Background
Colorectal cancer has emerged as one of the most prevalent malignancies of the digestive system [
1], and in 2020, more than 1.9 million new colorectal cancer cases and 935,000 deaths were estimated to have occurred worldwide [
2]. In China, the morbidity and mortality of colorectal cancer have remained high in recent years [
3]. Although some approaches have been applied in the clinic, including colonoscopy, sigmoidoscopy and faecal-based testing, the early detection and diagnosis of colorectal cancer remain challenging [
4,
5]. Thus, there is an urgent need to identify novel biomarkers of colorectal cancer.
Exosomes, which are regarded as a valuable source of promising biomarkers, have a size range of 40-120 nm and exist stably in bodily fluids [
6,
7]. Abundant evidence has shown that exosomes act as messengers in cell-to-cell communication by delivering specific cargos derived from parent cells to recipient cells involved in disease development and progression [
8]. Recently, circular RNAs (circRNAs) have been shown to be definitively enriched in exosomes and are easy to measurement in the circulation, and with covalently closed loop structures without 5′ to 3′ polarity or a polyadenylated tail, they are highly stable compared with their parental linear RNAs [
9‐
11]. Moreover, circRNAs can be specifically and differentially expressed under various pathologic conditions across tissues and fluids, including colorectal cancer [
12,
13], thus serving as potential biomarkers of disease progression for use in diagnostics [
14,
15]. Accordingly, exosomal circRNAs are novel in the frontier of cancer biomarkers with promise for use in both clinical applications and studies of disease aetiology [
11]. Currently, few studies have reported on the diagnostic role of exosomal circRNA in colorectal cancer [
16,
17]. For example, Shang et al. found that exosomal circPACRGL plays an oncogenic role in colorectal cancer and could be a potential marker for colorectal cancer [
18]. However, the validation sample size of these studies is limited, and the role of exosomal circRNAs as a new biomarker for colorectal cancer has not been systematically studied.
In this study, we initially performed high-throughput RNA sequencing (RNA-Seq) of tissues, and the results were validated through tissue microarray, cell line and exosome analyses. Subsequently, we evaluated the diagnostic capabilities of target exosomal circRNAs in plasma obtained from cancer-free controls, precancer individuals, colorectal cancer patients and patients with other types of cancer. Furthermore, in vitro and in vivo experiments with cell and mouse models were conducted to comprehensively clarify the biological function of a target exosomal circRNA involved in colorectal tumorigenesis.
Materials and methods
Study populations
Patients with colorectal cancer (
n = 112) or polyps (
n = 28) were randomly selected from the Nanjing ColoRectal Cancer (NJCRC) cohort, which is a long-term follow-up clinical cohort [
19,
20]. Briefly, the NJCRC study followed inclusion criteria indicated that participants must have been recently diagnosed by clinical and histopathology and not treated for colorectal cancer patients during surgical resection. Cancer-free control individuals (
n = 60) were recruited from the same geographical region during the same period in which other study participants were recruited. Moreover, patients with other cancers were consecutively recruited from the Collaborative Innovation Center for Cancer Personalized Medicine of Nanjing Medical University, including patients with gastric carcinoma (GC,
n = 74), breast invasive carcinoma (BRCA,
n = 18), bladder urothelial carcinoma (BLCA,
n = 24), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC,
n = 32), kidney renal clear cell carcinoma (KIRC,
n = 19), and lung adenocarcinoma (LUAD,
n = 42). Corresponding tumor/normal tissues and peripheral blood samples were collected, and plasma was obtained by centrifuging peripheral blood at 3000 rpm at 4 °C for 10 min. We also collected plasma samples from 25 colorectal cancer patients before and after surgical treatment. Written informed consent was obtained from all subjects recruited for this study, and the study was authorized by the Institutional Review Board of Nanjing Medical University (approval No. 2017-515).
Transcriptomics and proteomic analysis
A total of 52 pairs of colorectal tumors and normal adjacent tissues (NATs) were subjected to RNA-Seq to identify differentially expressed circRNAs. The circRNAs profiled were sequenced and generated with an Illumina HiSeq 2500 platform for 150-bp sequencing in collaboration with Genesky Biotechnologies, Inc. (Shanghai, China). The Find_circ [
21] and CIRI2 [
22] algorithms were used to detect and identify circRNAs. The raw counts were normalized on the basis of transcripts per kilobase million (TPM), and the differential expression profiles of the circRNAs were obtained using the R package “edgeR”. The TPM normalization method was used to normalize the RNA-Seq data because this method respects the average invariance and prevents the biological signal compared with other normalization methods [
23‐
25]. In addition, a tandem mass tag (TMT)-based quantitative proteomic analysis (Novogene Co., Ltd., China) was performed to identify and quantitate proteins based on 25 pairs of colorectal tumors and NATs (the samples were from RNA-Seq samples).
Fluorescence in situ hybridization (FISH) and immunofluorescence
FISH with 5-carboxyfluorescein (5-FAM)-labelled probe sequences specific for circLPAR1 (GenePharma, China) was performed to detect the expression and visualize the localization of circLPAR1 in a tissue microarray with 79 paired colorectal tumor tissues and NATs. The probe was dropped onto slides, and hybridization was performed overnight at 60 °C in a moist chamber. After hybridization, the slides were washed twice once with 2 × saline-sodium citrate (SSC) for 5 min, once in 0.5 × SSC for 15 min, and twice in 0.2 × SSC for 15 min. The slides were then subsequently fluorescein isothiocyanate (FITC)-streptavidin-biotin complex (SABC) at 37 °C for 45 min. The immunofluorescence accumulation optical density (IOD) of circLPAR1 was evaluated by Image-Pro Plus 7.0 software (Media Cybernetics, Inc., USA) [
26]. In addition, RNA-FISH was carried out by using a FISH Kit (RiboBio Inc., China) to assess the location of circLPAR1 in colorectal cancer cells. To observe the colocalization of circLPAR1 and eukaryotic translation initiation factor 3 subunit h (eIF3h), DLD1 cells were transiently transfected with Cyanine3 (Cy3)-labelled circLPAR1 (GenePharma, China) and then incubated with anti-eIF3h antibody.
Exosome isolation, characterization, and internalization
Exosomes were isolated from the culture medium of FHC, HCT116 and DLD1 cells by using an ExoQuick TC kit (SBI, USA), and exosomes from the plasma of subjects were purified by using an ExoQuick Plasma Prep with Thrombin kit (SBI, USA). The size distribution of the exosomes was characterized by nanoflow cytometry using a U30 Flow NanoAnalyzer (NanoFCM, Inc., China) with technical assistance provided by KeyGEN Biotech Co. Ltd. (Jiangsu Province, China). The shape and size of the exosomes were observed by transmission electron microscopy (TEM) (Tecnai G2, FEI, USA). Moreover, the characterization of the exosomes was confirmed by the presence of exosomal protein markers TSG101 (# ab125011, Abcam, USA) and Alix (# 92880, CST, USA). The green fluorescent dye PKH67 (Umibio, China) was utilized to label exosomes isolated from the culture medium of cells. After dye was incubated with recipient cells for 3 h, fluorescence microscopy (Zeiss, Germany) was performed to visualize PKH67-labelled exosomes in recipient cells. The detailed procedures were described in our previous study [
27].
Stability determination of plasma exosomal circLPAR1
To determine the stability of circLPAR1, colorectal cancer cells were treated with 6 U/6 μg RNase R (Epicentre Biotechnologies, USA) at 37 °C for 10 min. Moreover, 2 μg/ml actinomycin D (AbMole, China) was added to colorectal cancer cells and incubated for 1 h, 3 h, 6 h, and 24 h to inhibit new RNA synthesis. To determine the stability of plasma exosomal circLPAR1, plasma samples from subjects (colorectal cancer patients, precancer individuals and cancer-free control individuals) were subjected to extreme conditions, including incubation at room temperature for 0 h, 4 h, 8 h, and 24 h and repeated cycles (0, 2, 4, and 8 cycles) of freezing and thawing between − 80 °C and room temperature.
CircLPAR1 pull-down assay and mass spectrometry analysis
The RNA-binding proteins (RBPs) associated with circLPAR1 were determined by using a circRNA pull-down assay with MS2-capturing protein (MS2-CP) [
28,
29]. Briefly, two overexpression vectors, one carrying circLPAR1-MS2 and the other carrying MS2-CP-Flag, were constructed and labelled with green fluorescent protein (GFP) and red fluorescent protein (m-Cherry), respectively (Geneseed, China). These two vectors were co-transfected into HCT116 cells to induce MS2-CP expression. After specific binding between MS2-labelled circLPAR1 and MS2-CP, the MS2-CP-MS2-circLPAR1 complex was pulled down by an anti-Flag antibody. Lysate derived from HCT116 cells without the MS2 flagging system was used as the control. Subsequently, the captured products were identified by real-time quantitative polymerase chain reaction (RT-qPCR) with a fluorescent reporter or Western blotting. Then, the circLPAR1 pull-down complex and its control were analysed by mass spectrometry (Q Exactive, Thermo Scientific, USA).
RNA immunoprecipitation (RIP) assay
The RIP assay was performed with a Magna RIP kit (Millipore Magna, USA) to determine the interaction between exosomal circLPAR1 and eIF3h. In brief, lysates of HCT116 and DLD1 cells with stable overexpression of circLPAR1 or exosomal circLPAR1 were incubated with RIP buffer containing anti-eIF3h antibody (# 3413, CST, USA)- or control IgG (# ab172730, Abcam, USA)-conjugated magnetic beads. The immunoprecipitated RNA was then analysed by RT-qPCR.
Effects of exosomal circLPAR1 on mouse models of colorectal cancer
A total of 24 female NCG mice (4–5 weeks old) were randomly divided into the following four groups: NC; circLPAR1; circLPAR1-Exos; and NC-Exos. On days 1, 3 and 5, 1 × 107 circLPAR1 or NC stably expressed DLD1 cell lines were resuspended in 0.1 ml of phosphate-buffered saline (PBS) and injected into the right flank of the mice. PBS was injected into mouse tumors 15, 17 and 19 days after mouse tumor induction, and these tumors were used as the NC and circLPAR1 groups, respectively. To assess the effect of exosomal circLPAR1 on colorectal cancer tumorigenesis, an equal DLD1 cell concentration was injected into the right flank of normal mice on days 1, 3 and 5. Exosomes isolated from DLD1 cells stably transfected with circLPAR1/NC vector were injected into mouse tumors 15, 17 and 19 days after tumor induction, and these tumors were used as the circLPAR1-Exos and NC-Exos groups, respectively. The width and length of the tumors in each group were measured once every 2 days, and the tumor volume was calculated using the following formula: tumor volume (cm3) = (L × W2)/2, where L is length and W is width. All animal studies were approved by the Institutional Animal Care and Use Committee of Nanjing Medical University (IACUC-2012051).
Statistical analysis
Quantitative data are presented as the means ± standard deviations (SDs) and the group comparison was evaluated by Student’s
t-test (two-tailed). Unconditional univariate Cox regression analysis was utilized to obtain hazard ratios (HRs) and Kaplan-Meier analysis was used to evaluate the effect of circLPAR1 on the overall survival of colorectal cancer patients. The receiver operating characteristic (ROC) curve analysis was used to calculate the area under the curve (AUC) value to assess the capability for discriminating colorectal cancer patients from cancer-free controls. The correlation between circLPAR1 and RBPs in human tissues was determined with Spearman’s correlation analysis. A two-sided
P < 0.05 was considered statistically significant. Statistical analyses were performed with R software (version 3.6.0). Additional experimental methods are described in the
Supplementary Materials and Methods.
Discussion
In this study, we identified a novel circRNA, circLPAR1, stably carried in plasma exosomes, and it emerged as a potential diagnostic biomarker specific for colorectal cancer. Mechanistically, exosomal circLPAR1 is internalized by colorectal cancer cells and binds eIF3h to reduce METTL3-eIF3h-dependent mRNA translation, thereby inhibiting
BRD4 expression and suppressing cellular proliferation, invasion, and migration (Fig.
5H).
Circular lysophosphatidic acid receptor 1 (circLPAR1, hsa_circ_0087960), has rarely been reported to be in tumors; however, one study on muscle-invasive bladder cancer showed that a decreased circLPAR1 level in bladder tumors was associated with bladder cancer prognosis [
35]. Similarly, according to the evidence we obtained from circRNA expression profiling and a FISH analysis, circLPAR1 was expressed at low levels in colorectal cancer tissues and that its lower expression was associated with poor survival in patients with colorectal cancer. These observations suggest that circLPAR1 may be a circRNA specific to colorectal cancer. Notably, circLPAR1 was significantly lower in plasma exosomes obtained from patients with bladder cancer than in those from cancer-free control individuals, which is consistent with the report of Lin et al. [
35].
The advantages of exosomal circLPAR1 as a biomarker are based on three aspects. First, circLPAR1 was resistant to actinomycin D and RNase R, indicating that it is more stable than its cognate linear LPAR1 mRNA; circLPAR1 was encapsulated in exosomes, which conferred even greater protection against degradation. These characteristics rendered circLPAR1 exceptionally stable in plasma exosomes in multiple samples, such as the samples obtained from colorectal cancer, precancer, or cancer-free control individuals. Second, plasma exosomal circLPAR1 could distinguish colorectal cancer patients from cancer-free control individuals with acceptable sensitivity and specificity. Intriguingly, the plasma exosomal circLPAR1 level in colorectal cancer patients was significantly higher than that in patients with other types of cancer. Third, exosomal circLPAR1 significantly suppressed colorectal cancer cellular phenotypes, which is helpful both for exploring the biological mechanism of colorectal cancer and for offering new therapeutic approaches. These results collectively support the supposition that exosomal circLPAR1 can be used as a specific biomarker for colorectal cancer diagnosis.
An increasing number of studies have revealed that exosomes derived from certain parent cells can reach corresponding recipient cells in which they release their cargos to influence biological functions and induce a series of phenotypic changes [
36,
37]. In this study, we found that circLPAR1 expression was dramatically elevated in exosomes isolated from colorectal cancer cell lines transfected with circLPAR1-expressing lentivirus or vector. In contrast, treatment with GW4869, a noncompetitive, specific, potent inhibitor of membrane neutral sphingomyelinase (nSMase) that blocks exosome production [
38], resulted in significantly reduced expression of circLPAR1 in cell culture medium. Moreover, we noticed that transfection of circLPAR1 overexpression vectors into parent cells promoted the capacity of exosomal circLPAR1 to inhibit the acquisition of colorectal cancer malignant phenotypes, which was consistent with the effect of the direct overexpression of circLPAR1. In addition, circLPAR1 expression was much higher in the tumor tissues of circLPAR1-Exos xenograft-bearing mice than in NC-Exos group. These observations indicate that extracellular circLPAR1 can inhibit the acquisition of multiple malignant phenotypes of colorectal cancer when it is incorporated into exosomes.
The eIF3h was reported as one of the subunits of the eIF3 complex, which moderates several steps of initiation and elongation during protein synthesis [
39]. Evaluating the MS2-CP circRNA pull-down assay, mass spectrometry analysis and proteomic profiling validation, we found that eIF3h directly binds to circLPAR1 and negatively correlates with circLPAR1. Notably, previous studies have shown that METTL3 can interact with the eIF3h subunit at the 5′-end of
BRD4 mRNA to enhance translation and promote oncogenesis [
34].
BRD4, a member of the bromodomain and extra-terminal domain (BET) family, plays a crucial role in the development of multiple cancers and is regarded as a novel cancer therapeutic target [
40,
41]. In addition,
BRD4 has been shown to promote the proliferation of colorectal cancer cells, while the
BRD4 inhibitor AZD5153 reversed this effect [
42]. Our results extended the findings showing that exosomal circLPAR1 binds eIF3h to inhibit the METTL3–eIF3h interaction, which decreases the translation of
BRD4, thereby suppressing colorectal cancer cell proliferation, invasion and migration, and experiments with in vivo models showed consistent results. Although Lin et al. found that circLPAR1 could bind to miR-762 and moderated the invasion of bladder cancer cells [
35], we found that miR-762 was not enriched with circLPAR1 pull-down products. This inconsistent finding might be a result of different molecular mechanisms of circLPAR1 action in different types of cancer. In addition, more flexible screening criteria will be used in further RNA-Seq analyses, such as survival analysis or functional analysis, to obtain more effective circRNAs.
In summary, this study highlighted a novel colorectal cancer-associated circRNA, termed exosomal circLPAR1, which was easily and reliably detected in exosomes obtained from plasma and might act as a promising noninvasive biomarker for population screening and early detection of colorectal cancer. Mechanistically, exosomal circLPAR1 expression level led to decreased BRD4 levels because it binds eIF3h and inhibits the METTL3–eIF3h interaction, which remarkably suppressed colorectal cancer development. This study not only reports a novel mechanism of exosomal circRNA in colorectal cancer, but also opens a new avenue for screening strategies and diagnostic approaches to colorectal cancer.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.