Complex Fluids Based on Methacrylated Hyaluronic Acid

Biomacromolecules 2010, 11, 769–775

769

Complex Fluids Based on Methacrylated Hyaluronic Acid Joseph E. Prata,† Tiffany A. Barth,‡ Sidi A. Bencherif,† and Newell R. Washburn*,†,‡ Department of Chemistry, and Department of Biomedical Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213 Received December 2, 2009; Revised Manuscript Received January 19, 2010

We present the preparation and characterization of viscoelastic formulations of hyaluronic acid functionalized with polymerizable methacrylate groups. We explored three different processing strategies for controlling microstructure and interchain interactions: lightly cross-linked near-gels, emulsion-cross-linked microspheres, and an elastic microgel formed through centrifuging the microspheres. The component structure and rheological properties of these formulations were compared to those of high molecular weight hyaluronic acid solutions, which displayed classical behavior of high molecular weight polymer solutions reported by other investigators. We demonstrate that these processing strategies allow the tuning of solution properties from strongly viscoelastic behavior, observed in lightly cross-linked near-gels and concentrated microsphere solutions to elastic behavior in elastic microgels, behaving like pseudoplastic liquids having a well-defined yield stress above which viscous behavior was observed. In the centrifuged microspheres, the hyaluronic acid degree of methacrylation was inversely proportional to the gel elasticity, and a mechanism based on failure due to microsphere brittleness is proposed to explain this behavior. These results suggest that processing methacrylated hyaluronic acid can lead to a diversity of solution properties, providing methods for delivering this biologically active polymer in a broad range of applications.

Introduction Viscoelastic characteristics are a critical requirement for use in many medical applications. Solutions are used in injectable systems for promoting tissue repair, while ointments and other formulations characterized by a yield stress are more appropriate for topical application. Many of the strategies for preparing viscoelastic formulations of therapeutics employ polymers as biocompatible yet inert drug-delivery vehicles. The use of biologically active polymers as a material in biomedical applications will rely heavily on the development of processing strategies that tune the material properties but still provide a formulation that delivers a controlled dose of the biopolymer. Development of new processing strategies for preparing the required viscoelastic properties will build a foundation for broader application of these materials. Hyaluronic acid (HA) is a glycosaminoglycan with intrinsic biological activities in promoting tissue-repair responses.1 Most mammalian cells express receptors for high-molecular weight HA, such as CD44 and RHAMM, and their activation induces motile phenotype2 and could improve healing times.3 HA is being used broadly as a biopolymer therapeutic in applications ranging from viscosupplementation of arthritic joints4 to dressings for cutaneous wound healing.5 Wound dressings based on HA have been developed through cross-linking with divinyl sulfone,6 while viscosupplementation is particularly effective with HA having a molecular weight of 6 MDa.7 Control over the rheological properties of HA is necessary for use in a broad range of applications. The rheological properties of native HA have been studied extensively.8,9 In buffer solution, the rheology of high molecular weight HA is similar to that of high molecular weight polymers * To whom correspondence should be addressed. E-mail: washburn@ andrew.cmu.edu. † Department of Chemistry. ‡ Department of Biomedical Engineering.

in good solvents. While appropriate for some applications, the range of viscoelastic behavior is relatively narrow, especially when compared with the broad range of viscoelastic phenomena exhibited by other complex fluids. In concentrated solutions, polymer chains form an entangled network for which the viscosity reaches a constant value only at shear rates below the reciprocal terminal relaxation time and scales with concentration with a power-law exponent of 4. At frequencies above this terminal relaxation time, a broad rubbery plateau was observed for entangled HA solutions. Elastic responses can help in viscosupplementation but complicate facile injection of HA to injury sites, and for certain applications, it would be useful to broaden the frequency range over which the material behaves elastically, while for others it hinders application. Strategies for tuning the viscosity and conditions under which viscous or elastic responses are observed need to be developed to broaden the use of this biologically active polymer. Complex fluids have viscoelastic properties that are determined by the characteristic size and interactions of the components.10 Shear-thinning liquids display power-law decreases in viscosity as a function of shear rate. While entangled polymer solutions are a good example of this,11 many additives, such as associating alginate gels and surfactant micelles, have viscosities that decrease continuously from 105 to 102 cP as the shear rate increases from 0.1 to 100 s-1.12 In contrast, pseudoplastic liquids behave as elastic solids due to strong interparticle association but exhibit a yield value above which they flow as liquids. Many ointments are pseudoplastic liquids having yield values on the order of 100 Pa but viscosities above the yield stress that are less than 10 cP.12 Imparting this range of rheological behavior in hyaluronic acid will require the development of strategies for tuning the solution microstructure. Covalent modification of HA with polymerizable methacrylate groups has been used extensively to prepare elastomeric hydrogels. Smeds and Grinstaff were the first to report the preparation of cross-linked HA gels via methacrylation chem-

10.1021/bm901373x  2010 American Chemical Society Published on Web 02/11/2010

770

Biomacromolecules, Vol. 11, No. 3, 2010

Prata et al.

Scheme 1. Methacrylation of HA by Reacting with Glycidyl Methacrylatea

a

The transesterification product is not shown in this scheme, which would also functionalize at the pendant hydroxyl group.

istry.13 Burdick et al. demonstrated that methacrylic anhydride could be used to functionalize HA.14 Extending the method of Leach et al.,15 Bencherif et al. demonstrated a range of degree of methacrylation, ranging from 14 to over 90% by monomer fraction,16 as shown in Scheme 1. These authors noted that the more highly methacrylated samples had an increased Young’s modulus but became increasingly brittle. In this work, we demonstrate that this technique may also be used to produce complex fluids based on hyaluronic acid glycidyl methacrylate (HAGM) having a broad range of viscoelastic characteristics.

Experimental Section Materials. Hyaluronic acid derived from Streptococcus equi (∼1.6 MDa), glycidyl methacrylate (GM), triethylamine (TEA), hexanes, Span 80, and Tween 80 were purchased from Sigma-Aldrich and were used as received. 2,2′-Azobis (2-amidino-propane) dihydrochloride thermal initiator was purchased from Wako Chemicals. All chemicals were of reagent grade and were used without further purification. Synthesis of HAGM. Methacrylate groups were added to HA to yield HA-glycidyl methacrylate (HAGM) conjugates. HAGM macromonomers with different degrees of methacrylation (DM) were prepared as reported previously.16 For example, the synthesis of HAGM with a degree of methacrylation of 8% was performed as follows: 2.5 g of HA were first dissolved in 500 mL of phosphate buffer saline (PBS, pH ∼ 7.4) and subsequently mixed with 9.38 g (16.53 g for 8% DM) of GM and 4.69 g (8.27 g for 8% DM) of TEA. After a 10 d reaction, the solution was precipitated twice in a large excess of acetone (20× the volume of the reaction solution), filtered, dried, and dialyzed for 3 d against H2O-d. Following dialysis, the samples were lyophilized for 3 d to dry the material. The degree of methacrylation was measured using 1H NMR and could be tuned from 5 to 32% as a fraction of functionalized monomers (see Supporting Information for 1H NMR characterization data). Synthesis of HAGM Near-Gels. A total of 1.5 mg of HAGM (14% degree of methacrylation) was dissolved in 1 mL of PBS solution. To this, 82 µL of a 0.1% (by mass) solution of Irgacure 2959 photoinitiator was added. The solution was exposed to UV light (365 nm, 300 µW/ cm2) for 15 min to complete photocuring. Synthesis of HAGM Microspheres. A series of stable HAGMbased emulsions were prepared with different DMs (5, 8, 14, and 32%). A typical procedure was as follows: 3% (wt/v) solution of HAGM in 2 mL PBS was prepared. 2,2′-Azobis(2-amidino-propane) dihydrochloride initiator (0.001288 g, 4.75 × 10-6 mol) was subsequently added to the reaction mixture in a 50 mL round-bottom flask at room temperature. The resulting clear solution was mixed with a solution of Span 80 (0.3167 g) and Tween 80 (0.0762 g) in hexanes (20 g), and the mixture was sonicated for 2 min in an ice bath at 0 °C to form a

stable inverse mini-emulsion. Nitrogen was then bubbled through the dispersion for 10 min. The flask was immersed in an oil bath preheated to 53 °C to start the polymerization. The polymerization was stopped after 24 h by exposing the reaction mixture to air. A stable turbid dispersion remained. An aliquot was taken and the hexanes were removed from the flask either by rotary evaporation or exposure to atmosphere. The cross-linked microspheres were purified by dialysis for 3 d against H2O-d by using a dialysis membrane with MWCO of 50000 before being freeze-dried. The microspheres were then taken up in PBS to form HAGM microgels. Preparation of HAGM Entangled Microgels. Following the 24 h reaction period, 25-30 mL of the cross-linked HAGM sample was taken and added to a centrifuge tube. To the tube, 3 mL of PBS was added. The solution was centrifuged for 10 min at 4 °C at 4400 rpm. Hexanes were removed from the mixture, and 5 mL of PBS were added to the centrifuge tube. The centrifugation cycle was repeated a second time under the same conditions as the first run. The remaining solvents were removed, and the microgel was stored at 4 °C. Dynamic Light Scattering. Particle size was measured using dynamic light scattering (DLS) by CONTIN analysis. A Malvern Particle Sizer HPPS 5001 was used with a He-Ne laser (wavelength 633 nm) working at 3.0 mW. Polymers were dissolved in PBS at a concentration of 1.0 mg/mL and introduced into a thermostatted scattering cell at 25.0 °C for measurement. The sizes were calculated by volume measurement and expressed as Dav ( S (average diameter ( standard deviation). Transmission Electron Microscopy. The specimens were dried in a lyophilizer then reconstituted in deionized water and sprayed onto a TEM grid using a nebulizer. Samples were imaged using a tungsten salt negative-staining technique, but stained positively as HA absorbed the salt solution. An Hitachi H-7100 transmission electron microscope was used to capture images. Rheological Measurements of HAGM Formulations. Testing of the aqueous solutions of HAGM samples was performed on a Bohlin Instrument Gemini 200 rheometer. The cone and plate (40 mm diameter, 4° cone angle), cup and bob, and double gap geometries were used. The vertical gap was set to 150 µm for all measurements. Cone and Plate Geometry. The cone-and-plate cell was used for more concentration samples with higher viscosities. Approximately 5 mL of sample solution was added to the plate of the rheometer. The cone was then lower to a distance of 150 µm, forcing the viscous material to the edge of the cone. A cover was placed over the geometry to prevent evaporation. Samples were initially sheared at a constant stress of 1 Pa for 180 s, with an equilibration time of 30 s. Following the preshear, the shear rate was varied between 0.001 and 1000 s-1. Repeat measurements were run sequentially without preshear application. Cup and Bob Geometry. The cup and bob cell was used for dilute samples with low viscosities. Approximately 20 mL of material was

Complex Fluids and Methacrylated Hyaluronic Acid


771

Figure 1. (a) Steady-shear viscosity of HA solutions as a function of frequency and concentration. (b) Log-log plot of specific viscosity as a function of concentration. The semidilute entangled and entangled regimes are covered in this concentration range.

poured into the well of the cup and bob geometry. The bob was lowered into the cup with a separation distance of 150 µm, forcing the liquid approximately 1 mm up the stem of the bob. A cover was placed over the system to prevent evaporation. Identical preshear conditions were applied using this geometry. Viscosity as a function of shear rate was measured using the same upper and lower limits as in the cone and plate experiments.

Results and Discussion Complex fluids are formed through preparation of molecular assemblies with characteristic size distributions and interaction parameters, resulting in materials with a spectrum of stress relaxation times.10 The best studied examples of complex fluids are polymer solutions, which display a rich diversity of viscoelastic behaviors,11,17 but emulsions and other types of molecular aggregates have also been investigated extensively. Aggregates are formed either by using components with heterogeneous structures that self-assemble, such as surfactants18,19 and block copolymers,20,21 or through processing components into assemblies that impart the necessary solution properties. Polysaccharides, such as HA, are difficult to convert into narrowly defined, self-assembled solution structures because their homogeneous repeat structure makes it difficult to prepare macromolecules with distinct regions having polar and nonpolar character. Processing these biopolymers is the best option for tuning their viscoelastic properties, and methacrylation provides a straightforward method for tuning aggregate size and interactions in tandem with other processing techniques. The solution structure and rheology of unmodified HA, lightly cross-linked HAGM near-gels, HAGM microspheres, and HAGM microgels are presented. Unmodified HA. The HA had a number average molecular weight of 1.6 MDa and dynamic light scattering (DLS) characterization indicated the hydrodynamic diameter of the polymer in buffer was 770 nm (data shown in Supporting Information). This is significantly larger than the dimensions measured for HA having Mw of 1.86 MDa, which was 190.7 nm at salt concentrations of 0.06 N.22 This suggests the unmodified HA used in these studies tends to aggregate, although shear forces may significantly affect this process. To validate the base material, steady-shear viscosities of a series of HA solutions with concentrations ranging from 1.4 to 40.8 mg/mL as a function of shear rate, as shown in Figure 1a. From each plot the zero shear viscosity (η0) was determined from the equation η0 ) limγ˙ f0η. A best-fit line was extrapolated from the linear region of the curve to the y-intercept to determine

the η0 for each concentration. The specific viscosity was then calculated using (η0 - η solv)/ηsolv, where ηsolv in this equation is the solvent viscosity, which was then plotted as a function of concentration of HA in PBS. A power-law increase in specific viscosity was observed as the concentration of HA in solution increased. As shown in Figure 1b, there were two distinct slopes observed with a transition occurring around 4.4 mg/mL. The first slope of 1.45 ( 0.06 corresponded to the relationship in the semidilute entangled region, and the second slope of 3.74 ( 0.06 corresponded to the relationship in the entangled region. In our experiments, the entangled region was extended to a concentration of 40.8 mg/mL, which falls below the line fitting the data at lower concentrations but still follows the general trend. So while it appears that in these experiments the HA molecules aggregate in semidilute solutions, the agreement with previous rheological studies8 is good. Aggregation of HA was inferred based on NMR data and computer simulations, and the data indicated that a β-sheet-type structure was forming.23 While this would suggest that the rheological behavior of the native biopolymer would reflect this tendency toward gelation, most rheological anomalies in HA solutions have been traced to the effects of contaminating proteins.24 One explanation for our evidence of aggregates is that the cohesive forces responsible for them are weak compared to the shear forces applied in the rheology experiments, so they are only observed under static conditions used in the DLS experiments. HAGM Near-Gels. Lightly cross-linked solutions of highmolecular weight polymers can form large aggregates with long stress-relaxation times that are characterized by power laws.25 These near-gels form below the percolation threshold but have elastic responses over a broad frequency range, making them attractive alternatives to high molecular weight polymer solutions because significant increases in viscosity can be realized with lower volume fractions of material. We prepared HAGM near-gels by tuning the concentration and degree of methacrylation to that slightly below the gelation line at which a percolating network was formed. The transition from solution to gel occurs over a narrow composition range, with low HAGM concentrations and degrees of methacrylation resulting in no significant changes in solution viscosity while high HAGM concentrations and degrees of methacrylation resulted in the formation of a monolithic gel. However, at concentrations near 1.5 mg/mL and degrees of methacrylation near 5%, a highly viscous solution formed following crosslinking. Significant changes in either concentration or degree of methacrylation did not lead to the formation of the highly

772


Prata et al.

Figure 2. (a) Steady-shear viscosity of HAGM near-gel as a function of shear rate. (b) Storage modulus (G′) and loss modulus (G′′) as a function of oscillation frequency for HAGM near-gel.

dissipation begins to decline sharply as a percolating network is formed. At the critical gel point, classical near-gels have viscoelastic spectra characterized by power laws for which the time- and frequency-dependent moduli are predicted to follow the forms

G(t) ) St-n G'(ω) )

Figure 3. Transmission electron micrographs of HAGM microspheres that form the HAGM microgels. The scale bar at top right represents 500 nm, and the faint speckles that form the background are imaging artifacts.

viscous near-gel (see sample data in Supporting Information), so because the rheological properties were insensitive to small changes in these parameters, only this composition is reported. The solution structure of the near-gels was measured using DLS. A single mode centered with effective hydrodynamic diameter of 2453 nm was measured, (data shown in Supporting Information), which is four-times greater than the hydrodynamic diameter we measured for the unmodified HA, suggesting that these cross-linked solutions have HAGM aggregates that are significantly larger than the unmodified HA. The results from measurement of steady-shear viscosity as a function of shear rate are shown in Figure 2a. The magnitude of the steady-shear viscosity at 1 s-1 for a 1.5 mg/mL solution was 2 × 104 cP, which is 1000× greater than that of the comparable HA solution. Shear-thinning was observed across the much of the frequency range, which was fit to a power law equation with an exponent of -0.70, indicating that the terminal relaxation time in these HAGM near-gels is quite long, consistent with the increase in aggregate size over unmodified HA. Frequency-dependent measurements of the storage and loss modulus, shown in Figure 2b, have G′ > G′′ for the entire frequency range measured, although the crossover frequency appeared to be near 0.01 s-1. Gelation is modeled through cluster formation,26 and the gel point is the composition beyond which the low-frequency

G''(ω) ) Γ(1 - n)cos(nπ/2)Sωn tan(nπ/2)

(1a) (1b)

where S is the gel strength and n is a characteristic exponent. The fractal nature of these aggregates results in the power-law dependence of the moduli and the gel strength having units of Pa · s1/n. For n < 0.5, this relationship predicts that G′ > G′′. In these HAGM near-gels, G′ > G′′ across the frequency range measured but the slope of G′ on the low-frequency regime of the log-log plot the values of G′ and G′′ were converging, both with slopes of 0.3. These results suggest that the materials formed in these experiments were not consistent with the predictions of gelation theory. Degradable hydrogels formed through polymerization of pendant methacrylate groups have been shown to have degradation kinetics that obey the predictions of rubber elasticity theory,27 and the kinetics of crosslinking these materials were expected to follow the predictions of gelation theory. It is possible that controlling the solution composition so that bulk gelation cannot occur results in the formation of an incipient network structure that is qualitatively different than that formed in critical gels. However, further characterization will be needed, including comparisons to polymerizing HAGM solutions that have the gelation arrested near the critical point. HAGM Microspheres. Weakly interacting solutions of hard microspheres display nearly frequency-independent viscosities up to particle volume fractions of 30%,28 above which packing effects and multiparticle interactions start to dominate the response. In the low-concentration regime, the viscosity is governed by the Einstein equation:

η ) ηsolv(1 + 2.5φ + 6.5φ2)

(2)

where η is the viscosity of the microsphere solution, ηsolv is the viscosity of the pure solvent, and φ is the volume fraction of microspheres.



773

Figure 4. (a) Steady-shear viscosity of HAGM microgels at volume fractions of 10 and 40%. (b) Storage and loss moduli as a function of frequency for HAGM microgels at the same volume fractions.

For deformable microspheres, the rheological behavior is qualitatively different. For example, poly(methyl methacrylate) microspheres were prepared by emulsion polymerization and dispersed in a better than θ solvent.29 These solutions displayed shear-thinning viscosities at volume fractions less than 0.1, and the zero-shear viscosity diverged asymptotically at concentrations that depended on the degree of cross-linking. Furthermore, aqueous dispersions of poly(N-isopropyl acrylamide) at volume fractions of 12.5% displayed shear thinning under steady flow conditions but an increase in the high-frequency storage modulus under oscillatory conditions, which suggests the tendency to form networks.30 Because of their complex rheological responses, concentrated solutions of deformable microspheres are commonly referred to as microgels. Cross-linking HAGM in inverse microemulsions resulted in the formation of a turbid, low-viscosity solution that could be reversibly formed following lyophilization. We did not observe a decrease in the viscosity of unmodified HA following sonication, suggesting that this processing technique does not result in further degradation of the native biopolymer or of chemically modified forms of it.31 Images of the HAGM microspheres were collected using transmission electron microscopy (TEM), shown in Figure 3. The microstructure is characterized by spherical objects with a broad size distribution centered around 30 nm. A single mode with mean size 308 nm was observed in the DLS experiments (data shown in Supporting Information), which indicates the swelling ratio of the HAGM microspheres was approximately 10, lower than the swelling ratio of 30 [(gel mass)/(polymer mass)] measured for HAGM with 23% degree of methacrylation.16 While this is a rough estimate, it is possible that polymerizing HAGM in microemulsions results in tighter cross-linking due to confining the reaction to a smaller volume, making their mesh size smaller and their modulus higher than the bulk gels. The steady-shear viscosities of the cross-linked microspheres were measured at volume fractions of 10% and 40% are shown in Figure 4a. At 10% volume fraction of microsphere we observed shear thinning at frequencies greater than 0.1 s-1 in which the viscosity drops from 70 cP to around 2 cP, suggesting that a drastic shear-thinning transition occurred at a characteristic frequency. At microsphere concentrations of 40% and higher we observed a smooth decrease in the viscosity as a function of shear rate across the entire frequency range. Interestingly, samples at different concentrations below 40% displayed qualitatively similar behavior under steady shear, while those at 40 and 60% did not display this sudden drop in viscosity but had a smooth decrease in viscosity with increasing shear rate

Figure 5. TEM image of centrifuged HAGM microspheres. The scale bar at top right represents 100 nm.

(data shown in Supporting Information). We interpreted these results as suggesting that shearing is capable of disrupting microgel aggregates at lower concentrations, resulting in an abrupt change in viscosity at a characteristic shear rate. At higher concentrations, the higher microsphere packing makes the gels resistant to this transition, but steady shear can align particles and lead to constant decreases in viscosity with increasing shear rate. The storage and loss moduli of the 10% and 40% microgels as a function of frequency are shown in Figure 4b. Similar to the poly(NIPAAm) microgels,30 we observed increases in G′ and G′′ over the frequency range measured, with the change in slope of G′ increasing at a characteristic frequency of 0.2 s-1 for the 10% microgel and 3 s-1 for the 40% microgel. This is qualitatively similar to the high-frequency modulus measured for charged polystyrene lattices, which were attributed to the formation of polycrystalline aggregates.32 These transition frequencies were nearly constant for the microgels below and above 40% (additional plots are shown in Supporting Information), which suggests that packing effects lead to a qualitative change in microstructure that results in this shift. Work is currently underway to elucidate the structure of these phases and correlate them with their rheological responses.

774


Prata et al.

Figure 6. (a) Yielding behavior of centrifuged HAGM microspheres. The yield stress scales with the degree of methacrylation, reaching a maximum of 61.9 Pa for the 32% sample, whereas that for the 8% methacrylated sample was 9.3 Pa. (b) Relationship between yield stress and degree of methacrylation. The solid line is used as a guide. Table 1. Rheological Properties of HA and HAGM Forms hydrodynamic zero-shear crossover yield diameter (nm) viscosity (cP) frequency (1/s) stress (Pa) hyaluronic acid

7.875 mg/mL 1.395 mg/mL HAGM near-gels 1.5 mg/mL microgel 10% 40% entangled microgel 5% 8% 14% 32%

770 2453 308

1040.12 10.79 v v v

238

HAGM Entangled Microgels. Strongly interacting emulsions form a diverse class of complex fluid that displays elastic behavior up to a yield stress.10 Emulsion interactions may be mediated through Coulombic32 or solvent forces,33 or through entanglements between micelles having long-chain grafts on their surfaces.34,35 Often referred to as pseudoplastic liquids, studies of the rheological properties of ketchup, mayonnaise, toothpaste, and other strongly interacting microgels have provided a framework for understanding their properties. Mechanically processing the HA microspheres by centrifugation resulted in the formation of an opaque, weakly cohesive solid. The structure of these microgels was characterized by TEM and DLS. The TEM data, shown in Figure 5, indicate that the centrifuged microgel was also composed of HAGM microspheres. The DLS data (shown in Supporting Information) had a broad distribution and were fit to an average hydrodynamic diameter of 238 nm, a reduction in the average size of 70 nm compared to the size of HAGM microspheres measured with DLS. Mechanical processing shifted the size distribution to smaller particles, and we hypothesize that the microspheres were damaged by the stresses in centrifugation. Centrifuged HAGM microgels with degrees of methacrylation ranging from 5-32% were tested at increasing shear stress to determine the yield stress of the material at different degrees of methacrylation. In the bulk HAGM hydrogels, higher levels of methacrylation lead to higher modulus but a greater tendency toward brittle fracture, suggesting that mechanically processing the HAGM microgels may lead to fracture of the microspheres.16 The mechanical response of the HAGM microgels displayed a marked stress dependence. The storage modulus is shown as a function of shear stress in Figure 6a. For microgels based on HAGM having degree of methacrylation of 32%, the storage modulus was nearly constant 170 Pa but dropped sharply above

0.0148 0.183 3.677 1.00 52.24 1.80 98.45

0.10 9.29 39.06 61.85

ηss 1/s (cP)

ηss G′ G′′ 10/s (cP) at 1/s (Pa) at 10/s (Pa)

1032.00 731.40 17.09 7.89 40400.00 7122.00 189.00 4.34 0.613 0.3866 42.05 6.167 0.0421 2968.00 2053.00 22.05 7575.00 777.50 200.20 139.80 76.49

243.70 2.9520 3.8840 56.75 227.80 144.60 117.30

a shear stress of 61.9 Pa. Above this yield stress, the response was primarily viscous, suggesting that these materials behave like pseudoplastic liquids. Interestingly, the yield stress was correlated with the HA degree of methacrylation, as shown in Figure 6b. At 5% degree of methacrylation, the microgels demonstrated completely viscous responses, and a yield stress of 0.10 Pa was measured for the sample having degree of methacrylation of 5%. Based on these results, we hypothesize that mechanically processing the microspheres results in fracture that leads to decreases in particle size but frees HA chains to form entanglements between the particles. In this model, more highly crosslinked microspheres were more brittle, resulting in a greater interactions and increased yield stress following processing. We propose that formation of such entanglements leads to the significant change in the viscoelastic properties from those of the original HA microspheres, leading to the formation of an elastic network that we term an entangled microgel. Comparison of HA and HAGM Forms. A comparison of some of the metrics of the different HAGM forms compared to unmodified HA is shown in Table 1. In this table we report the zero-shear viscosity η0, crossover frequency from oscillatory measurements, yield stress, the steady-shear viscosity ηss at 1 and 10 s-1, and the storage and loss moduli at 1 and 10 s-1. Note that no formulation displayed all these characteristics, and the terminal viscosity for some could not be measured, so these were represented with an arrow (v). Cross-linking via methacrylate groups can be used to tune the size of the constituents in the solution and their interactions. For near-gels, the cross-linked aggregates have a hydrodynamic diameter that is more than three times that measured for native HA, resulting in shear-thinning behavior across the entire range of shear rates that were measured. The viscosity of the near-


gels at 1 s-1 was more than 1000× greater than that of native HA solutions having similar concentrations, demonstrating the strong effect of polymer size on solution properties. Emulsion polymerization of HA resulted in the formation of particles having diameters of approximately 300 nm, but the weak interactions between aggregates resulted in a solution viscosity that was comparable to that of the native HA. Because of particle compliance, deformations were hypothesized to result in non-Newtonian behavior characterized by shear thinning under steady shear and concomitant increases in G′ and G′′ under increasing oscillatory shear at a frequency that depended on microsphere concentration. Centrifugation of the microspheres resulted in the formation of a pseudoplastic liquid that displayed elastic responses up to a yield stress. The yield stress correlated with the degree of methacrylation and, given the tendency of highly cross-linked HAGM gels to fracture, we hypothesized that mechanical processing of the microspheres resulted in fracture that increased interparticle interactions. These entangled microgels represent a new route to preparing pseudoplastic liquids based on HA. In the future, it will be critical to determine the extent to which cells are still capable of recognizing methacrylated HA. The canonical HA receptors CD44 and RHAMM have been shown to recognize as few as six HA repeat units,36 which suggests that lightly methacrylated HA should have regions that can be readily recognized.

Conclusions The methacrylation of HA can be used as the basis for preparing formulations of this biologically active polymer with a diverse range of viscoelastic properties. While native HA displays rheological properties consistent with those of high molecular weight polyelectrolytes in a good solvent, methacrylated HA can be processed into near-gels that have high viscosities at low concentrations of HAGM, microgels that display shear thinning under steady flow conditions but have a high-frequency modulus indicative of a network formation, and entangled microgels with yield stresses similar to pseudoplastic liquids. These results suggest that methacrylation provides a versatile method for tuning the viscoelastic properties of this important biopolymer and could lead to new formulations that form the basis for use in a broad range of applications. Acknowledgment. We gratefully acknowledge Annette Jacobson, Susana Steppan, and Joseph Suhan for assistance with these measurements, Guy Berry for helpful discussions, the Center for Molecular Analysis at Carnegie Mellon University (NSF CHE-9808188), and the NMR facility at Carnegie Mellon University (NMR instrumentation partially supported by NSF CHE-0130903). T.A.B. gratefully acknowledges support from the HHMI Undergraduate Science Education Program at CMU (52005865). This work was supported in part by W81XWH08-2-0032 (U.S. Army). Supporting Information Available. Results of NMR characterization of the starting material as well as rheological and particle-size characterization. This material is available free of charge via the Internet at http://pubs.acs.org.


775

References and Notes (1) Kogan, G.; Soltes, L.; Stern, R.; Gemeiner, P. Biotechnol. Lett. 2007, 29 (1), 17–25. (2) Savani, R. C.; Cao, G.; Pooler, P. M.; Zaman, A.; Zhou, Z.; DeLisser, H. M. J. Biol. Chem. 2001, 276 (39), 36770–36778. (3) Gailit, J.; Clark, R. A. Curr. Opin. Cell Biol. 1994, 6 (5), 717–725. (4) Strauss, E. J.; Hart, J. A.; Miller, M. D.; Altman, R. D.; Rosen, J. E. Am. J. Sports Med. 2009, 37 (8), 1636–1644. (5) Matsumoto, Y.; Arai, K.; Momose, H.; Kuroyanagi, Y. J. Biomater. Sci., Polym. Ed. 2009, 20 (7-8), 993–1004. (6) Larsen, N. E.; Pollak, C. T.; Reiner, K.; Leshchiner, E.; Balazs, E. A. J. Biomed. Mater. Res. 1993, 27 (9), 1129–1134. (7) Allard, S.; O’Regan, M. Clin. Ther. 2000, 22 (6), 792–795. (8) Krause, W. E.; Bellomo, E. G.; Colby, R. H. Biomacromolecules 2001, 2 (1), 65–69. (9) Milas, M.; Roure, I.; Berry, G. C. J. Rheol. 1996, 40 (6), 1155–1166. (10) Larson, R. G. The Structure and Rheology of Complex Fluids; Oxford University Press: New York, 1998; p 688. (11) Ferry, J. D. Viscoelastic Properties of Polymers; John Wiley & Sons: New York, 1980; p 641. (12) Laba, D. Rheological Properties of Cosmetics and Toiletries; Marcel Dekker, Inc.: New York, 1993; p 426. (13) Smeds, K. A.; Pfister-Serres, A.; Miki, D.; Dastgheib, K.; Inoue, M.; Hatchell, D. L.; Grinstaff, M. W. J. Biomed. Mater. Res. 2001, 54 (1), 115–21. (14) Burdick, J. A.; Chung, C.; Jia, X.; Randolph, M. A.; Langer, R. Biomacromolecules 2005, 6 (1), 386–91. (15) Leach, J. B.; Bivens, K. A.; Patrick, C. W.; Schmidt, C. E. Biotechnol. Bioeng. 2003, 82 (5), 578–589. (16) Bencherif, S. A.; Srinivasan, A.; Horkay, F.; Hollinger, J. O.; Matyjaszewski, K.; Washburn, N. R. Biomaterials 2008, 29 (12), 1739–1749. (17) Macosko, C. W. Rheology: Principles, Measurements, and Applications; Wiley-VCH: Weinheim, Germany, 1994; p 568. (18) Gerber, M. J.; Walker, L. M. Langmuir 2006, 22 (3), 941–948. (19) Schubert, B. A.; Kaler, E. W.; Wagner, N. J. Langmuir 2003, 19 (10), 4079–4089. (20) Seitz, M. E.; Burghardt, W. R.; Faber, K. T.; Shull, K. R. Macromolecules 2007, 40 (4), 1218–1226. (21) Watanabe, H.; Matsumiya, Y. Macromolecules 2005, 38 (9), 3808– 3819. (22) Fouissac, E.; Milas, M.; Rinaudo, M.; Borsali, R. Macromolecules 1992, 25 (21), 5613–5617. (23) Scott, J. E.; Heatley, F. Proc. Natl. Acad. Sci. U.S.A. 1999, 96 (9), 4850–5. (24) Lapcik, L.; Lapcik, L.; De Smedt, S.; Demeester, J.; Chabrecek, P. Chem. ReV. 1998, 98 (8), 2663–2684. (25) Winter, H. H.; Mours, M. AdV. Polym. Sci. 1997, 134, 165–234. (26) Stauffer, D.; Coniglio, A.; Adam, M. AdV. Polym. Sci. 1982, 44, 103– 158. (27) Mason, M. N.; Metters, A. T.; Bowman, C. N.; Anseth, K. S. Macromolecules 2001, 34 (13), 4630–4635. (28) Dekruif, C. G.; Vanlersel, E. M. F.; Vrij, A.; Russel, W. B. J. Chem. Phys. 1985, 83 (9), 4717–4725. (29) Wolfe, M. S.; Scopazzi, C. J. Colloid Interface Sci. 1989, 133 (1), 265–277. (30) Kiminta, D. M. O.; Luckham, P. F.; Lenon, S. Polymer 1995, 36 (25), 4827–4831. (31) Jia, X.; Yeo, Y.; Clifton, R. J.; Jiao, T.; Kohane, D. S.; Kobler, J. B.; Zeitels, S. M.; Langer, R. Biomacromolecules 2006, 7 (12), 3336–44. (32) Buscall, R. J. Chem. Soc., Faraday Trans. 1991, 87 (9), 1365–1370. (33) Woutersen, A. T. J. M.; Dekruif, C. G. J. Chem. Phys. 1991, 94 (8), 5739–5750. (34) Chen, H. W.; Lu, Y. J.; Shen, L.; Liu, X. B.; Zhang, G. Z. Polymer 2008, 49 (8), 2095–2098. (35) Gohy, J. F.; Antoun, S.; Jerome, R. Polymer 2001, 42 (21), 8637– 8645. (36) Turley, E. A.; Noble, P. W.; Bourguignon, L. Y. J. Biol. Chem. 2002, 277 (7), 4589–4592.

BM901373X

Complex Fluids Based on Methacrylated Hyaluronic Acid

Recommend Documents