A method to obtain nanofibers and complex nanostructures from single or multiple extracellular matrix proteins is described. This method uses protein-surface interactions to create free-standing protein-based materials with tunable composition and architecture for use in a variety of tissue engineering and biotechnology applications.
The extracellular matrix (ECM) in tissues is synthesized and assembled by cells to form a 3D fibrillar, protein network with tightly regulated fiber diameter, composition and organization. In addition to providing structural support, the physical and chemical properties of the ECM play an important role in multiple cellular processes including adhesion, differentiation, and apoptosis. In vivo, the ECM is assembled by exposing cryptic self-assembly (fibrillogenesis) sites within proteins. This process varies for different proteins, but fibronectin (FN) fibrillogenesis is well-characterized and serves as a model system for cell-mediated ECM assembly. Specifically, cells use integrin receptors on the cell membrane to bind FN dimers and actomyosin-generated contractile forces to unfold and expose binding sites for assembly into insoluble fibers. This receptor-mediated process enables cells to assemble and organize the ECM from the cellular to tissue scales. Here, we present a method termed surface-initiated assembly (SIA), which recapitulates cell-mediated matrix assembly using protein-surface interactions to unfold ECM proteins and assemble them into insoluble fibers. First, ECM proteins are adsorbed onto a hydrophobic polydimethylsiloxane (PDMS) surface where they partially denature (unfold) and expose cryptic binding domains. The unfolded proteins are then transferred in well-defined micro- and nanopatterns through microcontact printing onto a thermally responsive poly(N-isopropylacrylamide) (PIPAAm) surface. Thermally-triggered dissolution of the PIPAAm leads to final assembly and release of insoluble ECM protein nanofibers and nanostructures with well-defined geometries. Complex architectures are possible by engineering defined patterns on the PDMS stamps used for microcontact printing. In addition to FN, the SIA process can be used with laminin, fibrinogen and collagens type I and IV to create multi-component ECM nanostructures. Thus, SIA can be used to engineer ECM protein-based materials with precise control over the protein composition, fiber geometry and scaffold architecture in order to recapitulate the structure and composition of the ECM in vivo.
The extracellular matrix (ECM) in tissues is composed of multifunctional proteins involved in physical and chemical regulation of multiple cell processes including adhesion, proliferation, differentiation, and apoptosis1-3. The ECM is synthesized, assembled, and organized by cells and the constituent protein fibrils have unique compositions, fiber size, geometries and interconnected architectures that vary with tissue type and developmental stage. Recent work has demonstrated that the ECM can provide instructive cues to guide cells to form engineered tissues4, suggesting that recapitulating the ECM in terms of composition and structure might enable the development of biomimetic materials for tissue engineering and biotechnology applications.
A number of fabrication methods have been developed to engineer polymeric scaffolds that can mimic aspects of the ECM in tissues. For example, electrospinning and phase separation have both demonstrated the ability to form porous matrices of fibers with diameters ranging from tens of micrometers down to tens of nanometers5-7. Both techniques have also shown that highly porous matrices of nanofibers can support cell adhesion and infiltration into the scaffold8. However, these approaches are limited in the possible fiber geometries, orientations and 3D architectures that can be created. Electrospinning typically produces scaffolds with either randomly oriented or highly aligned fibers whereas phase separation produces scaffolds with randomly oriented fibers. There are also limitations on the materials, with researchers typically using synthetic polymers, such as poly(ε-caprolactone)8 and poly(lactic-co-glycolic acid)9, that are subsequently coated with ECM proteins to promote cell adhesion. Natural biopolymers are also used, including collagen type I10, gelatin11, fibrinogen12, chitosan13, and silk14, but represent only a small subset of the proteins found in native tissue. Most tissues contain a larger milieu of ECM proteins and polysaccharides including fibronectin (FN), laminin (LN), collagen type IV and hyaluronic acid that are difficult or impossible to fabricate nanofibers using existing methods.
To address this challenge, we have focused our research efforts on mimicking the way cells synthesize, assemble and organize ECM protein fibrils in their surroundings. While the specific fibrillogenesis process varies for different ECM proteins, typically a conformational change in the ECM protein molecule is triggered by an enzymatic or receptor-mediated interaction, which exposes cryptic self-assembly sites. Here we use FN as a model system to better understand the fibrillogenesis process. Briefly, FN homodimers bind to integrin receptors on the cell surface via the RGD amino acid sequence in the 10th type III repeat unit. Once bound, the integrins move apart via actomyosin contraction and unfold the FN dimers to expose cryptic self-assembly sites. The exposure of these FN-FN binding sites enables the FN dimers to assemble into an insoluble fibril right on the cell surface15. Work in cell-free systems has demonstrated that cryptic FN-FN binding sites can be revealed through unfolding using denaturants16 or surface tension at an air-liquid-solid interface17-19. However, the FN fibers created by these techniques are restricted to specific fiber sizes and geometries and are typically bound to a surface.
Here we describe an approach termed surface-initiated assembly (SIA) 20 that overcomes these limitations by utilizing protein-surface interactions to create free-standing insoluble nanofibers, nanofabrics (2D sheets) and other nanostructures composed of single or multiple ECM proteins (Figure 1). In this process, ECM proteins are adsorbed from a compact, globular conformation in solution and partially denatured (unfolded) onto a patterned, hydrophobic polydimethylsiloxane (PDMS) stamp. The ECM proteins are then transferred in this state onto a thermally responsive poly(N-isopropylacrylamide) (PIPAAm) surface through microcontact printing22. When hydrated with 40 °C water the PIPAAm remains a solid, but when cooled to 32 °C it passes through a lower critical solution temperature (LCST) where it becomes hydrophilic, swells with water and then dissolves, releasing the assembled ECM nanostructures off of the surface. The SIA method provides control over the dimensions with nanometer-scale precision. By controlling key parameters such as composition, fiber geometry, and architecture, it is possible to recapitulate many properties of the ECM found in vivo and to develop advanced scaffolds for tissue engineering and biotechnology applications.
1. Fabrication of Master Mold Using Photolithography
2. Making the PDMS Stamps
3. Microcontact Printing of ECM Patterns
4. Release of ECM Nanofibers and Nanostructures
SIA is capable of engineering ECM protein nanofibers with precise control over fiber dimensions. To demonstrate this, arrays of FN nanofibers with planar dimensions of 50 x 20 μm were patterned onto a PIPAAm coated coverslip (Figure 2A). Upon release, the fibers contracted because they were under an inherent pre-stress when patterned on the PIPAAm surface (Figure 2B). Analysis of the FN nanofibers revealed they were monodisperse pre-release with an average length of 50.19 ± 0.49 μm and width of 19.98 ± 0.17 μm (Figure 2C). Despite contracting appreciably, FN fibers post-release were still monodisperse with an average length of 14.15 ± 0.92 μm and width of 2.65 ± 0.32 μm (Figure 2D). Atomic force microscopy provided a higher resolution perspective of the fiber dimensional changes associated with the SIA release process. Notably, fibers pre-release had a uniform thickness of ~5 nm (Figure 2E) whereas fibers post-release had a heterogeneous thickness on the order of several hundred nanometers (Figure 2F).
Using the SIA process it is possible to engineer a variety of ECM protein nanostructures with tunable size, shape, and composition (Figure 3). For example, FN nanofibers initially 20 μm in width and 1 cm in length were patterned onto a PIPAAm-coated coverslip. Upon cooling and PIPAAm dissolution the nanofibers were released forming long threads (Figure 3A). Further, because the pattern is defined by the surface topography of the PDMS stamp used for microcontact printing, it is possible to engineer complex ECM protein nanostructures. As proof-of-concept, we created multi-armed FN stars that retained their general shape after release (Figure 3B). Interestingly, the arms of the FN star contracted like the nanofibers but the body of the star maintained its size. While FN is important, we also wanted to demonstrate that SIA works with other ECM proteins such as LN and that multiple ECM proteins can be incorporated into the same structure. For example, we engineered a 2-D nanofabric composed of orthogonal and interconnected nanofibers of FN and LN in a square lattice array (Figure 3C). Pre-release the FN nanofibers were 20 μm wide and the LN nanofibers were 50 μm wide. Upon release both types of nanofibers contracted but the overall interconnectivity and square lattice structure was maintained. These results demonstrate that SIA can be used to engineer ECM materials with a variety of compositions and structures.
Instances of failed SIA of ECM nanofibers are shown in Figure 4. One cause is improper release of an incomplete pattern due to poor transfer of ECM protein to the PIPAAm surface during microcontact printing (Figure 4A). The presence of holes, irregular edges, and other defects will create nanofibers and nanostructures that are incomplete and prone to breakage and fragmentation upon release. Rapid dissolution of PIPAAm can also cause poor pattern fidelity after release (Figure 4B). For example, using room temperature 20 °C DI water already below the LCST of PIPAAm instead of 40 °C water will cause the PIPAAm to rapidly swell and dissolve. This can cause two problems; (i) the rapid expansion can rip apart some ECM nanofibers and (ii) rapid expansion can cause disruption of the pattern arrangement after release.
Figure 1. Schematic of the SIA process. (A) A silicon wafer is spincoated with SU8 negative photoresist and exposed to UV light through a photomask. Non-exposed regions are developed away leaving a topographically patterned master mold. (B) PDMS prepolymer is poured over the master mold and cured for 4 hr at 65 °C. (C) After curing, a PDMS stamp is cut out. (D) The stamp is then incubated with an ECM protein solution where the proteins adsorb to the stamp in a partially unfolded conformation. (E) The stamp is rinsed to remove excess protein, dried, and placed in conformal contact with a PIPAAm-coated glass coverslip. (F) The stamp is then removed leaving behind patterned ECM protein on the PIPAAm coated coverslip, the pattern is dictated by the topography of the stamp. (G) The coverslip is then placed in a Petri dish and covered with 40 °C water and then cooled below the LCST of PIPAAm (~32 °C), which triggers the dissolution of PIPAAm and the release of assembled ECM protein nanofibers and/or nanostructures from the surface.
Figure 2. Using SIA to engineer populations of monodisperse nanofibers. (A) An array of FN rectangles, 50 μm in length and 20 μm in width were patterned onto the PIPAAm surface. (B) Addition of 40 °C DI water and subsequent cooling below the LCST of PIPAAm triggered the dissolution of PIPAAm and the release of the FN nanofibers. Upon release, the fibers contracted as they are under a pre-stress when patterned on the PIPAAm surface. (C) Analysis of nanofiber dimensions pre-release reveals the fibers are monodisperse with a length and width of 50.19 ± 0.49 μm and 19.98 ± 0.17 μm, respectively. (D) Upon release the nanofibers contracted but remained monodisperse with post-release length and width of 14.15 ± 0.92 μm and 2.65 ± 0.32 μm, respectively. (E) AFM showed that the pre-release nanofibers were ~5 nm thick. (F) AFM of post-release nanofibers showed that the thickness had increased to several hundred nanometers while the length and width decreased. -Scale Bars are (A) 50 μm and (B) 10 μm.
Figure 3. Using SIA to engineer ECM protein nanofibers and nanostructures with tunable size, shape, and composition. (A) FN nanofibers were patterned onto a PIPAAm-coated coverslip as 1 cm in length and 20 μm in width. Thermal release resulted in the SIA of long, intact FN nanofibers with a reduced width of ~3 μm. (B) An example of a more complicated multi-armed FN star, representative of the diverse nanostructures that can be created using SIA. Thermal release resulted in the contraction of the arms but not the central region of the star, where the arms joined together. (C) It is also possible to integrate multiple ECM proteins into the same structure. For example, orthogonal, interconnected 20 μm wide lines of FN (red) and 50 μm wide lines of LN (green) lines integrated into a 2D nanofabric were patterned and then released. Even after release, the pattern retained its initial geometry and interconnectivity. Scale bars are 50 μm.
Figure 4. Examples of problems that may prevent proper SIA and protein release from the PIPAAm surface. (A) Poor microcontact printing of ECM proteins onto the PIPAAm results in fragmentation and other defects of a 20 μm wide FN line prevents the SIA of a continuous FN nanofiber and instead results in the formation many smaller FN fragments. (B) Rapidly releasing the FN nanofibers from the PIPAAm substrate may also affect the final fiber arrangement. In this case water at 20 °C, already below the LCST of PIPAAm, was added and triggered rapid dissolution causing the nanofibers to snap back, break (30 sec) and form random, unorganized configurations (52 sec). Scale bars are 50 μm.
The SIA method presented here mimics cell-mediated matrix assembly and enables the engineering of ECM protein nanofibers and nanostructures with tunable size, organization and composition. While not identical to cell-generated ECM, SIA creates ECM composed of nanoscale protein fibrils20 that undergo reversible folding/unfolding during mechanical strain21 and can bind cells20. This provides a unique capability to build ECM protein materials that recapitulate many properties of the ECM found in vivo. For example, ECM nanofibers can be fabricated in specific lengths with a level of control infeasible with other techniques. We demonstrate the ability to create arrays of monodisperse nanofibers (Figure 2) with precise control of length and width pre-release (Figure 2C) and post-release (Figure 2D). These ECM nanofibers can be any length, as demonstrated by fabricating FN nanofibers in ~1 cm lengths (Figure 3A). In contrast, other fabrication techniques such as electrospinning and phase separation can create nanofibers but not with the precise control of length, producing mostly continuous fibers. These techniques are also limited in the fiber geometries and organization within a scaffold. SIA can be used to build ECM nanostructures with arbitrary planar geometry, such as a star (Figure 3B), and in 2-D sheets with arbitrary control of fiber organization, such as a nanofabric (Figure 3C). Further, other fabrication methods typically use synthetic materials or only a limited number of natural ones such as chitosan and fibrin. In comparison, SIA enables the assembly of nanofibers and nanostructures composed completely of ECM proteins such as FN and LN, which cannot be fabricated with these other methods.
The critical step in the SIA process is the thermally-triggered release from the PIPAAm surface. To ensure proper release, there are a few key steps that should be considered. First, after the microcontact printing step, the fidelity of the transferred ECM pattern should be inspected either using phase contrast microscopy or fluorescence microscopy (if the ECM proteins are fluorescently tagged). If there are any defects in the protein pattern on the PIPAAm surface, then the subsequent release will produce nanostructures that contain these defects (as observed in Figure 4A). If a PDMS stamp repeatedly produces protein patterns with defects, it is probable that the micropatterned side of the PDMS stamp has been scratched or contains defects and a new stamp and potentially a new master mold should be made. In addition, care should be taken when hydrating the protein patterned, PIPAAm-coated coverslip. The water should be at or near 40 °C and be allowed to cool slowly. If the water is not warm enough or cools too rapidly, the PIPAAm will swell and dissolve too quickly causing the ECM protein nanofibers and nanostructures to be potentially torn apart from the forces and haphazardly released such that any structural organization resembling the dry, pre-released state will be lost (Figure 4B).
ECM protein nanofibers, nanostructures and nanofabrics assembled using SIA have many applications in biomechanics, tissue engineering, and biotechnology. For example, recent evidence indicates that the mechanical properties of ECM protein fibrils governs their biological functionality23. SIA is ideally suited to create ECM protein nanofibers in purified form for mechanical analysis. To perform these experiments, released ECM protein nanofibers can be manipulated using precision micropositioners and then stretched. Deravi et al have recently used this approach to demonstrate that FN nanofibers can withstand up to 8-fold extensions and that they undergo elastic, plastic, and strain-stiffening regimes with increasing strain21. In tissue engineering, ECM protein based scaffolds in the form of gels (e.g., collagen type I and fibrin) or decellularized tissues4 are widely used. Compared to these techniques, the advantage of SIA is the ability to engineer scaffolds with well-defined architecture in 1-D fibers and 2-D sheets (Figure 3C). For example, we have previously shown that cardiomyocytes can be seeded onto FN nanofibers to form functional strands of cardiac muscle20. This shows that the FN nanofibers are cell binding and that they can direct the anisotropic assembly of cells into aligned tissue structures. The 2-D nanofabrics can mimic the composition and laminar structure of the basement membrane for application in the engineering of epithelial and endothelial tissues. Further, we are developing new ways to deploy these ECM nanofibers and nanofabrics in 3-D by embedding them within hydrogel matrices. From the process described in this article, it is possible to use SIA to engineer ECM protein-based, nanostructured materials for a wide variety of applications.
The authors have nothing to disclose.
Financial support was provided to J.M.S. from the NIH Biomechanics in Regenerative Medicine T32 Training Program (2T32EB003392), to Q.J. from the Dowd-ICES Fellowship and to A.W.F. from the NIH Director’s New Innovator Award (1DP2HL117750).
Poly(N-isopropylacrylamide) / PIPAAm | Polysciences | 21458-10 | 40,000 Mw |
Sylgard 184 Silicone kit (PDMS) | Dow Corning | Mix 10 parts base with 1 part curing agent. | |
Butanol | |||
Fibronectin | BD biosciences | 354008 | Human, 1mg |
Laminin | BD biosciences | 354239 | Ultrapure, mouse, 1mg |
Negative Photoresist | Microchem | SU8-2015 | |
SU8 Developer | Microchem | ||
Sonicator Branson M3510 | Branson Ultrasonic Corporation | CPN-952-318 | |
Thinky ARE-250 Mixer | Thinky Corporation | ||
Spincoater | Specialty Coating Systems | G3P-8 | |
Glass cover 25mm diameter, No 1.5 | Fisher Scientific | 12-545-86 |