A total of 56 HCC intratumor specimens and 56 paired peritumor specimens (as controls; obtained at a distance of >10 mm from the tumor edge) were included in the present study. All specimens were obtained from the pathological archives of Baotou Cancer Hospital (Batou, China) and had been obtained between May 2009 and June 2014 with informed consent of the patients. The HCC specimens had been obtained by surgical resection, immediately frozen in liquid nitrogen and stored at −80°C prior to radiotherapy or chemotherapy. Clinico-pathological characteristics of each patient are listed in Table I. The present study was approved by the Medical Ethics Committee of Baotou Cancer Hospital (Batou, China).

The MHCC97L, HepG2 and MHCC97H human HCC cell lines and the L02 control liver cell line were purchased from the cell resource center of the Chinese Academy of Medical Sciences (Beijing, China). Each cell line was cultured in Dulbecco's modified Eagle's medium (DMEM; Ameresco, Inc., Framingham, MA, USA) with 10% fetal bovine serum (FBS; Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) at 37°C in a humidified atmosphere containing 5% CO₂.

The open reading frame (ORF) of human XAF-1 (NM_017523) and TNF-α (NM_000594) was amplified by polymerase chain reaction (PCR) with primers that deleted the stop codon, and was overlapped with a sequence encoding a 2A peptide linker (24). The overlapped XAF-1 - 2A - TNF-α nucleotide was inserted into the pShuttle-cytomegalovirus (CMV) vector (Qbiogene, Inc., Irvine, CA, USA) to generate the recombinant pShuttle-CMV - XAF-1 - 2A - TNF-α. The adenovirus Ad-XAF-1&TNF-α and the Ad-control (Ad-con) virus were enveloped via co-transfecting the pShuttle-CMV - XAF-1 - 2A - TNF-α and the pAdeasy-1 (the viral DNA plasmid) into 293GPG retrovirus packaging cell line (Cell Resource Center of the Chinese Academy of Medical Sciences) using Lipofectamine™ 2000 (Invitrogen; Thermo Fisher Scientific, Inc.). To co-express XAF-1 and TNF-α in HCC cells, MHCC97L cells were infected with Ad-XAF-1&TNF-α at a multiplicity of infection (MOI) of 1 or 10 for 2 h, followed by culture in fresh DMEM containing 2% FBS.

Cellular mRNA was isolated from tissues or cell lines using TRIzol (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's manual, following homogenization of tissues. RT-qPCR was performed using the One Step SYBR^® Green RT-qPCR kit (Sigma-Aldrich, St. Louis, MO, USA) following the manufacturer's instructions. The PCR reaction conditions were as follows: Initial denaturation, 5 min at 95°C; 40 cycles of denaturation for 20 sec at 94°C, annealing for 20 sec at 61°C and extension for 20 sec at 72°C; and a final extension for 5 min at 72°C. The primers for XAF-1, TNF-α and β-actin were synthesized by Invitrogen; Thermo Fisher Scientific, Inc., and were as follows: Forward, 5′-CCCAGGGACCTCTCTCTAATC-3′ and reverse, 5′-ATGGGCTACAGGCTTGTCACT-3′ for TNF-α; forward, 5′-AGAATTCCCCATTCAGTAAG-3′ and reverse, 5′-GTGTAAGGAAGTGGTTCAGT-3′ for XAF-1; and forward, 5′-CATTAAGGAGAAGCTGTGCT-3′ and reverse, 5′-GTTGAAGGTAGTTTCGTGGA-3′ for β-actin. RT-qPCR was performed in an ABI PRISM 7000 (Applied Biosystems; Thermo Fisher Scientific, Inc.). Expression levels were normalized to the internal control β-actin, expressed as the fold change compared with the control and calculated using the ∆∆Ct method (25), subsequent to confirm the target PCR product with melting curve analysis.

Intratumor or peritumor specimens from HCC patients were homogenized prior to protein extraction. Lysis was then performed with a Cell Lysis and Protein Extraction kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions, followed by addition of protease inhibitor cocktail (Sigma-Aldrich). Proteins were quanitified using the BCA Protein assay reagent kit (Thermo Fisher Scientific, Inc.) and 25 μg of each sample was separated by 10% sodium dodecyl sulfate polyacrylamide gel (Thermo Fisher Scientific, Inc.) electrophoresis and then transferred onto a nitrocellulose membrane (EMD Millipore, Billerica, MA, USA). Non-specific binding was blocked with 2% bovine serum albumin (Ameresco, Inc.) overnight at 4°C, and membranes were subsequently probed with rabbit polyclonal antibody to XAF-1 (Abcam, Cambridge, MA, USA; cat. no. ab81353; 1:500 dilution) β-actin (Abcam; cat. no. ab8227; 1:200 dilution) or TNF-α (Cell Signaling Technology Inc., Danvers, MA, USA; cat. no. 3727; 1:500 dilution) at 4°C overnight. The membrane was finally incubated with goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (Promega Corp., Madison, WI, USA; cat. no. W4011) at 4°C for 2 h, and antibodies were visualized using an enhanced chemiluminescence detection system (GE Healthcare, Little Chalfont, UK) following the manufacturer's instructions. The images of the blots were captured on a UVP BioSpectrum 500 imaging system (UVP, LLC, Upland, CA, USA) and the bands were analyzed using Image J (imagej.nih.gov/ij/). The protein levels of XAF-1 or TNF-α were expressed as a percentage to β-actin.

The apoptosis of MHCC97L cells with or without infection (1 MOI, 24 h) with Ad-Con, Ad-XAF-1, Ad-TNF-α or Ad-XAF-1&TNF-α was examined with an ApoDETECT Annexin V-FITC kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Briefly, MHCC97L cells either without or post-infection were incubated at 37°C for 24 h, and then were harvested and suspended in binding buffer (5×10⁵ cells/ml). The suspended cells were mixed with 5 μl Annexin V-FITC and 10 μl of PI and incubated for 15 min in the dark at room temperature. The stained cells were detected using a FACScan flow cytometer (BD Biosciences, Franklin Lakes, NJ, USA). The results were calculated using the CellQuest™ Pro software (BD Biosciences) and were presented as the percentage of apoptotic cells to total cells.

The proliferation of HCC cells was evaluated using a cell counting assay and a colony formation assay. The cell counting assay was performed as follows: Cells (10³/ml) were seeded into 12-well plates and then infected with Ad-XAF-1&TNF-α or Ad-con virus at an MOI of 1 or 10 for 2 h, followed by further incubation in medium for 1, 3 or 5 days. The cells were trypsinized and the number of viable cells was counted using a hemocytometer (Reichert, Inc., Depew, NY, USA) following trypan blue (Thermo Fisher Scientific, Inc.) staining. For the colony formation assay, 1,000 cells were seeded into a 12-well plate and infected with Ad-XAF, Ad-TNF-α, Ad-XAF-1&TNF-α or Ad-con virus at an MOI of 1 or 10 for 2 h, followed by incubation in medium for another five days. The cell colonies were stained with 0.5% crystal violet (Sigma-Aldrich) in methanol for 10 min and colonies were counted on the plate by the naked eye.

Values are expressed as the mean ± standard error of the mean. Differences between two groups were evaluated using Student's unpaired t-test for the cell viability assay, and the paired-samples t-test was used for comparison of expression levels in the tumor and peritumor tissues or among the cell lines. Statistical analysis was performed using GraphPad Prism software (version 5; GraphPad Inc., La Jolla, CA, USA) and P<0.05 was considered to indicate a statistically significant difference between values.

To confirm the tumor suppressive role of XAF-1 in HCC, its expression was determined in 56 HCC specimens and paired peritumor tissues. As shown in Fig. 1A, compared to the levels in the peritumor tissues, the relative XAF-1 mRNA levels (to β-actin) in the HCC specimens were significantly reduced by ~40% (P<0.01). Furthermore, western blot analysis confirmed a ~50% reduction of XAF-1 protein expression in the HCC group compared with that in the peritumor samples (P<0.001) (Fig. 1B). Furthermore, the expression of XAF-1 in the HCC cell lines MHCC97L, HepG2 and MHCC97H was significantly reduced at the mRNA and protein level compared to that in the L02 human hepatic cell line (P<0.05 or P<0.01) (Fig. 1C and D). Thus, the significant downregulation of XAF-1 in HCC specimens and cell lines was confirmed.

To assess the association of the reduced XAF-1 with the malignant characteristics of HCC, the correlation of XAF-1 expression with clinico-pathological features, including tumor size, Barcelona clinic liver cancer (BCLC) stage, tumor -nodes -metastasis (TNM) stage and vascular invasion, was assessed. As shown in Table I, there was no significant difference in age, hepatitis B surface antigen positivity or alpha-fetoprotein levels between the groups with XAF-1 levels <1 and XAF-1 levels ≥1. However, XAF-1 expression was negatively associated with the tumor size, BCLC stage, TMN stage (P<0.05, respectively). However, the association of XAF-1 mRNA levels with vascular invasion was not significant (P>0.05). In conclusion, these results confirmed the association of reduced XAF-1 mRNA levels with the degree of malignancy of HCC.

To further identify the suppressive role of XAF-1 in HCC, an adenovirus co-expressing XAF-1 and TNF-α was constructed. XAF-1 and TNF-α cDNA were amplified by PCR and then linked with a 2A peptide coding sequence (24). The construction strategy of the recombinant adenovirus Ad-XAF-1&TNF-α was illustrated in Fig. 2A. The adenovirus expressing green fluorescence protein (Ad-con), XAF-1 (Ad-XAF-1) or TNF-α (Ad-TNF-α) were used as controls. Each recombinant adenovirus was enveloped via co-transfection of the respective adenoviral genomic plasmid and the shuttle plasmid into BJ5183 bacterial cells. The efficiency of the adenovirus to co-express XAF-1 and TNF-α was evaluated in MHCC97L cells at an MOI of 1 or 10. At 24 h post-infection, the mRNA levels of the two genes were significantly enhanced (P<0.01 or P<0.0001, respectively) (Fig. 2B). Furthermore, western blot analysis indicated that the protein levels of XAF-1 and TNF-α were significantly enhanced by the adenovirus (P<0.01, P<0.001 or P<0.0001) (Fig. 2C and D).

In addition, to compare the effects of Ad-XAF-1&TNF-α with those of XAF-1 or TNF-α alone, MHCC97L cells were infected with Ad-XAF-1 or Ad-TNF-α. As expected, XAF-1 was only overexpressed following infection with Ad-XAF-1, while TNF-α was only overexpressed following infection with Ad-TNF-α at the mRNA and protein level (P<0.001), while Ad-con had no effect (Fig. 3). While XAF-1 as well as TNF-α were significantly overexpressed following infection with Ad-XAF-1&TNF-α (P<0.001), their expression levels were significantly lower than those following infection with the respective mono-overexpression vectors.

In addition, apoptosis in the MHCC97L cells that were infected with 1 MOI Ad-Con, Ad-XAF-1, Ad-TNF-α or Ad-XAF-1&TNF-α was examined. MHCC97L cells without infection served as a blank control. As indicated in Fig. 3D and E), compared with the Ad-Con, Ad-XAF-1 or Ad-TNF-α induced a significantly increased level of apoptosis in MHCC97L cells (P<0.05 or P<0.01). Furthermore, the Ad-XAF-1&TNF-α infection induced more apoptotic cells than the infection with Ad-XAF-1 or Ad-TNF-α (P<0.01). Thus, the co-expression of XAF-1 and TNF-α synergistically induced apoptosis in MHCC97L cells.

The present study then investigated the effects of XAF-1 and TNF-α co-expression on the growth of HCC cells. The growth of MHCC97L cells was assessed in vitro using a cell counting assay and a colony formation assay. It was revealed that following infection with Ad-XAF-1&TNF-α, the proliferation of MHCC97L cells was reduced compared with that of the cells infected with Ad-con, Ad-XAF-1 or Ad-TNF-α at either 3 days (P<0.05 or P<0.001) or 5 days (P<0.05, P<0.01 or P<0.001) post-infection, while infection with Ad-XAF-1 or Ad-TNF-α also significantly inhibited the proliferation of MHCC97L cells (P<0.05 or P<0.01) (Fig. 4A and B). Similarly, the colony formation assay showed that the clonogenicity of MHCC97L cells following infection with Ad-XAF-1&TNF-α was significantly reduced compared with that following infection with Ad-XAF-1 or Ad-TNF-α (P<0.05; Fig. 4C and D), while mono-infection still significantly reduced the number of colonies formed (P<0.05 or P<0.01; Fig. 4C and D). These results indicated that co-expression of XAF-1 and TNF-α inhibited the growth of HCC MHCC97L cells more efficiently than either protein alone, even though their co-expression was lower than that following infection with Ad-XAF-1 or Ad-TNF-α.