Hydrolytically Degradable Polyrotaxane Hydrogels for Drug and Cell

Article pubs.acs.org/Biomac

Hydrolytically Degradable Polyrotaxane Hydrogels for Drug and Cell Delivery Applications Clementine Pradal,† Lisbeth Grøndahl,‡ and Justin J. Cooper-White*,†,§,∥ †

Tissue Engineering and Microfluidics Laboratory, Australian Institute for Bioengineering and Nanotechnology, ‡School of Chemistry and Molecular Biosciences, §The School of Chemical Engineering, The University of Queensland, St. Lucia, Queensland 4072, Australia ∥ CSIRO, Manufacturing Flagship, Clayton, Victoria 3168, Australia S Supporting Information *

ABSTRACT: Self-assembled pseudopolyrotaxane (PPR) hydrogels formed from Pluronic polymers and α-cyclodextrin (α-CD) have been shown to display a wide range of tailorable physical and chemical properties that may see them exploited in a multitude of future biomedical applications. Upon the mixing of both components, these self-assembling hydrogels reach a metastable thermodynamic state that is defined by the concentrations of both components in solution and the temperature. However, at present, their potential is severely limited by the very nature by which they form and hence also disassemble. Even if the temperature is kept constant, PPR hydrogels will dissociate and collapse within a few hours when immersed in a liquid (such as cell culture media) that contains a lower concentrations of, or no, Pluronic or α-CD due to differences in chemical potential driving dissolution. In this article, an enzymatically mediated covalent cross-linking function and branched eight-arm poly(ethylene glycol) (PEG) were thus introduced into the PPR hydrogels to improve their robustness to such environmental changes. The eight-arm PEG also acted as an end-capping group to prevent the dethreading of the α-CD molecules. The covalent cross-linking successfully extended the lifetime of the hydrogels when placed in cell culture media from a few hours to up to 1 week, with the ability to control the degradation rate (now initiated by hydrolysis of the introduced ester bonds and not by dissolution) by changing the amount of eight-arm PEG present in the hydrogels. Highly tunable hydrogels were obtained with an elastic modulus between 20 and 410 kPa and a viscous modulus between 150 Pa and 22 kPa by varying the concentrations of α-CD and eight-arm PEG. Sustained release of a model drug from the hydrogels was achieved, and viability of mouse fibroblasts encapsulated in these hydrogels was assessed. These self-assembling, hydrolytically degradable, and highly tunable hydrogels are seen to have potential applications in tissue engineering relying on controlled drug or cell delivery to sites targeted for repair.

■

elucidated their molecular behavior.8 Additionally, cyclodextrins can be functionalized through the pendant hydroxyl groups present on their outside rim, which makes PPRs attractive for biomedical applications. Indeed, PPR hydrogels have been proposed as drug or gene delivery vehicles: systems composed of PEO,9,10 PEO−poly(ε-caprolactone),11−13 PEO−poly[(R)3-hydroxybutyrate]−PEO,14 and Pluronic (F127)15,16 have been used with α-CD to form hydrogels for drug release, whereas Pluronic threaded with oligoethylenimine-grafted βCDs17,18 or Pluronic (F68) and poly(L-lysine) polymers threaded with α-CD19 has been used for DNA delivery. Self-assembly is an interesting gelation mechanism for materials targeted for tissue engineering applications because it avoids the use of chemicals or external stimuli (e.g., UV) to cross-link the hydrogels that can potentially be harmful to cells.

INTRODUCTION Self-assembling hydrogel systems, in which the components spontaneously form ordered aggregates,1 is driven by weak noncovalent interactions such as van der Waals interactions, the hydrophobic effect, and hydrogen bonding and can occur on the molecular and nanoscale2 levels (e.g., folding of an amino acid sequence into α-helices or β-sheets, micellar assembly of amphiphilic molecules) or on larger micro and macro scales, for example, in the self-assembly of synthetic hydrogel building blocks.3,4 α-Cyclodextrin (α-CD) and poly(ethylene oxide) (PEO) or PEO-based block copolymers, such as Pluronic polymers, are known to form self-assembled pseudopolyrotaxane (PPR) hydrogels, where α-CD preassembles into aggregates and threads onto the PEG units of the polymer.5−8 These pseudopolyrotaxanes, in turn, self-assemble into larger structures, leading to the formation of a hydrogel. Previous investigations into α-CD/Pluronic systems have highlighted their attractive tunable gelation and self-healing properties and © XXXX American Chemical Society

Received: November 4, 2014 Revised: December 2, 2014

A

dx.doi.org/10.1021/bm501615p | Biomacromolecules XXXX, XXX, XXX−XXX

Biomacromolecules

Article

with the aim of creating covalent cross-links between the PPR components of these hydrogels. In addition, a tyraminefunctionalized multiarm PEG was added into the hydrogels to generate a branched network and improve the stability of the hydrogels. These new generation hydrogel designs were then assessed for possible future tissue engineering applications relying on drug and cell delivery.

However, by nature, self-assembled systems are reversible, and perturbations to their thermodynamic equilibrium can bring them back to their dissociated state. For example, a Pluronic hydrogel formed by micellar entanglement due to the hydrophobic associations between the PPO units will revert to a liquid Pluronic unimer solution if a decrease in temperature or a decrease in the Pluronic concentration occurs.20 PPR hydrogels are not exempt from this phenomenon, and in fact, PPR hydrogels have been shown to dissociate rapidly if immersed in a liquid not containing the pseudopolyrotaxane molecules.9,14 This was assumed to be due to rapid interfacial dissolution and eventual failure of the hydrogel as the local concentration of the PPR components decreased below the gelation threshold, as a result of water diffusion into the hydrogel and diffusion of PPR components out of the hydrogel to balance the chemical potential of the system. This reversibility, although of interest in some applications, is not ideal when the end use of the hydrogel is targeted toward tissue engineering. Indeed, if the hydrogels are to be inserted in the body, then they will be exposed to interstitial fluid flow that will lead to the dissociation of these noncovalent hydrogels. From a laboratory perspective, if cells are to be cultured on or within the hydrogels, then the addition of culture media on top of the gels to supply the cells with nutrients will also result in their dissociation. Moreover, the rapid degradation of the self-assembled PPR hydrogels will likely release a large amount of α-CD. Large amounts of free αCD can be cytotoxic to cells, due to their ability to solubilize membrane lipids (e.g., cholesterol21), which disrupts the cell membrane and can cause cell death.22,23 Therefore, rapid localized degradation of these types of hydrogels is not a desirable characteristic from a tissue engineering perspective. To improve the stability of these PPR hydrogels for future tissue engineering applications, we decided to investigate the introduction of a covalent cross-linking function into the gels. To be suitable for cell encapsulation, the cross-linking process should avoid leaving residual cytotoxic components in the hydrogels and should not require exposure to harsh chemical or physical conditions. Peroxidase-catalyzed oxidation permits coupling of two phenolic moieties under mild conditions.24,25 In the presence of horseradish peroxidase (HRP) and hydrogen peroxide (H2O2), C−C and C−O bonds can be formed between two phenol groups, allowing the formation of hydrogels from polymers functionalized with phenolic moieties. This mechanism of cross-linking has been used with various synthetic polymers (eight-arm PEG,26 Tetronic27−30) and biopolymers (hyaluronic acid,24,31−34 chitosan,35,36 gelatin37−42). Although H2O2 is toxic to cells at high concentrations,43 it is consumed (and converted to H2O) during the cross-linking reaction, and its rate of consumption can be controlled by changing the kinetics of the reaction through changing the ratio of H2O2 to HRP.24 In this way, the exposure time of H2O2 to the cells can be minimized. H2O2/HRPmediated cross-linking has been successfully used for cell culture in 2D using mouse embryonic fibroblasts,35 mouse myoblasts,44,45 and human umbilical vein endothelia cells (HUVECs).30 In 3D, human mesenchymal stem cells (hMSCs) have been successfully encapsulated within hydrogels crosslinked with H2O2/HRP without compromising the viability of the encapsulated cells.26,37 In this article, an enzymatically mediated cross-linking function was thus introduced onto the Pluronic end groups of the PPRs through the use of a phenolic moiety (tyramine),

■

MATERIALS AND METHODS

Materials. Pluronic F68 (PEO76PPO29PEO76, Mw = 8400 g mol−1), Pluronic F127 (PEO100PPO65PEO100, Mw = 12 600 g mol−1), α-CD, 4dimethylaminopyridine (DMAP), triethylamine (TEA), N-hydroxysuccinimide (NHS), N,N′-dicyclohexylcarbodiimide (DCC), succinic anhydride (SA), tyramine (Tyr), HRP (113 U mg−1), tetrahydrofuran (THF), 1,4-dioxane (DO), diethyl ether, 6-aminofluorescein, chlorotrimethylsilane (CTMS), MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide), and dimethyl sulfoxide (DMSO) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Phosphate buffered saline (PBS) 10× stock was diluted with Milli-Q water (Merck Millipore, Billerica, MA, USA). H2O2 (30% solution) was obtained from Merck (Whitehouse Station, NJ, USA), eight-arm PEG (10 000 g mol−1, hexaglycerol core) was obtained from JenKem (Allen, TX, USA), high glucose Dulbecco’s modified Eagle’s medium (DMEM), batch tested fetal bovine serum (FBS), penincillin/ streptomycin, and the live/dead reduced biohazard cell viability (L7013) kit were purchased from Life Technologies (Carlsbad, CA, USA). All materials were used as received without further purification. Synthesis of Pluronic and PEG Functionalized with Tyramine. To synthesize tyramine-terminated Pluronic (F68-Tyr or F127-Tyr) or eight-arm PEG (PEG-Tyr), the polymer was first functionalized with a carboxylic group according to the following procedure, adapted from ref 46 (Scheme 1; note that only Pluronic is

Scheme 1. (A) Reaction Scheme for the Functionalization of Pluronic with Tyramine and (B) Cross-Linking Reaction between Two Phenols Present as End Groups on a Polymer Mediated by H2O2 and HRP

shown, but the reaction scheme is identical for the eight-arm PEG). The polymer was dissolved in anhydrous DO (1 g of polymer in 10 mL of DO) under a nitrogen atmosphere. DMAP and TEA were added in a 2.5 molar excess to the number of end groups. The reaction was stirred at room temperature for 15 min before addition of SA in a 2.5 molar excess to the number of end groups. The reaction was left to stir at room temperature for 24 h under a nitrogen atmosphere. The DO was then evaporated under vacuum, and the product was B


Biomacromolecules

Article

H2O2 was investigated in the concentration range between 7.8 and 45 mM and added to the solution, followed by a rapid mixing step (by pipetting the solution up and down), before the solution was either loaded onto a rheometer to study its gelation properties or pipetted into a mold or onto a glass slide to form gels. To form enzymatically cross-linked gels from the PPRs, these were first formed by mixing solutions of F68-Tyr or F127-Tyr with solutions of α-CD as described previously8 to obtain a final Pluronic concentration of 10% (w/v) with the final desired α-CD theoretical coverage. The theoretical coverage is defined as the percentage of PEO units covered by α-CD, using the fact that one molecule of α-CD covers two units of PEO.48 It should be noted that the theoretical coverage is an approximation of the actual coverage, as not all of the αCD molecules incorporated in the solution will thread onto the Pluronic. The PPR hydrogels were left to assemble for 4 h, at 20 °C for the gels formed from Pluronic F68-Tyr to ensure the absence of micelles, or at 37 °C for the gels formed from Pluronic F127-Tyr to obtain some micelles.8 After the PPR gelation step, the white opaque gels were sheared manually with a spatula and a PEG-Tyr solution (if necessary), HRP, and H2O2 solutions were added. Given their consistency, the PPR gels could not be pipetted and were manually loaded onto the rheometer or in a mold using a spatula before the enzymatic cross-linking took place. The nomenclature for the gels containing α-CD is as follows: type of Pluronic-type of end groupα% coverage-letter representing the ratio of PEG-Tyr to Pluronic, with the coverage being defined as the theoretical percentage of PEO units covered by α-CD molecules. For example, F68-Tyr-α10-L represents a gel composed of F68-Tyr and a ratio of one PEG-Tyr molecule for eight F68-Tyr molecules and containing enough α-CD to cover 10% of the PEO units in F68. The general schematic of the hydrogels with all of the components is presented Figure 1. The hydrogel based on F127-Tyr will have

precipitated with cold diethyl ether, collected by centrifugation, and dried under vacuum overnight. The yield was calculated to be 96%. To functionalize the polymers with tyramine, the following procedure, adapted from ref 46, was used: the carboxylated polymers were dissolved in anhydrous THF under a nitrogen atmosphere (1 g of Pluronic in 20 mL of THF). DCC and NHS were added in a 2.5 molar excess to the number of end groups. The reaction was stirred at room temperature for 1 h before addition of tyramine (dissolved in THF) in a 2.5 molar excess to the number of end groups. The reaction was left to stir at room temperature for 24 h under a nitrogen atmosphere. The THF was then evaporated under vacuum, and the product was precipitated with cold diethyl ether. The white powder was then dissolved in water, and some insoluble material was seen, corresponding to a dicyclohexylurea salt.47 The solution was dialyzed against water for 3 days before being filtered to remove the insoluble material and freeze-dried. The yield was calculated to be 90%. 1 H Nuclear Magnetic Resonance. Quantitative 1H NMR (500 MHz) spectra were acquired on a Bruker Avance III HP 500 NMR spectrometer (Supporting Information Figures 1−3). Spectra were recorded of the polymers dissolved in DMSO-d6, and the solvent peak at δ = 2.50 ppm was used as the internal reference. The degree of substitution (DS) of tyramine on the Pluronic molecules was calculated from the integral of the methyl protons resonance from the PPO units (δ = 1.04 ppm) and the aromatic resonances from tyramine (δ = 6.97 and 6.70 ppm). For PEG-Tyr, the methylene and methine protons were used (δ = 3.50 ppm). To ensure the peaks observed for tyramine were arising from tyramine covalently bound to Pluronic and PEG, pulsed gradient 1H NMR was employed at a gradient of 3 and 95%, and the intensity of the resonances was compared to the quantitative scans (Supporting Information Figures 1−3). The intensities of the methyl protons of the Pluronic polymers or methylene and methine protons of PEG compared to the aromatic protons of tyramine remained unchanged in the diffusion spectra, showing that all of the tyramine groups were covalently bound to the polymer backbone with no free material present. A high degree of functionalization of 96% (±4) was obtained for both F127-Tyr and F68-Tyr, which is equivalent to an average of 1.92 out of 2 end groups functionalized. For PEG-Tyr, the degree of functionalization was found to be 98% (±4), which is equivalent to 7.84 out of 8 arms functionalized per molecule, on average. Formation of Hydrogels Containing Eight-Arm PEG and αCD. Two types of gels were studied: gels without α-CD and gels containing α-CD threaded onto Pluronic F68 or F127 with tyramine end groups. Our previous investigation8 showed that PPRs formed from F68 and F127 polymers have different molecular structures, principally due to the absence of micelles in the F68-based systems, whereas micelles are present in the F127 systems. These insights prompted both types of Pluronic to be investigated in this work, aiming to determine whether the differences in the molecular structure influence the covalent cross-linking process and the behavior of the resulting hydrogels. Moreover, the type of Pluronic used influences the mechanical properties of the self-assembled PPR hydrogels and hence, by comparing both types of Pluronic, this enables one to see whether these differences in structure and mechanical properties translate to the covalently cross-linked system. To determine the ideal concentrations of HRP and H2O2 required for the cross-linking, gels were first formed without α-CD. Pluronic-Tyr was dissolved in PBS. In some cases, PEG-Tyr was added to the solution with varying ratios of PEG to Pluronic. The ratios used were as follows: no PEG-Tyr (0), one PEG-Tyr molecule for eight (low ratio, L), for four (medium ratio, M), or for two (high ratio, H) molecules of Pluronic-Tyr. Unless otherwise stated, all of the gels are covalently cross-linked through the phenol end groups using HRP/H2O2. The final concentration of Pluronic-Tyr was 10% (w/v) for all experiments. Samples are designated as follows: type of Pluronic-type of end group-letter representing the ratio of PEG-Tyr to Pluronic. For example, F68-TyrL represents a gel composed of F68-Tyr and a ratio of one PEG-Tyr molecule for eight F68-Tyr molecules. To form the gel, HRP was added and mixed into the solution. The range of concentration investigated for HRP was 0.125 U mL−1 to 0.625 U mL−1. Finally,

Figure 1. Schematic of the covalently cross-linked F68 PR hydrogels containing eight-arm PEG. Pluronic micelles instead of linear Pluronic molecules (as would be the case for the F68-Tyr system8). The introduction of PEG-Tyr into the hydrogels leads to the end capping of the PPRs, thus forming polyrotaxanes (PRs). Although not all of the end groups might be endcapped, the abbreviation PR will be used when the Pluronic/α-CD assemblies are cross-linked with PEG-Tyr, and the abbreviation PPR will be used for the purely self-assembled systems. Rheological Analysis. The shear elastic (storage) modulus (G′) and viscous (loss) modulus (G″) of the hydrogels were measured using an AR 1500ex rheometer (TA Instruments, New Castle, DE, USA) with a cone (diameter, 40 mm; angle, 2°; truncature distance, 56 μm) and plate geometry. Strain and frequency sweeps were performed C


Biomacromolecules

Article

Figure 2. Optimization of the concentrations of H2O2 and HRP for F68-Tyr-0 and F127-Tyr-0 hydrogels. The gelation time (gray circles) and elastic modulus G′ (after 1 h, black squares) are reported (mean of triplicates ± standard error of the mean), and the kinetics curves (limited to the first 30 min for clarity) are shown. (A, C) Varying [H2O2] for F68-Tyr-0 with [HRP] constant at 0.5 U mL−1; (B, D) varying [HRP] for F68-Tyr-0 with [H2O2] constant at 15 mM; (E, G) varying [H2O2] for F127-Tyr-0 with [HRP] constant at 0.5 U mL−1; (F, H) varying [HRP] for F127-Tyr-0 with [H2O2] constant at 11.7 mM. Mean of triplicates ± standard error of the mean. HRP at 37 °C with a strain of 0.1% and a frequency of 1 Hz to determine the amounts of H2O2 and HRP to form cross-linked gels with the desired kinetics of gelation, that is, a gelation time that allows for the manipulation of the solutions (loading onto the rheometer, cell encapsulation, etc.) before the gel is formed, and to determine their mechanical properties. The solutions were mixed as described above and immediately loaded onto the rheometer. The initial optimization

to determine the boundaries of the linear viscoelastic region of the systems studied. Time sweeps at 37 and 20 °C with a strain of 0.1% and a frequency of 1 Hz were performed to follow the gelation of Pluronic-Tyr with α-CD in order to ensure that the threading of α-CD onto Pluronic was not impaired by the presence of the tyramine end groups. For the enzymatic cross-linking, time sweeps were performed in triplicate for various compositions of Pluronic, PEG-Tyr, H2O2, and D


Biomacromolecules

Article

added last as described, and 50 μL drops of the suspension were pipetted onto the glass slide to form hemispheres and incubated at 37 °C until the gelation was complete. For the gels containing α-CD, the PPR gels were first sheared, and solutions of PEG-Tyr and HRP were added to the gels before the cell suspension was added into the mixture. H2O2 was then added to this mixture, and 50 μL drops of gel were deposited onto the glass and incubated at 37 °C until the gelation was complete. The hydrogel hemispheres were then transferred into a 48-well plate, and 500 μL of high-glucose DMEM was added on top of the hydrogels. A live/dead cell viability kit was used to stain the cells so that they could be observed as either live (green), dead (red), or dying (yellow/orange) cells within the hydrogels. Z-stacked images (100 μm, 10 μm slices) were obtained using a Zeiss LSM710 confocal laser scanning microscope. Images in at least two gel locations were taken. Statistical Analysis. Unpaired, two-tailed Student’s t tests and two-way ANOVA tests were performed using the software GraphPad Prism 6 (CA, USA) to determine the significance between the different samples for the mechanical properties, degradation properties, drug release properties, and cell viability experiments.

of H2O2 and HRP was performed without PEG-Tyr, and the elastic modulus G′ obtained after 1 h is reported, with this time frame being sufficient for the mechanical properties to reach a plateau for all of the conditions tested. For the optimization of the gelation conditions in the presence of PEG-Tyr, the elastic modulus G′ obtained after 1 h is reported. Once the optimal concentrations of H2O2 and HRP were determined, the gels containing α-CD were also studied. The solutions were prepared as described above and immediately loaded onto the rheometer. The elastic modulus G′ obtained after 30 min is reported (being enough time for the mechanical properties to reach a plateau). Frequency sweeps between 0.1 and 100 Hz and a strain of 0.1% were also performed after the gels were formed. Due to the substantial viscoelastic nature of these gels and the recent insights into the importance of such mechanical properties on stem cell behaviors,49,50 the time-dependent deformation response of these hydrogels to an imposed step stress (or creep response) was also tested. The linear viscoelastic regions of the gels were initially confirmed using oscillatory stress sweep tests by cross-linking the gels on the rheometer for 30 min (while monitoring the gelation using a time sweep), as described above, and then applying oscillatory stresses ranging from 1 to 1200 Pa at a frequency of 1 Hz and a temperature of 37 °C. To assess the creep behavior, the gels were cross-linked for 30 min on the rheometer as described above (while monitoring the gelation using a time sweep) followed by application of a constant stress of 1000 Pa (as a step function) to the sample for a period of 30 min at 37 °C and measuring the resultant observed strain as a function of time. This value of stress was chosen as representative of the stress that cells exert on their surroundings51−53 while being within the linear range of the gels. The samples were performed in triplicates. Degradation Study. Gels with and without α-CD were formed as described above and poured into cylindrical silicone molds with a diameter of 11 mm and a height of 3 mm. The use of silicone molds ensured that all the gels are of comparable volume and surface area. The gels were left to set for 30 min at 37 °C before being extracted from the mold, placed into 5 mL plastic tubes, and weighed. One milliliter of PBS was then added on top of the gels, and the tubes were incubated at 37 °C. Each day, PBS was removed, and the gels were gently blotted to remove the excess PBS and weighed. Every 3 days, the PBS was replaced with fresh PBS to mimic cell culture conditions, where the media is exchanged at regular intervals. The gels were weighed until complete degradation. Triplicates of each sample were performed. The percentage of the initial mass is reported. To study the mechanism of degradation of the hydrogels, the supernatant collected during each PBS change until complete degradation of the gels was kept and pooled. It was then dialyzed (cutoff: 3500 Da) against water for 3 days to remove any small degradation products. The resulting solution was freeze-dried, and the powder obtained was dissolved in DMSO-d6 and analyzed by NMR. Drug Release. As a model drug, a poorly water-soluble molecule, 6-aminofluorescein, was encapsulated in the gels. 6-Aminofluorescein was first dissolved in PBS at a concentration of 0.1 mM. This solution was then either mixed with the Pluronic solution, PEG-Tyr, H2O2, and HRP to form the gels without α-CD or was used to dissolve α-CD and then added to the Pluronic solution to form the PPR gel before the enzymatic cross-linking was performed. The model drug loading used was 12 μg of 6-aminofluorescein per gram of hydrogel. The gels were formed into a disc shape as per the degradation study and placed into 5 mL plastic tubes. One milliliter of PBS was added on top of the hydrogels. At regular time points, 100 μL of the PBS was taken from the supernatant, the absorbance was then measured at 490 nm, and the 100 μL was then added back to the rest of the supernatant. Triplicates were performed for each condition. The measurements were performed until complete disintegration of the hydrogels. Three-Dimensional Cell Encapsulation. NIH-3T3 mouse embryonic fibroblasts were encapsulated in the hydrogels at a density of 2 million cells mL−1 according to the following method: glass slides were hydrophobized by spreading CTMS on the surface, followed by thorough rinsing with distilled water. The hydrogel mixtures were prepared by resuspending the required number of cells in a solution composed of F68-Tyr or F127-Tyr, PEG-Tyr, and HRP. H2O2 was

■

RESULTS Optimization of Enzymatic Cross-Linking Conditions. The amounts of HRP and H2O2 for cross-linking of the PPR gels were first optimized with F68-Tyr-0 and F127-Tyr-0, that is, the tyramine-functionalized polymers in the absence of αCD and PEG-Tyr (Figure 2). The gels formed by this method would have end-to-end cross-linked Pluronic polymers. The aim was to determine the optimal H 2 O 2 and HRP concentrations that would lead to the highest mechanical properties and optimal cross-linking time. The highest mechanical properties were desired in order to mitigate the loss of mechanical properties due to the shearing of the PPR hydrogels preceding the enzymatic cross-linking. Optimal crosslinking time should allow for manipulation of the gels (cell encapsulation, transfer into a mold, etc.) and consequently cannot be less than a few seconds. The ratio of H2O2 to HRP also had to be optimized to avoid depletion of the H2O2, which manifested as a sudden plateau in the mechanical properties, as seen in Figure 2, panels C, for [H2O2] = 10 mM, and G, for [H2O2] = 7.8 mM. Theoretically, the formation of a bond between two phenols requires one molecule of H2O2.24 This amount served as the starting point for the optimization protocol. For both F68-Tyr-0 and F127-Tyr-0, increasing the H2O2 concentration led to an increase in gelation time (Figure 2). The highest mechanical properties were achieved using 1.2 to 1.5 molecules of H2O2 for 2 molecules of phenol. Larger amounts of H2O2 led to a decrease in mechanical properties. The decrease in mechanical properties and the increase in the gelation time upon increase of the H2O2 concentration might be due to inactivation of HRP by H2O254−57 and has been previously reported for other types of hydrogels.24,26 An increase in the HRP concentration decreased the gelation time while increasing the mechanical properties. However, above 0.63 U mL−1 for F127-Tyr (data not shown) and 0.5 U mL−1 for F68-Tyr, a further increase in the HRP concentration either led to gels with gelation kinetics that were too fast for the targeted applications or no further increase in mechanical properties. Overall, the best cross-linking conditions found for F68-Tyr-0 were 15 mM of H2O2 and 0.5 U mL−1 of HRP, and for F127-Tyr-0, the optimal conditions were 11.7 mM of H2O2 and 0.63 U mL−1 of HRP (listed in Table 1). PEG-Tyr was subsequently added in various ratios in order to create a branched network. The amounts of H2O2 and HRP were scaled up proportionally to the number of end groups E


Biomacromolecules

Article

Figure 3. Optimization of [H2O2] and [HRP] for F127-Tyr-L. The gelation time (gray circles) and elastic modulus G′ (after 1 h, black squares) are reported (mean of triplicates ± standard error of the mean) and the kinetics curves (limited to the first 30 min for clarity) are shown. (A, C) Fixed [HRP] at 0.96 U mL−1; (B, D) fixed [H2O2] at 18 mM.

present in the mixtures. As a proof of concept that a simple scale up still corresponded to the optimal cross-linking conditions, the amounts of H2O2 and HRP were varied for F127-Tyr-L. The scaled up amounts for F127-Tyr-L from the optimal conditions found with F127-Tyr-0 are 18 mM of H2O2 and 0.96 U mL−1 of HRP. Three other concentrations of H2O2 ([HRP] fixed) and 3 other concentrations of HRP ([H2O2] fixed) were studied, and the best conditions were found to be the scaled up amounts from the F127-Tyr-0 optimization (Figure 3). Scale up of the H2O2 and HRP concentrations was applied in a similar way to the other ratios of PEG-Tyr to F127Tyr and to F68-Tyr. The HRP and H2O2 concentrations used are summarized in Table 1. Figure 4. Evolution of the mechanical properties as a function of time during threading of α-CD onto Pluronic F127 and F127-Tyr at 37 °C. Black, G′; gray, G″. Continuous line, F127-α10-0; dotted line, F127Tyr-α10-0, before covalent cross-linking.

Table 1. Optimized H2O2 and HRP Concentrations for the Various Tyramine-Based Hydrogel Compositionsa

a

gel composition

optimal [HRP] (U mL−1)

optimal [H2O2] (mM)

F68-Tyr-0 F68-Tyr-L F68-Tyr-M F68-Tyr-H F127-Tyr-0 F127-Tyr-L F127-Tyr-M F127-Tyr-H

0.50 0.75 1.00 1.5 0.63 0.96 1.28 1.92

15 22.5 30 45 11.7 18 24 36

the PPR assembly, although a slower initial threading of the αCD onto F127-Tyr was observed. The final mechanical properties of the PPR hydrogels were not affected. Enzymatic cross-linking using HRP is a relatively fast process (a few minutes) compared to the formation of the PPRs (a few hours for the mechanical properties to reach a plateau). Sufficient time needs to be allowed for the PPR gels to form by threading of the α-CD onto the Pluronic polymers before covalently cross-linking the gels with H2O2 and HRP. In order to accommodate for these two very different time scales, the self-healing capacity of the PPR hydrogels8 was exploited. The PPR gels were allowed to form for 4 h before being sheared to allow the addition and mixing of PEG-Tyr, HRP, and H2O2. F127-Tyr-α8 does not form a gel;8 however, it was included in order to evaluate if the presence of α-CD influences the mechanical properties of the gels compared to the gels formed in the absence of α-CD. The frequency sweeps of the cross-

The same concentrations were used in the presence of α-CD.

Enzymatic Cross-Linking and Characterization of the PR Hydrogels. The threading of α-CD onto the tyramine functionalized Pluronic polymers was validated by following the elastic and viscous moduli over time of mixtures composed of F127-Tyr and α-CD with a coverage of 10%. As shown in Figure 4, comparing functionalized and nonfunctionalized F127, the presence of tyramine does not significantly impair F


Biomacromolecules

Article

Madin-Darby canine kidney (MDCK) cell has been reported to be 0.6 nN μm−2 (600 Pa), and the highest stress was found to be 3.8 ± 0.1 nN μm−2 (3800 Pa),51 whereas human foreskin fibroblasts have been reported to exert stress values of 5.5 ± 2 nN μm−2 (5500 Pa).53 In 3D, studies have found that mouse fibroblasts exert tractions in the range of 100 to 5000 Pa.58 Here, 1000 Pa was chosen to match the order of magnitude of stresses exerted by cells while being within the linear viscoelastic range of the gels. Figure 7 shows the creep response of F68-Tyr hydrogels with varying amount of PEGTyr and α-CD. Upon increasing the amount of PEG-Tyr, the total magnitude of strain observed in the samples post creepringing decreases (Figure 7A), as expected due to the increase in the elastic modulus of these gels. The rate of creep, however, also decreases with increasing PEG-Tyr ratios, although, when quantified across the number of samples tested for each gel type, these decreases did not appear to be statistically significant (Figure 7B). Increasing the amount of α-CD leads to decreasing levels of strain post creep-ringing (Figure 7C), again, an expected result due to increased values of elastic modulus across these gels. However, somewhat unexpectedly, the rate of creep does not vary with varying amounts of α-CD. Degradation of the Enzymatically Cross-Linked Hydrogels. To evaluate the stability of the hydrogels when immersed in a solution, the wet mass of several hydrogel compositions was followed over time. Figure 8 shows the result of a degradation study for F68-Tyr- and F127-Tyr-based hydrogels that were immersed in PBS at 37 °C. Self-assembled PPR hydrogels described in previous work8 cannot be handled and weighed due to their sensitivity to shear, especially during the dissolution process. However, similar size cylinders of the self-assembled PPRs were found to dissolve after 4 to 6 h under the conditions tested (visual assessment, data not shown). It can be seen from Figure 8 that the end-to-end coupling of the F68-Tyr or F127-Tyr molecules (which creates long F68 or F127 multiunit chains throughout the gel but not a covalently cross-linked network) using H2O2 and HRP within the selfassembled hydrogels increased the hydrogel lifetime to 3 days. Dissolution is still, however, expected to be the dominant mechanism for the loss of structure in these systems. The introduction of PEG-Tyr, which creates a covalently crosslinked network gel, slows the loss of gel structure, causing the gels with the high ratio of PEG-Tyr to remain stable for 8 days. The introduction of the PEG-Tyr will inhibit the dissolution of the PPR units and hence any loss of gel structure or mass is now likely through the degradation of the ester bonds introduced during the functionalization of the polymer with Tyr (discussed further below). The evolution of the mass of the hydrogels follows the same overall pattern but with different time frames, depending on the composition: swelling of the hydrogels (up to 200%) is observed over for up to 8 days before complete degradation of the hydrogels, usually within one subsequent day. The degree of swelling compared to the initial mass is more pronounced and statistically significant for most of the F68-Tyr compared to the F127-Tyr-based gels. Changing the PBS solution at days 3 and 6 for the gels still present at those time points seems to increase the degree of swelling, which is likely to influence the degradation process. The presence of α-CD appears to accelerate the degradation: for the same composition of Pluronic and PEG-Tyr, hydrogels containing α-CD are generally fully degraded 1 to 2 days before the hydrogels with no α-CD. A possible explanation for this effect is that the presence of α-CD on the Pluronic backbone

linked gels over the range of mechanical properties achieved by the hydrogels (Figure 5) confirm the attainment of a frequencyindependent hydrogel state.

Figure 5. Representative frequency sweeps. (A) F68-Tyr-based hydrogels: squares, F68-Tyr-α8-L; circles, F68-Tyr-α10-M; and triangles, F68-Tyr-α12-H. (B) F127-Tyr-based hydrogels: squares, F127-Tyr-α8-L; circles, F127-Tyr-α10-M; and triangles, F127-Tyr-α12H at 37 °C. Strain: 0.1%.

Figure 6 shows the mechanical properties of the gels crosslinked by the method described and containing various amounts of PEG-Tyr, α-CD, and F68-Tyr or F127-Tyr. It can be seen that both the elastic and loss moduli can be tuned by changing the amounts of PEG-Tyr and α-CD. For both F68Tyr and F127-Tyr, increasing the ratio of PEG-Tyr to Pluronic led to higher mechanical properties of the resulting gels, both with and without α-CD. Increasing the coverage of α-CD also led to an increase in the elastic modulus compared to that of gels with no α-CD. This increase was always statistically significant for coverages of 10 and 12% compared to no α-CD. For F127-Tyr-α8, which did not form a gel before the addition of HRP and H2O2, an increase in elastic and viscous moduli was observed for the resulting gel relative to the gel in which no αCD was present, although it is only significant for G″. Overall, the elastic modulus G′ ranged between 50 and 410 kPa for F68Tyr-based gels and between 20 and 180 kPa for F127-Tyrbased gels, and the loss modulus G″ ranged between 150 Pa and 22 kPa for F68-Tyr-based gels and between 800 Pa and 10 kPa for F127-Tyr-based gels. The creep response of the cross-linked hydrogels to the application of a constant stress of 1000 Pa was next assessed. Cells have been found to exert traction forces on their environment. In 2D, the average stress exerted by a single G


Biomacromolecules

Article

Figure 6. Elastic modulus G′ and loss modulus G″ for F68-Tyr (A, B) and F127-Tyr (C, D) based hydrogels containing various amounts of α-CD and PEG-Tyr. L, one PEG-Tyr per eight Pluronic molecules; M, one PEG-Tyr for four Pluronic molecules; and H, one PEG-Tyr for two Pluronic molecules. The number after α indicates the percent theoretical coverage. Mean of triplicates + standard error of the mean. Significance (two-way ANOVA): *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001; ****, p ≤ 0.0001.

Figure 7. Creep testing of the PR hydrogels. (A) Representative creep of PR gels at 1000 Pa applied shear stress with varying amounts of PEG-Tyr: circles, F68-Tyr-α12-L; squares, F68-Tyr-α12-M; triangles, F68-Tyr-α12-H. (B) Influence of PEG-Tyr on the average rate of creep. Mean of triplicates + standard error of the mean. (C) Representative creep of PR gels at 1000 Pa applied shear stress with varying amounts of α-CD: diamonds, F68Tyr-M; circles, F68-Tyr-α8-M; triangles, F68-Tyr-α10-M; squares, F68-Tyr-α12-M. (D) Influence of α-CD on the average rate of creep. Mean of triplicates + standard error of the mean.

H


Biomacromolecules

Article

renders the pseudopolyrotaxanes more hydrophilic than the unthreaded Pluronic due to the presence of the hydroxyl groups on α-CD, increasing the local density of interactions between the polymers and water (i.e., the overall hydration state) and thus accelerating the rate of hydrolysis. The tyramine-functionalized polymers of this study possess an ester group. To assess whether the degradation of the hydrogels was due to the hydrolysis of the ester group, the supernatant resulting from the degradation of F68-Tyr-L, F68Tyr-M, and F68-Tyr-H was dialyzed against water to remove fragments smaller than 3500 Da before being analyzed by NMR. Figure 9 shows the comparison of the degradation products of F68-Tyr-L (Figure 9C) with nonfunctionalized eight-arm PEG (Figure 9A) and F68 (Figure 9B). The NMR peaks observed for the F68-Tyr-L degradation products match the peaks observed for the nonfunctionalized eight-arm PEG and F68. None of the peaks observed for the functionalized F68 and eight-arm PEG (Supporting Information Figures 1 and 3) can be observed, showing that the degradation occurs at the ester bond and that the degradation products are the initial unmodified F68 and eight-arm PEG. The evaluation of the degradation products of F68-Tyr-M and F68-Tyr-H yields the same result (Supporting Information Figures 4 and 5). Model Drug Release from Tyramine-Based Gels. To assess whether the tyramine-based hydrogels could provide support for the sustained delivery of poorly water-soluble drugs, a model molecule, 6-aminofluorescein, was encapsulated within the hydrogels of various compositions. The release profile of the molecule is shown Figure 10. In the absence of PEG-Tyr, the release of 6-aminofluorescein was rapid (60 h) and with a potentially undesirable (depending on the actual target drug used) burst release after 48 h, causing up to 70% of the 6aminofluorecein to be released within 15 h. The addition of

Figure 8. Wet mass profile of tyramine-functionalized hydrogels immersed in PBS at 37 °C. (A) F68-Tyr-based hydrogels. (B) F127Tyr-based hydrogels. The last bar for each sample represents the last day before the sample was fully dissolved. Mean of triplicates + standard error of the mean. The asterisks show samples that have a mass that is statistically significant (two-way ANOVA) compared to the initial (day 0) mass (p ≤ 0.05).

Figure 9. 1H NMR Spectra of (A) eight-arm PEG, (B) F68, and (C) degradation products from F68-Tyr-L. I


Biomacromolecules

Article

PEG-Tyr led to a prolonged and sustained release that lasted up to 15 days for F68-Tyr-based gels and 17 days for F127-Tyrbased gels. To determine whether the presence of cyclodextrin affects the rate of release of the 6-aminofluorescein, various amounts of α-CD were incorporated. For the F68-Tyr-based gels, the release profile did not differ between the α-CD-containing hydrogels and the hydrogels without α-CD. For the F127-Tyrbased gels, the release was slower in the absence of α-CD (statistically different (p ≤ 0.05) for F127-Tyr-H compared to F127-Tyr-α8-H, F127-Tyr-α10-H, and F127-Tyr-α12-H between 135 and 238 h); however, a similar steady release was obtained both with and without the presence of α-CD. Overall, the covalently cross-linked tyramine-based hydrogels studied here provided a satisfactory platform for the sustained delivery of poorly water-soluble molecules over a period of 2 weeks. Cell Encapsulation in the Enzymatically Cross-Linked Hydrogels Containing PEG-Tyr. To evaluate the suitability of the covalently cross-linked gels for short-term cell encapsulation, mouse 3T3 fibroblast were encapsulated within the hydrogels, and a live/dead staining was performed after 24 h. Figures 11 and 12 show that the majority of cells are alive after 24 h for all of the conditions tested. Quantification of the live and dead cells shows between 74 and 83% live cells for all of the F68-Tyr-based gels and between 78 and 88% live cells for F127-Tyr, with no marked difference between the gels containing α-CD and gels without α-CD. Additionally, with

Figure 10. Release profile of 6-aminofluorescein from the hydrogels for various compositions. (A) F68-Tyr-based hydrogels. (B) F127-Tyrbased hydrogels. Mean of triplicates ± standard error of the mean. Lines connecting data points are provided to guide the eye.

Figure 11. Live/dead staining and quantification of viability (mean of duplicates + range) of 3T3 cells encapsulated within the F68-Tyr hydrogels after 24 h. Live cells are shown in green; dead cells are shown in red/yellow. L, one PEG-Tyr per eight Pluronic molecules; M, one PEG-Tyr for four Pluronic molecules; and H, one PEG-Tyr for two Pluronic molecules. The number after α indicates the percent theoretical coverage. Scale bar: 100 μm. J


Biomacromolecules

Article

Figure 12. Live/dead staining and quantification of viability (mean of duplicates + range) of 3T3 cells encapsulated within the F127-Tyr hydrogels after 24 h. Live cells are shown in green; dead cells are shown in red/yellow. L, one PEG-Tyr per eight Pluronic molecules; M, one PEG-Tyr for four Pluronic molecules; and H, one PEG-Tyr for two Pluronic molecules. The number after α indicates the percent theoretical coverage. Scale bar: 100 μm.

undesirable for biological applications since high local concentrations of α-CD could potentially cause cytotoxicity.22,23 In order to stabilize the pseudopolyrotaxane-based hydrogels presented in our previous work,8 a covalent crosslinking function was thus introduced within the hydrogels. Although some of the advantages of the self-assembly, such as the ability to form a hydrogel purely through physical crosslinks in the absence of chemicals or other stimuli, are lost, this method is a necessary step toward the creation of PR hydrogels that can be used for tissue engineering or drug delivery applications. The method of cross-linking chosen is based on the coupling of two phenolic moieties based on a peroxidase-catalyzed oxidation. The phenol moieties were introduced into both polymer types (Pluronic and branched PEG) via first reacting the hydroxyl end group on each chain with succinic anhydride and thereafter adding a tyramine end group using a DCC/NHS coupling reaction. The cross-linking of the phenolic groups introduced at the Pluronic end groups that constitute the

increasing amounts of PEG-Tyr, the concentration of H2O2 was increased from 15 to 45 mM for F68-Tyr and from 11.7 to 36 mM for F127-Tyr; however, this increase did not lead to an increase in cell death, showing that even at these levels of H2O2 the rate of consumption (and conversion to H2O) by HRP is high enough to ensure that the H2O2 is always below a toxic threshold.

■

DISCUSSION Self-assembled PPR hydrogels present numerous attractive features, including tunable gelation and self-healing properties. However, due to their self-assembled nature, they are more sensitive to environmental factors than covalently cross-linked systems and can dissociate rapidly if their equilibrium state is disrupted. Rapid dissolution of these metastable hydrogels can lead to the release of undesirable toxic products. In the case of PPR hydrogels, upon dissolution of the gel network and dissociation of the PPR into Pluronic and CD components, high local concentrations of α-CD may result, which is certainly K


Biomacromolecules

Article

variations in components is initially somewhat puzzling, especially considering the self-assembled nature of the components of these systems and, in particular, the possible contributions of assemblages of CD aggregates that have a degree of both rotational and translational freedom to a creep response. Creep is a function of the architecture of the system, which, in turn, is reflected in the relative ratios and magnitudes of the loss and elastic modulus. At a constant elastic modulus, increases in loss modulus will result in increased creep response due to the increased dissipative architectures throughout the system. Although an increase in the amount of PEG-Tyr leads to an increase in the loss modulus G″, there is also a respective increase in the elastic modulus. Furthermore, the relative ratio of change for each of these moduli for increasing amounts of PEG-Tyr is similar. The same situation exists for increasing amounts of α-CD. From the results, it appears that the elastic modulus is playing a more dominant role in the response of the material to a constant stress than any changes in dissipation throughout these self-assembled cross-linked networks. In addition to having tunable mechanical properties, the hydrogels presented here also have tunable degradation properties. The introduction of the covalent cross-linking successfully increased the stability of the PPR hydrogels. Compared to the physical hydrogels that degrade after a few hours when immersed in PBS, the enzymatically cross-linked hydrogels lasted for up to 8 days. The method of functionalization chosen for the coupling of tyramine to the polymers introduces an ester bond between the polymer backbone and the end group. The presence of this ester bond creates a hydrolytically degradable hydrogel network. The rate of hydrolysis of ester bonds is highly dependent on the surrounding chemical environment.60,61 Moreover, the rate of degradation of an ester-containing hydrogel depends on the number of ester groups present in the hydrogel. Hydrogels containing ester bonds have previously been reported with varying degradation kinetics. Poly(vinyl alcohol)-based hydrogels have been shown to degrade in 12 to 44 days, depending on the number of ester cross-links in the network.62 Similarly, the degradation of PEG hydrogels containing ester bonds has been modulated from a few hours to 6 days via changing the number of ester bonds present in the network as well as the local chemical composition surrounding the ester bond.63 In the present study, the rate of degradation was tuned by changing the number of eight-arm PEG molecules present in the hydrogels, which effectively is equivalent to increasing the number of ester groups in the system. Although not tested quantitatively, it is expected that the mechanical properties of the hydrogels will decrease as hydrolytic degradation of the network progresses. The hydrolytically degradable hydrogels presented here offer applications in areas such as drug or cell delivery, where a release spanning over a few days to a week is desirable. Delivery systems capable of prolonging the release of drugs are of high importance, especially for poorly water-soluble drugs that might precipitate once injected, thus reducing their bioavailability. Hydrogels have emerged as an attractive class of delivery systems, mostly due to the possibility to tailor the release rate of molecules by varying the physical (e.g., crosslinking density) or chemical (interactions with the drug) parameters.64 Purely self-assembled PPR systems have been investigated as drug release systems, but the rapid dissociation of the self-assembled network of α-CD and PEO leads to a fast release of encapsulated dextran-FITC.9 Previous attempts to

backbone of the PPRs results in long multiunit chains spanning multiple PPR aggregates, whereas the introduction of the phenol-functionalized eight-arm PEG permits the formation of a branched covalently cross-linked network, aiming at increasing the stability of these gels. The eight-arm PEG also serves as an end-capping group for the α-CD/Pluronic pseudopolyrotaxanes, converting them into polyrotaxanes (PRs) and preventing dethreading until the introduced crosslinks are degraded. This end capping was thus expected to substantially slow the dethreading of the α-CD molecules from the Pluronic backbone. The hydrogels detailed in this study display a wide range of mechanical properties that can be tuned as required for the intended application. Previous work8 has established that the mechanics of PPR hydrogels can be tuned by varying the type of Pluronic and the coverage of α-CD. These two parameters still act as tuning agents in the covalently cross-linked PR hydrogels, with an additional parameter being the amount of PEG-Tyr. The kinetics of gelation of the enzymatically crosslinked hydrogels is too rapid to allow for the whole PPR network to reform after shearing the PPR hydrogels, thus leading to overall lower mechanical properties than with the hydrogels formed with physical PPR gels, except for F127-Tyrα8, which does not form a gel in the absence of covalent crosslinking. For the same α-CD coverages as those studied here, the physical F68 PPR gels had an elastic modulus ranging between 450 kPa and 1.2 MPa,8 whereas covalently cross-linked gels have an elastic modulus between 50 and 410 kPa. For F127, the PPR gels have an elastic modulus between 46 and 250 kPa,8 whereas the covalently cross-linked gels range between 20 and 180 kPa. These hydrogels are easily tuned to possess different mechanical properties by simply varying the amounts of PEGTyr and α-CD, and although the effect of H2O2 and HRP were not specifically investigated on these PR gels, it is expected that the concentrations of HRP and H2O2 could also serve as tuning parameters26 since they affect the mechanical properties of F68Tyr-0 and F127-Tyr-0. The mechanical properties of the cross-linked PR hydrogels reported here are still very high compared to those of other published studies on covalently cross-linked PR-containing hydrogels, which have reached a maximum G′ of only ∼1000 Pa.35 Other studies on hybrid PR−chitosan enzymatically crosslinked hydrogels reported a compressive modulus of 20 to 45 kPa.59 The higher mechanical properties achieved in this current work are most likely due to the fact that, in the present system, the PPR network remains, and its influence can still be seen on the mechanical properties. Indeed, the mechanical properties without α-CD present in the covalently cross-linked hydrogels are lower than when α-CD is present and increase with increasing α-CD concentration (the elastic modulus of the covalent hydrogels increases up to 5 times through the addition of α-CD). The creep response of the gels, or the amount of timedependent deformation under constant applied stress, which has been shown to influence stem cell behavior,49,50 was also studied. For a given α-CD coverage, increasing the amount of PEG-Tyr in the hydrogels leads to seemingly lower rates of creep; however, they were not statistically different. In addition, variations in the amounts of α-CD, at constant F68-Tyr and PEG-Tyr, did not have a significant impact on the rate of creep. While the order of magnitude of creep for these systems is substantial, and certainly enough to elicit changes in stem cell behavior, based on previous work,49,50 the lack of change with L


Biomacromolecules

Article

presence of interactions between the α-CD molecules that originate from the initial PPR hydrogel formation. From a tissue engineering perspective, this property has the potential to generate interesting and unique cellular responses, where cells could modulate their local environment by modifying the physically assembled part of the hydrogels through spreading and migration, whereas the covalent network could provide a more durable scaffold. It has been shown previously that remodeling of the cellular microenvironment in hydrogels through cell traction and degradation could influence hMSC behavior.67 With the possibility to remodel the physically associated components of the hydrogel while maintaining the structure of the covalent network, the PR hydrogels presented here could offer a new tool to study this effect and its impacts on stem cell biology.

slow the release of drugs from self-assembled PPR hydrogels were based on the use of a block copolymer with a hydrophobic component (PEO−poly[(R)-3-hydroxybutyrate]−PEO) to help maintain the cohesion of the hydrogels. This resulted in a slower release that lasted for around 25 days.14 In the present study, the release of a poorly water-soluble molecule from the covalently cross-linked PR hydrogels was shown to be sustained over a period of around 2 weeks. The mechanism of release of 6-aminofluorescein from the hydrogels is likely to be due to erosion of the hydrogels since hydrogel degradation was observed concomitantly to the release of the molecule. Although it might seem contradictory that the drug release lasted for 2 weeks when the hydrogels were fully degraded after 8 days during the degradation experiment, the regular change of the PBS surrounding the hydrogels during the degradation study likely acted as an accelerating factor for the degradation when compared to the drug release study, where the same PBS solution was kept for the entire study. As the gels degrade over the 2 week period, free α-CD is released in the solution. Cyclodextrins have been shown to be able to complex hydrophobic drugs, thus improving their solubility, stability, and bioavailability.65 Consequently, the simultaneous release of α-CD and the encapsulated molecule from the gels could lead to the formation of inclusion complexes, which could further enhance the bioavailability of the molecule being delivered. Further investigation is required to confirm such utility. The hydrogels developed here can also be used for cell encapsulation: mouse 3T3 fibroblast cells remained viable for at least 24 h in the tyramine hydrogels, including the hydrogels containing α-CD, which shows that cytotoxic amounts of α-CD are not released from the hydrogels. Additionally, despite the seemingly high absolute concentrations of H2O2 used to crosslink the hydrogels (up to 45 mM, compared to 5−10 mM of H2O2 being cytotoxic for mammalian cells66), the viability of cells encapsulated within the hydrogels remains high, which shows that the H2O2 is consumed rapidly during the crosslinking reaction. More important than the absolute concentration of H2O2 in the hydrogel, the ratio of H2O2 molecules to tyramine groups is likely to be the determinant factor in the assessment of the cytotoxicity of the hydrogels. Park et al.46 previously noted that the viability of an osteoblast cell line (MC3T3-E1) encapsulated in hydrogels formed from enzymatically cross-linked tetronic (a 4-arm PEO−PPO block copolymer) was compromised when the molar ratio of tyramine to H2O2 was approximately 1:6. They also found that for a molar ratio of tyramine to H2O2 lower or equal to 1:1.2 the viability of the cells was maintained. In the present study, the highest molar ratio of tyramine groups to H2O2 is 1:0.75, which is below the threshold observed by Park et al. as being safe for cells. Although different cell types have different levels of tolerance toward chemicals, this is an indication that the concentrations of H2O2 used here are suitable for cell encapsulation, as long as the amount of tyramine groups to be cross-linked is sufficiently high to ensure the rapid consumption of the H2O2 molecules by HRP in the catalytic process. The covalently cross-linked hydrogels present a lower sensitivity to shear than purely self-assembled PPR hydrogels and are therefore easier to handle. Using PPRs as precursors for the hydrogels offers the possibility for further functionalization through the hydroxyl groups present on the CDs. Moreover, the use of PPR allows us to obtain hydrogels with dual properties: the covalent network formed through the enzymatic cross-linking of the tyramine end groups is reinforced by the

■

CONCLUSIONS A covalent network of enzymatically cross-linked PRs has been successfully developed, resulting in stable and tunable hydrogels. The hydrogels are hydrolytically degradable in 2−8 days due to the presence of ester bonds, making them suitable for short-term cell encapsulation (for example, for cell delivery) and for sustained release of poorly water-soluble molecules. The mechanical properties of the gels show signatures of both the cross-linked network and the self-assembled network since they can be tuned by changing the amounts of eight-arm PEG and α-CD, showing that both elements participate in the network. This dynamic double network, in which covalently bound PRs are still able to interact via physical bonds at the molecular level and participate in increasing the mechanical properties of the hydrogels, could see applications in tissue engineering, where its distinctive properties could elicit unique cellular responses.

■

ASSOCIATED CONTENT

S Supporting Information *

Diffusion 1H NMR spectra of the tyramine-modified polymers, NMR spectra of the degradation product of the hydrogels, live/ dead staining of 3T3 cells encapsulated within the F127-Tyr or F68-Tyr hydrogel after 24 h. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by Australian Research Council Discovery grant nos. DP1095429 and DP110104446. NMR experiments were conducted at the Centre for Advanced Imaging at the University of Queensland. The confocal imaging was performed at the Queensland node of the Australian National Fabrication Facility, a company established under the National Collaborative Research Infrastructure Strategy to provide nano- and microfabrication facilities for Australia’s researchers.

■

ABBREVIATIONS CD, cyclodextrin; CTMS, chlorotrimethylsilane; DCC, N,N′dicyclohexylcarbodiimide; DMAP, 4-dimethylaminopyridine; M


Biomacromolecules

Article

(28) Jun, I.; Kim, S. J.; Lee, J.-H.; Lee, Y. J.; Shin, Y. M.; Choi, E.; Park, K. M.; Park, J.; Park, K. D.; Shin, H. Adv. Funct. Mater. 2012, 22, 4060−4069. (29) Kim, D. W.; Jun, I.; Lee, T.-J.; Lee, J. h.; Lee, Y. J.; Jang, H.-K.; Kang, S.; Park, K. D.; Cho, S.-W.; Kim, B.-S.; Shin, H. Biomaterials 2013, 34, 8258−8268. (30) Kim, S. J.; Jun, I.; Kim, D. W.; Lee, Y. B.; Lee, Y. J.; Lee, J.-H.; Park, K. D.; Park, H.; Shin, H. Biomacromolecules 2013, 14, 4309− 4319. (31) Lee, F.; Chung, J. E.; Kurisawa, M. Soft Matter 2008, 4, 880− 887. (32) Wang, L.-S.; Lee, F.; Lim, J.; Du, C.; Wan, A. C. A.; Lee, S. S.; Kurisawa, M. Acta Biomater. 2014, 10, 2539−2550. (33) Toh, W. S.; Lim, T. C.; Kurisawa, M.; Spector, M. Biomaterials 2012, 33, 3835−3845. (34) Frith, J. E.; Cameron, A. R.; Menzies, D. J.; Ghosh, P.; Whitehead, D. L.; Gronthos, S.; Zannettino, A. C. W.; Cooper-White, J. J. Biomaterials 2013, 34, 9430−9440. (35) Tran, N.; Joung, Y.; Lih, E.; Park, K.; Park, K. Macromol. Res. 2011, 19, 300−306. (36) Lih, E.; Lee, J. S.; Park, K. M.; Park, K. D. Acta Biomater. 2012, 8, 3261−3269. (37) Wang, L.-S.; Chung, J. E.; Pui-Yik Chan, P.; Kurisawa, M. Biomaterials 2010, 31, 1148−1157. (38) Wang, L.-S.; Du, C.; Chung, J. E.; Kurisawa, M. Acta Biomater. 2012, 8, 1826−1837. (39) Wang, L.-S.; Du, C.; Toh, W. S.; Wan, A. C. A.; Gao, S. J.; Kurisawa, M. Biomaterials 2014, 35, 2207−2217. (40) Hwang, J. H.; Kim, I. G.; Piao, S.; Jung, A. R.; Lee, J. Y.; Park, K. D.; Lee, J. Y. Biomaterials 2013, 34, 6037−6045. (41) Park, K. M.; Lee, Y.; Son, J. Y.; Bae, J. W.; Park, K. D. Bioconjugate Chem. 2012, 23, 2042−2050. (42) Lee, Y.; Bae, J. W.; Oh, D. H.; Park, K. M.; Chun, Y. W.; Sung, H.-J.; Park, K. D. J. Mater. Chem. B 2013, 1, 2407−2414. (43) Gülden, M.; Jess, A.; Kammann, J.; Maser, E.; Seibert, H. Free Radical Biol. Med. 2010, 49, 1298−1305. (44) Jun, I.; Park, K.; Lee, D.; Park, K.; Shin, H. Macromol. Res. 2011, 19, 911−920. (45) Lee, J.; Jun, I.; Park, H.-J.; Kang, T. J.; Shin, H.; Cho, S.-W. Biomacromolecules 2013, 15, 361−372. (46) Park, K. M.; Shin, Y. M.; Joung, Y. K.; Shin, H.; Park, K. D. Biomacromolecules 2010, 11, 706−712. (47) Hermanson, G. T. Zero-length crosslinkers. In Bioconjugate Techniques, 2nd ed.; Academic Press: Boston, MA, 2008; Chapter 3, pp 213−233. (48) Harada, A.; Kamachi, M. Macromolecules 1990, 23, 2821−2823. (49) Cameron, A. R.; Frith, J. E.; Cooper-White, J. J. Biomaterials 2011, 32, 5979−5993. (50) Cameron, A. R.; Frith, J. E.; Gomez, G. A.; Yap, A. S.; CooperWhite, J. J. Biomaterials 2014, 35, 1857−1868. (51) du Roure, O.; Saez, A.; Buguin, A.; Austin, R. H.; Chavrier, P.; Siberzan, P.; Ladoux, B. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 2390− 2395. (52) Legant, W. R.; Miller, J. S.; Blakely, B. L.; Cohen, D. M.; Genin, G. M.; Chen, C. S. Nat. Methods 2010, 7, 969−971. (53) Balaban, N. Q.; Schwarz, U. S.; Riveline, D.; Goichberg, P.; Tzur, G.; Sabanay, I.; Mahalu, D.; Safran, S.; Bershadsky, A.; Addadi, L.; Geiger, B. Nat. Cell Biol. 2001, 3, 466−472. (54) Nakajima, R.; Yamazaki, I. J. Biol. Chem. 1987, 262, 2576−2581. (55) Arnao, M. B.; Acosta, M.; del Rio, J. A.; Varón, R.; GarcíaCánovas, F. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 1990, 1041, 43−47. (56) Mao, L.; Luo, S.; Huang, Q.; Lu, J. Sci. Rep. 2013, 3, 3126. (57) Baynton, K. J.; Bewtra, J. K.; Biswas, N.; Taylor, K. E. Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol. 1994, 1206, 272−278. (58) Legant, W. R.; Chen, C. S.; Vogel, V. Integr. Biol. 2012, 4, 1164− 1174. (59) Tran, N. Q.; Joung, Y. K.; Lih, E.; Park, K. M.; Park, K. D. Biomacromolecules 2010, 11, 617−625.

DMEM, Dulbecco’s modified Eagle’s medium; DMSO, dimethyl sulfoxide; DO, 1,4-dioxane; FBS, fetal bovine serum; G′, oscillatory shear elastic modulus; G″, oscillatory shear viscous modulus; H, high; H2O2, hydrogen peroxide; HRP, horseradish peroxidase; L, low; M, medium; MTT, 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; NHS, N-hydroxysuccinimide; NMR, nuclear magnetic resonance; PBS, phosphate buffer saline; PEG, poly(ethylene glycol); PEO, poly(ethylene oxide); PPO, poly(propylene oxide); PPR, pseudopolyrotaxane; PR, polyrotaxane; SA, succinic anhydride; TEA, triethylamine; THF, tetrahydrofuran; Tyr, tyramine

■

REFERENCES

(1) Whitesides, G. M.; Boncheva, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4769−4774. (2) Bishop, K. J. M.; Wilmer, C. E.; Soh, S.; Grzybowski, B. A. Small 2009, 5, 1600−1630. (3) Harada, A.; Kobayashi, R.; Takashima, Y.; Hashidzume, A.; Yamaguchi, H. Nat. Chem. 2011, 3, 34−37. (4) Qi, H.; Ghodousi, M.; Du, Y.; Grun, C.; Bae, H.; Yin, P.; Khademhosseini, A. Nat. Commun. 2013, 4, 3275. (5) Ceccato, M.; Lo Nostro, P.; Baglioni, P. Langmuir 1997, 13, 2436−2439. (6) Lo Nostro, P.; Lopes, J. R.; Cardelli, C. Langmuir 2001, 17, 4610−4615. (7) Travelet, C.; Schlatter, G.; Hébraud, P.; Brochon, C.; Lapp, A.; Hadziioannou, G. Langmuir 2009, 25, 8723−8734. (8) Pradal, C.; Jack, K. S.; Grøndahl, L.; Cooper-White, J. J. Biomacromolecules 2013, 14, 3780−3792. (9) Li, J.; Ni, X.; Leong, K. W. J. Biomed. Mater. Res., Part A 2003, 65, 196−202. (10) Ma, D.; Zhang, L.-M. Biomacromolecules 2011, 12, 3124−3130. (11) Li, X.; Li, J. J. Biomed. Mater. Res., Part A 2008, 86, 1055−1061. (12) Ma, D.; Zhang, L.-M.; Xie, X.; Liu, T.; Xie, M.-Q. J. Colloid Interface Sci. 2011, 359, 399−406. (13) Zhu, W.; Li, Y.; Liu, L.; Chen, Y.; Xi, F. Int. J. Pharm. 2012, 437, 11−19. (14) Li, J.; Li, X.; Ni, X.; Wang, X.; Li, H.; Leong, K. W. Biomaterials 2006, 27, 4132−4140. (15) Ma, D.; Zhang, H.-B.; Tu, K.; Zhang, L.-M. Soft Matter 2012, 8, 3665−3672. (16) Simões, S. M. N.; Veiga, F.; Torres-Labandeira, J. J.; Ribeiro, A. C. F.; Sandez-Macho, M. I.; Concheiro, A.; Alvarez-Lorenzo, C. Eur. J. Pharm. Biopharm. 2012, 80, 103−112. (17) Li, J.; Yang, C.; Li, H.; Wang, X.; Goh, S. H.; Ding, J. L.; Wang, D. Y.; Leong, K. W. Adv. Mater. 2006, 18, 2969−2974. (18) Yang, C.; Wang, X.; Li, H.; Tan, E.; Lim, C. T.; Li, J. J. Phys. Chem. B 2009, 113, 7903−7911. (19) Ma, D.; Zhang, H.-B.; Chen, D.-H.; Zhang, L.-M. J. Colloid Interface Sci. 2011, 364, 566−573. (20) Alexandridis, P.; Holzwarth, J. F.; Hatton, T. A. Macromolecules 1994, 27, 2414−2425. (21) Yancey, P. G.; Rodrigueza, W. V.; Kilsdonk, E. P. C.; Stoudt, G. W.; Johnson, W. J.; Phillips, M. C.; Rothblat, G. H. J. Biol. Chem. 1996, 271, 16026−16034. (22) Leroy-Lechat, F.; Wouessidjewe, D.; Andreux, J.-P.; Puisieux, F.; Duchêne, D. Int. J. Pharm. 1994, 101, 97−103. (23) Irie, T.; Uekama, K. J. Pharm. Sci. 1997, 86, 147−162. (24) Kurisawa, M.; Chung, J. E.; Yang, Y. Y.; Gao, S. J.; Uyama, H. Chem. Commun. 2005, 34, 4312−4314. (25) Kurisawa, M.; Lee, F.; Wang, L.-S.; Chung, J. E. J. Mater. Chem. 2010, 20, 5371−5375. (26) Menzies, D. J.; Cameron, A.; Munro, T.; Wolvetang, E.; Grøndahl, L.; Cooper-White, J. J. Biomacromolecules 2012, 14, 413− 423. (27) Park, K. M.; Jun, I.; Joung, Y. K.; Shin, H.; Park, K. D. Soft Matter 2011, 7, 986−992. N


Biomacromolecules

Article

(60) Rydholm, A. E.; Anseth, K. S.; Bowman, C. N. Acta Biomater. 2007, 3, 449−455. (61) Schoenmakers, R. G.; van de Wetering, P.; Elbert, D. L.; Hubbell, J. A. J. Controlled Release 2004, 95, 291−300. (62) Martens, P. J.; Bowman, C. N.; Anseth, K. S. Polymer 2004, 45, 3377−3387. (63) Zustiak, S. P.; Leach, J. B. Biomacromolecules 2010, 11, 1348− 1357. (64) Hoare, T. R.; Kohane, D. S. Polymer 2008, 49, 1993−2007. (65) Brewster, M. E.; Loftsson, T. Adv. Drug Delivery Rev. 2007, 59, 645−666. (66) Gardner, A. M.; Xu, F.-h.; Fady, C.; Jacoby, F. J.; Duffey, D. C.; Tu, Y.; Lichtenstein, A. Free Radical Biol. Med. 1997, 22, 73−83. (67) Khetan, S.; Guvendiren, M.; Legant, W. R.; Cohen, D. M.; Chen, C. S.; Burdick, J. A. Nat. Mater. 2013, 12, 458−465.

O


Hydrolytically Degradable Polyrotaxane Hydrogels for Drug and Cell

Recommend Documents