International Journal of Biological Macromolecules 79 (2015) 736–747

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules journal homepage: www.elsevier.com/locate/ijbiomac

Preparation and characterization of vanillin-crosslinked chitosan therapeutic bioactive microcarriers Qin Zou a , Junfeng Li b , Yubao Li a,∗ a b

Research Center for Nano-Biomaterials, Analytical & Testing Center, Sichuan University, Chengdu 610064, China Department of Materials Science & Engineering, Chengdu University of Technology, Chengdu 610059, China

a r t i c l e

i n f o

Article history: Received 27 December 2014 Received in revised form 14 May 2015 Accepted 21 May 2015 Available online 4 June 2015 Keywords: Vanillin-crosslinked chitosan microspheres Drug delivery Biocompatibility

a b s t r a c t Chitosan microspheres with diameter of 14.3–48.5 ␮m were prepared by emulsion method and using natural vanillin as cross-linking agent. The surface morphology and microstructure of the microspheres were characterized by scanning electron microscopy, X-ray diffraction and Fourier-transform infrared spectroscopy, etc. The hollow microspheres showed a well-defined spherical shape with median diameter of 30.3 ␮m and possessed a uniform surface with micro-wrinkles, which is in favor of the drug release. Interpenetrating network cross-linking mechanism might result from the Schiff base reaction and the acetalization of hydroxyl and carbonyl between chitosan and vanillin. Berberine, as a model drug, was loaded in the microspheres and released in a sustainable manner. The drug loading ratio could change from 9.16% to 29.70% corresponding to the entrapment efficiency of 91.61% to 74.25%. In vitro cell culture study using MG63 cells and in vivo implantation clearly showed that the microspheres could provide favorable cell attachment and biocompatibility. The results confirm that the drug-loaded vanillincrosslinked chitosan microspheres could be a worthy candidate either as carriers of drugs and cells, or as therapeutic matrix for bone repair and regeneration. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Osteomyelitis with bone loss presents a special challenge for treatment and reconstruction. A staged reconstruction technique consists of initial debridement and local–systemic antibiotic therapy, followed by bone reconstruction [1]. An ideal system for local delivery of antibiotics should provide sustained delivery of higher concentration of antibiotics by diffusion to avascular area and yet minimize the risk of systemic toxicity associated with traditional method of intravenous delivery [2]. Surgical removal of infected bone often results in considerable bone loss and skeletal deficiency. Small bone defects can spontaneously regenerate to the original anatomic configuration, however, regeneration of defects that exceed a certain size is difficult to reconstruct following resolution of the infection. In the final stage of reconstruction, bone substitutes have to be selected by clinicians to treat large defect cases [3]. To solve these problems, polymer based drug delivery systems have been developed. A major benefit of bioactive carrier system is the ability to regenerate bone in the defect while allowing local antibiotic release at a sustainable or controllable rate. This combination

∗ Corresponding author. Tel.: +86 28 85412847. E-mail address: [email protected] (Y. Li). http://dx.doi.org/10.1016/j.ijbiomac.2015.05.037 0141-8130/© 2015 Elsevier B.V. All rights reserved.

could eliminate the need for subsequent bone reconstruction to fill large bone defects [4]. At the moment, the microsphere technology is the core technology designed to fill the defects and to release therapeutic drugs in a predictable manner [5]. The microspherical particles possess high physicochemical integrity; loading of drugs can be achieved through chemical crosslinking, blending or simple adsorption [6], and bone-bonding bioactive apatite crystals can be incorporated during preparation. In particular, the microspheres can be refined by varying polymer molecular weight and concentration, degree of crosslinking, or by chemical modification of the polymer matrix to achieve sustained and controlled drug release [7]. Chitosan (CS) has presented great potential applications in adsorption and isolation of protein, catalytic carrier, enzyme immobilization, and controlled drug release in the form of fibers, membranes, microspheres, and capsules [8]. The most attractive properties of CS are related to its biodegradability and good biocompatibility, which makes CS and its derivatives be extensively used in biomedical fields, such as for wound healing, drug delivery and tissue engineering, particularly for developing nano-/microspheres as carrier systems [9]. As a co-polymer, CS is made up of linear ˇ-(1 → 4) glycosidic linkage which is similar in structure to cellulose [10]. In addition, CS has also been shown to facilitate cell adhesion and proliferation, and osteogenic differentiation of

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

mesenchymal stem cells [11]. CS has one primary amino and two free hydroxyl groups for each C6 unit. Cross-linking is a common way to modify the CS structure and improve its controlled-release and mechanical properties [12]. The available amino and hydroxyl groups on CS are active sites capable of forming a number of linkages [13]. In cross-linked CS, the polymer chains are interconnected by cross linkers, forming a three dimensional network structure. To date, various dialdehydes, such as glyoxal and glutaraldehyde [14], are used to perform the covalent cross-linking on NH2 sites, forming stable imine bonds between amine groups of CS and aldehyde groups via a Schiff reaction [15]. Researchers have evaluated cross-linking of chitosan microspheres with glutaraldehyde as well as glyoxyal for the controlled delivery of centchroman. Their studies demonstrated that drug release rates may be changed not only by the degree of cross-linking of the microspheres, but also by the type of cross-linker used [16]. However, the main drawback of dialdehydes is related to their toxicity [17]. For example, glutaraldehyde can cause irritation to mucosal membranes because of its toxicity. Therefore, considering the biocompatibility, many investigations have developed new reagents to cross-link chitosan. Researchers also reported that chitosan microspheres showed a sustained release for centchroman for 50 h with sodium hexameta polyphosphate (SHMP) as physical cross-linker [18]. However, physically crosslinked microspheres have exhibited inferior release properties due to their weak electrostatic interactions between anions and chitosan. Therefore, it is necessary and crucial to search for alternative cross-linkers for chitosan microspheres. Currently, vanillin (4-hydroxy-3- methoxybenzaldehyde) is a popular flavor extracts broadly used in food [19]. The aldehyde groups in vanillin and the amino groups in CS may undergo Schiff base reaction and form a network structure to favor the stabilization and controlled release [20]. Its biocompatibility when used for biomaterial preparation needs to be further assessed. The objective of the present study is to prepare the CS microspheres by emulsion method, using vanillin as the cross linker. The physicochemical properties and in vitro drug release behavior of the microspheres were characterized and tested. The biocompatibility of the vanillin cross-linked CS microspheres was evaluated via in vitro cell culture and in vivo animal experiment using glutaraldehyde cross-linked CS microspheres as the control. 2. Materials and methods 2.1. Materials CS was obtained from Jinan Haidebei Marine Bioengineering Co. Ltd. (Shandong, China) with 90% deacetylation degree. Aqueous acetic acid solution was used as the solvent for CS microspheres preparation. Vanillin and glutaraldehyde (Aladdin Co. Ltd., China) were used as the cross linkers. All chemicals and reagents (liquid paraffin, Span 80, petroleum ether (60–90 ◦ C), isopropyl alcohol, etc.) used in the experiments were of analytical grade.

microspheres were collected with centrifugation, and fully washed with petroleum ether followed by dimethyl carbinol to remove the residual liquid paraffin and acetone (Fig. 1). The vanillin crosslinked CS microspheres (VCM) and glutaraldehyde cross-linked CS microspheres (GCM) were dried at 50 ◦ C for 12 h in an oven in air. The drug loaded vanillin cross-linked CS microspheres (DVCM) and drug loaded glutaraldehyde cross-linked CS microspheres (DGCM) were prepared separately by mixing the berberine into CS solution, followed by a similar procedure mentioned above. 2.3. Characterization of CS microspheres 2.3.1. Particle size analysis The particle-size distribution of cross-linked CS microspheres was measured by laser diffractometry. The microspheres were redispersed in 500 ml distilled water and sized by laser particle sizer (BT-9300H, China). 2.3.2. Morphology The optical microscope (Nikon TE-2000, Japan) was used for shape determination of the as-prepared crosslinked CS microspheres. The scanning electron microscopy (SEM, JEOL, JEM-100CX, Japan) was used to observe the microspheres dried at 60 ◦ C, and observe the dried drug-loaded CS microspheres which were embedded in methyl methacrylate, cutting into sections of 1 mm in thickness. Samples for SEM examination were sputter coated with gold before observation. 2.3.3. X-ray diffraction analysis The phase composition and crystallinity of microsphere samples were analyzed by X-ray diffraction (XRD, Philips X’Pert Pro MPD, Netherlands). XRD patterns were recorded through a 2 range from 5◦ to 58◦ at a rate of 2◦ /min. 2.3.4. Fourier transform infrared spectroscopy The infrared spectra of pure CS and cross-linked CS microspheres were collected on an attenuated total reflectance Fourier transform infrared spectrometry (TENSOR27, BRUKER Co. Germany). All spectra were recorded by transmittance mode (100 times scanning, 400–4000 cm−1 ). 2.4. Swelling measurement The swelling property of the CS microspheres was determined by immersion method. 200 mg of dry microspheres without drug were put into a beaker containing 20 ml PBS (phosphate buffered solution, 0.l M, pH 7.4) and shaking occasionally at 37 ◦ C. At predetermined time intervals, the swollen microspheres were removed from the solution, collected by centrifugation and weighed immediately after removing surface water with filter paper. The degree of swelling (Sw ) was calculated using the following equation: Sw (%) =

2.2. Preparation of microspheres CS solution (3%, w/v) was prepared by dissolving 3 g of CS in 100 ml acetic acid solution (1%, v/v) at room temperature. The solution was stirred at 1000 rpm for 30 min to ensure complete dissolution of CS. Then the solution (20 g) was poured into the liquid paraffin suspension medium (60 ml), which contained span80 (0.9 g), tween80 (0.3 g) and magnesium stearate (0.1 g), heated at 50 ◦ C and stirred at 1800 rpm. The formed microspheres were chemically cross-linked by vanillin or glutaraldehyde, i.e., the cross linker acetone solution (10%, v/v) was added dropwise into the suspension mixture and stirred for 3 h. At the end, the CS

737

W − W  t 0 W0

× 100

(1)

where Wt and W0 represent the weight of the swollen and dry samples, respectively [21]. Each swelling test was done intriplicate. To simulate the flow of a biological liquid, the buffer solution was replaced at each time point. 2.5. In vitro drug release 2.5.1. Drug loading ratio and entrapment efficiency of microspheres The microspheres (10 mg) containing drug were added into 100 ml anhydrous alcohol and heated with reflux condensation at

738

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

Fig. 1. Schematic diagram for preparation of CS microspheres.

85 ◦ C for 8 h. After cooling down to room temperature and centrifugation, the amount of berberine released in solution was analyzed by means of UV–VIS spectroscopy with a spectrophotometer (SP2100UVPC, China). Absorbance values were taken at a wavelength of 421 nm at which the berberine showed an absorbance maximum. Each value represents an average of three runs. Values of drug concentration were calculated in mg/ml eluant using a conversion factor obtained from the gradient of a calibration curve of berberine, which was determined by taking absorbance vs. berberine concentration between 0.01 and 0.5 mg/ml as parameters. C = (A + 0.00062)/0.01221 (r = 0.99982). Where A is the absorbance, C is the concentration (mg/ml) and the r is correlation coefficient. Eluants from the control samples (crosslinked CS microspheres without drug) were read in the same way as test strips. The absorbance of control strip eluants was subtracted from the readings of the test eluants to show only drug change of specimens. The drug loading ratio and entrapment efficiency of microspheres were calculated using Eqs. (2)–(4) respectively. Drug loading ratio(%) =

W1 × 100 W0

Theoretical drug loading ratio(%) = Entrapment efficiency(%) =

(2) W2 × 100 W3

Drug loading ratio × 100 Theoretical loading ratio

(3) (4)

where Wl is the weight of berberine loaded in microspheres, W0 is the weight of microspheres. In theoretical drug loading (weight of initial drug/weight of both drug and polymer × 100), W2 is weight of initial drug and W3 is weight of both drug and polymer (CS) added in the formulation. 2.5.2. In vitro release behavior The release behavior of berberine was evaluated by the dialysis method. Dialysis bags containing 20 mg drug-loaded microspheres were immersed in a thermostated beaker containing 10 ml PBS (0.1 M, pH 7.4) at 37 ◦ C, and stirred at 100 rpm. At predetermined time intervals, the release medium from each sample were withdrawn and centrifugated, then an equal volume of fresh buffer was added immediately. Release measurements were described previously [22]. Each value represents an average of three runs. The cumulative release (%) was expressed as the percent of total berberine released over a period of 6 days. 2.6. Cell morphology and viability To evaluate the biological properties of the microspheres in vitro, we focused on cell morphology and cell activity assays by direct contact with MG63 cells. MG63 cell is a non-transformed cell line and exhibits an osteoblastic phenotype. The cells were routinely

grown in F12 medium supplemented with 10% volume fraction of calf serum, 1% penicillin/streptomycin and 1% l-glutamine. The cells were maintained at 37 ◦ C and 100% humidity with 5% CO2 . The media was replaced every 2–3 days. Cultures of 90% confluent cells were trypsinized, washed and suspended in fresh media. MG63 cells cultured in media for 3 days were seeded on the top of pre-wetted microspheres. The specimens were then placed in the wells of plastic dishes (24 well cell culture plate, Corning, USA) [23]. One milliliter of cross-linked CS microspheres suspension was transferred to Transwell-cell culture inserts (Corning, USA) and placed in 24-well plates such that the entire surface of the insert mesh was covered by a thin layer of microspheres, and the media was discarded before the cells were seeded. Subsequently, the MG63 cells were seeded on the microspheres at a seeding density of 1 × 104 cells per well followed by incubation at 37 ◦ C and 100% humidity with 5% CO2 [24]. The adhesion and morphology of MG63 cells were observed by SEM. Fixation in 2.5% glutaraldehyde, dehydration through a graded ethanol (50, 70, 90, and 100%) and drying of Critical Point were carried out in situ in the inserts. Afterwards, each insert was taken out of the well and the entire insert mesh was carefully cut out from the insert along the edge using a scalpel. The samples were sputter-coated with gold before SEM observation [25]. Cell viability on the microspheres was evaluated using the Live/Dead viability/cytotoxicity assay (Molecular Probes, Netherlands). After 1 day and 4 days post seeding, the samples were washed twice with PBS and incubated with Live/Dead solution (2 ␮mol Calcein-AM and 4 ␮mol ethidium homodimer) at room temperature in the dark. After combination dying for 30 min and washed twice with PBS, the samples were subsequently analyzed with an inverted microscope (TE2000-U, Nikon, Japan) with a 100 W Hg lamp and photographed. The viable cells (in green) and nonviable cells (in red) were distinguished under the fluorescence microscope [26].

2.7. In vivo experiments Eight randomly selected Sprague–Dawley (SD) rats were assigned to each group. After routine treatment, the microspheres (8–10 mg) sterilized by ethylene oxide were implanted into muscular incision of approximately 20 mm long on the rat back. The musculature and the skin were then closed with suture separately. At 1 week and 4 weeks postoperatively, the rats were sacrificed and the implanted microspheres along with their surrounding tissues were retrieved [27]. The samples used for the SEM examination were first fixed with 2% glutaraldehyde in 0.1 M of sodium cacodylate. Subsequently, the samples were dehydrated in graded ethanol solutions, critical-point dried with carbon dioxide. Samples for histological analysis were prepared by immersing

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

in 4% phosphate-buffered paraformaldehyde, followed by washing overnight in running water, dehydrated through alcohols, cleared in xylene, and embedded in paraffin wax. The cutting sections (5 ␮m in thickness) were stained with hematoxylin and eosin (H&E) and observed by microscope [28]. 2.8. Statistical analysis Data were collected in a Microsoft Excel 2007 database and the results were presented as means and standard deviations using the Origins 8.0 software. A Student’s t-test was performed to determine the statistical significance between experimental groups. A value of p < 0.05 was considered to be statistically significant. 3. Results and discussion 3.1. Particle size distribution and morphology The particle-size distribution analysis of GCM indicates that the microspheres is in the range of 9.9–31.9 ␮m (D10–D90), the median diameter (D50) is 19.5 ␮m (Fig. 2a). The VCM were found to be larger in size, in the range of 14.3–48.5 ␮m (D10–D90) with median diameter 30.3 ␮m (Fig. 2b). Preparation of microspheres by cross-link of CS in a W/O emulsion system is to utilize the reactive functional amine groups of CS to cross-link with the available reactive groups of cross linker. Firstly, the CS aqueous solution is emulsified in the oil phase, in the presence of a suitable surfactant stabilizing the aqueous droplets, followed by the use of a cross linker to harden the droplets. This method is helpful for controlling the size of microspheres via controlling the size of aqueous droplets. The mean size of the obtained microspheres is dependent on the type, concentration, cross-linking time and extent of cross-linkers [12]. When fixing the speed of stirring, the increase of cross-linker concentration and cross-linking time will increase the size of microspheres. Furthermore, increase of the polymer concentration and the cross-linking extent will also enhance the matrix density of the microspheres. Under the same condition, the GCM in Fig. 2c look smaller and more compact due to its stronger crosslinking reaction and shorter reaction time in comparison with VCM. The VCM show a well-defined spherical shape and possesses a uniform surface with micro-wrinkles (Fig. 2d), which are in favor of the drug release. The drug loading has no significant influence on the size of microspheres. The cross section images in Fig. 2(e) and (f) demonstrate that both GCM and VCM are hollow microspheres. Using microspheres either as carriers of drugs, growth factors and cells, or as basis of an injectable system or scaffold matrix, the size and morphology of microspheres have a great impact on the functionality, such as the controlling release profile and matrix properties [29]. 3.2. XRD analysis CS as a semi-crystalline polymer shows a strong diffraction peak at 2 = 20◦ and a weak diffraction peak at 10.7◦ (Fig. 3a) associated with the (0 0 1) and (1 0 0) planes [30]. The crystallinity of the cross-linked CS microspheres (GCM and VCM) is lower than the native CS powder. After cross-linking with glutaraldehyde, the GCM show only one broader or weak peak around 23◦ (Fig. 3b). The peak around 11◦ , which is assigned to CS chains aligned through intermolecular interaction, is absent. However, the VCM show still the main peak around 2=20◦ and a small peak situated at about 7–9◦ . Compared to the native CS, the weakened intensity and peak disappearance or shift of microspheres should result from the cross-linking reaction and the fall of CS intermolecular interaction [29,30]. In addition, the relatively stronger cross-linking reaction between CS and glutaraldehyde than between CS and vanillin

739

should be responsible for the crystalline and structure difference of GCM and VCM. Vanillin has a lower reactivity than glutaraldehyde, so it needs longer time to fulfill the cross-linking process which insures the smaller change of the CS phase structure. The peak shift to 7–9◦ may be related to crystal 1 in CS molecular chain, since it comprises functional groups like NH2 and OH and has undergone significant change after cross-linking [31]. 3.3. IR analysis The IR spectrum of CS in Fig. 4a suggests that the stretching vibration absorption center due to OH is observed at 3440 cm−1 , while the peaks at 2930 cm−1 and 2865 cm−1 are ascribed to C H stretching vibration of CH2 . The bands at 910 and 1162 cm−1 correspond to saccharine structure. The absorption peaks at 1629, 1595 and 1327 cm−1 are characteristic of CS for the amide I, II and III, respectively [32]. The broad band around 1082 cm−1 indicates the C O stretching vibration of CH2 OH in CS. The IR spectrum of GCM in Fig. 4b is similar to that of native CS except the shift of some peaks, e.g., the C O band at 1082 cm−1 shifts to 1072 cm−1 , and the amide II related peak at 1595 cm−1 shifts to 1569 cm−1 , also the intensity of amide III peak at 1327 cm−1 decreases, which provide proofs of cross-linking reaction through a Schiff base condensation between the amino groups of CS and the aldehyde groups of glutaraldehyde [16], as shown in Fig. 5a. The new peaks at 1513, 819 and 1287 cm−1 related to the benzene ring and OH of vanillin are also the outcome of cross-linking reaction, such as bonding between carbonyl groups of vanillin and amino groups of CS. The peak at 1643 cm−1 corresponding to characteristic stretching vibration of C N, which can be attributed to the Schiff base reaction between the aldehyde group of vanillin and amino group of chitosan (Fig. 5b(1)) [33–35]. The peak of OH of chitosan shifts from 3440 to 3423 cm−1 and its intensity has been reduced significantly after cross-linking, which may be due to the hydrogen bond interaction between chitosan and vanillin (Fig. 5b(1)). The peak at 1118 cm−1 is due to O C O vibration absorption, which is the result of acetalization reaction (Fig. 5b(2)) [36]. The benzene rings of vanillin have conjugative effect. In the one hand, aromatic base conjugate function results in passivation of the nucleophilic addition reaction of aldehyde group which could lower the activation energy of the Schiff base reaction between the aldehyde group of vanillin and amino group of chitosan. On the other hand, aromatic base conjugate function can stabilize the hemiacetal which is the intermediates of acetalization. 3.4. Degree of swelling The degree of swelling is a significant characteristic of crosslinked microspheres that controls the loading and release profile of the loaded drug. It can be seen from Fig. 6 that the cross-linked CS microspheres exhibit an ability to absorb large amounts of water, and both the pH and the cross linkers have a strong influence on the swelling degree. With the increase of absorbed water, the mobility of CS macromolecules increases and, thus, the free volume available for diffusion increases [37]. The two microspheres in PBS solution with pH 5.7 show a saturated swelling degree of about 260% for VCM and 250% for GCM within a period of 144 h, both values are higher than those in PBS solution of pH 7.4, i.e. the swelling ratio reaches a saturation value of 180% for VCM and 200% for GCM. The reason should be strongly related to the action of H+ in the acidic PBS which reacts more easily with the NH2 groups of uncrosslinked chitosan in the microspheres, forming a gel state and thus increasing the swelling degree. The swelling behavior of CS or CS microspheres is affected by various physical and

740

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

Fig. 2. Laser particle-size distributing graphs (a, b), SEM images (c, d) and sections of drug-loaded CS microspheres embedded in PMMA matrix (e, f).

chemical factors, such as bonding, structure, and the relative magnitudes of water diffusion and polymer-relaxation times [38]. Since the crosslinking action of vanillin is lower than glutaraldehyde due to conjugative effect of the benzene rings of vanillin, the crosslinked structure of VCM should be weaker than that of GCM, this may cause the swelling degree of VCM (260%) higher than GCM (250%). However, when the pH of PBS is 7.4, the swelling depends chiefly on the CS matrix and penetration of water. The saturated swelling degree of VCM (180%) is lower than GCM (200%) should be caused by the higher hydrophobicity of VCM. When CS matrix is crosslinked by aromatic vanillin with benzene rings, the hydrophobic property of CS can be remarkably improved in comparison to crosslinking by glutaraldehyde with a chain hydrocarbon structure. Consequently, the hydrophobic groups in CS molecular chains lead to decrease of swelling degree.

3.5. Drug loading and release 3.5.1. Loading ratio and entrapment efficiency The drug loading ratio and entrapment efficiency of DGCM and DVCM prepared with different ratios of berberine to CS are listed in Table 1. During preparation of microspheres, when the solvent in berberine/CS solution is dried to form solid microspheres, the hydrophilic berberine can be uniformly distributed in the CS matrix, it is different from the hydrophobic drug which will migrate with the solvent to the surface of microspheres. The practical drug loading ratios of the two microspheres are from 9.35% to 31.01% for DGCM and 9.16% to 29.70% for DVCM, which shows an increasing difference from the designed ratios of 10% to 40%. The drug entrapment efficiency is found to be high, from 77.52% to 93.49% for DGCM and from 74.25% to 91.61% for DVCM, although it is inversely

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

Fig. 3. XRD patterns of native CS powder (a), GCM (b) and VCM (c).

Fig. 4. IR spectra of native CS powder (a), GCM (b) and VCM (c).

741

742

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

Fig. 5. Schematic diagram of the reaction between CS and glutaraldehyde (a) or vanillin (b).

Fig. 6. Swelling curves of GCM and VCM in 0.1 M PBS at pH 5.7 and pH 7.4 (mean ± SD, n = 3).

Table 1 Drug loading ratio and entrapment efficiency of DGCM and VGCM (mean ± SD, n = 3). Ratio of drug to CS

Drug loading ratio (%) DGCM

1:9 2:8 3:7 4:6

9.35 17.76 24.42 31.01

Entrapment efficiency (%) DVCM

± ± ± ±

1.88 1.25 1.19 2.45

9.16 17.59 23.94 29.70

DGCM ± ± ± ±

1.37 1.66 1.02 1.97

93.49 88.79 81.41 77.52

DVCM ± ± ± ±

2.23 1.91 2.81 2.24

91.61 87.95 79.81 74.25

± ± ± ±

1.66 1.92 2.61 2.05

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

743

Table 2 Fitting results of in vitro release curve of DGCM and DVCM. Sample

n

r

Mt/M∞

DGCM DVCM

0.6184 0.6237

0.7710 0.8585

0.1070 × (t − 0.5)0.6184 + 0.1085 0.0987 × (t − 0.5)0.6237 + 0.1189

pathway [39]. The t50% value, the time required for the release of 50% encapsulated drug, is about 9.2 h for DGCM and 11.9 h for DVCM respectively. The diffusion of berberine occurs simultaneously with the swelling of CS polymeric matrix. Drug release kinetics was analyzed by plotting the cumulative release data vs. time by fitting to an exponential equation (4): Mt = k(t − t0)n + b M∞ Fig. 7. In vitro drug release curve of GCM and VCM at pH 7.4 (mean ± SD, n = 3).

related to the ratio of drug to CS. The high entrapment efficiency should be attributed to the characteristic property of CS matrix and the preparation technology which may be also suitable for loading growth factors and cells. 3.5.2. In vitro drug release In clinical therapy, it is necessary to maintain drug at a therapeutic serum concentration for specific period of time. So the microspheres with a minimal initial burst release and a relatively well-controlled sustained release are likely to be ideal drug carriers. The release behavior of berberine from microspheres in PBS at pH 7.4 is shown in Fig. 7. The initial burst release is because the drug is only physically entrapped in the microspheres, and does not chemically react with the CS polymer. It is associated with the diffusion of berberine dispersing in the surface region of microspheres. It can be seen that a sustained berberine release is realized, that is, the amount of cumulative release increases in a parabolic curve within 72 hours and then nearly reaches saturation state. About 11% of berberine is released from both microspheres at 0.5 h. The driving force of berberine dissolving out is from the concentration difference of berberine in CS matrix. The following release comes from the interior of the swollen CS matrix through a more convoluted

(5)

where Mt is the amount of drug released at time t; M∞, the total drug released over a long time period (almost 100%); t0, the time of burst release (0.5 h); b, the drug released at 0.5 h, k is a constant and n is the diffusional exponent reflecting the mechanism of drug release [40]. The results are presented in Table 2. The correlation coefficient ‘r’ represents the linear release of drug from microspheres. The value of n between 0.5 and 1.0 indicates anomalous transport kinetics, and the n value of approximately 0.5 indicates the pure diffusion controlled mechanism (Fickian diffusion). The smaller n value below 0.5 may be due to drug diffusion partially through a swollen matrix and water filled pores in the formulation [41]. The SEM images in Fig. 2(e) and (f) show that a cavity is present in the center of both microspheres. The release of berberine from the hollow microspheres should involve three different mechanisms [42]: (1) release from the surface of microspheres, (2) diffusion from the swollen CS matrix, (3) diffusion from the solution in the cavity. The release mechanism (3) is not present in normal CS microspheres [43]. So the values between 0.5 and 1.0 in Table 2 should be attributed to the anomalous type of diffusive transport, involving mechanism (1) and mechanism (2) in Fig. 8. It is clear that DVCM microspheres show better diffusive transport than DGCM. Therefore, the vanillin cross-linked microspheres can be a good drug microcarrier for localized and sustained delivery of appropriate drug at a desirable rate and concentration.

Fig. 8. Drug release mechanism of berberine loaded cross-linked CS hollow microspheres.

744

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

Fig. 9. Fluorescence microscopy images (a, b, e, f, live cells – green, dead cells – red, (a, e) 100× and (b, f) 40×) and SEM images (c, d, g, h) of MG63 cells attached to the GCM and VCM cultured on day 1 (a–d) and day 4 (e–h). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

745

Fig. 10. SEM images (a, b, e, f) and photomicrographs (c, d, g, h) of the tissues implanted with GCM and VCM at 1 week (a–d) and 4 weeks (e–h) postoperatively. The inflammation mediating cells stained purple and normal tissue stained pink by H.E. staining. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

746

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747

3.6. Cytotoxicity 3.6.1. Cell morphology A drug carrier for biomedical application should be noncytotoxic and have good biocompatibility to prevent any adverse effects on residing and recruited cells and neighboring tissues. In addition, the carrier can be formulated and own particular structure to facilitate cellular infiltration and growth [44]. The morphologies of MG63 cells cultured with GCM and VCM are shown in Fig. 9. On day 1 and day 4, the density of growing MG63 cells on VCM (b, f) is larger than on GCM (a, e), and the cells exhibit normal and healthy osteoblastic spindle-like morphology. The cells adhere to the VCM surface (d, h) much closer than to the GCM surface (c, g), demonstrating a better cytocompatibility for VCM. The VCM is almost completely covered by MG63 cells with assembly of filopodia and focal adhesion-like junctions. This means the VCM may also be favorable to the delivery of cellular aggregates. In fact, the chemical structure of CS is similar to that of glycosaminoglycan which is a key molecule in the extracellular matrix to modulate cell morphology and function [45]. Use of VCM can overcome the drawback of glutaraldehyde that is considered to be toxic [46]. Vanillin has been safely used in several types of food without any known toxicity [46]. The cells exhibiting the desired stretched morphology on VCM suggests the presence of a conducive environment for cell anchorage and growth [47]. This combined function of delivering drug along with potential delivery of cells can be considered ideal to achieve native-tissue equivalents by ex vivo culturing, as well as to help complete tissue formation after implantation [48].

3.6.2. Biocompatibility of the implanted CS microspheres Along with the in vitro cellular behavior, the in vivo tissue compatibility of the microspheres was assessed by implanting the microspheres in rat subcutaneous tissue. At 1 week and 4 weeks postoperatively, the two CS microspheres remain in good sphericity as shown in Fig. 10(a) (b) (e) (f). However, the tissue of GCM group (c, g) displays much stronger neutrophilic inflammatory reaction than VCM group (d, h), some macrophages can be observed on the surface of GCM. For chemically modified microspheres, the main problem is the potential toxic of the cross-linker or its remaining residues. Benign tissue response with the CS microspheres could be obtained by using safer cross-linker or a very low concentration of glutaraldehyde [27]. At 1 week, no inflammatory cells can be seen in the surrounding tissue of VCM (d), and only mild inflammation is present at 4 weeks comparing to the severe inflammatory response of GCM. The lower inflammatory reaction of VCM should be due to the lower toxicity of its remaining residues. For the future application of the microspheres, loading of drug, like berberine, is necessary to prevent and treat the infection and inflammation. The results from this study suggest that the VCM possess favorable properties for use in biomedical applications.

4. Conclusion Vanillin cross-linked CS microspheres (VCM) exhibited better cyto- and bio-compatibility than glutaraldehyde cross-linked CS microspheres (GCM). The use of vanillin as cross-linker successfully overcomes the toxic drawback of glutaraldehyde. The VCM microspheres show a well-defined spherical shape, and the cross-linking of CS with vanillin is found to be caused by Schiff base formation and acetalization. The VCM microspheres have a high drug loading and entrapment efficiency and moderate swelling ratio, presenting a sustained drug release behavior- via release from the surface of microspheres and diffusion from the swollen CS matrix. It can be a good candidate for localized and sustained

drug delivery, allowing for further biomedical investigation as therapeutic bioactive microcarrier.

Acknowledgements The authors thanks for the financial support of China NSFC project (No. 31370971), Sichuan project (No. 2012FZ0125) and Chengdu project (No. 12DXYB145JH-005).

References [1] C. Bibbo, Reverse sural flap with bifocal ilizarov technique for tibial osteomyelitis with bone and soft tissue defects, J. Foot Ankle Surg. 53 (2014) 344–349. [2] C. Bibbo, J. Spellman, Successful salvage of complicated calcaneal blastomycosis in disseminated disease with staged surgical reconstruction and local–systemic antifungal therapy, J. Foot Ankle Surg. 53 (2014) 472–479. [3] V. Uskokovic, T.A. Desai, Simultaneous bactericidal and osteogenic effect of nanoparticulate calcium phosphate powders loaded with clindamycin on osteoblasts infected with Staphylococcus aureus, Mater. Sci. Eng. C: Mater. Biol. Appl. 37 (2014) 210–222. [4] M.N. Rahaman, B.S. Bal, W. Huang, Review: emerging developments in the use of bioactive glasses for treating infected prosthetic joints, Mater. Sci. Eng. C: Mater. Biol. Appl. 41 (2014) 224–231. [5] M. Sokolsky-Papkov, K. Agashi, A. Olaye, K. Shakesheff, A.J. Domb, Polymer carriers for drug delivery in tissue engineering, Adv. Drug Deliv. Rev. 59 (2007) 187–206. [6] S. Ansboro, J.S. Hayes, V. Barron, S. Browne, L. Howard, U. Greiser, P. Lalor, F. Shannon, F.P. Barry, A. Pandit, J.M. Murphy, A chondromimetic microsphere for in situ spatially controlled chondrogenic differentiation of human mesenchymal stem cells, J. Control. Release 179 (2014) 42–51. [7] R.A. Perez, A. El-Fiqi, J.H. Park, T.H. Kim, J.H. Kim, H.W. Kim, Therapeutic bioactive microcarriers: co-delivery of growth factors and stem cells for bone tissue engineering, Acta Biomater. 10 (2014) 520–530. [8] S. Kim, Y. Kang, C.A. Krueger, M. Sen, J.B. Holcomb, D. Chen, J.C. Wenke, Y. Yang, Sequential delivery of BMP-2 and IGF-1 using a chitosan gel with gelatin microspheres enhances early osteoblastic differentiation, Acta Biomater. 8 (2012) 1768–1777. [9] A. Bernkop-Schnurch, S. Dunnhaupt, Chitosan-based drug delivery systems, Eur. J. Pharm. Biopharm. 81 (2012) 463–469. [10] S.K. Shukla, A.K. Mishra, O.A. Arotiba, B.B. Mamba, Chitosan-based nanomaterials: a state-of-the-art review, Int. J. Biol. Macromol. 59 (2013) 46–58. [11] R.A.A. Muzzarelli, Chitins and chitosans for the repair of wounded skin, nerve, cartilage and bone, Carbohydr. Polym. 76 (2009) 167–182. [12] S. Jose, J.F. Fangueiro, J. Smitha, T.A. Cinu, A.J. Chacko, K. Premaletha, E.B. Souto, Cross-linked chitosan microspheres for oral delivery of insulin: Taguchi design and in vivo testing, Colloids Surf. B: Biointerfaces 92 (2012) 175–179. [13] K.V. Harish Prashanth, R.N. Tharanathan, Crosslinked chitosan – preparation and characterization, Carbohydr. Res. 341 (2006) 169–173. [14] B.P. da Costa Neto, A.L.M.L. da Mata, M.V. Lopes, B. Rossi-Bergmann, M.I. Ré, Preparation and evaluation of chitosan–hydrophobic silica composite microspheres: role of hydrophobic silica in modifying their properties, Powder Technol. 255 (2014) 109–119. [15] J. Ma, Y. Sahai, Chitosan biopolymer for fuel cell applications, Carbohydr. Polym. 92 (2013) 955–975. [16] K.C. Gupta, F.H. Jabrail, Glutaraldehyde and glyoxal cross-linked chitosan microspheres for controlled delivery of centchroman, Carbohydr. Res. 341 (2006) 744–756. [17] J. Berger, M. Reist, J.M. Mayer, O. Felt, N.A. Peppas, R. Gurny, Structure and interactions in covalently and ionically crosslinked chitosan hydrogels for biomedical applications, Eur. J. Pharm. Biopharm. 57 (2004) 19–34. [18] K.C. Gupta, F.H. Jabrail, Preparation and characterization of sodium hexameta phosphate cross-linked chitosan microspheres for controlled and sustained delivery of centchroman, Int. J. Biol. Macromol. 38 (2006) 272–283. [19] M. Avila, M. Zougagh, A. Escarpa, A. Rios, Supported liquid membrane-modified piezoelectric flow sensor with molecularly imprinted polymer for the determination of vanillin in food samples, Talanta 72 (2007) 1362–1369. [20] H. Peng, H. Xiong, J. Li, M. Xie, Y. Liu, C. Bai, L. Chen, Vanillin cross-linked chitosan microspheres for controlled release of resveratrol, Food Chem. 121 (2010) 23–28. [21] P.V. Kulkarni, J. Keshavayya, V.H. Kulkarni, Effect of method of preparation and process variables on controlled release of insoluble drug from chitosan microspheres, Polym. Adv. Technol. 18 (2007) 814–821. [22] Q. Zou, Y. Li, L. Zhang, Y. Zuo, J. Li, J. Li, Antibiotic delivery system using nano-hydroxyapatite/chitosan bone cement consisting of berberine, J. Biomed. Mater. Res. Part A 89A (2009) 1108–1117. [23] Q. Zou, Y. Li, L. Zhang, Y. Zuo, J. Li, X. Li, Characterization and cytocompatibility of nano-hydroxyapatite/chitosan bone cement with the addition of calcium salts, J. Biomed. Mater. Res. Part B: Appl. Biomater. 90B (2009) 156–164.

Q. Zou et al. / International Journal of Biological Macromolecules 79 (2015) 736–747 [24] S.D. Nath, S. Son, A. Sadiasa, Y.K. Min, B.T. Lee, Preparation and characterization of PLGA microspheres by the electrospraying method for delivering simvastatin for bone regeneration, Int. J. Pharm. 443 (2013) 87–94. [25] N.J. Lakhkar, J.-H. Park, N.J. Mordan, V. Salih, I.B. Wall, H.-W. Kim, S.P. King, J.V. Hanna, R.A. Martin, O. Addison, J.F.W. Mosselmans, J.C. Knowles, Titanium phosphate glass microspheres for bone tissue engineering, Acta Biomater. 8 (2012) 4181–4190. [26] Z. Shi, K.G. Neoh, E.T. Kang, W. Wang, Antibacterial and mechanical properties of bone cement impregnated with chitosan nanoparticles, Biomaterials 27 (2006) 2440–2449. [27] F.-L. Mi, Y.-C. Tan, H.-F. Liang, H.-W. Sung, In vivo biocompatibility and degradability of a novel injectable-chitosan-based implant, Biomaterials 23 (2002) 181–191. [28] R. van Dijkhuizen-Radersma, S.C. Hesseling, P.E. Kaim, K. de Groot, J.M. Bezemer, Biocompatibility and degradation of poly(ether–ester) microspheres: in vitro and in vivo evaluation, Biomaterials 23 (2002) 4719–4729. [29] S. Bagheri-Khoulenjani, H. Mirzadeh, M. Etrati-Khosroshahi, M.A. Shokrgozar, Particle size modeling and morphology study of chitosan/gelatin/nanohydroxyapatite nanocomposite microspheres for bone tissue engineering, J. Biomed. Mater. Res. Part A 101 (2013) 1758–1767. [30] H.S. Samanta, S.K. Ray, Controlled release of tinidazole and theophylline from chitosan based composite hydrogels, Carbohydr. Polym. 106 (2014) 109–120. [31] D. Anjali Devi, B. Smitha, S. Sridhar, T.M. Aminabhavi, Pervaporation separation of isopropanol/water mixtures through crosslinked chitosan membranes, J. Membr. Sci. 262 (2005) 91–99. [32] P.A. Soares, A.I. Bourbon, A.A. Vicente, C.A. Andrade, W. Barros Jr., M.T. Correia, A. Pessoa Jr., M.G. Carneiro-da-Cunha, Development and characterization of hydrogels based on natural polysaccharides: policaju and chitosan, Mater. Sci. Eng. C: Mater. Biol. Appl. 42 (2014) 219–226. [33] E.S. Costa-Junior, E.F. Barbosa-Stancioli, A.A.P. Mansur, W.L. Vasconcelos, H.S. Mansur, Preparation and characterization of chitosan/poly(vinyl alcohol) chemically crosslinked blends for biomedical applications, Carbohydr. Polym. 76 (2009) 472–481. [34] D.A. Devi, B. Smitha, S. Sridhar, T.M. Aminabhavi, Pervaporation separation of isopropanol/water mixtures through crosslinked chitosan membranes, J. Membr. Sci. 262 (2005) 91–99. [35] A.P. Rokhade, S.A. Patil, T.M. Aminabhavi, Synthesis and characterization of semi-interpenetrating polymer network microspheres of acrylamide grafted

[36] [37]

[38] [39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47] [48]

747

dextran and chitosan for controlled release of acyclovir, Carbohydr. Polym. 67 (2007) 605–613. H. Michio (Ed.), Studies in Organic Chemistry, Elsevier, 1992, pp. v–vi. L. Lacerda, A.L. Parize, V. Favere, M.C. Laranjeira, H.K. Stulzer, Development and evaluation of pH-sensitive sodium alginate/chitosan microparticles containing the antituberculosis drug rifampicin, Mater. Sci. Eng. C: Mater. Biol. Appl. 39 (2014) 161–167. K.V. Harish Prashanth, R.N. Tharanathan, Crosslinked chitosan—preparation and characterization, Carbohydr. Res. 341 (2006) 169–173. K.T. Trifkovic, N.Z. Milasinovic, V.B. Djordjevic, M.T. Krusic, Z.D. Knezevic-Jugovic, V.A. Nedovic, B.M. Bugarski, Chitosan microbeads for encapsulation of thyme (Thymus serpyllum L.) polyphenols, Carbohydr. Polym. 111 (2014) 901–907. H. Kim, R. Fassihi, Application of binary polymer system in drug release rate modulation. 2: Influence of formulation variables and hydrodynamic conditions on release kinetics, J. Pharm. Sci. 86 (1997) 323–328. I. Armentano, M. Dottori, E. Fortunati, S. Mattioli, J.M. Kenny, Biodegradable polymer matrix nanocomposites for tissue engineering: a review, Polym. Degrad. Stab. 95 (2010) 2126–2146. J. Li, Y. Li, J. Xiao, Q. Zou, The mechanism of the release of DDP from DDP-loaded n-HA/CS composite hollow microspheres, J. Control. Release 172 (2013) e46. Y. Su, Q. Su, W. Liu, M. Lim, J.R. Venugopal, X. Mo, S. Ramakrishna, S.S. Al-Deyab, M. El-Newehy, Controlled release of bone morphogenetic protein 2 and dexamethasone loaded in core–shell PLLACL-collagen fibers for use in bone tissue engineering, Acta Biomater. 8 (2012) 763–771. S.-H. Lee, H. Shin, Matrices and scaffolds for delivery of bioactive molecules in bone and cartilage tissue engineering, Adv. Drug Deliv. Rev. 59 (2007) 339–359. A. Lahiji, A. Sohrabi, D.S. Hungerford, C.G. Frondoza, Chitosan supports the expression of extracellular matrix proteins in human osteoblasts and chondrocytes, J. Biomed. Mater. Res. 51 (2000) 586–595. D.P. Speer, M. Chvapil, C.D. Eskelson, J. Ulreich, Biological effects of residual glutaraldehyde in glutaraldehyde-tanned collagen biomaterials, J. Biomed. Mater. Res. 14 (1980) 753–764. P.X. Ma, Biomimetic materials for tissue engineering, Adv. Drug Deliv. Rev. 60 (2008) 184–198. T. Miao, K.S. Rao, J.L. Spees, R.A. Oldinski, Osteogenic differentiation of human mesenchymal stem cells through alginate-graft-poly(ethylene glycol) microsphere-mediated intracellular growth factor delivery, J. Control. Release 192C (2014) 57–66.