Rationally designed particle preloading method to improve protein delivery performance of electrospun polyester nanofibers

Rationally designed particle preloading method to improve protein delivery performance of electrospun polyester nanofibers

International Journal of Pharmaceutics 512 (2016) 204–212 Contents lists available at ScienceDirect International Journal of Pharmaceutics journal h...

3MB Sizes 0 Downloads 32 Views

International Journal of Pharmaceutics 512 (2016) 204–212

Contents lists available at ScienceDirect

International Journal of Pharmaceutics journal homepage: www.elsevier.com/locate/ijpharm

Rationally designed particle preloading method to improve protein delivery performance of electrospun polyester nanofibers Xue Suna , Ke Lia , Sainan Chena , Bing Yaoa , Yifa Zhoub , Sisi Cuib,* , Junli Hua,* , Yichun Liua,* a b

Key Laboratory of UV-Emitting Materials and Technology (Northeast Normal University), Ministry of Education, Changchun, Jilin, 130024, China School of Life Sciences, Northeast Normal University, Changchun, Jilin, 130024, China

A R T I C L E I N F O

Article history: Received 5 May 2016 Received in revised form 23 August 2016 Accepted 25 August 2016 Available online 25 August 2016 Keywords: Protein Polyester nanofibers Particle preloading Rational design

A B S T R A C T

Particle preloading method by first loading proteins onto nano- or microparticles and then integrating these particles into electrospun polyester nanofibers has been widely used to encapsulate therapeutic proteins into polyester nanofibers. However, poor method design has resulted in unsatisfactory protein delivery performance. For example, the harsh conditions involved in preloading procedures damage the bioactivities of proteins, the improper integration leads to an uneven distribution of particles in nanofibers or insecure attachment of particles to nanofibers, producing uncontrolled protein release profiles. This study aimed to improve the protein delivery performance of polyester nanofibers by rationally designing a particle preloading method. Positively charged chitosan nanoparticles (CNPs) were used as carriers to adsorb negatively charged proteins in mild conditions and as primary barriers for protein release. The polar CNPs were then homogeneously dispersed in a polar polyester solution and subjected to electrospinning. Microscope observations indicated that CNPs were homogeneously embedded within polyester nanofibers. In vitro release behaviour and cell studies showed that proteins retained their bioactivity and could release from polyester nanofibers in a sustained manner for more than 4 weeks without any initial burst. Epidermal growth factor encapsulated in polyester nanofibers enhanced diabetic wound healing in vivo, demonstrating an application potential in biomedicine. Other properties of the nanofibers, including composition, wettability, cytotoxicity, and cell adhesion and spreading, were examined in detail as well. ã 2016 Elsevier B.V. All rights reserved.

1. Introduction Aliphatic polyesters such as polylactic acid (PLA), poly(lacticco-glycolic acid) (PLGA) and polycaprolactone (PCL) represent an essential class of biomedical materials because of their biocompatibility, biodegradability, and tuneable physiochemical properties (Seyednejad et al., 2011; Cameron and Shaver, 2011). In particular, their electrospun nanofibers are widely used in the delivery of therapeutic protein for wound healing or tissue regeneration (Hu et al., 2014; Xie et al., 2008). Owing to the different solubilities of protein and polyester, strategies such as emulsion electrospinning (Li et al., 2006; Kim et al., 2007), coaxial electrospinning (Zhang et al., 2006; Liao et al., 2006), and chemical conjugation (Choi et al., 2008; Cho et al., 2010) have been used to

* Corresponding authors. E-mail addresses: [email protected] (S. Cui), [email protected] (J. Hu), [email protected] (Y. Liu). http://dx.doi.org/10.1016/j.ijpharm.2016.08.053 0378-5173/ã 2016 Elsevier B.V. All rights reserved.

encapsulate proteins in electrospun polyester nanofibers and control the release of proteins. However, the success of these strategies has been limited because of problems such as damaged bioactivity of proteins or complicated experimental setups. A new approach, the particle preloading method, has been developed, and it presents several advantages over other proteinloading strategies (Ionescu et al., 2010; Xie et al., 2013; Wei et al., 2007; Gungor-Ozkerim et al., 2014; Lai et al., 2014). The particle preloading method includes two steps: (1) load proteins onto nano- or microparticles; (2) integrate particles into polyester nanofibers. The particle matrix protects proteins from the harsh conditions involved in electrospinning (e.g., high voltage, organic solvents) and functions as a preliminary release barrier controlling the protein release rate. Various proteins have been encapsulated into polyester nanofibers via preloaded particles and applications in wound healing and tissue engineering have been explored (Ionescu et al., 2010; Xie et al., 2013; Wei et al., 2007; GungorOzkerim et al., 2014; Lai et al., 2014).

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

However, to achieve the ideal protein delivery performance, rational design is still lacking for current preloading methods. Some methods involved harsh conditions such as an oil-water interface and organic solvents in the preloading step, causing proteins to lose their bioactivities (Xie et al., 2013). Some methods integrated particles with polyester nanofibers simply by mixing the two components. The particles could easily shed off of nanofibers because there were no effective interactions between them, resulting in the loss of proteins (Gungor-Ozkerim et al., 2014). In others, to integrate the preloaded particles with polyester nanofibers, the authors dispersed particles in organic solutions of polyester and the suspensions were then subjected to electrospinning. Because the dispersion of particles in polyester solution was uneven, the distribution of particles in the polyester nanofibers was heterogeneous and the release behaviour of proteins was poorly controlled (Lai et al., 2014). All of these flaws greatly impaired the protein delivery performance of the particle-loaded polyester nanofibers. In this study, we intended to rationally design the particle preloading method to avoid these deficiencies and improve the protein delivery performance of polyester nanofibers. As shown in Fig. 1, to guarantee the activity retention of proteins in the preloading step, we used positively charged chitosan nanoparticles (CNPs) as carriers to adsorb negatively charged proteins in mild conditions. The nanoparticles were then dispersed in a solution of poly(lactic acid) (PLA) in 1,1,1,3,3,3,-hexafluoro-2-propanol (HFIP) and subjected to electrospinning. The nanoparticle matrix could prevent proteins from losing bioactivity during the electrospinning process. The nanoparticles’ small size and similarity of polarity to the electrospinning solution could enable their homogeneous dispersion in the solution, the subsequent homogeneous distribution within the electrospun polyester nanofibers, and the sustained release of proteins. The morphology, size, and wettability of the nanofibers along with the distribution of the nanoparticles within the polyester nanofibers were characterized in detail by a scanning electron microscope (SEM), contact angle measurements and a laser scanning confocal microscope (LSCM). The release behaviour and bioactivity of proteins were examined by in vitro incubation and a cell proliferation assay, respectively. Cell adhesion and

205

spreading on the polyester nanofibers were examined with cell culture experiments, as well as any cytotoxic effects. The application potential of the protein delivering polyester nanofibers was evaluated with in vivo wound healing experiments. 2. Materials and methods 2.1. Materials Chitosan (Mw  600 kDa, degree of deacetylation = 92.9%) was purchased from Sinopharm Chemical Reagents (Beijing, China). PLA (Mn = 265 kDa)was kindly gifted from Changchun Sinobiomaterials Co. Ltd. (Changchun, China). Bovine serum albumin fraction V (BSA), 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-Htetrazolium bromide (MTT), and the BCA protein assay kit were purchased from Dingguo Biotech (Beijing, China). Epidermal growth factor (EGF) was supplied by PeproTech (New Jersey, US). Sodium tripolyphosphate (TPP) and HFIP were purchased from Aladdin Reagents (Shanghai, China). Streptozotocin (STZ) was purchased from Meilune Biotech. (Dalian, China). All other chemical reagents were purchased from Xilong Chemicals (Guangzhou, China) and used as received. Deionized water was used at all times. 2.2. Methods 2.2.1. Preloading protein on chitosan nanoparticles (CNPs) Proteins were first loaded onto CNPs by the dipping method. First, chitosan nanoparticles were fabricated by the ionic gelation method and purified by centrifugation and repeated washing as previously reported (Hou et al., 2012). Then, a certain amount of model protein BSA or EGF was added to the chitosan nanoparticle suspension (20 mg/mL in water). The mixture was incubated at 4  C for 48 h to allow sufficient interactions between proteins and chitosan nanoparticles. The protein-loaded CNPs were collected by centrifugation (11,000 rpm, 30 min) and rinsed twice with deionized water. The supernatant was used to collect a UV absorbance measurement (280 nm) with a UV–vis spectrometer (Lambda 900, Perkin Elmer). The drug encapsulation efficiency

Fig. 1. Schematic illustration of the rationally designed particle preloading method to improve protein delivery performance of electrospun polyester nanofibers.

206

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

(EE) and loading content (LC) of protein were calculated using the following equations: EE ¼

mo  m1  100% m0

LC ¼

mo  m1  100% mo  m1 þ m2

where m0 , m1 , and m2 are the amount of the input protein, the amount of protein in the supernatant, and the amount of input chitosan nanoparticles, respectively. The size and zeta potential of chitosan nanoparticles before and after protein loading was measured with a Zetasizer Nano (MPT-2, Malvern). The CNP suspension was diluted to 0.1 mg/mL with PBS before these measurements. To study release behaviour of protein from CNPs, BSA-loaded CNPs (BSA@CNPs) (13 mg) were mixed with PBS (5 mL) and incubated at 37  C. At predetermined time points, the solution was centrifuged for 30 min. Supernatant (1 mL) was collected for protein concentration determination with the BCA assay, and fresh PBS (1 mL) was replenished for further incubation. The percentage of the protein released was calculated based on the weight of the loaded protein in the CNPs. 2.2.2. Embedding protein-loaded CNPs within PLA nanofibers through dispersion electrospinning A certain amount of protein-loaded CNPs were added into PLA solution (2 wt.% in HFIP) to achieve a final particle content in nanofibers between 6.7–29.7 wt.%. The mixture was stirred for 1–3 h to allow homogeneous dispersion of nanoparticles in the solution. The obtained dispersion was transferred to a 1-mL glass syringe with a steel needle connected to DC voltage with low current output (High DC Power Supply, Dalian Dingtong Technology). A round aluminium mesh with a 10-cm diameter was placed 15 cm away from the needle tip and used as a collector. All electrospinning was carried out at a voltage of 15 kV. 2.2.3. Characterizations 2.2.3.1. Morphology and size. The morphology of nanofibers was observed on a scanning electron microscope (SEM, XL-30 ESEM FEG, Micro FEI Philips) at an acceleration voltage of 20 kV after sputter-coating with gold for 120 s. The average nanofiber diameter in the SEM images was calculated by measurements from 50 nanofibers. 2.2.3.2. Distribution of nanoparticles in nanofibers. FITC-labelled BSA was used instead of normal BSA during the material preparation process. The electrospun nanofibers were collected on a clean glass slide and observed with laser scanning confocal microscope (LSCM, LSM 780, Carl Zeiss). 2.2.3.3. Composition. The Fourier transform infrared (FTIR) spectra of the nanofibers were collected with an FTIR spectrometer (IS10, Thermo Scientific) with an MCT detector. Samples were prepared by mixing nanofibers with FTIR grade KBr at a mass ratio of approximately 1:100 and pressing them into discs. Scans were carried out at a detective model of absorbance, resolution of 16 cm1 and range between 4000 and 600 cm1. Blank KBr discs were used for background reduction. 2.2.3.4. Wettability. The wettability of nanofiber mats was evaluated through water contact angle measurements (DSA100, KRUSS). A 5-mL drop of water was placed on the nanofiber mats, and the contact angle was measured with Drop Shape Analysis software.

2.2.3.5. In vitro protein release. Nanofibers (20 mg) were soaked in phosphate-buffered saline (PBS, pH 7.4, 5 mL) and incubated at 37  C. At predetermined time points, soaking solution (1 mL) was collected for the BCA protein assay. Fresh PBS (1 mL) was replenished for further incubation. 2.2.3.6. Cytotoxicity. The cytotoxicity of nanofibers was evaluated through the MTT assay of their leachate. Nanofibers (10 mg) were soaked in Dulbecco Minimum Essential Medium (DMEM, Gibco, 2 mL) at 37  C for 24 h. L929 mouse fibroblast cells (ATCC, US) were prepared using a 96-well tissue culture plate in DMEM supplemented with 10% foetal calf serum (FCS) at a density of 5000 cells/well. The cells were allowed to attach and grow for 24 h at 37  C in a humidified atmosphere containing 5% CO2. Leachate of nanofibers of different concentrations (0  3.75 mg/mL) was added into cell medium (n = 6). After cultured for 48 h, the cells were quantified by the MTT assay. 2.2.3.7. Cell adhesion, spreading and proliferation on nanofiber mats. The nanofiber mats were plated in sterile 24-well plates and fixed to the bottom with an alumina ring. L929 mouse fibroblast cells were seeded on nanofiber mats with DMEM supplemented with 10% FCS at a density of 20,000–80,000 cells/ well and cultured for 4 h (for the adhesion assay), 10 h (for SEM observation) or 7 days (for the proliferation assay) at 37  C in a humidified atmosphere containing 5% CO2. To calculate the cell adhesion rate of L929 cells on nanofiber mats, the 24-well plates were rinsed three times with D-Hank’s solution after incubation for 4 h. The adherent cells were trypsinized, collected, and counted using a flow cytometer (Accuri C6, BD). The ratio of the adherent cells to the total cell number was taken as cell adhesion rate. After culture for 10 h, the cells on nanofiber mats were fixed with 2.5% paraldehyde solution at 4  C for 1 h, washed three times with deionized water, dehydrated in a graded series of ethanol, and observed with SEM. After culture for 7 days, cell proliferation on nanofiber mats was evaluated by the MTT assay. 2.2.3.8. Application in wound healing of diabetic rats. Sprague Dawley rats were purchased from Changchun Institute of Biological Products Co., Ltd. All animal experiments were carried out in compliance with the Animal Management Rules of the Ministry of Health of People's Republic of China and approved by the Animal Care and Use Committee of Northeast Normal University. A type I diabetic rat model was established according to previous report (Yang et al., 2014). Briefly, 15 male Sprague Dawley rats (body weight 250  300 g) were divided randomly into two groups: blank PLA/CNP nanofiber mats and PLA/ EGF@CNP nanofiber mats. Each rat was intraperitoneally injected with STZ at a dose of 60 mg/kg, which was dissolved in sodium citrate buffer (pH 4.5). Whole-blood was obtained from the tail vein 5 days later, and the glucose level was monitored using a Blood Glucose monitor (Beijing Yicheng Bioelectronics Technology Co., Ltd., China). STZ-injected rats with glucose levels higher than 16.7 mM were considered diabetic. Diabetic rats were anesthetized with pentobarbital (35 mg/kg) through intraperitoneal injection. After clipping the dorsal hair by a pair of scissors, the dorsal area of rats was depilated by Na2S solution. Three full-thickness circular wounds were created on the upper back of each rat using a corneal trephine with a diameter of 9 mm. Blank PLA/CNP nanofiber mats and EGF encapsulated nanofiber mats were sterilized by UV irradiation for 24 h and then applied to the wounds of diabetic rats in the corresponding groups. Each treated wound was covered by antiseptic gauze and the edge of gauze was sutured to the skin around the wound area. After treatment for 7 days, wounds were photographed with a camera and measured with a ruler.

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

207

Fig. 2. Size and zeta potential of naked chitosan nanoparticles (CNPs, a and b) and BSA loaded chitosan nanoparticles (BSA@CNPs, c and d).

3. Results and discussion 3.1. Preloading protein onto chitosan nanoparticles (CNPs) Ionic gelation of chitosan by TPP occurs in aqueous and acidic conditions. Biomolecules can be loaded onto the formed chitosan nanoparticles (CNPs) either in situ or afterwards by dipping. The method has been used to deliver a wide range of biomolecules including proteins and antibiotics due to its simplicity and mild reaction conditions (Yang et al., 2011; Sipoli et al., 2015; Yang et al., 2014). Here we used these gel-like TPP crosslinked chitosan nanoparticles as preliminary nanocarriers of proteins. As shown in Fig. 2(a–b), the size and surface charge of the pristine CNPs were dependent on the amount of crosslinking reagent (TPP). As the mass ratio between TPP and chitosan increased from 1:10 to 2:10, the diameter of the nanoparticles markedly decreased from 580 nm to 240 nm and the zeta potential of the nanoparticles decreased slightly from +44 mV to +39 mV. The reduction in size and surface charge of CNPs corresponding to the increased amount of crosslinking reagents was probably due to the reduced electrostatic repulsion between the positively charged chitosan chains in nanoparticles. These repulsive forces lead to the relatively large size of the particles. As the negative charges of TPP could neutralize the positive charges of chitosan chains and reduce the repulsion, the amount of TPP affects the particle size. More TPP increased the neutralization effect, thus decreasing the repulsion forces and resulting in a smaller particle size. A similar CNP size dependency on TPP amount was observed by Alonso et al. (Calvo et al., 1997). These results indicated that in hydrated CNPs the chitosan molecular chains were not densely packed but were loosely crosslinked by TPP ions. The space between these loosely

crosslinked chitosan chains would allow proteins enter into the nanoparticles and facilitate the interaction between proteins and chitosan nanoparticles. Proteins could therefore be adsorbed not only on the surface but also inside of chitosan nanoparticles. Smaller nanoparticles were preferred in the protein preloading step because the nanoparticles will be embedded in nanofibers in the following step. The smallest CNPs (prepared at a TPP/chitosan mass ratio of 2:10) were selected for protein loading. BSA was used as a model protein. The size and zeta potential of nanoparticles after loading with various amounts of protein were examined. As shown in Fig. 2(c–d), loading of BSA dramatically reduced the size and surface charge of the particles. As the BSA/CNPs mass ratio increased from 1:10 to 3:10, the diameter of the nanoparticles gradually decreased from the 122 nm to 66 nm and the zeta potential of the nanoparticles gradually decreased from + 22 mV to approximately + 10 mV. The reduced size and zeta potential caused by BSA loading probably demonstrated that the negatively charged BSA entered into the space between positively charged chitosan chains, neutralized their charges, and dragged them closer. Table 1 lists the EE and LC of BSA in CNPs. The EE was above 90% at all BSA/ CNPs mass ratios. An LC as high as 21.7% could be obtained at a BSA/ CNPs mass ratio of 3:10. These results indicated a very high protein loading capacity of CNPs. Table 1 Encapsulation Efficiency (EE) and Loading Content (LC) of BSA in CNPs at Various BSA/CNPs Mass Ratios. BSA/CNPs mass ratio

1:10

2:10

3:10

EE (%) LC (%)

93.4  0.5 8.5  0.2

93.1  0.3 15.7  0.2

92.1  0.5 21.7  0.3

208

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

Fig. 3. SEM (a1–d1, a2–d2) and LSCM (a3–d3) images of PLA nanofibers embedding various contents of BSA@CNPs: (a) 6.7%; (b) 9.1%; (c) 16.7%; (d) 29.7%.

Therefore, using chitosan nanoparticles as nanocarriers, a protein could be loaded onto the nanocarriers under mild conditions with a high encapsulation efficiency and loading content and yielding nanoparticles of a very small size. 3.2. Embedding protein-loaded CNPs within PLA nanofibers through dispersion electrospinning BSA-loaded chitosan nanoparticles (BSA@CNPs), prepared at a BSA/CNPs mass ratio of 2:10 with a size of 80 nm and a protein LC of 15.7%, were dispersed in a PLA/HFIP solution and the suspension was subjected to electrospinning. No observable aggregation or precipitation occurred during the electrospinning process, demonstrating the homogeneous and stable dispersion of the nanoparticles in the suspension. As shown by SEM images (Fig. 3(a1–d1)), smooth and bead-free nanofibers were obtained at nanoparticle content of 6.7% and 9.1%. No nanoparticles were observed on the surface of the nanofibers, indicating that they were successfully embedded in the nanofibers at these two nanoparticle contents. Some rough regions appeared on the surface of the nanofibers as the content of nanoparticles increased to 16.7%. Nanofibers with completely rough surfaces as well as spindle and break regions were obtained as the content of nanoparticles further increased to 29.8%. The average diameter of the nanofibers was approximately 350 nm at the 6.7% and 9.1% content of nanoparticles, while it decreased to 309 nm and 273 nm as the content of nanoparticles increased to 16.7% and 29.8%, respectively, probably owing to the increased electric force resulted from the increased conductivity.

With FITC-labelled BSA as a fluorescent probe to label nanoparticles, we could image the distribution of nanoparticles within nanofibers with LSCM. As shown by Fig. 3(a3–d3), nanoparticles were homogeneously scattered in the centre of nanofibers containing 6.7%–16.7% nanoparticles. In contrast, for nanofibers containing 29.8% nanoparticles, the spindle regions of the SEM images correlated well with the fluorescent spindle regions in LSCM (Fig. 3(d2, d3)), indicating that these regions were formed by nanoparticles. These results suggested that a certain extent of nanoparticle aggregation could occur during the electrospinning process and impair the formation of smooth nanofibers and the homogeneous distribution of nanoparticles within nanofibers. At a low nanoparticle content of 6.7% and 9.1%, the extent of aggregation was very low. As the content of nanoparticles increased to 16.7%, the extent of aggregation increased and its effect on the nanofiber morphology was observed. As the content of nanoparticles was further increased to 29.7%, aggregation occurred to a large extent and impaired the formation of smooth nanofibers and the homogeneous distribution of nanoparticles within nanofibers dramatically. The above results indicated that the protein-loaded nanoparticles could be homogeneously embedded by polyester nanofibers at certain nanoparticle contents. As discussed in the introduction section, homogeneous and secure distribution of nanoparticles in the nanofiber matrix is a key factor for controllable protein release. Protein-loaded chitosan nanoparticles possess two important features to achieve this goal: (1) Small size and (2) Polarity. Electrospun PLA nanofibers normally have diameters ranging from 150 nm to 400 nm. The small size the

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

209

the nanoparticles and the dispersion solvent supports a homogeneous and stable dispersion of nanoparticles in suspension and thus guarantee the homogeneous distribution of the nanoparticles within the produced polyester nanofibers. As a result, the risk of shedding nanoparticles off the surface or uneven protein release from nanofibers was minimized, providing for controllable protein release. 3.3. Composition

Fig. 4. Representative FTIR spectrum of PLA nanofibers with embedded BSA@CNPs. Raw chitosan and PLA materials are shown for comparison.

Fig. 4 shows a representative FTIR spectrum of nanofibers containing 9.1% BSA @CNPs. Spectra from raw PLA and chitosan are shown for comparison. There was a characteristic peak at 1763 cm1 in the FTIR spectrum of PLA attributed to the C¼O stretch of its ester groups. A characteristic peak at 1635 cm1 in the spectrum of chitosan was attributed to the C¼O stretch of its amide groups. Both peaks (a red square for 1763 cm1 and a blue triangle for 1635 cm1) were observed in the spectrum of PLA/BSA@CNP nanofibers, demonstrating that the protein loaded chitosan nanoparticles were successfully integrated with PLA nanofibers. 3.4. Wettability

Fig. 5. Dependence of contact angle on the content of BSA@CNPs in PLA/BSA@CNP nanofiber mats.

nanoparticles (80 nm as reported above) allow them to be fully embedded in the nanofibers. Chitosan nanoparticles are polar, similar to HFIP, the solvent of PLA. This polarity similarity between

The wettability of nanofiber mats was examined by contact angle measurements (Fig. 5). As the content of BSA@CNPs in nanofibers increased from 0 to 29.7%, the water contact angle of nanofiber mats remained consistent at 130  3 , indicating the high hydrophobicity of the nanofiber mats. The integration of nanoparticles showed no significant effects on the wettability of the nanofiber mats. As discussed previously, nanoparticle aggregation occurred in PLA nanofibers at high nanoparticle content. The negligible effect of nanoparticle content on the water contact angle of nanofiber mats suggested that most nanoparticles, aggregated or not, were encapsulated within PLA nanofibers. Because PLA is intrinsically a hydrophilic polymer with a contact angle of approximately 80 , the increased hydrophobicity of the nanofiber mats was probably due to the exposure of the methyl groups and the porous surface morphology of nanofiber mats, as reported and suggested previously (Chang et al., 2016; Shao et al., 2015). The hydrophobicity of the nanofiber mats might be advantageous for tissue engineering scaffold applications because it would lead to improved protein adsorption upon contact with physiological fluid and thus result in improved cell adhesion (Anand et al., 2010; Wilson et al., 2005).

Fig. 6. Release behaviour of BSA from (a) CNPs and (b) PLA/BSA@CNP nanofibers.

210

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

3.5. Protein release behaviour Fig. 6(a–b) show the release profile of protein from CNPs and nanofibers, respectively. As shown by Fig. 6(a), 30% of the loaded BSA was released from CNPs after the first day of incubation. After the initial burst, BSA was released in a sustained manner, reaching 90% of total released in 2 weeks. The burst release on the first day is likely attributed to the loosely adsorbed BSA on the surface of CNPs. The following sustained release suggested that more BSA was encapsulated inside of the oppositely charged chitosan matrix. These results indicated that, in addition to functioning as a protective matrix in the electrospinning process, the CNPs also function as a preliminary release barrier for protein. As shown in Fig. 6(b), when the loading content of BSA fell into the range of 1.2%–2.9%, it was released from nanofibers in a sustained manner for more than 4 weeks without any initial burst. The BSA initial burst was observed for CNPs (Fig. 6(a)), so the absence of the BSA initial burst from nanofibers demonstrated the secondary barrier effect of PLA nanofibers, which was further demonstrated by the much slower release rate of BSA from nanofibers compared with that from CNPs in the 4-week time frame. The highly controllable release behaviour of BSA could be attributed to the homogeneous embedding of nanoparticles within nanofibers as previously demonstrated. The release rate of BSA varied significantly with its loading content. The slowest release was observed from nanofibers containing the least BSA (1.2%). The fastest release was observed from nanofibers containing a medium amount of BSA (1.6%). A medium release rate was achieved by nanofibers containing the most BSA (2.9%). The drug release rate from nanofibers is affected by the diameter and surface morphology of the nanofibers and the drug content in the nanofibers. Because we used BSA@CNPs with a fixed BSA loading content (15.7%) for fabrication of PLA nanofibers, the nanofibers containing 1.2%, 1.6% and 2.9% BSA corresponded to those containing 6.7%, 9.1% and 16.7% BSA@CNPs (Fig. 3a–c), respectively. As indicated previously, the average diameter of the

former two nanofibers was approximately 350 nm and that of the latter was 309 nm. The surface of the former two was smooth while that of the latter was slightly rough. The higher BSA release rate from nanofibers containing 1.6% in comparison with that from nanofibers containing 1.2% BSA might mainly be attributed to a higher number of diffusion routes at higher BSA content. For nanofibers with the highest BSA loading content of 2.9%, because of their smallest diameter and roughest surface, they possessed the highest surface area, which would allow more buffers to enter. In addition, because of the highest BSA loading content, the greatest number of BSA diffusion routes would form in these nanofibers. It is thus expected that BSA would release fastest from these nanofibers, which however was not true from the results. The unusual slow release at the highest BSA loading content was thus considered to be associated with the previously discussed nanoparticle aggregation. It is possible that a large extent of nanoparticle aggregation occurred at high BSA loading content, i.e., high nanoparticle content, and hindered the diffusion of the protein. To verify this speculation, the morphology of the nanofibers during protein release was imaged. As shown in Fig. 7, as incubation proceeded, the size of each individual nanofiber grew uneven, the surface turned rougher, and fusion between fibers occurred. More importantly, as indicated by the encircled red oval regions in the 4-week images, spindle regions appeared and increased in number with the prolonging of the incubation time. This phenomenon was particularly evident for the nanofibers with the highest BSA loading content (2.9%) and nanoparticle content (16.7%). The previous results (Fig. 3) suggested that the spindle regions were formed by the aggregated BSA@CNPs, so these results demonstrated that aggregation of BSA@CNPs did occur within nanofibers and its extent increased with the content of nanoparticles. The large extent of aggregation in PLA nanofibers with a high content of BSA@CNPs (16.7%) probably hindered the diffusion of BSA and resulted in a reduced release rate.

Fig. 7. Morphology of nanofibers during protein release.

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

211

3.6. Biocompatibility Fig. 8 shows the viability of L929 mouse fibroblast cells cultured in the presence of various concentrations of nanofiber leachate. At a leachate concentration between 0.24 and 3.75 mg/mL, more than 85% of the L929 mouse fibroblast cells were viable, indicating good cytocompatibility of the nanofibers. The good cytocompatibility of the nanofibers could be attributed to the high biocompatibility of the materials because chitosan and PLA are both known as biocompatible polymers. In addition, the fabrication process involved only one toxic reagent, HFIP, which is highly volatile and probably evaporated during the electrospinning and drying process. 3.7. Protein bioactivity

Fig. 8. Cytotoxicity of PLA/BSA@CNP nanofibers.

The overall released percentage of BSA in 4 weeks was 9.2%, 45.2%, and 19.4% for PLA nanofibers containing 1.2%, 1.6%, and 2.9% BSA, respectively. The remaining BSA might be close to the central regions of the nanoparticles/nanofibers and might be released with further incubations. The long release time of protein is particularly suitable for long time demanding tissue engineering applications such as bone regenerations.

To investigate the bioactivity of protein after the fabrication process, a bioactive cytokine, epidermal growth factor (EGF), was used as another model protein and encapsulated in PLA nanofiber mats. The adhesion, spreading and proliferation of cells on EGFencapsulated PLA nanofiber mats were evaluated with a flow cell counter, SEM observations, and an MTT assay, respectively. As shown in Fig. 9(a), when 50,000 cells were seeded on EGFencapsulated nanofiber mats, 76% of them successfully attached to the nanofiber mats after 4 h of incubation. As more cells (80,000 and 100,000) were seeded initially, the adhesion rate increased to over 80%. SEM observations (Fig. 9(b)) indicated that at 10 h after seeding, cells spread well on the nanofiber mats. The high hydrophobicity and roughness of the nanofiber mats probably

Fig. 9. Adhesion (a), spreading (b) and proliferation (c) of L929 mouse fibroblast cells on the PLA/EGF@CNP nanofiber mats. PLA/CNP blank nanofiber mats were used as controls.

Fig. 10. Representative images (a) and statistical areas (b) of diabetic skin wounds after being treated with PLA/EGF@CNP nanofiber mats for 7 days. PLA/CNP blank nanofiber mats were used as controls.

212

X. Sun et al. / International Journal of Pharmaceutics 512 (2016) 204–212

facilitated protein adsorption, which offered more binding sites to cell membrane receptors and thus enhanced cell adhesion and spreading (Anand et al., 2010; Wilson et al., 2005; Stevens and George, 2005). As shown by Fig. 9(c), EGF-encapsulated nanofiber mats induced 30% more cell proliferation than blank nanofiber mats, suggesting that the encapsulated model protein retained its biological activity. The retained protein activity can be attributed to the rational design of this particle preloading method. In the first preloading step, proteins were loaded on nanoparticles in fully aqueous and neutral conditions. In the second dispersion electrospinning step, proteins were protected by the particle matrix from the high voltage and organic solvent, both of which were reported to induce protein unfolding and aggregation and thus cause their activity loss (Sah, 1999; Zhao and Yang, 2010). The bioactivity of proteins was not damaged in each step and thus was well retained. 3.8. Application in wound healing As EGF is widely reported to facilitate wound healing, the EGFencapsulated polyester nanofiber mats were applied in diabetic wound healing experiments to examine their efficacy in vivo. Fig. 10 showed the representative images and statistical areas of wounds on the backs of diabetic rats after being treated with blank PLA/CNP nanofiber mats and EGF-encapsulated nanofiber mats for 7 days. The healing rate of the wounds treated with EGFencapsulated nanofiber mats was moderately faster than the ones treated with blank nanofiber mats. The size of the diabetic wounds was reduced from the initial 50 mm2 to 28 mm2 after being treated with blank nanofiber mats for 7 days and was reduced to 18 mm2 after treatment with EGF-encapsulated nanofiber mats for the same time. The higher recovery rate of the diabetic wounds induced by EGF-encapsulated polyester nanofiber mats indicated the therapeutic efficacy of EGF. In this in vivo experiment, the results demonstrated the application potential in biomedical fields of the protein-encapsulated polyester nanofibers fabricated with the rationally designed particle preloading method. 4. Conclusions Protein delivery performance of electrospun polyester nanofibers was improved by rationally designing the particle preloading and dispersion electrospinning steps in the particle preloading method. The obtained protein-encapsulated polyester nanofibers were of smooth surface and uniform size and contained preloaded nanoparticles homogeneously embedded within. Proteins retained their bioactivity and were released from nanofibers in a sustained manner for more than 4 weeks without any initial burst. The nanofiber mats showed no obvious cytotoxicity, could support cell adhesion and spreading, and induced cell proliferation efficiently. The EGF-delivering polyester nanofiber mats enhanced diabetic wound healing, demonstrating their application potential in biomedical fields. Acknowledgements This work was supported by the Science and Technology Development Plan of Jilin Province (No. 20150520021JH), the National Natural Science Foundation of China (Nos. 51503027, 81401516), and the 111 project (No. B13013). The authors thank Jie Chen and Zhantuan Gao from Changchun Institute of Applied Chemistry for great help.

References Anand, G., Sharma, S., Dutta, A.K., Kumar, S.K., Belfort, G., 2010. Conformational transitions of adsorbed proteins on surfaces of varying polarity. Langmuir 26, 10803–10811. Calvo, P., Remuñán-López, C., Vila-Jato, J.L., Alonso, M.J., 1997. Novel hydrophilic chitosan-polyethylene oxide nanoparticles as protein carriers. J. Appl. Polym. Sci. 63, 125–132. Cameron, D.J., Shaver, M.P., 2011. Aliphatic polyester polymer stars: synthesis, properties and applications in biomedicine and nanotechnology. Chem. Soc. Rev. 40, 1761–1776. Chang, Y., Liu, X., Yang, H., Zhang, L., Cui, Z., Niu, M., Liu, H., Chen, J., 2016. Nonsolvent-assisted fabrication of multi-scaled polylactide as superhydrophobic surfaces. Soft Matter 12, 2766–2772. Cho, Y.I., Choi, J.S., Jeong, S.Y., Yoo, H.S., 2010. Nerve growth factor (NGF)-conjugated electrospun nanostructures with topographical cues for neuronal differentiation of mesenchymal stem cells. Acta Biomater. 6, 4725–4733. Choi, J.S., Leong, K.W., Yoo, H.S., 2008. In vivo wound healing of diabetic ulcers using electrospun nanofibers immobilized with human epidermal growth factor (EGF). Biomaterials 29, 587–596. Gungor-Ozkerim, P.S., Balkan, T., Kose, G.T., Sarac, A.S., Kok, F.N., 2014. Incorporation of growth factor loaded microspheres into polymeric electrospun nanofibers for tissue engineering applications. J. Biomed. Mater. Res. A 102A, 1897–1908. Hou, Y., Hu, J., Park, H., Lee, M., 2012. Chitosan-based nanoparticles as a sustained protein release carrier for tissue engineering applications. J. Biomed. Mater. Res. A 100A, 939–947. Hu, X., Liu, S., Zhou, G., Huang, Y., Xie, Z., Jing, X., 2014. Electrospinning of polymeric nanofibers for drug delivery applications. J. Control Release 185, 12–21. Ionescu, L.C., Lee, G.C., Sennett, B.J., Burdick, J.A., Mauck, R.L., 2010. An anisotropic nanofiber/microsphere composite with controlled release of biomolecules for fibrous tissue engineering. Biomaterials 31, 4113–4120. Kim, T.G., Lee, D.S., Park, T.G., 2007. Controlled protein release from electrospun biodegradable fiber mesh composed of poly(e-caprolactone) and poly(ethylene oxide). Int. J. Pharm. 338, 276–283. Lai, H.J., Kuan, C.H., Wu, H.C., Tsai, J.C., Chen, T.M., Hsieh, D.J., Wang, T.W., 2014. Tailored design of electrospun composite nanofibers with staged release of multiple angiogenic growth factors for chronic wound healing. Acta Biomater. 10, 4156–4166. Liao, C., Chew, S.Y., Leong, K.W., 2006. Aligned core–shell nanofibers delivering bioactive proteins. Nanomedicine 1, 465–471. Li, C., Vepari, C., Jin, H.J., Kim, H.J., Kaplan, D.L., 2006. Electrospun silk-BMP-2 scaffolds for bone tissue engineering. Biomaterials 27, 3115–3124. Sah, H., 1999. Protein instability toward organic solvent/water emulsification: implications for protein microencapsulation into microspheres PDA. J. Pharm. Sci. Technol. 53, 3–10. Seyednejad, H., Ghassemi, A.H., Van Nostrum, C.F., Vermonden, T., Hennink, W.E., 2011. Functional aliphatic polyesters for biomedical and pharmaceutical applications. J. Control Release 152, 168–176. Shao, J., Tong, L., Tang, S., Guo, Z., Zhang, H., Li, P., Wang, H., Du, C., Yu, X.F., 2015. PLLA nanofibrous paper-based plasmonic substrate with tailored hydrophilicity for focusing SERS detection. ACS Appl. Mater. Interfaces 7, 5391–5399. Sipoli, C.C., Radaic, A., Santana, N., Jesus, M.B., Torre, L.G., 2015. Chitosan nanoparticles produced with the gradual temperature decrease technique for sustained gene delivery. Biochem. Eng. J. 103, 114–121. Stevens, M.M., George, J.H., 2005. Exploring and engineering the cell surface interface. Science 310, 1135–1138. Wei, G., Jin, Q., Giannobile, W.V., Ma, P.X., 2007. The enhancement of osteogenesis by nano-fibrous scaffolds incorporating rhBMP-7 nanospheres. Biomaterials 28, 2087–2096. Wilson, C.J., Clegg, R.E., Leavesley, D.I., Pearcy, M.J., 2005. Mediation of biomaterialcell interactions by adsorbed proteins: a review. Tissue Eng. 11, 1–18. Xie, J., Li, X., Xia, Y., 2008. Putting electrospun nanofibers to work for biomedical research. Macromol. Rapid Commun. 29, 1775–1792. Xie, Z., Paras, C.B., Weng, H., Punnakitikashem, P., Su, L.C., Vu, K., Tang, L., Yang, J., Nguyen, K.T., 2013. Dual growth factor releasing multi-functional nanofibers for wound healing. Acta Biomater. 9, 9351–9359. Yang, Y., Xia, T., Zhi, W., Wei, L., Weng, J., Zhang, C., Li, X.H., 2011. Promotion of skin regeneration in diabetic rats by electrospun core-sheath fibers loaded with basic fibroblast growth factor. Biomaterials 32, 4243–4254. Yang, Y., Wang, S., Wang, Y., Wang, X., Wang, Q., Chen, M., 2014. Advances in selfassembled chitosan nanomaterials for drug delivery. Biotech. Adv. 32, 1301– 1316. Zhang, Y.Z., Wang, X., Feng, Y., Li, J., Lim, C.T., Ramakrishna, S., 2006. Coaxial electrospinning of (fluorescein isothiocyanate-conjugated bovine serum albumin)-encapsulated poly(e-caprolactone) nanofibers for sustained release. Biomacromolecules 7, 1049–1057. Zhao, W., Yang, R., 2010. Experimental study on conformational changes of lysozyme in solution induced by pulsed electric field and thermal stresses. J. Phys. Chem. B 114, 503–510.