Circ-PRMT5 promotes gastric cancer progression by sponging miR-145 and miR-1304 to upregulate MYC

Abstract Background Gastric cancer (GC) is a global leading source of cancer-associated deaths. Circular RNAs (circRNAs) are a new type of non-coding RNA and promising biomarkers for diagnosis of multiple diseases such as cancer. Methods Circ-PRMT5 expression was validated in 90 GC patient tissues and 6 different GC cells by qRT-PCR. Sublocalization of circ-PRMT5 in GC cells was determined in isolated nuclear and cytoplasmic RNAs. CircInteractome and miRanda were used to predict binding sites between circ-PRMT5 with micRNAs, and micRNAs with target mRNA. The correlation between genes was determined by the Pearson correlation analysis. The molecular mechanism was demonstrated by RNA in vivo precipitation, point mutation, luciferase activity and rescue experiments. Results Circ-PRMT5 expression was significantly higher in GC than in adjacent normal tissues, and GC patients with circ-PRMT5 high expression had shorter survival times. Functionally, circ-PRMT5 silence inhibited GC cell growth and invasion. Mechanism analysis showed that circ-PRMT5 sponged miR-145/miR-1304 to upregulate MYC expression and GC development. Conclusion Our findings demonstrated that circ-PRMT5 function as an oncogene in GC patients by targeting miR-145/miR-1304/MYC axis. High circ-PRMT5 expression may provide a poor prognostic indicator of survival in GC patients and targeting circ-PRMT5/miR-145/miR-1304/MYC axis may be a novel therapeutic strategy for GC.


Introduction
Gastric cancer (GC) remains to be a main hazard to human healthiness, the fourth most universal cancer and the third leading cause of global cancer-associated deaths as said by global cancer statistics [1]. In spite of the application of various improvements in diagnosis and treatment, the GC prognosis continues to be quite poor, with a 5-year overall survival lower than 40% in many countries, due to the tumour invasion, metastasis and recurrence [2,3].
In the past decades, non-coding RNAs (ncRNAs), including microRNA (miRNA) and long non-coding RNA (lncRNA) have been found to be deregulated in GC patients, and have prospective clinical applications [4,5].
CircRNAs are a novel identified and distinct type of ncRNAs derived from introns, exons or intergenic regions, which is characterized by a covalently closed structure without 5 0 caps and 3 0 tails and protein coding ability. CircRNAs also exhibit tissue or cell-specific expression, and are conserved through species owing to their resistance to RNase R [6][7][8]. Compared with linear RNAs, circRNAs are really stable and primarily in the cytoplasm [9,10]. CircRNAs comprise miRNA response elements (MREs) that can be applied to investigate miRNA-specific antagonists. More and more evidences indicate that circRNAs interact with RNA binding proteins (RBPs) and act as miRNA sponges to regulate gene expression [11].
CircRNAs participate in many physiological and pathological procedures, such as transcription, RNA decay, mRNA splicing and translation, and their dysregulation results in abnormality of cellular functions and human illnesses [12].
Especially, the tumourigenesis functions of circRNAs are widely explored recently. It is found that many circRNA types are aberrantly expressed in GC, HCC, oesophageal squamous cancer, CRC, oral cancer and bladder cancer and related with cancer progression [13][14][15][16][17]. Therefore, it is crucial to recognize deregulated circRNAs, discern novel molecular mechanisms and therapeutic targets for GC treatment.
In current study, we aimed to explore the function, molecular mechanisms and clinical significances of the circ-PRMT5 in GC. Our results validated that circ-PRMT5 expression level was dramatically increased in GC tissues. High expression of PRMT5 was associated with the poor prognosis of patients with GC. More importantly, we found that PRMT5 functioned as a sponge of miR-145 and miR-1304 to upregulate the MYC expression and consequently promoted the tumourigenesis of GC.

Methods
Clinical data and patient tissues GC and paired adjacent normal gastric tissue samples from 90 GC patients were acquired from patients undergoing surgery at the Affiliated Hospital of Inner Mongolia University for Nationalities and Nantong Tumor Hospital. All specimens were snap frozen in liquid nitrogen upon collection and stored at À80 C until use. All patients were regularly followed up and overall survival (OS) time was from the date of surgery to the date of death.

Consent to participate and ethics approval
All samples were collected with written informed consents from the patients before the study was initiated. All experiments were approved by the Ethics Committee of Second Affiliated Hospital of Kunming Medical University and informed in accordance with Declaration of Helsinki.  Six GC cell lines (AGS, MKN-28, MKN45, BGC823, MGC803,  SGC7901) and a normal human gastric epithelial cell line (GES-1) were purchased from American Type Culture Collection (ATCC). The cells were cultured in DMEM containing 10% FBS, 100 lg/mL streptomycin, and 100 U/mL penicillin. All cells were cultured in a humidified atmosphere and 5% CO 2 at 37 C.

Analysis and target gene prediction
The sponged miRNAs by the circ-PRMT5 were predicted using the website tool CircInteractome (https://circinteractome.nia.nih.gov/index.html). The target mRNA of miR-145 and miR-1304 was predicted by the website tool miRanda (http://www.microrna.org/).

Quantitative real-time PCR (qRT-PCR)
Total RNA was isolated using RNAiso Plus. RNA purity and concentration were confirmed spectrophotometrically using the NanoDrop ND-1000 (Thermo Fisher Scientific, Wilmington, DE). OD260/OD230 ratios >1.8 but <2.1 were accepted. Reverse transcription was achieved using PrimeScrip TM RT Master Mix, and cDNA amplification was achieved using SYBR Green Premix Ex Taq TM II following the manufacturer's instructions.
MiRNA was isolated using a miRNeasy Mini Kit, reverse transcription was achieved using a miScript II RT Kit, and cDNA amplification was achieved using a miScript SYBR Green PCR Kit following the manufacturer's instructions. qRT-PCR was conducted on an AB7300 thermo recycler (Applied Biosystems, Carlsbad, CA) with primers and TaqMan Universal PCR Master Mix. GAPDH was used as the reference gene for mRNA and circRNA. U6 was used as an internal control for miRNA. Gene expression was quantified by 2 ÀDDCt . The primers are as follows:

Cell proliferation assay
Cell proliferation was detected using a CCK-8 assay kit. Cells were seeded in a 96-well plate (3000 cells/well). Then, 10 lL of CCK-8 reagent was directly added into the culture medium on the indicated time to measure the optional density (OD) at 450 nm after incubation for 2.5 h at 37 C.

Colony formation assay
Cells were plated in a 6-well plate (1 Â 10 3 cells/well) and incubated for 7 d at 37 C with medium change every 2 d. On day 7, cells were fixed in 4% paraformaldehyde and stained with a crystal violet solution. Cell colonies were then counted and analyzed.

Transwell migration and invasion assay
The 24-well Transwell (Corning Costar, Corning, NY, 8.0 lm pore size) insert chamber plates were used for cell migration and invasion assay without (for migration assay) or with (for invasion assay) coated Matrigel in the upper surfaces of the transwell filters. Cells (3 Â 10 5 ) in 200 lL of serum-free medium were loaded to the upper chamber. A total of 500 lL of medium with 10% FBS was added into the lower chamber. After 48 h incubation, the non-migrated and noninvaded cells on the upper side of the chamber were scraped off with a cotton swab, and the cells migrated or invaded to the opposite side were fixed in 4% paraformaldehyde and stained with a crystal violet solution. The cell numbers were then counted and analyzed.

Flow cytometric analysis
To identify cell apoptosis, digested GC cells were washed with cold PBS and re-suspended in binding buffer and fixed, then incubated with APC-AnnexinV and 7-AAD at room temperature for 15 min in the dark. Then, the fluorescence was measured with a BD FACSCalibur flow cytometer (BD) after Annexin V binding buffer was added to the mixture. Cell Quest software (Becton Dickinson, Franklin Lakes, NJ) was used to investigate the cell apoptosis.

Luciferase reporter assay
WT or MUT plasmid of circ-PRMT5 or MYC-3 0 UTR were coinfected with miR-145/miR-145 mimics into GC cells seeded in a 96-well plate. The miR-NC was used as negative control. The cells were collected after 48 h, the Firefly and Renilla luciferase activities were determined with a dual-luciferase reporter assay system. The relative luciferase activity was normalized to Renilla luciferase activity.

Cytoplasmic and nuclear RNA extraction
Cytoplasmic and Nuclear RNAs were extracted using the RNA subcellular isolation kit following the manufacturer's instructions.

Western blotting analysis
Total proteins from CG cells were purified using RIPA lysis buffer. Equal amounts of proteins were separated on 12% SDS-PAGE gels and transferred onto PVDF membranes. The primary antibodies were diluted 1:1000 according to the instructions and incubated overnight at 4 C. HRP rabbit IgG secondary antibodies were incubated at a dilution of 1:2000 at room temperature for 1 h. The membranes were washed three times with TBST and visualized using SuperSignal West Dura Extended Duration Substrate following the manufacturer's instructions.

Statistical analysis
Statistical analyses were performed using SPSS 20.0 (IBM, SPSS, Chicago, IL) and GraphPad Prism (GraphPad, La Jolla, CA). Student's t-test or chi-square test was used to evaluate the statistical significance for two group comparisons. OS curves were evaluated with the Kaplan-Meier method. Pearson correlation analysis was used for the association between genes. p < .05 was statistically significant.

High expression of circ-PRMT5 indicated poor prognosis in GC patients
To investigate the clinical significance of circ-PRMT5 expression in GC patients, we collected 90 pairs of GC and selfmatched adjacent normal tissues from the GC patients. The circ-PRMT5 expression in these 90 GC patients was measured by qRT-PCR. Compared with that in adjacent normal tissues, circ-PRMT5 expression was significantly increased in GC tissues ( Figure 1(A), p < .001). We further confirmed this result by comparing the circ-PRMT5 expression between the GC cells and the normal human gastric epithelial cells. As shown in Figure 1(B), circ-PRMT5 was dramatically higher in all tested GC cells (AGS, MKN-28, MKN-45, BGC-823, MGC803, SGC-7901) than that in the normal human gastric epithelial GES-1 cells (p < .001).
Next, we analyzed whether circ-PRMT5 acts as an independent prognostic factor for GC patients. The GC patients were divided into two groups, circ-PRMT5 high expression group and circ-PRMT5 low expression group, using the median circ-PRMT5 expression level in GC1 tissues as the cutoff value. The association between the circ-PRMT5 expression and the clinicalpathological parameters of GC patients in circ-PRMT5 low expression group (n ¼ 45) and circ-PRMT5 high expression group (n ¼ 45) was analyzed by the chi-square test. We found that the high circ-PRMT5 expression was closely related with the tumour size, the TNM stages, the degree of differentiation, the lymph node metastasis and the distant metastasis regardless of the patient's age and gender, and the difference was statistically significant ( Table 1, p < .001).
In addition, further statistical analyses with the Kaplan-Meier method showed that GC patients with low circ-PRMT5 expression had longer overall survival (OS) times than those with high circ-PRMT5 expression, which indicated that high circ-PRMT5 expression in GC patients was negatively associated with patient prognosis and displayed a poorer OS ( Figure 1, p < .001).

Circ-PRMT5 promoted EMT in GC cells
It is known that EMT is a critical process that induces aggressiveness in cancer cells. Therefore, the transwell migration and invasion assays were used to assess if circ-PRMT5 could enhance EMT phenotypes in CG cells. Our results showed that silence of circ-PRMT5 statistically significantly inhibited the migration (Figure 2(E)) and invasion (Figure 2(F)) abilities of the GC cells, the images (Figure 2(E,F)), left panel) and numbers of the migrated (Figure 2(E), middle and right panels) and invaded (Figure 2(F), middle and right panels) cells were acquired under the microscope.
Circ-PRMT5 was confirmed to sponge miR-145 and miR-1304 in GC cells As previously reported, circRNAs primarily function as miRNA sponges to regulate gene expression and our mRNA fractionation examination revealed that the majority of circ-PRMT5 localized in the cytoplasm (Figure 3(A)). Therefore, we examined the potential miRNAs associated with circ-PRMT5. CircInteractome was used to predict the potential target miRNAs that could bind with the circ-PRMT5 sequence, and miR-145 and miR-1304 were selected as the best potential targets of circ-PRMT5 (Figure 3 MYC was validated as a target gene of miR-145 and miR-1304 in GC cells Next, we used miRanda (http://www.microrna.org/) to predict the target gene of miR-145 and miR-1304 in GC cells; and miR-145 and miR-1304 binding sites were found in the 3 0 non-coding region of the classical oncogene MYC, thus showed the best potential (Figure 4(A)). GC cells co-infected with plasmid containing 3 0 -UTR-WT regions of MYC with miR-145/miR-1304 mimics had significantly less relative luciferase activity than the controls (miR-NC), while mutation of the potential miR-145/miR-1304 binding sites in the MYC 3 0 -UTR abolished this effect (Figure 4

Circ-PRMT5 sponged miR-145 and miR-1304 in GC cells to upregulate MYC
To determine whether circ-PRMT5 sponges miR-145 and miR-1304 to regulate MYC expression, we first measured MYC expression levels by qRT-PCR and Western Blotting. Our results found that circ-PRMT5 silence considerably downregulated MYC expression at both the mRNA and protein levels in GC cells, and these effects could be partially reversed by coinfection of the miR-145 or miR-1304 inhibitor ( Figure 5(A,B), p < .001). Pearson correlation analysis of the association between circ-PRMT5 and MYC expression in GC tissues from 90 GC patients (same samples as in Figure 1(C)) showed that circ-PRMT5 expression was statistically significantly and positively correlated with MYC expression (   circRNAs are rich in cells and show cell/tissue-specific expression, suggesting circRNAs may be vital in regulating gene transcription and expression, thus the essential functions of circRNAs in various biological and pathological procedures [19]. In recent years, researchers have focused extensive attention on circRNAs due to their important functions in many cellular mechanisms. CircRNAs also have the potential to be as encouraging biomarkers for diagnosis of many diseases such as cancer [20]. Therefore, the circRNA function in carcinogenesis has recently attracted researcher's attention. The gene expression regulation varies of CircRNAs in different types or different stages of cancer [19]. CircRNAs are well known to act as miRNA sponges to form the circRNA-miRNA-mRNA axis [8,13,14,16,[21][22][23]. Circ-PRMT5 is a newly identified circRNA, plays critical role in promoting cell EMT of urothelial carcinoma of the bladder (UCB) and may serve as an exploitable therapeutic target for UCB patients. However, the expression and function of circ-PRMT5 in GC remain unknown.
Here, we investigated the GC tissues and paired adjacent normal tissues from 90 GC patients and found that circ-PRMT5 expression was significantly higher in the GC tissues than in the matched adjacent normal tissues. Increased circ-PRMT5 expression was significantly correlated with the poor prognosis in GC patients. GC patients with low circ-PRMT5 expression had longer survival times than those with high circ-PRMT5 expression. Therefore, it was illustrated that circ-PRMT5 could serve as a biomarker to predict the survival in GC patients.
MiRNA, a key component of the ncRNA family with the length of less than 22 nucleotides, plays multifaceted roles in controlling cellular functions by degradation of the target genes. Large evidences have demonstrated that miRNA mediated growth, oxidative response, signalling transduction and cell apoptosis.
In this study, CircInteractome was used to predict the potential miRNAs binding to circ-PRMT5, and miR-145 and miR-1304 were found to be the best prospective candidates. MiR-145 has been reported as a tumour suppressor in many cancer types, such as colorectal cancer [24], cervical carcinoma [25], prostate cancer [26], oesophageal squamous cell carcinoma [27] and gastric cancer [28]. MiR-1304 has also been reported being sponged by circRNAs to promote cancer progression [29,30]. However, the possible mechanism might need additional exploration. In current study, we also found that miR-145 and miR-1304 were down-regulated in the GC tissues and acted as an oncosuppressor in GC patents. MiR-145 and miR-1304 expression was statistically significantly up-regulated when circ-PRMT5 was silenced in AGS and MKN-28 cells; overexpression of miR-145 or miR-1304 inhibited luciferase activity of the wild-type circ-PRMT5 reporter gene in AGS and MKN-28 cells, while this effect was disappeared when the predicted binding sites of circ-PRMT5 were mutated; circ-PRMT5 probe could pull down more miR-145/miR-1304 than the oligo probe; Pearson correlation coefficient analysis showed a negative correlation between the circ-PRMT5 and the miR-145/miR-1304 expression.
MYC is an oncogene involved in cell cycle regulation, cell adhesion, cell growth arrest, metabolism, protein synthesis, ribosome biogenesis and mitochondrial function, which has been described as a crucial element of numerous human carcinogenesis processes, such as GC [31][32][33][34]. MiRanda database showed binding sites of miR-145 and miR-1304 in the 3 0 noncoding region of MYC. Our further investigate showed that mutation of the potential miR-145/miR-1304 binding sites in the MYC 3 0 -UTR abolished MYC induced luciferase activity decrease; furthermore, knockdown/overexpression of miR-145 or miR-1304 increased/decreased MYC mRNA and protein expressions; the mRNA expression levels of MYC in GC tissues was also increased; Pearson correlation coefficient analysis showed a negative correlation between the miR-145/miR-1304 and the MYC expression, suggesting MYC could directly interacted with miR-145/miR-1304 to promote the progress of GC.
In summary, our findings revealed that circ-PRMT5 expression was significantly increased in GC patients and cells. High circ-PRMT5 expression was associated with a poor prognosis in GC patients. Functionally, circ-PRMT5 promoted GC cell growth, clone formation, migration and invasion, while inhibited apoptosis by sponging miR-145 and miR-1304 in GC cells to upregulate MYC expression, thus contributed to the GC tumourigenesis. Therefore, circ-PRMT5 may be used as a potential prognostic predictor and therapeutic target in GC patients.