Biodegradable, thermoresponsive PNIPAM-based hydrogel scaffolds for the sustained release of levofloxacin

Abstract

A series of novel thermoresponsive biodegradable hydrogels (TBHs) was prepared from N-isopropyl acrylamide (NIPAM) and two biodegradable crosslinkers, poly(ε-caprolactone) dimethacrylate (PCLDMA) and bisacryloylcystamine (BACy). The morphology, thermal behavior, swelling/deswelling kinetics, compression properties, in vitro drug delivery and biodegradation were investigated. The results indicated that the properties of the TBHs strongly depended on temperature and the feeding molar ratio of the PCLDMA to BACy components. Levofloxacin (LVF)-loaded hydrogels were prepared to explore their stimuli-responsive release process. The cumulative release profile of LVF-loaded TBHs exhibited a thermo-induced slow sustained drug release and a reduction-induced fast release. At a physiological pH, TBHs could be biodegraded slowly in glutathione (GSH) at 37 °C. Due to their homogeneous pore diameter, highly interconnected architecture, degradable chemistry and thermoresponsive properties, the TBHs developed herein are highly attractive with respect to tissue engineering scaffold applications.

---

Introduction

Smart or intelligent hydrogels, which exhibit a phase transition change in response to external stimuli, such as temperature, pH, light and ionic strength, are extensively used as soft tissue engineering scaffolds and drug delivery systems due to their excellent biocompatibility and high water content (i.e., ca. 90% w/v), closely resembling living tissues, having a soft and rubbery consistency, as well as flexible methods of synthesis and constituents.^1–10 In addition, hydrogels can protect a drug from hostile environments and control drug release when undergoing reversible volume phase transitions in response to environmental stimuli.¹¹ Besides, hydrogels can be used as scaffolds for prompting cell growth and proliferation.¹²

Thermosensitive hydrogels are of the most interest in drug delivery, tissue engineering and cell encapsulation due to their ability to respond with noticeable property changes to temperature stimuli in the presence of water or physiological fluids.¹³ Among this class of hydrogels, thermoresponsive poly(N-isopropylacrylamide) (PNIPAM) is one of the most widely investigated polymers due to its unique volume phase transition from a hydrated, expanded state to a collapsed state at its lower critical solution temperature (LCST) in water or physiological fluid around 32 °C.^13–16 The PNIPAM-based hydrogels undergo an abrupt, reversible swelling–deswelling transition below and above the LCST. In the process of drug delivery, PNIPAM hydrogels swell and load drugs when the environmental temperature is below the LCST, rearranging into a compact structure and releasing the loaded drugs when heated to a temperature above the LCST.^17–19 Furthermore, a high degree of controlled drug release can be easily achieved owing to the hydrogels being composed of hydrophilic and hydrophobic constituents and by adjusting the hydrophilic and hydrophobic feeding molar ratio. Several investigations, including our own work, have shown that the phase behavior and mechanical properties of PNIPAM hydrogels can be improved by the addition of more hydrophobic or hydrophilic monomers for desired drug delivery and tissue engineering scaffolds.^20–24

Despite their significant potential in biomedical applications, PNIPAM hydrogels have not achieved clinical success due to their hard degradation, which could lead to incomplete drug release.^25–27 In order to resolve this problem, thermoresponsive and biodegradable hydrogels (TBHs) have blossomed during the past decades with respect to sustained drug delivery and tissue engineering scaffolds.^16,28–32 Various types of strategies for the preparation of biodegradable PNIPAM-based hydrogels introduced different biodegradable cross-linkers or natural polymers, such as proteins, poly(amino acids), polysaccharides, polyurethanes and poly(α-esters).^32–36 Yang et al. developed a type of biodegradable, thermosensitive hydrogel (PPCN) with intrinsic antioxidant properties for delivery of therapeutics, based on sequential polycondensation and radical polymerization of citric acid, poly(ethylene glycol), and PNIPAM.³² PPCN hydrogel exhibited an intrinsic antioxidant feature to scavenge free radicals, chelate metal ions and inhibit peroxidation. Ultimately, the PPCN gels could be resorbed over time when administered as subcutaneous injections in rats. Also, disulfides provide an excellent choice for degradable crosslinkers. A number of disulfide bonds introduced into the hydrogels can control the biodegradation rate of the slow cleavage of the disulfide bond.^37–40 The main reducing agent inside cells is glutathione (GSH), which is present at mM concentrations in the intracellular fluid. Conversely, the extracellular GSH concentration is only at μM concentrations; therefore, this differential provides a unique and selective trigger to promote intracellular degradation. Phillips incorporated disulfide bonds into PNIPAM to obtain dual-responsive materials.³⁸ These polymers showed selective degradation in the presence of intracellular and extracellular concentrations of glutathione.

The aim of the current research was to increase temperature-responsive degradation of hydrogels, enhance their mechanical strength, or improve their responsiveness. However, it is hard to induce these properties simultaneously. To improve the PNIPAM-based hydrogel biodegradation, mechanical properties and tunable controlled release of LVF, a series of PNIPAM-based TBHs was synthesized using temperature-sensitive PNIPAM as the main body and biodegradable PCL and BACy as a cross-linking agent. PCL is a hydrophobic moiety so can undergo hydrolytic degradation, whereas BACy is a hydrophilic moiety and can undergo reducible degradation. Polyester PCL, which is approved by the Food and Drug Administration (FDA) and is biodegradable in humans as a contraceptive implant, has an excellent slow-release performance and its degradation products lower the pH value and inhibit bacterial inflammation.⁴¹ What is more, PCL can improve the mechanical properties of PNIPAM hydrogels. The disulfide linkage of BACy can be effectively cleaved into two thiol groups in the presence of reducing agents such as 1,4-dithiothreitol (DTT) or glutathione (GSH).^33,42 There is a relatively high redox potential due to a very low concentration of GSH in body fluids (e.g. blood) and in extracellular matrices, which could slowly degrade the TBHs.^37,43 Finally, degradation of the hydrogel matrix not only circumvents removal of the empty device, but can also be used to modulate the release of encapsulated drugs for a long period of time.⁴⁴ In a word, in this study, the synthetic TBHs were temperature-responsive, had favorable mechanical strength and were degradable in the physiological environment. Moreover, degradation would greatly improve the release of loaded drug.

Experimental

Chemicals

NIPAM was purchased from TCI (Shanghai, Chuo-ku, Tokyo, Japan), and recrystallized twice at 50 °C in n-hexane before use. PCL was obtained from Mitsubishi Chemical Corp. (Chiyoda-ku, Tokyo, Japan; weight-average molecular weight (M_w = 2000 Da)), and was dried under vacuum for 12 h at 80 °C. The following reagents were all purchased from Aladdin-Agent (Shanghai, China). Dichloromethane (CH₂Cl₂) was stirred with CaH₂ at room temperature overnight and was distilled before use. Azobisisobutyronitrile (AIBN) was recrystallized twice at 75 °C in anhydrous ethyl alcohol. Triethylamine (TEA) was dehydrated by distillation over the drying agent, CaH₂. Molecular sieves were used to treat 1,4-dioxane. Cystamine dihydrochloride, 4-dimethylaminopyridine (DMAP), tetrahydrofuran (THF), acryloyl chloride, methacryloyl chloride, hexane, ethyl acetate, and glutathione (GSH) were used as received.

Preparation of the bisacryloylcystamine (BACy)

N,N′-Bis-acrylcystamine (M_w = 260.38 Da) was synthesized following the method reported previously.^42,45 Briefly, one equivalent of cystamine dihydrochloride in 80 mL THF was added with 4.2 equivalents of TEA and the resulting solution was stirred for 24 hours at room temperature. Next, 2.2 equivalents of acryloyl chloride in 20 mL THF was added dropwise into the above solution at a rate of 1 mL min⁻¹. The resulting mixture was allowed to react at 40 °C under nitrogen atmosphere for 24 hours. After reaction, the formed solid was filtered off, and the remaining TEA and THF were evaporated using a rotary evaporator under reduced pressure. The raw product was dissolved with chloroform, and separated from the water layer. The extract was washed with 0.1 M HCl and saturated sodium chloride aqueous solution. The collected organic phases were dried over Na₂SO₄. After removal of the solvent under vacuum, the residue was separated by column chromatography with a gradient solution of hexane and ethyl acetate (30/70 to 0/100; v/v). The obtained BACy was further re-crystallized in hexane/ethyl acetate (30/70; v/v), and dried as a white solid (yield: 50%). IR (KBr, cm⁻¹): 1620 cm⁻¹ (s, C=C), 3248 and 1554 cm⁻¹ (s, N–H), 1653 cm⁻¹ (s, C=O). ¹H NMR (400 MHz, CDCl₃): δ 5.7–6.3 (s, 3H, CH₂=CH), 2.9–3.7 (s, 4H, CH₂–CH₂), 6.8 (s, 1H, N–H).

Preparation of poly(ε-caprolactone dimethacrylate) (PCLDMA)

PCLDMA was synthesized following the method reported previously.⁴⁶ Briefly, to the mixture of PCL (10 g, M_w = 2000 Da, 5 mmol) and TEA (2.2 mL, 15 mmol) in anhydrous dichloromethane (200 mL), methacryloyl chloride (1.5 mL, 15 mmol) was added dropwise at 0 °C. The reaction mixture was stirred at room temperature for 18 h. The formed solid was filtered off, and dichloromethane phase was washed with 0.1 N HCl, saturated NaHCO₃ and saturated sodium chloride aqueous solution, respectively. The organic phase was dried over anhydrous MgSO₄, filtered and evaporated under vacuum conditions to produce the crude PCLDMA, which was a slightly yellow powder. Purified PCLDMA was obtained by precipitation with cold hexane to afford the product as a white waxy solid, 7.5 g (yield 75%). IR (KBr, cm⁻¹): 1640 cm⁻¹ (m, C=C), 1735 cm⁻¹ (s, C=O), 1163 cm⁻¹ (m, C–O). ¹H NMR (400 MHz, CDCl₃): δ 5.6–6.1 (m, 2H, CH₂=C), 0.9 (s, 3H, CH₃), 1.3–4.1 (vs, CH₂). The degree of acylation was determined with ¹H NMR spectroscopy, examining the ratio between the signal under the peak at δ 6.1 (s, 2H, olefinic, cis) and 3.9 (m, 4H, –OCH₂CH₂OCH₂CH₂O–).

Preparation of TBHs

To prepare the thermoresponsive and biodegradable PNIPAM-based gels, a mixture of NIPAM, PCLDMA, BACy, AIBN, and 1,4-dioxane was put into a glass tube. After the tube was degassed and sealed, the polymerization was conducted at 60 °C for 36 h. To remove the unreacted monomer, the obtained gels were immersed in deionized water for 5 days at room temperature, and fresh water was replaced every day. Then, samples were lyophilized for 48 h. The contents of each component are summarized in Table 1.

Table 1 Feed composition and equilibrium swelling ratios (ESR) of TBHs

Sample code	NIPAM (g)	PCLDMA (g)	BACy (g)	Yield^a (%)	Molar-ratio (NIPAM/PCLDMA/BACy)^b	ESR (g g^−1)
a The composition (%) was calculated as follows: (W/W0) × 100%, where W is the weight of the dry gel and W0 is the sum of the weights of PCLDMA, BACy and NIPAM.b Mw of BACy was approximately 277 Da; Mw of PCLDMA was approximately 2100 Da and Mw of NIPAM was 113 Da.
PCLDMA7	0.4	0.0735	0	98	700 : 7 : 0	6.57
PCLDMA5–BACy2	0.4	0.0525	0.0027	96	700 : 5 : 2	7.45
PCLDMA4–BACy3	0.4	0.0420	0.0042	96	700 : 4 : 3	8.92
PCLDMA3–BACy4	0.4	0.0315	0.0055	95	700 : 3 : 4	10.75
PCLDMA1–BACy6	0.4	0.0105	0.0084	95	700 : 1 : 6	14.78
BACy7	0.4	0	0.0097	95	700 : 0 : 7	20.68

---

Structure measurement

¹H NMR spectra were obtained on a Bruker 400 MHz spectrometer with deuterated chloroform (CDCl₃) as the solvent, and chemical shifts were obtained relative to tetramethylsilane. Fourier transform infrared (FT-IR) measurements were carried out using a Nicolet 5100 spectrometer with KBr disks in the scanning range from 500 to 4000 cm⁻¹ with a resolution of 2 cm⁻¹.

Interior morphology

The swollen hydrogel samples, after reaching equilibrium swelling ratios in water at room temperature, were quickly frozen in liquid nitrogen and then lyophilized in a freeze drier under vacuum at −55 °C for 2 days. The lyophilized samples were then fractured carefully in liquid nitrogen, and the interior morphology of the hydrogel samples was studied with a scanning electron microscope (SEM, Bruker, S-3400N, and Hitachi, S-4800). Before SEM observation, the hydrogel specimens were coated with gold for 7 min.

LCST behavior

The LCST behavior of the hydrogel was determined by Differential Scanning Calorimetry (DSC Q200, TA). Each sample was immersed in distilled water at 10 °C and allowed to reach the equilibrium state. Five or 7 mg of the swollen sample was used for each analysis. The samples were placed in a hermetic sample pan and sealed. The samples were scanned from 10 °C to 40 °C at a rate of 2 °C min⁻¹ under nitrogen.

Equilibrium-swelling ratio (ESR)

The pre-weighed dried hydrogels (W_d) were immersed in 10 mL deionized water at 20 °C until swelling equilibrium was attained. Each gel was then removed from the water bath, tapped with filter paper to remove excess surface water, and weighed to obtain the swelling equilibrium weight (W_e). The equilibrium-swelling ratio (ESR, g g⁻¹) in H₂O was calculated from the following equation:where W_e is the weight of the gel at a certain temperature and W_d is the dry weight of the gel. All experiments were performed in triplicate and an average value of three measurements was recorded.

Swelling and deswelling kinetics by gravimetry

The swelling and deswelling kinetics were determined by measurement of the temporal weight change of the gels. For the swelling kinetic studies, the lyophilized, disk-shaped gels were first equilibrated in deionized water at a predetermined temperature (20 °C). The gels were weighed at each given time. The swelling kinetics were calculated as follows:where SR is the swelling ratio of the gels at a certain time, W_t is the weight of the swelling gel at a certain time and W_d is the weight of the dry gel. All experiments were performed in triplicate and an average value of three measurements was recorded.

The kinetics of the deswelling behavior of the hydrogels was measured at 50 °C. Before measuring the deswelling kinetics, the hydrogel reached swelling equilibrium in deionized water at 20 °C in advance (ESR). After wiping off surface water with filter paper, the weight of the gel was recorded (in triplicate) during the course of deswelling at each regular time interval. The deswelling ratio (DSR) is defined as:

where W_t is the weight of the deswelling gel at a certain time; W_d is the weight of the dry gel, and W_e is the weight of the fully swollen gel.

Compression properties

Compression testing of the specimens was performed using a Texture Analyzer TA-XT Plus (Stable Micro System). Unconstrained uniaxial compression was applied, while compressive force and displacement were recorded after the probe tip contacted the sample. Confined compression tests were performed on the cylindrical shaped hydrogels at room conditions, with a 10 N load cell. The samples were soaked in PBS and kept at 25 °C for at least 24 h before the test. A plate of 12.7 mm in diameter was used and samples were tested at 0.15 mm min⁻¹. The compressive modulus was defined as the linear regression between 0 and 30% strain.⁴⁷

Drug loading and in vitro release with or without GSH

Levofloxacin (LVF) was chosen as the model drug in the drug-release experiment. The lyophilized hydrogel samples were loaded with LVF by soaking in 10 mL LVF solution, with a concentration of 5 mg mL⁻¹, for 3 days at 4 °C. LVF was dissolved in a PBS solution (0.1 M, pH 7.4). The soaking process was protected from light. Thus, LVF was loaded into hydrogels up to equilibrium by swelling in the LVF-containing solution, and the LVF-loaded hydrogel samples were used for in vitro release kinetic study. The loaded samples were rinsed with PBS solution and dried with absorbent paper in order to remove excess PBS solution. The LVF content in the residual solution was determined by UV-vis spectrometry (Shimadzu, UV-2550) at 285 nm against the established calibration curve.

In vitro release with or without reductant was carried out at 37 °C to investigate the effect of special structures of TBHs on drug release profiles. LVF release experiments were conducted by immersing the above LVF-loaded hydrogel samples in a closed glass bottle filled with 20 mL PBS (0.1 M, pH 7.4) at 37 °C (above LCST) with or without GSH (10 mM). At a predetermined period in the in vitro release experiment, 10 mL aliquots of the buffer medium were removed from the glass tube and analyzed by UV-vis absorption. Then, 10 mL fresh preheated buffer solution with or without GSH (10 mM) was added back to the glass tube to maintain the same total solution volume. All release studies were carried out in triplicate and an average value of three measurements was recorded. The results were presented in terms of cumulative release as a function of time:

where V₀ is the amount of release media taken out each time (10 mL), C_i is the concentration of LVF released from the hydrogels at a displacement time of i, and M_∞ is the estimated amount of LVF loaded in the hydrogels and was calculated from the weight difference between the initial LVF solution concentration (before loading) and the remaining LVF solution concentration after loading (M_∞ (mg) = 5 − m_{residual solution}).

In vitro biodegradability

Partially degradable TBHs were added to PBS (0.1 M, pH 7.4) solution and 3 mM GSH in PBS (0.1 M, pH 7.4) solution overnight at 4 °C; the hydrogels were then added to 4 mL preheated PBS (pH 7.4) solution or 3 mM GSH in PBS (pH 7.4) solution at 37 °C in a 20 mL vial. The fresh PBS solution with or without GSH (3 mM) was replaced every day. The samples were extracted at scheduled intervals and then freeze-dried. The hydrolytic and reducible degradability of the TBHs was determined by observing the morphology before and after degradation using scanning electron microscopy (SEM, Bruker, S-3400N, and Hitachi, S-4800).

Results and discussion

Synthesis and characterization of TBHs

The chemical structures of NIPAM, PCLDMA, BACy, and PNIPAM-co-PCLDMA-co-BACy hydrogel and a schematic drawing of a hydrogel are shown in Scheme 1. TBHs with thermo-responsive monomer NIPAM and degradable cross-linkers BACy and PCLDMA possessed a similar transparency to the hydrogels prepared by a conventional method with a bis-type crosslinking agent, such as methylene bis-acrylamide. The hydrogel, which was prepared in a glass test tube, was removed from it conveniently.

Scheme 1 Synthesis route and schematic chemical structure of NIPAM, PCLDMA, BACy, and TBH (A), schematic of degradation of TBH (B), and drug loading and release of thermo- and reduction-responsive morphology transformation of TBH (C–E).

The chemical structure of the resulting TBHs was characterized by FTIR. As shown in Fig. 1, the spectrum of PCLDMA3–BACy4 showed characteristic peaks at 1735 cm⁻¹, 1654 cm⁻¹ and 1654 cm⁻¹, which are assigned to the carbonyl groups (C=O) from PCLDMA, BACy and NIPAM, respectively. Amides II of BACy and NIPAM appeared at 1542 cm⁻¹. The ν_N–H in the BACy and NIPAM units existed at 3251 and 3288 cm⁻¹, while for PCLDMA3–BACy4, the ν_N–H broadened at 3150–3650 cm⁻¹ because of the formation of intramolecular hydrogen bonds. What is more, NIPAM, BACy and PCLDMA had active sites (double bonds). However, the disappearance of the double bonds indicated that copolymerization was complete and the TBHs formed a crosslinked network structure, as shown in the FTIR spectra.

Fig. 1 FT-IR spectra of (A) NIPAM, (B) BACy, (C) PCLDMA, and (D) PCLDMA3–BACy4 hydrogel.

Interior morphology

The interior morphologies of the swollen TBHs were observed by SEM. Completely swollen TBHs were frozen in liquid nitrogen and then freeze-dried, which may maintain the original nature of the TBHs. The dramatic differences in morphology observed between TBHs were presumably of an intrinsic nature, since the fixation procedures were identical among all TBH specimens. The SEM images revealed a porous morphology with a homogeneous pore diameter, which is shown in detail in Fig. 2.

Fig. 2 SEM micrographs of swollen-dried TBHs.

The most distinctive feature in Fig. 2 is that a homogeneous smooth morphology and pores were observed, and the pore diameter of the TBHs increased as the PCLDMA content decreased. Due to the excellent miscibility of the three polymeric components in 1,4-dioxane, no obvious phase separation was observed in the TBH matrix. The interior microstructures of these TBHs indicated a network without concomitant phase separation because the gelation was completed within a homogeneous region. What is more, it was obvious that the pore size of PCLDMA7 gel (only about 1 μm) was much smaller than that of the other samples. This was because the PCL is a semicrystalline hydrophobic polymer,⁴⁸ and could disrupt the close-packed chain, producing a small hole in the semicrystalline area of the hydrogels, which was essential in the swelling and deswelling kinetics, drug loading and release of TBHs.

Based on the morphology determined by SEM, TBHs would be expected to have a different mechanical strength, swelling ratio and drug release properties compared to conventional PNIPAM hydrogels. The mechanical properties are discussed later and they indeed imply an increase in the compression moduli as the PCLDMA content increases.

LCST determination

The LCST is the key parameter of temperature-responsive polymers, e.g. PNIPAM. Below the LCST, PNIPAM is hydrophilic and in a swollen state in water. However, it becomes hydrophobic, shrinks, and displays an abrupt volume decrease when the environmental temperature is higher than the LCST.⁴⁹

The LCSTs of TBHs were determined by DSC and defined as the peak temperature of the endotherms, as shown in Fig. 3. The LCST of BACy7 in this paper was 32.61 °C, which is the typical LCST of PNIPAM polymers.^50,51 LCSTs of other TBHs decreased with increasing content of PCLDMA. The LCSTs for PCLDMA1–BACy6, PCLDMA4–BACy3 and PCLDMA7 were 31.03, 30.80 and 30.68 °C, respectively. In general, LCSTs of temperature-responsive hydrogels are usually governed by the relative hydrophobicity of the bulk.⁴⁶ As a result of the copolymerization of NIPAM with PCLDMA and BACy, the hydrophobic/hydrophilic balance was changed due to the incorporation of CL units and BACy into the TBHs backbone.⁵² The presence of the hydrophobic CL fragments increases the hydrophobicity of PCLDMA7, PCLDMA4–BACy3 and PCLDMA1–BACy6. It has been suggested that the PCLDMA4–BACy3 hydrogel has a more hydrophobic character than PCLDMA1–BACy6, since it was prepared with a higher molar percent of PCLDMA.

Fig. 3 DSC curves of (A) PCLDMA7, (B) PCLDMA4–BACy3, (C) PCLDMA1–BACy6 and (D) BACy7 hydrogels.

It is well known that the phase transition of PNIPAM hydrogels result from a balance between hydrophilicity and hydrophobicity in the polymeric backbone. When a hydrophilic moiety was copolymerized into the PNIPAM hydrogel network, its LCST shifted to a higher temperature. On the contrary, if a hydrophobic moiety was copolymerized into the polymeric chain, its LCST shifted to a lower temperature.⁴⁹ The phase-transition behavior was the result of the balance between hydrophobic and hydrophilic groups in the polymer chains. As the temperature increased, the hydrophobic PCLDMA enhanced the hydrophobic association interactions, which contributed to a decrease in the LCSTs.

Equilibrium swelling ratio

The temperature-dependent swelling ratio is one of the most crucial parameters to evaluate the temperature-sensitive properties of hydrogels.^2,3 The swelling ratios of the novel TBHs produced in our work at different temperatures from 20 to 50 °C are shown in Fig. 4. As the temperature increased, the swelling ratios of all the TBHs decreased, which is in accordance with thermoresponsive behavior described in previous research.^46,52 The data showed that these TBHs absorbed water and became swollen at temperatures below 30 °C. When the temperature increased, the TBHs lost water and gradually shrank in volume. The temperatures at which all the TBHs exhibited a sharp change were between 30 and 35 °C, which indicated that the LCSTs of the TBHs were within this temperature range, in accordance with the DSC results. Meanwhile, the swelling ratio decreased as the PCLDMA content increased, and BACy7 had the largest swelling ratio, while PCLDMA7 had the smallest. This observation can be explained by the fact that the introduction of PCL chains enhanced the hydrophobicity of the hydrogels and weakened their expansion. As is known, there were hydrophilic groups, –CONH–, and hydrophobic groups, –CH(CH₃)₂, in the monomer NIPAM, corresponding to the hydrophilic and hydrophobic regions, respectively.⁵³ In this study, an increase in the PCLDMA content increased the hydrophobicity of the hydrogel network. What is more, as a hydrophobic polymer, the entanglement of the PCL chains hindered further swelling of the gel network, causing TBHs containing PCLDMA to have a smaller swelling ratio and a less obvious phase transition. In addition, the pore sizes of TBHs incorporating PCLDMA were smaller and accommodated less water.

Fig. 4 ESR of TBHs determined as a function of temperature range from 20 to 50 °C after swelling equilibrium for 24 h at each temperature interval in deionized water.

Swelling and deswelling behaviors of TBHs

Hydrogels with high water swelling and moderate deswelling properties could directly influence the drug loading and release performance.⁵⁴ Dynamic water swelling processes were studied gravimetrically by monitoring the water imbibed into the TBHs at 20 °C, as shown in Fig. 5. As shown, all the TBHs took over 9 h to reach their equilibrium states due to the high hydrophilicity of the hydrogels at this temperature. It was found that the swelling ratio (SR, g g⁻¹) of all the TBHs increased quickly during the initial 100 min, then gradually increased, finally leveling off near 500–600 min. The swelling ratios for the TBHs shown in Fig. 5 indicated that the SRs of the gels were increased as the ratio of hydrophobic PCLDMA to hydrophilic BACy decreased, that is, PCLDMA7 (6.6 g g⁻¹) < PCLDMA5–BACy2 (7.5 g g⁻¹) < PCLDMA4–BACy3 (8.9 g g⁻¹) < PCLDMA3–BACy4 (10.8 g g⁻¹) < PCLDMA1–BACy6 (14.8 g g⁻¹) < BACy7 (20.7 g g⁻¹). The results were consistent with previous reports indicating that a hydrophobic moiety could lower the swelling ratio and rate of the gels.^55,56 PCL is a type of aliphatic polyester that has a strong hydrophobic effect, and curls into a close-packed state, leading to small holes in the TBHs. As a result, these TBHs absorbed less water than BACy7 gels, because small holes in the gels inhibited the interchange of water molecules between the hydrogel matrix and the external liquid. Despite having the same cross-linking density as the BACy7 hydrogel, which was a hydrophilic moiety like traditional hydrogels, they had a much higher swelling ratio and rate than that of other TBHs. However, the disadvantage of the BACy7 hydrogel was its poor mechanical strength, resulting in it being difficult to handle. The mechanical properties of the TBHs gradually improved as the content of PCLDMA increased, which will be discussed later in the text.

Fig. 5 Swelling kinetics curves (left) of TBHs (20 °C) and deswelling kinetics curves (right) of TBHs (from 20 to 50 °C) in deionized water.

Fig. 5 also shows the shrinking kinetics of the TBHs after transferring an equilibrated swollen specimen at 20 °C (below its LCST, two days) to hot distilled water at 50 °C (above its LCST). All TBHs shrank rapidly within ten minutes, the highest shrinkage being that of BACy7, to 58 wt%, while the least shrinkage was exhibited by PCLDMA5–BACy2, which shrank to 22 wt%. Therefore, BACy7 had the largest shrinkage ratio and rate. When studying the deswelling properties of TBHs, it was interesting to find that the TBHs containing PCLDMA shrank more slowly than the BACy7 gel, but the shrinkage rate was different due to the various ratios of hydrophobic PCLDMA to hydrophilic BACy, that is, PCLDMA5–BACy2 < PCLDMA3–BACy4 < PCLDMA1–BACy6. As a whole, the larger the ratio of hydrophobic PCLDMA to hydrophilic BACy, the slower the shrinkage. However, there was an exception in that the shrinkage rate of PCLDMA3–BACy4 was a little slower than that of PCLDMA4–BACy3. The reason for this may be that a synergistic interaction existed between the hydrophobic and hydrophilic moieties. Amazingly, during the deswelling kinetics test, it was noticed that the pure hydrophilic BACy crosslinked gel did not show slow deswelling behavior, while the slow-release performance of the fully hydrophobic PCLDMA crosslinked gel was not particularly outstanding; in fact, it was worse than those TBHs containing both PCLDMA and BACy, which is consistent with the literature.^57,58 For example, after 100 min in 50 °C hot water, PCLDMA5–BACy2, PCLDMA4–BACy3 and PCLDMA3–BACy4 shrank, losing about 32 wt%, 53 wt% and 51 wt% water, respectively. In contrast, about 71 wt% water was freed from BACy7 and 59 wt% water was freed from PCLDMA7 within the same time frame.

It is well known that PNIPAM-based hydrogels swell due to the hydrogen bonds between the amide group and water when the external temperature is below the LCST. When the temperature was elevated above the LCST, the hydrophobic interactions from the isopropyl group became dominant and the hydrogel chains suddenly contracted and aggregated, which resulted in shrinkage of the hydrogel volume. Moreover, the diffusion rate of the free water through the hydrogel networks determined the deswelling rate of the hydrogels. When BACy7 gel was immersed in hot water (above the LCST), it shrank and the volume reduced abruptly, which repelled most of the remaining water. However, TBHs containing rigid PCL prevented the hydrogels from collapsing to a certain extent and reserved relatively more water than BACy7 gels. With more PCL chains, the gels became more rigid and less sensitive to changes in temperature. As a result, the deswelling rate was slow and the water retention rate was relatively high. This feature is beneficial with respect to long-term controlled drug release.

Compression modulus

The mechanical properties of hydrogels are important design parameters in tissue engineering, as the gel must create and maintain a space for tissue development. In addition, the adhesion and gene expression of cells are closely related to the mechanical properties of the polymer scaffold.¹² The mechanical properties of hydrogels mainly depend on the original rigidity of polymer chains, the types of cross-linking molecules and the cross-linking density, and swelling as a function of the hydrophobic/hydrophilic balance.⁵⁹

The mechanical properties of TBHs were measured with a Texture Analyzer TA-XT Plus (Stable Micro System, SMS, UK) testing machine when hydrogels reached equilibrium swelling ratios at 25 °C in PBS (0.1 M, pH 7.4), and the results are shown in Fig. 6. The data confirm our original hypothesis that TBHs containing both PCLDMA and BACy would have better mechanical properties than the corresponding PCLDMA or BACy ones, being neither too soft nor too rigid.

Fig. 6 Compressive stress–strain curves of TBHs, (A) PCLDMA7, (B) PCLDMA5–BACy2, (C) PCLDMA4–BACy3, (D) PCLDMA1–BACy6 and (E) BACy7 at 25 °C.

As shown in Fig. 6, all of the TBHs had good mechanical properties and high elastic moduli (E) (75–470 kPa).⁶⁰ The elastic modulus of the TBHs increased with increasing molar ratio of PCLDMA to BACy, i.e., PCLDMA7 > CLDMA5–BACy2 > PCLDMA4–BACy3 > PCLDMA1–BACy6 > BACy7. The compressive moduli of PCLDMA5–BACy2 (184.8 kPa) and PCLDMA4–BACy3 (142.0 kPa) was about 2.45 and 1.88 times higher than that of BACy7 hydrogel (75.4 kPa), respectively. The improvement in the compression properties of TBHs was due to an increasing ratio of PCLDMA to BACy. And the hydrophobic moiety, PCL, could increase the compressive strength of TBHs. This may be attributed to increased hydrophobic interaction resulting in physical entanglement of thermally-responsive polymer chains that tended to form compact polymer clusters and increase compression moduli.⁶¹ The results were in accordance with the properties of the hydrogels observed during handling. During handling, it was clear that a larger ratio of PCLDMA to BACy improved the hydrogels' structural integrity and mechanical resistance. The hydrogels containing more PCLDMA kept their shape more effectively after being removed from the mold.⁴⁷

More interestingly, as shown in Fig. 6, the kinetics curves showed a pronounced deviation from linearity. The E values of all of the samples initially decreased and then increased. It can be seen that the more the PCLDMA in the TBHs, the more the deviation. It is possible that part of the hydration was destroyed initially, and then the frizzy PCL chains were stretched with increasing pressure, which could have led to an initial decrease in the compression moduli, followed by a later gradual increase. However, the deviation of the BACy gel was much smaller in this procedure.²⁰

The loading–unloading curves are shown in Fig. 6. The area of hysteresis reflects the ability of a hydrogel to dissipate energy per unit volume.²⁶ For a single-network hydrogel, the dissipating energy is mainly dependent on physical crosslinks. This energy can be evaluated using the following equation:

In the equation, V is the volume fraction of the hydrogel in the process zone, W_D is the mechanical energy dissipated per unit volume of a hydrogel element in the process zone in the undeformed state and is given by the areas of the hysteresis loops in stress–strain curves produced by deforming and undeforming the hydrogel element; h is the width of the process zone around the notch. Fig. 6 demonstrated qualitatively that the dissipating energy reached a maximum value at PCLDMA7 gel.⁶² while the dissipating energy of BACy7 gel showed the minimum value.

Drug loading and releasing

Levofloxacin was used in this study due to its hydrophilic features and its UV-vis absorbability. LVF was chosen because its molar mass is small (M_w = 370 Da) and its concentration can be easily detected by UV/vis spectroscopy without interference from degraded polymer materials. In order to evaluate the ability of the TBHs to effectively deliver LVF, in vitro LVF release from PCLDMA7, PCLDMA5–BACy2, PCLDMA4–BACy3, PCLDMA3–BACy4 and BACY7 gels at 37 °C with or without GSH was studied (Scheme 1). Firstly, LVF was loaded into hydrogels by incubating the freeze-dried hydrogels in an LVF (50 mg mL⁻¹) PBS buffer solution (0.1 M, pH 7.4) at 25 °C. The loading content of LVF to hydrogels was 9.1, 10.5, 11.7, 15.1 and 21.1 mg (LVF) g⁻¹ (dry gel) from PCLDMA7, PCLDMA5–BACy2, PCLDMA4–BACy3, PCLDMA3–BACy4 and BACy7, respectively. Then, after drug encapsulation, the drug release experiment was carried out at 37 °C above the LCSTs. Fig. 7 shows the release profiles of LVF from the TBH hydrogels as a function of time. TBHs with different ratios of PCLDMA to BACy demonstrated varying drug release profiles. As Fig. 7 shows, the release rate of the LVF increased with a decreasing molar ratio of PCLDMA to BACy, but all of the TBHs were weak in the release of LVF. The results indicated that LVF was released quickly at first; however, after the initial burst period, the release rate slowed down and gradually reached a plateau. However, all the TBHs exhibited a slow rate of release of LVF. The cumulative amount of LVF released in the first 9 h of the burst release was 7% for PCLDMA5–BACy2, 12% for PCLDMA4–BACy3 and 12% for PCLDMA3–BACy4. Although all the TBHs exhibited very similar release trends, their release rates and extents are different. The cumulative LVF release during almost a month was BACy7 (28%) > PCLDMA7 (20%) > PCLDMA3–BACy4 (17%) > PCLDMA5–BACy2 (15%). These results were in accordance with the internal morphology, ESR and swelling kinetics of the TBHs. The hydrophobic rigid structure of PCLDMA had the smallest TBH pore size, resulting in the minimum ESR and drug-loading capacity. Although the loaded amount of LVF in the PCLDMA7 gel was the lowest, the slowest release rate was exhibited by those TBHs simultaneously containing hydrophobic PCLDMA and hydrophilic BACy, rather than the PCLDMA7 gel, which identified with the deswelling curves of TBHs.

Fig. 7 In vitro cumulative release of levofloxacin from TBHs in PBS (pH 7.4) solution.

Fig. 8 exhibits the drug release behavior as a function of time for TBHs containing the disulfide cross-linker BACy, such as PCLDMA5–BACy2 and PCLDMA4–BACy3, in GSH (10 mM) PBS buffer solution (0.1 M, pH 7.4). The two hydrogels presented a similar drug release behavior in GSH containing PBS buffer solution to that of pure PBS buffer solution. However, the drug release rate increased much more quickly, which provided evidence for the breakage of the cross-linker in the presence of the reducing agent, GSH. As shown in Fig. 8, the LVF-loaded hydrogels, (PCLDMA4–BACy3 and PCLDMA5–BACy2), showed that complete release occurred after 120 h in 10 mM GSH PBS (0.1 M, pH 7.4) buffer solution. However, in pure PBS (0.1 M, pH 7.4) buffer solution, only 21% of LVF was released from PCLDMA4–BACy3 and the vast majority of LVF remained in the hydrogels and was difficult to release even though the releasing time was extended to one month. The release rate of PCLDMA4–BACy3 was faster than that of PCLDMA5–BACy2 due to it containing more BACy, which involved breaking more disulfide bonds. These results suggested that the hydrogels containing disulfide agents decomposed into water-soluble unimers in the presence of a reducing agent, such as GSH, which accelerated the drug release.^63–65 The reducible degradation of TBHs could destroy the dense surface of gels and trigger the release of encapsulated drug molecules, as well as facilitate the removal of empty vehicles. Furthermore, the release rate gradually increased with increasing BACy. As the content of the disulfide bond cross-linker, BACy, increased, more degradable cross-links could be broken, resulting in a higher release rate. After a burst release, the release rate underwent the process of initially slowing down and then releasing faster and faster, because the increase in osmotic pressure and the swelling pressure of the TBHs accelerated, squeezing LVF out as the TBHs released by degrading.⁶⁶

Fig. 8 In vitro cumulative release of levofloxacin from PCLDMA4–BACy3 and PCLDMA5–BACy2 in 0 and 10 mM GSH PBS (pH 7.4) solution, respectively.

Biodegradation in vitro

The controlled degradation of hydrogels is another crucial parameter in tissue engineering, whether the gels are derived from natural materials or are synthetically created. Typically, one desires to match the degradation rate of a scaffold to tissue development, and the time will be dependent on the tissue type to be engineered.⁶⁷ As shown in Scheme 1, the TBHs in this paper degraded in two ways: hydrolytic degradation (PCL is one of the most studied synthetic polyesters and has been proved to exhibit hydrolytic degradation in vitro and in vivo) and reducible degradation (BACy has the degradable disulfide cross-linker, which allows reducible biodegradation with thiol groups in the presence of GSH). Therefore, we discussed the degradation of TBHs in vitro at a physiological pH in the presence of GSH. It is known that DTT or GSH can cleave disulfide bonds efficiently. Many reports describe the usage of DTT or GSH as a water-soluble reducing agent to degrade disulfide-containing polymers into the corresponding thiols.^22,57,68,69 Surface and interior morphological observation by SEM of the hydrolytic and reducible degradation of TBHs was conducted at 37 °C in 3 mM GSH PBS solution (pH = 7.4, 0.1 M). Fig. 9 shows the degradation profile of PCLDMA7, PCLDMA1–BACy6, and BACy7 for 0, 10 and 60 days. Initially, the surface morphology of the hydrogel was smooth, uniform and neat. After 10 days of erosion, the pores increased in size by varying degrees. For example, the pore sizes of PCLDMA7 increased from 1 μm to 5–10 μm, those of BACy7 increased from 20 μm to 30–50 μm, while the largest gap appeared in PCLDMA1–BACy6 (>50 μm), much greater than the original 20 μm. Moreover, the edges were not as clear and smooth as in the original SEM morphology. After 60 days of erosion, the pores of the TBHs became larger. This was because cleavage of the disulfide bonds of BACy caused by the presence of GSH and the hydrolytic degradation of the ester bonds of PCL results in erosion of the hydrogel structure, speeding up degradation.

Fig. 9 SEM micrographs of degraded TBHs for 0 (A1–C1), 10 (A2–C2) and 60 (A3–C3) days.

Conclusion

A series of novel TBH scaffolds have been designed, synthesized and characterized. The TBHs were composed of PNIPAM as a thermoresponsive unit, PCLDMA as a hydrolytically degradable and hydrophobic unit, and BACy as a reducibly degrading and hydrophilic unit. Therefore, the merits of both thermoresponsive and biodegradable polymeric systems were combined in these novel TBH scaffolds. The successful syntheses of the hydrogels were confirmed by IR spectroscopy.

The morphology, thermal behavior and swelling/deswelling kinetics of TBHs could be engineered through changing the molar feeding ratio of PCLDMA to BACy. Measurement of the equilibrium swelling ratio revealed that the TBHs had a good thermoresponsive properties, with an LCST at around 32 °C producing a volumetric phase change. The phase transition became less sharp as the molar feeding ratio of PCLDMA to BACy increased. As the molar feeding ratio of PCLDMA to BACy increased, the pore sizes became smaller, the swelling ratio decreased, and the kinetics of swelling slowed down. Interestingly, it was found that the TBHs containing PCLDMA and BACy simultaneously showed slower deswelling kinetics than PCLDMA7 gel containing a hydrophobic cross-linker and BACy7 containing a hydrophilic one. Levofloxacin was used as a model drug in a simulated in vitro release, and the results indicated that all the TBHs showed sustained release in PBS buffer (0.1 M, pH = 7.4), less than 30% of the total amount of loading LVF releasing within one month. However, with a low concentration of the reductant, GSH, in the PBS solution, the LVF could be completely released in 120 h, and the release rate increased gradually during the degradation process. Furthermore, the degradation of TBHs incubated in GSH (3 mM) PBS solution showed that the hydrogels could be degraded in a reductive environment over two months. Meanwhile the TBHs could be quickly degraded in alkaline conditions at 25 °C. Because of the limitations of the experimental conditions, other properties, such as the cytotoxicity of the scaffold, cell adhesion and the cell loading related to the tissue engineering scaffold, were not obtained. Due to the good elasticity, fully degradable nature and thermo-responsive properties, the TBHs showed great potential for biomedical applications, including drug delivery and tissue engineering, and could be excreted after use.

Acknowledgements

The authors are grateful to acknowledge the financial supports from the Guangdong Natural Science Foundation, China (No. 2015A030313798), Zhujiang Science & Technology New-star Program of Guangzhou, China (No. 2013J2200016) and Guangdong Special Support Program-Youth Top-notch Talent (No. 2014TQ01C400).

References and notes

1. E. A. Appel, J. del Barrio, X. J. Loh and O. A. Scherman, Chem. Soc. Rev., 2012, 41, 6195–6214.
2. F. Liu and M. W. Urban, Prog. Polym. Sci., 2010, 35, 3–23.
3. E. S. Dragan, Chem. Eng. J., 2014, 243, 572–590.
4. V. A. Ganesh, A. Baji and S. Ramakrishna, RSC Adv., 2014, 4, 53352–53364.
5. R. Censi, P. Di Martino, T. Vermonden and W. E. Hennink, J. Controlled Release, 2012, 161, 680–692.
6. S. Van Vlierberghe, P. Dubruel and E. Schacht, Biomacromolecules, 2011, 12, 1387–1408.
7. M. A. Ward and T. K. Georgiou, Polymers, 2011, 3, 1215–1242.
8. M. C. Koetting, J. T. Peters, S. D. Steichen and N. A. Peppas, Mater. Sci. Eng., R, 2015, 93, 1–49.
9. Z. Tai, J. Yang, Y. Qi, X. Yan and Q. Xue, RSC Adv., 2013, 3, 12751–12757.
10. D. Das and S. Pal, RSC Adv., 2015, 5, 25014–25050.
11. Y. Qiu and K. Park, Adv. Drug Delivery Rev., 2012, 64, 49–60.
12. K. Y. Lee and D. J. Mooney, Chem. Rev., 2001, 101, 1869–1880.
13. M. T. Calejo, S. A. Sande and B. Nystrom, Expert Opin. Drug Delivery, 2013, 10, 1669–1686.
14. V. Tokárová, A. Pittermannová, J. Čech, P. Ulbrich and F. Štěpánek, Soft Matter, 2012, 8, 1087–1095.
15. X.-Z. Zhang, X.-D. Xu, S.-X. Cheng and R.-X. Zhuo, Soft Matter, 2008, 4, 385–391.
16. M. Patenaude and T. Hoare, ACS Macro Lett., 2012, 1, 409–413.
17. D. G. Barnes, M. Vidiassov, B. Ruthensteiner, C. J. Fluke, M. R. Quayle and C. R. McHenry, PLoS One, 2013, 8, e69446.
18. R. Hernández, M. Criado, A. Nogales, M. Sprung, C. Mijangos and T. A. Ezquerra, Macromolecules, 2015, 48, 393–399.
19. H. Cong, L. Li and S. Zheng, Polymer, 2013, 54, 1370–1380.
20. L. Miao, M. Lu, C. Yang, Y. Zhang, T. Shen and J. Ma, J. Appl. Polym. Sci., 2013, 128, 275–282.
21. Y. Li, H. Guo, J. Gan, J. Zheng, Y. Zhang, K. Wu and M. Lu, J. Polym. Res., 2015, 22, 91.
22. C. Zhao, X. Zhuang, P. He, C. Xiao, C. He, J. Sun, X. Chen and X. Jing, Polymer, 2009, 50, 4308–4316.
23. G. Iyer, L. M. V. Tillekeratne, M. R. Coleman and A. Nadarajah, Polymer, 2008, 49, 3737–3743.
24. G. Kali, S. Vavra, K. László and B. Iván, Macromolecules, 2013, 46, 5337–5344.
25. E. M. Anderson, M. L. Noble, S. Garty, H. Ma, J. D. Bryers, T. T. Shen and B. D. Ratner, Biomaterials, 2009, 30, 5675–5681.
26. T. L. Sun, T. Kurokawa, S. Kuroda, A. B. Ihsan, T. Akasaki, K. Sato, M. A. Haque, T. Nakajima and J. P. Gong, Nat. Mater., 2013, 12, 932–937.
27. M. R. Park, B. B. Seo and S. C. Song, Biomaterials, 2013, 34, 1327–1336.
28. A. Mellati, S. Dai, J. Bi, B. Jin and H. Zhang, RSC Adv., 2014, 4, 63951–63961.
29. M. K. Gupta, J. R. Martin, T. A. Werfel, T. Shen, J. M. Page and C. L. Duvall, J. Am. Chem. Soc., 2014, 136, 14896–14902.
30. G. P. Misra, R. S. Singh, T. S. Aleman, S. G. Jacobson, T. W. Gardner and T. L. Lowe, Biomaterials, 2009, 30, 6541–6547.
31. M. H. Park, M. K. Joo, B. G. Choi and B. Jeong, Acc. Chem. Res., 2012, 45, 424–433.
32. J. Yang, R. van Lith, K. Baler, R. A. Hoshi and G. A. Ameer, Biomacromolecules, 2014, 15, 3942–3952.
33. B.-y. Zhang, W.-D. He, L.-y. Li, X.-l. Sun, W.-t. Li and K.-r. Zhang, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 3604–3612.
34. S. Deshayes and A. M. Kasko, J. Polym. Sci., Part A: Polym. Chem., 2013, 51, 3531–3566.
35. X. Huang, G. P. Misra, A. Vaish, J. M. Flanagan, B. Sutermaster and T. L. Lowe, Macromolecules, 2008, 41, 8339–8345.
36. D.-Q. Wu, F. Qiu, T. Wang, X.-J. Jiang, X.-Z. Zhang and R.-X. Zhuo, ACS Appl. Mater. Interfaces, 2009, 1, 319–327.
37. M. H. Lee, Z. Yang, C. W. Lim, Y. H. Lee, S. Dongbang, C. Kang and J. S. Kim, Chem. Rev., 2013, 113, 5071–5109.
38. D. J. Phillips and M. I. Gibson, Biomacromolecules, 2012, 13, 3200–3208.
39. P. Du, J. Liu, G. Chen and M. Jiang, Langmuir, 2011, 27, 9602–9608.
40. W. Lv, S. Liu, W. Feng, J. Qi, G. Zhang, F. Zhang and X. Fan, Macromol. Rapid Commun., 2011, 32, 1101–1107.
41. K. A. Athanasiou, G. G. Niederauer and C. M. Agrawal, Biomaterials, 1996, 17, 93–102.
42. Y.-S. Lin, Y.-L. Huang, W.-F. Lee and C.-H. Lin, J. Chin. Chem. Soc., 2014, 61, 945–952.
43. F. Meng, W. E. Hennink and Z. Zhong, Biomaterials, 2009, 30, 2180–2198.
44. W. N. E. van Dijk-Wolthuis, J. A. M. Hoogeboom, M. J. van Steenbergen, S. K. Y. Tsang and W. E. Hennink, Macromolecules, 1997, 30, 4639–4645.
45. N. Grant, H. Wu and H. Zhang, Ind. Eng. Chem. Res., 2014, 53, 246–252.
46. A. Galperin, T. J. Long and B. D. Ratner, Biomacromolecules, 2010, 11, 2583–2592.
47. M. A. de Moraes, E. Paternotte, D. Mantovani and M. M. Beppu, Macromol. Biosci., 2012, 12, 1253–1264.
48. H. Sun, L. Mei, C. Song, X. Cui and P. Wang, Biomaterials, 2006, 27, 1735–1740.
49. X.-D. Xu, H. Wei, X.-Z. Zhang, S.-X. Cheng and R.-X. Zhuo, J. Biomed. Mater. Res., Part A, 2007, 81, 418–426.
50. X. Huang and T. L. Lowe, Biomacromolecules, 2005, 6, 2131–2139.
51. L. Klouda and A. G. Mikos, Eur. J. Pharm. Biopharm., 2008, 68, 34–45.
52. L.-F. Sun, R.-X. Zhuo and Z.-L. Liu, Macromol. Biosci., 2003, 3, 725–728.
53. S. W. Kim, Y. H. Bae and O. Teruo, Pharm. Res., 1992, 9, 283–290.
54. C. Ji and J. Shi, Mater. Lett., 2012, 72, 110–112.
55. X. Huang, B. R. Nayak and T. L. Lowe, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 5054–5066.
56. Y. Zhang and C. C. Chu, J. Biomed. Mater. Res., 2002, 59, 318–328.
57. C. H. Wang, Y. S. Hwang, P. R. Chiang, C. R. Shen, W. H. Hong and G. H. Hsiue, Biomacromolecules, 2012, 13, 40–48.
58. X. J. Loh, S. H. Goh and J. Li, Biomaterials, 2007, 28, 4113–4123.
59. K. Y. Lee, J. A. Rowley, P. Eiselt, E. M. Moy, K. H. Bouhadir and D. J. Mooney, Macromolecules, 2000, 33, 4291–4294.
60. R. Fei, J. T. George, J. Park, A. K. Means and M. A. Grunlan, Soft Matter, 2013, 9, 2912–2919.
61. S. Shekhar, M. Mukherjee and A. K. Sen, Polym. Bull., 2016, 73, 125–145.
62. Y. Zhang, Y. Li and W. Liu, Adv. Funct. Mater., 2015, 25, 471–480.
63. X. Z. Shu, Y. Liu, Y. Luo, M. C. Roberts and G. D. Prestwich, Biomacromolecules, 2002, 3, 1304–1311.
64. X. Wan, D. Wang and S. Liu, Langmuir, 2010, 26, 15574–15579.
65. Y. Wang, J. Nie, B. Chang, Y. Sun and W. Yang, Biomacromolecules, 2013, 14, 3034–3046.
66. B. G. Stubbe, W. E. Hennink, S. C. De Smedt and J. Demeester, Macromolecules, 2004, 37, 8739–8744.
67. K. Y. Lee, E. Alsberg and D. J. Mooney, J. Biomed. Mater. Res., 2001, 56, 228–233.
68. L. Zhang, W. Liu, L. Lin, D. Chen and M. H. Stenzel, Biomacromolecules, 2008, 9, 3321–3331.
69. S. S. Sivan, S. Roberts, J. P. G. Urban, J. Menage, J. Bramhill, D. Campbell, V. J. Franklin, F. Lydon, Y. Merkher, A. Maroudas and B. J. Tighe, Acta Biomater., 2014, 10, 1124–1133.

---