LncRNA LOXL1‐AS1 regulates the tumorigenesis and development of lung adenocarcinoma through sponging miR‐423‐5p and targeting MYBL2

Abstract Lung adenocarcinoma (LUAD) is the most common form of malignant tumor and closely correlated with high risk of death worldwide. Accumulating researches have manifested that long noncoding RNAs (lncRNAs) are deeply involved in the progression of multiple cancers. LncRNA LOXL1 antisense RNA 1 (LOXL1‐AS1) was identified as an oncogene in several cancers, nonetheless, its biological effect and regulatory mechanism have not been explained in LUAD. Our present study suggested that LOXL1‐AS1 expression was considerably increased in LUAD tissues and cells. Moreover, LOXL1‐AS1 deficiency notably hampered cell proliferation and migration as well as dramatically facilitated cell apoptosis. Through molecular mechanism assays, LOXL1‐AS1 was identified as a cytoplasmic RNA and acted as a sponge of miR‐423‐5p. Furthermore, MYBL2 was targeted and negatively modified by miR‐423‐5p. Rescue experiments revealed that MYBL2 knockdown could counteract miR‐423‐5p repression‐mediated enhancement on the progression of LOXL1‐AS1 downregulated LUAD cells. More importantly, MYBL2 was discovered to interact with LOXL1‐AS1 promoter, indicating a positive feedback loop of LOXL1‐AS1/miR‐423‐5p/MYBL2 in LUAD. These findings manifested the carcinogenic role of LOXL1‐AS1 and LOXL1‐AS1/miR‐423‐5p/MYBL2 feedback loop in LUAD, which could be helpful to explore effective therapeutic strategy for LUAD patients.

patients is tremendously poor. 5 Therefore, there is an exigent necessity to probe novel regulatory mechanism of LUAD and to investigate underlying therapeutic targets for LUAD patients.
Long noncoding RNAs (lncRNAs) are a type of transcripts that lengths are over 200 nucleotides and limited to encode proteins and related to biological processes. [6][7][8] A growing number of lncRNAs exerted their regulatory roles in various tumors. LncRNA CHRF, a sponge of miR-NA-10b, promoted cell proliferation and epithelial-mesenchymal transition (EMT) process in prostate cancer. 9 LncRNA ZEB1-AS1 mediated the miR-200c/141-ZEB1 axis to facilitate tumor growth and cell metastasis in glioma cancer. 10 LncRNA SNHG14, which was identified as an oncogene, boosted the development of cervical cancer via regulating miR-206 and targeting YWHAZ. 11 Previous investigations also indicated that a variety of lncRNAs were participated in the regulation of LUAD development. LINC00324 targeted miR-615-5p/AKT1 axis to promote the tumor process of LUAD. 12 LncRNA LINC00460 promoted tumor growth and resulted to a terrible prognosis in LUAD by completely sponging miR-302c-5p and upregulating FOXA1 expression. 13 Overexpressed SNHG6 in LUAD promoted cell proliferation and metastasis through regulating E2F7 expression and sponging miR-26a-5p. 14 LOXL1 antisense RNA 1 (LOXL1-AS1) has been confirmed as an oncogene and its regulatory mechanism in other cancers has been explored. In medulloblastoma, LOXL1-AS1 enhances its proliferation and metastasis via the activation of PI3K pathway. 15 Nevertheless, its biological role and molecule mechanism in LUAD remain largely unknown.
Here, we explored the functional role and potential mechanism of LOXL1-AS1 in LUAD, and the interaction between LOXL1-AS1 and MYBL2 feedback loop was also studied. It was found that LOXL1-AS1 accelerates tumor progression of LUAD through sponging miR-423-5p and upregulating MYBL2, which meant that LOXL1-AS1 could be a regulator in LUAD development.

| Cell tissues
Forty patients who underwent surgery at Affiliated Hospital of Hebei University from February 2016 to March 2017 provided the LUAD tissues and matched adjacent normal tissues. They all did not receive any preoperative treatment. For extracting RNA, the specimens were collected and promptly stored at −80°C after surgery. This research was conducted with the approval from ethics committee of Affiliated Hospital of Hebei University. Before surgery, all patients signed the written informed consents.

PCR)
RNAiso Plus (Takara) was adopted for the extraction of total RNA. Synthesis of cDNA was carried out afterwards. SYBR Premix Ex Taq II Kit (TaKaRa) with a Light Cycler System was utilized for conducting Quantitative PCR. GAPDH or U6 was applied to normalize the RNA input via 2 −ΔΔC t method.

| Colony formation assay
A549 or SPC-A1 cells after transfection were collected, and then seeded to 6-well plates at a density of 1 × 10 5 cells per well. After culturing for 5 days, cells were stained with 1% crystal violet (Sigma-Aldrich) for 20 minutes. After washing with phosphate buffer saline (PBS; Thermo Fisher Scientific) thrice, the efficiency of clone formation was calculated by DMi8 S Platform Live cell microscope (Leica Microsystems).

| TUNEL staining assay
Cell apoptosis was assessed by TUNEL staining assay using the In Situ Cell Death Detection Kit (Roche) following the recommended guidelines. Biologic coloring agent of DAPI (Haoran Biotechnology) or Merge (Gene-denovo) was applied to dye A549 or SPC-A1 cells. Relative fluorescence intensity was detected via an EVOS FL microscope (Thermo Fisher Scientific).

| Transwell migration assay
Transwell chambers were purchased from Corning Costar. Transfected A549 or SPC-A1 cells, incubated in complete serum-free medium, were planted into upper chambers, whereas the lower chamber were filled with complete medium with 10% FBS. Using a cotton swab, the nonmigrated cells were removed. Then, the migrated cells were separately fixed and stained with 4% paraformaldehyde (PFA; The BSZH Scientific) and 0.1% crystal violet. The Olympus 1X71 camera system (Leica) was used to photograph migrated cells.

| Separation of cytoplasm and nuclear RNA
A PARIS kit (Thermo Fisher Scientific) was adopted for isolating nuclear and cytoplasmic fractions of A549 or SPC-A1 cells as per its protocol. Reverse transcription and qRT-PCR were carried out with extracted RNA.

| RNA pull down assay
LOXL1-AS1 biotin probe and LOXL1-AS1 no-biotin probe were constructed by GenePharma. For developing probecoated beads, they were cultured with M-280 Streptavidin magnetic beads (Invitrogen). Then, cell lysates of A549 and SPC-AS1 cells were incubated with probe-coated beads at 4°C. Subsequently, RNA complexes combined with the beads were eluted. Levels of microRNAs (miRNAs) were detected with qRT-PCR.

| RIP assay
Based on the specification offered by the supplier, the EZ-Magna RIP kit (Millipore, Billerica, MA, USA) was employed to conduct RIP assay. A549 or SPC-A1 cells were collected via centrifugation and subsequently lysed in RIP lysis buffer (Thermo Fisher Scientific). Cell lysates were subjected to immunoprecipitation with antibody against Ago2 (Millipore) or IgG (Millipore). The immunoprecipitated RNA was harvested and then augmented with qRT-PCR using the Primer Script RT Master Mix kit (TaKaRa).

| Western blot
A549 or SPC-A1 cells after transfection were lysed in order to extract protein. The concentrations of protein samples were evaluated with the bicinchoninic acid assay kit (Pierce). All extracted proteins were isolated by 10% SDS-PAGE (Bio-Rad Laboratories) and later transferred to polyvinylidene fluoride membranes (Millipore). After being treated with primary antibody against MYBL2 (0.8 μg/mL, AV100748, Sigma-Aldrich) or GAPDH (1/20 000, ab128915, Abcam) and secondary antibodies, immunoreactive bands were exposed via ECL method.

| Chromatin immunoprecipitation
The Magna ChIP Kit (Millipore) was adopted for chromatin immunoprecipitation (ChIP) assay. Briefly speaking, after cross-linked chromatin was initially sonicated to 200-500-bp fragments, lysates were immunoprecipitated with antibody against MYBL2 or IgG. Precipitated chromatin DNA was analyzed via qRT-PCR.

| Statistical analysis
GraphPad Prism software (GraphPad) was used to conduct statistical analysis. Data were assessed under Student's t test (two groups) or one-way ANOVA (multiple groups), and shown as mean ± standard deviation. Statistical significance was considered when P-value was <.05. All experiments were conducted for at least three times.
The matched adjacent normal tissues were used as a control. Results displayed that LOXL1-AS1 was strongly overexpressed in LUAD tissues ( Figure 1B). Consistently, data from qRT-PCR manifested that LOXL1-AS1 expression in LUAD cell lines (H1299, H1975, A549, SPC-A1) was extremely higher than that in normal BEAS-2B line ( Figure  1C). Taken together, LOXL1-AS1 was upregulated in LUAD tissues and cells.

| LOXL1-AS1 facilitates cell proliferation and migration as well as suppresses cell apoptosis
To explore the biological role of LOXL1-AS1 in LUAD, shRNAs targeting LOXL1-AS1 (sh-LOXL1-AS1#1, sh-LOXL1-AS1#2) were firstly transfected into A549 and SPC-A1 cells. As a result, the expression of LOXL1-AS1 was remarkably downregulated in sh-LOXL1-AS1 transfected A549 and SPC-A1 cells (Figure 2A). Then, the effect of LOXL1-AS1 knockdown on the proliferation of LUAD cells was tested by conducting CCK-8 assay and colony formation assay. The results depicted that the proliferative ability of A549 and SPC-A1 cells was significantly restrained by LOXL1-AS1 silence ( Figure 2B,C). Through TUNEL assay, it was identified that LOXL1-AS1 deficiency efficaciously facilitated the apoptosis in A549 and SPC-A1 cells ( Figure  2D). Moreover, transwell assay was applied to measure cell migration, and the results revealed that the migratory capacity of LUAD cells was evidently inhibited with the transfection of sh-LOXL1-AS1 ( Figure 2E). In short, LOXL1-AS1 facilitates cell proliferation and migration as well as suppresses cell apoptosis in LUAD.

| LOXL1-AS1 acts as a sponge for miR-423-5p
To research the molecular mechanism of LOXL1-AS1 in LUAD, we first determined the location of LOXL1-AS1 in A549 and SPC-A1 cells by using nuclear-cytoplasmic fractionation. The results indicated that LOXL1-AS1 was mainly distributed in the cytoplasm of LUAD cells ( Figure 3A), indicating the regulatory role of LOXL1-AS1 in posttranscription. Therefore, we speculated that LOXL1-AS1 might function as a ceRNA to sponge certain miRNAs. Then, we searched potential miRNAs for LOXL1-AS1, and 26 miRNAs that could bind to LOXL1-AS1 were identified from starBase. Through RNA pull down assay, it was observed that miR-423-5p was strongly enriched in LOXL1-AS1 biotin probe, while no significant changes were demonstrated in other miRNAs ( Figure 3B). Besides, miR-423-5p expression was found to be prominently downregulated in LUAD cell lines with comparison to BEAS-2B cell ( Figure 3C). Besides, miR-423-5p expression was remarkably increased upon LOXL1-AS1 knockdown ( Figure 3D). In order to perform follow-up assays, we elevated miR-423-5p expression with the transfection of miR-423-5p mimics. Additionally, we also weakened the expression of miR-423-5p by transfecting miR-423-5p inhibitor ( Figure 3E). Furthermore, miR-423-5p was predicted to have two binding sites in LOXL1-AS1, revealing that miR-423-5p was highly matched with the sequence of LOXL1-AS1, which was subsequently verified by luciferase reporter assay ( Figure 3F). RIP assay further validated that LOXL1-AS1 and miR-423-5p were both significantly enriched in Ago2 antibodies, indicating the direct interaction between LOXL1-AS1 and miR-423-5p ( Figure 3F). These data delineated that LOXL1-AS1 acted as a sponge of miR-423-5p.

| MYBL2, a target gene of miR-423-5p, is modulated by LOXL1-AS1
To further support ceRNA hypothesis, we searched the downstream genes of miR-423-5p by bioinformatics analysis. Through RNA22, PicTar, TargetScan and miRmap, potential miR-423-5p target genes (MYBL2, PA2G4, SUPT6H, ZNF609, and SOX12) were presented by Venn diagram ( Figure 4A). Results form qRT-PCR analysis suggested that miR-423-5p potently affected the expression of MYBL2 among these candidate genes in A549 and SPC-A1 cells ( Figure 4B). According to western blot assay, we observed that MYBL2 protein level was decreased by miR-423-5p mimics and increased by miR-423-5p inhibitor ( Figure 4C). Subsequently, MYBL2 expression was detected in LUAD cell lines. As illustrated in Figure 4D, F I G U R E 2 LOXL1-AS1 facilitates cell proliferation and migration as well as suppresses cell apoptosis. A, qRT-PCR was applied to examine the efficiency of LOXL1-AS1 knockdown in A549 and SPC-A1 cells. B and C, The proliferation in LOXL1-AS1 silenced A549 and SPC-A1 cells were accessed. D, The effect of LOXL1-AS1 knockdown on the apoptosis of A549 and SPC-A1 cells was estimated through TUNEL assay. E, Transwell assay was utilized to determine cell migration in LUAD cells transfected with sh-LOXL1-AS1. Data were assessed under Student's t test. *P < .05, **P < .01 are statistical symbols that indicated the significant difference between groups. LOXL1-AS1, LOXL1 antisense RNA 1; LUAD, lung adenocarcinoma; qRT-PCR, quantitative real-time PCR MYBL2 mRNA expression was observably upregulated in LUAD cells in comparison with the control cell lines. Further, it was predicted that there was a binding site for miR-423-5p in MYBL2 3' UTR, and luciferase reporter assay confirmed that the luciferase activity of WT MYBL2 was lessened with the transfection of miR-423-5p mimics, and strengthened by miR-423-5p inhibitor, whereas no significant difference was observed in that of mutant type MYBL2 ( Figure 4E). In addition, RIP assay manifested that LOXL1-AS1 and MYBL2 combined with the RISC which formed by miR-423-5p ( Figure 4F). From Rescue assay, miR-423-5p inhibitor countervailed LOXL1-AS1 knockdown-mediated downregulation on MYBL2 expression ( Figure 4G). All the results manifested that MYBL2, the target gene of miR-423-5p, was modulated by LOXL1-AS1.

| DISCUSSION
In the past decades, accumulating numbers of studies have identified critical effects of lncRNAs on the carcinogenesis and progression of human tumors. LOXL1-AS1, which researched in this study, has been reported to regulate prostate cancer cell cycle through sponging miR-541-3p and F I G U R E 5 LOXL1-AS1/miR-423-5p/MYBL2 axis drives LUAD progression. A, MYBL2 expression in sh-MYBL2 transfected A549 and SPC-A1 cells was tested by the use of qRT-PCR. B and C, The proliferation of LOXL1-AS1 silenced A549 and SPC-A1 cells was evaluated after transfecting NC inhibitor, miR-423-5p inhibitor, miR-423-5p inhibitor + sh-MYBL2#1 through CCK-8 and colony formation assay. D, NC inhibitor, miR-423-5p inhibitor, miR-423-5p inhibitor+sh-MYBL2#1 were transfected to LOXL1-AS1 downregulated A549 and SPC-A1 cells for observing cell apoptosis with TUNEL assay. E, Cell migration in A549 cell/sh-LOXL1-AS1#1 and SPC-A1 cell/sh-LOXL1-AS1#1 transfected with NC inhibitor, miR-423-5p inhibitor, miR-423-5p inhibitor+sh-MYBL2#1 was estimated with the employment of transwell assay. Data were assessed under one-way ANOVA. *P < .05, **P < .01 are statistical symbols that indicated the significant difference between groups. CCK-8, cell counting kit-8; LOXL1-AS1, LOXL1 antisense RNA 1; LUAD, lung adenocarcinoma; qRT-PCR, quantitative real-time PCR targeting CCND1. 18 Moreover, it was also found to predict a poor prognosis and promote cell proliferation and metastasis in osteosarcoma. 19 Nonetheless, its role in LUAD remains extremely unclear. In our study, a remarkable upregulation of LOXL1-AS1 was discovered in LUAD tissues and cells. In addition, silenced LOXL1-AS1 could effectively hamper cell proliferation and migration as well as enhance cell apoptosis in LUAD. All these results indicated the oncogenic role of LOXL1-AS1 in LUAD.
MicroRNAs are small RNAs, which contain 21-25 nucleotides, 20 and correlated with the modulation of cellular processes, like cell proliferation, 21 differentiation, 22 and apoptosis. 23 Through years of research, the dysregulation of miRNAs has been observed in multiple cancers, including LUAD. 24 For example, miR-767-3p, an upstream gene of CLDN18, restrains cell growth and migration in LUAD. 25 Through regulating AKT2, miR-608 modulates the apoptosis of human LUAD cells. 26 We also noticed that miR-218 can inhibit EMT transition of LUAD cells via targeting Robo1 and Ecop. 27 Increasing evidence suggested that lncRNA could act as a sponge of target miRNA to modulate cancer progression. 28 For instance, lncRNA HNF1A-AS1 sponges miR-17-5p to boost cell growth in non-small cell lung cancer. 29 LncRNA HOXD-AS1, a sponge of miR-133a-3p, enhances cell proliferation and invasion via the activation of Wnt/β-catenin signaling pathway in ovarian cancer. 30 MiR-423-5p can be downregulated by lncRNA AFAP1-AS1 and thereby facilitated the metastasis of nasopharyngeal carcinoma. 31 However, its function and molecular mechanism in LUAD are largely obscure. Our present study indicated that miR-423-5p was considerably underexpressed in LUAD cells. Furthermore, its expression was negatively modulated by LOXL1-AS1. Interestingly, two binding sites of miR-423-5p were confirmed in LOXL1-AS1 sequence. Taken all together, LOXL1-AS1 functioned as a sponge of miR-423-5p. F I G U R E 6 Transcription factor MYBL2 targets LOXL1-AS1 promoter region. A, qRT-PCR was conducted to detect the relative expression of LOXL1-AS1 in sh-MYBL2 transfected LUAD cells. B, Luciferase activity of LOXL1-AS1 promoter upon MYBL2 knockdown was determined. C, The luciferase activity of P1, P2, P3, and P4 in LOXL1-AS1 promoter region was evaluated by luciferase reporter assay. D, ChIP assay was applied to further confirm the interaction between MYBL2 and LOXL1-AS1 promoter. E, MYBL2 was predicted to have a binding site in LOXL1-AS1 promoter, which was subsequently validated by luciferase reporter assay. Data were compared under Student's t test. **P < .01, ***P < .001 are statistical symbols that indicated the significant difference between groups. ChIP, chromatin immunoprecipitation; LOXL1-AS1, LOXL1 antisense RNA 1; LUAD, lung adenocarcinoma; qRT-PCR, quantitative real-time PCR Emerging investigations have implied the carcinogenic function of MYBL2 in cancer progression, 32 and it was also reported as a transcription factor in LUAD. 33 Our study revealed that MYBL2 was targeted by miR-423-5p and LOXL1-AS1 regulated the expression of MYBL2 by completely sponging miR-423-5p. Through rescue experiments, it was revealed that LOXL1-AS1/miR-423-5p/MYBL2 axis could facilitate LUAD progression. In addition, MYBL2 was found to positively regulate the expression of LOXL1-AS1 via inhibiting its promoter activity. This result indicated that MYBL2 could bind to LOXL1-AS1 promoter, suggesting a positive feedback loop between LOXL1-AS1 and MYBL2.
In conclusion, our data confirmed that lncRNA LOXL1-AS1 functioned as an oncogene in LUAD, and regulated LUAD development through sponging miR-423-5p and targeting MYBL2. More importantly, MYBL2 could transcriptionally activate the expression of LOXL1-AS1. These results disclosed LOXL1-AS1/miR-423-5p/MYBL2 feedback loop in LUAD, providing a meaningful revelation for the exploration of LUAD diagnostic and therapeutic target.

ACKNOWLEDGMENT
We are very grateful to all individuals and groups involved in this study.