**2.3 Linker bond**

298 Non-Viral Gene Therapy

ranging from C5:0 to C18:1, but oleyl, lauryl, myristyl, palmityl and stearyl, have been the most researched ones (Niculescu-Duvaz et al., 2003). A common variation is the use of branched (Ferrari et al., 2002), acetylenic (Fletcher et al., 2006) chains and *cis*-

It is commonly believed that cationic lipids have one to four hydrocarbon chains. Several studies have also shown that incorporating aliphatic chains with different numbers can improve transfection efficiency potentially by promoting endosomal escape (Felgner et al., 1987; Tang & Hughes, 1999a; Gaucheron at al., 2002; Zhi et al., 2010). Cationic lipids with double-chain hydrocarbons in the hydrophobic domain represent the majority of cationic lipids synthesized so far. Cationic lipids containing two aliphatic chains such as DOTMA and DOTAP, are among the most active lipids for systemic gene delivery. However, Tang et al. (Tang & Hughes, 1999a) demonstrated that 6-lauroxyhexyl ornithinate (LHON) with one tail was more efficient and of lower cytotoxicity compared to DOTAP. Generally speaking, for aliphatic chains, single-tailed and three-tailed cationic lipids are better known as surfactants because of their ability to form micelles in solution, but they are more toxic and less efficient than their double-tailed counterparts. Usually, cationic lipids with doubletailed hydrocarbons are capable of forming liposomes by themselves or with a helper phospholipid. Therefore, most of the aliphatic chains in the cationic lipids are double-tailed. It is generally agreed that the length and saturation of the aliphatic chains incorporated into cationic lipids significantly affect their transfection efficiency. In order to gain chain length– activity correlation, Felgner et al. (Felgner et al., 1994) studied a series of hydroxyethyl quaternary ammonium lipids with myristoyl (diC14:0, DMRIE), palmitoyl (diC16:0, DPRIE), stearoyl (diC18:0, DSRIE), and oleoyl (diC18:1, DORIE) chains. They observed that a comparison of vectors based solely on the lengths of the two aliphatic chains led to identify the order C14:0 > C18:1 > C16:0 > C18:0. Our study on double-chain cationic lipids also showed increasing transfection efficiency with decrease of the chain length (Liu et al., 2008). It was therefore proposed that cationic lipids with shorter chain length (for saturated chains) were generally important for acquiring high transfection efficiency, since they are responsible for membrane fluidity and good lipid mixing within the bilayer. Beyond that, the best chains in terms of benefit to transfection are frequently the unsaturated ones. The overwhelming majority of results showed that the unsaturated C18:1 oleyl was the optimal aliphatic chain,

which was frequently the best choice for good transfection (Fletcher et al., 2006).

et al., 2001a, 2001b, 2001c, 2001d).

Some new vectors were designed to covalently connect some special moieties in the hydrophobic chains, in order to get the relationship between hydrophobic chains and transfection efficiency. Jacopin et al. (Jacopin et al., 2001) synthesized a glycosylated analogue (Fig. 6.) of the dialkylamidoglycylcarboxyspermines, which formed stable particles at low charge ratio and was efficient for gene delivery. Many groups also reported a few other glycosylated cationic bolaamphiphiles similar to the compound (Fabio et al., 2003; Brunelle et al., 2009). In addition, the fluorinated part of the hydrophobic chain can also influence the transfection efficiency of cationic lipids *in vivo* and *in vitro.* Many varieties of fluorinated cationic lipids have been developed as transfecting agents, which are very efficient in compacting DNA and delivering genes into cells *in vivo* and *in vitro* (Gaucheron

In the steroid groups, cholesterol is by far the most frequently encountered and used as an alternative to aliphatic chains because of its rigidity, as well as its endogenous biodegradability and fusion activity. An example is cationic lipid 'GL-67' (Fig. 7.), which has been found to be particularly efficient for gene transfer to cultured cells and in murine lungs

monounsaturated alkyl chains (Bennett at al., 1997).

For lipids without a backbone, the linker bond that acts as a connector between the hydrophobic and cationic headgroup domains can affect the transfection efficiency, biodegradability and stability of cationic lipids. Linker bonds are commonly ether, ester (Leventis & Silvius, 1990), amide (Behr et al., 1989) or urethane (or carbamate) (Lee et al., 1996) groups (Koynova & Tenchov, 2010), but other groups such as redox-sensitive disulphide (Byk et al., 1998, 2000) have also been employed (Fig. 1.). Cationic lipids with ether bonds such as DOTMA in the linker domain generally render better transfection efficiency, but they are too stable to be biodegraded thus may cause higher toxicity. Compared with ether bonds, although cationic lipids with ester bonds such as DOTAP are more biodegradable and associated with less cytotoxicity in cultured cells (Leventis & Silvius, 1990; Choi et al., 2001), those with ester may also decrease the stability of liposomes in systemic circulation.

The chemistry of the linker has most often been of the carbamate or amide variety, both of which are chemically stable and biodegradable, and cationic lipids with these linkers could be used as efficient gene delivery carriers (Ren et al., 2001; Liu et al., 2005a, 2005b, 2008). A typical example of cationic lipid with carbamate linker is DC-Chol, which was the first lipid used in clinical trials because of its combined properties of transfection efficiency, stability, and low toxicity (Gao & Huang, 1995). As well known, when incorporating a carbamate group into the linker, it may therefore be hoped that the pH drop will act as a trigger,

Cationic Liposomes in Different Structural Levels for Gene Delivery 301

DOPE often presents a super synergistic effect when used in cationic liposomes, because DOPE destabilized lipid bilayers, and it was believed to be involved in endosomal disruption (Litzinger & Huang, 1992), allowing the release of DNA into the cytosol (Farhood et al., 1995) and leading to mixed bilayers (Scarzello et al., 2005). Most studies have shown that lipoplexes containing the non-bilayer-phase-preferring lipid DOPE or cholesterol would promote HCII organization (Zuhorn et al., 2005). A transition from the LCα phase to the HCII phase could be expected by increasing weight fraction of DOPE, via controlling the spontaneous radius of curvature "Ro" of the lipid layers, favored by the elastic free energy (Safinya, 2001). Another helper lipid, cholesterol, could also promote HIIC organization as DOPE. It has been proved that *in vivo* applications cholesterol was a more effective helper

Koltover et al. (Koltover et al., 1998) disclosed the reason in the level of phase transition through synchrotron small-angle X-ray scattering (SAXS) and optical microscopy to show the phase transition from LαC to HIIC induced by DOPE via controlling the spontaneous curvature *C*o = 1/*R*o of the lipid monolayer. It has been concluded that DOPE facilitates endosomal escape by forming an unstable inverted hexagonal phase at the endosomal pH that destabilizes both the complex and the endosomal membrane. But in a recent study (Leal et al., 2010), they developed CL-siRNA complexes with a novel cubic phase nanostructure exhibiting efficient silencing at low toxicity by using glycerol monooleate other than DOPE as the helper lipid. The inverse bicontinuous gyroid cubic nanostructure was unequivocally established from synchrotron X-ray scattering data, while fluorescence

Tubes of lipoplexes containing DOTAP/MOG, DOTAP or DOTAP/PC, and DOTAP/DOPE were observed in freeze-fracture electron micrographs. The tubes were extremely short and appeared bead-like in lipoplexes containing DOTAP/MOG, slightly longer in those containing DOTAP or DOTAP/PC, and extensively elongated in DOTAP/DOPE lipoplexes (Xu et al., 1999). The spaghetti-like structures, occurring at DNA: lipid concentrations which were used during transfection and their diameter came closest to the diameter of the nuclear pores, may be the active cationic lipoplexes (Zhdanov et al., 2002). In the study of the structure and morphology of DC-Chol–DOPE/DNA complexes it was found the existence of cluster-like aggregates made of multilamellar DNA/lipid domains coexisting with other multilamellar lipoplexes or, alternatively, with DNA-coated vesicles (Amenitsch et al., 2010). The further study showed that DC-Chol-DOPE/DNA lipoplexes preferentially used a raft mediated endocytosis, while DOTAP-DOPC/DNA systems were mainly internalized by not specific fluid phase macropinocitosys. Most efficient multicomponent lipoplexes, incorporating different lipid species in their lipid bilayer, can use multiple endocytic pathways to enter cells. Their data demonstrated that efficiency of endocytosis was

microscopy revealed colocalization of lipid and siRNA in complexes.

Fig. 8. Chemical structure of DOPE and DOPC.

lipid than DOPE (Lasic, 1997).

**3.1 DOPE** 

disconnecting the hydrophobic and hydrophilic portions of the lipoplex, and thereby to release DNA after entering endosomes in cell because of the pH decreasing (Liu et al., 2005a, 2008). We synthesized a series of carbamate-linked cationic lipids for liposome-mediated gene delivery, which proved to have good gene transfection properties (Fig. 4.).

It is familiar to chemists that compounds comprising redox-sensitive disulphide bonds is stable chemically as long as no reducing agents, and it is expected that these disulphidelinked lipids can keep stable in the circulation system while decomposing to release DNA after entering endosomes in cells (in a similar manner as the pH-sensitive systems) (Tang & Hughes, 1999b). Byk et al. (Byk et al., 1998, 2000) prepared a series of lipopolyamines that harbor a disulfide bridge within different positions in the backbone of the lipids as biosensitive function. They found that an early release of DNA during or after penetration into cells, probably promoted by reduction of a disulfide bridge placed between the polyamine and the lipid, implied a total loss of transfection efficiency.

In addition, structural variations at the linker region such as length, the specific type of chemical bonds and the relative position of the hydrocarbon chains can affect the transfection efficiency, biodegradability and stability of cationic lipids (Fujiwara et al., 2000). The level of hydration and toxicity of the lipid can also be determined by the length of the linker (Floch et al., 2000). In a word, the use of linkers incorporating functional groups that are cleavable on shorter time scales and under specific stimuli is however of emerging interest, as DNA release may here be facilitated by a triggered decomplexation mechanism.

From the chemistry point of view, the structure of cationic compounds is an important factor for their transfection activity and toxicity. Some common conclusions can be achieved by comparing the different structures and their transfection activity in the same family or different families of lipids. The transfection efficiency is not only determined by one domain of cationic lipids, but also depends on the combination of them (Tang & Hughes, 1999a). In general, it seems when researchers design cationic compounds for gene delivery, the balances between the opposite factors including fluidity and rigidity, symmetry and asymmetry, saturation and unsaturation, linearity and branching, short chain and long chain, hydrophilicity and lipophilicity of compounds should be taken into serious consideration.
