Research Article For reprint orders, please contact: [email protected]
Cancer-targeting siRNA delivery from porous silicon nanoparticles
Aims: Porous silicon nanoparticles (pSiNPs) with tunable pore size are biocompatible and biodegradable, suggesting that they are suitable biomaterials as vehicles for drug delivery. Loading of small interfering RNA (siRNA) into the pores of pSiNPs can protect siRNA from degradation as well as improve the cellular uptake. We aimed to deliver MRP1 siRNA loaded into pSiNPs to glioblastoma cells, and to demonstrate downregulation of MRP1 at the mRNA and protein levels. Methods: 50–220 nm pSiNPs with an average pore size of 26 nm were prepared, followed by electrostatic adsorption of siRNA into pores. Oligonucleotide loading and release profiles were investigated; MRP1 mRNA and protein expression, cell viability and cell apoptosis were studied. Results: Approximately 7.7 µg of siRNA was loaded per mg of pSiNPs. Cells readily took up nanoparticles after 30 min incubation. siRNA-loaded pSiNPs were able to effectively downregulate target mRNA (~40%) and protein expression (31%), and induced cell apoptosis and necrosis (33%). Conclusion: siRNA loaded pSiNPs downregulated mRNA and protein expression and induced cell death. This novel siRNA delivery system may pave the way towards developing more effective tumor therapies.
Yuan Wan1,2, Sinoula Apostolou3, Roman Dronov2, Bryone Kuss3 & Nicolas H Voelcker*,1 Mawson Institute, University of South Australia, Mawson Lakes, Adelaide, SA 5095, Australia 2 School of Chemical and Physical Sciences, Flinders University, Bedford Park, Adelaide, SA 5042, Australia 3 School of Medicine, Flinders University, Bedford Park, Adelaide, SA 5042, Australia *Author for correspondence: nico.voelcker@ unisa.edu.au 1
Original submitted 14 August 2013; Revised submitted 4 December 2013 Keywords: drug delivery • glioblastoma • MRP1 • porous silicon nanoparticles • siRNA • ToF-SIMS
dsRNA triggers sequence-specific gene silencing in eukaryotic cells by a process known as RNA interference (RNAi) [1] . Exogenous synthetic small interfering RNA (siRNA), which usually consists of 21–23 nucleotides, introduced into the cytoplasm binds to the RNA-induced silencing complex (RISC). This binding event induces unwinding and separation of strands, and is followed by association with target mRNA. The tagged mRNA is then degraded by Slicer (Argonaute-2), an enzyme residing within the RISC complex [2] , resulting in downregulation of cellular protein production [3] . This process is highly specific and has the potential to silence any gene of interest [4] . siRNA delivery is therefore a promising gene therapy approach with relevance to a range of diseases including cancer [5] . MRP1, located at the plasma or intercellular membranes, is a member of ATP-dependent
10.2217/NNM.14.12 © 2014 Future Medicine Ltd
transporter protein [6] . It plays a role in cellular defense and protects healthy cells from toxic or harmful agents [7] . However, in various tumor types including human glioblastoma, breast cancer, and so on, the overexpression of MRP1, which acts as a membrane efflux pump, can increase transport of drugs out of the cancer cells [8,9] . Moreover, MRP1 protein levels correlate with previous exposure to chemotherapeutic drugs and with the tumor grade [10] . High MRP1 activity is a predictor of poor response to chemotherapy and a reduced overall survival of patients [11] . Various strategies including targeting membrane transporters [12] , inhibition of cell survival pathways [13] , altering transcription factors [14] and silencing antiapoptotic pathways have been developed to suppress multidrug resistance [15] . A promising sensitization strategy is the use of siRNA to silence MRP1 gene expression. On one hand,
Nanomedicine (Lond.) (2014) 9(15), 2309–2321
part of
ISSN 1743-5889
2309
Research Article Wan, Apostolou, Dronov, Kuss & Voelcker effective blocking of MRP1 protein expression can enhance chemosensitivity [16] ; on the other hand, siRNA can cause significant apoptosis in neuroblastoma due to the unique effect of MRP1 on tumor biology [17,18] . Although siRNA has shown significant potential as a new gene therapy, siRNA-based therapies face a number of key challenges that need to be overcome [19] . The polyanionic nature of bare siRNA results in poor binding to and transfer across the negatively charged cell membranes [20] . In addition, exogenous siRNA undergoes degradation upon exposure to nucleases [21] . Furthermore, side effects due to off-target immune responses can result from exposure of siRNA in the circulation to immune cells [22] . Viral vectors, liposomes, cationic polymers, recombinant proteins, electroporation and hydrodynamic gene transfer have been used to facilitate delivery of siRNA to target cells [23] . However, due to issues around cytotoxicity, high cost, low efficiency, and time-consuming and labor-intensive processing, these vehicles and methods have not been introduced into clinical practice. Porous silicon nanoparticles (pSiNPs) are a new addition to the nanoparticle-based drug delivery vehicle field [24,25] . These nanoparticles are unique in that they combine biocompatibility, biodegradability and high payload capacity [26–29] . The porous silicon (pSi) base is produced by a simple one-step electrochemical fabrication procedure. It is totally inorganic and can be readily sterilized. Pore dimensions can be controlled, from micropores (50 nm) [30] . It has a high surface area (400–800 m2/g) [30–32] . The material completely degrades in aqueous solutions into non-toxic silicic acid, the major bioavailable form of silicon [30] . pSiNPs with sizes in the range of 100–300 nm avoid renal filtration, leading to a prolonged residence time in the bloodstream that enables more effective targeting of diseases tissues, but these particles also effectively penetrate narrow tight epithelial junctions in the blood–brain barrier and deliver therapeutic agents [18] . pSiNPs have been deployed for bioimaging [25,26] , drug delivery [33–35] , and photodynamic [36] and photothermal therapy [37] . Using pSi microparticles, the Ferrari group has shown effective siRNA delivery. We have recently reported that pSiNPs are promising vehicles for targeted delivery of hydrophobic chemotherapy drugs [38] , and we have demonstrated siRNA loading onto pSiNPs with small pore sizes [39] . However, in our previous studies, the inner space of pores was not available for siRNA loading due to the small pore size, thereby limiting the loading capacity. Residual siRNA on the nanoparticle surface is subject to enzyme-mediated degradation in vivo, and therefore loading of siRNA into the pores is preferred [40] . Loading of siRNA into mesoporous silica nano-
2310
Nanomedicine (Lond.) (2014) 9(15)
particles has also been reported [21,41] . However, these nanoparticles are not biodegradable, and the tuning of pore size for siRNA delivery is limited. In this study, we report loading of MRP1 into pSiNPs that were prepared by electrochemical etching and grafted with amino groups on the inner and outer surfaces. These MRP1 siRNA-loaded nanoparticles showed uptake by T98G glioblastoma cells after 30 min incubation. MRP1 mRNA and protein were downregulated to approximately 40 and 30%, respectively, and 27% of cells underwent apoptosis. These results demonstrate that pSiNPs hold potential as a siRNA delivery vehicle for cancer therapy. Experiments & materials Cell lines & culture
The T98G glioblastoma line was obtained from Flinders Medical Center and cultured according to instructions. In brief, T98G cells were cultured in 25 ml of RPMI media 1640 (Invitrogen, CA, USA) containing 10% fetal bovine serum (FBS), 10 µl of 200 mM l-glutamine, 100 units/ml of penicillin and 100 µg/ml of streptomycin at 37°C in a humidified 5% carbon d ioxide atmosphere. Small interfering RNAs
ABCC1 siRNA duplexes (sense: 5´-GAG GCU UUG AUC GUC AAG UTT-3´; antisense: 5´-ACU UGA CGA UCA AAG CCU CTT-3´) and negative control siRNA duplexes (sense: 5´-UUC UCC GAA CGU GUC ACG UTT-3´; antisense: 5´-ACG UGA CAC GUU CGG AGA ATT-3´) were synthesized commercially by GenePharma (Shanghai, China). siRNA was resuspended with DEPC-treated water (Invitrogen). pSiNP preparation & characterization
pSiNPs were prepared by electrochemical etching of highly doped, (100)-oriented, p-type silicon wafers (Siltronix, France) in a 3:1 solution of 48% aqueous hydrofluoric acid (HF)/ethanol at a constant current density of 200 mA/cm2 for 150s; the resulting porous layer was then detached by electropolishing in a 3.3% HF in ethanol solution for 250s at a current density of 4 mA/cm2. The porous film was ultrasonicated in ethanol over 16 h; the supernatant was then filtered through a 0.2-µm nylon filter membrane (Carl Roth, Germany); the filtered solution was centrifuged at 22,000 g for 30 min, and the pellet of pSiNPs at the bottom was then resuspended in absolute ethanol. Diluted pSiNPs solution was dropped onto a graphite substrate, allowed to dry at room temperature (RT) and examined by scanning electron microscopy (SEM) with a Philips XL30 (Philips Co., OR, USA) and TEM with a FEI Tecnai G2 (FEI Co., OR, USA). Twenty pores were randomly selected from etched porous layer
future science group
Cancer-targeting siRNA d elivery from porous silicon nanoparticles
before electropolishing for measurement of pore size with SEM. The nanoparticle size distribution was measured using a Malvern Zetasizer Nano (ATA Co., Australia). The surface area of pSiNPs was determined from the Brunnauer–Emmett–Teller (BET) theory. Preparation of amine-functionalized pSiNPs
Approximately 3 mg of pSiNPs were functionalized with 2% (3-aminopropyl)triethoxysilane (APTES) in toluene for 1 h at RT or at 80°C, followed by rinsing with toluene and ethanol three times, respectively, to remove excess APTES and toluene. An alternative functionalization method involved pretreatment of pSiNPs in ethanol with microwave at 600 Watts for 3 min; the silane functionalization and rinse procedures followed the same protocol described above. Time-of-flight secondary ion mass spectrometry analysis
pSi membranes (diameter: 15 mm; thickness: 80 µm) were prepared and functionalized with APTES descr ibed above. A total of 40 µl of 25 µM DNA oligomer (5´TTT AGT TGA TTT GTG CTT CAG TGT GCT3´) mixed with 60 µl of ethanol were added onto the pSi membrane overnight at 4°C. The membrane was rinsed with DI water three times to remove unabsorbed DNA oligonucleotide and ethanol, then dried and broken in half. Positive and negative secondary ion images and spectra of the cross-section were acquired using timeof-flight secondary ion mass spectrometry (ToF-SIMS; Phi, MN, USA) with a pulse 25 keV, 1.3 pA primary ion beam in high current bunched mode from 50 × 50 µm areas on the sample top area. The collected data were analyzed with WinCadenceN software. Loading & release of siRNA
In total, 40 µl of 25 µM siRNA duplexes were mixed with approximately 1 mg of amine-functionalized pSiNPs in 500 µl of ethanol. siRNA was allowed to be absorbed onto the surface and pores of pSiNPs by shaking the mixture for 1 h or overnight at RT or at 4°C. Before and after siRNA loading, the zeta potential of the nanoparticles was determined using a Malvern Zetasizer Nano (ATA Co., Australia). The measurement were made at room temperature, and samples were diluted with DI water to an appropriate concentration to yield count rate per second (KCps) in the range of 2500–3500 KCps. Free siRNA was separated from adsorbed siRNA by centrifugation using a membrane centrifugal filter unit (Amicon Ultra-0.5 ml, Millipore, MA, USA). The concentration of free siRNA was determined by measuring the UV absorbance at 260 nm using a spectrophotometer (Nanodrop 1000, Thermo Fisher Science, MA, USA). The loaded siRNA
future science group
Research Article
amount equals the total siRNA loading amount prior to loading minus the amount of siRNA left in the supernatant after incubation. siRNA-loaded pSiNPs were resuspended in DEPC-water for siRNA release studies in vitro. At 0.5, 1, 2, 3 and 4 h time points pSiNPs were centrifuged, and siRNA concentration in supernatant was measured using the Nanodrop. siRNA uptake into cells
T98G glioblastoma cells were plated in six-well cell culture plates in 2.5 ml of RPMI antibiotic free media at a concentration of 1 × 105 cell/well (30–40% confluence) for 24 h prior to exposure to siRNA-loaded pSiNPs. The medium was removed and approximately 200 µl of mixture of culture medium containing siRNA-loaded pSiNPs were added. After 30 min, additional 2 ml fresh culture medium was added to maintain cell proliferation. First, cells were transiently transfected with 200 pmol FAM-labelled siRNA-loaded pSiNPs or with FAM-labelled siRNA alone. After 18 h incubation, cells were rinsed with 1× PBS thrice to completely remove remnant pSiNPs and/or FAM-labelled siRNA, and watched with the Nikon Eclipse (Nikon, Tokyo, Japan). MRP1 mRNA knockdown experiments were performed in triplicate with varying amounts of siRNA. The final amount of loaded siRNA corresponded to 100, 200 and 300 pmol. Three controls (cells without transfection, cell incubated with unloaded pSiNPs, and cells incubated with a mixture of 300 pmol of siRNA and bare pSiNPs) and a negative control comprising 300 pmol of negative RNA were also included. Cells were harvested at 48 h after incubation; both viable cells and dead cells were counted. Cell viability was determined by measuring lactate dehydrogenase (LDH) levels in supernatants at a 48-h time point after MRP1 siRNA transfection. A total of 1 ml of culture medium was collected from each group and centrifuged at 600 g for 10 min. A total of 10 µl of supernatants were transferred to an opaque-walled 96-well plate in triplicates. The LDH-cytotoxicity assay reagents (Abcam, UK; ab65393) were added and incubated for 30 min at RT. The reaction was stopped by adding 10 µl of stop solution and absorbance was measured at 490 nm in a UV-visible spectroscopy system (Agilent 8453, CA, USA). Wells where cells had been lysed to generate maximum LDH release were used as positive controls. Wells containing cell culture medium only were used to generate background fluorescence values. LDH release percentage was then calculated using the formula given below: LDH release (%) =
(test sample LDH release - culture medium background) # 100 (maximum LDH release - culture medium background) (Equation 1)
www.futuremedicine.com
2311
Research Article Wan, Apostolou, Dronov, Kuss & Voelcker Reverse transcription PCR for analysis of MRP1 expression Total RNA was extracted from the harvested cells with Trizol reagent (Invitrogen) according to the manufacturer’s protocol and quantified by Nanodrop. cDNA was synthesized with M-MLV reverse transcriptase kit (Promega, WI, USA) according to the manufacturer’s instructions. Three sets of primers for MRP1 (exon 7&8, exon 8&9 and siMohb) and GAPDH control are shown in Table 1. Primers were designed using the Invitrogen Oligoperfect designer web tool (http:// tools.invitrogen.com), and primer sequences were screened using a BLAST search to confirm specificity. PCR amplification was performed at 94°C for 3 min, then 45 cycles at 94°C for 1 min, 53°C for 1 min and 72°C for 1 min. Following amplification, the reactions were subjected to a thermal melt from 55 to 95°C in 0.5°C increments, and FAM fluorescence was monitored at each increment. The expression level of products corresponding to MRP1 primers exon 7 and 8, exon 8 and 9 and siMohb in each of the transfected groups was compared with the respective expression level in the untransfected group (defined as 100% expression) using the standard curve method. MRP1 protein expression analysis by flow cytometry
A total of 48 h post-transfection, 3 × 105 harvested cells were fixed by BD Cytofix/Cytoperm solution (BD PharMingen) for 30 min, washed with 1× BD Perm/Wash buffer twice, and incubated with 2 µg of MRP1 (QCRL-1) primary antibody (Santa Cruz Biotechnology) for 90 min at RT in the dark. After washing twice with 1× BD Perm/Wash buffer to remove excess primary antibody, cells were incubated with 50 µl of 1/50 sheep anti-mouse Ig, R-Phyceorythrin
(PE) conjugated affinity purified F(ab’) secondary antibody for 30 min at RT, followed by washing with 1× PBS containing 1% fetal calf serum (FCS) and 0.2% sodium azide. Finally, cells were resuspended in 250 µl 1× PBS and analyzed by Accuri-C6 flow cytometry (BD, Accuri Cytometers, MI, USA). Cell apoptosis assay
5 × 105 harvested cells in 100 µl 1× binding buffer (10 mM Hepes/NaOH pH7.4, 140 mM NaCl2, and 2.5 mM CaCl2) were incubated with 10 µl propidium iodide (PI; 2.5 mg/ml) and 5 µl of fluorescein- conjugated Annexin V (Annexin V-FITC; 1 mg/ml) for 30 min at RT in the dark. Each sample was analyzed by Accuri-C6 flow cytometry. Results & discussion Characterization of the pSiNPs
pSiNPs were fabricated by electrochemical etching of single-crystal silicon wafers; electropolishing to lift up the resulting porous membrane; fracturing the membrane into particles by ultrasonication and finally filtration of the particles through a 0.2-µm membrane. The diameter of pores in the generated nanoparticles ranged from 13 to 42 nm, with an average size of 26.4 nm. Such pore size may provide accessibility for a variety of proteins or enzymes; however, compared with direct attachment of siRNA onto the pSiNPs surface siRNA loaded into porous materials can protect siRNA from degradation by nucleases to a certain extent [40] . pSiNPs had dimensions between 50 and 220 nm with a mean size of 106.4 nm (Figure 1), and a BET surface of 490.6 m2g-1. The size of nanoparticles for drug delivery should not exceed 400 nm, since the interendothelial gap defense in tumors can only allow extravasation of nanoparticles up to 400 nm [42] . On the other hand,
Table 1. Bioimaging techniques based on gold nanoparticles. Primer
Sequence
Exon number
Exon 7&8 (not including siRNA region exon 8) Forward
5´-GCA GTG ACC TCT GGT CCT TAA-3´
7
Reverse
5´-CTC CTC ATT CGC ATC CAC CT-3´
8
Exon 8&9 (not including siRNA region exon 8) Forward
5´-CTA CTT CCT CAT GAG CTT CTT C-3´
8
Reverse
5´-AGC AGA TGT GGA AGT ACT GGT-3´
9
siMohb (including siRNA region exon 8) Forward
5´-AAG GAA TGC GCC AAG ACT AG-3´
7
Reverse
5´-CCT TAA ACA GAG AGG GGT TC-3´
8
Forward
5´-GAA GGT GAA GGT CGG AGT C-3´
NA
Reverse
5´-GAA GAT GGT GAT GGG ATT TC-3´
NA
GAPDH
2312
Nanomedicine (Lond.) (2014) 9(15)
future science group
Cancer-targeting siRNA d elivery from porous silicon nanoparticles
nanoparticles will experience rapid renal clearance if they are smaller than 10 nm in diameter [43] . Therefore, pSiNPs prepared in this study are in principle suitable, in terms of size, for drug delivery applications, exploiting the enhanced permeability and retention (EPR) effect and avoiding renal clearance. Shape is also a fundamental property of nanoparticles that is critical for their biological functions. For example, cylindrical- shaped filamentous micelles can effectively evade nonspecific uptake by the reticulo-endothelial system [44] . In our case, after 16 h soniciation the pSiNPs have flake-like (or discoidal) and rod-like shapes, rather than the spherical shape common to mesoporous s ilica. Interestingly, rod-shaped nano particles have shown higher cell binding rates due to the tumbling motion and larger adhesion area once upon contact with cells compared with spherical ones [45] . In addition, discoidal nanoparticles can accumulate in most organs but liver [46] . Therefore, the shape of our pSiNPs may be beneficial to drug delivery.
Research Article
400 nm
siRNA loading & evidence for siRNA within channels of pSi film
It has recently been reported that a high volume percentage of ethanol in DI water can significantly increase siRNA loading into mesoporous silica nanoparticles due to lower polarity of ethanol and pore wetting [41] . Here, ethanol/DI water mixtures (60/40 v/v) were found to be optimal for siRNA loading. pSiNPs were oxidized and then functionalized with APTES in order to generate a positively charged surface and facilitate adsorption of polyanionic oligonucleotides. We compared different oxidation and APTES functionalization procedures in terms of the ability to load siRNA target ing the MRP1 protein. siRNA loading was assessed using UV absorbance measurements of the supernatant before and after incubation with pSiNPs (Figure 2) . Less than 1 µg siRNA was loaded per mg of oxidized pSiNPs (Figure 2A) ; prolonged incubation time slightly improved the loaded amount (Figure 2B) . APTES functionalization of pSiNPs significantly increased the amount of loaded siRNA, and was best applied in conjunction with extended siRNA incubation times (Figure 2C–F) . Traditionally postmodification dry baking can cause pSiNP aggregation, and dispersion of aggregated pSiNPs via sonication is difficult; therefore, increasing reaction temperature during modification can avoid dry baking and also stabilize APTES grafting. A heating (80°C) and cooling cycle of nanoparticles in 2% APTES solution prior to siRNA incubation further improved loading, presumably due to stabilization of the APTES layer on the surface (Figure 2D & E) [47] . The use of microwave heating was employed to rapidly oxidize the surface and produce silanol groups [48,49] . A
future science group
200 nm
200 nm
Figure 1. Morphology and size distribution of pSiNPs. (A) Scanning electron microscopy (SEM) photomicrograph of several pSiNPs; (B) SEM photomicrograph of a single pSiNP. 10 µl of diluted pSiNPs suspension in ethanol was added onto a polished graphite substrate followed by SEM characterization with an accelerating voltage of 20 kV. (C) Transmission electron microscopy (TEM) photomicrograph of a single pSiNP. 10 µl of diluted pSiNPs suspension in ethanol was added onto a standard TEM grid followed by TEM characterization operated at 200 kV.
short (3-min) microwave treatment was applied to avoid nanoparticle degradation at longer treatment times [48] . This was followed by APTES modification at 80°C for 1 h. This treatment produced the best loading of 7.7 µg (Figure 2F) . The zeta-potential of APTES- modified
www.futuremedicine.com
2313
Research Article Wan, Apostolou, Dronov, Kuss & Voelcker with APTES following microwave-loaded oxidation gave the best loading results, this procedure was hence used for further experiments.
Loaded siRNA (µg)
9
Cellular uptake of siRNA-loaded pSiNPs
6
3
0 Different surface treatment Figure 2. siRNA (µg) loaded into 1 mg pSiNPs (n = 4). (A) siRNA was incubated with bare pSiNPs at RT for 1 h; (B) siRNA was incubated with bare pSiNPs overnight; (C) siRNA was incubated with APTES-modified (at RT for 1 h) pSiNPs at RT for 1 h; (D) siRNA was incubated with APTES-modified (at RT for 1 h) pSiNPs overnight; (E) siRNA was incubated with APTES-modified (at 80°C for 1 h) pSiNPs overnight; (F) siRNA was incubated with microwave-treated (3 min) and APTES-modified (at 80°C for 1 h) pSiNPs overnight.
pSiNPs was 12.6 ± 1.7 mV; the positive potential have resulted from protonation of amino groups on the surface of APTES-modified pSiNPs. After siRNA loading, the zeta-potential of pSiNPs dropped to -4.5 ± 1.2 mV. We also observed that the addition of guanidine hydrochloride (a chaotropic salt used to enhance packaging of oligonucleotides into small pores) had no significant effect on siRNA loading into pSiNPs (data not shown). Guanidine hydrochloride weakens electrostatic repulsion between siRNA molecules has a dehydrating effect on siRNA and nanoparticles, and drives siRNA to enter into pores by protonation of amines in its acidic environment. It did show a significant effect on packaging siRNA into small pores (~3.7 nm) [41] . However, in our case pore diameters were approximately seven-times larger. ToF-SIMS confirmed that oligonucleotides in ethanol/DI water mixture penetrated into the pSi pores. Here, DNA oligonucleotides were used. In particular, we were able to detect thymidine fragments of C4H6NO +, in positive mode (Figure 3A) . The amino functionalization played an important role in the ability to adsorb oligonucleotides into the pores since, for the pSi samples without APTES modification, oligonucleotide loading into the pores was low (Figure 3B) . Similar results found in C5H8N2O2+, C4H3N2O2-, and C5H5N2O2- analysis in positive and negative mode, respectively, also confirmed the difference in oligo nucleotide loading between the APTES-modified and the unmodified pSi samples. Since pSiNP modification
2314
Nanomedicine (Lond.) (2014) 9(15)
Approximately 1 mg of pSiNPs loaded with the equivalent of 200 pmol FAM-labelled siRNA were added to 1 × 105 T98G glioblastoma cells to study internalization of the siRNA. After 30 min of incubation, cells began to show a slight but detectable fluorescence emission. Figure 4A shows that after 18 h incubation, 65% cells gave strong green fluorescence. By contrast, when the same amount of siRNA was added to cell culture medium in the absence of pSiNPs or in addition to unloaded pSiNPs, no fluorescence or only weak spontaneous fluorescence was detected (Figure 4B). It should be noted that the pSiNPs themselves were nonfluorescent. The poor transfection efficiency of bare siRNA was attributed to poor cross-membrane translocation. By contrast, FAM-labelled siRNA-loaded pSiNPs were taken up into the cytoplasm via clathrin- and/or caveoli-mediated endocytosis and/or pinocytosis [50] . T98G cells transfected with 300, 200 and 100 pmol MPR1 siRNA-loaded pSiNPs showed 44.3, 22.1 and 11.5% reductions in cell number, respectively, compared with untreated cells at the same time point (Figure 4C). MRP1 siRNA-loaded pSiNP (300 or 200 pmol) treated groups showed a significant reduction (p
Cancer-targeting siRNA delivery from porous silicon nanoparticles
Aims: Porous silicon nanoparticles (pSiNPs) with tunable pore size are biocompatible and biodegradable, suggesting that they are suitable biomaterials as vehicles for drug delivery. Loading of small interfering RNA (siRNA) into the pores of pSiNPs can protect siRNA from degradation as well as improve the cellular uptake. We aimed to deliver MRP1 siRNA loaded into pSiNPs to glioblastoma cells, and to demonstrate downregulation of MRP1 at the mRNA and protein levels. Methods: 50–220 nm pSiNPs with an average pore size of 26 nm were prepared, followed by electrostatic adsorption of siRNA into pores. Oligonucleotide loading and release profiles were investigated; MRP1 mRNA and protein expression, cell viability and cell apoptosis were studied. Results: Approximately 7.7 µg of siRNA was loaded per mg of pSiNPs. Cells readily took up nanoparticles after 30 min incubation. siRNA-loaded pSiNPs were able to effectively downregulate target mRNA (~40%) and protein expression (31%), and induced cell apoptosis and necrosis (33%). Conclusion: siRNA loaded pSiNPs downregulated mRNA and protein expression and induced cell death. This novel siRNA delivery system may pave the way towards developing more effective tumor therapies.
Yuan Wan1,2, Sinoula Apostolou3, Roman Dronov2, Bryone Kuss3 & Nicolas H Voelcker*,1 Mawson Institute, University of South Australia, Mawson Lakes, Adelaide, SA 5095, Australia 2 School of Chemical and Physical Sciences, Flinders University, Bedford Park, Adelaide, SA 5042, Australia 3 School of Medicine, Flinders University, Bedford Park, Adelaide, SA 5042, Australia *Author for correspondence: nico.voelcker@ unisa.edu.au 1
Original submitted 14 August 2013; Revised submitted 4 December 2013 Keywords: drug delivery • glioblastoma • MRP1 • porous silicon nanoparticles • siRNA • ToF-SIMS
dsRNA triggers sequence-specific gene silencing in eukaryotic cells by a process known as RNA interference (RNAi) [1] . Exogenous synthetic small interfering RNA (siRNA), which usually consists of 21–23 nucleotides, introduced into the cytoplasm binds to the RNA-induced silencing complex (RISC). This binding event induces unwinding and separation of strands, and is followed by association with target mRNA. The tagged mRNA is then degraded by Slicer (Argonaute-2), an enzyme residing within the RISC complex [2] , resulting in downregulation of cellular protein production [3] . This process is highly specific and has the potential to silence any gene of interest [4] . siRNA delivery is therefore a promising gene therapy approach with relevance to a range of diseases including cancer [5] . MRP1, located at the plasma or intercellular membranes, is a member of ATP-dependent
10.2217/NNM.14.12 © 2014 Future Medicine Ltd
transporter protein [6] . It plays a role in cellular defense and protects healthy cells from toxic or harmful agents [7] . However, in various tumor types including human glioblastoma, breast cancer, and so on, the overexpression of MRP1, which acts as a membrane efflux pump, can increase transport of drugs out of the cancer cells [8,9] . Moreover, MRP1 protein levels correlate with previous exposure to chemotherapeutic drugs and with the tumor grade [10] . High MRP1 activity is a predictor of poor response to chemotherapy and a reduced overall survival of patients [11] . Various strategies including targeting membrane transporters [12] , inhibition of cell survival pathways [13] , altering transcription factors [14] and silencing antiapoptotic pathways have been developed to suppress multidrug resistance [15] . A promising sensitization strategy is the use of siRNA to silence MRP1 gene expression. On one hand,
Nanomedicine (Lond.) (2014) 9(15), 2309–2321
part of
ISSN 1743-5889
2309
Research Article Wan, Apostolou, Dronov, Kuss & Voelcker effective blocking of MRP1 protein expression can enhance chemosensitivity [16] ; on the other hand, siRNA can cause significant apoptosis in neuroblastoma due to the unique effect of MRP1 on tumor biology [17,18] . Although siRNA has shown significant potential as a new gene therapy, siRNA-based therapies face a number of key challenges that need to be overcome [19] . The polyanionic nature of bare siRNA results in poor binding to and transfer across the negatively charged cell membranes [20] . In addition, exogenous siRNA undergoes degradation upon exposure to nucleases [21] . Furthermore, side effects due to off-target immune responses can result from exposure of siRNA in the circulation to immune cells [22] . Viral vectors, liposomes, cationic polymers, recombinant proteins, electroporation and hydrodynamic gene transfer have been used to facilitate delivery of siRNA to target cells [23] . However, due to issues around cytotoxicity, high cost, low efficiency, and time-consuming and labor-intensive processing, these vehicles and methods have not been introduced into clinical practice. Porous silicon nanoparticles (pSiNPs) are a new addition to the nanoparticle-based drug delivery vehicle field [24,25] . These nanoparticles are unique in that they combine biocompatibility, biodegradability and high payload capacity [26–29] . The porous silicon (pSi) base is produced by a simple one-step electrochemical fabrication procedure. It is totally inorganic and can be readily sterilized. Pore dimensions can be controlled, from micropores (50 nm) [30] . It has a high surface area (400–800 m2/g) [30–32] . The material completely degrades in aqueous solutions into non-toxic silicic acid, the major bioavailable form of silicon [30] . pSiNPs with sizes in the range of 100–300 nm avoid renal filtration, leading to a prolonged residence time in the bloodstream that enables more effective targeting of diseases tissues, but these particles also effectively penetrate narrow tight epithelial junctions in the blood–brain barrier and deliver therapeutic agents [18] . pSiNPs have been deployed for bioimaging [25,26] , drug delivery [33–35] , and photodynamic [36] and photothermal therapy [37] . Using pSi microparticles, the Ferrari group has shown effective siRNA delivery. We have recently reported that pSiNPs are promising vehicles for targeted delivery of hydrophobic chemotherapy drugs [38] , and we have demonstrated siRNA loading onto pSiNPs with small pore sizes [39] . However, in our previous studies, the inner space of pores was not available for siRNA loading due to the small pore size, thereby limiting the loading capacity. Residual siRNA on the nanoparticle surface is subject to enzyme-mediated degradation in vivo, and therefore loading of siRNA into the pores is preferred [40] . Loading of siRNA into mesoporous silica nano-
2310
Nanomedicine (Lond.) (2014) 9(15)
particles has also been reported [21,41] . However, these nanoparticles are not biodegradable, and the tuning of pore size for siRNA delivery is limited. In this study, we report loading of MRP1 into pSiNPs that were prepared by electrochemical etching and grafted with amino groups on the inner and outer surfaces. These MRP1 siRNA-loaded nanoparticles showed uptake by T98G glioblastoma cells after 30 min incubation. MRP1 mRNA and protein were downregulated to approximately 40 and 30%, respectively, and 27% of cells underwent apoptosis. These results demonstrate that pSiNPs hold potential as a siRNA delivery vehicle for cancer therapy. Experiments & materials Cell lines & culture
The T98G glioblastoma line was obtained from Flinders Medical Center and cultured according to instructions. In brief, T98G cells were cultured in 25 ml of RPMI media 1640 (Invitrogen, CA, USA) containing 10% fetal bovine serum (FBS), 10 µl of 200 mM l-glutamine, 100 units/ml of penicillin and 100 µg/ml of streptomycin at 37°C in a humidified 5% carbon d ioxide atmosphere. Small interfering RNAs
ABCC1 siRNA duplexes (sense: 5´-GAG GCU UUG AUC GUC AAG UTT-3´; antisense: 5´-ACU UGA CGA UCA AAG CCU CTT-3´) and negative control siRNA duplexes (sense: 5´-UUC UCC GAA CGU GUC ACG UTT-3´; antisense: 5´-ACG UGA CAC GUU CGG AGA ATT-3´) were synthesized commercially by GenePharma (Shanghai, China). siRNA was resuspended with DEPC-treated water (Invitrogen). pSiNP preparation & characterization
pSiNPs were prepared by electrochemical etching of highly doped, (100)-oriented, p-type silicon wafers (Siltronix, France) in a 3:1 solution of 48% aqueous hydrofluoric acid (HF)/ethanol at a constant current density of 200 mA/cm2 for 150s; the resulting porous layer was then detached by electropolishing in a 3.3% HF in ethanol solution for 250s at a current density of 4 mA/cm2. The porous film was ultrasonicated in ethanol over 16 h; the supernatant was then filtered through a 0.2-µm nylon filter membrane (Carl Roth, Germany); the filtered solution was centrifuged at 22,000 g for 30 min, and the pellet of pSiNPs at the bottom was then resuspended in absolute ethanol. Diluted pSiNPs solution was dropped onto a graphite substrate, allowed to dry at room temperature (RT) and examined by scanning electron microscopy (SEM) with a Philips XL30 (Philips Co., OR, USA) and TEM with a FEI Tecnai G2 (FEI Co., OR, USA). Twenty pores were randomly selected from etched porous layer
future science group
Cancer-targeting siRNA d elivery from porous silicon nanoparticles
before electropolishing for measurement of pore size with SEM. The nanoparticle size distribution was measured using a Malvern Zetasizer Nano (ATA Co., Australia). The surface area of pSiNPs was determined from the Brunnauer–Emmett–Teller (BET) theory. Preparation of amine-functionalized pSiNPs
Approximately 3 mg of pSiNPs were functionalized with 2% (3-aminopropyl)triethoxysilane (APTES) in toluene for 1 h at RT or at 80°C, followed by rinsing with toluene and ethanol three times, respectively, to remove excess APTES and toluene. An alternative functionalization method involved pretreatment of pSiNPs in ethanol with microwave at 600 Watts for 3 min; the silane functionalization and rinse procedures followed the same protocol described above. Time-of-flight secondary ion mass spectrometry analysis
pSi membranes (diameter: 15 mm; thickness: 80 µm) were prepared and functionalized with APTES descr ibed above. A total of 40 µl of 25 µM DNA oligomer (5´TTT AGT TGA TTT GTG CTT CAG TGT GCT3´) mixed with 60 µl of ethanol were added onto the pSi membrane overnight at 4°C. The membrane was rinsed with DI water three times to remove unabsorbed DNA oligonucleotide and ethanol, then dried and broken in half. Positive and negative secondary ion images and spectra of the cross-section were acquired using timeof-flight secondary ion mass spectrometry (ToF-SIMS; Phi, MN, USA) with a pulse 25 keV, 1.3 pA primary ion beam in high current bunched mode from 50 × 50 µm areas on the sample top area. The collected data were analyzed with WinCadenceN software. Loading & release of siRNA
In total, 40 µl of 25 µM siRNA duplexes were mixed with approximately 1 mg of amine-functionalized pSiNPs in 500 µl of ethanol. siRNA was allowed to be absorbed onto the surface and pores of pSiNPs by shaking the mixture for 1 h or overnight at RT or at 4°C. Before and after siRNA loading, the zeta potential of the nanoparticles was determined using a Malvern Zetasizer Nano (ATA Co., Australia). The measurement were made at room temperature, and samples were diluted with DI water to an appropriate concentration to yield count rate per second (KCps) in the range of 2500–3500 KCps. Free siRNA was separated from adsorbed siRNA by centrifugation using a membrane centrifugal filter unit (Amicon Ultra-0.5 ml, Millipore, MA, USA). The concentration of free siRNA was determined by measuring the UV absorbance at 260 nm using a spectrophotometer (Nanodrop 1000, Thermo Fisher Science, MA, USA). The loaded siRNA
future science group
Research Article
amount equals the total siRNA loading amount prior to loading minus the amount of siRNA left in the supernatant after incubation. siRNA-loaded pSiNPs were resuspended in DEPC-water for siRNA release studies in vitro. At 0.5, 1, 2, 3 and 4 h time points pSiNPs were centrifuged, and siRNA concentration in supernatant was measured using the Nanodrop. siRNA uptake into cells
T98G glioblastoma cells were plated in six-well cell culture plates in 2.5 ml of RPMI antibiotic free media at a concentration of 1 × 105 cell/well (30–40% confluence) for 24 h prior to exposure to siRNA-loaded pSiNPs. The medium was removed and approximately 200 µl of mixture of culture medium containing siRNA-loaded pSiNPs were added. After 30 min, additional 2 ml fresh culture medium was added to maintain cell proliferation. First, cells were transiently transfected with 200 pmol FAM-labelled siRNA-loaded pSiNPs or with FAM-labelled siRNA alone. After 18 h incubation, cells were rinsed with 1× PBS thrice to completely remove remnant pSiNPs and/or FAM-labelled siRNA, and watched with the Nikon Eclipse (Nikon, Tokyo, Japan). MRP1 mRNA knockdown experiments were performed in triplicate with varying amounts of siRNA. The final amount of loaded siRNA corresponded to 100, 200 and 300 pmol. Three controls (cells without transfection, cell incubated with unloaded pSiNPs, and cells incubated with a mixture of 300 pmol of siRNA and bare pSiNPs) and a negative control comprising 300 pmol of negative RNA were also included. Cells were harvested at 48 h after incubation; both viable cells and dead cells were counted. Cell viability was determined by measuring lactate dehydrogenase (LDH) levels in supernatants at a 48-h time point after MRP1 siRNA transfection. A total of 1 ml of culture medium was collected from each group and centrifuged at 600 g for 10 min. A total of 10 µl of supernatants were transferred to an opaque-walled 96-well plate in triplicates. The LDH-cytotoxicity assay reagents (Abcam, UK; ab65393) were added and incubated for 30 min at RT. The reaction was stopped by adding 10 µl of stop solution and absorbance was measured at 490 nm in a UV-visible spectroscopy system (Agilent 8453, CA, USA). Wells where cells had been lysed to generate maximum LDH release were used as positive controls. Wells containing cell culture medium only were used to generate background fluorescence values. LDH release percentage was then calculated using the formula given below: LDH release (%) =
(test sample LDH release - culture medium background) # 100 (maximum LDH release - culture medium background) (Equation 1)
www.futuremedicine.com
2311
Research Article Wan, Apostolou, Dronov, Kuss & Voelcker Reverse transcription PCR for analysis of MRP1 expression Total RNA was extracted from the harvested cells with Trizol reagent (Invitrogen) according to the manufacturer’s protocol and quantified by Nanodrop. cDNA was synthesized with M-MLV reverse transcriptase kit (Promega, WI, USA) according to the manufacturer’s instructions. Three sets of primers for MRP1 (exon 7&8, exon 8&9 and siMohb) and GAPDH control are shown in Table 1. Primers were designed using the Invitrogen Oligoperfect designer web tool (http:// tools.invitrogen.com), and primer sequences were screened using a BLAST search to confirm specificity. PCR amplification was performed at 94°C for 3 min, then 45 cycles at 94°C for 1 min, 53°C for 1 min and 72°C for 1 min. Following amplification, the reactions were subjected to a thermal melt from 55 to 95°C in 0.5°C increments, and FAM fluorescence was monitored at each increment. The expression level of products corresponding to MRP1 primers exon 7 and 8, exon 8 and 9 and siMohb in each of the transfected groups was compared with the respective expression level in the untransfected group (defined as 100% expression) using the standard curve method. MRP1 protein expression analysis by flow cytometry
A total of 48 h post-transfection, 3 × 105 harvested cells were fixed by BD Cytofix/Cytoperm solution (BD PharMingen) for 30 min, washed with 1× BD Perm/Wash buffer twice, and incubated with 2 µg of MRP1 (QCRL-1) primary antibody (Santa Cruz Biotechnology) for 90 min at RT in the dark. After washing twice with 1× BD Perm/Wash buffer to remove excess primary antibody, cells were incubated with 50 µl of 1/50 sheep anti-mouse Ig, R-Phyceorythrin
(PE) conjugated affinity purified F(ab’) secondary antibody for 30 min at RT, followed by washing with 1× PBS containing 1% fetal calf serum (FCS) and 0.2% sodium azide. Finally, cells were resuspended in 250 µl 1× PBS and analyzed by Accuri-C6 flow cytometry (BD, Accuri Cytometers, MI, USA). Cell apoptosis assay
5 × 105 harvested cells in 100 µl 1× binding buffer (10 mM Hepes/NaOH pH7.4, 140 mM NaCl2, and 2.5 mM CaCl2) were incubated with 10 µl propidium iodide (PI; 2.5 mg/ml) and 5 µl of fluorescein- conjugated Annexin V (Annexin V-FITC; 1 mg/ml) for 30 min at RT in the dark. Each sample was analyzed by Accuri-C6 flow cytometry. Results & discussion Characterization of the pSiNPs
pSiNPs were fabricated by electrochemical etching of single-crystal silicon wafers; electropolishing to lift up the resulting porous membrane; fracturing the membrane into particles by ultrasonication and finally filtration of the particles through a 0.2-µm membrane. The diameter of pores in the generated nanoparticles ranged from 13 to 42 nm, with an average size of 26.4 nm. Such pore size may provide accessibility for a variety of proteins or enzymes; however, compared with direct attachment of siRNA onto the pSiNPs surface siRNA loaded into porous materials can protect siRNA from degradation by nucleases to a certain extent [40] . pSiNPs had dimensions between 50 and 220 nm with a mean size of 106.4 nm (Figure 1), and a BET surface of 490.6 m2g-1. The size of nanoparticles for drug delivery should not exceed 400 nm, since the interendothelial gap defense in tumors can only allow extravasation of nanoparticles up to 400 nm [42] . On the other hand,
Table 1. Bioimaging techniques based on gold nanoparticles. Primer
Sequence
Exon number
Exon 7&8 (not including siRNA region exon 8) Forward
5´-GCA GTG ACC TCT GGT CCT TAA-3´
7
Reverse
5´-CTC CTC ATT CGC ATC CAC CT-3´
8
Exon 8&9 (not including siRNA region exon 8) Forward
5´-CTA CTT CCT CAT GAG CTT CTT C-3´
8
Reverse
5´-AGC AGA TGT GGA AGT ACT GGT-3´
9
siMohb (including siRNA region exon 8) Forward
5´-AAG GAA TGC GCC AAG ACT AG-3´
7
Reverse
5´-CCT TAA ACA GAG AGG GGT TC-3´
8
Forward
5´-GAA GGT GAA GGT CGG AGT C-3´
NA
Reverse
5´-GAA GAT GGT GAT GGG ATT TC-3´
NA
GAPDH
2312
Nanomedicine (Lond.) (2014) 9(15)
future science group
Cancer-targeting siRNA d elivery from porous silicon nanoparticles
nanoparticles will experience rapid renal clearance if they are smaller than 10 nm in diameter [43] . Therefore, pSiNPs prepared in this study are in principle suitable, in terms of size, for drug delivery applications, exploiting the enhanced permeability and retention (EPR) effect and avoiding renal clearance. Shape is also a fundamental property of nanoparticles that is critical for their biological functions. For example, cylindrical- shaped filamentous micelles can effectively evade nonspecific uptake by the reticulo-endothelial system [44] . In our case, after 16 h soniciation the pSiNPs have flake-like (or discoidal) and rod-like shapes, rather than the spherical shape common to mesoporous s ilica. Interestingly, rod-shaped nano particles have shown higher cell binding rates due to the tumbling motion and larger adhesion area once upon contact with cells compared with spherical ones [45] . In addition, discoidal nanoparticles can accumulate in most organs but liver [46] . Therefore, the shape of our pSiNPs may be beneficial to drug delivery.
Research Article
400 nm
siRNA loading & evidence for siRNA within channels of pSi film
It has recently been reported that a high volume percentage of ethanol in DI water can significantly increase siRNA loading into mesoporous silica nanoparticles due to lower polarity of ethanol and pore wetting [41] . Here, ethanol/DI water mixtures (60/40 v/v) were found to be optimal for siRNA loading. pSiNPs were oxidized and then functionalized with APTES in order to generate a positively charged surface and facilitate adsorption of polyanionic oligonucleotides. We compared different oxidation and APTES functionalization procedures in terms of the ability to load siRNA target ing the MRP1 protein. siRNA loading was assessed using UV absorbance measurements of the supernatant before and after incubation with pSiNPs (Figure 2) . Less than 1 µg siRNA was loaded per mg of oxidized pSiNPs (Figure 2A) ; prolonged incubation time slightly improved the loaded amount (Figure 2B) . APTES functionalization of pSiNPs significantly increased the amount of loaded siRNA, and was best applied in conjunction with extended siRNA incubation times (Figure 2C–F) . Traditionally postmodification dry baking can cause pSiNP aggregation, and dispersion of aggregated pSiNPs via sonication is difficult; therefore, increasing reaction temperature during modification can avoid dry baking and also stabilize APTES grafting. A heating (80°C) and cooling cycle of nanoparticles in 2% APTES solution prior to siRNA incubation further improved loading, presumably due to stabilization of the APTES layer on the surface (Figure 2D & E) [47] . The use of microwave heating was employed to rapidly oxidize the surface and produce silanol groups [48,49] . A
future science group
200 nm
200 nm
Figure 1. Morphology and size distribution of pSiNPs. (A) Scanning electron microscopy (SEM) photomicrograph of several pSiNPs; (B) SEM photomicrograph of a single pSiNP. 10 µl of diluted pSiNPs suspension in ethanol was added onto a polished graphite substrate followed by SEM characterization with an accelerating voltage of 20 kV. (C) Transmission electron microscopy (TEM) photomicrograph of a single pSiNP. 10 µl of diluted pSiNPs suspension in ethanol was added onto a standard TEM grid followed by TEM characterization operated at 200 kV.
short (3-min) microwave treatment was applied to avoid nanoparticle degradation at longer treatment times [48] . This was followed by APTES modification at 80°C for 1 h. This treatment produced the best loading of 7.7 µg (Figure 2F) . The zeta-potential of APTES- modified
www.futuremedicine.com
2313
Research Article Wan, Apostolou, Dronov, Kuss & Voelcker with APTES following microwave-loaded oxidation gave the best loading results, this procedure was hence used for further experiments.
Loaded siRNA (µg)
9
Cellular uptake of siRNA-loaded pSiNPs
6
3
0 Different surface treatment Figure 2. siRNA (µg) loaded into 1 mg pSiNPs (n = 4). (A) siRNA was incubated with bare pSiNPs at RT for 1 h; (B) siRNA was incubated with bare pSiNPs overnight; (C) siRNA was incubated with APTES-modified (at RT for 1 h) pSiNPs at RT for 1 h; (D) siRNA was incubated with APTES-modified (at RT for 1 h) pSiNPs overnight; (E) siRNA was incubated with APTES-modified (at 80°C for 1 h) pSiNPs overnight; (F) siRNA was incubated with microwave-treated (3 min) and APTES-modified (at 80°C for 1 h) pSiNPs overnight.
pSiNPs was 12.6 ± 1.7 mV; the positive potential have resulted from protonation of amino groups on the surface of APTES-modified pSiNPs. After siRNA loading, the zeta-potential of pSiNPs dropped to -4.5 ± 1.2 mV. We also observed that the addition of guanidine hydrochloride (a chaotropic salt used to enhance packaging of oligonucleotides into small pores) had no significant effect on siRNA loading into pSiNPs (data not shown). Guanidine hydrochloride weakens electrostatic repulsion between siRNA molecules has a dehydrating effect on siRNA and nanoparticles, and drives siRNA to enter into pores by protonation of amines in its acidic environment. It did show a significant effect on packaging siRNA into small pores (~3.7 nm) [41] . However, in our case pore diameters were approximately seven-times larger. ToF-SIMS confirmed that oligonucleotides in ethanol/DI water mixture penetrated into the pSi pores. Here, DNA oligonucleotides were used. In particular, we were able to detect thymidine fragments of C4H6NO +, in positive mode (Figure 3A) . The amino functionalization played an important role in the ability to adsorb oligonucleotides into the pores since, for the pSi samples without APTES modification, oligonucleotide loading into the pores was low (Figure 3B) . Similar results found in C5H8N2O2+, C4H3N2O2-, and C5H5N2O2- analysis in positive and negative mode, respectively, also confirmed the difference in oligo nucleotide loading between the APTES-modified and the unmodified pSi samples. Since pSiNP modification
2314
Nanomedicine (Lond.) (2014) 9(15)
Approximately 1 mg of pSiNPs loaded with the equivalent of 200 pmol FAM-labelled siRNA were added to 1 × 105 T98G glioblastoma cells to study internalization of the siRNA. After 30 min of incubation, cells began to show a slight but detectable fluorescence emission. Figure 4A shows that after 18 h incubation, 65% cells gave strong green fluorescence. By contrast, when the same amount of siRNA was added to cell culture medium in the absence of pSiNPs or in addition to unloaded pSiNPs, no fluorescence or only weak spontaneous fluorescence was detected (Figure 4B). It should be noted that the pSiNPs themselves were nonfluorescent. The poor transfection efficiency of bare siRNA was attributed to poor cross-membrane translocation. By contrast, FAM-labelled siRNA-loaded pSiNPs were taken up into the cytoplasm via clathrin- and/or caveoli-mediated endocytosis and/or pinocytosis [50] . T98G cells transfected with 300, 200 and 100 pmol MPR1 siRNA-loaded pSiNPs showed 44.3, 22.1 and 11.5% reductions in cell number, respectively, compared with untreated cells at the same time point (Figure 4C). MRP1 siRNA-loaded pSiNP (300 or 200 pmol) treated groups showed a significant reduction (p
