Article Info

Funding: This study was financially supported by grants from the Natural Science Foundation of Zhejiang Province (Y2111250), the Key Science and Technology Project of Huzhou City (2011GG14) and the Key New Drug Discovery Project of 12th Five-Years Plan (2013ZX09102051), the Ministry of Science and Technology, China.

BACKGROUND: Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide. Most of the patients with HCC lose the surgical opportunity at the time of diagnosis. Some novel therapeutic modalities, like gene therapy, are promising for the treatment of HCC. However, the success of gene therapy depends on two aspects: efficient gene materials and gene delivery vectors. The present study was to develop new chitosan-based nanoparticles for a midkine-siRNA (anti-HCC gene drug) delivery.

METHODS: The novel gene delivery vector (MixNCH) was synthesized by hybrid-type modification of chitosan with 2-chloroethylamine hydrochloride and N, N-dimethyl-2-chloroethylamine hydrochloride. The chemical structure of MixNCH was characterized by FT-IR and ¹HNMR. The cytotoxicity of MixNCH was determined by MTS assay. The gene condensation ability and size, zeta potential and morphology of MixNCH/MK-siRNA nanoparticles were measured. The in vitro transfection and gene knockdown efficiency of midkine by MixNCH/MK-siRNA nanoparticles was detected by qRT-PCR and Western blotting. Gene knockdown effect at the molecule level on the proliferation of HepG2 in vitro was determined by MTS assay.

RESULTS: MixNCH was successfully acquired by aminoalkylation modification of chitosan. The MixNCH could condense MK-siRNA well above the weight ratio of 3. The average size of MixNCH/MK-siRNA nanoparticles was 100-200 nm, and the surface charge was about +5 mV. Morphologically, MixNCH/MK-siRNA nanoparticles were in regular spherical shape with no aggregation. Regarding to the in vitro transfection of nanoparticles, the MixNCH/MK-siRNA nanoparticles reduced MK mRNA level to 14.03%±4.03%, which were comparable to Biotrans (8.94%±3.77%). MixNCH/MK-siRNA effectively inhibited the proliferation of HepG2 in vitro.

Hepatocellular carcinoma (HCC), one of the most common tumors worldwide, has a high mortality rate, especially in China.^[1] Gene therapy has gained great attention in recent years. Gene therapy which is to introduce genetic materials into patients' cells or tissues has shown the potential for the treatment of HCC.^{[2, 3]} Meanwhile, the RNA interference (RNAi) technology has also developed rapidly after gene therapy over the past decade.^[4] During the RNAi, double stranded small interfering RNA (siRNA) degrades target message RNA (mRNA), subsequently inhibiting protein synthesis of the associated target gene.^[5] Midkine (MK), a heparin-biding growth cytokine, has been reported to be generally over-expressed in malignant tumors. MK exhibits several cancer-related activities, including fibrinolytic, anti-apoptotic, mitogenic, transforming, angiogenic and chemotatic functions.^[6] In our study, MK-siRNA was found to suppress the growth of HepG2 cells in vitro (CN200810059365.6), but the siRNA delivery still has such problems as rapid degradation and low intracellular uptake. An effective gene delivery vector is still essential to gene therapy.

Although the viral gene vector has been widely used in clinical trials, the non-viral gene vector attracts greater attention from researchers because of its bio-safety and easy modification.^[7] Among the non-viral gene vectors, chitosan has been widely studied because of its properties such as the cationic charge, biodegradability, biocompatibility, mucoadhesiveness, and permeability-enhancing feature.^[8]

Chitosan from chitin is the second most abundant natural polymer. With low toxicity and biocompatibility, chitosan has been recommended as a gene vector candidate.^[9] A lot of chitosan/gene nanoparticles have been created and their transfection efficiency has been tested in vitro and in vivo.^{[10, 11]} However, the application of unmodified chitosan has been significantly limited by its poor solubility in physiological condition due to its pKa value (about 6.3-6.4). Moreover, the transfection ability of unmodified chitosan is not enough to gain a therapeutic effect. Chitosan structure modification is an effective way to improve its stability in biological fluids and enhance cell delivery. Up to now, water-soluble chitosan derivatives have been synthesized to improve transfection efficiency of chitosan at physiological pH.^{[12, 13]}

We hypothesized that introduction of amino residues to chitosan could improve its solubility at physiological pH and its stability to form a complex with negatively charged siRNA, thereby enhancing the cellular association and gene silencing efficiency. The present study synthesized a novel chitosan derivative (MixNCH). We combined 2-chloroethylamine hydrochloride and N, N-dimethyl-2-chloroethylamine hydrochloride with chitosan and chararcterized its physicochemical properties. We investigated the gene transfection efficiency of MixNCH/MK-siRNA nanoparticles in HCC cell line and its inhibitory effect on HCC cell proliferation.

Chitosan (5-20 mPa.s, 0.5% in 0.5% acetic acid at 20?��) was purchased from TCI (Tokyo, Japan). 2-chloroethylamine hydrochloride and N, N-dimethyl-2-chloroethylamine hydrochloride were purchased from Sinopharm Chemical Reagents (Shanghai, China). siRNA transfection media were bought from Santa Cruze Biotechnology (Dallas, Texas, USA). MK-siRNA (sense: 5'-GGA UUG CGG CGU GGG UUU Ctt-3'; antisense: 5'-GAA ACC CAC GCC GCA AUC Ctt-3') was synthesized by Sangon Biotech. Biotrans was bought from Changzhou Bio-generating Biotechnologies Co., Ltd. Dulbecco's modified Eagle's medium (DMEM), OPTI-MEM and fetal bovine serum (FBS) were purchased from Life Tech (CA, USA). CellTiter one solution cell proliferation assay was purchased from Promega (Madison, WI, USA). RNAiso Plus, PrimeScript® RT reagent kit with gDNA Eraser (Perfect Real Time), and Premix Ex Taq™ (Probe qPCR) were purchased from TAKARA (Tokyo, Japan). Cell lysis buffer for Western blotting and Immunoprecipitation, bicinchoninic acid (BCA) Protein Assay Kit and ECL Plus Kit were purchased from Beyotime (Nantong, China). GAPDH antibody was purchased from Bioworld (MN, USA). MK antibody was bought from Abcam (Cambridge, UK).

Chitosan (500 mg) was added in one portion to a stirred solution of 2-chloroethylamine hydrochloride (8.7?g) in H₂O (25 mL). The suspension was heated at 50?��, and a solution of NaOH (3.0 g) in H₂O (25 mL) was then added dropwise. Stirring was continued at 50?�� for 5 hours until a clear solution appeared. Dialysis against purified water for 2 days followed by lyophilization afforded the desired product (666 mg) as a cottony solid.

Purebred type chitosan modified by 2-chloroethylamine as prepared above (285 mg) was dissolved in a stirred solution of N,?N-dimethyl-2-chloroethylamine hydrochloride (5.6 g) in H₂O (13 mL). The solution was heated at 50?��, and a solution of NaOH (1.56 g) in H₂O (13 mL) was then added dropwise. Stirring was continued at 50?�� for 4 hours. Dialysis against purified water for 2 days followed by lyophilization afforded the desired product (779 mg) as a cottony solid.

MixNCH was fully characterized by FT-IR and ¹HNMR. FT-IR spectra were recorded with a NICOLET NEXUS 670 FT-IR spectrometer. All samples were prepared as potassium bromide disks. ¹HNMR was measured with a DMX-400 (Bruker Company) NMR spectrometer. Samples (10 mg) were dissolved in D₂O (0.5 mL), and all experiments were done at 300 K.

The cytotoxicity of MixNCH, unmodified chitosan, and PEI prepared at different concentrations were determined by MTS assay. Human HCC cells (HepG2) were grown in 96-well plates at an initial density of 5000 cells/well in 0.1 mL of complete medium and incubated for 48 hours before replacement by 0.1 mL of fresh DMEM medium containing different concentrations of polymers to each well. They were transfected at the concentrations of 10, 50, 100, 200, and 300 µg/µL, respectively. Each concentration was replicated in 3 wells. After incubation for 24 hours, 20 µL of MTS solution was added in each well. After further incubation for 4 hours, the absorbance at 490 nm of each well was recorded with the DYNEX Technologies MRXII. Polymer-untreated cells in media were used as controls.

Electrophoresis was performed to confirm the MK-siRNA condensation ability of the MixNCH. MixNCH/MK-siRNA nanoparticles were induced at various w/w ratios from 1 to 5, and the final volume with 10× agarose gel loading dye mixture was 10 µL. The complexes were loaded onto the 4% agarose gels with ethidium bromide (0.1 µg/mL) and assessed with Tris-Borate buffer at 100 V for 40 minutes. siRNA retardation was revealed under the UV light.

The mean particle size, zeta potential and polydispersity index (PDI) of MixNCH/MK-siRNA nanoparticles were measured by Zetasizer Nano ZS 90 (Malvern Instruments, Worcestershine, UK) after diluting appropriately with ddH₂O. Each sample was measured for three times, and data taken as the mean of three measurements.

MixNCH/MK-siRNA nanoparticles were analyzed with a transmission electron microscope to estimate their morphology. Nanoparticles were observed by negative staining with uranyl acetate. The prepared sample was observed under a JEM-2010 transmission electron microscope (Jeol, Tokyo, Japan) at an accelerating voltage of 80 kV.

For transfection in vitro, HepG2 was seeded in 12-well plates at an initial density of 10⁵ cells/well and incubated for 24 hours after transfection. To measure transfection efficiency, 30 pmol MK-siRNA and MixNCH at different weight ratio were vortexed together. They were incubated for 30 minutes at room temperature. The cells were washed once with the siRNA transfection medium (sc-36868). For each transfection, 0.8 mL siRNA transfection medium was added to each tube containing 0.2 mL MixNCH/MK-siRNA nanoparticles. They were mixed gently and overlaid onto the washed cells. They were incubated for 4 hours at 37?��. Four hours later, 1 mL DMEM showed doubled normal serum concentration without removal of nanoparticles. The cells were further incubated for 48 hours at 37?��.

Gene silencing of MK was measured at the mRNA level using the ABI 7500 real-time PCR system. Total RNA was extracted using RNAiso Plus (TAKARA). After RT, the cDNA was amplified with real-time PCR according to the protocols from Premix Ex Taq™ kit (TAKARA). The probes and primers were synthesized by Life Tech with the following sequences, for MK forward primer: 5'-GAC CAT CCG CGT CAC CA-3'; Reverse primer: 5'-TCC AGG CTT GGC GTC TAC CA-3'; Taqman probe: 5'-FAM-CAA AGG CCA AAG CCA AGA AAG GGA AG-TAMRA-3'. For endogenous control glyceraldehydes-3-phosphate dehydrogenase (GAPDH), Forward primer: 5'-GCC AGC CGA GCC ACA T-3'; Reverse primer: 5'-CTT TAC CAG AGT TAA AAG CAG CCC-3'; and Taqman probe: 5'-FAM-CCA AAT CCG TTG ACT CCGAC CTT CA-TAMRA-3'. All samples were analyzed in duplicate, and the values of sample copies were obtained after quantitative amplification and normalized to GAPDH using the 2^-??Ct method. Water was used as negative and quality controls, and each sample was measured in triplicate.

After 48-hour transfection, HepG2 cells were washed twice with ice-cold PBS, total proteins were extracted with cell lysis buffer for Western blotting and IP (Beyotime). Protein concentrations of the lysates were measured with a BCA protein assay kit (Beyotime). Thirty microgram proteins were loaded in each lane to the 12% SDS-PAGE gel and were transferred to a polyvinylidene fluoride membrane (Millipore) followed by blocking in 5% skim milk in Tris-buffered saline with Tween for 2 hours. Then, the membrane was incubated with rabbit anti-MK monoclonal antibody diluted in a 1:500 (v/v) ratio overnight at 4?�� and peroxidase-conjugated goat anti-rabbit IgG diluted in a 1:5000 (v/v) ratio for 1 hour at room temperature. As internal standard for the proteins, the other membrane was also incubated with a 1:5000 (v/v) dilution of rabbit anti-GAPDH polyclonal antibody and a 1:5000 (v/v) dilution of peroxidase-conjugated goat anti-rabbit IgG. The membranes were exposed to Kodark X-Omat film using the BeyoECL Plus kit (Beyotime) and the densitometric intensities of protein bands were determined.

HepG2 cells were seeded in a 96-well plate at a density of 5000 cells per well in 100 µL DMEM medium containing 10% FBS and incubated for 24 hours. Before transfection, the cells were washed once with the siRNA transfection medium (sc-36868). Then, different amounts of MixNCH/MK-siRNA nanoparticles [5.0, 7.5, 10.0, 12.5, 15.0 pmol/well, MixNCH:MK-siRNA (w/w)= 160:1] were added into each well and allowed to incubate for 48 hours. The final volume of each well should be ensured about 100 µL, and the lack of volume was complemented with the siRNA transfection medium. Each concentration was replicated in 5 wells. After incubation for 48 hours, 20 µL of MTS solution was added into each well. After further incubation for 4 hours, the absorbance at 490 nm of each well was recorded with the DYNEX Technologies MRXII.

MixNCH was synthesized in two steps (Fig. 1). First, the purebred type chitosan was modified by 2-chloroethylamine. The ¹HNMR spectrum of intermediate product of MixNCH was detected (Fig. 2). The peaks appeared at around 4.4-4.7 ppm for anomeric protons and around 3.4-4.0 ppm for other protons on the carbohydrate ring in chitosan. As evidence of the reaction, complex signals were found between 2.5-3.3 ppm for introduced methylene groups. A minor peak was also observed at 2.0 ppm for the methyl groups due to incomplete deacetylation. FT-IR was further used to verify the chemical structure of the intermediate product (Fig. 3). In the FT-IR spectrum of the intermediate product, the adsorptions at 3700-2600/cm, particularly the C-H stretching vibration peaks at 3072 and 2921/cm, were stronger and broader than commercially available chitosan, indicating the introduction of CH₂-CH₂ side chains and N-H functionalities. It should be also noted that the C-O stretching vibration peaks of the secondary alcohol in the modified chitosan at 1069/cm is significantly stronger than that corresponding to primary alcohol at 1025/cm, whereas the difference in the IR spectrum of chitosan is small. This result indicated that alkylation mainly occurred in those primary hydroxyls at C-6. In addition, the IR spectrum also showed the introduction of various amino groups by a broader band at around 566/cm, which is the out-plane flexural vibration of N-H bonds.

Second, the intermediate product was further modified by N, N-dimethyl-2-chloroethylamine hydrochloride. In the FT-IR spectrum of MixNCH (Fig. 3), the C-H stretching vibration peaks at 3010 and 2964/cm were stronger and broader than the intermediate product, indicating the introduction of more CH₂-CH₂ and CH₃ side chains. It should also be noted that the C-O stretching vibration peaks of the secondary alcohol in the modified chitosan at 1066/cm are not so much stronger than those corresponding to primary alcohol at 1028/cm, whereas the difference is greater for the derivative modified by 2-chloroethylamine. This result revealed that second-stage alkylation modification occurred in the secondary hydroxyls at C-3 in addition to the amino coat. The asymetric stretching vibration of C-O-C at 1107/cm revealed the increasing number of ether groups. Moreover, the ¹HNMR spectrum of MixNCH in D₂O at 300 K showed complex signals between 4.70 and 2.75 ppm, wherein the protons on the carbohydrate ring and introduced alkyl groups overlapped. It is very difficult to assign these signals. A minor peak was also observed at 2.0 ppm in the methyl groups because of incomplete deacetylation (Fig. 2).

Fig. 4 shows the viability of HepG2 cells assessed by the MTS assay after incubation for 48 hours with MixNCH, unmodified chitosan, and PEI (Mn=25 000) at various concentrations. PEI has been widely used as a DNA condensing agent and a transfection vector and is the standard to which new polymeric vectors are often compared. Unfortunately, PEI is associated with significant level of cytotoxicity. In our study, fewer than 50% of cells treated with PEI remained viable at a polymer concentration of 0.01 µg/µL, and cell viability was as low as 13% at a high concentration of PEI under otherwise identical conditions. MixNCH and unmodified chitosan exhibited improved biocompatibility as compared with PEI. As shown in Fig. 4, cells incubated with unmodified chitosan remained nearly non-toxic relative to controls. MixNCH also showed good biocompatibility with the cell viability of over 85%, and a dose-dependent cytotoxicity was observed with increasing concentrations of polymers. Thus, the results suggest that MixNCH is a good candidate as gene delivery vector for further study.

MixNCH/MK-siRNA nanoparticles were prepared by simple mixing of MixNCH and MK-siRNA solution through electrostatic interactions between amine groups of MixNCH and phosphate groups of siRNA. Some simple physicochemical properties of the nanoparticles such as condensation ability, size, zeta potential and morphology were investigated.

A successful gene vector should effectively condense and negatively charge siRNA into nanosized particles. To confirm the formation of the MixNCH/MK-siRNA complex, we checked the retardation of siRNA mobility by agarose gel electrophoresis. Compared with the naked MK-siRNA control, MixNCH showed the ability of MK-siRNA condensation as less MK-siRNA ran into the gel. MixNCH could efficiently package MK-siRNA at the weight ratio of 3 and above, as no mobility of MK-siRNA was observed at these ratios (Fig. 5).

The size and surface charge of nanoparticles are important issues which can significantly affect cell-surface approach, uptake, and their transfection efficiency. The particle sizes of MixNCH/MK-siRNA nanoparticles at various w/w ratios from 40 to 160 were measured by dynamic light scattering (Table). They ranged from 166.8 nm to 124.3 nm with a constant decrease followed by an increased w/w ratio. The zeta potential of all the nanoparticles were around +5 mV. These surface charges increased slightly with an increased w/w ratio of MixNCH to siRNA.

The MK gene was targeted by MK-siRNA using MixNCH nanoparticles and Biotrans. The knockdown efficiency was determined both at the mRNA and protein level. The mRNA expression was evaluated by qRT-PCR (Fig. 7A). Compared to no-treatment cells (100%), naked MK-siRNA reduced mRNA levels of MK to 96.23%±0.31%, MK-siRNA delivered by MixNCH to 14.03%±4.03%-45.37%±2.23% and Biotrans to 8.94%±3.77%. Based on the statistical analysis, mRNA expression of MK-siRNA delivered by both MixNCH and Biotrans nanoparticles was significantly reduced when compared to the naked MK-siRNA group (P<0.001). There was no difference in the MK mRNA level shown by MK-siRNA delivered by MixNCH (160:1) nanoparticles comparable to Biotrans (P=0.194). Moreover, in parallel with mRNA expression, the MixNCH/MK-siRNA (160:1) nanoparticles and Biotrans/MK-siRNA nanoparticles provided the great inhibitory effect on protein level as detected by Western blotting (Fig. 7B).

MTS assay was also used to determine the effect of MK knockdown by MK-siRNA on HepG2 cell growth. Five groups (5.0, 7.5, 10.0, 12.5, 15.0 pmol/well) were set for investigation. Compared to the control group, naked MK-siRNA showed no inhibition on the HepG2 cell growth. In contrast, when compared to the control, MixNCH/MK-siRNA nanoparticles significantly suppressed the HepG2 cell growth and this inhibition was dose-dependent (Fig. 8).

HCC, one of the primary hepatic neoplasmas, is considered the sixth most common cancer type with the highest prevalence worldwide.^[14] The current treatment, such as surgical resection, liver transplantation, and chemotherapy, have limited cure effect.^[15] Gene therapy is a potential treatment for patients with HCC.^[16]

MK is over-expressed in many human carcinomas including HCC, and believed to contribute to tumorigenesis and tumor progression.^[17] Therefore, the inhibition of the synthesis or action of MK contributes to cancer therapy. In our laboratory, we screened one MK-siRNA (CN200810059365.6) that could inhibit the growth of HepG2 cells. At that time, the in vitro transfection reagents we used was the commercial product, Oligofectamine 2000, which was not suitable for in vivo application because of its toxicity.

Chitosan is a biocompatible, biodegradable, and mucodhensive cationic polysaccharide with minimum immunogenicity and low cytotoxicity, which is most extensively studied for siRNA delivery.^[18] Thus, we decided to adopt chitosan derivative as the safer delivery tool which may have great potential in clinical application. Chitosan has a primary amine group with a pKa value of approximately 6.5 and is therefore only weakly charged and barely soluble at neutral pH or above. If modifications of chitosan could improve the solubility at physiological pH and increase the positive surface charge, the new gene delivery vector would obviously enhance the transfection ability. The novel chitosan derivative we designed was based on this concept. After the introduction of more amino groups into chitosan, the new chitosan derivative exhibited better results in the transfection efficiency. The introduction of various kinds of amino groups into the chitosan has three advantages. First, the increased chitosan solubility can enhance the colloidal stability of chitosan/siRNA nanoparticles for a longer circulating time.^[19] Second, the increased positive surface charge allows it to interact with negatively charged siRNA to form nanoparticles, and enhance the cellular transport through cell membrane.^[20] Third, the increase of amino groups could enhance the role of proton sponge, and help the dissociation of MixNCH and MK-siRNA when in the cytoplasm.^[21]

Cytotoxicity is a major hurdle for clinical feasibility of polycationic gene vectors. A perfect gene vector should have a low toxicity with minimum effect on cell growth or viability, and the MTS assay is generally used as an initial indicator of cytotoxicity. The cytotoxicity of MixNCH on HepG2 cells increased as the concentration of MixNCH rose. MixNCH as a chitosan's derivative is known for its good bio-safety, and our study showed that its enhanced cationic property affects the cell viability as compared to unmodified chitosan. Unlike the widely used cationic gene vector PEI, MixNCH still exhibited better biocompatibility.

Like most cationic-polycations used to package siRNA, MixNCH/MK-siRNA nanoparticles are prepared by fast addition and mixing of MixNCH to siRNA solution. The gel retardation assay is a method to qualitatively assess the ability of MixNCH to condense siRNA. The binding affinity of MixNCH for siRNA determines its ability to protect the siRNA as well as nanoparticles stability, which in turn greatly determines nanoparticles transfection efficiency. When the weight ratio of MixNCH and MK-siRNA was 3 and above, the MixNCH could condense all of the MK-siRNA into nanopaticles. Koping-Hoggard showed that formulations that are very stable are much less efficient at transfecting cells than their less stable counterparts.^[22]

The particle size of nanoparticles is a critical factor that determines cellular uptake rate. The DLS analysis showed that the nanoparticles in the present study were 100-200 nm. Generally, smaller particles are more efficient in transfecting cells,^[23] and our nanoparticles with this kind also showed efficient transfection into HepG2 cells. The surface charge density of MixNCH/siRNA nanoparticles is also an important parameter to affect the colloidal stability of nanoparticles. The zeta potential of MixNCH/MK-siRNA nanoparticles (w/w from 40 to 160) was 3.5 to 4.3 mV in a slight increase, which was effective for dissociation of MixNCH and siRNA when the nanoparticles entered into the cytoplasm.

On the basis of characterization of MixNCH/MK-siRNA nanoparticles, we evaluated their transfection efficiency in HepG2 cells in the present study. The experiments showed that MK-siRNA delivered by MixNCH nanoparticles could significantly reduce both mRNA and protein levels of MK, leading to a strong suppression of cell proliferation in HepG2 cells. Thus, MixNCH is shown as an efficient and safe vector in siRNA delivery. Moreover, the success of gene silencing activities by MixNCH/MK-siRNA nanoparticles should represent an effective and low toxic approach for the treatment of HCC.