days, mice were killed and tumour tissues were excised ( Figure 7A,E). The results showed that overexpression of ATF3 could promote the growth of tumours (P < .05, Figure 7B,C), while silence of ATF3 could inhibit the development of tumours (P < .05, Figure 7F,G). In addition, ATF3 attenuated the effect of irradiation in vivo (P < .05, Figure 7B,C), while silence of ATF3 enhanced the sensitivity of irradiation of breast cancer (P < .05, Figure 7F,G). On the other hand, qRT-PCR was carried out to detect the level of ATF3 in tumour. It revealed that the level of ATF3 was up-regulated followed irradiation compared with control group (P < .05, Figure 7D,H). All these indicated that ATF3 was able to enhance radiotherapy tolerance in vivo as well.

| DISCUSSION
In this present study, we detected that breast cancer radioresistance was enhanced when the ATF3 expression was up-regulated, resulting in less breast cancer cell apoptosis, decreased G2/M phase block and more activated PI3K/Akt signalling pathway. On the contrary, if ATF3 expression in breast cancer cells was silenced by the siATF3 transfection or was suppressed by the signalling pathway inhibitor LY204002, breast cancer radioresistance could be relieved to a great extent and the radiation therapy efficiency can be also improved.
Overwhelming evidence showed that ATF3 was highly expressed and was closely related to cell migration in breast cancer. 6 F I G U R E 4 Apoptosis rate of breast cells significantly rose after treating with irradiation. A, The apoptosis rate was remarkably improved after the radiation therapy and was evidently higher than that of the negative control group; B, The apoptosis rate was remarkably increased after the radiation therapy and was evidently lower than that of the negative control group; C, Apoptosis-related protein cleaved-caspase3 was upregulation after the radiation therapy in MCF7 cells transfected with siATF3; D, The activity of caspase3 was significantly improved after the radiation therapy and was evidently higher than that of the negative control group in MCF7 cells; E, Apoptosis-related protein cleaved-caspase3 was also up-regulation after the radiation therapy in SUM159 cells transfected with LZRS-IRES-zeo-ATF3; F, The activity of caspase3 was significantly improved after the radiation therapy and was evidently lower than that of the negative control group. siNC: cells transfected with independent nucleic sequence with the same base number. siATF3: cells transfected with interference nucleotide sequence. NC: cells transfected with empty plasmid LZRS-IRES-zeo. ATF3: cells transfected with plasmid LZRS-IRES-zeo-ATF3. *P < .05, compared with NC or siNC group Nakai could resist cancer by up-regulating ATF3 repression in colorectal cancer. 19 In our study, differentially expressed genes were analysed by microarray. As it was found out, ATF3 expression was significantly higher in breast cancer tissues and cells after radiation therapy. Therefore, the following experiment was focused on the relationship between ATF3 expression and radioresistance of breast cancer. Previous studies concentrating on the relationship between differentially expressed genes and breast cancer cell radioresistance, such as one studying the expression of ABT-737, Bcl-2 and Bcl-xL in breast cancer, 20     ATF3 could be expressed in these signalling pathways to influence the effects of breast cancer radioresistance. All of the limitations would like to be involved in the next research.
In conclusion, throughout the research, ATF3 expression could exert significant influences on various spheres of the breast cancer radioresistance that is the efficiency of radiation therapy. When ATF3 expressed highly in breast cancer, the cancer radioresistance was enhanced and radiation therapy could not well function in causing breast cancer cell apoptosis or in remaining the G2/M phase cell cycle block. To increase the radiotherapy efficiency, methods such as silencing ATF3 expression and inhibiting the activation of the PI3K/ Akt signalling pathway were recommended.

CONF LICT OF I NTEREST
The authors declare that they have no conflict of interest.

ETHICAL APPROVAL
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

INF ORMED CONSENT
Informed consent was obtained from all individual participants included in the study.