Hsa_circRNA_002144 promotes growth and metastasis of colorectal cancer through regulating miR-615-5p/LARP1/mTOR pathway

Abstract CircRNAs (circular RNAs), recently identified as a critical regulator in tumorigenesis, participate in CRC (colorectal cancer) growth. However, the role of hsa_circRNA_002144 in CRC was poorly understood. Firstly, hsa_circRNA_002144 showed significantly elevation in both of CRC tissues and cell lines, and suggested closely associated with poor prognosis in patients. Secondly, data from functional assays revealed that silence of hsa_circRNA_002144 inhibited CRC progression with reduced cell viability, proliferation, migration and invasion, while enhanced cell apoptosis. In addition, in vivo CRC growth and metastasis were also suppressed by knockdown of hsa_circRNA_002144. However, CRC progression was promoted with over-expression of hsa_circRNA_002144. Thirdly, hsa_circRNA_002144 colocalized with miR-615-5p in the cytoplasm of CRC cells, and decreased miR-615-5p expression. Moreover, miR-615-5p could target LARP1 (La ribonucleoprotein 1, translational regulator). Lastly, the suppressive effects of hsa_circRNA_002144 knockdown on CRC progression were reversed by LARP1 over-expression. In conclusion, hsa_circRNA_002144 could sponge miR-615-5p to promote CRC progression through the regulation of LARP1, providing a therapeutic target for cancer intervention.


Introduction
Colorectal cancer (CRC) is one of the most common cancers in the world (1). Although, the recovery rate is high after early CRC resection through multidisciplinary therapy (radiotherapy and systemic chemotherapy), high recurrence rate and distant metastasis still threaten a large proportion of patients (2). Efforts have been made to elucidate the genetic and molecular characteristics of CRC in order to predict prognosis and response to targeted therapy (3,4), however, tumor heterogeneity suggests the need for new CRC markers and therapeutic targets (5).
CircRNAs, highly conserved and stable covalently closed RNA transcript, formed by the reverse splicing of a single pre-mRNA with gene regulatory potential (6). Cumulative evidences suggest that circRNAs are closely associated with human diseases, especially cancer, and may be a better biomarker due to its abundance in tumor tissues or plasma and stability (7,8).
Recently, circRNAs are involved in metabolic dysregulation and tumor development of CRC (9). For example, hsa_circ_001988 was decreased in CRC and closely associated with the clinical characteristics of patients (10). CircHIPK3 (11) or hsa_circ_0000069 (12) could promote CRC growth and metastasis, while circRNA CBL.11 suppressed CRC proliferation (13). Therefore, circRNAs could function as potential biomarkers and therapeutic targets for CRC. In addition, researches on more circRNAs associated with CRC progression may contribute to clinical value.
Hsa_circRNA_002144, a new circRNA produced by RPPH1 gene, was localized at chr14:20811282-20811431-and firstly found to be upregulated in hepatocellular carcinoma (14). Data from CircBase showed that hsa_circRNA_002144 was also upregulated in CRC plasma (15) and tissues (16). However, the detailed role of hsa_circRNA_002144 on CRC progression remains enigmatic. Generally, circRNAs could function as microRNA sponges to sponge miRNAs via the miRNAs response elements, and miRNAs could also target mRNAs to form a network responsible for CRC progression (17). Hsa_circRNA_002144 could sponge miR-326-mediated ETS transcription factors to promote cervical cancer progression (18). Hsa-miR-615-5p was reported to be a binding target of hsa_circRNA_002144 in hepatocellular carcinoma (14). MiR-615-5p was reported to suppress angiogenesis to regulate tumor microenvironment of CRC (19), and highly miR-615-3p expression was associated with poorly differentiated CRC (20). However, whether miR-615-5p was involved in hsa_circRNA_002144-mediated CRC progression, and the target mRNAs of miR-615-5p in CRC, remains elusive.
This study firstly verified the oncogenic role of hsa_ circRNA_002144 in CCRC progression, the potential miR-615-5p-mediated mRNA target was then validated and suggested that hsa_circRNA_002144 functioned as a microRNA sponge to enhance CRC progression, providing the potential therapeutic target.

Tumor tissues collection
Protocols were approved by the Ethics Committee of Hainan Affiliated Hospital of Hainan Medical University and in accordance with those of the 1964 Helsinki Declaration and its later amendments for ethical research involving human subjects. Sixty pairs of CRC and adjacent non-tumorous tissues were obtained from patients at Hainan Affiliated Hospital of Hainan Medical University. Patients with written informed consents were assigned with stage I, II, III and IV based on TNM (tumor-node-metastasis) classification.

Cell proliferation
T84 or LoVo cells with different treatment (5 × 10 2 cells/well) were seeded and cultured with RPMI 1640 medium for 2 weeks. Cells were fixed in 10% formaldehyde and then stained with 0.1% crystal violet. Cell colonies were visualized under microscope (Olympus, Tokyo, Japan).

Wound healing
T84 or LoVo cells with different treatment (5 × 10 3 cells/well) were seeded for 24 h and scratched via a plastic pipette tip. Washed with phosphatebuffered saline buffer, cells were cultured for another 24 h and calculated the wound width under microscope (Olympus).

Transwell assay
T84 or LoVo cells with different treatment (5 × 10 4 cells) were incubated on upper chambers (BD Biosciences, San Jose, CA, USA) with serum-free medium coated with Matrigel. Medium with 10% fetal bovine serum was added to the lower chambers. Filters were removed 8 h later, and the invasive cells to the lower chambers were fixed in 100% methanol 24 h later. After staining with 0.1% crystal violet, cells were counted under microscope (Olympus).

Dual luciferase reporter assay
Wild-type or mutant sequences of hsa_circRNA_002144 or 3′-UTR of LARP1 were subcloned into pmirGLO luciferase reporter vector (GenePharma, Suzhou, China). T84 or LoVo cells were co-transfected miR-615-5p mimics or NC mimic with the vectors. Luciferase activities were performed 48 h after transfection.

Mouse xenograft assay
Animal study was approved by the Ethics Committee of Hainan Affiliated Hospital of Hainan Medical University and in accordance with the National Institutes of Health Laboratory Animal Care and Use Guidelines. Twelve 5-week-old female BALB/c nude mice (20-25 g) were randomly divided into two groups (n = 6). T84 cells (5 × 10 6 /0.1 mL PBS) with stable knockdown of hsa_circRNA_002144 or shNC were subcutaneously injected into right flank of nude mice. Ten days later, tumor volume was calculated every two days. Twenty days later, the mice were killed with 40 mg/kg sodium pentobarbital, the tumor tissues were isolated, weighed and photographed. For detection of metastasis, the T84 cells were injected into NOD/SCID mice by tail vein. After intraperitoneal injection of luciferin (4 mg/50 μL phosphate buffered saline; Promega) for 10 min, the metastases were photographed by IVIS Lumina II system (Caliper LifeSciences, Hopkinton, MA, USA). Lung tissues were also isolated, and fixed in 10% formalin and embedded in paraffin. The sectioned lung tissues were subjected to H & E staining, and the representative images were observed via microscope (Olympus).

Immunohistochemistry
Formalin-fixed and paraffin-embedded tumor sections from the mice were incubated in 3% H 2 O 2 and then blocked in 5% bovine serum albumin. The sections were incubated overnight with anti-LARP1, anti-Ki67, anti-E-cadherin, anti-N-cadherin or anti-mTOR antibodies (Abcam). After incubation with horseradish peroxidase labeled secondary antibody, the slides were examined under a light microscope (Olympus) with counterstaining with hematoxylin.

Statistical analysis
Data were expressed as mean ± SEM, and processed by SPSS 19.0 software. The statistical analyses were determined by Student's t-test, one-way analysis of variance, Pearson's chi-squared test, non-parametric test and the survival curves were analyzed by Kaplan-Meier method and log-rank test. Value of P < 0.05 was considered to be statistically significant.

Expression of hsa_circ_002144 in CRC
According to circRNA dataset of GEO database (GSE126095; 10 of patients and 10 of normal) (21), hsa_circ_002144 was significantly elevated in CRC patients ( Figure 1A). qRT-PCR analysis confirmed the up-regulation of hsa_circ_002144 in 60 CRC tissues ( Figure 1B). Group of highly hsa_circ_002144 expression (n = 30) was closely associated with shorter overall survival compared with group of lowly has_circ_002144 (n = 30) (P = 0.0357) through Kaplan-Meier curves analysis ( Figure 1C), tumor size (P = 0.004), lymph node metastasis (P = 0.004), distant metastasis (P = 0.028) and TNM stage (P = 0.004) (Supplementary Table S1 is available at Carcinogenesis Online). Moreover, Univarite COX regression analysis demonstrated that hsa_circ_002144 was related to prognosis in patients with colorectal cancer (Supplementary  Table S2 is available at Carcinogenesis Online). However, multivariate Cox regression analysis showed that hsa_circ_002144 was not an independent prognostic biomarker for colorectal cancer (Supplementary Table S2 is available at Carcinogenesis Online). These results suggested that hsa_circ_002144 might be involved in metastatic property of CRC. A significant up-regulation of hsa_circ_002144 was also verified in CRC cell lines (T84, LoVo, SNU-175, CL-34) ( Figure 1D), and the circular nature of hsa_circ_002144 was validated in both of T84 and Lovo cells followed by RNase R digestion ( Figure 1E). As shown in Figure 1F, hsa_circ_002144 was mainly expressed in cytoplasm of T84 and Lovo cells, suggesting that hsa_circ_002144 might function as miRNA sponger to regulate target genes in the nucleus of CRC cells.

Hsa_circ_002144 promoted malignant behaviors of CRC
Function roles of hsa_circ_002144 on CRC progression were then determined in T84 transfected with specific shRNAs targeting hsa_circ_002144 or Lovo transfected with pcDNA-hsa_circ_002144. The transfection efficiency was validated by qRT-PCR in Figure 2A. Over-expression of hsa_circ_002144 promoted cell viability ( Figure 2B) and proliferation ( Figure  2C) of Lovo cells, while knockdown of hsa_circ_002144 suppressed the cell viability ( Figure 2B) and proliferation ( Figure  2C). In addition, cell apoptosis was substantially decreased in Lovo cells ( Figure 2D), while the apoptosis was significantly promoted in T84 cells ( Figure 2D). Protein expression of caspase-3 and cleaved caspase-3 were decreased in T84 cells by knockdown of hsa_circ_002144, while increased in Lovo cells by over-expression of hsa_circ_002144 ( Figure 2E). The metastatic capacities of T84 and Lovo were also suppressed by knockdown of hsa_circ_002144 ( Figure 3A and B), while over-expression of hsa_circ_002144 promoted the metastatic capacities ( Figure 3A and B)B. Protein expression of E-cadherin was increased in T84 cells with knockdown of hsa_circ_002144 ( Figure 3C), and decreased in Lovo cells with over-expression of hsa_circ_002144 ( Figure 3C). However, hsa_circ_002144 demonstrated a reversed effect on protein expression of N-cadherin compared with E-cadherin ( Figure 3C). All these results indicated the suppressive effects of hsa_circ_002144 knockdown on malignant behaviors of CRC.

Knockdown of hsa_circ_002144 suppressed CRC tumor growth
To further assess effect of hsa_circ_002144 on in vivo CRC tumor growth, xenograft mouse model was established and injected with T84 cells stably silence of hsa_circ_002144. Hsa_circ_002144 was significantly enhanced whereas miR-615-5p was reduced in sh 1#-hsa_circ_002144 group mice compared with shNC group (Supplementary Figure S1A  Immunohistochemical analysis showed decrease of LARP1, Ki-67, N-cadherin and mTOR, while increase of E-cadherin, in tissues from mice injected with T84 cells stably silence of hsa_ circ_002144 compared with shNC (Supplementary Figure S1E is available at Carcinogenesis Online). These results showed that knockdown of hsa_circ_002144 suppressed CRC tumor growth and metastasis.

Discussion
Increasing evidence has suggested that circRNAs could function as remarkable prognostic biomarkers of CRC (22). Moreover, the circular nature of circRNAs demonstrates resistance to RNAase, thus showing longer half-life than linear RNA and predicting as an attractive biomarker in CRC (23). Recently, circRNAs have also been verified to be associated with CRC development. ciRS-7-A not only functions as a promising prognostic biomarker in CRC, but also provides as a potential therapeutic target for CRC (24). Notably, a novel circRNA, hsa_circ_002144, was up-regulated in CRC (15,16), and hsa_circ_002144 could promote proliferation and invasion of cervical cancer, while suppressing apoptosis and autophagy (18). Nevertheless, the role and mechanism of hsa_circ_002144 on CRC growth required further exploration. In line with previous reports (15,16), hsa_circ_002144 was dramatically up-regulated in CRC tissues, and this study for the first time revealed that highly hsa_circ_002144 expression was closely associated with metastatic properties of CRC, such as tumor size, lymph node metastasis, distant metastasis and TNM stage. Moreover, highly hsa_circ_002144 expression was also closely associated with shorter overall survival of CRC patients, suggesting that hsa_circ_002144 might serve as a potential biomarker for CRC prognosis.
Oncogenic role of hsa_circ_002144 in cervical cancer has been reported before (18), the present study showed that hsa_ circ_002144 promoted cell viability, proliferation, migration and invasion of CRC cells, while induced the cell apoptosis. Moreover, the oncogenic role of hsa_circ_002144 in the invasion of CRC was associated with the promotion of epithelial-mesenchymal transition via an increase of N-cadherin and E-cadherin. Reports before have shown that CRC cells undergo epithelial-mesenchymal transition during local invasion (25,26). Reduce of E-cadherin and enhance of N-cadherin was shown to be associated with lymph node metastasis in CRC, and contribute to malignant progression of CRC (27). Results from in vivo tumor model revealed that silence of hsa_circ_002144 enhanced E-cadherin and reduced N-cadherin, thus inhibiting epithelial-mesenchymal transition of CRC. Moreover, knockdown of hsa_circ_002144 inhibited in vivo tumor growth and metastasis of CRC, suggesting potential clinical application of hsa_circ_002144 in CRC.
A potential circRNA-miRNA-mRNA network has been reported to be involved in CRC development (16). Hsa_circ_002144 has been shown to sponge miR-326 to regulate ETS transcription factor during cervical cancer progression (18). The underlying miRNA-mRNA network involved in hsa_circ_002144-mediated CRC progression was then clarified. Consistent with previous research that miR-615-5p was a potential binding target of hsa_circ_002144 (14), data from this study confirmed the direct interaction between hsa_circ_002144 and miR-615-5p in CRC. MiR-615-5p was widely considered as a tumor suppressor in hepatocellular carcinoma (28), ovarian cancer (29), non-small cell lung cancer (30), pancreatic ductal adenocarcinoma (31) and esophageal squamous cell carcinoma (32). Considering the angioregulatory role of miR-615-5p in CRC (19), miR-615-5p might also function as CRC tumor suppressor. Our results showed that miR-615-5p could target LARP1 to regulate CRC progression.
LARP1 is a conserved RNA-binding protein to control ribosome biogenesis (33), and participates in cancer cell survival (34), thus representing cancer therapeutic target (35). LARP1 has been reported to be prognosis biomarker of CRC, and silence of LARP1 inhibited cell proliferation of CRC (36). Here, this study showed that over-expression of LARP1 could counteract the suppressive effects of hsa_circ_002144 silence on CRC development. Therefore, hsa_circ_002144-miR-615-5p-LARP1 network was closely associated with CRC progression. Moreover, LARP1 could regulate mRNA stability and translation of mTOR to regulate cell growth and proliferation (37,38). LARP1/ mTOR has been shown to contribute to epithelial cancer progression (39). Here, our results showed that knockdown of hsa_ circ_002144 could promote miR-615-5p expression to decrease LARP1/mTOR during the suppression of CRC progression. mTOR is regulated by phosphatidylinositol 3-kinase/Akt pathway to participate in proliferation, angiogenesis and metastasis of CRC (40) and inhibition of mTOR could emerge as a potential strategy for cancer intervention of CRC (41)(42)(43). Whether phosphatidylinositol 3-kinase/Akt pathway was involved in hsa_circ_002144-miR-615-5p-LARP1-mTOR axis-mediated CRC progression should be investigated in further study. In summary, knockdown of the oncogene, hsa_circ_002144, inhibited the development of HCC via sponging miR-615-5pmediated LARP1-mTOR axis. This finding provided potential application of hsa_circ_002144 in CRC.

Supplementary material
Supplementary data are available at Carcinogenesis online. Figure S1.  Table S1. Relationship between hsa_circ_002144 and clinicopathological parameters of CRC patients. Table S2. COX analysis of hsa_circ_002144 in colorectal cancer.

Funding
Not applicable.