Biomaterials 31 (2010) 641–647
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
Hierarchical scaffolds via combined macro- and micro-phase separation Peter A. George, Katie Quinn, Justin J. Cooper-White* School of Engineering and Australian Institute for Bioengineering and Nanotechnology, University of Queensland, Bld 75, Cnr Cooper and College Rds, QLD 4072, Australia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 June 2009 Accepted 29 September 2009 Available online 17 October 2009
Recent advances in biomaterial surface engineering have shown that surface biomechanical, spatial and topographical properties can elicit control over fundamental biological processes such as cell shape, proliferation, differentiation and apoptosis. Along these lines, we have very recently shown that the selfassembly of block copolymers into thin films can be used as an extremely labile method to precisely position cellular adhesion molecules, at nanometre lateral spacings, to effect control over cell attachment and morphology. Here, we extend our work in 2-dimensional block copolymer films into the production of 3-dimensional porous block copolymer scaffolds. The reported method combines macro-scale temperature induced phase separation and micro-phase separation of block copolymers to produce highly porous scaffolds with surfaces comprised of nano-scale self-assembled block copolymer domains, representing a significant advance in currently available scaffold engineering technologies. The phase behaviour of these polymer–solvent systems is described and potential mechanisms leading to the observed structure formation are presented. The nano-domains have thereafter been functionalised with CGRGDS peptides throughout the scaffold and shown to effect changes in cell attachment and spreading, in agreement with previous 2-dimensional studies. These multi-scale, functional scaffolds are easy to manufacture and scaleable, making them ideal candidates for tissue engineering applications. Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
Keywords: Nanotechnology Integrin clustering Self assembly Surface modification Block copolymer
1. Introduction The development of 3-dimensional scaffolds for the controlled generation of authentic tissue structures and their delivery to the body in order to replace, or augment, the behaviour of diseased or damaged tissues is the ultimate goal of tissue engineering. These scaffolds define the shape of the resultant tissue and should ideally provide both a porous 3-dimensional structure capable of facilitating and ultimately controlling the attachment, growth and directing cell fate decisions, such as, lineage specification [1,2]. Unfortunately, despite offering significant promise, synthetic polymeric scaffolds have thus far had limited success in tissue engineering applications [3]. One potential reason for the limited success of polymeric scaffolds is that current synthetic polymers are not biochemically familiar to cells. In vivo, cells interact predominantly through integrins, as well as other receptors, with an extracellular matrix made mostly of collagen, elastin, glycoproteins, proteoglycans, laminin and fibronectin [4]. This interaction is controlled with incredible precision, for example, collagen fibres interact with
* Corresponding author. Tel.: þ61 7 3346 3858; fax þ61 7 3346 3973. E-mail address:
[email protected] (J.J. Cooper-White).
a spatial periodicity of 67 nm with tissue cells in vivo [5], and is critical for many cellular functions, such as regulation of cell morphology, migration, growth, differentiation and apoptosis. The recent outstanding success of de-cellularised tissues as a 3-dimensional scaffold for regeneration highlights the need for enhanced biomimicry in synthetic polymer scaffolds [6]. However, naturally derived materials can still elicit an immune response, are prohibitively expensive to produce, and also suffer from batch to batch variations. There thus remains a significant need for the development of a scaffold processing methodology for synthetic scaffolds that a) can present molecular signals throughout a scaffold structure in appropriate spatial and temporal manners, so that individual cells are motivated to form desired tissue structures, and b) can be carried out reproducibly, economically, and on a large scale [7]. In this work we present a method of scaffold manufacture that combines macro and micro-phase separation processes to generate scaffolds where the precise nano-scale positioning of adhesion motifs (or other biological molecules) can be controlled throughout the 3-dimensional porous construct. The process is simple and reproducible, as it relies on self-assembly, as well as scaleable and economical, thus making it an ideal technology for large-scale tissue engineering applications. By combining temperature induced phase separation (TIPS) with block copolymer self-
0142-9612/$ – see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2009.09.094
642
P.A. George et al. / Biomaterials 31 (2010) 641–647
assembly, highly porous scaffolds are produced that have incredibly complex surface structures (up to 1013 discrete domains per ml of scaffold volume). In this study we have utilised model nondegradable polystyrene based polymers, however, the behaviour described can be extended to biodegradable polymeric systems. This work allows for the translation of recent insights (in 2-dimensional formats) into the effects of cell shape, adhesive ligand spacing, and focal adhesion formation on controlling cell behaviour, such as engineered 2-dimensional surfaces for directing stem cell differentiation [8–13], into porous 3-dimensional constructs. 2. Materials and methods 2.1. Materials Polystyrene-block-polyethylene oxide (PS–PEO), polystyrene-block-polyacrylic acid (PS–PAA) and polystyrene (PS) polymers were purchased from Polymer Source Pty. Ltd. (Montreal, PQ, Canada) (properties listed in Table 1). HPLC grade toluene was purchased from Aldrich and used as supplied. AR grade dioxane, dimethyl carbonate and cyclohexane were purchased from Aldrich and stored over molecular sieve prior to use. Silicon wafers (100 orientation, Boron doped) were purchased from Micro Materials and Research Consumables Pty. Ltd (Victoria, Australia). N-(pmaleimidophenyl) isocyanate (PMPI) was purchased from Pierce (Rockford, IL, USA). CGRGDS peptide was supplied by Peptide 2.0 Inc. (Chantilly, VA, USA). Dulbecco’s modified Eagle’s medium (DMEM), phosphate buffered saline (PBS), foetal bovine serum (FBS), penicillin, streptomycin and trypsin-ethylenediaminetetraacetic acid (EDTA) were purchased from Gibco (Mt Waverly, Australia). Hoechst, fluorescein isothiocyanate labeled phalloidin (FITC-phalloidin) and Triton-X 100 were sourced from Sigma–Aldrich. All other materials were purchased from Sigma and used as supplied.
measured cloud point temperatures. This information was then used with the Flory– Huggins model to determine the polymer-solvent phase diagram. The relationships for the binodal and spinodal curves are described by Eq. (1) and (2) respectively. vGm 0 vGm 00 ¼ : vf vf v2 Gm vf2
¼
1 1 þ 2c ¼ 0: m1 f1 m2 f2
(1)
(2)
where, Gm is Gibbs free energy of mixing, f1 & f2 are the volume fractions of the solvent and polymer respectively, and c is the Flory–Huggins interaction parameter. 2.4.1. Temperature induced phase separation Highly porous inter-connected scaffolds were generated by temperature induced phase separation of a polymer–solvent solution. Scaffold formation occurs via the separation of the solution into a polymer rich and polymer poor phase. For dioxane solutions, this process is driven by solvent crystallization, whereas for dimethyl carbonate and cyclohexane solutions, the process involves a liquid-liquid separation. 0.075 volume fraction solutions of polymer in solvent were heated to 50 C before being placed in a controlled temperature water bath at 10 C and held for 10 min. The solutions were then transferred to liquid nitrogen for 2–3 min to freeze in the structure and stored in the freezer at 20 C before further processing. The solvent was removed from the scaffold by vacuum drying at w103 mbar for 6 h.
2.5. RGD immobilisation to functional nano-domains A cysteine terminated GRGDS peptide (CGRGDS), containing the well-known RGD adhesive sequence present in many extracellular matrix molecules (such as fibronectin and collagens), was bound to surface maleimide groups on the PS–PEOMal scaffold. 100 ug/ml of peptide in coupling buffer (0.1 M sodium phosphate, 0.15 M sodium chloride, 10 mM EDTA, pH 7.2) was incubated on the surface for 2 h at room temperature. Surfaces were then thoroughly rinsed in PBS.
2.2. Maleimide functionalisation of PS–PEO polymer Following the method of George et al. [8], 100 mg of LMW PS–PEO copolymer and 8 mg of PMPI (greater than 20 fold molar excess) were dissolved by heating in 2 ml of dry N,N-dimethylformamide (DMF) (pH 8.5) and left to react for 1 h. The solvent was then removed by rotor-evaporation; the solid product was re-dissolved in chloroform and precipitated into methanol. The precipitate was recovered by vacuum filtration and washed extensively with methanol and water to remove any un-reacted PMPI, before being dried in a vacuum oven. The presence of the maleimide end group was confirmed using 1H NMR as previously reported [8]. 2.3. Determination of phase diagrams and scaffold manufacture 2.3.1. Cloud point measurements Cloud point measurements were performed to determine the boundary of phase separation (binodal). Cloud points were determined by visual turbidity. Polymer solutions for samples of PS and both PS–PEO and PS–PAA at each of three homopolymer/copolymer ratios (0:100, 50:50, 95:5) and at each of five total polymer concentrations (0.5, 2, 5 and 10% wt/vol) in each of three different solvents (dioxane, cyclohexane, dimethylcarbonate) were heated to 50 C in sealed glass vials. The solutions were then placed into a controlled temperature water bath and the temperature was decreased by 1 C every 10 min. The turbidity of the solutions was recorded at each temperature. Dioxane-polymer solutions crystallized prior to clouding. 2.4. Determination of phase diagram The full phase diagram for the polymer–solvent systems was estimated using a Flory–Huggins lattice model [9]. The temperature dependence of the interaction parameter for each system was determined by the line of best fit of the Flory– Huggins polymer–solvent interaction parameter (c) vs. 1/Tcloud point from the Table 1 Polymer properties. Polymer Label
PS PS–PAA LMW PS–PEO MMW PS–PEO HMW PS–PEO
Diblock copolymer Mn (kDa) PS-block
PEO or PAA-block
Total
202 42 51 102 190
– 4.5 11.5 34 48
202 46.5 62.5 136 238
Polydispersity index
1.05 1.18 1.05 1.18 1.07
2.6. Thin film generation Thin films of polymer were prepared by spin casting onto silicon wafers as previously described [10]. Silicon wafers were cut into 2 cm squares before being rinsed thoroughly in acetone, followed by isopropanol and dried under a stream of nitrogen. The wafers were then exposed to UV/ozone for 10 min to remove any remaining organics and to generate a uniform silicon oxide (SiO2) surface layer. Wafers were then rendered hydrophobic by boiling in benzyl alcohol (BnOH) for 4 h, rinsed thoroughly in isopropanol and dried under a stream of nitrogen. Thin films were generated by spin casting polymer solutions containing 1% wt/vol total polymer concentration in toluene onto the treated wafers at 3000 rpm.
2.7. Scaffold and thin film characterisation 2.7.1. Scanning electron microscopy (SEM) SEM imaging was performed using a JEOL JSM6300F microcscope. Prior to imaging scaffolds were coated with a 10 nm layer of platinum. 2.7.2. Atomic force microscopy (AFM) Thin block copolymer films spin cast onto silicon wafers were assessed via AFM. AC mode AFM was performed using an Asylum Research (Santa–Barbara, USA) Molecular Force Probe (MFP-3D). Imaging in air was obtained using cantilevers from Budget Sensors (Bulgaria) with a nominal spring constant of 42 N/m. Films were imaged at room temperature. The AFM is mounted on an anti-vibrational table (Herzan) and operated within an acoustic isolation enclosure (TMC, USA).
2.8. Cell culture, staining and imaging NIH-3T3 fibroblasts were cultured in growth media, consisting of DMEM þ 10% FBS supplemented with penicillin and streptomycin, and passaged using standard techniques. Prior to cell culture experiments 12 mm diameter scaffolds were cut into 1 mm thick sections, sterilized by immersion in 70% ethanol for 72 h, and rinsed thoroughly in PBS and DMEM. For cell culture experiments cells were seeded on polymer scaffolds at a density of 2,000,000 cells/cm3. To ensure efficient seeding of the scaffolds cells were delivered directly to the top surface of the scaffold in 100 ml of media and allowed to attach for 2 h before additional media was placed in the well-plate. This process is utilized in our lab to maximize the cell infiltration into the scaffold. Cells were cultured on the scaffolds for 24 h in growth media. Cells were fixed in 4% para-formaldehyde in PBS for 10 min at room temperature, permeated with Triton X-100 and stained with Hoechst and FITC-phalloidin. Cells were imaged using an Olympus BX61 fluorescent microscope.
P.A. George et al. / Biomaterials 31 (2010) 641–647
3. Results and discussion 3.1. Block copolymer–solvent phase behaviour Diblock copolymers are two polymers covalently linked together. As the interactions between the polymer segments increases, either by lowering the temperature or increasing the block copolymer concentration, the two polymers self-assemble into ordered 10–100 nm domains of each component polymer. The ordering process occurs to minimize the free energy of the system. This process is a trade-off between a reduction in the number of unfavourable enthalpic contacts between component blocks and the accompanying reduction in entropy due to chain ordering [11]. In this study we have selected two strongly segregating block copolymer systems, PS–PEO, a non-ionic polymer, and PS–PAA, where the acrylic acid block is negatively charged. Both polymers have an asymmetric structure, such that the polystyrene block occupies a larger volume than the corresponding PEO or PAA block. This leads to the formation, via self-assembly, of cylindrical or spherical domains of the minor component (PEO or PAA). Three solvents were selected based on the criteria that (1) they had high melting points, and thus could be easily incorporated into a temperature induced phase separation (TIPS) process, and (2) they were sufficiently volatile to be removed by vacuum drying. Based on these criteria dioxane, dimethyl carbonate and cyclohexane were selected (Properties listed in Table 2). The temperature versus polymer volume fraction phase diagrams were constructed by applying a Flory–Huggins lattice model to measured cloud points of the solutions. The Flory–Huggins lattice model was used specifically to predict the binodal and spinodal curves for each system (Fig. 1). Importantly, to ensure the best possible prediction of the spinodal envelope, cloud points were taken over a concentration range which allowed elucidation of the critical point. Block copolymers were blended with PS homopolymer, resulting in mixtures of polymer containing 0, 5, 50 and 100% block copolymer. No cloud points were observed in the polymer–dioxane systems prior to solvent freezing and hence their phase behaviour in terms of a binary phase diagram is not shown. PS–PAA in cyclohexane was clouded at all temperatures below the boiling point of cyclohexane and therefore its phase diagram is also not shown. Fig. 1 indicates that the phase behaviours of the shown polymer solvent systems are strongly affected by the addition of block copolymer. The PS–PEO-dimethyl carbonate phase boundary moves to decreasing temperatures with increasing addition of block copolymer, whereas for both the PS–PEO-cyclohexane and PS–PAA-dimethyl carbonate systems, the phase boundary moves to increasing temperatures with the increasing addition of block copolymer. A potential explanation for this dual behaviour is due to the different interaction parameters between the added PEO and PAA polymers and the solvent. The polymer–solvent interaction parameters can be estimated based on Hildebrand solubility parameters [12], and these values are summarized in Fig. 1 (Note: parameters are approximate and do not consider the effects of hydrogen bonding). Depending on the nature of the interactions, during phase separation events block copolymers can segregate to
Table 2 Solvent properties. Solvent
Melting point ( C)
Boiling point ( C)
Vapour pressure (mbar)
Cyclohexane Dimethyl carbonate 1,4-Dioxane
6.5 2–4 11.8
80.7 90 101.1
130 26.7 53.3
643
the newly formed interfaces between polymer rich and polymer poor phase separated regions. At these interfaces they act as compatibilisers, reducing the interfacial tension between the phases, and reducing the driving force for phase-separation. The lowering of the phase boundary for the PS–PEO-dimethyl carbonate system suggests that PS–PEO copolymers are segregating to the polymer rich-polymer poor (solvent) interface, compatibilising the two-phases and thus reducing the driving force for phase separation. In contrast, for the PS–PEO-cyclohexane and PS–PAA-dimethyl carbonate systems, the addition of block copolymer is driving phase-separation of the system. This is likely due to the minor component of the block copolymers (PEO and PAA) having strong non-favourable interactions with the solvent and during cooling of the polymer–solvent solution, the minor component polymers are forming sites for the nucleation and growth of the phase-separated structure (this phenomenon has been previously reported [13]). Whether this particular change in phase behaviour affects the capability of the PS–PEO or PS–PAA block copolymers to self assemble at the polymer rich-polymer poor interface (which is required for TIPS produced structures in these particular solvents), will be discussed in the proceeding section. 3.2. Scaffold structure Utilising the determined phase diagrams for each system, scaffolds of diblock copolymers PS–PEO and PS–PAA, as well as blends of 50% block copolymer and 50% polystyrene homopolymer, were manufactured by TIPS from 0.075 total polymer volume fraction solutions. The polymer volume fraction of 0.075 was selected for scaffold manufacture as this concentration was close to the critical point on the phase diagram for all systems, resulting in direct spinodal decomposition upon cooling (Fig. 1). Spinodal decomposition ensures the scaffolds have a highly inter-connected porous structure suitable for tissue engineering applications. Scaffolds were manufactured from the three chosen solvents by inducing either a solid–liquid phase separation (dioxane solutions) or a liquid-liquid phase separation (dimethyl carbonate, cyclohexane), followed by freezing the resultant polymer structure in place by quenching in liquid nitrogen. The solvent was thereafter removed under vacuum. 3.2.1. Solid–liquid phase separation Fig. 2 shows electron micrographs of the scaffolds manufactured from dioxane solutions. The resultant macro-structure of the scaffolds is highly porous with inter-connected pores. At higher magnification, the surface structure of the walls of the scaffold can be seen to consist of highly ordered arrays of micro-phase separated block copolymer for all polymer–dioxane systems investigated. For 100% PS–PEO the surface is decorated with highly ordered micro-phase separated nano-domains of PEO having a mean diameter of 14.8 nm (dark circles). This architecture is consistent with that observed in 2-dimensional thin films [10,14]. The 50% PS–PEO surface shows less dense, but larger PEO domains (mean diameter 26.4 nm). This structure is not consistent with rapidly self-assembled thin films of the same polymer system produced via spin casting. In the 2-dimensional films the size of the PEO domains decreases with the addition of PS homopolymer. This result suggests that although the resultant structures are similar, the kinetics of structure formation is different in this solid–liquid phase separation process compared to thin film formation. The PS–PAA block copolymer scaffolds also show spherical nano-sized domain morphology of PAA in a background of PS (dark circles are PAA domains, mean diameter of 100% PS–PAA is 14.0 nm and 50% PS–PAA is 12.4 nm). Similar PS–PAA polymers have been
644
P.A. George et al. / Biomaterials 31 (2010) 641–647
Fig. 1. Polymer–solvent phase diagrams. Solid symbols represent cloud point measurements. Solid lines are binodal curves, dashed lines are spinodal curves. Polymer–solvent interactions parameters estimated from Hildebrand coefficients d and molar volume V are summarised in table (lower right).
shown to generate spherical domains within the bulk of spin cast films [14], however, they are not presented at the surface as observed here. Interestingly, the size of the PAA domains displayed on the surfaces of these scaffolds does not seem to be significantly perturbed (14.0 nm vs 12.4 nm domain size) by the addition of 50% PS homopolymer to the PS–PAA block copolymer solution. The results from the scaffolds manufactured from dioxane indicate that although the structures formed by the block copolymer self-assembly process, which are governed by the intrinsic
properties of the polymers, such as molecular asymmetry, are similar to those observed in 2-dimensions, there are differences in the presentation and kinetics of structure formation in this combined macro–micro-phase separation process. The mechanism for structure formation in this solid–liquid phase separation process is believed to be as follows; as the solvent is crystallizing the polymer remains in the ever diminishing liquid phase, leading to increased local concentration of block copolymer and thus to increasing non-favourable interactions between the minor
Fig. 2. SEM micrographs of scaffolds manufactured from dioxane solutions. Top panel is low magnification images showing porous macro-scale structure of scaffolds. Lower panel is high magnification image showing surface structure of scaffolds.
P.A. George et al. / Biomaterials 31 (2010) 641–647
645
Fig. 3. SEM micrographs of scaffolds manufactured from dimethyl carbonate solutions. Top panel is low magnification images showing porous macro-scale structure of scaffolds. Lower panel is high magnification image showing surface structure of scaffolds.
component block (PEO or PAA) and the major component block (PS) and/or the homopolymer (PS). These interactions actively promote displacement of the block copolymer towards the nascent, but maturing polymer–solvent interface, resulting in a higher localized concentration of block copolymer at these interfaces prior to the solidification of the macro-structures at the completion of solvent crystallisation. This dynamic process thus results in the formation of a highly porous 3-dimensional scaffold structure displaying on its surface highly ordered block copolymer nano-domains. 3.2.2. Liquid-liquid phase separation and quenching Fig. 3 shows SEM micrographs of scaffolds manufactured from dimethyl carbonate solutions. The macro-architecture is obviously different from that observed in the dioxane scaffolds due to the difference in solvent properties and manufacturing process. The PS–PEO scaffolds again show the PEO nano-domain morphology, however the PS–PAA scaffolds have no obvious surface structure. Fig. 4 shows SEM micrographs of scaffolds manufactured from cyclohexane solutions. Again, the macro-architecture shows
significant differences from the other solvent systems, however all scaffolds, including PS–PEO, show no obvious surface structure. Recalling the phase diagrams from Fig. 1, the PS–PEO-dimethyl carbonate system was the only system to display a reduction in temperature of the phase boundary with increasing addition of block copolymer. As mentioned previously, this is likely due to the PS–PEO reducing the interfacial energy between the polymer– solvent phases, which prolongs the existence of a single phase mixture, but also preferentially locates the block copolymer at the interface between the developing polymer-rich and polymer-poor phases. In the TIPS process, as the spinodal interface matures, the local concentration of block copolymer at the surface is obviously high enough to induce ordering, resulting in the presentation of a micro-phase separated surface structure. The exact mechanism responsible for this remarkable structure formation is difficult to isolate from these experiments, and further experiments utilizing small angle neutron scattering and similar techniques will be required to elucidate the dominant kinetic events in this process. Regardless, this outcome shows that in order for nano-scale self-assembled features to be displayed throughout the surface of
Fig. 4. SEM micrographs of scaffolds manufactured from cyclohexane solutions. Top panel is low magnification images showing porous macro-scale structure of scaffolds. Lower panel is high magnification image showing surface structure of scaffolds.
646
P.A. George et al. / Biomaterials 31 (2010) 641–647
Fig. 5. Top panel shows AFM images of the surface of thin films of different molecular weight PS–PEO block copolymers. Middle panel shows low magnification SEM micrographs of scaffolds porous macro-scale structure. Bottom panel shows high resolution SEM micrographs showing surface structure of the scaffolds.
macro-porous structure formed through TIPS via a liquid-liquid phase separation pathway, the phase diagram must be perturbed to lower temperatures. That is, the block copolymer must compatibilise the system.
3.3. Controlling nano-domain presentation using copolymer molecular weight In order to further demonstrate that the observed surface structures are indeed the result of block copolymer micro-phase separation, and that they are discretely tuneable, scaffolds were manufactured from a series of different molecular PS–PEO block
copolymers in dioxane (polymer properties listed in Table 1). The molecular weight of the block copolymers determines the size of the resultant micro-phase separated domains, thus the size of the nano-domains on the scaffold should scale with block copolymer molecular weight. Fig. 5 shows atomic force microscopy images of 2-dimensional thin films as well as SEM micrographs of scaffolds generated from the different molecular weight copolymers. Note that due to the highly porous nature of the scaffolds, SEM was used to image 3-dimensional structures, as AFM was not possible. It is evident that the surface structure observed in the 3-dimensional scaffolds follows very closely the trend observed for the 2-dimensional thin films. Fig. 6 quantifies the PEO domain size based on image analysis of the 2-dimensional AFM and 3-dimensional SEM images. The slight reduction in the average domain size observed for the 3-dimensional samples is believed to be due to differences in image generation between the AFM and SEM instruments. Regardless of this small (consistent) discrepancy, it is clear that the domain sizes are determined by the molecular weight of the PS–PEO copolymer, proving that the nano-domain structures at all surfaces are tuneable and further, are generated as a result of block copolymer micro-phase separation. 3.4. Cellular interaction with nano-functionalised scaffolds
Fig. 6. Comparison of PEO domain size for 2-dimensional thin films and surface of highly porous 3-dimensional scaffolds indicating that the PEO domain size is controlled by polymer molecular weight.
To highlight the utility of this macro-micro-phase separated scaffolds for tissue engineering; the RGD adhesion motif was bound to the PEO nano-domains on the surface of a LMW PS–PEO scaffold (produced from dioxane). The terminal alcohol group on the PS–PEO block copolymer can be derivatized to a maleimide group (PS–PEO-mal, as previously described [8]). The addition of a maleimide group does not affect the surface structure formation, but allows for the conjugation of adhesion motifs, such as cysteine tagged RGD, or other biologically relevant molecules via the covalent bonding of maleimide to cysteine [8].
P.A. George et al. / Biomaterials 31 (2010) 641–647
647
Fig. 7. Fluorescence microscopy images of cells cultured inside scaffolds for 24 h. Nuclei stained blue, Actin stained green. Scale bar: 50 microns.
Fig. 7 shows fluorescent images of NIH-3T3 cells after 24 h of culture inside a control PS scaffold, a non-modified LMW PS–PEO scaffold, as well as an RGD modified PS–PEO-maleimide scaffold. The results are in agreement with our previous findings for similar surfaces presented as 2-dimensional films. NIH-3T3 fibroblast cells cannot adhere to, or spread on, the LMW PS–PEO scaffold, but can adhere to the PS and adhere strongly to the RGD modified PS–PEO surface [8,10]. From the morphology of the cells in the PS–PEO scaffold it is obvious that they have not adhered to the scaffold, and those that have been imaged are merely caught within the scaffolds pores. However, the introduction of the RGD group to the PEO nano-domains presented on the surface of the scaffold results in significant differences in cellular attachment and spreading compared even to the PS scaffold, confirming the utility of the system for 3-dimensional presentation of cellular binding motifs. 4. Conclusion In this paper we have demonstrated that through the controlled self-assembly of tailored block copolymers undergoing a combined macro-micro-phase separation process, it is possible to control the density and size of ‘islands’ of two distinctly different chemical functionalities, PEO and PAA, over the entire surface of three dimensional porous scaffolds with nanometer precision. Further, the nano-domains of these block copolymer can be further modified to present, on complex 3-dimensional surfaces, adhesion motifs (in this case RGD), or other biological molecules, with precise control over the lateral spacing and number density (presenting up to 1013 discrete nanometer-sized domains per ml of scaffold volume), enabling explicit control over cellular behaviours such as attachment and spreading in a purely synthetic 3-dimensional porous construct. Tissue engineering scaffolds manufactured via this combined macro–micro-phase separation methodology fulfill many of the ideal scaffold design criteria. Once functionalised with peptides, or recombinantly produced molecules, the entire construct is synthetic, can be manufactured cheaply and reproducibly, and yet offers complexity and precision of design previously unavailable to the tissue engineer. Acknowledgments This work was performed in part at the Queensland node of the Australian National Fabrication Facility, a company established
under the National Collaborative Research Infrastructure Strategy to provide nano and microfabrication facilities for Australia’s researchers.
Appendix Figures with essential colour discrimination. Figs. 1, 5 and 7 of this article have parts that are difficult to interpret in black and white. The full colour images can be found in the online version at doi:10.1016/j.biomaterials.2009.09.094
References [1] Sachlos E, Czernuszka JT. Making tissue engineered scaffolds work. Review on the application of solid freeform fabrication technology to the production of tissue engineered scaffolds. European Cells and Materials 2003;5: 29–40. [2] Vacanti JP, Morse MA, Saltzman WM, Domb AJ, Perez-Atayde A, Langer R. Selective cell transplantation using biorabsorbable artificial polymer matrices. Journal of Pediatric Surgery 1988;23. [3] Vacanti CA. The history of tissue engineering. Journal of Cellular and Molecular Medicine 2006;10(3):569–76. [4] Alberts B, Bray D, Lewis J, Raff M, Roberts K, Watson JD. Molecular biology of the cell. 3rd ed. New York: Garland Publishing; 1994. [5] Poole K, Khairy K, Friedrichs J, Franz C, Cisneros DA, Howard J, et al. Molecularscale topographic cues induce the orientation and directional movement of fibroblasts on two-dimensional collagen surfaces. Journal of Molecular Biology 2005;349(2):380–6. [6] Macchiarini P, Jungebluth P, Go T, Asnaghi MA, Rees LE, Cogan TA, et-al. Clinical transplantation of a tissue-engineered airway. The Lancet 372(9655): 2023–2030. [7] Griffith LG. Polymeric biomaterials. Acta Materialia 2000;48(1):263–77. [8] George PA, Doran M, Croll T, Cooper-White J. Nano-scale presentation of cell adhesive molecules via block copolymer self-assembly. Biomaterials 2009. [9] Flory PJ. Statistical thermodynamics of semi-flexible chain molecules. Proceedings of the Royal Society of London 1956;234(1196):60–73. [10] George PA, Donose BC, Cooper-White JJ. Self-assembling polystyrene-blockpoly(ethylene oxide) copolymer surface coatings: resistance to protein and cell adhesion. Biomaterials 2009;30(13):2449–56. [11] Bates FS, Fredrickson GH. Block copolymer thermodynamics - theory and experiment. Annual Review of Physical Chemistry 1990;41:525–57. [12] Van Krevelen DW. Properties of polymers: their correlation with chemical structure; their numerical estimation and prediction from additive group contributions. 3rd ed. New York: Elsevier; 1990. [13] Kim HD, Bae EH, Kwon IC, Pal RR, Nam JD, Lee DS. Effect of PEG-PLLA diblock copolymer on macroporous PLLA scaffolds by thermally induced phase separation. Biomaterials 2004;25(12):2319–29. [14] Boontongkong Y, Cohen RE. Cavitated block copolymer micellar thin films: lateral arrays of open nanoreactors. Macromolecules 2002;35(9):3647–52.
