EIF4A3-Mediated circ_0008126 Inhibits the Progression and Metastasis of Gastric Cancer by Modulating the APC/β-Catenin Pathway

Abstract

Background: Mounting evidence exhibits circRNAs as critical regulators in the progression of many tumors. The regulatory function and potential mechanism by which circ_0008126 in gastric cancer (GC) is unknown.

Methods: To validate and analyze the expression levels and clinical values of circ_0008126 in GC patients, the biological phenotypes of circ_0008126 in GC were investigated in vitro and in vivo. The roles and effects of circ_0008126 on miR-502-5p, EIF4A3, and APC in GC cells were explored using rescue experiment, RNA stability assay, RNA pull-down, dual-luciferase reporter, RNA immunoprecipitation (RIP), RNA FISH, immunofluorescence (IF), and TOP/Flash and FOP/Flash assays.

Results: Circ_0008126 expression levels were prominently down-regulated in GC tissues and cells. Importantly, low expression of circ_0008126 was relevant to the more lymphatic metastasis, advanced TNM stage, and poor survival period in patients with GC. Functionally, circ_0008126 inhibited GC cell proliferative activity, metastatic ability, and epithelial-mesenchymal transition (EMT) in vitro and vivo. Mechanistically, we verified that EIF4A3 can mediate the formation of circ_0008126, and circ_0008126 could competitively bind miR-502-5p and alleviate its role and effect on APC, thus inactivating the β-catenin pathway in GC. Additionally, circ_0008126 was determined to increase the stability of APC mRNA by interacting with cytoplasmic EIF4A3 protein and then enhancing the APC expression.

Conclusions: These data demonstrate that EIF4A3-mediated circ_0008126 could regulate the APC expression and inactivate the β-catenin pathway partly by binding to miR-502-5p and EIF4A3, thus inhibiting the tumorigenesis and development of GC.

Keywords: APC; EIF4A3; circ_0008126; gastric cancer; miR-502-5p.

Figure 1

Circ_0008126 is verified and characterized in GC samples. (A). The tissue-specific circRNAs in five paired human GC tissues and noncancerous tissues were validated by using a clustered heat map. Hsa_circ_0008126 is marked with a red box. (B,C). The expression of circ_0008126 in 35 pairs of GC and adjacent tissue samples was detected by using qRT-PCR. (D). The expression of circ_0008126 between GC and normal gastric epithelial cell lines was assessed by using qRT-PCR. (E). The expression and location of circ_0008126 in GC tissues and adjacent tissues were evaluated by using FISH assays. (F). Circ_0008126 was formed by head-to-tail splicing of the 2 and 3 exons of STX12 using Sanger sequencing. (G). STX12 levels were validated in different stages of GC tissues from the TCGA database. (H). The correlation of circ_0008126 expression levels with the prognosis of GC patients was analyzed using the Kaplan–Meier plotter. (I,J). The expression of circ_0008126 and STX12 in MKN-45 cells (I) and MGC-803 (J) was explored by using qRT-PCR after treatment with RNaseR. (K,L). The expression of circ_0008126 and STX12 at different time points in MKN-45 cells (K) and MGC-803 (L) was determined by using qRT-PCR after treatment with actinomycin D. (M). The expression and location of circ_0008126 in the cytoplasm of GC cells were confirmed by using FISH assays. Values are shown as the mean ± standard error. * p < 0.05, ** p < 0.01.

Figure 2

EIF4A3-induced circ_0008126 suppresses the proliferation, metastasis, and EMT ability of GC cells. (A). The binding sites for EIF4A3 in the flanking sequences of the STX12 pre-mRNA transcript were predicted using CircInteractome. (B). The putative binding sites of EIF4A3 in STX12 pre-mRNA (circ_0008126 precursor RNA flanking region) were demonstrated via the RIP assays. (C). Western blot analysis of A1, A2, A3, and A4 protein with MS2-RIP assay. Vector and GAPDH were used as negative controls. (D,E). The expression of circ_0008126 was detected in MGC-803 cells (D) and AGS cells (E) after EIF4A3 down-regulation or up-regulation by using qRT-PCR. (F). The expression of circ_0008126 in MGC-803 cells stably transfected with LV5 lentiviral-circ_0008126 was detected by using qRT-PCR. (G). The expression of circ_0008126 in AGS cells treated with siRNA was examined by using qRT-PCR. (H). The proliferation of MGC-803 cells was measured by using CCK-8 assays after the overexpression of circ_0008126. (I,J). The DNA synthesis of MGC-803 cells was tested via EdU assays after the overexpression of circ_0008126. (K). The proliferation of AGS cells was explored using CCK-8 assays after the silencing of circ_0008126. (L,M). The DNA synthesis of AGS cells was observed using EdU assays after the silencing of circ_0008126. Values are shown as the mean ± standard error. * p < 0.05, ** p < 0.01.

Figure 3

EIF4A3-induced circ_0008126 suppresses the proliferation, metastasis, and EMT ability of GC cells. (A,B). Colony formation ability was verified in MGC-803 cells after the overexpression (A) and knockdown (B) of circ_0008126. (C,D). Wound healing assays examined the migration ability of cells after the overexpression (C) and knockdown (D) of circ_0008126. (E,F). Transwell invasion assays verified the invasion ability of cells after the overexpression (E) and knockdown (F) of circ_0008126. (G). The expression of vimentin and E-cadherin after circ_0008126 overexpression or knockdown was validated by qRT-PCR assays. (H,I). The fluorescence intensities of vimentin and E-cadherin after circ_0008126 overexpression or knockdown were detected by using IF assays. Values are shown as the mean ± standard error. * p < 0.05, ** p < 0.01.

Figure 4

Circ_0008126 directly interacts with miR-502-5p in GC. (A,B). The expression changes for candidate miRNAs after circ_0008126 overexpression or knockdown were determined using qRT-PCR. (C). The expression levels of miR-502-5p in AGS, MGC-803, MKN-45, and GES-1 were examined using qRT-PCR. (D) The differential expression of miR-502-5p between 25 paired GC and adjacent tissue samples was detected using qRT-PCR. (E). Pearson correlation analysis indicated circ_0008126 and miR-502-5p expression levels in 25 GC tissues. (F). Schematic illustration of circ_0008126 -WT and circ_0008126 -MUT luciferase reporter vectors. (G,H). The relative luciferase activity of the circ_0008126-WT and circ_0008126-MUT were observed in MGC-803 (G) and AGS (H) cells. (I). The expression levels of miR-502-5p in MGC-803 and AGS cells were pulled down by circ_0008126 or oligo probe. (J). The co-localization of circ_0008126 and miR-502-5p in GC cells was validated by FISH assays. (K,L). The expression levels of circ_0008126 in each group were tested using qRT-PCR. MGC-803 cells were transfected with OE-NC, OE-circ, OE-circ + mimics-NC, or OE-circ + miR-502-5p mimics (K). AGS cells were transfected with Si-NC, Si-circ, Si-circ+ inhibitors-NC, or Si-circ + miR inhibitors (L). (M). The expression levels of miR-502-5p in OE-NC, OE-circ, OE-circ+ mimics-NC, or OE-circ + miR-502-5p mimics groups were detected using qRT-PCR. (N). The expression levels of miR-502-5p in Si-NC, Si-circ, Si-circ + inhibitors-NC, or Si-circ + miR inhibitors groups were verified using qRT-PCR (N). Values are shown as the mean ± standard error. * p < 0.05, ** p < 0.01.

Figure 5

Circ_0008126 reduced GC progression via targeting miR-502-5p. (A). The proliferation ability of the OE-NC, OE-circ, OE-circ + mimics-NC, or OE-circ + miR-502-5p mimics groups in MGC-803 cells was measured using CCK-8 assay. (B). The proliferation ability of the Si-NC, Si-circ, Si-circ+ inhibitors-NC, or Si-circ + miR inhibitors groups in AGS cells was determined using CCK-8 assays. (C,I). The DNA synthesis of the OE-NC, OE-circ, OE-circ+ mimics-NC, or OE-circ + miR-502-5p mimics groups in MGC-803 cells were observed using EdU assays. (D,J). The DNA synthesis of the Si-NC, Si-circ, Si-circ+ inhibitors-NC, or Si-circ + miR inhibitors groups in AGS cells was examined using EdU assays. (E,K). The migration ability of the OE-NC, OE-circ, OE-circ+ mimics-NC, or OE-circ + miR-502-5p mimics groups in MGC-803 cells were verified using wound healing assays. (F,L). The migration ability of the Si-NC, Si-circ, Si-circ+ inhibitors-NC, or Si-circ + miR inhibitors groups in AGS cells was tested using wound healing assays. (G,M). The invasion ability of the OE-NC, OE-circ, OE-circ+ mimics-NC, or OE-circ + miR-502-5p mimics groups in MGC-803 cells were validated using transwell invasion assays. (H,N). The invasion ability of the Si-NC, Si-circ, Si-circ+ inhibitors-NC, or Si-circ + miR inhibitors groups in AGS cells was confirmed by the transwell invasion assays. Values are shown as the mean ± standard error. * p < 0.05.

Figure 6

Circ_0008126 inhibits GC progression through miR-502-5p/APC/β-catenin signaling. (A). The miRWalk, Target Scan, miRDB, mirDIP, and microT-CDS databases predicted the potential target molecules of miR-502-5p. (B). The miR-502-5p predicted target molecules were subjected to KEGG enrichment analysis, and the top 20 signaling pathways with significant enrichment were labeled. (C). Schematic representation of the potential binding sites of circ_0008126 to miR-502-5p and miR-502-5p to the 3′UTR of APC. (D). The expression levels of APC after overexpression of circ_0008126 and miR-502-5p inhibitors were determined using qRT-PCR. (E). The expression levels of APC after the silencing of circ_0008126 and miR-502-5p mimics were detected using qRT-PCR. (F). The expression levels of APC after the overexpression or silencing of circ_0008126 were verified using WB. (G,H). The fluorescence intensity of APC after the overexpression or silencing of circ_0008126 was tested using the IF assays. (I). Schematic illustration of APC-WT and APC-MUT luciferase reporter vectors. (J). The relative luciferase activity of the APC-WT and APC-MUT were examined in AGS (I) and MGC-803 (J) cells. (K). The expression of APC, β-catenin and intranuclear β-catenin after OE-NC, OE-circ, OE-circ+ mimics-NC, OE-circ + miR-502-5p mimics groups and Si-NC, Si-circ, Si-circ+ inhibitors-NC, or Si-circ + miR inhibitors groups were determined using WB. (L,M). The dual-luciferase assays showed the effect on TOP/FOP reporter activity in GC cells after transfection with mimics NC, miR-502-5p mimics, inhibitors NC, miR-502-5p inhibitors, circ_0008126 overexpression or siRNA, and miR-502-5p mimics + circ_0008126 overexpression or miR-502-5p inhibitors + siRNA. Values are shown as the mean ± standard error. * p < 0.05.

Figure 7

Circ_0008126 interacts with the cytoplasmic EIF4A3 and enhances the stability of APC in GC. (A). EIF4A3 protein was pulled down by biotin-labeled circ_0008126 antisense or sense probes, as detected by WB analysis. (B). The association of EIF4A3 and circ_0008126 (upper) was shown via RIP assays. IgG was used as the negative control. (C). The subcellular location of circ_0008126 and EIF4A3 in GC cells was evaluated using FISH and IF assays. Nuclei were stained with DAPI. (D). The expression of EIF4A3 in GC cells transfected with oe-NC or EIF4A3 overexpression or Si-NC or Si-EIF4A3 was validated using qRT-PCR assays. (E). The expression of EIF4A3 in GC cells transfected with oe-NC or EIF4A3 overexpression or Si-NC or Si-EIF4A3 was validated using WB analysis. (F,G). The expression of APC in GC cells transfected with oe-NC or EIF4A3 overexpression or Si-NC or Si-EIF4A3 was verified using qRT-PCR assays. (H). The expression of APC in GC cells transfected with oe-NC or EIF4A3 overexpression or Si-NC or Si-EIF4A3 was determined via WB analysis. (I). The interaction probabilities of EIF4A3 with the 3′UTR of APC mRNA were predicted by using the RNA-Protein interaction prediction (RPISeq) website. Predictions with probabilities > 0.5 were considered to represent that the corresponding RNA and protein are likely to interact. (J). The association of EIF4A3 and APC was tested using RIP assays. IgG was used as the negative control. (K,L). RNA stability assays revealed the remaining levels of APC mRNA in GC cells transfected with the overexpression of EIF4A3 (K) or siRNA specifically targeting EIF4A3 (L). (M,N). The luciferase activities of APC 3′UTR were measured after the overexpression (M) or knockdown (N) of EIF4A3 in GC cells, respectively. (O). The luciferase activities of APC 3′UTR were measured in GC cells transfected with si-NC, si-EIF4A3 #1, or si-EIF4A3 #2, and those co-transfected with OE-NC or OE-circ_0008126. (P). The expression of APC in GC cells transfected with si-NC, si-EIF4A3, OE-NC, or OE-EIF4A3 and those co-transfected with si-NC, si-circ_0008126, OE-NC or OE-circ_0008126. Values are shown as the mean ± standard error. * p < 0.05, ** p < 0.01.

Figure 8

Circ_0008126 suppresses GC tumorigenesis and metastasis in vivo. (A). The subcutaneous tumor tissues in the circ_0008126 overexpression and OE-NC groups were excised and photographed. (B,C). The volume and weight of subcutaneous tumor tissues in the circ_0008126 overexpression and the OE-NC groups were regularly observed. (D). The expression levels of E-cadherin, vimentin, Ki-67, APC, and PCNA in the circ_0008126 overexpression and the OE-NC groups were verified by the IHC assays. (E). The expression levels of circ_0008126, miR-502-5p, APC, E-cadherin, vimentin, PCNA, and Ki-67 in the circ_0008126 overexpression and the OE-NC groups were detected by qRT-PCR. (F,G). Representative images of lung metastatic nodules in the circ_0008126 overexpression and the OE-NC groups are indicated by arrows. (H). The H&E staining of lung metastatic nodules in the circ_0008126 overexpression and the OE-NC groups was observed. (I,J). Representative images of liver metastatic nodules in the circ_0008126 overexpression and the OE-NC groups are indicated by arrows. (K). The H&E staining of liver metastatic nodules in the circ_0008126 overexpression and the OE-NC groups was observed. (L). The schematic diagram describes the mechanism by which EIF4A3-mediated circ_0008126 inhibits the proliferation and metastasis of GC through the regulation of APC by sponging miR-502-5p and interacting with EIF4A3, respectively. Values are shown as the mean ± standard error. * p < 0.05, ** p < 0.01.