Long non-coding RNA CASC2 suppresses epithelial-mesenchymal transition of hepatocellular carcinoma cells through CASC2/miR-367/FBXW7 axis

Seventy-five HCC tissues and adjacent normal tissues were collected from patients, who underwent hepatectomy in the Department of Hepatobiliary Surgery, the First Affiliated Hospital of Xi’an Jiaotong University during January 2009 to December 2011. All patients did not receive any embolotherapy or chemotherapy before surgical operation and were pathologically diagnosed post-operation.

Six HCC cell lines (MHCC-97L, Hep-3B, HepG2, Huh7, SMMC-7721 and MHCC-97H) and the human immortalized normal hepatocyte cell line (LO2) were cultured in DMEM (Gibco, USA) with 10% fetal bovine serum (FBS, Gibco), 100 μg/mL streptomycin and 100 U/mL penicillin (Sigma, USA). All of the cells were maintained in an incubator (5% CO₂, 37 °C). According to the product description, cell transfection were performed using Lipofectamine 2000 Reagent (Invitrogen Life Technologies, USA). Scrambled siRNA (siControl) and CASC2 siRNAs (siCASC2#1 and siCASC2#2), as well as pcDNA3.1-Control (pcDNA/Control) and pcDNA3.1-CASC2 (pcDNA/CASC2) were purchased from Invitrogen (USA). MiR-367 inhibitors (anti-miR-367), inhibitors negative control (NC), miR-367 mimics (miR-367) and mimics negative control (NC) were purchased from Genecopoeia (Guangzhou, China). FBXW7 Human cDNA ORF Clone and Control were purchased from OriGene (OriGene Technologies, Inc., USA).

Bioinformatics tools (microRNA.​org, Starbasev2.0, miRcode et al.) were applied to analyze the miR-367 binding sites on CASC2 gene as well as the FBWX7 binding sites on miR-367 gene. Then a dual luciferase reporter assay was conducted according to the protocols described in our previous study [16]. Subsequently, the luciferase reporter assay system (Promega, Madison, WI, USA) was applied to examine the luciferase activity. All of the above experiments were conducted repeatedly for three times.

The EZ-Magna RIP Kit (Millipore, USA) was applied to conduct the RIP assay according to the product specification. Firstly, cells were collected and lysed in complete RIP lysis buffer. Then, the cell extract was incubated with RIP buffer containing magnetic beads conjugated to a human anti-Ago2 antibody (Millipore, USA). Samples were incubated with proteinase K with shaking to digest proteins and the immunoprecipitated RNA was isolated. Subsequently, the NanoDrop spectrophotometer was used to measure the concentration of RNA, and the purified RNA was subjected to real-time PCR analysis.

The DNA fragment with the full length CASC2 sequence or negative control sequence was PCR amplified using a T7-containing primer and then cloned into GV394 (Genechem, Shanghai, China). The resultant plasmids DNAs were linearized using restriction enzyme XhoI. Biotin-labeled RNAs were reversely transcribed using Biotin RNA Labeling Mix (Roche, USA) and T7 RNA polymerase (Takara Biomedical Technology). The products were treated with RNase-free DNase I (Roche, USA) and purified with the RNeasy Mini Kit (Qiagen, USA) and RNA were extracted for real-time PCR.

Based on the manufacturer’s protocol of Trizol (Invitrogen, Carlsbad, CA, USA), total RNA from cells and tissues were isolated.Then the cDNA was obtained according to the protocols described in our previous study [16], and RNA levels were measured by real-time PCR analysis. Primers for CASC2, miR-367 and FBXW7 were purchased from Genecopoeia (Guangzhou, China). Then the results were calculated applying the 2^-ΔΔCt method.

Cells were harvested and then lysed via radioimmunoprecipitation assay buffer (RIPA buffer, Thermo Fisher Scientific, MA,USA) to obtain proteins. Subsequently, proteins were separated by SDS–PAGE and transferred to PVDF membranes. Then the above PVDF membranes were incubated with corresponding primary antibodies. E-cadherin (1:1000, Cell Signaling, MA, USA), Vimentin (1:500, Santa Cruz, CA, USA), FBXW7 (1:1000, Abcam, MA, USA) and β-actin (1:500, Santa Cruz, CA, USA). Then the membranes were incubated with HRP-conjugated secondary antibody for 2 h at room temperature (ZSGB-BIO, China). Detection was performed by enhanced chemiluminescence kit (Amersham, Little Chalfont, UK).

The upper chambers of transwell inserts (8 μM pore-sized, Nalge Nunc Intl., NY, USA) were coated with matrigel (1:8, BD, NJ, USA) for invasion assay, while migration assay without matrigel. Then the subsequent experiments were conducted according to the protocols described in our previous study [16]. After 24 h, cells were fixed using 4% paraformaldehyde, and permeabilized with methanol. After removing these cells still in the upper chambers, the cells in lower chambers were stained with 0.3% crystal violet dye and washed by PBS. The images were obtained and cells were counted using a light microscope (×100).

The experiments was conducted according to the protocols described in our previous study [16]. The sections were dewaxed, dehydrated, and rehydrated. Then, citrate buffer was used for retrieve antigen, and hydrogen peroxide (3.0%) was used for blocking the endogenous peroxidase activity. After being blocked by 10% goat plasma, the primary antibodies, including E-cadherin (1:400, Cell Signaling) and Vimentin (1:200, Santa Cruz) were added to the sections and incubated at 4 °C overnight. The biotinylated secondary antibodies (Goldenbridge, Zhongshan, China) applied for detecting the primary antibodies. Counterstained was performed using hematoxylin.

In order to assess the effect of CASC2 on metastatic ability of HCC cells in vivo, the lentivirus-based system was applied to establish CASC2 stably overexpressing MHCC-97H cells and CASC2 down-regulating Hep-3B cells. Then, these MHCC-97H cells (1 × 106) and Hep-3B cells (1 × 106) were injected into the tail vein of nude mice. Eight weeks later, the lungs of nude mice were gathered and fixed with dampen formaldehyde solution. Then paraffin-embedded, routine hematoxylin-eosin (H&E) staining and morphology observed by microscope were performed to evaluate the lung metastases. The protocols for these animal experiments were approved by the Ethics Review Committee of Xi’an Jiaotong University.

Firstly, real-time PCR was applied to measure CASC2 expression in 75 paired HCC tissues and adjacent non-tumor tissues. The results revealed that CASC2 expression in HCC tissues was dramatically decreased, compared with adjacent non-tumor tissues (P < 0.001, Fig. 1a). Furthermore, CASC2 expression was markedly downregulated in aggressive HCC tissues compared with non-aggressive HCCs (P < 0.01, Fig. 1b). CASC2 expression was notably decreased in HCC tissues with recurrence compared to those without recurrence (P < 0.001, Fig. 1c). Meanwhile, The data from R2: Genomics Analysis and Visualization Platform (http://​r2.​amc.​nl) including GEO and TCGA database showed that the expression of CASC2 was significantly downregulated in HCC tissues compared to normal liver tissues (P < 0.05, respectively, Additional file 1: Fig. S1), which were consistent with our results. Additionally, our data revealed that significantly underexpressed CASC2 could be observed in all the six HCC cell lines, compared to LO2 cells (P < 0.05, respectively, Fig. 1d). Hep-3B cells had the highest expression of CASC2 and MHCC-97H cells showed the lowest level (Fig. 1d), thus these two cell lines were used for loss- and gain-of-function experiments. In conclusion, these data suggested that CASC2 expression was markedly decreased in HCC.

Next, we attempted to investigate the functional effects of CASC2 on HCC cells. Firstly, MHCC-97H cells were transfected with pcDNA/CASC2, while Hep-3B cells were transfected with CASC2 siRNAs (siCASC2#1 and siCASCC2#2). Real-time PCR analysis revealed that the expression of CASC2 was significantly modulated by pcDNA/CASC2 and CASC2-siRNAs in MHCC-97H and Hep-3B cells (P < 0.01, respectively Fig. 2a and b). Functionally, Transwell assays demonstrated that overexpressed CASC2 could significantly diminish the migration and invasion abilities of MHCC-97H cells (P < 0.01, respectively Fig. 2c). On the other hand, CASC2 silencing significantly potentiated the cell migration and invasion of Hep-3B cells (P < 0.01, respectively Fig. 2d). Furthermore, we tried to confirm the functional effects of CASC2 on HCC cells in vivo. The results revealed that CASC2 stably overexpressing MHCC-97H cells appeared lower probability of lung metastasis, while CASC2 underexpressing Hep-3B cells presented much more lung metastasis compared to control group (P < 0.05, respectively Fig. 2e). These data indicated that CASC2 could repress migration and invasion of HCC cells both in vitro and in vivo.

EMT progression is of great importance for migration and invasion of HCC cells. Then, we attempted to explore whether CASC2 could exert any influence on EMT progression of HCC cells. Immunohistochemical staining analysis was conducted in 75 HCC samples to explore the associations between CASC2 and EMT markers (Vimentin and E-cadherin). Subgroups (CASC2 low/high) were divided according to the cutoff values of CASC2, which were defined as the median of the cohort. The results indicated that E-cadherin expression was significantly decreased while Vimentin expression was obviously increased in low CASC2 group compared to high CASC2 group (P < 0.01, respectively Fig. 3a). Consistently, Western blot results revealed that CASC2 overexpression could significantly elevate the expression of E-cadherin and repress the expression of Vimentin in MHCC-97H cells (P < 0.05, respectively Fig. 3b). CASC2 silencing reduced E-cadherin expression and increased Vimentin in Hep-3B cells (P < 0.01, respectively Fig. 3c). These results demonstrated that CASC2 inversely regulated EMT progression in HCC cells.

To explore the potential mechanisms of CASC2 functions in HCC cells, several bioinformatics tools (microRNA.​org and miRBase) were employed to analyze the potential targets of CASC2. Bioinformatics tools analysis revealed the potential binding site for miR-367 on CASC2 gene (Fig. 4a). And the data of real-time PCR showed that the expression of miR-367 was markedly elevated in HCC tissues (P < 0.01, Fig. 4b). Pearson correlation analysis revealed that a negative association between miR-367 and CASC2 expression was observed in HCC tissues (r = −0.7243, P < 0.001, Fig. 4c). In addition, we detected the expression of miR-367 in LO2, MHCC-97H and Hep-3B cells. The data indicated that both MHCC-97H and Hep-3B cells showed a higher miR-367 expression compared to LO2 cells (P < 0.05, respectively Fig. 4d). Meanwhile, MHCC-97H had the highest expression of miR-367 (P < 0.05, Fig. 4d). Next, MHCC-97H and Hep-3B cells were transfected miR-367 inhibitors and mimics, respectively. Then, miR-367 overexpression and knockdown were confirmed by real-time PCR (P < 0.01, respectively Fig. 4e). The luciferase reporter assays manifested that miR-367 significantly suppressed the luciferase activity that carried wild type (WT) but not mutant (Mut) 3′-UTR of CASC2 (P < 0.05, respectively Fig. 4f). In addition, it has been well identified that miRNAs exert its function via binding to Ago2, a core component of the RNA-induced silencing complex (RISC) that is required for miRNAs-mediated gene silencing, and potential targets of miRNAs can be isolated from this complex after Ago2 co-immunoprecipitation. Consistently, results of RIP assays confirmed that miR-367 was a target of CASC2 in HCC cells (P < 0.01, respectively, Additional file 2: Figure S2A and 2B). Subsequently, the biotin-labeled pulldown system was applied to further confirm that CASC2 could directly interact with miR-367. We observed a significant amount of CASC2 and miR-367 in the CASC2 pulled down pellet compared with control group as measured by real-time PCR (P < 0.01, respectively, Additional file 2: Figure S2C and 2D). Furthermore, our data revealed that CASC2 could negatively regulate the expression of miR-367 in HCC cells (P < 0.01 and P < 0.001, respectively, Fig. 4g). On the other hand, CASC2 expression was significantly elevated or decreased by modulating miR-367 level in HCC cells (P < 0.01, respectively, Fig. 4h). Taken together, these data demonstrated that CASC2 could directly bind to miR-367 in HCC cells, and showed a reciprocal repression of CASC2 and miR-367.

Next, we were aimed to confirm the roles of miR-367 in migration and invasion of HCC cells. And the results revealed that the migration and invasion abilities of HCC cells were markedly weakened by miR-367 silencing, and significantly enhanced by miR-367 restoration (P < 0.05, respectively, Fig. 5a and b). Next, immunohistochemical staining was conducted in 75 HCC samples to explore the associations between miR-367 and EMT markers (Vimentin and E-cadherin). Subgroups (miR-367 low/high) were divided according to the cutoff values of miR-367, which were defined as the median of the cohort. The results indicated that E-cadherin expression was significantly decreased while Vimentin expression was obviously increased in high miR-367 group compared to low miR-367 group (P < 0.01, respectively, Fig. 5c). Additionally, Western blot analysis revealed that miR-367 silencing increased E-cadherin expression and decreased Vimentin in MHCC-97H cells (P < 0.05, respectively, Fig. 5d). miR-367 overexpression facilitated the EMT process of Hep-3B cells (P < 0.05, respectively, Fig. 5e). Therefore, we concluded that miR-367 could promote migration, invasion and EMT of HCC cells.

To elucidate the underlying molecular mechanism about how miR-367 exerted its effects on HCC cells, we searched for candidate targets of miR-367 by using bioinformatics tools (microRNA.​org and miRBase). As shown in Fig. 6a, binding sequences of miR-367 were identified in 3’UTR of FBXW7 mRNA. Previous study in non-small cell lung cancer has reported that FBXW7 is a downstream target of miR-367 [14]. Besides, it has been confirmed that FBXW7 is a tumor suppressor and suppresses EMT in HCC cells [15, 17]. Moreover, FBXW7 could inhibit the EMT by down-regulating the RhoA signaling pathway in gastric cancer [18]. Then we focused on FBXW7 and assumed that miR-367 could directly target FBXW7 in HCC. Based on our previous studies about FBXW7 [17, 19]. Subsequently, the results of dual luciferase reporter assays, real-time PCR and western blot analysis revealed that miR-367 could directly target FBXW7 and negatively modulate the expression of FBXW7 in HCC cells (P < 0.01, respectively, Fig. 6b-g). Next, we explored whether CASC2 could regulate the expression of FBXW7. Our data showed that FBXW7 expression could be positively regulated by CASC2 in HCC cells (P < 0.01, respectively, Fig. 6h-k). Thus, we concluded that FBXW7 was a direct target of miR-367 and positively modulated by CASC2 in HCC cells.

Next, we tried to investigate whether CASC2 could exert its inhibitory effects on migration, invasion and EMT progression of HCC cells through CASC2/miR-367/FBXW7 axis. Notably, miR-367 overexpression reduced the expression of FBXW7 in CASC2-overexpressing MHCC-97H cells and miR-367 knockdown abolished the inhibitory effects of CASC2 silencing on FBXW7 expression in Hep-3B cells (P < 0.01, respectively, Fig. 7a and b, Additional file 3: Figure S3A). miR-367 expression was significantly reversed in cells co-transfected with pcDNA/CASC2 and miR-367 mimics compared to pcDNA/CASC2 transfection alone (P < 0.01, Additional file 3: Figure S3B). Functionally, Transwell assays indicated that miR-367 overexpression promoted migration and invasion of CASC2-overexpressing MHCC-97H cells (P < 0.05, respectively, Fig. 7c, Additional file 3: Figure S3C). Furthermore, the pro-metastatic effects of CASC2 silencing on migration and invasion of Hep-3B cells could be dramatically reversed by miR-367 knockdown (P < 0.05, respectively, Fig. 7d). Furthermore, the data from Western blotting indicated that miR-367 could also notably reverse the inhibitory effects of CASC2 on EMT progression of HCC cells (P < 0.05, respectively, Fig. 7e and f, Additional file 3: Fig. S3D). Otherwise, FBXW7 overexpression restrained EMT process of Hep-3B cells (Additional file 4: Figure S4A). FBXW7 restoration abrogated promoting effects of miR-367 on migration, invasion and EMT progression of Hep-3B cells (P < 0.01, respectively, Additional file 4: Fig. S4B and 4C). Thus, we demonstrated that CASC2 could exert its inhibitory effects of HCC metastasis via CASC2/miR-367/ FBXW7 axis.

For each cohort, different subgroups were plotted according to the cutoff values of CASC2 and miR-367, which were defined as the median of the cohort. As shown in Table 1, the statistical results indicated that CASC2 underexpression was dramatically associated with venous infiltration (P = 0.004), high Edmondson-Steiner grading (P = 0.043) and advanced TNM tumor stage (P = 0.003). Meanwhile, miR-367 overexpression was notably correlated with large tumor size (P = 0.015), venous infiltration (P = 0.025) and advanced TNM tumor stage (P = 0.020). Therefore, we suggested that aberrant expression of CASC2 and miR-367 were closely correlated with the metastasis-associated clinicopathologic features of HCC. Subsequently, 3-year overall survival (OS) and disease-free survival (DFS) were analyzed by Kaplan–Meier survival curves. We found that HCC patients in low CASC2 group presented a significant poorer outcome compared to those in high CASC2 group (P < 0.001, respectively. Fig. 8a). Meanwhile, miR-367 high-expressing HCC patients showed a shorter OS and DFS compared to miR-367 low-expressing cases. (P < 0.01, respectively. Fig. 8b). Among these four subgroups, HCC patients with high CASC2 expression and low miR-367 expression had the best prognosis, while CASC2 low-expressing and miR-367 high-expressing HCC patients showed the poorest prognosis (Fig. 8c). These data indicated that CASC2 and miR-367 expression, especially their combination, may be ponderable and promising factors for predicting the prognosis of HCC patients.