In Vitro Gene Delivery Mediated by Asialofetuin-Appended Cationic Liposomes Associated with γ-Cyclodextrin into Hepatocytes

The purpose of this study is to evaluate in vitro gene delivery mediated by asialofetuin-appended cationic liposomes (AF-liposomes) associating cyclodextrins (CyD/AF-liposomes) as a hepatocyte-selective nonviral vector. Of various CyDs, AF-liposomes associated with plasmid DNA (pDNA) and γ-cyclodextrin (γ-CyD) (pDNA/γ-CyD/AF-liposomes) showed the highest gene transfer activity in HepG2 cells without any significant cytotoxicity. In addition, γ-CyD enhanced the encapsulation ratio of pDNA with AF-liposomes, and also increased gene transfer activity as the entrapment ratio of pDNA into AF-liposomes was increased. γ-CyD stabilized the liposomal membrane of AF-liposomes and inhibited the release of calcein from AF-liposomes. The stabilizing effect of γ-CyD may be, at least in part, involved in the enhancing gene transfer activity of pDNA/γ-CyD/AF-liposomes. Therefore, these results suggest the potential use of γ-CyD for an enhancer of transfection efficiency of AF-liposomes.


Introduction
The principle of somatic gene therapy is that genes can be introduced into selected cells in the body in order to treat genetic or acquired diseases. The liver may be potentially an important target for gene therapy, because crucial diseases such as amyloidosis, primary biliary cirrhosis, familial hypercholesteremia, phenyl ketonuria, and virus hepatitis occur in this organ [1]. In addition, the liver has the ability to synthesize a wide variety of proteins, to perform various posttranslational modifications, and to secrete them into the blood.
Of various nonviral methods, the lipofection method, by which cationic lipids (cationic liposomes) are used for transfection and interact with plasmid DNA (pDNA) to give a lipoplex, has recently attracted attention [2]. Cationic liposomes have great advantages as gene delivery carriers such as (1) low cytotoxicity and immunogenicity [3], (2) regulation of the pharmacokinetics through the modification of particle size or lipids components of liposomes [4], (3) entrapment of pDNA into inner water phase of liposomes and suppression of DNA degradation by DNase [5], and (4) delivery of gene to target cells by the addition of target ligands and/or antibody [6].
Asialofetuin (AF) is a glycoprotein that possesses three asparagine-linked triantennary complex carbohydrate chains with terminal N-acetylgalactosamine residues. The protein displays affinity to asialoglycoprotein receptor (ASGP-R) on hepatocytes and enters the cells through the receptor [7,8]. Thus, AF has been used as a ligand to deliver drugs to hepatocytes and a competitive inhibitor to ASGP-R [9,10]. In fact, the widespread use of AF-appended liposomes (AFliposomes) as a hepatocyte-selective gene transfer carrier has been reported [11,12].
Cyclodextrins (CyDs) have recently been applied to gene transfer and oligonucleotide delivery [13][14][15][16]. CyDs are 2 Journal of Drug Delivery cyclic (α-1,4)-linked oligosaccharides of α-D-glucopyranose containing a hydrophobic central cavity and hydrophilic outer surface, and they are known to be able to act as novel host molecules by chemical modification [17]. Davis and his colleagues reported that the ternary complex of a water-soluble β-CyD polymer with 6 A ,6 D -dideoxy-6 A ,-6 D -di-(2-aminoethanethio)-β-CyD and dimethylsuberimidate (βCDP6), galactosylated, or transferrin polyethylene glycol conjugates with adamantane, and pDNA possesses higher transfection efficiency in hepatoma or leukemia cells, respectively, through receptor-mediated endocytosis [18,19]. Recently, we reported the potential use of PAMAM dendrimer functionalized with α-CyD (α-CDE) [20] and lactosylated α-CDE (Lac-α-CDE) as a hepatocyte specific gene delivery in vitro and in vivo [21]. Meanwhile, Lawrencia et al. reported that lipoplex transfection of pDNA with DOTAP (N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate) in the presence of cholesterol, which is solubilized by methyl-β-cyclodextrin (methyl-β-CyD), has significantly improved transfection efficiency in urothelial cells due to change in membrane fluidity by methyl-β-CyD [22]. In addition, we previously demonstrated that intravenous injection of the pegylated liposomes entrapping the doxorubicin (DOX) complex with γ-CyD in BALB/c mice bearing Colon-26 tumor cells showed DOX accumulation in tumor tissues and the potent antitumor effect, compared with those of DOX solution and pegylated liposomes entrapping DOX alone [23]. These lines of evidence suggest that transfection efficiency and pharmacokinetics of pDNA can be altered by the association of CyDs with AF-liposomes.
Based on these backgrounds, the purpose of this study is to evaluate in vitro gene delivery of AF-liposomes associated with CyDs as a hepatocyte-selective nonviral vector in HepG2 cells. In addition, the mechanisms by which γ-CyD enhanced transfection efficiency of pDNA/AF-liposomes were investigated in the view of a receptor recognition, physicochemical properties (particle size, ζ-potential, and encapsulation ratio), membrane fluidity, cellular uptake, and cytotoxicity of AF-liposomes.

Preparation of pDNA/CyDs/AF-Liposomes.
Preparation of pDNA/CyD/AF-liposomes was performed according to the method reported by Hara et al. with some modifications [12]. Briefly, 2 μL of the solution containing pDNA (1 μg/μL) and CyDs (1 μM/μM lipids) dissolved in TE buffer were added to AF-or N-liposomes suspension. After mixing, the solution was freeze-dried. Then, the sample was rehydrated with THBS (10 mM, pH 7.5) for 30 min. After freezingthawing for three times, the vesicles were extruded through PVDF membranes (Nucleopore, Plesanton, CA) with pores of diameter 450 and 200 nm. The filtrates were used for further experiments as pDNA/CyDs/AF-liposomes or pDNA/CyDs/N-liposomes. The entrapment ratios of CyDs were evaluated by the anthrone-sulfuric acid method [26]. Briefly, 3 mL of anthrone reagent was added to 0.5 mL of the suspension containing CyDs/liposomes. The tube was covered with a glass ball and was heated for 10 min in boiling water. After quenching with cold water, absorbance of the suspension was measured by a U-2000A spectrophotometer (Hitachi, Tokyo, Japan) at 620 nm. The encapsulation ratios of pDNA were determined by a fluorescent spectrometer F-4500 (Hitachi, Tokyo, Japan). Briefly, 350 μL of the suspension containing pDNA/CyD/AF-liposomes in 10 mM THBS (pH 7.5) were mixed with 200 times diluted Picogreen dsDNA reagent (350 μL). After incubation for 30 min at 25 • C, fluorescent intensity (F po ) was determined. Next, after addition of 20% of Triton-X (20 μL) to the sample, fluorescent intensity (F pt ) was determined and the encapsulation ratio of pDNA was calculated as follows: encapsulation ratio (%) = [(F pt · r − F po )/F pt · r] · 100, where r is compensation coefficient (r = 1.03). Particle size and ζ-potential value of liposomes in 10 mM THBS (pH 7.5) were measured by a submicron particle analyzer N4 Plus (Beckman Coulter, Fullerton, CA) and ELS-8000 (Otsuka Electronics, Osaka, Japan), respectively.

Interaction of CyDs with AF-Liposomes.
AF-liposomes encapsulating calcein were prepared by the freezing and thawing method after addition of 0.1 mM calcein in 10 mM THBS (pH 7.5). The vesicles were extruded through two stacked polycarbonate membranes (Nucleopore, Plesanton, CA) with pores of diameter 1 μm. The sample was subjected to 10 passes through the filter at 40 • C. The filtrates were extruded through the polycarbonate membranes (pore size 0.2 μm) as described above. Two milliliters of CyDs solution adjusted at the appropriate concentration (5-20 mM) using 10 mM phosphate buffer were added to 20 μL of the liposomal suspension, and then the resulting suspension was incubated for 30 min at 25 • C. The fluorescence intensity of calcein (F t ) was measured with a fluorophotometer (Hitachi F-4500, Tokyo, Japan) at 25 • C; excitation and emission wavelengths were 490 and 520 nm, respectively. After addition of 20 μL of cobalt chloride solution (10 mM) to the sample to quench the fluorescence of nonencapsulated calcein, the intensity of fluorescence of encapsulated calcein (F in ) was also determined. Then, the liposomes were completely disrupted by the addition of 20 μL of Triton X-100 (20%) solution, and the intensities of fluorescence after quenching by cobalt chloride (F q ) were measured. Calcein encapsulation ratio was calculated by the equation as follows: encapsulation ratio 2.5. Cell Culture. HepG2 cells, a human hepatocellular carcinoma cell line, A549 cells, a adenocarcinomic human alveolar basal epithelial cells, and NIH3T3 cells, a mouse embryonic fibroblast cell line, were obtained from Riken Bioresource Center (Tsukuba, Japan). HepG2, A549, and NIH3T3 cells were grown in DMEM, containing 1 × 10 5 U/L of penicillin, 0.1 g/L of streptomycin supplemented with 10% FCS at 37 • C in a humidified 5% CO 2 and 95% air atmosphere.

In Vitro Gene
Transfer. In vitro transfection of the pDNA/CyDs/AF-liposomes was performed utilizing the luciferase expression of pDNA (pRL-CMV-Luc or pGL3control vector) in HepG2, A549, and NIH3T3 cells. The cells (2 × 10 5 cells per 24 well plate) were seeded 6 h before transfection and then washed twice with 500 μL of serum-free medium. Two hundred μL of serum-free medium containing pDNA/CyDs/AF-liposomes in the absence and presence of AF as a competitor protein or BSA as a control protein were added to each dish and then incubated at 37 • C for 3 h. After washing HepG2 cells with serum-free medium twice, 500 μL of medium containing 10% FCS were added to each dish and then incubated at 37 • C for 21 h. After transfection, the gene expression was measured as follows: Renilla and firefly luciferase contents in the cell lysate were quantified by a luminometer (Lumat LB9506, EG&G Berthold Japan, Tokyo, Japan) using the Promega Renilla and firefly luciferase assay reagent (Tokyo, Japan), respectively. It was confirmed that CyDs and AF-liposomes had no influence on the luciferase assays under the present experimental conditions. Total protein content of the supernatant was determined by Bio-Rad protein assay kit (Bio-Rad Laboratories, Tokyo, Japan). EGFP-expressing cells were determined by a confocal laser scanning microscopy (CLSM, Olympus FV300-BXCarl Zeiss LSM-410, Tokyo, Japan) with an argon laser at 488 nm after fixation. Briefly, the cells (2 × 10 5 cells per 35 mm glass bottom dish) were seeded 6 h before transfection and then washed twice with 500 μL of serum-free medium. Transfection with pEGFP N1 DNA was performed using the same protocol as described above. The EGFP expression ratio was determined by the number of EGFP-expressing cells per 100 cells. To observe the cellular uptake of Rhodamine-labeled AF-liposomes (RH-AFliposomes) and Alexa-labeled pDNA (Alexa-pDNA), HepG2 cells (2 × 10 5 cells/dish) were incubated with the Alexa-pDNA/CyDs/RH-AF-liposomes for 3 h. After incubation, the cells were rinsed with PBS (pH 7.4) twice and fixed in methanol at 4 • C for 5 min prior to observation by a CLSM.

Interaction between AF-Liposomes and CyDs.
CyDs have been reported to interact with cell membrane constituents such as cholesterol and phospholipids, resulting in the induction of hemolysis of human and rabbit red blood cells at high concentrations of CyDs [28][29][30]. In addition, CyDs are well known to disrupt liposomal membranes, depending on CyD cavity sizes and membrane components [31,32]. Then, we evaluated the interaction of CyDs with AF-liposomes  These results suggest that the interaction of γ-CyD and HP-CyDs with AF-liposomes was weaker than that of α-CyD and DM-β-CyD.

In Vitro Gene Delivery of AF-Liposomes Associated with
CyDs. Next, we evaluated in vitro gene transfer activity of pDNA/CyDs/AF-liposomes in HepG2 cells, ASGP-R positive cells ( Figure 3). Here, we used pRL-CMV (CMV promoter) as pDNA and the charge ratio (AF-liposomes/pDNA) of 1.6 optimized by our previous study (data not shown). The gene transfer activity of pDNA/γ-CyD/AF-liposomes in HepG2 cells was significantly higher than that of pDNA/HP-α-, HP-β-, and HP-γ-CyDs/AF-liposomes ( Figure 3). Next, we measured the Renilla luciferase mRNA level after transfection of pDNA/CyDs/AF-liposomes in HepG2 cells by the RT-PCR method ( Figure 4). As predicted, γ-CyD significantly increased the luciferase expression, but HP-γ-CyD did not, suggesting that γ-CyD is involved in the enhancing effect on luciferase expression at or prior to a transcription process.
To evaluate the enhancing effects of γ-CyD on gene transfer activity of AF-liposomes associating pDNA encoding EGFP controlled by a CMV or SV40 promoter, we examined EGFP and firefly luciferase gene expression after transfection of pDNA/γ-CyDs/AF-liposomes in HepG2 cells ( Figure 5).   The charge ratio of AF-liposome/pDNA was 1.6. γ-CyDs were added to AFliposome suspension before freeze-drying. The concentrations of γ-CyDs were 1 μM/μM lipids. The ζ-potential was measured by a light-scattering method. The particle size was determined using a photon correlation spectroscopic analyzer. Each value represents the mean ± SEM of 3 experiments. * P < .05 versus without CyD.
The extent of EGFP-expressing cells in the pDNA/AFliposomes system without CyDs was found to be 8%, while that with γ-CyD and HP-γ-CyD were 19% and 10%, respectively ( Figure 5(a)). Additionally, the enhancing effect of γ-CyD was observed in the pGL3-control vector encoding firefly luciferase and having a SV40 promoter ( Figure 5(b)). These results suggest that the enhancing effect of γ-CyD on gene transfer activity of AF-liposomes is a geneand promoter-independent manner. To confirm whether pDNA/γ-CyD/AF-liposomes have ASGP-R-mediated gene transfer activity, we performed transfection experiments in HepG2 cells in the presence and absence of AF, as an ASGP-R competitive inhibitor. Here, we confirmed that ASGP-R are expressed in HepG2 cells by the RT-PCR method (data not shown), which is consistent with previous findings [33][34][35]. As shown in Figure 6, gene transfer activity of AF-liposomes was markedly inhibited by the addition of AF, but not BSA, a control protein. These results suggest that AF-liposomes had the ASGP-R-mediated gene transfer activity. Interestingly, the similar enhancing effects of γ-CyD on Renilla luciferase protein expression after transfection of pDNA/AF-liposomes were, however, observed in A549 cells and NIH3T3 cells, ASGP-R negative cells (Figure 7). These results suggest that γ-CyD enhances the transfection efficiency of pDNA/AFliposomes in ASGP-R-independent manner.
Although the cell viability after treatment with pDNA/AFliposomes and pDNA/CyDs/AF-liposomes for 3 h slightly decreased as a charge ratio of AF-liposomes/pDNA was increased in both HepG2 and NIH3T3 cells, that is, more than 80% of cell viability after application of pDNA/CyDs/AF-liposomes was observed at the charge ratio of 1.6 used in the transfection study as described above. These results suggest that pDNA/CyDs/AF-liposomes have great advantages as a nonviral vector, that is, superior transfection efficiency and less cytotoxicity, and the enhancing  effect of γ-CyD on gene transfer activity of pDNA/AFliposomes is not associated with cytotoxicity.
Next, we examined the effects of γ-CyDs on encapsulation ratios of pDNA into AF-liposomes ( Table 3). The encapsulation ratio of pDNA in pDNA/γ-CyD/AF-liposomes was significantly higher than those of pDNA/HP-γ-CyD/AFliposomes and pDNA/AF-liposomes. Meanwhile, the encapsulation ratios of γ-CyD and HP-γ-CyD into AF-liposomes were approximately 10.2% and 11.0%, respectively. These results suggest that γ-CyD improves the encapsulation of pDNA into AF-liposomes, although the extent of γ-CyD encapsulation into AF-liposomes is not high. In addition, both the encapsulation ratio of pDNA into AF-liposomes and gene transfer activity of pDNA/AF-liposomes were raised, as the number of the freeze-thaw cycle was increased, suggesting that the encapsulation of pDNA in AF-liposomes is correlated with the gene transfer activity of pDNA/AFliposomes (Table 4).

Effects of γ-CyD on Membrane Fluidity of Liposomes.
It is known that membrane fluidity of liposomes affects the release profiles as well as a retention time of drug encapsulated into liposomes. Therefore, we investigated the effects of γ-CyD on membrane fluidity of AF-liposomes using a Microcal MC2 scanning calorimeter. In the present study, we utilized N-liposomes to eliminate the effects of the heat degeneration of AF in a DSC thermograph. Figure 9 shows the effects of γ-CyD and HP-γ-CyD on DSC thermograms of N-liposomes (5 mM of total lipids). The peak derived from gel-to-fluid state transition in N-liposomes was observed at 55 • C, while new peak was appeared at 42 • C in the presence of 50 mM of γ-CyD (Figure 9(a)). On the other hand, no significant change in DSC thermographs was observed in the presence of HP-γ-CyD (Figure 9(b)). These results suggest that γ-CyD may affect the membrane fluidity of Nliposomes.
Generally, membrane fluidity and phase transition of liposomes are determined by the intensity of lipid-lipid interactions such as hydrophobic interaction, van der Waals forces, and hydrogen bond. Therefore, to evaluate the effects of γ-CyD on lipid-lipid interactions in AF-liposomes, we performed DSC analysis of DLPC-liposomes in the presence and absence of γ-CyD. The reason why we used DLPC-liposomes is due to clear observation of lipid-lipid interaction using DLPC composed of AF-liposomes. Figure 10   The charge ratio of AF-liposomes/pDNA was 1.6. pDNA/AF-liposome was prepared using a freeze-thaw method which was repeated from 0 to 5 cycles. Cells were incubated with pDNA/γ-CyDs/AF-liposome for 3 h in FCS-free medium. After washing twice, the cells were incubated for 21 h in culture medium supplemented with 10% FCS. The luciferase activity in cell lysates was determined using a luminometer. Each value represents the mean ± SEM of 3 experiments.
effects of γ-CyD and HP-γ-CyD on DSC thermograms, phase transition temperature, and enthalpy (ΔHcal) of DLPCliposomes (5 mM of total lipids). The phase transition temperature (Tc) of DLPC-liposomes in the presence of γ-CyD was shifted to high temperature as the concentration of γ-CyD was increased (Figures 10(a) and 10(c)). The ΔHcal value of DLPC-liposomes in the γ-CyD system was drastically elevated at 20 mM of γ-CyD (Figure 10(d)).
Meanwhile, in the case of HP-γ-CyD, there was no significant change in DSC thermograms, phase transition temperature, and the ΔHcal values (Figures 10(b), 10(c) and 10(d)). Taken together, these results strongly suggest that γ-CyD enhances the lipid-lipid interaction of DLPC-liposomes, leading to the membrane stabilization of DLPC-liposomes.
3.6. Cellular Uptake of pDNA/AF-Liposomes. Next, we examined the cellular uptake of pDNA/γ-CyD/AF-liposomes into HepG2 cells using a CLSM. Figure 11 shows the CLSM images for distribution of RH-AF-liposomes and Alexa-pDNA in HepG2 cells at 3 h after transfection. The strong fluorescence derived from RH-AF-liposomes and Alexa-pDNA in the presence of γ-CyD was mainly observed in cytoplasm of HepG2 cells. Meanwhile, the fluorescence RH-AF-liposomes and Alexa-pDNA in the absence of CyD and with HP-γ-CyD was mainly observed on cell surface. Hence, these results suggest that pDNA/γ-CyD/AF-liposomes can be internalized into HepG2 cells to a larger extent, compared to pDNA/AF-liposomes and pDNA/γ-CyD/AF-liposomes.

Discussion
In this study, we clarified that pDNA/γ-CyD/AF-liposomes have potent hepatocyte-selective gene transfer activity and negligible cytotoxicity, compared to pDNA/AF-liposomes and pDNA/HP-CyDs/AF-liposomes. In cationic liposome-mediated gene transfection, lipid composition and lipid type are the most important physicochemical factors, because they affect not only the interaction with pDNA but also the affinity to target cells [36,37]. In the present study, we prepared AF-liposomes with a lipid composition of TMAG/DOPE/DLPC/DPPE (2/4/3/1, molar ratio). DOPE is known to enhance the endosomal escape of pDNA due to its structural change into hexagonal II form at pH 5-6, an endosomal pH range, resulting in destabilizing endosomal membranes [38,39]. TMAG, a cationic lipid, makes it possible to interact with pDNA in AF-liposomes. Additionally, DPPE was used as a binding lipid with AF. Actually, cationic liposomes composed of TMAG/DOPE/DLPC (1/2/2, molar ratio) are commercially available transfection reagents as GeneTransfer, which have already been utilized in a clinical trial for a nonviral vector to deliver the interferon-β gene for the treatment of brain tumor in Japan [40]. Therefore, we used AF-liposomes composed of these lipids in the present study.
The most important finding found in the present study is that γ-CyD enhances transfection efficiency of pDNA/AFliposomes in HepG2 cells (Figures 3-5) with negligible cytotoxicity (Figure 8). The enhancing mechanisms of γ-CyD presumed are discussed as follows.
In the present study, we revealed that transfection efficiency of the pDNA/γ-CyD/AF-liposomes, not N-liposomes, was inhibited by the addition of AF in HepG2 cells ( Figure 6). Meanwhile, in NIH3T3 cells, transfection efficiency of the pDNA/AF-liposomes was not suppressed by the addition of AF. These results strongly suggest that pDNA/γ-CyD/AFliposomes can be entered HepG2 cells through ASGP-Rmediated endocytosis, consistent with Aramaki and his colleague's report [41], and the enhancing effect of the γ-CyD may by associated with the ASGP-R-mediated endocytosis. However, γ-CyD also enhanced gene transfer activity of pDNA/AF-liposomes even in A549 cells and NIH3T3 cells, ASGP-R negative cells (Figure 7). These results suggest that the enhancing effect of γ-CyD on transfection efficiency of pDNA/AF-liposomes is in an ASGP-R-independent manner. The charge ratio of AF-liposomes/pDNA was 1.6. γ-CyDs were added to AF-liposomes suspension before freeze-drying. The concentrations of γ-CyDs were 1 μM/μM lipids. Cells were incubated with Alexa-pDNA/γ-CyDs/RH-AF-liposomes for 3 h in FCS-free medium. After washing twice, the cells were observed using a confocal laser scanning microscopy.
The particle sizes and encapsulation ratio of pDNA/γ-CyD/AF-liposomes should be involved in the enhancing effect of γ-CyD on gene transfer activity of AFliposomes. The particle size of pDNA/γ-CyD/AF-liposomes was decreased in the presence of γ-CyD, although the ζpotential value of pDNA/γ-CyD/AF-liposomes was almost equivalent to that of pDNA/AF-liposomes and pDNA/HP-γ-CyD/AF-liposomes ( Table 2), suggesting that γ-CyD inhibits the aggregation of pDNA/AF-liposomes, because the particle size shows more than 200 nm, despite the fact that the liposomes were extruded through a filter membrane having a pore size of 200 nm. Meanwhile, the encapsulation ratio of pDNA was significantly increased by adding γ-CyD to pDNA/AF-liposomes (Table 3). Thus, these lines of evidence speculate that addition of γ-CyD enhances cellular uptake of pDNA/AF-liposomes. In fact, the CLSM study demonstrated that cellular uptake of pDNA/γ-CyD/AF-liposomes was higher than that of pDNA/γ-CyD/AF-liposomes ( Figure 11). Furthermore, we confirmed that transfection efficiency of pDNA/γ-CyD/AF-liposomes was increased, as the encapsulation ratio of pDNA into pDNA/γ-CyD/AF-liposomes was augmented (Table 4). In view of the findings, the particle size of pDNA/γ-CyD/AF-liposomes and encapsulation ratio of pDNA into pDNA/γ-CyD/AF-liposomes are crucial role for enhancing transfection efficiency of pDNA/AF-liposomes.
The important question regarding the enhancing effect of γ-CyD on transfection efficiency of pDNA/AF-liposomes still remains, because three types of HP-CyDs did not have the enhancing effect. To address this question, the DSC analysis was performed. This study indicated that γ-CyD, but not HP-γ-CyD, changed membrane fluidity and stabilized the N-liposomal membranes (Figure 9). In fact, γ-CyD increased the Tc value of DLPC-liposomes, although HPγ-CyD did not increase anymore ( Figure 10). This increase in the Tc value induced by γ-CyD could be attributed to a compactness of lipid bilayer of DLPC liposomes. It is thereby possible that γ-CyD may increase Tc values of the other liposomes such as DMPC, DPPC, and DSPC. Here, it is well known that γ-CyD is highly hydrophilic and surface inactive [17]. Therefore, we presumed that γ-CyD encapsulates the aqueous compartment of liposomes rather than in the bilayer of liposomes. Anyhow, it is clear that the magnification of the interaction of liposomal membranes containing DLPC with HP-γ-CyD is weaker than that with γ-CyD, possibly due to the steric hindrance of the HP group in a HP-γ-CyD molecule. Taken together, it is likely that the stabilizing effects of γ-CyD on AF-liposomal membrane may lead to inhibition of pDNA leakage from AF-liposomes and increase in cellular uptake of pDNA, eventually leading to the enhancement of in vitro transfection efficiency of pDNA/γ-CyD/AF-liposomes in cells.
Finally, we investigated the role of free γ-CyD on transfection efficiency of pDNA/γ-CyD/AF-liposomes in HepG2 cells. The physical mixture of pDNA/AF-liposomes and γ-CyD in culture medium had no enhancing effect on transfection efficiency of pDNA/AF-liposomes (data not shown). Therefore, encapsulation of γ-CyD into AF-liposomes may be pivotal for enhancing gene transfer activity. To reveal the detailed mechanism for the enhancing effect of γ-CyD associated in AF-liposomes on transfection efficiency of pDNA/AF-liposomes, further elaborate study should be necessary.

Conclusions
In the present study, we demonstrated that γ-CyD enhanced gene transfer activity of pDNA/AF-liposomes in not only HepG2 cells but also A549 and NIH3T3 cells, probably due to various effects of γ-CyD on AF-liposomes such as inhibition of aggregation of the liposomes, high encapsulation of pDNA into the liposomes, and stabilization of lipid bilayer of the liposomes. Consequently, the potential use of γ-CyD could be expected as an enhancer of gene transfer activity of AFliposomes. Also, these data may be useful for design of cellspecific cationic liposomes as a nonviral vector.