circRAPGEF5 Contributes to Papillary Thyroid Proliferation and Metastatis by Regulation miR-198/FGFR1.

The circular RNA RAPGEF5 (circRAPGEF5) is generated from five exons of the RAPGEF5 gene and abnormal expression in papillary thyroid cancer (PTC). However, whether circRAPGEF5 plays a role in PTC tumorigenesis remains unclear. The aim of the present study was to investigate the role of circRAPGEF5 in PTC. The results showed that circRAPGEF5 was upregulated in PTC tissues and cell lines. circRAPGEF5 knockdown inhibited cell proliferation, migration, and invasion in vitro; and circRAPGEF5 silencing downregulated fibroblast growth factor receptor 1 (FGFR1) expression by "sponging" miR-198, suppressing the aggressive biological behaviors of PTC. Luciferase reporter assays confirmed that circRAPGEF5 interacted with miR-198 and that miR-198 interacted with the 3' UTR of FGFR1 to downregulate its expression. Xenograft experiments confirmed that circRAPGEF5 knockdown suppressed FGFR1-mediated tumor growth by promoting miR-198 expression. circRAPGEF5 acts as a tumor promoter via a novel circRAPGEF5/miR-198/FGFR1 axis, providing a potential biomarker and therapeutic target for the management of PTC.

Introduction

The incidence of thyroid cancer has increased rapidly (>5% per year in both men and women) over the past several decades, especially papillary thyroid carcinoma (PTC).1, 2 Despite advances in the clinical treatment of thyroid cancer, including surgery and/or radiotherapy or chemotherapy, it remains associated with high rates of recurrence, indicating that the treatment is not sufficiently effective.3, 4 The aim of the present study was to identify novel targets and diagnostic biomarkers for thyroid cancer, particularly for PTC, and investigate the underlying molecular mechanism.

Circular RNA (circRNA) is a novel class of endogenous non-coding RNAs formed by a covalently closed loop. circRNAs are characterized by covalently closed loop structures with neither 5′-to-3′ polarity nor a polyadenylated tail. They are highly stable in vivo compared with their linear counterparts. Two mechanisms, “exon skipping” and “direct back-splicing,” have been proposed to form mammalian exonic circRNA.7, 8, 9 Increasing studies have found that abnormal expression of circRNAs is correlated with the progression of several cancers including hepatocellular carcinoma, non-small-cell lung cancer, osteosarcoma, and gastric cancer. Previous studies also found that abnormal expression of circRNAs is associated with the progression of PTC.14, 15 Previously reported sequencing analysis showed that the expression of hsa_circ_0001681 (circRNA RAPGEF5 or circRAPGEF5) in PTC was aberrantly increased; however, whether circRAPGEF5 plays a role in the progression of PTC remains unclear.

MicroRNAs (miRNAs) are an evolutionarily conserved group of small regulatory non-coding RNAs that are involved in various biological functions.16, 17 circRNAs act as “miRNA sponges” to regulate gene expression by suppressing miRNA activity. However, whether circRAPGEF5 acts as a “miRNA sponge” in PTC remains largely unknown.

The present study examined the expression of circRAPGEF5 in PTC and explored the biological roles of circRAPGEF5. The results showed that circRAPGEF5 was upregulated in PTC tissues and cell lines, and its upregulation promoted cell proliferation and migration. The present data provide novel evidence that may be useful for the development of therapeutic strategies against PTC.

Results

circRAPGEF5 Is Significantly Upregulated in PTC Tissues and Relative PTC Cell Lines

A previous high-throughput microarray assay showed that circRNAs are abnormally expressed in PTC, although the specific regulatory mechanism remains unclear. Analysis of the sequencing results showed that hsa_circ_0001681 was aberrantly expressed. Bioinformatics analysis demonstrated that hsa_circ_0001681 is cyclization with 5 exon, which derived from the RAPGEF5 gene. The genomic length of RAPGEF5 is 26,863 bp, and the spliced mature sequence length is 516 bp. It is located at chr7: 22330793–22357656. Therefore, hsa_circ_0001681 was named circRNA RAPGEF5 (circRAPGEF5) (Figure 1A). To determine whether circRAPGEF5 expression was altered in PTC, 30 pairs of PTC tissues and matched adjacent normal tissues were analyzed by qRT-PCR. The results showed that circRAPGEF5 expression was significantly higher in PTC tissues than in matched adjacent normal tissues (Figure 1B). The expression of circRAPGEF5 was higher in the PTC cell lines BCPAP, KTC-1, and K1 than in normal PTC cells (e.g., Nthy-ori 3-1) (Figure 1C). circRAPGEF5 expression was highest in BCPAP cells, and this cell line was, therefore, selected for subsequent experiments. The results of fluorescence in situ hybridization (FISH) showed that circRAPGEF5 was predominately localized in the cytoplasm (Figure 1D). Taken together, these results suggest that circRAPGEF5 plays a role in the progression of PTC.

Figure 1

The Expression of circRAPGEF5 Was Increased in Both PTC Tissues and PTC Cell Lines

(A) The genomic loci of the RAPGEF5 gene and circRAPGEF5. Arrow indicates back-splicing. (B) Analysis of 30 paired tumor tissue samples and adjacent non-tumor tissue samples showed that the expression of circRAPGEF5 was increased in PTC tissues compared with adjacent normal tissues (n = 30). Data are indicated as the mean ± SD. ***p < 0.001 versus normal tissues. (C) The expression of circRAPGEF5 in PTC cells (BCPAP, KTC-1, and K1 cells) and normal PT cells (Nthy-ori 3-1) was detected by qRT-PCR. Data are indicated as the mean ± SD. ***p < 0.001 versus normal PTC cells. (D) Fluorescence in situ hybridization (FISH) assay was conducted to determine the subcellular localization of circRAPGEF5. Green indicates DAPI, and red indicates circRAPGEF5. Scale bars, 30 μm.

Knockdown of circRAPGEF5 Suppresses BCPAP Cell Proliferation, Invasion, and Migration

To examine the role of circRAPGEF5 in PTC, BCPAP cells were transfected with small interfering RNA (siRNA) against circRAPGEF5 (sicircRNA or sicircRAPGEF5). The results of qRT-PCR showed that the expression of circRAPGEF5 was downregulated in BCPAP cells compared with the negative control (NC) group or control group after transfection with sicircRNA for 48 h (Figure 2A). Western blot analysis showed that fibroblast growth factor receptor 1 (FGFR1) was downregulated in response to circRAPGEF5 silencing (Figure 2B). Increasing evidence shows that FGFR kinases are promising therapeutic targets in multiple cancers. Although several FGFR kinase inhibitors have entered clinical trials, the regulatory mechanism remains unclear,20, 21 We speculated that FGFR1 may be regulated by circRAPGEF5.

Figure 2

Knockdown of circRAPGEF5 Suppresses BCPAP Cell Proliferation, Invasion, and Migration

(A) qRT-PCR detection of the expression of circRAPGEF5 after transfection with siRNA against circRAPGEF5 or negative control (NC) for 48 h. Data are indicated as the mean ± SD. n = 3. ***p < 0.001 versus NC. (B) Western blot detection of the expression of FGFR1. (C) Cell proliferation was measured by the CCK-8 assay. Data are indicated as the mean ± SD. n = 3. ***p < 0.001 versus NC. (D and E) Wound healing assays showed that downregulation of circRAPGEF5 prevents the closing of scratch wounds (D). Data are indicated as the mean ± SD (E). ***p < 0.001 versus NC. Scale bars, 100 μm. (F and G) Cell invasion was determined in BCPAP cells using the Transwell assay (F). Data are indicated as the mean ± SD (G). n = 3. ***p < 0.001 versus NC. Scale bars, 30 μm.

Cell Counting Kit-8 (CCK-8) detection showed that downregulation of circRAPGEF5 significantly decreased the proliferation of BCPAP cells compared with the NC group (Figure 2C). The results of the wound healing assay showed that circRAPGEF5 silencing resulted in a slower closing of scratch wounds compared with that in the control group (Figures 2D and 2E). Transwell invasion assays indicated that the invasive abilities of BCPAP cells were decreased in response to circRAPGEF5 silencing (Figures 2F and 2G). These results suggested that circRAPGEF5 knockdown inhibited PTC cell growth, migration, and invasion, but the exact mechanism remains unclear.

We also used K1 cells to study the effect of circRAPGEF5 on PTC cell proliferation, invasion, and migration. The results indicated that downregulation of circRAPGEF5 also suppressed cell proliferation with CCK-8 (Figure S1A) and a cloning formation assay (Figures S1B and S1C). circRAPGEF5 silencing was also shown to suppress cell migration (Figures S1D and S1E) and invasion (Figures S1F and S1G) with Transwell assays and wound healing assays.

miR-198 Is a Target of circRAPGEF5

Bioinformatics analysis identified miR-198 as a potential target of circRAPGEF5 (Figure 3A), and this was confirmed using the bifluorescein reporter experiment. After co-transfection with miR-198 mimics or the wild-type or mutated (Mut) type of circRAPGEF5, the luciferase activity of wild-type circRAPGEF5 was dramatically inhibited by miR-198 overexpression; however, the alterations of the luciferase activity were abolished by the mutation in the predicted miR-198 binding site (Figure 3B). To determine whether miR-198 plays a role in PTC progression, BCPAP cells were transfected with miR-198 inhibitor or circRAPGEF5 silencing vector alone or in combination. The results showed that miR-198 was upregulated in response to circRAPGEF5 silencing, suggesting that circRAPGEF5 targets and suppresses miR-198 expression under normal conditions. Treatment with a miR-198 inhibitor suppressed miR-198 expression, even in the circRAPGEF5 silencing condition (Figure 4A). Western blot detection showed that miR-198 inhibitor treatment reversed the inhibition of FGFR1 expression induced by circRAPGEF5 silencing (Figures 4B and 4C). The CCK-8 assay showed that miR-198 inhibition rescued circRAPGEF5-silencing-induced suppression of cell proliferation (Figure 4D). The wound healing assay showed that miR-198 inhibition rescued the circRAPGEF5-silencing-induced delay in the closing of scratch wounds (Figures 4E and 4F). Transwell invasion assays indicated that the effect of circRAPGEF5 silencing on decreasing the invasive ability of BCPAP cells was restored by miR-198 inhibition (Figures 4G and 4H). These results suggested that miR-198 was the target of circRAPGEF5 and that miR-198 exerted antitumor effects.

Figure 3

miR-198 Is a Target of circRAPGEF5

(A) Bioinformatics analysis predicted that the sequence of miR-198 matched the sequence of circRAPGEF5. The sequences of circRAPGEF5, including the wild-type (WT) and mutated (MUT) sequences (containing the miR-198 target sites), were cloned into the pmirGL0 vector. (B) The relative luciferase activity was determined at 48 h after transfection with circRAPGEF5 WT-Mut sequences and miR-198 mimics-NC in HEK293T cells. Data are indicated as the mean ± SD. n = 3. ***p < 0.001.

Figure 4

Downregulation of miR-198 Reversed the circRAPGEF5 Silencing-Induced Inhibition of Cell Proliferation, Invasion, and Migration in PTC Cells

(A) qRT-PCR detection of the expression of miR-198 after transfection with siRNA against circRAPGEF5 and miR-198 inhibitor alone or in combination. (B and C) Western blot detection of the expression of FGFR1 (B). The relative protein levels were analyzed, and data are indicated as the mean ± SD (C). n = 3. ***p < 0.001 versus NC; ###p < 0.001 versus sicircRNA. (D) Cell proliferation was measured by the CCK-8 assay. Data are indicated as the mean ± SD. n = 3. ***p < 0.001 versus NC; ###p < 0.001 versus sicircRNA. (E and F) Wound healing assay showed that downregulation of circRAPGEF5 prevents the closing of scratch wounds (E). Data are indicated as the mean ± SD (F). ***p < 0.001 versus NC. Scale bars, 100 μm. (G and H) Cell invasion was determined in BCPAP cells using the Transwell assay (G). Data are indicated as the mean ± SD (H). n = 3. ***p < 0.001 versus NC; ###p < 0.001 versus sicircRNA. Scale bars, 30 μm.

FGFR1 Is a Target of miR-198

Bioinformatics analysis identified FGFR1 as a potential target of miR-198, and this was confirmed by bifluorescein reporter experiments (Figure 5A). The relative luciferase activity was decreased in HEK293T cells transfected with the wild-type FGFR1 3′ UTR and miR-198 mimics for 48 h (Figure 5B). To determine whether FGFR1 plays a role in PTC progression, BCPAP cells were transfected with miR-198 mimics or FGFR1 overexpression vector alone or in combination. The results showed that miR-198 was upregulated in response to miR-198 mimics, similar to the effect of co-transfection with the FGFR1 overexpression vector. This suggested that exogenous expression of FGFR1 had no effect on miR-198 expression under normal conditions (Figure 6A). Western blot detection showed that miR-198 upregulation suppressed FGFR1 expression. However, the expression level of FGFR1 was rescued by transfection with FGFR1, and the expression level of FGFR1 was increased significantly compared with that in the NC group (Figures 6B and 6C). Because exogenous FGFR1 cannot interact with the 3′ UTR, miR-198 could not interact and cause the degradation of FGFR1 at the mRNA level. CCK-8 analysis results showed that overexpression of FGFR1 rescued the miR-198-induced suppression of cell proliferation (Figure 6D). The wound healing and Transwell invasion assays showed that overexpression of FGFR1 rescued the miR-198-induced suppression of BCPAP cell migration and invasion (Figures 6E–6H). These results suggested that FGFR1 was the target of miR-198.

Figure 5

FGFR1 Is a Target of miR-198

(A) Bioinformatics analysis predicted that the sequence of miR-198 matched the sequence of the 3′ UTR of FGFR1. The sequences of the FGFR1 3′ UTR, including the wild-type (WT) and mutated (MUT) sequences (containing the miR-198 target sites), were cloned into the pmirGL0 vector. (B) The relative luciferase activity was determined at 48 h after transfection with FGFR1 3′ UTR WT-Mut sequences and miR-198 mimics-NC in HEK293T cells. Data are indicated as the mean ± SD. n = 3. ***p < 0.001.

Figure 6

FGFR1 Overexpression Rescues the Effect of miR-198 on Suppressing Cell Proliferation, Invasion, and Migration in PTC Cells

(A) qRT-PCR detection of the expression of miR-198 after transfection with miR-198 mimics and FGFR1 overexpression vector alone or in combination. n = 3. Data are indicated as the mean ± SD. ***p < 0.001 versus NC; ###p < 0.001 versus mimic. (B and C) Western blot detection of the expression of FGFR1 (B). The relative protein levels were analyzed, and data are indicated as the mean ± SD (C). n = 3. ***p < 0.001 versus NC; ###p < 0.001 versus mimic. (D) Cell proliferation was measured by the CCK-8 assay. Data are indicated as the mean ± SD. n = 3. ***p < 0.001 versus NC; ###p < 0.001 versus mimic. (E and F) Wound healing assays showed that downregulation of circRAPGEF5 prevents the closing of scratch wounds (E). Data are indicated as the mean ± SD (F). n = 3. ***p < 0.001 versus NC. Scale bars, 100 μm. (G and H) Cell migration and invasion were determined in BCPAP cells using the Transwell assay (G). Data are indicated as the mean ± SD (H). n = 3. ***p < 0.001 versus NC; ###p < 0.001 versus mimic. Scale bars, 30 μm.

Silencing of circRAPGEF5 Suppresses Tumor Growth in Xenograft Mice

To further assess the effect of circRAPGEF5 silencing in vivo, we established xenograft mouse models by subcutaneous injection of equal amounts of BCPAP cells (n = 6 for each group). The results showed that circRAPGEF5 silencing significantly decreased tumor volume compared with the wild-type BCPAP group (Figures 7A and 7B). qRT-PCR detection showed that the expression of miR-198 was upregulated in the circRAPGEF5 silencing group (Figure 7C). However, FGFR1 was downregulated in response to circRAPGEF5 knockdown (Figures 7D and 7E).

Figure 7

Silencing of circRAPGEF5 Suppressed Tumor Formation in Xenograft Mice

(A) Representative image of nude mice with tumors formed by injection with BCPAP cells (n = 6). (B) Tumor volume of mice was measured every week. Data are indicated as the mean ± SD. ***p < 0.001 versus NC. (C) qRT-PCR detection of the expression of miR-198. Data are indicated as the mean ± SD. n = 6. ***p < 0.001 versus NC. (D and E) Western blot detection of the expression of FGFR1 (D). Data are indicated as the mean ± SD (E). n = 6. ***p < 0.001 versus NC.

Discussion

Increasing evidence indicates that circRNAs play crucial regulatory roles in different types of tumors, including hepatocellular carcinoma, breast cancer, gastric cancer, pancreatic cancer, osteosarcoma, and thyroid cancer. However, underlying mechanisms remain largely unclear. In the present study, we showed that circRAPGEF5 was significantly upregulated in PTC tissues and cell lines. Downregulation of circRAPGEF5 suppressed cell growth, migration, and invasion. In addition, circRAPGEF5 silencing upregulated the expression of miR-198, which plays an important role in tumorigenesis. Previous studies showed that miR-198 induces apoptosis and inhibits the proliferation, migration, and invasion of gastric cancer cells by downregulating TLR4 expression. miR-198 functions as a tumor suppressor in breast cancer by targeting CUB domain-containing protein 1. However, the function of miR-198 in thyroid cancer remains unclear.

In the present study, overexpression of miR-198 suppressed cell growth, migration, and invasion. Downregulation of miR-198 with a miR-198-specific inhibitor reversed the circRAPGEF5 silencing-induced inhibition of cell growth, migration, and invasion. These findings suggested that circRAPGEF5 acts as “miRNA sponge,” interacting with miR-198 and suppressing the activation of miR-198, as confirmed by the luciferase reporter assay.

Bioinformatics analysis showed that miR-198 interacted with the 3′ UTR of FGFR1 and suppressed FGFR1 expression at the post-transcriptional level, which was confirmed by the results of the luciferase reporter assay. Overexpression of FGFR1 reversed the inhibitory effect of miR-198 on cell growth, migration, and invasion. A recent study showed that high FGFR1 expression promoted epithelial-mesenchymal transition (EMT) and led to a poor prognosis of gastric cancer patients. Downregulation of FGFR1 inhibits glycolytic metabolism, proliferation, and invasion in breast cancer. These results suggest that FGFR1 has a tumor-promoting function. Taken together, these findings indicate that the miR-198/FGFR1 pathway plays a tumor-suppressor role in PTC.

In summary, the present findings provide robust evidence that circRAPGEF5 serves as a novel oncogenic circRNA by sponging miR-198 and promoting FGFR1 expression, which could be a promising prognostic biomarker in PTC. The present data also suggest that targeting the circRAPGEF5/miR-198/FGFR1 axis is a potential strategy for the treatment of PTC.

Materials and Methods

Animal Ethics Statement

Four week-old BALB/c nude mice weighing 15–20 g were purchased from SLARC (Shanghai, China). All animal experiments were approved by the Animal Research Committee of Tongren Hospital of Shanghai Jiaotong University School of Medicine.

Patients and Thyroid Tissue Collection

PTC tissues and paired adjacent noncancerous thyroid tissues were collected from 30 patients diagnosed with thyroid cancer between January 2017 and July 2018 at Tongren Hospital. None of the patients had chemotherapy or radiotherapy before surgery. The study was approved by the Ethics Committee of Tongren Hospital, and informed consent was obtained from all patients.

Cell Lines and Cell Culture

The human thyroid normal cell line Nthy-ori 3-1; the PTC cell lines BCPAP, KTC-1, and K1; and HEK293 cells were purchased from ATCC (Manassas, VA, USA). Cells were cultured in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS; Invitrogen) and maintained in a humidified incubator with 5% CO2 at 37°C.

RNA Extraction and Real-Time qPCR

Total RNA was isolated from cell lines and patient tissues using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. RNA samples were reverse-transcribed using a Prime Script RT Reagent Kit (Takara, Dalian, China). Real-time PCR was performed on an AB 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA) using SYBR Premix Ex Taq (Takara). U6 was used as an internal control for miR-198, and GAPDH was the internal control for circRAPGEF5 and FGFR1. The primers for miR-198, forward: 5′-GGUCCAGAGGGGAGAUAGGUUC-3′; primers for circRAPGEF5, forward: 5′-TGCCTCTCATTCCTGCCAGA-3′ and reverse: 5′-TCTTGATAGAGTCGCAGATGTTAGA-3′; primers for FGFR1, forward: 5′-CAGTCGACAATCACTCCTGGGAAGATCTCATTG-3′ and reverse: 5′-GGTGGATCCAGGGCCACAAGGTGGACAATCGG-3′; primers for GAPDH, forward: 5′-CAGGAGGCATTGCTGATGAT-3′ and reverse: 5′-GAAGGCTGGGGCTCATTT-3′; and primers for U6, forward: 5′-CTCGCTTCGGCAGCACA-3′ and reverse, 5′-AACGCTTCACGAATTTGCGT-3′.

FISH

Specific probes for the circRAPGEF5 sequence (Dig-5′-GCGTCTAACATCTGCGACTCTATC-3′-Dig) were used for in situ hybridization as previously described. In brief, cy5-labeled probes specific for circRAPGEF5 were designed. Nuclei were counterstained with DAPI. All procedures were performed according to the manufacturer’s instructions (GenePharma, Shanghai, China).

Bioinformatics Analysis

The hsa_circ_0001681 miRNA target genes were predicted using the website https://circinteractome.nia.nih.gov. The miRNA target genes were predicted using the website https://www.genecards.org.

CCK-8 Assay

Transfected cells were seeded in 96-well plates, and at 0, 24, 48, and 72 h post-treatment, 10 μL CCK-8 solution (Beyotime, Beijing, China) was added into each well, followed by incubation for 2 h according to the manufacturer’s instructions. Optical density (OD) was measured at 450 nm using a microplate reader (BioTek Instruments, Winooski, VT, USA).

Luciferase Reporter Assay

The binding sites of circRAPGEF5 and 3′ UTR-FGFR1, called circRAPGEF5-Wild, circRAPGEF5-Mut, FGFR1 3′ UTR-Wild, and FGFR1 3′ UTR-Mut, were inserted into the KpnI and HindIII sites of pGL3 promoter vector (Realgene, Nanjing, China) in a dual-luciferase reporter assay. Then 80 ng plasmid, 5 ng renilla luciferase vector pRL-SV40, 50 nM miR-198 mimics, and NC was transfected into HEK293T cells by applying Lipofectamine 2000 (Invitrogen, Shanghai, China). At 48 h after transfection, firefly and renilla luciferase activities were measured consecutively using the dual luciferase reporter assay system (Promega, Boston, MA, USA). The luminescence ratios of firefly to renilla luciferase were calculated, and each assay was repeated in three independent experiments.

Cell Transfection

For miR-198 overexpression, miR-198 mimic or the corresponding NC (miR-NC) was obtained from GenePharma. BCPAP cells were transfected with either miR-198 mimic or miR-NC at a final concentration of 50 nM using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol. Cells were used for miR-198 expression analysis or other experiments after 48 h of transfection. For miR-198 inhibition, BCPAP cells were treated with miR-198 inhibitor for 48 h before miR-198 expression analysis or other experiments.

For circRAPGEF5 expression analysis, siRNA against circRAPGEF5 (5′-GCGACUCUAUCAAGAAGAAUU-3′) was constructed by GenePharma using the CMV-copGFP-T2A-puro-H1-MCS vector. BCPAP cells were transfected with circRAPGEF5 silencing vector at a final concentration of 50 nM using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol.

For FGFR1 overexpression, the FGFR1 overexpression vector (pcDNA3.1) was constructed by GenePharma, and BCPAP cells were transfected with either FGFR1 overexpression vector or NC at a final concentration of 50 nM using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol.

Western Blot Analysis

Protein was extracted from cells using RIPA lysis buffer containing proteinase inhibitor (Sigma-Aldrich, St. Louis, MO, USA). Protein concentrations were determined using the BCA Protein Assay Kit (Vigorous Biotechnology Beijing, Beijing, China). Equal amounts of protein lysates (20 μg per lane) were resolved on 10% SDS-PAGE gels and electroblotted onto nitrocellulose membranes (Millipore, Madison, WI, USA). Membranes were blocked for 2 h with 5% non-fat dry milk in Tris-buffered saline containing 0.1% Tween 20 and incubated overnight at 4°C with primary antibodies. GAPDH was used as the internal protein loading control. Membranes were incubated with secondary antibodies for 1 h at room temperature. Immune complexes were detected by enhanced chemiluminescence (Cell Signaling Technology, Danvers, MA, USA). Integrated densities of bands were quantified by Quantity One software (Bio-Rad).

Invasion Assays

Invasion assays were performed using Transwell inserts (Corning Life Sciences, Bedford, MA, USA) and Matrigel (BD Biosciences) as previously reported. Cells were stained and counted in four random fields per Transwell insert. Mean values of migratory or invasive cells were expressed as percentages relative to the control. Data were based on three independent experiments.

Mouse Xenograft Model

BCPAP cells stably infected with circRAPGEF5 silencing vectors or NC were used for in vivo experiments. A total of 5 × 106 viable cells were injected into the right flanks of nude mice as described in our previous study. Tumor sizes were measured using a Vernier caliper every 5 days starting on the fifth day after injection, and tumor volume was calculated using the following formula: volume = 1/2 × length × width. At 30 days after implantation, mice were sacrificed for qRT-PCR analysis.

Statistical Analysis

Continuous variables were expressed as the mean ± SD. One-way ANOVA or Student’s t test was performed for multiple comparisons using GraphPad Prism software, v5.0 (GraphPad, La Jolla, CA, USA). p values ⩽ 0.05 indicated a statistically significant difference.

Author Contributions

W.L. and J.Z. generated and analyzed data. M.J. and M.Z. designed experiments and drafted the manuscript. All authors approve the version of the manuscript to be published.

Conflicts of Interest

The authors declare no conflict of interest.