Background
Pancreatic cancer is a common malignant digestive system disease with poor prognosis and delayed diagnosis [
1]. The incidence and death rate of pancreatic cancer cases are increasing worldwide, which causes a heavy burden to human health [
2]. Although gemcitabine-based chemotherapy is the conventional therapeutic manner, the 5-year survival rate of pancreatic cancer sufferers is still low with only 6% [
3]. As we known, early diagnosis and therapy are for reducing mortality, which can be achieved by seeking and developing new reliable targets.
Circular RNA (circRNA) is a novel RNA without protein coding capacity, featured by stability, conservatism and specificity in expression [
4]. Researchers have revealed that circRNAs are highly expressed in the cytoplasm owing to their ability in absorbing microRNAs (miRNAs), which commonly function at the post-transcriptional level [
5]. Disease-specific expression of circRNAs made them available as molecular markers for the treatment of cancers [
6]. Additionally, circRNAs were enrolled in the evolution of diverse cancers [
7], such as gastric carcinoma [
8], breast cancer [
9], lung carcinoma [
10] and bladder cancer [
11]. As previously reported, circ_0066147 contributed to cell growth and metastasis via upregulating p21-activated kinase 1 (PAK1) through acting as a sponge for miR-330-5p in pancreatic cancer [
12]. Zhu and his colleagues indicated that circ_0006215 upregulation promoted the viability and migration of pancreatic cancer cells [
13]. In this study, it was found that the expression of circ-membrane bound O-acyltransferase domain containing 2 (circ-MBOAT2) was significantly increased in pancreatic cancer specimen and cells; however, the underlying mechanism was still unclear.
MiRNAs are class of noncoding RNAs and are about 20 nucleotides in size [
14]. MiRNAs act as oncogenes or anti-oncogenes in cancer progression by regulating cellular biological activities, such as cell proliferation, metastasis, apoptosis and glutamine catabolism [
15,
16]. MiR-433-3p has been unveiled to repress tumor process. For example, miR-433-3p inhibited cell proliferation and metastasis, but contributed to cell apoptosis by binding to F-box protein 22 (FBXO22) in osteosarcoma [
17] and through interacting with CREB in glioma [
18]. MiR-433-3p inhibited cell epithelial-mesenchymal transition in colon cancer via targeting annexin A2 [
19]. Bi et al. also found that miR-433 (miR-433-3p) sponged by LINC00657 restrained cell growth and facilitated cell apoptosis through restraining the expression of p21 (RAC1) activated kinase 4 in pancreatic ductal cancer [
20]. In this experiment, it was found that miR-433-3p contained the binding sites of circ-MBOAT2. Whether circ-MBOAT2 could mediate pancreatic cancer growth via absorbing miR-433-3p was further explored in this study.
Glutamic-oxaloacetic transaminase 1 (GOT1) plays a vital role in the production of non-essential amino acids in glutamine metabolism, which is necessary for cell proliferation of various cancers [
21,
22]. Research explained GOT1 served as a tumor promoter in pancreatic cancer growth [
23]. Of note, starbase online database (
http://starbase.sysu.edu.cn/agoClipRNA.php?source=mRNA) showed GOT1 contained the binding sequence of miR-433-3p. These data demonstrated GOT1 might be implicated in miR-433-3p-mediated pancreatic cancer progression. Thus, we further explored the relationship between miR-433-3p and GOT1 in modulating pancreatic cancer development.
In this paper, the impacts of circ-MBOAT2 silencing on cell proliferation, metastasis, apoptosis and glutamine metabolism were revealed. Whether circ-MBOAT2 was a sponge of miR-433-3p and miR-433-3p targeted GOT1 were disclosed. In addition, rescue experiments were employed to illustrate the effects between miR-433-3p and circ-MBOAT2 or GOT1 on tumor evolution and glutamine metabolism. Furthermore, the influences of circ-MBOAT2 knockdown on tumor formation in vivo were unveiled.
Materials and methods
Specimen collection and ethics committee
34 pairs of human pancreatic cancer and matched normal tissues were obtained from patients with pancreatic cancer from Union Hospital, Tongji Medical College, Huazhong University of Science and Technology. The collected pancreatic cancer tissues included stage I-II pancreatic cancer tissues (N = 22) and stage III-IV pancreatic cancer tissues (N = 12), and the clinical stage was confirmed by at least two pathologists. Collected tissues were stored at − 80 °C in a refrigerator. The Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology allowed this research. Relevant patients signed the written informed consents before surgery.
Cell purchase and culture
EK-Bioscience Co., Ltd. (Guangzhou, China) provided human pancreatic cancer cell lines (AsPC-1, BxPC-3, PANC-1 and SW1990) and human pancreas ductal epithelial cell line HPDE. HPDE, AsPC-1 and BxPC-3 cells were grown in Roswell Park Memorial Institute-1640 (RPMI-1640; Procell, Wuhan, China). PANC-1 and SW1990 cells were cultivated in Dulbecco’s modified Eagle’s medium (DMEM; Procell) and L15 media (Procell), respectively. Media were supplemented with 10% fetal bovine serum (FBS; Procell) and 100 μg/mL penicillin-streptomycin solution (Procell). Cells were grown at 37 °C in a humid incubator with 5% CO2.
Plasmid construction and cell transfection
Small interfering RNA targeting circ-MBOAT2 and GOT1 (si-circ-MBOAT2 and si-GOT1), miR-433-3p mimics and inhibitors (miR-433-3p and anti-miR-433-3p) and controls (si-NC, si-con, sh-NC, miR-NC and anti-miR-NC) were synthesized by GenePharma (Shanghai, China). The small hairpin RNA against circ-MBOAT2 was built by inserting top strand (5′-GATCCGAGAACATGCACAAGTCAACTCAAGAG AGTTGACTTGTGCATGTTCTCTTTTTG-3′) and bottom strand (5′-AATTCAAAAA GAGAACATGCACAAGTCAACTCACTTCAGTTGACTTGTGCATGTTCTCG-3′) into the pCDH-U6-MCS-EF1-GreenPuro vector, and named as sh-circ-MBOAT2. The overexpression plasmids of circ-MBOAT2 (circ-MBOAT2) and GOT1 (GOT1) and controls (pCD5-ciR and pcDNA) were provided by Geneseed (Guangzhou, China). Cell transfection was conducted with Lipofectamine 2000 (Thermo Fisher, Waltham, MA, USA). Si-circ-MBOAT2, si-GOT1, miR-433-3p, anti-miR-433-3p, circ-MBOAT2, GOT1 and their controls were used to determine that whether circ-MBOAT2 regulated pancreatic cancer development and glutamine catabolism by miR-433-3p/GOT1 axis. Sh-circ-MBOAT2 and sh-NC were employed to reveal circ-MBOAT2-mediated impacts on tumor formation in vivo. The synthesized oligonucleotide sequences were si-circ-MBOAT2 5′-GAGAACATGCACAAGTCAACT-3′, miR-433-3p 5′-AUCAUGAUGGGCUCCUCGGUGU-3′, anti-miR-433-3p 5′-ACACCGAGGAGCCCAUCAUGAU-3′, si-NC 5′-CCTCTACCTGTCGCTGAGCTGTAAT-3′, miR-NC 5′-UUUGUACUACACAAAAGUACUG-3′ and anti-miR-NC 5′-CAGUACUUUUGUGUAGUACAAA-3′.
Quantitative real-time polymerase chain reaction (qRT-PCR)
Collected tissues or cultured cells were isolated using miRNeasy Mini Kit (Qiagen, Valencia, CA, USA). The concentration of obtained RNAs was detected with NanoDrop-1000 apparatus (Thermo Fisher). Then, cDNA was synthesized with a High-Capacity cDNA RT Kit (Thermo Fisher) or MicroRNA RT Kit (Thermo Fisher). For determining relative levels of RNAs/mRNAs, a SuperReal PreMix Color kit (Tiangen, Beijing, China) was employed with a Mx3000P system (Stratagene, Santa Clara, CA, USA). Obtained data were analyzed with the 2-∆∆Ct method, and U6 and β-actin acted as references. The sense and antisense primers were circ-MBOAT2 5′-ATGCCTTACACTTTCTTG-3′ and 5′-GAGCTAGTTTTGCTTGAA-3′; linear MBOAT2 5′-GATGTTTCGGAAGGATGA-3′ and 5′-TTGTAAGAGCAAAGTGGG-3′; miR-433-3p 5′-ACACTCCAGCTGGGATCATGATGGGCTCCT-3′ and 5′-TGGTGTCGTGGAGTCG-3′; miR-144-3p 5′-ACACTCCAGCTGGGTACAGTATAGATGA-3′ and 5′-TGGTGTCGTGGAGTCG-3′; GOT1 5′-CTGGGAGTGGGAGCATAT-3′ and 5′-CAAGGGCAAGACGAGAAG-3′; U6 5′-CTCGCTTCGGCAGCACA-3′ and 5′-AACGCTTCACGAATTTGCGT-3′; glyceraldehyde 3-phosphate dehydrogenase (GAPDH) 5′-GGTCACCAGGGCTGCTTT-3′ and 5′-GGAAGATGGTGATGGGATT-3′; β-actin 5′-CACCATTGGCAATGAGCGGTTC-3′ and 5′-AGGTCTTTGCGGATGTCCACGT-3′.
RNase R treatment assay
Cultured PANC-1 and SW1990 cells were collected and lysed. RNAs were extracted and then incubated with RNase R (3 U/μg RNA, Geneseed) at 37 °C for 30 min. After that, RNAs were purified using RNeasy MinElute Cleaning Kit (Qiagen). Finally, circ-MBOAT2 expression was determined by qRT-PCR. Linear MBOAT2 functioned as a control.
Cytoplasmic and nuclear circ-MBOAT2 analysis
Cytoplasmic and nuclear circ-MBOAT2 were assessed with a PARIS™ Kit (Thermo Fisher). In short, fresh PANC-1 and SW1990 cells were harvested and suspended in 200 μL ice-cold Cell Fractionation Buffer (Thermo Fisher), which were then centrifuged at 450 rpm for 4 min. The cytoplasmic fraction and nuclear pellet were carefully separated and lysed. Finally, RNAs were isolated, and circ-MBOAT2 expression was determined by qRT-PCR. U6 and GAPDH served as references.
3-(4,5)-dimethylthiahiazo (−z-y1)-3,5-di-phenytetrazoliumromide (MTT) assay
Cell viability was illustrated by a MTT kit (Solarbio, Beijing, China). In short, after performing various treatments, PANC-1 and SW1990 cells were cultured in 96-well plates for 24, 48 and 72 h, respectively. Fresh media (Procell) were mixed with MTT solution (Solarbio) and severally added into culture plates. 4 h later, dimethyl sulfoxide (Sigma, St. Louis, MO, USA) was added into wells to dissolve formazan. Samples were assessed by microplate reader (Thermo Fisher), and results were revealed by analyzing the output of the wavelength at 490 nm.
PANC-1 and SW1990 cells were seeded in 6-well plates. Media were renewed every 3 days during culture. Two weeks later, cell supernatant was removed, and proliferative colonies were incubated with paraformaldehyde (Sigma) and crystal violet (Sigma), respectively. Cell colony-forming ability was illustrated via counting cell numbers. A colony was deemed when cell numbers > 50.
Flow cytometry analysis
Cell apoptosis was determined by an Annexin V-fluorescein isothiocyanate (Annexin V-FITC) detection kit (Solarbio). Briefly, cultured cells were digested with trypsin (Thermo Fisher), and suspended in binding buffer (Solarbio). Following that, Annexin V-FITC (Solarbio) and propidium iodide (PI; Solarbio) were severally incubated with cells in dark. Finally, samples were analyzed with flow cytometer (Thermo Fisher).
Transwell invasion assay
The invasion of PANC-1 and SW1990 cells was investigated with transwell chambers with Matrigel (Corning, Madison, New York, USA). Shortly, cells were seeded in the upper chambers with FBS-free DMEM or L15 media (Procell). And DMEM or L15 media containing 15% FBS (Procell) were added into lower chambers. At 24 h after culture, cell supernatant was discarded and cells were singly incubated with methanol (Sigma) as well as crystal violet (Sigma). Results were analyzed by figuring up the cell numbers in the lower chambers under microscope (Nikon, Tokyo, Japan) at a 100(×) magnification.
Wound-healing assay
PANC-1 and SW1990 cells were cultured in 6-well plates until their confluence reached about 100%. Then, cell wounds were created and cells were washed using phosphate buffer solution (PBS; Thermo Fisher). Cells were continued to be cultured in serum-free media for 24 h. Finally, cell migratory ability was determined via calculating the width of wounds under microscope (Nikon) with a 100(×) magnification.
The determination of glutamine uptake and α-KG production
Glutamine uptake and α-KG production were determined by glutamine and α-KG assay kits (Abcam, Cambridge, UK), respectively. In brief, cultured cells were harvested and then suspended in assay Buffer (Abcam). After that, samples were centrifuged at 9000 rpm for 12 min, and supernatant was collected. Perchloric acid (Abcam) and potassium hydroxide (Abcam) were used to incubate with supernatant, respectively. After performing centrifugation at 12,000 rpm for 12 min, supernatant was collected and analyzed with the microplate reader (Thermo Fisher) with the wavelength at 450 nm for glutamine uptake assay or at 570 nm for α-KG production assay.
The detection of glutamate production
Glutamate production was detected with glutamate assay kit (Abcam). Briefly, cells were collected when their confluence reached about 80%. Then, harvested cells were incubated with lysis buffer (Abcam) for 15 min. After that, Assay buffer (Abcam), Enzyme mixture (Abcam) and NADP Stock Solution (Abcam) were mixed with lysates. Finally, samples were analyzed with the microplate reader (Thermo Fisher) with the wavelength at 570 nm.
In vivo assay
Charles River (Beijing, China) provided 5-week old nude mice (N = 16), and mice were fed in pathogen-free environment. Nude mice were averagely divided into 2 groups. 4 × 106 SW1990 cells stably transfected with sh-circ-MBOAT2 or sh-NC were diluted in 0.2 mL PBS, and then injected into the right flank of the back of mice. Eight days later, tumor volume was measured every 3 days. Twenty-three days later, nude mice were killed and tumor weight was determined. Additionally, a part of every tumor was reserved for further analyzing in circ-MBOAT2 expression. The Animal Care Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology approved this study.
Dual-luciferase reporter assay
According to the results predicted by circinteractome online database (
https://circinteractome.nia.nih.gov/index.html) or starbase online database (
http://starbase.sysu.edu.cn/agoClipRNA.php?source=mRNA), the wild-type (WT) plasmids and mutant vectors of circ-MBOAT2 and GOT1 3′-untranslated region (3’UTR) were provided by Geneseed Co., Ltd., and severally named as WT-circ-MBOAT2, GOT1 3’UTR-WT, MUT-circ-MBOAT2 and GOT1 3’UTR-MUT. Constructed plasmids were transfected into PANC-1 and SW1990 cells with miR-433-3p or miR-NC after mixed with DharmaFECT 4 (Thermo Fisher) according to the manufacturer’s instructions. Luciferase activities were detected by Dual-Lucy Assay Kit (Solarbio) with
Renilla luciferase activity as a control.
Western blot analysis
Collected tissues and cultured cells were severally lysed with RIPA buffer (Beyotime, Shanghai, China). Then, lysates were mixed with loading buffer (Thermo Fisher), and were then boiled in boiling water for 8 min. Lysates were loaded on 12% bis-tris-acrylamide gel (Thermo Fisher) to separate protein. After that, protein bands were electrotransferred onto polyvinylidene fluoride membranes (Millipore, Bradford, MA, USA), which were then immersed in nonfat milk (Solarbio). Membranes were incubated with anti-GOT1 (1:1000; Affinity, Nanjing, China) and anti-β-actin (1:8000; Affinity) overnight at 4 °C, and were then incubated with horseradish peroxidase-marked secondary antibody (1:8000; Affinity) at 37 °C for 2 h. Protein bands were presented by eyoECL Plus Kit (Beyotime). β-actin acted as a reference.
Statistical analysis
Data from 3 independent duplicate tests were assessed with SPSS 21.0 software (IBM, Somers, NY, USA), and were presented as means ± standard deviations (SD). The linear relationship between miR-433-3p and circ-MBOAT2 or GOT1 was compared with Spearman’s correlation test. Two-tailed Student’s t-tests or Wilcoxon rank-sum test was employed to assess the significant differences between the two groups, and one-way analysis of variance (ANOVA) was performed to compare the difference among 3 groups or more groups. Statistical significance was deemed when P value < 0.05.
Discussion
CircRNAs, a novel RNA, are commonly dysregulated and play vital parts in pancreatic cancer development [
24]. For example, circ-LDLRAD3 could serve as a diagnostic biomarker of pancreatic cancer as its close correlation with venous and lymphatic invasion as well as metastasis [
25]. The immune escape of pancreatic cancer cells induced via hypoxia-inducible factor 1-alpha could be modulated by circ_0000977 [
26]. In addition, circ_001653 downregulation hindered cell proliferative and invasive properties through sponging miR-377 [
27]. In this research, we found circ-MBOAT2 silencing suppressed tumor progression and glutamine catabolism via downregulating GOT1 through sponging miR-433-3p.
Previous study has illustrated that circ-MBOAT2 acted as an oncogene in tumor growth in prostate cancer [
28]. In addition, Yang et al. showed circ-MBOAT2 was highly expressed in pancreatic ductal cancer specimens, and it could increase metastasis–related matrix metallopeptidase 7 expression [
29]. Herein, circ-MBOAT2 was also shown to be apparently upregulated in the tissue and cell samples of pancreatic cancer, and circ-MBOAT2 expression was higher in stage III-IV pancreatic cancer tissues than in stage I-II. It was found that circ-MBOAT2 silencing restrained cell proliferation, migratory and invasive abilities and glutamine catabolism, and circ-MBOAT2 overexpression abolished these effects. Additionally, circ-MBOAT2 knockdown upregulated cell apoptotic rate, and repressed tumor growth in vivo. The results from cytoplasmic and nuclear circ-MBOAT2 analysis displayed that circ-MBOAT2 was mainly located in cytoplasm, suggesting that circ-MBOAT2 chiefly functioned at the post-transcriptional level. Considering circRNAs serving as sponge of miRNAs to regulate cancer evolution [
30], we further sought the miRNA associated with circ-MBOAT2.
Subsistent data unveiled that miR-433-3p participated in regulating the pathogenesis of various cancers with acting as an anti-oncogene. For instance, miR-433-3p hindered cell proliferative and metastatic capacities in esophageal squamous [
31] and oral squamous cell cancer [
32]. MiR-433-3p led to apparent repression on cell proliferation and metastasis, and promotion on cell apoptosis in lung cancer [
33]. In this paper, qRT-PCR results showed that miR-433-3p expression was reduced in pancreatic cancer tissue samples and cell lines. In addition, miR-433-3p also be found to inhibit malignant properties of pancreatic cancer via repressing cell proliferation and metastasis as well as inducing cell apoptosis, which was consistent with the findings of Bi et al [
20]. Beyond that, the role of miR-433-3p in glutamine catabolism was revealed in this paper. Data showed miR-433-3p repressed glutamine catabolism. Furthermore, miR-433-3p was negatively related to circ-MBOAT2 in expression, and miR-433-3p inhibitors attenuated si-circ-MBOAT2-mediated impacts on tumor development and glutamine catabolism. These evidences confirmed circ-MBOAT2 regulated malignant properties of pancreatic cancer by sponging miR-433-3p.
The downstream gene of miR-433-3p was continued to be explored. Results presented miR-433-3p contained the putative binding sites of GOT1, and the interaction between miR-433-3p and GOT1 was further identified by dual-luciferase reporter assay. Besides, our data displayed notably high GOT1 expression in PANC-1 and SW1990 cells. GOT1 ectopic expression restrained miR-433-3p-mediated influences on tumor development and glutamine catabolism, suggesting GOT1 could hinder malignant properties of pancreatic cancer, which was approved by the data from Wang et al [
23]. Meanwhile, the above data demonstrated miR-433-3p modulated pancreatic cancer evolution via targeting GOT1.
The relationship between circ-MBOAT2 and GOT1 was further revealed. Data showed circ-MBOAT2 silencing decreased GOT1 expression, whereas miR-433-3p inhibitors impaired this impact. This finding implied circ-MBOAT2 could modulate GOT1 expression via sponging miR-433-3p.
All in all, we found circ-MBOAT2 silencing repressed pancreatic cancer progression via regulating cell proliferation, apoptosis, migration, invasion through repressing glutamine catabolism, and the underlying mechanism was that circ-MBOAT2 induced GOT1 expression via absorbing miR-433-3p (Figure S
6). This finding not only provides a novel mechanism for studying circ-MBOAT2-based effects on the evolution of pancreatic cancer, but also provides a new potential biomarker for the diagnosis and therapy of pancreatic cancer.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.