In vitro degradation and release profiles for electrospun polymeric fibers containing paracetanol

Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

In vitro degradation and release profiles for electrospun polymeric fibers containing paracetanol Hongsen Peng, Shaobing Zhou ∗ , Tao Guo, Yanshan Li, Xiaohong Li, Jianxin Wang, Jie Weng Key Laboratory of Advanced Technologies of Materials, Ministry of Education, School of Materials Science and Engineering, Southwest Jiaotong University, Chengdu 610031, PR China

a r t i c l e

i n f o

Article history: Received 28 March 2008 Received in revised form 19 May 2008 Accepted 20 June 2008 Available online 9 July 2008 Keywords: Biodegradation PLA Electrospun Drug delivery Scaffolds

a b s t r a c t In the paper, the poly(d,l-lactide) (PDLLA) and poly(ethylene glycol)-co-poly(d,l-lactide) (PELA) fibers with and without paracetanol drug loading were prepared with an electrospinning method. The morphology of the fibers was observed by scanning electronic microscope (SEM). Their glass transition temperatures (Tg ) were measured with differential scanning calorimetry (DSC). The water contact angle (CA) measurement was also performed to characterize surface properties of fibers. At 37 ◦ C in a PBS buffer solution (pH 7.4), in vitro matrix degradation profiles of these fibers were characterized by measuring their weight loss, the molecular weight decrease, and their morphology change. The result showed that the effects of fiber diameter and porosities on the degradation of the electrospun scaffolds might exceed the effects of the molecular weight and the PEG contents, which was different from the polymeric microspheres degradation. In vitro paracetanol release profiles were also investigated in the same condition. The result showed that the drug burst release behaviour was mainly related with the drug–polymer compatibility and the followed sustained release phase depended on polymer degradation. © 2008 Elsevier B.V. All rights reserved.

1. Introduction In recent years, electrospun fibers with diameters ranging from the nano- to microscale, using natural and synthetic polymer as matrix, have received increased interest for use as scaffolds for tissue engineering, carriers for drug delivery system, and wound dressing materials [1–3]. Electrospinning uses electrostatic forces as the driving force to spin fibers. An electrospinning apparatus is very simple, and consists of three parts namely a high voltage supply, a capillary/syringe with a needle, and a grounded collector. In the electrospinning process by solution, a polymer solution held by its surface tension at the end of a capillary tube is subjected to an electric field. The technique has been proved to be convenient and effective for producing continuous nanofibres from most polymers. Biodegradable polymers, such as polylactide (PLA), polyglycolide (PGA), and their copolymer polylactide-co-glycolide (PLGA) find wide applications in the pharmaceutical industry as matrices for drug delivery [4–7] and in medicine as material for bone implants and bone fixation devices [8], surgical sutures, and anastomotic devices owing to their excellent biodegradation, biocompatibility and nontoxic degradation products. Used as fiber

∗ Corresponding author. Tel.: +86 28 87634023; fax: +86 28 87634649. E-mail addresses: [email protected], [email protected] (S. Zhou). 0927-7765/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfb.2008.06.021

matrix, besides above advantages, these synthetic biodegradable polymers also provide the necessary mechanical properties, such as viscoelasticity and strength, and their degradation rate can be controlled as needed [9]. Drug delivery with polymer nanofibers is based on the principle that the dissolution rate of a drug particulate increases with increased surface area of both the drug and the corresponding carrier if necessary. The main advantage of the electrospun fibrous scaffolds as drug carriers is that they offer site-specific delivery of any number of drugs from the scaffold into the body. In addition, the drug can be capsulated directly into fibers with different sizes, and the electrospun scaffolds can be cut to almost any size. Until now, the research that use of electrospun biodegradable polymeric fibers for drug delivery system for pharmaceutical application is quite limited [10–13]. It is well known that poly(ethylene glycol) (PEG), as a typical hydrophilic polymer, has been approved by the Food and Drug Administration (FDA) of USA for inner-body use [14]. Low molecular weight PEG is readily excreted through the kidney [15]. In our previous study, we found that poly(ethylene glycol)-co-poly(d,llactide) (PELA) microspheres were more potential as carriers for protein, peptide and gene delivery systems than commonly used PLA and PLGA due to the introduce of hydrophilic PEG segments into PLA backbones [16–18]. In addition, PELA copolymer also displays a good biocompatibility in vivo, improved degradation rate, decreased acidity of the degradation products, and increased

H. Peng et al. / Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

hydrophilicity compared to pure poly(d,l-lactic) and thus it exhibits good potential in formulating drug delivery systems [19–23]. Therefore, in the paper we focused on investigation of PELA copolymers with different amount of PEG in form of drugloaded fibers. Paracetanol (acetaminophen, N-(4-hydroxy-phenyl) acetamide), which is a widely used analgesic and antipyretic drug, was chosen as the model drug. Electrospun PELA fibers with a predetermined drug loading amount were prepared. The effects of both PEG amount in PELA copolymer and fiber diameter on characteristics of electrospun fibers were investigated. In vitro matrix degradation profiles of these fibers were characterized by measuring their weight loss, the molecular weight decrease, and their morphology change. In vitro paracetanol release profiles were also investigated in the same instrument. 2. Experimental 2.1. Materials Poly(d,l-lactide) (PDLLA, or PELA-0) and PELA with PEG weight ratios of 10% (PELA-10), 15% (PELA-15) and 20% (PELA-20) were synthesized by ring-opening polymerization in our laboratory. Their molecular weights and their distributions were determined by gel permeation chromatography (GPC, waters 2695 and 2414, Milford, MA) using polystyrene as a standard. The column used was a Styragel HT 4 (7.8–300 mm). The mobile phase consisted of tetrahydrofuran (THF) using a regularity elution at a flow rate of 1.0 ml/min. Paracetanol was obtained from Kangquan Pharmaceuticals Inc., China. All other chemicals and solvents were of reagent grade or better. 2.2. Preparation of electrospun fiber mats Preweighed PDLLA and PELA were dissolved in acetone solvent, respectively, as shown in Table 1. 5 wt.% of paracetanol with respect to the polymer used was added into the polymer solution. The solution was then immediately electrospun. The electrospinning set-up was described in our previous report [13]. The polymer solution mixed with paracetanol was transferred to a 5 ml syringe with a right angle-shaped needle of 0.4 mm in inner diameter attached to it. A pressure was applied to the solution in syringe to maintain a steady flow of the solution from the needle outlet in the range of 1.0–2.0 ml/h. The electric field strength was 1.33–1.67 kV/cm. The distance between the needle tip and the grounded target was 15 cm. In order to remove the residual acetone, the fiber mats collected were dried at room temperature under vacuum for about 48 h and stored at 4 ◦ C. 2.3. Fiber characterization The surface morphology and the fiber size distribution of the electrospun fibers were investigated by a scanning electron microscope (SEM, Quanta 200) whose accelerating voltage was 20 kV. Micrographs from the SEM analysis were digitized and analyzed by Image Tool 2.0 to determine the average fiber diameter of the mats

207

produced. The porosity of the fiber mats was calculated according to the following equation: Porosity =



1−

 0



× 100,

where  is the density of electrospun scaffold and 0 is the density of bulk polymer [24]. The water contact angle (CA) was measured using a sessile drop method at room temperature with the contact angle equipment (DSA 100, KRUSS, Germany). The contact angle was determined at 5th and 10th second after the distilled water droplet contacted on the surface of samples. The CA was an average value of five tests on different locations of the surface. Differential scanning calorimetry (DSC, Netzsch STA 449C, Bavaria, Germany) was used as a rapid test method to investigate the glass transition temperatures (Tg ) of the polymer fibers. The samples were analyzed in perforated and covered aluminum pans under a nitrogen purge. Approximately 1 mg of fiber sample was heated from 25 to 150 ◦ C with a heating rate of 10 ◦ C/min. 2.4. In vitro degradation study The dried electrospun nanofiber non-woven mat was cut into 2 cm × 2 cm pieces. Each cut specimen was measured for initial weight (∼30 mg), and was then placed in a test tube containing 20 ml of phosphate-buffered saline (PBS, 0.1 M, pH 7.4) for in vitro degradation study. The tubes were placed in a thermostated shaking air bath (Hua Li Da Laboratory Equipment Company, China) that was maintained at 37 ◦ C and 120 cycles/min and were observed for 6 weeks. At different time intervals, triplicate specimens for each sample condition were removed from the tubes. The degraded samples were first rinsed several times with distilled water and vacuum dried at room temperature for 1 week. The mass loss was determined gravimetrically by comparing the dry weight remaining at a specific time with the initial weight. The recovered and dried samples were dissolved in THF and filtered to eliminate insoluble residues. The molecular weight of recovered matrix polymer was determined using GPC as described above. The morphology of fibers was observed with SEM as described above. 2.5. In vitro drug release As described above, the electrospun non-woven mats with drug entrapment were first sectioned into 2 cm × 2 cm, and the drug content was determined as a function of scaffold weight. Each square sample was incubated into 20.0 ml of PBS and placed in the same shaking air bath as the in vitro degradation. Samples of 1.0 ml released solution were taken from the dissolution medium at 3, 6, 12, 24, 48, 72, 96, 120, 144, 288, 336, 432, 456, and 504 h after incubation, while an equal amount of fresh PBS was added back to the incubation solution. The amount of paracetanol was detected with a UV–vis spectrophotometer (UV-2550, Shimadzu, Japan), and a maximal absorption peak of 243 nm was observed for freshly prepared paracetanol in PBS and released within the

Table 1 The electrospinning parameters and the fiber characteristics Sample

Mw (kDa)

PEG content (wt.%)

Diameter (nm) No drug

PELA-0 PELA-10 PELA-15 PELA-20

100.86 50.98 41.45 33.01

0 10 15 20

610 720 770 1010

± ± ± ±

310 200 230 500

Porosity (%) Drug (5%) 770 790 880 1670

± ± ± ±

280 80 210 570

No drug 89 86 77 68

± ± ± ±

3.2 2.5 2.1 3.6

Drug (5%) 83 81 76 65

± ± ± ±

3.4 2.6 3.1 3.9

208

H. Peng et al. / Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

designed period. For standard samples with a concentration from 0 to 40 ␮g/ml, a linear correlation ( 2 = 0.9999) was determined between the absorption strength and paracetanol concentration. The percentage of the released drug in triplicate samples was then calculated based on the initial weight of the drug incorporated in the electrospun scaffold.

3. Results and discussion 3.1. Morphology of electrospun fiber mats Firstly, the electrospun scaffolds of PDLLA and PELA without drug were obtained because they are easy to be electrospun into

Fig. 1. SEM photographs of PELA electrospun fibers without (a–d) and with (e–h) drug entrapment. PEG content: (a, e) 0 wt.%; (b, f) 10 wt.%; (c, g) 15 wt.%; (d, h) 20 wt.%.

H. Peng et al. / Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

nanofibers. The SEM micrographs of the fibers obtained are shown in Fig. 1(a–d). The average diameters of PELA-0, PELA-10, PELA-15, PELA-20 fibers are 610, 720, 770 and 1010 nm, respectively, increasing with the content of PEG increase. It was found that the as-spun scaffolds had a porosity ranging from 65% to 89%. As shown in Table 1, the porosity of the electrospun scaffold decreased with the fiber diameter increasing. The different porosities can be obtained by adjusting the electrospinning parameters. Furthermore, as shown in Fig. 1, the PELA fibers are fused or bonded together at their contact sites, leading to a three-dimensional network. Especially, more nodes were appeared among the fibers with PEG content increase in PELA copolymer. It indicated that certain amount of PEG presented in the copolymers could increase the adherence among the fibers due to their hydrophilicity increase compared with that of PLA. There existed some hydrogen bonding interaction among PEG molecules, which resulted in the adherence. Due to paracetanol excellent solubility in acetone, the polymer/drug mixture solution was stable and homogeneous, and the electrospinning was successful. The surfaces of the fibers with drug entrapment are smooth and no drug crystal is detected, as shown in Fig. 1(e–h). The average diameters of the PELA-0, PELA-10, PELA-15, PELA-20 drug loading fibers are 770, 790, 880 and 1670 nm, respectively, also increasing with the content of PEG increase in agreement with the PELA fibers without drug. All these indicate that the drug is finely incorporated into the electrospun fibers, and the introduction of the small molecule drug (paracetanol) into polymeric matrices made the electrospun fibers more straight and uniform. The possible reason for this improvement is that addition of drugs disturbed the polymer solution, lowered the surface tension, and thus enhanced the bending instability [24]. Therefore, the paracetanol was an important factor to influence the characteristics of the electrospun fibers.

209

Fig. 2. DSC traces of electrospun fiber non-woven mats without (a: PELA-0, and c: PELA-20) and with (b: PELA-0, and d: PELA-20) drug entrapment.

3.3. DSC analysis of electrospun scaffolds The variation of Tg of the electrospun polymeric scaffolds with or without paracetanol was shown in Fig. 2, as measured through DSC. As can be seen, there was a significant decrease in Tg values between the PDLLA electrospun scaffolds without (55.6 ◦ C) and with (49.9 ◦ C) paracetanol entrapment. Similarly, the Tg of medicated PELA-20 electrospun scaffold (47.2 ◦ C) was lower than that of pure PELA-20 electrospun scaffold (51.4 ◦ C). The results from the Tg s change of other PELA fibers were almost identical. The phenomena were ascribed to the addition of the drug into the electrospun polymer fibers. The small molecule drug acted on the molecular chains and made the molecular chains move easily, leading to a lower Tg [22]. 3.4. In vitro degradation of electrospun scaffolds

3.2. The water contact angle (CA) analysis The changes of the water contact angle values at 5th and 10th second after the droplet contacted with the electrospun scaffolds are collected in Table 2. The CA (114.8 ± 2.5◦ ) of PDLLA fibers was much higher than PELA with PEG weight ratios of 10–20% fibers containing drug loading or not. Some difference for CA among PELA fibers with PEG content increasing from 10% to 20% existed at 5th second, whereas the difference was disappeared and the all CA values were 0◦ at 10th second. The results further showed that PDLLA fibers were very hydrophobic but PELA fibers were hydrophilic due to introducing PEG into PDLLA chains. As shown in Table 2, the initial CA values of all medicated electrospun fibrous mats were larger than that of the corresponding mats without drug loading. The possible reason is that addition of drugs is hydrophobic. The result also further indicated the morphological difference among the fibers came from Fig. 1. Furthermore, the CA values of the electrospun scaffolds based on pure PELA or medicated PELA were all about below 32◦ at 5th second and 8.9◦ (even 0◦ ) at 10th second, indicating that the water spread well on the hydrophobic PELA scaffold surface.

Table 2 Wetting behaviours of electrospun fibers with and without drug entrapment Sample

CA (◦ ) (no drug)

PELA-0 PELA-10 PELA-15 PELA-20

114.8 27 13 11

Contact 5 s ± ± ± ±

2.5 1.8 1.2 1.6

CA (◦ ) (with drug) Contact 10 s

Contact 5 s

105.4 ± 1.7 0 0 0

129.2 32 23 18

± ± ± ±

2.7 1.6 1.3 1.5

Contact 10 s 123.0 ± 1.8 8.9 ± 1.2 0 0

After incubation in the degradation medium, the non-woven mats changed from shrinking to puffing bigger than before. The morphologies of degraded PDLLA and PELA-20 fibers with 5.0% paracetanol inoculations after 3 and 6 weeks were shown in Fig. 3. It indicated that the electrospun scaffolds still remained their fibrous structure after 6 weeks of degradation. As shown in Fig. 3, most of the fibers were swollen compared with the original formation shown in Fig. 1, which was owing to chain relaxation of the matrix polymer after incubation into the medium. And it was found that part of fibers were broken down, which was own to the matrix breakdown. Gravimetric evaluation of the loss of electrospun scaffolds during incubation is summarized in Fig. 4. The mass loss of the electrospun scaffolds in the first 3 weeks may result from a lower molecular part of the polymers (Fig. 4) and small molecule drug on the fiber surface dissolving into the degradation medium (Fig. 4b). After 3 weeks degradation, the slight weight loss may be caused by the polymer degrading (Fig. 4) and drug diffusing out (Fig. 4b). Note that mass loss of the medicated electrospun fibers over the incubation period was larger than that of the polymer fibers without the drug. The possible reason for this phenomenon is that dissolution of small molecule drug on the fiber surface increased the contacted surface area between the polymer and water, consequently accelerated the matrix breakdown. However, it was somewhat unexpected that the mass loss of the medicated PELA-20 electrospun fibers was smaller than that of the medicated PELA-10 and PELA-15 fibers. And there were total 42%, 36%, and 35% mass loss for medicated PELA-10, PELA-15, and PELA-20 electrospun fibers, respectively. As shown in Fig. 5, the molecular weight of polymer matrix of fiber mats decreased gradually with incubation time. For PDLLA

210

H. Peng et al. / Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

Fig. 3. Morphological changes of PELA-0 (a, b) and PELA-20 (c, d) electrospun fibers containing 5.0% paracetanol during in vitro degradation at 37 ◦ C in a PBS buffer solution; (a, c) after 3 weeks; (b, d) after 6 weeks.

Fig. 4. Percent mass loss versus time for electrospun fibers without (a) and with (b) drug entrapment.

Fig. 5. Percent molecular weight loss versus time for electrospun fibers without (a) and with (b) drug entrapment.

H. Peng et al. / Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

211

electrospun scaffolds no apparent decrease in molecular weight was detected during the incubation period, while for PELA electrospun scaffolds a decrease in molecular weight was detected due to increased hydrophilicity as shown in Table 2. Also, the molecular weight loss of the medicated electrospun scaffolds over the incubation period Fig. 5b was larger than that of the polymer scaffolds without the drug (Fig. 5a), and the molecular weight loss of the medicated PELA-20 electrospun scaffolds was smaller than that of the medicated PELA-10 and PELA-15 scaffolds. The result corresponded with the mass loss shown in Fig. 4. It was well known that more PEG content in the PELA fibers should result in faster matrix breakdown. However, here it was found that the mass loss and molecular weight loss of the PELA10 electrospun scaffolds were larger than those of the PELA-15 and PELA-20 scaffolds (Figs. 4 and 5). It indicated that there were other important factors such as hydrophilicity, diameter, and porosity to influence the degradation of the electrospun polymeric scaffolds besides the chemical composition and molecular weight [7,14,22–24]. As shown in Table 2, it was found that all of the PELA electrospun scaffolds were hydrophilic at the PEG contents of 20, 15 and 10 wt.%. And seen in Table 1, the smaller fiber diameter is, the higher porosity of the PELA electrospun scaffolds could be obtained. So it can increase the contacted surface area between the polymer and water and further accelerate the matrix breakdown [13,14]. Therefore it was deduced that their effect on the degradation of the electrospun scaffolds might exceed the effects of the molecular weight and the PEG contents. The reason can be used to explain the results of Figs. 4 and 5. Therefore, we can draw a conclusion that the polymer degradation can also be controlled by adjusting the electrospun fiber diameter and its porosity.

fibers, but it may be incompatible with hydrophilic PELA fibers. So the burst release extent of medicated PDLLA electrospun scaffolds was much lower than that of other PELA scaffolds. In the followed 150 h, all these samples took on a sustained release phase. The release behaviour mainly depended on polymer matrix degradation and drug diffusion. Therefore, we can see that the drug release rate increase with PEG content decrease in electrospun PELA fibers (Fig. 6). So the drug release rate can also be controlled by adjusting the electrospun fiber diameter and polymer hydrophilicity.

3.5. In vitro drug release

This work was partially supported by National Natural Science Foundation of China (50773065), Programs for New Century Excellent Talents in university, Ministry of Education of China (NCET-07-0719) and Sichuan Prominent Young Talent Program (08ZQ026-040).

The drug release profiles of the electrospun PELA scaffolds containing 5.0% paracetanol are shown in Fig. 6. In the initial 6 h, paracetanol burst release from the medicated PDLLA, PELA10, PELA-15, and PELA-20 electrospun scaffolds is 3, 62, 51 and 45 wt.%, respectively. We think the reason is the same as drug from biodegradable polymeric microspheres [16], namely, the drug burst release behaviours are closely related to paracetanol on the fiber surface dissolved in buffer solution quickly. Due to paracetanol hydrophobicity, it can be dispersed into hydrophobic PDLLA

Fig. 6. In vitro paracetanol release from electrospun fibers with 5.0% of drug entrapment.

4. Conclusions In the paper, the characterization of biodegradable polymer electrospun fibers was investigated in detail. The result from SEM images showed that a better morphology of the fiber could be obtained by adding hydrophobic drug into PELA fibers. There was an apparent decrease in Tg from DSC analysis due to the paracetanol acted on the molecular chains of polymer. The in vitro degradation result showed that the effects of fiber diameter and porosities on the degradation of the electrospun scaffolds might exceed the effects of the molecular weight and the PEG contents. The drug burst release behaviour was mainly related with the drug–polymer compatibility and the followed sustained release phase depended on polymer degradation. So the polymer degradation and drug release behaviours could be controlled by adjusting fiber diameter and its porosity besides the polymer chemical composition and molecular weight. Acknowledgements

References [1] W.J. Li, C.T. Laurencin, E.J. Caterson, R.S. Tuan, F.K. Ko, J. Biomed. Mater. Res. 60 (2002) 613–621. [2] Y.K. Luu, K. Kim, B.S. Hsiao, B. Chu, M. Hadjiargyrou, J. Controlled Release 89 (2003) 341–353. [3] J.A. Matthews, G.E. Wnek, D.G. Simpson, G.L. Bowlin, Biomacromolecules 3 (2002) 232–238. [4] J. Zeng, L.X. Yang, Q.Z. Liang, X.F. Zhang, H.L. Guan, X.L. Xu, X.S. Chen, X.B. Jing, J. Controlled Release 105 (2005) 43–51. [5] J. Zeng, X.Y. Xu, X.S. Chen, Q.Z. Liang, X.C. Bian, L.X. Yang, X.B. Jing, J. Controlled Release 92 (2003) 227–231. [6] K. Kim, M.K. Yu, X.H. Zong, J. Chiu, D.F. Fang, Y.S. Seo, B.S. Hsiao, B. Chu, M. Hadjiargyrou, Biomaterials 24 (2003) 4977–4985. [7] K. Kim, Y.K. Luu, C. Chang, D.F. Fang, B.S. Hsiao, B. Chu, M. Hadjiargyrou, J. Controlled Release 98 (2004) 47–56. [8] G.B. Wei, Q.M. Jin, W.V. Giannobile, P.X. Ma, Biomaterials 28 (2007) 2087–2096. [9] M. Hakkarainen, Adv. Polym. Sci. 157 (2002) 113–138. [10] X.H. Zong, K. Kim, D.F. Fang, S.F. Ran, B.S. Hsiao, B. Chu, Polymer 43 (2002) 4403–4412. [11] H. Jiang, D.F. Fang, B.S. Hsiao, B. Chu, W. Chen, J. Biomater. Sci. Polym. Ed. 15 (2004) 279–296. [12] Y.Z. Zhang, X. Wang, Y. Feng, J. Li, C.T. Lim, S. Ramakrishna, Biomacromolecules 7 (2006) 1049–1057. [13] W.G. Cui, X.H. Li, X.L. Zhu, G. Yu, S.B. Zhou, J. Weng, Biomacromolecules 7 (2006) 1623–1629. [14] K.J. Zhu, X.Z. Lin, S.L. Yang, J. Appl. Polym. Sci. 39 (1990) 1–9. [15] K.S. Choi, D.J. Kim, J. Appl. Polym. Sci. 83 (2002) 435–445. [16] S.B. Zhou, X.M. Deng, X.H. Li, J. Controlled Release 75 (2001) 27–36. [17] S.B. Zhou, X.M. Deng, S.Y. He, X.H. Li, W.X. Jia, J. Pharm. Pharmacol. 54 (2002) 1287–12192. [18] S.B. Zhou, X.Y. Liao, X.H. Li, X.M. Deng, H.M. Li, J. Controlled Release 86 (2003) 195–205.

212

H. Peng et al. / Colloids and Surfaces B: Biointerfaces 66 (2008) 206–212

[19] H. Younes, P.R. Nataf, D. Cohn, Y.J. Appelbaum, G. Pizov, G. Uretzky, Biomater. Artif. Cells Artif. Organs 16 (1988) 705–719. [20] Y.H. Bae, K.M. Huh, Y. Kim, K.H. Park, J. Controlled Release 64 (2000) 3–13. [21] P. Cerral, M. Tricoli, L. Lelli, G.D. Guerra, J. Mater. Sci., Mater. Med. 5 (1994) 308–313.

[22] X.L. Xu, X.S. Chen, X.Y. Xu, T.C. Lu, X. Wang, L.X. Yang, X.B. Jing, J. Controlled Release 114 (2006) 307–316. [23] X.L. Xu, L.X. Yang, X.Y. Xu, X. Wang, X.S. Chen, Q.Z. Liang, J. Zeng, X.B. Jing, J. Controlled Release 108 (2005) 33–42. [24] J.H. Lee, A.K. Go, S.H. Oh, K.E. Lee, S.H. Yuk, Biomaterials 26 (2005) 671–678.