Fabrication and Characterization of Recombinant Silk-elastinlike Protein Fibers for Tissue Engineering Applications

Transcription

1 Fabrication and Characterization of Recombinant Silk-elastinlike Protein Fibers for Tissue Engineering Applications Item Type text; Electronic Dissertation Authors Qiu, Weiguo Publisher The University of Arizona. Rights Copyright is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author. Download date 21/06/ :27:45 Link to Item

2 FABRICATION AND CHARACTERIIZATION OF RECOMBINANT SILK-ELASTINLIKE PROTEIN FIBERS FOR TISSUE ENGINEERING APPLICATIONS By Weiguo Qiu A Dissertation Submitted to the Faculty of the AEROSPACE AND MECHANICAL ENGINEERING DEPARTMENT In Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOHPY WITH A MAJOR IN MECHANICAL ENGINEERING In the Graduate College THE UNIVERSITY OF ARIZONA 2011

3 2 THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Weiguo Qiu entitled Fabrication and Characterization of Recombinant Silk-elastinlike Protein Fibers for Tissue Engineering Applications and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy Date: XiaoyiWu Date: Pak Kin Wong Date: Jeong-Yeol Yoon Date: Yitshak Zohar April 07, 2011 April 07, 2011 April 07, 2011 April 07, 2011 Date: Final approval and acceptance of this dissertation is contingent upon the candidate s submission of the final copies of the dissertation to the Graduate College. I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement. Date: April 07, 2011 Dissertation Director: XiaoyiWu

4 3 STATEMENT BY AUTHOR This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library. Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author. SIGNED: Weiguo Qiu

5 4 Reproduced in part with permission from [1], Copyright [2009] American Chemical Society and from [2], Copyright [2010] American Chemical Society. 1. Qiu, W., et al., Wet-spinning of recombinant silk-elastin-like protein polymer fibers with high tensile strength and high deformability. Biomacromolecules, (3): p Qiu, W., et al., Complete recombinant silk-elastinlike protein-based tissue scaffold. Biomacromolecules, (12): p

6 5 ACKNOWLEDGEMENT The last four and half years are like a dream. When I wake up, everything is gone, but with fruitful harvest left. The result of this work and the accomplishments of my graduate years, will surely owe to the guidance of my academic advisor, the support of my family and friends. I am grateful to Drs. Xiaoyi Wu, Pak Kin Wong, Yitshak Zohar and Jeong-Yeol Yoon for serving on the committee and providing me with many invaluable advices, comments and questions that made this dissertation possible. I would especially like to thank my advisor Xiaoyi Wu, for his guidance, encouragement, doubts and questions in all aspects of research and academic life. He has been an irreplaceable role model for my academic career, and the knowledge he has provided me extends beyond what I can found in any textbook. I extend my thanks to those who helped and trained me on the research. Without their support, I could not finish my work. These include: Dale Drew in the machine shop, who helped me make the spinneret and every tiny parts of the experiment setup; Ali Kemal Yetisen, who made the electrospinning house cover for me; Professor Chorover, in department soil, water and environment science for the use of Fourier transform infrared spectrometer, and the freeze drier; Professor Downs in department of geosciences for the access to the Raman spectrometer; Yiding Huang, for her specialties in cell biological experiments to complete the cell study for me; and the imaging facility team in the Marley building. I would further thank my wife, for her support and love. No matter what happens, she will stand nearby.

7 6 I dedicate this dissertation to the people who suffered and are suffering from the heart disease.

8 7 TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES ABSTRACT CHAPTER 1: INTRODUCTION Background Tissue Engineering Biomaterials for Scaffolds Recombinant Silk-elastinlike Proteins Scaffold/Fiber Preparation Techniques in Tissue Engineering Key Properties of Scaffolds/Fibers and Characterization Methods Motivations and Scope of Work CHAPTER 2: WET-SPINNING OF MICROFIBERS Introduction Material and Methods Sample Preparation Wet-spinning: Experiment Design Crosslinking Surface Morphology by Scanning Electron Microscopy Secondary Structure Characterization by Raman Spectroscopy Mechanical Characterization Results and Discussions Wet-spinning of SELP-47K Microfibers... 52

9 8 TABLE OF CONTENTS - Continued Secondary Structures Analysis by Raman Spectroscopics Mechanical Properties Analysis Summary CHAPTER 3: ELECTROSPINNING OF NANOFIBERS Introduction Materials and Methods Experiment Setup Electrospinning: Design of Experiment Scaffold Post-fabrication Treatment Surface Morphology and Secondary Structures Characterizations Mechanical Characterization Cell Culture Experiment on Nanofibrous Scaffolds Results and Discussions Electrospinning Experiment Study Post-fabrication Treatment Effects on the Microstructures Fourier Transform Infrared Spectroscopic Secondary Structure Analysis Mechanical Properties Analysis Nanofibrous Scaffolds Cell Biocompatibility Study Summary

10 9 TABLE OF CONTENTS - Continued CHAPTER 4: FINAL REMARKES AND FUTUER WORK Final Remarks Future Work REFERENCES

11 10 LIST OF TABLES Table 1.1. Biomaterials for tissue applications with selected publications Table 1.2. Protein-based biomaterial from the nature for tissue engineering Table 1.3. Recombinant protein-based biomaterial for tissue engineering Table 1.4. Study of Silk-elastinlike proteins for tissue engineering since Table 1.5. Mechanical properties studied for protein-based biomaterials in hydrated state Table 2.1. Band Assignment of Raman Spectrum Table 2.2. Secondary structural contents (%) estimated by Raman Amide I/III spectral analysis Table 2.3. Secondary structural contens (%) of natural fiber Table 2.4. Mechanical Properties of as-spun dry fibers Table 2.5. Mechanical properties of hydrated, crosslinked SELP-47K fibers Table 3.1 Mechanical properties of fully hydrated SELP-47K nanofibrous scaffolds before and after mechanical preconditoning a

12 11 LIST OF FIGURES Figure 1.1. Tissue engineering approaches. Cells can be isolated from an individual and cultivated in vitro to differentiate, and eventually implanted back to the same individual (pathway 1) ). These same cells are can also be grown into a tissue with (2b)/without (2a) being seeded in a matrix, and implanted back (pathway 2)). The third approach is the introduction of tissue induced substance (pathway 3)). Tissue induce substance can be directly introduced to the body (3a) or introduced to promote cells growth in vitro and then infused to the body (3b) or encapsulated within a membrane (3c) to infuse to the body. Finally, tissue regeneration can also be achieved by using approaches that could be considered the combination of the three pathways [13] Figure amino acid SELP-47K sequence with a molecular weight of Daltons. It is composed of a head and tail sequence, and a series of silk-like (GAGAGS) and elastin-like (GVGVP) repeats (in bold). Abbreviation Key: A = alanine; D = aspartic acid; E = glutamic acid; F = phenylalanine; G = glycine; H = histidine; K = lysine; L = leucine; M = methionine; N = asparagine; R = arginine; P = proline; Q = glutamine; S = serine; T = threonine; V = valine; W = tryptophan; Y = tyrosine. [62]... 31

13 12 LIST OF FIGURES - Continued Figure 1.3. SELP-47K amino acid sequence produced by Genencor International Inc. [60] Figure 1.4. Schematic view of wet-spinning Figure 1.5. Schematic view of electrospinning Figure 2.1. Experiment setup of wet-spinning Figure 2.2. Wet-spun fiber placed in a vacuumed dessicator, the bottom is filled with liquid glutaradehyde Figure 2.3. Wet-spun SELP-47K fibers. Micro-diameter SELP-47K fibers of meters in length were collected on a glass spool (A); SEM micrographs of SELP-47K fibers with different diameters and surface morphologies (B, C); at the point of contact, two fibers can merge and form inter-fiber bonding before being air-dried (D, E); some fibers are partially hollow (F) Figure 2.4. Morphology of the end surface of a wet-spun fiber Figure 2.5. Surface morphology of wet-spun SELP-47K fibers that had been hydrated in 1 PBS for one year Figure 2.6. The diameter of the resultant fibers has a linear relationship with the diameter of the capillary Figure 2.7. Raman spectrum of crosslinked wet-spun fiber... 58

14 13 LIST OF FIGURES - Continued Figure 2.8. Raman spectrum of wet-spun SELP-47K microfibers Figure 2.9. Tensile stress-strain analysis of dry SELP-47K fibers Figure Stress-strain comparison between SELP-47K fibers to other natural silk fibers [111] Figure The Young s modulus of dry fibers is a function of fiber diameter Figure The fracture surface of dry SELP-47K fibers Figure Representative tensile stress-strain analysis of SELP-47K fibers fully hydrated in 1 PBS at 37 o C Figure Tensile stress-strain analysis of glutaraldehyde-crosslinked SELP-47K fibers fully hydrated in 1 PBS at 37 o C with comparison to human saphenous vein and aortic elastin (Table 1.5) Figure 3.1. Electrospinning experiment horizontal setup Figure 3.2. Thickness measurement of thin sample: schematic illustration (A) and image under optical microscope (B) Figure 3.3. A representative force-clamp gap curve (A) with a zoom-in view at small deformations (B). The clamp gap was initially 3 mm and reduced to 1.5 mm after a sample was hydrated to prevent any hydration-induced tension. B suggests that the true gauge length of samples was close to 3 mm

15 14 LIST OF FIGURES - Continued Figure 3.4. Preconditioning displacement (A), force measurement (B), and resistant force exerted by the PBS on the clamp, which was measured by running the device without a sample (C). A sample of about 3 mm in length was cyclically stretched between 1.5 mm and 4.0 mm. This leads to a preconditioning strain of 33% (1 mm/3 mm), and an off-load time of about 5 to 6 minutes between loading cycles Figure 3.5. A water droplet is dropped on the surface of the SELP-47K scaffold for contact angle measurement Figure 3.6. SEM images of SELP-47K nanofibers electrospun from solutions containing protein polymer at concerntrations of 50 (A), 150 (B), and 250 mg/ml (C). The diameter distribution (D) of SELP-47K nanofibers shown in (B) Figure 3.7. SEM images of SELP-47K nanofibers from solutions containing protein polymer at concerntrations of 150 mg/ml: fibers closer to the glass slide (A) and fibers closer to the needle (B) Figure 3.8. Fiber branching or bifurcation was examined in the electrsopinning of a 150 mg/ml SELP-47K solution in formic acid

16 15 LIST OF FIGURES - Continued Figure 3.9. SEM images of SELP-47K nanofiberous scaffolds electrospun from 200 mg/ml SELP in distilled de-ionized water: A- ribbon-like nanofibers; B- the illustration showing the mechanism of flat ribbon formation Figure SEM images of SELP-47K nanofiberous scaffolds electrospun from aqueous solutions under ambient room temperature and humidity: 300 mg/ml (A, B), 350 mg/ml (C, D) and 400 mg/ml (E, F). Scale bar: 50 μm (A, C, E) and 20 μm (B, D, F) Figure SEM images of SELP-47K nanofiberous scaffolds electrospun from 200 mg/ml in distilled de-ionized water under 18% (A), 24% (B), 35% (C) and 48% (D) humidity and 23 o C room temperature. Scale bar: 10 μm Figure SEM Images of as-spun (A), MeOH- (B), GTA- (C), and MeOH-GTA-treated (D) SELP-47K fibrous scaffolds electrospun from a 15% w/v protein solution. Scale bars: 5 µm Figure SEM images of SELP-47K nanofiberous scaffold electrospun from 150 mg /ml SELP in 98% formic acid under ambient room environment: A-before autoclave, B-after autoclave. Scale bar: 5 μm Figure FTIR spectra of as-spun nanofiberous scaffolds electrospun from 150 mg/ml SELP in 98% formic acid (a), and 200 mg/ml SELP in distilled de-ionized water (b), under ambient room temperature

17 16 LIST OF FIGURES - Continued Figure FTIR spectra of SELP-47K nanofiberous scaffolds in the cm -1 range (A) and cm -1 Amide I range (B) and cm -1 range (C), the numbers from 1 to 6, are corresponding to the spectra from bottom to top. Scaffolds were electrospun from 150 mg/ml SELP in 98% formic acid under ambient room temperature and humidity with a gauge 32 needle Figure Representative tensile stress-strain curves of MeOH- (1), GTA- (2), and MeOH-GTA-treated (3) SELP-47K nanofibrous scaffolds Figure Representative preconditioning behavior of MeOH- (1), GTA- (2), MeOH-GTA-treated (3), and 60 min autoclaved SELP-47K nanofibrous scaffolds: A - curves were vertically shifted for better clarity; B the 60 min autoclaved scaffolds; C- the preconditioning cycles were removed, and comparison with human saphenous vein and aortic elastin (Table 1.5) Figure Resilience of SELP-47K nanofibrous scaffolds as a function of the number of preconditioning cycles (curves were horizontally shifted for better clarity) Figure Water contact angles of methanol treated scaffolds (I), GTA-treated scaffold (II), methanol-gta-treated scaffolds (III), methanol soaked film (IV) and GTA-treated film (V)

18 17 LIST OF FIGURES - Continued Figure Fluorescent staining for cell viability of 3T3 fibroblasts grown on MeOH- (A), GTA- (B), and MeOH-GTA-treated (C) electrospun SELP-47K scaffolds for 5 days. Living cells were in green and dead cells were in red. Scale bars: 50 µm Figure SEM images revealed that 3T3 fibroblasts interacted with MeOH- (A) and GTA-treated (B) SELP-47K nanofibrous scaffolds, and with each other (C) on GTA-treated scaffolds. Images were taken 6 days after the initial cell seeding. Scale bars: 10 µm Figure Proliferation profiles of 3T3 fibroblasts grown on electrospun SELP-47K nanofibrous scaffolds up to 7 days Figure SEM images showed more than one layer of 3T3 fibroblasts grown on MeOH-GTA-treated (A) and ECM nanofilaments regenerated by cells on GTA-treated scaffolds (B). Images were taken 6 days after the initial seeding. 113

19 18 ABSTRACT The integration of functional and structural properties makes genetically engineered proteins appealing in tissue engineering. Silk-elastinlike proteins (SELPs), containing tandemly repeated polypeptide sequence derived from natural silk and elastin, are recently under active study due to the interesting structure. The biological, chemical, physical properties of SELPs have been extensively investigated for their possible applications in drug/gene delivery, surgical tissue sealing and spine repair surgery. However, the mechanical aspect has rarely been looked into. Moreover, many other biomaterials have been fabricated into fibers in micrometer and nanometer scale to build extracellular matrix-mimic scaffolds for tissue regeneration, but many have one or mixed defects such as: poor strength, mild toxicity or immune repulsion etc. The SELP fibers, with the intrinsic primary structures, have novel mechanical properties that can make them defects-minimized scaffolds in tissue engineering. In this study, one SELP (SELP-47K) was fabricated into microfibers and nanofibers by the techniques of wet-spinning and electrospinning. Microfibers of meters long were formed and collected from a methanol coagulation bath, and later were crosslinked by glutaraldehyde (GTA) vapor. The resultant microfibers displayed higher tensile strength up to 20 MPa and higher deformability as high as 700% when tested in hydrated state. Electrospinnig of SELP-47K in formic acid and water

20 19 resulted in rod-like and ribbon-like nanofibrous scaffolds correspondingly. Both chemical (methanol and/or GTA) and physical (autoclaving) crosslinking methods were utilized to stabilize the scaffolds. The chemical crosslinked hydrated scaffolds exhibit elastic moduli of MPa, ultimate tensile strength of MPa, and deformability of %, closely matching or exceeding the native aortic elastin; while the autoclaved one had lower numbers: 1.0 MPa elastic modulus, 0.3 MPa ultimate strength and 29% deformation. However, the resilience was all above 80%, beyond the aortic elastin, which is 77%. Additionally, Fourier transform infrared spectra showed clear secondary structure transition after crosslinking, explaining the phenomenon of scaffold water-insolubility from structural perspective and showed a direct relationship with the mechanical performance. Furthermore, the in vitro biocompatibility of SELP-47K nanofibrous scaffolds were verified through the culture of NIH 3T3 mouse embryonic fibroblast cells.

21 20 CHAPTER 1: INTRODUCTION 1.1 Background Tissue Engineering The therapeutic strategy of autograft and allograft is the most common approach to repair or reconstruct diseased tissues. However, the use of autografts and allografts is limited by organ donation shortages and health concerns including immune repulsion or donor site morbidity. Thus, the regeneration or reconstruction of a tissue is the most promising approach. The advancement of tissue engineering will revolutionize the present health care methods to improve the quality of a patient s life [1]. Tissue engineering was first termed in 1987 when Dr. Y.C. Fung presented his concept during a National Science Foundation (NSF) meeting [2]. It was defined the following year as the application of principles and methods of engineering and life science to obtain a fundamental understanding of structure-function relationships in novel and pathological mammalian tissues and the development of biological substitutes to restore, maintain and improve tissue functions [3]. The widespread interested in tissue engineering within the scientific community was triggered by reviews from two groups in the early years: one by Nerem [4] on cellular engineering and another by Langer and Vacanti [5] on tissue engineering. After that, combined efforts of material scientists, cell biologists, engineers and clinicians

22 21 contributed to a great degree for the growth of tissue engineering. Now this field has grown into a highly interdisciplinary area with a fast pace. It has attempted to regenerate or reconstruct almost all types of human tissues or organs including skin [6], bone [7], liver [8], blood vessel [9], urethra [10], cornea [11] and lung [12]. Regeneration of these tissues/organs has been proposed by mainly four approaches including: 1) Infusion of isolated cells to replace or make up the cells that are needed; 2) Introduction of tissue induced substance like growth factors or cytokines; 3) Utilization of cells on a matrix to develop a tissue; and 4) the combination of the three (Figure 1.1)[13]. Figure 1.1. Tissue engineering approaches. Cells can be isolated from an individual and cultivated in vitro to differentiate, and eventually implanted back to the same individual (pathway 1) ). These same cells are can also be grown into a tissue with (2b)/without (2a) being seeded in a matrix, and implanted back (pathway 2)). The third approach is the introduction of tissue induced substance (pathway 3)). Tissue induce substance can be directly introduced to the body (3a) or introduced to promote cells growth in vitro and then infused to the body (3b) or encapsulated

23 22 within a membrane (3c) to infuse to the body. Finally, tissue regeneration can also be achieved by using approaches that could be considered the combination of the three pathways [13]. Among all of the approaches, the preparation of a matrix that can house cells is the most extensively investigated approach[1]. In this approach, biodegradable biomaterials are usually used to develop supporting matrices or scaffolds for cell integration. This matrix acts as a guide to the regeneration of new tissue and as an appropriate structural support to mimic the structure and functions of natural extracellular matrix (ECM). The scaffold along with cells then is implanted back to the body to regenerate or remodel the damaged tissue. This technique has led to the development of many tissues in laboratory such as bone, ligament, tendon, heart valves, blood vessels and trachea [1]. However several engineering and biological challenges remain for the successful translation from laboratory products to the clinical application. These include complex structure and biology of the ECM using biomaterials, controlling cell-scaffold interactions, vascularization of cell-scaffold construct, and efficient bioreactor for in vitro cultivation [14]. Therefore, tissue engineers are making great efforts to overcome these problems, especially the first one, as it is demonstrated by Mina Bissell along with others [15] that the biomaterial plays an important role in tissue regeneration.

24 Biomaterials for Scaffolds In tissue engineering, biomaterials are developed and fabricated into matrices or scaffolds to mimic the ECM. As is known, abundant protein fibers in nano-diameter are present in the natural ECM or tissues like muscles, working as a structural support with appropriate mechanical properties and as a medium for cell signal transduction. Collagen fibers are the most abundant in animal ECM, and give structural support with appropriate mechanical strength, the diameter of collagen fibers is down to 50 nm. While ealstin, on the other hand, give elasticity to tissues, allowing them to stretch when needed and return to their original state. These elastin fibers have a high value of resilience which is defined as the ratio of unloading energy to the loading energy within one cycle of stretch and relax. Other protein fibers like fibronectin and laminin work to aid cell adhesion and facilitate cell movement. Beside these nano-diameter fibers in ECM, micro-diameter fibers are usually needed to work as a structural support in engineered scaffolds. Thus, in order to mimic these properties of the natural ECM and to be more specific for tissue engineering, biomaterials are made into scaffolds to meet the strict criteria as described by T.C Holmes [16]: (1) basic units that are amenable to design and modification; (2) a controlled rate of biodegradation of the materials; (3) lack of cytotoxicity; (4) properties that specifically promote or inhibit cell material interactions; (5) elicitation of minimal immune responses and inflammation; (6) easy

25 24 and scalable material production, purification and processing, and (7) chemical compatibility with aqueous solutions and physiological conditions. Failure to meet any of these criteria will result in limitations of the potential biomaterial. Thus, various biomaterials such as PGA, PCL, Alginate and collagen (Table 1.1) have been developed to meet some of the criteria. These biomaterials, as summarized by Stock and Vacanti [17], can fall into two major groups: 1) the biologically derived which includes decellularized xenogenic and collage based materials and 2) the synthetically or bacterially derived which contains PGA, PLA, and recombinant proteins etc. Synthetically derived ones have a great flexibility in design and modification, thus molecular structure and material properties including physical, chemical and mechanical properties are under a high degree of control. As a result, a great number of synthetic biomaterials have been fabricated into scaffolds that can meet the specific requirement like appropriate mechanical properties. However, the biodegradation byproducts of these synthetic polymers may pose biocompatibility and toxicity risks, thereby compromising the long-term performance of tissue scaffolds. For instance, when PGA scaffolds are used to grow a human artery, undifferentiated phenotypes of smooth muscle cells are observed in proximity to residual PGA fragments [18].

26 25 Table 1.1. Biomaterials for tissue applications with selected publications Applications Example biomaterials References Vascular graft PGA G.Matsumura [19] Vascular graft PCL/PLA G. Matsumura [19] Pulp & dentin PLGA R.M El-Backly [20] Bone PCL J.M Williams [21] Cartilage Chitosan-based polysaccharide J.K. Francis Suh [22] Liver Alginate M.Dvir-Ginzberg [23] Skin Caprine (Goat) collagen I. Banerjee [24] Urethra Submucosa collagen from porcine bladder R.E De Filippo [25] Lung PDLLA Y.M Lin [26] Cornea Type I and III collagen from bovine cornium X. Duan [27] Heart valves Porcine aortic valve with poly(hydroxybutyrate) C. Stamm [28] Nerve PLLA F. Yang [29] PGA: Polyglycolic acid; PLA: Polylactic acid; PDLLA:Poly-DL-lactic acid; PLGA: Poly(lactic-co-glycolic acid); PLLA: Poly-L-Lactides, PCL: Polycaprolactone

27 26 On the other hand, protein-based biomaterials either in the biologically derived or bacteria derived group are more attractive for their improved biocompatibility. After degradation, the fragments of amino acid chains can be utilized as cells source proteins. Moreover, the endless possibilities of amino acids combinations make these possible protein-based polymers an attractive biomaterial. In the biologically derived group, the decellularization of xenogenic donor organs such as heart, liver and lung provides an acellular naturally occurring 3D biological scaffold that subsequently can be seeded with cells. Formation of functional tissues in short-term preclinical animal models is realized [30]. However complications including the method of organ decellularization, endothelizalization of the donor matrix vasculature, and host immune rejection still remain. Other biologically derived materials (Table 1.2) under studying include gelatin/collagen, elastin and fibronectin from mammals, silks from spiders and silkworms, proteins from soybean [31-32], corn [33], bamboo, cotton and wood [34]. Among them, collagen and elastin, which are present in majority in the mammals connective tissues, have been successfully reconstructed into fibrous scaffolds [35-37]. However, the resulting mechanical strength is compromised due to the harsh chemicals used or the inability of repeating the natural process in laboratory, thus future application may not be possible.

28 27 Table 1.2. Protein-based biomaterial from the nature for tissue engineering Proteins Source References Casine Cow milk J.B. Xie [38] Gelatin Bovine skin, type B powder M.Y. Li [39] Collagen Calf skin, type I powder M.Y. Li [39] α-elastin Soluble bovine M.Y. Li [39] Tropoelastin Elastogenic cells M. Y. Li [39] Zein Maize C. Yao [40] Fibrinogen Human and bovine G.E. Wnek [41] Silkworm Silk Bombyx. mori silkworm cocoons H.J. Jin [42] Spider Silk Agelena labyrinthica spider S.B. Zhou [43] Soybean Soy flour M. Santin [44] Wheat protein Wheat gluten D. L. Woerdeman [45] Research on another material in this biologically derived group: silk, has significantly advanced, because its superior mechanical properties have attracted intensive attention. In vivo animal experiments of silk-based material have been performed by several groups [46-47]. Recently, Zhang et al. electrospun a tubular scaffold from silk and grew a vascular graft in a bioreactor [48]. The silk thus appears as a very promising candidate for scaffold materials in tissue engineering. However its low resilience may impose a risk on future implantation. Admittedly, the low resilience of the silk follows the nature s demand: spiders create a web with a high efficiency in energy dissipation for capturing a flying insect. But the low resilience

29 28 of an engineered tubular scaffold presents a challenge to the repair process after implantation [49]. Another issue unaddressed issue about the silk scaffold is the thrombogenesis. Nevertheless, efforts are being made to improve the use and/or design of protein-based materials. For instance, cell binding proteins can be deposited on the surface of protein-based scaffolds for recruiting vascular endothelial cells and thus creating endothelium. The bacterially derived materials are referred to the recombinant proteins, which have drawn significant interest, as the recombinant DNA techniques has enabled scientists to genetically design and biologically engineer new proteins with desired functional and structural properties. Examples of these engineered proteins include: elastin-mimetic proteins, RGD-silk proteins (here note specifically which amino acids are noted by which letter, RGD) and collagen-like proteins (Table 1.3). Silk-elastinlike proteins (SELPs) are a group of specific engineered protein that is gaining interest within the tissue engineering community.

30 29 Table 1.3. Recombinant protein-based biomaterial for tissue engineering Proteins Reference Elastin-like protein K.Nagapudi [50], Collagen-like protein C.L. Du [51], RGD-silk E. Bini [52], NcDS and [(SpI) 4 /SpII) 1 ] 4 S. Arcidiacono [53], Hybrid silk protein K. Ohgo [54], Tropoelastin S.M.Mithieux [55], Elastin fibronectin E.R.Welsh [56], Silk-elastinlike protein R.Nagarajan [57], Recombinant Silk-elastinlike Proteins The primary structure of SELPs is composed of tandemly repeated units of a silklike block (GAGAGS) that is derived from fibrous silks and an elastinlike block (GVGVP), which originates from mammalian elastins. SELPs have novel mechanical properties due to the uniqueness of the parent proteins. Silk provides these proteins with remarkable strength and toughness while elastin provides incredible resilience and elasticity. These properties make SELPs a promising category of synthetic proteins within the realm of tissue engineering. So far, a series of SELPs have been designed and produced in laboratory, with various ratios of silklike blocks to elastinlike blocks. Furthermore, lysine modified pentapeptide sequences GKGVP have been introduced in the chain, acting as crosslinking sites for subsequent stabilization purposes. Thus with different ratios of silklike, elastinlike and lysine

31 30 modified blocks, a seriers of SELPs were produced, including: SELP-47K, SELP-815K and SELP-67K. The investigations of these materials were first leaded by the Protein Polymer Technologies, Inc., and were later collaborated with several university labs, then recently joined by the Genencor International, Inc. Major work was focused on the characterization of the physical, chemical and thermal properties to aim the applications in drug/gene delivery and in surgical sealants. J. Cappello first obtained gels from SELP-27K, SELP-47K and SELP816, and started SELPs gel research in drug/gene delivery [58]. These gels have been explored as drug carriers for anticancer therapies. However, the study towards tissue engineering application was slow in the early years. The mechanical properties of SELPs have rarely been reported until our study. Since 2007, SELPs microfibers, nanofibers and films were fabricated with different techniques, thus made the mechanical characterization possible. In particular, Teng et al. examined the SELP-47K thin film by performing tensile test and creep/relaxation test, and demonstrated that the crosslinked films display elastic properties comparable to elastin [59]. R.Nagarajan manufactured nanoscale fibers from SELP-67K but didn t check the mechanical properties [57], while, Y. Ner, generated ribbon-like nanofibers from SELP-47K aqueous solution via electrospinning. The methanol-treated SELP-47K structures possess an ultimate

32 31 strength of 30.8 MPa, Young s modulus of 0.88 GPa in dry state [60]. Now in 2011, J. Chan used nanomechanical stimulus to accelerate the self-assembly of SELP-815K nanofibers in aqueous solution [61]. Though the mechanical properties have not been studied, these nanofibers were believed to be different from the electrospun ones, as electrospun nanfibers were water-soluble, while these were formed in water. At the same period during these years, SELP-47K fibers were generated by both wet-spinning and electrospinning, which is my topic in this work. The publications of SELP study in tissue engineering since 2007 were listed in Table 1.4, showing a rapid progress. The protein polymer SELP-47K, provided by Protein Polymer Technologies Inc (PPT), has a monomer structure of (E)4(S)4(EK)(E)3, in which S is the silklike sequence GAGAGS, E is the elastinlike sequence GVGVP, and EK is the lysine incorporated pentapeptide sequence: GKGVP. The complete amino acid sequence of SELP-47K is illustrated in Figure 1.2 [62]. We focused in working on this protein. Figure amino acid SELP-47K sequence with a molecular weight of Daltons. It is composed of a head and tail sequence, and a series of silk-like (GAGAGS) and elastin-like (GVGVP) repeats (in bold). Abbreviation Key: A = alanine; D = aspartic acid; E = glutamic acid; F = phenylalanine; G = glycine; H = histidine; K = lysine; L = leucine; M = methionine; N = asparagine; R = arginine; P = proline; Q = glutamine; S = serine; T = threonine; V = valine; W = tryptophan; Y =

33 32 tyrosine. [62] While the same term SELP-47K was used for the protein produced by Genencor International Inc (GI). The two blocks (silklike block and elastinlike block) have the same amino acid sequences as the one by the PPT, and the whole protein has the same silk-to-elastin ratio, while the bock sequence is different. Moreover, the information about the amino acid sequence of EK was not released, plus head and tail groups were different (Figure 1.3)[60]. Figure 1.3. SELP-47K amino acid sequence produced by Genencor International Inc. [60] As listed in Table 1.4, several techniques, including wet-spinning, electrospinning, solvent cast and self-assembly, have been used to fabricate SELP into different forms such as films, microfiber, nanofibers and nanoribbons. In parallel, numerous post-fabrication treatment methods are available to stabilize SELP structures.

34 33 Table 1.4. Study of Silk-elastinlike proteins for tissue engineering since 2007 SELP name SELP form Fabrication method Studied by Year SELP-67K Nanofibers Electrospinning R.Nagarajan 2007 [57] SELP-47K 1* Microfibers Wet-spinning W. Qiu 2009 [63] SELP-415K 1 Nanofibers Self-assembly W. Hwang 2009 [64] SELP-47K 1 Films Solvent cast W. Teng 2009 [59] SELP-47K 2 Nanoribbons Electrospinning Y. Ner 2009 [60] SELP-47K 1* Nanofibers Electrospinning W. Qiu 2010 [65] SELP-815K 1 Nanofibers Self-assembly J. Chang 2011 [61] SELP-47K 1 Film Sovent cast W. Teng 2011 [66] 1 : provided by Protein Polymer Technologies Inc.(PPT) 2 : provided by Genencor International Inc.(GI) * : This study Scaffold/Fiber Preparation Techniques in Tissue Engineering As stated above, fibers both in micro-diameter scale and nano-diameter scale are needed to engineer a scaffold with the properties of appropriate mechanical features and to aid the cell signal transductions. To date, various techniques have been developed to form scaffolds with fibers or directly to form fibers. These include wet-spinning, dry-spinning, melting spinning, extrusion spinning, gel spinning, electrospinning, self-assembly and computer-aided 3D printing [67-68]. Among them, the wet-spinning, an old technology used in the textile industry, is

35 34 now widely employed to generate fibers down to micro-diameter scale (Figure 1.4). It starts with a polymer solution driven through a spinneret that submerged under a coagulant bath, and as it emerges, the solution precipitates and solidifies into fibers, that were subsequently drawn by mechanically pulling. One example is O. Liivak s work: the wet-spinning of silk fibers by a microfabricated apparatus [69]. Fibers in tens of micro-diameter can be achieved and these fibers can be prepared into fibrous scaffold with different techniques like heat bonding [70]; however, much smaller diameter fibers are hard to obtain due to current limitation of wet-spinning technique like the difficult to get spinneret with a smaller diameter hole. Figure 1.4. Schematic view of wet-spinning Electrospinning, on the other hand, can generate fibers from micro-diameter to nano-diameter scale (Figure 1.5). In the electrospinning process a high voltage is used to create an electrically charged jet of polymer solution. One electrode is placed into the spinning solution and the other attached to a collector. Electric field

36 35 is subjected to the end of a stainless needle that contains the polymer fluid held by its surface tension. This induces a charge on the surface of the liquid. Mutual charge repulsion causes a force directly opposite to the surface tension. As the intensity of the electric field is increased, the hemispherical surface of the fluid at the tip of the needle elongates to form a conical shape known as the Taylor cone. With increasing field, a critical value is attained when the repulsive electrostatic force overcomes the surface tension and a charged jet of fluid is ejected from the tip of the Taylor cone. The discharged polymer solution jet undergoes a whipping process [71]wherein the solvent evaporates, leaving behind a charged polymer fiber, which lays itself randomly on a grounded collecting metal screen. Studies in the early years since its invention were focused on its theory and experiment setups, and it started drawing huge attentions from academia when the tissue engineering research emerged in 1987 [72]. Rodlike fibers, ribbons and tubes were successfully fabricated by this method. For instance, J.T. McCann produced hollow nanofibers by two-fluid electrospinning [73]. Other techniques such as computer aided 3D printing were also used to generate scaffolds for tissue engineering, thus offered engineers with a variety of options. In this study, both wet-spinning and electtrospinning were utilized to generate SELP fibers.

37 36 Figure 1.5. Schematic view of electrospinning Though scaffolds produced by some of the techniques had very fine and ideal morphologies comparable to native ECM, their physical or mechanical properties were far from expectation, for example, most of the electrospun nanofibers were water-soluble, hence prevented them from future application in aqueous physiological conditions. Stabilization or reinforcement was often needed as a post-fabrication treatment. The purpose of post-fabrication treatment is to preserve the scaffold/fiber as it is and further improve or add other features: water resistance, mechanical properties or functional additives. Alcohol (methanol or ethanol) is often used as a coagulant to treat weak and unstable proteins and convert the amorphous structures into crystal configurations by extracting excessive water molecules. Glutaraldehyde s aldehyde

38 37 ends can be easily covalently bonded with the free NH end from the lysine residues, through which, proteins are crosslinked and strengthened. Other methods like the ultraviolet (UV) crosslinking of the light sensitive agents that were pre-added in the proteins [74]. At the same time, chemical free or environment friendly stabilization strategies were also sought. These include mechanical preconditioning, water vapor annealing and thermal heating. In this work, both chemical crosslinking and autoclaving or steam sterilization are used to stabilize the SELP fibers. As stated in this part, the preparation process will decide or change the properties of the scaffolds/fibers, including the diameter of the fibers and mechanical property. However, these are some of the important properties that need to be considered when preparing the scaffolds/fibers. Other property like the secondary structure conformation is also important to understand the scaffold s mechanical performance Key Properties of Scaffolds/Fibers and Characterization Methods When engineer a scaffold, several key properties are usually investigated to evaluate the specific usage of the scaffold. These include physical structure features, secondary structure conformations, and mechanical behaviors. For the structure properties, both micro level and macro level features are studied. Under the micro level, the fibers diameter, pore size and organization of the microstructure like the

39 38 orientation of the fibers are studied. For instance, D. Gupta et al engineered a scaffold with aligned nanofibers via electrospinnning gelatin on a rotating disk. These aligned fibers thus can guide the nerve cell growing [75]. Under the macro level, different 3D architectures have been engineered and studied. For example, D. Zhang constructed 3D nanofibrous tubes with controllable architectures: many multiple interconnected tubes [76]. These tubes can be designed as the same shape as the branched abdomen aorta. Beside these structural features, the secondary structure conformations of the protein-based biomaterial are also vigorously studied, because the content of the secondary structures plays an important role in the scaffold s stability, degradation rate and mechanical behaviors. For example, C. Dicko et al, found a great conformations change during the spider spinning [77], which explained the reason why the silk s mechanical properties are totally different between the silk in the spider s gland and the silk fibers spun by the spider. Actually, the silk structure in the spun fibers is called silk II, and was determined to be β-sheet, which is mostly found in the crystal part of the natural silk fibers. And the silk stored in the spider s gland is named as silk I, the structure of which was poorly understood. Currently several secondary structures including α-helix and β-turn are proposed to the silk I structure [78]. Both silk I and silk II have the same silk, however their secondary structure conformations are totally different, and the resulting physical properties and

40 39 mechanical performances are significantly different. Thus, studying the secondary structure conformations is vital. Tools to study the secondary structure usually are Fourier transform infrared spectroscopy (FTIR), Raman infrared spectroscopy, X-ray, nuclear magnetic resonance spectroscopy (NMR), circular dichroism (CD), and vibrational circular dichroism (VCD). Due to the nature of the fibers in solid state, CD is not applicable, as sample for CD is usually in solution state. And NMR is a powerful tool to study protein structure in solution state, while low signal/noise ratio and the introduction of 13 C and 15 N to the protein structure made this tool complicated when in solid state. X-ray has a good capability to identify uniform oriented crystal structures, while poor at seeing the amorphous structures. VCD has been shown more sensitive in determining secondary structures than the FTIR and Raman. However due to the availability of this instrument in hand, only FTIR and Raman were used in this work to study the secondary structure conformations of SELP fibers. Particular, micro-diameter (>15 μm) fibers was scanned by Raman, and nanofibrous scaffolds with around 5 μm in thickness were scanned by transmission FTIR, as great light absorbance of micro-diameter fibers resulted in spectra with zero transmitted signal. Another important property studied is the mechanical property of the scaffold, as this plays a decisive role both in macro level to adjust the host responses, and in micro level to regulate cell behaviors. In macro level, for example, arteries near the

41 40 heart are much muscular to bear the high blood pressure from the left ventricle. Thus a strong tubular scaffold is always needed to replace the damaged/diseased part. In micro level, cell behaviors are influenced by the micro bio-environment, especially the local biomechanics. Thus mechanical property of scaffolds either in bulk or in cell scale is an important factor to be considered when engineering a scaffold. Mechanical characterization methods including uniaxial tensile test, biaxial tensile test, preconditioning test, relaxation/creep test and other dynamic tests are usually used to quantify the materials elasticity, resilience, viscoelasticity and other parametric properties. Due to the sample nature in fiber form, and in order to obtain the most basic parameters, in this study, only uniaxial tensile test and preconditioning test were performed to obtain the elasticity and resilience. To date, mechanical properties of many protein-based biomaterials have been characterized. Table 1.5 lists some protein-based biomaterials elasticity and resilience parameters in hydrated state (similar to the physiological environment). The mechanical properties of the major ECM proteins (elastin and collagen) are also listed. From this table, we can see that, different preparations resulted in different mechanical properties even with the same material, for example, the ultimate stresses of the collagen film prepared by solvent casting and collagen fibers prepared by electrospinning and GTA crosslinking were KPa and 7.4 ± 1.2 MPa respectively. And reconstructed native materials usually have poorer mechanical

42 41 performance than the native ones. So there must be some internal microstructure changes, which determine the mechanical properties. From the studies of silk and other proteins, the secondary structures β-sheets, which form crystals, are responsible for the high strength of natural silk. Actually, the mechanical properties of silk are determined by the hierarchical structural assembly at all length scale. This include the level of crystallinity, the size and distribution of the crystallites, and the conformational and orientation organization at the molecular level [79]. Thus the analysis of secondary structure conformation is only one of the aspects. The investigations of the secondary structures of the SELPs were started since its introduction, while the mechanical properties were rarely reported. Hence, studying the mechanical properties along with the secondary structures might allow us a partial understanding of these SELPs, and thus further SELPs in tissue engineering.

43 42 Table 1.5. Mechanical properties studied for protein-based biomaterials in hydrated state Biomaterial Preparation Methods E (MPa) 1 ultimate stress (MPa) failure strain (%) 3 R (%) 2 EP [80] Self-assembly 0.25 ± ± ± ± 7 EP20-24 [80] Self-assembly 0.25 ± ± ± ± 10 SELP-47K film [59] SELP-47K fim [59] Methanol casting Methanol casting and GTA crosslinked 1.66 ± ± 0.14 N/A 86 ± ± ± 0.11 N/A 88 ± 1 Collagen [81] Solvent casting ~ ~ ~71 n/a Collagen [36] Electrospinning, GTA crosslinked n/a 7.4 ± 1.2 n/a n/a Fibrinogen [82] Electrospinning, MeOH treated 0.19~ ~ ~110 n/a B. mori. Silk [49] Electrospinning, MeOH treated n/a n/a Spinder Silk[83] Native as is 2~250 n/a ~35000 n/a Collagen [84] Native as is 2.9 ± 0.3 a / b n/a n/a n/a Aortic elastin [80] Native as is ± 2 Human coronary artery [85-86] Native as is 1.1~4.1 d 0.4~1.4 d 160~190 d n/a Human saphenous vein [87] Native as is n/a a: shear modulus; b: bending modulus; 1: Young s modulus; 2: resilience; d: Mechanical properties depend on patient s age and gender and artery s tissue composition (i.e., adventitia, media, intima) and orientation (i.e., circumferential, longitudinal)

44 Motivations and Scope of Work The recombinant DNA technology has enabled scientists to genetically design and biologically synthesize various proteins with precisely controlled amino acid sequences and peptide chain lengths [88], thus make it possible to produce high performance biomaterials integrated with cellulous functions and different chemical, physical and mechanical properties. Silk-elastinlike proteins, with the repeating sequence of silklike blocks and elastinlike blocks, have novel mechanical properties, which may make them promising candidates as biomaterials in tissue engineering. SELPs have been rigorously studied for their applications in drug/gene delivery and surgical operations since their introduction nearly two decades ago. Their biological, physical and chemical properties have been well characterized. However, the mechanical features are rarely reported, and thus prevented further investigation towards application in tissue engineering. Protein-based fibers, which are fundamental building blocks of extracellular matrix, play an important role in structure support, scaffolding, stabilization, and the protection of cells, tissues and organisms. Fabricating performance protein-based fibers has been extensively pursued in tissue engineering. Thus, fabricating SELPs fibers to engineer a scaffold is highly demanded. Advance in the study of natural silk and elastin has illuminated the relationship

45 44 between the mechanical performance and the internal crystalline microstructures found within these biomaterials. Thus, such an understanding of the microstructures found in SELPs would allow a full understanding of the mechanical performance of the materials, particularly with understanding the mechanical properties that derived from the silk and elastin. As SELPs are a potentially powerful biomaterial candidate, they provide an excellent research focus. Thus, the overall goal of this dissertation is to detail the SELP fiber fabrication process, to study the physical structures, to confirm the secondary structures and finally to understand the mechanical properties of SELPs. Therefore, in this work, a specific SELP is under investigation. SELP-47K, with a monomer structure of (S) 4 (E) 4 (EK)(E) 3, in which S is the silklike block, E is the elastinlike block, and EK is a lysine modified pentapeptide sequence: GKGVP, was fabricated into microfiber and nanofibers. SELP-47K microfibers were initially generated via wet-spinning and were chemically reinforced by crosslinking the fiber with glutaraldehyde vapor. The surface morphologies, mechanical properties, secondary structures were examined or characterized by scanning electron microscopy (SEM), uniaxial tensile test, and Raman spectroscopy. The results of these studies provided a thorough understanding of SELP microfiber. Resulting information gave insight into SELP-47K properties including the toughness and elasticity of SELPs and the underlying relationship to

46 45 the contents of secondary structures. And SELP-47K was further fabricated into nanofibrous scaffolds using another technique electrospinning. The solvent selection, process parameters and relative humidity were widely investigated. And the resultant structures before and after post-fabrication treatment were studied by Scanning Electron Microscopics and Fourier Transform Infrared Spectroscopics. Again, mechanical and secondary structure analyses were performed to obtain the tensile strength, elastic modulus, deformability, resilience and qualitative contents of secondary structures. Additionally, NIH 3T3 mouse embryonic fibroblast cells were seeded on the scaffolds to check the cell-scaffold interactions and cell proliferations which demonstrated the in vitro biocompatibility of such SELP engineered tissue. The outcome of the characterization and cell assessment studies will further strengthen the use of SELPs in tissue engineering.

47 46 CHAPTER 2: WET-SPINNING OF MICROFIBERS 2.1 Introduction In this part, SELP-47K proteins were fabricated into microfibers via wet-spinning. Raman spectral analysis revealed the formation of anti-parallel β-sheet crystals of the silk-like blocks. Dry SELP-47K fibers display the dependence of mechanical properties such as Young s modulus on fiber diameter, suggesting more oriented and crystallized molecular chains in small-diameter fibers. Additionally, a brittle fracture mode was identified for dry fibers by SEM analysis of fracture surfaces. Hydration dramatically influenced the mechanical behavior of SELP-47K fibers. In contrast to the high tensile strength and limited strains to failure of dry fibers, fully hydrated SELP-47K fibers possessed strains to failure as high as 700%. Furthermore, upon chemical crosslinking, a tensile mechanical strength up to 20 MPa was achieved in hydrated fibers without compromising their high deformability. 2.2 Material and Methods Sample Preparation Frozen SELP-47K aqueous solution at a concentration of 13% w/w was generously provided by Protein Polymer Technologies, Inc. (San Diego, CA). The SELP-47K solutions were lyophilized and re-dissolved in 98% formic acids (VWR) at a concentration of 200 and 250 mg/ml for fiber spinning. Original aqueous solution

48 47 was also tried for wet-spinning Wet-spinning: Experiment Design A custom-made wet-spinning device was used for the fabrication of SELP-47K microfibers (Figure 2.1). Briefly, a wet-spinning procedure consists of extrusion of protein solutions, formation of a protein thread, coagulation of the protein thread into a fiber, washing extra coagulant remaining in the fiber, air-drying, and collecting the fiber. In wet-spinning, an apparatus housing a ceramic capillary (Small Precision Tools, CA) of 28 µm, 38 µm or 100 µm in diameter is wired to a syringe pump (model R99-FM, Braintree Scientific, MA). SELP-47K fibers were spun into a coagulant bath of methanol/water, air-dried, and then collected by custom-made rotating Teflon mandrels driven by a DC-powered motor (Sterling Instrument, NY). The distance between the collector and bath is fixed at 30 cm and the pump feed rate is adjusted to 0.5 ml/hr. The motor rotation speed is adjusted to an optimal speed as long as there are no excess fibers entangled in the bath.

49 48 Figure 2.1. Experiment setup of wet-spinning Crosslinking SELP-47K fibers were crosslinked overnight by glutaraldehyde (GTA) (Mallinckrodt Baker) vapor in a vacuumed dessicator (Figure 2.2). The bottom of the dessicator was filled with 25% GTA in water while SELP-47K fibers collected on a glass spool were placed on a ceramic plate with holes, which separated the bottom chamber with GTA from the top chamber. Figure 2.2. Wet-spun fiber placed in a vacuumed dessicator, the bottom is filled with

50 49 liquid glutaradehyde Surface Morphology by Scanning Electron Microscopy The microstructures (e.g., fiber diameter, surface morphology) of air-dried SELP-47K fibers were examined using a Hitachi-S3400 scanning electron microscope (SEM). In order to examine the fiber stability, some wet-spun fibers were extensively rinsed and then stored in 1 PBS at room temperature. After one year, fibers were taken out and air-dried for a SEM analysis. Likewise, the surface morphology of fractured dry fibers was also examined using a SEM Secondary Structure Characterization by Raman Spectroscopy Raman spectra of the wet-spun SELP-47K fibers were recorded on a Thermo Nicolet Almega microraman system (Thermo Scientific). A solid-state laser with the wavelength of 532 nm was used as the exciting source. The non-negative least-squares (NNLS) analysis of Raman amide I and III spectra that was first proposed by Williams and Dunker [89] and subsequently refined by Williams [90-91] was pursued to estimate the secondary structural contents, including helix, β-strand, β-turn, and undefined conformation, as follows: Ax b (Equation 2.1) f = Fx (Euqation 2.2) where A is a p n matrix containing the normalized Raman amide I or III spectra of n

51 50 reference proteins at p wavenumbers, column vector b contains the normalized Raman amide I or III spectrum of wet-spun SELP-47K fibers at the same p wavenumbers, column vector x is used to calculate percentage contents of secondary structures for SELP-47K fibers as a weighted sum of the secondary structures of the n reference proteins, F is an m n matrix representing m classes of secondary structures of the n reference proteins, and column vector f then provides percentage contents of secondary structure for SELP-47K fibers. As detailed elsewhere [91], 16 and 13 reference proteins were used in Raman amide I and III spectral analysis, respectively. Moreover, the A and F matrices in equations (1.1) and (1.2) were directly taken from Williams work [91]. Still, following the Williams procedure, the amide I band intensities of SELP-47K fibers from 1615 to 1710 cm -1 at an interval of 5 cm -1, and the amide III band intensities from 1225 to 1245 cm -1, from 1275 to 1310 cm -1 at an interval of 5 cm -1, and again at 1235 and 1285 cm -1 were tabulated into column vector b. Unlike proteins analyzed by Williams, SELP-47K contains less than 0.4% Tyr, Phe and Trp, and thus subtraction of side chain bands due to these residues is unnecessary, evident by the minimal intensity of the Trp band at 1555 cm -1. Additionally, subtraction of the spectrum of buffer is not relevant as the Raman spectra were collected on air-dried SELP-47K fibers. Nevertheless, base correction was performed using Grams 8.0 over the spectral region of 1800 to 1200 cm -1. Likewise, the NNLS analysis of the Raman amide I and III spectra of SELP-47K fibers was implemented

52 51 using MATLAB Mechanical Characterization The mechanical properties of both dry and hydrated (in 1 PBS at 37 o C) SELP-47K fibers were characterized using a high-resolution dynamic mechanical analyzer (DMA, PerkinElmer). Due to difficulty to handle single fibers, only fibers with diameter greater than 20 µm were mechanically analyzed. Dry fibers and hydrated fibers were elongated to break at a constant cross-head speed of 20 µm/min and 250 µm/min, respectively. The gauge length of dry fiber samples was 20 mm. Short samples with a gauge length of 5 mm were also analyzed for comparison. Due to the high extensibility of hydrated SELP-47K fiber samples and the limitation of the maximum travel distance of the drive shaft of DMA, short samples of 3 or 5 mm in length were used in mechanical analysis of hydrated samples. Data were successfully collected on 28 dry fiber samples with diameters ranging from 20~60 µm. Despite more than 10 trials, stress-strain analysis was only obtained on two fully hydrated fiber samples because of the limitation of the force measurement (~10 µn) of the current facility. GTA crosslinking greatly enhanced the mechanical stability of fully hydrated SELP-47K fibers. As a result, 9 fiber samples were successfully analyzed when fully hydrated in 1 PBS at 37 o C. The Young s modulus, ultimate tensile strength, and strain at failure of SELP fibers were obtained from stress-strain analysis.

53 52 The calculation equations are listed here. Area = π 2 4 σ = F Area (Equation 2.3) (Equation 2.4) ε = L L 0 (Equation 2.5) E dry = σ ε ϵ(0 1%) E wet = σ ε ϵ(0 20%) (Equation 2.6) (Equation 2.7) Where Φ is diameter of the fiber, σ is the engineer stress, ε is engineer strain, E dry is the Young s modulus of dry fiber, E wet is the Young s modulus of hydraged fiber. 2.3 Results and Discussions Wet-spinning of SELP-47K Microfibers The formation of continuous fiber needs several requirements: 1) solution with a lower surface tension, so that solution stream can be prevented from breaking up into droplets: the formic acid solution comparing to aqueous solution has a lower surface tension, thus wet-spinning of original aqueous solution resulted in droplets in the methanol bath, while fibers were formed when SELP solution in formic acid was tried; 2) enhanced molecular chain entanglement, so that continuous fibers without breaking can be obtained: 200 mg/ml SELP in formic acid resulted in short segments fibers while 250 mg/ml SELP in formic acid resulted in continuous fibers; 3) coagulant bath: methanol is an ideal alcohol that has a strong capability to remove

54 53 excess water molecules from the protein solution, thus protein solution was cipitated and solidified. Therefore protein solutions of SELP-47K copolymers at a concentration of 250 mg/ml in formic acid were successfully spun into fibers of meters in length (Figure 2.3 A). A SEM study demonstrated the formation of uniform, clean microfibers (Figure 2.3 B, C). Interestingly, before they are air-dried, two fibers can merge together and form inter-fiber bonding at the point of contact (Figure 2.3 D, E). Additionally, the formation of hollow fibers was observed (Figure 2.3 F). A B C D E F

55 54 E F Figure 2.3. Wet-spun SELP-47K fibers. Micro-diameter SELP-47K fibers of meters in length were collected on a glass spool (A); SEM micrographs of SELP-47K fibers with different diameters and surface morphologies (B, C); at the point of contact, two fibers can merge and form inter-fiber bonding before being air-dried (D, E); some fibers are partially hollow (F). The wet-spinning of SELP-47K with a monomer structure of (S) 4 (E) 4 (EK)(E) 3 results in the formation of fibers with smooth surfaces (Figure 2.3 and Figure 2.4). In contrast, wet-spun SELP5 fibers with a monomer structure of (S) 8 (E) 16 displayed horizontal lines on the surface along the fiber axis [92]. Likely, the fewer silk-like blocks per repeat and the incorporation of lysine in SELP-47K are responsible for the dramatic difference in the fiber surface morphology. Our prior biodegradation studies revealed that a SELP5 film implant in rats retained nearly 100% of its original mass over the course of 7 weeks while a SELP8 film with a monomer structure of (S) 4 (E) 8 lost over 80% of its original mass over the same time period [93]. Given the similar molecular weights and the same ratio of the silk- to elastin-like blocks in SELP5 and SELP8, the doubled number of silk-like blocks per repeat in SELP5 thus dramatically enhances the stability of protein polymer structures, likely due to greater crystallization. SELP-47K has a monomer structure very similar to SELP8, except

56 55 that the second valine of the pentapeptide GVGVP is replaced by a lysine in one of every eight elastin-like blocks. Compared to SELP5, the reduced number of silk-like blocks per repeat in SELP-47K leads to less crystallization of wet-spun fibers. Additionally, the incorporation of lysine residues, which are positively charged in formic acid, further inhibits the formation of large, continuous fibrils of β-sheet crystals in the wet-spun SELP-47K fibers. The small texture features on the end surface of the SELP-47K fibers (Figure 2.4) are due to coagulation effects. Figure 2.4. Morphology of the end surface of a wet-spun fiber Figure 2.5. Surface morphology of wet-spun SELP-47K fibers that had been hydrated in 1 PBS for one year When hydrated in 1 PBS at room temperature, no significant change of the SELP-47K fiber morphology occurred over a time period of 30 days (data not shown),

57 56 suggesting excellent material stability of the wet-spun fibers. Significantly, the wet-spun fibers remained stable in 1 PBS at room temperature over a time period of one year (Figure 2.5). Surprisingly, all samples that were examined using a SEM displayed smooth horizontal lines on the surface along the fiber axis. In contrast, no horizontal lines were observed in fresh fibers (Figure 2.3). The mechanism of the dramatic morphological changes of wet-spun SELP-47K fibers hydrated in 1 PBS for one year was not well understood. In this wet-spinning experiment, three capillaries with different inner diameters (28, 38 and 100 µm) were used for fiber spinning. Under the same collecting condition (collecting height, and collecting rotation speed), the diameter of the resulting fibers has a direct dependence on the opening size of the capillary (Figure 2.6). Fibers with diameters of 64 ± 20 µm, 29 ± 5 µm and 22 ± 4 µm were obtained with 100, 38, and 28 µm ID capillaries respectively. When 28 µm ID capillary was used, the fluid pressure built-up in the capillary resulted in spinneret breaking up, thus lower feeding was attempted, however, the coagulation effect made the spinneret easily clogged. Therefore, wet-spinning with 28 µm ID capillary was not in pursuit. 38 and 100 µm ID capillaries were utilized for fiber spinning.

58 57 64 ± 20 µm 22 ± 4 µm 29 ± 5 µm Figure 2.6. The diameter of the resultant fibers has a linear relationship with the diameter of the capillary Secondary Structures Analysis by Raman Spectroscopics The Raman spectrum of crosslinked fibers has a lower signal-to-noise ratio (Figure 2.7), which is due to the absorbance of C=N bond at wavelength of 300 nm and 325 nm [94], thus, this spectrum was not analyzed.

59 58 Figure 2.7. Raman spectrum of crosslinked wet-spun fiber However, the Raman spectrum of as-pun dry wet-spun SELP-47K fibers is collected and given in Figure 2.8, and the band assignment is detailed in Table 2.1. In the cm -1 spectral region, three major bands are ascribed to the methyl CH stretching doublet at 2972 and 2875 cm -1 and to the anti-symmetric stretching of CH 2 groups at 2935 cm -1. The CH 2 symmetric stretching band is missing, likely due to an overlap with that of the CH 3 symmetric stretching. Two weak bands at 2771 and 2729 cm -1 are assigned to the stretching vibration of C-(CH 3 ) 2 groups of valine and to that of CH-CH 3 groups of alanine, respectively. Typically, Raman bands between cm -1 are neglected in spectroscopic analysis owing to their significantly weak intensities relative to those bands above 2800 cm -1. However, Lawson et al.[95] demonstrated that these bands are particularly useful in analyzing the relative locations of the methyl groups and the presence of CH 2 C and CH 3 C structural features. A broad band at 3287 cm -1 is assigned to the stretching of hydrogen-bonded NH groups. Furthermore, a broad shoulder observed at 3400 cm -1 suggests that the OH

60 59 groups of serine (S) are largely hydrogen-bonded, because the characteristic band of free OH groups should be at ~3500 cm -1. A similar red shift of ν(oh) stretching bands induced by hydrogen bonding also has been reported in L-serine crystals [96]. Nevertheless, the stretching of free NH groups also contributes to the broad shoulder at 3400 cm -1. The NH and OH stretching bands, which are extremely sensitive to the changes of the hydrogen bonding structures, have been recently used to study the ph-induced conformational transitions in the poly-l-lysine model system [97]. Analysis of the cm -1 spectrum of the SELP-47K fibers reveals the marker bands of Silk II, including amide I at 1665 cm -1, amide III at 1228 cm -1, CC skeletal β-sheet stretching at 1085 cm -1, CH 3 rocking at 973 cm -1, and CH 2 rocking at 879 cm -1. These characteristic bands of Silk II have been identified in silk processed from wild and cultivated cocoons [98], as well as in model peptides comprising the repeating GAGAGS hexamers [99]. The observed marker bands of Silk II in the SELP-47K wet-spun fibers suggests that the silk-like blocks, (GAGAGS) 4, although segregated by the elastin-like blocks, form anti-parallel β-sheets. Additionally, a band at 1162 cm -1 ascribed to νcc and δc-oh has been observed at 1158 cm -1 in Silk II-Cp but at 1173 cm -1 in Silk I-Cp. This further confirms the existence of Silk II in the SELP-47K wet-spun fibers. However, the relatively weak bands at 1410 cm -1 (δc α H 2 ), 1320 cm -1 (amide III of the β(ii)-turn conformation) [99], 1119 cm -1 (νcc skeletal stretching,[100] and ρc(ch 3 ) 2 [101]) and 956 cm -1 (ρch 3 ) indicate the presence of

61 60 Silk I in the SELP-47K fibers, too. Noteworthy is the non-symmetric amide III band at 1228 cm -1, which may overlap other characteristic amide III bands of Silk I at high wave-numbers such as 1245 and 1270 cm -1. Interesting, these marker bands of Silk I have also appeared in elastin-like proteins (ELP), including cyclo-(vpgvg) 3 and poly(vpgvg) [102]. Likely, the β(ii)-turn conformation adopted by both Silk I[103] and ELP[102] results in the same marker bands. Thus, the elastin-like blocks of the SELP-47K fibers certainly result in the weak marker bands of Silk I. The silk-like blocks of Silk I may also contribute to the appearance of these bands. But it remains challenging to differentiate the contributions from the elastin-like blocks and the silk-like blocks in the Silk I form. Figure 2.8. Raman spectrum of wet-spun SELP-47K microfibers.

62 61 Table 2.1. Band Assignment of Raman Spectrum wave-numbers a approximate assignment of vibrational mode b 3400 br, sh ν(oh) stretching, H-bonded, and free ν(nh) stretching 3287 m, br ν(nh) stretching, H-bonded 2972 s ν as (CH 3 ) 2935 s ν as (CH 2 ) 2875 s ν s (CH 3 ) 2771 vw ν(c(ch 3 ) 2 )[95] 2729 w ν(ch-ch 3 ) aliphatic[95] 1665 s ν(co) amide I 1557 vw ν(cc) 1450 s δ(ch 2 ) scissoring 1410 w CH 2 wagging[101] 1334 m δ(ch 3 ) 1320 m δ(ch 2 ) 1228 s ν(cn) amide III β-sheet[104], δ(ch 2 )[98] 1162 m ν(cc), δ(coh)[104] 1119 w ν(cc) skeletal stretching[100], ρ(c(ch 3 ) 2 )[101] 1085 m ν(cc) skeletal random coil[104], β-sheet[78] 1024 m ν(cc) skeletal[104] 1010 m ν(cc) skeletal[104], ν(cn)[101] 973 s ρ(ch 3 )[104] 955 s ρ(ch 3 )[104] 936 s ν(cc)[101] 879 m ρ(ch 2 )[104] 854 m ν(cc) of Pro ring[105] 839 m Pro ring[106] 753 w, br ρ(ch 2 ) in-phase[104] 513 w, br δ(ccc) of Val[107]

63 62 a br: broad; sh: shoulder; s: strong; m: medium; w: weak; vw: very weak. b ν: stretching; δ: bending; ρ: rocking. The percentage contents of an α-helix, β-strain, β-turn and undefined structure may be estimated using the NNLS analysis of Raman amide I and III spectra, which was first proposed by Williams and Dunker [89] and subsequently refined by Williams [90-91]. Because there are lots of errors including the slight difference of the vibration frequencies between the reference protein and this protein, and errors from the reference proteins, thus this method is a semi-quantitative approach. Nevertheless, analyses of the amide I and III spectral led to consistent estimations of the β-turn content and of the undefined conformation content (Table 2.2). And an error of around 10% was observed in estimations of the helix contents from the amide I and III analyses. Likewise, estimations of the β-sheet content from the Raman amide I and III analyses led to around 12% error. Despite the large discrepancy in the estimations of the helix and β-sheet contents, both the Raman amide I and III spectral analyses suggest that β-sheet are the dominated conformers of wet-spun SELP-47K fibers. This could explain why the SELP-47K fibers are partially stable in PBS for one year. Table 2.2. Secondary structural contents (%) estimated by Raman Amide I/III spectral analysis undefined β-turn helix β-sheet 12/10 # 21/21 18/8 49/61 # a/b: a results from the Amide I, b results from the Amide III

64 63 An attempt was also made to quantify the amount of Silk I in the SELP-47K I fibers by the 1410 I1450 intensity ratio. Monti et al.[104] proposed to use the I 1415 I 1455 intensity ratio as a semi-quantitative measurement of the amount of Silk I, which is largely comprised of the β(ii)-turn conformation. If compared to Silk I chymotryptic precipitate (Cp), liquid silk, and film, the methylene bending and wagging bands were red shifted from 1455 cm-1 to 1450 cm-1 and from 1415 cm-1 I to 1410 cm-1, respectively, in the SELP-47K fibers. Thus, the 1410 I1450 intensity ratio was calculated for the SELP-47K fibers after the baseline correction. The intensity ratio is 0.21, which is much lower than that of 1.12 in Silk I Cp. Noting I that the 1415 I1455 intensity ratio is not equivalent to the percentage content of β-turns. In fact, an infrared spectral analysis from the same group revealed 55% β-turn conformation in Silk I Cp.[108] Compared to an estimation of 21.5% β-turn from the NNLS analysis of the amide I and III spectra (Table 2.2), the intensity ratio seems to underestimate the amount of β-turn conformation in the SELP-47K fibers. This may be due to the fact that ELP also displays a strong band at 1450 cm-1, although it largely adopts β(ii)-turn conformation.[102] Therefore, the intensity ratio of the CH2 wagging and scissoring modes seems inappropriate for semi-quantifying the amount of β(ii)-turn conformation in the SELP-47K fibers. Comparing to the natural protein fibers including silk fibers and elastin fibers (Table 2.3), the wet-spun SELPL-47K fibers have a β-sheet content close to the silk

65 64 (49/61% to 37%, 50%), while far from the elastin (49/61% to 16%). This shows a hint that the mechanical performance of the SELPL-47K fibers might be close to the silk while far from the elastin. Table 2.3. Secondary structural contens (%) of natural fiber undefined β-turn helix β-sheet N. Clavipes [79] B. mori [79] Aortic elastin ([109]) Mechanical Properties Analysis The tensile analysis of dry SELP-47K fibers revealed a breaking elongation of less than 2% and an ultimate tensile strength of 20 to 80 MPa (Figure 2.9). The variance of this data is due to several possible errors including error of protein concentration during preparation, collecting speed variance of the motor and diameter measurement error. Nonetheless, the test data of all the samples fall in the same range (Table 2.4).

66 Figure 2.9. Tensile stress-strain analysis of dry SELP-47K fibers 65

67 66 Table 2.4. Mechanical Properties of as-spun dry fibers Sample ID a Diameter φ (μm) Young s Modulus E(GPa) Ultimate Stress σ T (MPa) Ultimate Strain ε F (%) D D D D D D D D D D20-21* D D20-27* D D D20-20* D D D05-08* D D20-08* D20-19* D D D D05-06* D05-09* D D20-24* MAX MIN AVE STDEV a D20-XX indicates an initial sample gauge length of 20 mm while D05-XX indicates a sample length of 5 mm. *Samples broke at the clamps.

68 67 While compared to the natural silk fibers (Figure 2.10), the wet-spun fibers are more brittle and much weaker. The significant reduction in the tensile strength and extensibility may be due to the incorporation of elastin-like blocks, because dry elastin is very brittle but displays good elasticity when hydrated [110]. Figure Stress-strain comparison between SELP-47K fibers to other natural silk fibers [111] As compared to the recombinant elastin with GVGVP blocks in dry state which has an ultimate strength of 35MPa, and an ultimate strain of 4.2%, this SELP-4K has mechanical properties that are close to elastin. As stated in the secondary structure conformation part, this SELP-47K fibers have β-sheet contents close to the native silk. It seemed that the high content of β-sheet is not relevant to the mechanical strength. Actually, the mechanical properties of silk are determined by the

69 68 hierarchical structural assembly at all length scale. This include the level of crystallinity, the size and distribution of the crystallites, and the conformational and orientation organization at the molecular level [79]. Therefore, high contents of β-sheet do not directly link to the mechanical properties. Further study like quantifying the orientation of the secondary structures by the polarized Raman scan [79] is needed to correlate to the mechanical properties. But at least, the result of the secondary structural analysis could explain the reason why the fiber is partially stable in PBS for one year. Besides the strain-stress, the Young s modulus is of great interest, as it reveals the stuffiness of the material. The Young s modulus of the SELP-47K fibers is in the range of 1 to 5 GPa (Figure 2.11), comparable to those of tendon collagen and dragline silk [112]. And moreover, the Young s modulus of the fibers has a strong dependence on the fiber diameter (Figure 2.11). Typically, fibers of smaller diameter possessed higher Young s modulus and thus higher tensile strength. Liivak et al [69], also reported an increase in the maximum stress of wet-spun silk fibers when the fiber diameters decreased. Such trends in fibers have been established as partially the result of an increased fiber orientation in smaller fibers, which may be obtained using smaller spinnerets and/or through larger post-fabrication stretching. In smaller spinnerets, a protein solution thread experiences greater shear. Together with the elongational stress induced by the post-fabrication stretching, the enhanced shear

70 69 stress acting on the SELP-47K solution thread causes the silk-like blocks to better crystallize, resulting in the formation of protein fibers with more orientated filaments. The fracture surface of dry SELP-47K fibers was analyzed by SEM (Figure 2.12). Consistent with the low deformability revealed by the tensile analysis, the smooth fracture surfaces suggest brittle facture as a failure mode for the dry SELP-47K fibers. Additionally, small kinks (marked by letters and arrows in the upper and lower figures) on the fracture surfaces indicate non-homogenous microstructures of the dry fibers. We further speculate that the deformation of a less perfect round fiber may be localized and become un-symmetric under tensile strain. A crack may be first initiated in region A (Figure 2.12, left), and propagate as catastrophic brittle failure. However, as the crack propagates, the accumulated strain energy will be gradually released. When the weakened crack front encounters small crystallized phases, the crack will propagate along the soft interface of the amorphous and crystal phases in the fiber-axial direction. Because no large, continuous crystal fibrils were formed in the wet-spinning of SELP-47K fibers, it is anticipated that the crack propagation will be back to its original direction, leaving small kinks (markers B, C, D, and E in Figure 2.12, left and right) on the fracture surfaces. This is dramatically different from the split longitudinal or fibrillation fracture of collagen fibers, which are composed of continuous fibrils [113].

71 70 Figure The Young s modulus of dry fibers is a function of fiber diameter Figure The fracture surface of dry SELP-47K fibers. In contrast to dry SELP-47K fibers of high tensile strength but limited strain to failure, hydrated fibers display greater strain to failure but reduced tensile strength (Figure 2.13). These properties suggest that hydration, likely due to the plasticizing effect of water, decreases the protein chain crystallinity in the fibers. Actually, as explained by Zhengzhong Shao [114] that when silk fiber is hydrated in water, water molecular enters the some of the secondary structures and breaks down part of the hydrogen bonding, thus significantly decreases the mechanical strength. When the

72 71 SELP-47K fibers were hydrated in 1 PBS at 37 o C, a reduction in Young s modulus of more than 1,000-fold was observed. The deformability of hydrated fibers, however, increased several hundred- fold, compared to that of their dry counterparts (Figure 2.11). Accordingly, the ultimate tensile strength of SELP fibers decreased from several tens of MPa to about 1 MPa. Because the hydrated SELP-47K fibers were very soft (e.g., Young s modulus of around 0.2 MPa), low signal-to-noise ratio was a big issue and a reproducible tensile stress-strain analysis was difficult to obtain. When crosslinked by glutaraldehyde, the hydrated SELP-47K fibers were strengthened dramatically, possessing an ultimate tensile strength as high as 20 MPa (Figure 2.14 and Table 2.5). Although moderate reduction in their deformability was observed, the crosslinked SELP-47K fibers were stretched to over 200% strain without breaking. Strain to failure of some fibers even reached 700% strain. Impressively, the ultimate tensile strength of SELP fibers closely matched that of native collagen fibrous networks in bone tissue[115]. And the ulatimate strength and strain of the native elastin and human saphenous vein are much smaller than the wet-spun crosslinked fiber (Figure 2.14). The significant increase in mechanical strength is due to the reduction of the free end-to-end molecular chains as a result of GTA crosslinking.

73 72 Figure Representative tensile stress-strain analysis of SELP-47K fibers fully hydrated in 1 PBS at 37 o C. Figure Tensile stress-strain analysis of glutaraldehyde-crosslinked SELP-47K fibers fully hydrated in 1 PBS at 37 o C with comparison to human saphenous vein and aortic elastin (Table 1.5).

74 73 Table 2.5. Mechanical properties of hydrated, crosslinked SELP-47K fibers Sample ID φ (μm) E (MPa) σ F (MPa) ε T (%) φ: fiber diameter; E: Young s modulus; σ T : ultimate tensile strength; ε T : ultimate tensile strain. * Modulus are calculated in the region of 10% strain 2.4 Summary SELP-47K was spun into robust, clean microfibers using a wet-spinning technique. Raman spectroscopic analysis revealed the co-existence of structures consisting of Silk I of a β-turn conformation and Silk II in the form of anti-parallel β-sheet. An estimation of 50% to 60% β-sheet structures, together with the morphology analysis of wet-spun fibers and fiber fracture surfaces, indicate that no large, continuous fibrils were formed during the wet-spinning. Hydrated SELP-47K fibers were mechanically weak (e.g., Young s modulus of 0.2 MPa and tensile strength of 1.2 MPa) but displayed large strains to failure (e.g., 700%). In contrast, dry SELP-47K fibers possessed high tensile strength (e.g., 40 to 60 MPa) but limited strains to

75 74 failure (i.e., less than 2%). When chemically crosslinked using glutaraldehyde, the mechanical strength of hydrated SELP-47K fibers was enhanced dramatically. High tensile strength up to 20 MPa and deformability of 200% to 700% make the chemically crosslinked SELP-47K fibers very appealing for potential applications in tissue engineering. These wet-spun fibers are in the diameter range of tens of micrometer, thus preparing into a self-standing fibrous scaffold is possible[70]. On the other hand, fibers in nano-diameter scale are needed to construct a scaffold that can aid cell signal transduction. Therefore further SELP fiber fabrication is proceeded via electrospinning.

76 75 CHAPTER 3: ELECTROSPINNING OF NANOFIBERS 3.1 Introduction In this chapter, electrospun nanofibers were discussed from the fabrication process, post-fabrication treatment, secondary structural analysis and mechanical test, to in vitro cell culture. The technique, electrospinning was utilized to fabricate SELP-47K into nanofibes to form a completely protein-based, mechanically robust tissue scaffold. Many factors, both fabrication process and post-fabrication treatment, will affect the resulting scaffolds. Factors in the fabrication process include: (1) the solution property: the polymer itself, solvent and concentration; (2) the processing parameters: the needle size, infuse rate, voltage and travel distance; and (3) the environment: the ambient room temperature, relative humidity and air pressure. Thus, SELP-47K was dissolved in both formic acid solution and aqueous solution, and was electrospun into nanofibrous scaffolds under different conditions, which were stabilized via different chemical vapor treatments (i.e., methanol, glutaraldehyde, combined methanol and glutaraldehyde) or physical treatment: autoclave. The silklike sequence of SELPs is capable of crystallizing to provide the SELP structures mechanical strength. Treatment using methanol or other non-solvents can accelerate this crystallizing process. Additionally, lysine residues present in SELP-47K permit chemical crosslinking of the elastinlike blocks using glutaraldehyde. Alternatively, like frying an egg, autoclave is a chemical-free environment friendly approach, with

77 76 the employment the temperature, pressure and water steam to stabilize the SELP-47K scaffolds. Changes in the secondary structures of the SELP scaffolds due to these post-fabrication treatments were examined using Fourier transform infrared (FTIR) spectroscopy. The mechanical properties of the SELP scaffolds, including tensile strength, elastic modulus, deformability, and resilience, were evaluated. For chemical crosslinked nanofibrous scaffolds, 3.4~13.2 MPa of modulus, 5.7~13.5 MPa of ultimate stress and % of failure strain were obtained. However the autoclaved scaffold had lower numbers: 1.0 MPa elastic modulus, 0.3 MPa ultimate strength and 29% deformation. Additionally, the scaffolds from the formic acid solution were first evaluated for the in vitro cell biocompatibility through NIH3T3 mouse embryo fibroblast cell culture. 3.2 Materials and Methods Experiment Setup A custom-built electrospinning device, comprised of a Gama High Voltage DC power, a syringe pump (Braintree Scientific, MA) and a m plastic cubic house with a dryer or humidifier inside, was utilized to electrospin. The major part of the electrospinning part was shown in Figure 3.1. Briefly, protein solutions of SELP-47K copolymer formed a droplet at the orifice of a stainless steel needle. The

78 77 droplet was then stretched and splayed into a series of fine filaments under a high-voltage electric field. The generated SELP-47K nanofibers were collected on aluminum foil, silicon die, or glass slides, which were placed on a grounded steel plate as a secondary collector. Optimal parameters were fixed and maintained; these include: 20KV DC voltage, 12 cm travel distance between the needle tip and collector, 0.1 ml per hour of solution feed rate and a 32 gauge of stainless needle (ID 0.10 mm) for formic acid solution electrospinning or a 23 gauge (ID 0.33 mm ) needle for aqueous solution fabrication. Figure 3.1. Electrospinning experiment horizontal setup Electrospinning: Design of Experiment Again, the same frozen SELP-47K aqueous solutions at a concentration of 13% w/w were provided by Protein Polymer Technologies, Inc. (San Diego, CA). The SELP- 47K solution was lyophilized and re-dissolved for electrospinning.

79 78 Two groups of electrospinning experiments were carried out based on the solvents: 98% formic acid (VWR) and distilled de-ionized (DD) water (VWR). For the first, SELP-47K was dissolved in formic acid into various concentrations: 50, 75, 150, 200 and 250 mg/ml and were electrospun under ambient room temperature and relative humidity which were 23 ± 2 o C and 24 ± 2% respectively. One scaffold of about 10 µm in thickness was first obtained from the electrospinning of 150 mg/ml SELP-47K solution for 9 hours, and chemically crosslinked for subsequent characterizations and cell study. Another scaffold of about 200 µm in thickness, later, resulted from a 2-day electrospinning from a 150 mg/ml solution, was for the autoclave study. As for the DD water part, SELP-47K aqueous solutions were prepared into different concentrations: 70, 200, 250, 300,400 and 450 mg/ml, and were electrospun under ambient room temperature and relative humidity to study the microstructures dependence on the concentrations. Additionally, to study the influence of relative humidity, 200 mg/ml aqueous solution was electrospun under different relative humidity (ie., 18, 24, 35 and 48%). Both solvents have their advantages and disadvantages: the solution of SELP-47K in formic acid will preserve itself in solution form under room temperature at least for one week, thus continuous fabrication is possible, while the formic acid is a harsh chemical which is not environment friendly. In contrast, SELP-47K in DD

80 79 water will undergo irreversible gelation, making it impossible for further electrospinning, however, DD water is an environment friendly solvent and usually proteins functions will be preserved in water, hence make this choice very promising Scaffold Post-fabrication Treatment For the chemical treatment, SELP-47K nanofibrous scaffolds (from the 150 mg/ml acid solution) along with the steel collector, were treated for 48 hrs with methanol (MeOH), glutaraldehyde (GTA) or combined MeOH and GTA vapor (48 hrs with MeOH and 48 hrs with GTA) in a vacuumed dessicator. A petri dish containing 10 ml of 25% GTA [116] [117] or 99.9% MeOH was placed at the bottom of the dessicator. SELP-47K scaffolds collected on aluminum foil or glass slides were placed on a ceramic plate with holes, which separated the bottom chamber with chemicals from the top chamber. Treated SELP-47K nanofibrous scaffolds were then peeled off the aluminum foil for Scanning electron microscopy (SEM), FTIR and mechanical analysis, or kept on the glass slides for further treatment and cellular studies. For the autoclave treatment, fabricated scaffolds (from the 150 mg/ml acid solution) along with the steel plate were autoclaved at 134 o C, 2 bar for 30 or 60 min in the chamber of Tuttnauer 2340M autoclaver. After autoclaving, the scaffolds were cooled, dried and stored in a vacuumed desiccator.

81 Surface Morphology and Secondary Structures Characterizations Electrospun scaffolds before and after treatment were examined via both scanning electron microscope (SEM) and Fourier Transform Infrared Spectroscopy (FTIR). Vacuum-dried scaffolds were first coated with platinum for 30 s and then inspected under a Hitach-4800s field emission SEM to check the surface microstructures. Very thin scaffolds (< 5 μm in thickness) were prepared and oven heated overnight at 80 0 C before FTIR scanning. The contents of secondary structures were estimated by analyzing the spectra obtained from a Magna-IR 560 Nicolet spectrometer (Madison, WI) equipped with a CsI beam splitter and DTGS-detector. After purging CO 2 free dry air for 30 min, the scaffold was scanned in the range of cm -1 for 400 times at the resolution of 4 cm -1. Three replicate points were scanned. Data was collected and processed using Omnic 6.0 (Nicolet) Mechanical Characterization Because of the future application under physiological conditions, this time, hydrated SELP-47K scaffolds was tested on a PerkinElmer dynamic mechanical analyzer (DMA). For the rough 10 µm scaffolds, after chemical crosslinking, were first cut into rectangular strips, the width and thickness of which were measured using an optical

82 81 microscope (Figure 3.2). Samples were mounted onto the DMA with a gap of 3 mm between the clamps, which was further reduced to 1.5 mm to prevent any stress buildup induced by the shrinkage of samples upon soaking. Samples were immersed in a jacketed beaker filled with 1XPBS at 37 o C, equilibrated for 1 hr before test, and evaluated by each of two mechanical test protocols. (1) Monotonic stress-strain. Samples were stretched to break, and the Young s modulus, ultimate tensile strength, and strain at failure of SELP-47K scaffolds were obtained. Due to imperfect sample loading, a sample may be slightly buckled and its gauge length may not be exactly 3 mm. The true gauge length of samples, which was close to 3 mm, was determined from the force-extension curve, as illustrated in Figure 3.3. (2) Mechanical preconditioning. Samples of about 3 mm in length were cyclically stretched for 6 cycles between 1.5 mm and 4.0 mm (Figure 3.4A), and then stretched to break. This will lead to preconditioning strain of about 33% (displacement of 1 mm over the original sample length of 3 mm) and an off-loading period of 5 min between cycles. Throughout this study, loading and unloading were controlled by displacement at a fixed rate of 250 μm/min. Figure 3.2. Thickness measurement of thin sample: schematic illustration (A) and

83 82 image under optical microscope (B) Figure 3.3. A representative force-clamp gap curve (A) with a zoom-in view at small deformations (B). The clamp gap was initially 3 mm and reduced to 1.5 mm after a sample was hydrated to prevent any hydration-induced tension. B suggests that the true gauge length of samples was close to 3 mm. Figure 3.4. Preconditioning displacement (A), force measurement (B), and resistant force exerted by the PBS on the clamp, which was measured by running the device without a sample (C). A sample of about 3 mm in length was cyclically stretched between 1.5 mm and 4.0 mm. This leads to a preconditioning strain of 33% (1 mm/3 mm), and an off-load time of about 5 to 6 minutes between loading cycles. For the rough 200 µm scaffolds, after 60 min autoclave, were cut into rectangular strips, soaked in 37 o C 1 PBS for 1 hour and imaged under optical

84 83 microscope for dimensional determination (at this thickness level, strips can freely stand for cross-section imaging). The wetted strip was then mounted onto the DMA with a 3 mm gap of grip distance, immersed into a jacket beaker filled with 37 o C 1 PBS, equilibrated for 1 hour before test. For the test, only the mechanical preconditioning was performed. The tensile test procedure was similar. But due to the strain failure around 0.3, the strain range for preconditioning was from 0 to 0.2. So, a total of 5 hydrated strips were subjected to 6 loading and unloading cycles with subsequent elongation to failure at a constant rate of 250 um/min. Young s modules, ultimate stretch and strain were obtained from this test. Resilience was calculated from the loading and unloading curves as follows. % resilience = 100*area under unloading curve/area under loading curve (Equation 3.1) Reproducibility was examined using five replicate samples for each type of SELP47K nanofibrous scaffolds, including MeOH-, GTA-, MeOH- & GTA-treated and 60min autoclaved scaffolds Cell Culture Experiment on Nanofibrous Scaffolds Scaffold Preparation for Cell Culture Chemical crosslinked scaffolds were prepared for cell biocompatibility study. After post-fabrication treatment, SELP-47K nanofibrous scaffolds along with a glass slide

85 84 were vacuum oven cured at 80 0 C overnight to remove excess GTA, formic acid, methanol and trapped air bubbles, as these might affect the result of the next step cell culture. In order to check whether there are any trapped air bubbles inside, contact angle measurement was performed by dropping several 5 ul of water drops on the surface of the scaffolds and compared with the contact angles of the air bubble-free SELP-47K films (Figure 3.5). If there is a big difference in contact angle between the scaffolds and the films (~30 o ), there must be trapped air bubble within the scaffold. Figure 3.5. A water droplet is dropped on the surface of the SELP-47K scaffold for contact angle measurement After confirmed that there is not trapped air bubble in the scaffolds, the scaffolds were further sterilized by immersion in 70% ethanol overnight, ultrasonically degassed to remove possible air bubbles when immersion in ethanol, incubated in culture medium for 4 hrs to promote cell attachment, peeled off from the glass slide, and punched into circular disks of 8 mm in diameter. The circular scaffold samples were rinsed three times with cell culture grade 1 PBS, transferred to a 48-well cell culture plate, seeded with 10 µl of NIH/3T3 fibroblast suspension at a density of 300 cells/μl, and incubated in an incubator at 37 C with 5% CO2 for 7 days or until cells

86 85 reached 100% confluence. As a control study, cells were also cultured in a tissue culture polystyrene (TCPS) well without SELP-47K nanofibrous scaffolds. Scaffold samples 6 days after cell culture were prepared for SEM analysis. Briefly, samples were soaked in a primary fixation solution (4% formaldehyde, 1% GTA in 0.1 M phosphate buffer), fixed by 1% osmium tetroxide, rinsed with buffer, dehydrated through a series of graded alcohol solutions and then dried with a Polaron critial point drier using liquid carbon dioxide. The specimens were coated with platinum and viewed by a Hitachi-S4800 SEM. Cell Viability Study-Live/Dead Assay Cell viability was examined 5 days after cell culture using the LIVE/DEAD Viability/Cytotoxicity Kit (Invitrogen), according to the manufactuer s instructions. After removing culture medium, samples were rinsed three times with cell culture-grade 1 PBS and then incubated for 30 minutes with 120 µl of staining solution containing 2 µm calcein AM and 4 µm EthD-1. Live and dead cells were imaged under a fluorescent microscope (Leica DMI 4000 B) after removing the staining solution and rinsing cells once with 1 PBS. Cell Proliferation Study Cell proliferation was analyzed by the MTS assay (CellTiter 96 AQueous Assay, Promega) on days 1, 3, 5, and 7. Live cells react with a tetrazolium salt in the MTS reagent producing a soluble formazan dye, which has absorbance at a wavelength of

87 nm. Within the linear region of the absorbance curve, cell number is proportional to the absorbance intensity, measured using a Nanodrop UV-Vis spectrophotometer (Thermo Scientifc). Statistical Analysis Cell attachment, viability and proliferation were analyzed on multiple replicates, and measurements were expressed as mean ± standard deviation (SD). The student s t-test analysis of variance (SigmaPlot) was employed to assess statistical significance of the results. Difference was considered statistically significant when p < Results and Discussions Electrospinning Experiment Study Electrospinning of SELP-47K in Formic Acid Many factors may influence the micro-architecture of an electrospun scaffold, such as its fiber diameter, surface morphology and pore size. Adjustments of factors such as voltage, the distance between the spinneret and collector, and feed rates were made in order to establish conditions yielding good quality nanofibers (data not shown). One factor, the influence of SELP solution concentration on fiber micro-architecture was of particular interest. At low concentrations (e.g., 50 mg/ml), evident was the formation of SELP-47K filament of 10 to 30 nm in diameter and submicron droplets (Figure 3.6A). Interestingly, charge accumulation on a micro-droplet drastically

88 87 altered the local deposition of SELP-47K nanofilaments. More concentrated SELP-47K solutions would enhance polymer chain entanglement and thus solution elasticity. Consistent with a study by Yu et al [118], the enhanced elasticity of concentrated SELP-47K solutions (e.g., 150 and 200 mg/ml) suppressed the Rayleigh instability and arrested the break-up of the jet into droplets, and led to the formation of bead-free nanofibers (Figure 3.6B). However, if the viscosity of SELP-47K solution (e.g., 250 mg/ml) is too high, some jets of SELP-47K solutions may not be fully stretched by the electrostatic force, leading to the generation of beads-on-a-string structures (Figure 3.6C). The formation of SELP-47K nanofibers suggests that formic acid was removed during the electrospinning process and the jet of SELP-47K solutions was solidified before reaching the collector. Nevertheless, SELP-47K nanofibers electrospun from a 150 or 200 mg/ml solution have non-uniform diameters (Figure 3.6D). While the majority of fibers were 50 to 200 nm in diameter, some larger fibers of a few hundred nanometers in diameter were also observed (Figure 3.6B, D). A B C D

89 88 Figure 3.6. SEM images of SELP-47K nanofibers electrospun from solutions containing protein polymer at concerntrations of 50 (A), 150 (B), and 250 mg/ml (C). The diameter distribution (D) of SELP-47K nanofibers shown in (B) Nanofibers electrospun from a 150 mg/ml solution were examined at both sides of the SELP-47K scaffolds. At the side closer to the glass slide, two populations of nanofibers were observed: small fibers of 10 to 30 nm in diameter and large fibers of 100 to 300 nm in size (Figure 3.7A). On the side closer to the needle, only fibers of 100 to 300 nm in diameter were observed (Figure 3.7B). Likely, small fibers of 10 to 30 nm in size were first formed and deposited on the collector, leading to the build-up of charges on the collector surface. The charge build-up may alter the local electrostatic field, affecting the subsequent fiber formation and branching. Figure 3.7. SEM images of SELP-47K nanofibers from solutions containing protein polymer at concerntrations of 150 mg/ml: fibers closer to the glass slide (A) and fibers closer to the needle (B).

90 89 SEM analysis of SELP-47K nanofibers collected on silicon dies at a low fiber density revealed fiber branching or bifurcation (Figure 3.8). In particular, thin lateral fibers were branched from the primary fibers, appearing to be more prone to bending instability during the electrospinning process. Yarin et al. proposed that for spinnable polymer solutions, the combination of the electric Maxwell stress and surface tension may induce undulation of the solution jet, the instability of which can lead to the emanation of lateral branches [119]. This mechanism may explain the branching of SELP-47K nanofibers. Moreover, a laterally branched fiber observed in the electrospinning of SELP-47K nanofibers oftern took a sharp turn and gradually wound down, or terminated. Although the underlying mechanism behind this phenomena is not fully understood, fiber branching are likely responsible for the non-uniform diameter distribution of electrospun SELP-47K nanofibers. While large fibers may possess enhanced mechanical properties, small fibers of large surface areas can promote cell adhesion and growth. Therefore, non-uniformity of SELP-47K nanofibers will not compromise their applications in tissue engineering. Recently, hybrid electrospinning was designed to fabricate micro/nanometer poly(d,l-lactide-co-glycolide) (PLGA) composites in order to achieve improved mechanical and cellular properties [120].

91 90 Figure 3.8. Fiber branching or bifurcation was examined in the electrsopinning of a 150 mg/ml SELP-47K solution in formic acid. Electrospinning of SELP-47K in distilled de-ionized water Solvent system had a direct impact on the scaffold microstructures. Rod-like nanofibers were obtained from the acid solution, while as reported by Ner, Y. et [60] and repeated here, ribbon-like nanofibers were formed from the aqueous solution (Figure 3.9A). The mechanism for the ribbon formation is under debate; however, the well accepted explanation was introduced by S. Koombhongse et al in 2001 [121]. The explanation is that during spinning, a hard skin forms first due to the evaporation of the solvent near the surface, however, the core with liquid inside is soft; as more solvent evaporates, hollow structure will form, and because the hard skin is not easy to shrink, it collapses and form a ribbon (Figure 3.9B). The width of the ribbon is usually larger than the diameter of the fiber if under the same condition. The fast evaporation of the formic acid resulted in a continuous shrinkage in fiber diameter, thus smaller diameter of fibers are formed.

92 91 Figure 3.9. SEM images of SELP-47K nanofiberous scaffolds electrospun from 200 mg/ml SELP in distilled de-ionized water: A- ribbon-like nanofibers; B- the illustration showing the mechanism of flat ribbon formation The surface area of these ribbon-like fibers was smaller than the rod-like fibers provided the same size (width VS diameter). However, under this sub-micron level, porous mesh structures could also be obtained for cell attachment and integration, thus they are still a promising candidate for tissue engineering [60]. For the concentration effect, Ner, Y. et [60] tired concentrations from 6% to 18% w/w, corresponding to 63.8 and mg/ml, and got bead-free nanoribbons, but not tried higher ones. Here we prepared a concentration at a low level first 70 mg/ml and obtained droplets filled scaffolds (data not shown). So higher concentrations from 200 mg/ml up to 450 mg/ml, were attempted. And as a result, ribbon width was in the sub-micrometer to micrometer range: 0.79 ± 0.15 (μm) from 200 mg/ml squeous solution (Figure 3.9A), 0.88 ± 0.32 (μm) from 300 mg/ml solution (Figure 3.10A, B), 0.90 ± 0.56 (μm) from 350 mg/ml solution (Figure 3.10C, D) and 1.54 ± 0.52 (μm) from 400 mg/ml solution (Figure 3.10E, F). Higher concentration like 450 mg/ml was tried, but failed due to the incomplete dissolving.

93 92 A B C D E F Figure SEM images of SELP-47K nanofiberous scaffolds electrospun from aqueous solutions under ambient room temperature and humidity: 300 mg/ml (A, B), 350 mg/ml (C, D) and 400 mg/ml (E, F). Scale bar: 50 μm (A, C, E) and 20 μm (B, D, F). The spin ability of the solution was determined by three direct factors: viscosity, surface tension and conductivity, which were decided by polymer molecular weight, solvent and concentration. Polymers with higher molecular weight usually were prepared with a low concentration to get a low viscous solution for electrospinning. Under the molecular level, the viscosity is decided by the degree of molecular chain interactions. The chains of SELP-47K were fully or near-fully extended in a high

94 93 hydrogen-filled formic acid solution, and the molecular interactions increased to a great degree when the concentration was increased, hence, beaded fibers were formed from the viscous solution. While the chains of SELP-47K were not fully extended when dissolved in water, and might form globular structures with hydrophilic parts outward, and the degree of molecular interactions was lowered. Thus they can be prepared to a high concentration still with low viscosity. This is the reason why SELP-47K can still be electrospun from a concentrated solution up to 400 mg/ml. Other than the processing parameters, the environment is very important for the electrospinning. Relative humidity is one of the key factors in this category. Medeiros, E.S., et al. obtained nanofibers with many bifurcations by altering the humidity [122] and Casper, C.L., et al. got nanofibers with nanoholes on when increasing the humidity [123]. In this study, the resulting microstructures changed greatly when the humidity increased. Rare bifurcations and bigger ribbons were found when the humidity was 18% (Figure 3.11A); while more bifurcations and smaller ribbons were seen when the humidity increased to 24% (Figure 3.11B) or 35% (Figure 3.11C). Spindle-like beads along the fiber and prevailing triangle-shape like bifurcations were present when the humidity was 48% (Figure 3.11D). The phenomenon was explained by S. Tripatanasuwan et. al. [124]: the charged jet had a longer time to elongate when the evaporation rate was low due to the increased humidity, so that smaller ribbons were obtained; and the charge per unit area on the surface of the jet decreased when

95 94 the surface area increased because of the same reason, so beaded fibers and bifurcations appeared. The surface morphology on a single ribbon was not examined here but under further investigation. A B C D Figure SEM images of SELP-47K nanofiberous scaffolds electrospun from 200 mg/ml in distilled de-ionized water under 18% (A), 24% (B), 35% (C) and 48% (D) humidity and 23 o C room temperature. Scale bar: 10 μm. The SEM images showed that the relative humidity had a direct impact on the resulting microstructures when electrospun SELP-47K from aqueous solutions. The triangular-shape like bifurcations along with the smaller ribbon size might result in a structure with both bigger pore sizes and higher surface areas. In summary, SELP-47K proteins were electrospun into fibers/ribbons from both acid solutions and aqueous solution. The resultant mcirostructures had a different degree of dependence on the solvent selection, concentration and relative humidity.

96 Post-fabrication Treatment Effects on the Microstructures In the wet-spun microfiber chapter, we demonstrated that SELP-47K could be chemically crosslinked by either covalent or non-covalent means. In each monomer of SELP-47K (i.e., (E)4(S)4(EK)(E)3), a lysine residue replaces one of the valine residues in the 8 repeats of the elastinlike pentapeptide sequence. Glutaraldehyde can react with lysine residues to covalently crosslink SELP-47K. Likewise, we have shown that methanol can induce the crystallization of the silklike blocks, essentially forming non-covalent crosslinks in SELP-47K. Electrospun fibrous scaffolds of SELP-47K were treated with MeOH, GTA or a combination of MeOH and GTA, resulting in excellent inter-fiber bonding and the formation of an integrated, porous, fibrous scaffold (Figure 3.12). Pore size was 1.40 ± 0.87, 1.37 ± 0.83, 1.05 ± 0.60, and 1.09 ± 0.54 µm for as-spun, MeOH-, GTA-, and MeOH-GTA-treated scaffolds, respectively. While MeOH-treatment alone induced little changes in pore size, GTA- and MeOH-GTA treatment reduced pore size up to 25%.

97 96 Pore size: 1.40 ± 0.87µm Pore size: 1.37 ± 0.83µm Pore size: 1.05 ± 0.60µm Pore size: 1.09 ± 0.54µm Figure SEM Images of as-spun (A), MeOH- (B), GTA- (C), and MeOH-GTA-treated (D) SELP-47K fibrous scaffolds electrospun from a 15% w/v protein solution. Scale bars: 5 µm. Towards the environment friendly approach, autoclave was used here to stabilize SELP-47K scaffolds and the resultant scaffolds were all water-insoluble after this treatment for 30 or 60 min. In the autoclave chamber, pressure, temperature and water steam might all or partially did the work. To know which one is the leading factor, three more methods were tried: water vapor annealing under room temperature over night, oven baking over night at 134 o C, and super-heated steam (about 120 o C water steam) annealing for roughly 1 second. The former two didn t work; while after the steam, the scaffolds were water-insoluble. Thus the steam is the key factor. The pressure and temperature might work as an aid, but it needs further investigation.

98 97 Pore size: 1.28 ± 0.54µm A Pore size: 1.07 ± 0.42µm B Figure SEM images of SELP-47K nanofiberous scaffold electrospun from 150 mg /ml SELP in 98% formic acid under ambient room environment: A-before autoclave, B-after autoclave. Scale bar: 5 μm. In order to follow the standard sterilization procedure in our lab, 60 min of autoclaving was selected. And scaffolds after this time of treatment were stored for mechanical characterizations. Little difference existed between the scaffolds after 30 min and 60 min treatment, in terms of the morphology, fiber diameter and pore size (data not shown). However, as seen in the SEM images, the scaffolds had a dramatic morphology change comparing to the as-spun scaffolds (Figure 3.13). Dense layers and inter-fiber bondings were found in Figure 3.13B. Though the fiber diameters of scaffolds had little change: from 182 ± 105 to 180 ± 67 nm, while the pore sizes decreased from 1.28 ± 0.54 to 1.07 ± 0.42 μm. And a significant thickness decrease, from roughly 200 to 57.0 ± 5.0 μm (based on the five strips for mechanical test), was the result of the high pressure during the treatment Fourier Transform Infrared Spectroscopic Secondary Structure Analysis Both the fabrication process and the postfabrication treatment can alter the proteins conformations to different extensions. To study the solvent effect, FTIR spectra of

99 98 scaffolds electrospun from both aqueous and acid solutions were plotted and compared (Figure 3.14). The two spectra were similar except two peak shifts in Amide A and Amide I region. The peak in Amide A shifted from 3314 to 3302 cm -1, indicating the influence of the acid solvent. But this is a localized bond, telling little information about the protein backbone. In Amide I, broad bands of the two spectra overlapped a lot with a slight peak position shift from 1661 to 1656 cm -1, indicating the dominant presence of the same conformations: β-turns and unordered structures [125] for both, while a bit high of unordered structures in the scaffolds from the aqueous solution. These two conformations contributed to the water-soluble silk I form [126], which was confirmed by the phenomenon that both as-spun scaffolds were dissolved upon water contact. In summary, the solvent had a slight influence on the structural confirmations, and the resultant scaffolds had a lot of conformations in common. Figure FTIR spectra of as-spun nanofiberous scaffolds electrospun from 150 mg/ml SELP in 98% formic acid (a), and 200 mg/ml SELP in distilled de-ionized water (b), under ambient room temperature. Post-fabrication treatment (methanol and/or glutaraldehyde, or autoclave) may

100 99 induce the crystallization of the silklike blocks and the crosslinking of the elastinlike blocks, leading to changes in the secondary structure of SELP-47K fibrous scaffolds. Such secondary structural changes were examined (Figure 3.15). As stated above, non-treated SELP-47K scaffolds had dominance of β-turn and unordered conformations. Upon treatment, the amide I band (Figure 3.15B) of SELP47K scaffolds split into two: one at 1653 cm -1, that is the characteristic of unordered structure and the other at 1628~1635 cm -1 typical of anti-parallel β-sheets. The emerged band at 1628~1635 cm -1 suggested that post-fabrication treatment resulted in a partial transition from Silk I structure dominated by unordered conformations to Silk II structure characterized by β-sheet conformation. This was confirmed by the appearance of a shoulder at 1700 cm -1, and consistent with the shifting of the amide A band to a lower wavenumber (from 3310 cm -1 to 3290~3278 cm -1 ), indicating more involvement of NH bonds in hydrogen bonding. Besides, the intensities of the two peaks revealed the relative amounts of the unordered structures and the anti-parallel β-sheets. For the autoclaved scaffolds, the peak at 1628~1635 cm-1 is higher than the peak at 1653 cm -1, while for the chemical crosslinked ones, there is no significant difference between the two peaks, which implied that autoclaved scaffolds had a higher amount anti-parallel β-sheets than the chemical crosslinked ones. Further, between the two autoclaved scaffolds, the intensity of the peak at 1628 cm -1 increased comparing to the peak at 1653 cm -1, when the autoclave time

101 100 increased from 30 min to 60 min (Figure 3.15B). It suggested that a great amount of β-turns and intermediate unordered structures were changed to anti-parallel β-sheets. These findings were consistent with the steam sterilized Bombyx mori silk film [127]. On the other hand, the double amide I bands and their breadth revealed the coexistence of other conformations. Indeed, Silk I markers at 1410 cm -1 (CαH2 stretching [128]), 1330 cm -1 (CH3 symmetric stretching), and 1387 cm -1 (CH3 stretching [129]), identified by Taddei and Monti [108], were all displayed by the treated SELP-47K nanofibrous scaffolds (Figure 3.15C). The partial Silk I to Silk II conversion is associated with the enhanced stability of treated SELP-47K scaffolds.

102 101 Figure FTIR spectra of SELP-47K nanofiberous scaffolds in the cm -1 range (A) and cm -1 Amide I range (B) and cm -1 range (C), the numbers from 1 to 6, are corresponding to the spectra from bottom to top. Scaffolds were electrospun from 150 mg/ml SELP in 98% formic acid under ambient room temperature and humidity with a gauge 32 needle Mechanical Properties Analysis Mechanical properties of SELP-47K nanofibrous scaffolds were analyzed when samples were fully hydrated in 1x PBS at 37 o C. The width and length of hydrated MeOH-, GTA-, MeOH-GTA-treated and autoclaved samples were 94 ± 2, 102 ± 1, 108 ± 2% and 100% of their dry counterparts, respectively. However, their thicknesses were 143 ± 8, 68 ± 4, 70 ± 4 and 49 ± 13% of their dry counterparts, respectively. Here, stress calculation is based on the dimensions of hydrated samples. Uniaxial tensile analysis of MeOH-treated scaffolds revealed an ultimate tensile strength of 7.2 ± 2.3 MPa, and strain at failure of 130 ± 30%, and Young s modulus of 3.7 ± 1.2 MPa that was measured in the first 20% strain region (Figure 3.16, Table

103 ). GTA-treated SELP-47K scaffolds displayed a Young s modulus of 18.2 ± 3.3 MPa, ultimate tensile strength of 14.1 ± 3.8 MPa, and strain at failure of 130 ± 40%. Likewise, MeOH-GTA-treated scaffolds possessed mechanical properties comparable to scaffolds treated by GTA alone. Tensile test of the autoclaved scaffolds was not included here. Compared to the bulk mechanical properties of SELP-47K films under the same chemical treatments [59], the Young s modulus and tensile strength of SELP-47K nanofibrous scaffolds were 3 to 5 fold greater, while their deformability was comparable. Among others [69], as stated in the previous chapter, wet-spun microfibers of smaller diameter possessed higher Young s modulus and tensile strength. It is believed that an increased molecular orientation in smaller fibers and mechanical properties can be obtained using smaller spinnerets and/or through larger post-fabrication stretching.

104 103 Figure Representative tensile stress-strain curves of MeOH- (1), GTA- (2), and MeOH-GTA-treated (3) SELP-47K nanofibrous scaffolds. Mechanical preconditioning is often used to stabilize the microstructure of soft tissue materials, permitting the measurement of repeatable mechanical properties [130]. The influence of mechanical preconditioning on the material behavior of electrospun SELP47K nanofibrous scaffolds was examined by subjecting hydrated samples to 6 repetitive cyclic strains of about 34% (for chemical crosslinked scaffolds) and or 20% (for 60 min autoclaved ones) (Figure 3.17). Mechanical preconditioning had little influence on the mechanical behavior of chemical crosslinked SELP-47K scaffolds after the preconditioning strains (Table 3.1). By comparing the mechanical data between this SELP-47K (Table 3.1) and other proteins (Table 1.5), chemical crosslinked SELP-47K scaffolds possess ultimate tensile strengths comparable to those of GTA-cross-linked collagen

105 104 scaffolds [36] and far exceed those of fibrinogen scaffolds [82]. And the elasctic modulus ( MPa) are significantly higher than the native aortic elastin (0.8 MPa) [80], although their deformability is comparable to that of native elastin. Additionally, the elastic moduli of chemical crosslinked SELP-47K nanofibrous scaffolds are comparable to those of human coronary arteries and veins [85-86], yet lower than that of human saphenous vein [87] (Figure 3.17). In terms of ultimate tensile strength, they are stronger than human coronary arteries and saphenous vein. Still, their deformability is similar to that of native arteries and silk scaffolds.

106 105 Figure Representative preconditioning behavior of MeOH- (1), GTA- (2), MeOH-GTA-treated (3), and 60 min autoclaved SELP-47K nanofibrous scaffolds: A - curves were vertically shifted for better clarity; B the 60 min autoclaved scaffolds; C- the preconditioning cycles were removed, and comparison with human saphenous vein and aortic elastin (Table 1.5).

107 106 Table 3.1 Mechanical properties of fully hydrated SELP-47K nanofibrous scaffolds before and after mechanical preconditoning a E (MPa) σ T (MPa) ε F r (%) MeOH-treated 3.7±1.2/3.4± ±2.3/5.7± ±30/110± ±4.4/86.9±2.8 GTA-treated 18.2±3.3/10.6± ±3.8/8.0± ±40/70± ±3.6/83.8±6.5 MeOH-GTA-treated 13.8±3.7/13.2± ±5.3/13.5± ±30/110± ±6.7/80.6± min autoclaved 1.26 ± 0.36/1.0 ± /0.3 ± /30 ± ± 9.0/90.7 ± 3.4 Wet-spun (GTA) / / / a Values separated by a slash / are the mechanicalproperties of SELP-47K scaffolds before and after preconditioning, respectively. Abbreviations: E, elastic modulus measured in the strain range of 0 to 0.2; σ T, ultimate tensile strength; ε F, strain at failure; r, resilience.

108 107 The mechanical properties of the autoclaved scaffolds (Figure 3.17) are inferior to the methanol treated ones. As stated in the secondary structure analysis part, the autoclaved scaffolds had higher amount of anti-parallel β-sheets than the methanol treated one in dry state. It could be explained that the disoriented β-sheets crystallites in the autoclaved scaffold are greatly destroyed by water molecule when hydrated, while less amount of oriented β-sheets crystallites in methanol treated scaffold are broken down. Thus, the autoclaved scaffolds exhibited lower strength. However, the elastic modulus 1.0~1.3 MPa (Table 3.1) of the autoclaved scaffolds is stronger than the self-assembled recombinant elastin (0.25 MPa), native aortic elastin (0.81 MPa)[80] and electrospun fibrinogen (0.19~0.55 MPa) [82], and the ultimate stress (0.3 MPa) is in the same range or a little lower. Even though the ultimate strain is lower, effort could be made like autoclaveing process control (temperature, pressure and time adjustment) to improve the mechanical properties. Another parameter resilience is of great interest, as it characterizes the capacity for shape and energy recovery under mechanical loading, which is evaluated using formula (3.1). Without preconditioning, GTA-, MeOH-, MeOH-GTA-treated and autoclaved scaffolds possessed resilience of 45.9 ± 3.6%, 60.4 ± 4.4%, 59.5 ± 6.7%, and 66.6 ± 9.0% respectively (Figure 3.18). The enhanced resilience of MeOH-, MeOH-GTA-treated and autoclaved scaffolds over that of the GTA-treated scaffold suggests that the crystallization of the silklike blocks induced by methanol treatment

109 108 is more effective in limiting the chain rearrangement and thus reducing the hysteresis of electrospun scaffolds than is GTA treatment. Over the six loading cycles, the resilience of GTA-, MeOH-, MeOH-GTA-treated and autoclaved scaffolds increased to 83.8 ± 6.5%, 86.9 ± 2.8%, 80.6 ± 5.4%, and 90.7 ± 3.4% respectively (Figure 3.18, Table 3.1). The greatest increase in resilience largely occurred after the first cycle, presumably due to stabilization of deformation-induced changes in scaffold microstructure. Significantly, the resilience of the SELP-47K scaffolds after preconditioning (80~91%) approached that of native aortic elastin, which is 77 ± 2% [80]. Figure Resilience of SELP-47K nanofibrous scaffolds as a function of the number of preconditioning cycles (curves were horizontally shifted for better clarity).

110 109 To summarize, chemical crosslinked electrospun SELP-47K nanofibrous scaffolds display mechanical properties (e.g., elastic modulus, ultimate tensile strength, deformability) comparable to or exceeding those of native human arteries, collagen fibrils and electrospun collagen and fibrinogen scaffolds, possessing great potential for tissue engineering applications. Autoclaved scaffolds have mechanical properties inferior to the crosslinked ones, while this autoclaved method is effective to stabilize the electrospun SELP-47K fibers. Combined with the chemical crosslinking methods, scaffolds may be further strengthened Nanofibrous Scaffolds Cell Biocompatibility Study Contact Angle Analysis Electrospun nanofibrous scaffolds usually trap air bubbles, thus complete sterilization is not possible, and therefore, removal of these trapped air bubbles is critical. Vacuum oven heating [131], ultrasonic degassing [132] and manually squeezing are usually utilized for the removal of air bubble. Water contact angle measurement is a direct tool to check the effectiveness of these methods. So here, contact angles of both scaffolds and film are measured and compared (Figure 3.19), after the scaffolds were vacuum oven heated over night. The contact angles of the scaffolds fall in the range of o which is in the same range as the films: o. This can demonstrate that there is no trapped air bubble in the scaffolds.

111 110 Figure Water contact angles of methanol treated scaffolds (I), GTA-treated scaffold (II), methanol-gta-treated scaffolds (III), methanol soaked film (IV) and GTA-treated film (V) Scaffolds Cell Biocompatibility An ideal scaffold material should have excellent cytobiocompatibility and promote cell attachment and growth, and other cellular functions. Revealed by a subcutaneous implantation study in rats, SELPs in the form of thin films displayed a protein sequence-dependent degradation/resorption and excellent in vivo biocompatibility [133]. Recently, their use for gene therapy was also explored [134]. In this study, the NIH/3T3 fibroblasts were used as a model system to evaluate the in vitro biocompatibility of SELP-47K nanofibrous scaffolds. Figure Fluorescent staining for cell viability of 3T3 fibroblasts grown on MeOH- (A), GTA- (B), and MeOH-GTA-treated (C) electrospun SELP-47K scaffolds for 5 days. Living cells were in green and dead cells were in red. Scale bars: 50 µm. Cell viability was assessed using the LIVE/DEAD assay. After staining, live cells

112 111 produce an intense green fluorescence while dead cells emit a bright red fluorescence. The cell viability assay revealed that most of the cells seeded on all the three types of SELP-47K nanofibrous scaffolds were live (Figure 3.20). Cell viability was quantified by counting the number of live cells versus the total cells in ten digital microfluorescence images of stained cells on each type of scaffold. The percent viability of 3T3 fibroblasts on MeOH-, GTA-, and MeOH-GTA treated scaffolds were determined to be 99.6 ± 0.4%, 99.2 ± 0.6% and 93.5 ± 2.5%, respectively. The high viabilities of 3T3 fibroblasts suggested that electrospun SELP-47K nanofibrous scaffolds are highly biocompatible. Figure SEM images revealed that 3T3 fibroblasts interacted with MeOH- (A) and GTA-treated (B) SELP-47K nanofibrous scaffolds, and with each other (C) on GTA-treated scaffolds. Images were taken 6 days after the initial cell seeding. Scale bars: 10 µm. SEM analysis revealed more detailed cell-material and cell-cell interactions on

113 112 SELP47K nanofibrous scaffolds. Images of individual 3T3 fibroblasts cultured on MeOH- and GTA-treated scaffolds for 6 days were presented in Figure Cells cultured on MeOH-GTA-treated scaffolds for 6 days completely covered the scaffolds (Figure 3.23), preventing a direct analysis of cell-material interaction. As shown in Figure 3.21A, many filopodia were formed on the surface of the cell, linking it to the SELP-47K scaffold (Figure 3.21A). Interestingly, a cell lamellipodium managed to squeeze into a scaffold, interacting with finer SELP-47K nanofilaments (Figure 3.21B). Cells interacted not only with SELP-47K nanofibrous scaffolds, but also with each other (Figure 3.21C). Figure Proliferation profiles of 3T3 fibroblasts grown on electrospun SELP-47K nanofibrous scaffolds up to 7 days.

114 113 Figure SEM images showed more than one layer of 3T3 fibroblasts grown on MeOH-GTA-treated (A) and ECM nanofilaments regenerated by cells on GTA-treated scaffolds (B). Images were taken 6 days after the initial seeding To evaluate the ability of SELP-47K nanofibrous scaffolds to support cell proliferation, 3T3 fibroblasts were seeded at the same density on all three types of scaffolds and in TCPS wells as a control. Their proliferation profiles were assessed using the MTS assay. During a period of 7 days, 3T3 fibroblasts cultured on SELP-47K scaffolds continuously grew in number, suggesting the ability of cells to proliferate on the scaffolds (Figure 3.22). The proliferation of 3T3 fibroblasts slowed on day 7, as the cells approached 100% confluence on all three types of scaffolds. While MeOH-treated SELP47K scaffolds demonstrated cell proliferation slightly inferior to the culture control on TCPS, GTA- and MeOH-GTA-treated scaffolds displayed comparable or even better proliferation than the control. A close examination of SEM images suggests that more than one layer of cells were grown on MeOH-GTA-treated scaffolds (Figure 3.23A). Cells were grown less confluently on GTA-treated scaffolds than those seeded on MeOHGTA-treated scaffolds, permitting more detailed analyses of cell-cell interactions. A SEM analysis revealed the formation of ECM nanofilaments and filament sheets between cells (Figure 3.23B).