Background
Colorectal cancer (CRC) is a common malignant tumour in digestive system, which remains a substantial public health challenge across the globe in the last 30 years [
1]. The incidence and mortality rate of CRC are the third and second in all malignant tumours, respectively [
2]. A large number of epidemiological data showed that the number of CRC patients were increasing by years in developing and developed countries, and the progress of CRC seriously threatens the survival of patients [
3‐
5]. However, due to the lack of diagnostic biomarkers and therapeutic targets, the treatment of CRC is not ideal and further improvement is needed. Therefore, it is critical to further explore the molecular mechanisms underlying the progression of CRC.
Tumour cells are highly metabolized, and their continuous growth depends on adequate nutrient supply. In the early stage of tumorigenesis, nutrients can penetrate through the tissue to maintain its growth, but when the diameter exceeds 2 mm, the tumour must form new blood vessels to provide nutrition [
6‐
10]. In addition, blood metastasis is one of the main ways of tumour metastasis to distant organs [
11,
12]. Therefore, angiogenesis is essential for tumour growth and metastasis. Recently, circRNA has attracted great research interest due to its regulatory role in diseases especially for oncology. Many explorations confirmed that circRNA is closely related to the angiogenesis of CRC [
13‐
15]. The covalently closed loop structure makes circRNA conserved and stable. CircRNA were widely and diversely present in eukaryotic cells, with certain tissue specificity, timing and disease specificity which means it has the potential to be biomarker for tumour diagnosis. The competing endogenous RNA (ceRNA) is the most classic mechanism and have been widely reported in various types of cancer [
16‐
24]. Besides the ceRNA mechanism, RNA-binding proteins and functional proteins coding are main way to function [
25‐
28].
In this study, we initially investigated the function and molecular mechanism of hsa_circ_0001821 (designated as circ3823 through RNA-seq) in CRC, and explored its potential as a diagnostic biomarker. This research revealed that circ3823 promoted the expression of TCF7 and TCF7 downstream MYC and CCND1 via inhibition of miR-30c-5p, resulting in proliferation, metastasis and angiogenesis of CRC. In addition, we found that N6-methyladenosine (m6A) modification exists on circ3823. And the degradation rate of circ3823 was regulated by m6A recognition protein YTHDF3 and demethylase ALKBH5. Our results indicate that circ3823 exerts oncogenic potential and it may be a candidate in diagnosis marker and therapeutic target of CRC.
Materials and methods
Tissue, serum and paraffin section sources
Tissue and serum samples which were collected within the last 4 years were obtained from the First Affiliated Hospital of Zhengzhou University, Henan, China. After the tissues and serums were separated from the human body, they were quickly transferred to liquid nitrogen for storage. Since the tissue and serum samples were not repeatedly frozen and thawed, the experimental data truly reflected the RNA level in the body.
Paraffin-embedded tissue sections were also obtained from the First Affiliated Hospital of Zhengzhou University. Samples from patients with CRC treated between 2012 and 2016 were selected. This study was approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University, and all patients signed informed consent forms.
RNA isolation, library synthesis and RNA sequencing
Total RNA was isolated and purified using Trizol reagent (Invitrogen, Carlsbad, USA) following the manufacturer’s procedure. The RNA amount and purity of each sample was quantified using NanoDrop ND-1000 (NanoDrop, Wilmington, USA). The RNA integrity was assessed by Agilent 2100 with RIN number > 7.0. Approximately 5 μg of total RNA was used to deplete ribosomal RNA according to the manuscript of the Ribo-Zero rRNA Removal Kit (Illumina, San Diego, USA). After removing ribosomal RNAs, the left RNAs were fragmented into small pieces using divalent cations under high temperature. Then the cleaved RNA fragments were reverse-transcribed to create the cDNA, which were next used to synthesise U-labeled second-stranded DNAs with E. coli DNA polymerase I, RNase H and dUTP. An A-base is then added to the blunt ends of each strand, preparing them for ligation to the indexed adapters. Each adapter contains a T-base overhang for ligating the adapter to the A-tailed fragmented DNA. Single-or dual-index adapters are ligated to the fragments, and size selection was performed with AMPureXP beads. After the heat-labile UDG enzyme treatment of the U-labeled second-stranded DNAs, the ligated products are amplified with PCR by the following conditions: initial denaturation at 95 °C for 3 min; 8 cycles of denaturation at 98 °C for 15 s, annealing at 60 °C for 15 s, and extension at 72 °C for 30 s; and then final extension at 72 °C for 5 min. The average insert size for the final cDNA library was 300 bp (±50 bp). At last, we performed the paired-end sequencing on an Illumina Hiseq 4000 (LC Bio, China) following the vendor’s recommended protocol.
Cell culture
HCT116 cells were obtained from iCell Bioscience Inc. (Shanghai, China). SW480, DLD-1 and HT29 cells were obtained from the Biotherapy Center of The First Affiliated Hospital of Zhengzhou University. All cells were cultured in DMEM/high-glucose (HyClone, Logan, Australia) with 10% fetal bovine serum (Biological Industries, Cromwell, USA) at 37 °C and 5% CO2 under saturated humidity.
Total RNA isolation and real-time quantitative PCR (qRT-PCR) assay
Total RNA was isolated from serum and tissues using RNAiso Plus (Takara, Dalian, China) according to the manufacturer’s instructions. The integrity and purity of the extracted total RNA were measured using NanoDrop One (Thermo Fisher Scientific, Waltham, USA) ultra-micro UV spectrophotometer. Reverse transcription was performed using the PrimeScript RT reagent Kit (Takara, Dalian, China) with gDNA Eraser. After removing the genomic DNA at 42 °C for 2 min, the tissue RNA was reverse transcribed into cDNA under the following conditions: 37 °C for 15 min and 85 °C for 5 s. Serum RNA was reverse transcribed into cDNA using a RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, USA) under the following conditions: 25 °C for 5 min, 42 °C for 60 min, and 70 °C for 5 min. The product was immediately stored at − 80 °C until use.
The qRT-PCR was performed on a QuantStudio 5 Real-Time PCR System (Applied Biosystems, Foster City, USA) using a Hieff qPCR SYBR Green Master Mix kit (Yeasen, Shanghai, China). The qRT-PCR reaction was performed 95 °C for 5 min, followed by 40 cycles of 95 °C for 10 s and a primer-specific annealing temperature of 60 °C for 30 s. The qRT-PCR primer sequences were provided in Table
S2. The relative quantification values for RNA were calculated by the 2
−ΔΔCt method using GAPDH as an internal reference.
Vector construction and cell transfection
The full-length of circ3823 was cloned into over expression vector pcDNA3.1 (Hanbio Biotechnology, Wuhan, China), while the mock vector with no circ3823 sequence served as a control. SiRNAs targeting the back-splice junction site of circ3823 and siRNA-NC were synthesized by RiboBio (RiboBio, Guangzhou, China), efficiency detected by qRT-PCR. The mimics and inhibitors of miR-30c-5p were purchased from RiboBio (RiboBio, Guangzhou, China). Lipofectamine 3000 (Invitrogen, Carlsbad, USA) were used to cell transfections. The sequences of siRNAs were listed in Table
S5.
Western blot analysis
Total protein of CRC tissues or cell lines were extracted by RIPA with PMSF and determined via BCA Protein Assay Kit (Solarbio, Beijing, China). Equal amounts of protein were separated on SDS-PAGE gels and then transferred to PVDF membrane (Millipore, Massachusetts, USA). After blocked with 5% BSA in TBST, PVDF membrane were incubated with primary antibodies against CCND1 (1:1000), MYC (1:1000) (Abways, Shanghai, China), TCF7 (1:500) (Santa Cruz, CA, USA) and GAPDH (1:5000) (Proteintech, Wuhan, China) at 4 °C overnight and then hybridized with a secondary antibody at 37 °C for 1 h. The intensity of the bands was analysed with a chemiluminescence kit (Millipore, Massachusetts, USA).
In situ hybridization
The in-situ hybridization probes and kits were designed and synthesized by Wuhan Servicebio company. The probe sequences were listed in Table
S4. After prehybridization at 37 °C for 2 h, tissue sections were hybridized with specific DIG-labelled circ3823 probes at 37 °C overnight, and stained by DAB and hematoxylin (Solarbio, Beijing, China). Slides were photographed with a fluorescence microscope (Olympus, Tokyo, Japan). ISH was independently evaluated at 200× magnification using light microscopy by two pathologists who were blinded to the clinicopathological data. A semiquantitative evaluation of circ3823 was performed using a method described in the previous works [
23,
29]. The evaluation procedure was based on staining intensity and extent of staining as follows: staining intensity for circ3823 was scored as 0 (negative), 1 (weak), 2 (moderate) and 3 (strong). Staining extent was scored as 0 (0), 1 (1–25%), 2 (26–50%), 3 (51–75%) and 4 (76–100%), depending on the percentage of positively stained cells. The product of the staining intensity and the staining extent scores were regarded as ISH scores. The agreement between the two evaluators was 90%, and all scoring discrepancies were resolved by discussion between the two evaluators.
RNase R treatment
Total RNA (2 μg/group) of SW480 cells were incubated for 0 min, 10 min, 20 min, 30 min at 37 °C with 5 U/μg RNase R (Epicentre Technologies, Madison, USA), and subsequently the abundance of linear RNA and circular RNA were analysed by qRT-PCR.
Actinomycin D assay
SW480 cells were exposed to 100 ng/ml actinomycin D (Merck, Darmstadt, Germany) at 0 h, 4 h, 8 h, 12 h, 24 h. Then the cells were harvested, and total RNA was extracted. The stability of circ3823 and PVT1 mRNA were analysed using qRT-PCR.
Nuclear and cytoplasmic extraction and fluorescence in situ hybridization (FISH)
The nuclear and cytoplasmic fractions of RNA were extracted with a PARIS™ kit (Invitrogen, Thermo Fisher Scientific, Waltham, USA). Place the cell slide at the bottom of the 24-well plate and cultivate an appropriate number of cells (6 × 10
4/well). Before the experiment, make the cell confluence reach 60–70%. FISH assay was executed to observe the location of circ3823 and miR-30c-5p in CRC cells. Briefly, after prehybridization at 37 °C for 30 min, cell climbing piece were hybridized with 2.5 μL 20 μM specific Cy3-labelled circ3823 probes and FAM-labelled miR-30c-5p probes (Servicebio, Wuhan, China) at 37 °C overnight, and dyed with DAPI. The probe sequences were shown in Table
S4. Slides were photographed with confocal laser scanning microscopy (Zeiss, Jena, Germany).
Cell Counting Kit-8 (Dojindo Laboratories, Kumamoto, Japan), Cell-Light EdU Apollo567 In Vitro Kit (RiboBio, Guangzhou, China) and Colony formation assays were used to detect proliferation of CRC cells according to the manufacturer’s instructions. HCT116 cell apoptosis was detected by Annexin V-FITC/propidium iodide (PI) apoptosis detection kit (Beyotime, Shanghai, China). The cells were harvested and double stained with FITC and PI after transfection, followed analysed on a flow cytometer (Becon Dickinson FACSCalibur, NY, USA). As for tube junction forming assay of HUVEC, the culture supernatants of HCT116 and SW480 cells transfected with plasmid or siRNA were collected, followed incubate with 3 × 104 HUVEC cells per well in a 96-well culture plate precoated with matrix (Corning, NY, USA) for 6 h. Finally, the tube junction formation was observed under an optical microscope (Olympus, Tokyo, Japan).
Migration and invasion assays
After transfection, 2 × 104 CRC cells were seeded into the upper chambers without Matrigel (Corning, NY, USA), and 500 μL complete medium was added into the bottom chambers (Corning, NY, USA) for migration assays. For invasion assays, 1 × 105 CRC modified cells were seeded into the upper chambers with Matrigel (Corning, NY, USA). After 36 h, the cells on the compartment were fixed in 4% paraformaldehyde (Beyotime, Shanghai, China) and stained by crystal violet (Solarbio, Beijing, China), then photographed and counted with an optical microscope (Olympus, Tokyo, Japan).
Animal experiments
Female BALB/c nude mice (4-week-old) were housed under standard conditions. In the tumour growth xenograft model, 5 × 106 HCT116 cells with LV-circ3823 or LV-NC were suspended in 100 μL serum-free DMEM and subcutaneously injected into the right flank of each mouse. The volumes of tumours were measured every 3 days and calculated as 0.5 × length×width2. After 30 days the mice were sacrificed and the tumours were removed for further analysis. In the “tail vein–lung metastasis” nude mouse models, 2.5 × 106 HCT116 cells with LV-circ3823 or LV-NC were injected into the tail vein of mice. For imaging tumours in live animals, D-Luciferin, Potassium Salt D (Yeasen, Shanghai, China) was dissolved in sterile distilled water (final concentration: 15 mg/ml). Mice were anaesthetized with isoflurane and injected intraperitoneally with 100 μl of the luciferin solution. After 10 mins, images were acquired with the IVIS Lumina series III (PerkinElmer, Waltham, Massachusetts, USA). The mice were killed after 49 days, and all the lungs were surgically removed.
The competing endogenous RNA (ceRNA) hypothesis reveals an interaction mechanism between RNAs. We predicted the target microRNAs of circ3823 with TargetScan and regRNA. More reliable microRNAs were screened based on the intersection of TargetScan (
http://www.targetscan.org/vert_72/) and RegRNA 2.0 (
http://regrna2.mbc.nctu.edu.tw/). The target mRNA of each microRNA was predicted by TargetScan.
Immunohistochemistry (IHC) and immunofluorescence (IF)
For IHC assay, paraffin sections were incubated with primary antibodies against TCF7 (1:50) (Santa Cruz, CA, USA), MYC (1:100), CCND1 (1:100) (Abways, Shanghai, China), CD34 (1:800), CD31(1:1500), α-SMA (1:2000), collagen IV (1:1000) and ki67 (1:10000) (Proteintech, Wuhan, China) at 37 °C for 60 mins, secondary antibodies at 37 °C for 15 mins and horseradish enzyme labelled streptavidin solution for 10 min, then stained by DAB and hematoxylin. For IF analysis, cell climbing pieces were incubated with TCF7 antibodies overnight at 4 °C and AF488-conjugated secondary antibodies (Biolegend, San Diego, USA), then dyed by DAPI and observed by a fluorescence microscope (Olympus, Tokyo, Japan).
Luciferase activity assays
The sequences of circ3823, TCF7–3’UTR and their miR-30c-5p binding sites mutant versions were synthesized and add to luciferase reporter vector psiCHECK2 (Hanbio Biotechnology, Wuhan, China), named circ3823-WT, circ3823-Mut, TCF7-WT and TCF7-Mut, respectively. The experimental steps follow the manufacturer’s protocols of Dual Luciferase Assay Kit (Hanbio Biotechnology, Wuhan, China). The relative luciferase activity was examined by Full-wavelength multifunctional enzyme label tester SpectraMax M5e (Molecular Devices, Shanghai, China).
RNA immunoprecipitation (RIP)
According to the manufacturer’s instructions of Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore), RIP experiments were conducted with AGO2 antibody (Abcam, Burlingame, USA) and anti-m6A antibody (Synaptic Systems, Goettingen, Goettingen). Co-precipitated circ3823 was applied to qRT-PCR.
Biotin-coupled miRNA capture
The 5′ biotinylated miR-30c-5p pulldown probe or NC pulldown probe were designed and synthesized by Genepharma (Shanghai, China). HCT116 cells were lysed with RIPA and PMSF (Solarbio, Beijing, China) and incubated with miR-30c-5p probes or NC probe. Then cell lysates were incubated with streptavidin-coated magnetic beads (Invitrogen, Thermo Fisher Scientific, Waltham, USA) to pull down the biotin-labelled RNA complex. The RNA was purified with TRIzol (Takara, Dalian, China). Then the abundance of circ3823 and TCF7 was analysed by qRT-PCR.
Statistical analysis
All findings were shown as mean ± standard deviation (SD) and analysed with IBM SPSS Statistics 21 (Chicago, IL, United States), and graphs were generated by GraphPad Prism 7.0. Continuous data analysis was performed with Student’s t-test to analyse differential expression between two groups. ROC curves were used to assess the sensitivity and specificity of circRNA as a diagnostic biomarker. Survival curves were compared with the log-rank test. All statistical tests were 2 sided, and P < 0.05 was considered statistically significant.
Discussion
CircRNA expression profiling is a prerequisite for the identification of novel oncogenic circRNAs and tumour suppressors, as well as in elucidating their mechanisms and functions [
30]. Here, we applied RNA-seq to obtain the expression profiles of circRNA as well as mRNA in CRC tissues and normal tissues. Subsequently, we identified a novel circRNA termed circ3823 which was obviously highly expressed in CRC tissues. And particularly valuable discovery is the expression of circ3823 in serum has high sensitivity and specificity for detecting CRC which means circ3823 has the potential to be used as liquid diagnostic biomarker. Liquid biopsy refers to the analysis of tumours using biomarkers circulating in liquids such as blood, urine, ascites and cerebrospinal fluid [
31]. Circulating tumour cells, exosomes, and nucleic acids (DNA and RNA) have become attractive candidates for liquid biopsy because they have many key characteristics of ideal biomarkers [
32]. The ability to detect and characterize tumours in this minimally invasive and reproducible manner has considerable clinical significance [
33]. Due to the non-invasive and low-cost advantages of liquid diagnostics, circ3823 have the potential to be used for large-scale population screening of CRC in the future. GO functional annotation and KEGG enrichment analysis showed that the enriched function and pathway were related to apparent characteristics of CRC. The apparent characteristics of CRC include: tumour cell proliferation, apoptosis, infiltration and migration, stromal cell adhesion, angiogenesis, and tumour immunity, etc. Among them, angiogenesis is an indispensable inducement for tumour cell growth and metastasis. Our experiments confirmed that circ3823 significantly increase tumour neovascularization both in vivo
and in vitro.
The classic regulatory mechanisms of circRNA include ceRNA, RNA-binding proteins and functional proteins coding. Among these mechanisms, ceRNA hypothesis represents complex post-transcriptional regulatory network. Growing evidences indicated that circRNA regulates mRNA expression by competing for shared MRE [
34]. For example, circCRIM1 upregulating FOXQ1 to promote metastasis and docetaxel chemoresistance of nasopharyngeal carcinoma [
35]. Besides, circ-Erbin mediated HIF-1α activation by miR-125a-5p/4EBP-1 axis [
13]. Moreover, circSOD2 inhibits miR-502-5p expression and rescues the target gene DNMT3a [
36]. In this study, circ3823 exerted its function as a ceRNA that competitively bound to miR-30c-5p, then abolished the endogenous suppressive effect of miR-30c-5p on the target gene TCF7.
TCF7 is a key transcriptional effector of Wnt signaling pathway, which plays a very important regulatory role in tumorigenesis and development. TCF7 protein was identified to be highly expressed in various cancers, such as lung cancer, pancreas cancer, and breast carcinomas [
37‐
40]. In this study, we verified that TCF7 was highly expressed in CRC tissues. Circ3823 significantly increases both the mRNA and protein levels of TCF7 in CRC. In addition, elevated TCF7 promote the expression of MYC and CCND1 in CRC, leading to activation of Wnt signaling pathway. Revealing that circ3823 drive CRC cell proliferation, metastasis and angiogenesis depends on the highly expressed TCF7 in CRC.
We noticed that although the overexpression of circ3823 did not effectively introduce to increase the RNA levels of TCF7, MYC, and CCND1, those proteins were detected clearly in IHC of the xenograft. This result prompted us to think about the reasons for the discrepancy between RNA and protein levels. By consulting literatures, we learned that gene expression is divided into two levels, transcription and translation, that is mRNA level and protein level. First of all, there is a spatial and temporal interval between transcription and translation of eukaryotic genes [
41,
42]. It is possible that the express of protein is still increasing when the mRNA reaches its peak, or that the mRNA is already degraded when the protein level reaches its peak. Secondly, after transcription, there is post-transcriptional processing, such as degradation of transcription products and post-translational modifications. Hence, the transcription level and translation level are not exactly the same [
43‐
45]. In addition, there are negative feedback regulatory mechanisms in the organism, and the cells of the organism need to maintain homeostasis [
46‐
48]. When protein levels are elevated, it is a stress for the cell which inevitably decreases gene transcription in order to maintain homeostasis in the body. Conversely, when protein levels are low, the cell itself may promote transcription.
Currently, most explorations focus on the downstream mechanism of circRNA in tumour progression. However, the upstream mechanisms of circRNA generation, splicing, and degradation are rarely explored. Increasing evidence indicates that m6A modification was involved in regulating mRNA degradation [
49‐
53]. Therefore, we guessed whether circRNA degradation is also regulated by m6A modification? Previous research revealed that YTHDF2 recruits the adaptor protein HRSP12 and RNase P/MRP complex to the cleavage circRNA by reading m6A site. Resulting in the closed loop structure of circRNA was destroyed and rapidly degraded [
54]. The most recent article firstly proposed that N6-methyladenosine modification at specific sites promote the stability of circRNA, but the specific mechanism has not been further explored [
55]. Based on our results of RIP experiments, we found that m6A antibodies specifically enrich circ3823. Moreover, after transfection of si-YTHDF3 and si-ALKBH5, the expression of circ3823 was significantly up-regulated compared with the control group. In addition, according to TCGA and GEO database, we found that YTHDF3 and ALKBH5 were significantly reduced in CRC and YTHDF3/ALKBH5 were positively correlated with YTHDF2. By consulting literatures, we learned that YTHDF3 binds m6A in cells and shares mRNA targets with YTHDF2, YTHDF3 cooperate with YTHDF2 to promote mRNA degradation [
50]. Therefore, we speculate whether YTHDF3 and ALKBH5 cooperate with YTHDF2 to promote the degradation of circ3823. The above conjecture needs to be verified by rigorous experiments, which is the subject of our next exploration.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.