Long intergenic non-coding RNA LINC00485 exerts tumor-suppressive activity by regulating miR-581/EDEM1 axis in colorectal cancer

Long non-coding RNAs (lncRNA) play a vital role in colorectal cancer (CRC) progression. To investigate the role of long intergenic non-coding RNA LINC00485 in CRC, we performed in vitro functional experiments. LoVo tumor-bearing and liver metastasis mice were used as in vivo models. We found that LINC00485 expression was significantly lower in CRC tissues and cancer cells than in paired normal samples and human normal colonic epithelial cells. Lower expression of LINC00485 predicted poor prognosis in CRC patients. LINC00485 knockdown promoted the proliferation, migration, and invasion of FHC cells, while LINC00485 overexpression weakened these abilities of LoVo cells. MicroRNA miR-581 was the downstream target of LINC00485, which was downregulated in CRC samples and cancer cells compared to normal tissues and normal colonic epithelial cells. MiR-581 overexpression induced proliferation, migration, and invasion of FHC cells, while miR-581 antagomir treatment produced opposite results. MiR-581 directly targeted the 3'UTR of EDEM1 and inhibited its expression and induction of epithelial-mesenchymal transition of CRC. In mouse models, LINC00485 knockdown or down-regulation of miR-581 significantly repressed CRC cell growth and prevented CRC liver metastasis. Overall, LINC00485 suppressed CRC tumorigenesis and progression by targeting the miR-581/EDEM1 axis. LINC00485 may be a potential therapeutic target for CRC.


INTRODUCTION
Colorectal cancer (CRC) is a common digestive system neoplasm (DSN) that threatens human health globally [1]. Epidemiological statistics have indicated that the morbidity and mortality of CRC rank third and fourth among malignancies, respectively [2]. The pathogenesis of CRC is complex and has not been fully elucidated. AGING the expression of downstream genes [9]. LincRNAs have been widely identified to regulate tumorigenesis and cancer progression [10][11][12]. LINC00485 is a newly discovered lincRNA. A genome-wide association study identified a new locus on 12q23.2 (LINC00485), which was associated with uterine leiomyoma [13]. In addition, studies in lung adenocarcinoma (LAC) cells have shown that LINC00485 directly binds to miRNA-195 to upregulate checkpoint kinase 1 (CHEK1) expression, contributing to LAC cell proliferation and cisplatin resistance [14]. By contrast, we found that the expression of LINC00485 in CRC tissues was significantly downregulated when compared to adjacent normal tissues. MiR-581, a predicted target for LINC00485, was aberrantly upregulated in CRC samples. A significant negative correlation was found between LINC00485 and miR-581 expression. Interestingly, down-regulated miR-581 interfered with hepatocellular carcinoma development [15]. However, it remains unclear whether LINC00485 directly targets miR-581 and how it exerts biological activities against CRC progression. Hence, we performed a series of in vitro and in vivo experiments to uncover the role of LINC00485 and miR-581 and to provide new insight into the treatment and diagnosis of CRC.
In this study, we found that LINC00485 worked as a competing endogenous RNA (ceRNA) against miR-581 thereby up-regulating EDEM1 expression in CRC cells, thus promoting proliferation, migration, invasion, and the epithelial-mesenchymal transition (EMT) process of CRC cells. LINC00485 can be used as a therapeutic target for CRC treatment.

LINC00485 is down-regulated in CRC tissues and cells
To understand the role of LINC00485 in CRC, we collected normal paracancerous tissue and tumor tissues from 52 patients with CRC. We found that the expression of LINC00485 in CRC tissues was significantly lower than that in adjacent normal tissues ( Figure 1A). There was no significant correlation between LINC00485 expression and clinical parameters such as patient's age and sex (data not shown), but LINC00485 expression was strikingly reduced with tumor stage ( Figure 1B). Next, lower and higher LINC00485 expression groups were defined according to median LINC00485 expression values. The data indicated that CRC patients with higher expression of LINC00485 survived longer than patients with lower expression of LINC00485 ( Figure 1C). To determine the subcellular localization of LINC00485 in CRC cells, cells were examined by FISH assay using fluorescent probes. We found that LINC00485 was mainly expressed in the cytoplasm of CRC cells (SW480, LoVo) ( Figure 1D) and human normal colorectal epithelial cells (FHC) (Supplementary Figure 1). Moreover, LINC00485 levels in CRC cells (LoVo, SW460, HCT8 cell lines) were markedly lower than in human normal colorectal epithelial cell lines (FHC, NCM460, CCD-18co cells) ( Figure 1E). We further demonstrated that shRNA-mediated knockdown of LINC00485 could enhance the migratory ability of FHC cells, while overexpression of LINC00485 significantly attenuated the migration of LoVo cells ( Figure 1F-1I). These findings suggested that LINC00485 was implicated in CRC progression.

LINC00485 directly targes microRNA-581
To determine the mechanism underlying the involvement of LINC00485 in CRC progression, the biological prediction website DIANA-LncBase v2 [16] was used to predict targets of LINC00485. The results showed that miR-581 shared complementary binding sites with LINC00485 ( Figure 2A). Dual-luciferase reporter assays and RNA immunoprecipitation (RIP) analysis confirmed the interaction between LINC00485 and miR-581 ( Figure 2B-2D). MiR-581 was highly expressed in CRC tissues compared with normal tissues ( Figure 2E) and was significantly associated with tumor stage ( Figure 2F). CRC patients with miR-581 levels above median values have significantly worse prognosis ( Figure 2G). We observed a negative correlation between LINC00485 and miR-581 in CRC ( Figure 2H). Further, our data showed that miR-581 levels were significantly higher in CRC cell lines and specimens than in human normal colorectal epithelial cell lines and normal samples ( Figure 2I-2K). In addition, LINC00485 knockdown significantly increased the expression of miR-581 in FHC cells, and the overexpression of LINC00485 decreased miR-581 levels in LoVo cells ( Figure 2L, 2M). Treatment with miR-581 mimics or antagomir strongly reduced or elevated miR-581 levels, neither of which had any effect on LINC00485 expression; however, miR-581 knockdown significantly enhanced the cell viability of FHC cells, whereas the overexpression of miR-581 caused decreased proliferative activity of LoVo cells compared to the matched control cells ( Figure 2N-2S).

LINC00485 regulates cell proliferation, migration, and invasion by acting on miR-581
To determine how LINC00485 exerts its functions by regulating miR-581, we performed a series of functional experiments in vitro. Our findings demonstrated that overexpression of miR-581 promoted proliferation, migration, and invasion of human normal colorectal epithelial cells (FHC), whereas miR-581 down-regulation AGING suppressed the proliferative, invasive, and migratory abilities of CRC cells (LoVo) when compared with the matched control group (Figure 3A-3H; Supplementary Figure 2). Furthermore, LINC00485 silencing significantly elevated the proportion of Ki-67-positive cells, upregulated the protein expression level of proliferative markers PNCA and Ki-67, arrested cell cycle in Go/G1 phase (Supplementary Figure 3), enhanced colony formation, cell migration, and invasion capabilities of FHC cells in comparison to control cells, which could be partially reversed by miR-581 silencing ( Figure 4A, 4C, 4E, 4G). Additionally, LINC00485-overexpressed LoVo cells demonstrated a significant reduction in the percentage of Ki-67-positive cells, the expression levels of PNCA and Ki-67, and in colony formation, proliferative, migratory, and invasive capacities compared with the control groups; however, overexpression of miR-581 partially abolished LINC00485 overexpression-induced inhibition of cell proliferation, migration, and invasion in LoVo cells ( Figure 4B, 4D, 4F, 4H; Supplementary Figure 3). These findings suggested that LINC00485 regulated CRC progression by acting via miR-581. To explore downstream targets of miR-581, we performed bioinformatics analysis using TargetScan (http://www.targetscan.org/) to predict messenger RNA (mRNA) targets. We identified the complementary sites between miR-581 and EDEM1 ( Figure 5A). Dualluciferase reporter assays revealed that wild-type (WT) LINC00485, but not the mutant LINC00485,  AGING could directly bind to miR-581. When miR-581 was co-transfected with the WT 3'-untranslated region (3'UTR) of EDEM1 into 293T cells, the luminescence activity was significantly reduced, while when miR-581 was co-transfected with mutant 3'UTR of EDEM1 into 293T cells, no differences were observed, indicating that miR-581 targeted the WT 3'UTR of EDEM1, but not the mutant 3'UTR of EDEM1 ( Figure 5B). RIP and real time quantitative polymerase chain reaction (RT-qPCR) experiments further confirmed the binding relationship between miR-581 and 3'UTR of EDEM1 mRNA ( Figure  5C, 5D). Moreover, by searching The Cancer Genome Atlas (TCGA) dataset, we found that expression of EDEM1 was significantly down-regulated in CRC tissues compared to adjacent normal tissues (Supplementary Figure 4A). EDEM1 expression also showed significant differences in correlation with the patient's sex, tumor type, and stage (Supplementary Figure 4B-4D). A further search of the Human Protein Atlas (HPA) database revealed that the overall survival of CRC patients with low EDEM1 expression was significantly lower than those with high EDEM1 expression (Supplementary Figure 4E). In our study, levels of EDEM1 were markedly decreased in CRC tissues compared with normal tissues ( Figure 5E), an association that correlated with tumor stage ( Figure 5E-5H). EDEM1 expression was positively correlated with LINC00485 levels ( Figure 5I), while it inversely correlated with miR-581 expression ( Figure 5J). EDEM1 expression in vitro was significantly reduced in CRC cell lines compared to human normal colorectal epithelial cells at both the transcriptional and translational levels ( Figure 5K, 5L).

LINC00485/miR-581/EDEM1 axis regulates epithelialto-mesenchymal transition in CRC
To unravel whether the LINC00485/miR-581 axis exerts its biological function by modulating EDEM1 expression, small interfering RNA (siRNA) targeting EDEM1 were used to inhibit EDEM1 expression in LoVo cells, which resulted in a significant reduction of EDEM1 expression compared to the untransfected cells ( Figure 6A). In LINC00485-overexpressing LoVo cells, miR-581 levels were significantly down-regulated; however, further treatment with siEDEM1 had no effect on miR-581 expression ( Figure 6B), but markedly reduced EDEM1 expression as expected ( Figure 6C). Consistent with previous results, LINC00485 knockdown exhibited significant suppressive activity in LoVo cells. Transfection siEDEM1 partially counteracted LINC00485 knockdown-induced reduction of cell proliferation ( Figure 6D), colony formation ( Figure 6E), migration ( Figure 6F), and invasion ( Figure 6G) abilities of LoVo cells compared with the untransfected LINC0048-knockdown cells. These results confirmed that LINC00485 exerted its effects by regulating the EDEM1/miR-581 axis. EMT contributes to the pathogenesis of cancer metastasis [17]. In this cellular context, the expression of epithelium-related genes (cytokeratin, E-cadherin) was significantly upregulated, while that of mesenchymal-associated genes (Ncadherin, vimentin) was downregulated in LINC00485overexpressed LoVo cells compared to control cells. Further knockdown of EDEM1 using siRNA, could partially reverse the changes caused by LINC00485 overexpression ( Figure 6H).
In addition, treatment with miR-581 antagomir increased expression of EDEM1 in LoVo cells compared to the untreated cells. However, EDEM1 expression was strikingly attenuated in LoVo cells cotransfected with miR-581 antagomir and siEDEM1 compared with LoVo cells transfected with miR-581 antagomir alone ( Figure 7A). In line with the aforementioned results, the knockdown of miR-581 expression suppressed the proliferation, migration, and invasion of LoVo cells in comparison to the control cells. Finally, treatment with siEDEM1 was capable of partially rescuing the phenotypes caused by miR-581 knockdown in LoVo cells ( Figure 7B-7E).
Meanwhile, miR-581 antagomir treatment enhanced the expression of epithelial markers E-cadherin and cytokeratin, but down-regulated the expression of mesenchymal markers N-cadherin and vimentin in LoVo cells at both the transcriptional and translational levels ( Figure 7F-7K). Of note, our rescue experiments showed that siEDEM1 treatment could partially reverse the above-mentioned changes in gene expression induced by the knockdown of miR-581 ( Figure 7F-7K). Cumulatively, these data further supported our findings that the LINC00485/miR-581/EDEM1 axis regulated CRC progression.

Overexpression of LINC00485 or miR-581 knockdown attenuated CRC cell growth and liver metastasis in vivo
We established a xenograft nude mouse model and liver metastasis model to investigate the effects of LINC00485 and miR-581 on CRC cell growth and metastasis in vivo. As described in the Methods section, tumor growth was monitored weekly. When the diameter of the largest tumor reached 1 cm, the nude mice met the humane endpoints of the study. All animals were sacrificed under anesthesia. Tumors were excised and the volumes were measured. We found that overexpression of LINC00485 significantly decreased tumor diameter, tumor volume at day 42, and the number of hepatic nodules in comparison to the control group, suggesting that the up-regulation of LINC00485 suppressed tumor growth and liver metastasis of CRC AGING cells ( Figure 8A-8D). Additionally, miR-581 antagomir injection also contributed to the decrease of tumor diameter, tumor volume at day 42, and the number of metastatic nodules of LoVo cells, indicating that miR-581 knockdown restricted CRC cell growth and liver metastasis in vivo ( Figure 8F-8I). Furthermore, the expression levels of epithelium markers (cytokeratin, E-cadherin), mesenchymal markers (N-cadherin, vimentin), and cell proliferation markers (CK-19, Ki-67) were evaluated by RT-qPCR in metastatic nodules of nude mice. We found that the results were consistent with in vitro observations ( Figure 8E, 8J). Overall, it can be concluded that the LINC00485/miR-581/EDEM1 axis may represent an important mechanism involved the malignant progression of CRC.

DISCUSSION
Herein, we reported the role of LINC00485 in CRC progression. LINC00485 is downregulated in CRC tissues and cancer cells (LoVo, SW480, HCT8) compared with paired normal samples and human   AGING normal colonic epithelial cells (FHC, NCM460, CCD-18co). The LINC00485/miR-581/EDEM1 regulatory axis promotes the proliferation, migration, invasion, and EMT of CRC cells. Further, using the xenograft nude mouse model, we found that LINC00485 knockdown or downregulation of miR-581 not only significantly repressed CRC cell growth, causing a decrease in tumor volume, but also prevented CRC liver metastasis. Our findings suggested that LINC00485 plays an important role in CRC progression by regulating the miR-581/EDEM1 axis.
CRC is a serious threat to human health, and leads to hundreds of thousands of deaths annually [1,18]. Although there are many studies on the pathogenesis of CRC [19][20][21][22], the molecular mechanisms underlying CRC are still unknown. Hence, elucidating the biological functions of genes and molecules involved in the occurrence and development of CRC is of benefit for drug development and targeted treatments.
Cancer cell proliferation is one of the ten malignant hallmarks of malignancy [23], and tumor metastasis is the main cause of death in patients with cancer [24]. Mounting evidence has shown that non-coding RNAs, including lincRNA [25] and miRNA [26,27], exhibit antitumor effects by inhibiting the proliferation of CRC cells. In the present study, we found that low expression of LINC00485 promoted the proliferation, migration, and invasion of CRC cells, which are distinct features of malignant tumor cells [23]. Moreover, LINC00485 overexpression significantly restricted tumor growth and liver metastasis of colorectal carcinoma in vivo. These results indicated that aberrant expression of LINC00485 in CRC may play an important role in CRC progression. Conversely, it has been reported that LINC00485 is overexpressed in LAC cells and that it exerted its functional activity by targeting the miR-195/CHEK1 axis [14]; therefore, we suspected that LINC00485 may have a characteristic tissue-specific expression. Further work is required to elucidate the molecular mechanisms underlying this phenomenon.
LincRNAs are mostly found in the cytoplasm [14,28] and exert their biological activities by sponging miRNA [8]. Herein, our results confirmed that LINC00485 was predominantly localized within the cytoplasm and acted directly on miR-581. Lower expression of LINC00485 contributed to the upregulation of miR-581 in CRC, promoting the proliferation, migration, and invasion of CRC cells, which was consistent with previous studies indicating that the up-regulation of miR-581 regulated the SMAD7/TGFβ signaling pathway, driving CRC metastasis [29]. Further, miRNAs regulate the translation of mRNAs through direct base pairing with specific sites present in the 3'UTR of mRNA [30,31].
In this study miR-581 directly targeted the 3'UTR of EDEM1 mRNA. EDEM1 is a crucial regulator of the endoplasmic reticulum (ER)-associated degradation (ERAD) pathway, which recognizes N-glycans on misfolded proteins through the mannosidase-like domain (MLD), resulting in the degradation of misfolded glycoproteins [32][33][34]. The accumulation of misfolded glycoproteins was shown to induce ER stress followed by the activation of autophagy [35]. Autophagy is an important mechanism that the cell utilizes to sustain tumor cell metabolism [36].
The EMT process is closely linked to tumor progression and metastasis [17,37,38], and involves epithelial cells losing epithelial characteristics and gaining a mesenchymal phenotype. Based on our results, overexpression of LINC00485 significantly downregulated the expression of mesenchymal markers Ncadherin and vimentin, whereas the expression levels of epithelial markers E-cadherin and cytokeratin were upregulated by directly modulating the miR-581/EDEM1 axis, suggesting that cancer cells lose their malignant phenotype. LINC00485 has multiple roles in the pathogenesis of CRC. Other mechanisms including CRC development and progression merit further investigation.
In conclusion, our findings identified that LINC00485 was downregulated in CRC tissues in comparison with adjacent normal tissues. Low expression of LINC00485 was associated with poor prognosis of CRC patients. Downregulation of LINC00485 contributed to decreased expression of EDEM1 by sponging of miR-581, thereby facilitating CRC cell proliferation, migration, invasion, and the processed of EMT. We show that the LINC00485/miR-581/EDEM1 regulatory axis is implicated in regulating the malignant phenotypes of CRC cells both in vivo and in vitro. Our findings contribute to the elucidation of the molecular pathogenesis underlying CRC.

Patient samples
We collected 52 matched samples (surgically excised tumor tissues and adjacent healthy tissues) from CRC patients with no other medical illnesses, comorbidities, or undergoing neoadjuvant chemoradiotherapy prior to surgery. The detailed information on the enrolled CRC patients is listed in Table 1. Tissues were frozen and rapidly stored in liquid nitrogen after surgical removal. Written informed consent was obtained from patients enrolled in this study. This study was approved by the Ethics Committee of Nanjing First Hospital, Nanjing Medical University.

Establishment of cell lines overexpressing or silencing LINC00485
To construct the LINC00485-overexpressing cell line, mRNA of LINC00485 was reverse transcribed into cDNA and then cloned into the multicloning site of the pLVX plasmid (Sangon Biotech Co., Ltd., Shanghai, China) using cloning primers that incorporated an XhoI site (5'-CCCTCGACTGCGCCCGAGAGGCAGCG-3') and a PstI site (5'-AACTGCAAACTAGGTCATCTG TTTATT-3'). The construct was subsequently, cotransfected with the packaging plasmids (Sangon Biotech Co., Ltd.) into 293T cells to produce virus particles using the Lipofectamine 3000 transfection reagent (Invitrogen). The virus was collected and mixed in the presence of polybrene (Sangon Biotech Co., Ltd.) to infect LoVo cells. Afterwards, 10 μg/mL of puromycin was used to screen stably expressing cells for 3 days. For the construction of the LINC00485 knockdown cell line, FHC cells were infected with the LINC00485 RNAi lentivirus (Genechem Incorporation, Shanghai, China). The shRNA sequence targeting LINC00485 was as follows: 5'-AATAACCAACCCTA TAAACAT-3'.

Construction of reporter vectors and luciferase assay
Dual-luciferase reporter assays were performed using the Dual-Luciferase Assay Kit (TransGen Biotech Co., Ltd., Beijing, China), and Renilla luciferase activity served as an internal control for transfection efficiency. The WT 3'UTR of EDEM1 or WT LINC00485 (the binding sites with miR-581) was cloned into pGL3 luciferase reporter vector. Primers for LINC00485 were AGING as follows: the forward primer incorporated the PstI restriction site (5'-AACTGCAGAGAGGCAGCGCTC AGACAGC-3'); the reverse primer incorporated the EcoRI restriction site (5'-CGGAATTCTCCGACTTC CAGTTGGGTTC-3'). The primers for cloning the EDEM1 3'UTR fragment were established as follows: the forward primer incorporated the XbaI restriction site (5'-CTAGTCTAGATTCCTGTCCATTCCCAGCAT-3'); the reverse primer incorporated the XhoI restriction site (5'-CTAGTCTAGACGTAAGTTTTCCCAGAGT TTCCA-3'). The amplified product (4047 bp) of EDEM1was cloned into the pGL3 vector to construct the 3'UTR reporter vector.
Additionally, a high-fidelity amplification kit (Fast Mutagenesis System, #FM111-01, TransGen Biotech Co., Ltd., Beijing, China) and point mutation primers were used to amplify the above plasmid to obtain mutated EDEM1 3'UTR vectors. N, N'dimethyltryptamine (DMT) was applied to screen the mutated EDEM1 3'UTR reporter plasmid. Empty plasmid pGL3 was used as the control. Luciferase reporter vectors containing mutated LINC00485 were also constructed. Afterwards, as the experiment design showed, the corresponding vectors were co-transfected with miR-581 mimics into 293T cells. Luciferase activity was measured by CLARIOstar ® Microplate Reader (BMG Labtech Inc., Germany) at 48 h post transfection.

Fluorescence in situ hybridization assay
The Fluorescence in situ hybridization (FISH) assay was used to analyze the subcellular localization of RNA. The corresponding probes were synthesized by Genechem (Shanghai Genechem Co., Ltd., China) based on the sequences of LINC00485 and miR-581. Tissue sections or culturing LoVo cells on sterile glass slides were fixed by immersing slides in 4% formaldehyde for 10 min at room temperature. Subsequently, the slides were treated with pre-chilled PBS containing 0.4% Triton X-100 and 10 μM vanadyl ribonucleoside complexes (VRC) for 30 min at room temperature following hybridization with Cy5-(red, indole dicarboxylic cyanide) labeled LINC00485 probes or incubation with FAM-labeled miRNA-specific probes (green, carboxy fluorescein) at 37° C for 12 h. After staining the nuclei with 4',6-diamidino-2-phenylindole (DAPI), images were captured by fluorescence microscopy (Olympus, Japan).

RNA immunoprecipitation experiment
After lysis of the cells according to the instructions of the Magna RIP™ RNA-binding protein Immunoprecipitation Kit (Millipore, Stafford, VA, USA), the cell lysates were incubated with biotin-labeled LINC00485 for 2 h at 4° C. Biotin-labeled nonsense oligonucleotides were used as negative controls. Streptavidin magnetic beads were incubated for 12 h at 4° C to capture the hybridized nucleic acids. After precipitation by magnetic separation, TRIzol was used to dissolve the precipitate for RNA extraction. Reverse transcription and RT-qPCR assays were then performed to determine the expression levels of miR-581.

Cell counting kit-8 assay
The cells were cultured using a 96-well plate with a transparent bottom (Corning Inc., USA). Cell proliferation was determined following the manufacturer's guidelines. Briefly, after incubation with 10% (CCK-8) reagent (Beyotime Biotechnology, Shanghai, China) at 37° C, absorbance at 450 nm was detected at different time points. The absorbance value was used to evaluate cell proliferative ability.

Colony formation assay
Cells (1 × 10 3 ) were seeded into 6-well dishes, and then cultured for 10 days. After being fixed in 4% paraformaldehyde, the cells were stained with 0.1% Crystal Violet Staining Solution (Beyotime Biotechnology) for 30 min at room temperature followed by observation under the microscope (Leica, Germany). The number of cell colonies were quantified using ImageJ v1.8.0 from ten randomly selected fields.

Detection of Ki-67 by flow cytometry
Cells were collected and fixed with 4% paraformaldehyde for 30 min following treatment with 0.2% Triton X-100 for 30 min at room temperature. Thereafter, the cells were incubated with anti-Ki-67 antibody (dilution 1:300, #9129S, Cell Signaling Technology, USA) prepared by 5% bovine serum albumin (BSA) solution overnight at 4° C. After coupling with the fluorescent-labeled secondary antibody (dilution 1:500, #4412S, Cell Signaling Technology), fluorescence activity was detected by flow cytometry. Results were analyzed using Flow Jo 7.6.1 (Becton, Dickinson and Company, USA).

RNA preparation and quantitative PCR (qPCR)
Total RNA was isolated and purified using TRIZol reagent (Thermo Fisher Scientific, Inc., USA) according to the manufacturer's instructions. Mir-X miRNA First-Strand Synthesis Kit (Catalog No. 638315, Takara Biomedical Technology (Beijing) Co., Ltd., China) and AGING Mir-X miRNA quantitative real time polymerase chain reaction (qRT-PCR) TB Green ® Kit (Catalog No. 638316, Takara Biomedical Technology (Beijing) Co., Ltd.) were used for miRNA detection. The thermocycling conditions for to detect miRNA expression were as follows: 95° C for 30 s, followed by 40 cycles at 95° C for 15 s, 60° C for 60 s, and 72° C for 10 s. PrimeScript™ RT Master Mix (Perfect Real Time) (Catalog No. RR036A, Takara Biomedical Technology (Beijing) Co., Ltd.) and One Step TB Green ® PrimeScript™ RT-PCR Kit (Perfect Real Time) (Catalog No. RR066A, Takara Biomedical Technology (Beijing) Co., Ltd.) were used to evaluate gene expression. RT-qPCR thermocycling parameters to detect gene expression and lncRNA levels were 95° C for 30 s, followed by 40 cycles at 95° C for 15 s, 60° C for 30 s, and 72° C for 10 s. U6 and β-actin served as the internal controls for miRNA levels and gene expression, respectively.

In vitro scratch assay
Cells were seeded into 6-well plates (1 × 10 6 cells/well). When the cell confluence reached 90%, a pipette tip was used to scratch a gap on the cell monolayer and cultures were observed at 0, 24, 48, and 72 h. The rate of closure was assessed from three independent experiments.

Transwell migration and invasion assays
For cell migration and invasion (the upper chamber was coated with Matrigel) assays, a 24-well Transwell plate system (Cell BioLabs, Inc., San Diego, CA, USA) was used to measure the migration and invasive capabilities of both CRC and FHC cells. Matrigel was obtained from Gibco. The experimental protocol has been described previously in detail [27]. After seeding, waited 12 hours before subsequent fixation, staining, and photographing.

Animal experiments
For the tumor xenograft model, LoVo cells overexpressing LINC00485 (1 × 10 6 cells in 50 μL of PBS) or control cells (equal loading volumes) were implanted subcutaneously into the flank regions of 8week-old female nude BALB/c mice (a total of 12 mice; 6 mice per group). In a different set of experiments, LoVo cells (1 × 10 6 cells in 50 μL) were subcutaneously injected into the dorsal surface of nude mice (a total of 12 mice; 6 mice per group), in which six randomly selected animals were intratumorally injected with miR-581 antagomir (13 μg in a 100 μL volume per mouse) plus an equal volume of Lipofectamine 2000 once weekly. The mock group (n=6) received an equal volume of phosphate buffered saline (PBS) solution. Tumor diameters were measured every 7 days. The animals were sacrificed on day 42, and tumors were meticulously excised. The tissue volume was calculated using the formula: 0.5 × width 2 × length. To examine the ability of tumor cells to metastasize to the liver, LoVo cells (1 × 10 6 cells in 50 μL) were injected into the spleen of mice (a total of 12 mice; 6 mice per group) under anesthesia. MiR-581 antagomir (13 μg in volume of 100 μL per mouse) plus an equal volume of Lipofectamine 2000 was administered via tail vein once weekly. Another six animals received an equal volume of PBS plus Lipofectamine 2000. Additionally, we inoculated LoVo cells or LoVo cells overexpressing LINC00485 (1 × 10 6 cells in 50 μL) into the spleen of nude mice (a total of 12 mice; 6 mice per group) as mentioned above. All animals were sacrificed on day 28, liver tissues were carefully removed and fixed with 10% formalin solution. Paraffin sections (5 μm) were prepared for hematoxylin-eosin (H&E) staining and all metastatic nodules were counted in the section. Mice used in this study were purchased from Model Animal Research Center of Nanjing University. Animal experiments were approved by the Institutional Animal Care and Use Committee of Nanjing First Hospital.

Statistical analysis
Differences between two groups were assessed by applying student's t-test. Multiple comparison was analyzed using the one-way analysis of variance (ANOVA) with Fisher's least significant difference (LSD) test. Survival analysis was analyzed using Kaplan-Meier survival curves with a log-rank test. Data are expressed as mean ± standard deviation (SD). P < 0.05 indicated that the difference was statistically significant. All in vitro experiments were performed in three independent experiments.