*3.8. In Vivo Gene Therapy in KB Tumor Xenografts*

Among the FA-PEG-modified siRNA lipoplexes, LP-HAPC-2mol%FA-PEG2000 lipoplexes strongly suppressed the gene silencing effect in KB cells (Figures 2A and 7). Therefore, we evaluated the anti-tumor effect by direct injection into KB tumor xenografts with LP-HAPC-2mol%FA-PEG2000 lipoplexes of PLK1 siRNA. Injection of LP-HAPC-2mol%FA-PEG2000 lipoplexes with Cont siRNA or PLK1 siRNA was performed a total of three times, with 2 days between each injection. The anti-tumor effect on KB tumor xenografts was evaluated by measurement of tumor volume (mm3) and weight (mg) (Figure 10A,B). Intratumoral injections of LP-HAPC-2mol%FA-PEG2000 lipoplexes with PLK1 siRNA inhibited tumor growth compared with the Cont siRNA, although the difference was not significant. However, intratumoral injections of LP-HAPC-2mol%PEG2000 lipoplexes with PLK1 siRNA also inhibited growth compared with Cont siRNA, indicating that the in vivo anti-tumor effect by PEG-modified siRNA lipoplexes was not correlated with the in vitro one (Figure 8). This result suggested that the optimal formulation of PEG- or FA-PEG-modified cationic liposomes in FR-selective gene silencing effect might be different between in vitro and in vivo transfection studies.

**Figure 10.** In vivo siRNA therapy of KB tumor xenografts with FA-PEG-modified PLK1 siRNA lipoplexes in mice. When the average volume of the xenograft tumors reached 80 mm<sup>3</sup> (day 0), LP-HAPC-2mol%PEG2000 and LP-HAPC-2mol%FA-PEG2000 lipoplexes with 10 μg of Cont siRNA or PLK1 siRNA were injected directly into the tumor three times (day 0, 2, and 4). Tumor volume was measured at days 0, 2, 4, 6, and 8 (**A**). Tumor volume (%) was calculated as relative to each tumor volume at the day 0. The mice were sacrificed at day 8, and then the excised tumors were weighed (**B**). Each result represents the mean + SD (*n* = 3–4).

#### **4. Discussion**

Several studies have investigated the application of FA-PEG-modified cationic liposomes for siRNA delivery into FR-expressing cells [20–23], although, many studies have reported on the use of FA-PEG-modified liposomes for the delivery of anticancer drugs and plasmid DNA [36,37]. For siRNA delivery with FA-PEG-modified cationic liposomes, cationic lipids such as dioctadecyldimethylammonium chloride (DODAC) [20,21], 1,2-dioleoyl-3-trimethylammoniumpropane (DOTAP) [23], and 3β-(*N*-(*N* ,*N* -dimethylaminoethane)-carbamoyl)cholesterol (DC-Chol) [22] have been used, and FA-PEG-lipids were included at 0.5–2.5 mol% in the liposomal formulations. Previously, we reported that inclusion of 1 mol% FA-PEG2000-DSPE into lipid-based cationic nanoparticles composed of OH-Chol and Tween80 could enhance siRNA delivery into tumor cells [30]. However, there have been no reports about the effect of cationic lipid type in FA-PEG-modified cationic liposomes on FR-targeting. Therefore, in this study, we used three types of cationic cholesterol derivatives for the preparation of FA-PEG-modified siRNA lipoplexes, and we examined the effect of cationic lipid type in FA-PEG-modified cationic liposomes on FR-mediated siRNA transfection in tumor cells.

Cationic cholesterol derivatives contain the following parts: a cholesteryl skeleton, a cationic head group, and a linker bound between the cholesteryl skeleton and the cationic head group. The linker between the hydrophilic and hydrophobic parts influences the gene delivery efficacy of cationic liposomes [38]. Furthermore, the introduction of a hydroxyethyl group into the cationic head group of cationic cholesterol derivatives can enhance in vitro gene transfection by cationic liposomes [39–41]. Previously, we synthesized HAPC-Chol, OH-Chol, and OH-C-Chol as cationic cholesterol derivatives with a hydroxyethyl group in the cationic head group, and demonstrated

that cationic liposomes composed of their cationic cholesterol derivatives exhibited effective siRNA transfection activity [12–14,42]. Therefore, in this study, we used their cationic cholesterol derivatives, and prepared three types of FA-PEG-modified siRNA lipoplexes for the evaluation of FR-mediated siRNA transfection.

In in vitro transfection studies, FA-PEG-modification of LP-HAPC, LP-OH, and LP-OH-C could increase the gene silencing effect of siRNAs via efficient cellular uptake, compared with PEG-modification (Figures 2 and 4). In OH-Chol-based cationic liposomes, the increase in FA-PEG2000-DSPE in LP-OH-FA-PEG2000 increased the amount of siRNA taken up by the cells (Figure 4B). However, it also decreased the gene silencing activity (Figure 2C), indicating that increased FA-PEG-modification of LP-OH could increase cellular uptake of the siRNA lipoplexes via FR, but siRNA might not be efficiently delivered into the cytoplasm or not easily released from the siRNA lipoplexes after FR-mediated endocytosis. In LP-OH-C lipoplexes, the increase in FA-PEG2000-DSPE in LP-OH-C-FA-PEG2000 decreased the cellular uptake of the siRNA lipoplexes (Figure 4C), but it did not greatly decrease the gene silencing activity (Figure 2D), indicating that the LP-OH-C-1–3mol%FA-PEG2000 might deliver siRNA efficiently into cytoplasm after the endocytosis, although the increase of FA-PEG-modification in LP-OH-C decreased the cellular association. In contrast, the increase in FA-PEG2000-DSPE in LP-HAPC-FA-PEG2000 did not greatly affect the amount of siRNA taken up by the cells (Figure 4A) and the gene silencing effect (Figure 2A), suggesting that FA-PEG-modification of LP-HAPC could mediate cellular uptake via FR, and deliver siRNA efficiently into cytoplasm after the endocytosis. These results indicated that the cellular association via FR and gene silencing efficiency by FA-PEG-modified siRNA lipoplexes were strongly affected by the cationic lipid type in cationic liposomes, and HAPC-Chol will be a better cationic lipid in FR-mediated transfection by cationic liposomes, compared with OH-Chol and OH-C-Chol. However, it was not clear why the cationic lipid type in FA-PEG-modified cationic liposomes affected cellular association via FR and the gene silencing effect. The linker groups of cationic cholesterol derivatives control the flexibility of the cationic head groups. HAPC-Chol and OH-C-Chol have a carbamate-type linker, and OH-Chol has a carboxamide-type linker. Previously, we reported that the difference in the linker group between carboxamide and carbamate in cationic cholesterol derivatives affected cellular association with siRNA nanoplexes [42]. In addition, HAPC-Chol has a slightly longer linker compared with OH-Chol and OH-C-Chol. Therefore, we speculated that the difference in the linker group of these cationic cholesterol derivatives might affect the interaction between the cationic head group on cationic liposomes and the anionic cell membrane and/or endosomal escape after transfection with FA-PEG-modified siRNA lipoplexes.

In our study, the inclusion of PEG2000-DSPE or FA-PEG2000-DSPE into the formulations of LP-HAPC, LP-OH, and LP-OH-C were optimal at 2 mol%, 1 mol%, and 2 mol%, respectively, for selective FR-mediated gene silencing (Figures 2 and 7). In addition, LP-HAPC-2mol%FA-PEG2000, LP-OH-1mol%FA-PEG2000, and LP-OH-C-2mol%FA-PEG2000 lipoplexes with PLK1 siRNA selectively suppressed the growth of tumor cells via down-regulation of PLK1 mRNA (Figures 8 and 9), indicating that their FA-PEG-modified siRNA lipoplexes were selectively taken up by the FR-expressing cells, and then suppressed the expression of target genes in the cells. However, in in vivo transfection, both LP-HAPC-2mol%PEG2000 and LP-HAPC-2mol%FA-PEG2000 lipoplexes with PLK1 siRNA inhibited tumor growth compared with Cont siRNA, indicating that the in vivo anti-tumor effect of PEG-modified siRNA lipoplexes was not correlated with the in vitro one. We speculated that siRNA lipoplexes after intratumoral injection were subjected to a strict environment surrounded by tumor cells, resulting in induction of non-specific uptake into tumor cells. Therefore, in in vivo transfection, PEG-modification of LP-HAPC with more PEG2000-DSPE might be needed for suppression of the anti-tumor effect by PEG-modified lipoplexes with PLK1 siRNA. However, we reported previously that inclusion of 1 mol% FA-PEG2000-DSPE into lipid-based cationic nanoparticles composed of OH-Chol and Tween 80 could efficiently deliver siRNAs into KB cells, and FA-PEG-modified nanoplexes of HER-2 siRNA introduced by intratumoral injection significantly inhibited the tumor growth of KB xenografts compared with

Cont siRNA, but PEG-modified nanoplexes did not [30], indicating that the in vivo anti-tumor effects by PEG-modified and FA-PEG-modified siRNA nanoplexes were well correlated with the in vitro ones. These findings suggested that the FR-selective gene silencing by PEG- and FA-PEG-modified cationic liposomes might also be affected by the liposomal formulation. Further studies must be performed to investigate liposomal formulations to improve FR-selective transfection in vivo. In this study, we selected an intratumoral injection as a route of administration for FA-PEG-modified siRNA lipoplexes; however, in future study, we will need to evaluate anti-tumor effect after intravenous injection of FA-PEG-modified siRNA lipoplexes.
