Gene delivery to Her-2+ breast cancer cells using a two-component delivery system to achieve specificity

Abstract

Current liposomal gene delivery systems predominately utilize cationic lipids, which efficiently bind and deliver DNA plasmid, but also result in nonspecific gene expression in lung and liver tissue. To improve specificity, a two-component delivery strategy employing neutral liposomes was used to target breast cancers positive for the human epidermal growth factor receptor 2 (Her-2). The first component consisted of plasmid DNA condensed with cationic polyethylene glycol (PEG) modified polylysine (PL/DNA). The second component was a neutral Her-2 targeting liposome conjugated to the pore-forming protein, Listeriolysin O (LLO). Independently, PL/DNA delivery resulted in low expression of plasmid DNA. However, when PL/DNA and LLO/liposomes co-localized within an endosome, LLO disrupted endosome integrity, leading to cytoplasmic delivery and expression of the plasmid. When used to deliver a plasmid encoding the luciferase gene, this two-component system resulted in gene expression that was 268-fold greater in Her-2 positive cells than in Her-2 negative cells.

Background

Liposomal delivery of nucleic acids typically involves the use of cationic lipids.¹ The positively charged liposomes interact with negatively charged nucleic acids, compacting the DNA or RNA and protecting it from enzymatic degradation in vivo. Once introduced into the body, cationic liposomes interact strongly with cell membranes, internalize into endosomes and disrupt the endosomal membrane, allowing for nucleic acid delivery to the cytoplasm.² The ability of cationic liposomes to bind, protect and deliver nucleic acids has pushed the field of liposomal gene delivery away from neutral and negatively charged liposomes which do not possess these attributes. While delivery is efficient from cationic liposomes, gene expression after systemic delivery is almost always observed predominantly in the lungs and liver with much lower expression in tumor tissue.^{3, 4} Attempts to minimize lung-liposome interactions by shielding particles with polyethylene glycol (PEG) do not always improve biodistribution and can greatly reduce gene expression.^{3, 5}

Although liposomes can be injected directly into a solid tumor, systemic administration through intravenous injection will most likely be necessary to treat metastatic cancer. Many studies employing cationic liposome and viral gene delivery systems have relied upon intratumoral injection to avoid nonspecific gene expression in the lung and liver tissue.⁶ However, if a solid tumor is accessible for injection, the best treatment option is often surgical removal of the tumor. Liposomal gene delivery would be most useful for targeting metastatic tumors that cannot be easily located and excised, and which would require a tumor-specific liposomal gene delivery system that can be systemically administered. In order to aggressively target distant metastases with potent genes that lead to cell death, it will be necessary to achieve specific delivery to the tumor cells. Genes such as IL-12 or TNF-α which result in immune activation or cell apoptosis, respectively, lead to intolerable side effects if delivered nonspecifically.^7–9 Given that cationic liposomes accumulate heavily in the lung and liver, other options must be developed.

We propose a delivery system that utilizes two components in order to increase specificity for tumors that are positive for the human epidermal growth factor receptor 2 (Her-2), a growth receptor overexpressed in 20–30 % of breast cancers.¹⁰ The first component is PEGylated cationic polylysine which condenses the DNA (PL/DNA) and internalizes nonspecifically into cells. The second component is a neutral liposome conjugated to trastuzumab and bound to the pore-forming protein, Listeriolysin O (LLO) via a disulfide bond (LLO/liposomes). LLO/liposomes bind with high specificity to Her-2 positive cells and are subsequently endocytosed.¹¹ The reducing conditions of the endosome release disulfide bound LLO from the liposomes, leading to LLO-induced pore formation in the endosomal membrane.¹² If LLO/liposomes co-localize with internalized PL/DNA inside an endosomal compartment, DNA should be able to access the cytoplasm through the pores created by LLO. Without LLO, much of the PL/DNA will travel through the endocytotic pathway eventually being broken down in the lysosomes. Although PL/DNA internalization is non-specific, selective Her-2 targeting liposomes and the requirement that both components must co-localize to allow gene expression could confer specificity to the delivery system.

In this paper, the parameters of the two component system were varied in vitro to characterize and maximize delivery to Her-2 positive cells. The delivery system was also tested on a panel of cells that included both Her-2 positive and negative cells. Finally, using fluorescent microscopy, the internalization of labeled PL/DNA and LLO/liposomes was visualized. High throughput imaging and analysis were utilized to compare the degree of endosome co-localization with gene expression levels in an attempt to elucidate the mechanism of delivery.

Materials and Methods

Materials

1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2-dioleoyl-3-trimethylammoniumpropane (DOTAP) and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[PDP-poly(ethylene glycol)-2000] (DSPE-PEG(2000)-PDP) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). 1,2-Distearoyl-sn-glycero-3-phosphoethanolamine-n-[poly(ethylene glycol)-3400]-N-hydroxy succinamide (DSPE-PEG(3400)-NHS), and N-hydroxy succinamide PEG (5,000) (NHS-PEG) were purchased from Nanocs, Inc. (New York, NY, USA). Lissamine rhodamine B 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine (RhoPE) and Oregon Green 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine were purchased from Life Technologies (Grand Island, NY, USA). Luciferase plasmid DNA (5.9 kb) with a CMV promoter was a generous gift from Valentis Inc. (Burlingame, CA, USA).¹³ Column purification reagent, CL-4B Sepharose gel, was purchased from Amersham Biosciences (Uppsala, Sweden) and Dulbecco’s phosphate-buffered saline (DPBS, pH 7.4) from Gibco BRL (Gaithersburg, MD, USA). Dr. Hung Mien-Chie (M.D. Anderson Cancer Center, Houston, TX, USA) kindly supplied the two cell lines, MCF-7 and MCF-7/Her18. Trastuzumab was generously donated by Dr. Virginia Borges (UC Denver, CO, USA). LLO-pEt29-DP-E3570 transfected Escherichia coli were provided by Dr. Dan Portnoy, who introduced a His-tag onto the LLO vector and transfected the BL21(DE3) E. coli clone with the plasmid, (University of California Berkeley, CA, USA).¹⁴ Polylysine Hydrobromide (MW: 37,000, degree of polymerization: 177) was purchased from Sigma Chemicals Company (St Louis, MO, USA). All other chemicals and reagents, unless otherwise specified, were purchased from Thermo Fisher Scientific (Pittsburgh, PA, USA).

Liposome Preparation

Her-2 targeting neutral liposomes were prepared as previously described using the film hydration–extrusion method with slight modifications.¹¹ DPPC/RhoPE/DSPE-PEG(3400)-NHS/DSPE-PEG(2000)-PDP in chloroform were mixed at a molar ratio of 92:1.2:4.5:2.3. Before conjugation of trastuzumab to liposomes, the Fc region was removed using a Pierce (Fab’2) preparation kit (Thermo Scientific). For experiments utilizing rhodamine labeled polylysine, liposomes were formulated using Oregon Green PE instead of RhoPE. The lipids were dried with nitrogen for one hour and put under vacuum for 2 hours to remove any traces of chloroform. The dried lipid film was hydrated with 0.5 mL of water at 47 °C, and extruded once through a filter of 100 nm (Avestin, Ottawa, ON, Canada). Liposomes were quickly mixed with a 0.25 mL aliquot of trastuzumab Fab’2 in water at a concentration of 1.5 mg/mL which allows for conjugation between the trastuzumab amine and the DSPE-PEG(3400)-NHS. After at least 1 hour, liposomes were extruded 9 times through a 100 nm filter at 47 °C. After cooling to room temperature, the liposomes were column purified using a CL-4B Sepharose column hydrated in DPBS, pH 7.4. Column purified liposomes were diluted to a lipid concentration of 0.875 mg/mL or 0.935 mM. Cationic liposomes were prepared with the film hydration-extrusion method using DOTAP and cholesterol at a 1:1 molar, extruded at 100 nm and diluted in water to reach a final lipid concentration of 0.935 mM. To form a lipoplex, equal volumes of plasmid DNA encoding the luciferase gene (0.07 µg/µL) and cationic liposomes were mixed to form particles with a positive:negative charge ratio of either 4:1 or 8:1.

Conjugating LLO to the Liposome through Disulfide Bonding

LLO was synthesized in E. coli, and subsequently purified following the methods of Glomski et al. and diluted to a concentration of 1 mg/ml.¹⁴ To remove dithiothreitol (DTT) prior to disulfide bond formation, 100 µL of LLO was dialyzed for 12 h against 1L of storage buffer without DTT (50 mM phosphate buffer, pH 6.0, 1 M NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA)). One hundred microliters of column purified liposomes were mixed with a 5 µL aliquot of this dialyzed LLO and the mixture was incubated for 15 h at 25 °C at pH 7.4 to allow for disulfide conjugation.¹⁵

Formation of Polylysine/DNA Particles

NHS-PEG in chloroform was dried down under a stream of nitrogen for one hour. The NHSPEG was rehydrated with poly-L-lysine at a concentration of 1 mg/ml. Using FITC labeled NHS-PEG and purification with G25 sephadex mini columns, fluorescence analysis of collected fractions showed that the conjugation of NHS-PEG to polylysine occurs at approximately 44% efficiency. Based on this efficiency and the molar ratio of NHS-PEG added to polylysine we created samples that had PEG bound to approximately 0.33%, 0.66%, 1.3%, 2.6%, or 5.3% of the lysine residues. For all experiments other than that of Figure 3b, the polylysine with 1.3% PEG was used. The PEGylated polylysine was then added to the luciferase plasmid DNA at a concentration of 1 mg/ml to achieve positive:negative charge ratios of 1:1, 2:1, 4:1, 8:1, or 16:1. Charge ratio of PL/DNA was calculated assuming complete conjugation of DNA and polylysine and using a molecular weight of 146 g/mol for each positive lysine and 330 g/mol for each negative DNA base. For all other experiments other than that corresponding to Figure 3a, the charge ratio used was 4:1. After incubating for 30 minutes at room temperature to allow for particle formation, PL/DNA complexes were diluted in DPBS to a DNA concentration of 0.07 µg/µL. PL/DNA particles were labeled with rhodamine, by diluting NHS-Rhodamine to 10 mg/ml in DMSO and adding it to the polylysine prior to DNA binding at a concentration such that 5% of the lysine residues would be labeled. PL/DNA sizing was determined using a nanoparticle size analyzer, LM20 (Nanosight, Amesbury, Wiltshire, United Kingdom) (Table 1).

Figure 3.

Effect of charge ratio and PEG concentration on luciferase expression. The positive:negative charge ratio of PL:DNA particles was varied from 1:1 to 16:1. PL/DNA was administered with LLO/liposomes to MCF-7/Her18 and MCF-7 cells and luciferase expression was compared with delivery from standard DOTAP/cholesterol cationic liposomes (A). The percentage of lysines that are PEGylated within polylysine was varied from 0.33% to 5.3%, and PL/DNA was co-administered at a charge ratio of 4:1 with LLO/liposomes to MCF-7/Her18 cells and MCF-7 cells (B). DNA concentration was held constant at 1.75 ng/ µL and PL concentration was varied to achieve the appropriate charge ratio. Data are expressed as mean ± SE (n=5).

Table 1.

PL/DNA particle size at varying charge ratio

Charge Ratio of PL/DNA	Size (nm)
1:1	437 ± 354
2:1	231 ± 170
4:1	223 ±141
8:1	214 ± 106
16:1	228 ± 81
LLO/Liposomes	112 ± 27

The size of PL/DNA particles at different charge ratios and the size of the LLO/liposomes were measured after a 30 min incubation in 50% serum at 37°C.

Cell Culture Experiments

MCF-7, MCF-7/Her18, and SKBR-3 cells were grown in DMEM/F12 (10% FBS and 1% penicillin–streptomycin), HCC-1954 cells in RPMI (10% FBS and 1% penicillin–streptomycin) and MCF-12a cells in DMEM/F12 (supplemented with cholera toxin, EGF, insulin, 5% horse serum and 1% penicillin–streptomycin). Cells were split at least 24 hours prior to experimentation, and were plated in a 96 well plate so that they were between 50 –70% confluent when used. Liposomes and PL/DNA were first incubated in 50% serum for 15 minutes, diluted 5-fold with media so that cells would be exposed to 10% serum, and plated onto the cells with 100 µL per well. Final liposome and DNA concentrations are indicated in the figure legends. Cells were exposed to the liposomes and PL/DNA for 3 hours at 37 °C and 5% CO₂ before removing and rinsing cells twice with media. In studies involving fluorescent imaging the cells were analyzed at this 3 hour time point.

For luciferase studies, cells were then incubated at 37 °C and 5% CO₂ for 24 h before cells were harvested. Luciferase signal was measured as previously described using Cell Lysis Buffer (Promega, Madison, WI, USA), Luciferase Assay Reagent (Promega) and a Monolight 2010 luminometer (Analytical Luminescence Laboratory, Ann Arbor, MI, USA) to detect output luminescence.¹⁶ To determine the co-localization of liposomes and PL/DNA particles within cellular endosomes, cells were imaged using an Operetta high content imaging system with Harmony cellular analysis software (Perkin-Elmer, Santa Clara, CA, USA). The software requires a nuclear stain and membrane stain to locate cells for analysis. Therefore, cells were incubated in 100 µL of DPBS with Deep Red Cell Mask and Hoechst DAPI stain (3 µg/mL) for 5 min (Invitrogen). Cells were then rinsed and covered with 4% paraformaldahyde for fixing and imaging. Five wells with 5 images per well, randomly chosen by the software, were analyzed to determine the co-localization of Oregon Green-labeled liposomes and rhodamine-labeled PL/DNA. The Operetta software was programmed to quantitate both signals as well as the overlap of the two signals in each cell.

Results

Two component system

The two-component DNA delivery system characterized in this study consists of a LLO/liposome component and a Polylysine/DNA component that must co-localize in an endosome for gene expression. In the reducing conditions of an endosome, the disulfide-bound LLO should detach from the liposome and form pores in the endosomal membrane, thereby allowing plasmid DNA access to the cytoplasm.¹² DNA delivered to the cytoplasm can reach the nucleus upon cell division and breakdown of the nuclear membrane, resulting in expression of the plasmid DNA.² Although the PL/DNA internalizes into most cell lines, this treatment strategy should confer specificity since co-localization of PL/DNA and Her-2 targeting LLO/liposomes will be largely restricted to Her-2 positive cells.

To determine if endosomal co-localization of the PL/DNA and LLO/liposomes does occur more often in a Her-2 positive cell line, MCF-7 and MCF-7/Her18 cells were used. These cell lines are isogeneic except that MCF-7/Her18 cells have been stably transfected with Her-2 DNA and therefore overexpress Her-2 cell surface receptor by 45-fold.¹⁷ Fluorescence microscopy, using labeled PL/DNA and LLO/liposomes, provided evidence that co-localization of the two components was indeed more prevalent in the Her-2 positive cells (Fig 1). While PL/DNA internalized into both cell lines as can be seen by the red signal, the green signal shows LLO-liposomes were much more readily taken up by the Her-2 positive cell line. Analysis of 25 images showed that 68% of the PL/DNA co-localized with LLO/liposomes in the MCF-7/Her18 cells while there was only 0.7% co-localization in the control MCF-7 cells (Table 2). If colocalization does indeed lead to increased gene expression, this two-component system should result in specific expression in Her-2 positive cells.

Figure 1.

Co-localization of PL/DNA and LLO/liposomes. PEGylated liposomes are first conjugated to trastuzumab through an amide linkage and then disulfide bound to LLO (A). PL/DNA particles were labeled with red rhodamine (first column) and LLO/liposomes were labeled with Oregon Green (second column) (B). The top panel shows distribution of particles in Her-2 positive MCF-7/Her18 cells and the bottom panel shows Her-2 negative MCF-7 cells. The third column of pictures shows computer analysis of co-localization between PL/DNA and LLO/lipsomes with cell borders outlined in white. If PL/DNA co-localized with LLO/liposomes, the particles have been labeled yellow, whereas PL/DNA not co-localized with LLO/liposomes has been labeled purple. Lipid and DNA concentration administered to cells were 47 µM and 3.5 ng/µL respectively. The PL:DNA charge ratio was 4:1.

Table 2.

Internalization of PL/DNA, LLO/liposomes, co-localization, and luciferase expression in cell panel

	PL/DNA per cell (RFU)	LLO/Lipos ome per cell (RFU)	PL/DNA co-localized with LLO/liposomes per cell (RFU)	Percentage of PL/DNA co-localized with LL0/Liposomes (%)	Luciferase Expression (RLU/mg protein ÷1000
MCF-7/Her18	2308 ± 73	9357 ± 406	1583 ± 74	68 ± 1	7970 ± 2127
HCC-1954	678 ± 12	2415 ± 116	76 ± 12	11 ± 1	276 ± 181
SKBR-3	387 ± 12	2683 ± 111	155 ± 4	40.0 ± 0.4	218 ± 81
MCF-7	539 ± 39	172 ± 10	3.7 ± 0.5	0.7 ± 0.14	30 ± 6
MCF-12a	3523 ± 198	155 ± 9	31 ± 2	0.9 ± 0.1	5.2 ± 0.7

A panel of Her-2 positive and Her-2 negative cells were evaluated for internalization of PL/DNA, LLO/liposomes and luciferase expression as a result of the two-component delivery system. Percentage of co-localization shown in column 5 indicates the probability of a PL/DNA particle co-localizing with LLO/liposomes. Data are in relative fluorescence units (RFU) and relative luminescence units (RLU) per mg of protein and expressed as mean ± SE (n=3).

Utilization of a luciferase reporter plasmid allowed quantification of gene expression upon delivery of PL/DNA. When MCF-7/Her18 and MCF-7 cells were treated exclusively with PL/DNA, a moderate preference for uptake in the Her-2 positive cells was observed that resulted in 5.3-fold greater luciferase expression (Fig. 2). If PL/DNA was delivered with Her-2 targeting liposomes lacking LLO, Her-2 positive cells expressed about 18-fold more luciferase than the control MCF-7 cells (Fig. 2). Statistically, this expression was not significantly different from treatment with only PL/DNA for either cell line. Adding LLO to the liposomes increased total luciferase expression in the MCF-7/Her18 cells by 58-fold compared to delivery without LLO. The more effective delivery triggered by addition of LLO increased specificity for Her-2 positive cells, resulting in a 268-fold greater amount of luciferase expression in MCF-7/Her18 cells compared to MCF-7 cells (Fig. 2). In contrast, a standard cationic liposome delivery system consisting of DOTAP/Cholesterol (1:1 mole ratio) was non-specific and actually had slightly more luciferase expression in MCF-7 control cells compared to Her-2 positive cells (Fig. 2).

Figure 2.

Her-2 specific luciferase expression was increased by targeted LLO/Liposomes. Luciferase expression in MCF-7/Her18 cells was compared with expression in MCF-7 cells after delivery from PL/DNA, PL/DNA and liposomes without LLO, PL/DNA and LLO/liposomes, or cationic liposomes. Lipid and DNA concentration administered to cells were 47 µM and 3.5 ng/µL respectively. PL:DNA and cationic liposome:DNA charge ratios were 4:1. Data are in relative luminescence units (RLU) per mg of protein and expressed as mean ± SE (n=5).

Polylysine/DNA characterization

PL/DNA interacts strongly with cell membranes because it is positively charged. To explore the effect of PL/DNA charge ratio on delivery, the positive:negative charge ratio of PL:DNA was varied from 1:1 up to 16:1. As expected, increasing the charge ratio resulted in higher expression of the luciferase plasmid (Fig. 3A). With increasing charge ratio, specificity remained impressive with over 1000-fold greater luciferase expression in MCF-7/Her18 cells versus MCF-7 cells for the 4:1 and 16:1 ratios. The remaining experiments in this paper used PL/DNA at a 4:1 charge ratio because it utilized less polylysine, a cationic polymer which can be toxic at high concentrations in vivo.¹⁸

Preliminary experiments with polylysine and DNA showed that the particles quickly aggregated in buffer. PEGylating the PL/DNA created a polymer shield that stabilized the particles, preventing aggregate formation in buffer and 50% serum conditions (Table 1). However, it is known that excessive PEGylation can lead to decreased internalization of particles and therefore reduced gene expression ³. To optimize PEGylation, the percentage of polylysine residues bound to PEG was varied between 0.33% and 5.3%. As the percentage of PEG was increased, delivery to Her-2 positive cells decreased (Fig. 3B). Peak specificity for the Her-2 positive cells was observed at a PEG level of 2.6%. However, when PEGylation was decreased from 2.6% to 1.3%, luciferase expression increased by two-fold and specificity remained high. Given the detrimental effects of PEG on transfection, particles with 1.3% PEG and a charge ratio of 4:1 were used in subsequent experiments.

Delivery from decreasing particle concentration

Cells are easily exposed to high concentrations of LLO/liposomes and PL/DNA in vitro, but in vivo particles are cleared from the circulation quickly and endothelial barriers present a challenge. To test how this delivery system would perform if limited particles reached the tumor cells, the levels of LLO/liposomes and PL/DNA were reduced and delivery was compared to a standard cationic system. With both systems, there was a non-linear decrease in gene expression in response to reduced particle concentrations (Fig. 4A). An 8-fold reduction in particle concentration resulted in a 240-fold reduction in gene expression from the two-component system and a 115-fold reduction in expression from the cationic system.

Figure 4.

Dependence of luciferase expression on PL/DNA and LLO/liposome concentration. The expression in MCF-7/Her18 cells was measured after delivery from either DOTAP/cholesterol cationic liposomes or PL/DNA plus LLO/liposomes at decreasing concentrations (A). Either PL-DNA or LLO/liposome concentration was reduced from their starting concentrations (B). At each concentration in Fig. 4B, the bars on the left had LLO/liposome concentration held constant while PL/DNA was reduced and the bars on the right had PL/DNA concentration held constant while LLO/liposomes was reduced. At a 1× concentration, lipid and DNA concentration administered to cells were 47 µM and 3.5 ng/µL respectively. Data are expressed as mean ± SE (n=5).

To further explore the relationship between particle concentration and delivery, the concentration of either the PL/DNA or LLO/liposomes was held constant while the concentration of the other component was reduced (Fig. 4B). In both cases, there was a nonlinear decrease in luciferase expression as one component was reduced. These data indicate that both components seem to be associated with the dramatic decrease in expression that corresponds with reduced particle concentrations.

Delivery to a cell panel

The versatility of the two component delivery system was tested on two additional Her-2 positive cell lines, HCC-1954 and SKBR-3 as well as a non-cancerous Her-2 negative cell line, MCF-12a. Luciferase expression in the HCC-1954 cells was similar after delivery from the two-component and cationic liposome systems (Fig. 5). However, the cationic liposomes were more efficient in delivery to the SKBR-3 cells, as measured by luciferase expression. Importantly, the two component system resulted in significantly less luciferase expression in the Her-2 negative MCF-7 and MCF-12a cells when compared to treatment with the cationic liposomes. With the PL/DNA and LLO/liposome delivery system, all of the Her-2 positive cell lines had significantly more luciferase expression than the Her-2 negative cell lines (Fig. 5), showing that delivery is selective.

Figure 5.

Cell panel luciferase expression after delivery from PL/DNA and LLO/liposomes or from cationic liposomes. Luciferase expression was measured in Her-2 positive cell lines, MCF-7/Her18, HCC-1954 and SKBR-3 and was compared with expression in Her-2 negative cell lines, MCF-7 and MCF-12a after delivery either from PL/DNA and LLO/liposomes or from cationic liposomes. Lipid and DNA concentration administered to cells were 47 µM and 3.5 ng/µL respectively. Data are expressed as mean ± SE (n=5).

Using fluorescent microscopy to study the internalization and co-localization of PL/DNA and LLO/liposomes in the cell panel gives us further insight into the mechanism of the two component delivery system. Although the PL/DNA was internalized into all the cell lines, the amount of PL/DNA that co-localized with LLO/liposomes was significantly greater in all of the Her-2 positive cells when compared to the Her-2 negative cells (Fig. 6). The percentage of co-localization between the two components seems to be a good indication of the resulting luciferase expression (Table 2). Even though the MCF-12a cells internalized a large number of the PL/DNA particles, the lack of co-localization with LLO/liposomes resulted in very limited gene expression in this noncancerous Her-2 negative cell line.

Figure 6.

Co-localization of PL/DNA and LLO/liposomes in a cell panel. Her-2 positive cells, MCF-7/Her18, HCC-1934, SKBR-3 and Her-2 negative cells, MCF-7 and MCF-12a were studied to characterize the internalization of PL/DNA and LLO/liposomes. PL/DNA labeled with red rhodamine and LLO/liposomes labeled with Oregon Green were evaluated for internalization in the first and second columns respectively. The third column shows a computer evaluation of PL/DNA, with yellow indicating co-localization with LLO/liposomes, purple indicating no co-localization and a white outline showing the cell borders. Lipid and DNA concentration administered to cells were 47 µM nM and 3.5 ng/µL respectively. The amount of co-localization and luciferase expression were quantified and are shown in Table 2.

Discussion

Achieving specific gene expression from a cationic delivery system is challenging because of non-specific binding and internalization of cationic particles into most cell lines.¹⁹ Despite this drawback, anionic and neutral particles are rarely used for delivery since the negatively-charged DNA does not spontaneously complex with the particles.²⁰ To exploit the benefits of both neutral and cationic particles, we have created a two-component delivery system. Cationic particles consisting of DNA and polylysine deliver large quantities of DNA to cells while neutral Her-2 specific liposomes containing LLO facilitate endosome escape and impart specificity for Her-2 positive breast cancer cells.

Gene delivery from PL/DNA alone resulted in limited gene expression in both Her-2 positive and negative cells. The low expression was likely do in part to the addition of PEG on the particle, which both stabilizes PL/DNA and limits the interaction of the cationic polymer with the endosomal membrane.^{21, 22} To confer specificity to the gene delivery system, we utilized a Her-2 specific liposome that has trastuzumab and LLO conjugated to its surface. Her-2 positive cells treated with pore forming LLO bound liposomes in conjunction with PL/DNA showed a remarkable increase in gene expression. The corresponding increase in gene expression was much smaller in Her-2 negative MCF-7 cells, which internalized few of the Her-2 targeting liposomes. Given the necessity of endosomal escape for gene expression, it seems likely that the co-localization of Her-2 specific LLO/liposomes and PL/DNA allowed for cytoplasmic delivery of PL/DNA via LLO formed pores which in turn led to an increase in gene expression. This hypothesis is supported by mechanistic data which demonstrated a significantly greater colocalization of PL/DNA and Her-2 targeting LLO/liposomes in MCF-7/Her18 cells compared to Her-2 negative MCF-7 cells.

Decreasing the level of PEG or increasing the positive charge of PL/DNA led to a non-specific increase in gene expression from both the Her-2 positive and negative cells. This increase in gene expression could be the result of high cellular PL/DNA concentrations since particle internalization is augmented as PEG concentration decreases and as charge ratio increases.³ All variations of PEG and charge ratio, excluding a 1:1 charge ratio of polylysine to DNA, resulted in significant specificity for Her-2 positive cells. This is an improvement on a standard cationic liposome system which demonstrated no selectivity for MCF-7/Her18 cells. Ultimately, this system will require in vivo optimization to determine the PEG concentration and charge ratio that result in maximal tumor accumulation.

Both cationic liposomes and the two-component system lost efficiency as concentration of the administered particles was decreased. Furthermore, when either one of the two components was decreased in concentration while the other was held constant, this non-linear reduction in gene delivery was also apparent. For cationic liposomes, it has been hypothesized that a critical amount of cationic lipid is needed to neutralize the negative lipids of the endosome, which allows for membrane fusion and delivery of liposomal contents.²³ As the number of cationic liposomes per endosome was reduced perhaps this critical concentration was not reached and less DNA was delivered to the cytoplasm. Likewise it is thought that pore formation in the endosomal membrane requires 33–50 monomers of the LLO protein, and as less LLO/liposomes were internalized, it is likely that some endosomes did not contain the requisite LLO to form pores.²⁴ It is not immediately clear why reducing PL/DNA resulted in a drastic decrease in gene expression; LLO/liposomes levels were held constant, so the amount of DNA that reached the cytoplasm should have been roughly proportional to the concentration of PL/DNA. Perhaps the cell has a mechanism for clearing plasmid DNA from the cytoplasm which is efficient at low concentrations and overwhelmed as plasmid concentrations increase. A study into the mechanism of the reduction in gene expression would be relevant given the certainty that achieving high particles concentrations at in vivo tumor sites will be more difficult than in vitro.

Delivery to a cell-panel of Her-2 positive cells and Her-2 negative cells showed that the level of gene expression differs widely between cell lines. In all cases, the Her-2 positive cells had a higher level of gene expression than the Her-2 negative cells, demonstrating that the two-component delivery system is selective. On the contrary, cationic liposomes were not selective, resulting in higher gene expression from the Her-2 negative MCF-7 cells than the Her-2 positive HCC-1954 and SKBR-3 cell lines. The two-component system also had a lower level of gene expression in both of the Her-2 negative cell lines when compared to the cationic system. Although specificity was achieved, analysis of gene expression in the Her-2 positive cell lines showed much lower expression in the HCC-1954 and SKBR-3 cell lines than in the MCF-7/Her18 cells. Using fluorescently labeled LLO/liposomes and PL/DNA it was possible to measure the amount of each component internalized and the level of co-localization of the components within the endosomes. From these data, it is apparent that the level of gene expression can be related to the amount of co-localization of the two components. Both the HCC-1954 and SKBR-3 cells had less co-localization than the MCF-7/Her18 cells and the two Her-2 negative cell lines had even less co-localization. While it can be difficult to compare transfection in different cell lines, the fact that the MCF-7/Her 18 and MCF-7 cell lines are isogeneic except for the expression of Her-2 clearly demonstrates the specificity of delivery (268-fold greater expression in the former cell line) with this two-component system. This technique, using fluorescently labeled particles and microscopy, could be a valuable tool for determining if treatment would result in a sufficient level of gene expression for a given tumor cell line.

A review of the literature reveals a handful of systems that have achieved a certain degree of specificity in vivo, using cationic liposome systems.^25–30 These systems primarily use PEGylation to minimize non-specific gene delivery coupled with ligand targeting to increase delivery to the tumor cells.^25–28 Other techniques include the encapsulation of cationic polymer/DNA complexes inside of an anionic membrane or the use of tumor-specific promoters to increase level of gene expression at the tumor site.^{29, 30} When gene expression is restricted to the tumor tissue, therapeutic delivery of genes that activate the immune system or promote apoptosis can produce significant and impressive reduction in tumor growth.^26–28 Given the potential of a specific gene therapy and the few options for achieving specificity, it is vital that more effective delivery systems be designed. The experiments described herein show the initial in vitro characterization of a two-component system, which utilizes a new technique to achieve specificity for Her-2 positive breast cancer cells. Future studies will assess whether splitting the mechanism of gene delivery into these two separate components will lead to increased specificity in vivo.