2.1. HCC Patient Data and Tissue Specimens

A total of 188 HCC tissue specimens were obtained from patients undergoing hepatectomy at the First Affiliated Hospital, Sun Yat-sen University (Guangzhou, China) between January 2006 and December 2008. Another 40 pairs of matched HCC and adjacent noncancerous liver tissue samples were obtained during the operation between January 2009 and May 2013. The samples were snap frozen in liquid nitrogen and stored in a −80°C freezer for subsequent experiments, such as RNA extraction and formalin fixation and paraffin-embedded immunohistochemistry. Disease-free survival (DFS) and overall survival (OS) were used to assess efficacy. The follow-up of the patients was performed as described in our previous work [16]. The deadline for follow-up was 31 October 2015. We collected the patients’ complete clinical and pathological features. The study followed the Ethical Guidelines of the 1975 Helsinki Declaration, which were revised in 2013. All patients received informed consent on the use of their clinical specimens for medical research. The use of human body materials was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University.

2.2. Cell Lines and Cell Culture

The human HCC cell line HepG2 was purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA). The human HCC cell lines HepG3B, HCCLM3, SMMC-7721, and Bel7402 and the normal hepatocyte cell line L02 were all obtained from the Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences and verified.

Cell lines were cultured in low-glucose Dulbecco’s Modified Eagle Medium (DMEM) containing 10% foetal bovine serum (FBS), 100 U/ml penicillin, and 0.1 mg/ml streptomycin at 37°C and 5% CO₂.

2.3. Immunohistochemical Analysis of Gelsolin

Immunohistochemistry (IHC) was performed using Gelsolin antibodies (G4896, Sigma) as described in our previous work [17]. Immunohistochemical analysis was performed by two independent researchers who were blinded to patient data. Shimizu criteria [18] were used to score the expression of Gelsolin protein in HCC specimens from 0 to 3+. Patients with a score of 0 or 1+ were defined as the low expression group, while those with a score of score of 2 or 3+ were defined as the high expression group.

2.4. Producing High- and Low-GSN-Expressing HCC Cell Lines

We purchased a GSN overexpression plasmid from Forevergen Biosciences (Guangzhou, China). The sequence of the GSN-specific antisense oligonucleotide is 5′-UUCAGAACAAAGGCAUCGdTdT-3′. The sequence of the control oligonucleotide is 5′-UUCCGAACGUGUCACGUdTdT-3′. HCCLM3 cells were transfected with a GSN small hairpin RNA (shRNA) or a scramble shRNA, and as previously described, SMMC-7721 cells were transfected with the GSN overexpression plasmid or a vector control [19]. Next, 0.5 mg/mL purinomycin was used to select the transfected cells. Quantitative real-time polymerase chain reaction (qRT-PCR) and western blot (WB) analysis were used to confirm the stably transfected clones.

2.5. qRT-PCR and WB Analysis

qRT-PCR and WB analysis were performed as previously described [20]. qRT-PCR was used to analyse the transcripts of β-actin (control) and GSN. The primers were purchased from Invitrogen (Thermo Fisher Scientific, Inc.). The primer sequences were as follows: β-actin, forward 5′-AGCGAGCATCCCCCAAAGTT-3′ and reverse 5′-GGGCACGAAGGCTCATCATT-3′ and GSN, forward 5′-GGTGTGGCATCAGGATTCAAG-3′ and reverse 5′-TTTCATACCGATTGCTGTTGGA-3′.

2.6. Transwell Migration and Invasion Test

For the transwell migration assay, serum-free DMEM containing 5 × 10⁴ cells was inoculated into the upper chamber of 8 μm transwell inserts (BD Biosciences, Franklin Lakes, NJ), and DMEM containing 10% bovine serum albumin (BSA) was added to the lower chamber. After incubation at 37°C for 24 hours, the cells in the upper chamber were carefully removed, and the cells attached to the lower side of the transwell membrane were fixed in 20% methanol and stained with 0.1% crystal violet. An inverted microscope (Nikon, Chiyoda-Ku, Japan) was used to examine the number of cells. For the transwell invasion assay, all procedures were the same as those described for the transwell migration assay, except that the upper chamber was coated with Matrigel (BD Biosciences, Franklin Lakes, NJ).

2.7. In Vivo Lung Metastasis Model

In vivo metastasis determination was performed in male BALB/c thymus-free nude mice (3-4 weeks old) obtained from the Animal Center of Guangdong Province (Guangzhou, China). The mice were randomized and treated in a nonblinded manner. The use of laboratory animals was based on guidelines from the National Institutes of Health. For the lung metastasis model, HCC cells (1 × 10⁶) suspended in 200 μL of phosphate-buffered saline (PBS) were injected intravenously through the tail vein (5 mice in each group). At the point of termination, the mice were euthanized. Lungs were excised, and metastasis was confirmed by haematoxylin-eosin (HE) staining. The animal experiments were approved by the Life Animal Use Committee of the First Affiliated Hospital of Sun Yat-sen University.

2.8. Statistical Analysis

SPSS software (19.0; SPSS, Inc., Chicago, IL) was used for statistical analysis. The measurement data are expressed as the mean ± standard deviation (SD). Student’s t-test was used for intergroup comparisons. Categorical data were analysed by the chi-square test or Fisher’s exact test. Kaplan–Meier’s method and the logarithmic rank test were used to analyse the survival rate. A Cox regression model of multivariate competition risk was used to further analyse the factors identified as significant (P < 0.1) in the univariate analysis to identify significant independent predictors of DFS and OS in HCC patients. A forward step-by-step programme was used for variable selection and the final multivariable model. P < 0.05 was considered statistically significant.

3.1. GSN Is Overexpressed in HCC Tissues

To investigate the potential clinical relevance of GSN and HCC progression, GSN mRNA and protein levels in paired HCC specimens and adjacent noncancerous liver tissues were measured. The levels of both GSN mRNA and protein were significantly upregulated in HCC tissues compared with adjacent normal liver tissues (Figures 1(a) and 1(b)). Immunohistochemical analysis indicated that GSN mainly localizes within the cytoplasm of HCC tissues, in accord with previous reports, and the intensity of GSN-positive staining was markedly increased in HCC tissues compared with that in noncancerous tissues (Figure 1(c)). We also evaluated the difference in GSN expression between metastatic HCC tissues and HCC tissues in situ (Figure 1(d)). Because GSN expression is higher in HCC tissues, the DAB color development time of the IHC was 20 minutes as in Figure 1(c). In order to show the difference between metastatic and nonmetastatic tissues, in the IHC assay of Figure 1(d), we used a shorter IHC chromogenic time (5 minutes), so the IHC results in the nonmetastatic group were lower and the staining was shallow.

3.2. High GSN Expression Is Associated with a Poor Prognosis in HCC Patients

We assessed whether GSN expression was associated with clinicopathological factors in patients with HCC. The 188 patients with HCC were divided into two groups according to the IHC results: the GSN high expression group (n = 95) and the GSN low expression group (n = 93). The results showed that high GSN expression in HCC was positively correlated with an advanced Edmondson grade (P = 0.022), capsulation (P = 0.006), and multiple tumour lesions (P = 0.012); however, high GSN expression in HCC was not significantly associated with age, gender, a family history of HCC, alpha-fetoprotein (AFP), cirrhosis, or tumour size (P > 0.05) (Table 1).

Furthermore, we explored the prognostic value of GSN expression. We found that the 3- and 5-year DFS rates (15.8% and 11.6% vs. 48.4% and 46.2%, P < 0.001) and OS rates (40.0% and 27.4% vs. 61.3% and 57.0%, P < 0.001) of HCC patients in the high GSN expression group were worse than those in the low GSN expression group (Figures 2(a) and 2(b)). Kaplan–Meier analysis showed that AFP, Edmondson grade, tumour size, envelope, tumour number, vascular invasion, GSN expression, and tumour-node-metastasis (TNM) stage were risk factors for DFS and OS (Table 2). From the multivariate Cox regression analysis, high GSN expression was found to be an independent prognostic factor, with low DFS (hazard risk (HR) = 1.744, 95% confidence interval [CI]: 1.205–2.523, P = 0.003) and OS (HR) = 1.620, 95% CI: 1.105–2.373, P < 0.001) (Table 3). These results indicate that high GSN expression is significantly associated with a poor prognosis, suggesting a potential role for GSN in liver tumourigenesis.

3.3. GSN Is Associated with HCC Migration and Invasion

To examine GSN expression in HCC cells, qRT-PCR and WB analysis were performed in five hepatoma cell lines, namely, SMMC7721, HepG2, HCCLM3, HepG3B, and Bel7402 and the liver cell line L02 as a normal control. The mRNA levels of GSN in HepG2 and HCCLM3 cells were higher than those in the L02 cell line, while the GSN levels in the other three cell lines were lower (Figure 3(a)). The protein expression levels of GSN were consistent with the mRNA levels (Figure 3(b)).

GSN is frequently reported to be critical for cell migration and invasion [8, 19, 21, 22]. Considering the correlation between GSN expression and metastasis, we performed an immunohistochemistry assay to investigate GSN as a potential modulator of cell motility in HCC. To determine how GSN affects the motility of HCC cells, we overexpressed GSN in SMMC7721 cells and knocked down GSN in HCCLM3 cells (Figure 3(c)). Then, we measured the migration and invasion abilities of these two new cell lines. Compared with vector-transfected cells, the relative number of migrating and invading SMMC7721 cells transfected with GSN was significantly higher (Figures 3(d) and 3(e)). In contrast, the knockdown of GSN resulted in deceased migration and invasion in HCCLM3 cells (Figures 3(f) and 3(g)). These results suggest that GSN functions as a positive modulator of HCC metastasis that promotes the migration and invasion of HCC cells.

3.4. GSN Promotes EMT in HCC Cells

Tumour metastasis, which involves the migration and invasion of tumour cells, is always associated with EMT. We next wanted to determine whether GSN affects HCC cell motility by modulating EMT and performed western blot analysis after GSN overexpression and GSN knockdown. In GSN-overexpressing SMMC7721 cells, we observed the downregulation of E-cadherin, an epithelial marker, and the upregulation of N-cadherin, a mesenchymal marker. Vimentin was also upregulated (Figure 4(a)). In contrast, E-cadherin was increased, and N-cadherin was decreased in GSN-knockdown HCCLM3 cells. Vimentin was downregulated (Figure 4(c)). Matrix metalloproteinase (MMP)2 and MMP9, which are known as zinc-dependent endopeptidases, are tightly associated with tumourigenesis and metastasis and are essential for mesenchymal cell invasion. We found that MMP2 and MMP9 mRNAs were significantly upregulated in SMMC7721 cells with GSN overexpression and downregulated in HCCLM3 cells with GSN knockdown (Figures 4(b) and 4(d)). These results show that GSN promotes EMT in HCC cells, thereby enhancing cell migration and invasion.

3.5. Knockdown of GSN Attenuates HCC Metastasis In Vivo

Furthermore, to explore the biological importance of GSN in HCC, we examined the growth of tumours in xenotransplantation experiments. Human tumour cells were injected into nude mice by tail vein injection, and the growth of tumours was monitored. After 8 weeks, the incident rate and the number of pulmonary metastatic nodules decreased significantly in the sh-GSN group compared with the sh-NC group. Representative histomorphological gross and HE staining images are shown in Figures 5(a) and 5(b). These results indicate that GSN knockdown effectively reduces the metastatic lung tumour burden. Targeting GSN could effectively decrease HCC metastasis in vivo.