842
Biomacromolecules 2008, 9, 842–849
An Approach to Modulate Degradation and Mesenchymal Stem Cell Behavior in Poly(ethylene glycol) Networks Gregory A. Hudalla,† Timothy S. Eng,† and William L. Murphy*,†,‡,§ Departments of Biomedical Engineering, Pharmacology, Materials Science and Engineering, University of Wisconsin, 1550 Engineering Drive, Madison, Wisconsin 53706 Received October 24, 2007; Revised Manuscript Received December 22, 2007
A simple, sequential approach for creation of hydrolytically degradable poly(ethylene glycol) (PEG) hydrogels has been developed and characterized. The chemistry involves an initial step growth polymerization reaction between PEG-diacrylate and dithiothreitol (DTT) to form acrylate-terminated (-PEG-DTT-)n PEG chains, followed by photocross-linking to form a hydrogel network. Varying the extent of step growth polymerization prior to photocross-linking allowed for control over the equilibrium swelling ratio, degradation, and erosion of PEG hydrogels. Hydrogel degradability had a significant effect on behavior of human mesenchymal stem cells (hMSCs) encapsulated within PEG hydrogels, both in the presence and absence of an RGDSP cell adhesion ligand. In particular, enhanced network degradability resulted in enhanced hMSC viability and spreading during in vitro culture. Comparison of degradable and nondegradable hydrogels with similar physical properties (e.g., equilibrium swelling ratio) demonstrated that hMSC viability and spreading were dependent on network degradability. This study demonstrates that hydrolytically degradable PEG hydrogels can be formed via a sequential step growth polymerization and photocross-linking process and the resulting materials may serve as promising matrices for 3-dimensional stem cell culture and tissue engineering applications.
Introduction Hydrogels have emerged as an important component of several tissue engineering strategies that aim to regenerate natural tissue structure and function using combinations of materials, cells, and soluble signals.1,2 Recent strategies to design hydrogels for these applications have focused on bioinspired approaches in which specific properties of a natural biological environment, such as cell adhesion ligands or soluble growth factors, are included within a material to influence cell behavior.3,4 These approaches often achieve a degree of control over the cell’s environment by using a base material that can serve as a “blank slate” for presentation of specific biological moieties (e.g., cell adhesion peptides) without interacting nonspecifically with other molecules (e.g., serum proteins). Therefore, polymer chains that do not intrinsically interact with biological macromolecules, but can be covalently conjugated to specific biological moieties, are of significant interest. Notable examples include poly(ethylene glycol) (PEG),3,5–9 agarose,10 alginate,1 and poly(2-hydroxyethyl methacrylate).11 PEG chains derivatized on each end with vinyl-containing functional groups have been a particularly prevalent base polymer, as they can be readily processed into a hydrogel network via photocross-linking6,9 and can be derivatized via simple chemistries to generate networks that contain specific peptides and proteins.3,7,8 These properties have led to widespread use of PEG hydrogels as matrices for 3-dimensional (3-D) culture of a wide variety of cell types, including osteoblasts,12 chondrocytes,13,14 neural cells,15 and multiple stem cell types.9,16–18 * To whom correspondence should be addressed. E-mail: wlmurphy@ wisc.edu. Telephone: 608-262-2224. Fax: 608-265-9239. † Department of Biomedical Engineering, University of Wisconsin. ‡ Department of Pharmacology, University of Wisconsin. § Department of Materials Science and Engineering, University of Wisconsin.
A particularly important parameter dictating the suitability of PEG hydrogel networks in 3-D cell culture applications is their degradation characteristics. The rate of degradation of a hydrogel influences the network mesh size over time, which affects nutrient and waste diffusion to and from cells embedded within the network. Network degradation can also influence the ability of cells to migrate through the network, form junctions, and synthesize their own extracellular matrix during new tissue growth. Control over network degradation is particularly important in tissue engineering approaches, which typically aim to encourage degradation in concert with new tissue growth. Thus, there is significant interest in developing strategies to modulate the degradation of 3-D cell culture matrices, particularly the aforementioned PEG hydrogel networks. Previous studies have varied hydrolytic degradation by generating block copolymer networks comprising stable PEG blocks and hydrolytically labile poly(lactic acid) (PLA) blocks,19,20 and the degradability of these materials has proven useful in a variety of biomedical applications, including minimally invasive implantation21 and prevention of postoperative adhesions.22 In an alternative approach, Hubbell and co-workers have generated PEG networks that contain enzyme-labile peptide sequences as part of the polymer backbone, and the networks therefore degrade in response to specific cell-secreted proteases.3,23,24 Taken together, these clever approaches have demonstrated the potential of controlled PEG network degradation in 3-D cell culture and tissue engineering.12,13 Here we describe a simple, alternative approach to create hydrolytically labile PEG networks. The approach involves reacting excess PEG-diacrylate chains with a dithiol, specifically dithiothreitol (DTT), to form water-soluble (-PEG-DTT-)n PEG polymer chains. Because of the stoichiometric imbalance in favor of the PEG-diacrylate, this step growth polymerization reaction results in chains that are acrylate-terminated and can therefore be photocross-linked into a hydrogel network using
10.1021/bm701179s CCC: $40.75 2008 American Chemical Society Published on Web 02/21/2008
Degradation and Mesenchymal Stem Cells in PEG Networks
Figure 1. Schematic representation of (A) Michael-type addition reaction between poly(ethylene glycol) diacrylate and dithiothreitol to introduce hydrolytically labile linkages into the polymer chain, (B) photocross-linking of acrylate-terminated polymers to form hydrogels containing hydrolytically labile bridges, (C) incubation of hydrogels in buffer simulating a physiological environment, leading to hydrogel degradation over time, and (D) labile bond and degradation products.
standard protocols. The resulting network contains both photoinduced cross-links, which are not readily degradable over months in aqueous solutions at physiologic temperature and pH, and dithiol “bridges”, which are hydrolytically labile (Figure 1). The hydrolytic lability of DTT bridges is due to the presence of a thioether bond proximal to the acrylate ester bond. Specifically, the presence of this proximal thioether group establishes a more positive atomic charge on the carbonyl carbon of the ester, thereby enhancing its reactivity toward nucleophilic hydroxyl anions in the primary step of base-catalyzed ester hydrolysis.25 Therefore, the extent of network degradation and erosion can be modulated in our networks by simply varying the ratio of DTT bridges to photoinduced cross-links, while the initial network properties are defined by the polymer molecular weight and cross-linking density, similar to standard PEG networks. The ease of processing these networks facilitates incorporation of viable human mesenchymal stem cells (hMSCs), which show enhanced viability and spreading with enhanced hydrogel network degradability. Because of its simplicity and high level of control over degradation, this approach could prove useful in a variety of 3-D cell culture applications, including tissue engineering.
Materials and Methods PEG-Diacrylate Synthesis. DTT, triethylamine, acryloyl chloride, and PEG (Mw ) 8, 12, 20, and 35 kDa) were obtained from SigmaAldrich (St. Louis, MO). PEG-diacrylate chains were synthesized using standard methods described previously.26 Briefly, alcohol-terminated PEG chains were exposed to a 4× molar excess of triethylamine base and reacted with 4× molar excess of acryloyl chloride in azeotropically
Biomacromolecules, Vol. 9, No. 3, 2008 843 distilled benzene. The reaction product was filtered to remove triethylamine salts and precipitated with chilled diethyl ether to give a solid product. Completion of this reaction was confirmed by NMR spectroscopy. PEG-DTT Reaction. To characterize the PEG-diacrylate (PEGDA) reaction with DTT, we added DTT (625 µM) to a range of buffered (pH 7.4) 8 kDa PEGDA solutions (0–625 µM). Solutions were allowed to react for 60 min at 37 °C. The amount of free thiol remaining in solution after each reaction was characterized using the Measure-iT thiol assay (Invitrogen, Carlsbad, CA), which forms a fluorescent complex upon reaction with free thiol. The fluorescence intensity of these samples was measured on a Biotek Synergy plate reader using a 480/20 nm excitation and 520/15 nm emission filter set and related to free thiol concentration via a standard curve relating fluorescence intensity to known DTT concentrations in solution. This reaction was further characterized via spectrophotometric analysis of acrylate and thiolate species in solution. Lutolf et al. have previously described a spectrophotometric method to monitor the reaction between acrylate and thiolate species, as these species show strong absorbance at 233 nm.8 The PEGDA reaction with DTT was characterized using this same method by measuring the absorbance at 233 nm of a buffered solution (pH 7.4) containing 1.5 mM PEGDA and 1.5 mM DTT at 37 °C over 5 min intervals for 60 min to determine changes in the concentrations of the thiolate and acrylate groups. Hydrogel Formation. Hydrogels were formed from acrylateterminated PEG chains via a photocross-linking reaction. 15% w/v (0.15 g/mL) PEGDA solutions with various concentrations of DTT were prepared in phosphate-buffered saline (PBS) containing 0.05% w/v (0.0005 g/mL) Irgacure 2959 (I2959) photoinitiator by incubating at 37 °C for 2 h. After 2 h, aliquots of the PEG-diacrylate solutions were placed between two glass plates with a 1 mm spacer. The plates were exposed to UV radiation (4.5 mW/cm2, λ ) 365 nm) for 7 min to achieve photocross-linking. This resulted in free-standing hydrogel networks, which were then allowed to swell in PBS for 24 h at 37 °C. In each condition, the mass equilibrium swelling ratio (Qm) of the hydrogel networks was determined using the measured mass of the network after 24 h incubation in PBS (swollen mass, MS) and the measured mass of the same network after copious DI H2O rinsing and freeze-drying (dry mass, MD) using the relation Qm ) MS/MD. Degradation and Erosion Characteristics. Degradation and erosion of PEG hydrogels containing various amounts of DTT bridges was characterized by examining changes in wet and dry mass of hydrogels during incubation in aqueous buffer. Measurements made after 24 h of hydrogel swelling in PBS were designated day zero measurements. Degradation experiments entailed measuring the wet mass of hydrogels at 7 day intervals for 28 days in PBS at 37 °C. An increase in wet hydrogel mass was indicative of hydrolytic scission of polymer chains and consequent increase in network pore size. Erosion experiments entailed measuring the dry mass of hydrogels after 0, 7, 14, and 28 days of incubation in PBS at 37 °C. At the conclusion of each erosion time point, samples were incubated in DI H2O to remove buffer salts and freeze-dried using a Labconco freeze-dry system. A decrease in dry gel mass was indicative of erosion, which entails degradation and loss of polymer chains from the network. Human Mesenchymal Stem Cell Incorporation. Human mesenchymal stem cells (hMSCs) and media (MSC growth medium, MSCGM) were purchased from Cambrex (Baltimore, MD). During expansion, cells were cultured on tissue culture polystyrene plates according to the protocol provided by the supplier. At passage 6, cells were harvested from the plate, suspended in serum-free medium, and counted using a hemacytometer. PEG-diacrylate solutions were prepared in two stages: (i) preparation of PEGDA-DTT bridges by incubating sterile PEGDA (15 wt %), DTT (0, 2.5, 5, or 10 mM), and I2959 (0.05 wt %) in serum-free medium at 37 °C for 90 min, and (ii) incubating the resulting (-PEG-DTT-)n-PEG solution with a CGGRGDSP peptide at 37 °C for 90 min.27 Because a difference in equilibrium swelling ratio can be thought of as a difference in gel volume, the
844 Biomacromolecules, Vol. 9, No. 3, 2008 RGDSP concentration in each hydrogel was normalized based on differences in initial equilibrium swelling ratio to ensure that the density of RGDSP ligands presented within the networks remained constant across all conditions.28 After RGDSP conjugation to polymer chains, cells were added to the polymer solution. The cell density in each solution was also normalized based on differences in initial equilibrium swelling ratio using the same method described for RGDSP normalization.28 In a sterile environment, hMSC suspensions in PEG solution were placed between two glass plates separated by a 1 mm spacer and exposed to UV radiation (4.5 mW/cm2, λ ) 365 nm) for 5 min to form hydrogel networks. Hydrogels were transferred from the glass plates to individual wells of a 24-well plate containing 1 mL of hMSC media and incubated at 37 °C, 5% CO2. Medium was replenished every 48 h. At various time points, cell viability was qualitatively analyzed by staining cells with the Live/Dead Cell Viability assay (Invitrogen, Carlsbad, CA) as per the manufacturer’s protocol and imaged on an Olympus IX51 inverted microscope using a FITC or rhodamine fluorescence cube for live or dead staining, respectively. The number of viable cells in hydrogels was quantitatively analyzed using the CellTiter Blue assay (Promega, Madison, WI) in which a nonfluorescent indicator dye, resazurin, is converted to a fluorescent dye, resorufin, by metabolically active cells. To perform the assay, the indicator dye was added to culture medium and the resulting solution was analyzed for fluorescence using a BioTek Synergy plate reader equipped with a 560/20 nm excitation filter and a 590/35 nm emission filter. Cell spreading within the materials was visualized qualitatively by brightfield microscopy using an Olympus IX51 inverted microscope. To characterize whether unreacted DTT was likely to be present during cell culture, we quantified the amount of unreacted thiol present in solution after reacting 15 wt % PEGDA (Mw ) 8 kDa) with concentrations of DTT that were used in cell culture experiments (e.g., 2.5, 5.0, 7.5, 10, and 12.5 mM) using the Measure-iT thiol assay described previously. To ensure that replacing 1× PBS with serumfree medium does not lead to significant changes in the physical properties of the hydrogels, we also determined the equilibrium swelling ratio of hydrogels formed in serum-free medium using the aforementioned method.
Hudalla et al.
Figure 2. Extent of reaction between PEGDA and DTT analyzed by (A) reacting 625 µM DTT with variable concentrations of PEGDA (0–625 µM) for 60 min and measuring the concentration of unreacted thiol with the Measure-iT thiol assay, and (B) measuring changes in thiolate and acrylate absorbance at 233 nm over time in a buffered solution (pH 7.4) containing 1.5 mM PEGDA and 1.5 mM DTT at 37 °C.
Results PEGDA-Dithiol Reaction. The reaction of PEGDA with DTT was rapid and efficient. When DTT (625 µM) was introduced into buffered solutions containing 8 kDa PEGDA over a range of concentrations (0–625 µM) and allowed to react for 60 min, the decrease in thiol concentration with increasing acrylate concentration suggested a near-quantitative reaction (Figure 2A). When DTT and 8 kDa PEGDA were mixed in a buffered solution at equimolar concentrations (1.5 mM), the decrease in absorbance at 233 nm over time further supported a near-quantitative reaction within ∼40 min between acrylate and thiolate species (Figure 2B). Absorbance measurements at 283 nm indicated that no DTT was converted from the reactive reduced state to the cyclic oxidized state during a 60 min reaction, validating our assertion that decreases in the thiol concentration are due to a thiol-acrylate reaction. On the basis of on these results, the extent of the acrylate-thiol reaction was assumed to be nearly complete (p ) 0.99) in theoretical approximations of resulting number-averaged molecular weight of polymer chains after step growth polymerization of PEGDA and DTT (see Discussion Section, Figure 3). Network Formation. Hydrogel networks containing DTT bridges can be formed via photocross-linking, and the presence of DTT has a significant influence on initial hydrogel swelling. Hydrogels were prepared from solutions in which [PEGDA, Mw ) 8000 Da] ) 18.75 mM and 0 mM e [DTT] e 17.5 mM. Note that in this manuscript we will refer to DTT-containing networks by the concentration of the DTT present in solution
Figure 3. Carothers theory approximation of number-averaged molecular weight of -(-PEG-DTT-)n-PEG- polymers formed in a 99% complete reaction between 15 wt % PEGDA, Mw ) 8 kDa and various concentrations of DTT.
during the step growth polymerization prior to hydrogel formation. The equilibrium swelling ratio of hydrogels with 2.5 mM DTT (Qm ) 11.2 ( 0.6) is significantly larger than that of hydrogels containing no DTT (Qm ) 9.8 ( 0.2). This swelling ratio increases substantially with increasing DTT incorporation, culminating in a Qm ) 114.2 ( 9.1 for hydrogels containing 17.5 mM DTT (Figure 4A). This trend of increasing equilibrium swelling ratio can be attributed to an increase in the average molecular weight of polymer chains as a result of the PEGDA step growth reaction with DTT prior to photocross-linking.
Degradation and Mesenchymal Stem Cells in PEG Networks
Biomacromolecules, Vol. 9, No. 3, 2008 845
Figure 5. Mass loss during degradation of hydrogels. Hydrogels were prepared using PEGDA chains with Mw ) 8 kDa (15 wt %), and the concentration of DTT bridges in the network was varied (dashed line represents complete hydrogel erosion).
Figure 4. PEG hydrogels containing DTT bridges have varying mass equilibrium swelling ratios (Qm) and degrade in a controllable time frame in physiologic buffer. Changes in Qm over time during incubation in PBS, pH ) 7.4, T ) 37 °C, of 15 wt % PEG hydrogels formed by photocross-linking after reacting 8 kDa (A), 20 kDa (B), or 35 kDa (C) PEGDA with varying concentrations of DTT (* indicates significant difference relative to hydrogels containing 0 mM DTT, p < 0.05; ^ indicates significant difference relative to Qm at t ) 0 days, p < 0.05).
It is noteworthy that, as the [DTT]/[PEGDA] ratio approached unity, there was a noticeable increase in viscosity of the solution, and at the highest [DTT] examined (17.5 mM), the solutions were difficult to process. This viscosity increase can be attributed to an increase in the molecular weight of growing polymer conjugates. Solutions prepared with [DTT] > 17.5 mM were too viscous for further processing. These observations indicate that our approach is limited to [DTT]/[PEGDA] ratios
