Circular RNA hsa_circ_0004396 acts as a sponge of miR‐615‐5p to promote non‐small cell lung cancer progression and radioresistance through the upregulation of P21‐Activated Kinase 1

Abstract Backgrounds CircRNA hsa_circ_0004396 has been confirmed to be upregulated in human non‐small cell lung cancer (NSCLC). The aim of his study was to evaluate its mechanism in the radioresistance and progression of NSCLC. Methods Hsa_circ_0004396, miR‐615‐5p, and P21‐Activated Kinase 1 (PAK1) were measured by reverse transcription quantitative real‐time polymerase chain reaction (RT‐qPCR). The binding between miR‐615‐5p and hsa_circ_0004396 or PAK1 was predicted by circinteractome or Targetscan, as verified by dual‐luciferase reporter assay and RIP assay. Proliferation, clonogenicity capacity, cell cycle progression, apoptosis, migration, and invasion were assessed by CCK‐8, colony formation, flow cytometry, and Transwell assay. Bcl‐2, Bcl‐2 associated protein X (Bax), MMP‐2, and PAK1 protein levels were detected using western blot assay. In addition, in vivo function of hsa_circ_0004396 was evaluated by tumor xenograft assay. Results Hsa_circ_0004396 and PAK1 levels were upregulated, while miR‐615‐5p was declined in NSCLC. Hsa_circ_0004396 silencing inhibited NSCLC cell malignant behavior and induced radiosensitivity. Hsa_circ_0004396 functions as a molecular sponge of miR‐615‐5p to regulate PAK1 expression. Moreover, hsa_circ_0004396 knockdown inhibited NSCLC tumor growth in vivo. Conclusion Our findings demonstrated that hsa_circ_0004396 promoted NSCLC development and radioresistance through the miR‐615‐5p/PAK1 axis, which might provide a new therapeutic target for NSCLC treatment.


| INTRODUC TI ON
Lung cancer is a highly lethal cancer worldwide, with non-small cell lung cancer (NSCLC) accounting for 80%. 1 Most advanced-stage NSCLC sufferers are attributed to lacking an effective diagnostic method, which makes cancer hard to surgically remove and has a poor prognosis, while treatment failures are caused by radio-or chemoresistance. 2,3 Thus, it has important implications for further understanding the mechanisms in therapeutic resistance of NSCLC.
Accordingly, the regulatory mechanism of NSCLC progression and radioresistance needs further exploration.
Circular RNA (circRNA) can be divided into noncoding circRNAs and coding circRNAs. 4 Furthermore, circRNAs are classified as noncoding circRNAs due to their highly conserved structure, characterized by covalently closed loops lacking 5′ caps and 3′ poly tails. 5 Although circRNAs has been discovered nearly 40 years, their biological function in many human diseases is only beginning to understand on account of the development of deep RNA sequencing (RNA-seq) technologies and novel bioinformatic approaches. 6,7 Besides, circRNAs function as miRNA sponges, competing with pre-mRNA splicing and serving as circRNA-protein interactions. [8][9][10][11] In NSCLC, abnormal expression of circRNAs closely correlates with tumorigenesis and progression. 12 CircMTDH.4 functioning as a miR-630 sponge to boost growth, metastasis, and radioresistance of NSCLC cells. 13 Also, circ_000867290 deficiency could repress cell growth and enhance the radiosensitivity of NSCLC cells via mediating the miR-375/SPIN1. 14 In sum, circRNAs might play vital roles in NSCLC progression and radiosensitivity. Hsa_circ_0004396jm promoted colorectal cancer cell proliferation and migration. 15 Moreover, another research indicates that hsa_circ_0004396 is significantly upregulated in NSCLC, suggesting a potential attractive biomarker in NSCLC. 16 However, the elaborate role and mechanism of hsa_circ_0004396 in NSCLC progression and radioresistance are unclear.
CircRNAs act as competing endogenous RNAs (ceRNAs) to sponge their target miRNAs and fulfill their regulatory function via interaction with miRNAs, regulating the miRNA-targeted gene expression. [17][18][19] Here, we first discovered by bioinformatics analysis that hsa_circ_0004396 shared binding sequence miR-615-5p. Moreover, some studies exhibited the suppression of miR-615-5p in ovarian cancer and pancreatic ductal cancer. 20,21 Also, miR-615-5p blocked NSCLC development through downstream mRNAs. 22,23 Interestingly, miR-615-5p was previously reported to be related to the radiosensitivity of human cancer. 24 Of note, some studies suggested that P21 (RAC1) activated kinase 1 (PAK1) is an oncogenic serine/threonine kinase that is dysregulated in various cancers. 25 Furthermore, the overexpression of PAK1 could boost chemoresistance and the development of NSCLC cells via different pathways. 26,27 In this study, we aimed to explore whether the regulatory role of hsa_circ_0004396 on NSCLC cell malignant behavior and radiosensitivity could be mediated by miR-615-5p/ PAK1 axis.

| Cell transfection
For stable knocking down of hsa_circ_0004396, short hairpin RNA against hsa_circ_0004396 was applied. Both sh-hsa_circ_0004396 and sh-NC (GenePharma) were constructed into the lentiviral vector (Invitrogen), followed by introduction into A549 cells and cultured in cell medium with 2 μg/ml puromycin.

| RT-qPCR
TRIzol reagent (Invitrogen) was applied to extract the total RNA from tissues and cells. For circRNA expression measurement, total RNA was purified with an RNeasy Mini kit, followed by synthesis cDNA using SuperScript™ IV First-Strand Synthesis System (Thermo Fisher Scientific). After that, qRT-PCR procedures were performed according to AceQ Universal SYBR qPCR Master Mix (Vazyme). Then, cDNA and reaction solutions were mixed according to protocols, and the results were recorded by Thermo Fisher QuantStudio 3 QS3 (Thermo Fisher Scientific). Mir-X™ miRNA First-Strand Synthesis Kit (catalog no 638315, Takara) was used for miRNA quantitation. After being normalized GAPDH and U6, data were examined through the 2 −ΔΔC t method. For circRNAs stability measurement, total RNA was incubated with 5 U/μg RNase R for 20 min at 37°C or treated with actinomycin D (Act D) 2 μg/ml for 12 h to degrade the linear RNAs.
The primers used in this study are shown in Table 1.

| Colony formation assay
The NSCLC cells (5 × 10 2 cells/well) in the logarithmic growth phase were inoculated into the six-well-plates at the densities 1 × 10 3 . After incubation for 24 h, the cells were irradiated with single-dose X-ray A colony formation assay was conducted at least three times.

| Flow cytometry assay
Cell apoptosis was examined by Annexin V Apoptosis Detection Kit I (Solarbio). Briefly, the residual medium was washed with phosphate buffer saline (Hyclone), and A549 and SPC-A1 cells were resuspended in 100 µl binding buffer with Annexin V and Propidium iodide (PI). Ten minutes later under darkness, the samples were determined via flow cytometry.
Cell cycle progression was assessed using An FxCycle PI/RNAse Solution (F10797, Thermo Fisher Scientific). After being fixed with ethanol (75%), the cells were incubated with PI complement with DNase-free RNase A for 1 h in darkness. Data analysis was conducted through FlowJo 10. The experiment was conducted thrice.

| Transwell assay
For migration assay, NSCLC cells (5 × 10 4 cells/well) were seeded in the upper chambers with 200 µl serum-free cell medium, and 500 µl medium containing 10% FBS was added into the lower chamber.
Then, the cells were incubated at 37°C for 24 h. Subsequently, crystal violet (MedChemExpress) was used to stain the cells for 30 min.
The stained cells were observed under an inverted microscope (CarlZeiss). To detect cell invasion, 1 × 10 5 cells were introduced into the upper counterpart pre-covered with Matrigel Matrix, and the subsequent steps were the same as cell migration assay.

| Dual-luciferase reporter assay
The relationship between miR-615-5p and hsa_circ_0004396 or PAK1 was predicted by Starbase. The fragments of wild-type (hsa_circ_0004396-WT) and mutant (hsa_circ_0004396-MUT) hsa_circ_0004396 containing presumptive sites for miR-615-5p were generated by Ribobio. Then, the above sequences were con-

| RNA immunoprecipitation (RIP) assay
Non-small cell lung cancer cells were lysed by RIP lysis buffer (Beyotime), followed by a mixture with magnetic beads and anti-Ago2 or anti-IgG (Millipore). After digestion, immunoprecipitated RNA was analyzed.

| In vivo tumor formation assay
All animal experimental procedures were approved by the Institutional Animal Care and Use Committee of Gansu Provincial Tumor Hospital. For the establishing of the xenograft model, A549 cells (1 × 10 7 ) with sh-NC or sh-hsa_circ_0004396 were subcutaneously injected into the nude mice (4 weeks old, 5 mice/group). At the same time, mice were checked daily with normal activities such as eating, drinking, eliminating, and ambulating. Tumor size was measured every 5 days, followed by calculation following the formula that length × width 2 /2. After injection for 30 days, the mice were sacrificed and the tumor weight was measured. Immunohistochemical staining assay was carried out as described previously, 28 and tissue sections from the obtained tumors were incubated with antibodies specific for Ki67 (0.5 μg/ml, ab15580; Abcam) and (1:100, ab223849; Abcam).

| Statistical analysis
Data were exhibited as mean ± standard deviation (SD). Significance of the variance was analyzed using Student's t test or one-way analysis of variance (ANOVA). Significant differences were defined as p < 0.05.

| Hsa_circ_0004396 was upregulated in NSCLC
QRT-PCR showed a significant increase of hsa_circ_0004396 in NSCLC ( Figure 1A,B). Moreover, hsa_circ_0004396 low expression (divided according to the median) tends to have much better survival rates within a 5-year statistical period than the hsa_circ_0004396 high expression group ( Figure 1C). Meanwhile, our result exhibited that hsa_circ_0004396 was originated from exons 21, 22, and 23 of the HUWE1 gene ( Figure 1D). CircRNAs  are very stable due to their unique covalent loop structure. As expected, the RNase R treatment resulted in the obvious decrease in HUWE1 mRNA, whereas the hsa_circ_0004396 was unaffected in NSCLC cells ( Figure S1A,B). The transcription inhibitor actinomycin D (Act D) allowed an assessment of RNA decay over time. As shown in Figure S1C,D, hsa_circ_0004396 showed more outstanding resistance than HUWE1 mRNA. The above results further confirmed that hsa_circ_0004396 was a stable circRNA. Furthermore, Table 2 displayed patients' clinicopathologic parameters. The expression of hsa_circ_0004396 correlates with the TNM stage and Lymph node metastasis, implying the potential role of hsa_circ_0004396 in the forming and developing of NSCLC.

| Hsa_circ_0004396 deficiency suppressed NSCLC cell malignant behavior and increased radiosensitivity
Small interfering RNA against hsa_circ_0004396 (sh-hsa_ circ_0004396#1, sh-hsa_circ_0004396#2, and sh-hsa_ circ_0004396#3) and sh-NC were transfected into NCCLC cells, followed by verification knockdown efficiency using qRT-PCR ( Figure 2A). Due to significant knockdown efficiency, sh-hsa_ circ_0004396#1 was chosen for further experiments. Cell viability was significantly suppressed by sh-hsa_circ_0004396#1 in NSCLC cells ( Figure 2B-E). Apart from that, we also detected whether hsa_circ_0004396 is associated with the radiotherapy effect. The results showed that the hsa_circ_0004396 knockdown group has a higher survival fraction in a dose-dependent manner ( Figure 2F,G).
Hsa_circ_0004396 downregulation induced G0/G1 arrest and apoptosis in NSCLC cells ( Figure 2H,I). Furthermore, migration and invasion were both constrained via hsa_circ_0004396 knockdown ( Figure 2J,K). Then, protein levels of Bcl-2 (a pro-apoptosis factor), Bax (an anti-apoptosis factor), and MMP-2 (a pro-migration and invasion factor) were assessed. Data indicated that the decreased protein level of Bcl-2 and the increased protein levels of Bax and MMP-2 were viewed due to hsa_circ_0004396 downregulation ( Figure 2L).
Taken together, downregulation of hsa_circ_0004396 suppressed cell malignant behavior and increased radiosensitivity in NSCLC cells in vitro.

| MiR-615-5p was a target of hsa_ circ_0004396
Circinteractome predicted the combinative sequence between hsa_circ_0004396 and miR-615-5p, and binding sites are shown in Figure 3A. Luciferase activity in HEK293T cells was significantly decreased in the hsa_circ_0004396-WT group, while the luciferase activity was changed little in the hsa_circ_0004396-MUT group, suggesting hsa_circ_0004396 sponged miR-615-5p ( Figure 3B,C). Then, enrichments of hsa_circ_0004396 and miR-615-5p were greatly increased when incubated with Anti-Argonaute-2 antibody (anti-Ago2) using Ago2 RIP assay ( Figure 3D,E). Apart from that, the expression level of miR-615-5p was downregulated in NSCLC tissue samples ( Figure 3F) and negatively correlated with the expression of hsa_circ_0004396 ( Figure 3G).

F I G U R E 3
interacted with miR-615-5p.

| PAK1 was directly targeted by miR-615-5p
Considering the binding between miR-615-5p and PAK1 ( Figure S3A), their interaction in HEK293T cells was verified. As exhibited in Figure S3B, the luciferase activity was significantly blocked in the wild-type group, while there was no significant difference in the mutant group. Also, miR-615-5p and PAK1 were enriched in the Ago2 group vs. the IgG group using RIP assay ( Figure S3C,D). Besides, PAK1 expression was remarkably increased in NSCLC cells ( Figure S3E,F). Meanwhile, miR-615-5p was significantly decreased by a miR-615-5p inhibitor ( Figure S3G).
Furthermore, compared with the normal tissues, the levels of PAK1 protein ( Figure S3L) and mRNA ( Figure S3M) were distinctly increased in NSCLC tissues ( Figure S3J). Interestingly, in NSCLC tissues, PAK1 level was negatively correlated with miR-615-5p and positively associated with hsa_circ_0004396 level ( Figure S3N,O), implying the relationship of hsa_circ_0004396, miR-615-5p, and PAK1 in NSCLC. Overall, these data indicated that PAK1 was targeted by miR-615-5p.

| Knockdown of hsa_circ_0004396 inhibited tumor growth in NSCLC in vivo
To further investigate the function of hsa_circ_0004396 in NSCLC in vivo, xenograft formation assay was performed in nude mice. Tumor growth was significantly suppressed by hsa_circ_0004396 knockdown ( Figure 7A,B). Subsequently, qRT-PCR results exhibited that hsa_ circ_0004396 expression was reduced, while miR-615-5p expression was increased in the sh-hsa_circ_0004396 group compared with the  it has been confirmed that miR-615-5p can take part in the regulation of radiosensitivity in prostate cancer cells. 24 In agreement with these reports, miR-615-5p could inhibit NSCLC cell progression and radioresistance in this article.
Moreover, we further confirmed that miR-615-5p targeted the PAK1. PAK1, an oncogene, encodes a family member of serine/ threonine p21-activating kinases. Numerous studies indicated that PAK1 participated in cancers progression and was related to poor prognosis in gastric cancer 33,34 and thyroid cancer. 35,36 Additionally, PAK1 correlates with the radiosensitivity of cancers.
β-elemene was able to enhance the radiosensitivity of gastric cancer cells by inhibiting PAK1 activation. 37 PAK1 tyrosine phosphorylation is necessary to induce epithelial-mesenchymal transition and radioresistance in lung cancer cells. 38 Our research demonstrated that PAK1 was targeted by miR-615-5p in NSCLC. Also, miR-615-5p downregulation or PAK1 upregulation partly reversed hsa_circ_0004396 knockdown-mediated NSCLC cell progression and radioresistance repression in this work, and further verifying hsa_circ_0004396 could regulate the progression and radioresistance in NSCLC by targeting the miR-615-5p/PAK1 axis. However, we failed to detect the downstream signal transduction mechanisms and we are interested in molecular mechanisms by which the hsa_circ_0004396 /miR-615-5p/PAK1 axis exert their function in

ACK N OWLED G M ENT
None.

CO N FLI C T O F I NTE R E S T
The authors have no interests to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.