High hsa_circ_0020123 expression indicates poor progression to non-small cell lung cancer by regulating the miR-495/HOXC9 axis

Circular RNAs (circRNAs) belong to non-protein-coding RNAs that regulate different pathophysiological procedures. Upregulation of hsa_circ_0020123 is found in non-small cell lung cancer (NSCLC); however, its activity and functions are not clear. In this study, the results showed that hsa_circ_0020123 expression increased in both tumor tissues and NSCLC cells. A higher hsa_circ_0020123 expression also led to poor prognoses among NSCLC patients assayed via FISH. The data of FISH also confirmed that hsa_circ_0020123 primarily had a cytoplasmic location. Hsa_circ_0020123 knockdown caused a significant decrease in nude mouse xenograft growth. Bioinformatics analyses and dual luciferase reporter assays confirmed that hsa_circ_0020123 was an miR-495 sponge and that the HOXC9 gene was a miR-495 target. The miR-495 downregulation reversed cell migration and proliferation inhibition induced by hsa_circ_0020123 silencing in vitro. HOXC9 overexpression reversed miR-495-induced inhibition of cell migration and proliferation. The dual luciferase reporter assay demonstrated that hsa_circ_0020123 interacted with miR-495 by binding to the HOXC9 3′-UTR to suppresses post-transcriptional HOXC9 expression. Taken together, our study found that hsa_circ_0020123 functioned like a tumor promoter via a novel hsa_circ_0020123/miR-495/HOXC9 axis, highlighting its possibility as a new NSCLC therapeutic target.


INTRODUCTION
Non-small cell lung cancer (NSCLC) includes around 80% of lung cancers, which is a main factor regarding cancer-associated mortality in China [1,2]. Despite treatment advances, overall NSCLC 5-year survival remains low [3]. PET-CT imaging and pathological examination are usually used as diagnostic criteria for NSCLC. Accuracy and speed are severely constrained.
Thus, the identification of diagnostic markers that improve early detection will expand the opportunities for surgical intervention in NSCLC patients.
The current study found that hsa_circ_0020123 had abnormal expression in NSCLC in patients. However, the function and regulatory mechanism of hsa_circ_0020123 as a ceRNA in NSCLC progression are still unclear. Therefore, we examined the biological function of hsa_circ_0020123 and it underlying molecular mechanisms in NSCLC development from multiple viewpoints.

Hsa_circ_0020123 downregulation suppresses tumor growth in nude mouse xenografts
Previous microarray profiles revealed that hsa_circ_0020123 was upregulated in NSCLC tissues [14]. Our present research results also found that hsa_circ_0020123 expression increased in NSCLC tumor tissues compared with adjacent normal tissues using both qPCR and FISH detection assays (P<0.001) ( Figure 1A and 1B). FISH detection also showed that hsa_circ_0020123 was predominantly located in the cytoplasm. We divided samples into relatively high (above the adjacent normal tissue levels; n = 38) and relatively low (below the adjacent normal tissue levels; n = 42) expression levels. There was no correlation between hsa_circ_0020123 expression and clinical factors, such as sex (males and females) and patient age (≤ 60 years and > 60 years). However, the high hsa_circ_0020123 expression was positively correlated to lymph node invasiveness, TNM stage, and tumor size (Table 1). Also, the Gehan-Breslow-Wilcoxon Test survival curves validated that NSCLC patients with high hsa_circ_0020123 expression in NSCLC exhibited low overall survival (P<0.05) ( Figure 1C). The results suggested that hsa_circ_0020123 expression had an indispensable function in NSCLC progression.
We also found that hsa_circ_0020123 was derived and cyclized from a portion of the PDZD8 gene exon and was located at chr10:119042605-119049859 ( Figure  1D). qRT-PCR confirmed that hsa_circ_0020123 expression increased in NSCLC cell lines (H1299, A549, H1975, PC9, and H1650), compared with normal human lung epithelial cells (BEAS-2B) (P<0.001) ( Figure 1E). The highest expression was in H1299 and A549 cells, which were used in the subsequent assays. FISH showed that hsa_circ_0020123 was predominantly located in cytoplasm ( Figure 1F).
Lentivirus-mediated silencing decreased hsa_circ_002 0123 expression (P<0.001) (Figure 2A), and we used A549 cells transfected with hsa_circ_0020123 lentiviral interference vectors to assay tumor formation in nude mouse xenografts. Tumor volumes were measured 5 days after grafting with a vernier caliper. hsa_circ_0020123 knockdown clearly resulted in reduced xenograft volume and weight (P<0.001) ( Figure 2B-2D). Immunohistochemical staining revealed a decrease in the Ki67positivity, consistent with suppression of tumor growth by downregulation of hsa_circ_0020123 ( Figure 2E). qRT-PCR found that miR-495 expression increased in tumor tissues with hsa_circ_0020123 silencing (P<0.001) ( Figure 2F). Knockdown of hsa_circ_0020123 inhibited HOXC9 protein expression as shown in western blots (P<0.001) ( Figure 2G and 2H), consistent with a NSCLC tumorigenesis role.

hsa_circ_0020123 knockdown suppresses NSCLC cell invasion and proliferation via miR-495 sponge activity
Statistical analysis showed that miR-495 was the downstream target of hsa_circ_0020123. qRT-PCR verified that hsa_circ_0020123 expression was suppressed significantly by lentivirus-mediated silencing. miR-495 suppression had no effects on hsa_circ_0020123 expression ( Figure 3A), and hsa_circ_0020123 downregulation significantly promoted miR-495 expression. The miR-495 inhibitor significantly reduced miR-495 expression ( Figure 3B). The CCK8 and colony formation assays verified that hsa_circ_0020123 silencing suppressed A549 and H1299 cell proliferation and miR-495 inhibition reversed hsa_circ_0020123 silencinginduced inhibition of proliferation ( Figure 3C-3F). The A549 and H1299 cell migration was inhibited by hsa_circ_0020123 silencing but restored by decreased miR-495 expression ( Figure 3G, 3H). The overall results showed that hsa_circ_0020123 silencing inhibited NSCLC cell migration and proliferation via miR-495 promotion, but the mechanism remains to be established.

Interactions of hsa_circ_0020123, miR-495, and HOXC9
Bioinformatics analysis of the interactions of the miR-495, hsa_circ_0020123, and HOXC9 indicated that hsa_circ_0020123 targeted miR-495. The results from luciferase reporter assays indicated that miR-495 was a downstream hsa_circ_0020123 binding target (site: 986-1009 of hsa_circ_0020123) ( Figure 5A) and hsa_circ_0020123 inhibited luciferase activity in WT but not in mutated cell lines ( Figure 5B). Bioinformatics analysis also indicated that HOXC9 was a miR-495 target, showing that miR-495 interacted directly with HOXC9 3'-UTR (site: 594-616 of HOXC9 3'-UTR) to suppress expression of its mRNA ( Figure 5C). In the luciferase reporter assay, miR-495 inhibited luciferase activity in WT but not mutant (MUT) cell lines ( Figure 5D). Combined AGING results indicated that hsa_circ_0020123 silencing inhibited NSCLC invasiveness and growth with targeting of the miR-495/HOXC9 axis.

DISCUSSION
CircRNAs have gained attention due to their regulation and functional roles in human cancers. The hsa_circ_0020123 expression increased in both NSCLC cell lines and tissues. High hsa_circ_0020123 expression correlated with a poor prognosis. In this investigation we suggested that hsa_circ_0020123 expression increased in NSCLC tissues. Higher expression predicted a higher grade, higher aggression (lymph node invasiveness), and tumor size, suggesting that hsa_circ_0020123 functioned in NSCLC regulation of invasiveness and proliferation. Further investigations found that hsa_circ_0020123 silencing suppressed tumor growth in nude mouse xenografts, suggestive of a role in tumorigenesis; it also promoted miR-495 expression in vitro. The miR-495 inhibition reversed the cell proliferation and migration inhibition that resulted from hsa_circ_0020123 silencing. The results supported the hypothesis that hsa_circ_002 0123 played a role as a molecular sponge to suppress the biological activity of miR-495; a relationship that was further validated in a dual luciferase reporter assay.

AGING
MiRNAs are expressed abnormally in a number of human cancers and they function in tumorigenesis and tumor development [19][20][21]. Downregulation of miR-495 has previously been reported in melanoma cell lines and tissues, and miR-495 overexpression has been associated with suppression of cell proliferation and invasiveness of melanoma cells [22]. miR-495 was found to suppress hepatocellular carcinoma cell invasion and proliferation through targeting directly insulin-like growth factor receptor-1 [23]. Previous studies have found that miR-495 targets MTA3 in the regulation of lung cancer growth and migration [24]. miRNAs are post transcriptional inhibitors of gene expression, binding to the mRNAs 3′-UTRs. In this study, miR-495 suppressed HOXC9 expression, and HOXC9 overexpression reversed miR-495-induced inhibition of cell migration and proliferation, validating that HOXC9 was a miR-495 target, which was confirmed with the dual luciferase reporter assay.
HOX, or homeobox, genes act as transcription factors. In humans, 39 HOX genes have been identified and organized into 13 paralogous groups and four clusters, including HOXB, HOXC, HOXA, and HOXD. HOXC9 is upregulated in breast cancer, whose downregulation suppresses tumor cell proliferation and invasiveness [25,26]. The present results are consistent with former findings that HOXC9 promotes the invasiveness and proliferation of NSCLC cells. miR-495 bound to the HOXC9 3'-UTR region to suppress its expression, suggesting a role as a tumor suppressor in the miR-495/HOXC9 pathway in NSCLC.

CONCLUSIONS
Hsa_circ_0020123 promoted oncogenesis via functioning as a miR-495 sponge, and may be a newly identified prognostic marker in NSCLC. Targeting the novel hsa_circ_0020123/miR-495/HOXC9 axis represents a potential therapeutic strategy for NSCLC.

Tissue samples
In summary, we collected 92 fresh NSCLC tissues as well as paired adjacent noncancerous lung tissues after obtaining informed patient consent at the Xinhua Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China. We evaluated pathological and histological diagnostics of NSCLC patients following the Revised International System for Staging Lung Cancer (RISSL). Patients did not receive any chemotherapy or radiotherapy before tissue sampling (Table 1). We snap-froze samples in liquid nitrogen and stored them at -80°C prior to RNA extraction. The Ethics Committee in Xinhua Hospital, Shanghai Jiaotong University School of Medicine approved the investigation.

Bioinformatics analysis
We performed analysis and hsa_circ_0020123-miRNAtarget gene activity predictions with the Circular RNA Interactome. Prediction of the miR-495 mRNA target gene was conducted with TargetScanHuman. Expression of HOXC9 gene expression and patient prognosis were estimated using GEPIA.

Clone formation assays and cell proliferation
We assayed cell proliferation with a Cell Counting Kit-8 (CCK-8; Invitrogen, Carlsbad, CA, USA). We seeded transfected cells into plates with 96-wells in triplicate at AGING 2,000 cells per well and assayed viability at 0, 24, 48, 72, and 96 h after seeding them, following the standard procedures of the kit. Colony formation was assayed in transfected cells seeded into plates with 6 wells at 2,000 cells/well and grown in DMEM with 10% FBS for 10 days. We counted and photographed colonies after fixing and staining.

Cell migration assay
We assayed cell migration in Transwell chambers with 24 wells with 8 μm pore-size membranes (BD Biosciences, Franklin Lakes, USA). In brief, we added 1 × 10 5 cells in 200 µL serum-free medium to the upper chamber. We filled lower chamber with 500 µL complete medium as a chemoattractant. We fixed the cells that had invaded the lower chamber after 1 d with 4% paraformaldehyde for 0.5 h and stained them with Crystal Violet for 10 minutes.

Invasiveness assay and tumor xenograft formation
We injected a total of 1 × 10 7 viable wild-type (WT) or siRNA A549 cells into nude mice right flanks [17]. We measured tumor sizes every 5 days with a vernier caliper and calculated the volume as V = 0.5 × length × width 2 . Mice were euthanized and samples prepared for qRT-PCR and Ki67 immunohistochemical staining 30 days after implantation.
For invasiveness analysis, WT and siRNA A549 cells (2×10 5 ) were transfected with luciferase expression vectors, and the cells were injected intravenously into the tails of mice. After 30 d, A549 cell invasiveness was analyzed by bioluminescence imaging following an intravenous injection of luciferin (150 mg luciferin/kg body weight) into the tails.

Western blot assays
We extracted protein from cells or tissues with RIPA lysis buffer. Western blot assays were performed following a previous procedure [18]. We separated protein samples (20 μg) using a 10% sodium dodecyl sulfate polyacrylamide electrophoresis (SDS-PAGE) gel and transferred the proteins to a polyvinylidene fluoride (PVDF) membrane (Bio-Rad Laboratories, Hercules, CA). We blocked membranes with nonfat dry milk (Yili Milk Company, Inner Mongolia, China) at room temperature for 2 hours, and then hybridized with anti-HOXC9 (1:1000 dilution; SC-81100; Santa Cruz Biotechnology, CA.USA) and anti-GAPDH antibodies (1:2000 dilution; SC-365062; Santa Cruz Biotechnology) at 4°C overnight. We next added secondary biotin-conjugated antibodies, and developed immunoreactive bands with the enhanced chemiluminescence kit (Thermo Fisher Scientific). We used GAPDH as a loading control.

Dual luciferase reporter assay
We produced reporter plasmids through inserting circRNA or the HOXC9 3'-UTR sequence into a pmirGLO vector (Promega, Madison, USA). We cotransfected reporter plasmids and miR-495 mimics into human embryonic kidney (HEK) 239T cells through using Lipofectamine 2000. After cultivation for 2 d, we determined firefly and Renilla luciferase activities with a Dual Luciferase Reporter Assay System (Promega, Sunnyvale, USA) following standard protocols.

Statistical analyses
We reported data as the means ± standard deviation (SD). We used GraphPad Prism version 5.0 (GraphPad, AGING La Jolla, USA) to compare differences between groups. P-values ≤ 0.05 were considered as statistical significance.

Ethics approval
The Animal Research Committee of Xinhua Hospital, Shanghai Jiaotong University School of Medicine (Shanghai, China) approved all experimental protocols and surgical procedures.