Crosslinking of collagen using a controlled molecular weight bio-crosslinker: β-cyclodextrin polyrotaxane multi-aldehydes

Abstract

In this study, a series of β-cyclodextrins (β-CD) polyrotaxane multi-aldehydes with different molecular weight were prepared and then were used as crosslinker for collagen. The properties of crosslinked collagen were characterized by Fourier transform infrared spectrophotometry (FTIR), collagenase degradation assay, mechanical testing, scanning electron microscope (SEM) and in vitro cell studies. Compared to crosslinked by conditional crosslinkers, such as glutaraldehyde (GA) and N-ethyl-N-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), the collagen crosslinked by the novel biocrosslinker has been improved effectively in biodegradation rate, compress modulus and show lower cytotoxicity. Besides, the modified collagen was suitable for the cell adhesion and proliferation due to the appropriate swelling property, porosity and pore size. Considering the excellent properties, β-cyclodextrins (β-CD) polyrotaxane multi-aldehydes is a novel crosslinker for collagen and the obtained collagen gel shows great potential to use as tissue engineering scaffold.

---

1. Introduction

As the most widely distributed protein in mammalian tissue,¹ collagen possesses a series of advantages, such as the excellent biocompatibility, biodegradability and other unique biologic properties. It is widely used in wound dressing, artificial skin, cosmetic, drug release, and cell cultivation.^2–4 Many researches have used collagen as tissue engineering scaffold for cartilage repair and showed a bright future.^5–8 However, the insufficient mechanical strength, high biodegradation rate, low thermal stability and facile shrinkage became the biggest problems and restricted many applications.⁹ One of the effective ways to improve the defects mentioned above is the cross-linking of collagen. Generally, physical methods for collagen cross-linking, such as ultraviolet irradiation,¹⁰ γ-ray irradiation,¹¹ dehydrothermal treatment,¹² hardly bring cytotoxicity while result in inadequate degree of crosslinking. For primary used chemical crosslinkers, glutaraldehyde (GA) shows high cototoxicity,¹³ N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) can't effectively improve the anti-enzymolysis;¹⁴ while for genipin, chromogenic reaction¹⁵ and high price are the main problems. Therefore, a novel crosslinker which can improve the property of collagen more effectively is always required.

In recent years, cyclodextrin (CD)-based polyrotaxane supramolecular materials have attracted tremendous attention and represent a very active theme of science and technology because of the tailored properties as well as the potential to serve as molecular devices,¹⁶ molecular machines,¹⁷ biomaterials related to drug or gene delivery,^1,18–20 tissue engineering materials²¹ and so on. Some researches about biodegradability and toxicity of the modified CD-based polyrotaxane supramolecular materials are reported,^22,23 while rarely report to use as bio-crosslinkers. In this study, we successfully prepared CD-based polypseudorotaxane (PPR) by threading 4,6,8,10 units of β-CD onto a poly(propylene glycol)bis(2-aminopropylether) (PPG-NH₂, M_w ≈ 2000) chain (marked as PPR-1, PPR-2, PPR-3, PPR-4 for PPR with 4,6,8,10 units β-CD) and capped the resultant PPR by β-CD mono-aldehyde as stoppers. After reduction with NaBH₄, it was directly oxidized in high yield to the corresponding multiple aldehydes (marked as PR-CHO-1, PR-CHO-2, PR-CHO-3, PR-CHO-4 for corresponding PPR with 4,6,8,10 units β-CD) by cyclized 2-iodoxybenzoic acid (IBX) in dimethylsulfoxide (DMSO). The obtained PR-CHO-4 was used as cross-linking agent to crosslink collagen.

2. Experimental

2.1 Materials

β-CD (Sinopharm Chemical Reagent Company, China) was recrystallized three times before use. Poly(propylene glycol)bis(2-aminopropyl ether) (PPG-NH₂) (M_w = 2000) was obtained from Aldrich. Cyclized 2-iodoxybenzoic acid (IBX) was purchased from Shanghai Jinmao Chemical Ltd. Type I collagen was derived from bovine tendon according to the method previously reported,²⁴ type I collagenase was purchased from sigma chemical company used in the study. Dulbecco's modified Eagle medium (DMEM) was purchased from Gibco (U.S.A.). Cell count kit-8 (CCK-8) was from Shanghai Beyotime institute.

2.2 Preparation of β-Cyclodextrins polyrotaxane multi-aldehydes with different molecular weight

The β-cyclodextrins polyrotaxane multi-aldehydes was synthesized as we reported previously²⁵ and some modifications were made to prepared the β-CD polypseudorotaxane with different numbers of β-CD unit threaded onto PPG-NH₂ chain. Briefly, 2.5 g β-CD was dissolved in 125 mL deionized water to get a saturated aqueous solution, then different amount PPG-NH₂ were put into the aqueous solution and stirred at different temperature for 1 hour. After that, they were allowed to stand overnight at room temperature. The precipitated products were collected by centrifugation, then washed twice with deionized water and dried under vacuum at 100 °C to get the β-CD polypseudorotaxane (PPR).

2.3 Preparation of crosslinked collagen hydrogel

PR-CHO-4 solution with concentration of 0, 0.0015, 0.003, 0.006 and 0.012 mmol mL⁻¹ were freshly prepared by dissolving PR-CHO-4 in 1% (v/v) aqueous acetic acid. These were mixed with an aqueous collagen solution of 6 mg mL⁻¹, respectively (Fig. 1). The PR-CHO-4/collagen weight ratios were 1 : 100, 1 : 50, 1 : 25, 2 : 25 and 0 : 100 (pure collagen). The mixture was stirred at 4 °C for 48 h, and then stirred at 37 °C for 15 min. Degassed to remove the bubbles, and then poured into sealed molds and frozen at −20 °C for 48 h. At last, the frozen samples was thawed at 25 °C to yield the hydrogel.²⁴

Fig. 1 Scheme illustration for the reaction of collagen with PR-CHO-4.

2.4 Porosity test and characterization of morphology

The porosity of collagen hydrogel was determined according to the known method.¹² Briefly, the weight of collagen hydrogel prepared by 0.012 mmol mL⁻¹ PR-CHO-4 solution was m₀, a 50 mL bottle which filled up ethanol was m₁, the collagen hydrogel was immersed in the ethanol and then degassed by supersonic treatment to confirm the pore in hydrogel was filled by ethanol, then the weight was m₂, when removed the hydrogel gently, the weight of left part was m₃. The porosity (P) of collagen was calculated by the formula, P = (m₂ − m₀ − m₃)/(m₁ − m₃) × 100%.

The morphology and distribution of pore size were characterized by using NOVA NANOSEM 430, samples were preliminarily treated by lyophilization and cut into 1 mm-thin slices. The sputter-coated with gold using EM SCD 500 (LEICA, Germany).

2.5 Measurements of PR-CHO

The synthesized products were identified by ¹H NMR, X-ray diffraction (XRD). XRD measurements were performed with powder samples using a (X'Pert Pro, PANalytical Co., the Netherlands). X-ray diffractometer samples were placed on a sample holder and scanned from 10 to 60° in 2θ at a speed of 10° min⁻¹. ¹H NMR was recorded on a AVANCE Digital 400 MHz NMR spectrometer (Bruker, Germany), using DMSO-d6 as solvent and tetramethylsilane (TMS) as internal standard.

2.6 FTIR measurements of crosslinked collagen

The FTIR spectra of lyophilized pure collagen and hydrogels were performed using a Bruker Vector 33 FTIR spectrometer at room temperature in the range between 4000 and 500 cm⁻¹ with a step width of 4 cm⁻¹ by KBr method.

2.7 Swelling properties test

The relative crosslink densities of the collagen were determined directly by ascertaining their swelling ratio. The swelling capacity studies were performed at 25 °C by immersing the weighed (W_d) lyophilized samples of cylinder with 2.5 cm in diameter and 1 cm in height in pH = 7.4 phosphate buffered saline (PBS). After soaking for 1, 2, 3, 5, 8, 12, 20, 50, 90 min, the samples were fished out and the water on the surface was wiped with filter paper, and weighed (W_w) immediately.²⁶ The swelling ratio (SR) was calculated using the formula, SR= (W_w − W_d)/W_d. The experimental plot was obtained from average of three samples.

2.8 Enzymatic stability

Enzymatic stability of collagen samples was tested with collagenase digestion experiments in vitro. Each sample (n = 3) was incubated in 1 mL of 0.1 M tris–HCl (pH = 7.4) containing Clostridial bacterial type I collagenase (200U, Sigma) and 0.05 M CaCl₂ for 24 h at 37 °C. The reaction was subsequently stopped by the addition of 0.2 mL of 0.25 M ethylene diamine tetraacetic acid (EDTA). The mixture was centrifuged for 10 min at 10 000 rpm at 10 °C and the supernatant was collected. After autoclaved, it was analyzed for hydroxyproline content with ultraviolet spectroscopy according to the previously described.⁹ The biodegradation degree is defined as the percentage of the released hydroxyproline from the collagen to the completely degraded one with the same composition and weight. The untreated collagen was used as a control.

2.9 Mechanical testing

The compressive strength of the cylindrical collagen hydrogel sample (8 mm in diameter, 10 mm in height) was determined by ELF3200 (BOSE, America) with a cylindrical piston of 10 mm diameter that compressed the samples with a speed of 1.0 mm min⁻¹ at 22 °C after at least 2 hours of incubation in PBS-buffer at room temperature. As a characteristic value the required force for compression to 70% of the initial height was used.

2.10 Cytotoxicity evaluation

The cytotoxicity of different concentration of PR-CHO-4 solution crosslinked collagen hydrogel was evaluated by culturing L929 fibroblasts in 10% FBS containing medium. Crosslinked collagen were immersed in 75% (v/v) ethanol for 12 h for sterilization, followed by washing with PBS (pH = 7.4) five times. The samples were then incubated in media for 24 h in a 5% CO₂ humidified incubator at 37 °C to obtain the extracts according to GB/T16886.5-2003/ISO 10993-5:1999. L929 were cultured in 96-well plates at a density of 5 × 10³ cells per well and incubated for 4 h. After that, the culture medium was replaced by the extracts, and incubated for 1, 3, 5 days at 37 °C in 5% CO₂. 0.25% glutaraldehyde crosslinked collagen was used as a control. The extracts were changed every day. Cell viability was determined using CCK-8 (Dojindo Co., Ltd., Kumamoto, Japan) according to the manufacture's instructions. Briefly, 10 μL of CCK-8 solution was added to each well of the 96-well plate and it was incubated at 37 °C under 5% CO₂ for 2 h. The optical density was read on a microplate reader (uQuant; Bio-Tek Instruments Inc.) at a wavelength of 450 nm.²⁷ The corresponding fluorescence microscopic images were obtained by Eclipse Ti-U (Nikon, Japan).

2.11 Statistical analysis

All quantitative data were presented as mean ± standard deviation from three separate experiments. Statistical analysis was performed by one-way analysis of variance (ANOVA). A value for p < 0.05 was considered statistically significant.

3. Results and discussions

3.1 ¹H NMR measurements

The ¹H NMR spectra of the PPR with different molecular weight are shown in Fig. 2a. The integral area ratios between methyl protons of PPG-NH₂ (a molecule of PPG-NH₂ 2000 contains 102 methyl protons, δ = 1.03–1.05 ppm) and C1–H of β-CD protons (a molecule of β-CD contains 7C1–H protons, δ = 4.83 ppm) were (26.02 : 7.0), (17.15 : 7.0), (12.76 : 7.0), (10.22 : 7.0) respectively in PPR-1, PPR-2, PPR-3, PPR-4. Then we could calculate that there were 4 β-CD units in PPR-1, 6 units in PPR-2, 8 units in PPR-3 and 10 units in PPR-4. The PPR with different molecular weight were then capped, reduced and oxidized to obtain the corresponding β-CD polyrotaxane multi-aldehydes (PR-CHO), the ¹H NMR spectra of PR-CHO with different molecule weight are shown in Fig. 2b. The ratio between new signal of an anomeric proton at δ = 4.93 and δ = 4.83 was 1 : 6, which indicated the mono-oxidation of β-CD unit threaded onto PPG-NH₂, the result was consistent with our previous study,²⁵ Jing Hu²⁸ and Cornwell.²⁹ According to the above discussion, the molecular weight of PR-CHO-1, PR-CHO-2, PR-CHO-3, PR-CHO-4 was about 8720, 10 960, 13 200, 15 440 respectively.

Fig. 2 The ¹H NMR spectra of (a) PPR-1, PPR-2, PPR-3, PPR-4 (b) PR-CHO-1, PR-CHO-2, PR-CHO-3, PR-CHO-4.

3.2 XRD measurements

The X-ray diffraction patterns of β-CD and PPR with different molecule weight are shown in Fig. 3. The three strong peaks of pure β-CD appear at 10.6°, 12.5° and 19.5° indicates a cage-type crystal structure. The diffractogram of PPR shows a diffraction pattern different from that of β-CD, which has three strong peaks appeared at 5.7°, 11.5° and 17.8°. This indicated that a different crystal type was formed. According to literature,³⁰ PPR exhibited characteristic diffraction peaks of 2θ at around 11.5° and 17.8° in its XRD patterns due to a channel-type crystal structure formed from the self-assembly of β-CDs with a linear polymer. These channel were arranged along the direction of the chain of PPG-NH₂ and linked with neighboring hydroxyl groups of β-CD. The XRD diffraction peaks of 2θ at around 5.7° (d = 1.524 nm) reflected the head to head packed structure³¹ as shown in the image inserted in Fig. 3. Besides, according to the diffractogram of PR-1, PR-2, PR-3, PR-4, the intensity of three strong peaks increase gradually, which indicated the crystallinity of PPR was better with the increase of units of β-CD.

Fig. 3 The XRD spectra of β-CD and PPR-1, PPR-2, PPR-3, PPR-4 and the stacking mode of PPR.

3.3 FTIR measurements of crosslinked collagen

Fig. 4 shows the FTIR spectra of collagen before and after the cross-linking by PR-CHO-4. After crosslinked with PR-CHO-4, neither the position nor the intensity of amide I band at around 1658 cm⁻¹ related to the triple helix conformation have changed with the increasing PR-CHO-4 content, indicating the collagen triple helix in the hydrogel was not destructed and preserved after the introduction of PR-CHO-4. In other words, as crosslinker, PR-CHO-4 would not destroy the back-bone structure of collagen. It is generally accepted that the hydrogen bonding plays a dominate role in the stabilization of collagen. When mixed with PR-CHO-4 under acidic conditions, the side chain of collagen, that's, hydroxyl, carboxyl, amino and amide groups provided the potential interacting sites for the formation of hydrogen bonds with the abundance of hydroxyl groups of PR-CHO-4. In Fig. 4, the intensity of amide II and III bands decrease with the increasing PR-CHO-4 content, indicating that the amount of –NH₂ group in collagen decreases and transforms to C=N groups. Namely, collagen was reacted with PR-CHO-4 to generate Schiff base. However, the peak at about 1660 cm⁻¹ for C=N was not observed in FTIR spectra because of the stronger amide I bands masks the band.

Fig. 4 FTIR spectra of collagen/PR-CHO-4 hydrogel with different contents of PR-CHO-4, (a) (0), (b) (0.0015), (c) (0.003), (d) (0.006), (e) (0.012).

3.4 Swelling properties of PR-CHO-4 interacted with collagen hydrogel

The swelling ratios of the hydrogel with different PR-CHO-4 content are shown in Fig. 5. It shows that during the process, the water uptake sharply increases at the initial stage, and then reaches to balance after 15 min. Meanwhile, the swelling ratios decrease as the concentration of PR-CHO-4 increases, indicating that the swelling behaviors of the collagen may depend on the number of hydrophilic free amino groups. As a result, hydrogel with higher crosslinking density holds less water than that with lower crosslinking density.

Fig. 5 Swelling ratios of the different contents of PR-CHO-4 crosslinked collagen (n = 3).

3.5 The porosity and SEM of collagen hydrogel

The porosity of tissue engineering scaffold plays an important role in cell adhesion and proliferation. A scaffold with high porosity ensures the effective area of cell contacts, which is convenient for the excretion of cell metabolin and supply of cell nutrition. Besides, the scaffold material with high porosity also ensures the water-holding capacity and then impedes the erosion of nutrition and body fluid.³² Generally speaking, the porosity of scaffold material for cell cultivation is at least 75%.³² As shown in Table 1, the average porosity of crosslinked collagen is about 90.27%, which can facilitate cell adhesion and proliferation. The SEM images of collagen crosslinked by PR-CHO-4 are shown in Fig. 6. It indicates that after crosslink, the collagen has an intensive, orderly and highly porous structure, the pores are interconnection and the diameters are ranged from 100 to 300 μm, which is similarly beneficial to the cell adhesion and proliferation.

Table 1 The porosity of collagen crosslinked with PR-CHO-4

Group	Shape	Porosity (%)
1	Rectangle	88.96
2	Rotundity	90.03
3	Irregularity	91.82

---

Fig. 6 The SEM of collagen modified by PR-CHO-4.

3.6 Collagenase resistance of PR-CHO-4 crosslinked collagen

Another effect of crosslinking is the increased resistance to enzymatic degradation by bacterial collagenase, which preferentially cleaves X–Gly bond of the –Gly–Pro–X–Gly–Pro–X– sequence in the nonpolar regions of the collagen molecule.³³ The results of collagenase biodegradation of the hydrogel are shown in Fig. 7. Significant reduction in the degradation of collagen was observed for the cryogel treated with PR-CHO-4 compared to pure collagen. The enzyme stability of the collagen increases with the increase of PR-CHO-4. The relative degradation could decrease to 28.68% when the concentration of PR-CHO-4 is 0.60%, which is lower than that treated with GA. It is well known that the degradation depended on the crosslinking density. The decreasing of relative degradation indicates the increasing of crosslinking density. The stability of collagen treated with PR-CHO-4 caused by protection of the active sites in collagen recognized by collagenase. In summary, the significant improved enzyme stability offered by PR-CHO-4 could be due to the effectiveness of the later in exhibiting better interaction with collagen through covalent and hydrogen bonded crosslinks.

Fig. 7 The relative degradation degree of collagen modified by PR-CHO-4, EDC and GA (n = 3).

3.7 Mechanical properties

The mechanical property is crucial for scaffolds used in tissue engineering, three-dimensional hydrogel scaffolds may undergo compression stresses during handling.¹¹ Fig. 8 shows the compressive modulus of collagen hydrogels crosslinked with varies crosslinkers of different concentration. It is clear that the compressive modulus increase with the increase of concentration of PR-CHO-4, when the concentration of the PR-CHO-4 is 0.012 mmol mL⁻¹, the compressive modulus is 116.07 kPa, which was much higher than that of pure collagen and that of crosslinked collagen with the same concentration of EDC and GA, which were 26.59 kPa and 90.23 kPa, respectively. It is ascribed to the increased covalent crosslinks, special structure and the pulley effect of PR-CHO-4. When collagen gel is squeezed from the outside, the pressure could be evenly dispersed by pulley effect, so the mechanical property of crosslinked collagen was increased significantly.

Fig. 8 The compress modulus of collagen modified by PR-CHO-4, EDC and GA (n = 3, *indicates a significant difference when compared with low concentrations of PR-CHO-4, EDC and GA (p < 0.05)).

3.8 L929 cell cytotoxicity

In our previous study,²⁵ we have demonstrated the cytotoxicity of PR-CHO. Fig. 9a shows the results of cytotoxicity on collagen hydrogel extracts prepared using PR-CHO-4, EDC and GA after incubation for 1 day, 3 days and 5 days, while the matrices for cell culture was used as a control and the corresponding fluorescence microscopic images are shown in Fig. 9b. The number of L929 on collagen hydrogel crosslinked with PR-CHO-4, GA and EDC extracts was nearly the same after incubation for 1 day, which were less than that of control group. EDC is known as a low cytotoxicity crosslinker since it is not incorporated into the crosslinked structure; however, urea derivatives such as N-acylurea and O-acylisourea groups derived from the crosslinking reaction process show cytotoxicity. Therefore, the L929 adhesion on AlCol–EDC was inhibited due to the cytotoxicity of these urea derivatives in AlCol–EDC. Meanwhile, in the case of AlCol–GA, it was thought that unreacted aldehyde groups in AlCol caused cytotoxicity. Moreover, PR-CHO-4 bearing AlCol would show little cytotoxicity even if it remained in AlCol–PR-CHO-4, as shown in Fig. 9a. After culture for 3 days and 5 days, the growth of L929 on AlCol–PR-CHO-4, AlCol–EDC and AtCol–GA extracts were observed. The L929 number on AlCol–PR-CHO-4 extracts was significantly increased, which was comparable to AlCol–EDC and was higher than the control group. However, there was no significant increase in L929 number on AtCol–GA extracts. In AlCol–GA, it was suggested that the release of GA-related molecules or hydrolyzed cytotoxic monomers from AlCol–GA caused the inhibition of cell growth at day 3 and day 5. Results from Fig. 9 indicates that AlCol–PR-CHO-4 has excellent cytocompatibility as compared to what AlCol–EDC and AtCol–GA exhibit. Therefore, this β-CD derivative crosslinker can be successfully applied for collagen crosslinking to produce more cytocompatible collagen gel with the potential for use in tissue-engineering scaffolds.

Fig. 9 (a) Proliferation of L929 fibroblasts exposed to various crosslinkers crosslinked with collagen (n = 3, *indicates a significant difference when compared with GA (p < 0.05)) (b) corresponding fluorescence microscopic images.

4. Conclusion

We have successfully prepared PPR-1, PPR-2, PPR-3 and PPR-4 and the corresponding PPR-CHO-1, PPR-CHO-2, PPR-CHO-3, and PPR-CHO-4 in a convenient and efficient method. Then the PR-CHO-4 was used to crosslink with collagen, and the properties of cross-linked collagen have been improved significantly. The biodegradation rate was improved a lot and the relative enzymatic degradation was 23.1%. The compression strength has been significantly improved and the maximum compression modulus of collagen was 116.07 kPa. The other properties of cross-linked collagen, including swelling behavior, porosity and the diameter of the pores, are greatly suitable for cell adhesion and proliferation. The cytotoxicity of collagen when crosslinked with PR-CHO-4 was also almost same as the EDC modified collagen. From these results, the PR-CHO-4 is a very excellent crosslinker, which can effectively improve the property of collagen and avoid some deficiencies caused by conditional crosslinkers. Considering the excellent properties, PR-CHO with different molecular weight can be used as bio-crosslinkers to produce more collagen gel derivatives, which show great potential to apply as scaffolds in tissue engineering.

Acknowledgements

This work was supported by the National Key Technologies R&D Program (2012BAI17B02), National Natural Science Foundation of China (no. 51232002, 51273072) and the Natural Science Foundation of Guangdong Province (2012A080800015, 2012A080203010), which are gratefully acknowledged.

Notes and references

1. D. G. Wallace and J. Rosenblatt, Adv. Drug Delivery Rev., 2003, 55, 1631–1649.
2. Y. Peng, V. Glattauer, J. A. Werkmeister and J. A. M. Ramshaw, J. Cosmet. Sci., 2004, 55, 327–341.
3. C. H. Lee, A. Singla and Y. Lee, Int. J. Pharm., 2001, 221, 1–22.
4. Q. J. Lu, K. Ganesan, D. T. Simionescu and N. R. Vyavahare, Biomaterials, 2004, 25, 5227–5237.
5. C. R. Lee, A. J. Grodzinsky and M. Spector, Biomaterials, 2001, 22, 3145–3154.
6. M. W. Lee, H. J. Chen and S. W. Tsao, Carbohydr. Polym., 2010, 82, 920–926.
7. J. L. C. van Susante, J. Pieper, P. Buma, T. H. van Kuppevelt, H. van Beuningen, P. M. van Der Kraan, J. H. Veerkamp, W. B. van den Berg and R. P. H. Veth, Biomaterials, 2001, 22, 2359–2369.
8. L. Zheng, H. S. Fan, J. Sun, X. N. Chen, G. Wang, L. Zhang, Y. J. Fan and X. D. Zhang, J. Biomed. Mater. Res., Part A, 2010, 93, 783–792.
9. L. Castaneda, J. Valle, N. Yang, S. Pluskat and K. Slowinska, Biomacromolecules, 2008, 9, 3383–3388.
10. K. S. Weadock, E. J. Miller, E. L. Keuffel and M. G. Dunn, J. Biomed. Mater. Res., 1996, 32, 221–226.
11. X. M. Zhang, L. Xu, X. Huang, S. C. Wei and M. L. Zhai, J. Biomed. Mater. Res., Part A, 2012, 100, 2960–2969.
12. K. S. Weadock, E. J. Miller, L. D. Bellincampi, J. P. Zawadsky and M. G. Dunn, J. Biomed. Mater. Res., 1995, 29, 1373–1379.
13. J. M. McPherson, S. Sawamura and R. Armstrong, J. Biomed. Mater. Res., 1986, 20, 93–107.
14. M. J. B. Wissink, R. Beernink, J. S. Pieper, A. A. Poot, G. H. M. Engbers, T. Beugeling, W. G. van Aken and J. Feijen, Biomaterials, 2001, 22, 151–163.
15. L. L. H. Huang, H. W. Sung, C. C. Tsai and D. M. Huang, J. Biomed. Mater. Res., 1998, 42, 568–576.
16. L. Garcia-Rio, F. J. Otero-Espinar, A. Luzardo-Alvarez and J. Blanco-Mendez, Curr. Top. Med. Chem., 2014, 14, 478–493.
17. S. J. Li, G. H. Weng, W. Lin, Z. B. Sun, M. Zhou, B. Zhu, Y. Ye and J. Wu, Polym. Chem., 2014, 5, 3994–4001.
18. B. L. Tardy, H. H. Dam, M. M. J. Kamphuis, J. J. Richardson and F. Caruso, Biomacromolecules, 2014, 15, 53–59.
19. H. Zou, W. Guo and W. Z. Yuan, J. Mater. Chem. B, 2013, 1, 6235–6244.
20. J. Li and X. J. Loh, Adv. Drug Delivery Rev., 2008, 60, 1000–1017.
21. J. J. Li, F. Zhao and J. Li, Appl. Microbiol. Biotechnol., 2011, 90, 427–443.
22. D. D. Hou, X. M. Tong, H. Q. Yu, A. Y. Zhang and Z. G. Feng, Biomed. Mater., 2007, 2, S147–S152.
23. C. Moon, Y. M. Kwon, W. K. Lee, Y. J. Park, L. C. Chang and V. C. Yang, J. Biomed. Mater. Res., Part A, 2008, 84, 238–246.
24. C. D. Mu, F. Liu, Q. S. Cheng, H. L. Li, B. Wu, G. Z. Zhang and W. Lin, Macromol. Mater. Eng., 2010, 295, 100–107.
25. S. Liu, J. Cai, L. Ren, L. Wang and Y. J. Wang, RSC Adv., 2014, 4, 18608–18611.
26. A. Kishida, N. Kwangwoo and T. Kimura, Biomaterials, 2007, 28, 1–8.
27. H. Saito, S. Murabayashi, Y. Mitamura and T. Taguchi, J. Mater. Sci.: Mater. Med., 2008, 19, 1297–1305.
28. H. J. Schneider, F. Hacket, V. Rudiger and H. Ikeda, Chem. Rev., 1998, 98, 1755–1785.
29. M. J. Cornwell, J. B. Huff and C. Bieniarz, Tetrahedron Lett., 1995, 36, 8371–8374.
30. G. Wenz, B. H. Han and A. Muller, Chem. Rev., 2006, 106, 782–817.
31. K. Harata, Chem. Rev., 1998, 98, 1803–1827.
32. D. W. Hutmacher, Biomaterials, 2000, 21, 2529–2543.
33. R. E. Galardy and D. Grobelny, Biochemistry, 1983, 22, 4556–4561.

---