Biomaterials 35 (2014) 7929e7939

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Silk fibroin layer-by-layer microcapsules for localized gene delivery Linhao Li a, b, Sebastian Puhl a, Lorenz Meinel a, Oliver Germershaus a, * a

Institute for Pharmacy and Food Chemistry, University of Wuerzburg, Am Hubland, 97074 Wuerzburg, Germany Key Laboratory of Biorheological Science and Technology, Ministry of Education, College of Bioengineering, Chongqing University, 400030 Chongqing, PR China

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 April 2014 Accepted 21 May 2014 Available online 13 June 2014

Herein, we describe the delivery of plasmid DNA (pDNA) using silk fibroin (SF) layer-by-layer assembled microcapsules. Deposition of fluorescently labeled SF onto polystyrene (PS) template particles resulted in increasing fluorescence intensity and decreasing surface charge in correlation to SF layer number. After removal of the PS core, hollow, monodisperse, and structurally stable SF microcapsules of variable size and shell thickness were obtained. Plasmid DNA encoding for enhanced green fluorescent protein (eGFP) was loaded onto 1 or 4 mm capsules, either by incorporation of pDNA within the innermost layer of the shell or by adsorption to the microcapsules surface, and in vitro pDNA release, cytotoxicty and eGFP expression were studied. Sustained pDNA release over 3 days was observed using both loading techniques, being accelerated in the presence of protease. DNA loaded SF microcapsules resulted in efficient cell transfection along with low cytotoxicity after 3 days incubation compared to treatment with pDNA/ branched polyethylenimine complexes. Among the tested conditions highest transfection efficiencies were achieved using 1 mm capsules where pDNA was adsorbed to the capsule surface. Our results suggest that SF microcapsules are suitable for the localized delivery of pDNA, combining low cytotoxicity and high transfection efficiency. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Gene delivery Silk fibroin Microcapsules Cell transfection Cytotoxicty

1. Introduction The delivery of nucleic acid aims at introducing exogenous plasmid DNA (pDNA), antisense oligonucleotides, or small interfering RNA (siRNA) into host cells to influence protein expression. Despite the high hopes that have been pinned on nucleic acid delivery to treat a variety of inherited and acquired diseases, safe and efficient delivery still is a major challenge [1,2]. While viral vectors are superior to non-viral, synthetic delivery systems regarding gene transfer efficiency, significant safety concerns limit their usability [3]. Non-viral vectors on the other hand are easy to manufacture and to modify and are associated with fewer safety concerns [4,5]. Non-viral gene delivery systems have undergone significant development over the past decade, resulting in improved transfection efficiency and specificity [6e11] but substantial toxicity, limited physicochemical stability and incomplete protection of

* Corresponding author. Institute for Pharma Technology, University of Applied Sciences Northwestern Switzerland, Gruendenstrasse 40, 4132 Muttenz, Switzerland. Tel.: þ41 61 467 44 48. E-mail address: [email protected] (O. Germershaus). http://dx.doi.org/10.1016/j.biomaterials.2014.05.062 0142-9612/© 2014 Elsevier Ltd. All rights reserved.

encapsulated nucleic acids under physiological conditions still limit their use as viable clinical therapies [12e14]. The majority of non-viral vectors is currently developed for systemic application, requiring complex systems to achieve efficient and specific transfection of target cells. Localized delivery represents an alternative strategy avoiding several major obstacles encountered with systemic delivery, allowing for simpler delivery system design. To date, localized delivery often relies on naked pDNA or simple colloidal preparations, with limited potential for sustained or controlled release [15,16]. However, the application potential for localized gene delivery is immense with regard to tissue engineering as well as local therapy. While scaffold-based delivery of proteins requires the stabilization and controlled release of significant quantities of these fragile molecules [17,18], nucleic acid delivery needs smaller drug quantities and involves molecules being inherently more stable [15]. These considerations and the perceived lack of suitable delivery systems led us to design and characterize a system for the controlled localized delivery of nucleic acids based on silk fibroin (SF). Silk proteins are a particularly promising biomaterial due to the unique combination of biocompatibility, biodegradability, selfassembly, mechanical stability and control over structure and morphology [19,20]. Silk-based biomaterials have been shown to

7930

L. Li et al. / Biomaterials 35 (2014) 7929e7939

safeguard the activity of sensitive biomolecules in harsh environments, allowing delivering molecules that otherwise quickly lose efficacy [21e23]. While SF as a bulk biomaterial already has interesting properties, the controlled assembly of ultrathin SF films is particularly interesting as it allows further control of the architecture of the drug delivery system and associated properties such as drug release [24,25]. Layer-by-layer (LbL) assembly is a well-established technique for the fabrication of structurally diverse materials, including drug delivery systems, by sequential deposition of complementary species onto a template [26,27]. By using a sacrificial template and subsequent dissolution of the core, hollow microcapsules can be prepared possessing high loading capacity compared to nanoparticles and being able to accommodate a diverse range of molecules [28,29]. Microcapsules further allow incorporation of drugs such as nucleic acids either through preloading, i.e. incorporation of drugs into the capsule shell during shell fabrication or postloading, i.e. ad or absorption of payload after shell assembly [30e32]. The aim of the present study was to develop pDNA loaded SF microcapsules for local application allowing tunable, sustained release of nucleic acids resulting in efficient cell transfection and reduced cytotoxicity. We describe the generation of SF capsules with variable shell consisting of 2e10 layers of SF and establish preand post-loading techniques for pDNA incorporation. The effect of loading scheme as well as presence of protease on pDNA release is investigated and the biological response in vitro regarding cell viability and transfection efficiency is described. 2. Materials and methods 2.1. Materials and reagents Bombyx mori (silkworm) cocoon was obtained from Trudel Ltd. (Zurich, Switzerland). Polystyrene (PS) particles of 1 mm (Polybead® Microspheres 1.00 mm, CV 3%, Polysciences Europe GmbH, Eppelheim, Germany) and 4 mm diameter (Te0400, CV 5%, BS-Partikel GmbH, Wiesbaden, Germany) were used as templates, respectively. Cellulose ester dialysis tubing (Spectra/Por®, MWCO 3.5e5 kDa) were purchased from Spectrum Europe B.V. (DG Breda, The Netherlands). Dulbecco's Modified Eagle's Media (DMEM), fetal bovine serum (FBS), antibiotics and QuantiT™ PicoGreen® dsDNA reagent, used in the pDNA quantification were purchased from Life Technologies GmbH (Darmstadt, Germany). 3-(4,5-Dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide (MTT) and 40 -6-diamidino-2-phenylindole (DAPI) were obtained from SigmaeAldrich (Schnelldorf/Taufkirchen, Germany). The plasmid was labeled with the Cy5 fluorescent probe (pDNA-Cy 5) using Label IT® nucleic acid labeling kit from Mirus Bio (Madison, WI). Unless otherwise specified, all other reagents were purchased from SigmaeAldrich (Taufkirchen, Germany) and were at least of analytical quality. 2.2. Purification and fluorescent labeling of silk fibroin Cocoons from B. mori were boiled twice in an aqueous solution of 0.02 M Na2CO3 for 1 h, rinsed several times with ultrapure water (Millipore Milli-Q system) and dissolved in 9.3 M LiBr at 60  C to generate a 10% (w/v) solution. This solution was dialyzed (MWCO 3.5e5 kDa) against ultrapure water for 3 days by changing water daily to remove the ions and other impurities. The solution was collected, filtered and stored at 4  C. Fluorescein isothiocyanate (FITC)-labeled SF was prepared as described before with slight modifications [33]. SF was diluted to 2% (w/v) with ultrapure water and dialyzed against 500 ml of 0.1 M 2-(morpholino) ethanesulfonic acid (MES) buffer, pH 5.6 containing 150 mM NaCl. The buffered SF solution was mixed under stirring with 80 mg of 1-ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) (2 mM) and 220 mg N-hydroxysuccinimide (NHS) (5 mM) and the reaction was continued for 15 min. To quench the reaction, 20 mM a-mercaptoethanol was added. Activated carboxyl groups of SF were modified with 1.5 mg ethylenediamine for 2 h. After dialysis against 500 ml 0.1 M MES buffer, 10 mg of FITC was added to the solution, yielding a molar ratio of about 40:1 between fluorescent probe and SF. The reaction progressed for 2 h under slow stirring at room temperature. Finally, the solution was dialyzed against ultrapure water, yielding a fluorescent SF concentration of approx. 1% (w/v). 2.3. Fabrication of SF LbL microcapsules According to the literature on SF-based LbL coating processes, the maximum deposition was achieved once the protein concentration reached 1 mg/ml [24,34]. Therefore, all the coating experiments presented herein were performed at an SF

concentration of 1 mg/ml. Two hundred ml of a 5% w/v suspension of 1 or 4 mm PS particles was centrifuged at 10,000 g or 1500 g, respectively, for 2 min and the supernatant was removed and replaced with fresh ultrapure water. The pellet was agitated and subjected to a centrifugation and redispersion wash cycle. Prior to SF deposition, 1 ml of a 0.5 mg/ml aqueous solution of branched poly(ethylene imine) 25 kDa (bPEI25) was added to the particle suspension and incubated for 15 min at room temperature. After being washed three times, the particles were agitated to obtain a homogeneous suspension and incubated with 1 mg/ml SF solution for 15 min at 4  C. Assembly at low temperature was used to ensure SF protein stability and to decrease protein precipitation. After incubation, the particles were washed as described before. However, 1 mm particles required ultrasonication at 10% amplitude for 10 s to redisperse particles after centrifugation. The SF coated PS particles were immersed in 90% methanol for 15 min to induce silk crystalline b-sheet structure formation. The washed particles were then dried by a flow of nitrogen gas at 20 kPa (Turbovap LV, Zymark, Hopkinton, MA) and subjected to the next coating procedure. When a desired number of layers were deposited, the particles suspensions were washed several times with ultrapure water. To produce hollow microcapsules, PS cores were dissolved by shaking the dispersion for 4 h in THF solution. The dispersion of the microcapsules was then dialyzed against ultrapure water for 2 days. 2.4. Field emission scanning electron microscopy (FESEM) The morphology of SF microcapsules was characterized by FESEM (JSM-7500F, JEOL, Japan) with an accelerating voltage of 5 kV after coating with gold. For sample preparation, a drop of microcapsules suspension was applied to a mica film and dried under ambient conditions overnight. 2.5. Fourier-transform infrared spectroscopy (FTIR) The LbL assembly and structure analysis of SF microcapsules in dry state were performed by FTIR spectroscopy over a range of 4000e400 cm1. FTIR spectra of different samples were obtained by an FTIR-6100 spectrometer system (FTIR-6100, JASCO, Gross-Umstadt, Germany). Background measurements were taken twice with an empty cell and subtracted from sample readings. The samples were analyzed without any further preparation. 2.6. Confocal laser scanning microscopy (CLSM) and fluorescence microscopy Fluorescence images of SF microcapsules were obtained with a Leica DMI6000 (Wetzlar, Germany) confocal microscope or a Zeiss Axio Observer Z1 (Oberkochen, Germany) fluorescence microscope. Hollow microcapsules for microscopy were prepared as described above using FITC-labeled SF. Samples were prepared by embedding 20 ml of the capsules suspension using standard glass slide and coverslips. Samples were stored for at least 30 min at room temperature prior to imaging to allow sedimentation of capsules. 2.7. Flow cytometry To follow film buildup on PS particles, FITC-labeled SF was deposited as described before. The fluorescence intensity with different layers was monitored on a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ) using an 488 nm laser line of an argon-ion laser and emission bandpass filter 530/30 nm for detection of FITC fluorescence. Cy5 was excited using a 635 nm diode laser and emission was detected using a bandpass filter between 653 and 667 nm. Approximately 10,000 particles were analyzed in each experiment. 2.8. Zeta-potential measurements Zeta potentials of unmodified and SF coated 1 and 4 mm PS particles were measured in aqueous solutions using a Delsa Nano HC (Beckman Coulter, Brea, CA). Results were obtained at ambient conditions and by averaging three independent measurements of 6 sub-runs each. 2.9. Loading of pDNA using the pre- and post-loading approach Two different approaches to loading pDNA onto microcapsules, preloading and postloading, were evaluated [32]. Plasmid DNA was labeled with Cy5 using Mirus Label IT Nucleic Acid Labeling Kit according to the manufacturer's instructions. The Cy5-labeled pDNA was purified through a microspin column to remove any unbound dye. The recovered labeled plasmid was stored protected from light at 20  C until use. For the preloading method, 200 ml of 5% (w/v) suspension of bPEI25 modified 1 or 4 mm PS particles in ultrapure water were incubated with 1 ml of 10 mg/ml Cy5labeled pDNA solution for 20 min. Subsequently, 4 layers of FITC-labeled SF were deposited. Finally, the PS cores were dissolved using THF to produce capsules containing pDNA. For postloading, 4 layers of FITC-labeled SF were deposited on bPEI25 modified PS particles and then incubated with 0.1 mg/ml bPEI25 solution, resulting in a positive surface charge of particles. After core removal microcapsules were incubated with 1 ml of 10 mg/ml Cy5-labeled pDNA for 20 min while shaking at 1200 rpm (Thermomixer comfort, Eppendorf, Wesseling-Berzdorf, Germany) to avoid sedimentation and agglomeration. Subsequently, the capsules were washed

L. Li et al. / Biomaterials 35 (2014) 7929e7939 with ultrapure water three times. An inverted epifluorescence microscope Axio Observer Z1 equipped with a HXP120C lamp and an AxioCam MRm camera (Zeiss, Oberkochen, Germany) was used for image acquisition. The specimen was imaged using a Plan-Apochromat 63/1.40 oil objective and fluorescence excitation at 470/ 40 nm, dichroic beam splitter at 495 nm and emission bandpass of 525/50 nm for FITC fluorescence. Cy5 was imaged using excitation at 640/30 nm, dichroic at 660 nm and emission filter of 690/50 nm. In the case of 1 mm microcapsules, a PlanApochromat 100/1.40 oil objective was used. 2.10. Loading efficiency The loading efficiency was analyzed by quantification of pDNA in the supernatant using PicoGreen assay. For preloading, bPEI25 modified PS particles were washed three times in ultrapure water by centrifugation and redispersion. A defined number of particles (~1.68  108; 200 ml of 5% suspension) were incubated with increasing amounts of pDNA (2.5, 5, 12.5, 15, and 20 mg) to determine the amount of plasmid required to saturate the particles surface. The samples were incubated for 20 min while shaking at 1200 rpm (Thermomixer comfort, Eppendorf, WesselingBerzdorf, Germany). After centrifugation at 1500 g for 2 min, a known volume of supernatant was collected. The samples were subsequently washed with ultrapure water three times and the supernatants were collected separately. To quantify the actual amount of pDNA that adsorbed onto the particles, the amount of pDNA in supernatant was subtracted from the initial amount of pDNA added to the microcapsules. SF film deposition was done as described before and the supernatant after each adsorption step was collected. pDNA binding efficiency was calculated as follows: Efficiencyð%Þ ¼

initial amount of DNAeDNA in supernatant  100 DNA in supernatant

For postloading, after loading of pDNA onto microparticles as described above, the samples were washed with ultrapure water three times and the supernatants were collected separately for analysis. 2.11. Release studies To characterize the release behavior after pDNA preloading or postloading, 1 and 4 mm microcapsules were suspended in PBS with a gentle agitation at 37  C for 3 days. The fluorescence intensity of Cy5-labled pDNA associated with SF microcapsules was monitored using flow cytometry. SF degradation was studied in PBS in the presence and absence of 0.5 mg/ml Protease XIV at 37  C using fluorescence microscopy and flow cytometry. 2.12. Cell culture NIH/3T3 fibroblasts were cultured in DMEM supplemented with 10% FBS, and 1% penicillin/streptomycin. Cells were seeded in 24-well plates at a density of 1  104 cells per well and cultured at 37  C, 5% CO2 and 95% humidity. Before cell seeding, the pDNA loaded SF microcapsules were prepared by suspension in 70% ethanol for 30 min, washed 3 times with PBS for 20 min each and incubated with DMEM (2% serum) for 1 h. 2.13. Evaluation of cytotoxicity Cytotoxicity in the presence of microcapsules was determined using MTT assay. NIH/3T3 fibroblasts were seeded into 96-well plates at a density of 5000 cells per well. After 24 h medium was aspirated and replaced by 200 ml of microcapsule suspension in cell culture medium with FBS at three capsule/cell ratios (10:1, 100:1, 1000:1). The cells were then incubated with the particles for 3 days. Afterwards, medium was replaced by DMEM with 0.5% serum containing 100 ml of 0.5 mg/ml MTT solution. After 4 h incubation at 37  C in the dark, medium was aspirated and formazan crystals were dissolved in 200 ml dimethylsulfoxide per well. Measurement was performed using spectrophotometric microplate reader (SpectraMax 250, Molecular Devices, Sunnyvale, CA) at a wavelength of 570 nm. Cytotoxicity was calculated relative to positive (wells with untreated cells; set to 100% relative absorption at 570 nm) and negative controls (wells without cells; set to 0% relative absorption at 570 nm). 2.14. Cell transfection For transfection experiments 5  104 NIH/3T3 fibroblasts were seeded per well in a 24-well plate. After 24 h of growth, medium was replaced by 0.5 ml of fresh cell culture medium. As control, bPEI25epDNA complexes were prepared in a final volume of 100 ml (containing 2 mg pDNA) per well according to Ref. [35]. The pDNA pre- and post-loading SF microcapsules of two different sizes (1 or 4 mm) were incubated with cells for 1 day or 3 days. The capsule/cell ratio was 1000:1 containing approx. 2.5 mg pDNA. All experiments were performed in the presence of serum in cell culture medium. After 1 day or 3 days incubation period, the cells were detached and collected in 4% paraformaldehyde in PBS. EGFP expression was quantified by flow cytometry with excitation at 488 nm and using an emission filter of 530/30 nm (FACSCalibur, Becton Dickinson, Franklin Lakes, NJ). In addition, eGFP expression was

7931

also observed by using an inverted fluorescence microscopy after 1, and 3 days incubation. 2.15. Statistical analysis Results are expressed as means ± standard deviations. Statistical analysis was performed using Student's t-test as well as one-way analysis of variance (ANOVA) followed by the Tukey HSD test for post hoc comparison (Origin 7, OriginLab, Northampton, MA). Difference was considered statistically significant if p < 0.05.

3. Results 3.1. Fabrication of SF microcapsules by LbL deposition We used 1 and 4 mm PS particles as a sacrificial template and SF as a shell component to fabricate microcapsules by LbL coating, varying the number of SF layers between 2 and 10 layers (Fig. 1A). Due to the anionic surface charge of PS particles, direct coating with anionic SF proved to be inefficient (data not shown). Therefore, particles were first coated with bPEI25 to generate a positively charged surface allowing efficient deposition of SF. SF shell formation on PS template particles as well as appearance after core removal was followed by SEM (Fig. 1B). Upon coating with 4 layers of SF, surface roughness of initially smooth PS particles increased considerably, signifying deposition of SF film. Removal of the PS core resulted in hollow microcapsules, which collapsed during drying of the microcapsule suspension on the mica wafer for SEM imaging. 3.2. Composition and morphology of SF microcapsules Plain PS core particles, SF coated PS particles and SF microcapsules after PS core removal were characterized by FTIR spectroscopy (Fig. 2). The spectrum of unmodified PS particles was dominated by peaks at 2920 and 2850 cm1 from methylene stretching vibrations, aromatic ring breathing modes at 1600, 1490 and 1450 cm1 and aromatic CeH stretching vibrations at 3080, 3059 and 3024 cm1 (Fig. 2a). After coating of PS particles with SF and methanol treatment additional characteristic peaks were observed at 1625 cm1 (Amide I; carbonyl stretching), 1516 cm1 (Amide II; secondary NeH bending), and 1230 cm1 (Amide III; CeN stretching) indicating SF b-sheet formation (Fig. 2b) [36]. Subsequently, PS cores were removed by treatment with 90% THF, resulting in considerable reduction of the characteristic peaks of PS, while the secondary structure of SF remained unchanged (Fig. 2c). Microcapsules composed of 4, 6, 8, and 10 SF layers using 1 or 4 mm PS template particles were prepared and further characterized by SEM and CLSM (Fig. 3). Upon drying of the microcapsules suspension on a mica wafer for SEM imaging the capsules collapsed, exhibiting multiple random folds likely due to local instabilities and large wrinkles resulting from capillary forces acting during drying [25]. The morphology of 1 mm SF microcapsules appeared to be similar for all preparations and was characterized by a porous and granular structure (Fig. 3A). In the case of 4 mm SF microcapsules the surface roughness seemed to increase with increasing number of SF layers and distinct globular structures were observed on the microcapsule surface (Fig. 3B). Imaging of microcapsules prepared using FITC-labeled SF in aqueous solution using CLSM confirmed that capsules were uniform in shape and size with an average diameter of 4 mm as defined by the PS core particles and without signs of capsule collapse (Fig. 3C). 3.3. Monitoring of LbL deposition process SF shell formation was followed using FITC-labeled SF during LbL coating of PS particles by fluorescence microscopy and flow cytometry. The increase of overall fluorescence intensity of

7932

L. Li et al. / Biomaterials 35 (2014) 7929e7939

Fig. 1. Fabrication of silk fibroin layer-by-layer microcapsules. (A) Schematic illustration of SF microcapsules fabrication using PS particles as a template. The process involves coating bPEI25 (a polycationic polymer) onto PS particles, adsorbing SF (a polyanionic protein), inducing b-sheet formation of SF by treatment with methanol and nitrogen gas, multiple repeated deposition cycles and finally core removal. (B) Representative SEM images showing surface morphology of 4 mm PS particles, after 4 layers of SF coating and the dried microcapsules after core removal. Scale bar: 1 mm.

particles with increasing layer number was evident in fluorescence microscopy (Fig. 4A). However, fluorescence of particles upon deposition of the first two layers of SF was observed to be inhomogeneous, and only a fraction of particles was coated with FITC-SF

while others showed no fluorescence (Fig. S1). At higher layer numbers, homogeneous coverage of the particles with FITC-SF was observed. Flow cytometry was applied to quantify SF deposition per layer (Fig. 4B and C). In agreement with the results of fluorescence microscopy, a linear increase of fluorescence intensity with increasing number of SF layers was observed after the second layer (Fig. 4C). The initial geometric mean fluorescence intensity of bPEI25-modified PS particles of 3.8 ± 0.1 showed a statistically significant increase to 1228 ± 9 after coating with two layers of FITC-SF. Coating with further SF layers lead to a statistically significant increase of approx. 350 units per two additional SF layers until a final geometric mean fluorescence of 2715 ± 57 was reached after the tenth SF layer. SF deposition onto PS particles was further characterized by zeta potential measurements (Fig. 5). We found that both, 1 and 4 mm unmodified PS particles initially showed a negative zeta potential of 42.2 ± 2.4 mV and 39 ± 2 mV, respectively. Deposition of bPEI25 as a first layer led to an increase of the zeta potential to þ58 ± 1.1 mV and þ63.1 ± 0.52 mV, respectively and deposition of subsequent SF layers resulted in a gradual reduction of the zeta potential in both cases. 3.4. Plasmid DNA loading using pre- and post-loading approaches

Fig. 2. FTIR spectra of PS particles (a), PS particles with 10 layers of SF coating (b) and microcapsules with 10 layers of SF after core removal (c). The dotted lines indicated characteristic absorption peaks of SF.

Within this study two approaches to drug loading of SF microcapsules were established and compared with regards to loading and transfection efficiency as well as cytotoxicity. Plasmid DNA was

L. Li et al. / Biomaterials 35 (2014) 7929e7939

7933

Fig. 3. Representative SEM images of the dried SF microcapsules of 1 mm (A) and 4 mm (B) with different number of SF layers (4, 6, 8, and 10 layers). Scale bar: 1 mm. (C) Representative bright-field (left) and CLSM (right) images of 4 mm SF microcapsules (10 SF layers, SF was labeled with FITC) resuspended in aqueous solution. Scale bar: 20 mm. The inset image shows a higher magnification of the microcapsules. Scale bar: 10 mm.

either incorporated during preparation of the capsules shell, i.e. as the first, innermost layer (designated as preloading) or pDNA was adsorbed onto the microcapsules after shell formation (designated as postloading). To achieve efficient pDNA loading during postloading, bPEI25 was coated onto SF multilayers as a last layer, followed by subsequent core removal. This step aimed at introducing a positive surface charge resulting in efficient adsorption of negatively charged pDNA by electrostatic interaction (Fig. 6A). To optimize pDNA adsorption onto bPEI25 coated PS particles and subsequent SF deposition for efficient preloading, we determined the amount of pDNA that was adsorbed onto particles in relation to the initial pDNA concentration per fixed number of particles during incubation (Fig. S2A). The adsorbed amount of pDNA increased with pDNA concentration up to a concentration of 15 mg/ml per 1.68  108 particles. Above a concentration of 15 mg/ml per 1.68  108 particles, surface saturation was achieved. Complete saturation of the particle surface with pDNA negatively affects the subsequent polymer adsorption and capsule stability and was therefore avoided [31]. A concentration of 10 mg/ml pDNA per 1.68  108 particles (approx. 75% saturation) was used for all subsequent experiments. Plasmid DNA coding for enhanced green fluorescent protein (eGFP) was incorporated using both pre- and post-loading approaches into microcapsules of 1 and 4 mm and microcapsule morphology as well as pDNA distribution was investigated by fluorescence microscopy using FITC-labeled SF and Cy5 labeled pDNA revealing non-uniform distribution of pDNA for samples prepared using the postloading approach as compared to preloading samples. (Fig. 6B and Fig. S3).

We further determined the pDNA loading efficiency resulting after applying the preloading or postloading approach and comparing 1 and 4 mm microcapsules (Fig. S2B). Preloading of 4 mm microcapsules was significantly less efficient (82.5 ± 1.56%) than preloading of 1 mm microcapsules (90.2 ± 3.55%) and postloading of 4 and 1 mm microcapsules (93.4 ± 2.56% and 95 ± 1.78%, respectively). No significant differences in pDNA loading efficiency were found between 1 and 4 mm microcapsules prepared using the postloading approach. 3.5. In vitro pDNA release SF as a natural protein can be proteolytically degraded and resorbed in vivo over a long-term period [37]. To investigate whether this mechanism can be successfully exploited to release pDNA, microcapsules consisting of 4 layers of SF and prepared by the pre- or post-loading approach were incubated in PBS or in PBS with Protease XIV, respectively. The release and degradation behavior over time were studied by following the morphology and fluorescence intensities of Cy-5 labeled pDNA and FITC-labeled SF using fluorescence microscopy and flow cytometry, respectively. FITC-labeled SF microcapsules exhibited a slight deformation and partial degradation of the capsules shell as compared to fresh microcapsules and distribution of Cy-5 labeled pDNA incorporated using the preloading approach was less uniform after 3 days of incubation in PBS (Fig. 7A). Incubation of the same particles in PBS with Protease XIV for 3 days resulted in extensive disassembly of SF shells and the amount of Cy-5 labeled pDNA associated with microcapsules was strongly reduced (Fig. 7A). Microcapsules prepared

7934

L. Li et al. / Biomaterials 35 (2014) 7929e7939

Fig. 4. (A) Fluorescence microscopy images of 4 mm PS particles with 2, 4, 6, 8, 10 layers of FITC-SF. Scale bar: 10 mm. (B and C) Flow cytometry analysis of the fluorescence intensity of 4 mm PS particles with 2, 4, 6, 8, 10 layers of FITC-SF. Blue (I) and green (II) dotted square areas indicate the initial two layers of SF and the subsequent eight layers, respectively. Data is presented as mean ± SD, n ¼ 3. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

using the postloading approach showed similar Cy-5 fluorescence reduction after 3 days incubation in PBS (Fig. S4). Plasmid DNA release was investigated by quantification of Cy-5 fluorescence of particles over time using flow cytometry (Fig. 7B). We observed a decrease of geometric mean fluorescence intensity of particles in all samples after 3 days incubation. In the PBS groups, 1 mm microcapsules showed a higher residual pDNA content (56.8% for preloading and 60.2% for postloading) compared to 4 mm microcapsules (36.7% for preloading and 45.7% for postloading). After incubation with protease, no major differences between 1 mm (13.0% for preloading and 17.1% for postloading) and 4 mm microcapsules (19.0% for preloading and 11.5% for postloading) were observed. 3.6. Studies on cytotoxicity and transfection efficiency Fig. 5. Zeta-potential analysis of 1 and 4 mm bare PS particles, after coating with bPEI25, and after coating with different number of SF layers. Data is presented as mean ± SD, n ¼ 6, two experiments.

Microscopic evaluation of cell morphology of NIH/3T3 fibroblasts in the presence of SF microcapsules revealed a tendency to agglomeration of microcapsules and strong association with cells

L. Li et al. / Biomaterials 35 (2014) 7929e7939

7935

Fig. 6. Preparation of plasmid DNA loaded SF microcapsules. (A) Schematic illustration of pDNA pre- and post-loading approaches. Preloading: The pDNA was adsorbed onto bPEI25 functionalized PS particles; LbL assembly of SF was performed onto the pDNA-coated particles; SF was stabilized and the core removed. Postloading: PS particles were coated with bPEI25; LbL assembly of SF was performed onto bPEI25-coated PS particles; following bPEI25 coating again, the pDNA was adsorbed onto the bPEI25eSF shell and the core was removed. (B) Representative fluorescence microscopy images of 4 mm SF microcapsules prepared by pDNA pre- or post-loading. SF was labeled with FITC (green) and pDNA was labeled with Cy5 (red). Scale bar: 10 mm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

(Fig. 8A). The cell morphology was unaffected when cells were incubated with microcapsules at a capsule-to-cell ratio of 100:1 for 24 h. The cytotoxic effects of pDNA loaded SF microcapsules with sizes of 1 or 4 mm were further studied by MTT assay applying increasing capsule-to-cell ratios (Fig. 8B). The results demonstrate that at a capsule-to-cell ratio of 100:1 the metabolic activity after incubation for 96 h was not significantly reduced compared to untreated cells. When the capsule-to-cell ratio was increased to 1000:1, cytotoxicity was significantly increased in the case of postloaded microcapsules. Based on these results SF microcapsules of different size (1 and 4 mm) and prepared by both pre- and post-loading approach can be considered to be non-toxic to NIH/3T3 fibroblasts up to a capsule-to-cell ratio of 100:1. On the other hand, in the presence of bPEI25/pDNA complexes NIH/3T3 cytotoxicity was significantly increased with metabolic activity decreasing to about 50% of untreated control after incubation for 96 h. The transfection efficiency was assessed using pre- and postloaded microcapsules of 1 and 4 mm by the percentage of cells

expressing eGFP after 24 or 96 h by flow cytometry (Fig. 9). Branched PEI25epDNA complexes used as a positive control transfected 51.9 ± 12.8% of cells after incubation for 24 h, which was significantly higher than the transfection efficiency of all SF microcapsule preparations. However, the transfection efficiency of SF microcapsule preparations significantly increased with incubation time. Microcapsules of 1 mm prepared using the postloading approach showed similar transfection efficiency as bPEI25/pDNA treated NIH/3T3 fibroblasts after 3 days (57.2 ± 3.1% and 51.8 ± 7.2%, respectively). Transfection efficiency of microcapsules of 1 mm was significantly higher than that of 4 mm microcapsules, using both pre- and post-loading approach and pDNA incorporation by postloading resulted in significantly higher transfection efficiencies compared to incorporation by preloading. The transfection efficiency and cell morphology were in addition qualitatively assessed by fluorescence microscopy (Fig. S5), confirming the results obtained by flow cytometry and indicating that despite efficient cell transfection, SF microcapsules did not negatively affect cell morphology.

7936

L. Li et al. / Biomaterials 35 (2014) 7929e7939

Fig. 7. Release behavior of pDNA from SF microcapsules. (A) Representative fluorescence images of preloaded 4 mm SF microcapsules incubated in PBS or 0.5 mg/ml Protease XIV in PBS at 37  C for 3 days. SF was labeled with FITC (green) and pDNA was labeled with Cy5 (red). Scale bar: 20 mm. (B) Flow cytometry analysis of the geometric mean fluorescence intensity of Cy5 labeled pDNA loaded onto 1 or 4 mm microcapsules by pre- or post-loading. SF microcapsules were incubated in PBS and Protease XIV in PBS at 37  C for 1 day and 3 days. At least 3500 individual capsules were analyzed in each experiment (****p  0.0001). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

4. Discussion Platforms enabling the efficient and sustained delivery of nucleic acids are of high interest in various fields including tissue engineering and local or topical therapy. For such applications, nucleic acid delivery is expected to be superior to protein delivery due to the higher structural, chemical and thermodynamic stability of nucleic acids compared to therapeutic proteins as well as the inherently limited loading capacity of scaffolds. Nevertheless, only a limited number of nucleic acid delivery systems have been developed for this purpose, mostly relying on natural materials due to their superior biocompatibility [38e40]. We herein provide evidence that SF microcapsules are efficient carriers for the sustained, localized delivery of nucleic acids. SF with its proven biocompatibility, low immunogenicity and high mechanical stability is a material that seems optimally suited as a scaffold for sustained nucleic acid delivery [19,41]. Efficient non-viral gene delivery using SF-based systems was shown for block copolymers composed of recombinant SF, poly-L-lysine and a peptide sequence [42e44]. Similarly, sustained release of plasmid DNA or therapeutic viruses was achieved using recombinant silkeelastin like protein polymer gels, resulting in improved transfection efficiency compared to free plasmid DNA and significantly reduced toxicity compared to treatment with free virus [45,46]. The preparation of SF microcapsules by LbL coating was straightforward using PS particles as sacrificial template. Deposition of negatively charged SF onto plain PS particles proved to be inefficient due to the negative surface charge of PS particles (Fig. 5). After coating of PS particles with polycationic bPEI25, reversal of the surface charge was achieved and SF was efficiently deposited. Interestingly, efficiency of SF deposition was high for the first two layers (approx. 1200 fluorescence units per 2 SF layers) and dropped thereafter to approx. 350 fluorescence units per 2 SF layers (Fig. 4C). Zeta potential measurements revealed that particles were positively charged after coating with 2 layers of SF but zeta

potential approached zero after 4 layers of SF. These observations led us to conclude that SF deposition by LbL coating as performed herein relies on both electrostatic interactions between cationic particle surface and anionic SF, which are dominant during deposition of the first 2 SF layers as well as hydrophobicehydrophobic interactions controlling deposition of subsequent layers [24,25]. In addition, SF shells are stabilized by b-sheets induced by methanol treatment after deposition of each layer providing physical crosslinks between layers [47,48]. The presence of stabilizing b-sheets after shell formation as well as after PS core removal was confirmed by FTIR spectroscopy (Fig. 2). Scanning electron microscopy of SF microcapsules further revealed an apparent increase of the number of aggregates of approx. 100e500 nm visible on the microcapsule surface after coating of several SF layers (Fig. 3), being consistent with globular structures observed in SF films by several authors [49,50], and being attributed to the conversion of the amorphous to beta-pleated SF structure [50,51]. Incorporation of pDNA into SF microcapsules was performed by pre- and post-loading in an attempt to identify an optimal delivery system with regards to drug encapsulation efficiency, cytotoxicity, drug release and transfection efficiency. Preloading was achieved through electrostatic adsorption of pDNA onto bPEI25 coated PS particles prior to SF deposition such that nucleic acid/bPEI25 complexes were entrapped inside of the microcapsules, where the SF multilayer shell constituted both a protective as well as a diffusion barrier (Fig. 6A, B-Preloading). In the postloading approach, the cargo is loaded into or onto pre-fabricated capsules by adsorption and/or diffusion through the capsule shell (Fig. 6A, BPostloading) [52]. Both strategies have been successfully applied in several studies [31,53e57] and used to tailor drug release characteristics via the loading process and shell thickness [58]. Plasmid DNA encapsulation efficiency was found to be significantly higher for 4 mm microcapsules after postloading while the loading technique had no significant influence on encapsulation efficiency for 1 mm microparticles (Fig. S3B). We hypothesize that

L. Li et al. / Biomaterials 35 (2014) 7929e7939

7937

Fig. 9. Effect of pDNA pre- and post-loading onto SF microcapsules of different size (1 and 4 mm) on cell transfection efficiency in NIH/3T3 fibroblasts. Flow cytometry analysis of eGFP expression presented as percentage of eGFP positive cells of untreated cells, cells incubated with PEI-pDNA, pDNA pre- and post-loaded SF microcapsules (1 and 4 mm) for 1 day and 3 days (*p  0.05; data is presented as mean ± SD, n ¼ 3, two experiments). At least 8000 events were analyzed in each experiment.

Fig. 8. In vitro cellemicrocapsule interactions. (A) Fluorescence microscopy images of the Cy5-pDNA (red) preloaded 4 mm SF microcapsules incubated with NIH/3T3 fibroblasts at capsule/cell ratio 100:1 for 24 h. Scale bar: 50 mm. (B) Cytotoxicity in NIH/3T3 fibroblasts without treatment and after incubation with PEI-pDNA, pDNA pre- and post-loading SF microcapsules (1 and 4 mm) at different capsule/cell ratios (10:1, 100:1 and 1000:1) for 3 days, as measured by an MTT assay. Cytotoxicity is normalized to untreated cells (100% OD 570). (*p  0.05; data is presented as mean ± SD, n ¼ 3, two experiments). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the porous and granular morphology encountered by pDNA during postloading of 4 mm microcapsules allows more efficient binding than the smooth surface of PS particles encountered during preloading (Fig. 3). However, as bPEI25 is known for its cytotoxicity, we further assumed that bPEI25/pDNA complexes located within the microcapsules (preloading approach) would result in lower cytotoxicity compared to a situation was bPEI/pDNA complexes are primarily located on the microcapsule surface (postloading approach). Sustained release of pDNA as studied by the decline of microcapsule-associated Cy5-pDNA fluorescence was observed for all groups, being considerably higher in the presence of Protease XIV than in pure PBS. Release is thought to occur through diffusion and desorption in the case of preloaded microcapsules while pDNA release after postloading likely depends on desorption (Fig. 10A). Lu et al. reported that during SF degradation by protease, hydrophilic

blocks are first degraded, leaving behind blocks of high crystallinity which finally move into solution as particles [59]. SF degradation by protease resulting in increased shell porosity therefore significantly improved pDNA release and desorption for both, pre- and postloaded microcapsules. Achieving a proper balance between cytotoxicity and transfection efficiency is one of the key challenges during development of non-viral gene delivery systems [60]. We therefore compared the cytotoxicity and transfection efficiency of pDNA loaded SF microcapsules with that of pDNA/bPEI25 complexes, being a well-known standard in non-viral gene delivery. However, direct comparison of these delivery systems must be made with caution as the different morphology and release characteristics of these systems significantly affect cytotoxicity. Metabolic activity of cells after incubation with microcapsule preparations was not significantly different from untreated cells up to a microcapsule/cell ratio of 100:1 and cell morphology was unaffected (Fig. 8). However, at a ratio of 1000:1, microcapsules prepared by postloading resulted in significantly increased cytotoxicity whereas preloaded microcapsules were still non-toxic, confirming our assumption that bPEI25 coating at the last step of the postloading negatively affects cell compatibility. The significant reduction of cytotoxicity of bPEI25/ pDNA complexes after encapsulation into and/or adsorption to SF microcapsules is assumed to primarily result from the gradual, prolonged release of complexes from microcapsules, reducing the actual complex concentration and hence their cytotoxicity. Therefore, and as shown for elastin-like polypeptide nanocapsules as well as collagen microcapsules before, the alteration of the release pattern using natural polymer based reservoir systems significantly improves the biocompatibility of well-known transfecting agents in vitro [61,62]. In agreement with the observed sustained release of pDNA/ bPEI25 complexes from microcapsules, transfection efficiency in vitro was found to increase over time for all microcapsule preparations (Fig. 9). We further observed that transfection efficiency of microcapsules of 1 mm was significantly higher than that of 4 mm microcapsules, both after pre- and post-loading. We

7938

L. Li et al. / Biomaterials 35 (2014) 7929e7939

Fig. 10. Schematic illustration of the effects of the different loading approaches and sizes of pDNA loaded SF microcapsules on release behavior and cellecapsule interactions. (A) Preloading: pDNAebPEI complexes are released by diffusion out of the SF capsules. Postloading: Complexes are released by desorption from the surface of SF capsules. (B) Different sizes of gene carriers incorporating the same total amount of pDNA result in different distribution density on the cell surface, affecting cell viability and transfection.

hypothesize that this result may be related to the density of microcapsules on the cell surface, resulting in a higher probability of interactions and a closer contact with the cell surface with reducing capsule size (Fig. 10B). While this observation requires further detailed investigations, it potentially represents another important parameter allowing balancing cytotoxicity and transfection efficiency as the effect of capsule size on cytotoxicity was less distinct than its effect on transfection efficiency. Apart from capsule size, the loading scheme significantly affected transfection efficiency with postloading resulting in significantly higher transfection efficiency than preloading in all cases. Since pDNA release profiles were similar for pre- and post-loaded microcapsules, we assume that the surface charge of particles being negative for preloaded capsules and positive for postloaded capsules affects transfection efficiency. Nonspecific binding is expected for gene carriers with a positive surface potential, enhancing the cell adhesion, internalization and transfection [60]. In contrast, pDNA encapsulation within the microcapsule by the preloading approach resulted in particles with negative zeta potential reducing celleparticle interactions. 5. Conclusions We herein describe an SF LbL assembly process allowing the formation of SF shells of adjustable thickness on PS particles that, after template removal, yield monodisperse and stable SF microcapsules of variable size. These microcapsules were loaded with pDNA by pre- and post-loading techniques. Plasmid DNA loading efficiency depended on the mode of pDNA loading and pDNA release was primarily affected by the presence of proteolytic enzymes in release medium. Plasmid DNA loaded SF microcapsules efficiently transfected NIH/3T3 fibroblasts, while significantly reducing cytotoxic effects compared to pDNA/bPEI25 complexes. The presented results highlight the potential of SF microcapsules as a potent carrier for controlled, localized gene delivery. Acknowledgments L. Li was supported by “China Scholarship Council 201206050045” and “Academic Award for Excellent Ph.D.

candidates funded by Ministry of Education of China”. This work was supported by the IZKF Würzburg, grant number D-218. The generous gift of cocoons by Trudel Inc., Zürich is acknowledged. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biomaterials.2014.05.062. List of abbreviations bPEI25 CLSM DAPI DMEM dsDNA EDC eGFP FBS FITC FTIR LbL MES MTT NHS PBS pDNA PS SEM SF siRNA THF

branched polyethylenimine 25 kDa confocal laser scanning microscopy 40 ,6-diamidino-2-phenylindole Dulbecco's modified Eagle's medium double stranded deoxyribonucleic acid 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide enhanced green fluorescent protein fetal bovine serum fluorescein isothiocyanate Fourier transform infrared spectroscopy layer-by-layer 2-(N-morpholino)ethanesulfonic acid 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide N-hydroxysuccinimide phosphate buffered saline plasmid deoxyribonucleic acid polystyrene scanning electron microscopy silk fibroin small interfering ribonucleic acid tetrahydrofuran

References [1] Edelstein ML, Abedi MR, Wixon J. Gene therapy clinical trials worldwide to 2007 e an update. J Gene Med 2007;9:833e42. [2] Sheridan C. Gene therapy finds its niche. Nat Biotechnol 2011;29:121e8. [3] Lundstrom K. Latest development in viral vectors for gene therapy. Trends Biotechnol 2003;21:117e22.

L. Li et al. / Biomaterials 35 (2014) 7929e7939 [4] Germershaus O, Merdan T, Bakowsky U, Behe M, Kissel T. Trastuzumab-polyethylenimine-polyethylene glycol conjugates for targeting Her2-expressing tumors. Bioconjug Chem 2006;17:1190e9. [5] Germershaus O, Mao SR, Sitterberg J, Bakowsky U, Kissel T. Gene delivery using chitosan, trimethyl chitosan or polyethylenglycol-graft-trimethyl chitosan block copolymers: establishment of structure-activity relationships in vitro. J Control Release 2008;125:145e54. [6] Mintzer MA, Simanek EE. Nonviral vectors for gene delivery. Chem Rev 2009;109:259e302. [7] De Laporte L, Cruz Rea J, Shea LD. Design of modular non-viral gene therapy vectors. Biomaterials 2006;27:947e54. [8] Scholz C, Wagner E. Therapeutic plasmid DNA versus siRNA delivery: common and different tasks for synthetic carriers. J Control Release 2012;161:554e65. [9] Germershaus O, Neu M, Behe M, Kissel T. HER2 targeted polyplexes: the effect of polyplex composition and conjugation chemistry on in vitro and in vivo characteristics. Bioconjug Chem 2008;19:244e53. [10] Neu M, Germershaus O, Behe M, Kissel T. Bioreversibly crosslinked polyplexes of PEI and high molecular weight PEG show extended circulation times in vivo. J Control Release 2007;124:69e80. [11] Neu M, Germershaus O, Mao S, Voigt KH, Behe M, Kissel T. Crosslinked nanocarriers based upon poly(ethylene imine) for systemic plasmid delivery: in vitro characterization and in vivo studies in mice. J Control Release 2007;118:370e80. [12] Putnam D. Polymers for gene delivery across length scales. Nat Mater 2006;5: 439e51. [13] Zhang Y, Satterlee A, Huang L. In vivo gene delivery by nonviral vectors: overcoming hurdles? Mol Ther 2012;20:1298e304. [14] Nguyen J, Szoka FC. Nucleic acid delivery: the missing pieces of the puzzle? Acc Chem Res 2012;45:1153e62. [15] Bonadio J. Tissue engineering via local gene delivery: update and future prospects for enhancing the technology. Adv Drug Deliv Rev 2000;44:185e94. [16] Kim HS, Yoo HS. MMPs-responsive release of DNA from electrospun nanofibrous matrix for local gene therapy: in vitro and in vivo evaluation. J Control Release 2010;145:264e71. [17] Zhao H, Heusler E, Jones G, Li L, Werner V, Germershaus O, et al. Decoration of silk fibroin by click chemistry for biomedical application. J Struct Biol 2014;186(3):420e30. [18] Germershaus O, Werner V, Kutscher M, Meinel L. Deciphering the mechanism of protein interaction with silk fibroin for drug delivery systems. Biomaterials 2014;35:3427e34. [19] Altman GH, Diaz F, Jakuba C, Calabro T, Horan RL, Chen JS, et al. Silk-based biomaterials. Biomaterials 2003;24:401e16. [20] Wenk E, Merkle HP, Meinel L. Silk fibroin as a vehicle for drug delivery applications. J Control Release 2011;150:128e41. [21] Lu Q, Wang XQ, Hu X, Cebe P, Omenetto F, Kaplan DL. Stabilization and release of enzymes from silk films. Macromol Biosci 2010;10:359e68. [22] Wenk E, Murphy AR, Kaplan DL, Meinel L, Merkle HP, Uebersax L. The use of sulfonated silk fibroin derivatives to control binding, delivery and potency of FGF-2 in tissue regeneration. Biomaterials 2010;31:1403e13. [23] Zhang JE, Pritchard E, Hu X, Valentin T, Panilaitis B, Omenetto FG, et al. Stabilization of vaccines and antibiotics in silk and eliminating the cold chain. Proc Natl Acad Sci U S A 2012;109:11981e6. [24] Wang X, Wenk E, Hu X, Castro GR, Meinel L, Wang X, et al. Silk coatings on PLGA and alginate microspheres for protein delivery. Biomaterials 2007;28: 4161e9. [25] Shchepelina O, Drachuk I, Gupta MK, Lin J, Tsukruk VV. Silk-on-silk layer-bylayer microcapsules. Adv Mater 2011;23:4655e60. [26] Tang ZY, Wang Y, Podsiadlo P, Kotov NA. Biomedical applications of layer-bylayer assembly: from biomimetics to tissue engineering. Adv Mater 2006;18: 3203e24. [27] Hammond PT. Building biomedical materials layer-by-layer. Mater Today 2012;15:196e206. [28] Johnston APR, Cortez C, Angelatos AS, Caruso F. Layer-by-layer engineered capsules and their applications. Curr Opin Colloid 2006;11:203e9. [29] De Koker S, De Cock LJ, Rivera-Gil P, Parak WJ, Velty RA, Vervaet C, et al. Polymeric multilayer capsules delivering biotherapeutics. Adv Drug Deliv Rev 2011;63:748e61. [30] Cavalieri F, Postma A, Lee L, Caruso F. Assembly and functionalization of DNApolymer microcapsules. ACS Nano 2009;3:234e40. [31] Ng SL, Such GK, Johnston APR, Antequera-Garcia G, Caruso F. Controlled release of DNA from poly(vinylpyrrolidone) capsules using cleavable linkers. Biomaterials 2011;32:6277e84. [32] Becker AL, Orlotti NI, Folini M, Cavalieri F, Zelikin AN, Johnston APR, et al. Redox-active polymer microcapsules for the delivery of a survivin-specific siRNA in prostate cancer cells. ACS Nano 2011;5:1335e44. [33] Wang XQ, Wenk E, Matsumoto A, Meinel L, Li CM, Kaplan DL. Silk microspheres for encapsulation and controlled release. J Control Release 2007;117: 360e70. [34] Wang XY, Kim HJ, Xu P, Matsumoto A, Kaplan DL. Biomaterial coatings by stepwise deposition of silk fibroin. Langmuir 2005;21:11335e41.

7939

[35] Hsu CYM, Uludag H. A simple and rapid nonviral approach to efficiently transfect primary tissue-derived cells using polyethylenimine. Nat Protoc 2012;7:935e45. [36] Hu X, Kaplan D, Cebe P. Determining beta-sheet crystallinity in fibrous proteins by thermal analysis and infrared spectroscopy. Macromolecules 2006;39:6161e70. [37] Wang Y, Rudym DD, Walsh A, Abrahamsen L, Kim HJ, Kim HS, et al. In vivo degradation of three-dimensional silk fibroin scaffolds. Biomaterials 2008;29: 3415e28. [38] Krebs MD, Jeon O, Alsberg E. Localized and sustained delivery of silencing RNA from macroscopic biopolymer hydrogels. J Am Chem Soc 2009;131:9204e6. [39] Vinas-Castells R, Holladay C, di Luca A, Diaz VM, Pandit A. Snail1 downregulation using small interfering RNA complexes delivered through collagen scaffolds. Bioconjug Chem 2009;20:2262e9. [40] Rujitanaroj PO, Wang YC, Wang J, Chew SY. Nanofiber-mediated controlled release of siRNA complexes for long term gene-silencing applications. Biomaterials 2011;32:5915e23. [41] Meinel L, Hofmann S, Karageorgiou V, Kirker-Head C, McCool J, Gronowicz G, et al. The inflammatory responses to silk films in vitro and in vivo. Biomaterials 2005;26:147e55. [42] Numata K, Kaplan DL. Silk-based gene carriers with cell membrane destabilizing peptides. Biomacromolecules 2010;11:3189e95. [43] Nurnata K, Reagan MR, Goldstein RH, Rosenblatt M, Kaplan DL. Spider silkbased gene carriers for tumor cell-specific delivery. Bioconjug Chem 2011;22:1605e10. [44] Numata K, Mieszawska-Czajkowska AJ, Kvenvold LA, Kaplan DL. Silk-based nanocomplexes with tumor-homing peptides for tumor-specific gene delivery. Macromol Biosci 2012;12:75e82. [45] Gustafson JA, Price RA, Greish K, Cappello J, Ghandehari H. Silk-elastin-like hydrogel improves the safety of adenovirus-mediated gene-directed enzymeprodrug therapy. Mol Pharm 2010;7:1050e6. [46] Megeed Z, Haider M, Li DQ, O'Malley BW, Cappello J, Ghandehari H. In vitro and in vivo evaluation of recombinant silk-elastin like hydrogels for cancer gene therapy. J Control Release 2004;94:433e45. [47] Jiang CY, Wang XY, Gunawidjaja R, Lin YH, Gupta MK, Kaplan DL, et al. Mechanical properties of robust ultrathin silk fibroin films. Adv Funct Mater 2007;17:2229e37. [48] Hu X, Shmelev K, Sun L, Gil ES, Park SH, Cebe P, et al. Regulation of silk material structure by temperature-controlled water vapor annealing. Biomacromolecules 2011;12:1686e96. [49] Lv Q, Cao CB, Zhang Y, Ma XL, Zhu HS. Preparation of insoluble fibroin films without methanol treatment. J Appl Polym Sci 2005;96:2168e73. [50] Putthanarat S, Zarkoob S, Magoshi J, Chen JA, Eby RK, Stone M, et al. Effect of processing temperature on the morphology of silk membranes. Polymer 2002;43:3405e13. [51] Jin HJ, Kaplan DL. Mechanism of silk processing in insects and spiders. Nature 2003;424:1057e61. [52] Vergaro V, Scarlino F, Bellomo C, Rinaldi R, Vergara D, Maffia M, et al. Drugloaded polyelectrolyte microcapsules for sustained targeting of cancer cells. Adv Drug Deliv Rev 2011;63:847e63. [53] Kakade S, Manickam DS, Handa H, Mao GZ, Oupicky D. Transfection activity of layer-by-layer plasmid DNA/poly(ethylenimine) films deposited on PLGA microparticles. Int J Pharm 2009;365:44e52. [54] Wang ZJ, Qian L, Wang XL, Yang F, Yang XR. Construction of hollow DNA/PLL microcapsule as a dual carrier for controlled delivery of DNA and drug. Colloid Surf A 2008;326:29e36. [55] Jewell CM, Zhang JT, Fredin NJ, Lynn DM. Multilayered polyelectrolyte films promote the direct and localized delivery of DNA to cells. J Control Release 2005;106:214e23. [56] Zelikin AN, Becker AL, Johnston APR, Wark KL, Turatti F, Caruso F. A general approach for DNA encapsulation in degradable polymer microcapsules. ACS Nano 2007;1:63e9. [57] Jewell CM, Lynn DM. Multilayered polyelectrolyte assemblies as platforms for the delivery of DNA and other nucleic acid-based therapeutics. Adv Drug Deliv Rev 2008;60:979e99. [58] She Z, Antipina MN, Li J, Sukhorukov GB. Mechanism of protein release from polyelectrolyte multilayer microcapsules. Biomacromolecules 2010;11: 1241e7. [59] Lu Q, Zhang B, Li M, Zuo B, Kaplan DL, Huang Y, et al. Degradation mechanism and control of silk fibroin. Biomacromolecules 2011;12:1080e6. [60] Lv HT, Zhang SB, Wang B, Cui SH, Yan J. Toxicity of cationic lipids and cationic polymers in gene delivery. J Control Release 2006;114:100e9. [61] Dash BC, Mahor S, Carroll O, Mathew A, Wang W, Woodhouse KA, et al. Tunable elastin-like polypeptide hollow sphere as a high payload and controlled delivery gene depot. J Control Release 2011;152:382e92. [62] Browne S, Fontana G, Rodriguez BJ, Pandit A. A protective extracellular matrix-based gene delivery reservoir fabricated by electrostatic charge manipulation. Mol Pharm 2012;9:3099e106.