Design of a Multicomponent Peptide-Woven Nanocomplex for Delivery of siRNA

We developed and tested a multicomponent peptide-woven siRNA nanocomplex (PwSN) comprising different peptides designed for efficient cellular targeting, endosomal escape, and release of siRNA. To enhance tumor-specific cellular uptake, we connected an interleukin-4 receptor-targeting peptide (I4R) to a nine-arginine peptide (9r), yielding I4R-9r. To facilitate endosomal escape, we blended endosomolytic peptides into the I4R-9r to form a multicomponent nanocomplex. Lastly, we modified 9r peptides by varying the number and positions of positive charges to obtain efficient release of siRNA from the nanocomplex in the cytosol. Using this step-wise approach for overcoming the biological challenges of siRNA delivery, we obtained an optimized PwSN with significant biological activity in vitro and in vivo. Interestingly, surface plasmon resonance analyses and three-dimensional peptide models demonstrated that our designed peptide adopted a unique structure that was correlated with faster complex disassembly and a better gene-silencing effect. These studies further elucidate the siRNA nanocomplex delivery pathway and demonstrate the applicability of our stepwise strategy to the design of siRNA carriers capable of overcoming multiple challenges and achieving efficient delivery.


Introduction
RNA interference (RNAi) is a natural biological system by which RNA molecules silence gene expression by inhibiting protein translation. RNAi using synthetic, small interfering RNAs (siRNAs) has powerful and broad potential as a therapeutic strategy for suppressing target proteins through sequence-specific degradation of their corresponding mRNAs. Research on siRNA delivery has made great progress in recent decades, and at least 22 siRNA-based strategies for treating disease have reached clinical trials [1][2][3][4].

Gel retardation assay and RNase stability
Peptide-woven siRNA nanocomplexes (PwSN) were prepared by mixing siRNA in diethylpyrocarbonate (DEPC)-treated water (100 pmoles) with peptide carrier with a molar ratio of 1:5 to 1:50 (siRNA: peptide) for 30 min at room temperature. The samples were subjected on 2% agarose gels electrophoresis to examine siRNA encapsulation. To test the stability of the siRNA nanocomplex against RNase A, the nanocomplex containing 100 pmol of siRNA mixed with 20-fold molar excess of peptides, then was incubated in the presence or absence of RNase A (0.2 μg) for up to 6 h at 37°C. After incubation, the RNase A activity was inhibited by addition of 2% sodium dodecyl sulfate (SDS) and subjected to agarose gel electrophoresis [25].

Measurement of shape, size, and surface charge
Peptide-woven siRNA nanocomplexes were prepared as described above with 1:20 molar ratio (siRNA:peptide). The hydrodynamic radii and zeta potential of nanocomplexes were determined using the zeta-potential and dynamic light scattering (DLS) instrument (ELS-Z, Otzuka Electronics, Japan). The morphology of nanocomplex was observed by transmission electron microscopy (TEM, CM30 Electron Microscope, Philips, CA). Nanoparticles deposited on the grid were negatively stained with 2 wt.% uranyl acetate solution [26].

Cell culture and MTT assay
HeLa cells were obtained from ATCC (Manassas, VA) and cultured in Dulbecco's modification of Eagle's medium (DMEM) with 10% fetal bovine serum, penicillin (100 IU/ml) and

In vitro Gene Silencing analysis
HeLa cells (1.5×10 5 ) were plated 24 h prior in 35 mm dishes and allowed to reach at least 70% in confluency. Nanocomplexes were added at 200 nM siRNA/ 2uM peptide (I4R-9r/siGAPDH and I4R-9r/siControl) over the cells for 4 h at 37°C in serum-free DMEM. Cells were then washed and incubated with fresh culture medium containing 10% FBS for an additional 48 h before being lysed and examined by western blot with specific antibodies [28]. As a positive control, transfection with Lipofectamine 2000 (Invitrogen, Carlsbad, CA) was performed in accordance with the manufacturer's instructions.

Endosomal Escape
HeLa cells were plated 48 h prior in eight chamber slides and allowed to reach at least 70% confluence. The cells were treated with the early endosome marker (CellLight Early Endosome-RFP, Molecular Probe) for 12 h before the transfection of I4R-9r/siRNA-FITC nanocomplex [29,30]. After 4 hours after the transfection, the cells were fixed with 4% paraformaldehyde and observed under the laser scanning confocal microscope to monitor the endosomal escape of the nanocomplexes.

In vivo gene silencing efficacy analysis
This study strictly followed the recommendations of National Institute of Health (NIH) for the Care and Use of Laboratory Animals. Animal experiments were reviewed and approved by the Committee on the Ethics of Animal Experiments of the Korea Institute of Science and Technology (KIST). All efforts were made for minimizing animal suffering. Male nude mice (BALB/ c nude; body weight, 20±2 g; n = 9) were housed in a specific pathogen-free environment at 22±2°C, 55±5% relative humidity with light. Prior to the in vivo imaging, tumor bearing mice were anesthetized under inhalational anesthesia (1%, w/v, isofurane in 2 L oxygen), then D-luciferin (150mg/kg) was then injected intraperitoneally. 3 mice per group was ued to minimize unnecessary sacrifice.
To monitor the feasibility of in vivo gene silencing efficacy of I4R-9r(A1Ã)/sHGP-9r-PwSN, HT29-luc human colorectal adenocarcinoma cells (1 X 10 7 cells) were injected subcutaneously in the left flank of mouse. After 3 weeks, mice received the I4R-9r(A1Ã)-PwSN (40μg of siRNA/mouse) via intra-tumoral injection every other day. Bio-luminescence signals of subcutaneous tumor were monitored using the IVIS Spectrum imaging system (Caliper Life Sciences, Hopkinton, MA) during the experiment period.

Surface plasmon resonance (SPR) analysis
The dynamics of PwSN formation were studied by SPR analysis (SR7500 DC, Reichert Inc, Depew, NY). 5 0 -Biotinylated siRNAs were diluted to 1 μM in HBS buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 3.4 mM EDTA, 0.005% Tween20), heated at 85°C for 10 min, cooled to room temperature, diluted to 20 fmol/μl in running buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 62.5 μg/ml bovine serum albumin, 125 μg/ml tRNA, 1 mM dithiothreitol, 0.05% Tween20), and injected at 10 μl/min. 50-100 response units of siRNA were coated per flow channel on streptavidin-coated sensor chip (NeutrAvidin Surface chip; Reichert Analytical Instrument, Depew, NY). 2-min injections of four different peptide concentrations (10, 5, 2.5, and 1.25 μ M) were injected over the sensor chip surface at 25°C with a flow rate of 25 μl/min. The surface was regenerated by removing bound peptides with a 180-s 2 M NaCl injection and 20-s 0.1% SDS followed by a binding buffer injection to rinse the needle. Sensorgrams were fit to kinetic titration model using Clamp program. To monitor dissociation between peptide and siRNA upon heparan sulfate competition, The 2.5 μM of peptides were injected to reach saturated binding and different concentrations of heparan sulfate were injected.

Statistical analysis
Data were presented as mean ± standard deviation. Statistical significances were determined using the ANOVA analysis and Student`s t-test for pairwise comparison. The significance was accepted for P values < 0.05.

Tumor-specific cellular uptake of I4R-9r PwSN
To develop a tumor-specific siRNA delivery system, we synthesized a tandem peptide carrier composed of an N-terminal tumor targeting I4R peptide (CRKRLDRNC) and C-terminal 9 D-arginine repeats (Fig. 1A). The I4R peptide homes to tumors through specific binding to the IL-4 receptor, which is expressed on the surface of tumor cells [20,25]. To determine the amount of peptide needed to fully encapsulate free siRNAs into nanocomplexes, we mixed siRNA with tandem peptides at increasing molar ratios (Fig. 1B). Gel retardation assays showed that, upon nanocomplex formation, free siRNA decreased and most of the siRNA was encapsulated by I4R-9r at siRNA:peptide molar ratios between 1:20 and 1:50. The biocompatibility of the siRNA nanocomplex was evaluated by testing cytotoxicity toward HeLa cells using MTT assay. No significant decrease in cellular viability was observed, indicating that I4R-9r PwSN was not toxic at the concentrations examined ( Figure A in S1 File). To gain insight into the ability of I4R-9r PwSN to deliver siRNA, we assessed intracellular uptake of siRNA, the first critical barrier in siRNA delivery, using confocal microscopy and flow cytometry ( Fig. 1C and D). The FITC-conjugated siRNA was encapsulated by I4R-9r tandem peptides, forming I4R-9r PwSN, and incubated with HeLa cells. Fluorescent siRNAs in the nanocomplex were observed in the cytoplasm of HeLa cells, whereas no signals of free siRNA were detected, indicating the I4R-9r tandem peptides efficiently delivered fluorescent labeled siRNA into the cytosol (Fig. 1C). A FACS analysis also showed specific cellular uptake of I4R-9r PwSN; this uptake appeared to be reinforced by specific interaction of the I4R peptides with IL-4 receptors since I4R-9r PwSN exhibited stronger binding to HeLa cells than 9r-PwSN (Fig. 1D). To determine whether the cellular uptake of I4R-9r PwSN was mediated by direct interaction with the IL4 receptor, we performed blocking experiments using a series of concentrations of an anti-IL-4R antibody (Fig. 1E). Preincubation with an anti-IL-4R antibody specifically diminished cellular uptake in a concentration-dependent manner compared to a control antibody, suggesting that, because of the presence of I4R, the tandem presentation of 9R leads to specific uptake by cells expressing IL-4 receptors.
We next investigated the in vitro gene-silencing activity of siRNAs delivered by I4R-9r PwSN. HeLa cells were treated with siRNA against human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) bound to the I4R-9r tandem peptide or lipofectamine, and then analyzed for GAPDH knock down by Western blot. As shown in Fig. 1F, I4R-9r PwSN carrying siGAPDH induced a slight, but statistically insignificant, decrease in GAPDH protein expression in HeLa cells, whereas siGAPDH delivered by lipofectamine dramatically decreased GAPDH expression. The insignificant silencing effect of I4R-9r PwSN, despite its specific and effective cellular targeting, prompted us to develop strategies for overcoming barriers after entry of the nanocomplex into cells.

sHGP-9r tandem peptide for endosomal escape
After cellular internalization of the nanocomplex via receptor-mediated endocytosis, the next critical step in the delivery process is escape from the early endosome [11,[36][37][38]. Fluorescent siRNAs delivered by I4R-9r PwSN were present in punctuate vesicular structures, consistent with sequestration in endosomes (Fig. 1C). To directly confirm the intracellular localization of siRNAs, we applied nanocomplexes to HeLa cells pretreated with the early endosome marker-RFP and monitored FITC-conjugated siRNA by confocal microscopy ( Fig. 2A). Fluorescently labeled siRNA colocalized with the early endosome marker, suggesting that I4R-9r PwSN carried siRNA payloads into the cells via endocytosis and was initially sequestered in endosomes.

Design and characterization of I4R-9r variant PwSNs
To achieve gene knockdown, the siRNA cargo must ultimately dissociate from the carrier after escaping from the endosome to allow the incorporation of siRNA into the RISC machinery. Thus, we focused on siRNA release from the nanocomplex in the cytoplasm. It has been reported that CPPs with long poly-arginine peptides (12R-15R) do not readily release siRNA, likely owing to the large number of positively charges, which retard the unpacking of siRNA [13]. In our case, the siRNA complex was formed mostly by electrostatic interactions between the positively charged 9r and negatively charged siRNA. We thus modified 9r by replacing selected arginines with alanines, which would be expected to alter electrostatic interactions. We designed and created four tandem peptide candidates-I4R-9r(A1), I4R-9r(A1Ã), I4R-9r(A2) and I4R-9r(A3) (Table 1)-that differed with respect to the number and position of alanine substitutions. We then formed I4R-9r-variant PwSNs and validated their biochemical characteristics as well as in vitro gene-silencing efficacy compared with that of I4R-9r PwSN. First, we tested the encapsulation efficiency of the I4R-9r tandem peptide variants using gel-retardation assays. As was the case with the I4R-9r peptide, all variants were able to encapsulate siRNA at molar ratios of 1:20 and 1:30. Complex formation was not affected by the presence or absence of the sHGP-9r peptide (Fig. 3A). The physicochemical characteristics of I4R-9r-variant PwSNs (1:20 siRNA-to-peptide molar ratios) were not significantly changed by the variation of 9r. Using TEM and dynamic light scattering (DLS), we observed that all I4R-9r variants assembled with siRNA to form PwSNs with a round structure around 200 nm in diameter, which is the typical size of a CPP-based siRNA nanocomplex ( Fig. 3B and Table 1). Zeta potentials ranged from approximately +15 to +20 mV. The presence of sHGP-9r in the PwSN slightly reduced size and surface charge in comparison with those of sHGP-9r-free PwSNs (I4R-9r). Moreover, siRNAs carried by I4R-9r-variant PwSNs were protected from RNases A for at least 6 h at 37°C, whereas free siRNA was completely degraded (Fig. 3C). One of the variant PwSNs, I4R-9r (A3), in which three arginines were substituted with alanines, showed 50% protection of its siRNA payload. FACS analysis of the cellular uptake of I4R-9r-variant PwSNs containing different concentrations of siRNA (31.25 to 250 nM) showed that, with the exception for I4R-9r(A3), more than 80% of siRNA was transferred into cells by nanocomplexes at siRNA concentrations greater than 100 nM ( Figure B in S1 File).
Prior to animal experiments, the gene-silencing activity of I4R-9r (A1Ã) PwSN against luciferase was tested in HT29-luc cells, which stably express luciferase (Figures C and D in S1 File). I4R-9r (A1Ã) PwSN was efficiently taken up by HT29-luc cells and exhibited gene-silencing activity by reducing luciferase proteins by~52%. We next developed a mouse xenograft model using these HT29-luc cells. The bioluminescence signals of the tumor were compared before and after intra-tumoral injection of the nanocomplex. As shown Fig. 4B and C, luminescent signals in tumors of control mice increased approximately 3-fold after 4 d, whereas those of the I4R-9r(A1Ã) PwSN-treated mice increased by only 1.25-fold, suggesting that the siRNA carried by I4R-9r(A1Ã) PwSN successfully silenced its specific target in vivo. The siRNA carried by the unmodified I4R-9r PwSN exhibited marginal activity, consistent with its insignificant effect in vitro.

Dynamic interactions and structures of I4R-9r-variant/siRNA nanocomplexes
One possible explanation for the improved efficiency of I4R-9r(A1Ã) PwSN despite the shared biochemical characteristics of I4R-9r variant PwSNs described above involves the rate of nanocomplex dissociation. To test this, we assessed the dissociation of siRNA from nanocomplexes by performing heparan sulfate competition assays using gel-retardation assays. The encapsulated siRNAs began to dissociate from their corresponding nanocomplex upon addition of more than a 5-fold excess of heparan sulfate; siRNA release was similar for all variants except I4R-9r (A3) PwSN, which showed increased release of siRNA (Fig. 5A). To further explore the dynamic interactions of each carrier with siRNA, we examined the kinetics of I4R-9r tandem peptide variant binding with siRNA using SPR analysis. In these experiments, I4R-9r peptide variants with sHGP-9r (19:1 molar ratios) were injected at increasing concentrations onto immobilized siRNA, and association and dissociation were measured (Fig. 5B). The I4R-9r tandem peptide showed the strongest binding affinity for siRNA, presumably because of its higher number of positive charges. The single alanine-substituted variants, I4R-9r(A1) and I4R-9r(A1Ã), showed moderate affinity, and the double and triple alanine-substituted variants, I4R-9r(A2) and I4R-9r(A3), respectively, showed the lowest affinity, as expected by their diminished overall positive charge. In the end, we selected the most efficient binding candidates, I4R-9r, I4R-9r(A1) and I4R-9r(A1Ã), and tested their dissociation from siRNA in the presence of excess heparan sulfate (Fig. 5C). Heparan sulfate induced a concentration-dependent dissociation of siRNA from I4R-9r-variant PwSNs, providing a possible indication of complex disassembly in the presence of anionic components in the cytosol. Interestingly, siRNA dissociated more rapidly from I4R-9r(A1Ã) PwSN than from the other variants, suggesting that the improved gene-silencing activity of I4R(A1Ã) PwSN is attributable to rapid dissociation of siRNA in the cytosol. Similar results were observed for nanocomplexes lacking sHGP-9r ( Figures E and F in S1 File). To understand why I4R-9r(A1Ã) PwSN was more efficient in unpacking siRNA than I4R-9r(A1) and other variants, we investigated the molecular structures of I4R-9r-variant tandem peptides using 3D modeling (Fig. 5D). The 3D structural models showed that the I4R-9r(A1Ã) tandem peptide adopted a "U"-shaped structure, whereas other tandem peptides adopted an "L"shaped structure. Thus the structural changes induced by the position of alanines could affect the efficiency of nanocomplex unpacking, which may in turn affect the availability of siRNA within the cytosol and contribute to RNAi-mediated gene silencing.

Discussion
To trigger RNAi activity, siRNA molecules must be delivered to the interior of specific cells and become loaded into RISCs in the cytosol. Although the mechanistic understanding of the delivery process is limited, there are general guidelines for optimal delivery systems. An optimized siRNA carrier is one that is able to assist specific uptake by target cells and overcome intracellular barriers posed by escape from endosomes or lysosomes and liberation of siRNA from nanocomplexes in the cytoplasm [10][11][12]. In this study, we started with a simple 9r-CPP to encapsulate siRNA, and then implemented a sequential trouble-shooting strategy to resolve problems encountered in the siRNA-delivery process, ultimately developing an optimized multicomponent PwSN with specific and effective gene-silencing ability (Fig. 6).
Arginine-or lysine-rich CPPs efficiently encapsulate siRNA through electrostatic interactions and form complexes of approximately 200 nm in diameter, which is an ideal nanoparticle size: large enough to avoid renal filtration, but small enough to evade phagocytic clearance [39]. The main drawback of CPP-based siRNA carriers is their lack of target specificity. To enhance the specificity of CPPs, we fused the tumor-homing peptide, I4R, to 9r-CPP [20]. Results obtained in other delivery systems have shown that endogenous targeting ligands are often beneficial for improving uptake by specific cells [25]. Efficient targeting may also diminish the chance of nonspecific interaction with serum proteins, including nucleases, as well as non-targeted cells.
After specific targeting and subsequent receptor-mediated endocytosis, I4R-9r PwSN was sequestered within endosomes. Generally, siRNA complexes in early endosomes are transferred to the late endosome and then to the lysosome, where various nucleases ultimately degrade the siRNA. Technologies for promoting endosomal release have recently been developed, including the use of fusogenic lipids, polymers with high buffering capacity, and membrane-interacting peptides [19,37,38,40]. To resolve the endosome entrapping problem, we blended a small amount of an endosomolytic tandem peptide (sHGP-9r) with I4R-9r. Addition of the sHGP-9r peptide facilitated endosomal escape of nanocomplexes in the cytosol, likely by imparting of the lipophilicity of the nanoparticle. The slight reduction in the size and charge of PwSN containing an HGP-9r tandem peptide can be explained by improved lipophilicity.
Achieving gene-silencing activity ultimately requires that the siRNA cargo be liberated from the delivery complex after escaping from the endosome to allow the incorporation of siRNA into the RISC machinery. Since the positively charged, arginine-rich CPP binds the negatively charged backbone of the siRNA, the efficacy of siRNA delivery can be enhanced and sustained by changing the number of arginines in the CPP [38][39][40]. In an effort to improve siRNA unpacking, we modified the number and positions of the positive charges in the 9r domain in the tandem and analyzed complex stability in the presence of heparan sulfate using SPR spectroscopy. Addition of heparan sulfate mimics encounters with polyanionic molecules in the cytosol. Among I4R-9r-variant PwSNs, I4R-9r (A1Ã) PwSN, which most efficiently liberated siRNA, exerted a significant silencing effect both in vitro and in vivo. Our results indicate that the peptide charges affect the efficiency of nanocomplex unpacking and in turn contribute to RNAi-mediated gene silencing. The less positively charged 9r variants, such as A2 and A3, exhibited poor binding and encapsulating capability compared with more highly charged 9r variants, likely because the fewer cationic charges available for binding the negatively charged siRNA reduces the strength of the interaction with siRNA and provides insufficient charge shielding for membrane translocation. However, strong binding affinity between the siRNA and its carrier is a double-edged sword: it is advantageous in that it confers a high encapsulating capacity and shields siRNA during transport to target cells, but it may also prevent the release of siRNA in the cytosol [1,11]. Accordingly, in designing the carrier, we must consider not only complex assembly in the extracellular space but also complex disassembly in the intracellular space [11]. The degree of complexation between siRNA and carrier is determined by multiple factors, including electrostatic interactions, van der Waals forces, and nonpolar salvation [11,41,42]. Our data suggest that SPR analyses can be used to reveal the dynamics of siRNA complexation with a resolution not possible using gel-retardation assays.
The optimized multicomponent nanocomplex, I4R-9r(A1Ã)/sHGP-9r PwSN, had a diameter of~150 nm, was stable against RNase A treatment, was specifically internalized into target cells, escaped from the endosome, and efficiently liberated RNA to initiate the silencing pathway. An interesting feature of this complex is the unique tandem peptide structure of I4R-9r (A1Ã) revealed by a computer structural model. The "U"-shape of the I4R-9r(A1Ã) peptide could affect dynamic interactions with siRNA, suggesting that differences in the 3D structure of this peptide carrier could be an important contributor to its rapid dissociation and thus enhanced biological activity. The conclusion is consistent with recent reports indicating that carrier structure is a critical parameter in determining nanocomplex formation, although the delivery materials were different in these previous studies [41,43,44].
A comprehensive understanding of the molecular properties of the siRNA nanocomplex is important in designing an effective siRNA delivery system. As our understanding of the molecular properties of nanocomplexes grows, new criteria for efficient delivery systems will likely emerge. Recent studies of CPP-siRNA complexes also indicate that a comprehensive understanding of molecular properties is essential for more accurately predicting the gene-silencing effects [45]. Our sequential trouble-shooting strategy and observations of the dynamic characteristics of complexes based on binding affinity and molecular structure could provide useful guidelines for designing future multicomponent peptide-delivery vehicles. Additional studies will be required to overcome remaining challenges, such pharmacokinetics and evasion of the immune system, to enable clinical applications of this multicomponent PwSN.
Supporting Information S1 File. Figure A. The Cytotoxicity of I4R-9r PwSN. The concentrations of siRNA ranged from 31.25 nM to 500 nM were treated onto the HeLa cells. After 48 hrs, the cell viability was evaluated with MTT assay. Figure B. Cellular uptake of I4R-9r variant PwSNs. Uptake by HeLa cells with different concentrations of siRNA (31.25 to 250 nM) was assessed by flow cytometry for I4R-9r variant PwSNs. Figure C. The specific cellular uptake of I4R-9r.siRNA nanocomplex in HT 29-luc cells. HT 29-luc cells were incubated with nanocomplexes carrying siRNA labeled with FITC. Uptake by HT 29-luc cells was measured by FACS for free siRNA (black line), 9r/siRNA (blue line) and I4R-9r/siRNA nanocomplex (red line). Figure D. The in vitro gene silencing efficiency of I4R-9r.siRNA nanocomplex in HT 29-luc cells. In vitro gene silencing efficacy was quantified by the Luciferase assay. Error bars represent SD from cumulative data of three independent experiments. Figure E. SPR analysis of I4R-9r variant and siRNA in the absence of sHGP-9r. The kinetics of association of siRNA with I4R-9r variant using SPR analysis. Increasing concentrations of I4R-9r variants were injected to associate with 5`-biotinylated siRNA on the streptavidin chip. Figure F. Dissociation of I4R-9r variant and siRNA in the absence of sHGP-9r. The kinetics of disassociation of siRNA with I4R-9r variant using SPR analysis. After association of 5`-biotinylated siRNA with I4R-9r variants (2.5 μM) different concentrations of heparan sulfate were injected to analysis the dissociation kinetics.