Background
Prostate cancer (PCa) is a commonly diagnosed cancer in men and the leading cause of cancer-related death in western countries [
1]. Localized prostatic cancer can be cured by radical prostatectomy. However, once a tumor migrates outside of the gland, incurable metastatic disease is inevitable. Metastasis of PCa is the main cause of death, which greatly reduces the lifespan and quality of life of patients [
2,
3]. The current treatment option is limited and ineffective for patients with advanced stage metastatic PCa [
4‐
6]. Therefore, identifying novel biomarkers and therapeutic targets for metastatic PCa is a clinical imperative.
Circular RNAs (circRNAs) are single-stranded, covalently closed RNA molecules, formed by back-splicing of pre-mRNAs, and originally considered to be a by-product of splicing errors [
7]. However, many studies have revealed that circRNAs are essential regulators of gene expression and implicate in multiple biological processes of tumorigenesis via acting as microRNA (miRNA) sponges, RNA binding protein (RBP)-binding molecules, transcriptional regulators, or templates for protein translations [
8]. As reported, circRNA has multiple functions in regulating cell development, differentiation, and metabolism. In particular, some studies claim that the abnormal expression of circRNA is related to the invasion, migration and metastasis of cancer cells in distant organs. For example, circCSNK1G3 promotes PCa cell growth by interacting with miR-181 [
9]. CircNOLC1 promotes PCa progression by sponging miR-647 [
10]. CircRHOBTB3 inhibits gastric cancer growth via miR-654-3p [
11]. CircACTN4 boosts the progression of breast cancer by complexing with FUBP1 [
12]. CircSPARC enhances the progression of colorectal cancer by recruiting FUS [
13]. These tremendously large number of reports indicated that circRNA may act as an ideal biological target for diagnosis and treatment, including PCa. However, multiple of circRNAs and their specific the biological roles in the metastasis of PCa stills worth to be explored.
In our study, we found a circRNA, which is derived from exons 4 and 5 of the ZNF652 (circBase ID: hsa_circ_0003258), as a novel oncogene in PCa. We found that hsa_circ_0003258 was significantly up-regulated in PCa tissues, and associated with the metastasis of PCa cells. Mechanistically, hsa_circ_0003258 directly interacted with the RNA-binding protein IGF2BP3, and enhanced the stability of histone deacetylase 4 (HDAC4) mRNA, consequently resulting in the aggressive nature of PCa. Moreover, hsa_circ_0003258 may upregulate the expression of Rho GTPase activating protein 5 (ARHGAP5) to promote tumor progression by sponging miR-653-5p. This work indicates the essential roles of hsa_circ_0003258 in PCa progression.
Materials and methods
Patients and samples
Plasma and tissue samples were obtained from 18 PCa patients of the Nanfang Hospital (Guangzhou, China). Inclusion criteria: 1. Patients who were diagnosed as PCa by pathological diagnosis before operation; 2. Patients signed an informed consent form. Exclusion criteria: 1. Patients with second primary tumors, HIV or syphilis virus, severe liver, kidney or other systemic diseases, other malignant diseases; 2. Patients who received preoperative chemotherapy or radiotherapy before surgery.
Nine healthy individuals who were in good health determined by physical examinations in the Nanfang Hospital were included as controls. The research protocol and the use of human tissues and plasma samples were approved by the Ethics Committee of Southern Medical University and all participants signed informed consents. A PCa tissue microarray (TMA) was purchased from Shanghai Outdo Biotech Co. Ltd. and contained both normal prostate tissues and PCa tissues along with each patient’s age, clinical stage, Gleason score, and metastasis status and tumor node metastasis (TNM) classification, that were recorded and archived in the National Engineering Center for Biochip. The detailed clinic features of enrolled patients were summarized in Supplementary Table S
2. The TMA consisted of 134 cancer cases and were used for staining of has_circ_0003258.
CircRNA Array analysis
The circRNAs from the plasm of PCa patients and control individuals were collected for microarray analysis. Methods of extracting plasma RNA was based on the previous protocols [
14]. Sample preparation was performed according to the Arraystar’s standard protocols, as described previously [
15]. Briefly, Using Rnase R to digest linear RNA and retain circRNAs. Then, RNAs were amplified for complementary RNA and labeled with an Arraystar Super RNA Labeling Kit (Kangcheng Biotechnology, Shanghai, China). Finally, these labeled RNAs were hybridized using Arraystar mouse circRNA Array (V1.0; Arraystar), and scanned by the Agilent Scanner G2505C (ANDbio, Temecula, CA). We defined the statistical criteria for selecting differentially expressed circRNAs using |fold changes| ≥ 2.0 with
p values < 0.05.
Quantitative real-time polymerase chain reaction (qRT-PCR) and genomic DNA extraction
TRIzol reagent (Invitrogen) was used to extract RNA from PCa cells and tissues. TB-Green PCR Master Mix Kit (Takara) and PrimeScript RT reagent Kit (TaKaRa) were used in qRT-PCR. Data were normalized to Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and used the 2
-ΔΔCt method to calculate. Genomic DNA (gDNA) is extracted from cells by the Easy Pure Genomic DNA kit (Transgen Biotech, Lot#L61221). All primers were obtained from Sangon Biotech and listed in Supplementary Table S
1.
Nuclear-cytoplasmic fractionation
Nuclear and cytoplasm of cells were separated by Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, United States) Following the manufacturer’s instructions. Finally, Data were normalized to GAPDH (cytoplasmic control) and U6 (nuclear control) and used the 2-ΔΔCt method to calculate.
RNase R and Actinomycin D treatment
Total RNA (2 μg) was incubated with or without 3 U/mg of RNase R (Epicentre Technologies, Madison, WI, USA) for 15 min at 37 °C. The expression of hsa_circ_0003258 and other RNA were detected by qRT-PCR. PCa cells were treated with Actinomycin D or DMSO (Sigma Aldrich, St. Louis, MO, USA) to evaluate the stability of hsa_circ_0003258 and its linear gene ZNF652. The stability of RNA was detected by qRT-PCR.
Fluorescence in situ hybridization (FISH)
The location of hsa_circ_0003258 was detected by FISH assay using a Cy3-labeled probes.
FISH kit (RiboBio, China) was used to examine the signals following the manufacturer’s instruction. The Nikon AISi Laser Scanning Confocal Microscope (Nikon instruments Inc., Japan) was used to visualize the images.
Cell culture and transfection
PCa cell lines (LNCaP, C4–2, 22Rv1, PC3 and DU145) and human epidermal cell (RWPE-1) were purchased from the Stem Cell Bank, the Chinese Academy of Sciences. PCa cells lines were grown with RPMI-1640 medium (Gibco BRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Gibco, Australia). RWPE-1 cells were cultured in Keratinocyte Serum Free Medium (KSFM, Gibco). Cells were incubated at 37 °C in 5% CO
2. Use Lipofectamine 3000 (Invitrogen, Carlsbad, CA) to transfect cells with designated nucleotides or plasmids according to the manufacturer’s instructions. Small interfering RNAs (siRNAs) targeting hsa_circ_0003258, IGF2BP3, HDAC4, ARHGAP5 and negative control (NC) siRNA were provided by RiboBio Co. (Guangzhou, China). All siRNA sequences were listed in Supplementary Table S
1.
RNA sequence and data analyses
RNA-seq (RNA sequencing) between the control DU145-NC and sh-has_circ_0003258 cell was performed as previously described [
3]. RNA-seq is performed by BGISEQ platform. The other analyses including heatmap, gene set enrichment analysis was completed by BGI Dr. Tom system.
Western blotting
Western blot analysis was performed as previously described [
3]. In brief, PCa cells were added in radio immunoprecipitation assay buffer (RIPA, KeyGEN, BioTECH, China), and BCA Protein Assay Kit (KeyGEN, BioTECH, China) was used for protein determination. Then, the proteins were separated by 10% SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, USA). The membranes were blocked in no fat milk and then incubated with primary antibodies overnight at 4 °C. The primary antibodies included anti-Tubulin antibody (Cell Signaling Technology, USA, 1:1000), anti-ARHGAP5 antibody (ABclonal, China, 1:1000), anti-HDAC4 antibody (ABclonal, China, 1:1000), anti-IGF2BP3 antibody (Abcam, USA, 1:1000), anti-ZEB-1 antibody (Cell Signaling Technology, USA, 1:1000), anti-E-Cadherin antibody (Cell Signaling Technology, USA, 1:1000), anti-Vimentin antibody (Cell Signaling Technology, USA, 1:1000), and phospho-Erk1/2-T202/Y204 antibody (ABclonal, China, 1:1000). Subsequently, the membranes were immersed with the secondary antibody for 1 h. Finally, enhanced chemiluminescence (ECL) kit (Pierce Biotechnology, USA) was used to detect the level of protein.
Transwell migration assay
Transwell (Costar, Corning, USA) with a multipolar (8.0 μm) polycarbonate membrane was utilized to conducted Cell migration experiments. Cells (5 × 104) were mixed with serum-free medium and added into the upper chamber of the insert. Then, 800 μl complete medium was added to the bottom chamber. The cells in the chamber were mixed in 4% paraformaldehyde for 10 min and stained with Giemsa (Boster Ltd., Wuhan, China) at different time points. After that, cells in the upper chamber were removed and the number of cells on the bottom surface were observed under a microscope and counted using Image J software.
Wound-healing assay
PCa cells were grown to confluence in six-well plates before making a wound. Using a 10 μl pipette tip to draw a gap and keep the cells in RPMI-1640 medium for 24 h. The images of the wound area were taken using the microscope at 0, 12 and 24 h. Wound widths were measured using the ImageJ software.
Immunofluorescence
PCa cells were grown on coverslips (Corning, USA), and mixed with 4% paraformaldehyde for 30 min. Then the cells were permeabilized with 0.1% TritonX-100. Subsequently, the cells were incubated with Tris-buffered saline containing 5% bovine serum albumin (BSA) for 30 min. Afterwards, samples were incubated with antibodies specific for IGF2BP3 (Abcam, USA, 1:100) at 4 °C overnight. Finally, coverslips were treated with the fluorescent secondary antibody Alexa Fluor 488-conjugated goat anti-rabbit IgG (#4412S, Cell Signaling Technology) (1250 dilution) and DAPI (300 nmol/L) staining. The images were photographed under a Nikon AISi Laser Scanning Confocal Microscope (Nikon instruments Inc., Japan).
Luciferase reporter assay
Dual-Luciferase Reporter Assay kit (Genecopoaie, China) was used to verify the relationships between hsa_circ_0003258 and miR-653-5p. Wild-type (WT) or mutant (MUT) 3′-UTR of hsa_circ_0003258 were cloned into the firefly-tagged pGL3 promoter luciferase vector (GeneCopoeia, Rockville, USA). Using Lipofectamine 3000 (Invitrogen) to transfect HEK-293 cells with miR-653-5p mimic hsa_circ_0003258 or its mutant plasmid following the manufacturer’s instructions. Luciferase activities were detected by a dual luciferase assay system (GeneCopoeia) after 48 h. The experiment was performed with three replicates.
Biotinylated RNA pull-down assay
Biotin-labeled hsa_circ_0003258 and oligonucleotide probes (RiboBio Co. Ltd. Guangzhou, China) were mixed with streptavidin magnetic beads (Beaver, China) in RIP buffer for 4 h. Subsequently, the DU145 cell lysate was incubated with the probes complex for 12 h at 4 °C. After purification, enriched hsa_circ_0003258 and miRNAs were quantified by qRT-PCR. Meanwhile, the bound proteins were identified by Western blotting.
RNA immunoprecipitation assay
RIP assay was performed by using EZ-Magna RIP™ Kit (#17–701, Merck Millipore) with antibodies specific for IGF2BP3 following the manufacturer’s instructions. The immunoprecipitated RNAs were detected by qRT-PCR to measure the level of hsa_circ_0003258 and HDAC4. Total RNAs (input) and isotype antibody (IgG) were applied as controls.
Animal models
All experiments’ animal procedures were approved by the Animal Care Committee of Southern Medical University. For tumor metastatic studies in vivo, DU145 cells (1 × 107 cells per mouse) transfected with NC or sh-has_circ_0003258 were injected through lateral tail vein of BALB/c nude male mice (n = 5 for each group). Lung tissues were collected and examined for metastasis. After 40 days, the tumors in vivo were evaluated by fluorescence imaging using the IVIS (PerkinElmer, USA). The presence of cancer cells was confirmed by H&E (hematoxylin and eosin) staining. At the same time, immunohistochemistry (IHC) staining was conducted using antibodies against ARHGAP5 (Abcam, USA, 1:100) and HDAC4 (Abcam, USA, 1:100).
Statistical analyses
Statistical analyses were performed using the SPSS software version 20.0 (SPSS, Inc., Chicago, IL) or GraphPad Prism 7.0 (GraphPad Software, USA). All values are expressed as mean ± standard deviation (SD). Differences in mean values between groups were analyzed using ANOVA and Student’s t tests. The correlation between hsa_circ_0003258 expression and clinicopathological properties was analyzed using a χ2 test. P value < 0.05 was considered statistically significant. (*, P <0.05. **, P ≤ 0.01. ***, P ≤ 0.001. ****, P ≤ 0.0001).
Discussion
CircRNAs are spliced form pre-mRNAs and play crucial roles in regulating chromatin dynamics and protein expression [
20]. Recently, more and more circRNAs have been noticed dysregulated in different kinds of malignancies, which participated in diversified physiological and pathological processes of tumors cells, including proliferation, cell cycle progression, and metastasis [
9]. They can be potential diagnostic biomarkers and targets for tumor therapy. However, the functions of most circRNAs are still to be explored for there are just limited experimental evidence reveal the role they played in different kinds of tumors. Liquid biopsy is a non-invasive method that uses body fluids, such as plasma, urine and serum, to detect disease states [
21,
22]. Besides, some research has confirmed that circRNAs play critical roles in liquid biopsy [
23‐
26]. They are often detected not only in tumor tissues and cells but also in body fluids from cancer patients. Interestingly, our previously study identified circRNAs from patients’ serum and proved that cirZMIZ1 could facilitate the growth of PCa cells and might act as a liquid biomarker for PCa [
14]. This study implied that circRNAs in the body fluids may be promising biomarkers and even potential therapeutic targets for human malignancies, including PCa.
In order to further identify potentially valuable circRNAs in body fluids of PCa patients, circRNAs microarray scanning and bioinformatics analyses were performed on plasma extracted from peripheral blood specimen. Interestingly, we found a novel potential tumor metastasis promoting circRNA termed hsa_circ_0003258. Previously, hsa_circ_0003258 has been reported to promote metastasis of hepatocellular carcinoma by sponging miR-203/miR-502-5p [
27]. Furthermore, exosomal circ-0003258 promotes hepatocellular carcinoma progression via miR-29a-3p/GUCD1 Axis [
28]. However, whether hsa_circ_0003258 is involved in the development and progression of PCa is still elusive. Here, we identified hsa_circ_0003258 was upregulated in PCa tissues and cells, and which is associated with patient’s clinical parameters. Results of functional assays revealed that ectopic expression of hsa_circ_0003258 significantly increased cell migration. On the contrary, knockdown of hsa_circ_0003258 produced the opposite effects. Finally, through in vivo experiments, we further confirmed that hsa_circ_0003258 could promote the lung homing capacity of PCa cells. Interestingly, our results demonstrated that silencing of hsa_circ_0003258 largely increased the expression of E-cadherin, but not vimentin, which was consistent with previous reports [
3,
29]. Notably, changes in epithelial markers and changes in mesenchymal markers do not always occur at the same time during the EMT process, loss of E-cadherin was reckoned as the crucial step to initiate EMT that sustained tumor metastasis. Furthermore, our GO and KEGG analysis showed that the expression of hsa_circ_0003258 was correlated with cell adhesion and focal adhesion. Actually, E-cadherin not only acts as an epithelial marker during EMT, but also knows to be implicated in cell adhesion. It is well known that reduced cell adhesion is considered to be the initiation of the elial-EMT process [
30]. Collectively, these results suggested that hsa_circ_0003258 facilitates metastasis of PCa via including the EMT pathway.
Accumulating evidence shown that the mechanisms through which circRNAs exert biological functions are diverse, including serving as microRNA (miRNA) sponge, interacting with RBPs, and even translating proteins. One of the most extensively studies is circRNA terminating the regulation of miRNA to its target gene by binding to the miRNA as a competing endogenous RNA (ceRNA) through the base complementary pairing principle. For instance, CircHIPK3 promotes colorectal cancer via miR-7 [
31], circEPSTI1 regulates ovarian cancer via miR-942 [
32], circANKS1B regulates breast cancer via miR-148/152 [
33] and circ-AMOTL1L regulates PCa via miR-193a-5p [
34]. Here, we performed the bioinformatic analysis to predict the potential binding miRNAs of hsa_circ_0003258, and found that miR-653-5p is potentially regulated by hsa_circ_0003258. Then the direct binding relationship between hsa_circ_0003258 and miR-653-5p was validated by RNA pull-down assay and luciferase reporter experiments. Our results showed that hsa_circ_0003258 may act as the sponge for miR-653-5p and inhibit the metastasis of PCa.
Previous studies have shown conflicting results that miR-653-5p may be involved in tumor progression as an oncogene or tumor suppressor. MiR-653-5p has been identified as an oncogene in PCa [
35] and gastric cancer [
36], whereas acted as a tumor suppressor gene in cervical cancer [
37], non-small cell lung cancer [
38] and melanoma [
39]. On the contrary, we found that miR-653-5p plays a critical tumor suppressor role in the process of inhibiting PCa metastasis by suppressing EMT. It is well known that miRNAs can inhibit translation or reduce mRNA stability by directly targeting gene 3′-UTR, thereby acting as a post-transcriptional regulator. In this study, we explored that hsa_circ_0003258 may function as an endogenous miRNA sponge to inhibit the expression of miR-653-5p by binding to miR-653-5p in PCa cells, leading to diminish the repression of miR-653-5p on ARHGAP5 3′-UTR and eventually facilitate cell metastasis. Moreover, upregulation of miR-653-5p significantly reversed the expression of ARHGAP5 and the increased metastasis induced by overexpression of hsa_circ_0003258. Therefore, these data exhibit that hsa_circ_0003258 partially regulates the stability of ARHGAP5 by acting as a ceRNA in PCa.
The tertiary structure of circRNAs leads to higher protein adsorbing capacity than linear RNA sequences. Thus, the circRNA-interacting RBPs act as an important molecular mode of action in the occurrence, translation, transcription regulation and extracellular transport of target genes [
40]. For example, circFOXK2 complexed with YBX1 and hnRNPK to promote PDAC progression [
41]. CircFOXP1 contributed to gallbladder cancer progression by interacting with PTBP1 and sponging miR-370 [
42]. CircNDUFB2 in complex with IGF2BPs inhibited non-small cell lung cancer progression [
43]. Considering that it is a function of circRNA to interact with RBPs, we further explored the potential RBPs of hsa_circ_0003258 to understand its mechanism. Our results demonstrated that hsa_circ_0003258 could directly bind to IGF2BP3 to enhance the stability of HDAC4 mRNA and thereby promote the EMT in PCa. IGF2BP3 is a well-known RBP which prevents mRNA decay by recognizing m6A-modified mRNA [
44]. Numerous studies have shown that IGF2BP3 is ubiquitously expressed in eukaryotic tissues and is often up-regulated in human cancers [
45]. Previous study showed that IGF2BP3 could form a ternary complex of circFNDC3B-IGF2BP3-CD44 mRNA to increase CD44 expression in the translation process, finally promoting the migration and invasion of gastric cancer cells [
46]. Herein, we hypothesized that hsa_circ_0003258 might function its role in PCa metastasis by the same mechanism. RNA immunoprecipitation assays and RNA pull-down assays confirmed that hsa_circ_0003258 increased HDAC4 expression via the formation of a ternary complex of hsa_circ_0003258-IGF2BP3-HDAC4 mRNA, thereby increasing HDAC4 mRNA stability. Intriguingly, knockdown of IGF2BP3 abolished the effects of hsa_circ_0003258 on the protein expression of HDAC4, which indicates HDAC4 is the downstream effector molecule of hsa_circ_0003258. HDAC4, which belongs to class IIa of the HDAC family, may contribute to tumor development and progression through multiple mechanisms [
47,
48]. However, its biological roles in the tumor metastasis of PCa remain elusive. Here, our results demonstrated that HDAC4 silencing reduced the cell migration of PCa. After all, from the perspective of circRNA regulation, we have revealed a unique function, that is, circRNAs can interact with RBPs to improve the stability of mRNA.
The Ras-Mitogen-Activated Protein Kinase (MAPK) pathway that comprise of the extracellular signal-related kinase1/2 (ERK1/2), JNK and p38 MAPK which play key roles in regulating multiple cellular processes like cell proliferation, apoptosis and migration [
49]. ERK activates the downstream RAS/RAF/MEK/ERK and results in cancer progression [
50]. Inhibiting RAS and related GTPases will result inhibition of targeting the downstream effectors of RAS signaling, including the RAF-MAPK/ERK kinase pathway and the PI3K-AKT-mTOR kinase pathway [
51,
52]. In this study, we observed that when silencing of hsa_circ_0003258 the differentially expressed genes were enriched in the MAPK signaling pathway. Then our further validation experiments revealed that hsa_circ_0003258 could upregulate the expression of p-ERK. These results suggested that hsa_circ_0003258 could accelerate PCa metastasis through ERK signaling pathway.
However, it is interesting that hsa_circ_0003258 was downregulated in the plasma level, but our results demonstrate that hsa_circ_0003258 was significant up-regulated in PCa cells and tissues. Besides, hsa_circ_0003258 acted as a promising biomarker for metastasis of PCa. These contradictory results between cell and plasma levels suggested that there could be more molecular mechanisms for hsa_circ_0003258 in regulating PCa progression. Based upon, we assumed that hsa_circ_0003258 may be affected by the tumor micro-environment during the process of being released into the plasma, and may be taken up by cells in the micro-environment. As far as we know, plasma is such a mixture of the secreta of the whole-body cells. In addition, some studies focusing on the tumor micro-environment also confirm our hypothesis. For example, Loss of miR-203 promotes tumor growth in some tumors [
53] but some evidences show that high expression of exosomal miR-203 was associated with a higher TNM stage in colorectal cancer patients [
54]. In addition, it was found that compared with adjacent normal tissues, circRIP2 was reduced in bladder cancer tissues. More CD3/CD8 cells in higher circRIP2 expressed patients were identified when compared with lower circRIP2 expression. These findings suggest that the oncogenic circRIP2 may play a tumor suppressive role by communicating with the micro-environment comprehensively. But circRIP2 can promote bladder cancer progression via inducing EMT by activating miR-1305/Tgf-β2/smad3 pathway in vitro. This inconsistency between clinic and biological level suggests that there could be more molecular mechanisms for circRIP2 in regulating bladder cancer behaviors [
55]. Taken together, these results indicate that the involvement of tumor micro-environment must be considered if we want to fully understand the function of hsa_circ_0003258.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.