Long noncoding RNASEH1‐AS1 exacerbates the progression of non‐small cell lung cancer by acting as a ceRNA to regulate microRNA‐516a‐5p/FOXK1 and thereby activating the Wnt/β‐catenin signaling pathway

Abstract Background Till now, no study has focused on the functions of RNASEH1 antisense RNA 1 (RNASEH1‐AS1) in non‐small cell lung cancer (NSCLC). Accordingly, we measured the expression of RNASEH1‐AS1 in NSCLC and characterized its functions in detail. Finally, our research elucidated the mechanisms that occurred downstream of RNASEH1‐AS1. Methods RNASEH1‐AS1 expression was examined utilizing TCGA database and qRT‐PCR. Functional experiments were conducted to study the tumor‐associated functions of RNASEH1‐AS1. The targeting relationship among RNASEH1‐AS1, microRNA‐516a‐5p (miR‐516a‐5p), and forkhead box K1 (FOXK1) was revealed utilizing RNA immunoprecipitation and luciferase reporter assays. Results Utilizing TCGA database and our own cohort, we found a significantly increased level of RNASEH1‐AS1 in NSCLC. The high level of RNASEH1‐AS1 was markedly related with poor clinical outcomes. Knockdown of RNASEH1‐AS1 expression inhibited NSCLC cell growth, metastatic capacities, and epithelial‐mesenchymal transition and promoted the apoptosis in vitro, whereas RNASEH1‐AS1 overexpression exerted the opposite effects. Additionally, knocking down RNASEH1‐AS1 expression suppressed tumor growth in vivo. RNASEH1‐AS1 was confirmed to act as a miR‐516a‐5p sponge, consequently upregulating FOXK1 expression in NSCLC cells. As revealed by the subsequent rescue experiments, the miR‐516a‐5p/FOXK1 axis served as a downstream effector of RNASEH1‐AS1. In addition, by controlling the miR‐516a‐5p/FOXK1 axis, RNASEH1‐AS1 was capable of activating the Wnt/β‐catenin pathway. Conclusion RNASEH1‐AS1 exacerbated the oncogenicity of NSCLC by affecting the miR‐516a‐5p/FOXK1 axis and consequently promoting the activation of Wnt/β‐catenin pathway. Our newly identified RNASEH1‐AS1/miR‐516a‐5p/FOXK1/Wnt/β‐catenin network may offer an interesting foundation for NSCLC treatment in the clinic.


| INTRODUCTION
Lung cancer is the most frequently diagnosed cancer and the leading cause of tumor-associated death around the world. 1 Moreover, the worldwide incidence and mortality rates of lung cancer have continued to increase for many years. 2 More than one-third of new NSCLC cases are diagnosed in China, and this rate of diagnosis causes a notably heavy public health burden. 3 Lung cancer includes two different histological subtypes: non-small cell lung cancer (NSCLC) and small cell lung cancer; the former responsible for approximately 80%-85% of all lung cancer cases. 4 In clinical practice, only a relatively small number of NSCLC patients is diagnosed in the early stage and are eligible to undergo surgical excision. If it is not detected at an early stage, the disease can progress to a middle or late stage with local or distant metastasis, and the best opportunity for treatment is lost. 5 Despite the substantial progress in the development of therapeutic strategies in recent decades, patients with NSCLC often experience poor clinical outcomes. 6 Therefore, to improve therapeutic efficiency, it is necessary to reveal the molecular events that occur during NSCLC pathogenesis, which are currently largely unknown.
Long noncoding RNAs (lncRNAs) are a class of RNA molecules that are 200 nucleotides in length and lack protein-coding capacity. 7 Previously, lncRNAs were considered to be useless byproducts of genome transcription; however, an increasing amount of evidence has shown the important roles of lncRNAs in physiological and pathological biological processes. [8][9][10] Recently, increasing numbers of studies have revealed the dysregulation of lncRNA expression in almost all types of human cancer, and their aberrant expression is required for inducing malignancy and tumorigenesis. 11 In NSCLC, lncRNAs play cancer suppressive or carcinogenic roles and thus affect the onset and progression of NSCLC. 12 MicroRNAs (miRNAs) are a family of endogenous, single-stranded, and noncoding RNA molecules that consist of approximately 17-25 nucleotides. 13 By complementary interacting with the 3′-UTRs of downstream targets, miRNAs decrease gene expression via translation attenuation or mRNA degradation. 14 MiRNAs are closely associated with NSCLC pathogenesis, and they perform crucial actions in controlling tumor-related gene expression, thus regulating the aggressive phenotypes of NSCLC. 15 The competitive endogenous RNA (ceRNA) theory has been described and used to construct a key regulatory pathway through which lncRNAs can adsorb specific miRNAs and thereby modify the levels of downstream target mRNAs. 16 Untill now, no study has illustrated the expression and functions of RNASEH1 antisense RNA 1 (RNASEH1-AS1) in NSCLC. Our research revealed that RNASEH1-AS1 is overexpressed in NSCLC. Additionally, functional experiments were conducted to identify its tumor-related roles.

| Subjects
This research was approved by the Research Ethics Committee of The Fourth Hospital of Changsha. We collected NSCLC tissues from 59 patients in our hospital, and all of these subjects provided written informed consent. None anticancer therapies were administered to these participants. All fresh tissues were preserved in liquid nitrogen. The exclusion criteria were: patients who had been treated with anticancer therapies, patients with other types of human cancer, and patients with severe liver and kidney dysfunction.

| Cell lines
Three NSCLC cell lines, namely, the H460, H522, and H1299 cell lines, were maintained in PMI 1640 medium (Gibco; Thermo Fisher Scientific, Inc.). The NSCLC cell lines A549 and SK-MES-1 were, respectively, grown in F-12K and MEM medium. BEAS-2B cell line was cultured in bronchial epithelial cell growth medium. Furthermore, 10% fetal bovine serum (FBS) was added to all the cell culture media. All the aforementioned cell lines (ATCC) were grown in a cell incubator at 37°C in 5% CO 2 .

| Quantitative real-time polymerase chain reaction (qRT-PCR)
The extraction of total RNA was achieved utilizing TRIzol Universal Reagent (Tiangen). After quantification via a Nanodrop2000 (OD260) (Thermo Fisher Scientific), detection of miR-516a-5p was implemented applying miRcute miRNA First-Strand cDNA Synthesis Kit and miRcute miRNA qPCR Detection Kit SYBR Green (both from Tiangen). U6 served as the reference gene for normalization. For the measurement of RNASEH1-AS1 and FOXK1 levels, PrimeScript reagent Kit was utilized for the reverse transcription reaction, while quantitative PCR was conducted with the help of PrimeScript™ RT Master Mix (both from Takara). RNASEH1-AS1 and FOXK1 levels were normalized to GAPDH expression. Gene expression was calculated with the 2 −ΔΔCq method. 17 The sequences of all primers are presented in Table 1.

| Subcellular fractionation experiment
The segregation of the nuclear and cytoplasmic fractions was realized with a Cytoplasmic and Nuclear RNA Purification Kit (Norgen). After RNA extraction, the expression distribution of RNASEH1-AS1 was investigated utilizing qRT-PCR.

| Fluorescence in situ hybridization
The in situ localization of RNASEH1-AS1 was proven with the Fluorescent In Situ Hybridization Kit (RiboBio). After receiving the fixation, utilizing 4% paraformaldehyde, NSCLC cells were incubated with probes specifically targeting RNASEH1-AS1 at 37°C overnight. The whole process was performed in the dark. Immediately thereafter, DNA was stained with Hoechst solution. Finally, a confocal laser scanning microscope (Leica) was used for imaging.

| Colony formation assay
The 6-well plates were seeded with 500 transfected cells suspended. The cells were grown at 37°C for 2 weeks to form colonies. We replaced the culture medium every 3 days. Finally, 100% methanol was used to fix the colonies with over 50 cells, which were then stained with 0.1% crystal violet. The colonies were captured with an inverted microscope (Olympus Corp.).

| Flow cytometry analysis
Transfected cells were harvested with trypsin and collected by centrifugation. We then resuspended the cells in 195 μl Annexin V-FITC binding buffer, which was came from an Annexin V-FITC Apoptosis Detection Kit (Beyotime). Totally 5 µl Annexin V-FITC and 10 µl PI was added into cell suspension. Cell apoptosis was examined with a flow cytometer (BD Biosciences).

| Transwell migration and invasion experiments
We harvested transfected cells at 48 h post transfection and resuspended them in FBS-free medium. For the cell The lower compartments were filled with complete culture medium. Twentyfour hours later, 0.1% crystal violet solution was used to stain the migrated cells. After extensive washing, the numbers of migrated cells were counted under a light microscope (×200 magnification). For the cell invasion experiment, the Transwell inserts were pre-coated with Matrigel, and the experimental procedures followed that of the migration experiment.

| Tumor xenograft experiments
The in vivo tumor xenograft experiments were implemented under the permission from the Animal Care and Use Committee of the Fourth Hospital of Changsha. To develop a stable RNASEH1-AS1 depletion cell line, shRNA specific for RNASEH1-AS1 (sh-RNASEH1-AS1), and NC shRNA (sh-NC) were cloned into a lentiviral vector (GenePharma Co., Ltd). The sh-RNASEH1-AS1 sequence was 5′-CCGGCCCAATTGATCTAGTAGTAAAGTCTCG AGACTTTACTACTAGATCAATTGGGTTTTTG-3′, and the sh-NC sequence was 5′-CCGGCACGATAAGACAA TGTATTTCTCGAGAAATACATTGTCTTATCGTGTTT TTG-3′. SK-MES-1 cells were transfected with lentiviral vectors, and cells with stable RNASEH1-AS1 knockdown were selected with puromycin. Four-week-old BALB/c nude mice (HFK Bioscience) were subcutaneously inoculated with 2 × 10 6 SK-MES-1 cells stably overexpressing sh-RNASEH1-AS1. Each group contained five nude mice, and all mice were housed under specific pathogen-free conditions at 25°C and 50% humidity, with a 10:14 light/dark cycle and ad libitum access to food and water. After 1 week, the sizes of the xenografts were measured every 3 days. Tumor volumes were analyzed using the following formula: volume = length × width 2 /2. Thirty-one days later, the mice were euthanized, and the tumor xenografts were removed for weighing and immunohistochemistry.
Kaplan-Meier Plotter (https://kmplot.com/analy sis/) was utilized to unveil the association between RNASEH1-AS1 level and overall survival of patients with NSCLC.

| RNA immunoprecipitation
The RNA immunoprecipitation assay was achieved applying an EZ-Magna RIP™ RNA Binding Protein Immunoprecipitation Kit (Millipore). NSCLC cells were lysed by incubation with complete RIP buffer. Next, RIP buffer supplemented with magnetic beads conjugated to anti-Ago2 or normal mouse IgG (Millipore) was incubated with the cell lysates overnight at 4°C. After rinsing with RIP washing buffer, the proteins were digested via Proteinase K buffer treatment. The immunoprecipitated RNA was detected by qRT-PCR.

| Statistical analysis
All the experiments were repeated three times, and the obtained data presented as mean ±standard deviation were analyzed using SPSS software 19.0 (SPSS). Student's t-test and one-way analysis of variance with Tukey's test were applied for comparing data between groups. Pearson's correlation coefficient was used to assess gene expression relation. p < 0.05 indicated statistical significance.

| RNASEH1-AS1 exacerbates the malignant biological phenotype of NSCLC cells
First, we conducted comparative analyses with the TCGA database and investigated the expression of RNASEH1-AS1 in lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). LUAD and LUSC tissues clearly overexpressed RNASEH1-AS1 ( Figure 1A). Additionally, NSCLC tissues from our own cohort exhibited enhanced expression of RNASEH1-AS1 ( Figure 1B). According to the median value of RNASEH1-AS1, all patients with NSCLC were divided into low-or high-RNASEH1-AS1 expression groups. Using Kaplan-Meier Plotter (https://kmplot.com/analy sis/), we found that NSCLC patients with high RNASEH1-AS1 level displayed shorter overall survival in contrast to patients with low RNASEH1-AS1 level ( Figure 1C), which is consistent with the Kaplan-Meier survival analysis from our cohort ( Figure 1D).
Before we explored the specific functions of RNASEH1-AS1, RNASEH1-AS1 level in NSCLC cell lines was measured, and the upregulation of RNASEH1-AS1 in all the tested NSCLC cell lines was observed (Figure 2A). The H522 cell line was used for loss-of-function experiments because it expressed the highest levels of RNASEH1-AS1 of the five NSCLC cell lines. To prevent off-target effects, two siRNAs, namely, si-RNASEH1-AS1#1 and si-RNASEH1-AS1#2, were used to knock down RNASEH1-AS1. Moreover, SK-MES-1 cells, which expressed the lowest levels of RNASEH1-AS1, were treated with pcDNA3.1-RNASEH1-AS1 for gainof-function experiments ( Figure 2B). The proliferation and colony-forming capacities were severely impaired after RNASEH1-AS1 knockdown but enhanced after RNASEH1-AS1 overexpression ( Figure 2C,D). RNASEH1-AS1-knockdown H522 cells exhibited obvious increase in apoptosis compared with si-NC-transfected cells. Similarly, the apoptosis rate was notably decreased in SK-MES-1 cells overexpressing RNASEH1-AS1 ( Figure 2E).
The regulatory effect of RNASEH1-AS1 on metastasis was investigated. The knockdown of RNASEH1-AS1 expression decreased the motility of the NSCLC cells, while pcDNA3.1-RNASEH1-AS1 treatment yielded the opposite result ( Figure 3A,B). Furthermore, the western blotting results verified that RNASEH1-AS1 knockdown decreased N-cadherin and Vimentin expression while increasing E-cadherin level in H522 cells. In contrast, N-cadherin and Vimentin levels were increased, whereas E-cadherin level was decreased, in the pcDNA3.1-RNASEH1-AS1-transfected SK-MES-1 cells ( Figure 3C). Overall, RNASEH1-AS1 exerted carcinogenic effects in NSCLC cells.
Luciferase reporter assay was then conducted to further verify the targeting site between RNASEH1-AS1 and miR-516a-5p ( Figure 4I). MiR-516a-5p overexpression hindered the luciferase activity of RNASEH1-AS1-wt; however, when the targeting sequences were mutated, this inhibitory effect on the luciferase activity was abrogated ( Figure 4J). Ago2 is an important element of RNA-induced silencing complex (RISC) and required for miRNA-mediated gene silencing. It initiates the degradation of target genes by means of its catalytic activity in gene silencing processes guided by miRNAs. To address whether RNASEH1-AS1 relates with RISC complex, RIP assay was performed, and we measured RNASEH1-AS1 and miR-516a-5p in the Ago2 immunoprecipitates. RNASEH1-AS1 and miR-516a-5p were evidently enriched in the Ago2 immunoprecipitates ( Figure 4K). Thus, miR-516a-5p was sequestered by RNASEH1-AS1.
Then, we used bioinformatics tool to find potential targets of miR-516a-5p. FOXK1 harbored possible binding sequences for miR-516a-5p ( Figure 6A) and was thoroughly investigated due to its important oncogenic roles in NSCLC. FOXK1 expression was verified to be overexpressed in NSCLC tissues, and a negative relationship was affirmed between FOXK1 expression and miR-516a-5p levels ( Figure 6B,C). The miR-516a-5p-overexpressing NSCLC cells exhibited decreased luciferase activity triggered by FOXK1-wt, but not of FOXK1-mut ( Figure 6D). In addition, miR-516a-5p mimic-transfected NSCLC cells exhibited decreased FOXK1 expression levels ( Figure 6E,F).
pc-FOXK1 and si-FOXK1 were also used in rescue experiments, and western blotting confirmed the transfection efficiencies ( Figure 8A). Reintroduction of FOXK1 reversed the regulatory effects of si-RNASEH1-AS1. In addition, the increased growth and motility as well as the decreased apoptosis of NSCLC cells overexpressing RNASEH1-AS1 were reversed by FOXK1 knockdown (Figure 8B-F). In short, RNASEH1-AS1 affected NSCLC malignancy by controlling miR-516a-5p/FOXK1 axis.

| DISCUSSION
To date, substantial evidence has revealed that the onset and progression of NSCLC are affected not only by proteincoding genes, but also by non-protein-coding genes. [20][21][22] Thus, an intensive assessment of the detailed roles and working mechanisms of lncRNAs in NSCLC would be conducive to exploit promising theoretical evidence to develop targeted therapy. Many lncRNAs exist in the human genome; however, their roles in NSCLC pathogenesis are not completely understood. Herein, we determined the RNASEH1-AS1 expression in NSCLC and characterized its clinical value. Concurrently, functional experiments were conducted to study the tumor-associated roles of RNASEH1-AS1. Ultimately, our research also WT1-AS, 27 and MEG3 28 is identified in NSCLC and can lower the aggressive events of NSCLC. However, the functions of RNASEH1-AS1 in NSCLC have rarely been explored. Herein, utilizing TCGA database and our own cohort, we revealed a significant increase in the level of RNASEH1-AS1 in NSCLC. RNASEH1-AS1 upregulation was markedly related with poor clinical outcomes. RNASEH1-AS1 knockdown inhibited the growth, metastatic capacities and ETM, and promoted the apoptosis of NSCLC cells in vitro, whereas RNASEH1-AS1 overexpression exerted the opposite actions. Additionally, knocking down RNASEH1-AS1 expression inhibited tumor growth and ETM in vivo. In short, these F I G U R E 9 RNASEH1-AS1 affects the Wnt/β-catenin pathway. H522 cells were transfected with si-NC, si-RNASEH1-AS1+NC inhibitor, or si-RNASEH1-AS1+ miR-516a-5p inhibitor. SK-MES-1 cells were transfected with pcDNA3.1, pc-RNASEH1-AS1+NC mimic, or pc-RNASEH1-AS1+ miR-516a-5p mimic. After transfection, expression levels of βcatenin, cyclin D1 and c-Myc were examined. **p<0.01 observations enhance the understanding of the complex mechanism that occur during NSCLC pathogenesis.
The deeper molecular events underlying the anticancer activities of RNASEH1-AS1 knockdown are still unknown. To address this problem, the location of RNASEH1-AS1 in NSCLC cells was initially examined because the working mechanisms of RNASEH1-AS1 are likely determined by its subcellular distribution. Our data showed that most RNASEH1-AS1 was located in the cytoplasm of NSCLC cells. Over the years, the ceRNA theory authenticated that cytosolic lncRNAs can sequester miRNAs and thereby alleviate the miRNA-mediated decrease in target mRNA levels, consequently exerting regulatory effects. 29 A bioinformatics tool revealed that miR-516a-5p contained complementary bases that might pair with RNASEH1-AS1. Subsequently, RNASEH1-AS1 was shown to act as a miR-516a-5p sponge. Furthermore, utilizing bioinformatics prediction, luciferase reporter assay and molecular detection, FOXK1 was certified as a downstream target of miR-516a-5p. FOXK1 expression was further revealed to be positively modulated by RNASEH1-AS1, and the effect was realized by sponging miR-516a-5p. Altogether, three RNAs, namely, RNASEH1-AS1, miR-516a-5p, and FOXK1, form a newly identified ceRNA pathway in NSCLC cells.
MiR-516a-5p exhibits different expression and functional patterns in different human cancers. For instance, miR-516a-5p is overexpressed in bladder cancer and exerts pro-oncogenic effects. 30 However, miR-516a-5p is decreased in hepatocellular carcinoma 31,32 and papillary thyroid cancer, 33 and acknowledged as an anti-oncogenic miRNA. In NSCLC, downregulated miR-516a-5p expression is clearly related with aggressive clinicopathological parameters. 34 Furthermore, miR-516a-5p is an independent prognostic factor for predicting cancer relapse. 34 Functionally, miR-516a-5p participates in the regulation of multiple malignant properties, 34 which is consistent with our data. In mechanistic studies, FOXK1, a member of a subclass of Forkhead transcription factors, was a downstream target of miR-516a-5p in NSCLC. FOXK1 induces the Wnt/β-catenin pathway and consequently regulates tumor genesis and progression. 18 Here, our data further proved that the regulatory actions of RNASEH1-AS1 on the Wnt/β-catenin pathway were achieved by targeting miR-516a-5p/FOXK1. According to the subsequent rescue experiments, the miR-516a-5p/FOXK1 axis was further confirmed to be a downstream effector of the carcinostatic activities of RNASEH1-AS1 in NSCLC cells. Overall, RNASEH1-AS1 exacerbated the aggressiveness of NSCLC cells by controlling the miR-516a-5p/FOXK1/Wnt/βcatenin pathway.
Our study had two limitations. First, we illustrated the detailed functions of RNASEH1-AS1 in NSCLC and the downstream mechanisms; but, the underlying molecular events caused RNASEH1-AS1 dysregulation were not unveiled. Second, the regulatory actions of RNASEH1-AS1 on NSCLC cell metastasis in vivo were not explored. We will tackle the limitations in the near future.

| CONCLUSION
The central observations of our research confirmed that high RNASEH1-AS1 expression in NSCLC cells were related to poor patient prognosis. RNASEH1-AS1 accelerated the malignant progression of NSCLC cells, and its effects on promoting NSCLC progression were achieved by sequestering miR-516a-5p and overexpressing FOXK1, which, in turn, activated the Wnt/β-catenin pathway. Our newly identified RNASEH1-AS1/miR-516a-5p/FOXK1/ Wnt/β-catenin network may offer an interesting foundation for NSCLC treatment in the clinic.

ACKNOWLEDGEMENT
The Fourth Hospital of Changsha provided financial support.