miR-146a and miR-146b promote proliferation, migration and invasion of follicular thyroid carcinoma via inhibition of ST8SIA4

Follicular thyroid carcinoma (FTC) is a more aggressive form of thyroid cancer than the common papillary type. Alpha-2,8-sialyltransferase (ST8SIA) family members are expressed in various cancers and may be associated with FTC progression. In this study, we measured ST8SIA family expression in two FTC cell lines with different invasive potentials (FTC-133 and FTC-238) and Nthy-ori 3-1 cell lines, as well as FTC and normal thyroid tissues. ST8SIA4 was downregulated in the highly invasive FTC-238 cells and FTC tissues. Additionally, ST8SIA4 inhibited proliferation, migration and invasion of FTC both in vitro and in vivo. miR-146a and miR-146b were previously shown to be upregulated in thyroid carcinoma, and bioinformatics analyses indicated that miR-146a and miR-146b inhibit ST8SIA4. We found that miR-146a and miR-146b were significantly upregulated in FTC and promoted tumour progression. Furthermore, ST8SIA4 restoration decreased the invasiveness of miR-146a/b-overexpressing FTC-133 cells, and ST8SIA4 suppression reversed the effects of miR-146a/b inhibition in FTC-238 cells. We showed that miR-146a/b activated the PI3K-AKT-mTOR signalling pathway at least partially via suppression of ST8SIA4. Thus, our results demonstrate that miR-146a and miR-146b promote proliferation, migration and invasion of FTC via inhibition of ST8SIA4.


INTRODUCTION
Thyroid cancer is the most common endocrine neoplasm; it accounts for approximately 1% of all new malignant diseases, and its annual incidence is increasing worldwide [1]. Thyroid cancer is classified into four types: papillary (85% prevalence), follicular (10% prevalence), medullary (3-4% prevalence), and anaplastic thyroid cancer (1-2% prevalence) [2,3]. Compared with papillary thyroid cancer (PTC), follicular thyroid cancer (FTC) is more aggressive and harder to diagnose [4]. Although most FTCs can be cured with surgery, the overall 10-year survival rate of FTC is lower than that of PTC [5]. Therefore, understanding the molecular mechanisms of FTC tumours is important for developing better diagnostic strategies and improving the therapeutic outcome of FTC.
Alpha-2,8-sialyltransferase (ST8SIA I-VI) mediates the transfer of sialic acid with an alpha 2,8-linkage [6]. Previous reports demonstrated that sialic acid is involved in various biological processes, including cell-cell adhesion, immune defence, tumour cell metastasis, and inflammation [7][8][9][10]. Increasing evidence has indicated that the ST8SIA family is associated with several human tumours. For example, ST8SIA4 is involved in the development of multidrug-resistant neoplasms in acute myeloid leukaemia cells [11]. In addition, ST8SIA2 promotes the invasive properties and chemosensitivity of human hepatocellular carcinoma [12]. However, whether the ST8SIA family has biological functions in FTC is poorly understood.
MicroRNAs (miRNAs) are a class of small noncoding RNAs that activate or inhibit target genes via

Research Paper
translational repression or mRNA degradation [13]. miRNAs regulate many biological processes, such as the cell cycle, senescence, DNA repair and tumorigenesis [14,15]. Several reports have indicated that miRNAs are associated with thyroid cancer. For example, miR-34a promotes cell proliferation and inhibits apoptosis in PTC [16], whereas miR-200 inhibits epithelial-mesenchymal transition in anaplastic thyroid cancer [17]. However, whether miRNAs suppress ST8SIA and thereby affect the tumour characteristics of FTC remains unknown.
Activation of the PI3K-AKT-mTOR signalling pathway is implicated in the progression of thyroid cancer [18][19][20], and the PI3K-AKT signalling pathway is involved in the regulation of the sialyltransferase family [11]. Moreover, miRNAs have been reported to stimulate the PI3K-AKT-mTOR pathway in various diseases [21][22][23]. Therefore, we sought to determine whether PI3K-AKT-mTOR signalling is regulated by the ST8SIA family and miRNAs in FTC. This hypothesis has not been verified in previous studies.
In this study, we first investigated the expression and detailed role of the ST8SIA family variants in different invasive FTC cells. To elucidate the potential underlying mechanisms, we determined the role of miRNAs in the regulation of ST8SIA and FTC tumour characteristics. This work aimed to identify better preoperative markers for the diagnosis of FTC and new potential targets for its treatment.

RESULTS
Within the ST8SIA family, ST8SIA4 and ST8SIA6 are significantly altered in FTC Real-time PCR analyses were performed to evaluate the expression of ST8SIA family members in FTC-238 cells (highly invasive), FTC-133 cells (poorly invasive) and Nthy-ori 3-1 cells (non-invasive). As shown in Figure 1A, the expression of ST8SIA4 was significantly downregulated in the FTC-238 cells compared with that of the FTC-133 and Nthy-ori 3-1 cells (*p < 0.05). Additionally, ST8SIA6 was significantly increased in the FTC-238 cells compared with that of the FTC-133 and Nthy-ori 3-1 cells (*p < 0.05). However, in all three cell lines, ST8SIA1 and ST8SIA5 did not show significant differences in expression, and ST8SIA2 and ST8SIA3 were not expressed (*p > 0.05) ( Figure 1A).
To further verify these findings in cells, we examined the expression of the ST8SIA gene family in FTC and normal thyroid tissues. As expected, mRNA expression of ST8SIA4 was significantly lower in the FTC tissues than that of the normal thyroid tissues (*p < 0.05) ( Figure 1B). However, no significant differences were found in the remaining ST8SIA family members in the tissue samples, whereas the expression of ST8SIA6 was significantly higher in the FTC tissues than that of the normal thyroid tissues (*p < 0.05).
We measured ST8SIA4 and ST8SIA6 protein levels in three cell lines using western blotting ( Figure 1C). Similar to the mRNA levels, ST8SIA4 protein expression was lower in FTC-238 cells than that of the FTC-133 and Nthy-ori 3-1 cells, and ST8SIA6 protein expression was higher in the FTC-238 cells than that of the FTC-133 and Nthy-ori 3-1 cells (*p < 0.05). These results indicate that ST8SIA4 and ST8SIA6 regulate FTC development.

ST8SIA4 mediates the proliferation, migration and invasion of FTC cells both in vitro and in vivo
Based on the significant alterations in ST8SIA4 and ST8SIA6 expression in the highly invasive FTC cells, we investigated whether these two ST8SIA members could affect the tumour properties of FTC cells. First, we transfected FTC-133 and FTC-238 cell lines with ST8SIA4specific shRNA or ST8SIA4 expression vectors to establish specific knockdown and overexpression cell lines, respectively (*p < 0.05) (Figure 2A, 2B). CCK-8 assays demonstrated that ST8SIA4 inhibition promoted FTC-133 cell proliferation, and ST8SIA4 overexpression significantly inhibited FTC-238 cell proliferation compared with that of the negative control cells (*p < 0.05) ( Figure 2C). We also examined the colony formation capacity of the FTC-133 and FTC-238 cells. When ST8SIA4 was inhibited, the FTC-133 cells formed more colonies than those of the control groups, and when ST8SIA4 was overexpressed, FTC-238 cells formed fewer colonies than those of the control groups (*p < 0.05) ( Figure 2D). To assess the effect of ST8SIA4 on cellular motility, we conducted wound-healing assays and transwell assays to measure the migration and invasion of FTC-133 and FTC-238 cells. Cell migration and invasion were increased in the ST8SIA4-silenced FTC-133 cells and decreased in ST8SIA4-overexpressing FTC-238 cells (*p < 0.05) ( Figure 2E, Supplementary Figure 1A). However, overexpression or inhibition of ST8SIA6 did not affect the proliferation, migration and invasion of FTC cells in vivo (data not shown); therefore, we focused on the role of ST8SIA4 in subsequent experiments.
We further assessed whether ectopic expression of ST8SIA4 inhibited tumour growth in vivo. A significant decrease in tumour weight was observed in ST8SIA4overexpressing tumours, and tumour weights were increased in the ST8SIA4-inhibited tumours compared with those of the controls (*p < 0.05) ( Figure 2F). These data suggest that ST8SIA4 is an important negative regulator that mediates the proliferation, migration and invasion of FTC cells both in vitro and in vivo. These results prompted us to determine the molecular mechanisms underlying ST8SIA4 downregulation.

miR-146a/b can directly target and inhibit the expression of ST8SIA4
miRNAs are potent gene regulators in a wide range of diseases. Because ST8SIA4 may be suppressed by miRNAs, we searched for potential targets of ST8SIA4 using the bioinformatics algorithms TargetScan (http:// www.targetscan.org), miRanda (http://www.microrna.org/ microrna/home.do) and mirbase (http://www.mirbase.org). We identified several miRNAs that may target ST8SIA4, including miR-146a, miR-146b, miR-664-3p, miR-216a-5p, and miR-380-3p (data not shown). Notably, miR-146a and miR-146b were reportedly upregulated in thyroid cancer and promote the progression of thyroid cancer [24][25][26]. Figure 3A shows the sequence of the miR-146a and miR-146b seed regions and their 3ʹUTR target sites for ST8SIA4. To confirm that ST8SIA4 is a direct target of miR-146a and miR-146b in FTC, we transfected a luciferase fusion construct containing either the wild-type or mutated ST8SIA4 3ʹUTR into FTC-238 cells. As shown in Figure 3B, the luciferase activity of the wild-type (WT) 3′UTR reporter gene was significantly decreased, whereas the luciferase activity of the mutant reporter gene was not affected (*p < 0.05). These data indicate that miR-146a/b binds to the 3′UTR of the ST8SIA4 gene.
To further explore the expression of miR-146a/b in FTC, we examined the expression of miR-146a and  miR-146b in cells and tissues ( Figure 3C). As shown in Figure 3C, miR-146a and miR-146b levels increased in the FTC-238 cells compared with those of the FTC-133 or Nthy-ori 3-1 cells (*p < 0.05). In addition, the expression of miR-146a and miR-146b was higher in the FTC tissues than that of the normal thyroid tissues ( # p < 0.05) ( Figure 3C). We further examined the expression of miR-146a and miR-146b in FTC-133 cells transfected with miR-146a or miR-146b mimics and the FTC-238 cells transfected with miR-146a or miR-146b antagomirs ( Figure 3D).
The ST8SIA4 protein levels were downregulated in the FTC-133 cells transfected with miR-146a and/or miR-146b mimics compared with those of the NC cells (*p < 0.05) ( Figure 3E). Moreover, the ST8SIA4 protein levels were increased in the FTC-238 cells transfected with miR-146a and/or miR-146b antagomirs compared with those of the NC cells (*p < 0.05) ( Figure 3F). Taken together, these results demonstrate that miR-146a and miR-146b are negative regulators of ST8SIA4 in FTC.

miR-146a/b regulates the proliferation, migration and invasion of FTC-133 cells and FTC-238 cells both in vitro and in vivo
To further elucidate the role of miR-146a/b in FTC, we investigated the proliferation, migration and invasion of miR-146a/b-silenced FTC-238 cells or miR-146a/boverexpressing FTC-133 cells. CCK-8 assays demonstrated that miR-146a and/or miR-146b overexpression significantly promoted FTC-133 cell proliferation relative to that of the NC cells, whereas miR-146a and/or miR-146b inhibition reduced FTC-238 cell proliferation compared with that of the NC cells (*p < 0.05, # p < 0.05) ( Figure 4A). The colony formation assays indicated that FTC-133 cells overexpressing miR-146a and/or miR-146b formed more colonies than those of the control groups, and miR-146a-and/or miR-146b-inhibited FTC-238 cells formed fewer colonies than those of the control groups (*p < 0.05, # p < 0.05) ( Figure 4B). As shown in Figure 4C, miR-146a and/or miR-146b overexpression increased the migration and invasion of FTC-133 cells, whereas the inhibition of miR-146a and/or miR-146b reduced the migration and invasion of FTC-238 cells (*p < 0.05, # p < 0.05).
Consistent with the results of the above assays, the nude mouse xenograft model showed an increase in tumour weight relative to that of the NC cells when the FTC-133 cells were transfected with miR-146a and/ or 146b mimics (*p < 0.05) ( Figure 4D) and a decrease in tumour weight relative to the NC cells when the FTC-238 cells were transfected with miR-146a and/or miR-146b antagomirs ( # p < 0.05) ( Figure 4E, Supplementary Figure 1B). These data suggest that miR-146a/b promotes the proliferation, migration and invasion of FTC both in vitro and in vivo.

ST8SIA4 suppresses the effects of miR-146a/b in FTC
To investigate whether the effects of miR-146a/b in FTC are mediated by ST8SIA4, we rescued ST8SIA4 expression in the miR-146a/b-overexpressing FTC-133 cells and suppressed ST8SIA4 expression in the miR-146a/ b-inhibited FTC-238 cells. As shown in Figure 5A-5D, ST8SIA4 restoration decreased the proliferation, migration and invasion of miR-146a/b-overexpressing FTC-133 cells and partially decreased the migration and invasion capacity of these cells compared with those of the NC cells (*p < 0.05). Furthermore, ST8SIA4 suppression partially increased the proliferation, migration and invasion of miR-146a/b-inhibited FTC-238 cells compared with those of the NC cells (*p < 0.05). Importantly, the FTC-133 cells transfected with the ST8SIA4 expression vector reversed the tumour growth induced by miR-146a/b in vivo, whereas the FTC-238 cells transfected with the ST8SIA4-specific shRNA reversed the tumour growth inhibitory effect of miR-146a/b inhibition in vivo ( Figure 5E, Supplementary Figure 1C). These results suggest that ST8SIA4 is an important participant in the regulation of miR-146a/b and suppresses FTC proliferation, migration and invasion.

DISCUSSION
FTC is the second most common malignancy originating from thyroid follicular cells. In this study, we provided the first evidence of the following four findings: (1) ST8SIA4 expression is lower in FTC tissues and cultured highly invasive FTC cells than that in normal tissues and cells and suppressed the proliferation, migration and invasion of FTC both in vitro and in vivo; (2) miR-146a/b is significantly upregulated in the highly invasive FTC cells, which are direct targets of ST8SIA4; (3) miR-146a/b inhibition attenuates the proliferation, migration and invasion of FTC by repressing ST8SIA4 both in vitro and in vivo; and (4) miR-146a/b suppresses ST8SIA4 in FTC partly via regulation of the PI3K/AKT/ mTOR signalling pathway. To the best of our knowledge, this is the first study to explore the inhibitory role of ST8SIA4 in regulating the tumour characteristics of FTC and the relationship of ST8SIA4 and miRNAs in FTC.
Sialyltransferases catalyse the transfer of sialic acid to proteins and lipids and participate in the synthesis of the core structure of oligosaccharides [27]. Native sialylation is critical for the function of therapeutic proteins because it affects the physical, chemical and immunogenic properties of glycoproteins [28]. Importantly, altered sialylation is a hallmark of cancer, and abnormally expressed sialylated glycans are biomarkers of cancer [29]. In this study, we found that the expression of ST8SIA4, a type of sialyltransferase, was lower in highly invasive FTC cells (FTC-238 cells) than that in minimally invasive FTC cells (FTC-133 cells). Then, we performed functional analyses and found that ST8SIA4 downregulation may contribute to the aggressive properties of FTC. These findings prompted us to further explore the detailed mechanisms underlying ST8SIA4 downregulation in FTC.
miRNAs are potent gene regulators that have been implicated in multiple pathophysiological processes and may play a crucial role in the progression of thyroid cancer [17,[30][31][32]. Thus, we sought to determine whether ST8SIA4 expression is controlled by miRNAs. Bioinformatics analysis and luciferase activity assays confirmed miR-146/b binding to the 3′UTR of the ST8SIA4 mRNA. According to these results, we suggest that miR-146a and miR-146b are potential regulators of ST8SIA4.
Previous reports have shown that miR-146a and miR-146b are positively regulated in thyroid cancer [24,26,33]. We found that miR-146a and miR-146b promoted the tumour characteristics of FTC both in vivo and in vitro. Importantly, our gain-and loss-of-function analyses showed that the effects of miR-146a/b knockdown in the FTC-238 cells were reversed by inhibition of ST8SIA4 and vice versa. Thus, miR-146a/b suppresses ST8SIA4 in FTC, which consequently promotes the proliferation, migration and invasion of FTC cells.
Ma et al. found that the PI3K-AKT pathway is located downstream of ST8SIA4 [11]. In this study, we further verified that the PI3K-Akt-mTOR signalling pathway is involved in the regulation of miR-146a/b and ST8SIA4 in FTC. In addition, the expression of the PI3K-AKT-mTOR pathway mimics the expression of miR-146a/b in FTC-238 cells and FTC-133 cells, which are negatively regulated by ST8SIA4. Therefore, we suggest that PI3K-Akt-mTOR pathway activation is correlated with the tumour properties of FTC.
A single miRNA may target several transcripts [34]; therefore, additional miR-146a and miR-146b targets may be involved in regulating the tumour characteristics of FTC. PTTG1, which is targeted by miR-146a and miR-146b, is also regulates thyroid cancer [35]. However, PTTG1 expression was not significantly altered in FTC-133 and FTC-238 cells when miR-146a and miR-146b were inhibited (data not shown). PTEN, another target of miR-146a and miR-146b, may also suppress thyroid cancer [36][37][38]. However, whether this gene participates in FTC progression requires further investigation.
In this study, we found that miR-146a and miR-146b promote the proliferation, migration and invasion of FTC via inhibition of ST8SIA4 and regulation of the PI3K-AKT-mTOR signalling pathway. Our findings further elucidate the molecular mechanisms underlying FTC progression and provide candidate targets for the prevention and treatment of FTC. Although additional studies are needed to explore these therapeutic targets, we confirmed that ST8SIA4 and miR-146a/b play key roles in FTC.

Tissue samples
After informed consent was obtained, resected thyroid cancer tissue samples were collected from 110 patients diagnosed with FTC (48 men and 62 women; median age, 58; range, 37-76). Normal thyroid tissue samples were collected from 110 patients diagnosed with thyroid nodules (41 men and 69 women; median age, 41; range, 24-69). All patients underwent surgical resection between January 2012 and June 2015 at the Second Affiliated Hospital of Dalian Medical University. The investigation and informed consent protocol were certified and approved by the Ethics Committee of the Second Affiliated Hospital of Dalian Medical University. The pathological diagnosis of the collected samples was confirmed according to the Union for International Cancer Control (UICC) criteria. The clinicopathological data were extracted from the patient's medical records. None of the patients had received chemotherapy or radiation therapy. The collected tissues were snap-frozen in liquid nitrogen and stored at -80°C until use.

RNA isolation and real-time PCR
Total RNA was isolated from the cell lines using an RNeasy Mini Kit (Qiagen, Valencia, CA), and cDNA was synthesized using a QuantiTect Reverse Transcription Kit (Qiagen, Valencia, CA) according to the manufacturers' protocol. For the mRNA expression analysis, 3 μg of total RNA was converted into cDNA using M-MLV Reverse Transcriptase (Invitrogen). The ST8SIA mRNA was quantified using a SYBR Green Quantitative Real-time PCR mix (TaKaRa, Otsu, and Shiga, Japan) and normalized to GAPDH (glyceraldehyde-3-phosphate dehydrogenase). The expression of miR-146a/b was determined using a mirVana Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR) microRNA Detection Kit according to the manufacturerʹs protocol (Ambion Inc., Austin, TX, USA) and normalized relative to U6-small nuclear RNA using the 2 −ΔΔCT method. For reverse transcription, 5 μg of total RNA was converted to cDNA using a TaqMan MicroRNA Reverse Transcription Kit (Fermentas) according to the manufacturer's protocol. The resulting cDNA was diluted 1:10 and used for PCR with 4 μL of miR-146a/b or U6 TaqMan primers and SYBR Green/Fluorescein qPCR Master Mix (Fermentas) with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA). Next, 10 ng of total RNA was reverse transcribed using a High-Capacity cDNA Archive Kit (Applied Biosystems Inc., Foster City, CA) and then amplified on a ABI 7500 Real-Time PCR System (Applied Biosystems Inc., Foster City, CA). Real-time PCR was performed for 2 min at 95°C and then for 40 cycles of amplification for 20 s at 95°C and 1 min at 60°C.

Western blot analysis
Whole cell proteins were electrophoresed under reducing conditions on a 10% polyacrylamide gel. The separated proteins were transferred to a polyvinylidene difluoride membrane. After the membranes were blocked with 5% skim milk in PBS containing 0.1% Tween 20 (PBST), they were incubated with the appropriate antibody (1:1000; Abcam, Cambridge, UK) overnight at 4°C and then incubated with peroxidase-conjugated anti-rabbit IgG (1:10000; GE Healthcare UK Ltd., Little Chalfont, U.K.). GAPDH was used as a control. Band intensity was evaluated using an ECL Western Blotting Substrate Kit (Amersham Biosciences, Buckinghamshire, UK) according to the manufacturer's protocol. The bands were analysed using LabWorks (ver4.6, UVP, BioImaging Systems).

Transfection and luciferase activity assay
Mimics, antagomirs and negative control oligonucleotides for hsa-miR-146a/b were obtained from RiboBio Co., Ltd. (Shanghai, China). Plasmids containing WT Luc-ST8SIA4, mutant Luc-ST8SIA4, and lenti-miR-146a/b were synthesized. FTC cells were transfected using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instruction. Luciferase activity was measured 48 h after transfection using the Dual-Luciferase Reporter Assay System (Promega). Firefly luciferase activity was normalized to Renilla luciferase activity for each sample. The mean of the results from the cells transfected with the miR-control was set at 1.0. The data are presented as the mean value ± SD for three repeated experiments.

Downregulation of ST8SIA4 by RNAi
FTC-133 cells were incubated in the appropriate antibiotic-free medium with 10% foetal bovine serum. The cells were then transferred to a 6-well plate prior to incubation in a CO 2 incubator at 37°C to obtain 60-80% confluency. ST8SIA4 shRNA was mixed with Lipofectamine 2000. The cells were then harvested for further analysis. The transfected cells were cultured at 37°C for 6 h and then incubated with complete medium for an additional 24 h. Thereafter, the cells were harvested for further examination.

Overexpression of ST8SIA4
The human ST8SIA4 coding sequences obtained from TaKaRa (Dalian, China) were inserted into the pEGFP-N2 vector (Invitrogen, Carlsbad, CA) at the EcoRI and XhoI sites. The cells were transfected with 5 μg of the target gene expression vector or empty vector (EV) in 100 mm dishes using PolyFect Transfection Reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. After 4 weeks of screening, the cell lines stably expressing ST8SIA4 (overexpression of ST8SIA4 in FTC-238 cells) and the corresponding empty vector (FTC-238/ mock) were established and verified using western blotting.

Cell proliferation assay (CCK8 assay)
For the proliferation assay, the transfected cells and negative control (NC) cells were seeded at 5 × 10 3 cells/ well in 96-well plates and incubated at 37°C for 4 days. A 10 μl aliquot of Cell Counting Kit-8 reagent (Dojindo, Japan) was added to the cells. Following incubation for 3 h, absorbance was measured at 450 nm using a spectrophotometer (Bio-Rad, USA).

In vitro invasion assay
Cell invasion in vitro was demonstrated using 24well Transwell units (Corning, NY, USA) with a 8 μm pore size polycarbonate filter coated with extracellular matrix gel (EC Matrix gel) (Millipore Chemicon, MA, USA) to form a continuous thin layer. The cells (3 × 10 5 ) were harvested in serum-free medium containing 0.1% BSA and then added to the upper chamber. The lower chamber contained 500 μl DMEM. The cells were incubated for 24 h at 37°C in an incubator containing 5% CO 2 . After incubation, the cells on the upper surface of the filter were completely removed by wiping with a cotton swab. The filters were then fixed in methanol and stained with Wright-Giemsa stain. Cells that had invaded the Matrigel and reached the lower surface of the filter were counted using a light microscope at a magnification of 40×.

Wound-healing assay
For the wound-healing assays, confluent monolayers of FTC-133 and FTC-238 cells were treated with 2 μg/ml of mitomycin for 2 h (Sigma-Aldrich). The transfected cells were seeded onto 6-well plates at 1 × 10 5 cells/l and then cultivated in DMEM to maintain adherent cell growth for 6 h. Scratching (wounding) was performed using a 10 μl Eppendorf tip. The cells were cultured in 5% CO 2 at 37°C for 24 h after they were washed in serumfree medium 3 times. Images were taken to determine the number of migrating cells between the scratches using Image-Pro Plus 6.0 software. The mean values and standard deviations were calculated for the intra-group comparisons. The experiment was repeated 3 times.

Colony formation assays
The transfected cells were seeded in 6-well (100 cells/well) culture plates (BD Biosciences). The cells were incubated at 37°C with 5% CO 2 for 7-12 days until visible colony formation was observed in the dish. Subsequently, the culture medium was removed, and the wells were washed twice with phosphate-buffered saline (PBS). The colonies were fixed with 10% methanol for 20 min, dried and stained with 0.1% crystal violet solution for 15 min. Each plate was then washed three times with 9% sodium chloride solution. Cell colonies with > 50 cells were counted and imaged.

In vivo tumourigenicity assay
Athymic nude mice (5-week-old) were obtained from the Animal Facility of Dalian Medical University and provided with sterilized food and water. Approximately 1 × 10 7 cells were injected subcutaneously into the right flank of each nude mouse. After palpable tumours were identified (approximately 4 weeks after tumour cell inoculation), the mice were sacrificed, and their tumours were isolated and weighed.

Statistical analysis
The data are expressed as the mean ± standard deviation (SD). SPSS 17.0 software was used for the statistical analysis, and Student's t-test was used to determine the significance of differences among the examined groups. Normally distributed data were compared using a one-way ANOVA followed by the Student-Newman-Keuls test. A value of p < 0.05 was considered statistically significant.