SMMC-7721 cell line used in the experiment was from Institute of Biochemistry and Cell Biology (Shanghai, China); six-week-old female immune-deficient nude mice (BALB/c-nu) were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai Laboratory Animal Center of Chinese Academy Sciences). Adenovirus-mediated HMGB1 small hairpin RNA vector, negative control vector and virion-packaging elements were from Genechem (Shanghai, China). The primers of HMGB1, p-AKT, Ki-67 and MMP-2 were synthesized by ABI Co., Ltd. (USA). All antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

3-(4,5)-Dimethylthiahiazo(-z-yl)-3,5-di-phenytetrazolium bromide (MTT) was from Dingguo biology (Shanghai, China); Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were from Thermo Fisher Scientific Inc. (Waltham, MA, USA); TRIzol Reagent and Lipofectamine 2000 were from Invitrogen (Carlsbad, CA, USA); M-MLV Reverse Transcriptase was from Promega (Madison, WI, USA); SYBR Green Master Mix was from Takara (Otsu, Japan); Cell cycle analysis kit and apoptosis kit (Propidium Iodide (PI), RNase A, Annexin V-FITC) were from KeyGen Biology (Nanjing, China). ECL-PLUS/kit was from GE Healthcare (Piscataway, NJ, USA).

Forty freshly resected liver cancer and normal liver tissue samples were collected at the Department of General Surgery of Shanghai Rui Jin Hospital during 2010 and were classified according to American Joint Committee on Cancer (AJCC) TNM staging system. Tissues and clinical information were obtained as part of an approved study at Shanghai Jiao Tong University School of Medicine. There were 30 cases of liver cancer tissues and 10 cases of normal liver tissues. A portion of each tissue sample was stored in liquid nitrogen for RT-PCR and western blot examination. All tumors and normal tissues were diagnosed by two independent gastroenterologists.

To quantitatively determine the mRNA expression level of HMGB1, p-AKT, Ki-67 and MMP-2 in liver cancer, normal liver tissues and SMMC-7721 cells, RT-PCR was used. Total RNA of each clone was extracted with TRIzol according to the manufacturer’s protocol. Reverse-transcription was carried out using M-MLV and cDNA amplification was carried out using SYBR Green Master Mix kit according to the manufacturer’s protocol. The genes were amplified using specific oligonucleotide primer and human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was used as an endogenous control. The PCR primer sequences were: HMGB1, 5′-ATA TGGCAAAAGCGGACAAG-3′ and 5′-AGGCCAGGATGTT CTCCTTT-3′; p-AKT, 5′-GGAGAUCAUGCAGCAUCGC dtdt-3′ and 5′-GCGAUGCUGCAUGAUCUCCdtdt-3′; Ki-67, 5′-CTTTGGGTGCGACTTGACG-3′ and 5′-GTCGACCC CGCTCCTTTT-3′; MMP-2, 5′-GGCCCTGTCACTCCTGA GAT-3′ and 5′-GGCATCCAGGTTATCGGGGA-3′; GAPDH, 5′-CAACGAATTTGGCTACAGCA-3′ and 5′-AGGGGTCTA CATGGCAACTG-3′. Data were analyzed using the comparative Ct method (2^−ΔΔCt). Three separate experiments were performed for each clone.

Liver cancer, normal liver tissues and SMMC-7721 cells were harvested and extracted using lysis buffer (Tris-HCl, SDS, mercaptoethanol, glycerol). Cell extracts were boiled for 5 min in loading buffer and then equal amount of cell extracts were separated on 15% SDS-PAGE gels. Separated protein bands were transferred into polyvinylidene fluoride (PVDF) membranes and the membranes were blocked in 5% skim milk powder. The primary antibodies against HMGB1, p-AKT, Ki-67 and MMP-2 were diluted according to the instructions of antibodies and incubated overnight at 4°C. Then, horseradish peroxidase-linked secondary antibodies were added at a dilution ratio of 1:1000, and incubated at room temperature for 2 h. The membranes were washed with PBS for three times and the immunoreactive bands were visualized using ECL-PLUS/Kit according to the manufacturer’s instructions. The relative protein level in different cell lines was normalized to GAPDH concentration. Three separate experiments were performed for each clone.

SMMC-7721 cells were cultured in DMEM medium supplemented with 10% heat-inactivated FBS, 100 μg/ml of penicillin and 100 μg/ml of streptomycin. They were all placed in a humidified atmosphere containing 5% CO₂ at 37°C. Recombinant adenovirus vector rAd5-HMGB1 and negative control rAd5-GFP were transfected into SMMC-7721 cells. Cells were subcultured at a 1:5 dilution in 300 μg/ml G418-containing medium. Positive stable transfectants were selected and expanded for further study. The clone in which the rAd5-HMGB1 virus vectors transfected was named as rAd5-HMGB1 group, the negative control vectors transfected was named as GFP group and SMMC-7721 cells were named as CON group.

Cell proliferation was analyzed with the MTT assay. Briefly, cells infected with rAd5-HMGB1 were incubated in 96-well-plates at a density of 1×10⁵ cells per well with DEME medium supplemented with 10% FBS. Cells were treated with 20 μl MTT dye at 0, 24, 48, 72 h and then incubated with 150 μl of DMSO for 5 min. The color reaction was measured at 570 nm with enzyme immunoassay analyzer (Bio-Rad, Hercules, CA, USA). The proliferation activity was calculated for each clone.

Transwell filters were coated with matrigel (3.9 μg/μl, 60–80 μl) on the upper surface of a polycarbonic membrane (diameter 6.5 mm, pore size 8 μm). After incubating at 37°C for 30 min, the matrigel solidified and served as the extracellular matrix for analysis of tumor cell invasion. Harvested cells (1×10⁵) in 100 μl of serum-free DMEM were added into the upper compartment of the chamber. A total of 200 μl conditioned medium derived from NIH3T3 cells was used as a source of chemoattractant, and was placed in the bottom compartment of the chamber. After 24 h of incubation at 37°C with 5% CO₂, the medium was removed from the upper chamber. The non-invaded cells on the upper side of the chamber were scraped off with a cotton swab. The cells that had migrated from the matrigel into the pores of the inserted filter were fixed with 100% methanol, stained with hematoxylin, and mounted and dried at 80°C for 30 min. The number of cells invading through the matrigel was counted in three randomly selected visual fields from the central and peripheral portion of the filter using an inverted microscope (magnification, ×200). Each assay was repeated three times.

To detect cell apoptosis, cells infected with rAd5-HMGB1 were trypsinized, washed with cold PBS and resuspended in binding buffer according to the instruction of the apoptosis kit. FITC-Annexin V and PI were added to the fixed cells for 20 min in darkness at room temperature. Then, Annexin V binding buffer was added to the mixture before the fluorescence was measured on FACSort flow cytometer. The cell apoptosis was analyzed using the CellQuest software (Becton Dickinson, USA). Three separate experiments were performed for each clone.

To detect cell cycle variation, cells infected with rAd5-HMGB1 were trypsinized, washed by PBS and fixed with 80% cold ethanol overnight at −20°C. After PBS washing, the fixed cells were stained with PI in the presence of RNase A for 30 min at room temperature in darkness. Each sample was filtered through a 50 μm nylon filter to obtain single-cell suspension. The samples were then analyzed on FACsort flow cytometer (Becton Dickinson, Mountain View, CA, USA). ModFit3.0 software (Verity Software House, Topsham, ME, USA) was used for cell cycle analysis. Three separate experiments were performed for each clone.

Two mice were injected subcutaneously with 1×10⁸ SMMC-7721 cells in 50 μl of PBS pre-mixed with an equal volume of matrigel matrix (Becton Dickinson). Mice were monitored daily, and two mice developed subcutaneous tumors. When the tumor size reached ~5 mm in length, they were surgically removed, cut into 1–2 mm³ pieces, and re-seeded individually into 12 other mice. When tumor size reached ~5 mm in length, the mice were randomly assigned to control (CON), GFP and rAd5-HMGB1 groups. In GFP and rAd5-HMGB1 groups, 15 μl of adenovirus was injected into subcutaneous tumors using a multi-site injection format. Mice in CON group received 15 μl of PBS only. Injections were repeated on the third day after initial treatment. The tumor volume every three days was measured with a caliper, using the formula volume = (length × width)²/2.

The result of each experiment was shown as mean ± SD when applicable. Statistically significant difference in each assay was determined by SPSS version 11.5. Difference in each group was tested for significance using t-test and ANOVA analysis of variance. P<0.05 was considered significant.

The expression of HMGB1 in liver cancer was evaluated using RT-PCR and western blot assays. As shown in Fig. 1A and B, the mRNA expression level of HMGB1 was significantly increased in liver cancer tissues in comparison with normal liver tissues. Also, as shown in Fig. 1C, the protein expression level of HMGB1 was also significantly increased in liver cancer tissues compared with the normal liver tissues.

The relationship between HMGB1 mRNA expression and various clinical and pathologic features was analyzed. As shown in Table I, no significant correlation was found between HMGB1 expression with age, pathological type and classification, and serum AFP levels. However, the significant correlation was found between HMGB1 expression with TNM, pathological grade and distant metastasis of liver cancer.

In order to efficiently knockdown the expression of HMGB1 in liver cancer SMMC-7721 cells, an adenovirus-mediated RNAI approach was used to construct the rAd5-HMGB1 vector. In pilot studies, the transfection efficiency of rAd5-HMGB1 (MOI=100) in SMMC-7721 cells was greater than 85.0%. After rAd5-HMGB1 was transfected into SMMC-7721 cells, real-time PCR and western blot assays were performed at 48 h recovery to measure HMGB1 and p-AKT expression. As shown in Fig. 2A and B, an obvious inhibition of HMGB1 and p-AKT expression was observed in rAd5-HMGB1 group compared with the GFP group and CON group.

Deregulated cell proliferation is a hallmark of cancer (21). In order to test the effect of rAd5-HMGB1 on liver cancer cell proliferation, we investigated the proliferative activities of SMMC-7721 cells by MTT. As a result, it was found that knockdown of HMGB1 could significantly reduce the proliferative activities of SMMC-7721 cells in a time-dependent manner compared with GFP group and CON group (Fig. 3A). In addition, Ki-67 is at the very heart of many essential cellular processes and determines the tumor progression and the outcome of anticancer treatment. To determine whether knockdown of HMGB1 suppressed endogenous Ki-67 through translational repression, the expression of Ki-67 was examined by Real-time PCR and western blot assays. It was shown that the amount of Ki-67 expression was decreased in rAd5-HMGB1 group compared with GFP group and CON group (Fig. 3B and C), suggesting that knockdown of HMGB1 inhibits liver cancer cell proliferation through down-regulation of the Ki-67 expression.

To determine the effect of rAd5-HMGB1 on liver cancer cell metastasis, transwell assay was carried out. The results indicated that the invasive and metastatic potential was determined on the basis of the ability of cells to invade a matrix barrier containing laminin and type IV collagen, the major components of the basement membrane. Representative micrographs of Transwell filters can be seen in Fig. 4A. The invasive capability of SMMC-7721 cells was distinctly decreased in rAd5-HMGB1 group compared with GFP group and CON group (Fig. 4B). In terms of the important role of MMP-2 in tumor metastasis, Real-time PCR and western blot assays were performed to investigate the effect of rAd5-HMGB1 on expression of MMP-2. As shown in Fig. 4C and D, the expression of MMP-2 was significantly reduced in rAd5-HMGB1 group compared with GFP group and CON group, indicating that knockdown of HMGB1 inhibits liver cancer cell metastasis through down-regulation of the MMP-2 expression.

To determine whether knockdown of HMGB1 affected SMMC-7721 cell apoptosis and cycle distribution, flow cytometry with PI/FITC-Annexin V staining was performed. The results showed that the apoptotic index of SMMC-7721 cells in rAd5-HMGB1 group was markedly higher than the GFP and CON groups (Fig. 5A and C). The cycle distribution of SMMC-7721 cells was also analyzed and cell cycle kinetics showed that the G₀/G₁ phase fraction was increased, S-phase fraction was decreased and cell cycle was arrested in G₀/G₁ phase in rAd5-HMGB1 group compared with GFP and CON groups (Fig. 5B and D). Therefore, knockdown of HMGB1 can induce liver cancer cell apoptosis and block cell cycle progression.

Our in vitro experiments demonstrated that knockdown of HMGB1 efficiently inhibited the growth and metastasis of SMMC-7721 cells. Therefore, it is necessary to further investigate the effect of knockdown of HMGB1 on xenograft tumor growth in vivo. The mean volume of tumors in all experimental mice before treatment was 67.01±14.38 mm³. During the whole tumor growth period, the tumor growth activity was measured. Tumors treated with rAd5-HMGB1 grew relatively slowly compared with the CON and GFP groups (Fig. 6A and B). When the tumors were harvested, the average weight of tumors in group rAd5-HMGB1 was much less than that of the CON and GFP groups, respectively (Fig. 6C). This result in vivo indicated that knockdown of HMGB1 inhibited liver cancer cell growth.