Vinh X.
Truong
a,
Matthew P.
Ablett
b,
Hamish T. J.
Gilbert
b,
James
Bowen
c,
Stephen M.
Richardson
*b,
Judith A.
Hoyland
*b and
Andrew P.
Dove
*a
aDepartment of Chemistry, University of Warwick, Coventry, CV4 7AL, UK. E-mail: a.p.dove@warwick.ac.uk; Tel: +44 (0)24 7652 4107
bCentre for Tissue Engineering and Repair, Institute of Inflammation and Repair, Faculty of Medical and Human Sciences, The University of Manchester, Manchester, M13 9PT, UK. E-mail: Judith.hoyland@manchester.ac.uk; Tel: +44 (0)161 275 5425
cSchool of Chemical Engineering, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
First published on 4th October 2013
A water-soluble azide-functionalised chitosan was crosslinked with propiolic acid ester-functional poly(ethylene glycol) using copper-free click chemistry. The resultant hydrogel materials were formed within 5–60 min at 37 °C and resulted in mechanically robust materials with tuneable properties such as swelling, mechanical strength and degradation. Importantly, the hydrogels supported mesenchymal stem cell attachment and proliferation and were also non-toxic to encapsulated cells. As such these studies indicate that the hydrogels have potential to be used as injectable biomaterials for tissue engineering.
It displays excellent biocompatibility in addition to other unique properties such as being hemostatic, mucoadhesive and having high affinity to anionic biologically active molecules.1,2 Consequently, chitosan is an attractive precursor for the preparation of hydrogels for a range of biomedical applications such as wound healing,3,4 drug delivery5,6 and tissue engineering.7,8
Chitosan-based hydrogels have been widely studied, largely on account of their ability to form thermally-responsive, injectable gels that are capable of directing the differentiation of human mesenchymal stem cells (MSCs).9 Such gels are prepared from chitosan and glycerophosphate9–11 or PEG-grafted-chitosan,12–14 which remain liquid at ambient temperature and form gels within 10–15 minutes at 37 °C. Other injectable chitosan-based hydrogels have been reported using pH-responsive nanostructure transformations15,16 in which gelation occurs when the precursor solution is subjected to a change in pH. However, despite their attractive properties for biological applications, these hydrogels possess poor biomechanical properties and fast degradation rates that result from weak physical or ionic crosslinking interactions within the gel, thus limiting their potential for tissue engineering applications.
Several attempts to increase the strength of chitosan-based hydrogels have been reported by covalently crosslinking chitosan precursors using a range of chemistries such as thiol–ene Michael addition,17,18 enzyme-catalysed crosslinking (horseradish peroxidase),8,19 Schiff base formation,20 and by crosslinking with the natural crosslinker genipin.21,22 While in general, these processes led to hydrogels with increased mechanical properties in comparison to the ionic/physically crosslinked networks, some of the crosslinking methods resulted in a very fast gelation process (within seconds) which coupled with a low degree of control over the gelation time8,19 limits the potential application of these systems as injectable hydrogels. To this end, we sought to identify a system that was able to produce a wide range of chitosan-based hydrogels with tuneable mechanical properties and gelation times as well as demonstrate their ability to support MSC growth.
In recent years, so called ‘click’ reactions that are fast, selective and high yielding,23,24 have been used to prepare macrostructures with diverse and complex architecture in both organic and aqueous media. Notably, a selection of click reactions have even been shown to be carried out in complex biological media and in the presence of living cells. Such reactions are termed bioorthogonal click chemistry and have been successfully used in applications such as cell labelling25 and imaging26 and cell surface modification.27 Bioorthogonal click reactions are extremely attractive for fabrication of hydrogels to be used as cell culture scaffolds because they allow cell encapsulations and at the same time provide 3D environment which mimics the extracellular environment of natural tissues.28,29 Michael addition reactions are probably the most common bioorthogonal crosslinking chemistries utilised to prepare in situ-forming hydrogels on account of the mild reaction conditions, which require only basic buffer environment, and the availability of the functional precursors.28,30 More recently, several other click reactions have been introduced to the expanding bioorthogonal click toolkit for hydrogel preparation in biological media. They include reaction of an azide with a ring-strained alkyne (SPAAC),31,32 oxime click reaction with glutaraldehyde,33 tetrazine–norbonene reaction,34 and the tetrazole–alkene click reaction.35
Among these chemistries, SPAAC is gaining widespread attention for use in preparation of in situ-forming hydrogels31,32 as a consequence of the inert virtue of the functional groups which allow its use in conjunction with other bioorthogonal click chemistries to provide precise spatial and temporal manipulation of the microenvironment within the hydrogel scaffold. However, the synthetic challenges of both the multi-step preparation of the ring-strained alkyne and its incorporation in chitosan-based gels led us to focus on the more simple click reaction between organic azides and propiolic acid esters.23,36 In common with SPAAC, this reaction occurs readily at ambient temperature and does not require the addition of any catalytic species. Herein we demonstrate the preparation of robust chitosan-poly(ethylene glycol) (PEG) hydrogels by crosslinking an azide-functionalised chitosan with a water soluble 3-arm PEG containing an activated ester alkyne group. The tuneable nature of the gelation and gel properties are also demonstrated.
In order to determine the water uptake in phosphate-buffered saline (PBS) the prepared gels were allowed to swell in PBS for 1 day so that the swelling reached equilibrium. The surface water was then removed with soft tissue paper and the weight was recorded (Ws). The gels were then lyophilized and the weight recorded (Wd). The equilibrium water uptake was expressed as:
(Ws − Wd)/Wd × 100% |
For 2D MSC culture, hydrogels were prepared in tissue culture inserts (24-well plate inserts, BD Biosciences) as described above, but under sterile conditions and transferred to a 37 °C incubator until gelation had occurred. Cells were seeded at 5 × 104 cells per cm2 onto the gels with media changes every 3–4 days. For cell encapsulation (3D culture), MSCs were resuspended in CS-azide at 4 × 106 cells per mL before mixing with PEG-alkyne and pippetting into tissue culture inserts. Once gelation had occurred, media was added to the culture inserts and wells.
To assess cell viability both in 2D and 3D, gels were removed from tissue culture inserts after either 1 or 7 days in culture, washed in PBS and stained using Live/Dead® Viability/Cytotoxicity Kit for mammalian cells (Invitrogen) following the manufacturer's recommended protocol.
For morphological analysis, gels were removed from cell inserts after 7 days in culture, washed in PBS and fixed in 4% paraformaldehyde for 15 minutes at 4 °C. Cells were treated with 0.25% Triton-X in PBS (for 5 min) and then blocked in 1% bovine serum albumin (BSA) for 1 h at room temperature. Gels were incubated with 1:100 (in 1% BSA) mouse monoclonal vinculin primary antibody (Abcam) for 1 h at 4 °C followed by goat-anti mouse secondary antibody conjugated to Alexa Fluor® 488 (Invitrogen) at 1:300 dilution for 1 h, room temperature. 1 Unit of phalloidin per sample (Alexa Fluor 594, Invitrogen) was then added at room temperature for 20 min. Gels were then incubated with a 5 μM solution of DRAQ5™ (Biostatus) at 37 °C for 15 min to visualise the nuclei, after which the gels were washed and resuspended in PBS for imaging.
Images were collected on a Leica TCS SP5 AOBS upright confocal using a 20× dipping objective. ImageJ software was used to create maximum intensity projections of each 100 μm optical stack gathered using the confocal software.
Fig. 1 1H NMR spectra (in D2O–CD3COOD, 400 MHz) (A) and IR spectra (B) of the starting materials and chitosan-azide product. |
Initial hydrogel network formation was studied in nanopure water by crosslinking CS-azide (2 wt%) and PEG-alkyne (1:1 molar ratio to CS-azide) by metal-free, low temperature, alkyne–azide cycloaddition (Scheme 1). In contrast to the more commonly-applied copper catalysed alkyne–azide cycloaddition, this process does not require elevated temperatures or metal-based catalysts as a consequence of the activation of the alkyne by the adjacent electron withdrawing group.36 Gelation of these precursors was observed within 90 minutes at ambient temperature and 15 minutes at 37 °C as assessed by the vial tilt method. Such performance in pure water solutions demonstrates that gelation occurs in this system through the desired crosslinking reaction, eliminating the possibility of ionic crosslinking. To further demonstrate the biorthogonality of this reaction however, gelation in solutions that better mimic the cellular microenvironment i.e. those that include a range of components including salts, sugars, amino acids etc., we sought to evaluate the gelation process in PBS solution and cell culture media (MEM). Remarkably, in both solutions at 37 °C, gelation was complete within approximately the same time period, validating the hypothesis that the alkyne–azide cycloaddition remains bioorthogonal.
The ability to control both gelation times and gel properties is important in the design of injectable hydrogel systems for tissue engineering and regenerative medicine applications. Ideally, mechanically-stable gels should form at a rate that enables mixing and uniform distribution of cells or proteins before injection into the body, but is sufficiently rapid to prevent leaching of the active biological components or cell sedimentation. A range of hydrogels, with the total weight percentage of polymer precursors in solution of 2–15 wt%, were prepared in PBS with increasing CS-azide concentrations from 1 to 3 wt% and [alkyne]:[azide] ratios between 0.25 to 2.5:1 with gelation times measured in triplicate using the vial tilt method (Fig. 2A). At the highest CS-azide and PEG-alkyne concentration and [alkyne]:[azide] ratio, gels were found to form within 4 minutes; however at lower concentrations of either component, gelation times were observed to be reduced significantly such that at 1 wt% CS-azide and a 0.25:1 molar ratio of [alkyne]:[azide], a period of 55 minutes was required for gelation to occur, most likely on account of the decrease in the concentration of the functional group of the gelation system.
The viscoelastic profiles of selected chitosan-PEG hydrogels during gelation were evaluated by rheological studies. The kinetics of gelation for a CS-azide 1 wt% solution with a [alkyne]:[azide] ratio of 0.5 was also followed by rheological analysis (Fig. 2B). Gelation was shown to have commenced within 1 minute at 37 °C, as characterized by the crossover point of the rheological storage (G′) and loss moduli (G′′).21 Complete gelation was observed after approximately 45 minutes which is in agreement with the gelation time observed by vial tilt method. Notably, the G′ value at the gel formation time was 44.2 ± 1.2 kPa, which indicates a robust gel network structure was formed. This value is higher than the G′ value of chitosan hydrogel formed by chitosan/glycerol phosphate system (G′ value of ca. 5 kPa)10 and other chemically crosslinked chitosan networks in which G′ values in the range of 10–20 kPa have been reported.8,17,21 Higher G′ values (87–103 kPa) at the gel point were observed with increasing CS-azide wt% (Fig. S3 in ESI†). The fast gelation time observed is the result of two factors: the elevated temperature (37 °C) used for crosslinking; and the high functionality of the CS-azide. By using the generalization of the Arrhenius’ equation, the reaction rate should be double at the elevated temperature compared to at ambient temperature which consequently increases the rate of the crosslinking. In addition, the average functionality of the CS-azide is ca. 19.5 (by estimation of the degree of substitution by 1H NMR spectroscopy), and thus gelation is proposed to occur rapidly because the critical conversion is likely to be low as a consequence of the high average functionality of the chitosan-azide.
When the viscoelastic properties of the mixture were monitored for a longer period of time, the storage modulus of the network continued to increase slowly to an upper limit before decreasing significantly with extreme fluctuation in the collected data. This phenomenon is related to the syneresis properties of the cross-linked network, which occur when the network pressure exceeds the osmotic pressure generated by water bound within the gel, which in turn leads to gel contraction and water expelled to the environment.21 In turn, the gel shrinks away from the cone of the oscillatory rheometer causing a reduction in the recorded values. Similar rheological profiles were also reported for other chitosan hydrogel systems.21,40
The gel fraction, which is the percentage of the weight of the dried gel over the total weight of the polymer precursors, was measured as an indication of the efficiency of the gelation process. In these systems, the gel fraction was found to vary from 0.6 to 0.95 (Table 1) and increase with increasing concentration of CS-azide in the precursor solution, which indicates that the crosslinking is more efficient at higher CS-azide concentrations. Moreover, the gel fraction was found to be highest with the [alkyne]:[azide] ratio of 1:1 at all CS-azide concentrations, which shows that an equimolar ratio of [alkyne]:[azide] produced the most efficient crosslinking. The gels all displayed a high water sorption capacity, measured by equilibrium water content (EWC) of the prepared CS-PEG hydrogels found to be in the range of 85–95% at ambient temperature (Fig. 3). The EWC was found to decrease with increasing CS-azide concentration which indicates that the gels formed from higher CS concentrations have a higher crosslinking density which reduces the water sorption capacity. These values are in agreement with the gel fraction values, which showed higher crosslinking efficiency at higher CS-azide concentration. In addition, the EWC was found to decrease with increasing [alkyne]:[azide] ratio, i.e. increasing the [PEG-alkyne] while keeping the [CS-azide] constant, which is also a result of the increase in crosslinking density.
Fig. 3 Equilibrium water content of CS hydrogel as a function of functionalized CS with different [alkyne]:[azide] ratios. |
CS-azide wt% | [Azide] mM | PEG-alkyne wt% | [Alkyne] mM | Gel fraction |
---|---|---|---|---|
1 | 53 | 0.5 | 13 | 0.68 ± 0.06 |
1 | 53 | 1 | 26 | 0.63 ± 0.06 |
1 | 53 | 2 | 53 | 0.84 ± 0.05 |
1 | 53 | 5 | 133 | 0.71 ± 0.13 |
2 | 106 | 1 | 26 | 0.81 ± 0.02 |
2 | 106 | 2 | 52 | 0.76 ± 0.05 |
2 | 106 | 4 | 106 | 0.90 ± 0.01 |
2 | 106 | 10 | 266 | 0.85 ± 0.09 |
3 | 159 | 1.5 | 39 | 0.89 ± 0.06 |
3 | 159 | 3 | 78 | 0.82 ± 0.01 |
3 | 159 | 6 | 159 | 0.96 ± 0.01 |
3 | 159 | 15 | 399 | 0.91 ± 0.07 |
Uniaxial compression tests were undertaken on the CS-PEG gels at concentrations of 1, 2 and 3 wt% of CS-azide (with [alkyne]:[azide] ratio of 1:1) after gelation was complete in order to evaluate their mechanical strength (Fig. 4). The gels were shown to display a rupture compression stress of 81 ± 11, 160 ± 15 and 192 ± 20 kPa and compressive modulus of 42 ± 4, 61 ± 7 and 77 ± 9 kPa respectively. The compressive strength of the CS-PEG hydrogels increased with increasing polymer precursor concentration, most likely a consequence of the CS-PEG gels formed from higher polymer concentration having higher crosslinking density, as evidenced in the swelling studies described above. It should be noted that these hydrogels, while being robust in terms of mechanical strength, have high water sorption capacity with the water contents of 97 wt%, 94 wt%, and 91 wt% respectively. These strengths and moduli values are higher than the values obtained for chitosan-PEG hydrogels by thiol–ene Michael reaction, which have the highest previously reported compressive moduli for chitosan-based hydrogels of 35–50 kPa,17,18 and comparable to the mechanical strength of the gels formed by step growth polymerization such as poly(carbonate)-PEG hydrogels formed by photo-catalyzed thiol–ene reaction, which has the compressive strength of 2–450 kPa.41 The high mechanical integrity observed from these gels is attributed to the step-growth crosslinking progress which results in highly heterogeneous network structure comprised of multiple network chains being able to distribute mechanical stress across the gel network.42
Fig. 4 Stress–strain curves under uniaxial compression for CS gels formed from different CS-azide concentrations with the [azide]:[alkyne] ratio of 1:1. |
The in vitro stability of these hydrogels in PBS solution at pH 7.4 was also investigated by examining the mass of a range of hydrogels over time (Fig. 5). Specifically, gels comprised from CS-azide concentration of 1 wt%, 2 wt% and 3 wt% and PEG-alkyne with the [alkyne]:[azide] ratio of 1:1 were studied. The mass of the gels were observed to increase until it reached a critical point where sudden mass loss and collapse of the entire cleavage of the ester bonds, which in turn reduces the network occurred. The initial increase in mass is attributed to the crosslinking density and increase the internal volume of the network, allowing more water uptake. This degradation profile is consistent with bulk-degradation profile which results from hydrolysis of the ester linkage within hydrogel networks.43,44 CS gels formed from CS 1 wt% and 2 wt% solutions completely degraded in 50 days and 60 days respectively, while gels formed from CS 3 wt% solution were stable for more than 70 days. This slower degradation rate with increasing CS wt% is consistent with the higher crosslinking density of the hydrogel network. In the presence of lysozyme, an enzyme that can hydrolyze the glycosidic bond of the acetylated unit within the chitosan chain,45 faster degradation of hydrogels was observed where CS 3 wt% gel completely degraded in 70 days while the CS 1 wt% gel and CS 2 wt% gel degraded in 45 and 50 days respectively.
Fig. 5 Degradation profiles of CS-PEG gels at different CS-azide concentrations in (A) PBS solution pH 7.4, and (B) PBS solution pH 7.4 with addition of lysozyme (0.1 mg mL−1). |
Tissue engineering scaffolds must also be able to support cellular function in vivo. As such, the ability of these hydrogels to support the growth of MSCs was assessed. MSCs were encapsulated in the gels in vitro and cell viability assessed after 24 h in culture. MSCs demonstrated high viability (>95%, Fig. 6A and B). MSCs were also seeded onto formed gels and viability and morphology assessed after 7 days in culture. The MSCs adhered to the surface of the gels and no cytotoxic response was observed, with the majority of cells remaining viable (Fig. 6C and D). Furthermore, morphological analysis of the MSCs revealed a typical fibroblastic morphology similar to that observed during in vitro culture on tissue culture polystyrene, with defined F-actin filaments and vinculin staining (important cytoskeletal components for cell adhesion and spreading) (Fig. 6E). These data confirm that the chitosan hydrogels prepared by the activated alkyne–azide cycloaddition chemistry are not cytotoxic and can support human MSCs.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3bm60159e |
This journal is © The Royal Society of Chemistry 2014 |