Biomaterials from Ultrasonication-Induced Silk FibroinâHyaluronic

3178

Biomacromolecules 2010, 11, 3178–3188

Biomaterials from Ultrasonication-Induced Silk Fibroin-Hyaluronic Acid Hydrogels Xiao Hu,†,‡ Qiang Lu,† Lin Sun,§ Peggy Cebe,‡ Xiaoqin Wang,† Xiaohui Zhang,† and David L. Kaplan*,† Departments of Biomedical Engineering, Physics and Astronomy, and Chemical and Biological Engineering, Tufts University, Medford, Massachusetts 02155, United States Received September 5, 2010; Revised Manuscript Received September 19, 2010

We report formation of biocompatible hydrogels using physically cross-linked biopolymers. Gelation of silk fibroin (from B. mori silkworm) aqueous solution was effected by ultrasonication and used to entrap blended, un-crosslinked, hyaluronic acid (HA) without chemical cross-linking. HA was formed into silk/HA blended hydrogels with different mixing ratios, forming homogeneous materials with stable swelling behavior when the HA content was less than 40 wt %. This is a novel approach to HA hydrogel systems, which otherwise require chemical cross-linking. Further, these systems exploit the beneficial material and biological properties of both polymers. Differential scanning calorimetry (DSC), temperature modulated DSC, and thermal gravimetric analysis were used to show that well-blended silk/HA hydrogel systems formed without macrophase separation. Fourier transform infrared spectroscopy was used to determine secondary structures from the amide I region of silk protein by spectral subtraction and Fourier-self-deconvolution. The β-sheet crystal fraction of the silk protein increased with increase of HA content (26-35 wt %), which resulted in stable, crystalline features in the blend hydrogel materials, favorable features to support human mesenchymal stem cell attachment and proliferation. Scanning electron microscopy was used to characterize morphology. β-Sheet content controlled the stability of the silk/HA hydrogel systems, with a minimum crystalline content needed to maintain a stable hydrogel system of ∼26 wt %. This value is close to the β-sheet content in pure silk fibroin hydrogels. These novel nonchemically cross-linked blend hydrogels may be useful for biomedical applications due to biocompatibility and the widespread utility of hydrogel systems. The attributes of HA in combination with the features of silk, offer a useful suite of properties, combining the mechanical integrity and slow degradation of silk with the control of water interactions and biological signaling of HA.

Introduction Hydrogels are insoluble three-dimensional polymer chain networks that swell in aqueous solutions and that can hold or entrap liquid components.1-13 Through the balance between the osmotic force and the entropic retroactive force in the network, hydrogels exhibit solid-like mechanical behavior, with high compliance and high elastic-strain, while consisting mostly of liquid.1-13 The formation of hydrogel structures is due to the connectivity of the polymer chains as a result of “crosslinking”.1-13 There are two main types of cross-links in these types of hydrogel systems: chemical and physical.1-13 In chemically cross-linked hydrogels, networks are formed by chemical reactions or polymerization to stitch together the starting materials (such as monomers or polymers) via crosslinkers.5 The polymerization of hydrogels is often induced by irradiation with UV light, γ-rays, X-rays, or electron beams.1-13 In addition, unpurified residues, such as reactive agents, polymerized clusters, and nonterminated growing chains may remain in chemically cross-linked hydrogels after purification,1-13 which can lead to complications for biomedical applications. In contrast, physically cross-linked hydrogels avoid these complications. Crystallization, liquid-liquid phase separation, ionic interactions, hydrogen bonding, or even topological crosslinking can all be used to form the cross-linker regions in these * Corresponding author: [email protected]. † Department of Biomedical Engineering. ‡ Department of Physics and Astronomy. § Department of Chemical and Biological Engineering.

hydrogels.1-13 Therefore, physically cross-linked hydrogels are attractive for entrapping sensitive molecules, such as bioactive compounds like cytokines or cells, because toxic reactive molecules are avoided in the cross-linking process.1-13 Silk fibroin proteins are biosynthesized in the glands of domesticated silkworm (Bombyx mori).14-21 The highly repetitive primary sequence GAGAGS forms antiparallel β-sheet crystalline regions. Exposure to heat, physical shear, or some organic solvents16,17,14-30 can induce the formation of insoluble crystallized structures.14-30 Silk fibroin aqueous solutions form hydrogels directly through self-assembly, with the rate of this sol-gel transition dependent on protein concentration, temperature, metal ions, and pH.31,32 The mechanism of gelation is the self-assembly of the protein chains into physically crosslinked β-sheet crystals.32-34 Silk hydrogels have been used for biomedical applications because of their biocompatibility and adjustable mechanical properties.31-37 Recently, the encapsulation of compounds in silk hydrogel systems has been studied. For example, genipin was used to chemically cross-link chitosan with silk fibroin for cartilage tissue engineering.38 Genetically engineered silk-elastin-like protein hydrogels for tissue engineering and drug delivery have been studied for drug delivery and other applications.39,40 Cannas et al. coated polyelectrolyte modified HEMA (2hydroxyethyl methacrylate) hydrogels on silk fibroin for biomedical applications.41 In a previous study34 we developed an ultrasonication method to accelerate the sol-gel transition of silk by inducing rapid physical cross-links via β-sheet crystal

10.1021/bm1010504  2010 American Chemical Society Published on Web 10/13/2010

Biomaterials from Silk Fibroin/HA Hydrogels

formation. This method is controllable without the addition of chemical processes, based on the input of energy to the system from the sonication method. Mesenchymal stem cells within these gels retained viability and proliferated in culture conditions, which illustrate the biocompatible nature of the process.34 HA (hyaluronic acid) is an important glycosaminoglycan (GAG) in cartilage42-48 and is also found in many other tissues in the body. It provides the swelling pressure for cartilage function, which has a high capacity of lubrication, water sorption, and water retention.42-48 HA is a naturally occurring linear polysaccharide, which is comprised of β-1,4-linked D-glucuronic acid and β-1,3 N-acetyl-D-glucosamine disaccharide with a 1:1 ratio in the repeating units.42-48 HA is present widely in the extracellular matrix of higher animals and plays critical roles in cell differentiation and cell motility.42-48 Due to the important functions and applications of HA, it has been widely used in biomedical applications such as in scaffolds for wound healing, ophthalmic surgery, arthritis treatments, and as a component in implant materials.45-58 Many approaches have been developed to form the HA hydrogels or modified hydrogels.45-54 However, most HA-based hydrogels are chemically synthesized, which may result in residual components less attractive to bioactivity, or limit the range of components that can be encapsulated if the reaction conditions also effect the bioactive components of interest.45-54 The goal of the present study was to develop a new physical method that can be used to entrap other biopolymers with silk during the sol-gel transition. The ability to generate such composite hydrogels would expand the range of properties and utility for such systems in many areas of medical applications. The focus of the present study was on hyaluronic acid (HA) or hyaluronan. The attributes of HA in combination with the features of silk would offer a useful suite of properties, combining the mechanical integrity and slow degradation of silk with the control of water interactions and biological signaling of HA. This combination of material and biological properties would expand these hydrogel systems related to needs with biomaterials and tissue scaffolding. Further, the ability to combine these two polymers without the need for chemical cross-linking, simplifies materials preparation, avoids residual compounds that could cause undesirable biological impact, and provides direct control of material properties to correlate to biological responses, as reported here.

Experimental Section Materials. Silk fibroin preparation has been reported previously.14,22-26 Briefly, B. mori silkworm cocoons were boiled in a 0.02 M Na2CO3 solution to extract the glue-like sericin proteins.14 The remaining silk fibroin was dissolved in a 9.3 M LiBr solution at 60 °C for 4-6 h and then dialyzed with distilled water using dialysis cassettes for 2 days. After centrifugation and filtration to remove insoluble residues, a 6 wt % silk fibroin aqueous solution was obtained. Sodium hyaluronate powder was kindly supplied by Genzyme Corp. (Cambridge, MA) with a viscosity 15 dL/g. For comparison, a sodium HA powder with a purity >95% was also purchased from Acros Organics Corporation (Belgium). Both preparations showed similar properties in terms thermal and spectral analysis. The HA powders were first dissolved in distilled water to form a 1.0 wt % HA aqueous solution, then slowly mixed with the silk fibroin aqueous solution using a pipet to avoid protein aggregation during fast mixing in solution. The final solutions obtained were based on a mass ratio of silk/HA ) 95:5 (SH95), 90:10 (SH90), 80:20 (SH80), 66:33 (SH66), 50:50 (SH50), 33:66 (SH33), with pure silk (SH100) and pure HA (SH0) used as controls. The hydrogels generated were placed prefrozen in liquid

Biomacromolecules, Vol. 11, No. 11, 2010

3179

nitrogen and immediately lyophilized in -80 °C to avoid structural changes. The lyophilized hydrogels and aggregates of silk/HA were used for swelling and structural analysis. Ultrasonication. A Branson 450 Sonifier (Branson Ultrasonics Co., Danbury, CT), which consisted of a power supply (Model 450), a converter, an externally threaded disruptor horn, and a 1/8” (3.175 mm) diameter-tapered microtip, was used to sonicate the silk/HA solutions and induce gelation. A total of 1 mL of silk/HA (water) solutions (∼3 wt %) with varied weight ratios (SH100, SH95, SH90, SH80, SH66, SH50, SH33, and SH0. The number after SH means the weight percentage of silk fibroin in the silk/HA blends.) were each loaded in a 5 mL tube and were ultrasound sonicated twice through the microtip at a 20% amplitude for 20 s with specific power setting. Solutions were incubated at 37 °C after sonication and the sol-gel transition was monitored. It has been reported that several physical factors such as local temperature increase, mechanical shear, and increased air-liquid interfaces could induce cross-linking between polymer/protein chains and affect the process of gelation. For each solution, the temperature T increases as a function of time during ultrasonic propagation in silk protein/HA solution was measured. Based on the temperature versus time data, the absolute ultrasonic power P is given as59-62

P ) mcp · (dT/dt)t)0

(1a)

where cp is the heat capacity of the solution (protein/water) and m is the total mass of solution. The intensity of the ultrasonic power dissipated from the probe microtip (with a radius of r) is thus given by59,60

I ) P/(πr2) ) [mcp · (dT/dt)t)0]/(πr2)

(1b)

Therefore, the temperature increase (dT; energy input) in a certain polymer/protein solution with constant total heat capacity (mcp) can be directly controlled by the vibration intensity (I) and time chosen during sonication (dt), which results in the formation of hydrogel crosslinks. Differential Scanning Calorimetry (DSC). The dried silk/HA scaffolds (each about 5 mg) were encapsulated in Al pans and heated in a TA Instruments (New Castle, DE) Q100 DSC, with purged dry nitrogen gas flow (50 mL/min), and equipped with a refrigerated cooling system. The instrument was calibrated with indium for heat flow and temperature. Standard mode DSC measurements were performed at a heating rate of 2 K/min. Temperature-modulated differential scanning calorimetry (TMDSC) measurements were also performed at a heating rate of 2 K/min with a modulation period of 60 s and temperature amplitude of 0.318 K. Aluminum and sapphire reference standards were used for calibration of the heat capacity. The heat capacity measurements consisted of three runs, as described in our earlier work.24-27 In TMDSC, the “reversing heat capacity”, which represents a reversed heat effect within the temperature range of the modulation, were measured and calculated.24-27,63 Thermal Gravimetric Analysis. Thermal gravimetric Analysis (TGA, TA Instruments Q500) was used to measure changes in weight of silk/HA samples with increasing temperature. TGA curves were obtained under nitrogen atmosphere with a gas flow of 50 mL/min. The instrument was run at a heating rate of 2 K/min. Fourier Transform Infrared Spectroscopy (FTIR). FTIR analysis of Silk/HA scaffold samples was performed with a Jasco (Japan) FT/ IR-6200 Spectrometer, equipped with a deuterated triglycine sulfate detector and a multiple reflection, horizontal MIRacle ATR attachment (using a Ge crystal from Pike Tech (Madison, WI)). The instrument was continuously purged by nitrogen gas to eliminate the spectral contributions of atmospheric water vapor. For each measurement, 128 scans were coadded with resolution 4 cm-1, and the wavenumber ranged from 400-4000 cm-1.

3180


Hu et al.

Figure 1. Change in swelling ratio of different silk/HA hydrogels with time. The swelling times are shown in a log scale. The number after SH means the weight percentage of silk fibroin in the silk/HA blends, e.g., SH80 is the sample with silk (wt)/HA (wt) ) 80:20. Each data point is averaged from three samples with error bars as shown in the figure, when larger than the size of data symbols.

Fourier self-deconvolution (FSD) of the infrared spectra covering the amide I region (1595-1705 cm-1) was performed with Opus 5.0 software from Bruker Optics Corp. (Billerica MA) as we have described previously.24 Generally, Fourier self-deconvolution (FSD) is a signalprocessing tool that allows resolution of a band contours comprising several overlapping bands.24,64-68 Using a high pass filter, the broad and indistinct amide I bands (CdO stretching bonds in protein backbones) can be narrowed synthetically to provide a deconvoluted spectrum with better peak resolution.64-68 The deconvoluted spectra are better suited for subsequent Gaussian curve-fitting.64-68 Finally, the deconvoluted amide I spectra were area-normalized, and the relative areas of the single bands were used to determine the fraction of the secondary structural elements. Scanning Electron Microscopy (SEM). The surface morphologies of different silk/HA hydrogel scaffold samples were imaged using a Supra 55 VP SEM (Zeiss Corp.). Samples were first fractured in liquid nitrogen and sputtered with platinum. Cross-section images were then investigated. Swelling Ratio. The hydrogels were first freeze-dried to form hydrogel scaffolds (SH100, SH95, SH90, SH80, SH66, and SH50) and then approximately 2 g of the dried hydrogel samples was immersed into 40 mL of aqueous solution. Different time intervals were selected to measure the weight of the hydrated hydrogels after swelling. During weight measurements, the hydrogels were first removed from the aqueous solution and surface water was removed on blotting paper. The swelling ratio, S, was calculated by using the equation1-13

S ) [(Wt - W0)/W0]

tion with 70% ethanol solution for 48 h, the samples were preconditioned with cell culture medium overnight before cell plating. The sterilization treatment did not alter the secondary structure of the samples based on FTIR analysis. hMSCs were trypsinized and replated on the silk films at a density of 5000 cells/well. Cell attachment and spreading on the samples was observed after 3 h incubation in 5% CO2 at 37 °C. Images were captured by a phase contrast microscope with a TCS SP2 scanner (Leica Microsystems, Manheim/Wetzlar, Germany). The proliferation of hMSCs on the silk/ HA samples was evaluated using alamarBlue assay. AlamarBlue (BioSource International, Inc., Camarillo, CA) is a nontoxic dye that measures the metabolic activity of cells by fluorometric analysis.22 High levels of metabolic activity indicate cellular viability and proliferation, and low levels imply cellular toxicity.22 At each predetermined time point, the cells in culture were incubated with 100 µL 10% alamarBlue in culture medium at 37 °C for 1 h, followed by spectrofluorimetric analysis by excitation at 540 nm and measuring fluorescence at 590 nm in a SpectraMax/Gemini Em fluorescence microplate reader (Molecular Devices, Sunnyvale, CA). A set of empty wells with alamarBlue culture medium was used as background and controls. Statistical Analysis. Statistical differences were determined by a Mann-Whitney U test (Independent t test, SPSS). Statistical significance was assigned as **p < 0.01 and *p < 0.05, respectively.

(2)

where W0 is the initial weight of the dried hydrogel scaffold at time t ) 0, and Wt is the weight of the hydrated hydrogel at selected time, t. Each data point (in Figure 1) was obtained by averaging values from three samples under the same conditions, with error bars shown. Cell Attachment and Proliferation. Human mesenchymal stem cells (hMSCs) were obtained from bone marrow aspirates (Cambrex Bio Science Walkersville Inc., Walkersville, MD) of a 25 year old healthy male following procedures we have previously described.13 hMSCs were separated from hematopoietic stem cells (HSCs) on the basis of their adherence to tissue culture plastic. Cells were maintained in a humidified incubator at 37 °C with 5% CO2, with medium replenished every 4 days. Trypsin (0.25%)/1 mM ethylenediaminetetraacetic acid (EDTA) were used to harvest or passage hMSCs, and cells at passage 2-4 were used for all the experiments. For the stem cell attachment experiments, a thin layer (∼500 µm) of the different silk/HA hydrogels (SH100, SH95, SH90, SH80, SH66; n ) 8 for each set) were dried on 24-well, tissue culture treated plates. After steriliza-

Figure 2. (a) Standard DSC scans of the silk/HA hydrogels (SH100, SH95, SH90, SH80, SH66, SH50) and controls (SH0 and SH33) without gelation. The samples were heated at 2 °C/min from -30 to 290 °C. (b) Reversing heat capacities of the hydrogels (SH100, SH95, SH90, SH80, SH66, SH50) and controls (SH0 and SH33) from -30 to 260 °C, measured by TMDSC with a 2 °C/min heating rate, a modulation period of 60 s, and a temperature amplitude of 0.318 K.



3181

Figure 3. (a) Weight percentage change of the silk/HA hydrogels (SH100, SH95, SH90, SH80, SH66, SH50) and controls (SH0 and SH33) measured by TGA during heating from room temperature to 450 °C at 2 °C/min. (b) First derivative of the weight percentage function in Figure 4a.

Results and Discussion Swelling Behavior. The swelling behavior of the silk/HA blend hydrogels was determined in aqueous solution. The swelling ratios of the silk/HA hydrogels with different swelling times (5, 20, 40, 100, 360 (6 h), 1140 (19 h), 7200 (120 h), and 11520 min (8 days)) are plotted in Figure 1. The trend of swelling ratio with time was plotted on a log scale to show long times for the swelling behavior. At short swelling times (5 min), most of the hydrogels absorbed more than five times their weight of water, thus, the swelling ratios increased rapidly during the first 40 h. After this, the increase in swelling ratio became more stable from days 2 to 8, with only slight increases thereafter. The swelling ability of silk/HA hydrogels increased with an increase in HA content. A pure silk fibroin hydrogel only swelled about 15 times over 8 days. However, with 33% HA in the scaffold, sample SH66 swelled more than 30 times over the same time frame.24 These results indicate that HA improves the swelling ability of the silk/HA hydrogels. However, a high amount of HA may also reduce the stability of the hydrogel, as HA is not chemically cross-linked with the silk fibroin chains. An example is the SH50 hydrogel sample. Although it formed hydrogels during sonication, the swelling ratio of this hydrogel decreased with time because of the dissolution of the hydrogel structure. Therefore, silk/HA hydrogels with more than 40% HA tend to be not stable, due to the lower β-sheet content in the structures. The stability of HA in the 8 day swelling silk/HA hydrogels was also examined by FTIR after freeze-drying. No significant spectral differences were observed for the samples before and after swelling, which indicates that the silk/HA blend hydrogels were stable in aqueous solution.

Figure 4. FTIR absorbance spectra of silk/HA hydrogels (SH100, SH95, SH90, SH80, SH66, and SH50) and controls (SH33 and SH100) for (a) 800-1800 cm -1 and (b) 2700-3750 cm -1.

DSC Analysis. The thermal properties of the dried silk/HA hydrogels were examined by standard DSC and temperature modulated DSC (TMDSC). Figure 2a shows standard DSC scans of the silk or silk/HA hydrogels (SH100, SH95, SH90, SH80, SH66, SH50) and the aggregate samples (SH33, SH0). All of these samples showed water peaks around 80 °C, similar to silk-bound water films studied previously.25,26 After the glass transition temperature regions of the silk/HA samples, a strong nonisothermal crystallization peak appeared around 210∼230 °C and shifted from 228 °C (SH100) to 212 °C (SH0), which is related to the thermal transition of the silk-HA blend.24-26,56 No separate individual transition peak for the HA (228 °C) or silk (212 °C) components was observed in the blends, which indicates that silk and HA were well blended to a homogeneous system without macrophase separation. After crystallization, a degradation peak between 260 and 275 °C was shown for all silk containing samples. The pure HA (SH0), however, did not show any degradation peak along the baseline until 300 °C. To fully understand the phase transitions of the Silk/HA samples, temperature modulated DSC (TMDSC) was used to measure the reversing thermal properties of the silk/HA samples. Figure 2b shows the reversing heat capacity of the samples from -30 to 260 °C. All samples showed an endothermic peak around 50-70 °C, with slight shift from 55 °C in SH0 to 70 °C in

3182


Hu et al.

Figure 5. (a) Subtraction strategy to obtain the spectrum of silk in silk/HA blends. The vibration group of alcohol group in HA between 900 and 1200 cm-1 is fit based on the zero baseline between 2000 and 1750 cm-1. The remaining spectrum after the subtraction is mainly from the vibrational contribution of silk. (b) The spectrum of amide I, amide II, and amide III regions of the silk component from all scaffold samples after the HA subtraction. (c) The Fourier self-deconvoluted Amide I region of the silk fibroin components spectra in Figure 5b. The peaks are assigned and divided as side chain (S), β-sheet (B), random coil (R), R-helix (A), and turn (T) regions.24 (d) A curve fitting example of FSD amide I spectra using the SH100 sample. The fitted peaks are shown by dashed lines.

SH100. These downward shift curves in the reversing heat capacity traces were due to bound water removal in the system.26 Above 130 °C, the bound water was completely removed and the baseline for heat capacity became stable. A clear glass transition appeared for all samples. With the increase of HA content, the glass transition temperatures (Tg) of the hydrogels increased gradually from 180 to 210 °C, but the heat capacity steps during the Tg, ∆Cp, decreased rapidly. The homogeneous glass transition regions for the silk/HA hydrogels indicate a stable silk/HA macrophase mixing during gelation. The glass transitions of the samples SH33 and SH0 occurred at higher temperatures (>230 °C) and were partially obscured by the degradation processes, so they could not be clearly identified in the reversing heat capacity traces. TGA Analysis. To assess the thermal degradation of the silk/ HA hydrogels, TGA was performed on samples SH100, SH95, SH90, SH80, SH66, SH50, with pure HA (SH0) and SH33 as controls. Figure 3a shows the weight percentage change during heating from room temperature to 450 °C. Figure 3b shows the first derivative of the weight percentage function in each sample,

which reveals the degradation rates and middle degradation temperatures of each component in the silk/HA hydrogels. During the initial heating between room temperature and 120 °C, bound water was removed (evaporated) from all samples as shown in the DSC study. The pure HA sample (SH0) showed the largest amount of the bound water (about 10 wt %). With the addition of the silk, the percentage of bound water decreased, with 9 wt % in SH50, 7.5 wt % for SH66, 5.5 wt % for SH80 and SH90, and 5 wt % for the pure silk hydrogel (SH100). The samples showed stable weights in the temperature region from 120 to about 210 °C. Above 210 °C, thermal degradation started. The pure HA sample showed the most rapid weight loss (Figure 3a). A sharp peak around 210 °C was shown in the first derivative of weight loss in Figure 3b, and then a small shoulder appeared around 225 °C. After 250 °C, sample degradation became homogeneous and no peaks appeared in the firstderivative. With increasing silk content, the stability of the hydrogels improved and the degradation process became slower with flatter slopes (Figure 3a). The first HA degradation derivative peak completely disappeared (Figure 3b) for samples



3183

Table 1. Percentage of Different Secondary Structuresa in the Silk with Change of HA Content in the Hydrogels or Aggregate Controlsa silk-HA samples

silk in blend system (wt%)

β-sheet (B) in silk (%)

R-helix and random coils (A + R) in silk (%)

turns (T) in silk (%)

side chains (S) in silk (%)

crystallinity in total blend system (%)

SH100 SH95 SH90 SH80 SH66 SH50 SH33 SH0

100 95 90 80 66 50 33 0

25.9 35.8 37.3 41.8 45.6 48.0 49.4 N/A

41.1 32.2 31.2 30.7 27.5 26.5 25.8 N/A

31.2 31.4 28.6 25.2 24.8 23.2 22.7 N/A

1.8 6.4 2.9 2.3 2.1 2.2 2.0 N/A

25.9b 34.1c 33.6c 33.4c 30.1c 24.0d 16.3e 0e

a All numbers have same unit (wt%) with a (1.5% error bar. aggregates that serve as control samples.

b

Pure silk hydrogel.

Figure 6. Percentage change of different secondary structures in silk with change of HA content in silk/HA hydrogels (left vertical axis, in black symbols); the β-sheet crystal fraction to the total with change of HA content (right vertical axis, in red symbols). Numerical values of the left and right vertical axes are the same. The red dashed line demonstrates that only hydrogels with a beta-sheet content in the blend systems higher than pure silk hydrogel (∼25%, the red point with vertical double-headed arrow) are stable.

SH50, SH66, SH80, SH90, and SH95. Instead, the shoulder became a peak around 240∼250 °C, decreased and shifted gradually with increase of silk content. These results indicate that thermal stability of HA chains can be improved by the addition of the silk protein molecules. After the first peak of the HA component, a strong degradation derivative peak from silk followed at around 260∼290 °C. And this peak shifted to high temperature with increase of silk content. A derivative peak around 320 °C was also shown for pure silk hydrogel (SH100) and the sample SH95. However, this peak was not clear in other hydrogel blends. Degradation became homogeneous for all samples above 400 °C (Figure 3b). The TGA results demonstrated that the silk and HA components were well blended in the hydrogel systems and, therefore, did not follow their original individual thermal stability profiles, as seen in the control samples SH100 and SH0. FTIR Analysis. FTIR was performed to measure the structural changes of the silk/HA hydrogels during blending and gelation. Figure 4a,b show the FTIR absorbance spectra of the silk/HA hydrogels (SH100, SH95, SH90, SH80, SH66, and SH50) for the wavenumber region (a) 800-1800 cm-1 and (b) 2700-3700 cm-1, along with the aggregate samples SH33 and SH100. The FTIR spectra of HA has been well studied.44,53,55 Generally, the intense band between 950 and 1100 cm-1 is assigned to C-O stretching vibrations (νC-OH) in alcohols, including alcohol I (1010 cm-1), alcohol II (1030-1080 cm-1), and some νC-O-C ring vibration mode (1090 cm-1).44,53,55 The shoulder around 1145-1165 cm-1 corresponds to the antisym-

c

Stable silk/HA hydrogel.

d

Unstable silk/HA hydrogel.

e

Silk/HA

metric νC-O-C vibration in glycosidic groups.44 The vibrational νC-O modes of carboxyl COOH or COO- group for hyaluronic acid bands are around 1220-1415 cm-1.44 The strong band between 1500-1600 cm-1 arises from amide II vibrational mode, mainly contributed from δN-H of amide N-H group. The intense group of bands that extends from 1600 to 1700 cm-1 is the amide I band in HA,44 which is mainly from various carbonyl and carboxyl νCdO stretching vibrations. After a long linear baseline without any vibration in the HA structure, a band from C-H stretching vibration was observed around 2900 cm-1.44 Then a broad and intense band that extends from 3000 to 3600 cm-1 appeared, which belongs to stretching vibration modes of N-H (3100-3275 cm-1) and O-H (3300-3600 cm-1) groups engaged in the hydrogen bonds of the HA structure.44 We have previously described the FTIR spectrum for pure silk.24,26 The IR spectral region within 1700-1500 cm-1 was assigned to the peptide backbone of amide I (1700-1600 cm-1) and amide II (1600-1500 cm-1) absorptions, and the amide III region was from 1350 to 1200 cm-1.24,26 The amide I region mainly comes from the CdO stretching vibration (>80%).24,26 Therefore, the amide I vibration directly depends on the secondary structure of the protein backbone and is most commonly used for the quantitative analysis of different secondary structures. The out-of-phase combination of the CN stretching and the NH in-plane bending vibrations results in the amide II bands.26,65 The microenvironment and the conformation of protein side chain groups can easily affect the amide II region.26,65 The amide II′ mode (1490-1460 cm-1) is a larger CN stretching vibration mode converted by N-deuteration.26,65 The fingerprint region26,65 was located within 1330-1000 cm-1 with low absorbance intensity. The peak at 1437 cm-1 corresponds to the CH2 bending band of protein (methylene scissoring mode); the peak near 1405 cm-1 corresponds to the major deformation of CH2 and CH3 group and the minor COOsymmetric stretching band.26,65 The peaks ranging from 1350 to 1200 cm-1 are associated with NH bending vibration modes of amide III band.26,65 In the study of hydrogen bonds in the polymer IR spectrum, the amide A region (3310-3270 cm-1) is also considered when analyzing the N-H stretching vibration of materials.26,65 The amide A region is part of the Fermi resonance doublet connected with its second component amide B, which absorbs weakly in 3100-3030 cm-1.26,65 For the spectra of silk/HA hydrogels, with the increase of HA content, the broad peak including the alcohol group (900-1100 cm-1) and the shoulder from C-O-C vibration in glycosidic groups (around 1145 cm-1) increased significantly. Because the silk fibroin protein does not have any strong vibration mode within this region, this region can be set as an indicator of peaks of HA components in the silk/HA blends, which has shown no peak shift in the different samples. The

3184


Hu et al.

Figure 7. SEMs of the silk/HA hydrogels: (a) SH100; (b) SH95, (c) SH90, (d) SH80, (e) SH66, and (f) SH50, and non-cross-linked pure HA (SH0) in (g). Each sample is shown in three scales (scale bars are 50, 10, and 1 µm, respectively, in moving from left to right with the images). The green box shows the observed region of the next image to the right.

amide I and II regions (1700-1450 cm-1) of silk and HA components strongly overlapped (Figure 4a) because silk and HA both have vibration modes in this region. However, with the increase in silk content, the amide II region increased significantly, mostly contributed from the side chain vibrations of the silk protein structure. The amides A and B from N-H stretching in 3100-3300 cm-1 did not provide clear information, as both HA and silk have strong vibrations in these regions. In the amide I region, the main peak position shifted from 1640 to 1620 cm-1 with the decrease of silk content. As mentioned previously,33,34 the β-sheet crystals in silk play a key role for silk to form hydrogel networks. Thus, it is important to obtain

the range of secondary structures (including β-sheet crystals) in silk during the change of HA content. Therefore, a strategy was used here to subtract the HA component spectra from the total overlapped hydrogel spectra, which is similar to the background subtraction technique used in most sample/air or sample/solution studies. Figure 5a shows the subtraction strategy using the sample SH50 and the pure HA sample (SH0) spectra as an example. The vibration group from HA between 900 and 1200 cm-1 did not change during the physical blending and it is therefore proportional to the HA percentage in the hydrogels. Considering the zero baseline between 2000 and 1750 cm-1 for all samples, we can fit the spectrum of alcohol groups in


pure HA to each silk/HA blend spectrum, and the remaining spectrum after the subtraction is mainly from the vibrational contribution of the silk. This method assumed that the silk chains had no strong effect on the HA vibration mode during physical hydrogel blending, which has been shown in the unchanged C-OH stretching vibrations in alcohol groups and, therefore, is reasonable for the CdO stretching vibrations in amide I region. Figure 5b shows the amide I, II, and III regions of the silk component of the spectra from all hydrogel samples after HA spectrum subtraction. The β-sheet crystal peak (around 1620 cm-1) increased with increase of HA content, as seen in the amide I region. The amide II and III regions, however, remained stable for all spectra. To quantify the percentage of the secondary structures in the silk component, FSD was performed in the amide I region. Figure 5c shows the FSD deconvoluted amide I spectra of the silk components in Figure 5b, and a curve fitting example from silk/HA (100/0) is shown in Figure 5d. In Figure 5c, the dash lines show that the deconvoluted peaks are equal and located in the same region of the spectra (wave numbers) for all silk spectra when compared to the pure silk spectrum SH100. This result indicates that the subtractions were performed successfully without adding or missing any vibrational mode in silk spectra. The peak position and their related secondary structures have been detailed in our previous work24,26 and assigned in Figure 5d as side chain (S), β-sheet (B), random coil (R), R-helix (A), and turn (T) regions. Briefly, the region from 1600 to 1640 cm-1 is related to the intermolecular and intermolecular β-sheet bands, which increase during silk crystallization. The region between 1640 and 1660 cm-1 includes contributions from random coils and R-helices, which will decrease and transform to the β-sheet structures during crystallization. The remaining part of the amide I region from 1660 to 1690 cm-1 is mainly from turns, with a small β-sheet band peak in the region 1690-1705 cm-1 (30 min) sonication. Figure 7a shows typical interconnected macroporous morphologies for pure silk hydrogels with pore sizes ranging from 4 to 9 µm. Close observation in Figure 7a shows that the surface morphology is much rougher than the HA samples. Many small spherical structures were also seen, which may relate to micelle structures.21 With a low HA (5%) content in the silk/HA hydrogels, sponge-like pore structures were formed as seen in SH95 (Figure 7b). The sizes of the pores is around 7-14 µm. With an increase in HA content, thicker cross-linking networks were observed, such as in SH90 (Figure 7c), with an average diameter about 0.7-1 µm. With larger amplitude, many rougher spherical structures appeared on the

surface, which may indicate more silk β-sheet crystal formation. With higher HA content, as in SH80 (Figure 7d), the diameter of the networks increased to 8-13 µm. For SH66 in Figure 7e, the hydrogel tended to be more open with pore sizes around 10-15 µm. SH50 (Figure 7f) shows a large number of un-crosslinked sheets. The porous structures became larger and inhomogeneous, the limit of stability of silk/HA gels. Cell Attachment and Proliferation. hMSC attachment and spreading were utilized to evaluate cell responses to the silk/ HA samples (SH100, SH95, SH90, SH80, and SH66). After 3 h cell seeding and incubation, all samples showed good initial cell attachment (Figure 8a), similar to tissue culture plastic (TCP). Many hMSCs formed elongated morphologies on surfaces of samples after 3 h. Viability and proliferation of the hMSCs on the different silk/HA samples were evaluated by alamarBlue for 3 h, 5 days, 10 days, and 15 days (Figure 8(b)), along with tissue culture plastic (TCP) as controls. Cells survived and proliferated on the samples during the 15 days. The differences in proliferative activity for 3 h versus 5 days, and 5 versus 10 days were significant (**p < 0.01) in each sample set (SH100, SH95, SH90, SH80, and SH66). More than a four times increase of cell number was found during the 10 to 15 days for all samples. In addition, after 10 days of cell proliferation, the cell numbers of the high HA content group (SH66 and SH80) were significantly (*p < 0.05) higher than that of the low HA content group (SH95 and SH100). However, after 15 days of cell proliferation, the trend changed; the cell numbers of the low HA content group (SH95 and SH100) tended to be


significantly (*p < 0.05) higher than that of the high HA content group (SH66 and SH80). This result indicates that high HA content promoted the growth of stem cells in initial stages (

Biomaterials from Ultrasonication-Induced Silk FibroinâHyaluronic

Recommend Documents

Biomaterials from Ultrasonication-Induced Silk FibroinâHyaluronic