miR‐126‐5p enhances radiosensitivity of lung adenocarcinoma cells by inhibiting EZH2 via the KLF2/BIRC axis

Abstract Radiotherapy is a common method for the treatment of lung adenocarcinoma, but it often fails due to the relative non‐susceptibility of lung adenocarcinoma cells to radiation. We aimed to discuss the related mechanisms by which miR‐126‐5p might mediate radiosensitivity of lung adenocarcinoma cells. The binding affinity between miR‐126‐5p and EZH2 and between KLF2 and BIRC5 was identified using multiple assays. A549 and H1650 cells treated with X‐ray were transfected with miR‐126‐5p mimic/inhibitor, oe‐EZH2, or si‐KLF2 to detect cell biological functions and radiosensitivity. Finally, lung adenocarcinoma nude mouse models were established. miR‐126‐5p and KLF2 were poorly expressed, while EZH2 and BIRC5 were upregulated in lung adenocarcinoma tissues and cells. miR‐126‐5p targeted EZH2 to promote the KLF2 expression so as to inhibit BIRC5 activation. Both in vitro and in vivo experiments verified that elevated miR‐126‐5p inhibited cell migration and promoted apoptosis to enhance the sensitivity of lung adenocarcinoma cells to radiotherapy via the EZH2/KLF2/BIRC5 axis. Collectively, miR‐126‐5p downregulated EZH2 to facilitate the sensitivity of lung adenocarcinoma cells to radiotherapy via KLF2/BIRC5.

for lung adenocarcinoma. 9 However, little information concerning the impact of miR-126-5p on the radiosensitivity of lung adenocarcinoma cells. Moreover, the bioinformatics analysis in the current study predicted that miR-126-5p might target enhancer of zeste homolog 2 (EZH2). EZH2 is capable of affecting the occurrence and progression of human cancers by modulation of cell biological processes. 10 Existing literature has reported that EZH2 is dysregulated in lung adenocarcinoma. 11 The roles of EZH2 in prostate cancer cell resistance to radiotherapy have been explored. 12 However, the mechanism by which miR-126-5p affects radiosensitivity of lung adenocarcinoma cells by targeting EZH2 is still poorly understood.

| Ethic statement
This study was performed with the approval of the Ethics Committee

| Bioinformatics analysis
Lung adenocarcinoma-related miRNA expression dataset GSE13 5918 was downloaded through the Gene Expression Omnibus database. The dataset contains five adjacent normal tissues (normal) and five lung adenocarcinoma tissues. The R language 'limma' package was used for differential analysis of the gene expression, with |logFoldChange| >2, and p < 0.01 as the screening criteria. Gene Expression Profiling Interactive Analysis 2 (GEPIA2) tool was used for differential analysis of the gene expression of lung adenocarcinoma tissues in the The Cancer Genome Atlas database compared to normal samples. The bioinformatics target gene prediction tool StarBase was used to predict the target genes of miR-126-5p, and the lung adenocarcinoma-related genes were searched through the GeneCards database. The Venn tool was employed to identify the overlap between target genes and disease-related genes. In order to further screen target genes, the Coexpedia website was utilised to analyse the co-expression relationship network between candidate target genes, and to determine candidate genes based on the co-expression relationship score. The binding site of miR-126-5p and target gene was obtained through StarBase website. Next, the downstream regulatory factors of the target gene were predicted.
The interacting proteins were searched in the GeneCards database, and the jvenn tool was used to find the overlap between the interacting proteins and the significantly downregulated differential genes in lung adenocarcinoma. Similarly, the STRING website was employed to analyse the interaction network among genes, and the Cytoscape 3.5.1 software was employed to visualise the interaction network. The GEPIA2 tool was used to further analyse the correlation among genes, and the JASPAR website was utilised to predict the possible binding sites of transcription factors in gene promoter regions.  Table   S1. With the telephone follow-up and outpatient information inquiry, the physical conditions of postoperative patients at different periods were reviewed. This follow-up started from the diagnosis of lung adenocarcinoma until death, and the follow-up period was 2-30 months. The Kaplan-Meier method was employed to analyse the survival rate of the patient.

| In Situ Hybridisation (ISH)
ISH assay was implemented according to the manufacturer instructions of miRNA ISH miR-126-5p optimisation Kit (Exiqon, LNA™). 13 The samples were scanned and photographed with the help of digital pathological scanning equipment (P250 FLASH, 3DHISTECH) and analysed by Image Pro Plus (version 5, Media Cybernetics). The positive rate was calculated as the integrated optical density/sum of product.

| Radiotherapy treatment
The cells were irradiated with X-Ray with the help of the Astrophysics Torrex X-ray Inspection System (Model 120D, Scanray Corporation) at 115 kVp and 5 mA. The cells (1 × 10 5 cells/ml) in the logarithmic growth phase were inoculated in DMEM containing 10% FBS and 100 U/ml penicillin-streptomycin in an incubator at 37°C with 5% CO 2 . Irradiation was performed 24 h after transfection if needed.
Cells were irradiated with X-Ray of 10 Gy (dose rate: 0.341 Gy/min; radiation area: 10 × 10 cm 2 ), and further cultured for 24 h for subsequent experiments.
The above-mentioned plasmid mixed with lipo2000 (11668027, Thermo Fisher Scientific) were added to a 12-well plate. The cells were cultured for 6 h and the medium was renewed for another 48-h culturing.

| Transwell assay
Transwell migration assay was implemented based on previous method 14 with basement membrane matrix (50 mg/L, Sigma-Aldrich). The images were photographed under an inverted microscope (CKX41SF, Olympus).

| RT-qPCR
For the detection of genes EZH2, KLF2 and BIRC5, RNAiso plus (1 ml, Takara) was used for total RNA extraction. Nanodrop 2000 (Thermo Fisher) was utilised to detect the concentration of mRNA.
RT of mRNA and cDNA synthesis was conducted using PrimeScript acted as an internal reference, and TaqMan miRNA assays (Applied Biosystems) were used for miR-126-5p quantification. All RT-qPCR analysis was implemented using Applied Biosystems 7900HT Fast Real-Time PCR System (Applied Biosystems). The 2 −ΔΔCt method was used to quantify relative expression levels of target genes (Table S2).

| Dual-luciferase reporter gene assay
The bioinformatics target gene prediction tool StarBase (http://starb ase.sysu.edu.cn/) was used to predict the target gene of miR-126-5p, among which EZH2 was highly expressed in lung adenocarcinoma.

| Tumour formation in vivo
A total of 40 BABL/c male nude mice (ageing 4-6 weeks, weighing 20.1 ± 3.6 g, Shanghai SLAC Laboratory Animal Co., Ltd.) were used in this experiment. Nude mice were kept at a humidity of 45%-50% with a temperature of 25-27°C, free access to food and water.
A549, A549/lv-NC and A549/lv-miR-126-5p cells in logarithmic growth phase (the titre of the lentiviral vector was 2 × 10 9 TU/ml, Genechem) were resuspended in 0.2 ml PBS. The cell concentration was adjusted to 1 × 10 7 cells/ml. Cells were inoculated subcutaneously into the right posterior abdomen of nude mice. The tumour volume was measured every 3 days, and when the tumour volume reached 50 mm 3 , radiotherapy was conducted. Nude mice were randomly divided into four groups (n = 10): Control group (subcutaneous injection of A549 cells without radiotherapy), Radiotherapy group (subcutaneous injection of A549 cells and then underwent 20 Gy X-ray radiation treatment), Radiotherapy +lv-NC group (subcutaneous injection of A549/lv-NC cells and then underwent 20 Gy X-ray radiation treatment), Radiotherapy +lv-miR-126-5p group (subcutaneous injection of A549/lv-miR-126-5p cells and then underwent 20 Gy X-ray radiation), radiotherapy was performed every 5 days and repeated three times. During this period, the volume and weight of the transplanted tumour were observed. On the 21st day after tumour cell inoculation, nude mice were killed, and tumour specimens were collected and stored at −80°C for RT-qPCR and Western blot analysis.

| Statistical analysis
Data analysis was performed using the GraphPad Prism 8.0. The measurement data are presented as mean ± standard deviation and analysed by unpaired t-test or one-way analysis of variance (ANOVA) with post hoc Tukey's test was used. The Kaplan-Meier method was utilised to calculate the survival rate of patients. Values of p < .05 were considered significant.

| miR-126-5p is poorly expressed in lung adenocarcinoma tissues and cells, and is associated with poor prognosis of patients with lung adenocarcinoma
The differential analysis of the miRNA expression dataset GSE13 5918 related to lung adenocarcinoma was performed, revealing that differential expression multiple of hsa-miR-145-5p, hsa-miR-30c-5p, hsa-miR-431-5p and hsa-miR-126-5p were the highest and significantly downregulated ( Figure 1A). It has been reported that miR-126-5p was under-expressed in NSCLC, 15 which was consistent with the results of dataset GSE13 5918 ( Figure 1B), but there are few studies on its effect on lung adenocarcinoma and its possible mechanism. Therefore, miR-126-5p was selected for the experiments.
The follow-up period of enrolled patients was 2-30 months and the survival rate was analysed by Kaplan-Meier method. We found that the overall survival of patients with lung adenocarcinoma with high expression of miR-126-5p was obviously longer than those with low expression of miR-126-5p ( Figure 1C). Besides, lower miR-126-5p expression was seen in lung adenocarcinoma tissues ( Figure 1D); and miR-126-5p expression in patients with lung adenocarcinoma at the advanced stage was lower than those at early stage ( Figure 1E). ISH assay confirmed low miR-126-5p expression in the lung adenocarcinoma tissues ( Figure 1F).
It was further verified in five lung adenocarcinoma cell lines and human embryonic lung fibroblast cell lines by RT-qPCR that miR-126-5p expression in lung adenocarcinoma cells was obviously lower than that in human embryonic lung fibroblasts. Among the five lung adenocarcinoma cell lines, the miR-126-5p expression was relatively lower in the A549 cell line, and relatively higher in the H1650 cell line, and these two cell lines were selected for subsequent experiments ( Figure 1G).

| miR-126-5p targets and inhibits EZH2
In order to explore the downstream target genes regulated by miR-126-5p in lung adenocarcinoma, the differentially expressed genes in lung adenocarcinoma were obtained using the GEPIA2 tool ( Figure 2E). A total of 1188 target genes of miR-126-5p were predicted by the StarBase and the lung adenocarcinoma-related genes were obtained from the GeneCards database, and then 27 candidate target genes were collected after intersection ( Figure 2F). The Coexpedia was used to obtain the co-expression relationship network diagram of the candidate target genes ( Figure 2G), in which EZH2 and CHEK1 had the highest scores on the website (Score >20). Existing literature indicated that EZH2 was upregulated in lung adenocarcinoma cells. 11 Moreover, EZH2 was significantly highly expressed in lung adenocarcinomas through the GEPIA2 tool ( Figure 2H). Thus, EZH2 was selected for the experiments.

| EZH2 inhibits KLF2 expression by inducing KLF2 H3K27me3 modification
Next, 2462 interacting proteins of EZH2 were found in the GeneCards database, which were intersected with the significantly lowly-expressed DEGs in lung adenocarcinoma, then 53 genes were collected ( Figure 3A). In the interaction network diagram of 53 genes, 10 genes were at the core position (Degree ≥5; Figure 3B), among which KLF2 was obviously downregulated in lung adenocarcinoma tissues ( Figure 3C), and there was a significant correlation between EZH2 and KLF2 in lung adenocarcinoma ( Figure 3D). A prior study has suggested that EZH2 can be enriched in the KLF2 promoter region to inhibit KLF2 transcription, leading to the promotion the proliferation of NSCLC cells. 16 Therefore, KLF2 was selected for the following experiments. RT-qPCR and immunohistochemical staining ( Figure 3E-G) exhibited that KLF2 expression decreased in lung adenocarcinoma tissues and cells ( Figure 3F).
ChIP-PCR ( Figure 3H) showed that the EZH2 and H3K27me3 was enriched in the KLF2 gene promoter region in A549 and H1650 cells transduced with oe-EZH2. Next, A549 and H1650 cells were transduced with oe-EZH2, and we observed that the protein levels of EZH2 and H3K27me3 increased and KLF2 protein level decreased in A549 and H1650 cells transduced with oe-EZH2 ( Figure 3I,J and Figure S1B). It could be concluded that EZH2 could be enriched in the KLF2 promoter region to trigger H3K27me3 modification, thereby inhibiting the KLF2 expression.

| Transcription factor KLF2 inhibits BIRC5 transcription
It was found that KLF2 and BIRC5 were negatively correlated in lung adenocarcinoma through GEPIA2 ( Figure 4A), and BIRC5 was significantly highly expressed in lung adenocarcinoma ( Figure 4B).
It is known that BIRC5, a proto-oncogene encoding survivin protein, can regulate the apoptosis and senescence of NSCLC cells, and its expression is significantly related to the survival rate of patients with lung adenocarcinoma. 17 Therefore, we speculated whether KLF2 played a role in lung adenocarcinoma by regulating BIRC5 expression.
RT-qPCR ( Figure 4C,E) showed that BIRC5 expression increased in lung adenocarcinoma tissues and cells, and the same result was found by the immunohistochemical staining ( Figure 4D). The JASPAR website predicted the possible binding site of KLF2 in the BIRC5 promoter ( Figure 4F), and ChIP-PCR ( Figure 4G) displayed that the enrichment F I G U R E 2 miR-126-5p expression affects the radiosensitivity of lung adenocarcinoma cells. A549 and H1650 cells were treated with 10 Gy of X-ray and transduced with miR-126-5p mimic or inhibitor. Next, flow cytometry ( Figure 5D and Figure S2C), immunofluorescence staining ( Figure 5E and Figure S2D), and Transwell assay ( Figure 5F and Figure S2E) showed that elevated miR-126-5p promoted A549 and H1650 cell apoptosis, increased γ-H2AX focus number, and inhibited cell migration, while simultaneous elevated miR-126-5p and silencing KLF2 exerted the opposite effects on the A549 and H1650 cells. These findings proved that that miR-126-5p downregulated the EZH2 expression and increased the KLF2 expression to inhibit the BIRC5 activation, ultimately enhancing radiosensitivity.

| miR-126-5p increases radiosensitivity via the EZH2/KLF2/BIRC5 axis in vivo
Finally, the above mechanism was analysed in vivo. It was found that the volume ( Figure 6A) and weight ( Figure 6B,C) of transplanted tumour in nude mice decreased in nude mice treated with X-ray alone or combined with lv-miR-126-5p. Additional lv-BIRC5 increased the volume and weight of tumours in presence of lv-miR-126-5p after radiotherapy. An enhancement of miR-126-5p and KLF2 while reductions in EZH2 and BIRC5 were seen in tumour tissues of nude mice treated with X-ray alone or combined with lv-miR-126-5p.

| DISCUSS ION
Although great achievement has made in treatment strategies of lung adenocarcinoma, the progressive tolerance of lung adenocarcinoma cells in radiotherapy often leads to local recurrence and poor prognosis. 18  We identified low miR-126-5p expression in lung adenocarcinoma tissues and cells, which was associated with poor prognosis.
It is reported that many miRNAs, emerged as promising biomarkers for cancer treatment, are involved in the progression of multiple cancers, and have become promising biomarkers. 5 miRNAs are reported to be dysregulated in lung adenocarcinoma. 19 Meanwhile, similar result was seen in previous study that miR-126-5p was downregulated in lung adenocarcinoma. 8 Moreover, elevated miR-126-5p enhanced the radiosensitivity of lung adenocarcinoma cells. Growing evidence has demonstrated the important role of miRNAs in regulation of the radiosensitivity of cancer cells. 9 Partly in line with our study, a recent study also demonstrated the promotive effect of miR-126-5p in the sensitivity of clinical DDP treatment of NSCLC through negatively regulating ADAM9. 20 In addition, the findings in this study proved that miR-126-5p targeted and inhibited EZH2 expression to facilitate the radiosensitivity of lung adenocarcinoma cells. Similarly, EZH2 is identified as a direct target of miR-126. 21 Multiple evidences have suggested that EZH2 is aberrantly expressed in lung adenocarcinoma. 22,23 EZH2 is capable of inducing cell proliferation, invasion and migration so as to enhance cancer progression. 10 Upregulated EZH2 promotes lung adenocarcinoma cell invasiveness and metastasis in contribution to the progression of lung adenocarcinoma. 11 EZH2 plays a role in metastatic disease recurrence following radiotherapy. 12 At the same time, it has been explored that EZH2 contributes to the radioresistant phenotype of lung adenocarcinoma F I G U R E 6 miR-126-5p affects the radiosensitivity of lung adenocarcinoma cells via the EZH2/KLF2/BIRC5 axis in vivo. Nude mice were treated with 20 Gy X-ray alone or combined with lv-miR-126-5p. (A) Tumour volume of nude mice. (B) Xenograft tumour in nude mice. (C) Tumour weight of nude mice. (D) miR-126-5p expression in tumour tissues of nude mice determined by RT-qPCR. (E) Protein levels of EZH2, KLF2, and BIRC5 in tumour tissues of nude mice measured by Western blot analysis. (F) Molecular mechanism of miR-126-5p involved in regulating the EZH2/KLF2/BIRC5 axis to improve the radiosensitivity of lung adenocarcinoma cells. Data are shown as the mean ± standard deviation of three technical replicates. Data comparisons between two groups were analysed by unpaired t-test. Data comparisons among multiple groups were analysed by the one-way ANOVA with Tukey's post hoc test. *p < 0.05; **p < 0.01; ***p < 0.001 cells, providing the novel insights for investigating clinical treatment targets. 24 Moreover, the target relation between miR-126 and EZH2 in chemosensitivity in gastric cancer cells has been proved previously. 21 These findings support that miR-126-5p facilitated the radiosensitivity of lung adenocarcinoma cells by targeting EZH2.
Furthermore, this study also indicated that EZH2 inhibited KLF2 expression to repress the radiosensitivity of lung adenocarcinoma cells. As a catalytic component of the polycomb repressive complex 2, EZH2 is a histone methyltransferase to suppress the specific gene expression. 25 A prior study has proved that EZH2 can inhibit KLF2. 26 KLF2 is poorly expressed in various cancers and possesses tumour-suppressive features by repressing cell proliferation. 27 Downregulated KLF2 is proved to promote the progression of NSCLC by binding to EZH2. 28 Additionally, KLF2 suppressed BIRC5 to promote the radiosensitivity of lung adenocarcinoma cells. BIRC5 has been demonstrated to act as an oncogene to modulate the growth, migration and invasion of various cancer cells. 29 Silencing of BIRC5 inhibited cell proliferation and colony formation but enhanced apoptosis and radiosensitivity of lung adenocarcinoma cells. 30 Therefore, it can be concluded that miR-126-5p facilitated proliferation, invasion and radiosensitivity of lung adenocarcinoma cells and repressed apoptosis by targeting EZH2 via interaction between KLF2 and BIRC5.
To sum up, our study demonstrated that miR-126-5p inhibited EZH2 to elevate KLF2 expression and reduced BIRC5 expression, all of which leads to the repression of cell proliferation, migration and enhancement of radiosensitivity of lung adenocarcinoma cells and apoptosis ( Figure 6F). Our findings identify an effective therapeutic strategy against radio-resistance of lung adenocarcinoma cells. Due to the limited known researches, the roles of miR-126-5p, EZH2, KLF2, BIRC5 as well as their interaction in the radio-resistance of lung adenocarcinoma cells should be more clearly investigated.

ACK N OWLED G EM ENTS
Not applicable.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.