Formation of silk fibroin matrices with different texture and its cellular response to normal human keratinocytes

International Journal of Biological Macromolecules 34 (2004) 223–230

Formation of silk fibroin matrices with different texture and its cellular response to normal human keratinocytes Byung-Moo Mina , Lim Jeongb , Young Sik Namb , Jin-Man Kima , Jin Young Kima , Won Ho Parkb,∗ a

b

Department of Oral Biochemistry and Dental Research Institute, IBEC and BK21 HLS, College of Dentistry, Seoul National University, Seoul 110-749, Republic of Korea Department of Textile Engineering, College of Engineering, Chungnam National University, Daejeon 305-764, Republic of Korea

Abstract Three forms of silk fibroin (SF) matrices, woven (microfiber), non-woven (nanofiber), and film form, were used to perform a conformational analysis and cell culture using normal human oral keratinocytes (NHOK). To obtain the SF microfiber (SF-M) matrix, natural grey silk was degummed, while the SF film (SF-F) and nanofiber (SF-N) matrices were prepared by casting and electrospinning the formic acid solutions of the regenerated SF, respectively. For insolubilization, as-prepared SF-F and SF-N matrices were chemically treated with an aqueous methanol solution of 50%. The conformational structures of as-prepared and chemically treated SF matrices were investigated using attenuated total reflectance infrared spectroscopy (ATR-IR) and solid-state 13 C CP/MAS nuclear magnetic resonance (NMR) spectroscopy. The as-cast SF-F matrix formed a mainly ␤-sheet structure that was similar to the SF-M matrix, whereas the as-spun SF-N matrix had a random coil conformation as the predominant secondary structure. Conformational transitions from random coil to ␤-sheet of the as-spun SF-N occurred rapidly within 10 min following aqueous methanol treatment, and were confirmed by solid-state 13 C NMR analysis. To assess the cytocompatibility and cells behavior on the different textures of SF, we examined the cell attachment and spreading of NHOK that was seeded onto the SF matrices, as well as the interaction between the cells and SF matrices. Our results indicate that the SF nanofiber matrix may be more preferable than SF film and SF microfiber matrices for biomedical applications, such as wound dressings and scaffolds for tissue engineering. © 2004 Elsevier B.V. All rights reserved. Keywords: Silk fibroin; Nanofiber; Film; Microfiber; Matrix; Keratinocyte

1. Introduction Silk fibroin (SF) is the protein that forms the silkworm silk filaments and gives silk its unique physical and chemical properties [1]. Depending on its application, SF can also be found in various forms, such as gels, powders, fibers, or membranes [2–5]. Recently, researchers have investigated the potential of SF as a candidate material with biomedical applications, because of its distinctive biological properties that include good biocompatibility, good oxygen and water vapor permeability, biodegradability, and minimal inflammatory reaction [6–8]. In practice, SF has already been used for various applications, such as cosmetics, food additives, and medical materials. ∗

Corresponding author. Fax: +82 42 823 3736. E-mail address: [email protected] (W.H. Park).

0141-8130/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ijbiomac.2004.08.004

On the other hand, there have been considerable efforts over the last decade to develop scaffolds for tissue engineering that use biodegradable and biocompatible polymers. With respect to using SF scaffolds for cell culture, some researchers [9–13] have investigated the effects of the SF film matrix on the culture of fibroblasts and osteoblasts, and concluded that SF has positive effects on cell adhesion, viability, growth, and differentiated functions that are comparable to collagen. As well, Cai et al. [14,15] conducted surface modifications of poly(d,l-lactic acid) (PDLLA) using SF, to examine its effect on the culture of osteoblasts. They suggested that the SF-modified PDLLA surface can improve the interaction between osteoblasts and PDLLA films. The matrix texture, as well as the nature of the biomaterial, was also reported to control cell adhesion, proliferation, shape, and function [16–18]. The effects of the scaffold’s surface microstructure on modulating the spatial organiza-

224

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

tion and functions of cells have been investigated, specifically for endothelial cells [16,17,19], nerve cells [16], hepatocytes [20,21], and skin fibroblasts [18]. However, very little is known about the textural effects of the fibrous matrix on tissue engineering, although the three-dimensional structure has profound effects on cell morphology, proliferation, migration, differentiation, and function [22]. The highly specific surface area and highly porous three-dimensional structure of woven or non-woven fabrics places them among the most promising material forms used in tissue engineering applications and are quite desirable for high-density cell and tissue cultures. Fibrous materials offer a potentially wide range of suprastructures, created by changing the fiber diameter, orientation, porosity, and fabric (woven, knitting and nonwoven) characteristics. Until now, many synthetic polymers have been fabricated into fibrous forms for tissue engineering applications, and poly(glycolic acid) and its copolymers have also been extensively studied in a non-woven matrix [23,24]. Previously [25,26], we reported that the non-woven matrix of the electrospun SF nanofibers proved to be very effective in the cell attachment and spreading of normal human keratinocytes and fibroblasts. However, it remains to be confirmed whether the usefulness of the SF nanofiber matrix originates from the nature of SF, or the non-woven texture of the SF matrix. To our knowledge, the effects of the SF matrix texture for the purposes of tissue engineering have not been studied, although SF has been processed into foams, films and fibers, for other biomedical material applications that include wound dressings, artificial blood vessel, and artificial skin. In the present study, we prepared three types of SF matrices, including a matrix woven from SF microfibers, a nonwoven matrix of SF nanofibers, and a matrix of SF film. Then, we applied the chemical treatment to crystallize the SF film and non-woven matrix. Solid-state 13 C CP/MAS nuclear magnetic resonance (NMR) and attenuated total reflectance infrared spectroscopy (ATR-IR) were used to investigate the conformational structure of the SF matrices. In addition, we examined the effect of the three types of SF matrix on the cell adhesion and spreading of normal human epithelial cells. To further evaluate the effects of extracellular matrix (ECM) proteins in relation to the differing types of SF matrices, we investigated the cellular responses on SF matrices, using type I collagen, fibronectin, or laminin, as substrates, which were adsorbed onto the SF matrices.

2. Material and methods 2.1. Preparation of SF microfiber matrix Silk fibroin (SF) microfiber matrix (sample code, SF-M) was prepared by degumming grey silk twice with 0.5% (w/w) NaHCO3 solution at 100 ◦ C for 30 min and then rinsed with warm distilled water.

2.2. Preparation of SF film and nanofiber matrices Degummed SF was dissolved in a ternary solvent system of CaCl2 –CH3 CH2 OH–H2 O (1:2:8 in mole ratio) at 70 ◦ C for 6 h. After dialysis with cellulose tubular membrane (2507u; Sigma) in distilled water for 3 days, the SF solution was filtered and lyophilized to obtain the regenerated SF sponges. The SF solutions were prepared by dissolving the regenerated SF sponges in 98% formic acid (Aldrich) for 3 h. The SF film matrix (sample code, SF-F) was prepared by casting 3% concentration of SF in formic acid onto a Teflon plate. The SF nanofiber matrix (sample code, SF-N) was prepared by electrospinning at the concentration of 8 wt.%. The electrospun SF nanofibers were collected on a target drum which was placed at a distance of ∼7 cm from the syringe tip. A voltage of 15 kV was applied to the collecting target by a high voltage power supply. The flow rate of the polymer solution was 1.5 ml h−1 . 2.3. Chemical treatment of SF film and nanofiber matrices The as-prepared SF film and nanofiber matrices were crystallized and insolubilized by immersion in a 50% (v/v) aqueous methanol solution for 10–60 min at room temperature, and drying under vacuum for 24 h. 2.4. Characterization A scanning electron microscope (SEM, Hitachi S-2350) was used to investigate the macroscopic morphology/surface texture of the SF matrices. The ATR-IR spectra of the SF matrices were obtained using a Travel IR in the spectral range of 4000–400 cm−1 . WAXD experiments were performed on a Rigaku fine-focus fixed tube generator with Ni-filtered Cu K␣ radiation (λ = 0.154 nm), and a flat-film camera with pinhole collimation. Irradiation conditions were 40 kV and 30 mA. The 13 C solid-state CP/MAS NMR spectra of SF matrices were obtained on a Bruker DSX 400 MHz solid NMR spectrometer using a cross-polarization pulse sequence and using magic-angle spinning at 5 or 6.5 kHz. 2.5. Cells and cell culture Primary normal human oral keratinocytes (NHOK) were prepared from human gingival tissue specimens and maintained as described previously [27]. The tissue samples were obtained from three healthy individuals with ages in the range of 21–30 years, while they were undergoing surgery. Briefly, samples were thoroughly washed with calcium- and magnesium-free Hanks’ balanced salt solution (CMF-HBSS; GibcoBRL). To separate the epithelium from the underlying mucosa, the tissues were incubated in CMF-HBSS containing collagenase (type II, 1.0 mg/ml; Sigma) and dispase (grade II, 2.4 mg/ml; Boehringer-Mannheim) for 90 min at

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

37 ◦ C in 95% air and 5% CO2 . Epithelial cells were isolated from the epithelial tissue by trypsin digestion. These cells were seeded onto 60-mm Petri dishes and allowed to proliferate until reaching 60–70% confluence. Primary NHOK were serially subcultured, being passaged at every 70% confluence level; second passage keratinocytes were used in the described experiments. The cells were cultured in keratinocyte basal medium containing 0.15 mM calcium and a supplementary growth factor bullet kit (KGM; Clonetics). 2.6. Cell adhesion assay and cell spreading analysis Cell adhesion was assayed using a modification of the method of Mould et al.’s method [28]. Briefly, SF matrices (SF-M, methanol-treated SF-F, and methanol-treated SF-N) were cut out using a punch (14-mm in diameter) and placed onto the 24-well culture plates (Corning). These 24-well culture plates containing the SF matrices were coated with 300 ␮l/well of ECM proteins, in this case type I collagen (50 ␮g/ml), fibronectin (1 ␮g/ml), or laminin (10 ␮g/ml) in phosphate buffered saline (PBS) by overnight adsorption at room temperature. The wells were then washed with PBS, and the unbound sites were blocked with 10 mg/ml of bovine serum albumin (BSA) in PBS. Cells were detached by trypsin digestion, then 300 ␮l of a cell suspension containing 1 × 105 cells was placed in each well, and the cells were allowed to settle/adhere for 1 h at 37 ◦ C in an atmosphere of 5% CO2 . Loosely adherent or unbound cells from experimental wells were removed by aspiration, the wells were washed once with PBS, and the remaining bound cells were fixed with 10% formalin in PBS for 15 min. Once the fixative was aspirated, the wells were washed twice with PBS, and the cells attached to the SF matrices were stained with hematoxylin and eosin. The wells were gently washed three times with dH2 O. The SF matrices were mounted, and the cells attached to the matrices were photographed. To ensure a representative count, each SF matrix was divided into quarters and two fields per each quarter were photographed with an Olympus BX51 microscope at 100×. Average percentages and standard deviations were calculated from three independent experiments. Cell spreading was analyzed with photographs that were taken in the cell adhesion assay. To ensure a representative count, each SF matrix was divided into quarters and two fields per each quarter were photographed with an Olympus BX51 microscope at 100×. The cells that adopted a flattened, polygonal shape, with filopodia- and lamellipodia-like extensions were regarded as spreading cells. In contrast, the cells that resisted washing and remained tethered to the plate surface were regarded as non-spreading cells. The percentage of cells displaying the spread morphology was quantified by dividing the number of spread cells by the total number of bound cells. Average percentages and standard deviations were calculated from three independent experiments.

225

2.7. Statistical analysis Cell adhesion and spreading results were analyzed using the Student’s t test. Statistical significance was defined as P < 0.01.

3. Results and discussion 3.1. Morphology of SF-M and SF-N matrices Electrospinning generally produces non-woven matrices with randomly arranged, ultrathin fibers that have nanometerscale diameters. Fig. 1 shows a SEM micrograph (magnification 500×) of the woven SF microfiber and as-spun SF nanofiber matrix. From the image analysis, the SF nanofibers have an average diameter of 80 nm and their diameters ranged from 30 to 120 nm, while the diameter of the SF microfiber was 11 ␮m. The average diameter of SF nanofibers corresponds to about two orders of magnitude smaller than a SF microfiber. 3.2. Secondary structure of SF matrices The physiological properties of SF matrices strongly depend on its molecular conformation and surface texture. SF

Fig. 1. SEM image of woven SF microfibers and nonwoven SF nanofibers.

226

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

exhibits at least three crystalline forms: silk I, silk II, and alpha-helix. All three conformations can be formed from the appropriate preparation conditions and each is interchangeable under certain conditions [1]. Crystallization of SF involved the conformational transition that can be easily induced by simple physical (thermal), mechanical or chemical treatments. The most common method to convert the distorted conformation (random coil or silk I) of SF into the more stable ␤-sheet (silk II) conformation is to treat the SF film with an organic solvent. It is well known that organic solvents, particularly methanol, are highly effective in the crystallizing SF from a distorted conformation to a ␤-sheet. Therefore, we investigated the influence of the methanol treatment on the secondary structure of as-prepared SF-F and SF-N matrices, by means of ATR-IR, WAXD, and solid-state CP/MAS 13 C NMR. Infrared spectroscopy (IR) has been often applied to study the molecular conformation of silk fibroin fibers or films. The sensitive absorption bands on the IR spectrum are located in the spectral regions of ∼1625 cm−1 (amide I), ∼1528 cm−1 (amide II), ∼1230 cm−1 (amide III), and ∼700 cm−1 (amide V). To characterize the structure of SF matrices, we examined the ATR-IR spectra in the amide I and II regions. The IR spectra of SF matrices are shown in Fig. 2(a)–(c). The as-spun SF-N matrix was characterized by absorption bands at 1651 cm−1 (amide I) and 1528 cm−1 (amide II), attributed to the random coil conformation, as shown in Fig. 2a. The methanol-treated SF-N matrix showed strong ␤-sheet absorptions at 1622 and 1514 cm−1 within a methanol-treating time of 10 min, indicating that the random coil conformation of the SF nanofibers rapidly converted into a ␤-sheet structure (Fig. 2b). On the contrary, both the as-prepared SF-F and SF-M matrices showed strong ␤sheet absorptions at 1622 and 1514 cm−1 without methanol treatment, as shown in Fig. 2c and d. As previously reported, the raw SF had highly ordered intermolecular hydrogen bonds (␤-sheet), and the SF film took a mainly ␤-

Fig. 2. ATR-IR spectra of the as-prepared and methanol-treated SF matrices: (a) as-spun SF-N; (b) methanol-treated SF-N; (c) as-cast SF-F; (d) SF-M.

sheet conformation, when cast from the formic acid solution [29]. Solid-state 13 C NMR has been shown to be a more effective analytical tool for demonstrating the formation of ␤sheets in polypeptides and proteins, because the isotropic 13 C NMR chemical shifts of carbon atoms in proteins are sensitive to the ␤-sheet’s secondary structure. It is well established that SF conformations are dependent upon the species of silkworms and conditions of the sample preparation [30]. In particular, it has been reported that fibroin from Bombyx mori adopts two dimorphic structures, silk I and silk II [31]. The silk II form is identified by the 13 C chemical shifts of glycine (Gly), serine (Ser), and alanine (Ala) that are indicative of ␤-sheets, while the silk I form produces chemical shifts that are associated with a loose helix or distorted ␤-turn [31]. However, when compared with silk II, the less stable silk I shows a relatively unresolved structure, and the conformation of the soluble form of SF rapidly undergoes a transition to the insoluble silk II conformation [1]. In 13 C CP/MAS NMR structural analyses of B. mori silk fibroins, the two crystalline forms of silk fibroins, silk I and silk II (␤-sheet), have been distinguished by the conformation-dependent 13 C chemical shifts of the respective amino acid residues [32]. Fig. 3 shows

Fig. 3. 13 C CP/MAS NMR spectra (a) and the expanded 13 C NMR spectra of the methyl alanine region (b) of SF matrices. The dotted arrow indicates the chemical shift of alanine in the ␤-sheet conformation.

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

227

solid-state 13 C NMR spectra of the as-prepared SF-N and SFF matrices, together with that of the SF-M. Assignments were made according to the literature [32]. The chemical shifts of the Ala residue in Fig. 3a were 17.6 ppm for Ala C␤ , 49.9 ppm for Ala C␣ , and 173.5 ppm for Ala C O carbon. The observed 13 C NMR chemical shifts of the peaks suggest that the Ala residue of the SF-N matrix takes a mainly non ␤-sheet conformation (random coil and silk I). The vertical dashed line and arrow in Fig. 3 show the chemical shift of Ala in the ␤-sheet conformation. A shoulder was detected at ∼20.4 ppm and assigned to Ala C␤ in a ␤-sheet conformation, indicating that as-spun SF-N take some ␤-sheet conformations. The ␤-sheet structure of as-spun SF nanofibers could be formed partially by the elongational forces during the electrospinning process [25]. However, the characteristic resonances of SF-F matrix, especially the Ala C␤ at 20.4 ppm and Ala C␣ at 48.9 ppm in Fig. 3b, are similar with that of the SF-M matrix (Fig. 3c). Alanine (Ala) is a major constituent of SF [30], and among these 13 C peaks, the chemical shifts of Ala are sensitive to both the conformation and intermolecular arrangement. The methyl groups of Ala residues (Ala C␤ ) have been especially well established for Ala. Fig. 4 shows the expanded views of the Ala methyl regions for SF-N and SF-F matrices before and after methanol treatments, respectively. The peak centered at

∼20.4 ppm was assigned to Ala C␤ in a ␤-sheet conformation, while the one at 17.6 ppm is attributed to side chains in non ␤-sheet structures [32]. As shown in Fig. 4b, the asspun SF-N has a strong signal from the region near 17.6 ppm with a shoulder at 20.4 ppm, while the next two spectra have higher signal at 20.4 ppm. These data clearly indicate that the Ala residues in the SF molecular chain took ␤-sheet structure following methanol treatment. Therefore, the sequence of the expanded spectra region shows that SF molecules undergo significant conformational transitions during aqueous methanol treatment. On the other hand, the as-cast SF-F has a strong signal from the region near 20.4 ppm with shoulder at 17.6 ppm, irrespective of 10 or 60 min methanol treating time. From their relative peak areas in the Ala C␤ region, the relative portions of the ␤-sheet structure were determined by curve fitting, assuming the two Gaussians. The ratio of the area of the peak at 20.4 ppm to the total area of the peaks at 20.4 and 17.6 ppm correlated to the content of Ala residues in the SF that reside in the ␤-sheet conformations. Fig. 5 displays the relationship between the ␤-sheet content of Ala residues and the methanol treating time for SF-N and SFF. From Fig. 5b, it can be concluded that the conformational transition to ␤-sheet form in the SF-N matrix occurred rapidly

Fig. 4. Variation in the expanded 13 C NMR spectra of the methyl alanine region of SF-N (a) and SF-F (b) matrices with methanol-treating time.

Fig. 5. Relation between the ␤-sheet content of alanine residues and methanol-treating time for SF-N and SF-F matrices.

228

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

Table 1 Structural characteristics of SF matrices used for cell test Sample code

␤-Sheet content (%)

Matrix type

Average fiber diameter

SF-M

87.8

11 ␮m

SF-F SF-N

76.5 74.2

Woven (microfiber) Film Non-woven (nanofiber)

– 80 nm

within 10 min, whereas Fig. 5a shows that the ␤-sheet content in SF-F matrix increased slightly from 70 to 77% over 60 min. 3.3. Attachment and spreading of NHOK on three types of SF matrix, alone or in conjunction with ECM protein coating Studying the cytocompatibility of various types of SF matrices, alone or in conjunction with ECM protein coating, is especially interesting in relation to its possible use as wound dressing materials and scaffolds for tissue engineering. The structural characteristics of the SF matrices used for the cell test are summarized in Table 1. SF matrices, alone or in conjunction with ECM protein coating, were seeded with NHOK, to assess cell adhesion and spreading. We evaluated the adhesion of cultured keratinocytes using the cell adhesion assay in serum-free medium, with type I collagen (50 ␮g/ml), fibronectin (1 ␮g/ml), or laminin (10 ␮g/ml) as substrates which were adsorbed onto the SF matrices. We tested the effect of 0.1–50 ␮g/ml of type I collagen, fibronectin, and laminin on NHOK adhesion and spreading to the culture plate surface and found that 50 ␮g/ml type I collagen, 1 ␮g/ml fibronectin, or 10 ␮g/ml laminin showed the most comprehensive effects on human oral keratinocytes (data not shown).

Rapidly proliferating NHOK that adhered to SF matrices were microphotographed during the adhesion assay, after washing, fixing, and staining these cells with hematoxylin and eosin. The NHOK had significantly higher adhesion to the methanol-treated SF-F and SF-N matrices than to the SFM matrix. Particularly, the SF-M showed extremely low adhesion activity under our conditions, when compared to the polystyrene alone (Figs. 6 and 7a). Type I collagen promoted the adhesion of proliferating NHOK on all of the SF matrices tested, and the cell adhesion activity of both SF-F and SF-N was significantly higher than that of SF-M (Fig. 7a). Laminin- and fibronectin-coatings onto the SF-F and SF-N using cultured oral keratinocytes showed generally similar adhesion activity profiles, but these were notably lower than that of type I collagen. In contrast, laminin inhibited, rather than promoted, the adhesion of oral keratinocytes on the SFM when compared to the polystyrene alone, SF-M matrix alone, or BSA control (Fig. 7a). These results support our view that SF-F and SF-N have better matrix textures than SF-M in terms of cell adhesion. As well, type I collagen, one of the integrin ligands, is functionally active in promoting cell adhesion of normal human epithelial cells onto the SF-F and SF-N. We repeated this experiment three times and no gross difference was observed. To further evaluate the adhesion of type I collagen, fibronectin, and laminin, we determined whether adherent cells were tethered to the substrate or spreading over the substrate. Extremely low cell spreading for cultured oral keratinocytes was observed on either BSA- or ECM protein-coated SF-M (Fig. 7b). On type I collagen, 52 and 61% of the proliferating NHOK showed a spreading morphology on the SF-F and SF-N, respectively, i.e., they adopted a flattened, polygonal shape, with filopodia- and lamellipodia-like extensions (Figs. 6 and 7b). The remaining non-spreading cells on in-

Fig. 6. Microphotographs of the cell adhesion and spreading of NHOK to SF matrices, alone or in conjunction with ECM proteins: PS, polystyrene surface only; matrix only, SF-M, SF-F, or SF-N matrix only; BSA, bovine serum albumin-coated SF matrices; Col I, type I collagen-coated SF matrices; FN, fibronectin-coated SF matrices; LN, laminin-coated SF matrices.

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

229

goal of scaffold design is the producing an ideal structure that can replace natural ECM proteins, until the host cells can repopulate and re-synthesize a new natural matrix, SF-N matrix may be a better candidate for tissue engineering scaffolds. The SF nanofibers provide a higher level of surface area for cells to attach to due to its three-dimensional features (high porosity and high surface area-to-volume ratio) and less brittle structure than the SF film matrix.

4. Conclusions

Fig. 7. The level of cell adhesion (a) and the incidence of cultured cells spreading (b) of NHOK to SF matrices, alone or in conjunction with ECM proteins: PS, polystyrene surface only; matrix only, SF-M, SF-F, or SF-N matrix only; BSA, bovine serum albumin-coated SF matrices; Col I, type I collagen-coated SF matrices; FN, fibronectin-coated SF matrices; LN, laminin-coated SF matrices. *Significantly different from the matrix only within the groups at P < 0.01. **Significantly different from the SF-M counterparts at P < 0.01.

tegrin ligands resisted washing and remained tethered to the SF-F and SF-N surfaces. Laminin and fibronectin had no effect on cell spreading onto the SF-F in normal human keratinocytes, however, it functioned as a minor spreading stimulus on the SF-N matrix. These results indicate that type I collagen functions as spreading stimulus on the SF-F and SF-N matrices in the cells tested. As we described above, the cell attachment and spreading onto the SF-F and SF-N matrices, using as type I collagen, as a substrate, showed promising results in the human mucosal keratinocytes tested. Although the reason for the increased initial cell attachment and spreading of normal human epithelial cells onto the SF-F and SF-N matrices remains unknown from this study, it may be associated with the ␤-sheet contents of the SF matrices or the fiber diameters. With respect to the fiber diameter of the matrix, a surface texture having a submicron range may benefit from mimicking the topography of aligned fibrillar ECM [33]. In addition, the electrospun SF-N matrix may provide a three-dimensional structure for cell attachment and growth. This idea is supported by another study, which demonstrated that the electrospun poly(␧caprolactone) nanofiber, types of the biodegradable scaffolds are capable of supporting cell attachment and proliferation of human bone marrow-derived mesenchymal stem cells, and maintain their phenotypic shape, and guide growth according to the nanofiber orientation [34]. In addition, surface modifications of biomaterials with ECM proteins or peptide have been extensively tested. From the point of view that the final

We prepared three types of SF matrices, including a woven matrix from SF microfibers, a matrix of SF film, and a nonwoven matrix of SF nanofibers, then used chemical treatment to crystallize the SF film and non-woven matrix. The influence of the methanol treatment on the secondary structure of as-prepared SF film and nanofiber matrices was investigated using ATR-IR and solid-state CP/MAS 13 C NMR. In comparison with the regenerated SF film matrix, the conformational change from the as-spun SF nanofibers matrix into ␤-sheet occurred and was almost completed within 10 min in the case of the 50% aqueous methanol solution used. In the cell activity assessment, the SF nanofiber matrix was found to promote cell adhesion and spreading of type I collagen better than SF film matrix because the SF nanofiber matrix provides a higher level of surface area for cells to attach to due to its high porosity and surface area-to-volume ratio.

Acknowledgment This work was supported by the Korea Science and Engineering Foundation (KOSEF) through the Intellectual Biointerface Engineering Center at Seoul National University.

References [1] Kaplan D, Adams WW, Farmer B, Viney C, editors. Silk polymers: materials science and biotechnology. Washington DC: ACS Symposium Series 544, p. 1994. [2] Li M, Lu S, Wu Z, Tan K, Minoura N, Kuga S. Int J Biol Macromol 2002;30:89. [3] Takeshita H, Ishida K, Kamiishi Y, Yoshii F, Kume T. Macromol Mater Eng 2000;283:126. [4] Yao J, Masuda H, Zhao C, Asakura T. Macromolecules 2002;35:6. [5] Putthanarat S, Zarkoob S, Magoshi J, Chen JA, Eby RK, Stone M, et al. Polymer 2002;43:3405. [6] Sakabe H, Ito H, Miyamoto T, Noishiki Y, Ha WS. Sen-i Gakkaishi 1989;45:487. [7] Park WH, Ha WS, Ito H, Miyamoto T, Inagaki H, Noishiki Y. Fibers Polym 2001;2:58. [8] Santin M, Motta A, Freddi G, Cannas M. J Biomed Mater Res 1999;46:382. [9] Altman GH, Horan RL, Lu H, Moreau J, Martin I, Richmond JC, et al. Biomaterials 2002;23:4131.

230

B.-M. Min et al. / International Journal of Biological Macromolecules 34 (2004) 223–230

[10] Yohko G, Masuhiro T, Norihiko M. J Biomed Mater Res 1998;39:351. [11] Yohko G, Masuhiro T, Norihiko M, Yohji I. Biomaterials 1997;18:267. [12] Inouye K, Kurokawa M, Nishikawa S, Tsukada M. J Biochem Biophys Methods 1998;37:159. [13] Minoura N, Aiba S, Gotoh Y, Tsukada M, Imai Y. J Biomed Mater Res 1995;29:1215. [14] Cai K, Yao K, Cui Y, Yang Z, Li X, Xie H, et al. Biomaterials 2002;23:1603. [15] Cai K, Yao K, Lin S, Yang Z, Li X, Xie H, et al. Biomaterials 2002;23:1153. [16] Patel N, Padera R, Sanders GHW, Cannizzaro SM, Davies MC, Langer R, et al. FASEB J 1998;12:1447. [17] Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE. Science 1997;276:1425. [18] van Kooten TG, Whitesides JF, von Recum AF. J Biomed Mater Res 1998;43:1. [19] Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE. Biotechnol Prog 1998;14:356. [20] Singvi R, Kumar A, Lopez GP, Stephanopoulos GN, Wang DIC, Whitesides GM, et al. Science 1994;264:696. [21] Folch A, Toner M. Biotechnol Prog 1998;14:388.

[22] Edgington SM. Bio/Technology 1992;10:855. [23] Mikos AG, Bao Y, Cima LG, Ingber DE, Vacanti JP, Langer R. J Biomed Mater Res 1993;27:183. [24] Kim BS, Moony DJ. J Biomed Mater Res 1998;41:322. [25] Kim SH, Nam YS, Lee TS, Park WH. Polym J 2003;35:185. [26] Min BM, Lee G, Kim SH, Nam YS, Lee TS, Park WH. Biomaterials 2004;25:1289. [27] Min BM, Woo KM, Beak JH, Lee G, Park NH. Int J Oncol 1995;7:249. [28] Mould AP, Askari JA, Humphries MJ. J Biol Chem 2000;275: 20324. [29] Um IC, Kweon HY, Park YH, Hudson S. Int J Biol Macromol 2001;29:91. [30] Saito H, Tabeta R, Asakura T, Iwanaga Y, Shoji A, Ozaki T, et al. Macromolecules 1984;17:1405. [31] Asakura T, Yao J, Yamane T, Umemura K, Ulrich AS. J Am Chem Soc 2002;124:8794. [32] Asakura T, Demura M, Date T, Miyashita N, Ogawa K, Williamson MP. Biopolymers 1997;41:193. [33] Clark P, Connolly P, Curtis AS, Dow JA, Wilkinson CD. J Coll Sci 1991;99:73. [34] Yoshimoto H, Shin YM, Terai H, Vacanti JP. Biomaterials 2003;24:2077.