The Mahlavu and SK-HEP-1 human hepatocellular carcinoma lines were used in this study. The cell lines were maintained in DMEM and 10% fetal bovine serum at 37°C in a humidified atmosphere of 5 % CO₂ in air.

Targeting SP94 (SFSIIHTPILPL) peptides were synthesized and purified by reverse-phase high-performance liquid chromatography to >95% purity by the Peptide Synthesis Core Facility, Institute of Cellular and Organismic Biology, Academia Sinica. The predicted mass was confirmed by mass spectrometry.

A total of 8.5 mg of NHS-PEG-DSPE [N-hydroxysuccinimido-carboxyl-polyethyleneglycol (MW, 3400)-derived distearoylphosphatidyl ethanolamine] (NOF Corp.) dissolved in 0.25 ml of dichloromethane (Sigma-Aldrich) was added to 0.25 ml of DMSO (Sigma-Aldrich) containing 3.1 mg of peptide. This was then mixed with 0.011 ml of triethylamine (Sigma-Aldrich) to catalyze the reaction. The stoichiometric molar ratio of peptide and NHS-PEG-DSPE was 1.1:1. The reaction was carried out for 72 h at room temperature. The peptide-PEG-DSPE conjugates were purified by dialysis with a 2-kDa cut-off membrane (Spectrum), and were then dried through lyophilization.

A lipid film hydration method was used to prepare PEGylated liposomes composed of distearoylphosphatidylcholine, cholesterol, and mPEG2000-DSPE, which were then used to encapsulate doxorubicin (3:2:0.3 molar ratio) or vinorelbine (3:2:0.15 molar ratio). The lipid films were hydrated at 60°C in 250 mM ammonium sulfate or 300 mM ammonium salts of 5-sulfosali-cylic acid solution, and were extruded through polycarbonate membrane filters with a pore size of 0.1 µm using high-pressure extrusion equipment (Lipex Biomembranes, Vancouver, BC, Canada) at 55°C. Doxorubicin or vinorelbine were encapsulated by a remote loading method, at concentrations of 1 mg or 3.5 mg per 10 µmol of phospholipid, respectively. The final concentration of liposome was estimated by phosphate assay. The peptide-PEG-DSPE was subsequently incorporated into pre-formed liposomes by co-incubation at 60°C, the transition temperature of the lipid bilayer, for 0.5 h with gentle shaking. Sepharose 4B (GE Healthcare) gel filtration chromatography was used to remove released free drug, unconjugated peptides, and unincorporated conjugates. Doxorubicin concentrations in the fractions of eluent were determined by measuring fluorescence at λEx/Em = 485/590 nm using a spectrofluorometer (Spectra Max M5, Molecular Devices). Vinorelbine concentrations were determined using the HPLC method.

M13 bacteriophages were amplified in Escherichia coli, and phage titers were determined according to published procedures. Following titer determination, the bacteriophages were simultaneously labeled with succinimidyl esters of HiLyte Fluor™ 750 using a modified procedure (21,22). Briefly, HiLyte Fluor 750 succinimidyl ester was added to M13 bacteriophage in 100 µM bicarbonate buffer, pH 8.3. The resulting solution was incubated for 1 h at room temperature in the dark. Following incubation, the labeled bacteriophage was precipitated by addition of a PEG-8000/2.5 M NaCl solution, and then left to stand on ice for 30 min. The bacteriophage was pelleted by centrifugation at 10,000 × g for 15 min. After removal of the supernatant, the pellet was resuspended in DPBS buffer. Typical dye labeling using this procedure resulted in 400–500 copies of each dye per bacteriophage particle.

A total of 4×10¹¹ pfu of HiLyte Fluor 750-labeled PC94 phage was diluted in 100 µl saline solution. The saline solution was injected i.v. into mice bearing subcutaneous Mahlavu tumors. A control solution of 4×10¹¹ pfu of the HiLyte Fluor 750-labeled control phage was injected into NOD.CB17-Prkdc^scid/J mice bearing subcutaneous Mahlavu tumors. In vivo imaging was performed using a Xenogen IVIS^® Imaging System 200. The animal was imaged at 0.1, 0.5, 6, 24, and 48 h post-injection using a Indocyanine Green (ICG) Filter set (excitation 710–760 nm, emission 810–875 nm). Organs were dissected and imaged 48 h after injection of conjugate.

The dorsolateral flanks of severe combined immunodeficient mice, NOD. CB17-Prkdc^scid/J (4–6 weeks old), were injected s.c. with 5×10⁶ Mahlavu or SK-HEP-1 cells. Tumors were measured with calipers, and mice were weighed twice weekly. The tumor volumes were calculated using the following formula: length × (width)² × 0.52. All animals were cared for in a specific pathogen-free room and treated in accordance with the animal care protocol approved by the Academia Sinica Animal Committee (approval no. 11-06-190).

Plasma, brain, heart, lung, liver, kidney, or tumor samples were processed using the extraction procedure, and then analyzed using the method described by Laginha et al (23). Result concentrations were determined relative to the respective calibration curves.

Frozen tumor tissue sections were incubated with TUNEL reaction mixture (Roche Diagnostics) at 37°C for 1 h. The slides were counterstained with Hoechst 33258 (Molecular Probes) and mounted with mounting medium (Vector Laboratories). The slides were then visualized under a fluorescent microscope. The sections were analyzed using automated cell acquisition (TissueGnostics), and TUNEL-positive areas were quantified using MetaMorph software (Molecular Devices).

The frozen tumor tissue sections were fixed with methanol/acetone (1:1), washed with PBS, and immersed in blocking buffer (1 % bovine serum albumin in PBS), followed by incubation with rat anti-mouse CD31 (BD Pharmingen). The sections were washed with PBST0.1 (0.1 % Tween-20 in PBS), and then incubated with rabbit anti-rat antibody (Stressgen) and immersed in rhodamine-labeled goat anti-rabbit antibody solution (Jackson ImmunoResearch). The slides were counter-stained with Hoechst 33258, mounted with mounting medium, and visualized under a fluorescent microscope.

Blood was collected by venipuncture of the right submandibular vein of unanesthetized mice with a lancet. Samples were drawn into plastic K2 EDTA blood-drawing tubes. Complete blood counts were performed using the Abbott CELL-DYN 3700 Hematology Analyzer (Abbott Diagnostics, Abbott Park, IL, USA). The parameters assessed were as follows: white blood cell (WBC) count, neutrophil absolute count (NEU), lymphocyte absolute count (LYM), red blood cell/erythrocyte count (RBC), hemoglobin concentration (HGB), hematocrit (HCT), red cell distribution width (RDW), platelet/thrombocyte (PLT), mean platelet volume count (MPV), plateletcrit (PCT), neutrophil percentage (NEU %), lymphocyte percentage (LYM %), monocyte percentage (MONO %), monocyte absolute count (MONO), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and platelet distribution width (PDW).

Mahlavu cells were seeded in 96-well plates and allowed to attach overnight. Cells were treated in 96-well format in triplicate for each drug concentration combination, and viability was assessed after 3 days of treatment using the MTT-based assay. The cytotoxicity of each drug and of their combinations was assessed by drug response matrix using the Combenefit software (24). Loewe synergy score and HSA synergy score were calculated and plotting by SynergyFinder software (25).

NOD.CB17-Prkdc^scid/J mice were used for HCC implantation. SK-HEP-1 cells were infected with Lenti-luc virus (lentivirus containing the luciferase gene). The mice were anesthetized via i.p. injection of Avertin, 2,2,2-Tribromoethanol (Sigma Chemical Co.) at a dose of 250 mg/kg. Prior to orthotopic implantation, a 1-cm laparotomy was performed, and orthotopic human hepatocellular carcinoma (HCC) was established by intrahepatic injection of 10⁵ SK-HEP-1-Luc cells (luciferase-expressing cells) suspended in 30 µl DMEM into the left liver lobe. Post-injection bleeding and tumor cell escape were avoided by short-term local compression. The abdomen was closed using an absorbable 5-0 vicryl suture, and the skin was closed with a 5-0 proline suture. For orthotopic therapeutic studies, implanted mice were treated with different formulations of anticancer drugs. Tumor progression was monitored by bioluminescence quantification. Mouse body weight and survival rate were measured. Animal care was carried out in accordance with the guidelines of Academia Sinica, Taiwan. The experimental protocols were approved by the Academia Sinica Institutional Animal Care and Utilization Committee (approval no. 11-06-190).

All data were expressed as the mean ± standard error (SEM) of at least 3 independent experiments. Statistical significance was assessed using a two-tailed Student's t-test with P<0.05 considered to be statistically significant. Non-compartmental pharmacokinetic analysis of plasma concentrations versus time was performed with pharmacokinetic parameters using WinNonlin software version 5.2 (Pharsight, Mountain View, CA, USA).

Intracellular trafficking is a determining factor in the therapeutic efficiency of nanoparticle-based drug delivery. In the present study, transmission electron microscopy was used to investigate the mechanism of cellular uptake of SP94-modified liposomes. TEM images of SK-HEP-1 cells incubated with SP94-LD or LD revealed that liposomes were internalized by vesicular transport, and partially escaped to the cytosol at the perinuclear region at 37°C (Fig. 1A and B). After SP94-LD exposure, SK-HEP-1 cells exhibited numerous coated pit structures (Fig. 1C and D) and endocytotic vesicles in the cell membrane and cytoplasm (Fig. 1E–I). In contrast, few endocytotic vesicles were observed at the same magnification in cells incubated with LD (Fig. 1B). Analysis of the TEM images revealed that cells incubated with SP94-LD showed a high amount of internalized vesicles (Fig. 1J–M). The total vesicle area per SK-HEP-1 cell for SP94-LD was ~8.8-fold greater than that for LD (Fig. 1M).

Molecular imaging plays a critical role in oncological drug development, as stated in the FDA's Critical Path Initiative documents (26). To date, no molecular imaging method has been shown to accurately detect, characterize, or monitor the response of HCC to treatment. Labeling of phages with a near-infrared fluorescence tag enables the distribution of the phages to be efficiently tracked in vivo. We evaluated the in vivo distribution of PC94 phage using a previously described optical imaging method (22). Mice bearing subcutaneous Mahlavu tumors were intravenously injected with HiLyte Fluor 750-labeled PC94 phage or control phage, and phages were monitored at 0.1 and 0.5 h at the initial stages, and again at 6, 24, and 48 h post-injection. Fig. 2A shows representative near-infrared images taken at 0.1, 0.5, 6, 24 and 48 h post-injection from two subjects. Mice injected with HiLyte Fluor 750-labeled PC94 phage exhibited higher fluorescent signals in tumors, as compared to tumors in animals injected with control phage (Fig. 2A and B).

We proceeded to further evaluate the distribution profiles of PC94 phage in mice. The accumulation of phages in Mahlavu tumor peaked at ~24 h and then decreased gradually, but >80% of the peak level was retained in the tumor by 48 h (Fig. 2B). The organ accumulation of phages in Mahlavu tumors was evaluated by near-infrared fluorescence imaging after sacrifice at 48 h post-injection, revealing that the SP94 peptide enhanced accumulation in tumor sites, but not healthy organs including brain, heart, lung, liver, kidney, pancreas, spleen and intestine (Fig. 2C and D).

For pharmacokinetic analysis, SP94-LD and LD were administered to NOD.CB17-Prkdc^scid/J mice at matched 2 mg doxorubicin/kg by tail vein injection. Blood samples were withdrawn at selected time-points, and were examined for doxorubicin content using a validated fluorescent quantitative method. The blood profiles of both SP94-LD and LD were similar throughout the study, declining progressively over time (Fig. 3A). The time course of SP94-LD and LD accumulation in tumor and various organs are shown in Fig. 3B and C. The maximum total doxorubicin concentration in tumor was 1.53±0.47 µg doxorubicin/g tumor, which occurred at 24 h after SP94-LD administration; 1.01±0.18 µg/g remained at 72 h (Fig. 3B). Maximum tumor accumulation of LD at 2 mg/kg (1.03±0.69 µg/g) occurred at 48 h post-injection, and gradually decreased with time to 0.89±0.69 µg/g at 72 h post-treatment. At 24 h, tumor accumulation of SP94-LD was 1.6-fold higher than that of LD. In contrast, distribution of SP94-LD in all non-malignant tissues was similar to that of LD at all time-points (Fig. 3C). The tumor doxorubicin AUC_0–72 for SP94-LD was 81.75 µg h/g and the LD AUC_0–72 was 58.23 µg h/g, representing a 1.4-fold increase in doxorubicin AUC_0–72 for mice treated with the targeted drug.

To determine the amount of bioavailable drug in tumor cells, we performed whole body perfusion through the left ventricle of the heart with DPBS before analyzing biodistribution. This operation can eliminate blood and liposomes remaining in vessels and the interstitial space of tumors. After whole body perfusion, the tumor mass, brain, heart, lung, liver, and kidneys were harvested, and doxorubicin was quantified. We used intracellular accumulation of doxorubicin as an indicator of bioavailability of the liposomal drug. SP94-LD and LD were administered to tumor (SK-HEP-1)-bearing mice at matched 1 mg doxorubicin/kg by i.v. injection. The doxorubicin levels were measured in the blood at different time-points using a fluorescent quantitative method. The blood profiles of both SP94-LD and LD were similar (Fig. 4A). Tumor uptake of SP94-LD at 1 mg/kg gradually increased, before peaking at 0.99±0.24 µg/g at 24 h post-injection, while 0.70±0.13 µg/g remained at 48 h (Fig. 4B). Maximum tumor uptake of LD at 1 mg/kg (0.55±0.11 µg/g) occurred 24 h post-injection, and experienced almost no change over time, remaining at 0.55±0.19 µg/g at 48 h post-injection. At 24 h, a 1.79-fold higher uptake of SP94-LD was observed as compared to LD. The tumor doxorubicin AUC_0–48 for SP94-LD was 34.45 µg h/g and the LD AUC_0–48 was 21.27 µg h/g, representing a 1.62-fold increase in doxorubicin AUC_0–48 when conjugated to SP94. The uptake of both drugs in most non-malignant tissues was low at 1 h post-injection; uptake of both drugs by liver and kidney gradually increased by 24 h, and then slowly declined thereafter (Fig. 4C). Interestingly, the retention of SP94-LD in liver and kidney were lower than that of LD at 24 h post-injection, although these differences were not statistically significant (P=0.14 and P=0.06 respectively).

To evaluate the antitumor efficacy of systemically-administered SP94-LD as compared to LD, NOD.CB17-Prkdc^scid/J were inoculated s.c. with SK-HEP-1 tumors. Mice bearing hepatocellular carcinoma xenografts (~100 mm³) were assigned into four groups for different treatments: A, SP94-LD; B, free doxorubicin (FD); C, LD; and D, PBS. Treatments were administered through tail vein injection, 1 mg/kg every 3.5 days, for eight doses with a total cumulative dose of 8 mg/kg. By day 28, administration of SP94-LD had significantly inhibited tumor growth by 74.6% (P<0.01), whereas treatment with LD and FD inhibited tumor growth by 56.1 (P<0.01) and 11.6% (P>0.05), respectively (Fig. 5A and B), as compared to untreated controls. SP94-LD-mediated inhibition of growth was more significant than that mediated by LD (P=0.013). The SP94-LD and LD groups did not exhibit significant changes in body weight (Fig. 5C) or complete blood counts (Table I) during the treatment period.

Histological analysis of the SK-HEP-1 tumors revealed that FD, LD, and SP94-LD induced apoptosis (Fig. 5D and E), with SP94-LD treatment resulting in the greatest proportion of apoptotic cells (Fig. 5E). The CD31 (angiogenesis) index (Fig. 5F and G) of SP94-LD treated tumors was lower compared to that of tumors treated with FD, LD, or saline.

We have previously observed that SP94-peptide-targeted liposomal doxorubicin (SP94-LD) inhibits the proliferation of small tumors. Nevertheless, in many cancer cases in humans, the tumors are detected when they are large. Thus, here we examined the activities of FD, LD, and SP94-LD against large tumors. Luciferase activity in live cells of mice bearing hepatocellular carcinoma xenografts was measured after i.v. injection with FD, LD, or SP94-LD (Fig. 6A). We observed a statistically significant reduction of tumor growth after i.v. treatment of hepatocellular carcinoma with SP94-LD, as compared with tumors in untreated and FD or LD-treated mice (Fig. 6A and B). SP94-LD significantly inhibited tumor growth as evidenced by a 91.47% reduction in luminescence (P<0.01), whereas treatment with LD reduced luminescence by 96.01% (P<0.01) compared with untreated controls. SP94-LD inhibition of growth was more significant than that by LD (P=0.011). Neither SP94-LD nor LD had a significant effect on body weight during the treatment period (Fig. 6C).

Previous study has shown that doxorubicin can inhibit HCC growth, but the clinical response was low. Combinations of anticancer agents with different mechanisms of action may be effective treatment regimens. For this reason, we attempted to develop an effective treatment for HCC patients by combining two or more anticancer drugs. In this study, HCC cells were treated with doxorubicin or vinorelbine alone, or in combination using a dose-matrix approach to evaluate the drug combination effects at various drug ratios. To determine the combination response (additivity, synergy, or antagonism), the Loewe Additivity score (27) and the Highest Single Agent (HAS) score (28) were calculated for the drug combination using the software Combenefit (24) and SynergyFinder (25). Fig. 7A presents a dose-response matrix between 6 concentrations of doxorubicin and 6 concentrations of vinorelbine in a 4-fold dilution scheme. The cytotoxicity of doxorubicin or/and vinorelbine over a wide range of molar ratios for HCC cells is shown in Fig. 7B. Vinorelbine significantly enhanced the effect of doxorubicin on the viability of HCC cells. Similarly, doxorubicin also enhanced the cytotoxicity of vinorelbine on HCC cells. The drug combination responses at various dose levels were quantified by computational methods. The dose-matrix combination showed the combination modeling between doxorubicin and vinorelbine by using the Loewe Additivity model (Fig. 7C) and HAS model (Fig. 7D). The synergy heatmaps showed that doxorubicin and vinorelbine have additive/synergistic effects (red areas in the model graph) on inhibiting cell proliferation at a wide range of drug combination ratio. These results suggest that the combined use of doxorubicin and vinorelbine may add an advantage to the current therapeutic regimens in liver cancer.

Models based on the subcutaneous injection of cancer cell lines may not accurately reproduce the biology of human HCC. To study the influence of the liver microenvironment upon response to therapy, we developed an orthotopic liver cancer model to recapitulate the tumor growth pattern seen in liver cancer patients. We investigated the antitumor potential of i.v. administration of both SP94-targeted liposomal doxorubicin and liposomal vinorelbine to SK-HEP-1-Luc tumors stably expressing firefly luciferase. Orthotopic tumor growth was non-invasively monitored by bioluminescence imaging. Prior to the first therapeutic injection (4 days after tumor cell implantation), growing orthotopic tumors were found to be localized mainly at the liver (Fig. 8A).

Mice were treated with either vehicle alone (PBS), FD (1 mg/kg) + free vinorelbine 11 (FV) (2 mg/kg), LD (1 mg/kg) + stable liposomal vinorelbine (sLV) (2 mg/kg), SP94-LD (1 mg/kg) + SP94-sLV (2 mg/kg), or sorafenib alone (30 mg/kg) every other day for sixteen days. Bioluminescence was examined weekly to monitor tumor burden. Bioluminescence images revealed significant inhibition of tumor growth in the SP94-LD (1 mg/kg) + SP94-sLV (2 mg/kg)-treated group compared with the vehicle control group, the FD (1 mg/kg) + FV (2 mg/kg) group, or the sorafenib (30 mg/kg) group (Fig. 8A and B). Body weight was not affected by any treatment regimen (Fig. 8C). Kaplan-Meier survival curves of all groups are shown in Fig. 8D. At the end of the study, the median survival times for the PBS, FD + FV, LD + sLV, SP94-LD + SP94-sLV, and sorafenib treatment groups were 39.5, 20, 69, 75 and 41.5 days, respectively (Fig. 8D and E). A survival analysis with a log-rank (Mantel-Cox) test revealed that SP94-LD + SP94-sLV treatment significantly extended animal survival as compared with PBS, FD + FV, LD + sLV, or sorafenib treatment (Fig. 8F).