1,1-Dimethylbiguanide hydrochloride (metformin, #D150959-5G), chloroquine (CQ, #C6628) and 3-methylad-enine (3MA, #M9281) was purchased from Sigma-Aldrich (St. Louis, MO, USA); sorafenib (C21H16ClF3N4O3C7H8O3S, CAS 475207-59-1) was purchased from Selleck Chemicals (Houston, TX, USA); and the cell counting kit-8 (CCK-8, KGA317), the Annexin V-FITC apoptosis detection kit (KGA108), the terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate (dUTP) nick-end labeling (TUNEL) assay kit (KGA707), the cell cycle detection kit (KGA512) and the ROS detection kit (KGT010) were purchased from KeyGen Biotech (Nanjing, China). The Histostain™-Plus kits (IgG/Bio, S-A/HRP, DAB) were purchased from Zhongshan Golden Bridge Co., Ltd. (Beijing, China).

The HCC cell lines Bel-7402 and HepG2 were purchased from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). The cell lines were cultured in RPMI-1640 supplemented with 10% fetal bovine serum (FBS) (both from Gibco, Carlsbad, CA, USA) and 100 µg/ml each of penicillin and streptomycin (Invitrogen, Carlsbad, CA, USA) in 5% CO₂ at 37°C.

The following antibodies were used in western blot analysis: β-actin (sc-47778, diluted 1:1,000) was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). AKT (ab126811, diluted 1:1,000) and phosphorylated AKT (phospho-Ser473, ab66138, diluted 1:1,000) were from Abcam (Cambridge, MA, USA). Caspase-3 (AP7563C, diluted 1:1,000), cleaved caspase-3 (AJ1131b, diluted 1:1,000), CDK4 (AP7520b, diluted 1:1,000) and cyclin D1 (AP2612c, diluted 1:1,000) were from Abgent (San Diego, CA, USA). AMPKα (#2532, diluted 1:1,000) and phosphorylated AMPKα (phospho-Thr172, #2535, diluted 1:1,000), mTOR (#2983, diluted 1:1,000), phosphorylated mTOR (phospho-Ser2448, #5536, diluted 1:1,000), phosphorylated mTOR (phospho-Ser2481, #2974, diluted 1:1,000), phosphorylated Raptor (phospho-Ser792, #2083, diluted 1:1,000), phosphorylated p70 S6 kinase (phospho-Thr389, #9234, diluted 1:1,000), phosphorylated 4E-BP1 (phospho-Thr37/46, #2855, diluted 1:1,000), PARP (#9532, diluted 1:1,000), cleaved PARP (#5626, diluted 1:1,000), ERK (#4696, diluted 1:2,000), phosphorylated ERK (phospho-Thr202/Tyr204, #4370, diluted 1:2,000), Ki-67 (#9449, diluted 1:1,000), LC3-I/II (#12741, diluted 1:1,000) and p62 (#8025, diluted 1:1,000) were from Cell Signaling Technology, Inc. (Danvers, MA, USA). Goat anti-rabbit and goat anti-mouse IgG peroxidase-conjugated secondary antibodies (31460 and 31430, both diluted 1:10,000) were from Thermo-Pierce (Rockford, IL, USA). The following antibodies used in the IHC assays were as described above: phosphorylated mTOR (phospho-Ser2448, diluted 1:200), phosphorylated mTOR (phospho-Ser2481, diluted 1:200), phosphorylated AKT (phospho-Ser473, diluted 1:200), phosphorylated ERK (phospho-Thr202/Tyr204, diluted 1:200), cleaved caspase-3 (diluted 1:200), and Ki-67 (diluted 1:200).

Cell viability was determined using the CCK-8 assay according to the manufacturer's instructions. Cells (5×10³) were seeded into a well of a 96-well plate and cultured in 100 µl of RPMI-1640 supplemented with 10% FBS, 100 µg/ml penicillin and 100 µg/ml streptomycin. After 24 h, the agents (10 mmol/l metformin and/or 5 µmol/l sorafenib) was added to the culture medium. After the cells were incubated at 37°C for different times (24, 48 or 72 h), the medium was changed for 100 µl of RPMI-1640 and 10 µl of CCK-8 reagent. The cells were incubated for 2 h at 37°C. Finally, the optical density was measured using an EnSpire™ 2300 Multilabel Reader (Perkin-Elmer, Waltham, MA, USA) at 450 nm. Five replicates were prepared for each condition. The mean values were calculated and growth curves were drawn.

The inhibitory effect of metformin and sorafenib on HCC cell proliferation was also determined by clonogenic assay. Logarithmic-phase HCC cells were trypsinized and collected, and the resuspended cells were seeded into 6-well plates in triplicates at a density of 500 cells/well in 2 ml of medium containing 10% FBS. After a 24 h incubation, the cultures were replaced with fresh medium containing 2% FBS and 5 mmol/l metformin, 2.5 µmol/l sorafenib or their combination and grown for 10 days. The cell clones were stained with a solution containing 1% crystal violet and 25% methanol for 2 min. The excess dye was removed by gently rinsing with tap water for 15 min. The average surface of the clones was determined by Image Pro Plus 6.0 software.

Cells after different treatments or tumor tissues from xenografts were harvested and lysed in RIPA buffer (KGP702) supplemented with 1 mM phenylmethylsulfonyl fluoride (PMSF; KGP610) and 1 mM phosphatase inhibitor cocktail (KGP602; all from KeyGen Biotech). The mixture was centrifuged at 12,000 × g for 20 min and the supernatant was collected. The protein concentration was determined using the BCA assay kit (KGPBCA) and each sample contained 30 µg protein per 10 µl. The protein samples were mixed with loading buffer (KGP101) and the proteins were separated using 6, 8 or 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by transfer to polyvinylidenedifluoride (PVDF) membranes (Bio-Rad Laboratories, Hercules, CA, USA). After soaking in blocking buffer for 2 h, the blots were incubated with the primary antibody with gentle agitation overnight at 4°C and were subsequently incubated at 37°C for 1 h with the HRP-conjugated secondary antibody. The bands were visualized by chemiluminescence, imaged using a ChemiDoc XRS and analyzed using Image Lab (both from Bio-Rad).

To evaluate the effects of metformin on cell cycle arrest, induction of apoptosis and intracellular ROS, the cells were examined using the cell cycle detection kit, the Annexin V-FITC apoptosis detection kit and the ROS detection kit, respectively, according to the manufacturer's protocols. Bel-7402 and HepG2 cells were seeded into 6-well plates (5×10⁴ and 1×10⁵ cells/dish for analysis of cell cycle arrest and apoptosis, respectively) and incubated for 24 h. For cell cycle analysis, after treatment with agents for 48 h, a total of 1×10⁶ cells was pelleted by centrifugation and washed twice with PBS. Then, the cell pellets were resuspended in 500 µl of ice-cold 70% ethanol and incubated at 4°C overnight. The fixed cells were centrifuged and the pellets were washed with PBS. After incubation with 100 µl RNase A (10 µg/ml) for 30 min at 37°C in the dark, the cells were resuspended in 400 µl propidium iodide (PI) (50 µg/ml) and placed at 4°C in the dark for 30 min. The stained cells were analyzed using an Accuri C6 flow cytometer (Accuri Cytometers Inc., Ann Arbor, MI, USA).

For the apoptosis analysis, the cells were trypsinized, washed with cold PBS and suspended in PBS. Then, the cells were stained using the Annexin V-FITC reaction reagent (5 µl of Annexin V-FITC, 5 µl of PI) at 37°C for 30 min in the dark. The stained cells were analyzed using an Accuri C6 flow cytometer (Accuri Cytometers Inc.).

For the detection of intracellular ROS, an oxidation-sensitive fluorescent probe (DCFH-DA) was used. After treatment with agents for 24 h, a total of 1×10⁶ cells was trypsinized and pelleted by centrifugation and washed twice with PBS. Then, the cell pellets were resuspended in 1 ml RPMI-1640 (serum-free) with 10 µm/l DCFH-DA and incubated for 20 min at 37°C. The positive control was treated with Rosup according to the manufacturer's protocol (data not shown). The stained cells were analyzed using an Accuri C6 flow cytometer (Accuri Cytometers Inc.). The FL1-A received the fluorescence induced by DCF. For each sample, 20,000 events were collected.

To investigate the anti-proliferative effect of metformin and sorafenib in combination on HCC cells in vivo, a nude mouse model bearing HCC cell xenografts was established. Five-week-old male athymic nude mice were obtained from the Animal Facility of Dalian Medical University. The mice were maintained under pathogen-free conditions and were provided with sterilized food and water. First, 5×10⁶ Bel-7402 cells were injected subcutaneously into the right flank near the hind leg of each nude mouse. When the mice bore palpable tumors (the tumor volume was ~100 mm³), they were randomly divided into control [100 µl normal saline (NS) by intraperitoneal injection plus 100 µl 1% dimethyl sulfoxide (DMSO)] and 0.5% carboxymethyl cellulose [(CMC)-Na sterile water), metformin (200 mg/kg/day by intraperitoneal injection plus 100 µl 1% DMSO and 0.5% CMC-Na sterile water)], sorafenib (30 µg/kg/day by intragastric administration plus 100 µl NS by intraperitoneal injection) and combination (metformin, 200 mg/kg/day by intraperitoneal injection plus sorafenib 30 µg/kg/day by intragastric administration) groups (n=6 animals/group). The treatments were performed for 4 weeks, 5 times/week. The tumor volume was detected every week and was calculated by the following formula: volume = 1/2 (length x width²). After 4 weeks, the mice were euthanized and the tumors were isolated.

The tumors isolated from the mice were paraffin-embedded and cut into 10 µm sections in a microtome cryostat (HM 500 OM; Carl Zeiss, Jena, Germany). Immunohistochemical staining was conducted according to the manufacturer's protocols for the Histostain™-Plus kits. Primary antibodies (as described in the antibodies section) were incubated at 4°C overnight. Images were captured with a light microscope (Axiolab; Carl Zeiss) and 5 images/sample were prepared. The Image-Pro Plus 4.5 software was used to analyze the staining data.

In situ detection of apoptotic cells in the tumors isolated from the mice was performed by TUNEL assay. The tumors were paraffin embedded and cut into 10 µm sections in a microtome cryostat (HM 500 OM; Carl Zeiss). The TUNEL assay was conducted according to the manufacturer's protocols. 3,3-diaminobenzidine (DAB) was used as the substrate for the peroxidase. The images were captured with a light microscope (Axiolab; Carl Zeiss) and 5 images/sample were prepared. Image-Pro Plus 4.5 software was used to analyze the staining data.

HCC cells were harvested by trypsinization and fixed by immersion in 2.5% glutaraldehyde in 0.1 ml phosphate buffer (pH 7.4). The cells were then post-fixed in 1% osmium tetroxide, dehydrated in a graded series of ethanol and embedded in epoxy resin. The cells were sliced into 1 µm sections and stained with 0.2% lead citrae and 2% uranyl acetate. Representative areas were chosen for ultra-thin sections to view with a Tecnai Spirit electron microscope.

SPSS 13.0 statistical software was used for the statistical analysis. The values are presented as the mean ± SD. Statistical analyses were performed using Student's t-test. The analysis of multiple groups was performed with ANOVA with an appropriate post-hoc test.

The effect of a combination treatment of metformin and sorafenib on HCC proliferation was investigated in Bel-7402 and HepG2 cells using CCK-8 and clonogenic assays. The CCK-8 assay revealed a much stronger inhibitory effect of the combination than any single drug treatment for 3 days (Fig. 1A and B). A comparable defect was detected in single cell colony formation (Fig. 1C and D). Over a 10-day treatment, although metformin and sorafenib suppressed single cell colony progression, the combination almost eliminated colony formation. Next, we performed cell cycle analysis. A significant increase in the number of cells arrested in G0/G1 phase was observed in Bel-7402 and HepG2 cells with a combined treatment of metformin and sorafenib compared to the single drug treatment (Fig. 1E and F). Consequently, cyclin D1 and CDK4, the key regulators for G0/G1 to S phase transition, were further reduced in Bel-7402 and HepG2 cells (Fig. 1G and H) by the combined treatment when compared to the any single treatment. Remarkably, 10 mmol/l metformin and 5 µmol/l sorafenib were used for the CCK-8 assay and 5 mmol/l metformin and 2.5 µmol/l sorafenib were used for the clonogenic assay which are described in the figure legends. We had used the same drug concentration of CCK-8 assay in clonogenic assay, but no clones grew (data not shown). Thus, we think the lower cell concentration may increase the sensitivity of HCC cells to metformin and sorafeinib and we decreased the drug concentration in clonogenic assay. So the different drug concentration may be responsible for the different results between CCK-8 and clonogenic assay. Collectively, these results revealed that a combined treatment of metformin and sorafenib remarkably promotes anti-proliferative effects and G0/G1 cell cycle arrest in HCC cells.

Apoptosis is an underlying antitumor mechanism for both metformin and sorafenib. Here, we investigated the effects of the combined treatment on apoptosis in Bel-7402 and HepG2 cells. The apoptosis assay (Annexin V/PI staining), as shown in Fig. 2A and B, revealed a significant increase in the number of apoptotic cells observed in Bel-7402 and HepG2 cells with a combined treatment of metformin and sorafenib compared to the single drug treatment. To confirm these results, we performed western blot analysis and found increased expression of cleaved PARP and cleaved caspase-3 in both metformin- and sorafenib-treated Bel-7402 and HepG2 cells. Consistent with the greater apoptotic events, the combined treatment led to higher levels of cleaved PARP and cleaved caspase-3 in both Bel-7402 and HepG2 cells compared to single drug treatment alone (Fig. 2C and D). As sorafenib can induce mitochondrial-dependent ROS production, which elicits cell death in hepatomas (13), we next determined the influence of metformin on sorafenib-induced ROS production. As expected, data in Fig. 2E and F present that metformin (10 mmol/l) was readily to promote sorafenib (5 µmol/l) induced ROS production.

To analyze the potential molecular mechanism, we investigated the effects of metformin and sorafenib on the AMPK/ERK and mTOR pathways in HCC cells (Fig. 3). Both metformin (10 mmol/l) and sorafenib (5 µmol/l) activated AMPK and inhibited ERK compared to untreated HCC cells. At concentrations of 0, 5 and 10 mmol/l, metformin significantly inhibited the mTORC1 (p-mTOR Ser2448, p-p70S6K Thr389, p-Raptor Ser792, p-4E BP1 Thr37/46) but activated the mTORC2 (p-mTOR Ser2481, p-Rictor Thr1135, p-AKT Ser473) (Fig. 3) pathway. Inconsistent results were obtained after sorafenib treatment, which suppressed the activation of both the mTORC1 and the mTORC2 pathway in HepG2 and Bel-7042 cells (Fig. 4A and B). The consequences of the combination of metformin and sorafenib on mTOR1/2 networks was further investigated, revealing that sorafenib reversed the activation status of mTORC2 induced by metformin and enhanced the suppression of mTORC1 by metformin in HCC cells (Fig. 4C and D). This result may explain the additional reduction in proliferation upon combined treatment.

To further validate our in vitro results showing the anti-proliferative effects of combined treatment of metformin and sorafenib, we treated male athymic nude mice bearing palpable tumors (~100 mm³) of Bel-7402 xenografts with control (vehicle-treated mice), metformin (200 mg/kg/day), sorafenib (30 µg/kg/day) and their combination (Fig. 5A–C) for 28 days. The tumor volumes of the combined treatment group were significantly reduced compared to the groups treated with either metformin or sorafenib. Consistent with the in vitro data, immunohistochemistry and TUNEL analyses (Fig. 5D and E) of the xenograft tumors revealed that the metformin and sorafenib combination effectively inhibited the expression of Ki-67, a marker for representing tumor proliferation. Moreover, metformin facilitated the sorafenib induced apoptosis in xenograft tumors as evaluated by the cleaved caspase-3 staining and TUNEL assay.

Immunohistochemistry and western blot analysis of the xenograft tumors revealed that metformin and sorafenib inhibited the activation of ERK in the in vivo xenograft model as well (Fig. 6A–C). Similarly, xenograft tumors were further examined for the status of p-mTOR (Ser2448 and Ser2481) and p-AKT (Ser473) after treatment with the control, metformin, sorafenib and their combination (Fig. 6). Consistent with the in vitro results, sorafenib significantly reversed the activation status of mTORC2 and the combined treatment of metformin and sorafenib synergistically abrogated activity of mTORC1.

Since combined treatment of metformin and sorafenib profoundly inhibit the mTOR pathway, which is known to closely regulate autophagy, the potent influences of both on HCC autophagy was further determined. Transmission electron microscopy (TEM) and western blot analysis were used to monitor the autophagic status in HCC cells and clearly demonstrated that accelerated autophagy occurs after 48 h treatment with metformin and sorafenib (Fig. 7A and B). The results of TEM revealed increased autophagic vacuoles in HCC cells treated with metformin and sorafenib combination compared to single drug or control. Microtubule-associated protein light chain 3 (LC3) is a specific marker for autophagy initiation. When autophagy occurred, the cytoplasmic form of LC3 (LC3-I) converted to the pre-autophagosomal and autophagosomal membrane-bound form of LC3 (LC3-II) and p62 was degraded (14). However, two autophagy inhibitors CQ and 3MA could both promote apoptotic death induced by metformin and sorafenib in HCC cells (Fig. 7C–E). CQ could prevent fusion of endosomes and lysosomes and inhibit the formation of autophagosomes. Thus, accumulated LC3-II and p62 could be detected by western blot analysis (Fig. 7B and E) The 3MA is a phosphoinositide 3-kinase inhibitor that inhibits autophagy before the formation of autophagosomes. In conclusion, these results revealed that combined treatment of metformin and sorafenib triggered autophagy in HCC cells, which may be cytoprotective as inhibition of autophagy by CQ or 3MA could promote metformin and sorafenib induced apoptotic cell death. Especially, for the LC3-II band intensity in Fig. 7E, we found that when the cells were treated with CQ, LC3-II displayed strong increase compared to the other groups. The band intensity would be overtested in group with CQ if the other groups could be displayed as in Fig. 7B. So, the bands in the other groups were too weak for accurate quantification.

Collectively, we propose a potential molecular mechanism where metformin and sorafenib inhibit HCC development via modulating AMPK and the mTOR1/2 pathway (Fig. 8).