Silk Fibroin-Based Complex Particles with Bioactive Encrustation for

Nov 13, 2013 - ABSTRACT: Application of bone morphogenetic protein 2. (BMP-2) currently faces its challenges, and its efficacy of delivery has to be i...
1 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/Biomac

Silk Fibroin-Based Complex Particles with Bioactive Encrustation for Bone Morphogenetic Protein 2 Delivery Pujiang Shi,†,‡,§ Sunny A. Abbah,‡,§,# Kushagra Saran,† Yong Zhang,† Jun Li,† Hee-Kit Wong,‡,§ and James C. H. Goh*,†,‡,§,∥ †

Department of Biomedical Engineering, National University of Singapore, Singapore 117575 NUS Tissue Engineering Programme, National University of Singapore, Singapore 117510 § Department of Orthopedic Surgery, National University of Singapore, Singapore 119028 ∥ Life Sciences Institute, National University of Singapore, Singapore 117456 # Network of Excellence for Functional Biomaterials (NFB), National University of Ireland, Galway, IDA Business Park, Dangan, Galway, Ireland ‡

S Supporting Information *

ABSTRACT: Application of bone morphogenetic protein 2 (BMP-2) currently faces its challenges, and its efficacy of delivery has to be improved. The proper dosage of the powerful bioactive molecule is still under discussion and needs to be investigated further. In this work, pure silk fibroin particles and particles with calcium carbonate encrustation (complex particles) are designed, developed, and functionalized by BMP-2. These are used to deliver the bioactive molecule to mesenchymal stem cells (MSCs) to induce osteogenic differentiation. Results are compared with those of control groups of BMP-2 carriers under the same condition. Silk fibroin-based particles with size and component variations are prepared by self-assembly, desolvation, and soft template formation to improve BMP-2 loading efficiency. Results show that the particles significantly enhance osteogenic differentiation of MSCs, which is evident in the high ALP enzyme activity as well as the increased level of expression of osteogenic genes. Specifically, the combination of calcium compound and BMP-2 in the silk fibroin−calcium carbonate complex particles synergistically enhances osteogenesis. Release tests and mathematical modeling are applied to describe BMP-2 dissolution profiles, and the release mechanism is based on diffusion and polymer chain relaxation. In summary, the particles show high efficacies of BMP-2 delivery, and introduction of the complex particle can progressively enhance osteogenesis.

■

INTRODUCTION Silk fibroin is a natural material with excellent biocompatibility and biodegradability.1 From this perspective, silk fibroin can be considered to have an outstanding potential as a carrier for controlled release of cargo. In particular, the slow biodegradability of silk fibroin particles could be tailored to prolong release.1 Silk fibroin has been previously evaluated for loading and delivery of bioactive molecules.2 Several methods of preperation, including lipid template fabrication, spray drying, self-assembly, and oil emulsion, have been used to fabricate silk fibroin particles.3 Some of these methods have been used to manufacture the functionalized particle-loaded bioactive molecule and shown to enhance tissue regeneration.4 Specifically, bone morphogenetic protein 2 (BMP-2)-loaded silk particles have been utilized to induce ectopic bone formation in Sprague-Dawley rats, with bone density increasing in parallel with amplification of BMP-2 dosage.5 Moreover, BMP-2-loaded silk fibroin particles show superior osteoinduction compared with that of BMP-2 alone.2 This underlines the importance of this carrier in increasing the efficacy of the © 2013 American Chemical Society

loaded BMP-2. Although it has been recognized that the osteoinductivity of BMP-2 is powerful, the delivery and bioactivity of BMP-2 need to be controlled to minimize the risk of complications, including peri-implant tissue swellings, nerve damage, immunological responses, and even cancer.6 Therefore, it is important to appropriately restrain BMP-2 in its carrier to minimize its diffusion and burst release and optimize the dosage. This implies that the carrier must provide stable fixation to the bioactive molecule during applications. From a therapeutic efficacy perspective, an optimal dose of BMP-2 that would ensure consistent fusion with minimal complications has not been described. Although a larger dose of BMP-2 has been shown to accelerate bone regeneration, excessive inflammation, premature consolidation of the newly formed bone tissue, and structural abnormalities have been reported.7 A number of studies have indicated that smaller Received: September 16, 2013 Revised: November 6, 2013 Published: November 13, 2013 4465

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

doses may yet facilitate healing of hard tissue defects.8 Hence, controlling the release of osteoinductive BMP-2 from carriers remains an important goal, in order to coordinate the release profile with the rate of bone tissue regeneration. Calcium compounds have been widely applied in bone regeneration as artificial bone grafts, and the compound has also been used to improve the osteoconductivity of orthopedic devices.9 Specifically, recent studies have demonstrated the importance of calcium carbonate as a bone filling material because of its biocompatibility and osteoconductivity.10 The bioactivity of this material is based on its ability to generate calcium phosphate by reaction with phosphates.11 The transformative process of calcium carbonate to calcium phosphates may enhance interactions and coordinate with surrounding tissues to promote neo-tissue regeneration. Moreover, calcium carbonate shows outstanding properties for the formation of organized structures on removable templates.12 Although silk fibroin particles are excellent carriers of BMP-2, the influences of preparative methods, particle size, and components of the particles on the efficacy of BMP-2 loading and delivery are yet to be fully elucidated. In this work, representative methods, including oil emulsion, self-assembly, desolvation, and template-guided self-assembly, were employed to manufacture BMP-2-loaded silk fibroin-based particles. Material characterization and in vitro tests were subsequently performed to evaluate the particles. The silk fibroin−calcium carbonate complex particles that can be considered as efficient carriers of the bioactive molecule and building blocks of the organic−inorganic composite scaffold have been invented and reported here. The complex particles retain the potential for use in diverse bone regeneration applications, such as bone tunnel regeneration and spinal fusion.

■

PVA [2 mL, 2% (w/v)], and vortexed for 10 s. The mixture was frozen at −25 °C for 24 h. This resulted in BMP-2-loaded silk particles. All the supernatants were preserved for testing and calculation of loading efficiencies (eq a). Fabrication of BMP-2-Loaded Silk Fibroin−Calcium Carbonate Complex Particles. LSB and SDBS were mixed in an equimolar ratio to form a surfactant cocktail. Silk fibroin nanoparticles were slowly added to the surfactant cocktail under mild agitation. The mixture was stirred for 0.5 h before the addition of calcium chloride and sodium carbonate solutions to initiate the reaction. Finally, the silk fibroin−calcium carbonate complex particles were formed. The BMP-2 solution (10 μL, 1.5 μg/μL) was added to 1 mL of double distilled water comprising 1.5 mg of complex particles. The mixture was mildly agitated at 4 °C for 24 h. Subsequently, the particles were collected by centrifugation and supernatants were analyzed to calculate the loading efficiency. Fabrication of Plain Calcium Carbonate and Plain Silk Particles. The plain CaCO3 particles were manufactured according to the protocol of complex particle preparation without the addition of silk fibroin nanoparticles. The plain silk fibroin nanoparticles (Grp #3) were also prepared for future application. Preparation and Loading of FITCBMP-2. Fluorescein-5(6)carboxamidocaproic acid N-succinimidyl ester (10 μg, FITC) was added to 1 mL of the BMP-2 (1.5 mg/mL, clinical grade; Medtronic) solution. The mixture was held at 4 °C for 24 h before dialysis. Finally, FITC-labeled BMP-2 (FITCBMP-2) was collected through the concentrator (10K, Pierce) and reconstituted to 1 mL. FITCBMP-2 was loaded into the particles using the protocols mentioned above. Electron Microscopy. Silk fibroin particles fabricated by oil emulsion, self-assembly, and desolvation were directly added on top of conductive tapes mounted on sample stubs separately. Meanwhile, the original complex particles and the complex particles in phosphatebuffered saline (PBS) were dried and mounted on sample stubs. The samples were sputtered with platinum and observed with a scanning electron microscope (Philips XL-30 FEGSEM). The original silk fibroin nanoparticles and complex particles were diluted in distilled water, and then one drop (approximately 10 μL) was placed on the copper mesh grid and dried before being observed via transmission electron microscopy (TEM) (JEOL JEM-1010). Fourier Transform Infrared Spectroscopy (FTIR). The original and transformed complex samples were washed with double distilled water several times. Then the samples were freeze-dried and analyzed by FTIR (IR Prestige, Shimadzu, Japan). Particle Analysis. The particle size and ζ potential of all the samples were tested with Zetasizer Nano ZS instrument (Malvern Instruments). Loading Efficiency and Release Study. The BMP-2-loaded samples (1.5 mg) were reconstituted with 1 mL of PBS. The supernatant was periodically aspired and replaced with fresh PBS. The amount of BMP-2 in the supernatants was quantified and recorded to plot the release profiles. The BMP-2 loading efficiency and release were quantified by an enzyme-linked immunosorbet assay (Quantikine, R&D Systems). The loading efficiency and accumulated release were calculated according to the following formulas:

MATERIALS AND METHODS

Materials. Bombyx mori silk (Silk Innovation Center, Maha Sarakham, Thailand) was degummed in a 0.02 M NaHCO3 solution at 90 °C for 1.5 h to completely remove sericin. Then the silk solution was obtained by dissolving sericin-free silk fibers in a CaCl2/ CH3CH2OH/H2O ternary solvent system (1:2:8 molar ratio) followed by dialysis. Finally, a 6% (w/v) silk solution was prepared and stored at 4 °C for further applications. Sodium carbonate, calcium chloride, ethanol, polyvinyl alcohol (PVA, 85000−124000 Da), sodium dodecylbenzenesulfonate (SDBS), and lauryl sulfobetaine (LSB) were procured from Sigma-Aldrich. Clinical grade BMP-2 (INFUSE Bone Graft) was purchased from Medtronic Sofamore Danek Inc. (Minneapolis, MN). Fabrication of BMP-2-Loaded Silk Particles. Silk particles were manufactured by several methods, including oil emulsion, selfassembly, and desolvation. During oil emulsion, the silk solution (2%, 1 mL) was added dropwise into soybean oil (10 mL) containing 1% Tween 20. The whole emulsion was agitated at 500 rpm for 30 min, and then ethanol (500 μL) was added to the emulsion. The final mixture was agitated overnight to allow solidification of the silk particles. Subsequently, the silk fibroin particles were thoroughly washed with toluene, ethanol, and water before lyophilization. Silk fibroin particles with diameters of ∼250 nm were fabricated by desolvation. Briefly, the silk fibroin solution (2%, 1 mL) was slowly added to isopropyl alcohol (100 mL) to form the silk nanoparticles, and then these particles were washed with water several times before lyophilization. The silk fibroin particles (1.5 mg) made by oil emulsion (Grp #1) and desolvation (Grp #3) were individually incubated with the BMP-2 solution (1 mL, 15 μg/mL). Finally, the particles were centrifuged, washed with water, and stored in a freezer. To fabricate self-assembled BMP-2-loaded silk fibroin particles (Grp #2), the silk fibroin solution (100 μL, 15 μg/μL) was mixed with a concentrated BMP-2 solution (10 μL, 1.5 μg/μL), mixed with ethanol (40 μL) and

BMP‐2 loading efficiency (w/w) = [(actual amount of BMP‐2 in the particles)/(initial loading)] (a)

× 100

accumulated release (%) = [(total amount of released BMP‐2) /(actual loading of BMP‐2)] × 100

(b)

Mathematical Modeling. The release profiles were examined by the following models: zero-order (eq 1), first-order (eq 2), Higuchi (eq 3), Korsmeyer−Peppas (eq 4), and Weibull (eq 5).

F = K 0t 4466

(1) dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

Figure 1. Electron micrograph of silk fibroin-based particles. (a−d) Scanning electron microscopy images of silk fibroin particles made by oil emulsion (a, Grp #1) and self-assembly (b, Grp #2) and silk fibroin particles fabricated by desolvation (c, large, Grp #3). TEM images of silk fibroin particles fabricated by desolvation (c, small, Grp #3) and silk fibroin−calcium carbonate complex particles (d, Grp #4). The yellow arrows point to a silk nanoparticle. Scale bars of 2 μm [a, b, and c (large)], 50 nm [c (small)], and 0.5 μm (d).

log Q t = log Q 0 +

Kt 2.303

F = KH t

(3)

Mt = Kt n M∞

(4)

⎡ − (t − t )b ⎤ Mt lag ⎥ = 1 − exp⎢ ⎥⎦ ⎢⎣ M∞ tscale

(5)

R adjusted 2 = 1 −

n−1 (1 − R2) n−p

cells were detached from the flask with 0.05% trypsin and subcultured when the cells reached 70−80% confluence. MSCs at passage 3 were used in all in vitro experiments. Three-dimensional (3D) models were used to evaluate the BMP-2 delivery efficacies of each type of carrier. Briefly, the plain particles and particle aggregates loaded with 300 ng of BMP-2 were centrifuged to the bottom of 15 mL tubes, and then 1 mL of medium containing 0.5 million cells was gently added to the tubes, which were centrifuged again. Finally, the MSCs were seeded on all the groups of particles, and they were cultured in osteogenic medium that consisted of High Glucose Dulbecco’s modified Eagle’s medium (HDMEM) (Invitrogen), 10% FBS, and 1% penicillin/ streptomycin, supplemented with L-ascorbic acid (50 mM) and βglycerophosphate (10 mM). Meanwhile, cell pellets were also cultured in osteogenic medium (containing 300 ng/mL soluble BMP-2) as a control. To test the biocompatibility of the BMP-2-loaded particles, an Alamar Blue (Invitrogen) assay (n = 6) was performed. The test medium (1 mL) consisted of Alamar Blue (10%) and FBS (5%); the medium was incubated with all samples for 2 h, and the absorbance of Alamar Blue was measured according to the vendor’s protocol. For testing alkaline phosphatase activity (ALP), the cell lysate was analyzed by using p-nitrophenyl phosphate disodium salt. For gene expression, total RNA was extracted by using Trizol reagent and purified by using the RNeasy Mini Kit (Qiagen, Valencia, CA). The RNA concentration was determined by using Nano Drop, and complementary DNA synthesis was conducted by using total RNA (500 ng) according to the manufacturer’s protocol for iScript cDNA (Bio-Rad, Hercules, CA). Real-time polymerase chain reaction was conducted using cDNA (4 μL), SYBR Green Supermix (12.5 μL), and the optimized concentration of primers, with the volume increased to 25 μL with nuclease-free water. All reactions were performed on the real-time IQ5 (Bio-Rad). The thermal cycling program for all polymerase chain reactions was as follows: 95 °C for 3 min, followed by 45 cycles of amplification, which consisted of a denaturation step at 95 °C for 10 s, an annealing step at 58 °C for 30 s, and an extension step at 72 °C for 30 s. The genes analyzed were Runt-related transcription factor 2 (RUNX2) and osteocalcin for osteogenesis (Table S1 of the Supporting Information). The level of expression of the target gene, normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Table S1 of the Supporting Information), was then calculated using the 2−ΔΔCt formula with reference to the respective control group.2 Further, the pellets were fixed in 10% formalin and embedded in paraffin. Sections (10 μm) were prepared and stained with Alizarin

(2)

(6)

The zero-order model describes the system in which the drug release rate was independent of the concentration of the dissolved solute.13 The first-order formula was used to describe the drug release rate depending on its concentration.13 The Higuchi formula was used to describe the drug release rate based on the diffusion process following Fick’s law, and this model could be used to describe the release of a drug from several types of pharmaceutical forms. Further, F represented the portion of drugs released over time t, and KH was the Higuchi dissolution constant.14 The Korsmeyer−Peppas model was applied to describe the profile of the release of a drug from swelling-controlled systems, where Mt/M∞ is the fractional release at time t, k is a constant, and n is the diffusional exponent.14 In the Weibull model, tlag represents the delay of drug release, tscale is the duration of the release process, and b indicates the shapes of release curves (briefly, b values greater than, equal to, and less than 1 indicate S-shaped, exponential, and parabolic curves, respectively).14 Further, Radjusted2 was applied here to verify the accuracy of each model (eq 6), where p is the number of parameters and R2 is the coefficient of determination.1 Cell Isolation, Expansion, and Differentiation. MSCs were isolated according to our previously published protocol.15 Briefly, 10 mL of bone marrow was aspirated from the posterior iliac crest of New Zealand white rabbits and suspended in 20 mL of Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) (HyClone, Logan, UT), Lglutamine (580 mg/L), and penicillin/streptomycin (100 units/mL). Cultures were incubated at 37 °C and in 5% CO2. After 72 h, nonadherent cells were removed by changing the medium. Adherent 4467

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

Table 1. BMP-2 Loading Efficiencies, Particle Sizes, Polydispersities, and ζ Potentials of Silk Fibroin-Based Particles Fabricated by Various Methods sample no. Grp Grp Grp Grp

#1 #2 #3 #4

sample name

BMP-2 loading efficiency (%)

silk fibroin particles (oil emulsion) silk fibroin particles (self-assembly) silk fibroin particles (desolvation) complex particles

0 95.5 ± 0.6 89.3 ± 0.1 92.1 ± 0.1

particle size (nm) 2406.67 1101.5 254.3 2199

± ± ± ±

428.09 51.6 47.5 105

polydispersity index (PDI) 0.69 0.07 0.17 0.26

± ± ± ±

0.47 0.04 0.07 0.03

ζ potential (mV) −17.4 −3.2 −8.8 −17.2

± ± ± ±

1.4 0.3 0.7 0.5

Figure 2. Fluorescent images and controlled release of BMP-2 from the BMP-2-loaded particles. (a−c) The self-assembled silk particles are well dispersed (a, Grp #2); nanoparticles are aggregated (b, Grp #3), and the complex particles are similar to capsule structures (c, Grp #4). (d) Slow release of BMP-2 from the three groups of particles in PBS. Statistical Analysis. All data are expressed as means ± the standard deviation. Statistical analysis was performed by pairwise comparison of experimental categories using a two-tailed, unpaired Student’s t test and multiple comparisons using single-factor analysis of variance (ANOVA) and post hoc Tukey tests, using SPSS Statistics version 22.0, and p < 0.05 was considered statistically significant.

Red S. The results were evaluated by three individuals who were blinded to treatments. Fluorescent Microscopic Images. The FITCBMP-2-loaded particles were washed three times using double distilled water, collected and put on a glass slide for further observation, to investigate the effects of the 3D model on the morphology of MSCs. MSCs were seeded on the particle aggregates and cultured in osteogenic medium. At week 1, each sample was fixed with formaldehyde (4%, 1 mL) for 30 min, permeabilized with 0.1% Triton X-100 in PBS for 3 min at room temperature, and washed three times with PBS. The samples were stained with DAPI and Alexa Fluor 594 Phalloidin (Invitrogen) for 30 min and then washed three times with PBS. Images were taken under a Carl Zeiss apotome microscope (Carl Zeiss, Jena, Germany). Silk Degradation. Grp #2 and Grp #3 silk particles (50 mg each) were incubated in PBS and PBS containing 1.5 mg/mL protease XIV (Sigma-Aldrich). The samples were kept in a water bath at 37 °C, and the supernatants with and without protease XIV was replenished every 3 days. At each time point, the samples were washed and centrifuged, and then the precipitate was dried and weighed. The weight losses of the samples were calculated according to the formula

■

RESULTS AND DISCUSSION The quality of silk fibroin particles prepared by oil emulsion is erratic. The stirring speed, silk fibroin concentration, and solvent alter the quality of the particles (Figure 1a). The oil emulsion procedures are aggressive, normally requiring the addition of organic solvents, heating, and high-speed agitation. Further, the particles are washed using other solvents that could be toxic;16 subsequently, eliminating the oil/organic phase is challenging. Therefore, the bioactivity of molecules could be lost in these procedures. Because the risk of residual oil phase/ organic solvent inside the particles cannot be fully eliminated, the loading efficiencies of desired cargo may be compromised. The loading efficiency of BMP-2 is very low in Grp #1 (Table 1), which can be attributed to the possible existence of organic

weight loss (%) = {[50 mg − remnant of silk particles (mg)] /(50 mg)} × 100 4468

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

Figure 3. Scanning electron microscopy image and FTIR analysis of original complex particles (a and c1) and calcium phosphate spikes on the surface of the complex particles after incubation in PBS for 24 h (b and c2). Scale bars of (a) 2 and (b) 5 μm.

residues that could hamper protein absorption. In Grp #2, the particles are produced at a low temperature, and silk fibroin self-assembles to form particles. These mild fabrication conditions are beneficial with respect to the preservation of BMP-2 bioactivity.2 BMP-2 and silk fibroin solutions are mixed gently, followed by addition of ethanol to induce formation of a β-sheet structure of silk fibroin (solid state). BMP-2 loading of the silk fibroin particles is also achieved using PVA that prevents particle aggregation. The self-assembled particles are of fairly uniform size distribution (Figure 1b), and the loading of BMP-2 is visualized on the fluorescent image (Figure 2a). The loading efficiency is observed to be 95.5%. This is presumably due to the fact that BMP-2 is mixed with the silk fibroin solution during the first stage of particle fabrication, incorporating large amounts of BMP-2 into the silk fibroin matrix, leading to a high loading efficiency. In Grp #3, particles are formed as the solvent is exchanged immediately after mixing. The particles made by this method are approximately 250 nm in size (Table 1). The handling of particles in Grp #3 is relatively easy, and the solvent employed is 2-propanol, which evaporates relatively easily. After lyophilization, the particles are incubated with a BMP-2 solution, leading to a modest loading efficiency [89.3% (Table 1)]. The nanoparticles have a high surface:volume ratio,15 which is good for BMP-2 absorption. In the last group (Grp #4), calcium carbonate has been shown to be effective in enhancing osteogenesis, and it is a precursor of calcium phosphates, including biphasic calcium phosphate, tricalcium phosphate, and hydroxyapatite.9 The particle enhances the efficacy of BMP-2 delivery, while the calcium carbonate encrustation provides osteoconductive influence to synergistically enhance osteogenesis. As shown in Figure 1d,

the silk fibroin nanoparticles can be observed around the porous calcium carbonate encrustation. This could account for the slight increase in BMP-2 loading efficiency to 92.1% compared with that of plain silk nanoparticles as the calcium carbonate could absorb extra BMP-2. The absorption of BMP-2 is confirmed by fluorescent images (Figure 2b,c). The complex particles are capsule-like in structure; hence, this fabrication protocol can be adopted for the assembly of calcium carbonate capsules and/or templates. The particles in Grp #1 cannot be used for protein release analysis because of the low loading efficiency (Table 1). The samples in the other three groups are subjected to release analyses conducted in PBS. As shown in Figure 2d, BMP-2 is slowly released from sample particles in all groups. Specifically, 0.85 are considered evidence of super case II transport.1,14 In Grp #2 and Grp #3, the values of n are 0.623 and 0.74, respectively, indicating the release kinetics is anomalous release. According to the data of weight loss and release kinetics of the silk particles, diffusion of BMP-2 and degradation of silk fibroin contribute to release of BMP-2. A higher BMP-2 loading efficiency is observed in self-assembled silk fibroin particles than that in silk nanoparticles (Table 1). Although BMP-2 release profiles of the two groups are different (Figure 2d, Grp #2 and Grp #3), they have similar release mechanisms. Interestingly, the release mechanism of the complex particle (Grp #4) is Fickian release (n = 0.39). The complex particles have hierarchical structure and different components. We confirmed in this study that BMP-2 is tethered on the components. In line with the previous description, the Fickian release of BMP-2 from the complex particles may be attributable to the slow release of BMP-2 from calcium carbonate during material transformation. However, further release of BMP-2 from plain silk fibroin nanoparticles may be restricted by the calcium compounds. In the Higuchi model, the KH (Higuchi dissolution constant) of each group fits its own release profile. The self-assembled particles have the highest Higuchi dissolution constant, in accord with highest accumulated release. Meanwhile, the KH of Grp #4 is higher than the value in Grp #3, which is probably powered by the extra BMP-2 on the encrustation. Nevertheless, it should be noted that the zero-order and first-order models are employed

rate of burst release of BMP-2 from complex particles compared to that from plain nanoparticles may be caused by the faster release of BMP-2 from the inorganic components of the complex particles during material transformation. Subsequently, the rate of release of BMP-2 may be decreased by the surrounding inorganic compounds that may obstruct the release of BMP-2 from loaded silk fibroin plain nanoparticles. As a result, the release profiles of both groups begin to approach each other. Figure 4 indicates that weight loss of silk particles in PBS containing protease XIV is greater than that in PBS. The plain nanoparticles have a relatively higher ratio of surface area to volume, which drives the rapid weight loss under both conditions. According to the work of Kaplan et al.,17 the mechanism of silk fibroin weight loss may be caused by degradation of hydrophilic parts of silk, which is followed by dissolution of the hydrophobic remnant of the silk. Hence, the plain nanoparticles that have a larger surface area are both degraded and dissolved, resulting in a more rapid weight loss in PBS containing protease XIV. Meanwhile, the rate of weight loss of the particle without the influence of enzyme may be induced predominantly by dissolution of the silk fibroin molecule.17 The release profiles are further analyzed by zero-order, firstorder, Higuchi, Korsmeyer−Peppas, and Weibull equation models (Table 2). The release data of BMP-2 can be best explained by the Weibull model (Radjusted2 > 0.99), followed by the Korsmeyer−Peppas model (Radjusted2 varying between 0.97 and 0.99). The explanations using other mathematical models such as the Higuchi model (Radjusted2 varying between 0.93 and 0.97), the first-order model (Radjusted2 varying between 0.20 and 0.94), and the zero-order model (Radjusted2 varying between 0.18 and 0.94) could be used in that order. As indicated by the Weibull model (b < 1), all the release patterns are parabolic 4470

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

fibroin and calcium carbonate) in this study are not known to be cytotoxic and have been successfully applied in tissue engineering applications.15,18 The particles show good biocompatibility in vitro and are subsequently used in cellular differentiation. MSCs are cultured alone and seeded on the particle aggregation with and without BMP-2. On the basis of a previous report, 300 ng/mL BMP-2 is the effective dosage for inducing osteogenic differentiation, and silk alone will not stimulate osteogenic differentiation.2 In this study, we evaluated the level of alkaline phosphatase (ALP) activity, an early marker of osteogenesis.19 In Figure 6a, the ALP activities of cells on BMP-2-loaded particle groups and the plain calcium carbonate particle group are significantly higher than those on the other groups. This demonstrates the osteoinductive potentials of the BMP-2 and calcium carbonate particles. The ALP activity of MSCs cultured on plain silk particles is relatively low. Besides, ALP is also known as an indicator of the maturity of mineralized tissue, as the enzyme increases the local concentration of mineralization promoters and restrains the concentration of inhibitors of mineralization.19 The enzyme is located in the membrane of cells, and its higher activity in particle groups is presumably caused by the proximity of the particles to the cellular membrane providing continuous stimulation. Simultaneously, the particles can protect the BMP2 from degradation and internalization.2 As long as there is BMP-2 in the medium, the ALP activity of MSCs is significantly increased (Figure S2 of the Supporting Information). Furthermore, the relative gene expressions of osteocalcin and

here to explain the release profiles based on ideal conditions, which is indeed inappropriate for analysis of the particles (the values of Radjusted2 are in disparity). Alamar blue reduction is used to evaluate cell viability (higher reduction corresponding to higher cell viability). For all groups, reduction is around 50% on day 1 (Figure 5). There are

Figure 5. Cell viability of MSCs. Cells were cultured with soluble BMP-2 (control), BMP-2-loaded self-assembled silk fibroin particles (Grp #2), BMP-2-loaded silk fibroin nanoparticles (Grp #3), and BMP-2-loaded complex particles (Grp #4).

not significant differences among the four groups. The materials employed in the fabrication of the complex particles (i.e., silk

Figure 6. Expressions of osteogenic markers of MSCs. (a) ALP activity of MSCs cultured with soluble BMP-2, BMP-2-loaded particles, calcium carbonate, and plain silk particles. (b and c) Relative gene expressions of osteocalcin (b) and RUNX2 (c) after week 1 (*p < 0.05). 4471

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

RUNX2 are clearly upregulated in complex particles. Gene expressions of RUNX2 and osteocalcin are critical in early stages of bone regeneration,20 and plain silk particles have no effect on the expression of either genes (Figure 6b,c). The expressions of osteocalcin in Grp #4 and plain calcium carbonate particles are higher than that of the other groups in week 1 (Figure 6b). Osteocalcin is a noncollagenous protein secreted by osteoblasts,21 and the protein mainly binds to and remodels calcium phosphate compounds.22 Calcium carbonate is bioactive because it can transform itself into calcium phosphates in osteogenic medium. As the cells come in contact with the BMP-2 releasing complex particles, osteogenesis is induced. Meanwhile, the particles spontaneously start to transform to synergistically stimulate osteogenesis (Figures 3c, 6b, and 8d and Figure S1 of the Supporting Information). Moreover, RUNX2 (Figure 6c) is linked to osteoblast proliferation and differentiation, and it regulates skeletal genes and exists and becomes stronger from stem cell to osteocyte.23 Remarkably, RUNX2 expressions in the three BMP-2-loaded particle groups are significantly higher than that of the other groups (p < 0.05). Although calcium carbonate can enhance osteocalcin expression, it has a limited influence on RUNX2 expression, and only the BMP-2-loaded particles can stimulate RUNX2 expressions. Hence, the BMP-2-loaded complex particles are suitable for bone regeneration, taking advantage of osteoinductive BMP-2, biocompatibility of silk fibroin, and bioactivity of calcium carbonate. With respect to the release profiles and mechanism, the significantly slower release of Grp #3 and Grp #4 will have a significant effect in enhancing osteogenesis, which is shown by data obtained in vitro. Further, saturation of BMP receptors is also vital for osteogenic differentiation. The BMP-2-loaded particles on the cell surface may help to activate BMP receptors, including BMPR-IA, BMPR-IB, and BMPR-II, and therefore stabilize the activity of the whole assembled complex to initiate osteogenic differentiation.2 In Grp #4, particles with hierarchical structure consisted of biodegradable polymer and bioactive inorganic components have the potential to enhance osteogenesis through biochemical and physical stimulations. In Figure 7, the strong staining in Grp #4 is probably caused by the calcium component in the complex particles. Moreover, it could be evident that the particles also provided calcium as a source to enhance osteogenic differentiation of cells in addition to controlling the release of BMP-2. Importantly, we observed the growth of calcium phosphate spikes on silk−calcium carbonate complex particles after incubation in PBS (Figure 3). The calcium phosphate spikes may be internalized by the cells, which could lead to a high concentration of cytoplasmic calcium, enhancing osteogenic differentiation.24 Further, the relative gene expression of osteocalcin is higher in the presence of calcium carbonate (Figure 6b); hence, the complex particles are good for osteogenesis. Of similar importance is the observation from F-actin staining, used to investigate cellular responses of MSCs on particle aggregation. In Figure 8a, the MSCs grow normally on the two-dimensional plate, and their F-actin stress fibers are well organized and polarized. The Factin stress fibers and lamellipodium extend in the direction of cell migration, and cell mobility is not compromised. The morphologies of cells are different in silk fibroin micro- and nanoparticle groups (Figure 8b,c). The F-actin stress fibers are scattered in all directions and stabilized. This phenomenon probably indicates that the cell motility has been prevented;25 hence, silk fibroin nano- and microparticle stimulations of cells

Figure 7. Alizarin Red staining of MSCs. Cells were cultured with soluble BMP-2 (a, control), BMP-2-loaded self-assembled silk particles (b, Grp #2), BMP-2-loaded silk particles made by desolvation (c, Grp #3), and BMP-2-loaded complex particles (d, Grp #4) after week 1. Scale bar of 50 μm.

Figure 8. F-Actin staining of MSCs. (a−d) Cells are growing on a twodimensional plate (a), BMP-2-loaded silk fibroin microparticles (b, Grp #2), nanoparticles (c, Grp #3), and complex particles (d, Grp #4). Nuclei are counterstained by DAPI (blue): lamellipodia extending in the direction of cell migration (yellow arrows), white arrows indicating the direction of cellular movement (a), nonpolarized lamellar extensions around the entire cell periphery (yellow arrows, b and c), and formation of radial protrusions resembling lamellipodia that contain a dense meshwork of peripheral actin filaments and few stress fibers (yellow circle, d).

can be directly observed. In the last group [complex particles (Figure 8d)], cells are round, with the formation of radial protrusions resembling lamellipodia. Meanwhile, almost no Factin stress fibers can be observed in these cells. MSCs seeded on the complex particles have rarely stabilized F-actin stress fibers and lost almost all their cell polarity subsequently; cell migration may have been fully prevented.25 The cells respond differently to complex particles, the behavior of which is compared to the cellular behavior of soluble BMP-2 and BMP2-loaded silk fibroin nano- and microparticles separately. Generally, the lack of a polarized phenotype suggests that cell mobility has been prevented in all particle groups. The subtle cellular events cannot be observed in normal scanning electron 4472

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

Notes

microscopy images, in which deformation of cells is unavoidable because of the processing protocol and only a macroscopic view of cell−particle interaction can be observed (Figure S1 of the Supporting Information). Finally, this interaction might have influenced the osteogenic differentiation of MSCs and needs to be investigated further. The BMP-2-loaded particles are in persistent contact with MSCs, stimulating the cells irrespective of a change in medium, BMP-2 degradation, or internalization. The presence of the particles on the cellular membrane is believed to work with adhesion receptors of the cells, triggering profound osteogenic cascade effects, for example, BMP-2 receptor−integrin collaboration, cytoskeletal dynamics, and biochemical signaling pathways.2 Most importantly, the special bioactive complex material can achieve a higher efficacy of BMP-2 delivery and enhance osteogenic differentiation in material and biological aspects.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work is supported by the Biomedical Engineering Programme (BEP), A*STAR, and the National University of Singapore.

■

(1) Shi, P.; Goh, J. C. H. Int. J. Pharm. 2011, 420 (2), 282−289. (2) Shi, P.; Chen, K.; Goh, J. C. Adv. Healthcare Mater. 2013, 2 (7), 934−939. (3) Shi, P.; Goh, J. C. H. Powder Technol. 2012, 215−216, 85−90. (4) Pritchard, E. M.; Kaplan, D. L. Expert Opin. Drug Delivery 2011, 8 (6), 797−811. (5) Bessa, P. C.; Balmayor, E. R.; Hartinger, J.; Zanoni, G.; Dopler, D.; Meinl, A.; Banerjee, A.; Casal, M.; Redl, H.; Reis, R. L.; Van Griensven, M. Tissue Eng., Part C 2010, 16 (5), 937−945. (6) Boraiah, S.; Paul, O.; Hawkes, D.; Wickham, M.; Lorich, D. G. Clin. Orthop. Relat. Res. 2009, 467 (12), 3257−3262. (7) (a) Sailhan, F.; Gleyzolle, B.; Parot, R.; Guerini, H.; Viguier, E. Injury 2010, 41 (7), 680−686. (b) Zara, J. N.; Siu, R. K.; Zhang, X.; Shen, J.; Ngo, R.; Lee, M.; Li, W.; Chiang, M.; Chung, J.; Kwak, J.; Wu, B. M.; Ting, K.; Soo, C. Tissue Eng., Part A 2011, 17 (9−10), 1389− 1399. (8) (a) Abbah, S. A.; Lam, C. X.; Hutmacher, D. W.; Goh, J. C.; Wong, H. K. Biomaterials 2009, 30 (28), 5086−5093. (b) Wang, M.; Abbah, S. A.; Hu, T.; Toh, S. Y.; Lam, R. W.; Goh, J. C.; Wong, H. K. Spine 2013, 38 (17), 1452−1458. (c) Abbah, S. A.; Liu, J.; Goh, J. C.; Wong, H. K. Tissue Eng., Part A 2013, 19 (3−4), 350−359. (9) Bhatt, R. A.; Rozental, T. D. Hand Clin. 2012, 28 (4), 457−468. (10) Fujihara, K.; Kotaki, M.; Ramakrishna, S. Biomaterials 2005, 26 (19), 4139−4147. (11) Ma, M. Y.; Zhu, Y. J.; Li, L.; Cao, S. W. J. Mater. Chem. 2008, 18 (23), 2722−2727. (12) Wu, G. X.; Ding, J.; Xue, J. M. J. Mater. Res. 2008, 23 (1), 140− 149. (13) Liu, G.; Zhou, J. Macromolecules 2003, 36 (14), 5279−5284. (14) Costa, P.; Sousa Lobo, J. M. Eur. J. Pharm. Sci. 2001, 13 (2), 123−133. (15) Shi, P.; Teh, T. K.; Toh, S. L.; Goh, J. C. Biomaterials 2013, 34 (24), 5947−5957. (16) (a) Baimark, Y. Asia-Pac. J. Chem. Eng. 2012, 7 (Suppl. 1), S112−S117. (b) Baimark, Y.; Srihanam, P. J. Appl. Sci. 2009, 9 (21), 3876−3881. (c) Baimark, Y.; Srihanam, P.; Srisuwan, Y.; Phinyocheep, P. J. Appl. Sci. 2010, 118 (2), 1127−1133. (d) Imsombut, T.; Srisuwan, Y.; Srihanam, P.; Baimark, Y. Powder Technol. 2010, 203 (3), 603−608. (17) Lu, Q.; Zhang, B.; Li, M.; Zuo, B.; Kaplan, D. L.; Huang, Y.; Zhu, H. Biomacromolecules 2011, 12 (4), 1080−1086. (18) (a) Chen, K.; Teh, T. K. H.; Ravi, S.; Toh, S. L.; Goh, J. C. H. Tissue Eng., Part A 2012, 18 (17−18), 1902−1911. (b) He, P.; Ng, K. S.; Toh, S. L.; Goh, J. C. H. Biomacromolecules 2012, 13 (9), 2692− 2703. (c) He, P.; Sahoo, S.; Ng, K. S.; Chen, K.; Toh, S. L.; Goh, J. C. H. J. Biomed. Mater. Res., Part A 2013, 101 (2), 555−566. (d) Teh, T. K. H.; Toh, S. L.; Goh, J. C. H. Tissue Eng., Part A 2013, 19 (11−12), 1360−1372. (19) Golub, E. E.; Boesze-Battaglia, K. Current Opinion Orthopedics 2007, 18 (5), 444−448. (20) Lin, Z.; Rios, H. F.; Volk, S. L.; Sugai, J. V.; Jin, Q.; Giannobile, W. V. J. Periodontol. 2011, 82 (7), 1007−1017. (21) Lumachi, F.; Camozzi, V.; Tombolan, V.; Luisetto, G. Ann. N.Y. Acad. Sci. 2009, 1173, E64−E67. (22) Sandberg, M. M.; Aro, H. T.; Vuorio, E. I. Clin. Orthop. Relat. Res. 1993, 289, 292−312. (23) Stein, G. S.; Lian, J. B.; Van Wijnen, A. J.; Stein, J. L.; Montecino, M.; Javed, A.; Zaidi, S. K.; Young, D. W.; Choi, J. Y.; Pockwinse, S. M. Oncogene 2004, 23 (24), 4315−4329.

■

CONCLUSIONS Using particles to deliver bioactive molecule BMP-2 is an interesting topic and an effective method for reducing costs and risks. This article provides a detailed report of complex particles and an overview of advantages and disadvantages of the preparation of several silk fibroin-based particles for bioactive molecule delivery and tissue engineering. Large particles with a wide size distribution made by oil emulsion are not suitable for loading BMP-2. The silk fibroin nano- and microparticles have relatively high loading efficiencies that can enhance the efficacies of BMP-2 delivery. The complex particles fulfill intrinsic requirements of a material for delivery, including excellent biocompatibility, biodegradability, and bioactivity. Meanwhile, manufacturing techniques, sizes, and components of the materials will also have significant effects on the efficacy of delivery. The novel complex particles can enhance osteogenesis through material transformation and biological stimulations. Meanwhile, the encrusted particles can deliver the bioactive molecule properly and sustainably, in which calcium carbonate is transformed to hydroxyapatite, an important component in hard tissue. Further, the complex particles can significantly enhance the expression of osteogenic markers. The results indicate that cellular differentiation can be enhanced by application of the complex particles and the spatial organization of BMP-2-loaded particles. This concept of using complex particles to deliver BMP-2 and any other bioactive molecules (such as epidermal growth factor and vascular endothelial growth factor, etc.) will be useful for answering queries about cell−material interactions and growth factor signaling, which may breed a new generation of functional biomaterials.

■

ASSOCIATED CONTENT

S Supporting Information *

Primer sequences designed for real-time reverse transcriptase− polymerase chain reaction (Table S1), scanning electron microscopy images of MSCs on silk nanoparticles (Grp #3) and complex particles (Grp #4) after week 1 (Figure S1), and ALP activity of MSCs cultured with soluble BMP-2 with gradient concentrations (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*Telephone: +65 65161911. Fax: +65 68723069. E-mail: [email protected]. 4473

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474

Biomacromolecules

Article

(24) Santos, C.; Gomes, P. S.; Duarte, J. A.; Franke, R. P.; Almeida, M. M.; Costa, M. E. V.; Fernandes, M. H. J. R. Soc. Interface 2012, 9 (77), 3397−3410. (25) McHardy, L. M.; Warabi, K.; Andersen, R. J.; Roskelley, C. D.; Roberge, M. Mol. Cancer Ther. 2005, 4 (5), 772−778.

4474

dx.doi.org/10.1021/bm401381s | Biomacromolecules 2013, 14, 4465−4474