Gene Vector Laboratory, Division of Cellular and Molecular Research, National Cancer Center, 11 Hospital Drive, Singapore 169610

Correspondence to: K M Hui

The synthesis of a series of novel cationic lipids through the systematic substitution of cholesterol derivatives that could greatly enhance the delivery and expression of plasmid DNA in vitro and in vivo is described. Two of the newly synthesized lipids, designated as NCC4 and NCC10, were chosen to be studied in detail and gave much higher levels of gene expression than that which could be obtained with some of the conventional cationic polymers and cationic liposomes. In vivo studies with both NCC4 and NCC10 also showed better ability in delivering the reporter gene to the target cells through intrasplenic injection. In addition, by varying the DNA/lipid charge ratios, NCC4 and NCC10 can withstand serum inactivation in vitro. However, this does not correlate with the corresponding increase in the level of gene expression following systemic gene delivery with NCC4 and NCC10 in vivo. Gene Therapy (2001) 8, 855-863.

cationic liposomes; cancer gene therapy; gene delivery; heterocyclic cholesterol; serum inactivation

Introduction

Gene therapy represents a promising approach for the treatment of inherited or acquired diseases.^1,2,3,4 However, one of the most difficult hurdles in achieving effective gene therapy is the requirement for the use of efficient vehicles to deliver the gene of interest into target cells. A diverse spectrum of gene delivery vehicles ranging from replication-incompetent viruses to DNA formulated with various delivery vehicles has been utilized.^5,6,7 Administration of DNA alone has also yielded successful gene transfer for a number of isolated applications but with a limited spectrum of organ-specific expression.⁸ However, the majority of applications requires the assistance of a delivery vehicle to facilitate gene transfer.

Essentially, two approaches have been adopted to introduce exogenous DNA into cells. These are viral vector- based and plasmid DNA-based systems. Each system has its advantages and disadvantages and all vehicles have been reported to achieve some level of gene delivery. Virus-based vectors have attracted most interest because of their expected high efficiency at mediating gene transfer.^9,10 Currently, the two most popular viral vectors for gene transfer are the replication-defective retroviruses and adenoviruses. The retroviral vectors provide high degrees of gene transfer but can transduce only dividing cells.¹¹ Although adenoviral vectors can transfect non-dividing cells, they induced inflammatory and immune responses and therefore limited the duration of expression and efficacy of subsequent re- administration.^12,13,14 Most nonviral gene delivery systems are based on cationic compounds. These include either cationic polymers or cationic lipids that spontaneously complex with a plasmid DNA construct by means of electrostatic interactions to yield a condensed form of DNA that shows increased stability towards nucleases. Several features of the nonviral systems offer many advantages over viral systems, such as the ease of manufacture, safety, stability, lack of vector size limitation, low immunogenicity and the potential to incorporate targeting ligands.^15,16,17

Among the nonviral gene delivery systems, much interest has been directed towards cationic liposomes since they can potentially overcome the problems associated with most viral vectors. Several different cationic lipids, capable of achieving a reasonable level of gene delivery, have been synthesized.^{18,19,20,21,22,23,24,25,26} The basic structure of most of the cationic lipids employed for gene delivery is made up of four chemical functional domains: (1) a positively charged head-group that allow interactions with negatively charged DNA; (2) a spacer arm of varying length; (3) a hydrophobic anchor group of either a cholesteryl or oleoyl residue and allow interactions with cell membrane; and (4) a linker bond that bridges the amino head-group and the hydrophobic anchor residue together.²⁷ These vary from lipids having monocation head-groups, such as DOTMA, DMRIE, DOTAP, and DC-chol, to lipids having polycation head-groups. The latter include DOSPA and DOGS. Except for DOGS which was reported to have optimal gene delivery activity on its own,²⁸ most of the cationic lipids are formulated as an aqueous liposome suspension with the helper lipid, DOPE or cholesterol.²⁹ Most of the existing cationic lipids are the derivatives of long-chain fatty acids or cholesterol.

Although the uptake of cationic liposomes has been suggested to be via endocytosis and the endosomes,^30,31 the exact mechanism of cationic lipid-mediated gene transfer has not been elucidated. Much of the recent development on new cationic lipid formulations has been empirical and requires the systematic screening for optimal activity in the respective target cells/tissues. In this report, we describe the synthesis of a series of novel cationic lipids that can greatly enhance the expression of plasmid DNA in vitro and in vivo through systematic substitution of cholesterol derivatives. The structures and charges of the various cationic lipids synthesized were also employed to correlate with the effectiveness of gene delivery in an attempt to elucidate how cationic liposomes interact with DNA molecules and the cell membranes.

Results

A variety of amphipathic amines or polyamines were synthesized by acylation (Figure 1). These newly synthesized lipids were evaluated systematically for their abilities to effect gene delivery. Ten representative cationic lipids having different chemical structures and charges, as shown in Figure 2, have been chosen for study in detail.

In our laboratory, we have employed the cationic lipid DC-chol as a gene delivery vehicle.^32,33 It has been confirmed that the gene delivery activity of DC-chol is optimal when used in conjunction with the neutral lipid DOPE at the molar ratio of 6:4 (DC-chol:DOPE).¹⁹ When the newly synthesized cationic lipids were compared with DC-chol and DOPE at the molar ratio of 6:4 for their ability to deliver the reporter gene pCMV-luciferase into human HepG2 cells, it was observed that NCC 3, 4, 5, 8 and 10 all gave significantly higher activities than DC-chol (Figure 3). The increase was even more pronounced for NCC4, NCC5 and NCC10 when compared with DC-chol. NCC4 and NCC10 gave an overall increase of more than six- and three-fold, respectively, in the luciferase activities when compared with DC-chol 24 h following gene delivery into HepG2 cells (Figure 3). On the other hand, NCC2 and NCC6 demonstrated a reduction in luciferase activity in comparison with DC-chol under the same conditions following gene delivery to HepG2 cells (Figure 3). NCC4 and NCC10 were chosen to be studied further since they are relatively easy to synthesize and demonstrated a good level of gene expression following delivery.

When compared with Lipofectamine (Gibco-BRL, Gaithersburg, MD, USA) and the cationic polymer PEI, NCC4 and NCC10 gave a more than two-fold increase in the luciferase gene activity following introduction of the pCMV-luciferase DNA into HepG2 cells (Figure 4). Among the various viral vectors available for gene delivery, adenovirus is one of the most efficient. We have therefore compared the ability of adenoviruses to infect the human HepG2 hepatoma cells with that of gene delivery mediated by the cationic liposomes containing NCC4, NCC10 and DC-chol (Figure 5). The infection conditions used have been optimized for adenovirus and different viral particles were employed for the infection (Figure 5). It could be demonstrated that NCC4 and NCC10 gave a better or at least a comparable level of gene expression to that obtained with 100 adenoviral particles.

The efficiencies of NCC4 and NCC10 to mediate gene transfer to various other tumor cell lines were further compared with that of DC-chol. These included KZ2 (melanoma), CNE2 (nasopharyngeal carcinoma), MCF7 (breast carcinoma), A549 (lung carcinoma), PA1 (ovarian carcinoma). For most of the cell lines studied, NCC4, NCC10 and DC-chol gave comparable results to that obtained with human hepatoma cell HepG2 (Figure 6). For the cell lines MCF-7 and SW837, DC-chol gave a higher level of gene expression in comparison to NCC4 (Figure 6). However, NCC10 gave the highest level of gene expression compared with NCC4 and DC-chol (Figure 6).

One of the major disadvantages with current cationic liposomes as mediators for gene transfer is their low efficiency of transfection, especially when employed for in vivo gene delivery.¹⁸ This observation has been demonstrated to be due to the inactivation of the DNA- liposome complexes in vivo.^34,35,36 In the present study, we evaluated the ability of NCC4, NCC10 and DC-chol to withstand serum inactivation at various charge ratios using HepG2 cells in the presence of different concentrations of FBS and mouse serum. In the presence of FBS, NCC10 showed resistance to serum inactivation at all the serum concentrations tested (Figure 7). At the charge ratio of 0.4, the efficiency of transfection mediated by NCC10 was enhanced by more than 276 fold in the presence of 1% FBS. It is apparent that NCC10 became more sensitive at higher serum concentrations at the charge ratios of 0.4 and 1.0 (Figure 7). However, at the charge ratios of 2.6 and 4.0, the efficiency of transfection mediated by NCC10 was not affected by the increase in serum concentrations (Figure 7). In comparison to NCC10, NCC4 was basically ineffective in mediating gene transfer at the charge ratio of 0.4 (Figure 7). However, the efficiency of transfection of NCC4 markedly increased with the gradual increase in the charge ratios to 1.0, 2.6 and 4.0 (Figure 7). The efficiencies of gene delivery obtained with liposomes produced with DC-chol remained relatively low for all the conditions studied (Figure 7).

All the three lipids were tested for their ability to transfect the human cell line HepG2 at various charge ratios in the presence of 1% and 5% of mouse serum (Figure 8). Compared with FBS, it was apparent that mouse serum is a more potent inhibitor of cationic liposome-mediated gene delivery (Figure 8). Although the efficiency of transfection mediated by NCC4 increased with the increase in charge ratios in the presence of 1% mouse serum, NCC10 gave the best consistent resistance to mouse serum inactivation at all the conditions tested (Figure 8).

To study if the differences in the efficiencies of transfection obtained for NCC4, NCC10 and DC-chol in the presence of sera resulted from differences in the zeta potentials of the liposomes at the different serum concentrations, the zeta potentials of the liposomes prepared from the various lipids were determined. As shown in Table 1, the overall charge of all of the liposome preparations changed from a positive value to a negative value when measured in the presence of serum. However, no significant difference in zeta potential could be obtained between NCC4, NCC10 and DC-chol at different concentrations of FBS (Table 1). Therefore, it is unlikely that the differences in the efficiencies of transfection obtained for NCC4, NCC10 and DC-chol in the presence of serum is due to differences in the zeta potentials of the liposomes.

We have also compared the ability of the newly synthesized cationic lipids NCC4 and NCC10 with DC-chol in their ability to mediate gene transfer in vivo. DNA liposome complexes, prepared either from NCC4, NCC10 or DC-chol, at charge ratios of 0.4, 1.0, 2.6 as well as 4.0 were injected directly into the spleens of mice and assayed for luciferase activities 1 day following the injection. At the charge ratio of 0.4, both NCC4 and NCC10 gave the highest expression of the reporter gene pCMV-luciferase (Table 2). The ability of the various cationic liposomes including NCC4, NCC10, DC-chol and DOTAP to mediated gene delivery were then compared under the pre-determined optimal conditions for each of the respective cationic liposome. The published optimal conditions for DOTAP³⁷ were followed. Under these conditions, it was observed that NCC4 and NCC10 were much better mediators for gene delivery in comparison to DOTAP and DC-chol (Table 2).

To try to understand why NCC4 and NCC10 did not produce a high level of in vivo gene expression following systemic delivery of the DNA/liposome complexes, we have repeated our in vitro gene transfer experiments to include the serum concentration of 20% and 55% (Figure 9). At the liposome/DNA charge ratio of 2.6 and a serum concentration of 20%, NCC4 and NCC10 gave a marginally better level of gene expression in comparison with DC-chol (Figure 9). At the liposome/DNA charge ratio of 4.0 and a serum concentration of 20%, NCC4 gave a high level of gene expression (Figure 9). However, at the liposome/DNA charge ratios of 2.6 and 4.0 and the serum concentration of 55%, it was demonstrated that the level of gene expression obtained for NCC4 and NCC10 was low, but still higher than that of DC-chol (Figure 9). A serum concentration of 55% would be similar to the serum concentration in vivo, this could therefore be one of the possible explanations for the low level of gene expression obtained following systemic injection of the NCC4-, NCC10-DNA complexes. Therefore, it appears that there is a direct correlation of the ability of NCC4 and NCC10 to mediate gene transfer in vitro and in vivo.

Discussion

Structurally, there are four functional domains that could be identified in cationic lipids as described in the previous part of this article. It has been reported that these four domains play a role in determining their abilities to transfect cells as well as their toxicities to target cells.^15,23,38 The anchor residues could be either cholesterol or diacyl chains. Usually, it has been reported that lipids with cholesterol residues gave better efficiency of gene delivery and relatively less toxic to target cells in comparison to lipids having diacyl residues in their hydrophobic anchor.²⁹ The linker within the cationic lipid could be in the form of a urea, amine, amide, ether or ester bond.²² The linker bond has been found to have certain correlation with reference to the stability of the cationic liposomes. It has been studied and reported that when carbamoyl bond is employed as the linker bond, the lipids derived would be degradable and therefore would be potentially less toxic to the target cells both in vitro and in vivo.¹⁹ A spacer arm of between three and six atoms between the amino group and the linker bond was also reported to provide the optimal distance for efficient gene delivery activity.²³ The positively charged head-group of a cationic lipid appears to be the most important domain in determining the overall efficiency of gene delivery characteristics for the particular cationic lipid. Lipids bearing linear amines or polyamines as the positively charged head-group exhibit good gene delivery activity.²³ This is especially true for cationic lipids that demonstrate an overall T-shape configuration.²³ Therefore, when the orientation of the amine or polyamine head group is structurally perpendicular in relation to the lipid anchor, the efficiency of the lipid to mediate DNA gene delivery will be enhanced.

In the present paper, we have synthesized a new series of cationic lipids using cholesterol and carbamoyl as the hydrophobic domain and linker bond, respectively. This is to exploit the fact that lipids containing carbomoyl linker bond are degradable and could potentially be less toxic in vivo. In comparison to the cationic lipid DC-chol, the main difference in the structure of our newly synthesized cationic lipids is the presence of heterocycles as the amino group for most of them. Chemically, these include morphiline (NCC4), imidazole (NCC3), pyridine (NCC6), and piperazine (NCC1, NCC5, NCC9, NCC10). To study the effect of these chemical changes on the overall efficiency of gene delivery of the cationic lipids, cationic lipids containing linear amine (NCC2) or polyamine (NCC7 and 8) as the head-group were also synthesized and compared with lipids having heterocyclic head groups for their ability to act as gene delivery vehicles (Figure 2). Cationic liposomes prepared from the cationic lipids NCC1, NCC3, NCC4, NCC5 and NCC10 that contain heterocycle as the head-group gave better or similar efficiency of gene transfer in comparison with DC-chol. The only exception is NCC6 which gave a poorer efficiency in comparison with DC-chol (Figure 2).

It appears therefore that cationic lipid with linear primary amines or polyamines as the head group were less active than lipids having heterocycles as the head-group in their structure with reference to their ability to deliver DNA into target cells. Within the group of cationic lipids having heterocycles in their structure, lipids with piperazine (NCC1, 5, 9 and 10) and morphiline (NCC4) are relatively more active (Figure 3). Furthermore, NCC4 and NCC10 are the most active of the 10 newly synthesized cationic lipids. Although NCC9 has a piperazine group as the head-group, it is not efficient in gene delivery. This may due to the fact that NCC9 contains two cholesteryl groups and as a result, it is possible that its overall structure might be too bulky to interact with DNA (Figure 2). In comparison, cationic lipids with pyridine as their head-group, for example NCC6, are less active. The positively charged head-group was generally believed to allow interactions between the cationic lipids, the negatively charged DNA, as well as cell membrane through charge/charge interactions.^39,40 The presence of nitrogen and oxygen atoms in the heterocyclic ring might further contribute to this charge/charge interaction of liposome and plasmid DNA and stabilize the binding between the cationic liposomes and DNA.

Of all the newly synthesized cationic lipids in this report, NCC4 and NCC10 are the most efficient gene delivery vehicles. This is also true when they were employed to transfect cell lines such as HepG2 (human liver cancer cell line) and KZ2 (human melanoma cell line) that are generally very difficult to transfect with other reagents including DC-chol, PEI, and Lipofectamine (Figure 6). When first studied, the cationic lipid NCC5 gave high efficiency of transfection with HepG2. However, its activity decreased very sharply on storage. A likely explanation is that NCC5 is not stable in such a formulation. In addition, all the newly synthesized cationic lipids reported contain the same carbamoyl bond in their structure as that reported for DC-chol and they should be hydrolyzed by esterases and therefore be degraded once administered into cells. Judging also from the appearance and the amount of extractable protein recovered from the cells following transfection, the newly synthesized lipids exhibited no higher toxicity in comparison to DC-chol (Gao and Hui, personal observation).

One of the major disadvantages with cationic liposome as a mediator for gene transfer is their low efficiency of transfection, especially when employed for in vivo gene delivery.⁴¹ One of the explanations is that serum proteins interact with the membrane bilayers of liposomes and thus destabilize the DNA-liposome complexes in vivo.^34,35,36 Attempts have been made to develop serum-resistant lipids and new formulations of DNA-liposome complexes to prevent the inactivation of cationic lipoplexes by serum.^42,43 It has been demonstrated that the charge ratios of cationic liposomes to DNA play a critical role in determining the efficiency of gene delivery in the presence of serum.⁴⁴ It was also suggested that the role of charge ratio in serum sensitivity is dependent on the chemical structures of the cationic lipids. For example, with the cationic lipid Lipofectamine, it is inactivated by serum at all the charge ratios studied. In contrast, cationic liposomes produced with the cationic lipid DOTAP and the helper lipid DOPE showed a lower charge ratio at which it can completely overcome the serum effect. In the present studies, NCC10 could overcome the serum effect at lower charge ratio than NCC4. At charge ratio of 0.4, the presence of FBS dramatically increased the gene expression of NCC10, whereas NCC4 was ineffective at that charge ratio. However, when the charge ratio was increased from 0.4 to 2.6, the gene expression mediated by both NCC4 and NCC10 was substantially increased by inclusion of FBS in the transfection complexes (Figure 7). Compared with FBS, mouse serum is a more potent inactivator of cationic liposomes. The addition of 5% mouse serum completely inactivated the transfection activity of NCC4 and NCC10 (Figure 8). The mechanism by which low concentrations of serum could transiently increase the gene delivery activity of NCC4 and NCC10 is presently unknown. Nevertheless, it is possible that the uptake of the liposome/DNA complexes into cells by endocytosis may be less active in the absence of serum.

The approach to generate new cationic lipids by systematic modifications of the head-group of cationic lipid allows us to study the interactions between liposomes and cell membrane, as well as their contributions towards the overall efficiency of gene delivery. A series of experiments has been performed to optimize the conditions for gene expression following intrasplenic injection of DNA/liposome complexes. The efficiency of gene delivery mediated by NCC4 and NCC10 at charge ratios of 0.4, 1.0, 2.6, as well as 4.0 was compared. At the charge ratio of 0.4, both NCC4 and NCC10 gave the highest level of gene expression of the reporter gene (Table 2). This is different from the cationic lipid DOTAP that reported to have maximal efficiency of gene delivery at the charge ratio of 3.0. It was observed that both NCC4 and NCC10 gave higher efficiency of gene delivery into spleens when compared with DOTAP at its respective optimal conditions (Table 2).

We have observed that NCC4 and NCC10 can mediate efficient gene delivery in vitro and mediate efficient gene delivery followed intrasplenic injection in vivo. Moreover, it was demonstrated that both NCC4 and NCC10 could withstand serum inactivation in vitro by changing the DNA/lipid charge ratios. We have therefore attempted to achieve efficient gene delivery by systemic injection of NCC4 and NCC10 via the tail vein of mice. At the charge ratio of 0.4, both NCC4 and NCC10 demonstrate lower levels of gene expression following tail vein injection (data not shown). It can therefore be concluded that the conditions employed for intrasplenic injection could not be directly employed for systemic gene delivery and the conditions established in vitro to demonstrate serum sensitivity (Figures 7 and 8) could not be applied directly in vivo. In addition, this suggests that the mechanisms of serum inactivation of DNA-liposome complexes in vitro may differ from that in vivo.

Materials and methods

Chemical synthesis

In this article, a series of novel cationic lipids for gene delivery was designed. Ten of these lipids, whose structures are shown in Figure 2, were synthesized. They are derivatives of cholesterol.

For all the synthesis, cholesteryl chloroformate was allowed to react either with amines or polyamines to produce the respective cationic lipids (Figure 2). Essentially, two different protocols of synthesis were employed: (1) Synthesis of 3{[4-(3-aminopropyl) morpholine]-carbamoyl) cholesterol chloride (HCl salt of NCC4). For this reaction, cholesteryl chloroformate (1 g, 2.2 mmol) was dissolved in 15 ml of anhydrous chloroform and stirred under N₂. 4-(3-Aminopropyl) morpholine (0.457 ml, 2.2 mmol) was added to the reaction mixture over a 20-min period. When the addition of 4-(3-aminopropyl) morpholine was complete, the mixture was further stirred at room temperature overnight. The solvent was then evaporated under vacuum and the residue was recrystallized with anhydrous ethanol and acetonitrile. The desired product was obtained as a white powder with an estimated yield of 49.2%. Proton NMR (300 MHz, CDCl₃): 5.66(s, 1H), 5.38(s, 1H), 4.43(bs, 1H), 4.12(bs, 4H), 3.31(t, 2H), 3.05(t, 6H), 2.36-0.62(m, 45H). The HCl salts of NCC1, NCC2, NCC3, NCC6, NCC9 and NCC10 were similarly synthesized. (2) Synthesis of 3{[N, N-bis(3-aminopropyl)-1,3-propanediamine]-carbamoyl} cholesterol chloride, (HCl salt of NCC8). N, N-bis (3-aminopropyl)-1,3-propanediamine (0.445 ml, 2.2 mmol) was dissolved in 10 ml anhydrous chloroform and stirred under N₂. Cholesteryl chloroformate (1 g, 2.2 mmol) was dissolved in 10 ml anhydrous chloroform and added to the reaction mixture over a 30-min period. The mixture was further stirred at room temperature for 2 h. The solvent was evaporated under vacuum and the dry powder was redissolved in 60 ml of anhydrous ethanol and filtered. HCl vapor was passed into the solution until no further precipitate formed. The white powder obtained was then dried under vacuum. The yield for this synthesis is approximately 73.6%. Proton NMR (300 MHz, CDCl₃): 5.44(s, 1H), 4.40(bs, 1H), 3.40-0.77(m, 67H), 0.68(s, 3H). The HCl salts of NCC5 and NCC7 were similarly synthesized.

Preparation of cationic liposomes

Cationic liposomes were prepared by mixing stocks of the various cationic lipids synthesized and DOPE (dioleoyl phosphatidylethanolamine; Avanti Polar- Lipids, Alabama, USA) at the molar ratio of 6:4 under a gentle stream of N₂ and dried under vacuum overnight as described earlier.³² The dried film of lipid was hydrated in 1 ml of 20 mM sterile HEPES buffer (pH 7.8). For lipids that did not dissolve readily in HEPES, 0.7 ml HEPES buffer was first added and the pH was then adjusted using 0.1 N sterile HCl until all the lipids had dissolved. The final volume was diluted to 1 ml with the addition of extra HEPES buffer. The liposomes were vortexed for 1 min and hydrated overnight at 4°C. All the cationic liposome preparations were further sonicated before use as described earlier.³²

Preparation of liposomes and DNA complexes

The liposomes/DNA complexes were prepared as earlier described.³² Briefly, various amounts of liposomes and DNA were mixed together in 1 ml sodium lactated Ringer's buffer (B Braun Melsungen AG, Melsungen, Germany) and the complexes were allowed to form for a minimum of 15 min at room temperature before being used for gene delivery.

Transfection of cancer cell lines in vitro

The human cancer cell lines A549, HepG2, CNE-2, KZ2, and SW837 were grown in RPMI 1640 supplemented with 10% fetal calf serum. The human breast cancer cell line MCF-7 was propagated in RPMI 1640 with 10% FBS and 10 g/ml bovine insulin. For transfection, 1 ´ 10⁵ cells were seeded into each well of a six-well plate (Nunc, Roskilde, Denmark). After culturing for 12 h, cells were washed twice with PBS. Freshly prepared liposome/ DNA complexes were added into each well and incubated at room temperature for 15 min. For transfection performed in the presence of various concentrations of serum, FBS (Hyclone Laboratories, Logan, UT, USA) or mouse serum (obtained from BALB/c mice) was added to the DNA-liposome solution to the final concentration as indicated in the text. One ml of sodium lactated Ringer's buffer was then added into each well and incubated for an additional 2 h at 37°C. At the end of 2 h, 2 ml of 20% FBS was added into each well and the cells were assayed after 24 h.

Infection of HepG2 cells in vitro

The adenovirus containing the luciferase gene was obtained from Dr Matt Cotton (IMP, Vienna, Austria). The adenoviruses were amplified using human 293 cells. For infection with the adenoviruses, 1 ´ 10⁵ cells were seeded into each well of a six-well plate 24 h before infection and either 10, 100 or 1000 virus particles per cell were then added to each well. Cells were harvested 48 h following infection. Luciferase activity and protein assays were performed as described below.

Gene transfer in vivo

The direct injection of DNA-liposome complexes into the spleens of mice has been previously described.³² Briefly, 20 g of pCMV-luciferase DNA was injected directly into the spleen of each mouse. Expression of the luciferase gene in the spleens of the recipient mice were assayed 24 h following injection as previously described. The charge ratio (liposome to DNA) used is 0.4.

Assay for luciferase activity

Cells were harvested from six-well plates 24 h following transfection, washed, resuspended in 120 l Tris-HCl (pH 7.8), and freeze-thawed three times. Cell debris was discarded following centrifugation at 14 000 g at 4°C for 10 min. One hundred l of the supernatant collected was used for assaying luciferase activity using the Auto-Lumat LB952 luminometer (EG&G Berthold, Bad Wildbad, Germany). Five l of the collected supernatant was employed for the determination of protein concentration using the BioRad protein assay dye reagent (BioRad Laboratories, Hercules, CA, USA) with the Ultrospec 3000 UV/visible spectrophotometer (Pharmacia Biotech, Cambridge, UK).

Determination of zeta potential and size of cationic liposomes

The zeta potential of the newly synthesized cationic liposomes was determined by the ZetaSizer 3000HS (Malvern Instruments, Worcestershire, UK). Calibration was established using the -50 mV DTS50/50 standards from Malvern Instruments, as recommended by the manufacturer. Experimental samples taken from the 1 ml liposome stock of the various lipids following sonication for 5 min (prepared as described under the preparation of cationic liposome above) were diluted in 20 mM HEPES buffer (pH 7.8) to give a count rate of particles about 2500 KCps. The zeta potential of the sample was measured 10 times at 144 volts and 25°C using capillary cell.

Similarly, to measure the size of the liposomes, experimental samples taken from the 1 ml liposome stock of the various lipids following sonication for 5 min (prepared as described under the preparation of cationic liposome above) were diluted in 20 mM HEPES buffer (pH 7.8). Each sample was measured 10 times using a detector angle of 90° and the size obtained was analyzed using monomodal. The 204 nm ± 6 nm nanosphere size standards polystyrene polymer (Duke Scientific, CA, USA) was employed as standard to calibrate the instrument.

Acknowledgements

We thank Professor Leaf Huang for his comments on the manuscript. This research is supported by grants from the National Medical Research Council of Singapore and the Singapore National Science and Technology Board.

1 Smith AE. Gene therapy - where are we? Lancet 1999; 354 (Suppl.): 1-4,

2 Anderson WF. Human gene therapy. Nature 1998; 392 (Suppl.): 25-30, Article MEDLINE

3 Ledley FD. Pharmaceutical approach to somatic gene therapy. Pharm Res 1996; 13: 1595-1614, MEDLINE

4 Romano G et al. Latest developments in gene transfer technology: achievements, perspectives, and controversies over therapeutic applications. Stem Cells 2000; 18: 19-39, MEDLINE

5 Scherman D et al. Application of lipids and plasmid design for gene delivery to mammalian cells. Curr Opin Biotechnol 1998; 9: 480-485, MEDLINE

6 Mountain A. Gene therapy: the first decade. TIBTECH 2000; 18: 119-128,

7 Smith KT, Shepherd AJ, Boyd JE, Lees GM. Gene delivery systems for use in gene therapy: an overview of quality assurance and safety issues. Gene Therapy 1996; 3: 190-200, MEDLINE

8 Wolff JA et al. Direct gene transfer into mouse muscle in vivo. Science 1990; 260: 926-932,

9 Gilboa E. Retroviral gene transfer: applications to human therapy. Prog Clin Biol Res 1990; 352: 301-311, MEDLINE

10 Rosenfeld MA et al. Adenovirus-mediated transfer of a recombinant alpha 1-antitrypsin gene to the lung epithelium in vivo. Science 1991; 252: 431-434, MEDLINE

11 Miller DG, Adam MA, Miller AD. Gene transfer by retrovirus vectors occurs only in cells that are actively replicating at the time of infection. Mol Cell Biol 1990; 10: 4239-4242, MEDLINE

12 Herz J, Gerard RD. Adenovirus-mediated transfer of low density lipoprotein receptor gene acutely accelerates cholesterol clearance in normal mice. Proc Natl Acad Sci USA 1993; 90: 2812-2816, MEDLINE

13 Simon RH et al. Adenovirus-mediated transfer to the CFTR gene to lung of nonhuman primates: toxicity study. Hum Gene Ther 1996; 4: 771-780,

14 Ali M, Lemoine NR, Ring CJA. The use of DNA virus as vectors for gene therapy. Gene Therapy 1994; 1: 367-384, MEDLINE

15 Zhou X, Huang L. DNA transfection mediated by cationic liposomes containing lipopolylycine - characterization and mechanism of action. Biochim Biophys Acta 1994; 1189: 195-203, MEDLINE

16 Nabel EG et al. Gene transfer in vivo with DNA-liposome complexes: lack of autoimmunity and gonadal localisation. Hum Gene Ther 1992; 3: 649-656, MEDLINE

17 San H et al. Safety and short-term toxicity of a novel cationic lipid formulation for human gene therapy. Hum Gene Ther 1993; 4: 781-788, MEDLINE

18 Felgner PL et al. Lipofection: a highly efficient, lipid-mediated DNA-transfection procedure. Proc Natl Acad Sci USA 1987; 84: 7413-7417, MEDLINE

19 Gao X, Huang L. A novel cationic liposome reagent for efficient transfection of mammalian cells. Biochem Biophys Res Commun 1991; 179: 280-285, MEDLINE

20 Hawley NP et al. Lipofectamine reagent: a new, higher efficient polycationic liposome transfection reagent. Focus 1993; 15: 73-79,

21 Felgner JH et al. Enhanced gene delivery and mechanism studies with a novel series of cationic lipid formulations. J Biol Chem 1994; 269: 2550-2561, MEDLINE

22 Solodin I et al. A novel series of amphiphilic imidazolinium compounds for in vitro and in vivo gene delivery. Biochemistry 1995; 34: 13537-13544, MEDLINE

23 Lee ER et al. Detailed analysis of structures and formulations of cationic lipids for efficient gene transfer to the lung. Hum Gene Ther 1996; 7: 1701-1717, MEDLINE

24 Budker V et al. Protein/amphipathic polyamine complexes enable highly efficient transfection with minimal toxicity. BioTechniques 1997; 3: 139-147,

25 Kikuchi A et al. Development of novel cationic liposomes for efficient gene transfer into peritoneal disseminated tumor. Hum Gene Ther 1999; 10: 947-955, Article MEDLINE

26 Choi YH et al. Characterization of a targeted gene carrier, lactose-polyethylene glycol-grafted poly-l-lysine, and its complex with plasmid DNA. Hum Gene Ther 1999; 10: 2657-2665, MEDLINE

27 Singhal A, Huang L. Gene transfer in mammalian cells using liposomes as carriers. In: Wolff JA (ed). Gene Therapeutics. Birkhauser: Boston, 1994, 118-142.

28 Wheeler C et al. Converting an alcohol to an amine in a cationic lipid dramatically alters the co-lipid requirement, cellular transfection activity and the ultrastructure of DNA-cytofectin complexes. Biochim Biophys Acta 1996; 1280: 1-11, MEDLINE

29 Lasic DD, Templeton NS. Liposomes in gene therapy. Adv Drug Delivery Rev 1996; 20: 221-266,

30 Litzinger D, Huang L. Phosphatidylethanolamine liposomes: drug delivery, gene transfer and immunodiagnostic applications. Biochim Biophys Acta 1992; 1113: 201-227, MEDLINE

31 Legendre JY, Szoka FC. Delivery of plasmid DNA into mammalian cell lines using pH-sensitive liposomes: comparison with cationic liposomes. Pharm Res 1992; 9: 1235-1242, MEDLINE

32 Hui KM et al. Induction of alloreactive sytotoxic T lymphocytes by intra-splenic immunization with allogenic class I major histocompatibility complex DNA and DC-chol cationic liposomes. J Liposome Res 1994; 4: 1075-1090,

33 Hui KM, Ang PT, Huang L, Tay SK. Phase I study of immunotherapy of cutaneous metastases of human carcinoma using allogeneic and xenogeneic MHC DNA-liposome complexes. Gene Therapy 1997; 4: 783-790, MEDLINE

34 Bonte F, Juliano RL. Interaction of liposomes with serum proteins. Chem Phys Lipids 1986; 40: 359-372, MEDLINE

35 Farhood H, Bottega R, Epand RM, Huang L. Effect of cationic cholesterol derivatives on gene transfer and protein kinase C activity. Biochim Biophys Acta 1992; 1111: 239-246, MEDLINE

36 Lewis JG et al. A serum-resistant cytofectin for cellular delivery of antisense oligodeoxynucleotides and plasmid DNA. Proc Natl Acad Sci USA 1996; 93: 3176-3181, Article MEDLINE

37 Li S et al. Dynamic changes in the characteristics of cationic lipidic vectors after exposure to mouse serum: implications for intravenous lipofection. Gene Therapy 1999; 6: 585-594, MEDLINE

38 Gao X, Huang L. Cationic liposomes mediated gene transfer. Gene Therapy 1995; 2: 710-722, MEDLINE

39 Jaaskelainen I, Monkkonen J, Urtti A. Oligonucleotide-cationic liposome interactions. A physicochemical study. Biochim Biophys Acta 1994; 1195: 115-123, MEDLINE

40 Mortimer I et al. Cationic lipid-mediated transfection of cells in culture requires mitotic activity. Gene Therapy 1999; 6: 403-411, MEDLINE

41 Zabner J et al. Cellular and molecular barriers to gene transfer by a cationic lipid. J Biol Chem 1995; 270: 18997-19007, MEDLINE

42 Brunette E, Stribling R, Debs R. Lipofection does not require the removal of serum. Nucleic Acids Res 1991; 20: 1151,

43 Hofland HE, Sphephard L, Sullivan SM. Formation of stable cationic lipid/DNA complexes for gene transfer. Proc Natl Acad Sci USA 1996; 93: 7305-7309, MEDLINE

44 Yang JP, Huang L. Overcoming the inhibitory effect of serum on lipofection by increasing the charge ratio of cationic liposome to DNA. Gene Therapy 1997; 4: 950-960, MEDLINE

Figure 1 Reaction scheme for the synthesis of NCC1-NCC10.

Figure 2 Chemical structures of the newly synthesized cationic lipids, NCC1-NCC10.

Figure 3 Comparison of the efficiencies of transfection on HepG2 cells in vitro mediated by the cationic lipids NCC1-NCC10. 3 g DNA was complexed with the cationic liposomes formulated with cationic lipids synthesized in this report and DOPE at the molar ratio of 6:4 per transfection. The values shown are expressed as means ± s.d. (n = 4).

Figure 4 Comparison of the efficiencies of transfection mediated by a variety of nonviral vectors on human HepG2 cells. The cationic liposomes derived from DC-chol, NCC4 and NCC10 containing DOPE at the molar ratio of 6:4. 3 g pCMV-Luciferase DNA was employed per transfection. Cells were harvested 24 h following DNA-mediated gene transfer with the various nonviral reagents and the luciferase activities shown were means ± s.d. of triplicates.

Figure 5 Comparison of the efficiencies of transfection on human HepG2 cells following gene delivery with cationic liposomes and adenoviruses. The luciferase activities shown here were means ± s.d. of triplicates (n = 3).

Figure 6 Comparison of the efficiencies of transfection mediated by DC-chol, NCC4 and NCC10 on various human tumor cell lines. The cationic liposomes derived from DC-chol, NCC4 and NCC10 containing DOPE at the molar ratio of 6:4. 3 g pCMV-luciferase DNA was employed per transfection. The DNA-liposome complexes in 1 ml lactate buffer were added to 5 ´ 10⁵ cells per well in six-well plates in the absence of serum. After 2 h, 20% FBS serum was added. The cells were harvested 24 h after transfection. The luciferase activities shown were means ± s.d. of duplicates.

Figure 7 Effect of FBS on the efficiencies of transfection mediated by cationic liposomes. Cationic liposomes prepared from NCC4, NCC10 or DC-chol at the charge ratios of 0.4, 1.0, 2.6 and 4.0 were employed to transfect the pCMV-luciferase DNA into human HepG2 cells in the presence or absence of FBS. The luciferase activities shown were means ± s.d. of triplicates (n = 3).

Figure 8 Effect of mouse serum on the efficiencies of transfection mediated by cationic liposomes. Cationic liposomes prepared from NCC4, NCC10 or DC-chol at the charge ratios of 0.4, 1.0, 2.63 and 4.0 were employed to transfect the pCMV-luciferase DNA into human HepG2 cells in the presence or absence of mouse serum. The mouse serum was obtained from normal BALB/c mice and pooled. The luciferase activities shown were means ± s.d. of duplicates (n = 2).

Figure 9 NCC4 and NCC10 were less active at high FBS concentrations. Cationic liposomes prepared from NCC4, NCC10 or DC-chol at the charge ratios of 2.6 and 4.0 were employed to transfect the pCMV-luciferase DNA into human HepG2 cells in the presence or absence of FBS. The luciferase activities shown were means ± s.d. of triplicates (n = 3).

Table 1 Determinations of the sizes and zeta potentials, in the absence or presence of FBS, for cationic liposomes derived from NCC4, NCC10 and DC-chol

Table 2 Expression of luciferase in spleens of mice following in vivo gene delivery with various formulations of cationic liposomes at different charge ratios