EpCAM-targeted liposomal si-RNA delivery for treatment of epithelial cancer

Abstract

Background: RNA interference (RNAi) technology using short interfering RNA (si-RNA) has shown immense potential in the treatment of cancers through silencing of specific genes. Cationic non-viral vectors employed for gene delivery exhibit toxic effects in normal cells limiting their widespread use, therefore, site-specific delivery using benign carriers could address this issue.

Objective: Design of a non-toxic carrier that enables site-specific delivery of si-RNA into the cancer cells is of prime importance to realize the promise of gene silencing.

Methods: In the present study, non-cationic liposomes encapsulating si-RNA against epithelial cell adhesion molecule (EpCAM) were developed and characterized for encapsulation efficiency, colloidal stability, in vitro and in vivo gene silencing efficacy.

Results: PEGylated liposomes containing phosphatidyl choline and phosphatidyl ethanolamine exhibited maximum si-RNA encapsulation efficiency of 47%, zeta potential of -21 mV, phase transition temperature of 51 °C and good colloidal stability in phosphate-buffered saline (PBS) containing bovine serum albumin (BSA) and plasma protein (PP) at 37 °C. Conjugation of epithelial cell adhesion molecule (EpCAM) antibody to the liposomes resulted in enhanced cell internalization and superior down-regulation of EpCAM gene in MCF-7 cell lines when compared with free si-RNA and the non-targeted liposomes. In vivo evaluation of immunoliposomes for their efficacy in regressing the tumor volume in Balb/c SCID mice showed about 35% reduction of tumor volume in comparison with the positive control when administered with an extremely low dose of 0.15 mg/kg twice a week for 4 weeks.

Conclusion: Our results exhibit that the nanocarrier-mediated silencing of EpCAM gene is a promising strategy to treat epithelial cancers.

• Colloidal stability
• gene delivery
• gene silencing
• non-cationic liposomes
• release kinetics

Introduction

Short interfering RNA (si-RNA) is a powerful tool that has the potential to down-regulate the expression of the target genes (Pai, 2005). However, gene silencing through RNA interference (RNAi) for the treatment of different cancers has not made significant progress due to several limitations. Most strategies to improve RNAi efficiency have not been successful due to the immune response and the additional barriers presented by the biological system (Whitehead et al., 2009). The major challenges in the use of RNAi are the poor stability of si-RNA in circulation, poor targeting to specific cells, inability to escape endosomal degradation, off-target effect and production of inflammatory response (Juliano et al., 2009). Therefore, the prime requisite for successful implementation of effective gene down-regulation in vivo is to enable target-specific delivery of si-RNA using an appropriate carrier.

Numerous approaches for delivery of si-RNA have been investigated including both viral and non-viral delivery systems (Gao & Huang, 2008). Viral vectors, such as adenoviral, lentiviral and retroviral systems exhibit high transfection efficacy (Raper et al., 2003; Stewart et al., 2003). But, despite their desirable transfection efficiency, use of viral vectors for gene delivery is restricted because of their ability to integrate into the host genome and produce severe inflammation (Barton & Medzhitov, 2002). Clinical trials conducted for the treatment of X-linked severe combined immune deficiency using retrovirus vectors revealed that the patients exhibited increased T-cell count because of the integration of viral genome at the first intron of LMO-2 (Hacein-Bey-Abina et al., 2003). These results demonstrate the risk associated with viral vectors. This has resulted in the development of non-viral vectors to improve gene therapy strategies. Majority of the carriers employ cationic moieties to facilitate better complexation of the negatively charged si-RNA through electrostatic associations with the carrier. Among the polymers commonly employed as non–viral si-RNA delivery systems, chitosan (Ma et al., 2014; Ozpolat et al., 2014; Xie et al., 2014), poly(ethylene imine) (Urban-Klein et al., 2005; Nabzdyk et al., 2014), poly(amidoamine) dendrimers (Patil et al., 2008; Conti et al., 2014) and poly(l–lysine) (Inoue et al., 2008; Kodama et al., 2014) have been extensively investigated.

Liposomal carriers, in particular, have received considerable attention over the years due to their ease of formation, tailorable surface functionalities and excellent cell internalization (Immordino et al., 2006; Kong et al., 2012; Nag & Awasthi, 2013). Cationic liposomes prepared with 1,2–dioleoyl–3–trimethylammonium propane (DOTAP) and dioleoyl phosphatidylethanolamine (DOPE) lipids have been found to be highly efficient in complexing the si-RNA and have been evaluated for their silencing efficacy in Many cancer cell lines (Ying & Campbell, 2014). Custom-made cationic lipids (Oh & Park, 2009) and cationic gemini lipids (Zhao et al., 2014) have also been explored as gene delivery vehicles. Despite their promise in in vitro trials, most of these systems are ineffective in vivo due to the toxicity of the cationic moieties and their non-specific interactions with non-target cells limit their widespread use. High-inflammatory responses to these systems have also restricted their use as gene delivery systems in vivo (Lonez et al., 2008). Therefore, attempts to reduce or mask the cationic moieties have been reported to reduce the adverse effects of these systems. Use of anionic lipids to form liposomes along with a divalent cation has also been attempted to overcome the drawbacks of cationic liposomes (Daniel & Godbey, 2011).

The present study intends to develop si-RNA-encapsulated PEGylated liposomes formed using non-cationic lipids. We attempt to achieve cancer cell-specific delivery through surface modification with EpCAM antibody. Epithelial cell adhesion molecule has been found to be over-expressed in epithelial cancers, cancer stem cells (Allard et al., 2005) as well as circulating cancer cells (Simon et al., 2013). The EpCAM found in normal cells are expressed in the basolateral side, while it is over-expressed on the apical side in cancer cells (Slanchev et al., 2009). This characteristic of EpCAM makes it an attractive target for site-specific si-RNA delivery. EpCAM-targeting liposomes encapsulating doxorubicin were found to be more effective in colon cancers when compared to free drug and the non-targeted carrier (Allen & Cullis, 2004). Synthetic amphiphiles (SAINT) were reported to effectively deliver antisense oligonucleotides to melanoma and colon cancer cells when conjugated to a monoclonal antibody targeting EpCAM (Van Zanten et al., 2004). However, significant reduction in cancer cell numbers was achieved only in colon cancer cells implying the need for further exploration of this targeting moiety and its role in different types of cancer. Delivery of bispecific antisense nucleotides against bcl-2 and bcl-xL has been accomplished using EpCAM antibody fragment as the targeting moiety (Hussain et al., 2006). The system was found to sensitize the epithelial cancer cells to doxorubicin.

EpCAM has also been recognized to possess oncogenic potential (Osta et al., 2004). The MCF-7 breast cancer cell line has been extensively studied for EpCAM over-expression and it has been found that EpCAM over-expression activates cell proliferation by stimulating gene expression of c–myc and JNK/AP–1 (Sankpal et al., 2011). The silencing of EpCAM gene has been demonstrated to down-regulate proliferation genes and up-regulate the expression of apoptotic genes, such as DRAM and cytochrome c. It also reduces the expression of tumor invasive genes like MMP2 and cdc42 (Mitra et al., 2010). Therefore, EpCAM gene silencing may help in controlling tumor growth and can be used for the treatment of epithelial cancers. However, reports on the use of an appropriate gene delivery system to achieve silencing of EpCAM gene through RNAi technology are scanty in literature. The realization of the therapeutic potential of EpCAM silencing through the use of targeted cationic lipid-free liposomal system therefore forms the crux of the present study.

Materials and methods

Materials

Egg phosphatidylcholine (EPC) was purchased from Avanti Polar Lipids, Alabaster, AL. N–(Carbonyl–methoxypolyethyleneglycol 2000)–1,2–distearoyl–sn–glycero–3–phosphoethanolamine sodium salt (PEG chain MW 2000 Da) (DSPE–PEG) were purchased from NOF, Grobbendonk, Belgium. EpCAM si–RNA and Alexa Fluor 488 labeled si–RNA were purchased from Qiagen, Valencia, CA. EpCAM antibody was procured from Santa Cruz Biotechnology Ltd., Dallas, TX. Bovine serum albumin (BSA) was purchased from Sigma Aldrich, St. Louis, MO. All other organic reagents of analytical grade were purchased from Merck Chemicals, NJ, USA.

Preparation and characterization of liposomes

Liposomes were formed using thin-film hydration. The solvent in the phospholipid solution was evaporated in a current of nitrogen to obtain a thin layer of phospholipids. An aqueous dispersion of the si-RNA complex in phosphate-buffered saline (PBS) of pH 7.4 was added to the thin film and stirred constantly for 30 min at 60 °C. The solution was extruded through polycarbonate membranes with a pore diameter of 200 nm for 10 cycles to obtain uniform sized liposomes and were freeze–dried (Christ alpha 2-4 LD Freeze dryer, Martin Christ, Osterode am Harz, Germany) and stored until further use.

For preparation of hybrid immunoliposomes, thiolated EpCAM antibody was used. The antibody was mixed with Traut's reagent (2–iminothiolane, sigma Aldrich, St. Louis, MO) at 1:20 ratio in HEPES-buffered saline (HBS) (25 mM HEPES, 140 mM NaCl, pH 8) and allowed to stand at room temperature for 1 h. The reaction buffer contains 2 mM EDTA that inhibits self-polymerization during the thiolation (Pan et al., 2007). The unreacted Traut's reagent was removed by dialysis in HBS of pH 7.5. The thiolated antibody was then mixed with liposomes incorporated with maleimide-terminated PEG lipid (DSPE–PEG–MAL) and the mixture was kept overnight at 4 °C for the formation of C–S covalent link between the antibody and PEGylated hybrid liposome. The unattached antibodies were removed using Centriprep® (Darmstadt, Germany) at 4 °C and linking was confirmed through Fourier transform infrared spectroscopy (FTIR) (Spectrum 100, Perkin Elmer, Waltham, MA). The reaction scheme is shown in Figure 1.

Figure 1. Schematic representation of establishment of immunoliposomes.

The encapsulation efficiency of the si-RNA in liposomes was quantified using Alexa Fluor–488 (Qiagen, Valencia, CA) linked fluorescent si-RNA. To quantify the encapsulation efficiency, the samples containing the liposomes and the unencapsulated si-RNA was subjected to centrifugation at 15 000 rpm for 30 min at −10 °C. The supernatant was analyzed using multimode reader at an excitation wavelength of 490 nm and emission wavelength of 525 nm for determining the amount of si-RNA present in the liposomes. The encapsulation efficiency was then calculated using the formula:

The particle size and zeta potential of the samples were measured using dynamic light scattering. The samples were dispersed in equal volume of deionized water and loaded in disposable capillary cells. The analysis was carried using Zetasizer (Nano–ZS, Malvern, UK) at room temperature.

Thermal behavior of the samples was analyzed in differential scanning calorimeter (DSC) (Q20, TA instruments, New Castle, DE). Five milligrams of liposome samples were weighed and taken in aluminum pan. Heat flow was set to 10 °C per min and samples were analyzed from 10 °C to 90 °C.

The colloidal stability of the samples was recorded using the laser diffraction method. The extruded samples were dispersed in 20 mL of PBS (pH 7.4) or PBS containing bovine serum albumin (BSA) or plasma protein (PP) and incubated for 24 h at 37 °C and then analyzed using particle size analyzer (Microtrac Bluewave, Montgomeryville, PA).

For evaluation of the magnitude of protein adsorption, liposomes were prepared using the thin-film hydration technique. The pellet having the liposomes was dispersed in PBS of pH 7.4. The reconstituted liposome samples were incubated for 2 h with PBS containing BSA or PP (2 mg/mL). After 2 h incubation, the liposomes were separated by centrifugation at 15 000 rpm for 30 min at −10 °C and the supernatant was analyzed for determining the non-adsorbed proteins using Bradford protein estimation assay.

The antibody conjugation to the liposomes was documented using Fourier transform infrared spectroscopy (FTIR). To prepare samples for the FTIR analysis, liposomes were mixed with KBr (IR grade, Merck, NJ, USA) and pelleted using a pelletizer. The spectra were recorded in a FTIR spectrometer (Spectrum 100, Perkin Elmer, Waltham, MA) between 4000 and 400 cm⁻¹.

In vitro studies

For in vitro experiments 2 × 10⁵ MCF-7 breast cancer cells were seeded per well in 24-well plates. Dulbecco's Modified Eagle's Medium (DMEM) and fetal bovine serum (FBS) (Gibco, NY, USA) were used for maintaining the culture. For cell uptake studies, MCF-7 cells were seeded 24 h prior to the experiment. After 24 h, the spent medium was replaced with fresh DMEM medium. The liposomes containing Alexa Fluor–488 linked si-RNA dispersed in 0.5 mL PBS of pH 7.4 were added to the respective wells. At pre-determined time points (0, 1, 2 and 4 h), the cells were counterstained with Hoechst and imaged using laser scanning confocal microscope (FV1000, Olympus, Tokyo, Japan). Cell uptake was also quantified by flow cytometry. Briefly, 2 × 10⁵ MCF-7 cells were treated with Alexa Fluor–488 linked si–RNA encapsulated liposomes. After 24 h, cells were collected, centrifuged and washed twice in PBS. The pellet was dispersed in sheath fluid and cells were analyzed using flow cytometer (FACS Calibur, BD Biosciences, New Jersey, USA).

The quantitative analysis of silencing of EpCAM gene was determined using qRT–PCR (AG22331, Eppendorf, Hamburg, Germany). About 2 × 10⁵ MCF-7 cells were incubated with immunoliposomes and liposomes without antibody containing 100 and 200 nM si-RNA. After 48 h of incubation, the cells were collected and the total RNA was isolated using RNeasy mini kit (Qiagen, Valencia, CA) following manufacturer's protocol and the RNA was quantified using UV-visible spectrophotometer (Nano Drop, Thermo Fisher Scientific, Waltham, MA). Complementary DNA (cDNA) was synthesized using Quantitect Reverse transcription kit (Qiagen, USA) and random primers. Then, the cDNA samples were subjected to amplification using specific EpCAM primers with the following sequence – Forward: CGCAGCTCAGGAAGAATGTG and Reverse: TGAAGTACACTGGCATTGACG. The gene expression was quantified employing the ΔC_t–ΔC_t method in real time RT-PCR using SYBR Green probe (Qiagen, USA).

Toxicity of immunoliposomes was evaluated using MTS assay (Cell Titer 96 AQueous one solution, Promega, Madison, MI). Ten thousand cells were seeded in a 96-well plate and incubated at 37 °C in 5% CO₂. After the cells attained confluency, the samples were added. At the end of each time point, the samples were washed with PBS solution to remove the non-adherent cells. MTS reagent (20 μl) and 200 μl of serum-free media were added to each of the samples and incubated at 37 °C for 2 h. The reaction was stopped by an addition of 25 μl of 10% sodium dodecyl sulfate (SDS) solution. The absorbance was measured at 490 nm using a multimode reader (Infinite 200M, Tecan, Grodig, Austria).

In vivo studies

Female BALB/c SCID mice aged 5–7 weeks were used for the study. Animal ethics clearance was obtained from the Institutional Animal Ethics Committee (IAEC approval no. 16/13–(6/4/2013)) to carry out the study. Mice were given irradiated rodent diet ad libitum (Pet car brand, Bangalore, India) and were maintained in a sterile setting at 22–25 °C with 12 h light/dark cycle. 48 h prior to the tumor induction, the mice were implanted subcutaneously with 0.18 mg slow release 17β–estradiol pellets (SE–121, Innovative Research of America, Novi, MI). About 5 × 10⁶ MCF-7 breast cancer cells in 100 μL matrigel (FACS Calibur, BD Biosciences, New Jersey, USA) were inoculated subcutaneously in the right flank of the mice. The animals were observed frequently for the tumor growth. Palpable tumors were visible from the second week after cancer induction. The mice were randomized after they attained a tumor volume of about 70–90 mm³, and grouped with five mice in each group. Liposomal formulations were given thorough injection to a site adjacent to the tumor at 0.15 mg/kg body weight of mice. The formulations were administered twice a week for four weeks. The tumor dimensions were measured twice a week using digital vernier calipers. The tumor volume was calculated using the relation: where, L is the length and W is the width of the tumor. After completion of four weeks, animals were sacrificed by CO₂ asphyxiation and the tumors were excised, weighed. For further trials, half of the tumor fragment was snap frozen in liquid nitrogen and stored at −80 °C. The other half of the tumor was fixed in 10% neutral-buffered formalin for histopathological analysis. The sections stored in −80 °C were weighed, crushed in 1 mL of lysis buffer. The lysate were used for the mRNA isolation following manufacturer's protocol. cDNA preparation and qRT-PCR for gene expression was carried out as mentioned for in vitro studies. Formalin-fixed tumor samples were used for histopathological analysis. Thinner sections 2–3 mm thick were cut and processed. Paraffin-embedded tissues were sectioned to 3–5 micron thick slices using microtome and stained with haematoxylin and eosin. The samples were then imaged using light microscope (Nikon Eclips Ti, Tokyo, Japan) and imaged.

Western blot analysis

Total protein was isolated from the tumor tissue using cell lysis buffer (10 mM Tris-HCl of pH 7.2), 2% sodium dodecyl sulphate (SDS), 10 mM dithiothreitol, 1% protease and protease inhibitors cocktail (Sigma Aldrich, USA). The concentration of protein was measured using Lowry's method. An aliquot (containing 50 µg protein) of lysate was used for 12% sodium dodecyl sulfate-polyacrylamide gel. The blocking of the membrane blots were carried for 1 h with blocking buffer (5% skimmed milk in Tris-buffered saline containing 0.1% Tween-20) and then incubated overnight in primary antibody (EpCAM antibody, dilution 1:150, Cell Signalling, Beverly, MA) at 4 °C. The blots were washed and incubated with appropriate anti-mouse horseradish peroxidase-conjugated secondary antibodies (dilution 1:2000, Cell Signaling, USA) for 3 h at room temperature. Tetramethyl benzidine/hydrogen peroxide (TMB/H₂O₂, Bangalore Genei, Bangalore, India) reagent was used for the visualization according to the manufacturer's instructions. Membranes were stripped in 100 mM glycine of pH 2 for 40 min, reblocked, and re-incubated in primary antibody for the housekeeping protein β-actin (Sigma Aldrich, USA). The UVP BioDoc-IT 220 Imaging System (Upland, CA) and ImageJ software (http://imagej.nih.gov/ij/) were used for the scanning and densitometry analysis of digitized images. After the background normalization, β-actin band intensity was used to normalize the band intensity of the corresponding EpCAM protein band.

Statistical analysis

Two-way analysis of variance (ANOVA) was employed to determine the statistical significance. The level of significance was determined using Bonferroni's test (p < 0.05).

Results

Liposomes were prepared using thin-film hydration method. The lipid composition was modified to achieve maximum encapsulation. The phospholipids used were egg phosphatidyl choline (egg PC), dioleoyl phosphatidyl ethanolamine (DOPE), cholesterol, distearoyl phosphatidyl ethanolamine-polyethylene glycol (DSPE-PEG) and distearoyl phosphatidyl ethanolamine-polyethylene glycol-maleimide (DSPE-PEG-mal).

Figure 2 shows the encapsulation efficiency liposomes prepared with different lipid ratios. It is evident from Figure 1 that the lipid composition influences the si-RNA encapsulation significantly. The least encapsulation efficiency of 19% was exhibited by liposomes prepared using egg PC only (LA). Introduction of the helper lipid DOPE improved the encapsulation efficiency. The liposomes with 8:2 ratio of egg PC and DOPE exhibited 23% efficiency, which increased to 35% when the ratio was changed to 7:3 egg PC:DOPE. Further changes in the DOPE content did not cause any significant increase in the encapsulation of si-RNA. PEGylation of liposomes was found to positively influence the encapsulation of si-RNA with the liposomes of composition 7:2:0:1 ratio of egg PC:DOPE:cholesterol:DSPE-PEG exhibiting the highest encapsulation of 47%. Introduction of cholesterol reduced the encapsulation efficiency of si-RNA. However, the 7:1:1:1 ratio of egg PC:DOPE:cholesterol:DSPE-PEG exhibited an encapsulation efficiency of 39%. Therefore, for further trials, liposomes with the ratio of egg PC:DOPE:cholesterol:DSPE-PEG 10:0:0:0, 7:3:0:0, 7:2:0:1 and 7:1:1:1 were chosen. These were designated as LA, LB, LC and LD, respectively.

Figure 2. si-RNA encapsulation efficiency of liposome formed by different lipid ratios. (* < 0.05).

Table 1 shows the particle size, zeta potential and phase transition temperature of the liposomes with four different lipid compositions (LA, LB, LC and LD).

Table 1. Particle size, zeta potential and phase transition temperature of liposomes.

	Blank liposomes			si-RNA liposomes
Liposomes type	Particle size (nm)	Zeta potential (mV)	Tm (°C)	Particle size (nm)	Zeta potential (mV)	Tm (°C)
LA	65 ± 0.126	−29.37 ± 1.7	47.09	95 ± 0.05	−24.37 ± 0.6	48.63
LB	73 ± 0.78	−17.53 ± 2.2	33.12	120 ± 0.264	−14.7 ± 1.2	39.69
LC	96 ± 0.138	−25.9 ± 1.2	47.25	138 ± 0.286	−21.7 ± 1.1	51.6
LD	103 ± 0.006	−32.1 ± 1.1	49.77	142 ± 0.121	−26.03 ± 1.1	51.66

It is observed from Table 1 that the encapsulation of si-RNA causes an increase in particle size of LA, LB, LC and LD liposomes when compared with their blank counterparts. The zeta potential exhibits a reverse trend where the negative zetapotential was shifted towards less negative values after the si-RNA encapsulation in all liposome types suggesting masking and redistribution of the surface charges owing to the incorporation of si-RNA. The presence of si-RNA in the liposomes is further confirmed from the shift in the phase transition temperatures of the blank liposomes with different lipid compositions after encapsulation of si-RNA. A general trend that is discernible from the results is that the si-RNA-loaded liposomes exhibit a higher phase transition temperature when compared with their corresponding blank counterparts.

Figure 3 shows the shift in zeta potentials of blank and si-RNA-loaded liposomes with time. Initially, the blank liposomes exhibit a higher negative zeta potential in all lipid compositions. These values are found to decrease progressively with time indicating a gradual reduction in the colloidal stability of the liposomes. A similar trend was observed in the case of si-RNA-loaded liposomes but the changes were much more subtle suggesting that encapsulation of si-RNA improves the colloidal stability of liposomes.

Figure 3. Zeta potentials of blank and si-RNA-loaded liposomes after 0, 30, 60 and 120 min of incubation at 37 °C (a) LA, (b) LB, (c) LC and (d) LD (* < 0.05).

Figure 4(a) and (b) show the percentage adsorption of BSA and PP on different liposomes after 24 h of incubation in PBS containing BSA and PP. It was observed that the adsorption of BSA and PP was less over PEGylated liposomes when compared to non-PEGylated liposomes. This indicates that the PEG chains confer protein-repelling property to the liposomal surface. Adsorption of plasma proteins was greater on the liposomal surface when compared with bovine serum albumin.

Figure 4. Protein adsorption and zeta potentials of blank and si-RNA-loaded liposomes (LA, LB, LC and LD) (a) Bovine serum albumin and (b) plasma protein adsorption after 24 h of incubation. (c and d) Zeta potential of the liposomes after 24 h incubation in PBS containing bovine serum albumin and plasma proteins (* < 0.05).

In all cases, adsorption of BSA on the surface of si-RNA-loaded liposomes was lesser than that recorded for their respective blank counterparts. But in the case of PP, no significant difference was observed between blank and si-RNA liposomes. Instead, variations in the magnitude of protein adsorption were found to be solely dependent on the lipid composition of the liposomes. Figure 4(c) and (d) show the zeta potential of the liposomes after 24-h incubation in PBS containing BSA and PP. Zeta potential values of the blank were decreased considerably than the si-RNA-encapsulated liposomes. PEGylated liposomes encapsulating si-RNA were found to exhibit highest zeta potentials reaffirming their superior stability when compared with other lipid compositions.

The introduction of a cargo in the liposomal carrier can either enhance or decrease its stability. This change can be monitored by measurement of the particle size with time. Figure 5 depicts the size distribution of different liposomes after their incubation in PBS, PBS with BSA and PBS with PP. It was observed that the size of the LA and LB liposomes progressively increased with time with appearance of a greater percentage of micron-sized population. However, PEGylated LC and LD liposomes maintained a narrow size distribution in the nanodimension even after 24 h. The change in size was more pronounced in LA, LB and LD liposomes incubated in the presence of plasma proteins. The si-RNA-encapsulated liposomes exhibited a slower change in their dimensions when compared with their blank counterparts in all the liposome combinations.

Figure 5. Colloidal stability of LA, LB, LC and LD liposomes in PBS, PBS containing BSA and PP after 24 h at 37 °C. BL: Blank and SL: siRNA-loaded liposomes in PES, BLA and SLA: in presence of BSA. BLPP and SLPP: in presence of Plasma protein.

Figure 6 shows the release of si-RNA from liposomes with different lipid compositions. It is seen that the LC liposomes (7:2:0:1 ratio of egg PC:DOPE:cholesterol:DSPE-PEG) exhibit the lowest burst release of about 20% and a sustained release of si-RNA owing, whereas cholesterol incorporated liposomes and LB liposomes (7:3 ratio of egg PC and DOPE) show about 50-70% burst release. The cholesterol containing liposomes display the fastest release when compared to the other liposomes. The release profile of LA liposomes is not shown in Figure 6 because of its low-encapsulation efficiency and poor stability.

Figure 6. si-RNA release profile of LB, LC and LD lipomses (* < 0.05).

As the PEGylated liposomes exhibit better colloidal stability, the LC liposomes with lipid composition of egg PC:DOPE:cholesterol:DSPE-PEG in the ratio 7:2:0:1 was used for in vitro studies. In addition, introduction of the targeting ligand is possible in PEGylated liposomes and hence LC liposomes with the maximum encapsulation of si-RNA and superior colloidal stability were employed for further trials. Conjugation of the targeting ligand EpCAM antibody on the surface of LC liposomes was carried out to obtain EpCAM antibody-tagged liposomes. The presence of EpCAM antibody in the liposomes was confirmed from the FTIR spectrum for the tagged liposomes. The significant vibration bands that appear in the FTIR spectra of LC liposomes without and with EpCAM antibody modification are summarized in Table 2.

Table 2. FTIR vibration bands of liposomes and immunoliposomes.

	Vibration frequency (cm^−1)
Bond	Liposomes	Immunoliposomes
Asymmetric and symmetric stretching  of –CH2–	2924, 2853	2926, 2852
–CH2– scissoring vibration	1466	1466
Rocking of the multiple –CH2–	720	720
Asymmetric and symmetric stretching  of the –CH3	2956, 2871	–
–OH stretch	Broad band at 3400	Broad band at 3400
Ester carbonyl stretch	1734	1736
P-O stretch	983	934, 1003
Amide carbonyl stretch	–	1634
C-N stretch	–	1188
Aromatic ring bend	–	750-900
C-S bond	–	593

It is evident from Table 2 that conjugation of EpCAM antibody to the liposome has resulted in the appearance of new bands at 1634 cm⁻¹ due to the amide carbonyl group and at 1188 cm⁻¹ due to C-N stretch of the peptide bond. Similarly, the band at 593 cm⁻¹ may be attributed to the conjugation of the thiolated antibody to the PEG chain on the liposome surface. The characteristic bending vibrations due to aromatic rings from aromatic amino acids present in the antibody also appear between 750 and 900 cm⁻¹ in the FTIR spectrum of LC liposomes conjugated with EpCAM antibody.

In vitro studies

Cell uptake of the LC liposomes without and with covalent linking of EpCAM antibody in MCF-7 breast cancer cell line was carried out using laser scanning confocal microscopy. Figure 7 shows the cell uptake of the LC liposomes with and without EpCAM antibody conjugation over a time period of 4 h. In order to ascertain if the EpCAM antibody modification directs the uptake of the liposomes in the cells through receptor-mediated endocytosis, the cells were pre-incubated with EpCAM antibody. The medium was then replaced with fresh medium to which the EpCAM antibody conjugated LC liposomes were added.

Figure 7. Confocal images show the uptake of LC liposomes encapsulating fluorescent si-RNA at various time points by MCF-7 cells. Panel a illustrates the uptake of LC liposomes; Panel b illustrates the cell uptake of LC immunoliposomes intervals and Panel c illustrates the uptake of LC immunoliposomes by MCF-7 cells that have been pre-incubated with anti-EpCAM [Nucleolus: Blue (Hoechst), si-RNA: Green (Alexa flour488)].

It is seen that the cell uptake of the unmodified liposomes gradually increases with time with more amount of liposomes (green spots) internalized at 4 h. The immunoliposomes (LC liposomes conjugated with EpCAM antibody) were found to exhibit rapid cell uptake and are discernible as intense green spots in the confocal images of cells even after 30 min of incubation. The fluorescence intensity was found to increase with time indicating higher percentage of internalization of the immunoliposomes. When the cells were pre-incubated with EpCAM antibody prior to incubation with the immunoliposomes, it was observed that the cell internalization of the immunoliposomes becomes negligible even after 4 h. This may be attributed to the saturation of the EpCAM receptors by its antibody during the pre-incubation. This result confirms that in the case of immunoliposomes, the cell internalization is mediated through binding to EpCAM receptors on the cell surface. Quantification of cell uptake using flow cytometry shows a mean fluorescence intensity 56.47 ± 2.5 and 79.17 ± 0.9 in cells treated with LC liposomes and LC immunoliposomes encapsulated with fluorescent si-RNA, respectively.

EpCAM gene silencing efficacy of the LC liposomes in MCF-7 cells after 48 h is shown in Figure 8. It was observed that the immunoliposomes containing 100 nM as well as those with 200 nM concentrations of si-RNA exhibited better silencing efficacy when compared with the LC liposomes containing the same concentrations of si-RNA. The immunoliposomes containing 100 nM and 200 nM si-RNA exhibit 1.7- and 3.9-fold down-regulation of EpCAM, respectively. In contrast, the LC liposomes show only 1.4- and 2-fold decrease in the EpCAM expression levels at the corresponding concentrations of si-RNA suggesting that higher cell internalization is a key player in modulating gene silencing efficacy. Figure 8(b) depicts the cell viability of LC liposomes and LC immunoliposomes and it reveals no toxic effect of LC liposomes in MCF-7 cells.

Figure 8. (a) EpCAM silencing efficacy of LC liposomes and immunoliposomes with 100 and 200 nM si-RNA concentration after 48 h (* < 0.05), (b) Cell viability of LC liposomes and LC immunoliposomes.

In vivo studies

In order to evaluate the potential of nanocarrier-mediated EpCAM silencing system on tumor regression, animal studies were performed using severely compromised immunedeficient (SCID) mice.

Figure 9 shows the results of the in vivo experiments in xenografted breast cancer SCID mice. Figure 9(a) and (c) show the change in the body weight of the animals and tumor volume during the treatment period, respectively. It is seen that there is no significant change in body weights of the treated mice during the period and they tend to decrease with time. The normal mice that served as the negative control, however, exhibited a slight increase in the body weight during the same period. Tumor volume was observed to increase gradually in the control group, whereas it was noticed that the mice treated with antibody-linked LC liposomes show 32% reduction in tumor volume when compared with positive control group after receiving eight doses of 0.15 mg/kg spread over 28 d.

Figure 9. (a) Body weight of mice from different groups during treatment period, (b) tumor images weight after necrosis, (c) tumor volume during treatment period and (d) qRT-PCR data of EpCAM expression in tumor tissue (*<0.05).

Figure 9(b) represents the tumor weight after necropsy of different groups and the results are in agreement with the tumor volume data. A significant reduction in the tumor weight is observed in the mice treated with immunoliposomes when compared with the positive control. EpCAM gene expression in excised tumor tissue was analyzed using qRT-PCR. The results (Figure 9d) reveal that antibody-linked LC liposomes (immunoliposomes) down-regulate EpCAM expression by 2-fold in comparison to LC liposomes where significant silencing of EpCAM was not observed.

Histopathological analysis

The tumor was excised and histopathological analysis was carried out after 28 days of treatment. Figure 10 shows the hematoxylin and eosin stained tumor tissue from negative control, positive control, mice treated with LC liposomes and antibody tagged LC liposomes. It is observed that the positive control group show large areas of necrosis and hemorrhage. In contrast, the tissue sections from the mice treated with LC liposomes as well as the group treated with immunoliposomes exhibit densely packed cuboidal to low columnar epithelial cells (arrow) and myoepithelium (arrowhead) cells with negligible hemorrhage and necrosis. Figure 9(e) shows the Western blot analysis of excised tumors from the control free si-RNA, LB liposome and LB immunoliposomes. Band intensity of immunoliposomes as well as LC liposomes was lesser than the control and the free si-RNA group.

Figure 10. Histopathological staining images of (A) normal skin from negative control mice, (B) tumor tissue from positive control, (C) mice treated with Liposomes, (C) liposomes and (D) antibody linked liposomes group. Western blot showing EpCAM protein bands and the band intensity calculated using imagej software.

Discussion

Liposomal carriers have been extensively investigated for the delivery of oligonucleotides and si-RNA for the treatment of cancer and many genetic disorders (Ozpolat et al., 2014). Colloidal stability and protein adsorption on the surface of liposomes, however, determine the cell uptake and circulation half-life of the liposomes in serum, which in turn influences the therapeutic efficacy of the system. The charge on the liposome determines its colloidal stability, the nature and magnitude of protein adsorption, which is responsible for the extent of immune response produced against the carrier and their clearance from the body (Ishida et al., 2002). Hence, there exists the need to develop a liposomal carrier that possesses adequate life-time in the body without compromising its stability and therapeutic efficacy to realize the true potential of gene therapy applications.

Confinement of a highly charged hydrophilic molecule like si-RNA into the liposome is chiefly dependent on the composition of lipids used and their interaction with si-RNA. DOPE possesses fusogenic property because of its ability to undergo phase transformation from lamellar to hexagonal phase in response to pH changes. This property of DOPE can be invaluable in gene delivery as it enables better cell internalization and facilitates endosomal escape (Li et al., 2014). Hence DOPE has been used as a “helper lipid” along with cationic lipids in many gene delivery systems (Farhood et al., 1995; Huang et al., 1995). However, use of cationic lipids tends to increase cytotoxicity and trigger inflammatory response in biological systems and hence cationic lipid-free liposomes were explored in this study. Among the nine different liposomes with different ratios of egg phosphatidyl choline, DOPE, cholesterol and DSPE-PEG, the liposomes with ratios 7:3:0:0 (LB), 7:2:0:1 (LC) and 7:1:1:1 (LD) exhibit significantly high encapsulation of si-RNA when compared to other lipid ratios. The highly polar si-RNA favors a polar environment for encapsulation, which is provided by the hydrophilic dynamic chains of polyethylene glycol present in LC and LD liposomes. However, introduction of cholesterol decreased the encapsulation due to its hydrophobic nature. Interestingly, increasing DOPE concentrations beyond 30% was found to decrease the encapsulation of si-RNA, which may be attributed to the structure of DOPE that does not aid close packing of the lipids thereby making the membrane leaky and fluid.

The increase in particle size of the liposomes after encapsulation confirms the presence of si-RNA in the liposomes. PEGylated liposomes display more negative zeta potential values as the PEG chains tend to mask the negative surface groups from being neutralized by the cations from the buffer (Ramana et al., 2012). The phase transition temperature of the liposomes exhibit a positive shift after encapsulation of si-RNA. This may be attributed to the increased associative forces in the liposomes due to electrostatic interactions between the phospholipids and si-RNA that restricts the mobility of the acyl chains. This is reflected in the increase in the phase transition of the liposomes after encapsulation with si-RNA. The presence of DOPE was found to decrease the phase transition temperature owing to its ability to hinder close packing of the lipid chains, while the cholesterol tends to increase the phase transition temperature of liposomes due to its stiff phenanthrene ring structure that impedes acyl chain mobility.

Colloidal stability of liposomes influences their life-time in the biological system as agglomeration will lead to immune recognition (Kapoor & Burgess, 2012). Agglomeration of liposomes will also be reflected through a change in surface charges. The zeta potential measurements of the liposomal samples with time shows a progressive shift to less negative values indicating neutralization of surface charges due to aggregation. However, the measurement of zeta potential in PBS has been reported to mask the surface charges and lower the value of the zeta potential (Moghaddam et al., 2011). The shift is less pronounced in the case of si-RNA-loaded liposomes when compared with their corresponding blank counterparts implying better colloidal stability. The shift in zeta potential is also found to be less steep in the case of PEGylated liposomes containing si-RNA, which may be attributed to the PEG chains on the liposomal surface that retard aggregation of liposomal carriers through steric hindrance.

Opsonins are a class of proteins which promote phagocytosis and subsequent removal of particles from the blood circulation (Ishida et al., 2002). Therefore, adsorption of these proteins on liposomes will have an important role in deciding the fate of the delivery system. Albumin, an abundant component of plasma, belongs to the opsonin class of proteins (Aggarwal et al., 2009). In the present study, the adsorption of albumin and plasma protein with time was found to be least in PEGylated liposomes when compared to the non-PEGylated liposomes. This may be attributed to the fast moving chains of PEG on the surface that tend to repel proteins. Interestingly, si-RNA encapsulated liposomes also display low levels of albumin adsorption which suggests that some si-RNA that is present on the surface of liposomes may electrostatically repel the anionic albumin thereby retarding its adsorption. In contrast, the levels of plasma protein adsorbed on the liposomal carriers over the same period of time were found to be significantly higher. This may be due to the presence of a large number of proteins with varying sizes and isoelectric points in the plasma, which can contribute to an accelerated adsorption of proteins (Chono et al., 2008; Ramana et al., 2010).

One of the contributing factors for liposome aggregation with time is a phenomenon called “Ostwald ripening” (Keck et al., 2012). Colloidal stability of liposomes systems can be explained by the Derjaguin-Landau-Verwey-Overbeek (DLVO) theory, which describes that the stability of the colloidal systems is determined by the net effect of repulsive electrical double layer and the attractive van der Waals forces that the particles experience as they contact one another (Sabín et al., 2006). Factors like adsorption of proteins and neutralization of surface charges promote van der Waal's associative forces and reduce the repulsive forces. Therefore, the energy barrier must not be overcome by van der Waal's forces and the repulsive forces should stay dominant to prevent the formation of aggregates. In the present study, effect of albumin and plasma protein adsorption on the particle size distribution of blank and si-RNA-loaded liposomes of all the four lipid combinations (LA, LB, LC and LD) shows interesting outcomes. PEGylated liposomes (LC and LD) with si-RNA incubated with albumin or plasma protein do not show significant changes in the particle size distribution with time, which may be attributed to the steric repulsion offered by the fast moving PEG chains on the surface. But, non-PEGylated liposomes formed from only egg PC (LA) and egg PC and DOPE (LB) show a discernible change in the particle size distribution with an appearance of distinct micron sized populations suggesting progressive adsorption of the proteins on the liposomal surface promoting their aggregation.

The release of si-RNA from the LC liposomes show a sustained release profile, which may be attributed to the presence of fast moving PEG chains on the surface that impede the diffusion of the si-RNA from the liposomes. The low burst release of si-RNA from LC liposomes also imply that most of the si-RNA is localized in the interior due to the presence of PEG chains on the surface. In contrast, the DOPE containing non-PEGylated liposomes (LB) show a pronounced burst release and a rapid release. This may be due to the presence of the unsaturated acyl chains of DOPE that hinder the close packing of lipids leading to creation of defects that facilitate the diffusion of si-RNA from the liposomes. Cholesterol containing liposomes (LD) also display a high-burst release and rapid release of si-RNA suggesting that the si-RNA is localized near the periphery of the liposomes owing to the presence of the rigid phenanthrene rings of cholesterol.

MCF-7 breast cancer cell line is a type of epithelial cancer that over-expresses the EpCAM receptor. Hence, LC liposomes linked with the EpCAM antibody exhibited better cell internalization when compared to the LC liposomes without antibody. When the cells were pre-incubated with EpCAM antibody, the internalization of the immunoliposomes was significantly reduced. This is because the surface EpCAM receptors are bound to the EpCAM antibody preventing the binding of the immunoliposomes. This result also proves that the internalization of the immunoliposomes is mediated through receptor-mediated endocytosis. The enhanced down-regulation of EpCAM gene expression levels after treatment with immunoliposomes is also a direct consequence of the better internalization of the immunoliposomes when compared with the unmodified LC liposomes. This is also reflected in the in vivo studies where the immunoliposomes-treated mice show better reduction in tumor volume with no significant change in body weight. This can be ascribed to the ability of the immunoliposomes to deliver its cargo more efficiently into the cancer cells when compared with the non-targeted liposomes as well as free si-RNA. The results also suggest that targeting cancer cells using EpCAM antibody can enhance the efficiency of the tumor treatment. The histopathological staining of tumor tissue depicts less necrotic regions in mice treated with immunoliposomes with presence of densely packed cuboidal to low columnar epithelial cells. Qualitative and quantitative analysis of EpCAM silencing in tumor samples are in agreement with the earlier in vitro inference that the immunoliposomes possess better cell internalization and higher EpCAM silencing efficiency when compared to its non-targeted counterpart and free si-RNA. It was also observed in PCR and Western blot analysis of tumor tissues from various treatment groups that the immunoliposomes significantly silence the EpCAM gene and reduces the EpCAM protein levels in cells. As the down-regulation of EpCAM has been associated with up-regulation of apoptotic genes and down-regulation of proliferation genes (Mitra et al., 2010), silencing of EpCAM through immunoliposome-mediated delivery resulted in a better tumor regression when compared to other treatment groups. The si-RNA concentrations used in this study (0.15 mg/kg) are among the lowest reported for gene silencing in vivo (Casals et al., 2003). Increase in the si-RNA concentrations or increase in the duration of treatment can result in complete regression of the tumor.

Conclusion

The development of cationic lipid-free liposomes targeting EpCAM expressed on the surface of epithelial cancers for silencing the EpCAM gene has been proved to be a feasible strategy to bring about tumor regression even at very low concentrations. This proof-of-concept could be taken forward to evaluate its efficacy against other types of epithelial cancers. The use of PEGylated nanoliposomes has enabled realization of the promise of gene silencing strategies for cancer therapy.

Declaration of interest

The authors have no conflict of interest to declare.

The authors wish to acknowledge funding from Department of Biotechnology Government of India (BT/PR11210/NNT/28/2008) and SASTRA University for infrastructural support.