Physicochemical properties and biocompatibility of PZL/PLGA/bioglass composite scaffolds for bone tissue engineering

Abstract

Polypeptides synthesized from α-amino acid N-carboxyanhydrides are considered ideal candidates as scaffold materials for bone tissue engineering due to their excellent biocompatibility and non-toxic degradation products. However, undesirable mechanical properties and osteoinductivity restrict their application in bone healing and regeneration. To overcome these limitations, we modified a synthetic polypeptide poly(N_ε-Cbz-L-lysine) (PZL) with poly(lactic-co-glycolic acid) (PLGA) having relatively high strength and osteoinductive bioglass (BG) particles, and fabricated foamy PZL/PLGA/BG composite scaffolds using a negative NaCl-templating method. The morphology, compression modulus, thermal and degradation behaviors of the PZL/PLGA/BG composite scaffolds were characterized, and the in vitro biocompatibility was evaluated with MC3T3-E1 cells using live/dead staining, MTT and ALP assays. The results indicated that the PZL/PLGA/BG composite scaffold with a weight ratio of 5 : 5 : 1 (PZL5PLGA5BG) had higher compression modulus and protein adsorption and mineralization abilities than other scaffolds, and was more conducive to the adhesion, proliferation and osteoblastic differentiation of MC3T3-E1 cells. The in vivo biocompatibility of the scaffolds was evaluated in both rat subcutaneous model and rabbit tibia defect model. The results of histological studies of subcutaneous implants, as confirmed by H&E, Masson's trichrome and CD34 staining assays, demonstrated that the PZL5PLGA5BG composite scaffolds allowed the ingrowth of tissue and microvessels and exhibited reduced inflammation response as compared to other scaffolds after 8 weeks of implantation. In experimental studies in bone defect model, the results of digital radiography and H&E staining confirmed that the PZL5PLGA5BG composite scaffolds after 8 weeks of implantation significantly improved in vivo osteogenesis. The newly formed bone tissue grew into the scaffolds along with the degradation of the materials. To sum up, the foamy PZL/PLGA/BG composite scaffolds had good comprehensive performances and would meet the needs of bone regeneration.

---

1 Introduction

In bone tissue engineering, the osteoinductivity of artificial grafts plays an increasingly important role in bone healing and regeneration processes.¹ A most widely used method to enhance osteoinductivity is to incorporate bioactive molecules like growth factors into porous scaffolds, such as TGF-β,² BMP,³ IGF,⁴ FGF,⁵ and VEGF.⁶ These growth factors can facilitate osteogenesis, bone tissue regeneration or ECM formation via recruiting and differentiating osteoprogenitor cells to specific lineages. The osteoinductivity is not intrinsic to the scaffolds themselves, but is rather the result of the ability of scaffolds to adsorb growth factors. On the other hand, the supraphysiological dose of growth factors used in clinic may cause potential side effects or adverse effects, such as undesirable immunological response, edema and cancer metastasis.^7–9 Therefore, in order to avoid the side effects, a promising alternative to growth factors is to develop biomaterials with favorable intrinsic osteoinductivity that can promote bone defect healing and regeneration.

Polypeptides in the form of homopolymer and copolymer, which are synthesized by ring-opening polymerization (ROP) of α-amino acid N-carboxyanhydrides (NCAs), have attracted increasing attention in various biomedical fields because of their ability to enzymatically degrade into non-toxic products with a near neutral pH, pendant functional groups, excellent biocompatibility, and non-immunogenicity.¹⁰ Nowadays, a variety of polypeptides, such as poly(γ-glutamic acid),¹¹ poly(L-lysine),¹² phenylalanine¹³ and polyaspartic acid,¹⁴ have been studied intensively and extensively for drug/gene delivery systems.¹⁵ However, they just sprouted in the field of tissue engineering. The most likely reason is that high molecular weight polypeptides are not easily synthesized, which leads to undesirable mechanical performances of porous scaffolds. In particular, there are very few reports on the application of polypeptides as porous scaffolds for bone tissue engineering.¹⁶ As to the porous structure of bone scaffolds, foamy scaffolds fabricated from negative NaCl bulk templates that are obtained by sintering NaCl particles have attracted increasing attention due to their merits of high porosity, easy-preparation, no residual salt, controllable pore size and high pore connectivity.¹⁷ It was reported that high porosity could facilitate oxygen and nutrient diffusion.¹⁸ To endow polypeptide-based scaffolds with desirable mechanical properties and bioactivity, an efficient method is to prepare composite materials by mixing with another biopolymer with high mechanical strength and bioactive inorganic particles.¹⁹

Poly(lactic-co-glycolic acid) (PLGA) has attracted considerable interest for biomedical applications due to its good biocompatibility, excellent mechanical strength and controllable degradation kinetics.²⁰ Hydroxyapatite (HAp) and β-tricalcium phosphate (β-TCP) are typical bioactive inorganic materials used for bone repair. However, many studies have indicated that the dissolution of HAp is too low to achieve rapid bone healing,²¹ while the dissolution rate of β-TCP is too fast for bone bonding.²² In contrast, bioactive glasses (bioglass, BG) have been proved to have controllable biodegradability and can stimulate bone repair and vascularization.^23,24

Herein, we proposed bioactive foamy polymer/inorganic composite scaffolds composed of poly(N_ε-Cbz-L-lysine) (PZL), PLGA and BG particles. The composite scaffolds were fabricated using negative NaCl templates. As shown in Fig. 1, we first characterized the physicochemical properties of the scaffolds, including morphology, mechanical performances, degradation kinetics, thermal behaviors, protein adsorption and release profiles, mineralization properties, and ion release behaviors. The in vitro biocompatibility of the scaffolds was evaluated with MC3T3-E1 cells using live/dead viability, MTT and alkaline phosphatase (ALP) activity assays. The in vivo biocompatibility of the scaffolds was evaluated in SD rat subcutaneous model and rabbit tibia defect model. Histological analysis including H&E, Masson's trichrome and CD34 staining as well as digital radiography were carried out to reveal the tissue response, blood vessel ingrowth and osteogenesis of the scaffolds. The results demonstrated that the PZL/PLGA/BG composite scaffolds provided a favorable substrate for MC3T3-E1 cells and showed good biocompatibility and osteogenesis, holding great potential for bone tissue engineering application.

Fig. 1 Schematic illustration of physicochemical properties of PZL/PLGA/BG composite scaffolds, in vitro culture of MC3T3-E1 cells/scaffold constructs and in vivo implantation in SD rat subcutaneous and rabbit tibia bone defect models.

2 Experimental

2.1 Materials

L-Lysine hydrochloride, triphosgene and triethylamine were purchased from Aladdin Reagent Inc. (Shanghai, China). Ethylenediaminetetraacetic acid disodium salt (EDTA-2Na) and albumin from bovine serum (BSA) were purchased from Sinopharm Chemical Reagent (Xi'an, China). PLGA (LA/GA = 75 : 25, M_w = 300 kDa) was purchased from Jinan Daigang Biomateiral Co., Ltd. (Jinan, China). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) was purchased from Sigma-Aldrich. The live/dead viability/cytotoxicity kit for mammalian cells was purchased from Invitrogen (Carlsbad, CA, USA). Minimum essential medium α (MEMα) and fetal bovine serum (FBS) were purchased from Procell (Wuhan, China). ALP kit was purchased from Beyotime (Shanghai, China). Hematoxylin and eosin (H&E) and Masson's trichrome staining kits were purchased from BOSTER (Wuhan, China). CD34 monoclonal anti-rat antibody was purchased from Abcam (Cambridge, MA, USA). MC3T3-E1 cells, Sprague Dawley (SD) rats and adult white New Zealand rabbits were supplied by Medical Center of Xi'an Jiaotong University (Xi'an, China). Tetrahydrofuran and n-hexane were dried by refluxing over sodium wire and distilled before use, while N,N-dimethylformamide, diethyl ether and triethylamine were dried by refluxing over calcium hydride and distilled before use. Simulated body fluid (SBF) was prepared by following Kokubo's protocol²⁵ and used to investigate the mineralization potential of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds in vitro. All other chemicals were of analytical grade, purchased from Sinopharm Chemical Reagent (Xi'an, China) and used without further purification. All aqueous solutions were prepared using ultrapure water with a resistance of 18.25 MΩ.

2.2 Preparation of PZL/PLGA/BG composite scaffold

To prepare scaffolds, PZL and BG particles were synthesized according to the reported processes in the literature.^26,27

PZL was synthesized as follows. N_ε-Benzyl formateoxycarbonyl-L-lysine-NCA was dissolved in anhydrous DMF under nitrogen atmosphere. Triethylamine (0.1% molar ratio of monomers) as an initiator was added and the mixture solution was stirred for 72 h at 30 °C. After the reaction had been completed, most of the solvent was removed by distillation under vacuum and the residue was poured into excessive ice-cold diethyl ether. Purified product as a pale yellow solid was obtained and dried under vacuum. (¹H NMR (CDCl₃, δ ppm): 1.35–1.5 (m, CH₂), 3.15 (m, CH and NH), 3.91 (s, NHCHCO), 5.1 (s, C₆H₅CH₂), 7.2 (s, C₆H₅)).

The typical procedure to synthesize BG powder was as follows. Tetraethyl orthosilicate and triethyl phosphate were mixed with 1 M HNO₃ solution and hydrolyzed for 60 min under vigorous stirring. Then, calcium nitrate tetrahydrate and sodium nitrate were gradually added into the mixture. After being stirred for 6 h, the resultant BG sol was allowed to age in a sealed container and then dried for 12 h, resulting in the formation of gel. After that, the gel was heat-treated at 550 °C for 2 h in order to remove the organic substances. The obtained BG was ball milled to obtain micro-scale particles. Since the content of SiO₂ in the BG was about 45 wt%, the BG was called 45S5 BG.

Negative NaCl templates were obtained by sintering NaCl particles with a size range of 100 to 400 μm in cylindrical alumina molds (15 mm inner diameter, 18 mm external diameter and 10 mm height) at 550 °C for 2 h.

To obtain PZL/PLGA/BG composite scaffolds, the templates were completely immersed into a mixture of PZL, PLGA and BG particles at different weight ratios in anhydrous CHCl₃ in a beaker. The beaker was kept under −0.08 MPa for 30 min and then the infiltrated templates were taken out, air dried, desalted with distilled water and dried under vacuum. When the weight ratios of PLGA and PZL were 0 : 1, 3 : 7 and 5 : 5, respectively, the obtained scaffolds were denoted as PZL, PZL7PLGA3 and PZL5PLGA5 scaffolds. In the case of BG addition to PZL5PLGA5, the scaffold was called as PZL5PLGA5BG composite scaffold. The weight ratio of BG particles and polymers was fixed at 1 : 10.

2.3 Characterization of PZL/PLGA/BG composite scaffolds

The chemical structure of the as-synthesized PZL was characterized by ¹H NMR spectroscopy with a Bruker 400 MHz spectrometer (Bruker, Germany) using CDCl₃ as the solvent and tetramethylsilane as the internal reference at 25 °C. Differential scanning calorimetry (DSC) curves were recorded from −40 to 150 °C on a differential scanning calorimeter (Netzsch DSC200 F3, Germany) in nitrogen atmosphere with a gas flow of 40 mL min⁻¹. DSC curves were used to determine glass transition temperature (T_g). The compression moduli of the scaffolds were measured using a universal testing instrument (Instron 5943, Norwood, MA, USA) at a constant loading rate of 1 mm min⁻¹. The compression moduli were calculated from stress–strain curves within the strain range from 10% to 20%. The morphology of the scaffolds was observed by means of scanning electron microscopy (SEM, S-3400 N, Hitachi, Japan). In vitro degradation tests were carried out in phosphate buffer solution (PBS, pH 7.4) in triplicates at 37 °C, and the degradation medium was replaced every day. The porosity of the scaffolds was measured based on the Archimedes' principle¹⁶ and ethanol was used as liquid medium. The porosity was calculated via the following formula:

P = (m₂ − m₀)/(m₂ − m₁) × 100%

where m₀ is the dry weight of the scaffolds, m₁ is the weight of scaffolds suspended in ethanol and m₂ is the weight of scaffolds saturated with ethanol. Five samples were tested to calculate the average porosity.

2.4 In vitro cell culture

MC3T3-E1 cells were used to evaluate the cytocompatibility of the scaffolds. MC3T3-E1 cells were cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin in a humidified atmosphere with 5% of CO₂ at 37 °C for 1, 3 and 5 days. Prior to seeding with cells, the scaffolds were sterilized with ethylene oxide and then placed in a 24-well culture plate. MC3T3-E1 cells suspended in 0.2 mL medium were seeded onto the surface of scaffold samples at a concentration of 3 × 10⁴ cells per well. After 4 h of culture, 1 mL of culture medium was added. Culture medium was replaced every day.

2.4.1 Live/dead cell viability and proliferation assay. The viability of MC3T3-E1 cells cultivated on the scaffolds was evaluated using the live/dead viability/cytotoxicity assay. At days 1, 3 and 5, the scaffold/cell constructs were clearly rinsed with PBS and stained with calcein AM and ethidium homodimer whose concentrations were 10 μM. The living cells were stained green and the dead cells were stained red. The scaffold/cell constructs were visualized using a confocal laser scanning microscope (Leica, TCS SP5 II, Germany).

2.4.2 MTT assay. The metabolic activity of cells incubated on the scaffolds was evaluated in triplicates using the MTT assay. After 1, 3 and 5 days of culture, the medium was removed and 100 μL of MTT solutions were added to each well in a 48-well plate. The scaffold/cell constructs were incubated for 4 h at 37 °C. After that, the medium was removed and 200 μL of DMSO was added to dissolve the blue formazan crystals. The solution was then transferred to a 96-well plate and the optical absorbency was measured at 490 nm using a microplate reader (PerkinElmer, EnSpire, America). Non-seeded scaffolds were used as a negative control.

2.4.3 ALP activity assay. To assess the ALP activity of MC3T3-E1 cells grown on the scaffolds, 3 × 10⁴ cells were seeded on each scaffold and cultured in a 24-well plate for 1, 3 and 5 days. At the predetermined time points, medium was decanted and the cell layer washed gently three times with PBS and once in cold 50 mM Tris buffer, and then cells were lysed in 200 μL of 0.2% Triton X-100. Lysates were sonicated after being centrifuged at 14 000 rpm for 15 min at 4 °C. 50 μL of supernatant was mixed with 150 μL of working solution according to the manufacturer's protocol. The conversion of p-nitrophenylphosphate into p-nitrophenol in the presence of ALP was determined by measuring the absorbance at 520 nm using a microplate reader (Bio-Rad 680, USA). The ALP activity was calculated from a standard curve after normalizing to the total protein content which was determined by a MicroBCA protein assay kit (Pierce) at 570 nm with a microplate reader (Bio-Rad 680, USA). The results were expressed in μM of p-nitrophenol produced per min per mg of protein.

2.5 BSA adsorption and release

Equal amounts of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds were immersed in 1 mL of 1 mg mL⁻¹ BSA solution in PBS. The system was shaken in an incubator at a speed of 150 rpm at 37 °C for 24 h, and then the samples were taken out and washed three times with PBS to remove any non-adsorbed protein and centrifuged at 14 000 rpm for 10 min. The BSA concentration in the supernatant was analyzed by means of a UV spectrophotometer (Persee, TU-1810PC, China) at a wavelength of 280 nm. The amount of BSA adsorbed on the scaffolds was calculated using a mass balance before and after adsorption.

To reveal the in vitro release profiles of BSA from the scaffolds, the scaffold samples were soaked in 10 mL of PBS and incubated at 37 °C in a constant-temperature incubator shaking at 150 rpm. The release medium was replaced with fresh PBS at predetermined time intervals. The release evaluation was performed for 10 days, and the amount of released BSA was measured using the UV method. The release data were expressed as a function of the cumulative release percentage of BSA (%, w/w) versus incubation time. All the experiments were done in triplicates.

2.6 In vitro mineralization

Both PZL5PLGA5 and PZL5PLGA5BG composite scaffolds were immersed in SBF and incubated at 37 °C for 30 days. The SBF was replaced every 2 days. At the end of the incubation period, any loosely bound minerals were removed by washing the scaffolds with deionized water. The scaffolds were dried and their morphology was observed using observed using a scanning electron microscope. To determine the phases of the mineral, mineralized samples were analyzed by X-ray diffractometer (PANalytical X'Pert Pro, the Netherlands) in the 2θ range of 10–80°, employing CuKα radiation source operated at 40 kV and 40 mA.

2.7 Release kinetics of Ca²⁺ and Si⁴⁺ from the PZL5PLGA5BG composite scaffolds

The release rates of Ca²⁺ and Si⁴⁺ ions from the PZL5PLGA5BG composite scaffolds soaked in PBS under quasi-dynamic conditions were measured using the ICP-AES (ICAP700, Thermo, USA). Each scaffold was immersed in 50 mL of PBS and enclosed in a test tube. The tube was incubated in a water bath at 37 °C and shaken at 150 rpm. The release study was performed for 21 days. At predetermined time points, the concentrations of Ca²⁺ and Si⁴⁺ in the soaking solutions were determined by comparing with the standard curve. To define it, five point calibrations (25, 10, 5, 1 and 0.1 ppm) were performed by diluting certified standards (Carl Roth, Germany) with PBS. Given errors were estimated by linear regression analysis.

2.8 Scaffold implantation into SD rat subcutaneous model

To assess the tissue ingrowth, angiogenesis and immunogenicity of the PZL5PLGA5 and PZL5PLGA5BG composite scaffolds in vivo, the scaffolds were embedded subcutaneously in SD rats with a body weight of 190–200 g. Both rats and rabbits were purchased from the Medical Center of Xi'an Jiaotong University (Xi'an, China). All the animal experiments were approved by the Institutional Animal Investigation Committee of Xi'an Jiaotong University. The animals were cared for in accordance with the principles of the Guide for the Care and Use of Laboratory Animals (National Institute of Health). PZL5PLGA5 and PZL5PLGA5BG composite scaffolds were sterilized with ethylene oxide and then implanted into the subcutaneous pockets on the backs of SD rats that were anesthetized by chloral hydrate solution (10% wt/wt, 4 μL g⁻¹ body weight). Three rats were used for each experiment. After the surgeries, cefazolin (25 μg g⁻¹ body weight) and buprenorphine (0.1 μg g⁻¹ body weight) were administered and each rat was individually housed. After being carefully nurtured in sterilized environment for 4 and 8 weeks, all SD rats were euthanized by injecting a lethal dose of anaesthetic via intraperitoneal injection for histological analysis.

2.9 Scaffold implantation into rabbit tibia defect model

Rabbit tibia defect model was used to evaluate the in vivo osteogenesis of the scaffolds. Six adult white New Zealand rabbits (male, 2–2.5 kg body weight) were anesthetized by injecting 3% Nembutal (30 mg kg⁻¹) via the ear vein, and a longitudinal incision was made by scalpel in the rabbit tibia under rigorous aseptic conditions. Quadrate holes (3 mm width, 8 mm length and 8 mm deep) were drilled using a surgical electronic drill and thoroughly rinsed with physiological saline to remove shards of bone. The ethylene oxide-sterilized scaffolds were implanted into the bone defects and wounds were sutured with nylon thread. Three rabbits were used for each experiment. After the surgery, cefazolin (25 μg g⁻¹ body weight) and buprenorphine (0.1 μg g⁻¹ body weight) were administered and each rabbit was individually housed. After being carefully nurtured for 1 and 8 weeks, the healing state of all rabbits were observed using digital radiography. Sequential radiographs of the tibia bone were recorded at regular intervals in a digital X-ray imaging system (5000CPlus, Mednova, China) to reveal the status of implants and the interaction between host bone and scaffold material.

2.10 Histological staining

After being explanted from SD rats, the implants were fixed overnight at room temperature in buffered formalin (37%, pH 7.2–7.4, Sigma-Aldrich) and embedded in paraffin, and then 7 μm sections were prepared. H&E, Masson's trichrome and CD34 staining assays were performed on 7 μm thick sections to observe the tissue response to implants and the ingrowth of tissue and microvessels. The histological sections were observed using a Leica SCN400 Slide Scanner and the images were captured using Leica Digital Image Hub (Leica, Germany). After 8 weeks of implantation, the implants harvested from rabbit bone defects were fixed in 4% paraformaldehyde for 48 h at 4 °C and then decalcified with 15% EDTA-2Na for 4 weeks. The samples were dehydrated with graded ethanol washes and embedded in paraffin, and serial longitudinal sections (7 μm in thickness) were prepared for H&E staining.

2.11 Statistical analysis

Experimental data were expressed as mean ± standard deviation. The statistical analysis for significance between the groups was performed by means of Student's t-test using SPSS13.0 software. A confidence level of 95% (p < 0.05) was considered to be statistically significant. Each experiment was done in triplicates.

3 Results and discussion

3.1 Scaffold morphology

The microstructure of foamy PZL, PZL7PLGA3, PZL5PLGA5 and PZL5PLGA5BG composite scaffolds was observed by SEM, and the typical SEM images are displayed in Fig. 2. As indicated in Fig. 2 that all scaffolds displayed a highly interconnected and bi-modal pore structure. The macropores were 540 ± 32 μm in size while the micropores were 40 ± 5 μm in size. It has been verified that the pores less than 200 μm could support the formation of microvascular networks at high densities and poor penetration. In contrast, the pores larger than 200 μm favored the formation of macrovascular networks at low densities and deep penetration depth.²⁸ Recently, some researchers found that the scaffolds with multi-scale pores supported osteogenic differentiation better than the scaffolds only with macropores.²⁹ In addition, it can be seen from the high magnification image in Fig. 2d that there were many BG particles with the size of 0.2–2.5 μm evenly distributed on the pore walls of the PZL5PLGA5BG composite scaffolds. Besides pore size, the porosity of scaffolds is also an important parameter to influence cell behavior.³⁰ The porosity of all scaffolds in this study was up to 90% that is higher than that of those scaffolds fabricated by traditional salt-leaching method.³¹ It was also noted that the addition of BG had a negligible effect on the pore structure and porosity of scaffolds. The reason for this may be that the negative NaCl templates had high strut connectivity and avoided the existence of residual NaCl particles in polymeric matrix during the salt-leaching process which ensured open porosity. Such a high porosity has been shown to be suitable for cellular and vascular infiltration of scaffolds.³²

Fig. 2 SEM images of PZL (a), PZL7PLGA3 (b), PZL5PLGA5 (c) and PZL5PLGA5BG (d) scaffolds.

3.2 Physicochemical properties of scaffolds

The compression stress–strain curves of the scaffolds are shown in Fig. 3a. Obviously, the compressive modulus of the scaffolds increased with increasing content of PLGA. The compressive moduli of PZL, PZL7PLGA3, PZL5PLGA5 and PZL5PLGA5BG composite scaffolds over the strain range of 10–20% were determined to be 93.5 ± 2.8, 195.2 ± 5.1, 242.8 ± 6.6 and 446.5 ± 5.5 kPa, respectively. They were sufficient for non-load-bearing applications as bone grafts.³³ Interestingly, the PZL5PLGA5BG composite scaffold displayed a higher compressive modulus than the PZL5PLGA5 scaffold, since the addition of BG particles led to improved mechanical properties. The reinforcing effect of BG particles within the porous scaffolds was also observed by other researchers.³⁴

Fig. 3 Compression stress–strain curves (a), DSC curves (b) and degradation profiles (c) of PZL, PZL7PLGA3, PZL5PLGA5 and PZL5PLGA5BG composite scaffolds (n = 3).

The DSC curves of the different materials are shown in Fig. 3b. As indicated in Fig. 3b, all DSC curves displayed only one endothermic peak attributed to T_g. The T_g of PZL7PLGA3, PZL5PLGA5 and PZL5PLGA5BG was 38.4 ± 0.2, 43.9 ± 0.4 and 48.5 ± 0.3 °C, respectively, which were between T_g of PZL (34.2 ± 0.2 °C) and T_g of PLGA (about 50 °C). At the same time, the T_g of the mixture of PZL and PLGA increased with increasing PLGA content, due to the relatively high T_g of PLGA.³⁵ These results clearly indicated that the compatibility between PZL and PLGA was very good. Some researchers pointed out that the T_g of scaffold materials played an important role in the cell growth behavior.³⁶ In addition, the addition of BG particles to PZL5PLGA5 led to a higher T_g, which is similar to that reported in the literature.³⁷

The scaffolds were expected to be degradable. Fig. 3c showed the degradation behaviors of different scaffolds. It was found that the weight of all scaffolds decreased nearly linearly as degradation time was prolonged. This was closely related to the multi-scale porosity of the foamy scaffolds in which water could easily penetrate into the inside of scaffolds.³³ After 8 weeks soaking in PBS, the weight loss of PZL, PZL7PLGA3, PZL5PLGA5 and PZL5PLGA5BG composite scaffolds was 20.4 ± 1.2%, 15.5 ± 1.1%, 12.2 ± 0.8% and 18.2 ± 1.5%, respectively. For the former three kinds of scaffolds, the degradation rate decreased with increasing PLGA content, because the weight loss of the scaffolds was mainly attributed to the removal of protective carbobenzoxy groups in PZL. The faster degradation rate of PZL5PLGA5BG composite scaffolds as compared to PZL5PLGA5 scaffolds was attributed to the enhanced water absorption ability caused by the addition of hydrophilic BG particles.^38,39 Water first penetrated along the interfaces of polymer and BG particles, and then the interface fractured along with the dissolution of BG particles and the degradation of polymer matrix, which facilitated water penetration and the leaching of small polymer chains.^40,41

3.3 In vitro biocompatibility evaluation

To evaluate the in vitro biocompatibility of the scaffolds, the viability, mitochondrial activity and osteogenic differentiation of MC3T3-E1 cells grown on the different scaffolds were characterized using the live/dead staining, MTT assay and ALP assay, and the results are presented in Fig. 4. As shown in Fig. 4a, after 1 day of incubation, a few green fluorescent dots representing viable cells were uniformly distributed in the scaffolds, and there was no significant difference between different scaffolds. After 3 days of culture, more green fluorescent dots were observed, and some larger green fluorescent dots appeared. When the culture time was prolonged to 5 days, there was an obvious increase in the intensity of green fluorescence, which suggested that the population of live cells became very much. It was also noted that the number of the cells on PZL5PLGA5BG composite scaffolds was larger than those on PZL, PZL7PLGA3 and PZL5PLGA5 scaffolds. These results coincided with those of MTT assay (Fig. 4b). After 3 and 5 days of culture, there was statistically significant difference between PZL5PLGA5BG composite scaffold group and other groups. The results suggested that the incorporation of BG into PZL5PLGA5 scaffolds could enhance the cell viability of MC3T3-E1 cells. A possible reason is that the addition of BG particles improved the surface roughness of pore struts and provided a favorable substrate to cells, which enhanced the adhesion, spreading and proliferation of MC3T3-E1 cells.^42,43

Fig. 4 Confocal micrographs (a), MTT activity (b) and ALP activity (c) of MC3T3-E1 cells cultured on PZL, PZL7PLGA3, PZL5PLGA5 and PZL5PLGA5BG scaffolds for 1, 3 and 5 days (n = 3).

The osteogenic differentiation of MC3T3-E1 cells cultured on the scaffolds was also evaluated by quantifying ALP activity, and the results are shown in Fig. 4c. It was found that the ALP activity of MC3T3-E1 cells cultured on all scaffolds increased with increasing culture time, and the ALP activity on the PZL5PLGA5BG composite scaffolds increased most quickly. This is expected because ALP is generally considered an early biomarker for osteoblastic differentiation of pre-osteoblasts.⁴⁴ A higher level of ALP activity reflects a more differentiated stage towards mature osteoblasts. A low level of ALP activity was detected after 1 day culture, but no significant difference was observed between the different scaffolds (p > 0.05). At days 3 or 5 after seeding, the MC3T3-E1 cells grown on the PZL5PLGA5BG composite scaffolds exhibited a much higher level of ALP activity than those on the PZL, PZL7PLGA3 and PZL5PLGA5 scaffolds (p < 0.05), possibly due to the presence of BG particles which may induce the osteoblastic differentiation of MC3T3-E1 cells to osteoblasts in vitro.⁴⁵ The results demonstrated that the PZL5PLGA5BG composite scaffolds could support the osteogenic differentiation of MC3T3-E1 cells into mature osteoblasts.

3.4 BSA adsorption and release

The surfaces of scaffolds in biological environments would be coated rapidly with the proteins that mediate the interaction between scaffold materials and cells by regulating the final cell behavior through complex signaling pathways.⁴⁶ Therefore, the quantitative characterization on how microstructure to determine the amount, structure and distribution of adsorbed proteins is necessary for understanding cell-microstructured surface interaction. In this study, the adsorption and release kinetics of BSA on PZL5PLGA5 and PZL5PLGA5BG composite scaffolds were studied, since BSA may affect the adhesion and proliferation of cells on scaffolds.⁴⁷ Fig. 5a shows the capacities of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds to adsorb BSA. As shown in Fig. 5a, their BSA adsorption capacities were 33 ± 2.3 and 45 ± 3.5 mg g⁻¹, respectively, and a significant difference was observed between PZL5PLGA5 and PZL5PLGA5BG groups (p < 0.05). Since a little dose is required for bone scaffolds, such protein adsorption capacities are high enough.⁴⁸ The amount of BSA adsorbed onto the scaffolds was influenced by the surface roughness and crystallinity of scaffold materials.^46,49 The BSA adsorption capacity of the PZL5PLGA5BG scaffolds was higher than that of PZL5PLGA5 scaffolds, because the former had coarser strut surface than the latter due to the presence of BG particles, which increased the binding sites on the material surface for proteins.⁵⁰

Fig. 5 Adsorption capacities of (a) and cumulative release profiles of BSA from (b) PZL5PLGA5 and PZL5PLGA5BG composite scaffolds in PBS at 37 °C.

Fig. 5b displays the release profiles of BSA from PZL5PLGA5 and PZL5PLGA5BG scaffolds. It can be seen from Fig. 5b that the release profiles of BSA displayed two stages: an initial fast release followed by a slow release. After 1 day of incubation, approximately 38.1 ± 1.5% and 32.2 ± 1.5% of BSA was released from the PZL5PLGA5 and PZL5PLGA5BG groups, respectively, while the percentages of BSA released from the two kinds of scaffolds after 10 days were 80.1 ± 2.9% and 72.1 ± 2.1%, respectively. On the whole, the release rate of BSA from the PZL5PLGA5BG scaffolds was slightly lower than that from the PZL5PLGA5 scaffolds, though their porosity was similar. The reason might be that BSA was adsorbed onto the PZL5PLGA5BG composite scaffolds through the formation of hydrogen bonds between Si–OH groups of BG particle surfaces and negatively charged BSA, but was adsorbed on the surface of pore struts of the PZL5PLGA5 scaffolds via van der Waals forces.⁵¹ This led to the slower release from the PZL5PLGA5BG composite scaffolds compared to the PZL5PLGA5 scaffolds.

3.5 Biomineralization

The apatite-forming ability of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds in SBF was characterized by SEM and XRD. Fig. 6a presents the typical SEM images of strut surfaces of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds after being immersed into SBF for 1 month. No mineralized particles were observed on the strut surface of PZL5PLGA5 scaffolds, while the strut surfaces of PZL5PLGA5BG scaffolds were covered with a great number of small mineralized particles. These cluster-like particles were 1–2 μm in size and composed of hydroxyapatite crystals, as evident from the presence of characteristic diffraction peaks (25.9°, 32.2° and 49.5°) of hydroxyapatite in XRD pattern (Fig. 6b). The results indicated that PZL5PLGA5 itself had little or no osteoinductivity while the BG exhibited excellent osteoinductivity.⁵² The possible mineralization mechanism is that the BG particles exposed on the strut surfaces released Ca²⁺ or/and PO₄³⁻ which interacted with PO₄³⁻ or/and Ca²⁺ in SBF, leading to the deposition of hydroxyapatite.⁴³ In addition, the strut surfaces of the two kinds of scaffolds were very rough, indicating that the scaffolds had partially degraded.

Fig. 6 SEM images (a) and XRD patterns (b) of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds after 1 month of immersing in SBF; release behaviors (c) of Ca²⁺ and Si⁴⁺ from the PZL5PLGA5BG composite scaffolds soaked in PBS.

The simultaneous release of Ca²⁺ and Si⁴⁺ from BG can greatly promote osteoblast differentiation and biomineralization. In this study, the concentrations of Ca²⁺ and Si⁴⁺ released from the PZL5PLGA5BG composite scaffolds in PBS were monitored for 21 days (Fig. 6c). It was found that the concentrations of Ca²⁺ and Si⁴⁺ significantly increased within the first 14 days and then the ion concentrations became steady. By comparing with the release rates of Ca²⁺ and Si⁴⁺ from the materials reported in the literature,⁵³ the PZL5PLGA5BG composite scaffolds showed relatively lower release rates. This is attributed to the slower degradation rates of PZL5PLGA5.

3.6 H&E, Masson's trichrome and CD34 staining of implants in rat subcutaneous model

To explore the tissue response to PZL5PLGA5 and PZL5PLGA5BG composite scaffolds in SD rat model, histological analysis including H&E, Masson's trichrome and CD34 staining was carried out after 4 and 8 weeks of implantation, and the results are presented in Fig. 7. The transplantation and extraction processes of scaffolds were shown in Fig. 7a. Fig. 7b shows the H&E staining images of tissue/scaffold constructs. As displayed in Fig. 7b, the formation of mass tissue was observed in all scaffolds, and the presence of irregular blank areas suggested that the scaffolds had begun to degrade. Inflammatory cells infiltrated into the tissues surrounding the scaffolds after 4 weeks of implantation. At week 8 post-implantation, the increased proportion of blank areas meant that more scaffold materials degraded with increasing implantation time. The number of inflammatory cells in the scaffolds obviously decreased.

Fig. 7 Surgery photos of composite scaffolds subcutaneously implanted into SD rats (a); H&E (b), Masson's trichrome (c) and CD34 (d) staining of PZL5PLGA5 and PZL5PLGA5BG scaffolds after 4 and 8 weeks of implantation. Black arrows represent microvessels; M represents material; T represents tissue; brown dots represent inflammatory cells; orange dots represent erythrocytes.

Interestingly, the number of inflammatory cells in PZL5PLGA5BG scaffolds reduced more significantly than that in PZL5PLGA5 scaffolds. As shown in the Masson's trichrome staining images presented in Fig. 7c, after 4 weeks of implantation, the presence of inflammatory cells indicated the scaffolds activated the host immune response. At 8 weeks after the implantation, more inflammatory cells were observed on the PZL5PLGA5 scaffolds as compared to the PZL5PLGA5BG scaffolds. This may be due to the inflammation reaction induced by the acidic degradation products of PLGA.^54,55 However, in the case of PZL5PLGA5BG composite scaffolds, the reduction in pH value in the surrounding tissue caused by the decomposed products of PLGA could be buffered by the ions released from BG particles,^56,57 which reduced the tissue response.

Blood vessels ingrown in scaffolds will facilitate input of nutrients and oxygen to cells to survive and output of waste products from cell metabolism. The images of CD34-stained microvessels are shown in Fig. 7d. The positive expression of CD34 was observed for the two kinds of scaffolds, and the number of microvessels increased when the implantation time was prolonged from 4 to 8 weeks. Compared to the PZL5PLGA5 composite scaffolds, the PZL5PLGA5BG composite scaffolds showed more microvessels, demonstrating that the addition of BG facilitated microvessel ingrowth into the scaffolds.⁵⁸ The reason for this is that the Ca²⁺ and Si⁴⁺ released from BG could induce positive angiogenic or osteogenic responses and actively affect the function and growth behavior of endothelial cells.^59,60 Furthermore, BG could enhance the expression levels of endothelial gene markers such as CD34, CD31/PECAM1 and VEGFR2 in stromal cells.⁶¹

3.7 Digital radiography and H&E staining of implants in rabbit tibia defect model

The osteogenesis ability of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds in rabbit tibia defect model was investigated by digital radiography and histological analysis, and the results are displayed in Fig. 8. Fig. 8a shows the transplantation and extraction processes of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds. Fig. 8b presents the digital radiographs of the bony defects implanted with PZL5PLGA5 and PZL5PLGA5BG composite scaffolds after 1 week and 8 weeks post operatively. The radiographs at 1 week after implantation showed that the rectangular implants occupied the distal metaphysis of the rabbit tibia, resulting in a rectangular black area. After 8 weeks of implantation, the radiographs exhibited a homogeneous filling of bony structure. A clear irregular gap between the bone tissue and the implant was still observed for the PZL5PLGA5 scaffold. In contrast, the bone defects implanted with PZL5PLGA5BG composite scaffolds exhibited markedly higher defect filling, because BG could promote bone regeneration and bone defect healing by enhancing angiogenesis and osteogenesis.⁵⁸ Quantitatively, the average percentage of newly formed bone filling the defect region for the PZL5PLGA5BG scaffolds was 1.6 times higher than that for the PZL5PLGA5 scaffolds, respectively.

Fig. 8 Representative operation images (a); digital radiographs of tibia defects filled with PZL5PLGA5 and PZL5PLGA5BG composite scaffolds at 1 week and 8 weeks post-operatively (b); and H&E staining images of scaffold implants after 8 weeks of implantation (c).

Fig. 8c shows the histological evaluation results of the PZL5PLGA5 and PZL5PLGA5BG composite scaffolds implanted in the rabbit tibia bone defects for 8 weeks, respectively. In routine H&E staining, the newly formed bone and residual material are indicated in carmine and pink, respectively. After implantation for 8 weeks, the cells in marrow cavity including lipocytes, hemocytes and inflammatory cells had infiltrated into the inside of PZL5PLGA5 and PZL5PLGA5BG composite scaffolds. The diffusion footprints at the interface of the tissue and material were clearly seen, which demonstrated the good in vivo biocompatibility of the scaffolds. In addition, the bioresorption of the implants was observed. Compared with the PZL5PLGA5 scaffold implants, new bone tissues were found in the PZL5PLGA5BG scaffold implants, which further verified the osteoinductivity of BG.⁵² Some researchers considered that the ionic dissolution products of BG were capable of directly inducing the expression of genes relevant to osteoblast metabolism and the maintenance of extracellular matrix.⁶² In addition, BG has shown considerable potential in actively promoting osteoblast differentiation via enhancing gene expression of osteopontin, osteonectin, core binding factor alpha 1, osterix, collagen type I, alkaline phosphatase and bone sialoprotein.^63,64

4 Conclusions

A foamy PZL/PLGA/BG scaffold composed of PZL, PLGA and BG was successfully fabricated via a negative NaCl template method. Blending modification of synthetic polypeptide PZL with PLGA and BG particles effectively improved its mechanical performance and osteogenic ability. The obtained scaffolds had highly interconnected porosity, desirable compression modulus and T_g, appropriate degradation rate, good in vitro biocompatibility, and satisfactory protein adsorption and mineralization abilities. The results from studies of rat subcutaneous and rabbit tibia defect models demonstrated that the PZL5PLGA5BG composite scaffolds could effectively support tissue and vascular ingrowth and had tolerable inflammation response and that the scaffolds could efficiently promote bone regeneration through enhancing angiogenesis and osteogenesis. In a word, the novel PZL5PLGA5BG composite scaffolds had desirable physicochemical properties and excellent biocompatibility and osteogenesis ability, and hold great potential for bone tissue engineering applications.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (50773062, 50603020), the Natural Science Basic Research Plan in Shaanxi Province of China (2013K09-27) and the Fundamental Research Funds for the Central Universities (XJJ2014124, XJJ2013130).

Notes and references

1. L. Moroni, A. Nandakumar, F. B. Groot, C. A. van Blitterswijk and P. Habibovic, J. Tissue Eng. Regener. Med., 2015, 9, 745–759.
2. S. H. Kim, S. H. Kima and Y. Jung, J. Controlled Release, 2015, 206, 101–107.
3. E. Quinlan, A. Lopez-Noriega, E. Thompson, H. M. Kelly, S. A. Cryan and F. J. O'Brien, J. Controlled Release, 2015, 198, 71–79.
4. E. Vardar, H. M. Larsson, E. M. Engelhardt, K. Pinnagoda, P. S. Briquez, J. A. Hubbell and P. Frey, Acta Biomater., 2016, 41, 75–85.
5. R. A. Perez, S. J. Seo, J. E. Won, E. J. Lee, J. H. Jang, J. C. Knowles and H. W. Kim, Mater. Today, 2015, 18, 573–589.
6. A. Khojasteh, F. Fahimipour, M. B. Eslaminejad, M. Jafarian, S. Jahangir, F. Bastami, M. Tahriri, A. Karkhaneh and L. Tayebi, Mater. Sci. Eng., C, 2016, 69, 780–788.
7. J. Shen, O. Velasco, K. Khadarian, G. Asatrian, A. Hwang, Y. Zhang, J. Kwak, C. Chawan, K. Khalilinejad, M. Ajalat, T. Pritchard, X. Zhang, A. W. James and C. Soo, Int. J. Orthop., 2015, 2, 468–475.
8. D. Ben-David, S. Srouji, K. Shapira-Schweitzer, O. Kossover, E. Ivanir, G. Kuhn, R. Muller, D. Seliktar and E. Livne, Biomaterials, 2014, 35, 1869–1881.
9. J. J. Yin, K. Selander, J. M. Chirgwin, M. Dallas, B. G. Grubbs, R. Wieser, J. Massague, G. R. Mundy and T. A. Guise, J. Clin. Invest., 1999, 103, 197–206.
10. N. Cui, J. M. Qian, J. L. Wang, C. L. Ji, W. J. Xu and H. J. Wang, RSC Adv., 2016, 6, 73699–73708.
11. Z. X. Liao, S. F. Peng, Y. L. Chiu, C. W. Hsiao, H. Y. Liu, W. H. Lim, H. M. Lu and H. W. Sung, J. Controlled Release, 2014, 193, 304–315.
12. T. Liu, W. Xue, B. Ke, M. Q. Xie and D. Ma, Biomaterials, 2014, 35, 3865–3872.
13. H. Wu, L. Zhu and V. P. Torchilin, Biomaterials, 2013, 34, 1213–1222.
14. J. Zhang, Z. F. Yuan, Y. Wang, W. H. Chen, G. F. Luo, S. X. Cheng, R. X. Zhuo and X. Z. Zhang, J. Am. Chem. Soc., 2013, 135, 5068–5073.
15. H. L. Xu, Q. Yao, C. F. Cai, J. X. Gou, Y. Zhang, H. J. Zhong and X. Tang, J. Controlled Release, 2015, 199, 84–97.
16. J. M. Qian, X. Q. Yong, W. J. Xu and X. X. Jin, Mater. Sci. Eng., C, 2013, 33, 4587–4593.
17. K. A. Gross and L. M. Rodriguez-Lorenzo, Biomaterials, 2004, 25, 4955–4962.
18. X. L. Ding, X. Wei, Y. Huang, C. D. Guan, T. Q. Zou, S. Wang, H. F. Liu and Y. B. Fan, J. Mater. Chem. B, 2015, 3, 3177–3188.
19. M. Ionita, L. E. Crica, H. Tiainen, H. J. Haugen, E. Vasile, S. Dinescu, M. Costache and H. Iovu, J. Mater. Chem. B, 2016, 4, 282–291.
20. P. Gentile, V. Chiono, I. Carmagnola and P. V. Hatton, Int. J. Mol. Sci., 2014, 15, 3640–3659.
21. S. E. Smith, R. A. White, D. A. Grant and S. A. Grant, J. Nanosci. Nanotechnol., 2016, 16, 1160–1169.
22. M. Bohner, L. Galea and N. Doebelin, J. Eur. Ceram. Soc., 2012, 32, 2663–2671.
23. Q. Z. Chen, I. D. Thompson and A. R. Boccaccini, Biomaterials, 2006, 27, 2414–2425.
24. H. Y. Li, J. He, H. F. Yu, C. R. Green and J. Chang, Biomaterials, 2016, 84, 64–75.
25. T. Kokubo and H. Takadama, Biomaterials, 2006, 27, 2907–2915.
26. W. N. E. van Dijk-Wolthuis, L. van de Water, P. van de Weterung, M. J. van Steenbergen, J. J. Kettenes-van den Bosch, W. J. Schuyl and W. E. Hennink, Macromol. Chem. Phys., 1997, 198, 3893–3906.
27. W. Li, P. Nooeaid, J. A. Roether, D. W. Schubert and A. R. Boccaccini, J. Eur. Ceram. Soc., 2014, 34, 505–514.
28. S. W. Choi, Y. Zhang, M. R. MacEwan and Y. N. Xia, Adv. Healthcare Mater., 2013, 2, 145–154.
29. A. J. Wang, T. Paterson, R. Owen, C. Sherborne, J. Dugan, J. M. Li and F. Claeyssens, Mater. Sci. Eng., C, 2016, 67, 51–58.
30. Q. L. Loh and C. Choong, Tissue Eng., Part B, 2013, 19, 485–502.
31. C. Correia, S. Bhumiratana, L. P. Yan, A. L. Oliveira, J. M. Gimble, D. Rockwood, D. L. Kaplan, R. A. Sousa, R. L. Reisb and G. Vunjak-Novakovica, Acta Biomater., 2012, 8, 2483–2492.
32. E. Quinlan, S. Partap, M. M. Azevedo, G. Jell, M. M. Stevens and F. J. O'Brien, Biomaterials, 2015, 52, 358–366.
33. N. Cui, J. M. Qian, T. Liu, N. Zhao and H. J. Wang, Carbohydr. Polym., 2015, 126, 192–198.
34. A. Gantar, L. P. da Silva, J. M. Oliveira, A. P. Marques, V. M. Correlo, S. Novak and R. L. Reis, Mater. Sci. Eng., C, 2014, 43, 27–36.
35. H. Fouad, T. Elsarnagawy, F. N. Almajhdi and K. A. Khalil, Int. J. Electrochem. Sci., 2013, 8, 2293–2304.
36. T. Serra, J. A. Planell and M. Navarro, Acta Biomater., 2013, 9, 5521–5530.
37. M. Kouhi, M. Morshed, J. Varshosaz and M. H. Fathi, Chem. Eng. J., 2013, 228, 1057–1065.
38. A. Larranaga, P. Aldazabal, F. J. Martin and J. R. Sarasua, Polym. Degrad. Stab., 2014, 110, 121–128.
39. A. R. Baccaccini, J. J. Blaker, V. Maquet, R. M. Day and R. Jerome, Mater. Sci. Eng., C, 2005, 25, 23–31.
40. G. Vergnol, N. Ginsac, P. Rivory, S. Meille, J. M. Chenal, S. Balvay, J. Chevalier and D. J. Hartmann, J. Biomed. Mater. Res., Part B, 2016, 104, 180–191.
41. J. Rich, T. Jaakkola, T. Tirri, T. Narhi, A. Yli-Urpo and J. Seppala, Biomaterials, 2002, 23, 2143–2150.
42. S. F. Wang, D. H. R. Kempen, G. C. W. de Ruiter, L. Cai, R. J. Spinner, A. J. Windebank, M. J. Yaszemski and L. H. Lu, Adv. Funct. Mater., 2015, 25, 2715–2724.
43. Q. Chen, R. P. Garcia, J. Munoz, U. P. de Larraya, N. Garmendia, Q. Q. Yao and A. R. Boccaccini, ACS Appl. Mater. Interfaces, 2015, 7, 24715–24725.
44. J. M. Qian, W. J. Xu, X. Q. Yong, X. X. Jin and W. Zhang, Mater. Sci. Eng., C, 2014, 36, 95–101.
45. S. Zahid, A. T. Shah, A. Jamal, A. A. Chaudhry, A. S. Khan, A. F. Khan, N. Muhammad and I. Ur Rehman, RSC Adv., 2016, 6, 70197–70214.
46. P. E. Scopelliti, A. Borgonovo, M. Indrieri, L. Giorgetti, G. Bongiorno, R. Carbone, A. Podesta and P. Milani, PLoS One, 2010, 5, e11862.
47. D. Depan and R. D. K. Misra, Acta Biomater., 2013, 9, 6084–6094.
48. N. Li, C. Jiang, X. Zhang, X. Gu, J. Zhang, Y. Yuan, C. Liu, J. Shi, J. Wang and Y. Li, J. Mater. Chem. B, 2015, 3, 8558–8566.
49. W. H. Lee, A. V. Zavgorodniy, C. Y. Loo and R. Rohanizadeh, J. Biomed. Mater. Res., Part A, 2012, 100, 1539–1549.
50. M. Peter, N. S. Binulal, S. V. Nair, N. Selvamurugan, H. Tamura and R. Jayakumar, Chem. Eng. J., 2010, 158, 353–361.
51. X. D. Zhang, D. L. Zeng, N. Li, X. Q. Jiang, C. S. Liu and Y. S. Li, J. Mater. Chem. B, 2016, 4, 3916–3924.
52. A. El-Fiqi, J. H. Kim and H. W. Kim, ACS Appl. Mater. Interfaces, 2015, 7, 1140–1152.
53. H. Y. Li, J. He, H. F. Yu, C. R. Green and J. Chang, Biomaterials, 2016, 84, 64–75.
54. D. Jeong, C. S. Kang, E. Jung, D. Yoo, D. M. Wu and D. W. Lee, J. Controlled Release, 2016, 233, 72–80.
55. J. W. M. Hoekstra, J. L. Ma, A. S. Plachokova, E. M. Bronkhorst, M. Bohner, J. L. Pan, G. J. Meijer, J. A. Jansen and J. J. J. P. van den Beucken, Acta Biomater., 2013, 9, 7518–7526.
56. S. J. Chen, Z. Y. Jian, L. S. Huang, W. Xu, S. H. Liu, D. J. Song, Z. M. Wan, A. Vaughn, R. S. Zhan, C. Y. Zhang, S. Wu, M. H. Hu and J. S. Li, Int. J. Nanomed., 2015, 10, 3815–3827.
57. A. M. Haaparanta, P. Uppstu, M. Hannula, V. Ella, A. Rosling and M. KellomAki, Mater. Sci. Eng., C, 2015, 56, 457–466.
58. E. Quinlan, S. Partap, M. M. Azevedo, G. Jell, M. M. Stevens and F. J. O'Brien, Biomaterials, 2015, 52, 358–366.
59. H. Y. Li, J. He, H. F. Yu, C. R. Green and J. Chang, Biomaterials, 2016, 84, 64–75.
60. A. Hoppe, N. S. Guldal and A. R. Boccaccini, Biomaterials, 2011, 32, 2757–2774.
61. R. El-Gendy, J. Kirkham, P. J. Newby, Y. Mohanram, A. R. Boccaccini and X. B. Yang, Tissue Eng., Part A, 2015, 21, 2034–2043.
62. A. Hoppe, N. S. Guldal and A. R. Boccaccini, Biomaterials, 2011, 32, 2757–2774.
63. P. S. P. Poh, D. W. Hutmacher, B. M. Holzapfel, A. K. Solanki, M. M. Stevens and M. A. Woodruff, Acta Biomater., 2016, 30, 319–333.
64. O. Tsigkou, J. R. Jones, J. M. Polak and M. M. Stevens, Biomaterials, 2009, 30, 3542–3550.

---