Circular RNA circVAPA knockdown suppresses colorectal cancer cell growth process by regulating miR-125a/CREB5 axis

Samples of CRC tumor mucosal tissues and paired normal mucosal tissues (n = 42) were provided by patients who were diagnosed with CRC at China–Japan Union Hospital of Jilin University. Each CRC patient taking part in this research signed the written informed consent and the approval was endowed by the Ethics Committee of China-Japan Union Hospital of Jilin University.

Human normal colon mucosal epithelial cell line NCM460 and CRC cell lines (HCT116 and LOVO) were collected from Cell Bank of Chinese Academy of Science (Shanghai China) and American Type Culture Collection (ATCC, Manassas, VA, USA), respectively. Cells were maintained in a humidified incubator with 5% CO₂ at 37 °C with Roswell Park Memorial Institute 1640 medium (RPMI 1640, Transgene, Beijing, China), when RPMI 1640 cultured cells were added with 10% fetal bovine serum (FBS, Gibco, Carlsbad, CA, USA).

Total RNA was extracted from CRC tissues and cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), followed by incubation at 37 °C for 20 min with or without RNase R (Epicentre, Shanghai, China). After purification with phenol–chloroform (Solarbio, Beijing, China), total RNA was reversely transcribed into the complementary DNA (cDNA) by a PrimeScript™ RT Reagent Kit (Takara, Dalian, China) in accordance with the supplier’s direction. Whereafter, with the help of SYBR^® Premix Ex Taq™ reagent (TaKaRa), RT-qPCR was conducted on a Roche Light Cycler 480 Real-time PCR Amplifier following the operation manual. The relative levels of genes were calculated by the 2^− ΔΔCt method, normalizing with house-keeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or U6 small nuclear RNA. The primers used were presented as follows:

CircVAPA: 5′-TGGATTCCAAATTGAGATGCGTATT-3′ (sense), 5′-CACTTTTCTATCCGATGGATTTCGC-3′ (antisense); miR-125a: 5′- GGTCATTCCCTGAGACCCTTTAAC -3′ (sense), 5′-GTGCAGGGTCCGAGGT-3′ (antisense); CREB5: 5′-AGATGGTCCTCTGTTGGGAA-3′ (sense), 5′-TGGACACGGTTATGAGAATGA-3′ (antisense); VAPA mRNA: 5′-GAAGCTGTGTGGAAAGAGGC-3′ (sense), 5′-GAGCATTCC. CTGGTGGAGTT-3′ (antisense); GAPDH: 5′-GTCAACGGATTTGGTCTGTATT-3′ (sense), 5′-AGTCTTCTGGGTGGCAGTGAT-3′ (antisense); U6: 5′-GCTCGCTTCGGCAGCACA-3′ (sense), 5′-GAGGTATTCGCACCAGAGGA-3′ (antisense).

CircVAPA small interfering RNA (si-circVAPA), si-CREB5 and their negative control (si-con), miR-125a mimics (miR-125a), miR-125a inhibitor (anti-miR-125a) and their negative controls (miR-con and anti-miR-con) were provided by GenePharma (Shanghai, China). Moreover, pcDNA3.1-circVAPA (circVAPA), pcDNA3.1-CREB5 (CREB5) and pcDNA 3.1 empty vector (vector) were obtained from Genecreat (Wuhan, China). The oligonucleotide and vector were transfected in HCT116 and LOVO cells using Lipofectamine 3000 reagents (Invitrogen)based on the instruction guidelines.

Simply, transfected CRC cells were fixed with ice ethanol, followed by treatment with RNase A (Takara) and Propidium Iodide (PI) (Invitrogen). At 24 h after incubation, a FACSCalibur (Becton–Dickinson, Franklin Lakes, NJ, USA) was employed to detect the DNA content. Percentage of cell-cycle stages was analyzed with ModFit LT (Verity Software House, Topsham, ME, USA).

The abilities of migration and invasion were examined with Transwell chambers (Corning Incorporated, Corning, NY, USA) and matrigel (BD Biosciences, San Jose, CA, USA). In the assay, cells were seeded on the upper chambers with (for invasion assay) or without matrigel (BD Biosciences; for migration assay). As a chemoattractant, the medium containing FBS (Gibco) was added into the lower chambers. At 48 h after transfection, the cells on the lower surface were fixed and stained, and the cells remaining in the upper chamber were removed with cotton swabs. The staining cells were imaged and counted under an inverted microscope.

For this assay, ECAR was measured on Seahorse XF96 Glycolysis Analyzer (Seahorse Bioscience, North Billerica, MA, USA) with Seahorse XF Glycolysis Stress Test Kit (Seahorse Bioscience) based on the user’s guidebook. In short, transfected cells were plated into a Seahorse XF 96 cell culture microplate, followed by the measurements of baseline. At the indicated time points, Glucose, Oligomycin (oxidative phosphorylation inhibitor) and 2-DG (glycolytic inhibitor) were sequentially injected. Similarly, OCR was detected with Seahorse XF Cell Mito Stress Test Kit (Seahorse Bioscience) on this Analyzer. Oligomycin, FCCP (reversible inhibitor of oxidative phosphorylation) and Rote + AA (mitochondrial complex I inhibitor rotenone plus the mitochondrial complex III inhibitor antimycin A) were injected. Finally, the results were analyzed with the Seahorse XF96 Wave software (Seahorse Bioscience).

Glucose uptake, lactate and ATP production in CRC cells were examined by Glucose Uptake Colorimetric Assay kit (Biovision, Milpitas, CA, USA), Lactate Assay kit (Sigma St. Louis, MO, USA) and ATP Colorimetric Assay kit (Sigma), severally, in line with the supplier’s direction.

Briefly, the sequences of circVAPA or CREB5 3′ UTR containing putative binding sites of wild-type miR-125a were amplified and inserted into the luciferase report pGL3 vector (Promega, Madison, WI, USA), obtaining circVAPA-wild type (WT) or CREB5-WT reporter vector. Subsequently, the constructed reporter vector was transfected with miR-125a, miR-125a + vector, and miR-125a + CREB5 or miR-125a + circVAPA into HCT116 or LOVO cells using Lipofectamine 3000 reagents (Invitrogen). At 48 h after transfection, the luciferase activity was assessed with the dual-luciferase reporter assay kit (Promega).

RIP assay was implemented with a Magna RIP kit (Millipore, Bedford, MA, USA) as per the manufacturer’s protocol. In this assay, transfected CRC cells were harvested and lysed in complete RIP lysis buffer. Whereafter, magnetic beads conjugated with anti-Argonaute2 (Ago2, Millipore) or anti-immunoglobulin G (IgG, Millipore) were added into the buffer. And then, proteinase K was applied to separate the RNA–protein complexes from beads. At last, the levels of circVAPA, miR-125a and CREB5 in the precipitates were detected with RT-qPCR.

In short, radioimmunoprecipitation lysis buffer (RIPA; Beyotime, Ningbo, China) was utilized to harvest the lysates from CRC tissues and cells, followed by the separation with 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). Then, the proteins were transferred to a nitrocellulose membrane (Millipore), and blocked with 5% non-fat milk. After the incubation with specific primary antibody: CREB5 (1:1000, ab168928, Abcam, Cambridge, MA, USA) and β-actin (1:1000, ab8226, Abcam), the secondary antibody was probed with the membrane. Finally, protein signals were visualized with enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech, Piscataway, NJ, USA).

All values were presented as mean ± standard deviation (SD) at three independent assays. Statistical analysis was calculated with SPSS 20.0 software. Differences between groups were performed with Student’s t- test or one-way analysis of variance (ANOVA). P values of less than 0.05 were recognized as statistically significant.

To investigate the function of circVAPA in CRC, its expression level was detected by using RT-qPCR. Compared with 42 normal mucosal tissues, circVAPA level was highly expressed in 42 CRC tumor mucosal tissues (Fig. 1a). Then, we further confirmed that circVAPA level was evidently increased in CRC cell lines (HCT116 and LOVO) relative to human normal colon mucosal epithelial cell line NCM460 (Fig. 1b). Moreover, to verify that circVAPA was a circular RNA, HCT116 and LOVO cells were treated with RNase R. As displayed in Fig. 1c, d RNase R treatment effectively downregulated VAPA mRNA level in HCT116 and LOVO cells, but not that of circVAPA expression level, suggesting that circVAPA was resistant to RNase R digestion. In a word, as a circular RNA, circVAPA might be involved in CRC progression.

Considering the high expression of circVAPA in CRC, we knocked down circVAPA expression in HCT116 and LOVO cells. As shown in Fig. 2a, transfection of si-circVAPA obviously reduced circVAPA expression level, manifesting that si-circVAPA could be used for the subsequent loss-of-function assays. Flow cytometry assay exhibited that circVAPA deletion inhibited cell cycle progression in HCT116 and LOVO cells (Fig. 2b, c). Meanwhile, decreased migration and invasion were viewed caused by the downregulation of circVAPA (Fig. 2d, e). Overall, circVAPA silencing could hinder cell cycle progression, migration and invasion of CRC cells.

Glycolysis has been reported to exert the positive role in the progression of CRC [23], hence, we explored whether circVAPA has a function in glycolysis in CRC cells. Meanwhile, extracellular acidification rate (ECAR), a product of glycolysis, was largely determined by lactic acid release. The ECAR approximated the glycolysis flux rate [24], and oxygen consumption rate (OCR) was an indicator of mitochondrial oxidative respiration. Therefore, with the treatment of glucose, oligomycin or 2-DG, we detected the effect of circVAPA knockdown on the levels of ECAR and OCR in CRC cells. As illustrated in Fig. 3a, b circVAPA deletion led to a conspicuous decrease in ECAR in HCT116 and LOVO cells. Moreover, accelerated OCR (Fig. 3c, d) and the reduced level of glucose uptagke (Fig. 3e), lactate production (Fig. 3f) and ATP (Fig. 3g) in HCT116 and LOVO cells with si-circVAPA were viewed. Collectively, circVAPA deficiency impeded CRC cells glycolysis.

Given that circVAPA exerted the crucial role in the cell cycle, migration, invasion and glycolysis of CRC cells. Meanwhile, a substantial body of recent research has disclosed that circRNA could serve as miRNA sponge to regulate mRNA expression [25]. Therefore, through bioinformatics analysis, there was a potential binding between miR-125a and circVAPA or CREB5 3′UTR (Fig. 4a, b). Whereafter, we used the dual-luciferase reporter assay to further verify the prediction results in HCT116 and LOVO cells. Data suggested that the miR-125a notably dampened the luciferase activity of circVAPA-WT reporter vector, while re-introduction of CREB5 3′UTR effectively abrogated the effect in HCT116 cells (Fig. 4c). Meanwhile, our results also indicated that miR-125a could obviously reduced the luciferase activity of CREB5 WT reporter vector, which was abrogated by circVAPA upregulation in LOVO cells (Fig. 4d). Consistent with bioinformatics analysis and dual-luciferase reporter assay, RIP assay indicated that the levels of circVAPA, miR-125a and CREB5 were strikingly enriched in Ago2 pellets of HCT116 and LOVO cell extracts compared with normal IgG control group (Fig. 4e, f). Besides, miR-125a level was evidently reduced in HCT116 and LOVO cells transfected with circVAPA, while its level was notably increased in CRC cells transfected with si-circVAPA (Fig. 4g). Meanwhile, we found that miR-125a overexpression repressed the level of CREB5 in HCT116 and LOVO cells, and miR-125a downregulation promoted CREB5 level (Fig. 4h). Intriguingly, in HCT116 and LOVO cells, miR-125a upregulation strikingly impaired the promotion effect of circVAPA overexpression on CREB5 protein level, on the contrary, miR-125a inhibition remarkably overturned the negative effect of si-circVAPA on CREB5 protein level (Fig. 4i, j). All of these suggested that circVAPA could perform as a sponge of miR-125a to positively modulate CREB5 expression.

As mentioned above, we speculated that circVAPA could exert its function through miR-125a/CREB5 axis in CRC. Simultaneously, low expression of miR-125a and high expression of CREB5 were observed in HCT116 and LOVO cells relative to NCM460 (Fig. 5a, b). Thus, to prove the inference, we executed the rescue assay in HCT116 and LOVO cells. As presented in Fig. 5c, d cell cycle arrested by circVAPA silencing was partly abrogated by miR-125a knockdown or CREB5 overexpression in HCT116 and LOVO cells. Transwell analysis result testified that the suppression of migration and invasion triggered by circVAPA knockdown was abolished through anti-miR-125a or CREB5 in HCT116 and LOVO cells (Fig. 5e–h). In addition, our data suggested that CREB5 upregulation boosted cell cycle progression, migration, invasion and glycolysis in HCT116 and LOVO cells, however, the knockdown of CREB5 produced the opposite results (Additional file 1: Figure S1). All the data described above indicated that circVAPA could regulate cycle progression, migration and invasion through miR-125a/CREB5 axis in CRC cells.

Next, we further identify whether circVAPA could modulate glycolysis by miR-125a/CREB5 axis in CRC cells. As exhibited in Fig. 6a, b miR-125a knockdown or CREB5 overexpression significantly mitigated the inhibitory effect of circVAPA deletion on ECAR in HCT116 and LOVO cells. Additionally, we also found that anti-miR-125a or CREB5 obviously abolished circVAPA silencing-mediated increase in OCR (Fig. 6c, d), and decline in the levels of glucose uptake (Fig. 6e, f), lactate production (Fig. 6g, h) and ATP (Fig. 6i, j) in HCT116 and LOVO cells. Taken together, these results suggested that circVAPA knockdown could exert the inhibitory effect on glycolysis through miR-125a/CREB5 axis in CRC cells.