DOPE, Chol, RIPA buffer, bicinchoninic acid (BCA) kit were obtained from Sigma-Aldrich (St. Louis, MO, USA). 2-[-(2-hydroxyethyl) piperazinyl]-ethanesulphonic acid (HEPES), ethidium bromide (10 mg/ml), tris(hydroxymethyl)-aminomethane hydrochloride (Tris HCl), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), acridine orange (AO), ethylenediaminetetraacetic acid (EDTA), phosphate buffered saline tablets (PBS), sodium carbonate (Na ₂CO ₃) and sodium bicarbonate (NaHCO ₃) were supplied by Merck (Darmstadt, Germany). The siGENOME non-targeting siRNA #1 (D-001210-01-20), ON-TARGETplus SMARTpool Human MYC (4609) siRNA (L-003282-02-0020), 5× siRNA buffer (B-002000-UB-100, 0.3M KCl, 30mM HEPES, 1mM MgCl ₂, pH 7.5) and molecular grade RNase-free water (B-003000-WB-100) were purchased from Thermo Scientific Dharmacon Products (Lafayette, CO, USA). Ultrapure ^™ agarose powder and Lipofectamine ^™ 3000 reagent were procured from Invitrogen (Carlsbad, CA, USA). Sodium dodecylsulphate (SDS), iScript ^™ genomic DNA (gDNA) clear complementary DNA (cDNA) Synthesis Kit (1725035), SsoAdvanced ^™ Universal SYBR ^® Green Supermix (1725272), PrimePCR ^™ SYBR ^® Green Assay: MYC, Human (Unique Assay ID: qHsaCID0012921); PrimePCR ^™ SYBR ^® Green Assay: ACTB, Human (Unique Assay ID: qHsaCED0036269), PrimePCR ^™ Template for SYBR ^® Green Assay: MYC, Human; PrimePCR ^™ Template for SYBR ^® Green Assay: ACTB, Hard-Shell ^® 96-well PCR plates, 0.2 ml PCR tube strips, 10× Tris-buffered saline (TBS), 10% Tween 20, nonionic detergent; and blotting-grade blocker (non-fat dry milk), were supplied by Bio-Rad Laboratories (Richmond, CA). MCF 7 and HT-29 cells were obtained from the American Type Culture Collection (Manassas, VA, USA). Eagle’s Minimum Essential Medium (EMEM) with L-glutamine, Fetal bovine serum (FBS), trypsin-EDTA and penicillin/streptomycin mixture (10 000 U/ml penicillin, 10 000 μg/ml streptomycin) were purchased from Lonza BioWhittaker (Verviers, Belgium). All sterile plasticware were obtained from Corning Inc. (Corning, NY, USA). TRIzol ^® reagent (15596-026), c- MYC epitope tag antibody 9E11 (AHO0052), goat anti-mouse IgG2A secondary antibody, horseradish peroxidase (HRP) conjugate were obtained from Life Technologies (Carlsbad, CA, USA). β-actin antibody was purchased from Santa Cruz (Santa Cruz, CA, USA). Ultrapure 18 Mohm water was used throughout. All other reagents were of analytical grade.

MS09 was prepared from cholesterylformylhydrazide hemisuccinate (MS08) via an active ester intermediate according to the method previously published ²³ . Briefly, the N-hydroxysuccinimide ester of cholesterylformylhydrazidehemisuccinate (0.083 mmol) and dimethylpropylamine (0.35 mmol) were dissolved in a H ₂O:pyridine:DMF (13:7:10, v/v/v) mixture. The product (MS09) was monitored and purified by thin layer chromatography. For liposome formulation, the lipids were combined in quantities as in Table 1 and concentrated to a thin film in vacuo (Büchi Rotavapor R rotary evaporator). The film was rehydrated (4°C, 48h) in sterile HEPES buffered saline (HBS, 20 mM HEPES, 150 mM NaCl, pH 7.5; 0.5 ml), vortexed and, then sonicated in a Transsonic (T460/H) bath-type sonicator (Singen, Germany). Liposome preparations were stored at 4°C and sonicated prior to use.

The morphology of liposomes and siRNA lipoplexes were evaluated using cryogenic transmission electron microscopy (cryo-TEM), using a JEOL JEM.1010 transmission electron microscope (JEOL Ltd., Tokyo, Japan). Images were captured using an Olympus MegaView III digital camera in conjunction with SIS iTEM Universal Imaging Platform software (Shinjuku, Japan). Particle size and zeta (ζ) potential were determined using the Nanoparticle tracking analysis (NTA, NanoSight NS500, Malvern Instruments, Worcestershire, UK) fitted with a ZetaSight™ module and NTA 3.0 software.

Lipoplexes were prepared ( ^w/ _w) with varying amounts of the liposomes and a fixed quantity of siGENOME non-targeting siRNA (0.3 μg), as per Table 2. Lipoplexes (10 μl in HBS) were mixed with gel loading buffer (40% sucrose, 0.25% xylene cyanol, 0.25% bromophenol blue; 2.5 μl) and loaded onto 2% agarose gels. Electrophoresis was carried out for 30 min in a Mini-Sub ^® Cell GT electrophoresis cell (Bio Rad, Richmond, CA) containing tris-phosphate-EDTA (TPE) running buffer (36mM Tris-HCl, 30mM NaH ₂PO ₄, 10mM EDTA, pH 7.5) at 50 V. Gels were stained with ethidium bromide (0.1 μg/ml in water) for 30 min, and images were viewed and captured using a Vacutec Syngene G:Box gel documentation system, fitted with GeneSnap software, version 7.05.02 (Syngene, Cambridge, UK). Densitometric analysis was performed with the associated GeneTools software. The fluorescence intensities of unbound siRNA were expressed as a percentage against that of a naked siRNA control. The amount of liposome-associated siRNA at each MS09:siRNA ( ^w/ _w) ratio was determined as follows: % bound siRNA = 100 − % free siRNA

Lipoplexes each containing 0.3 μg non-targeting siRNA were assembled at ( ^w/ _w) ratios of 12:1–32:1. These were incubated with 10% FBS at 37°C for 4h. A control, 0.3 μg naked siRNA was treated similarly. Nuclease activity was terminated using EDTA (10mM), and complexes destabilized with 0.5% ( ^w/ _v) SDS (55°C, 25 min). This was followed by electrophoresis as described previously.

MCF-7 and HT-29 cells were seeded at densities of 4.0 × 10 ⁴ cells/well in 48 well plates, and maintained at 37°C for 24h, followed by treatment with lipoplexes containing non-targeting siRNA (14 nM) at ratios used in the nuclease protection assay. Controls with naked siRNA (14 nM) were included. Lipofectamine™ 3000 (LF3K), prepared according to the manufacturer’s protocol was used as a positive control. The LF3K transfecting complex (25 μl) had a final siRNA concentration of 25 nM. At 48h post-transfection, growth medium was aspirated, and cell viability was assessed by the MTT assay. Cells were incubated (37°C, 4h) with 200 μl medium containing 20 μl MTT (5 mg/ml in PBS) per well. Wells were then drained, and formazan crystals dissolved in DMSO (200 μl/well) to produce purple-colored solutions. Absorbance was read at 540 nm in a Mindray microplate reader, MR 96A (Vacutec, Hamburg, Germany), against a DMSO blank. Percentage cell viability was calculated as follows: [ A 540 nm ( treated cells ) − A 540 nm ( blank ) ] [ A 540 nm ( untreated cells ) − A 50 nm ( blank ) ] × 100

In all ensuing experiments, MS09:Chol and MS09:DOPE lipoplexes (MS09:siRNA ^w/ _w = 16:1) contained 12 nM siRNA. MCF-7 and HT-29 cells were seeded in 6-well plates at densities of 3.0 ×10 ⁵ cells/well and incubated as previously described. Cells were transfected with lipoplexes as previously described. Lipoplexes were assembled with 12 nM of either non-targeting siRNA or ON TARGETplus SMARTpool Human myc-siRNA (anti-c- myc-siRNA). Transfections were carried out in triplicate, followed by cell harvesting for total RNA and protein.

Total cellular RNA was extracted using TRIzol ^® Reagent, and cDNA synthesis was carried out using the gDNA Clear cDNA Synthesis Kit, as per manufacturer’s instructions. For a single reaction, 0.8 μg of the RNA sample was used. Reaction mixtures were prepared on ice and the reactions were carried out in a C1000 Touch™ Thermal Cycler (Bio-Rad Laboratories (PTY) Ltd., Richmond, USA). The reaction was performed as follows: DNA digestion (25 °C, 5 min), DNase inactivation (75 °C, 5 min), and samples were maintained at 4 °C for 10 min. The RT supermix (4 μl) was added and cDNA synthesis was carried out as follows: priming (25 °C, 5 min), reverse transcription (46 °C, 20 min), RT inactivation (95 °C, 1 min) and samples were held at 4 °C for 10 min. Two cDNA synthesis reactions per RNA isolate were performed in parallel i.e. one reaction containing the RT supermix and a no RT control in which the no-RT supermix was added instead. cDNA samples were diluted to a final concentration of 25 ng/μl in nuclease-free water and stored at 4 °C, for no more than a week. The product of each cDNA synthesis reaction was subjected to RT-qPCR. A single reaction mixture (20 μl) contained SsoAdvanced™ Universal SYBR ^® Green Supermix (10 μl); primers (1 μl) specific for either the target gene, c-MYC (PrimePCR™ SYBR ^® Green Assay: MYC, Human) or reference gene, β-actin (PrimePCR™ SYBR ^® Green Assay: ACTB, Human); cDNA sample (25 ng, 1μl) and nuclease-free water (8 μl). Reaction mixtures in which DNA templates (either PrimePCR™ Template for SYBR ^® Green Assay: MYC, Human; or PrimePCR™ Template for SYBR ^® Green Assay: ACTB, Human; 1 μl) were substituted for cDNA served as positive controls. Reaction mixtures where nuclease-free water (1 μl) was substituted for either primers or cDNA were included as negative controls. All reactions were performed in triplicate. Reaction mixtures were prepared, on ice, in Hard-Shell ^® 96 well plates, sealed, briefly centrifuged and then loaded in a C1000 Touch™ Thermal Cycler (CFX 96 Touch™ Real-Time PCR Detection System, Bio-Rad Laboratories (PTY) Ltd., Richmond, USA). Data was analyzed with CFX Manager™ Software version 3.0, and c-myc expression was normalized to β-actin using the ΔΔCq comparative quantification algorithm.

MCF-7 and HT-29 cells were seeded in 6-well plates at densities of 3.0 ×10 ⁵ cells/well and incubated as previously described. Cells were transfected with lipoplexes as previously described. The medium was removed, and cells washed twice with ice-cold PBS (1 ml/well). Cold (4°C) RIPA buffer (100 μl/well) was then added and cells placed on ice for 5 min with gentle shaking to dislodge cells. Cell suspensions were centrifuged (14 000 × g, 4°C, 15 min) and lysates/protein extracts were immobilized onto the wells of a 96-well plate with 50 mM carbonate-bicarbonate coating buffer (pH 9.6, at 4°C), overnight. Three replicates per isolate were performed. Each well received 10 μg protein in 100 μl coating buffer. The coating buffer was removed, and wells were rinsed twice with TBS (20 mM Tris-HCl, pH 7.5, 150 mM NaCl) containing 0.1% Tween 20 (TBST, 100 μl/well). Wells were then treated with 5% non-fat dry milk in TBST (100 μl) with gentle agitation to saturate unoccupied attachment sites. The blocking agent was removed, and wells rinsed twice with TBST (100 μl/well). Either c-myc (1:2000, in TBST) or β actin (1:10 000, in TBST) primary antibodies were added (100 μl/well) and incubated at room temperature for 1h. Primary antibodies were removed and wells washed with TBST (4x, 100 μl/well) for 5 min each, with agitation. The secondary antibody (1:2000, in TBST) was then added and incubated at room temperature for 1h. Wells were drained and washed with TBST, as previously. TMB (100 μl/well) was applied (room temperature, 30 min), and the reaction was terminated by the addition of 2M H ₂SO ₄ (100 μl/well). Absorbance was measured at 450 against a TMB (100 μl)/2M H ₂SO ₄ (100 μl) blank. Wells containing BSA (10 μg) served as negative controls. Antibody-free and substrate-free controls were included. Expression of c-myc was normalized to β actin and presented relative to untreated cells.

Live, apoptotic and necrotic cells were distinguished by the AO/ethidium bromide (EtBr) dual staining method ²⁵ . Lipoplexes were introduced to semi-confluent cells (8.0 ×10 ⁴ cells/well) in 24-well plates as previously described. After 48h, cells were rinsed with PBS (200 μl/well), stained with AO/EtBr solution (100 μg/ml AO and 100 μg/ml EtBr in PBS; 10 μl/well). Excess stain was removed by rinsing with PBS (100 μl/well). Cellular changes associated with apoptosis were observed under an inverted fluorescence microscope (CKX41, Olympus, Japan) at excitation and emission wavelengths of 540 nm and 580 nm, respectively. Images were acquired at 200× magnification using Analysis Five Software (Olympus Soft Imaging Solutions, Olympus, Japan). The % live/apoptotic/early apoptotic/late apoptotic/necrotic cells were calculated as below: [number of live/apoptotic/early apoptotic/late apoptotic/necrotic cells] [total cells counted] × 100

Statistical analyses were performed using one-way analysis of variance (ANOVA), followed by Tukey’s Multiple Comparison Test to compare between groups using GraphPad Prism version 5.04 (GraphPad Software Inc., USA). P values less than 0.05 were considered significant.

TEM presented MS09:DOPE and MS09:Chol formulations as round to irregular shaped unilamellar vesicles, respectively ( Figure 1a,b). The liposome-siRNA complexes ( Figure 1c,d) assumed structures that were different from the vesicles, emphasizing the heterogeneity and assembly of the liposome-siRNA complexes.

NTA results ( Table 3) show that MS09:DOPE and MS09:Chol liposome sizes were below 150 nm in size and may be classified as small unilamellar vesicles. The substitution of DOPE with Chol seems to have no significant effect on size or zeta potential, which remained high. It was observed that Chol is not likely to influence the electrical surface potential of liposomes because it does not bear an ionizable group ²⁶ . Lipoplexes were less than 150 nm in size ( Table 3), which is important for passive targeting of tumour cells via the enhanced permeability and retention effect.

Zeta potential measurements are based on the interaction of the particle with ions in the medium in which it is dispersed. In this study, liposomes were dispersed in HEPES buffer, which can influence the zeta potential of a bilayer depending on the orientation of the molecule’s ionizable groups relative to the membrane ^{27, 28} in the electrical double layer. Hence, the negative zeta potential values do not necessarily imply that the liposomes would be unable to associate with siRNA molecules. Lipoplexes similarly displayed negative zeta potentials ( Table 3). Although it is accepted that the net positive charge of lipoplexes allows for binding to anionic membrane-associated proteoglycans to initiate cellular uptake ²⁹ , it is also possible for siRNA lipoplexes with negative zeta potential to enter cells and successfully facilitate gene silencing.

The gel retardation assay or band shift assay is widely documented as the first step in assessing the siRNA-binding ability of cationic carriers ^{30– 32} . This assay is based on the premise that the migration of siRNA is retarded in an electric field when bound to a carrier due to the formation of electroneutral complexes that are unable to permeate the gel matrix. MS09:DOPE was capable of fully preventing the migration of siRNA ( Figure 2a), whereas Figure 2b shows that although there was a decrease in unbound siRNA with increasing liposome content for the MS09:Chol formulation, unbound siRNA was evident at all MS09:siRNA ( ^w/ _w) ratios. Densitometric analysis ( Figure 3) confirmed that the binding of siRNA by these liposome formulations increased until a point, whereupon free siRNA in the gel persisted despite addition of liposome. This MS09:siRNA ( ^w/ _w) ratio was taken as the end-point/optimum binding ratio of these liposomes.

Attempts at lipid-mediated siRNA delivery are often frustrated by adverse interactions with serum. Figure 4 shows that, while naked siRNA was entirely degraded under experimental conditions (lanes 2), intact siRNA bands, less intense than the untreated control, were clearly visible in all instances. This shows that the liposomal formulations partially protected siRNA at the respective MS09:siRNA (w/w) ratios. Densitometric analysis of gels ( Figure 5) provided further insight into the siRNA-protecting capabilities of the liposomes. Maximum intact siRNA of 73% afforded by MS09:Chol liposomes at the MS09:siRNA ( ^w/ _w) ratio of 24:1 was significantly less than that of MS09:DOPE. 25% of siRNA was likely to be surface associated, as it was so loosely bound that it was coaxed off during electrophoresis. Such siRNA is readily detached from the liposomal bilayer upon exposure to serum nucleases ³³ .

It was important that any growth inhibitory effects in the cancer cells be attributed solely to the delivered siRNA, and not due to any intrinsic harmful effect of the liposomal carrier. This assay is based on the principle that enzymes of the mitochondria of living cells reduce soluble MTT, a tetrazolium salt, to formazan crystals ³⁴ . The intensity of the resultant purple solution correlates with the extent of MTT reduction, and the number of viable cells ³⁵ . No significant reduction in viability was detected upon treatment of all cells with naked siRNA at concentrations contained in lipoplexes ( Figure 6). Cells retained viability of at least 88% with exposure to LF3K-siRNA complexes. In general, cell survival after exposure to the liposomal formulations were greater than 85% with no severe cytotoxicity evident.

Given that the initial RNAi effect is exerted at the mRNA level, the effect of transfection with anti- c-myc lipoplexes on c-myc transcripts in cancer cells was studied using RT-qPCR. Figure 7 shows a decrease in c-myc-mRNA in instances where a transfecting agent was used to deliver anti- c-myc-siRNA sequences. Quantification of cellular c-myc protein by ELISA ( Figure 8) showed that a decrease in c-myc mRNA levels was, in all cases, accompanied by a concomitant reduction in protein expression. Complexes assembled with non-targeting siRNA were without effect. This confirmed that the observed reduction in c-myc-mRNA contributed to the RNAi effect of the delivered anti- c-myc-siRNA. The fact that naked anti- c-myc-siRNA did not influence c-myc expression in any way, highlights the need for a delivery vehicle.

MS09:Chol and MS09:DOPE lipoplexes produced significant gene silencing compared to LF3K in both cell lines. In the MCF-7 cell line, MS09 lipoplexes achieved 8- and 3.5-fold greater knockdown of c-myc than LF3K at the mRNA and protein levels, respectively. In HT-29 cells, the decrease in c-myc-mRNA and protein was 5- and 2.8-fold more for MS09 lipoplexes. The superior performance of MS09 lipoplexes is highlighted by the fact that these contained at half the final siRNA concentration as LF3K.

A desirable feature is for cancer treatment to induce cancer cell death without causing harm to surrounding healthy tissue. Hence, several anticancer approaches exploit the natural mechanisms of cell death such as apoptosis, that normally eliminates damaged and/or harmful cells in a regulated fashion ³⁶ . The AO/EtBr method is based on the principle that AO enters cells with intact plasma membranes and binds to DNA to emit green fluorescence, while EtBr enters cells with defective membrane integrity and fluoresces red-orange when bound to DNA. Differentiation between normal, early apoptotic, late apoptotic and necrotic cells was made based on observations of nuclear morphology ( Figure 9 and Figure 10). Live cells were characterized by a bright green nucleus in the center of the cell. The nuclei of early apoptotic cells, with undamaged membranes, also stained green, but appeared to be fragmented or condensed. In contrast, the nuclei of late apoptotic cells, with compromised membrane integrity, stained orange with evidence of fragmentation or condensation. Finally, necrotic cells were characterized by an intact bright orange nucleus ³⁷ .

The major mechanism of cell death observed with these anti- c-myc lipoplexes was apoptosis, which is similar to other studies demonstrating that inhibition of c-myc in cancer cells leads to apoptosis ^{38– 40} . Importantly, necrosis, a non-specific form of cell death that is associated with an inflammatory response ⁴¹ , was negligible in all instances, accounting for less than 3% of total cells per sample. Hence, MS09:Chol and MS09:DOPE-mediated anti- c-myc-siRNA delivery is capable of destroying cancer cells without damaging healthy tissue. The application of anti- c-myc-siRNA on its own did not result in any anticancer activity. Hence we can conclude that the anticancer effects can be ascribed to the RNAi activity of liposome bound anti- c-myc-siRNA.

Zenodo: Anti-c-myc cholesterol based lipoplexes as onco-nanotherapeutic agents in vitro, http://doi.org/10.5281/zenodo.3946640 ⁶⁶ .

This project contains the following underlying data:

Uncropped/unedited electron microscopy images showing morphology of liposomes and siRNA lipoplexes

Zenodo: Anti-c-myc cholesterol based lipoplexes as onco-nanotherapeutic agents in vitro, http://doi.org/10.5281/zenodo.3940426 ⁶⁶ .

This project contains the following extended data:

Number of lipid molecules that constitute liposomal vesicles

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).