The human HCC cell lines MHCC97-Hand MHCC97-L were purchased from the Liver Cancer Institute of Fudan University (Shanghai, China). Cells were cultured in complete Dulbecco's modified Eagle's medium (DMEM; HyClone; GE Healthcare Life Sciences, Logan, UT, USA) containing 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) and 100 U/ml penicillin (Invitrogen; Thermo Fisher Scientific, Inc.), and incubated at 37°C in a humidified atmosphere containing 5% carbon dioxide.

Membrane proteins were extracted using a CelLytic MEM Protein Extraction kit (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany), following the manufacturer's protocol. A total between 1×10⁶ and 1×10⁷ cells were collected and resuspended in 600 µl ice-cold Lysis and Separation Working Solution, provided in the kit, containing 6 µl protease inhibitor cocktail. Cell suspensions were incubated on ice for 10 min and subsequently centrifuged in a pre-cooled (4°C) microcentrifuge at 10,000 × g for 5 min. The clarified lysates were transferred to a sterile microcentrifuge tube, incubated at 30°C for 3–5 min and centrifuged at room temperature (>20°C) at 3,000 × g for 3 min. The lower, hydrophobic phase, which is enriched with hydrophobic and lipid raft-associated proteins, was transferred to a new microcentrifuge tube and stored at −20°C until use.

The 8-plex iTRAQ kits were obtained from Applied Biosystems (Thermo Fisher Scientific, Inc.), and experiments were conducted according to the manufacturer's protocol. Briefly, protein concentrations were quantified by Bicinchoninic Acid assay (BCA; Qcbio Science & Technologies Co., Ltd., Shanghai, China) and ~100 µg of prepared proteins from each group were precipitated, dissolved in dissolution buffer, denatured, cysteine blocked, digested with 2 µg of sequencing grade modified trypsin, and labeled with the iTRAQ tags; MHCC97-L proteins were labeled with the iTRAQ-113 tag and MHCC97-H proteins were labeled with the iTRAQ-115 tags. A schematic view of the iTRAQ methodology is shown in Fig. 1; the experiments were repeated twice.

iTRAQ-labeled peptides were fractionated by immobilized pH gradient-isoelectric focusing (IPG-IEF). Briefly, the pooled iTRAQ-labeled peptide samples were solubilized in 300 µl of 8 M urea and 1% Pharmalyte solution, and rehydrated on an 18 cm-long IPG gel strips (pH 3–10) (both from Amersham; GE Healthcare Life Sciences, Little Chalfont, UK) for 14 h at 30 V. The peptides were subjected to IEF at 68 kV/h with an IPGphor IEF System (Amersham; GE Healthcare Life Sciences). Peptides were extracted by incubating the gel fractions in 2% acetonitrile and 0.1% formic acid for 1 h at room temperature. Fractions were purified and concentrated on a C-18 Discovery DSC-18 SPE column (Sigma-Aldrich; Merck KGaA). The desalted fractions were lyophilized and stored at −20°C until used in liquid chromatograph-tandem mass spectrometry (LC-MS/MS) analysis.

Mass spectrometric analysis was performed on a QStar Elite mass spectrometer (Applied Biosystems; Thermo Fisher Scientific, Inc.) coupled to an online capillary LC Dionex Ultimate 3000 system (Thermo Fisher Scientific, Inc.), as previously described (20,21). The peptide separation was performed on a PepMap C-18RP capillary column (Thermo Fisher Scientific, Inc.) with a flow rate of 0.3 µl/min using a linear gradient ranging between 2 and 100% Buffer B (0.1% formic acid; 98% acetonitrile) for 90 min, 2.1 kV, and flow rate of 300 nl/min. Data acquisition was performed in the positive ionization mode with a selected mass range of 300–1,800 m/z. The two most intensely charged peptides with >20 counts were chosen for MS/MS with a dynamic exclusion of 30 sec (22).

Protein identification and quantification were performed using ProteinPilot software v2.0 (Applied Biosystems/MDS Sciex Joint Venture; Thermo Fisher Scientific, Inc.; Nordion, Inc., Ottawa, ON, Canada). MS/MS data were analyzed with the UniProt database (http://www.uniprot.org). Differentially expressed proteins were indicated as having a protein score >1.3 (1.3/1) or <0.77 (1/1.3) (Table I), and >2 unique peptides with 95% confidence and P<0.05 were required for quantitation.

Total RNA was isolated from 1×10⁵–1×10⁷ cells using TRIzol (Invitrogen; Thermo Fisher Scientific, Inc.), according to the manufacturer's instructions. First strand cDNA was synthesized using 2 µg total RNA and the A3500 Reverse Transcription System (Promega, Madison, WI, USA). The resulting cDNA was amplified by PCR using an ABI 7300 system (Applied Biosystems; Thermo Fisher Scientific, Inc.), SYBR Premix Ex Taq II (Tli RNaseH Plus) and ROX plus reference dye (Takara Bio, Inc., Otsu, Japan). Gene-specific primer sequences are provided in Table II. PCR thermocycling conditions were as follows: 95°C for 15 sec followed by 35 cycles of 72°C for 30 sec and 60°C for 60 sec. PCR reactions were evaluated by melting curve analysis, and relative gene expression levels were calculated according to the 2^−ΔΔCq method and normalized to the internal reference gene GAPDH (23). Experiments were performed at least three times and similar results were obtained in each experiment.

Cells were lysed and fractionated using a CelLytic MEM Protein Extraction Kit (Sigma-Aldrich; Merck KGaA), as aforementioned. Total protein was quantified by BCA, and equal amounts of proteins (30 µg protein/lane) were separated by 10% SDS-PAGE and transferred to a polyvinylidene fluoride membrane (EMD Millipore, Billerica, MA, USA). Membranes were blocked for 1 h at room temperature in 5% skim milk powder and TBS + 0.5% Tween-20 buffer (pH 7.6) and incubated with primary antibodies at 4°C overnight. Primary antibodies against PAFAH1B2 (1:1,000; cat. no. ab157479), DPP2 (1:1,000; cat. no. ab226935), UBE3C (1:1,000; cat. no. ab226173), PLAA (1:1,000; cat. no. ab32354), ILF2 (1:1,000; cat. no. ab154169), EXOC1 (1:1,000; cat. no. ab118798), ACSL5 (1:500; cat. no. ab104892), GNAS (1:1,000; cat. no. ab83735), TGOLN2 (1:500; cat. no. ab2809), HAUS7 (1:500; cat. no. ab192616), EPCAM (1:1,000; cat. no. ab71916), AHNAK (1:1,000; cat. no. ab168104), NCSTN (1:500; cat. no. ab45425), SLC3A2 (1:1,000; cat. no. ab108300), SLC7A5 (1:500; cat. no. ab85226), CD9 (1:1,000; cat. no. ab92726), MUC13 (1:1,000; cat. no. ab124654), MISP (1:200; cat. no. ab175101), APOL2 (1:500; cat. no. ab196771), CD63 (1:1,000; cat. no. ab59479) and ATPase Na⁺/K⁺ transporting subunitβ2 (ATP1B2; 1:1,000; cat. no. ab176163) were purchased from Abcam (Cambridge, MA, USA). ITGB1 (1:1,000; cat. no. 4706) was purchased from Cell Signaling Technology, Inc. (Danvers, MA, USA) and CYB5B (1:250; cat. no. PA5-52482) was from Invitrogen (Thermo Fisher Scientific, Inc.). Subsequently, membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit (1:10,000; cat. no. ab205718) or goat anti-mouse (1:10,000; cat. no. ab6789) secondary antibodies (Abcam) for 1 h at room temperature, and protein bands were visualized with the Immobilon Western Chemiluminescent HRP substrate (Merck KGaA) on a ChemiDoc MP imaging system (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Target protein expression levels were analyzed by densitometric analysis with Image Lab v4.0 (Bio-Rad Laboratories, Inc.) and normalized to ATP1B2. Experiments were repeated at least three times.

CD9-targeted siRNA oligomer duplexes siRNA1 (cat. no. 10404) and siRNA2 (cat. no. 145966) and a non-specific, scrambled-siRNA was used as negative control (cat. no. 4390843) were purchased from Invitrogen (Thermo Fisher Scientific, Inc.). MHCC97-H cells (6×10⁵ cells/well) were seeded in 6-well plates 24 h at 37°C prior to transfection. Cells were transfected for 6 h at 5% CO₂ and 37°C with scrambled-siRNA, CD9-siRNA1 or CD9-siRNA2 using Lipofectamine^® 2000 Transfection Reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and Opti-MEM (Gibco; Thermo Fisher Scientific, Inc.); untreated MHCC97-H cells were used as a blank control. Subsequently, the transfection culture medium was removed and cells were further incubated with complete culture solution for 48 hat 5% CO₂ and 37°C. Transfected cells were harvested and the expression levels of CD9 mRNA and protein in each sample were confirmed by RT-qPCR and western blotting, respectively. All experiments were repeated at least three times.

The wound-healing assay was conducted following siRNA transfections to determine the effects on the migratory ability of cells. MHCC97-H cells were seeded in 6-well plates until they reached 100% confluence at 48 h post-transfection. A sterile P20 pipette tip was used to scratch a wound into the cell monolayer and the debris was removed by washing thrice with PBS. The cells were grown in serum-free medium at 37°C, and digital images of gap closure were captured under a Nikon TS100/TS100-F inverted light microscope (Nikon Corporation, Tokyo, Japan) at 0, 24 and 48 h. Experiments were repeated at least three times.

Invasion assays were conducted using a Cell Invasion Assay kit (Cell Biolabs, Inc., San Diego, CA, USA). Viable CD9-siRNA1/2 or scrambled-siRNA transfected MHCC97-H cells (1×10⁵ cells) were harvested and re-suspended in serum-free medium, and used to seed the upper chamber of a 24-well plate; membranes were precoated with Matrigel. A total of 500 µl DMEM + 10% FBS was loaded into the lower chamber. Cells were incubated for 24 h at 37°C. Subsequently, cells were washed with PBS and the cells in the upper chamber were wiped off with cotton swab. Cells in the lower chamber were fixed with 4% formaldehyde at room temperature for 30 min and stained with 0.1% crystal violet (Solarbio, Beijing, China) for 20 min at room temperature. Following three washes with PBS, the number of cells that invaded through the Matrigel was measured by Image-Pro Plus 6.0. At least three fields of views were randomly selected to count cells under an inverted light microscope (magnification, ×200). Experiments were repeated at least three times.

Statistical analyses were performed using SPSS software v16.0 (SPSS Inc., Chicago, IL, USA). Quantitative variables were presented as the mean ± standard deviation. Differences between two groups were analyzed by Student's t-test; multiple comparison between the groups were performed by analysis of variance followed by Student-Newman-Keuls post hoc test. P<0.05 was considered to indicate a statistically significant difference.

To investigate the mechanisms involved in HCC metastasis, iTRAQ-based MS was conducted to analyze the proteins that were isolated from MHCC97-H and MHCC97-L cell membranes. The ratio of 115-tagged protein expression to 113-tagged protein expressions in MHCC97-H and MHCC97-L samples was used to determine differentially expressed proteins. A total of 780 unique proteins were identified and quantified in each cell line, regardless of whether the P-value was <0.05 in the iTRAQ ratios. It has been reported previously that membrane proteins are closely related to malignant tumor cell invasion and metastasis, thus the membrane proteins were classified separately in 780 differentially expressed proteins: ~30 membrane proteins as being differentially expressed between MHCC97-L and MHCC97-H cells. Based on the 'Unused' column in Table I (that is, the score of protein confidence level), ~20 membrane proteins were verified by RT-qPCR and western blotting, which demonstrated similar mRNA and protein expression levels (Fig. 2). Therefore, 22 proteins were analyzed in this experiment. Of these 22 differentially expressed membrane proteins, 16 were upregulated and 6 were downregulated in the MHCC97-H cell line compared with MHCC97-L cells (Table I).

To verify the iTRAQ results, RT-qPCR and western blot analyses were performed. RT-qPCR analysis demonstrated that, compared with expression levels in MHCC97-L cells, the mRNA levels of PAFAH1B2, DPP2, UBE3C, PLAA, IFL2 and EXOC1 were significantly downregulated in MHCC97-H cells (Fig. 2A), whereas ACSL5, ITGB1, GNAS, CYB5B, TGOLN2, HAUS7, EPCAM, AHNAK, NCSTN, SLC3A2, SLC7A5, CD9, MUC13, MISP, APOL2 and CD63 mRNA expression levels were upregulated (Fig. 2B). Western blot analyses were performed to determine the expression levels of these membrane proteins, which revealed a similar pattern of down- and upregulated protein expression levels, as was identified for mRNA in the MHCC97-H cells compared with the MHCC97-L cell (Fig. 2C). The results of mRNA and protein expression were consistent with the iTRAQ results.

The tetraspanin membrane protein CD9 was demonstrated to be highly expressed in MHCC97-H cells (Fig. 2). CD9 had the lowest score in the 'Unused' column among 22 membrane proteins listed in Table I. The score indicated that CD9 was less accurate compared with other proteins and, therefore, was used to verify the accuracy of iTRAQ. CD9-targeted siRNA1 and siRNA2 were transfected into MHCC97-H cells to silence the expression of CD9. Subsequent western blot analysis demonstrated that CD9 protein expression was reduced in the CD9-siRNA-treated cells compared with the untreated control and scrambled-siRNA-treated groups (Fig. 3A and B). A significant decrease of CD9 mRNA expression levels was also observed intheCD9-siRNA-treated MHCC97-H groups compared with the untreated control and scrambled-siRNA-treated groups (Fig. 3C).

To further examine the role of CD9 in HCC cell motility, wound-healing and invasion assays were performed on untransfected MHCC97-H cells as well as MHCC97-H cells that were transfected with CD9-siRNA1, CD9-siRNA2 or scrambled-siRNA. Results from the wound-healing assay demonstrated that CD9 silencing significantly decreased the migratory ability of MHCC97-H cells transfected with CD9-siRNA1 and CD9-siRNA2 compared with the control groups (P<0.05) (Fig. 4). Similarly, the CD9-siRNA transfections resulted in significantly decreased invasive ability compared with the control groups (P<0.05) (Fig. 5).