Asialoglycoprotein receptor-magnetic dual targeting nanoparticles for delivery of RASSF1A to hepatocellular carcinoma

We developed a nanovector with double targeting properties for efficiently delivering the tumor suppressor gene RASSF1A specifically into hepatocellular carcinoma (HCC) cells by preparing galactosylated-carboxymethyl chitosan-magnetic iron oxide nanoparticles (Gal-CMCS-Fe3O4-NPs). After conjugating galactose and CMCS to the surface of Fe3O4-NPs, we observed that Gal-CMCS-Fe3O4-NPs were round with a relatively stable zeta potential of +6.5 mV and an mean hydrodynamic size of 40.1 ± 5.3 nm. Gal-CMCS-Fe3O4-NPs had strong DNA condensing power in pH 7 solution and were largely nontoxic. In vitro experiments demonstrated that Gal-CMCS-Fe3O4-NPs were highly selective for HCC cells and liver cells. In vivo experiments showed the specific accumulation of Gal-CMCS-Fe3O4-NPs in HCC tissue, especially with the aid of an external magnetic field. Nude mice with orthotopically transplanted HCC received an intravenous injection of the Gal-CMCS-Fe3O4-NPs/pcDNA3.1(+)RASSF1A compound and intraperitoneal injection of mitomycin and had an external magnetic field applied to the tumor area. These mice had the smallest tumors, largest percentage of TUNEL-positive cells, and highest caspase-3 expression levels in tumor tissue compared to other groups of treated mice. These results suggest the potential application of Gal-CMCS-Fe3O4-NPs for RASSF1A gene delivery for the treatment of HCC.


Physical and chemical analysis of Gal-CMCS-Fe 3 O 4 -NPs.
depicts the synthesis of Gal-CMCS-Fe 3 O 4 -NPs. Infrared spectrum analysis revealed that the primary absorption peak of Fe 3 O 4 -NPs was attributed to the vibration of Fe-O. The series of absorption peaks of CMCS-Fe 3 O 4 -NPs included the stretch vibration absorption peak of -NH 2 and -OH at 3423 cm −1 , the anti-symmetric vibration peak of -COO and the symmetric vibration absorption peak of -COO-at 1604 cm −1 and 1453 cm −1 , respectively, and the shoulder peaks of the sugar ring at 1091 cm −1 and 1219 cm −1 . The strengthening and widening of the absorption peak at 1601 cm −1 in Gal-CMCS-Fe 3 O 4 -NPs indicated that -NH 2 had linked with related groups. The stretching vibration of the C-N key at 1123 cm −1 also indicated the introduction of galactosyl, whereas the widening of peaks at 3462 cm −1 and 1045 cm −1 indicated that the introduction of galactosyl resulted in an increase in hydroxyl (Fig. 1B). The thermal gravimetric curve for Gal-CMCS-Fe 3 O 4 -NPs showed three weight loss events as follows: the first from the room temperature to 100 °C, which may be due to evaporation of water on the surface of Gal-CMCS-Fe 3 O 4 -NPs, the second from 100° to 200 °C, which can be attributed to the decomposition of galactose, and the third from 200° to 350 °C (Fig. 1C), which can be attributed to the decomposition of CMCS.
To evaluate the binding of galactose moieties on Gal-CMCS-Fe 3 O 4 -NPs to galactose-recognizing lectins, the aggregation of Gal-CMCS-Fe 3 O 4 -NPs induced by ricinus communis agglutinin I (RCA120) was measured by changes in turbidity over time. When RCA120 was added, Gal-CMCS-Fe 3 O 4 -NPs had a greater absorbance than CMCS-Fe 3 O 4 -NPs. Consistent with characteristics of RCA120 28 , when excess of a galactose competitive antagonist was added, Gal-CMCS-Fe 3 O 4 -NPs decomposed, and the absorbencies of Gal-CMCS-Fe 3 O 4 -NPs and CMCS-Fe 3 O 4 -NPs became similar (Fig. 1D).
Transmission electron microscope (TEM) showed that Gal-CMCS-Fe 3 O 4 -NPs had a relatively uniform round shape ( Fig. 2A). The average primary particle diameter of Gal-CMCS-Fe 3 O 4 -NPs was 20.0 ± 2.5 nm. Use of a magnet showed that Gal-CMCS-Fe 3 O 4 -NPs had good magnetic responsiveness (Fig. 2B,C). The saturation magnetization for Gal-CMCS-Fe 3 O 4 -NPs at room temperature was 38.23 emu/g using Lake Shore 7407 vibrating sample magnetometer. The mean zeta potential and hydrodynamic size of Gal-CMCS-Fe 3 O 4 -NPs in water were measured using Nicomp 380 ZLS and were found to be + 6.5 mV and 40.1 ± 5.3 nm (Fig. 2D), respectively. At pH 7, the hydrodynamic size (Fig. 2E) and zeta potential (Fig. 2F) of Gal-CMCS-Fe 3 O 4 -NPs were relatively stable across 5 days of observation in water. There was no statistically significant difference between water and Dulbecco's modified Eagle's medium (DMEM) with relevant to hydrodynamic size and zeta potential of Gal-CMCS-Fe 3 O 4 -NPs (P > 0.05).
Hemolysis assay and toxicity assessment. Hemolysis of the Gal-CMCS-Fe 3 O 4 -NPs was investigated for its hemocompatibility as shown in Fig. 3A. The degree of hemolysis of all the tested Gal-CMCS-Fe 3 O 4 -NP samples at different concentrations were below 2%.
An in vitro toxicity test showed that when exposed to a concentration of 200 μg/ml Gal-CMCS-Fe 3 O 4 -NPs, the viability of L02 cells was over 95% (Fig. 3B). As Gal-CMCS-Fe 3 O 4 -NP concentration approached 500 μg/ml, cell viability decreased but remained high at 80%. Therefore, a concentration of 200 μg/ml Gal-CMCS-Fe 3 O 4 -NPs was chosen for subsequent experiments.
Gal-CMCS-Fe 3 O 4 -NPs were injected into the tail vein of nude mice and serum was collected to assess liver function 1, 2, 3, 7, or 14 days later. Gal-CMCS-Fe 3 O 4 -NPs induced transient toxicity, as ALT, AST, and T-BIL Scientific RepoRts | 6:22149 | DOI: 10.1038/srep22149 levels were higher than in the control group on day 1 and 2 (P < 0.05) but returned to normal by day 3 (P > 0.05; Fig. 3C-E). After Gal-CMCS-Fe 3 O 4 -NPs injection, mice showed no signs of acute toxic reaction, discomfort, or fatigue and slowly gained weight over the 14-day observation period. After hematoxylin and eosin staining of paraffin-embedded sections, optical microscopy revealed no obvious differences in the morphology of primary organs between the normal saline (NS) and Gal-CMCS-Fe 3 O 4 -NPs groups (Fig. 3F).
Characterization of Gal-CMCS-Fe 3 O 4 -NPs/DNA complexes. Under a pH of 5, 7, or 9, the swimming speed of Gal-CMCS-Fe 3 O 4 -NP/DNA was lower than that of the plasmid-only group (Fig. 4A). In addition, a small amount of free DNA was observed under the three different pH conditions. However, the lowest level of free DNA was associated with a pH of 7 (close to the pH of the human body). At a pH of 7, as the concentration of Gal-CMCS-Fe 3 O 4 -NPs increased, the retention of DNA also increased (Fig. 4B). When the NPs/DNA mass ratio was 3:1, Gal-CMCS-Fe 3 O 4 -NPs retained all DNA. Therefore, a mass ratio of 3:1 was used in subsequent experiments. Electrophoresis of Gal-CMCS-Fe 3 O 4 -NPs/DNA after treatment with the digestive enzyme DNase I showed that DNA coated with NPs had no obvious fragments, whereas DNA fragments were observed with uncoated DNA (Fig. 4C). Figure 5A depicts the schematic diagram of the entry of Gal-CMCS-Fe 3 O 4 -NPs inside the nucleus of cell. To investigate the targeting specificity of Gal-CMCS-Fe 3 O 4 -NPs for HCC cells, NPs were used to transfect plasmids into HepG2, L02, GES-1, U87, and SPCA-1 cell lines (Fig. 5B). Seventy-two hours after transfection, strong green fluorescence was observed in liver cells (L02 and HepG2), whereas weaker fluorescence was observed in non-liver cells (GES-1, U87 and SPCA-1). Flow cytometry showed that the average transfection efficiency of pcDNA6.2mir-EGFP in L02 and HepG2 cells was 39.12 ± 2.56% and 35.23 ± 2.33%, respectively (P > 0.05), whereas transfection efficiency in SPCA-1, GES-1, and U87 cells was only 18.01 ± 1.97%, 18.89 ± 1.86%, and 16.99 ± 1.64%, respectively. This difference in Gal-CMCS-Fe 3 O 4 -NPs transfection efficiency between liver and non-liver cells was statistically significant (P < 0.01). Additionally, approximately 54.55 ± 4.27% of HepG2 cells were transfected with the aid of an external magnetic field, whereas 35.23 ± 2.33% of HepG2 cells were transfected without an external magnetic field (P < 0.05). The addition of galactose decreased the transfection efficiency of Gal-CMCS-Fe 3 O 4 -NPs in HepG2 cells from 35.23 ± 2.33% to 18.93 ± 1.96% (P < 0.05; Fig. 6). However, the transfection efficiency of CMCS-Fe 3 O 4 -NPs was similar with or without the addition of galactose (P > 0.05; Fig. 6).

Targeted transfection of HCC tissue by Gal-CMCS-Fe 3 O 4 -NPs in vivo.
After removing subcutaneous tumors composed of HepG2 cells from nude mice, orthotopic transplantation of the tumors under capsula fibrosa was performed. Two weeks later, Gal-CMCS-Fe 3 O 4 -NP/pcDNA6.2mir-EGFP compound was injected into the tail vein of mice. After 3 days, mice were sacrificed and livers, kidneys, spleen, heart, lungs, and orthotopically transplanted HCC tissue were removed (Fig. 7A). We observed green fluorescence in liver and HCC tissue sections. The average transfection efficiency of pcDNA6.2mir-EGFP in liver tissue was 32.6%, Furthermore, the average transfection efficiency was approximately 40.8% in HCC tissue with the aid of an external magnetic field, and 29.7% in HCC tissue without an external magnetic field (P < 0.01; Fig. 7B,C). No obvious fluorescence was observed in kidney, spleen, heart, or lung tissue sections (Fig. 7B).

Efficient delivery of the RASSF1A gene for HCC treatment by Gal-CMCS-Fe 3 O 4 -NPs combined with MMC.
Two weeks after the orthotopic HCC transplantation model mice received treatment, tumor volumes and weights were lower in treated mice than in control mice (group a, P < 0.01; group b, P < 0.01; group c, P < 0.05; group d, P < 0.05; Fig. 8A-C). Among the treatment groups, intravenous injection of RASSF1A-NPs and intraperitoneal injection of MMC with the aid of an external magnetic field (group a) inhibited tumor growth the most. The average percent of terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) positive cells in the four treatment groups and control group were 40.5%, 29.7%, 0.8%, 11.2%, and 0.5%, respectively (Fig. 8D). RASSF1A expression in tumor tissue was higher in groups a and c than in group b, whereas RASSF1A expression was not observed in group d or in the control group (group e; Fig. 8E). Caspase-3 expression in tumor tissue was higher in groups a, b, and d than in group c or the control group. There were no differences between groups in p53, p21, bcl-2, or bax expression.

Discussion
The purpose of targeted gene therapy is to assemble the desired genes with a suitable carrier that can target specific tissues and enable effective gene expression 29 . The design and development of NPs with high transfection efficiency and low cytotoxicity are critical for successful gene therapy 30 . Super paramagetic iron oxide NPs targeted to specific cells for magnetic resonance imaging, tissue repair, targeted drug delivery, and hyperthermia with a large number of polycations, including chitosan, polyethylenamine, polyamidoamine, and polyamines have been receiving considerable attention 31 . It is necessary to introduce functional ligands such as galactose, folic acid, epithelial cell adhesion molecule, and α -fetoprotein that can actively interact with the corresponding binding sites on the cell surfaces of HCC to further improve the binding of ligands to specific receptor targets 32 . However, the common challenge among these applications is to ensure sufficient uptake of NPs by HCC cells. Furthermore, the potential toxic effects of these NPs in vivo also remain unclear 33 .
Here, with the aim of enhancing targeted HCC gene therapy, we constructed Gal-CMCS-Fe 3 O 4 -NPs that could be used for transfection in vivo and in vitro, were safe and efficient, and could be used with an external magnetic field to target the liver. Examination with a laser particle size analyzer showed that vector particles had a diameter of approximately 40.1 nm, which is beneficial for a HCC-targeted gene carrier 34,35 . ASGP-R-mediated endocytosis of galactose-modified delivery systems is influenced by the size of NPs 36 , with NPs less than 50 nm in diameter efficiently targeting hepatocytes and NPs over 140 nm in diameter being more selective for Kupffer cells. Therefore, the Gal-CMCS-Fe 3 O 4 -NPs prepared in this study should be absorbed by HCC cells 37 .
We further investigated the chemical and structural properties of Gal-CMCS-Fe 3 O 4 -NPs using infrared spectrum and thermogravimetric analyses.  The hydrodynamic size of Gal-CMCS-Fe 3 O 4 -NPs is composed of three parts namely Fe 3 O 4 -NPs primary size, polymer-coated, and hydration layer thickness. The primary size of nuclear magnetic particle is obtained using TME. Hence, the hydrodynamic size (40.1 ± 5.3 nm) is bigger than the primary size (20.0 nm ± 2.5 nm) in our experimental results. Further experiments showed that Gal-CMCS-Fe 3 O 4 -NPs had good magnetic responsiveness in a magnetic field and exhibited strong DNA-binding capabilities in both acidic and alkaline environments, with the strongest binding force at pH 7, which is close to that of the human body. Moreover, the zeta potential which indicated Gal-CMCS-Fe 3 O 4 -NPs could combine with electronegative DNA 38 was stable across 5 days of observation in water and cell culture media containing DMEM at room temperature. These results suggest that Gal-CMCS-Fe 3 O 4 -NPs are stable at physiological pH, which allows for high transfection efficiencies of Gal-CMCS-Fe 3 O 4 -NPs in vivo 39 . Gel electrophoresis and DNA precipitation experiments at different mass ratios showed that the best mass ratio for NPs/DNA binding was 3:1, at which the binding rate of DNA reached 95% (data not shown). Through digestion with DNase I in vitro, we observed that Gal-CMCS-Fe 3 O 4 -NPs had an excellent protective effect on DNA. The hemolysis of Gal-CMCS-Fe 3 O 4 -NPs was below 2% . It was reported that up to 5% hemolysis is permissible for biomaterials 30    Although the diagnosis and treatment of HCC has greatly improved over the past two decades, transarterial chemoembolization or chemotherapy still plays an important role in its treatment 42 . Because most HCC patients have a medical history of posthepatitic cirrhosis and hepatic insufficiency, the appropriate dose, intensity, and mode of chemotherapy is difficult to determine 43 . Moreover, the lack of tumor suppressor gene expression in HCC cells can lead to defects in apoptosis-related signal transduction pathways and promote tolerance to chemotherapy 44,45 , thus limiting the efficacy of chemotherapy for HCC. Re-expressing an inactive tumor suppressor gene through a transgene vector can restore apoptosis in tumor cells, providing a new strategy for enhancing the sensitivity of HCC cells to chemotherapy. This would allow for reduced dosage of chemotherapy drugs and reduced toxic side effects 13 .
In this study, we successfully validated the antitumor efficacy of Gal-CMCS-Fe 3 O 4 -NPs/RASSF1A compound by observing the expression of RASSF1A protein in orthotopically transplanted HCC tissue and the inhibition of tumor growth in mice. Furthermore, the presence of an external magnetic field increased RASSF1A expression, slowed the growth of tumors, and enhanced the anti-proliferative effect of Gal-CMCS-Fe 3 O 4 -NPs/RASSF1A compound. Moreover, the increased expression of RASSF1A was associated with greater MMC-induced apoptosis of HCC cells, indicating that by inducing the apoptosis of targeted cells, the expression of RASSF1A can enhance the sensitivity of HCC cells to chemotherapy. To explore the mechanism by which RASSF1A regulates apoptosis, we used western blot analysis to measure changes in caspase-3, p53, p21, and bax levels in HCC tissue. The level of activated caspase3 in HCC tissue increased as the expression of RASSF1A increased. Therefore, in addition to the effect of MMC on HCC cells, the RASSF1A gene may activate caspase-3 through specific signaling pathways, thereby promoting apoptosis of HCC cells and increasing the sensitivity of HCC cells to chemotherapy 46 .

Synthesis of Gal-CMCS-Fe 3 O 4 -NPs.
Under the protection of nitrogen, FeCl 3 (10.8 g, 0.067 mol) and FeCl 2 (4 g, 0.031 mol) were added to 50 ml HCl (1.1 mol/l) and filtered through a 0.22-μm filter to remove bacterium. The solution (25 ml) was quickly poured into 250 ml NaOH (1.5 mol/l) and agitated for 1 h at 80 °C. The combined solution was then poured into a 500-ml beaker attached to a permanent magnet. Supernatants were discarded after the black material had completely precipitated. Double-distilled water was used for flushing until sedimentation no longer occurred. When the pH was approximately 8, Fe 3 O 4 -NPs were isolatd following cooling and drying.
The CMCS-Fe 3 O 4 -NPs was prepared in accordance with the literature with minor modification 48 . Briefly, under the protection of nitrogen. The separated Fe 3 O 4 -NPs (140 mg, 0.6 mmol) were re-suspended in 40 ml PBS with 120 mg EDC (1-ethyl-3-(3-dimethylaminopropyl) and 120 mg NHS (N-hydroxysuccinimide, Pierce, Rockford, USA), then 280 mg CMCS was added immediately. The solution was dispersed for 2 h at room temperature with ultrasonic waves. A permanent magnet was used to isolate the magnetic compound that was subsequently washed twice with with ethanol. Double-distilled water was added to a constant volume of 60 ml, and a colloid solution of CMCS-Fe 3 O 4 -NPs was obtained and evenly dispersed with ultrasonic waves at 37 °C. Lactose (336 mg, 3.7 mmol) and sodium cyanoborohydride (168 mg, 2.7 mmol) were slowly added, and the solution was agitated for 1 h at 37 °C. The magnetic compound was isolated with a permanent magnet, washed twice with ethanol (30 ml), freeze-dried in a vacuum, and preserved for later use. KBr as a diluting agent and scanned against a blank KBr pellet background. A thermogravimetric analyzer (Shimadzu TGA-50 Analyzer, Tokyo, Japan) was used to perform thermal analyses. The saturation magnetization for Gal-CMCS-Fe 3 O 4 -NPs was done using Lake Shore 7407 vibrating sample magnetometer (Lake Shore Cryotronics, Westerville, OH, USA).
Agarose gel retardation assay. The reporter pcDNA6.2mir-EGFP plasmid was purified using an EndoFree Plasmid Mega Kit (Qiagen Co. Ltd., Shanghai, China) according to the manufacturer's instructions. Gal-CMCS-Fe 3 O 4 -NPs and pcDNA6.2mir-EGFP were mixed at mass ratios of 3:1 in a 50-μl reaction system with pH values adjusted to 5, 7, or 9, or at mass ratios of 0.5:1, 1:1, 2:1, 3:1, or 4:1 with the pH adjusted to 7. After 1 h at room temperature, the reaction products were removed for electrophoresis on 0.5% agarose gel at 80 V for 2 h. Gels were imaged using a gel imaging system. At a pH of 7, Gal-CMCS-Fe 3 O 4 -NPs and pcDNA6.2mir-EGFP plasmid were mixed at the best mass mixture ratio. After 1 h at room temperature, 0.5 U DNase I or fresh mouse serum was added to the mixtures in a water bath for 1 h at 37 °C. The compounds were separated on 0.5% agarose gels.
Hemolysis assay. Hemolysis of red blood cells (RBCs) was examined as previously described 30 . Briefly, 1.5 mL of fresh rat RBCs were harvested by centrifuging at 1500 rpm for 10 mins. The resultant RBC suspension was washed three times with NS. Finally, the RBCs were resuspended in NS to a concentration of 2% (v:v). Then, 0.7 ml of diluted 2% RBC suspensions were added to varying concentrations of 0.1 mL of Gal-CMCS-Fe3O4-NPs solutions in NS (25,50,100,150,200,250, and 300 μg/ml). The resultant mixtures were incubated at 37 °C for 2 h and then centrifuged at 1500 rpm for 5 mins. The absorbance of the supernatant was measured for release of hemoglobin at 545 nm. The percentage of hemolysis was calculated as follows: % hemolysis = (OD t -OD n )/ (OD p -OD n ) × 100. Where, OD t , OD n , and OD p are the absorbance values of the test sample, negative control (NS), and positive control (water), respectively. All the hemolysis experiments were performed in triplicate.
Cytotoxicity assessment in vitro. L02 cells were seeded at a density of 5 × 10 3 cells/well in a 96-well microtiter plate and were cultured in DMEM with Gal-CMCS-Fe 3 O 4 -NPs at 0(as control), 10,25,50,100,150,200,250, or 500 μg/ml. After 72 h, cellular morphology was observed using an inverted microscope, and the growth of cells from four wells was assessed using a Cell Counting Kit-8 (CCK-8, Beyotime Institute of Biotechnology, Jiangsu, China) according to the manufacturer's instructions. Cell viability (%) was calculated as the mean optical density of treated wells/mean optical density of control wells ×100.
In vitro transfection. Galactose (1 ml, 100 mM) were added 15 min before transfection. HepG2 cells were co-incubated with Gal-CMCS-Fe 3 O 4 -NPs prepared with pcDNA6.2mir-EGFP plasmid (DNA content 2 μg) at an N/P ratio of 3:1. After a 72-h co-incubation, cells were fixed in 4% paraformaldehyde for 30 min followed by nuclear staining with 4′ ,6-diamidino-2-phenylindole (DAPI; Beyotime Biotech Inc.). An inverted fluorescence microscope was used to observe the transfected cells, and flow cytometry was used to determine the percentage of transfected cells. Every experiment repeated three times.

Cytotoxicity assessment in vivo.
Gal-CMCS-Fe 3 O 4 -NPs (90 μg, 100 μl) and the same volume of NS as control were were injected into the tail vein of BALB/C nude mice(n = 50). At 1, 2, 3, 7, or 14 days after injection, 5 mice of each group were sacrificed, and blood serum samples were collected , and levels of aspartate transaminase (AST), alanine transaminase (ALT), and total bilirubin (T-BIL) were measured. The influence of NPs on the morphology of various organs and tissues was also observed at the 14th day after injection. Orthotopic transplantation tumor model of HCC. HepG2 cells (1 × 10 6 ) were injected subcutaneously into the flanks of 4-week-old male BALB/C athymic nude mice. Tumorectomies were performed when the subcutaneous tumors grew to a diameter of 1 cm. A small piece (approximately 1-2 mm 3 ) of prepared fresh tumor tissue was implanted into the capsule of the liver lobe in nude mice at an angle of 20°. Absorbable sutures (7-0) were used for local stiffening. All of the animal protocols were approved by the Animal Care and Use Committee of Nantong University and the Jiangsu Province Animal Care Ethics Committee (Approval ID: SYXK (SU) 2007-0021), and the methods were carried out in accordance with the approved guidelines. Western blot. Proteins were extracted from the tissues using RIPA lysis buffer (Beyotime) containing phosphatase inhibitor (100:1). Then the lysates were centrifuged at 14 000 rpm for 20 minutes. Western blot analysis was performed as previously described 50 . Briefly, the protein was transferred into PVDF membrane after separating from 10% SDS-PAGE. Primary antibodies used included anti-RASSF1A (eBioscience, San Diego, CA, USA), anti-caspase-3, anti-p53, anti-bax, and anti-p21 (Santa Cruz Biotechnology, CA, USA). Antibodies were diluted according to the manufacturers' instructions.

In vivo transfection of Gal-CMCS-
Statistical analysis. Quantitative data are shown as mean ± standard error of the mean of at least three independent experiments. Statistical analysis was performed using t-tests with SPSS/Win13.0 software (SPSS, Inc., Chicago, IL, USA). A P-value < 0.05 was considered statistically significant.