The construction protocol has been published in our previous study (13). In short, CMV-IRES-dsred2 vector (Clontech, Mountain View, CA) was digested by Aquaspirillum serpens (Ase) I and Bacillus megaterium (Bmt) I then blunted by Klenow enzyme to form pIres-dsred2. The NF-κB responsive element isolated from pNF-κB-luc vector (Clontech) by Micrococcus luteus (Mlu) I and Haemophilus influenzae Rd (Hind) III was blunted by Klenow enzyme and then ligated into pIres-dsred2 to form the pNF-κB-ires-dsred2 vector. The luc2, which was isolated from pGL4-luc2 (Promega, Madison, WI) by Bacillus stearothermophilus (Bst) XI and Nocardia otitidis-caviarum (Not) I and blunted by Klenow enzyme, was inserted downstream of pNF-κB-ires, which was derived from pNF-κB-ires-dsred2 after being digested by NotI and Xanthomonas badrii (Xba) I, to form pNF-κB-ires-luc2. The HSV1-tk, which was isolated from pORF-HSV1-tk (InvivoGen, San Diego, CA) by HindIII and Escherichia coli RY13 (EcoR) I then blunted Klenow enzyme, was inserted into pNF-κB-ires-luc2 to form pNF-κB-tk-luc2 vector.

Two human HCC cell lines, Huh7 and Hep3B, were used in this study. Huh7 was kindly provided by Dr Jason Chia-Hsien Cheng at the Department of Radiation Oncology, National Taiwan University Hospital, Taipei, Taiwan. Hep3B was obtained from the American Type Culture Collection (ATCC, Gaithersburg, MD). Both cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with supplemental 10% fetal bovine serum (FBS) and cultured at 37°C in a humidified incubator containing 5% CO₂. The Huh7/NF-κB-tk-luc2/rfp stable clone (transfection procedure as described) was maintained in the same medium as Huh7 with additional 500 μg/ml of G418 (Calbiochem, Darmstadt, Hesse, Germany).

Huh7 cells (2×10⁶) were seeded in a 10-cm dia meter dish for 24 h prior to transfection. pNF-κB-tk-luc2 vector (8 μg) and 16 μl of jetPEI™ solution (Polyplus Transfection, Strasbourg, Alsace, France), diluted with 500 and 484 μl of 145 mM NaCl, respectively, were mixed and incubated for 30 min at room temperature to make the 1000 μl jetPEI™/DNA mixture. The mixture was then added to Huh7 cells and incubated at 37°C for 24 h followed by trypsinization, then cultured with DMEM containing 1 mg/ml G418 supplemented with 10% FBS for two weeks. The surviving clones were isolated and transferred to 96-well plates for cell growth. The expression of luc2 protein in each clone was assayed using BLI, and re-named as Huh7/NF-κB-tk-luc2 cell line.

CAG promoter (composed of CMV enhancer and β-actin promoter)-driven red fluorescent protein (RFP) vector (National RNAi Core Facility Platform, Academic Sinica, Taipei, Taiwan) was used to transfect Huh7/NF-κB-tk-luc2 cell line with the same strategy as the afore-mentioned protocol. Two weeks after RFP vector transfection, cells with red fluorescence were sorted and isolated by flow cytometry. The isolated cells were transferred to 96-well plates for growth. The RFP expression in each clone was assayed using an IVIS50 Imaging System (Xenogen, Alameda, CA). This bioluminescent cell clone with simultaneous red fluorescence was renamed as Huh7/NF-κB-tk-luc2/rfp cell line.

Huh7/NF-κB-tk-luc2 cells transfected with IκBα mutant vector (pI-κBαM, Clontech) was used as the positive control for the inhibition of NF-κB as compared with that suppressed by sorafenib treatment. In addition, Huh7/NF-κB-tk-luc2 cell line transfected with empty vector was used as the negative control for NF-κB inhibition.

The method for extraction of sorafenib was described in our previous study (21). Briefly, sorafenib was extracted from commercial Nexavar^® tablets (Bayer, Leverkusen, North Rhine-Westphalia, Germany) composed of 200 mg sorafenib. The tablet was ground to fine powder and transferred to a 100-ml conical flask. The powder was washed with 15 ml deionized distilled water three times to remove water-soluble components; 15 ml ethyl acetate was then used to extract the precipitate three times in order to recover the sorafenib. The organic phases were combined and dried over anhydrous sodium sulfate, followed by evaporation under reduced pressure. The residue was recrystallized with acetone and hexane to yield a white solid extract weighing about 122 mg (60% recovery). Nuclear magnetic resonance (NMR) spectra were recorded with a spectrometer (Varian Gemini 200; Oxford Instruments, Abingdon, Oxfordshire, UK) to determine the chemical structure of the sorafenib extract. High-performance liquid chromatography (HPLC) was conducted using a PU-2089 plus quaternary gradient pump, equipped with a UV-2075 Plus intelligent UV/VIS detector (Jasco, Tokyo, Kanto, Japan). The 1H-NMR spectrum of the recovered sorafenib was the same as the one reported before (22). More than 98% chemical purity was achieved for the recovered sorafenib (retention time, 16.2 min) as determined with HPLC.

Male nude mice (n=6 per group and repeated 3 times, total 72 mice, 4–6 weeks old, purchased from National Laboratory Animal Center, Taipei, Taiwan) were injected subcutaneously with 1×10⁷ Huh7/NF-κB-tk-luc2/rfp cells in the right hind flank. Mice were randomly divided into four groups (Fig. 5): control [vehicle treated with 100 μl phosphate-buffered saline (PBS) in 1% dimethyl sulfoxide (DMSO) daily by gavage], SAHA alone (25 mg/kg/day for 21 days by gavage), sorafenib alone (20 mg/kg/day for 21 days by gavage) and combination (SAHA plus sorafenib, 25 mg/kg/day plus 20 mg/kg/day for 21 days by gavage). All treatments were administered by gavage. Treatment was initiated when tumor volume reached about 50 mm³. Tumor volume was assayed with digital caliper twice a week and BLI once a week. Mice were sacrificed on day 21 post-treatment for ex vivo EMSA, ex vivo western blotting, and whole-body autoradiography. The animal study protocols complied with institutional animal care and use guideline of National Yang-Ming University, Taipei, Taiwan.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich, St. Louis, MO) was dissolved in PBS (145 mM NaCl, 1.4 mM KH₂PO₄, 4.3 mM Na₂HPO₄ and 2.7 mM KCl, pH 7.2). Huh7/NF-κB-tk-luc2/rfp cells were seeded into 96-well plates with 3×10⁴ cells/well for 24 h, then treated with different concentrations of SAHA (0–10 μM in 0.1% DMSO) or SAHA combined with 10 μM sorafenib for additional 48 h. After washing with fresh medium, 100 μl of 5 mg/ml MTT solution was added to each well followed by incubation at 37°C for 2 h. Then 100 μl DMSO was added to dissolve the MTT formazan, and the absorbance was determined with an enzyme-linked immunosorbent assay (ELISA) reader (Power Wave X340, Bio-Tek, Winooski, VT) using a wavelength of 570 nm for excitation.

Huh7/NF-κB-tk-luc2/rfp cells (1×10⁶) were seeded in 6-well plates and treated with vehicle (0.1% DMSO), 10 μM SAHA, 10 μM sorafenib or combination of both for 48 h. After treatment a genomic DNA purification kit (Axygen, Tewksbury, MA) was used to extract DNA following the instruction provided by the manufacturer. DNA laddering was analyzed with 1.5% agarose gel electrophoresis.

Forty-eight hours post-treatment with vehicle, 10 μM SAHA, 10 μM sorafenib and combination of SAHA and sorafenib, respectively, nuclear fractions of Huh7/NF-κB-tk-luc2/rfp cells were obtained using Nuclear Extraction kit (Chemicon, Temecula, CA). The NF-κB/DNA binding activity was evaluated using LightShift Chemiluminescent EMSA kit (Thermo Scientific, Rockford, IL). The procedure followed the protocol provided with the kit. In brief, DNA sequences were synthesized for NF-κB binding. Sense: AGTTGAGGGGACTTTCCCAGGC and antisense: GCCTGGGAAAGTCCCCTCAACT. Nuclear extracts were incubated with the biotin-labeled DNA probe for 20 min at room temperature. The protein-DNA complex was separated from free oligonucleotides on a 5% polyacrylamide gel, then transferred to a nylon membrane and cross-linked with UV light. The membrane was incubated with streptavidin-horseradish peroxidase, and detected by enhanced chemiluminescence (ECL, Thermo Scientific).

Mice were sacrificed on day 21 and tumors were removed for nuclear protein extraction by using Nuclear Extraction kit (Chemicon). The NF-κB/DNA binding activity was evaluated using LightShift Chemiluminescent EMSA kit (Thermo Scientific).

Huh7/NF-κB-tk-luc2/rfp cells (2×10⁶) were seeded into a 10-cm diameter dish (5 dishes/group) for 24 h prior to the treatment with vehicle, 10 μM SAHA, 10 μM sorafenib or combination of 10 μM SAHA and 10 μM sorafenib, respectively, for another 48 h. Cells were then lysed with 100 μl lysis buffer [50 mM Tris-HCl (pH 8.0), 120 mM NaCl, 0.5% NP-40, 1 mM phenylmethanesulfonylfluoride]. Total proteins (40 μg) were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), then transferred to a polyvinylidene difluoride membrane (Millipore, Billerica, MA), blocked with 5% non-fat milk in TBS-Tween buffer (0.12 M Tris-base, 0.15 M NaCl and 0.1% Tween-20) for 1 h at room temperature, and then incubated with primary antibody (Millipore) for XIAP, BCL-2, VEGF, cyclin D1, c-FLIP, caspase-3, caspase-8, cytochrome-C, extracellular signal-regulated kinase (ERK) and β-actin, respectively, overnight at 4°C, followed by incubation with secondary peroxidase-conjugated anti-rabbit antibody for 30 min at room temperature. The expressions of proteins were determined by ECL (Millipore). ImageJ software (National Institutes of Health, Bethesda, MD) was used for the quantitative analysis. A cytosol/nuclear extraction kit (Chemicon) was used to extract cytosolic cytochrome-C following the manufacturer’s protocol.

Mice were sacrificed on day 21 after treatments. Tumors were removed for protein extraction using T-PER kit (Thermo Scientific). Equal amounts of protein (40 μg) were loaded for SDS-PAGE, then transferred to nitrocellulose membranes. The membranes were incubated with primary antibodies specific for VEGF, XIAP, BCL-2, cyclin D1, c-FLIP, caspase-3 and caspase-8, cytochrome-C, ERK and β-actin, followed by incubation with horseradish peroxidase-conjugated secondary antibodies. The protein expression was determined by ECL. A cytosol/nuclear extraction kit (Chemicon) was used to extract cytosolic cytochrome-C following the manufacturer’s protocol.

Procedure for each modality of molecular imaging was described in our previous study (13). In brief, BLI was used for monitoring the NF-κB activation in Huh7/NF-κB-tk-luc2/rfp cells. A total of 3×10⁴ Huh7/NF-κB-tk-luc2/rfp cells were cultured in a 96-well plate for 48 h, then treated with various concentrations of SAHA (0–10 μM), 10 μM sorafenib, combination of SAHA and sorafenib, and 10 μM of various signaling molecule inhibitors for 48 h, respectively. D-luciferin (100 μl of 500 μM, Xenogen) was added to each well for imaging. Relative NF-κB activity was calculated as follows: ROI value of treatment group/ROI value of control group, and both ROIs were corrected by cell viability which was determined by MTT assay. BLI was also used in monitoring temporal change of NF-κB activation in the Huh7/NF-κB-tk-luc2/rfp tumor-bearing mice. Mice were anesthetized using 1–3% isoflurane and received 150 mg/kg D-luciferin via intraperitoneal injection 15 min before imaging. RFPI was used for monitoring viable tumor cell growth in the Huh7/NF-κB-tk-luc2/rfp tumor bearing mice. The photons emitted from tumors obtained by BLI and RFPI were detected with an IVIS50 Imaging System (Xenogen). The image acquiring periods for BLI and RFPI were 5 min and 30 sec, respectively. Regions of interest (ROIs) of the images were drawn conformally to the tumor contour and quantified as photons/sec using Living Image software (Version 2.20, Xenogen).

Mice were injected intravenously with 3.7×10⁶ Bq/0.2 ml ¹³¹I-1-(2-deoxy-2-fluoro-1-D-arabinofuranosyl)-5-iodouracil (FIAU) on day 21 post-treatments, and sacrificed 24 h later for whole-body autoradiography. Frontal sectioning was performed with thickness of 30 μm at −20°C with a cryostat microtome (Bright Instrument, Huntingdon, Cambridgeshire, UK). Sections were placed on the imaging plate (BAS-SR2040, Fuji Photo Film, Tokyo, Kanto, Japan) in the cassette (2040, Fuji Photo Film). After exposure, the imaging plate was assayed with a FLA5000 reader (Fuji Photo Film) to acquire the phosphor image. The parameters of the reader were: resolution of 10 μm, gradation of 16 bits, 635-nm laser light and 800 V of photomultiplier tube.

All data were shown as the mean ± standard error. Student’s t-test was used for the comparison between two groups. Differences between the means were considered significant at p<0.05 or less.

The viabilities of both Huh7/NF-κB-tk-luc2/rfp and Hep3B cells were decreased linearly (Fig. 1A and B) with increasing SAHA concentrations (2–10 μM) with or without 10 μM sorafenib concurrently. Either SAHA or sorafenib alone induced apoptosis through extrinsic (cleaved caspase-8) and intrinsic (cytochrome-C) pathways as unraveled by increased expressions of apoptosis-related proteins (Fig. 1E and F). Combination of SAHA and 10 μM sorafenib not only enhanced HCC cytotoxicity, but accompanied by obvious DNA fragmentation and significantly increased expression of pro-apoptotic proteins (Fig. 1).

After treatments with various concentrations of SAHA for 24 h, the NF-κB activity was inhibited significantly in HCC cells in a dose-dependent manner (Fig. 2A), and 10 μM SAHA reduced effector proteins expressions (Fig. 2B) to 50% of those of the controls. Notably, SAHA significantly induced NF-κB activity in a dose-dependent manner after 48 h treatments (Fig. 2C), and increased the expressions of NF-κB-regulated effector proteins by 1.5- to 2-fold except cyclin D1 as compared to those of the controls (Fig. 2D).

SAHA-induced NF-κB/DNA binding activity was inhibited with the transfection of pI-κBαM in Huh7/NF-κB-tk-luc2 cells (Fig. 3A). Cell viability after SAHA treatment was significantly decreased in Huh7/NF-κB-tk-luc2 cells transfected with pI-κBαM as compared with that transfected with empty vector (Fig. 3B). Apoptosis based on the DNA fragmentation assay was significantly increased in the combination group as compared with those of others groups (Fig. 3C). The transfection of pI-κBαM inhibited SAHA-induced expression of NF-κB-regulated downstream effector proteins and increased SAHA-induced apoptotic-related proteins (Fig. 3C and D).

Intrinsic NF-κB activation, SAHA-induced NF-κB activity and expressions of downstream effector proteins were suppressed significantly by sorafenib as shown in Fig. 4A–C. Among the inhibitors for signal transduction, i.e., AKT inhibitor, c-Jun N-terminal kinase (JNK) inhibitor, ERK inhibitor and P38 inhibitor, only ERK inhibitor shows similar effect as sorafenib in suppressing SAHA-induced NF-κB activation (Fig. 4D). Furthermore, SAHA-induced ERK phosphorylation in Huh7/NF-κB-tk-luc2/rfp cells was suppressed by sorafenib as shown in Fig. 4E.

Experimental design for the study in vivo is shown in Fig. 5. Though the tumor volumes of SAHA-treated mice were smaller as compared with those of the control, no significant difference was found. Both sorafenib-treated and combination groups had significantly smaller tumor volumes throughout the experimental period until day 21 as compared with those of the control. Notably, the highest tumor suppression was found in the combination group (Fig. 6A). The viable tumor cell growth was evaluated with RFPI once a week. The emitted photon flux correlates well with the number of viable tumor cells (13). The results demonstrated in Fig. 6B were similar to those shown in Fig. 6A. Since the NF-κB level could be increased by SAHA treatment for 48 h, the NF-κB activation inside the tumor was observed by BLI once a week. The higher photon flux represents the more intense NF-κB activity (20). SAHA significantly increased the NF-κB activity as compared with that of the control, but the increase of NF-κB activity could be suppressed by sorafenib as shown in the combination group (Fig. 6C). Simultaneous imaging of living tumor cells and NF-κB activity by RFPI and BLI, respectively, was shown in Fig. 6D. In addition, the uptake of ¹³¹I-FIAU in tumors, which represented the NF-κB activity, assayed by whole-body autoradiography (13) were also shown to be the most prominent in SAHA-treated mice. Lower uptake of ¹³¹I-FIAU in sorafenib-treated and combination groups than that of the control was found (Fig. 7A). Furthermore, sorafenib not only inhibited SAHA-induced NF-κB/DNA binding activity, but expression of oncogenic proteins as well, while enhanced expression of SAHA-induced pro-apoptotic proteins by ex vivo western blotting (Fig. 7C).