PEGylation potentiates hepatoma cell targeted liposome-mediated in vitro gene delivery via the asialoglycoprotein receptor

1 Introduction

Hepatocellular carcinoma (HCC) is the second leading cause of cancer-associated deaths worldwide [1] and is particularly prevalent in sub-Saharan Africa and East Asia [2], where hepatitis B, its main cause, is endemic [3]. Indeed, >80% of HCC patients are seropositive for hepatitis B virus [4]. This condition is associated with a particularly poor prognosis, and current treatment options include surgical resection, liver transplantation, radiofrequency ablation, trans-arterial chemoembolization and systemic targeted sorafenib treatment [5], [6]. However, novel treatment modalities involving intervention at the genetic level are being actively investigated. These include an elegant gene therapy approach in which suicide genes are delivered to hepatoma cells by replicating virus vectors [7]. Such approaches also address the requirement for expression to be tumor-selective, as impaired liver function is often associated with HCC, and transgene expression should be limited to tumor cells alone without affecting normal hepatocytes [8]. Transcriptional targeting has therefore emerged as a promising approach in achieving this goal. Thus, tumor-specific α-fetoprotein enhancers [9] and core promoters [7] have been employed to drive expression of genes of interest. The high efficiency achieved by viral vectors is, however, often tempered by strong immune responses [10] and insertional mutagenesis [11]. This has given impetus to the further development of, as yet, less efficient but generally safer non-viral vectors [12]. Liposomal vehicles account for most of the non-viral vectors currently undergoing clinical trials worldwide [13], and their formulation flexibility, ease of manufacture and relatively low cost have ensured their ongoing popularity.

The asialoglycoprotein receptor (ASGP-R), which is highly expressed on the basolateral side of the plasma membrane of hepatocytes and on hepatocyte-derived hepatoma cells and which displays high affinity for D-galactopyranosyl and N-acetylgalactosaminyl moieties [14], [15], [16], is a convenient target for the receptor-mediated uptake of lipoplexes carrying therapeutic genes [17]. The lipoplexes directed to the ASGP-R are internalized by clathrin-dependent receptor-mediated endocytosis into endosomes. The endosomal membrane anionic lipids interact with the cationic components of the lipoplexes forming charge-neutralized ion pairs, thereby displacing the cargo DNA and releasing it into the cytoplasm [18], [19]. Endosomal escape is also promoted by the helper lipid dioleoylphosphatidylethanolamine (DOPE) when present in lipoplexes in high concentration [18], [20]. The cone-shaped DOPE favors non-bilayer hexagonal phase formation, which destabilizes the endosomal membrane [21]. Although the ligand asialoorosomucoid, which displays D-galactopyranose units at the non-reducing ends of its heteroglycan appendages, has been tagged to liposomes for ASGP-R-mediated gene delivery [22], [23], several simpler constructs have emerged in which the galactopyranosyl moiety is affixed directly to one of the liposome components. Hence, liposomes that incorporate cholesteryl-β-D-galactopyranoside [24], lactosyl DOPE [25] and cholesteryl-3β-N-(4-aminophenyl-β-D-galactopyranosyl) carbamate [26] all target DNA lipoplexes to the human HCC cell line HepG2. Transfection experiments conducted in vivo with cationic liposomes containing cholesten-5-yloxy-N-(4-((1-imino-2-D-thiogalactosylethyl)amino)butyl formamide (Gal-C4-Chol) reveal that lipoplexes measuring 141 nm in diameter achieve transfection levels in parenchymal cells significantly higher than in non-parenchymal cells [27]. In related studies siRNA lipoplexes decorated with galactopyranosyl units attached to the distal end of DSPE-PEG₂₀₀₀ have successfully directed siRNA to hepatocytes in vivo [28], [29]. Lipoplexes (liposome-DNA complexes) readily associate with serum proteins [30] and are rapidly marked for elimination by the reticuloendothelial system [31], thus severely limiting the availability of complexes at the intended target tissue. This may be mitigated, to a large extent, by anchoring poly(ethylene glycol)-2000 (PEG₂₀₀₀) to the liposomal membrane bilayer [32]. The biocompatible, hydrophilic, protective layer provided by PEG₂₀₀₀ limits opsinization and increases particle circulation times [31]. Nevertheless, lower transfection efficiencies are commonly observed with untargeted PEGylated lipoplexes. This PEG dilemma arises largely through reduced membrane fusion and limited non-bilayer intermediate formation [33].

In the present study we have prepared galactosylated cationic liposomes constituted with the cytofectin Chol-T, DOPE, cholest-5-en-3-yl-2-[2-(2-{2-[4-(β-D-galactopyranosyl-1-oxymethyl)-1H-1,2,3-triazol-1-yl]ethoxy}ethoxy)ethoxy]ethylcarbamate (1 Figure 1), with and without 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-2000 (DSPE-PEG₂₀₀₀). Studies conducted in the ASGP-R-positive human hepatoma cell line, HepG2, indicate that PEGylation enhances ASGP-R-mediated transfection by galactosylated lipoplexes, while in the ASGP-R-negative human embryo kidney cell line HEK293 [34], transfection levels were significantly lower.

Figure 1:

(A) Galactopyranosyl cholesteryl derivatives 1, cholest-5-en-3-yl-2-[2-(2-{2-[4-(β-D-galactopyranosyl-1-oxymethyl)-1H-1,2,3-triazol-1-yl]ethoxy}ethoxy)ethoxy]ethylcarbamate; 2, cholest-5-en-3-yl-2-[2-(2-{2-[4-(2,3,4,6-tetra-O-acetyl-β-D-galactopyranosyl-1-oxymethyl)-1H-1,2,3-triazol-1-yl]ethoxy}ethoxy)ethoxy]ethylcarbamate. (B) Schema of galactosylated, PEGylated cationic liposome membrane bilayer.

2 Experimental

3 Results and discussion

The galactopyranosyl cholesteryl derivative cholest-5-en-3-yl-2-[2-(2-{2-[4-(β-D-galactopyranosyl-1-oxymethyl)-1H-1,2,3-triazol-1-yl]ethoxy}ethoxy)ethoxy]ethylcarbamate (1, Figure 1), which incorporates a hydrophilic 24.1 Å spacer element separating the sugar component from the sterol, was obtained by the deacetylation of its peracetylated precursor cholest-5-en-3-yl-2-[2-(2-{2-[4-(2,3,4,6-tetra-O-acetyl-β-D-galactopyranosyl-1-oxymethyl)-1H-1,2,3-triazol-1-yl]ethoxy}ethoxy)ethoxy]ethylcarbamate (2, Figure 1) with K₂CO₃ in methanol. This, in turn, had been synthesized by an alkyne-azide cycloaddition (CuAAC) ‘click’ reaction between the peracetylated galacto alkyne, 2,3,4,6-tetra-O-acetyl-1-(2′-propargyl)-β-D-galactopyranose, and cholest-5-en-3-yl 2-{2-[2-(2-azidoethoxy)ethoxy] ethoxy} ethylcarbamate.

Liposomes were formulated to contain 10 mol% of 1 or 2 with or without 5 mol% DSPE-PEG₂₀₀₀ (Table 1) and exhibited hydrodynamic diameters in the 80–200 nm range (Table 1 and Figure 2). PEGylated liposomes were almost 50% smaller than their nonPEGylated counterparts due to inter-bilayer repulsion induced by PEG on the liposome surface [40]. Cryo-TEM revealed that liposome preparations comprised spherical and ellipsoidal unilamellar vesicles (Figure 2). PEG₂₀₀₀ was selected for this study over larger or smaller molecular sizes of PEG to achieve longer circulation times and a balance between steric hindrance to prevent opsinization [41] and permitting ligand-receptor interaction for receptor-mediated uptake of lipoplexes. Inclusion in liposome formulations at 5 mol% was to ensure the formation of an effective steric barrier yet avoiding membrane disruption and possible micelle formation, which may occur at higher PEG concentrations [42], [43], [44]. Moreover, at 5 mol% PEG, chains are largely arranged in a ‘brush regime’, which affords thermodynamic stability to liposomes [45] and resistance to clearance by macrophages [46]. The galactosyl ligand was incorporated at the 10 mol% level to favor lipoplex interaction with the hepatocytic ASGP-R [19] over binding to a similar C-type lectin located on scavenger nonparenchymal Kupffer cells [47] of the liver, which is known to occur at higher galactosyl grafting densities [48].

Figure 2:

Cryo-transmission electron micrographs of liposomes and lipoplexes. Scale bar=100 nm.

The association between liposomes and plasmid DNA was followed by two separate and independent assays. Thus, in gel retardation assays, plasmid DNA was fully liposome-associated at N/P ratio=3:1 (Figure 3A–D), while in ethidium displacement experiments, cargo DNA was maximally condensed at N/P ratios=2:1–3:1 (Figure 4). Results also revealed that plasmid DNA was condensed to a higher degree by unPEGylated liposomes 1 and 2, with 70% ethidium displacement, than by PEGylated liposomes, which was accompanied by a 50% ethidium displacement (Figure 4). The integrity of liposome-bound plasmid DNA in the presence of HBS was monitored by agarose gel electrophoresis. All liposomes offered partial protection to the plasmid DNA in the N/P range 1:1–3:1 (Figure 5). Gels revealed that plasmid DNA was largely in the nicked, relaxed closed circular form after exposure to HBS. This topological form of pCMV-luc DNA has been shown to yield higher transfection activities than either the superhelical or linear conformers when delivered by cationic liposomes to cells in vitro [49]. It is also likely that superhelical DNA may be nicked during its intracellular journey to the nucleus [50].

Figure 3:

Gel retardation assay of binding between liposomes and plasmid DNA. Incubation mixtures in HBS (8 μL) contained pCMV-luc plasmid DNA (0.5 μg) and varying amounts of liposomes to achieve N/P ratios+1:1, 2:1 and 3:1 in lanes 2–4, respectively. Lane 1 contained plasmid DNA alone. (A) Liposomes 1, (B) 1PEG, (C) 2, and (D) 2PEG. Superhelical DNA is indicated (#).

Figure 4:

Ethidium bromide intercalation assay. Incubation mixtures in HBS (104 μL) containing EtBr (2 μg) and pCMV-luc plasmid DNA (0.5 μg) were treated with increasing amounts of liposome suspension (0.5 μg aliquots).

Figure 5:

Serum nuclease protection assay of lipoplex-associated DNA. Mixtures (10 μL) contained pCMV-luc DNA (0.5 μg), liposomes (N/P ratios+1:1, 2:1 and 3:1 in lanes 3–5 and 6–8) and FBS (10%, v/v) in HBS. Lane 1: untreated plasmid DNA; lane 2: plasmid DNA in the presence of FBS. Relaxed closed circular (*) and superhelical (#) plasmid DNA are indicated.

Cell viability studies after exposure to liposome-DNA complexes under transfection conditions (N/P=3:1) were conducted in HEK293 (ASGP-R-negative) and in HepG2 (ASGP-R-positive) cell lines. Results indicate that PEGylated lipoplexes were better tolerated (>70% viability) in both cell lines than nonPEGylated complexes (50–60% viability) (Figure 6). Transfection experiments revealed that both lipoplexes 1 and 1PEG achieved higher transfection activity in HepG2 cells than in the HEK293 cell line, with 1PEG transfection levels approximately 3× higher than those achieved by lipoplex 1, while transfection activities in the presence of the ASGP-R ligand, asialofetuin, were reduced to background levels. Lipoplexes 2 and 2PEG were less effective with no discernible differences between the two cell lines (Figure 7). This may be attributed largely to the acylated sugar hydroxyl groups, which may prevent correct binding of the galacto moiety to the ASGP-R. It has been shown that the specificity of ASGP-R for terminal D-galacto residues results from H-bonding between the galacto 3-OH and 4-OH groups with carboxylate and amide side chains in the ASGP-R binding site [17]. High transfection levels achieved by PEGylated lipoplex 1PEG in HepG2 cells may be attributable, in part, to the looser association of the cargo DNA, which may therefore be more readily released upon cell entry of the lipoplex by receptor mediation.

Figure 6:

Cytotoxicity assay of lipoplexes in HEK293 and HepG2 cells. Cells in 48-well plates bathed in 300 μL of complete medium containing antibiotics and FBS (10%) were treated with lipoplexes at N/P ratio=3:1 (0.5 μg plasmid). Data are presented as means±SD (n=3). **P<0.01 for indicated comparisons.

Figure 7:

Transfection studies of lipoplexes in HEK293 and HepG2 cells. Cells bathed in complete medium (300 μL, 48-well plates) were exposed to lipoplexes (N/P ratio=3:1) containing 0.5 μg pCMV-luc plasmid DNA. Control: untreated cells; Control+DNA: cells exposed to naked plasmid DNA (0.5 μg). In competition assays, wells contained 200 μg asialofetuin. Data are presented as means±SD (n=3). Differences between 1/1PEG and 1PEG/2PEG are statistically very significant (***P<0.001).

4 Conclusion

We report on PEGylated and nonPEGylated liposomes decorated with D-galacto and peracetylated D-galacto residues appended to 24.1 Å spacer elements and their vehiculation of pCMV-luc plasmid DNA in transfection experiments in vitro. Results confirm that PEGylated galacto liposomes very effectively direct reporter DNA to the human hepatoma cell line HepG2 largely by ASGP-R mediation. Physicochemical studies also indicate that lipoplexes fall within an acceptable size range for extravasation in vivo and that DNA is adequately protected from serum nucleases. Moreover, DNA appears to be more tenuously associated with PEGylated liposomes than with their unPEGylated counterparts. In summary, PEGylated ASGP-R-targeted liposomes reported herein possess the required characteristics for hepatotropic gene delivery and may be considered for further development in an in vivo environment.