Collagen-inspired self-assembling materals

The rapid increase of the quality of life together with the progress of medical science asks for the development of new, tuneable and controllable materials. For the same reason, materials used for biomedical applications have to be increasingly biocompatible, biodegradable and biofunctional. Most of the available systems, however, lack one property or the other. For example, conventional animal-derived gelatin that is often used in biomedicine, is susceptible to a risk of contamination with prions or viruses and has a risk of bringing out allergic reactions, particularly against the non helix-forming domains of collagen [1]. Furthermore, gelatin is composed of a variety of molecules and structures with different thermal stabilities and molecular sizes. This, in combination with the impossibility to change the molecular structure at will, limits the chances to elucidate the relation between the structure and function. On the other hand, synthetic materials that have a rather well-controlled size distribution often lack biocompatibility, biofunctionality or biodegradability. In addition to that, as their synthesis often requires toxic solvents, their application in the human body is restricted. All the drawbacks of the presently used materials have brought scientists towards a new approach in designing materials viz. genetic engineering. Rapid progress in recombinant techniques has led to new ways of producing molecules with well-defined composition and structure and with full control over the length and sequence of the biopolymer and its constituent blocks. These methods thus combine the advantages of natural and synthetic polymers. Using molecular biology tools, unique molecules can be created by merging in a desired manner naturally occurring self-assembling motifs such as elastin, silk or collagen [2-4], or entirely artificial fragments. As we show in this thesis, the precise control over the molecular design of these biotechnologically produced block polypeptides is extremely valuable as it also leads to control over their physicochemical properties. In this thesis we present a new class of monodisperse, biodegradable and biocompatible network-forming block polymers that are produced by genetically modified strain of yeast, Pichia pastoris (Chapter 2). Trimer-forming end blocks, abbreviated as T, consisting of nine Pro-Gly-Pro amino acid triplets, are symmetrically flanking a random coil-like middle block composed of four or eight repeats of highly hydrophilic R or P sequences (Figure 8.1). R and P are identical with respect to length (99 amino acids) and composition but have different amino acid sequences. The P block has a glycine in every third position (as in collagen) but does not form any supramolecular structures and maintains a random coil-like conformation at any temperature [5]. The R block is a shuffled version of the P block. Four recombinant gelatins are reported in this thesis, denoted as TR4T, TR8T, TP4T and TP8T (Figure 8.1). All of these were successfully produced with high yields (1-3 g/l of fermentation broth) by the Pichia pastoris GS115 strain transformed with a pPIC9 vector with the gene of interest in its expression cassette. Figure 8.1 Schematic representation of collagen-inspired telechelic polypeptides: TR4T, TR8T, TP4T and TP8T. In Chapter 3 we described the linear rheological properties of hydrogels formed by TR4T polypeptides. At a temperature of 50 °C, the solution does not show any viscoelastic response. However, upon cooling, the collagen-like trimer-forming domains (T) start to assemble into triple helical nodes and a well-defined network, with a node multiplicity of three, is formed. In the beginning of the gelation process, viscous properties are predominant, but as the network formation progresses, the elastic properties prevail. A plateau storage modulus is reached within a few hours. At this point the triple helices are in equilibrium with the free T blocks. An equilibrium or near-equilibrium state is reached, contrary to natural gelatin, because the collagen-like (T) assembling domains are relatively short and well-defined. The T blocks are solely responsible for the network formation. We have shown that a solution of the middle blocks only (i.e. R4) does not demonstrate any elastic response at any time and temperature. In addition, differential scanning calorimetry (DSC) (Chapter 5) proved that the collagen-like side blocks are near-quantitatively responsible for trimerization, as the observed melting enthalpies are in good agreement with values obtained by Frank et al. [6] for free (Pro-Gly-Pro)10 peptides. The equilibrium fraction of T blocks involved in triple helices shifts with temperature. By lowering the temperature, the fraction of triple helices increases, while the fraction of free ends decreases. There are two possibilities to form a triple helix. It can be formed either by three T blocks from three different chains, or by three T blocks from two different chains, so that two side blocks come from the same polypeptide. As a consequence, the network is composed of dangling ends, elastically active bridges and inactive loops (Figure 8.2). Because of the precisely-known junction multiplicity of three, we could develop an analytical model that links the internal structure of the gel, with dangling ends, loops, and bridges, to the physicochemical properties. This model uses a limited set of input parameters that can all be measured independently. It describes the experimental data quantitatively without further adjustable parameters. Using this model, we could show that the observed strong dependency of the storage modulus, the relaxation time and the viscosity on concentration and temperature is related to the changes in the number of loops, active bridges, and dangling ends in the gel matrix. Figure 8.2 Network formation by collagen-inspired telechelic biopolymers. In Chapter 4 we show that the number of intermolecular junctions and intramolecular loops depends not only on protein concentration and temperature but also on the length and the stiffness of the middle block. We synthesised new triblock copolymers with middle blocks, of different lengths and amino acid sequences, named TP4T, TR8T and TP8T (Figure 8.1). For all new proteins, there is a strong dependency of the storage modulus, the relaxation time and the viscosity on concentration and temperature (as for TR4T). However at comparable molar concentrations, the longer versions of polypeptides i.e. TR8T and TP8T show a significantly higher storage modulus and relaxation time than their counterparts TR4T and TP4T. This is because a longer middle block leads to a larger radius of gyration (Rg), which decreases the probability that two end blocks from the same molecule associate with each other, and form a loop. The consequence of fewer loops in the system is a higher storage modulus and a higher overall relaxation time. TR8T TP8T Besides the effect of polymer length, we also observed that the R series, i.e. TR4T and TR8T, show a higher storage modulus than their P counterparts, i.e. TP4T and TP8T, at the same concentration and temperature. This can be explained by differences in coil flexibility. Although the P and R blocks have exactly the same amino acid composition, their amino acid sequence is different. Fitzkee et al. [7] have shown that even a polypeptide chain that assumes a random coil conformation still has locally folded conformations that contribute to the overall flexibility of the chain. This apparently leads to a smaller radius of gyration for the P middle block than for the R middle block and thus to a higher probability of loop formation. Even though the melting behaviour obtained with DSC is the same for all four polypeptides (as the end blocks stay the same), the temperature at which the G0 value approaches zero and the gel completely loses its elastic properties varies with the length of the middle block. Shorter molecules, i.e. TP4T and TR4T, melt at lower temperatures. A solution of 1.2 mM TP4T melts at 298 K, while TP8T at a comparable molar concentration melts at a temperature which is 15 degrees higher. Furthermore, the R versions show slightly higher melting temperatures than the P versions. These differences in melting behaviour are related to the gel structure and the relative probabilities of forming intramolecular and intermolecular assemblies. We could account for these findings with the help of the analytical model presented in Chapter 3. The only parameter that had to be varied in the model was the coil size of the polymer, since the enthalpy and the melting temperatures of the triple helices did not change with the length of the middle block. The theoretical calculations clearly show that the molecules with smaller Rg form up to 30 % more loops than their bigger counterparts. Loops that act as gel stoppers do not contribute to the network elasticity and significantly lower the melting temperatures detected with rheology. The network junctions in our gels are solely formed by triple helices. The mechanism of junction formation by the T blocks can be well-described by a two-step kinetic model (Chapter 5). Prior to triple helix propagation, a trimeric nucleus has to be formed. For dilute systems, nucleation is the limiting step, giving an apparent reaction order of three. These results indicate that only triple helices are stable. For more concentrated solutions, when nucleation is relatively fast, propagation of triple helices becomes rate-limiting and the apparent reaction order is close to unity. The propagation of triple helices is probably limited by cis-trans isomerization of peptide bonds, in which proline residues are involved. Above overlap concentration (C*) the measured enthalpy for stable gels (~15 hours) indicates that almost 100 % of the T blocks are involved in triple helices. Values obtained by us are in good agreement with values obtained by Frank et al. [6] for single (Gly-Pro-Gly)10 peptides. Conversely, at concentrations below C*, the enthalpy per mole of protein is becoming less, suggesting that the fraction of free ends or mismatched helices becomes more pronounced. The apparent melting temperature increases slightly with increasing concentration. This can be explained on the basis of the reaction stoichiometry under equilibrium conditions [8, 9]. Except for the highest measured concentration (2.4 mM), the apparent melting temperature revealed a dependence on the scan rate, indicating that it was not possible to maintain equilibrium during the heating step. At a concentration of 2.4 mM concentration there is no scan rate dependence, since the melting occurs at a higher temperature, where the dissociation kinetics is faster [4, 10]. The kinetics of triple helix formation determines the rate of gel formation. The gelation starts when the first triple helical node is formed. At that time viscous properties (loss modulus) predominate, but as the network formation evolves the elastic response (storage modulus) becomes more pronounced. The storage modulus (G’) reaches a plateau value within a few hours. Changes in network structure and mechanical properties of the gel in time can be predicted from the kinetics of triple helix formation, using the model presented in Chapter 3. By comparing the kinetics obtained with rheology and with DSC we could see that for our system, the helix content is not simply proportional to the network progress and that the relation between the elastic properties (G’) and the helix content (pH) depends on the protein concentration. The reason for this concentration dependence is the formation of loops, which is more likely at low concentrations. The investigated hydrogels undergo time-dependent macroscopic fracturing when a constant shear rate or shear stress is applied (start-up and creep experiments, respectively) (Chapter 6). Observations with particle image velocimetry (PIV) showed that in the beginning of a start-up (or creep) experiment the sample flows homogenously. After some time, the gel fractures, and is separated into two fractions. The inner region moves at the same velocity as the moving bob, while the outer fraction does not move at all. From the rate-dependence of the fracture strength we can conclude that gel fracture is due to stress-activated rupture of the triple helical nodes in the network.When the deformation is taken away, the gel can heal (Chapter 6). The capacity of self-healing is due to the transient character of the network nodes with a finite relaxation time. Such behaviour, impossible for most permanent gels, is highly desired in many applications, as hydrogels are often subjected to deformations, which easily go beyond the linear regime. As we present in Figure 8.3, TR4T gels cut into small pieces (grey and transparent), can heal within 2 hours. As measured with rheology the broken gel can recover up to 100 % of its initial elastic properties, even after several fracturing cycles. Interestingly, the kinetics of healing differs from the kinetics of fresh gel formation (Chapter 5). The latter is characterized by a lag-phase before elastic properties start to appear. This lag-phase occurs because at low degrees of crosslinking there is not yet a percolated network, so that the storage modulus is undetectable. By contrast, the recovery of the gel after rupturing is much faster and does not show a lag-phase. The elastic modulus, depending on the rupturing history, comes back to its initial value within 1-5 hours. These findings indicate that outside the fracture zone, the network nodes have not dissociated significantly, so that healing only requires the reformation of junctions that connect the undamaged pieces of the network (gel clusters). Figure 8.3 Self-healing of TR4T hydrogels. (A) Pieces of broken gel. (B) Two gel pieces healed after 2 hours. In Chapter 7 we demonstrated the shape-memory effects in hydrogels formed by permanently crosslinked TR4T molecules. The programmed shape of these hydrogels was achieved by chemical crosslinking of lysine residues present in the random coil. The chemical network could be stretched up to 200 % and “pinned” in a temporary shape by lowering the temperature and allowing the collagen-like end blocks to assemble into the physical nodes. The deformed shape of hydrogel can be maintained, at room temperature, for several days, or relaxed within few minutes upon heating to 50 ºC or higher. The presented hydrogels could return to their programmed shape even after several thermo-mechanical cycles, hence indicating that they remember the programmed shape. We have studied in more detail the shape recovery process by describing our hydrogels by a mechanical model composed of two springs and a dashpot. With the help of this model we showed that above the melting temperature of the triple helices, the recovery is exponential and that the decay time is roughly ten times slower than the relaxation of the physical network. 1.Biomedical applications - perspectives and considerations The class of collagen-inspired self-assembling materials, which we present in this thesis are nice model systems for a systematic study of physical networks, but they also have a lot of potential for biomedical applications. In this section we discuss the possibilities for these self- assembling hydrogels in biomedicine. Drug delivery systems One of the major goals of modern medicine is to ensure that the required amount of an active substance is available at the desired time at the desired location in the body. Consequently, a lot of effort is put into designing delivery systems with precisely adapted release profiles, sensitive to external stimuli such as temperature or pH. A frequently used group of materials in this field are hydrogels, both chemically and physically crosslinked. In the case of covalently crosslinked networks the release of the drug is mostly via diffusion of the drug out of the gel particle after it swells. The rate of drug release is governed by the resistance of the network to volume increase [11]. Although permanent networks are widely used as drug carriers they have some disadvantages such as incomplete release of active substances and poor biodegradability in the body. The problem can be partially solved by introducing enzymatic cleavage or hydrolysis sites into the main chain, but still the hydrogel erosion cannot be precisely controlled and complete material degradation can not be guaranteed. These obstacles can be overcome by using physical hydrogels that are formed by weak interactions. These can dissociate in a controlled manner and completely release the active component. In contrast to chemical gels, erosion of physical gels occurs spontaneously. The erosion rate is determined by the life time of the junctions, but it depends also on the relative amount of intramolecular loops and intermolecular junctions, as demonstrated by Shen et al. [12]. These authors showed that, by using triblock polymers with dissimilar coiled-coil side domains rather than identical ones, loop formation could be suppressed, leading to a lower erosion rate [12]. The potential of our gels for drug delivery applications was tested by Teles et al. [13]. It was shown that trapped proteins (BSA) can be completely released from TR4T and TR8T gels, both at 37 ºC and 20 ºC. The release at 37 ºC from 20 % gels was completed within 48 hours while at 20 ºC it took about 5 times longer. At body temperature the release was mostly driven by dissociation of trimeric junction and dissolution of the separate polymer chains (gel erosion). At 20 ºC the junction life time was long so that erosion was slower and swelling and diffusion played a more important role. The observations of Teles are in agreement with studies of several groups that demonstrated the importance of hydrogel erosion for controlled release [12, 14]. The erosion rate of physical hydrogels is governed by the junction relaxation time. The mean relaxation time of transient networks can be manipulated either by varying the gel architecture (Chapter 3 and 4) or by changing the relaxation time of a single triple helix. The gel architecture (i.e. the number of loops and bridges) can be altered either as we show in Chapter 3 by changing the protein concentration or as we demonstrate in Chapter 4 by manipulating the design of the middle block. The number of loops becomes lower as the spacer length and stiffness increase (Chapter 4). The lifetime of a single node can be changed by enzymatic hydroxylation of proline to hydroxyproline, [15] which leads to more hydrogen bonds among adjacent T blocks, or by changing the length of the collagen-like T domains. Preliminary results showed that average relaxation time of the network is roughly hounded times higher for molecules with collagen-like domains composed of sixteen Pro-Gly-Pro repeats instead of nine (unpublished data). For these biotechnologically produced collagen-inspired polymers, the length or the composition of the blocks can be changed simply by changing the DNA template. This, in combination with the model elaborated in Chapter 3 and 4 that links the internal gel architecture with the physicochemical gel behaviour, gives ample possibilities to design materials with custom-desired release profiles of active components. Tissue engineering Materials for tissue engineering scaffolds have to mimic the in vivo extracellular matrix environment. They provide physical support, but also have to guarantee proper adhesion of cells and controlled release of growth factors. A very important role in scaffolds design is played by the mechanical properties of the matrix [16-18]. As shown by Engler et al. [17], the elasticity of the matrix directs stem cell development to different lineages. Soft networks (0.1-1 kPa), which mimic brain tissue, promote neuron development, stiffer scaffolds (8-17 kPa) are myogenic, while gels with an elastic modulus of 24-40 kPa promote growth of bone cells. The stiffness of the matrix affects focal-adhesions and the organization of the cytoskeleton structure, and thus contractility, motility and spreading [16, 18]. Another significant factor, which plays a role in tissue growth is the degradation rate of the scaffold. The degradation should be synchronized with cellular repair in such a way, that tissue replaces the material within the desired time interval. The scaffold disintegration also controls the release of growth factors. For naturally derived materials such as alginate, the degradation rate could be influenced by partial oxidation of the polymer chain or via a bimodal molecular weight distribution [19]. For synthetic polymers different degradation profiles can be realized by incorporating in the polymer backbone groups with different susceptibility to hydrolysis [19]. Presently the most widely used scaffolds for tissue engineering are natural polymers such as collagen, gelatin, and polysaccharides [20] or synthetic, biodegradable polymers such as poly (L-lactic acid) (PLLA), poly(glycolic acid) (PGA), and poly(ethylene glycol) (PEG). [21-23]. Although these materials show promising properties, their use is limited as they suffer from batch to batch variations, polydispersity, viral contamination, allergic reactions or toxic byproducts after degradation. Also their mechanical properties are poorly-controlled and it is difficult to relate the molecular structure to the resulting properties. Furthermore, in the case of synthetic polymers, there is no intrinsic mechanism to interact with cells and to propagate cell adhesion proliferation or migration. This problem can be partially solved by functionalizing synthetic materials with bioactive molecules, such as collagen [24] or short peptides (for example arginine-glycine-aspartic (RGD) or tyrosine-isoleucine-glycine-serine-arginine (YIGSR) [25]). It remains difficult, however, to precisely control the spatial distribution, of these biofunctional domains [25]. A very promising alternative for the currently used scaffolds are hydrogels formed by self-assembling protein polymers [2, 26-28], including the collagen-inspired polypeptides presented in this thesis. Our block polymers form physical gels with precisely controlled elastic properties. As discussed in Chapter 3 and 4, the gel structure and the resulting mechanical properties strongly depend on concentration, temperature and on the molecular design of the polymer. Within the investigated range of conditions our gels have an elastic modulus between 0.03 and 5 kPa. Thus they seem most appropriate for neuron cell growth [17]. Moreover, it is also possible to incorporate specific short adhesive peptide sequences (such as RGD) in the middle block to improve attachment and cell propagation. The presently investigated proteins, with T domains composed of nine Pro-Gly-Pro repeats, still need some enhancement in terms of stability. As shown by Teles et al. the currently available molecules erode within 2 days [13]. For tissue engineering applications, this is too fast. We therefore propose some strategies to stabilize our hydrogels. A first possibility alternative is to introduce amino acids which can form chemical bonds such as cysteines that can form disulfide bridges under oxidizing conditions [29], or lysines, which can be functionalized with acrylate and then photo-crosslinked with UV radiation [30-32]. However, one has to be aware that this additional procedure may have negative side effects such as toxic byproducts, incomplete polymer degradation in the body, or loss of responsiveness to external stimuli. Alternatively, the erosion can be moderately slowed down by increasing the relaxation time of the network (as discussed in section on drug delivery systems). Wound dressing materials Under normal circumstances wound healing is a very long process. In order to speed it up, so that bacterial infections or wound dehydration can be avoided, wound dressing materials are used [33-37]. These materials should fulfil several general requirements such as biocompability, ease of application and removal, proper adherence (to avoid fluid pockets, in which bacteria could proliferate), ease in gas exchange between tissue and environment, and controlled release of active components such as antimicrobial agents or wound repair agents (for example Epidermal Growth Factor (EGF)) [37]. All above-mentioned requirements can be fulfilled by the collagen-inspired hydrogels presented in this thesis. The advantage of our materials is that they can follow the contour of the wound and entirely fill it, thus forming an efficient barrier for microbes, but at the same time being permeable for water vapour and oxygen. Furthermore they can entrap active components and release them in a controlled way during the healing process, as discussed above. Depending on the circumstances, the release profile can be synchronised with the wound healing process. An additional advantage of our genetically engineered molecules is that adhesion domains can be introduced along the middle block, assuring better integration of the gel with the damaged tissue. 8.3 Final conclusions and outlook In this final chapter we have discussed the potential of our collagen-inspired materials in biomedical applications. They are biocompatible and biodegradable, whilst offering numerous possibilities to change the molecular design in order to meet the desired mechanical or biological properties. Furthermore, the well-defined nature of the triple helical junctions allows us to predict the mechanical properties of the gel from the molecular design of polypeptides. This exclusive feature of our system makes it unique and offers great flexibility to design custom biomedical materials. Biomedical needs, however, are very variable and often require an individual approach. That is why in our group we have created a family of genetically engineered block copolypeptides. Besides collagen we use other motifs present in nature, such as silk or elastin. We can combine these motifs in various ways in order to create unique stimuli-responsive (often multi-responsive) molecules that can meet individual application needs. The silk-like domains consist of (Gly-Ala-Gly-Ala-Gly-Ala-Gly-Xxx)n repeats. Position Xxx is occupied by charged amino acids such as histidine, lysine or glutamic acid. When the charge is screened, the molecules assemble, forming first β sheet-like secondary structures, and then long fibres. As shown by Martens et al. [38], block polymers comprising silk-like domains with glutamic acid or histidine in the Xxx position form fibre-like gels at a pH of 2 or 12, respectively. They also assemble when mixed with oppositely charged (coordination) polymers [3, 38]. Probably, the assembling conditions can be tuned even more precisely by adjusting the isoelectric point of the assembling domain. This will allow the production of hydrogels that are formed after being injected into the body, while they disassemble (releasing the drug) when exposed to the acidic or the alkaline conditions. The nanofibre gels are also stable enough to serve as scaffolds for tissue engineering [39-41]. Another motif that has been used is elastin. It consists of (Val-Pro-Gly-Xxx-Gly)n repeats and it self-assembles above a lower critical solution temperature (LCST). The transition temperature can be tuned by introducing more or less polar amino acid residues in position Xxx. By combining elastin-like or collagen-like blocks with silk-like blocks, thermo and pH responsive networks can be obtained. This may allow us toswitch from fibre-like gels to associative networks. Block polypeptides, produced using recombinant techniques, besides biocompability and biodegradability offer many possibilities to adjust the molecular design in will, to realize the desired mechanical or biological properties. Three dimensional structures with different thermal stabilities can be programmed by combining in a precise manner various amino acid sequences. The obtained materials can respond to external stimuli such as pH, ionic strength or temperature. They can also carry peptides fragments that can enhance cells adhesion and proliferation or induce crystallization. The new approach in material science, which we present in this thesis, opens a new world of polymers, in which the main constraint is imagination. References [1] European Commission, Updated opinion on the safety with regards to TSE risks of gelatine derived from ruminant bones or hides. (2003). [2] E.R. Wright, V.P. Conticello, Self-assembly of block copolymers derived from elastin-mimetic polypeptide sequences. Advanced Drug Delivery Reviews 54(8) (2002) 1057-1073. [3] A.A. Martens, J. van der Gucht, G. Eggink, F.A. de Wolf, M.A.C. Stuart, Dilute gels with exceptional rigidity from self-assembling silk-collagen-like block copolymers. Soft Matter 5(21) (2009) 4191-4197. [4] P.J. Skrzeszewska, F.A. de Wolf, M.W.T. Werten, A.P.H.A. Moers, M.A. Cohen Stuart, J. van der Gucht, Physical gels of telechelic triblock copolymers with precisely defined junction multiplicity. Soft Matter 5(10) (2009) 2057-2062. [5] M.W.T. Werten, H. Teles, A. Moers, E.J.H. Wolbert, J. Sprakel, G. Eggink, F.A. de Wolf, Precision Gels from Collagen-Inspired Triblock Copolymers. Biomacromolecules 10(5) (2009) 1106-1113. [6] S. Frank, R.A. Kammerer, D. Mechling, T. Schulthess, R. Landwehr, J. Bann, Y. Guo, A. Lustig, H.P. Bachinger, J. Engel, Stabilization of short collagen-like triple helices by protein engineering. Journal of Molecular Biology 308(5) (2001) 1081-1089. [7] N.C. Fitzkee, G.D. Rose, Reassessing random-coil statistics in unfolded proteins. Proceedings of the National Academy of Sciences of the United States of America 101(34) (2004) 12497-12502. [8] J. Engel, H.T. Chen, D.J. Prockop, H. Klump, Triple helix reversible coil conversion of collagen-like polypeptides in aqueous and non aqueous solvents - comparison of thermodynamic parameters and binding of water to (L-Pro-L-Pro-Gly)n and (L-Pro-L-Hyp-Gly)n Biopolymers 16(3) (1977) 601-622. [9] A.V. Persikov, Y.J. Xu, B. Brodsky, Equilibrium thermal transitions of collagen model peptides. Protein Science 13(4) (2004) 893-902. [10]S. Boudko, S. Frank, R.A. Kammerer, J. Stetefeld, T. Schulthess, R. Landwehr, A. Lustig, H.P. Bachinger, J. Engel, Nucleation and propagation of the collagen triple helix in single-chain and trimerized peptides: Transition from third to first order kinetics. Journal of Molecular Biology 317(3) (2002) 459-470. [11]P. Gupta, K. Vermani, S. Garg, Hydrogels: from controlled release to pH-responsive drug delivery. Drug Discovery Today 7(10) (2002) 569-579. [12]W. Shen, K.C. Zhang, J.A. Kornfield, D.A. Tirrell, Tuning the erosion rate of artificial protein hydrogels through control of network topology. Nature Materials 5(2) (2006) 153-158. [13]H. Teles, T. Vermonden, G. Eggink, W.E. Hennink, F.A. de Wolf, Hydrogels of collagen-inspired telechelic triblock copolymers for sustained release of proteins. Journal of Controlled Release 147(2) (2010) 298-303. [14]K.S. Anseth, A.T. Metters, S.J. Bryant, P.J. Martens, J.H. Elisseeff, C.N. Bowman, In situ forming degradable networks and their application in tissue engineering and drug delivery. Journal of Controlled Release 78(1-3) (2002) 199-209. [15]R.E. Rhoads, Udenfrie.S, Bornstei.P, In vitro enzymatic hydroxylation of prolyl residues in alpha1-CB2 fragment of rat collagen. Journal of Biological Chemistry 246(13) (1971) 4135-&. [16]D.E. Discher, P. Janmey, Y.L. Wang, Tissue cells feel and respond to the stiffness of their substrate. Science 310(5751) (2005) 1139-1143. [17]A.J. Engler, S. Sen, H.L. Sweeney, D.E. Discher, Matrix elasticity directs stem cell lineage specification. Cell 126(4) (2006) 677-689. [18]R.J. Pelham, Y.L. Wang, Cell locomotion and focal adhesions are regulated by substrate flexibility. Proceedings of the National Academy of Sciences of the United States of America 94(25) (1997) 13661-13665. [19]G. Chan, D.J. Mooney, New materials for tissue engineering: towards greater control over the biological response. Trends in Biotechnology 26(7) (2008) 382-392. [20]Y.C. Wang, L.B. Wong, H. Mao, Creation of a long-lifespan ciliated epithelial tissue structure using a 3D collagen scaffold. Biomaterials 31(5) 848-853. [21]M. Martina, D.W. Hutmacher, Biodegradable polymers applied in tissue engineering research: a review. Polymer International 56(2) (2007) 145-157. [22]B.S. Kim, D.J. Mooney, Engineering smooth muscle tissue with a predefined structure. Journal of Biomedical Materials Research 41(2) (1998) 322-332. [23]B.S. Kim, D.J. Mooney, Development of biocompatible synthetic extracellular matrices for tissue engineering. Trends in Biotechnology 16(5) (1998) 224-230. [24]A.J. Engler, M.A. Griffin, S. Sen, C.G. Bonnetnann, H.L. Sweeney, D.E. Discher, Myotubes differentiate optimally on substrates with tissue-like stiffness: pathological implications for soft or stiff microenvironments. Journal of Cell Biology 166(6) (2004) 877-887. [25]L.Y. Koo, D.J. Irvine, A.M. Mayes, D.A. Lauffenburger, L.G. Griffith, Co-regulation of cell adhesion by nanoscale RGD organization and mechanical stimulus. Journal of Cell Science 115(7) (2002) 1423-1433. [26]R.E. Sallach, W.X. Cui, F. Balderrama, A.W. Martinez, J. Wen, C.A. Haller, J.V. Taylor, E.R. Wright, R.C. Long, E.L. Chaiko, Long-term biostability of self-assembling protein polymers in the absence of covalent crosslinking. Biomaterials 31(4) (2010) 779-791. [27]W. Shen, J.A. Kornfield, D.A. Tirrell, Structure and mechanical properties of artificial protein hydrogels assembled through aggregation of leucine zipper peptide domains. Soft Matter 3(1) (2007) 99-107. [28]J.S. Guo, K.K.G. Leung, H.X. Su, Q.J. Yuan, L. Wang, T.H. Chu, W.M. Zhang, J.K.S. Pu, G.K.P. Ng, W.M. Wong, X. Dai, W.T. Wu, Self-assembling peptide nanofiber scaffold promotes the reconstruction of acutely injured brain. Nanomedicine-Nanotechnology Biology and Medicine 5(3) (2009) 345-351. [29]W. Shen, R.G.H. Lammertink, J.K. Sakata, J.A. Kornfield, D.A. Tirrell, Assembly of an artificial protein hydrogel through leucine zipper aggregation and disulfide bond formation. Macromolecules 38(9) (2005) 3909-3916. [30]S.A. Maskarinec, D.A. Tirrell, Protein engineering approaches to biomaterials design. Current Opinion in Biotechnology 16(4) (2005) 422-426. [31]N. Sanabria-DeLong, A.J. Crosby, G.N. Tew, Photo-Cross-Linked PLA-PEO-PLA Hydrogels from Self-Assembled Physical Networks: Mechanical Properties and Influence of Assumed Constitutive Relationships. Biomacromolecules 9(10) (2008) 2784-2791. [32]J.A. Benton, C.A. DeForest, V. Vivekanandan, K.S. Anseth, Photocrosslinking of Gelatin Macromers to Synthesize Porous Hydrogels That Promote Valvular Interstitial Cell Function. Tissue Engineering Part A 15(11) (2009) 3221-3230. [33]K.J. Quinn, J.M. Courtney, J.H. Evans, J.D.S. Gaylor, W.H. Reid, Principles of burn dressing. Biomaterials 6(6) (1985) 369-377. [34]S.B. Lee, Y.H. Kim, M.S. Chong, S.H. Hong, Y.M. Lee, Study of gelatin-containing artificial skin V: fabrication of gelatin scaffolds using a salt-leaching method. Biomaterials 26(14) (2005) 1961-1968. [35]S.R. Hong, S.J. Lee, J.W. Shim, Y.S. Choi, Y.M. Lee, K.W. Song, M.H. Park, Y.S. Nam, S.I. Lee, Study on gelatin-containing artificial skin IV: a comparative study on the effect of antibiotic and EGF on cell proliferation during epidermal healing. Biomaterials 22(20) (2001) 2777-2783. [36]B. Balakrishnan, M. Mohanty, P.R. Umashankar, A. Jayakrishnan, Evaluation of an in situ forming hydrogel wound dressing based on oxidized alginate and gelatin. Biomaterials 26(32) (2005) 6335-6342. [37]A. Schneider, J.A. Garlick, C. Egles, Self-Assembling Peptide Nanofiber Scaffolds Accelerate Wound Healing. Plos One 3(1) (2008). [38]A.A. Martens, G. Portale, M.W.T. Werten, R.J. de Vries, G. Eggink, M.A.Cohen Stuart, F.A. de Wolf, Triblock Protein Copolymers Forming Supramolecular Nanotapes and pH-Responsive Gels. Macromolecules 42(4) (2009) 1002-1009. [39]S.G. Zhang, F. Gelain, X.J. Zhao, Designer self-assembling peptide nanofiber scaffolds for 3D tissue cell cultures. Seminars in Cancer Biology 15(5) (2005) 413-420. [40]F. Zhang, G.S. Shi, L.F. Ren, F.Q. Hu, S.L. Li, Z.J. Xie, Designer self-assembling peptide scaffold stimulates pre-osteoblast attachment, spreading and proliferation. Journal of Materials Science-Materials in Medicine 20(7) (2009) 1475-1481. [41]F. Gelain, D. Bottai, A. Vescovi, S.G. Zhang, Designer Self-Assembling Peptide Nanofiber Scaffolds for Adult Mouse Neural Stem Cell 3-Dimensional Cultures. Plos One 1(2) (2006).