Preparation of keratin/chlorhexidine complex nanoparticles for long-term and dual stimuli-responsive release

Abstract

Nanoscale polyion complex formation via the electrostatic complexation of a polyelectrolyte and a charged drug is the most convenient method for building a drug delivery system that simultaneously realizes the carrier preparation and drug embedding. Herein, we prepared keratin/chlorhexidine complex nanoparticles (KCNPs) based on a drug-induced ionic gelation technique without using a crosslinker, organic solvent or surfactant. These KCNPs exhibited good stability even after having been long-standing for 14 days. The KCNPs were characterized using FTIR, DLS, SEM and TEM. It was found that these nanoparticles had a spherical morphology with a diameter of about 180 nm, and a negatively charged surface with a zeta potential of about −39.1 mV. The cell toxicity of the KCNPs at different dosage levels was evaluated using the MTT assay method, indicating their slight cytotoxicity at lower dosages. The antibacterial activity against E. coli and S. aureus was determined using the zone of inhibition method. It seemed that the KCNPs had better antibacterial activity against S. aureus than against E. coli. Drug delivery profiles showed that the chlorhexidine (CHX) loaded nanoparticles exhibited both pH- and glutathione-responsive character. These keratin-based complex nanoparticles can be regarded as a valuable stimuli-responsive strategy for the delivery of anticancer agents.

---

1. Introduction

Keratin is a chief component found in hair, skin, fur, wool, horn, and feathers. It has a high content of cysteine and serine, and a large number of hydroxyl amino acids. Keratin based biomaterials have emerged as potential candidates for many biomedical and biotechnological applications due to their intrinsic biocompatibility, biodegradability, mechanical durability, and natural abundance.¹ Keratin based films can be used as a wound dressing,^2,3 implantable device coating^4,5 and cell encapsulant,^6,7 and in ocular surface reconstruction.^8,9 Keratin-based hydrogels are neuroinductive and capable of facilitating regeneration in a peripheral nerve injury model in mice.^10,11 Keratin hydrogels can also act as a hemostatic agent in a rabbit model for lethal liver injury,¹² a carrier for rhBMP-2,¹³ and a platform for antibiotics.¹⁴ Keratin-based fibers have a potential use in tissue engineering scaffolds. Due to its poor mechanical property, keratin is usually blended with biodegradable polymers such as poly(3-hydroxybutyrate-co-3-hydroxyvalerate),¹⁵ poly(lactic acid),¹⁶ and poly(ε-caprolactone).¹⁷

Keratin should be an ideal drug carrier due to its advantages such as good biocompatibility, biodegradability, absorbability, and non-immunogenicity. In addition, keratin is richly negatively charged, which favors the electrostatic absorption of small positively charged molecules, such as acid salt typed drugs. Furthermore, the large amount of carboxyl groups within keratin makes it a pH responsive smart drug carrier candidate. However, little work has been done on keratin-based drug carriers. The van Dyke group has described the delivery and activity of the antibiotic ciprofloxacin delivered from a keratin hydrogel.¹⁴ The Wang group has prepared a keratin hydrogel and film for controlling drug release.^18,19 The Liu group has synthesized poly(ethylene glycol) (PEG) and poly(N-(2-hydroxypropyl)methacrylamide) (PHPMA) grafted keratin and used it as a dual reduction and enzyme responsive drug carrier.^20,21

Various methods have been used to prepare protein-based nanoparticles, such as desolvation,²² nanoprecipitation,²³ self-assembly,²⁴ and so on. Among them, the desolvation and nanoprecipitation procedures need the use of organic solvents or additives, which may cause protein denaturation. For self-assembly procedures, protein and drug solutions are directly mixed to allow their interaction based on specific forces such as hydrogen bonding and hydrophobic interactions. For our modified keratin, a large amount of carboxylate groups (–COO⁻) can strongly attract the cationic drug. This electrostatic complexation leads to the formation of nanoparticles. Based on electrostatic interactions, Malek et al.²⁵ have prepared cisplatin-incorporated nanoparticles using the drug-induced ionic gelation technique. The ionic gelation technique has been used widely to prepare chitosan nanoparticles.^26–32 To the best of our knowledge, this method has not yet been reported to prepare protein nanoparticles.

Chlorhexidine (CHX) is widely used to treat and prevent skin and mucosal infections with low toxicity. In dentistry, it has been recognized as a gold standard against plaque and gingivitis for three decades.³³ Due to its cationic charge, CHX is used as a drug model. The purpose of this study is to develop a facile method to fabricate stable keratin-based drug delivery systems without any chemical cross-linkers and surfactants. Herein, keratin/CHX complex nanoparticles (KCNPs) were prepared using the drug induced ionic gelation technique via electrostatic interactions (Fig. 1). Then, the size, zeta potential, and drug-loading capacity of the KCNPs were characterized using DLS, SEM, and TEM. An MTT assay was used to evaluate their cell toxicity. The pH dependent CHX delivery behavior was tested at pH 5.29, 7.4 and 9.18. The glutathione (GSH) sensitive release behavior was also tested at pH 7.4. These KCNPs have potential for long-term and controlled release of CHX for antibacterial applications.

Fig. 1 Schematic representation of the preparation of keratin/CHX complex nanoparticles (KCNPs) using the ionic gelation method.

2. Materials and methods

2.1 Materials

Chlorhexidine acetate (C₂₂H₃₀N₁₀Cl₂·2C₂H₄O₂) was supplied by Aladdin (Shanghai, China) with a purity of 98%. E. coli (ATCC 25922) and S. aureus (ATCC 25923) were provided by the Jiangsu Provincial Center for Disease Prevention and Control, China. All other chemicals were of analytical grade and were used without further purification.

2.2 Extraction of keratin^2,34,35

Human hair was washed with soap and 70% ethanol to remove surface oil, followed by being rinsed extensively with water and dried and then cut into short pieces. The pretreated hair was mixed with urea, sodium dodecyl sulfate (SDS), 2-mercaptoethanol and water. The mixture was kept under stirring for 48 h at 65 °C and then filtered. Subsequently, the filtrate was dialysed against deionized water for 48 h to afford a colorless solution. The dialysate was then allowed to react with iodoacetic acid for stability. Finally, this dialysate was dialysed again and lyophilized to obtain S-(carboxymethyl) keratin. The molecular weight of keratin was analyzed using SDS-polyacrylamide gel electrophoresis (PAGE).

2.3 Preparation of KCNPs

KCNPs were prepared using the ionic gelation process. Keratin and CHX were each dissolved in an aqueous solution to obtain a stock solution with a final concentration of 1 mg mL⁻¹. Then, the CHX solution was added dropwise at a speed of 1.5 mL h⁻¹ under constant stirring to the keratin solution according to the ratios listed in Table 1, followed by stirring for 60 min and storage at room temperature overnight. The mixture was dialyzed for 1 day to remove unstable adsorbed CHX. Subsequently, the free CHX dialysate was measured using an ultraviolet spectrometer to determine the CHX loading content and encapsulation efficiency. The dialysate was lyophilized for future studies. The drug loading content (LC) and drug encapsulation efficiency (EE) were calculated according to the following equations:

Table 1 Parameters of the as-prepared KCNPs

Samples	EE (%)	LC (%)
Weight ratio of keratin and CHX = 9 : 1	91.2	9.2
Weight ratio of keratin and CHX = 7 : 3	64.1	21.5
Weight ratio of keratin and CHX = 5 : 5	46.1	31.6

---

2.4 Characterization of the KCNPs

The infrared spectra of keratin and the KCNPs were recorded within the range of 4000–400 cm⁻¹ using a Perkin-Elmer FT-IR spectrometer in KBr. The hydrodynamic diameter distribution and zeta potentials of the nanoparticles were determined using the dynamic light scattering (DLS) method using a Malvern Zetasizer Nano ZS90 system (Malvern Instruments Company Limited, British), in which freeze-dried KCNPs were dispersed in different pHs of phosphate buffer solutions. All analyses were triplicated and the results were the average of the three runs. The morphology of the KCNPs was observed using transmission electron microscopy (TEM, JEM-100S, JEOL, Japan) and scanning electron microscopy (SEM, Hitachi SU 3500, Japan).

2.5 In vitro pH independent CHX release study

100 mg of KCNPs in 5 mL buffer solution was placed into a dialysis bag (MWCO 3500). Subsequently, the dialysis bag was dipped into a receptor compartment containing 200 mL dissolution medium, and was shaken gently at 37 ± 0.5 °C. Then, the medium was adjusted with pH 5.29 PBS, pH 7.4 PBS, and pH 9.18 sodium tetraborate buffer to test the release of CHX. The receptor compartment was closed to prevent evaporation losses from the dissolution medium. Periodically, 3 mL of the release medium was withdrawn and 3 mL of fresh PBS was added to the system. The CHX concentration in the sampled medium was determined using a UV spectrometer with absorption at the wavelength of 260 nm.

2.6 In vitro GSH sensitive CHX release study

Regarding the GSH sensitive delivery, a similar procedure to that described above was followed except for the presence of GSH. The concentrations of GSH both inside and outside the dialysis bag were kept at 10 μM and in pH 7.4 PBS solution during the release experiments to mimic the redox conditions in blood plasma. At certain time intervals, the CHX concentrations of the outer side of the dialysis bag were determined using a UV spectrometer with absorption at the wavelength of 260 nm.

2.7 Bacterial inhibition

The antimicrobial activity of the KCNPs was investigated against E. coli as the model for Gram-negative bacteria and S. aureus as the model for Gram-positive bacteria by the disc diffusion method. The bacteria were incubated in nutrient agar at 37 °C to reach 10⁶ CFU mL⁻¹. Agar plates were streaked with a sterile swab moistened with the KCNPs. Then, the bacteria were coated on the agar surface evenly and the plates were incubated overnight at 37 °C. The reaction of the microorganisms to the KCNPs was determined using the size of the inhibitory zone. When the materials have an excellent antibacterial activity, the inhibitory zones are very large.

2.8 Cytotoxicity assay using MTT

The cytotoxicity of the KCNPs against NIH-3T3 cells was evaluated using an MTT assay. The cells in a DMEM medium containing 10% fetal bovine serum (FBS, Gibco) and 1% penicillin G–streptomycin were seeded into 96-well flat-bottomed plates with a density of 1 × 10⁴ cells per well and then were incubated in CO₂ atmosphere at 37 °C for 12 h. Next, the medium in each well was removed and the samples (100 μL) with various concentrations in DMEM were added to the wells. Subsequently, the cells were incubated for 48 h under the same conditions. The medium was replaced with 90 μL of fresh medium with 10 μL of the MTT solution (5 mg mL⁻¹). After incubation for 4 h, the solution was removed, leaving the precipitate. Then, 100 μL of dimethyl sulfoxide (DMSO) was added to each well and was then oscillated for 30 min in the dark at room temperature. The cell viability was measured using a microplate reader (BioTek Synergy2) at a wavelength of 570 nm.

2.9 Statistical analysis

Results are displayed as means ± standard deviation (SD). Statistical analysis was performed using one-way analysis of variance (ANOVA) with a Tukey posthoc method, and a significance level of p < 0.05 was chosen for all the tests.

3. Results and discussion

3.1 Preparation of the KCNPs

Generally, keratin can be extracted using an oxidative method or a reduction method. Herein, we extracted keratin from human hair using the reduction strategy. The molecular weight of the extracted keratin was determined using SDS-PAGE analysis (Fig. 2). The patterns revealed two major, diffuse clusters of bands: 25 and 40 kDa. So, the molecular weight of keratin was concentrated on ca. 25 and 40 kDa.

Fig. 2 The electrophoresis separation patterns of the human hair keratin using an SDS-PAGE assay.

The ionic gelation method was used to prepare KCNPs with different ratios between keratin and CHX of 9/1, 7/3, and 5/5. The stability of the resulting KCNPs after 4 h and 14 days standing is shown in Fig. S1.† It shows that the KCNPs at the ratio of 9/1 are stable even when standing for as long as 14 days. As for the ratios of 7/3 and 5/5, the prepared nanoparticles are unstable after standing for 14 days. So, the ratio of 9/1 was fixed for the sample preparation.

The drug encapsulation efficiency (EE) and loading content (LC) of the KCNPs are listed in Table 1. Generally, a high drug concentration leads to a great drug loading content while a high keratin concentration tends to decrease drug encapsulation efficiency. Herein, the drug loading content was gradually increased while the encapsulation efficiency decreased with increasing mixed ratio between CHX and keratin. For keratin/CHX nanoparticles complexed at the ratio of 9 to 1, the EE and LC are 91.2% and 9.2%, respectively. From the above results, we can confirm that cationic drugs can be effectively entrapped in keratin molecules due to their strong electrostatic interaction.

3.2 Characterization of the KCNPs

3.2.1 FT-IR and UV-Vis spectrum analysis. The results of the FTIR spectra of keratin and KCNPs are shown in Fig. S2.† Regarding keratin, the characteristic absorption peaks near 1646, 1548 and 1245 cm⁻¹ are known as amide I, amide II, and amide III, respectively. Amide I, mainly related to C=O stretching, is useful for the analysis of the secondary structure of the proteins. The peaks at 990 cm⁻¹ and 614 cm⁻¹ are the characteristic absorption peaks of the C–S and S–S bonds.^36,37 In the case of KCNPs, there is no shift for the amide I peak, indicating the maintenance of the keratin conformation.

UV-Vis absorption measurement is a very simple method to explore the structural change and to know the complex formation. In the present study, we have recorded the UV-Vis absorption spectra of keratin and the keratin–CHX complex. In Fig. 3, the spectra showed a large blue shift in λ_max from 276 nm to 263 nm, which can be related to complex formation between CHX and keratin through hydrogen bonds and electrostatic interaction.^38,39

Fig. 3 UV-Vis spectra of keratin and keratin–CHX.

3.2.2 Size and morphology. The hydrodynamic diameters of the resulting KCNPs in medium with various pH values are investigated and presented in Table 2. It can be seen that the mean diameter of the nanoparticles increases from 163 to 195 nm when the pH value of the medium is changed from 5.29 to 9.18. It may be attributed to the fact that the keratin moiety of the KCNPs becomes more hydrated and swells because of deprotonation due to the increasing pH value.

Table 2 Sizes and zeta potentials of the KCNPs determined using DLS

pH value	Size distribution by intensity	PDI	Z-Average (d nm)	Zeta potential (mV)
7.0		0.260	180	−39.1 ± 1.5
5.29		0.394	163	−15.3 ± 1.5
7.4		0.365	185	−15.5 ± 0.6
9.18		0.299	195	−29.5 ± 1.4

---

A typical TEM image of the KCNPs is shown in Fig. 4a, from which it can be seen that the KCNPs are solid spheres with a diameter of about 176 nm, which is very close to the result obtained using DLS. Generally, protein molecules exist as a form of colloid in water. As a result, there is no obvious difference between the sizes of the nanoparticles in aqueous medium and in the dry state. The morphology of the KCNPs is also observed using SEM and the results are shown in Fig. 4b. A statistical study of the particle size gives an average diameter of 170 nm.

Fig. 4 TEM (a) and SEM (b) images of KCNPs.

3.2.3 Zeta potential measurements. Surface charges and thereby the stabilities of the prepared KCNPs are determined using zeta potential measurements. The zeta-potential directly relates to the net charges on the surface of the KCNPs. The zeta potential value for the KCNPs is found to be −39.1 mV. This value lies within the stable range, indicating that the KCNPs are stable and possess a negative surface charge. The negative charge results from the excess of carboxyl groups on keratin. The zeta potentials of the KCNPs at different pH values are given in Table 2. The zeta potentials of the KCNPs are −15.3 mV and −29.5 mV in the pH range from 5.29 to 9.18 due to the deprotonated carboxyl groups of keratin. As the pH reaches 5.29, the net charges on the KCNP surfaces are too low to stabilize the nanoparticles, resulting in precipitation (Fig. S3†).

3.3 In vitro pH independent CHX delivery profile

Keratin seems to be very appropriate to load CHX due to its rich carboxylic acid groups. Thanks to the electrostatic attraction between the cationic CHX molecules and the anionic keratin, CHX is accordingly loaded into the KCNPs with a satisfactory drug loading content of 9.2% and an encapsulation efficiency of 91.2% (Table 1). Fig. S4† shows the standard curve of CHX with a high goodness of fit as the function of concentration at 260 nm. To investigate the release characteristics of CHX from the KCNPs, the dialysis method is used. To demonstrate the pH-sensitive release, the release of CHX is studied over a period of 240 h under acidic (PBS, pH 5.29), 120 h under neutral (PBS, pH 7.4), and 120 h under basic (sodium tetraborate buffer, pH 9.18) conditions. Fig. 5 shows the whole release profiles of CHX from the KCNPs. No initial burst release occurs, indicating the strong complexation of CHX with keratin.

Fig. 5 Release profile of CHX from the KCNPs. (a) Release profiles at pH 7.4 and 9.18; (b) release profile at pH 5.29; and (c) linear regression curves of the release data fitting with a pseudo-second order kinetics mode for KCNPs.

In PBS with a pH value of 7.4, which corresponds to the physiological conditions of the blood stream, only ∼23% of CHX is delivered in the initial 100 h (Fig. 5a). The dialysate keeps clear during the whole release period. It indicates that the KCNPs are stable for a long time at physiological pH (Fig. S3†). The release of CHX is extraordinarily slow in PBS with a pH value of 9.18 (Fig. 5a). For example, only ∼8% of the loaded CHX is released within 100 h, suggesting that the CHX release from the KCNPs in basic environments is significantly suppressed. It is because the enhanced negative charges of the KCNPs adsorb CHX tightly, which inhibits CHX release.

The release of CHX is even slower at acidic pH 5.29 than at pH 7.4 in the initial 60 h (Fig. 5b). Depending on the protonation of keratin, the carboxylate groups (COO⁻) of the KCNPs can be converted into neutral carboxylic acid groups (COOH) by reducing the pH value; the hydrophilicity of KCNPs decreases. So, KCNPs become smaller and denser at pH 5.29 (Table 2), which suppresses the delivery. Thus, the release rate is faster than that at pH 9.18. It is due to the protonation of CHX increasing its hydrophilicity, which enhances the delivery. After releasing for the first 60 h, the release rate increases sharply. It is due to the synergetic effects of the dual protonation of CHX and keratin. The protonation of CHX increases the hydrophilicity of CHX. Meanwhile, the protonation of keratin weakens the electrostatic attraction of CHX. These two above synergetic effects both accelerate the delivery of CHX, resulting in this sharp release. It is interesting that the release profile shows a straight line, indicating a constant speed release of CHX. The pharmacokinetic release behavior follows a zero order equation. Fig. S3† shows the effect of the pH value on the stability of the KCNPs. The KCNPs tend to aggregate and deposit at acidic pH values, which are close to the iso-electric point of keratin. Such a pH-sensitive drug release behavior is desirable for tumor treatment since the ideal antitumor drug release should be slow in the neutral environment of the in vivo systemic circulation and relatively faster in the weak acidic environment of tumor tissues.⁴⁰ In our study, about 39.5% of CHX is released within 240 h at pH 5.29. These results suggest that keratin-based nanocarriers are satisfactory for long term anticancer drug loading and release. The possible process of drug release could be expected. The pH-sensitive antitumor drug-loaded KCNPs would accumulate and the drug would be released at the tumor site due to the enhanced permeability and retention (EPR) effect.

The release process of CHX may be described with pseudo-first-order kinetics or pseudo-second-order kinetics equations.⁴¹ The pseudo-first-order kinetics equation can be represented in a linear form as

ln(q_e − q_t) = ln q_e − k₁t (1)

where q_e and q_t are the equilibrium release amount and the release amount at any time (t), respectively, and k₁ is the rate constant of pseudo-first-order release kinetics. If the pseudo-first-order kinetics model is applicable, the plot of ln(q_e − q_t) vs. t will be linear, and the k₁ value can be obtained from the slope of the linear plot.

The pseudo-second-order kinetics equation can be represented in a linear form as

t/q_t = 1/(k₂q_e²) + t/q_e, (2)

where k₂ is the rate constant of pseudo-second-order release kinetics. If this kinetics model is applicable, the plot of t/q_tvs. t will be linear, which allows computation of k₂. Fig. 5c shows the plot of t/q_tvs. t for CHX release at pH 7.4 and 9.18. Fair straight lines are obtained and the correlation coefficients of the fitting curves reach at least 0.991. It is found that the pseudo-second-order model is more satisfactory for describing the release process for the simulation of the above two kinetics models for the release kinetics data.

3.4 In vitro GSH sensitive CHX release study

The release of CHX was also investigated at a GSH concentration of 10 μM corresponding to the level of GSH in blood plasma and pH 7.4 (Fig. 6a). It can be seen that about 45% of the loaded CHX can be finally released after reaching the plateau of 100 h. In comparison, only 23% of CHX can be released without GSH at the plateau of 70 h. These results could be attributed to the fact that the disulfide bonds of the KCNPs were partially cleaved by GSH through the reduction reaction, which resulted in an enhanced release of CHX. Fig. 6b shows the size of the KCNPs in the presence of 10 μM GSH at the delivery times of 0 h, 48 h, 72 h, and 120 h. The sizes of the KCNPs become larger with increasing delivery time, indicating the broken disulfide bonds of the KCNPs. The swollen KCNPs lead to a sufficient release of CHX as compared to that at pH 7.4 without GSH. Fig. 6c shows the accumulated release of CHX at pH 5.29 with and without 10 μM GSH as a function of time. The delivery is accelerated under the action of GSH as compared to that without GSH.

Fig. 6 (a) Release profiles at pH 7.4 with and without 10 μM GSH; (b) size of the KCNPs determined using DLS at pH 7.4 and 10 μM GSH with the delivery times of 0 h, 48 h, 72 h, and 120 h; and (c) release profiles at pH 5.29 with and without 10 μM GSH.

3.5 In vitro cytotoxicity using an MTT assay

An antibacterial material requires low toxicity to normal cells at appropriate concentrations for killing bacteria efficiently. The cytotoxicity of the KCNPs towards fibroblast cells is evaluated using the MTT method. As shown in Fig. 7, at a lower concentration (2.5 μg mL⁻¹ of KCNPs, i.e., 0.2 μg mL⁻¹ of CHX), the cell viability of the KCNPs is 66% as compared to the FBS control, indicating their slight cytotoxicity. However, as the concentration gets even higher, they show significant cytotoxicity. Ostad et al. have used endometrial cells to assess the cytotoxicity of CHX and CHX-releasing devices. The results indicate that CHX is toxic at the concentration of 1 μg mL⁻¹.⁴² It is also reported that 0.6 and 20 mg mL⁻¹ of CHX decrease cell metabolism and viability by 60 and 70%, respectively.⁴³ According to our drug loading content data (9.2%), 2.5 μg mL⁻¹ of KCNPs is the equivalent of 0.23 μg mL⁻¹ of CHX. It seems that the cytotoxicity of CHX is enhanced when CHX is complexed with keratin.

Fig. 7 Cell viability of the KCNPs at various concentrations towards NIH-3T3 cell growth determined using an MTT viability assay (* indicates a significant difference at the p < 0.05 level).

3.6 Antibacterial test

CHX is used worldwide as a cationic antimicrobial agent against oral microbiota. The antibacterial activity of the KCNPs against E. coli and S. aureus is determined using the zone of inhibition method. The disks produced a clear zone of inhibition (Fig. 8). The zone of inhibition for S. aureus looks larger than that for E. coli. It seems that the KCNPs have better antibacterial activity against S. aureus than E. coli. The above results are consistent with the known conclusion that CHX is more effective against Gram-positive microbes than against Gram-negative microbes.⁴⁴

Fig. 8 Antibacterial inhibition zones of the KCNPs against E. coli (a) and S. aureus (b).

4. Conclusion

In summary, keratin/CHX complex nanoparticles (KCNPs) are successfully prepared with a ca. 180 nm diameter. The CHX release behavior shows that these nanoparticles are both pH and GSH sensitive. Under acidic and blood plasma level GSH conditions, the KCNPs exhibit more rapid and long term drug release. Moreover, a low concentration of the KCNPs shows a slight cytotoxicity and simultaneously maintains their antibacterial property. In a word, keratin-based drug delivery systems can be regarded as a valuable pH-responsive strategy for the delivery of anticancer agents such as the doxorubicin hydrochloride salt (DOX·HCl).

Acknowledgements

This work was supported by the National Natural Science Foundation of China (21274063) and PAPD of Jiangsu Higher Education Institutions.

References

1. J. G. Rouse and M. E. van Dyke, Materials, 2010, 3, 999–1014.
2. J. Yuan, J. Geng, Z. Xing, K. J. Shim, I. Han, J. C. Kim, I. K. Kang and J. Shen, J. Tissue Eng. Regener. Med., 2015, 9, 1027–1035.
3. A. Vasconcelos, A. P. Pêgo, L. Henriques, M. Lamghari and A. Cavaco-Paulo, Biomacromolecules, 2010, 11, 2213–2220.
4. J. J. Martin, J. M. Cardamone, P. L. Irwin and E. M. Brown, Colloids Surf., B, 2011, 88, 354–361.
5. K. Yamauchi, M. Maniwa and T. Mori, J. Biomater. Sci., Polym. Ed., 1998, 3, 259–270.
6. K. Yamauchi and A. Khoda, Colloids Surf., B, 1997, 9, 117–119.
7. R. Silva, B. Fabry and A. R. Boccaccini, Biomaterials, 2014, 35, 6727–6738.
8. M. Borrelli, N. Joepen, S. Reichl, D. Finis, M. Schoppe, G. Geerling and S. Schrader, Biomaterials, 2015, 42, 112–120.
9. S. Reichl, M. Borrelli and G. Geerling, Biomaterials, 2011, 32, 3375–3386.
10. L. A. Koman, A. Atala, M. V. Dyke, P. Sierpinski, J. Garrett, J. Ma, P. Apel, D. Klorig and T. Smith, Biomaterials, 2008, 29, 118–128.
11. P. J. Apel, J. P. Garrett, P. Sierpinski, J. Ma, A. Atala, T. L. Smith, L. A. Koman and M. E. van Dyke, J. Hand Surg., 2008, 33, 1541–1547.
12. T. Aboushwareb, D. Eberli, C. Ward, C. Broda, J. Holcomb, A. Atala and M. van-Dyke, J. Biomed. Mater. Res., Part B, 2009, 90, 45–54.
13. C. J. Kowalczewski, S. Tombyln, D. C. Wasnick, M. R. Hughes, M. D. Ellenburg, M. F. Callahan, T. L. Smith, M. E. van Dyke, L. R. Burnett and J. M. Saul, Biomaterials, 2014, 35, 3220–3228.
14. J. M. Saul, M. D. Ellenburg, R. C. de Guzman and M. V. Dyke, J. Biomed. Mater. Res., Part A, 2011, 98, 544–553.
15. J. Yuan, Z. C. Xing, S. W. Park, J. Geng, J. Shen, W. Meng, K. J. Shim, I. S. Han, J. C. Kim and I. K. Kang, Macromol. Res., 2009, 17, 850–855.
16. J. Yuan, J. Shen and I. K. Kang, Polym. Int., 2008, 57, 1188–1193.
17. A. Edwards, D. Jarvis, T. Hopkins, S. Pixley and N. Bhattarai, J. Biomed. Mater. Res., Part B, 2015, 103, 21–30.
18. J. H. Guo, S. J. Pan, X. C. Yin, Y. F. He, T. Li, R. M. Wang, J. Appl. Polym. Sci. DOI:10.1002/app.41572.
19. X. C. Yin, F. Y. Li, Y. F. He, Y. Wang and R. M. Wang, Biomater. Sci., 2013, 1, 528–536.
20. Q. M. Li, L. J. Zhu, R. G. Liu, D. Huang, X. Jin, N. Che, Z. Li, X. Z. Qu, H. L. Kang and Y. Huang, J. Mater. Chem., 2012, 22, 19964–19973.
21. Q. M. Li, S. N. Yang, L. J. Zhu, H. L. Kang, X. Z. Qu, R. G. Liu and Y. Huang, Polym. Chem., 2015, 6, 2869–2878.
22. M. Rahimnejad, N. Mokhtarian and M. Ghasemi, Afr. J. Biotechnol., 2009, 8, 4738–4743.
23. S. A. Khan and M. Schneider, Macromol. Biosci., 2013, 13, 455–463.
24. S. Han, M. Li, X. Liu, H. Gao and Y. Wu, Colloids Surf., B, 2013, 102, 833–841.
25. S. J. Malek, R. Khoshchehreh, N. Goodarzi, M. R. Khoshayand, M. Amini, F. Atyabi, M. Esfandyari-manesh, S. Tehrani, R. M. Jafari, M. S. Maghazei, F. Alvandifara, M. Ebrahimi and R. Dinarvand, Colloids Surf., B, 2014, 122, 350–358.
26. A. L. de Pinho Neves, C. C. Milioli, L. Müller, H. G. Riella, N. C. Kuhnen and H. K. Stulzer, Colloids Surf., A, 2014, 445, 34–39.
27. J. Zhang, C. Tang and C. H. Yin, Biomaterials, 2013, 34, 3667–3677.
28. E. N. Koukaras, S. A. Papadimitriou, D. N. Bikiaris and G. E. Froudakis, Mol. Pharmaceutics, 2012, 9, 2856–2862.
29. Y. Hou, J. Hu, H. Park and M. Lee, J. Biomed. Mater. Res., Part A, 2012, 100, 939–947.
30. W. Fan, W. Yan, Z. Xu and H. Ni, Colloids Surf., B, 2012, 90, 21–27.
31. F. M. Goycoolea, G. Lollo, C. Remuñán-López, F. Quaglia and M. J. Alonso, Biomacromolecules, 2009, 10, 1736–1743.
32. Q. Gan, T. Wang, C. Cochrane and P. McCarron, Colloids Surf., B, 2005, 44, 65–73.
33. C. G. Jones, Periodontol 2000, 1997, 15, 55–62.
34. K. Yamauchi, A. Yamauchi, T. Kusunoki, A. Kohda and Y. Konishi, J. Biomed. Mater. Res., 1996, 31, 439–444.
35. P. M. M. Schrooyen, P. J. Dijkstra, R. C. Oberthür, A. Bantjes and J. Feijen, J. Agric. Food Chem., 2000, 48, 4326–4334.
36. J. Z. Shao, J. H. Zheng, J. Q. Liu and C. M. Carr, J. Appl. Polym. Sci., 2005, 96, 1999–2004.
37. S. W. Ha, A. E. Tonelli and S. M. Hudson, Biomacromolecules, 2005, 6, 1722–1731.
38. K. Abdi, Sh. Nafisi, F. Manouchehri, M. Bonsaii and A. Khalaj, J. Photochem. Photobiol., B, 2012, 107, 20–26.
39. S. Bakkialakshmi and D. Chandrakala, Spectrochim. Acta, Part A, 2012, 88, 2–9.
40. C. J. T. Hoes, J. Grootoonk, R. Duncan, I. C. Hume, M. Bhakoo, J. M. W. Bouma and J. Feijen, J. Controlled Release, 1993, 23, 37–53.
41. L. Dong, L. Yan, W. G. Hou and S. J. Liu, J. Solid State Chem., 2010, 183, 1811–1816.
42. S. N. Ostad and P. R. Gard, J. Pharm. Pharmacol., 2000, 52, 779–784.
43. F. C. R. Lessa, A. M. F. Aranha, I. Nogueira, E. M. Giro, J. Hebling and C. A. Costa, J. Appl. Oral Sci., 2010, 18, 50–58.
44. G. J. Christensen, J. Am. Dent. Assoc., JADA, 2002, 133, 229–231.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra16253j

---