Micropatterning of Nanoengineered Surfaces to Study Neuronal Cell Attachment in Vitro
Download 0.87 Mb. Pdf ko'rish
|
|
7 As x decreases, the protein D e is decreased because the protein has an increasing probability of encountering a polymer chain rather than freely diffusing in solution. For example, Cruise et al. found that the diffusion coefficient of myoglobin in PEG hydrogels with mesh sizes of 58 A ˚ or 34 A˚ was decreased two or four orders of magnitude, respectively, compared to the D e of myoglobin in water. 8 Another network parameter that affects the D e of a protein in a hydrogel is the charge, which can be controlled in protein-releasing hydrogels by varying the number of ionized groups present on the network backbone at physiological pH. For example, when Gu et al. encapsulated vascular endothelial growth factor (VEGF) (pI ¼ 8.5) in alginate hydrogels (pI ¼ 3.5) there was significant electrostatic attraction between the protein and the hydrogel, and only $10% of the total loaded VEGF was released over five days. Addition of CaCl 2 to the buffer surrounding the alginate hydrogels disrupted the electrostatic interaction between VEGF and alginate, resulting in $100% VEGF release within 3 days. 9 Based on these studies and others, it is clear that the physicochemical properties of a hydrogel network strongly influence protein diffusion and, in turn, protein release. In this study we hypothesized that dynamic variation in hydrogel network properties via dynamic protein conformational changes could provide a novel mechanism to modulate release of a therapeutic protein. Recently there has been a great deal of interest in developing ‘‘smart’’ hydrogels for dynamic, controlled protein delivery. 7 Specifically, previous studies have demonstrated that the rate of protein delivery from a hydrogel can be dynamically varied over time by creating hydrogels that change their x, charge, or internal fluid pressure in response to a stimulus. One mechanism used to a The University of Wisconsin, Department of Biomedical Engineering, Madison, WI 53706, USA. E-mail: wlmurphy@wisc.edu b The University of Wisconsin, Department of Materials Science and Engineering, Madison, WI 53706, USA c The University of Wisconsin, Department of Pharmacology, Madison, WI 53706, USA † This paper is part of a Soft Matter issue highlighting the work of emerging investigators in the soft matter field. This journal is ª The Royal Society of Chemistry 2009 Soft Matter, 2009, 5, 2399–2406 | 2399 PAPER www.rsc.org/softmatter | Soft Matter dynamically change x involves creating hydrogel networks that change their solubility in response to a stimulus. This class of dynamic hydrogels has been designed to respond to physi- ochemical stimuli including pH, 10 temperature, 11 electric
current, 12 and light. 13,14 Another mechanism used to dynamically alter the protein release rate involves changing the internal fluid pressure in a hydrogel network. For example, Lee et al. mechanically loaded alginate hydrogels impregnated with VEGF, and found that the release rate of VEGF was increased due to the increased fluid pressure inside the hydrogel. 15 These approaches indicate that changes in the physicochemical prop- erties of a hydrogel network can be translated into changes in protein release characteristics. However, these approaches may be limited in the types of stimuli that can be used to induce changes in protein release, and they have focused on non-specific, physicochemical stimuli to date. Recently, hydrogels have been designed to respond to more specific biochemical stimuli by incorporating proteins that bind to specific ligands. 16–20
These binding events effectively change the number of crosslinks in a hydrogel network, which in turn decreases network x, resulting in volume changes of up to 20% when compared to the initial hydrogel volume. 17,19 For example, Miyata et al. created hydrogels in which antibodies and their specific antigens were covalently attached to the hydrogel network. Within the hydrogel network the antibodies were able to bind to the antigens, thereby increasing the crosslinking density and decreasing x. When free antigen was added to the solution bathing the hydrogel, the free antigen competitively replaced the network-bound antigen, resulting in a decreased crosslinking density and an increased x, leading to macroscopic volume shifts of approximately 10%. 17 This approach provides an important example of dynamic hydrogels that respond to biochemical cues, and motivates the need for hydrogels that undergo more substantial dynamic changes in response to specific biological molecules. We have previously designed dynamic hydrogels whose macroscopic dynamic properties are based on a nanometre-scale conformational change in the protein calmodulin (CaM). 21–23
Calmodulin undergoes a ‘‘hinge motion’’ upon binding of a variety of specific ligands, including small molecules, peptides, and proteins. 23 Therefore, hydrogels in which CaM is included as a functional network component have been shown to undergo reversible volume decreases of up to 75% of their initial volume in the presence of CaM-specific ligands. 21–23
In the current study, we hypothesized that the dynamic properties of CaM-based hydrogels could be tailored by changing: (1) the ratio of dynamic (PEG-CaM-PEG) to static (PEGDA) polymer in the hydrogel; and (2) the initial concentration of PEG-CaM-PEG present during hydrogel formation. Furthermore, we hypothesized that the volume decreases of PEG-CaM-PEG hydrogels in the pres- ence of a CaM ligand would be accompanied by fluid exclusion, which in turn would lead to enhanced release of a therapeutic protein molecule. More specifically, we hypothesized that transport of a growth factor out of a PEG-CaM-PEG hydrogel would be increased during hydrogel ‘‘collapse’’ when compared to typical transport via Fickian diffusion. We specifically focused on modulating release of VEGF in these studies, as VEGF is a potent therapeutic promoter of angiogenesis, the growth of new blood
vessels from
a pre-existing vascular supply.
24,25 Furthermore, in previous studies VEGF activity has been shown to be maintained after release from hydrogels. 15,26
Results herein indicate that uptake of VEGF into PEG-CaM-PEG hydrogels is proportional to the initial PEG-CaM-PEG concentration present during hydrogel formation, and that the VEGF release rate is significantly increased in environments that promote hydrogel collapse. These data provide a first demonstration that protein conformation changes can be used as a mechanism to modulate protein release from a material. Results PEG can be covalently conjugated to an engineered version of the protein calmodulin, and the resulting PEG-CaM-PEG conjugates can be purified without significant loss of protein activity. The Michael-Type addition reaction of PEGDA to the sulfhydryl side chains of CaM(T34C, T110C), shown schemati- cally in Fig. 1A, proceeded to completion as demonstrated by MALDI-ToF mass spectrometry (Fig. 1B). Phenyl sepharose column purification removed unreacted PEGDA from the reac- tion mixture as evaluated by HPLC (Fig. 1C). CaM(T34C, T110C) maintained nearly 100% of its original activity during phenyl sepharose purification, dialysis, and reaction with PEGDA. Lyophilization in the absence of the excipient raffinose decreased CaM(T34C, T110C)’s activity to 52.4 Æ 5.6% of its original activity. In contrast, when lyophilization was performed in the presence of the excipient raffinose 85.5 Æ 1.2% of original CaM activity was maintained (Fig. 1D). The initial mass fraction of polymer (4) and ratio of PEG- CaM-PEG3400 to PEGDA3400 influenced the static and dynamic properties of protein-based hydrogels. The initial Q m of
PEG-CaM-PEG3400/PEGDA3400 hydrogels was significantly greater than the Q m of the collapsed hydrogels in all conditions (p < 0.05). Interestingly, the Q m s of all hydrogels after exposure to the TFP ligand were not significantly different (p > 0.05) (Fig. 2A). PEG-CaM-PEG3400/PEGDA3400 hydrogels decreased their volume to a greater extent in response to TFP as the ratio of PEG-CaM-PEG3400/PEGDA3400 was increased. PEG-CaM-PEG3400/PEGDA3400 hydrogels with ratios of 7/3% and 8/2% underwent significant recovery when TFP was removed from the buffer, but did not recover to 100% of their swollen volume (p < 0.05) (Fig. 2B). Pure PEG-CaM-PEG3400 hydrogels of varying 4 decreased their volume to a greater extent with decreasing 4 (Fig. 2C). Hydrogels composed of constant 4 of PEG-CaM-PEG3400 and varying 4 of PEGDA3400 decreased in volume less in response to TFP as the 4 of PEGDA increased (Fig. 2D). Absorption and release of VEGF by dynamic, protein-based hydrogels is strongly influenced by the hydrogel physiochemical properties and hydrogel
dynamics. PEG-CaM-PEG3400 hydrogels absorbed more VEGF as 4 was increased (R 2 ¼ 0.9857) (Fig. 3A). Importantly, the VEGF release rate was increased when 10% PEG-CaM-PEG3400 hydrogels were exposed to TFP (Fig. 4A). 10% PEG-CaM-PEG3400 hydrogels could also be triggered to increase their VEGF release rate in response to TFP even after being initially maintained without TFP for 24 h (Fig. 4B, black line). 10% hydrogels undergo significant volume decreases when exposed to TFP (Fig. 4C), as demonstrated quantitatively in Fig. 2C. 2400 | Soft Matter, 2009, 5, 2399–2406 This journal is ª The Royal Society of Chemistry 2009
Interestingly, the release rate of VEGF increased in all PEG- CaM-PEG3400 hydrogels when the hydrogel underwent TFP- induced volume collapse, regardless of their initial polymer concentration (Fig. 5A–B). However, hydrogels with the highest 4 did not release the most VEGF over a 48 h time frame (Fig. 5A left side). 10% (w/v) PEG-CaM-PEG3400 hydrogels released 100.4
Æ 0.8% of their absorbed VEGF, whereas 15% (w/v) PEG- CaM-PEG3400 hydrogels released only 57.9 Æ 1.2% of their absorbed VEGF over 48 h (Fig. 5A right side). In hydrogels prepared with both PEG-CaM-PEG3400 and PEGDA3400, an increase
in the
PEG-CaM-PEG3400/PEGDA3400 ratio
increased the total amount of VEGF absorbed in the hydrogels, which in turn resulted in an increased amount of total VEGF released (Fig. 5B left side). Hydrogels prepared with PEG-CaM- PEG3400/PEGDA3400 ratios of 7/3%, 9/1%, and 10/0% each released more than 99% of their loaded VEGF in the presence of TFP (Fig. 5B right side). Discussion The synthesis and purification approach used here may be used for synthesis of other dynamic, protein-based materials. CaM (T34C, T110C) was reacted with PEGDA via a Michael-Type addition reaction between cysteines on the CaM(T34C, T110C) mutant and acrylate groups on PEGDA molecules (Fig. 1A–B). Many proteins have surface-exposed cysteine residues, which could be modified in a similar manner. Here we removed excess PEGDA after the PEGDA-protein reaction by purifying PEG- CaM-PEG conjugates using hydrophobic interaction chroma- tography (Fig. 1C). This form of column chromatography has been optimized for several proteins, and could be used to purify other protein–polymer conjugates without extensive method development. 27 The activity of CaM was monitored during PEG- CaM-PEG synthesis and purification using a modified assay for calcineurin activity, which is dependent on the presence and concentration of CaM. 28 PEG-CaM-PEG was found to maintain its activity during each processing step, with the exception of lyophilization. However, the presence of the sugar raffinose during lyophilization decreased the loss of CaM activity during lyophilization (Fig. 1D). These results are consistent with previous reports, which have shown that CaM is partially denatured during lyophilization and that raffinose has a protec- tive effect during lyophilization. 29 It is noteworthy that dynamic proteins often undergo conformation changes to hide or expose autoinhibitory domains that regulate another protein’s enzy- matic activity. 30 Thus, enzymatic activity assays—similar to the calcineurin-based assay performed herein—could be used to assess the activity of dynamic proteins in other applications of dynamic, protein-based materials. 31 The Q m of dynamic hydrogels can be tuned by varying 4 of PEG-CaM-PEG3400 or the ratio of PEG-CaM-PEG3400 to PEGDA3400. Increasing the 4 of PEG-CaM-PEG3400 hydro- gels led to a decrease in Q m , which may be attributed to an increased concentration of acrylate groups and increased cross- linking density. In hydrogels prepared with both PEG-CaM- PEG3400 and PEGDA3400, increasing the 4 of PEGDA3400 Fig. 1
(A) Schematic representation of the Michael-Type addition reaction between CaM(T34 C, T110C) and PEGDA to form PEG-CaM-PEG conjugates. (B) MALDI-MS spectra of CaM(T34C, T110C) before and after reaction with PEGDA 575. (C) HPLC chromatogram of CaM(T34C, T110C) and PEG-CaM-PEG575. (D) The calmodulin-dependent phosphatase activity of calcineurin was used to measure the activity of CaM as purchased, or the CaM mutant CaM(T34C, T110C) (CaMmut) after different processing steps in the synthesis of PEG-CaM-PEG molecules (n ¼ 2). This journal is ª The Royal Society of Chemistry 2009 Soft Matter, 2009, 5, 2399–2406 | 2401 while decreasing the 4 of PEG-CaM-PEG3400 produced hydrogels with decreasing Q m . This observation could be attributed to the increased crosslinking density of hydrogel networks
prepared with
higher 4 of PEGDA3400, as PEGDA3400 molecules have a higher number of acrylate groups per unit mass compared to PEG-CaM-PEG3400 molecules (Fig. 2A). These results are consistent with classical Flory– Rehner theory. 32 The dynamic properties of hydrogels were dependent upon 4 of PEG-CaM-PEG3400 hydrogels and the ratio of PEG-CaM- PEG3400 to PEGDA3400 in the initial hydrogel formulation. Pure PEGDA3400 hydrogels underwent no volume change in response to the addition or removal of TFP, which is consistent with previous studies by Sui et al. (Fig. 2B). 22,23
Hydrogels prepared
with mixtures
of PEG-CaM-PEG3400 and PEGDA3400 underwent greater magnitude volume collapses in response to TFP with increasing fractional mass of PEG-CaM- PEG3400. This phenomenon could be attributed to the greater Q m
ratios, which likely permits greater magnitude volume collapses compared to hydrogels with lesser Q m (Fig. 2A–B). A similar phenomenon was observed in the 100% PEG-CaM-PEG3400 hydrogels. Specifically, increasing the 4 of PEG-CaM-PEG3400 from 10–20% produced hydrogels with a decreased magnitude of Fig. 2
Characterization of the influence of the initial concentrations of PEG-CaM-PEG3400 and PEGDA3400 during hydrogel formation on hydrogel properties. (A) The influence of PEG-CaM-PEG3400 4 on the equilibrium swelling ratio (Q m ) of swollen and collapsed 5 ml hydrogels. [*-Significant difference between swollen and collapsed hydrogels (n ¼ 3) p < 0.05]. (B) The dynamic properties of 10% (w/v) hydrogels composed of different initial 4 of PEG-CaM-PEG3400/PEGDA3400 [*-Significant difference between swollen and collapsed hydrogels, † denotes significant difference between collapsed and recovered hydrogels (n ¼ 3) p < 0.05]. (C) Volume decreases and subsequent recovery of PEG-CaM-PEG3400 hydrogels with different 4 PEG-CaM-PEG hydrogels [*-Significant difference between swollen and collapsed hydrogels, † denotes significant difference between collapsed and recovered hydrogels (n ¼ 3) p < 0.05]. (D) Volume decreases of PEG-CaM-PEG3400/PEGDA3400 hydrogels with a constant 4 of PEG-CaM-PEG3400 [*-Significant difference between swollen and collapsed hydrogels, † denotes significant difference between collapsed and recovered hydrogels (n ¼ 3)
p < 0.05]. Fig. 3
Absorption of VEGF into protein-based, dynamic hydrogels. (A) VEGF absorption into hydrogels prepared with different 4 of PEG-CaM- PEG3400, and incubated in a 444 ng/ml VEGF solution. The inlayed schematic represents a PEG-CaM-PEG3400 hydrogel absorbing VEGF from solution. (B) The crystal structures and isoelectric points of VEGF (PDB : 2VPF) and CaM (PDB : 3CLN). Positively charged amino acids are shown in red on VEGF and negatively charged amino acids are shown in red on CaM (analyzed with Rasmol Software). 2402 | Soft Matter, 2009, 5, 2399–2406 This journal is ª The Royal Society of Chemistry 2009
dynamic volume change (Fig. 2C). Again, increasing the 4 of these hydrogels produces higher crosslinking densities and a decreased Q m , which limits the magnitude of hydrogel volume collapse. Importantly, hydrogels in which the 4 of PEG-CaM- PEG3400 was held constant and the 4 of PEGDA3400 was increased underwent smaller magnitude volume collapses with increasing PEGDA3400 (Fig. 2D). These results further support the hypothesis that the magnitude of hydrogel collapse is modulated by the initial crosslinking density and associated Q m .
PEG3400/PEGDA3400 hydrogels recovered to 100% of their swollen volume. This observation could be attributed to inho- mogeneous CaM incorporation. Upon hydrogel collapse, local- ized regions of high [CaM] within the hydrogel may become insoluble and form aggregates. Another possible explanation could involve CaM multimerization. CaM has been previously shown to non-covalently dimerize with an equilibrium dissocia- tion constant (K d ) of 570
Æ 70 mM. 33 Although this is a relatively Fig. 4 Release of VEGF is modulated by protein conformational changes in dynamic, protein-based hydrogels. (A) The total VEGF mass released from 10% PEG-CaM-PEG3400 hydrogels incubated with or without the ligand TFP, which induces a CaM conformational change. (B) The total VEGF mass released from 10% PEG-CaM-PEG3400 hydrogels, with TFP added after 24 h to ‘‘trigger’’ VEGF release. (C) Representative micrographs of 10% 4 PEG-CaM-PEG3400 hydrogels in their swollen, collapsed, and recovered state (scale bar ¼ 1 mm). Fig. 5
Release of VEGF is modulated by protein conformational changes in dynamic, protein-based hydrogels. (A) The total VEGF mass released (left side) and % of total VEGF released (right side) from PEG-CaM-PEG3400 hydrogels prepared with different 4s. (B) The total VEGF mass released (left side) and % of total VEGF released (right side) from PEG-CaM-PEG3400/PEGDA3400 hydrogels with different 4s. This journal is ª The Royal Society of Chemistry 2009 Soft Matter, 2009, 5, 2399–2406 | 2403 low affinity interaction, there could be significant increases in the K d within the hydrogels due to the avidity effects resulting from the mM concentrations of CaM used in PEG-CaM-PEG hydrogels. The resulting non-covalent interactions could prevent the hydrogels from recovering to 100% of their swollen volume. The physiochemical properties of PEG-CaM-PEG3400 hydrogels modulated VEGF absorption, and hydrogel dynamics had a clear influence on VEGF release. The isoelectric point of VEGF has been reported to be 8.5 25 and the isoelectric point of CaM has been reported to be 4.3 (Fig. 3B). 34 Thus, at the pH of our studies (pH ¼ 6.3) these molecules are likely to undergo significant charge–charge attraction. This interaction may largely explain the observation that VEGF absorption is linearly related to the 4 of PEG-CaM-PEG3400 hydrogels (Fig. 3A). Impor- tantly, the TFP-induced conformational change in PEG-CaM- PEG3400 molecules resulted in hydrogel collapse and
a concomitant increase in the VEGF release rate (Fig. 4A). The rate of VEGF release from 10% PEG-CaM-PEG3400 hydrogels was also increased when TFP was added 24 h after VEGF absorption, compared to hydrogels that were not exposed to TFP. This result indicates that an increase in VEGF release can be ‘‘triggered’’ via exposure to the TFP ligand (Fig. 4B). Interestingly, increased VEGF release in the presence of the TFP ligand was observed in every hydrogel formulation that included PEG-CaM-PEG3400 (Fig. 5A–B). Furthermore, PEG- CaM-PEG hydrogels released increasing total amounts of VEGF when the hydrogels could undergo greater dynamic hydrogel collapse (Fig. 2B–2C, 5A–5B). The increase in release may be attributed to an increase in fluid pressure within the hydrogel network upon volume collapse. This conclusion is supported by Download 0.87 Mb. Do'stlaringiz bilan baham: 
|