Dextran from Leuconostoc spp. (average molecular weight [MW] approximately 6,000), hexamethylene diamine [HMDA], N,N-dicylohexylcarbodiimide [DCC], and N-hydroxysuccimide [NHS] were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sorafenib was purchased from LC Laboratories (Woburn, MA, USA). Spectra/Por™ dialysis membranes (MW cutoff [MWCO] = 2,000 g/mol and 8,000 g/mol) were purchased from Spectrum Labs (Rancho Dominguez, CA, USA). Poly(DL-lactic acid-co-glycolic acid) (PLGA-5005, MW = 5,000 g/mol) were purchased from Wako Pure Chemicals (Osaka, Japan).

DexbLG copolymer was synthesized as reported previously 13. Aminated dextran was prepared as follows. Dextran (180 mg) dissolved in dimethylsulfoxide [DMSO] was mixed with sodium cyanoborohydride and stirred for 24 h. After that, 10 equivalents of HMDA were added and stirred for 24 h at room temperature. The resulting aminated dextran was obtained by dialysis against deionized water and was lyophilized. N-hydroxysuccimide PLGA [PLGA-NHS] was prepared by reaction with DCC and NHS. DexbLG copolymer was prepared by dissolving 120 mg of aminated dextran and 100 mg of PLGA-NHS in DMSO and undergoing reaction for 2 days. Reactants were dialyzed to remove unreacted dextran (MWCO of dialysis membrane = 8,000 g/mol), and the product was lyophilized. The resulting white powder was dissolved in chloroform to remove unreacted PLGA. Yield of the final product was about 89% (w/w).

The sorafenib-incorporated DexbLG nanoparticles were prepared by the nanoprecipitation-dialysis method as follows. DexbLG copolymer dissolved in 3 ml of DMSO was mixed with sorafenib in 2 ml of DMSO. This solution was added dropwise to 15 ml of deionized water for over 10 min to form nanoparticles. The solvent was removed by dialysis against deionized water for 1 day. Empty nanoparticles of DexbLG copolymer were prepared by the same procedure, omitting sorafenib. To evaluate the drug contents and loading efficiency, 5 mg of sorafenib-incorporated nanoparticles were distributed into the mobile phase (acetonitrile/methanol/1% acetic acid in a ratio of 35:38:27) and stirred overnight. Drug concentration was determined with high-performance liquid chromatography [HPLC]. The drug content (in percent) was calculated using the following equations:

Drug content =  Drug weight in the nanoparticles Weight of the nanoparticles × 100

and

Loading eficiency =  Residual drug in the nanoparticles Initial feeding amount of drugs × 100 .

The characterization of nanoparticles was performed in DMSO-d₆ or D₂O using 500 MHz¹H nuclear magnetic resonance [NMR] spectroscopy (500 MHz superconducting FT-NMR spectrometer; Varian Unity-Inova 500; Agilent Technologies, Foster City, CA, USA). The morphology of nanoparticles was observed by transmission electron microscopy [TEM] using a JEM-2000 FX II microscope (JEOL, Tokyo, Japan). One drop of nanoparticle solution containing phosphotungstic acid (0.05% w/w) was placed onto a carbon film coated on a copper grid for TEM. Observation was done at an accelerating voltage of 80 kV. The particle size and zeta potential were measured by the Nano-ZS apparatus (Malvern Instruments, Malvern, UK). A sample solution prepared by dialysis was used to determine the particle size.

The release experiment was carried out in vitro. A sample solution prepared by dialysis was used directly. This solution was introduced into the dialysis membrane. Next, the dialysis membrane was placed in a 200-ml bottle with 100 ml of phosphate buffered saline [PBS] containing 1% (v/v) Tween 80 [PBST]. This bottle was placed in a shaking incubator with a stirring speed of 100 rpm at a temperature of 37°C. At specific times, the PBST was sampled for analysis of drug concentration. After each sampling, the entire volume of PBST was replaced with fresh PBST to prevent drug saturation. The concentration of the released sorafenib was determined by HPLC.

The Flexar HPLC system (PerkinElmer, Waltham, MA, USA) was equipped with a Solvent Manager 5-CH degasser, an autosampler, a quaternary LC pump, a column oven, and a UV-visible detector. Chromatography was performed on a guard column (SecurityGuard^® Guard Cartridge Kit; Phenomenex, Torrance, CA, USA) and on a C18 column (Brownlee C18^®, 5 μm, 150 × 4.6; PerkinElmer) at 37°C. Sorafenib was eluted isocratically with mobile phase (acetonitrile/methanol/1% acetic acid at a ratio of 35:38:27) at a flow rate of 1 ml/min and monitored at 254 nm. The chromatograms were recorded and integrated with the Chromera 2.1 system software (PerkinElmer).

HuCC-T1 cells maintained in RPMI 1640 (10% fetal bovine serum, 5% CO₂ at 37°C) were used to evaluate the antitumor activity of sorafenib-incorporated nanoparticles. Viability of tumor cells was evaluated by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide [MTT]-based cell proliferation assay. HuCC-T1 cells were seeded at a density of 2 × 10³ cells/well in 96-well plates with 100 μl of medium before addition of polymeric micelles. Next, free sorafenib, sorafenib-incorporated polymeric micelles, or empty polymeric micelles were added to 96-well plates at 100 μl. Controls were treated with 0.1% (v/v) of DMSO. Cells were incubated for 48 h, and cell viability was then measured in triplicate using an established MTT assay protocol.

Dextran is an appropriate macromolecule for block copolymerization because it has one reductive end. DexbLG copolymer was synthesized by the coupling of aminated dextran and PLGA as described previously (Figure 1b) 13. In the block copolymer, dextran acts as a hydrophilic domain and PLGA acts as a biodegradable hydrophobic domain. ¹H NMR revealed the theoretical [MW_T] and experimental [MW_E] MWs of DexbLG copolymer as 10,100 and 9,580 g/mol, respectively, while the MW of PLGA was 4,920 g/mol (M_w) and 4,780 g/mol (M_n) in gel permeation chromatography (data not shown). Because the DexbLG copolymer is amphiphilic, it may form core-shell type nanoparticles in an aqueous environment 2513. In this configuration, dextran comprises the hydrophilic outer shell, while PLGA comprises the hydrophobic core that is the actual reservoir of sorafenib.

Sorafenib (Figure 1a) is a relatively novel class of angiogenesis inhibitor, which was selected as an anticancer drug due to its poor aqueous solubility. Nanoparticles are regarded as an ideal candidate for these kinds of drugs because the abundant microvascular structure of the tumor tissue is a useful target for nanoparticulate drug delivery system via the EPR effect.

Nanoparticles of DexbLG copolymer were prepared by nanoprecipitation-dialysis method 13. Nanoprecipitation-dialysis method was superior over direct dialysis in minimizing particle size of nanoparticles (data not shown) 14. Empty nanoparticles of DexbLG had a spherical shape with a diameter < 100 nm (Figure 2a). As summarized in Table 1, their average particle size was 46 nm. Sorafenib-loaded DexbLG nanoparticles were increased in size, with their size being related to the quantity of drug loaded (Figure 2b,d). Among them, nanoparticles with a polymer/dry weight ratio of 40:2 and 40:5 showed a relatively uniform size distribution in contrast to 40:7 nanoparticles. Particle size analysis of sorafenib-incorporated DexbLG nanoparticles also showed an increased average particle size according to the contents of drug in the nanoparticles (Table 1). Zeta potential was not significantly changed according to drug incorporation. Drug contents were increased according to the feeding amount of drug (Table 1). Drug loading efficiency was lowest at 40:2, while 40:5 and 40:7 nanoparticles showed a loading efficiency > 35%.

To investigate the core-shell structure of nanoparticles and drug incorporation, ¹H NMR was adapted to measure nanoparticles in DMSO-d₆ or D₂O (Figure 3). Sorafenib displayed intrinsic peaks in its¹H NMR spectrum between 0.5 and 9.5 ppm (Figure 3a). When sorafenib-incorporated nanoparticles were reconstituted in D₂O, only dextran peaks between 2.5 and 5.0 ppm were observed, while specific sorafenib and PLGA peaks at 1.45, 3.35, and 4.9 ppm disappeared (Figure 3c). However, all the peaks of sorafenib, dextran domain, and PLGA domain were apparent when nanoparticles were dissolved in DMSO (Figure 3b). These results indicated that sorafenib was successfully entrapped into the PLGA core of the nanoparticles and that the dextran domain constituted the hydrophilic outer shell.

Sorafenib release study was performed in vitro. Sorafenib release rate from 40:5 nanoparticles into PBST was significantly low, with < 10% of the total incorporated drug being released over 2 weeks (Figure 4). Therefore, PBST was used to maintain a sink condition in subsequent experiments. The sorafenib release rate changed according to the drug contents; increased drug incorporation produces slower release rates (Figure 4b). Notably, an initial burst of release from 40:2 nanoparticles was observed until 6 h, followed by a sustained release of sorafenib for 2 weeks. Unexpectedly, 40:5 and 40:7 nanoparticles showed a very low drug release for 2 weeks. This phenomenon might be due to the hydrophobic interaction at higher drug contents in the core of nanoparticles 213. At higher drug contents, the hydrophobic drug might crystallize in the solid core of the nanoparticles, which would markedly hinder drug released 2. However, when the absolute amount of released drug was compared, 40:2 and 40:5 nanoparticles showed an almost equal amount of drug release until 96 h (Figure 4a).

Antitumor activity of sorafenib-incorporated DexbLG nanoparticles was tested using HuCC-T1 cells. Sorafenib itself showed a dose-dependent antiproliferative effect against tumor cells that was almost the same as that produced by sorafenib-incorporated DexbLG nanoparticles, while empty DexbLG nanoparticles did not significantly inhibit cell viability (Figure 5). These results indicate that DexbLG copolymer has no significant cytotoxicity against tumor cells, and sorafenib-incorporated DexbLG has at least similar antitumor activity in vitro. Nanoparticles or nanomatrix was known to improve bioavailability 11 and antitumor activity in vivo 12 with a decreased hemolysis or myelosuppression effect. Our results demonstrate that DexbLG nanoparticles are appropriate vehicles for sorafenib transport and release and that the nanoparticles are superior candidates for antitumor drug delivery.