Circular RNA hsa_circ_0000117 accelerates the proliferation and invasion of gastric cancer cells by regulating the microRNA-337-3p/signal transducer and activator of transcription 3 axis

ABSTRACT Circular RNA hsa_circ_0000117 is reportedly increased in Gastric cancer (GC), however, its role is unexplored. Hsa_circ_0000117 expression and function in GC was investigated using standard cell phenotypic and expression assays. Pull-down and luciferase reporter assays also elucidated hsa_circ_0000117 mechanisms. In the present study, we observed increased hsa_circ_0000117 and signal transducer and activator of transcription 3 (STAT3) expression, while microRNA-337-3p (miR-337-3p) was decreased in GC cells. Depleted hsa_circ_0000117 decreased GC proliferation and invasion. Hsa_circ_0000117 was also identified as a miR-337-3p sponge. Also, STAT3 was identified as a miR-337-3p target. Similarly, rescue assays indicated STAT3 overexpression (or miR-337-3p inhibition) reversed hsa_circ_0000117 effects in GC progression. Thus, our data suggested hsa_circ_0000117 exhibited oncogene properties in combination with the hsa_circ_0000117/miR-337-3p/STAT3 axis in GC, potentially providing a new therapeutic target for GC. Abbreviations GC: gastric cancer; STAT3: Signal transducer and activator of transcription 3; circRNA: Circular RNA; miRNA: microRNA; DMEM: Dulbecco’s modified Eagle’s medium; FBS: fetal bovine serum; PVDF: polyvinylidene fluoride; CCK-8: Cell counting kit-8; qRT-PCR: Quantitative real-time PCR; SDS-PAGE: sodium dodecyl sulfate polyacrylamide gel electrophoresis; TNM: TNM Classification of Malignant Tumors; mTOR: mechanistic target of rapamycin; ANOVA: one-way analysis of variance

Gastric cancer (GC) is a common cancer, and is the third most common cause of cancer death [1]. In 2018, GC culminated in 783000 worldwide deaths and was diagnosed in approximately 1,000,000 individuals [2]. Several factors are implicated in GC etiology, including, Helicobacter pylori infection, eating behaviors, smoking, and being overweight [3]. In recent years, while diagnostic progress and GC standardized treatments have occurred, the prognosis for patients with advanced GC stages remains unsatisfactory [4,5]. Therefore, the molecular mechanisms behind GC tumor progression must be investigated to improve GC treatments.
Circular RNA (circRNA) is a non-coding RNA, and is more stable than linear RNA [6]. Studies have shown circRNAs are dynamically expressed in human cancers and regulate gene expression, suggesting a close association with tumor occurrence and invasiveness [7,8]. It was reported circMTO1 suppressed cancer cell properties by targeting Wnt/β-catenin signaling in colorectal cancer [9]. Also, circPRRC2A promoted cancer cell characteristics via epithelial-mesenchymal transition (EMT), and increased TRPM3 levels in renal cell carcinoma [10]. Similarly, circRNA hsa_-circ_0008285 increased cervical cancer cell properties by activating SOX4, and inhibiting miR-211-5p [11]. Equally, hsa_circ_0000117 was highly abundant in GC [12]. However, hsa_circ_0000117 roles and its underlying mechanism in GC are still unclear.
MicroRNAs (miRNAs) are another crucial class of non-coding RNAs with approximately 21-23 nucleotides in length [13]. Recently, miR-337-3p has been established to be a tumor inhibitor in various human cancers, including GC. For instance, loss of miR-337-3p expression in GC was correlated with lymph node metastasis [14]. Also, miR-337-3p suppressed GC cells proliferation and invasion of GC cells through repressing MZF1-facilitated expression of MMP14 [15]. Similarly, miR-337-3p reduced GC cells growth and invasion through regulating ARHGAP10 [16]. Increasing studies have demonstrated that circRNAs could function as direct sponges for miRNAs to play their essential biological functions in eukaryotic cells [17,18]. However, the relationship between hsa_circ_0000117 and miR-337-3p has not been studied.
Here, we explored the biological functions and molecular mechanisms of hsa_circ_0000117. We identified that the hsa_circ_0000117/miR-337-3p/ STAT3 feedback loop enhanced GC cell proliferation and invasion, suggesting a new treatment strategy for GC.

Patient tissues
Thirty-nine (39) tumor specimen pairs and matched normal samples were obtained from the Affiliated Hospital of North Sichuan Medical College. GC surgery is provided at this facility. All patients were informed and signed consent forms. The research was approved by the local Ethics Committee. Participants received no preoperative treatment prior to study commencement, and were diagnosed with GC by two independent pathologists. Tissue specimens were stored at −80°C.

Quantitative real-time PCR
Total RNA was isolated from tissue or cells using TRIzol® (Invitrogen) and cDNA synthesized using the miScript Reverse Transcription Kit (Qiagen, Germany). After reverse transcription, qPCR was conducted using the miScript SYBR Green PCR Kit (Qiagen) to quantify miR-337-3p levels with respect to U6 small nuclear RNA. To quantify hsa_circ_0000117 and STAT3, cDNAs were generated by the PrimeScript RT-reagent kit (Takara, Japan), and levels standardized to GAPDH expression. Data were processed using the 2 −ΔΔCt approach.

Transwell assay
2 × 10 5 cells were cultured in an upper transwell chamber containing Matrigel (BD Biosciences, NJ, USA) in 200 μl medium lacking serum. The lower chamber contained 600 μl complete medium. After a 48 h incubation, the upper chamber was removed, and 4% paraformaldehyde and 0.5% crystal violet sequentially added to fix and stain cells, respectively, in the lower chamber. A light microscope (Nikon, Japan) was used to count cells.

Western blotting
The extracted proteins were measure by BCA Protein Assay Kit (Thermo, USA), separated by SDS-PAGE, transferred to PVDF membrane, and blocked in 5% fat-free milk for 1 h. Then, the membranes were incubated with primary antibodies against STAT3 (Santa Cruz, CA, USA). Secondary antibody was added and incubated for 2 h at 37 . Finally, the immunofluorescence of protein was visualized using an ECL detection system.

Pull-down assay
Pull-down assay was used as previous study [10]. Briefly, hsa_circ_0000117 probe was used for incubation with M-280 Streptavidin magnetic beads for 2 h at room temperature to generate probe-coated beads (Invitrogen). Cell lysate with hsa_circ_0000117 probe or oligo probe was incubated for a night at 4 . Then, the RNA mix bound to the beads was eluted and extracted with a RNeasy Mini Kit for qRT-PCR (Qiagen). SPSS 18.0 was used for statistical analysis. Data were represented as the mean ± SD. Student's t-test compared two groups, and ANOVA demonstrated differences between more than two groups. P < 0.05 indicated statistical significance.

Hsa_circ_0000117 is upregulated in GC
Previously, circRNA hsa_circ_0000117 expression was shown to be increased in GC [12]. However, the roles of hsa_circ_0000117 is still unclear in GC progression. In the present study, our expression data indicated hsa_circ_0000117 levels were indeed elevated in GC tissue and cell lines (Figure 1(a-c)). Additionally, correlation analyses showed this expression was positively correlated with advanced TNM stage, and lymphatic metastasis in GC patients (Figure 1(d-e)). Thus, hsa_-circ_0000117 appeared to play a critical role in GC development.

Hsa_circ_0000117 silencing reduces GC proliferation and invasion
To assess hsa_circ_0000117 mechanisms in GC tumorigenesis, AGS and BGC823 cells were transfected with si-circ_0000117 to silence expression (Figure 2(a)). CCK-8 and colony formation studies indicated this silencing decreased in vitro GC proliferation and colony formation (Figure 2(b-d)). Transwell assay data also indicated hsa_-circ_0000117 knockdown impeded GC cell invasion (Figure 2(e)). Combined, hsa_circ_0000117 silencing appeared to suppress key GC phenotypes.
Next, we examined miR-337-3p expression in GC tissue. Our qRT-PCR data indicated this expression was downregulated in GC tissue, and corroborated by the TCGA database (figure 3(fg)). Moreover, hsa_circ_0000117 silencing elevated miR-337-3p expression in GC cells (Figure 3(h)). Together, hsa_circ_0000117 appeared to act as a sponge for miR-337-3p in GC cells.

STAT3 is a downstream target of miR-337-3p
To investigate miR-337-3p target genes, mirTarBase and MicroT-CDS databases were searched. Intersection analyses identified 10 potential genes (Figure 4(a)) which were explored using the TCGA database, from which STAT3 was chosen ( Figure  4(b-c)). Immunohistochemistry data revealed STAT3 levels were increased in GC tissue (Figure 4  (d)). Kaplan-Meier analysis revealed elevated STAT3 expression was linked with poor overall survival in GC patients (Figure 4(e-f)). Moreover, function assays showed STAT3 inhibition decreased in vitro colony formation and invasion capacities in AGS cells (Figure 4(g-h)).

Discussion
Recent evidence has suggested circRNAs have crucial roles in GC tumor occurrence and advancement [19,20]. CircRBMS3 downregulation suppressed GC cell proliferation and invasion by sponging miR-153 and regulating SNAI1 expression [21]. Similarly, circNRIP1 acted as a miR-149-5p sponge to promote GC progression via the AKT1/mTOR pathway [22]. However, hsa_-circ_0000117 function in GC remains unclear. In this study, hsa_circ_0000117 was increased in GC tissue and associated with advanced clinical features, consistent with previous data [12]. We also showed hsa_circ_0000117 suppression significantly decreased in vitro GC cell proliferation and invasion, suggesting hsa_circ_0000117 is an important regulator in GC etiology.

Conclusion
Hsa_circ_0000117 was highly abundant in GC cells and manifested its oncogene functions by sponging miR-337-3p and increasing STAT3 expression. Hence, hsa_circ_0000117 may function as a therapeutic treatment target for GC.