- Email: [email protected]
Effect of silk fibroin molecular weight on physical property of silk hydrogel
Polymer 90 (2016) 26e33
Contents lists available at ScienceDirect
Polymer journal homepage: www.elsevier.com/locate/polymer
Effect of silk fibroin molecular weight on physical property of silk hydrogel H.H. Kim a, D.W. Song a, M.J. Kim a, S.J. Ryu a, I.C. Um b, C.S. Ki a, **, Y.H. Park a, * a b
Department of Biosystems and Biomaterials Science and Engineering, Seoul National University, Seoul 151-921, Republic of Korea Department of Bio-fibers and Materials Science, Kyungpook National University, Daegu 702-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 December 2015 Received in revised form 19 February 2016 Accepted 23 February 2016 Available online 26 February 2016
Silk hydrogel has recently received great attention for its excellent biocompatibility. The property of silk hydrogel is, however, not only hardly controlled but very limited due to non-variability of silk fibroin (SF) molecule. In this study, alkaline hydrolysis was utilized to manipulate the silk hydrogel properties. By regulating the hydrolysis time (10e180 min), a broad molecular weight range of SF was obtained (263.1 e82.7 kDa) Gel point increased with a decrease of SF molecular weight. The change of molecular weight of SF also greatly affected the physical properties (i.e., swelling ratio, shear modulus, transparency) as well as cell adhesion of SF hydrogels. As a result of structural analysis, the molecular weight of SF played a crucial role in the construction of microscopic structure of SF hydrogel. These findings indicate that SF hydrogels of variable physical properties can be fabricated based on molecular weight control for diverse purposes in biomedical engineering. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Silk hydrogel Alkaline hydrolysis Molecular weight Transparent hydrogel Microscopic structure
1. Introduction Silk fibroin (SF) is a major protein, which plays a critical role in both structural feature and mechanical property of silk cocoons. In raw silk fibers, two SF strands are covered with a gum-like protein, silk sericin (SS), which is easily removed by weak alkali hot water [1]. To date, it has been widely reported that SF shows excellent biocompatibility as well as superior mechanical property in biomedical applications [2e7]. Besides, the facile recrystallization by alcohol treatment inducing the physical cross-linking via intraand inter-molecular hydrogen bonding of hydrophobic segments of SF is generally exploited in the fabrication process [8,9]. Such a process is very useful in biomedical material fabrications (e.g., tissue engineering scaffold, drug carrier) since this process does not cause any harmful chemical species damaging living organisms. SF can be fabricated into various forms (e.g., film, nanofiber, sponge, hydrogel) [10]. Especially, SF hydrogel has been recently received great attention in tissue engineering field. The network structure of SF hydrogel retains the large amount of water with inter-connected porous structure. This structure provides the
* Corresponding author. ** Co-corresponding author. E-mail addresses: [email protected] (C.S. Ki), [email protected] (Y.H. Park). http://dx.doi.org/10.1016/j.polymer.2016.02.054 0032-3861/© 2016 Elsevier Ltd. All rights reserved.
physiologically stable condition for cell survival [11,12]. Also, porous three-dimensional (3D) structure of SF hydrogel is an appropriate microenvironment in which cells reside allowing exchanging nutrients and wastes. Therefore, SF hydrogel shows very excellent biological properties compared to synthetic polymer-based hydrogel (e.g., poly(ethylene glycol) [13], poly(vinyl alcohol) [14]) as a 3D cell niche. The properties of SF hydrogel are affected by a wide variety of processing parameters, such as concentration of SF solution [15,16], incubation temperature [15], vortexing time [7], and ultra-sonic power [12], The concentration is a major factor to determine physical properties of SF hydrogel [15,16]. Generally, the higher concentration results in the stiffer hydrogels with the shorter gelation time. The gelation behavior is also largely dependent on the incubation temperature [15]. For example, the SF hydrogel formed at a higher temperature (~50 C) mostly has a higher Young's modulus than that formed at a lower temperature (~4 C). The effect of amplitude of shear force was also investigated. Wang et al. has shown that the amplitude of ultra-sonication affects the mechanical properties of SF hydrogel [12]. Yucel et al. reported the vortexing time effect on the shear modulus of SF hydrogel [7]. Although the significant few studies to manipulate SF hydrogel properties were reported, none of processing variables could elicit the change of a wide range in mechanical as well as physical properties except the concentration of the precursor SF solution. Generally, the molecular weight of polymer composing the
H.H. Kim et al. / Polymer 90 (2016) 26e33
network structure of hydrogel is a crucial factor that determines the gel properties (i.e., modulus, swelling ratio, permeability) [17e19]. Both cross-linking density and mesh size of hydrogel are largely dependent on the polymer chain length regardless of cross-linking methods (i.e., chemical and physical cross-linking) [20,21]. Nevertheless, the effect of molecular weight of SF on hydrogel fabrication has not been explored. Hence, in this study, we conducted the heatalkaline treatment (HAT) during the SF dissolution step to control the molecular weight of SF as a one-pot process. This method is not only simple but also suitable to obtain a high yield of hydrolyzed SF. In contrast, proteinase treatment needs relatively delicate process including enzyme removing process and causes sever mass loss of product in spite of high efficiency [22]. Then, we formed SF hydrogel by ultra-sonication, followed by physical and mechanical property analyses to investigate the effect of molecular weight of SF on the hydrogel formation. Especially, we tried to focus on the change of the microscopic structure of SF hydrogel according to molecular weight variation, which allows the wide range property control of the hydrogel. Finally, it was demonstrated that such a property manipulation of SF hydrogel can be successfully utilized in tissue engineering field by human mesenchymal stem cell (hMSC) culture on SF hydrogels.
27
2.2. SF hydrogel fabrication To initiate the gelation of SF solution, ultra-sonication was performed on 3% (w/v) SF aqueous solution of different molecular weights using an ultra-sonic processor (VCX-130, SONICS, USA) at 25% amplitude for 3 min. The treatment was conducted in an ice chamber to prevent the heat elevation during ultra-sonication. Then, ultra-sonicated SF solution was filtered and incubated at 60 C for 2 days. 2.3. Gel filtration chromatography The molecular weight of SF was measured by gel filtration chromatography (GFC) (AKTA Purifier, GE Healthcare, USA) with Superdex 200 10/300 GL column (GE Healthcare, USA). 0.5 mL of 3% (w/v) SF solution was added to 4 mL of 6 M urea aqueous solution, followed by filtration with 0.2 mm membrane. For the measurement, 300 mL of sample solution was injected and 1.5 column volume of 4 M urea was eluted at a constant flow rate of 0.5 mL/min. The elution of SF was detected at 280 nm. The molecular weight of SF was determined by a calibration curve, which was obtained by a standard globular protein kit (Gel Filtration Cal Kit High Molecular Weight, GE Healthcare, USA).
2. Experimental 2.4. Visible light transmittance measurement
2.1. Materials To remove silk sericin, Bombyx mori cocoons were boiled in 0.3% (w/v) sodium oleate and 0.2% (w/v) sodium carbonate cocktail solution at 100 C for 1 h. The SF aqueous solutions were obtained by using two different dissolving methods. For LiBr-dissolution method, the degummed cocoons were dissolved in 9.3 M LiBr solution at 80 C for 30 min. To hydrolyze SF, 0.6 M sodium hydroxide aqueous solution was directly added to the SF solution at a volume ratio of 1-to-5. Then, the final concentration of sodium hydroxide became 0.1 M in the SF solution. To control SF molecular weight, the hydrolysis time was varied from 10 to 180 min, followed by subsequent dialysis against de-ionized water using cellulose acetate dialysis tube (MWCO: 12,000e14,000 Da) for 3 days. For CaCl2-dissolution method, the degummed cocoons were dissolved in a ternary solvent of a CaCl2/H2O/EtOH (molar ratio 1/8/ 2) at 80 C for 5 min instead of LiBr solution. The hydrolysis was directly performed at 80 C for 10e120 min. The detail preparation conditions and sample ID were listed in Table 1 and Table S1. After dialysis, SF solutions were centrifuged at 3000 g for 10 min to remove insoluble aggregates. The final concentrations of SF solutions were in the range of 3.5e4% (w/v) and each solution was diluted to 3% (w/v) concentration. The prepared SF solutions were stored at 4 C until gel fabrication. Sodium oleate, lithium bromide 1-hydrate, and calcium chloride were purchased from Tokyo Chemical Industry, Kanto Chemical, and Yakuri, respectively. The other chemicals were purchased from SigmaeAldrich without further purification.
Table 1 Sample ID and preparation conditions of alkali hydrolyzed SF solutions using LiBrHAT method. Sample Dissolution condition ID L0 L10 L30 L90 L180
Dissolution Hydrolysis time (min) condition
Solvent: 9.3 M 30 LiBrTemperature: 80 CLiquor ratio: 1 g/5 mL
Solvent: 0.1 M NaOH Temperature: 80 C
Hydrolysis time (min) 0 10 30 90 180
3% (w/v) SF solution was transferred into polystyrene UV/Vis spectrometry cuvette and each cuvette was completely sealed. The absorbance was measured in the range between 400 and 700 nm by using a UV/Vis spectrometer (OPTIZEN POP, Mecasys, Korea). The path length was fixed at 10 mm. To measure the turbidity, SF hydrogels were formed in the same cuvette by ultra-sonication and subsequent incubation at 60 C for 2 days. 2.5. Rheometry To determine the gelation time (gel point), shear elastic and loss moduli (G0 and G00 ) were measured by a rheometer (HAAKE MARS III, Thermo Fisher Scientific, Germany) over the SF solution incubation time after ultra-sonication. The measurement was performed using a time-sweep oscillatory mode (strain: 5%, frequency: 1 Hz, gap size: 0.1 mm) with a parallel plate geometry (Dia.: 35 mm). All the solutions were stored at 60 C before measurement. Gel point was determined when G0 surpassed G00 . To measure the equilibrium shear elastic modulus, SF hydrogel slabs were formed, followed by subsequent swelling in pH 7.4 PB S at 37 C for 24 h. The swollen SF hydrogel was punched out using a biopsy punch (8 mm). Then, shear elastic modulus (G0 ) was measured by the rheometer using a strain-sweep oscillatory mode (strain: 0.1e10%, frequency: 1 Hz, gap size: 2.5 mm) with a parallel plate geometry (8 mm). After the measurement, G0 was determined from the linear viscoelastic region. 2.6. Swelling behavior To determine gel fraction of SF hydrogel, each gel was cut into a square slab (5 mm 5 mm 0.1 mm) and dried at 60 C for 24 h in vacuum right after the fabrication, followed by original dry-weight (Wd0) measurement. The dried gels were incubated in de-ionized water at 37 C for 24 h and washed several times. The swollen gels were then re-dried in vacuum for 24 h and its dry-weight (Wd1) was measured. The gel fraction was obtained by following Eq. (1).
28
Gel fraction ð%Þ ¼
H.H. Kim et al. / Polymer 90 (2016) 26e33
Wd1 100 Wd0
2.11. Cell culture
(1)
For swelling ratio measurement, each SF hydrogel was cut into a square slab (5 mm 5 mm 0.1 mm) and incubated in pH 7.4 PB S at 37 C for 24 h. Then, the samples were weighted to obtain swollen weight of hydrogel (Ws). The swollen hydrogels were washed to remove the residual ions of PBS for 24 h using de-ionized water, followed by vacuum drying and dry weight (Wd) measurement. The equilibrium mass swelling ratio of SF hydrogel was defined as following Eq. (2).
Swelling ratio ðQ Þ ¼
Ws Wd
(2)
2.7. Wide angle X-ray diffraction Wide angle X-ray diffraction (WAXD) method was used for examining crystal structure of SF hydrogel. X-ray diffractogram was obtained by two theta (2q) scanning with a GADDS (general area detector diffraction system, Brucker-Axs, Germany) using Cu Ka Xray (1.5406A) at 40 kV and 45 mA irradiation conditions. 2.8. Small angle X-ray scattering
hMSCs were cultured in low glucose DMEM (Dulbecco Modified Eagle's Medium, Low glucose) supplemented with 10% (v/v) FBS (Fetal bovine serum), 1% (v/v) antibiotic-antimycotic and 1 ng/mL bFGF (basic fibroblast growth factor) at 5% CO2 and 37 C. Before cell seeding, the SF hydrogels of different SF molecular weights were prepared in 48 well-plate following the previously described method at ‘SF hydrogel fabrication’. For sample sterilization, SF aqueous solutions were filtered using 0.2 mm membrane before gel fabrication. Then, the cells were trypsinized with 0.25% trypsinEDTA solution and hMSC cells of 15,000 were seeded on the top of each hydrogel. The seeded hMSCs were cultured in medium for 24 h at 5% CO2, 37 C. Finally, the hMSCs on SF hydrogel were observed using a microscope (Olympus CKX41, Olympus, Japan). Metabolic activity of hMSCs cultured on SF hydrogel was measured by CellTiter-Blue® assay (Promega, USA). Briefly, CellTiter-Blue® 10 reagent was diluted into FBS-contained DMEM at 10% (v/v). Then, 500 mL of the 1 CellTiter-Blue® reagent was added into each well after removing old culture medium, followed by incubation at 5% CO2 and 37 C for 2 h. 200 mL of reduced 1 CellTiter-Blue® reagent was transferred to 96-well plate for fluorescence measurement (ex/em: 560/590 nm) using a microplate reader (Synergy HT, Bio-Tek instruments, USA). Fluorescence images of hMSCs on SF hydrogels were obtained using a confocal laser scanning microscope (CLSM) (SP8 X STED, Leica, Germany). The hMSCs were
The physical structure of SF hydrogel was analyzed by using a small angle X-ray scattering (SAXS) spectrophotometer (TVXAENIF1, Techvalley, Korea). Wetted-samples (original hydrogel formation) were used for SAXS analysis. The distance from sample to detector was 1 m and the X-ray wavelength was 0.154 nm. Each SAXS pattern was collected for 10,000 s and converted to the scattering function I(q) as a function of the magnitude of the scattering vector (q). At a low q2 region (q2 < 0.024 nm), radius of gyration (Rg) was derived from Guinier Eq. (3), where IG(q) is the asymptotic value of the Guinier intensity at q / 0. The Rg value was calculated from the slope value of Guinier plot (q2 versus ln(IG(q)).
. IGðqÞ ¼ IGð0Þexp q2 Rg 2 3
(3)
2.9. Fourier transform infrared spectroscopy The secondary structure content of SF hydrogel was analyzed by attenuated total reflectance Fourier transform infrared spectroscopy (ATR FT-IR) (Nicolet 6700, Thermo Scientific, USA). For sample preparation, SF hydrogels were freeze-dried to prevent the structural deformation. Then, dried samples were scanned in an amide I region (1800-1000 cm1) with 128 scan number at 8 cm1. Peak deconvolution was conducted using Origin Pro 8.0 and each content of secondary structure of SF (e.g., b-sheet, random coil) was calculated [23]. 2.10. Thioflavin T assay The b-sheet formation of SF hydrogel was monitored by thioflavin T assay. For sample preparation, 990 mL of ultra-sonic wave treated SF solutions were mixed with 10 mL of 2 mM thioflavin T aqueous solution and prepared in 96 well-plate. Fluorescence was measured (ex/em: 450/485 nm) using a microplate reader (Synergy HT, Bio-Tek instruments, USA). All the samples were stored at 60 C before measurement.
Fig. 1. (A) Gel filtration chromatograms; (B) number average molecular weight (Mn) and polydispersity index (PDI) of alkali hydrolyzed SF by HAT.
H.H. Kim et al. / Polymer 90 (2016) 26e33
visualized 1-day post-seeding. Cell-attached hydrogels were fixed in 4% paraformaldehyde at room temperature for 10 min. Then, samples were rinsed with PBS and seeded cells were permeabilized with 0.25% Triton X-100, followed by blocking with 1% BSA in PBS containing 0.05% Tween20 at room temperature for 30 min. Samples were subsequently stained using rhodamine phalloidin (Molecular Probes®, USA) and DAPI dihydrochloride (Molecular Probes®, USA) for 15 min and washed with PBS. 2.12. Statistics All experiments were triplicated. Data were presented as mean ± SD. One-way ANOVA test was performed to determine the statistical significance between the indicated groups (*p < 0.05). 3. Results and discussion 3.1. Molecular weight control of SF The solubilized SF molecules are able to form hydrogel network structure by a so-called self-assembly mechanism, in which the physical cross-linking occurs between hydrophobic segments of SF. Such a physical cross-linking can be initiated or accelerated by various external stimuli, such as shear force, electric current, and ultra-sonication in physiologically preferred condition [7,12,24]. Therefore, the fabrication process and properties of SF hydrogels have been widely explored especially for biomedical purposes. Nevertheless, the properties of SF hydrogel were not much variable due to the distinct characteristics of a natural polymer SF. In this study, to manipulate the physical properties of SF hydrogel, the depolymerization of SF molecules by heat-alkaline treatment (HAT) was applied, resulting in a change of average molecular weight as well as distribution of SF molecules. At first, we tried to evaluate the availability of HAT on generally-known two dissolution methods (i.e., LiBr and CaCl2/H2O/EtOH solvent systems). By regulating HAT
29
time, similar Mn and PDI values of hydrolyzed SF could be obtained from both solvent systems, 82.7e263.1 kDa for LiBr and 73.5e252.5 kDa for CaCl2 (Table S2). However, in case of CaCl2-HAT method, the resulted SF solution was not able to form a hydrogel (Table S3). Hence, the ‘LiBr-HAT’ dissolution-hydrolysis method was finally adopted to fabricate SF hydrogels of different molecular weights. Fig. 1A revealed molecular weight distributions of hydrolyzed SF with different HAT times. The intact SF (L0) showed a relatively narrower single peak at around 8.2 mL (~443 kDa), which is attributed to the molecular weight of SF heavy chain. As the hydrolysis time increased, the first shoulder peak at 9.5 mL (~200 kDa) appeared. With further time progression, the second. shoulder peak developed at 14.0 mL (~17 kDa) while the intact SF peak at 8.2 mL gradually decreased. During alkali hydrolysis, the number average molecular weight (Mn) of SF decreased from 263.1 to 82.7 kDa while the polydispersity index (PDI) increased up to 3.4 at 180 min (Fig. 1B). 3.2. Effect of hydrolysis time on gelation behavior Fig. 2A show shear moduli (G0 and G00 ) of 3% (w/v) SF solutions hydrolyzed for different time (0e180 min) right after the ultrasonication. G0 was higher than G00 in intact SF solution (L0), indicating the typical gel state (Fig. 2Aa). In contrast, G0 was slightly lower than G00 in hydrolyzed SF solution (L180), indicating that the SF solution still maintained the sol state (Fig. 2Ae). And the other samples (L10, L30, and L90) of different hydrolysis time also showed the sol state (Fig. 2Ab, 2Ac, and 2Ad). These mean that hydrolyzed SF solutions remained sol after the ultra-sonication. To determine the gel point, ratios of G0 to G00 were obtained as a function of incubation time in 60 C for all the SF solutions and the gel state was determined when the ratio (G0 /G00 ) surpassed 1 as shown in Fig. 2B. In all samples, G0 /G00 increased with incubation time. Especially for L0, the G0 /G00 was about 7 even without 60 C
Fig. 2. (A) Shear elastic and loss moduli (G' & G00 ) of (a) non-, (b) 10-min, (c) 30-min, (d) 90-min, and (e) 180-min hydrolyzed 3% (w/v) SF solutions immediately after ultrasonication. (B) G'/G00 ratio change of SF solutions with incubation time (a G0 /G'0 ratio larger than 1 indicates gel state, mean ± SD, n ¼ 3). (C) Gel point (time when G'/G'' ¼ 1) of SF aqueous solutions formed of different molecular weight of SF.
30
H.H. Kim et al. / Polymer 90 (2016) 26e33
incubation. However, G0 /G00 slowly increased as the hydrolysis time was prolonged. For example, L180 took 7 times longer time to reach the same value (G0 /G00 ¼ 3) than L10. The gel point of each hydrolyzed SF solution was presented in Fig. 2C. Such an increase of gelation time was mainly attributed to the destruction of hydrophobic segments of SF. Matsumoto et al. and Nagarkar et al. reported that free SF chains build physical cross-links via b-sheet structure formation by intermolecular hydrogen bonding during the solegel transition [25,26]. Because the hydrophobic segment (GAGAGX (G ¼ Gly, A ¼ Ala, and X ¼ Ser or Tyr)) of SF plays a key role in the cross-linking, the longer hydrolysis caused a significant damage on SF chain segment which forms b-sheet structure, resulting in a lower content of b-sheet (Fig. S1B and S2). Consequently, the lower Mn of SF did not efficiently form a hydrogel due to a lack of gel formation ability. 3.3. Effect of molecular weight of SF on hydrogel properties Fig. 3A and B presents gel fraction and equilibrium mass swelling ratio, respectively. The hydrogel formed with intact SF (L0) showed about 98% gel fraction while the SF hydrogel of the lowest Mn showed only 18%, indicating most part of SF molecules of low Mn did not participated in gel formation. On the contrary, the
Fig. 4. Shear elastic modulus of SF hydrogels formed of different molecular weights of SF (mean ± SD, n ¼ 3, *p < 0.01).
swelling ratio increased as Mn of SF decreased since the loose network of low Mn SF had larger space to capture more water. Fig. 4 shows equilibrium shear elastic dynamic modulus (G0 ) of SF hydrogels with different Mn. On the whole, the shear elastic modulus increased with Mn of SF. However, the Mn that formed the stiffest SF hydrogel (L10, G0 z 16.7 kPa) was not the highest Mn (~270 kDa) of the intact SF but 200 kDa). It might be due to too rapid gelation of L0 during ultra-sonication, resulting in inhomogeneous network formation. Comparing to 1-min ultra-sonicated L0 hydrogel, gel formation time of 3-min ultra-sonicated L0 was so fast and not sufficient to form the homogeneous gel structure. In addition, longer ultra-sonication treatment time increases the seeding density of SF aqueous solution [27], which consequently lowers the G0 of SF hydrogel (Fig. S3). Interestingly, we found that the hydrolyzed SF solution and hydrogel showed quite different light transmittance with molecular weight of SF. Fig. 5 shows
Fig. 3. (A) Gel fraction and (B) equilibrium swelling ratio of SF hydrogels formed of different molecular weights of SF (mean ± SD, n ¼ 3).
Fig. 5. (A) Images of SF hydrogels formed of different molecular weights of SF. The bottom letters depicts transparency of each hydrogel. (B) Transmittance of SF hydrogels at different wavelength of visible light (mean ± SD, n ¼ 3).
H.H. Kim et al. / Polymer 90 (2016) 26e33
31
Fig. 6. (A) SAXS curves of SF hydrogels with different molecular weights of SF. (B) Guinier plot curves. (C) Radius of gyration (derived from Guinier equation) of SF hydrogels with different molecular weights of SF.
Fig. 7. Schematic of microscopic structures of non-/180 min-hydrolyzed SF hydrogels.
transmittance of SF hydrogel formed with different molecular weights of SF. Generally, SF hydrogel was completely opaque (Fig. 5A, L0) due to the crystalline structure formation [25]. However, transparency of SF hydrogel gradually increased with an increase of hydrolysis time (Fig. 5A). In general, crystallinity is crucial factor affecting the optical property of polymeric materials. However, we could not observe noticeable difference in crystallinities of SF hydrogels by WAXD analysis. It might be due to a high water content (>95%) in hydrogel (Fig. S4). Alternatively, conformational analysis was conducted on the SF hydrogel by thioflavin T assay and FT-IR spectroscopy. As a result, b-sheet content of SF decreased with an increase of hydrolysis time (Figs. S1 and S2), indicating variation of b-sheet content influenced the transparency of SF hydrogel. However, it is not the only factor determining the transparency of hydrogel because hydrogel is basically a composite material of water phase and polymeric network. Therefore, hydrogels can be opaque even if the polymeric phase is completely amorphous. To quantitatively evaluate the optical property of SF hydrogels, the transmittance in the visible light wavelength range (400e700 nm) was measured using a spectrophotometer. As a result, the transmittance increased as the Mn of SF decreased at all wavelengths while L0 showed less than 10% (Fig. 5B). Interestingly,
an increment of transmittance with a decrease of Mn was much higher at a longer wavelength compared with a shorter wavelength. Although L180 exhibited about 80% at 700 nm, the transmittance at 400 nm was lower than 40%. It is noted that such different transmittances of SF hydrogels are closely related to structural cluster which forms physical network structure of polymer chains [28]. Herein, the transparency of SF hydrogel could increase when the cluster size became smaller than visible light wavelength range (400e700 nm), indicating the shorter SF chains (i.e., lower Mn) make smaller clusters. Therefore, it is speculated that the average cluster size of transparent SF hydrogel would be mostly smaller than 400 nm (Fig. 5B). 3.4. Effect of molecular weight on microscopic structure of SF hydrogels Fig. 6A shows SAXS curves of SF hydrogel. For further analysis of basic structure constituting SF hydrogel, the SAXS curves were converted into I(q) at the low q2 region (0.007e0.024) using Guinier equation (Eq. (3)) (Fig. 6B). The radius of gyration (Rg) obtained via Guinier plotting generally represents the gyration of a polymer chain in solution [29,30]. In case of a hydrogel, Rg indicates
32
H.H. Kim et al. / Polymer 90 (2016) 26e33
Fig. 8. (A) Phase-contrast images of hMSC cultured on SF hydrogels of different hydrolysis times 24-h post-seeding (scale: 100 mm). (B) Metabolic activity of hMSC cultured on SF hydrogels of different molecular weights 24-h post-seeding (mean ± SD, n ¼ 3). (C) Confocal images of hMSC on SF hydrogels 24-h post-seeding: (a) L0 and (b) L180 (F-actin: red, nucleus: blue).
the size of the polymer-rich solid-like domains (PRSDs) [29]. The Rg of SF hydrogel was in the range of 9.5e12.6 nm and also tended to increase with Mn of SF increment (Fig. 6C). S. Nagarkar et al. previously reported that aggregates of SF hydrogel made of thin strands of length less than 15 nm [26]. These results correspond to the size of SF hydrogel PRSDs value derived from Guinier equation. Accordingly, we hypothesized that this PRSDs was smallest base structural unit in SF hydrogel. Base on the previous results (Figs. 5 and 6), we depict the microscopic structures of SF hydrogels constituted of unhydrolyzed/hydrolyzed SF molecules, respectively, in Fig. 7. In the non-hydrolyzed (opaque) SF hydrogel (e.g., L0 of the highest Mn), SF chains formed PRSDs of 12e13 nm in diameter. After ultrasonication treatment, those domains were agglomerated together to form single domain network. Then, these single domain networks were further grown to form hydrogel cluster, which so we called aggregated particle connected network or fibril entangled network [26,27]. The size of cluster was larger than visible light wavelength range (400e700 nm) that visible light could not penetrate the hydrogel network structure and scattered at the cluster of hydrogel, consequently looks opaque. On the other hand, in the hydrolyzed (transparent) SF hydrogel (e.g., L180 of the lowest Mn), the fragmented SF chains formed relatively smaller PRSDs in the range of 9e10 nm and they built much smaller cluster of size smaller than visible light wavelength range. Consequently, such a hydrogel finally showed a higher visible light transmittance (Fig. 5). 3.5. Cell adhesion on SF hydrogels To evaluate cell adhesion on SF hydrogels, hMSCs were cultured on SF for 24 h. The hMSCs well attached and spread on SF hydrogels from relatively higher Mn SF hydrogels (i.e., L0, L10, and L30) while the cells maintained spherical shape and less cells were observed on the relatively lower Mn SF hydrogels (i.e., L90 and L180) (Fig. 8A). The metabolic activity of hMSCs cultured on SF hydrogels showed a corresponding decrease as the Mn of SF decreased (Fig. 8B). Fig. 8C depicts F-actin structure of hMSCs on SF 24-h post-seeding. It was shown that F-actin was well developed on the whole cytoplasm of hMSCs on the intact SF hydrogel (L0, Fig. 8Ca) while observed at round boundary near cell membrane (L180, Fig. 8Cb). It is generally known that cell attachment and spreading are governed by the
matrix stiffness [31,32]. As expected, a fewer hMSCs attached on L180 gel surface with less spreading while much more hMSCs remained and spread on the surface of stiffer SF hydrogels (i.e., L0) with well-developed actin cytoskeletons. Hence, it is proved that SF hydrogel properties could be altered in very wide ranges like other synthetic polymer-based hydrogels by molecular weight control of SF. Especially, the improved transparency of depolymerized SF hydrogel will be very useful for cell morphology observation and various biological assays. Also, this technique can expand the optical device application of silk-based materials.
4. Conclusions In conclusion, the gelling retardation of depolymerized SF solution was caused by lack of physical cross-linking formation of hydrolyzed SF molecular chains. The SF hydrogels prepared from depolymerized SF by HAT showed variable physical properties with different molecular weights. Especially, both shear elastic modulus and visible light transmittance could be controlled in very wide ranges (0.4e16.7 kPa and 9.7e79.8%), respectively. Such an alternation in hydrogel property was mainly due to a difference in microscopic structure of SF hydrogel. The shorter SF chains rendered smaller base structural units in hydrogel, resulting in relatively loose network structure with a higher porosity. Consequently, it is expected that the molecular weight control manner could be utilized to enhance silk hydrogel performance as well as other silk-based materials in biomedical engineering application.
Acknowledgement This work was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (NRF-2015R1A2A2A03002680).
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.polymer.2016.02.054.
H.H. Kim et al. / Polymer 90 (2016) 26e33
References [1] J.S. Ko, K. Yoon, C.S. Ki, H.J. Kim, D.G. Bae, K.H. Lee, Y.H. Park, I.C. Um, Effect of degumming condition on the solution properties and electrospinnablity of regenerated silk solution, Int. J. Biol. Macromol. 55 (2013) 161e168. [2] B. Kundu, N.E. Kurland, S. Bano, C. Patra, F.B. Engel, V.K. Yadavalli, S.C. Kundu, Silk proteins for biomedical applications: Bioengineering perspectives, Prog. Polym. Sci. 39 (2) (2014) 251e267. [3] B. Kundu, R. Rajkhowa, S.C. Kundu, X. Wang, Silk fibroin biomaterials for tissue regenerations, Adv. Drug Deliv. Rev. 65 (4) (2013) 457e470. [4] E.M. Pritchard, D.L. Kaplan, Silk fibroin biomaterials for controlled release drug delivery, Exp. Opin. Drug Deliv. 8 (6) (2011) 797e811. [5] A.B. Mathur, V. Gupta, Silk fibroin-derived nanoparticles for biomedical applications, Nanomedicine 5 (5) (2010) 807e820. [6] C.S. Ki, Y.H. Park, H.-J. Jin, Silk protein as a fascinating biomedical polymer: structural fundamentals and applications, Macromol. Res. 17 (12) (2009) 935e942. [7] T. Yucel, P. Cebe, D.L. Kaplan, Vortex-induced injectable silk fibroin hydrogels, Biophys. J. 97 (7) (2009) 2044e2050. [8] N. Kasoju, U. Bora, Silk fibroin in tissue engineering, Adv. Healthc. Mater 1 (4) (2012) 393e412. [9] Z.-H. Li, S.-C. Ji, Y.-Z. Wang, X.-C. Shen, H. Liang, Silk fibroin-based scaffolds for tissue engineering, Front. Mater. Sci. 7 (3) (2013) 237e247. [10] C. Vepari, D.L. Kaplan, Silk as a biomaterial, Prog. Polym. Sci. 32 (8) (2007) 991e1007. [11] N.E. Davis, L.N. Beenken-Rothkopf, A. Mirsoian, N. Kojic, D.L. Kaplan, A.E. Barron, M.J. Fontaine, Enhanced function of pancreatic islets coencapsulated with ECM proteins and mesenchymal stromal cells in a silk hydrogel, Biomaterials 33 (28) (2012) 6691e6697. [12] X. Wang, J.A. Kluge, G.G. Leisk, D.L. Kaplan, Sonication-induced gelation of silk fibroin for cell encapsulation, Biomaterials 29 (8) (2008) 1054e1064. [13] M.J. Mahoney, K.S. Anseth, Three-dimensional growth and function of neural tissue in degradable polyethylene glycol hydrogels, Biomaterials 27 (10) (2006) 2265e2274. [14] C.M. Hassan, N.A. Peppas, Structure and applications of poly (vinyl alcohol) hydrogels produced by conventional crosslinking or by freezing/thawing methods, Adv. Polym. Sci. (2000) 37e65. [15] U.-J. Kim, J. Park, C. Li, H.-J. Jin, R. Valluzzi, D.L. Kaplan, Structure and properties of silk hydrogels, Biomacromolecules 5 (3) (2004) 786e792. [16] K. Numata, T. Katashima, T. Sakai, State of water, molecular structure, and cytotoxicity of silk hydrogels, Biomacromolecules 12 (6) (2011) 2137e2144. [17] I. Noh, G.-W. Kim, Y.-J. Choi, M.-S. Kim, Y. Park, K.-B. Lee, I.-S. Kim, S.-J. Hwang, G. Tae, Effects of cross-linking molecular weights in a hyaluronic acid -
[18]
[19]
[20]
[21]
[22] [23]
[24] [25]
[26] [27]
[28]
[29]
[30]
[31] [32]
33
poly(ethylene oxide) hydrogel network on its properties, Biomed. Mater 1 (3) (2006) 116e123. Y.L. Zhang, C.C. Chu, The effect of molecular weight of biodegradable hydrogel components on indomethacin release from dextran and poly(DL)lactic acid based hydrogels, J. Bioact. Compat. Polym. 17 (2) (2002) 65e85. R. Fukae, M. Yoshimura, T. Yamamoto, K. Nishinari, Effect of stereoregularity and molecular weight on the mechanical properties of poly(vinyl alcohol) hydrogel, J. Appl. Polym. Sci. 120 (1) (2011) 573e578. H.Y. Zhou, X.G. Chen, M. Kong, C.S. Liu, D.S. Cha, J.F. Kennedy, Effect of molecular weight and degree of chitosan deacetylation on the preparation and characteristics of chitosan thermosensitive hydrogel as a delivery system, Carbohydr. Polym. 73 (2) (2008) 265e273. H.J. Kong, D. Kaigler, K. Kim, D.J. Mooney, Controlling rigidity and degradation of alginate hydrogels via molecular weight distribution, Biomacromolecules 5 (5) (2004) 1720e1727. M. Li, M. Ogiso, N. Minoura, Enzymatic degradation behavior of porous silk fibroin sheets, Biomaterials 24 (2) (2003) 357e365. H. Zhang, L. Deng, M. Yang, S. Min, L. Yang, L. Zhu, Enhancing effect of glycerol on the tensile properties of bombyx mori cocoon sericin films, Int. J. Mol. Sci. 12 (5) (2011) 3170. G.G. Leisk, T.J. Lo, T. Yucel, Q. Lu, D.L. Kaplan, Electrogelation for protein adhesives, Adv. Mater 22 (6) (2010) 711e715. A. Matsumoto, J. Chen, A.L. Collette, U.-J. Kim, G.H. Altman, P. Cebe, D.L. Kaplan, Mechanisms of silk fibroin sol-gel transitions, J. Pyhs. Chem. B 110 (43) (2006) 21630e21638. S. Nagarkar, T. Nicolai, C. Chassenieux, A. Lele, Structure and gelation mechanism of silk hydrogels, Phys. Chem. Chem. Phys. 12 (15) (2010) 3834e3844. H.J. Cho, C.S. Ki, H. Oh, K.H. Lee, I.C. Um, Molecular weight distribution and solution properties of silk fibroins with different dissolution conditions, Int. J. Biol. Macromol. 51 (3) (2012) 336e341. L. Yu, Z. Zhang, J. Ding, In vitro degradation and protein release of transparent and opaque physical hydrogels of block copolymers at body temperature, Macromol. Res. 20 (3) (2012) 234e243. L. Wang, G. Shan, P. Pan, A strong and tough interpenetrating network hydrogel with ultrahigh compression resistance, Soft Matter 10 (21) (2014) 3850e3856. L. Zhao, L. Li, G.-Q. Liu, X.-X. Liu, B. Li, Effect of frozen storage on molecular weight, size distribution and conformation of gluten by SAXS and SEC-MALLS, Molecules 17 (6) (2012) 7169e7182. R.G. Wells, The role of matrix stiffness in regulating cell behavior, Hepatology 47 (4) (2008) 1394e1400. R.G. Breuls, T.U. Jiya, T.H. Smit, Scaffold stiffness influences cell behavior: opportunities for skeletal tissue engineering, Open Orthop. J. 2 (2008) 103.
Contents lists available at ScienceDirect
Polymer journal homepage: www.elsevier.com/locate/polymer
Effect of silk fibroin molecular weight on physical property of silk hydrogel H.H. Kim a, D.W. Song a, M.J. Kim a, S.J. Ryu a, I.C. Um b, C.S. Ki a, **, Y.H. Park a, * a b
Department of Biosystems and Biomaterials Science and Engineering, Seoul National University, Seoul 151-921, Republic of Korea Department of Bio-fibers and Materials Science, Kyungpook National University, Daegu 702-701, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 December 2015 Received in revised form 19 February 2016 Accepted 23 February 2016 Available online 26 February 2016
Silk hydrogel has recently received great attention for its excellent biocompatibility. The property of silk hydrogel is, however, not only hardly controlled but very limited due to non-variability of silk fibroin (SF) molecule. In this study, alkaline hydrolysis was utilized to manipulate the silk hydrogel properties. By regulating the hydrolysis time (10e180 min), a broad molecular weight range of SF was obtained (263.1 e82.7 kDa) Gel point increased with a decrease of SF molecular weight. The change of molecular weight of SF also greatly affected the physical properties (i.e., swelling ratio, shear modulus, transparency) as well as cell adhesion of SF hydrogels. As a result of structural analysis, the molecular weight of SF played a crucial role in the construction of microscopic structure of SF hydrogel. These findings indicate that SF hydrogels of variable physical properties can be fabricated based on molecular weight control for diverse purposes in biomedical engineering. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Silk hydrogel Alkaline hydrolysis Molecular weight Transparent hydrogel Microscopic structure
1. Introduction Silk fibroin (SF) is a major protein, which plays a critical role in both structural feature and mechanical property of silk cocoons. In raw silk fibers, two SF strands are covered with a gum-like protein, silk sericin (SS), which is easily removed by weak alkali hot water [1]. To date, it has been widely reported that SF shows excellent biocompatibility as well as superior mechanical property in biomedical applications [2e7]. Besides, the facile recrystallization by alcohol treatment inducing the physical cross-linking via intraand inter-molecular hydrogen bonding of hydrophobic segments of SF is generally exploited in the fabrication process [8,9]. Such a process is very useful in biomedical material fabrications (e.g., tissue engineering scaffold, drug carrier) since this process does not cause any harmful chemical species damaging living organisms. SF can be fabricated into various forms (e.g., film, nanofiber, sponge, hydrogel) [10]. Especially, SF hydrogel has been recently received great attention in tissue engineering field. The network structure of SF hydrogel retains the large amount of water with inter-connected porous structure. This structure provides the
* Corresponding author. ** Co-corresponding author. E-mail addresses: [email protected] (C.S. Ki), [email protected] (Y.H. Park). http://dx.doi.org/10.1016/j.polymer.2016.02.054 0032-3861/© 2016 Elsevier Ltd. All rights reserved.
physiologically stable condition for cell survival [11,12]. Also, porous three-dimensional (3D) structure of SF hydrogel is an appropriate microenvironment in which cells reside allowing exchanging nutrients and wastes. Therefore, SF hydrogel shows very excellent biological properties compared to synthetic polymer-based hydrogel (e.g., poly(ethylene glycol) [13], poly(vinyl alcohol) [14]) as a 3D cell niche. The properties of SF hydrogel are affected by a wide variety of processing parameters, such as concentration of SF solution [15,16], incubation temperature [15], vortexing time [7], and ultra-sonic power [12], The concentration is a major factor to determine physical properties of SF hydrogel [15,16]. Generally, the higher concentration results in the stiffer hydrogels with the shorter gelation time. The gelation behavior is also largely dependent on the incubation temperature [15]. For example, the SF hydrogel formed at a higher temperature (~50 C) mostly has a higher Young's modulus than that formed at a lower temperature (~4 C). The effect of amplitude of shear force was also investigated. Wang et al. has shown that the amplitude of ultra-sonication affects the mechanical properties of SF hydrogel [12]. Yucel et al. reported the vortexing time effect on the shear modulus of SF hydrogel [7]. Although the significant few studies to manipulate SF hydrogel properties were reported, none of processing variables could elicit the change of a wide range in mechanical as well as physical properties except the concentration of the precursor SF solution. Generally, the molecular weight of polymer composing the
H.H. Kim et al. / Polymer 90 (2016) 26e33
network structure of hydrogel is a crucial factor that determines the gel properties (i.e., modulus, swelling ratio, permeability) [17e19]. Both cross-linking density and mesh size of hydrogel are largely dependent on the polymer chain length regardless of cross-linking methods (i.e., chemical and physical cross-linking) [20,21]. Nevertheless, the effect of molecular weight of SF on hydrogel fabrication has not been explored. Hence, in this study, we conducted the heatalkaline treatment (HAT) during the SF dissolution step to control the molecular weight of SF as a one-pot process. This method is not only simple but also suitable to obtain a high yield of hydrolyzed SF. In contrast, proteinase treatment needs relatively delicate process including enzyme removing process and causes sever mass loss of product in spite of high efficiency [22]. Then, we formed SF hydrogel by ultra-sonication, followed by physical and mechanical property analyses to investigate the effect of molecular weight of SF on the hydrogel formation. Especially, we tried to focus on the change of the microscopic structure of SF hydrogel according to molecular weight variation, which allows the wide range property control of the hydrogel. Finally, it was demonstrated that such a property manipulation of SF hydrogel can be successfully utilized in tissue engineering field by human mesenchymal stem cell (hMSC) culture on SF hydrogels.
27
2.2. SF hydrogel fabrication To initiate the gelation of SF solution, ultra-sonication was performed on 3% (w/v) SF aqueous solution of different molecular weights using an ultra-sonic processor (VCX-130, SONICS, USA) at 25% amplitude for 3 min. The treatment was conducted in an ice chamber to prevent the heat elevation during ultra-sonication. Then, ultra-sonicated SF solution was filtered and incubated at 60 C for 2 days. 2.3. Gel filtration chromatography The molecular weight of SF was measured by gel filtration chromatography (GFC) (AKTA Purifier, GE Healthcare, USA) with Superdex 200 10/300 GL column (GE Healthcare, USA). 0.5 mL of 3% (w/v) SF solution was added to 4 mL of 6 M urea aqueous solution, followed by filtration with 0.2 mm membrane. For the measurement, 300 mL of sample solution was injected and 1.5 column volume of 4 M urea was eluted at a constant flow rate of 0.5 mL/min. The elution of SF was detected at 280 nm. The molecular weight of SF was determined by a calibration curve, which was obtained by a standard globular protein kit (Gel Filtration Cal Kit High Molecular Weight, GE Healthcare, USA).
2. Experimental 2.4. Visible light transmittance measurement
2.1. Materials To remove silk sericin, Bombyx mori cocoons were boiled in 0.3% (w/v) sodium oleate and 0.2% (w/v) sodium carbonate cocktail solution at 100 C for 1 h. The SF aqueous solutions were obtained by using two different dissolving methods. For LiBr-dissolution method, the degummed cocoons were dissolved in 9.3 M LiBr solution at 80 C for 30 min. To hydrolyze SF, 0.6 M sodium hydroxide aqueous solution was directly added to the SF solution at a volume ratio of 1-to-5. Then, the final concentration of sodium hydroxide became 0.1 M in the SF solution. To control SF molecular weight, the hydrolysis time was varied from 10 to 180 min, followed by subsequent dialysis against de-ionized water using cellulose acetate dialysis tube (MWCO: 12,000e14,000 Da) for 3 days. For CaCl2-dissolution method, the degummed cocoons were dissolved in a ternary solvent of a CaCl2/H2O/EtOH (molar ratio 1/8/ 2) at 80 C for 5 min instead of LiBr solution. The hydrolysis was directly performed at 80 C for 10e120 min. The detail preparation conditions and sample ID were listed in Table 1 and Table S1. After dialysis, SF solutions were centrifuged at 3000 g for 10 min to remove insoluble aggregates. The final concentrations of SF solutions were in the range of 3.5e4% (w/v) and each solution was diluted to 3% (w/v) concentration. The prepared SF solutions were stored at 4 C until gel fabrication. Sodium oleate, lithium bromide 1-hydrate, and calcium chloride were purchased from Tokyo Chemical Industry, Kanto Chemical, and Yakuri, respectively. The other chemicals were purchased from SigmaeAldrich without further purification.
Table 1 Sample ID and preparation conditions of alkali hydrolyzed SF solutions using LiBrHAT method. Sample Dissolution condition ID L0 L10 L30 L90 L180
Dissolution Hydrolysis time (min) condition
Solvent: 9.3 M 30 LiBrTemperature: 80 CLiquor ratio: 1 g/5 mL
Solvent: 0.1 M NaOH Temperature: 80 C
Hydrolysis time (min) 0 10 30 90 180
3% (w/v) SF solution was transferred into polystyrene UV/Vis spectrometry cuvette and each cuvette was completely sealed. The absorbance was measured in the range between 400 and 700 nm by using a UV/Vis spectrometer (OPTIZEN POP, Mecasys, Korea). The path length was fixed at 10 mm. To measure the turbidity, SF hydrogels were formed in the same cuvette by ultra-sonication and subsequent incubation at 60 C for 2 days. 2.5. Rheometry To determine the gelation time (gel point), shear elastic and loss moduli (G0 and G00 ) were measured by a rheometer (HAAKE MARS III, Thermo Fisher Scientific, Germany) over the SF solution incubation time after ultra-sonication. The measurement was performed using a time-sweep oscillatory mode (strain: 5%, frequency: 1 Hz, gap size: 0.1 mm) with a parallel plate geometry (Dia.: 35 mm). All the solutions were stored at 60 C before measurement. Gel point was determined when G0 surpassed G00 . To measure the equilibrium shear elastic modulus, SF hydrogel slabs were formed, followed by subsequent swelling in pH 7.4 PB S at 37 C for 24 h. The swollen SF hydrogel was punched out using a biopsy punch (8 mm). Then, shear elastic modulus (G0 ) was measured by the rheometer using a strain-sweep oscillatory mode (strain: 0.1e10%, frequency: 1 Hz, gap size: 2.5 mm) with a parallel plate geometry (8 mm). After the measurement, G0 was determined from the linear viscoelastic region. 2.6. Swelling behavior To determine gel fraction of SF hydrogel, each gel was cut into a square slab (5 mm 5 mm 0.1 mm) and dried at 60 C for 24 h in vacuum right after the fabrication, followed by original dry-weight (Wd0) measurement. The dried gels were incubated in de-ionized water at 37 C for 24 h and washed several times. The swollen gels were then re-dried in vacuum for 24 h and its dry-weight (Wd1) was measured. The gel fraction was obtained by following Eq. (1).
28
Gel fraction ð%Þ ¼
H.H. Kim et al. / Polymer 90 (2016) 26e33
Wd1 100 Wd0
2.11. Cell culture
(1)
For swelling ratio measurement, each SF hydrogel was cut into a square slab (5 mm 5 mm 0.1 mm) and incubated in pH 7.4 PB S at 37 C for 24 h. Then, the samples were weighted to obtain swollen weight of hydrogel (Ws). The swollen hydrogels were washed to remove the residual ions of PBS for 24 h using de-ionized water, followed by vacuum drying and dry weight (Wd) measurement. The equilibrium mass swelling ratio of SF hydrogel was defined as following Eq. (2).
Swelling ratio ðQ Þ ¼
Ws Wd
(2)
2.7. Wide angle X-ray diffraction Wide angle X-ray diffraction (WAXD) method was used for examining crystal structure of SF hydrogel. X-ray diffractogram was obtained by two theta (2q) scanning with a GADDS (general area detector diffraction system, Brucker-Axs, Germany) using Cu Ka Xray (1.5406A) at 40 kV and 45 mA irradiation conditions. 2.8. Small angle X-ray scattering
hMSCs were cultured in low glucose DMEM (Dulbecco Modified Eagle's Medium, Low glucose) supplemented with 10% (v/v) FBS (Fetal bovine serum), 1% (v/v) antibiotic-antimycotic and 1 ng/mL bFGF (basic fibroblast growth factor) at 5% CO2 and 37 C. Before cell seeding, the SF hydrogels of different SF molecular weights were prepared in 48 well-plate following the previously described method at ‘SF hydrogel fabrication’. For sample sterilization, SF aqueous solutions were filtered using 0.2 mm membrane before gel fabrication. Then, the cells were trypsinized with 0.25% trypsinEDTA solution and hMSC cells of 15,000 were seeded on the top of each hydrogel. The seeded hMSCs were cultured in medium for 24 h at 5% CO2, 37 C. Finally, the hMSCs on SF hydrogel were observed using a microscope (Olympus CKX41, Olympus, Japan). Metabolic activity of hMSCs cultured on SF hydrogel was measured by CellTiter-Blue® assay (Promega, USA). Briefly, CellTiter-Blue® 10 reagent was diluted into FBS-contained DMEM at 10% (v/v). Then, 500 mL of the 1 CellTiter-Blue® reagent was added into each well after removing old culture medium, followed by incubation at 5% CO2 and 37 C for 2 h. 200 mL of reduced 1 CellTiter-Blue® reagent was transferred to 96-well plate for fluorescence measurement (ex/em: 560/590 nm) using a microplate reader (Synergy HT, Bio-Tek instruments, USA). Fluorescence images of hMSCs on SF hydrogels were obtained using a confocal laser scanning microscope (CLSM) (SP8 X STED, Leica, Germany). The hMSCs were
The physical structure of SF hydrogel was analyzed by using a small angle X-ray scattering (SAXS) spectrophotometer (TVXAENIF1, Techvalley, Korea). Wetted-samples (original hydrogel formation) were used for SAXS analysis. The distance from sample to detector was 1 m and the X-ray wavelength was 0.154 nm. Each SAXS pattern was collected for 10,000 s and converted to the scattering function I(q) as a function of the magnitude of the scattering vector (q). At a low q2 region (q2 < 0.024 nm), radius of gyration (Rg) was derived from Guinier Eq. (3), where IG(q) is the asymptotic value of the Guinier intensity at q / 0. The Rg value was calculated from the slope value of Guinier plot (q2 versus ln(IG(q)).
. IGðqÞ ¼ IGð0Þexp q2 Rg 2 3
(3)
2.9. Fourier transform infrared spectroscopy The secondary structure content of SF hydrogel was analyzed by attenuated total reflectance Fourier transform infrared spectroscopy (ATR FT-IR) (Nicolet 6700, Thermo Scientific, USA). For sample preparation, SF hydrogels were freeze-dried to prevent the structural deformation. Then, dried samples were scanned in an amide I region (1800-1000 cm1) with 128 scan number at 8 cm1. Peak deconvolution was conducted using Origin Pro 8.0 and each content of secondary structure of SF (e.g., b-sheet, random coil) was calculated [23]. 2.10. Thioflavin T assay The b-sheet formation of SF hydrogel was monitored by thioflavin T assay. For sample preparation, 990 mL of ultra-sonic wave treated SF solutions were mixed with 10 mL of 2 mM thioflavin T aqueous solution and prepared in 96 well-plate. Fluorescence was measured (ex/em: 450/485 nm) using a microplate reader (Synergy HT, Bio-Tek instruments, USA). All the samples were stored at 60 C before measurement.
Fig. 1. (A) Gel filtration chromatograms; (B) number average molecular weight (Mn) and polydispersity index (PDI) of alkali hydrolyzed SF by HAT.
H.H. Kim et al. / Polymer 90 (2016) 26e33
visualized 1-day post-seeding. Cell-attached hydrogels were fixed in 4% paraformaldehyde at room temperature for 10 min. Then, samples were rinsed with PBS and seeded cells were permeabilized with 0.25% Triton X-100, followed by blocking with 1% BSA in PBS containing 0.05% Tween20 at room temperature for 30 min. Samples were subsequently stained using rhodamine phalloidin (Molecular Probes®, USA) and DAPI dihydrochloride (Molecular Probes®, USA) for 15 min and washed with PBS. 2.12. Statistics All experiments were triplicated. Data were presented as mean ± SD. One-way ANOVA test was performed to determine the statistical significance between the indicated groups (*p < 0.05). 3. Results and discussion 3.1. Molecular weight control of SF The solubilized SF molecules are able to form hydrogel network structure by a so-called self-assembly mechanism, in which the physical cross-linking occurs between hydrophobic segments of SF. Such a physical cross-linking can be initiated or accelerated by various external stimuli, such as shear force, electric current, and ultra-sonication in physiologically preferred condition [7,12,24]. Therefore, the fabrication process and properties of SF hydrogels have been widely explored especially for biomedical purposes. Nevertheless, the properties of SF hydrogel were not much variable due to the distinct characteristics of a natural polymer SF. In this study, to manipulate the physical properties of SF hydrogel, the depolymerization of SF molecules by heat-alkaline treatment (HAT) was applied, resulting in a change of average molecular weight as well as distribution of SF molecules. At first, we tried to evaluate the availability of HAT on generally-known two dissolution methods (i.e., LiBr and CaCl2/H2O/EtOH solvent systems). By regulating HAT
29
time, similar Mn and PDI values of hydrolyzed SF could be obtained from both solvent systems, 82.7e263.1 kDa for LiBr and 73.5e252.5 kDa for CaCl2 (Table S2). However, in case of CaCl2-HAT method, the resulted SF solution was not able to form a hydrogel (Table S3). Hence, the ‘LiBr-HAT’ dissolution-hydrolysis method was finally adopted to fabricate SF hydrogels of different molecular weights. Fig. 1A revealed molecular weight distributions of hydrolyzed SF with different HAT times. The intact SF (L0) showed a relatively narrower single peak at around 8.2 mL (~443 kDa), which is attributed to the molecular weight of SF heavy chain. As the hydrolysis time increased, the first shoulder peak at 9.5 mL (~200 kDa) appeared. With further time progression, the second. shoulder peak developed at 14.0 mL (~17 kDa) while the intact SF peak at 8.2 mL gradually decreased. During alkali hydrolysis, the number average molecular weight (Mn) of SF decreased from 263.1 to 82.7 kDa while the polydispersity index (PDI) increased up to 3.4 at 180 min (Fig. 1B). 3.2. Effect of hydrolysis time on gelation behavior Fig. 2A show shear moduli (G0 and G00 ) of 3% (w/v) SF solutions hydrolyzed for different time (0e180 min) right after the ultrasonication. G0 was higher than G00 in intact SF solution (L0), indicating the typical gel state (Fig. 2Aa). In contrast, G0 was slightly lower than G00 in hydrolyzed SF solution (L180), indicating that the SF solution still maintained the sol state (Fig. 2Ae). And the other samples (L10, L30, and L90) of different hydrolysis time also showed the sol state (Fig. 2Ab, 2Ac, and 2Ad). These mean that hydrolyzed SF solutions remained sol after the ultra-sonication. To determine the gel point, ratios of G0 to G00 were obtained as a function of incubation time in 60 C for all the SF solutions and the gel state was determined when the ratio (G0 /G00 ) surpassed 1 as shown in Fig. 2B. In all samples, G0 /G00 increased with incubation time. Especially for L0, the G0 /G00 was about 7 even without 60 C
Fig. 2. (A) Shear elastic and loss moduli (G' & G00 ) of (a) non-, (b) 10-min, (c) 30-min, (d) 90-min, and (e) 180-min hydrolyzed 3% (w/v) SF solutions immediately after ultrasonication. (B) G'/G00 ratio change of SF solutions with incubation time (a G0 /G'0 ratio larger than 1 indicates gel state, mean ± SD, n ¼ 3). (C) Gel point (time when G'/G'' ¼ 1) of SF aqueous solutions formed of different molecular weight of SF.
30
H.H. Kim et al. / Polymer 90 (2016) 26e33
incubation. However, G0 /G00 slowly increased as the hydrolysis time was prolonged. For example, L180 took 7 times longer time to reach the same value (G0 /G00 ¼ 3) than L10. The gel point of each hydrolyzed SF solution was presented in Fig. 2C. Such an increase of gelation time was mainly attributed to the destruction of hydrophobic segments of SF. Matsumoto et al. and Nagarkar et al. reported that free SF chains build physical cross-links via b-sheet structure formation by intermolecular hydrogen bonding during the solegel transition [25,26]. Because the hydrophobic segment (GAGAGX (G ¼ Gly, A ¼ Ala, and X ¼ Ser or Tyr)) of SF plays a key role in the cross-linking, the longer hydrolysis caused a significant damage on SF chain segment which forms b-sheet structure, resulting in a lower content of b-sheet (Fig. S1B and S2). Consequently, the lower Mn of SF did not efficiently form a hydrogel due to a lack of gel formation ability. 3.3. Effect of molecular weight of SF on hydrogel properties Fig. 3A and B presents gel fraction and equilibrium mass swelling ratio, respectively. The hydrogel formed with intact SF (L0) showed about 98% gel fraction while the SF hydrogel of the lowest Mn showed only 18%, indicating most part of SF molecules of low Mn did not participated in gel formation. On the contrary, the
Fig. 4. Shear elastic modulus of SF hydrogels formed of different molecular weights of SF (mean ± SD, n ¼ 3, *p < 0.01).
swelling ratio increased as Mn of SF decreased since the loose network of low Mn SF had larger space to capture more water. Fig. 4 shows equilibrium shear elastic dynamic modulus (G0 ) of SF hydrogels with different Mn. On the whole, the shear elastic modulus increased with Mn of SF. However, the Mn that formed the stiffest SF hydrogel (L10, G0 z 16.7 kPa) was not the highest Mn (~270 kDa) of the intact SF but 200 kDa). It might be due to too rapid gelation of L0 during ultra-sonication, resulting in inhomogeneous network formation. Comparing to 1-min ultra-sonicated L0 hydrogel, gel formation time of 3-min ultra-sonicated L0 was so fast and not sufficient to form the homogeneous gel structure. In addition, longer ultra-sonication treatment time increases the seeding density of SF aqueous solution [27], which consequently lowers the G0 of SF hydrogel (Fig. S3). Interestingly, we found that the hydrolyzed SF solution and hydrogel showed quite different light transmittance with molecular weight of SF. Fig. 5 shows
Fig. 3. (A) Gel fraction and (B) equilibrium swelling ratio of SF hydrogels formed of different molecular weights of SF (mean ± SD, n ¼ 3).
Fig. 5. (A) Images of SF hydrogels formed of different molecular weights of SF. The bottom letters depicts transparency of each hydrogel. (B) Transmittance of SF hydrogels at different wavelength of visible light (mean ± SD, n ¼ 3).
H.H. Kim et al. / Polymer 90 (2016) 26e33
31
Fig. 6. (A) SAXS curves of SF hydrogels with different molecular weights of SF. (B) Guinier plot curves. (C) Radius of gyration (derived from Guinier equation) of SF hydrogels with different molecular weights of SF.
Fig. 7. Schematic of microscopic structures of non-/180 min-hydrolyzed SF hydrogels.
transmittance of SF hydrogel formed with different molecular weights of SF. Generally, SF hydrogel was completely opaque (Fig. 5A, L0) due to the crystalline structure formation [25]. However, transparency of SF hydrogel gradually increased with an increase of hydrolysis time (Fig. 5A). In general, crystallinity is crucial factor affecting the optical property of polymeric materials. However, we could not observe noticeable difference in crystallinities of SF hydrogels by WAXD analysis. It might be due to a high water content (>95%) in hydrogel (Fig. S4). Alternatively, conformational analysis was conducted on the SF hydrogel by thioflavin T assay and FT-IR spectroscopy. As a result, b-sheet content of SF decreased with an increase of hydrolysis time (Figs. S1 and S2), indicating variation of b-sheet content influenced the transparency of SF hydrogel. However, it is not the only factor determining the transparency of hydrogel because hydrogel is basically a composite material of water phase and polymeric network. Therefore, hydrogels can be opaque even if the polymeric phase is completely amorphous. To quantitatively evaluate the optical property of SF hydrogels, the transmittance in the visible light wavelength range (400e700 nm) was measured using a spectrophotometer. As a result, the transmittance increased as the Mn of SF decreased at all wavelengths while L0 showed less than 10% (Fig. 5B). Interestingly,
an increment of transmittance with a decrease of Mn was much higher at a longer wavelength compared with a shorter wavelength. Although L180 exhibited about 80% at 700 nm, the transmittance at 400 nm was lower than 40%. It is noted that such different transmittances of SF hydrogels are closely related to structural cluster which forms physical network structure of polymer chains [28]. Herein, the transparency of SF hydrogel could increase when the cluster size became smaller than visible light wavelength range (400e700 nm), indicating the shorter SF chains (i.e., lower Mn) make smaller clusters. Therefore, it is speculated that the average cluster size of transparent SF hydrogel would be mostly smaller than 400 nm (Fig. 5B). 3.4. Effect of molecular weight on microscopic structure of SF hydrogels Fig. 6A shows SAXS curves of SF hydrogel. For further analysis of basic structure constituting SF hydrogel, the SAXS curves were converted into I(q) at the low q2 region (0.007e0.024) using Guinier equation (Eq. (3)) (Fig. 6B). The radius of gyration (Rg) obtained via Guinier plotting generally represents the gyration of a polymer chain in solution [29,30]. In case of a hydrogel, Rg indicates
32
H.H. Kim et al. / Polymer 90 (2016) 26e33
Fig. 8. (A) Phase-contrast images of hMSC cultured on SF hydrogels of different hydrolysis times 24-h post-seeding (scale: 100 mm). (B) Metabolic activity of hMSC cultured on SF hydrogels of different molecular weights 24-h post-seeding (mean ± SD, n ¼ 3). (C) Confocal images of hMSC on SF hydrogels 24-h post-seeding: (a) L0 and (b) L180 (F-actin: red, nucleus: blue).
the size of the polymer-rich solid-like domains (PRSDs) [29]. The Rg of SF hydrogel was in the range of 9.5e12.6 nm and also tended to increase with Mn of SF increment (Fig. 6C). S. Nagarkar et al. previously reported that aggregates of SF hydrogel made of thin strands of length less than 15 nm [26]. These results correspond to the size of SF hydrogel PRSDs value derived from Guinier equation. Accordingly, we hypothesized that this PRSDs was smallest base structural unit in SF hydrogel. Base on the previous results (Figs. 5 and 6), we depict the microscopic structures of SF hydrogels constituted of unhydrolyzed/hydrolyzed SF molecules, respectively, in Fig. 7. In the non-hydrolyzed (opaque) SF hydrogel (e.g., L0 of the highest Mn), SF chains formed PRSDs of 12e13 nm in diameter. After ultrasonication treatment, those domains were agglomerated together to form single domain network. Then, these single domain networks were further grown to form hydrogel cluster, which so we called aggregated particle connected network or fibril entangled network [26,27]. The size of cluster was larger than visible light wavelength range (400e700 nm) that visible light could not penetrate the hydrogel network structure and scattered at the cluster of hydrogel, consequently looks opaque. On the other hand, in the hydrolyzed (transparent) SF hydrogel (e.g., L180 of the lowest Mn), the fragmented SF chains formed relatively smaller PRSDs in the range of 9e10 nm and they built much smaller cluster of size smaller than visible light wavelength range. Consequently, such a hydrogel finally showed a higher visible light transmittance (Fig. 5). 3.5. Cell adhesion on SF hydrogels To evaluate cell adhesion on SF hydrogels, hMSCs were cultured on SF for 24 h. The hMSCs well attached and spread on SF hydrogels from relatively higher Mn SF hydrogels (i.e., L0, L10, and L30) while the cells maintained spherical shape and less cells were observed on the relatively lower Mn SF hydrogels (i.e., L90 and L180) (Fig. 8A). The metabolic activity of hMSCs cultured on SF hydrogels showed a corresponding decrease as the Mn of SF decreased (Fig. 8B). Fig. 8C depicts F-actin structure of hMSCs on SF 24-h post-seeding. It was shown that F-actin was well developed on the whole cytoplasm of hMSCs on the intact SF hydrogel (L0, Fig. 8Ca) while observed at round boundary near cell membrane (L180, Fig. 8Cb). It is generally known that cell attachment and spreading are governed by the
matrix stiffness [31,32]. As expected, a fewer hMSCs attached on L180 gel surface with less spreading while much more hMSCs remained and spread on the surface of stiffer SF hydrogels (i.e., L0) with well-developed actin cytoskeletons. Hence, it is proved that SF hydrogel properties could be altered in very wide ranges like other synthetic polymer-based hydrogels by molecular weight control of SF. Especially, the improved transparency of depolymerized SF hydrogel will be very useful for cell morphology observation and various biological assays. Also, this technique can expand the optical device application of silk-based materials.
4. Conclusions In conclusion, the gelling retardation of depolymerized SF solution was caused by lack of physical cross-linking formation of hydrolyzed SF molecular chains. The SF hydrogels prepared from depolymerized SF by HAT showed variable physical properties with different molecular weights. Especially, both shear elastic modulus and visible light transmittance could be controlled in very wide ranges (0.4e16.7 kPa and 9.7e79.8%), respectively. Such an alternation in hydrogel property was mainly due to a difference in microscopic structure of SF hydrogel. The shorter SF chains rendered smaller base structural units in hydrogel, resulting in relatively loose network structure with a higher porosity. Consequently, it is expected that the molecular weight control manner could be utilized to enhance silk hydrogel performance as well as other silk-based materials in biomedical engineering application.
Acknowledgement This work was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (NRF-2015R1A2A2A03002680).
Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.polymer.2016.02.054.
H.H. Kim et al. / Polymer 90 (2016) 26e33
References [1] J.S. Ko, K. Yoon, C.S. Ki, H.J. Kim, D.G. Bae, K.H. Lee, Y.H. Park, I.C. Um, Effect of degumming condition on the solution properties and electrospinnablity of regenerated silk solution, Int. J. Biol. Macromol. 55 (2013) 161e168. [2] B. Kundu, N.E. Kurland, S. Bano, C. Patra, F.B. Engel, V.K. Yadavalli, S.C. Kundu, Silk proteins for biomedical applications: Bioengineering perspectives, Prog. Polym. Sci. 39 (2) (2014) 251e267. [3] B. Kundu, R. Rajkhowa, S.C. Kundu, X. Wang, Silk fibroin biomaterials for tissue regenerations, Adv. Drug Deliv. Rev. 65 (4) (2013) 457e470. [4] E.M. Pritchard, D.L. Kaplan, Silk fibroin biomaterials for controlled release drug delivery, Exp. Opin. Drug Deliv. 8 (6) (2011) 797e811. [5] A.B. Mathur, V. Gupta, Silk fibroin-derived nanoparticles for biomedical applications, Nanomedicine 5 (5) (2010) 807e820. [6] C.S. Ki, Y.H. Park, H.-J. Jin, Silk protein as a fascinating biomedical polymer: structural fundamentals and applications, Macromol. Res. 17 (12) (2009) 935e942. [7] T. Yucel, P. Cebe, D.L. Kaplan, Vortex-induced injectable silk fibroin hydrogels, Biophys. J. 97 (7) (2009) 2044e2050. [8] N. Kasoju, U. Bora, Silk fibroin in tissue engineering, Adv. Healthc. Mater 1 (4) (2012) 393e412. [9] Z.-H. Li, S.-C. Ji, Y.-Z. Wang, X.-C. Shen, H. Liang, Silk fibroin-based scaffolds for tissue engineering, Front. Mater. Sci. 7 (3) (2013) 237e247. [10] C. Vepari, D.L. Kaplan, Silk as a biomaterial, Prog. Polym. Sci. 32 (8) (2007) 991e1007. [11] N.E. Davis, L.N. Beenken-Rothkopf, A. Mirsoian, N. Kojic, D.L. Kaplan, A.E. Barron, M.J. Fontaine, Enhanced function of pancreatic islets coencapsulated with ECM proteins and mesenchymal stromal cells in a silk hydrogel, Biomaterials 33 (28) (2012) 6691e6697. [12] X. Wang, J.A. Kluge, G.G. Leisk, D.L. Kaplan, Sonication-induced gelation of silk fibroin for cell encapsulation, Biomaterials 29 (8) (2008) 1054e1064. [13] M.J. Mahoney, K.S. Anseth, Three-dimensional growth and function of neural tissue in degradable polyethylene glycol hydrogels, Biomaterials 27 (10) (2006) 2265e2274. [14] C.M. Hassan, N.A. Peppas, Structure and applications of poly (vinyl alcohol) hydrogels produced by conventional crosslinking or by freezing/thawing methods, Adv. Polym. Sci. (2000) 37e65. [15] U.-J. Kim, J. Park, C. Li, H.-J. Jin, R. Valluzzi, D.L. Kaplan, Structure and properties of silk hydrogels, Biomacromolecules 5 (3) (2004) 786e792. [16] K. Numata, T. Katashima, T. Sakai, State of water, molecular structure, and cytotoxicity of silk hydrogels, Biomacromolecules 12 (6) (2011) 2137e2144. [17] I. Noh, G.-W. Kim, Y.-J. Choi, M.-S. Kim, Y. Park, K.-B. Lee, I.-S. Kim, S.-J. Hwang, G. Tae, Effects of cross-linking molecular weights in a hyaluronic acid -
[18]
[19]
[20]
[21]
[22] [23]
[24] [25]
[26] [27]
[28]
[29]
[30]
[31] [32]
33
poly(ethylene oxide) hydrogel network on its properties, Biomed. Mater 1 (3) (2006) 116e123. Y.L. Zhang, C.C. Chu, The effect of molecular weight of biodegradable hydrogel components on indomethacin release from dextran and poly(DL)lactic acid based hydrogels, J. Bioact. Compat. Polym. 17 (2) (2002) 65e85. R. Fukae, M. Yoshimura, T. Yamamoto, K. Nishinari, Effect of stereoregularity and molecular weight on the mechanical properties of poly(vinyl alcohol) hydrogel, J. Appl. Polym. Sci. 120 (1) (2011) 573e578. H.Y. Zhou, X.G. Chen, M. Kong, C.S. Liu, D.S. Cha, J.F. Kennedy, Effect of molecular weight and degree of chitosan deacetylation on the preparation and characteristics of chitosan thermosensitive hydrogel as a delivery system, Carbohydr. Polym. 73 (2) (2008) 265e273. H.J. Kong, D. Kaigler, K. Kim, D.J. Mooney, Controlling rigidity and degradation of alginate hydrogels via molecular weight distribution, Biomacromolecules 5 (5) (2004) 1720e1727. M. Li, M. Ogiso, N. Minoura, Enzymatic degradation behavior of porous silk fibroin sheets, Biomaterials 24 (2) (2003) 357e365. H. Zhang, L. Deng, M. Yang, S. Min, L. Yang, L. Zhu, Enhancing effect of glycerol on the tensile properties of bombyx mori cocoon sericin films, Int. J. Mol. Sci. 12 (5) (2011) 3170. G.G. Leisk, T.J. Lo, T. Yucel, Q. Lu, D.L. Kaplan, Electrogelation for protein adhesives, Adv. Mater 22 (6) (2010) 711e715. A. Matsumoto, J. Chen, A.L. Collette, U.-J. Kim, G.H. Altman, P. Cebe, D.L. Kaplan, Mechanisms of silk fibroin sol-gel transitions, J. Pyhs. Chem. B 110 (43) (2006) 21630e21638. S. Nagarkar, T. Nicolai, C. Chassenieux, A. Lele, Structure and gelation mechanism of silk hydrogels, Phys. Chem. Chem. Phys. 12 (15) (2010) 3834e3844. H.J. Cho, C.S. Ki, H. Oh, K.H. Lee, I.C. Um, Molecular weight distribution and solution properties of silk fibroins with different dissolution conditions, Int. J. Biol. Macromol. 51 (3) (2012) 336e341. L. Yu, Z. Zhang, J. Ding, In vitro degradation and protein release of transparent and opaque physical hydrogels of block copolymers at body temperature, Macromol. Res. 20 (3) (2012) 234e243. L. Wang, G. Shan, P. Pan, A strong and tough interpenetrating network hydrogel with ultrahigh compression resistance, Soft Matter 10 (21) (2014) 3850e3856. L. Zhao, L. Li, G.-Q. Liu, X.-X. Liu, B. Li, Effect of frozen storage on molecular weight, size distribution and conformation of gluten by SAXS and SEC-MALLS, Molecules 17 (6) (2012) 7169e7182. R.G. Wells, The role of matrix stiffness in regulating cell behavior, Hepatology 47 (4) (2008) 1394e1400. R.G. Breuls, T.U. Jiya, T.H. Smit, Scaffold stiffness influences cell behavior: opportunities for skeletal tissue engineering, Open Orthop. J. 2 (2008) 103.