The Roles of MicroRNA-122 Overexpression in Inhibiting Proliferation and Invasion and Stimulating Apoptosis of Human Cholangiocarcinoma Cells

Our study investigated whether microRNA-122 (miR-122) played important roles in the proliferation, invasion and apoptosis of human cholangiocarcinoma (CC) cells. QBC939 and RBE cells lines were chosen and divided into five groups: miR-122 mimic group, anti-miR-122 group, negative control (NC) group, mock group and blank group. MiR-122 expression was measured by qRT-PCR. Roles of miR-122 in cell proliferation, apoptosis and invasion were investigated using MTT assay, flow cytometer and Transwell invasion assay, respectively. MiR-122 expression was lower in CC tissues and QBC939 cell than that in normal bile duct tissues, HCCC-9810 and RBE cells. In both QBC939 and RBE cells lines, miR-122 expression was higher in miR-122 mimic group than that in NC group, mock group and blank group; opposite results were found in anti-miR-122 group. Cell proliferation and invasion were remarkably inhibited in miR-122 mimic group after 48 h/72 h transfection, while apoptotic cells numbers were much greater in miR-122 mimic group; the opposite results were obtained from anti-miR-122 group (all P < 0.05). MiR-122 expression was significantly weaker in CC tissues, and miR-122 overexpression might play pivotal roles in inhibiting proliferation, stimulating apoptosis and suppressing invasion of CC cells, suggesting a new target for CC diagnosis and treatment.

Cells transfection and grouping. The miR-122 expression of different cell lines was measured through qRT-PCR, the result showed that the expression level of miR-122 in QBC93 cells was lower than that in HCCC-9810 and RBE cells, indicating that the QBC93 cell line was suitable for the following experiment. The reason might be that QBC93 cells were derived from liver metastasis foci of extrahepatic bile duct carcinoma, while HCCC-9810 and RBE cells were all from intrahepatic primary foci. Hence the malignant degree of QBC939 cells was comparatively high. One day before the transfection, RBE cells were seeded in 12-well plates at 0.5 × 10 5 cells/well. The cells were plated to 60% ~ 80% confluence within 24 hours, and 1 μ L green fluorescent protein (GFP) (0.5 μ g/μ l) were then added to the transfected cells. After 72 h incubation, the expression of GFP was observed under the fluorescence microscope with a fluorescence intensity at 350-420 nm. Detailed grouping information were as follows: the miR-122 mimic group was transfected with miR-122 mimics; the anti-miR-122 group was transfected with miR-122 inhibitors; the negative control (NC) group was transfected with miR-122-NC; the Mock group was transfected with reagent by using Lipofectin TM 2000; no transfection was performed on the Blank group. The sequences of transfected miR-122 mimics, miR-122 inhibitors and miR-122-NC (designed and synthesized by Shanghai GenePharma Co., Ltd) were 5′ -UGGAGUGUGACAAUGGUGUUUG-′ 3, 5′ -CAAACACCAUUGUCACACU-′ 3, and 5′ -UUCUCCGAACGUGUCACGUTT-′ 3 respectively.
Cell proliferation assay. The proliferation of QBC939 and RBE cells were examined at the logarithmic growth phase. The cells were then recovered by digestion with 0.25% trypsin for 20 min, then a volume of RPMI-1640 containing 10% (v/v) FBS was used to terminate the digestion and to suspend the cells with a concentration of 2 × 10 5 cells/mL. After the incubation in a 24-well plate (500 μ L/well), cells were grouped and transfected when these cell lines reached 80% confluency, and three parallel wells were measured in each group. After a 24 h transfection, 1 mL medium was used to culture the suspensions, and then the suspensions were transferred into the 96-well plates after a 1:5 dilution; six parallel wells were measured in each group. After 24 h, 48 h and 72 h of transfection, 10 μ l MTT (5 mg/mL) was added to each well, and then the cells were incubated for 4 h in a 5% CO 2 at 37 °C 19 . Subsequently, the crystallization (purple-red) was observed under the inverted microscope. After the supernatant was discarded, 100 μ l DMSO was added to each well, and the mix was shaken for 15 min. The concentrations after 24 h, 48 h and 72 h were measured by absorbance of 570 nm wavelength with a reference wavelength of 630 nm using enzyme-linked immunosorbent assays. All experiments were repeated for at least three times. The cell growth curve was drawn to show cell proliferation rate.
Cell apoptosis assay. The transfected QBC939 and RBE cells were collected after 48 h (or 72 h) of transfection. Cells were fixed in 70% ethanol for 24 h and centrifuged once. Then DNA-binding dye propidium iodide (DAPI) (10 μ g/ml) was added to the cell suspensions for staining, and the mixed solution was incubated at room temperature for 5 min in the dark. Then the samples were observed using the fluorescence microscope. Additionally, the transfected cells after 48 h were washed for three times by cold PBS and fixed in 70% ethanol. After fixation, the cell suspensions were stained in 500 μ l precooled 1× binding buffer containing 5 μ l annexin V/propidium iodide medium and 2.5 μ l propidium iodide. Cell apoptosis was examined on FACScan flow cytometer (Becton-Dickinson, Sunny-vale, CA, USA) 20 .

Cell invasion assay.
At 48 h after the cell transfection, cells were added to 1.5 ml EP tube and centrifuged at 2000 rpm for 5 min. Then the supernatant was discarded and the cells were re-suspended in 200 μ l serum-free Dulbecco's modified Eagle's medium. The prepared 1 × 10 5 cell was added to the upper chamber coated with Matrigel (8 μ m pore size) and polycarbonate membrane. RPMI1640 medium with 10% FBS was placed in the lower chamber 21 . After 24 h at 37 °C, the filters with the mixture of methanol and acetic acid (3:1) were fixed for 30 min, and the cells were stained for 15 min with crystal violet solution. Invasive cells (five high-power fields) were counted using a light microscope.

Statistical analysis.
All the experiments were repeated for three times independently. Enumeration data comparisons were conducted with I 2 test. Numerical data were presented as means ± standard deviation. The comparisons between groups were performed using t test; comparisons among multiple groups were conducted using One-way analysis of variance (ANOVA). Robust tests of homogeneity of variances were conducted before the analysis. Interactions among multiple groups were tested by Fishers-LSD t-tests. SPSS 19.0 software (SPSS, Chicago, IL) was applied for statistical analyses, and P < 0.05 was considered to be statistically significant. Fig. 1A, qRT-PCR results indicated that the expression of miR-122 was significantly lower in CC tissues than that in the normal bile duct tissues, and the relative expression levels (2 −ΔΔCt ) were (0.442 ± 0.051) and (0.990 ± 0.121), respectively (t = 14.10, P < 0.001). The relative expressions of miR-122 in HCCC-9810, RBE and QBC939 were (1.582 ± 0.088), (1.534 ± 0.047), and (0.968 ± 0.012), respectively. Furthermore, the expression level of miR-122 in QBC939 was significantly lower than those of the first two cell lines (both P < 0.05), while there was no significant difference in the expression levels of miR-122 between HCCC-9810 and RBE (P > 0.05) (Fig. 1B).

Association of miR-122 expression with the clinicopathological features of CC patients. The
associations between miR-122 expression and the clinicopathological features of the CC patients were presented in Table 1. The results showed that there was no significant association between miR-122 expression and gender, age, tumor size or differentiation (all P > 0.05). However, the expression of miR-122 was found significantly associated with lymph node metastasis and distant metastasis (both P < 0.05).
Transfection efficiency of miR-122 in CC QBC939 and RBE cell lines. The transfection efficiency of GFP was observed by fluorescence microscope for many times for this kind of transient transfection assay, and results indicated that the transfection efficiency of liposome miR-122mimics/inhibitor and Lip2000 in QBC939 and RBE cells were both better than those in the other cells, with the transfection efficiency reached to 70%-80% (Fig. 2).

MiR-122 expression level after transfection.
The amplification curve of target gene miR-122 was in accordance with the smooth curve with four characteristic stages (baseline period, exponential phase, exponential phase, platform stage). As shown in Fig. 3, the dissolution curve had only one peak and no impurity peak, suggesting that the gene was effectively amplified with a single amplification product, which met the experiment requirement. After the transfection with miR-122 mimics, the miR-122 expression levels in QBC939 and RBE cells of the NC group, the Mock group and the Blank group were significantly lower than that of miR-122 mimic group (all P < 0.05). However, the expression levels of miR-122 in QBC939 and RBE cells of the NC group, the Mock group and the Blank group were significantly higher than that of anti-miR-122 group (all P < 0.05). No significant difference in miR-122 level was found among the NC group, the Mock group and the Bland group (all P > 0.05) ( Fig. 4 and Table 2).
The role of miR-122 in inhibiting CC cell proliferation. At 24 h after co-culture, QBC939 and RBE cell proliferation had no significant difference among the groups (all P > 0.05). However, compared to the NC group, Mock group and Blank group, QBC939 and RBE cell proliferations were both significantly inhibited in the miR-122 mimic group at the time points of 48 h and 72 h after the transfection (P < 0.05), while the cell proliferation was remarkably promoted in the anti-miR-122 group (Tables 3-4  The role of miR-122 in stimulating CC cell apoptosis. As shown in Fig. 6, the rates of QBC939 cell apoptosis in the NC group, the Mock group and the Blank group were (8.76 ± 0.15) %, (8.72 ± 0.18) % and (8.57 ± 0.16) % respectively, with no significant difference between any two (all P > 0.05). Similar results were also observed with respect to the rate of RBE cell apoptosis in the NC group (11.91 ± 0.15) %, the Mock group (11.87 ± 0.11) % and the Blank group (12.01 ± 0.13) %, showing no apparent statistically difference (all P > 0.05) (Fig. 7). Besides, in the miR-122 mimic group, the rate of QBC939 cell apoptosis was (21.63 ± 0.60) %, and in the anti-miR-122 group, the rate of QBC939 cell apoptosis was (4.11 ± 0.25) %, suggesting that rates of apoptosis in the NC group, the Mock group and the Blank group were significantly different from those in the miR-122 and the anti-miR-122 group (all P < 0.05). Similarly, the rate of RBE cell apoptosis in the miR-122 mimics group and the anti-miR-122 group was (27.76 ± 0.31) % and (6.71 ± 0.10) %, respectively, which were significantly different from those in the NC group, the Mock group and the Blank group (all P < 0.05).
The role of miR-122 in suppressing CC cell invasion. The cell invasion of QBC939 were measured after 24 h using Transwell invasion assay and Matrigel. The number of QBC939 cells migrating into Matrigel in the miR-122 mimic group (33.47 ± 3.292) was significantly smaller than that in the NC group (64.87 ± 4.42), the Mock group (67.87 ± 3.502) and the Blank group (65.53 ± 3.796) (all P < 0.05). However, the number of invasive QBC939 cells into Matrigel in anti-miR-122 group was (101.02 ± 5.92), significantly greater than those in the NC group, the Mock group and the Blank group (all P < 0.05). The number of QBC939 cells migrating into Matrigel was not significantly different among the NC group, the Mock group and the Blank group (all P > 0.05) (Fig. 8). On the other side, the numbers of RBE cells migrating into Matrigel in the miR-122 mimic group and the anti-miR-122 group were (21.19 ± 3.292) and (71.21 ± 3.091), respectively. In contrast, the number of migrating cells in the miR-122 mimic group was significantly greater than that in the NC group (43.96 ± 2.567), the Mock group (45.76 ± 3.011) and the Blank group (43.42 ± 2.201), while the number in the anti-miR-122 group was significantly smaller than the latter three groups (all P < 0.05) (Fig. 9).

Discussion
MiR-122 serves as functional miRNA involved in regulating lipid metabolism, cell differentiation, hepatic metabolism and hepatitis C virus replication 22 . The design of the experiment constituted a knockout mouse model modeling the loss of miRNA in hepatocytes, and finally liver disruptions including steatosis, inflammation and hepatocyte apoptosis were identified, suggesting that miRNAs played a key role in the liver function and that miR-122 may be associated with hepatocarcinogenesis by influencing hepatocyte survival and tumor progression 23,24 . MiR-122 deficiency in CC patients may contribute to the dysregulation of mitochondrial functions related with liver function, so its loss of expression may lead to increased morbidity and mortality and may predict poor prognosis of CC patients 25  confirmed that the expression of miR-122 was significantly lower in CC tissues than that in normal bile duct tissues, indicating that the decreased expression of miR-122 might be closely related to the development of CC. Therefore, we suggested that regulating the level of miR-122 could be used in controlling the progression of CC, which was consistent with the previous study findings revealing that the modulation  of miR-122 level may be one of the targets for the prognosis prediction and the design of effective therapies for CC patients 26 . Furthermore, miR-122 participates in the tumor cell survival, proliferation, differentiation and migration as a tumor suppressor 27 . Previous evidence have shown that compared with adjacent benign liver, miR-122 appears to be down regulated in deficient liver, suggesting the potential of miR-122 as a novel biomarker for liver injury 28 . Further, the magnitude of decrease of the miR-122 level was previously proved to be helpful for the measurement the increase in the grade of hepatocellular injury 29 . Our study also confirmed that the decreased expression of miR-122 was also associated with lymph node metastasis and distant metastasis. Similarly, it was also previously reported that serum miR-122 level was an important indicator in the diagnosis of liver injury and could be used as a biomarker and therapeutic target for diagnosis and treatment of CC- 30 . In addition, we identified that the expression level of miR-122 in QBC939 was significantly lower than those of HCCC-9810 and RBE cells, while there was no significant difference between HCCC-9810 and RBE. These results supported that QBC939 cells   Table 2. Quantitative real-time polymerase chain reaction (qRT-PCR) testing microRNA-122 (miR-122) expression level after transfection in different groups. Note: NC: negative control. a compared to the blank group, P < 0.05. b compared to the mock group, P < 0.05. c compared to the NC group, P < 0.05. d compared to the miR-122 group, P < 0.05.  were promising for further in vitro experiment; the reason might be that QBC939 cells widely exist in metastatic foci in liver of extrahepatic bile duct carcinoma, while HCCC-9810 and RBE cells mainly exist in the primary foci in the liver, so QBC939 had a higher degree of malignancy 31 .
In our study, we found that the overexpression of miR-122 played pivotal roles in inhibiting proliferation, stimulating apoptosis and suppressing invasion of QBC939 and RBE cells. As the most abundant miRNA in the liver, miR-122 is well known for its biologic function in maintaining liver  homeostasis, as well as its role in regulating cell growth, differentiation, apoptosis and metabolism in the carcinogenesis in the liver, which is detrimental to normal liver function 32,33 . The over expression of insulin-like growth factor 1 receptor (Igf1R) in the development of tumors stimulates cell growth, survival and proliferation and regulates the initiation of cancer cell metastasis; the level of Igf1R is negatively associated with the level of miR-122 expression, implying that the overexpression of miR-122 can inhibit tumor cell growth and proliferation by suppressing Igf1R expression 34 . MiR-122 also functions as a key modulator of cyclin G1 expression and there is also a negative association between the levels of miR-122 and cyclin G1 (CCNG1) 35 . Decreased level of miR-122 was detected in the CC patients, leading to the increased level of CCNG1, which is associated with accumulation of tumor cells via affecting cell cycle (http://wenku.baidu.com/link? url= qineu6YlskaIZAh01hq9dV1Uw9rC6aU_ JUHsgmy_NaJyMuVaIYFt4BErfqFSganqod6GceBfAMuN5rOiL1NZAZ1yikvVs_mgqjoeFy232we). The imbalance between miR-122 and CCNG1 may help to inhibit the tumor cell proliferation of CC through triggering p53 tumor suppressor gene 36 .
A previous study demonstrated that the abnormal expression of miR-122 was responsible for hepatocarcinogenesis; the loss of miR-122 led to the down-regulation of tumor cell apoptosis 37 . MiR-122 expression in tumor cells is suppressed in the early phase of CC, resulting in severe metastasis of tumor cells, and the restoration of its expression may help to control tumor progression of CC patients 38 . As a vital apoptosis regulator, and the mechanism of miR-122 in CC cells involves suppressing Bcl-W mRNA and the protein level, consequently leading to large reduction of cell motility 39 . Bcl-W activity can inhibit cancer cell apoptosis, and the overexpression of miR-122 can inhibit the expression of Bcl-W and CCNG1 to induce cell apoptosis and cell cycle arrest 40 . Thus, down-regulated miR-122 is potential to be an independent predictor of the development and progression of CC characterized by the loss of anti-apoptotic effect 41 . Our study also found that the role of miR-122 in antitumor activity is manifested in suppressing tumor cell invasion. Based on previous study results, miR-122 down-regulation was identified to be associated with hepatic cell invasion, intrahepatic metastasis and reduction of tumor cell sensitivity to drug agent resulting in tumor aggressiveness 42 . As a tumor suppressor, miR-122 can inhibit intrahepatic invasion and migration of CC cells by suppressing angiogenesis through regulating the disintegrin and metalloprotease 17 activity 43 .