materials science & engineering c

Contents lists available at ScienceDirect

Materials Science & Engineering C

journal homepage: www.elsevier.com/locate/msec

Silk fibroin-poly(lactic acid) biocomposites: Effect of protein-syntheticpolymer interactions and miscibility on material properties and biologicalresponses

Fang Wanga,b, Hao Wuc, Venkat Venkataramanc, Xiao Hua,d,e,⁎

a Department of Physics and Astronomy, Rowan University, Glassboro, NJ 08028, USAb Center of Analysis and Testing, Nanjing Normal University, Nanjing 210023, Chinac Department of Cell Biology, Rowan University School of Osteopathic Medicine, Stratford, NJ 08084, USAdDepartment of Biomedical Engineering, Rowan University, Glassboro, NJ 08028, USAe Department of Molecular and Cellular Biosciences, Rowan University, Glassboro, NJ 08028, USA

A R T I C L E I N F O

Keywords:SilkPolylactic acidPolymer blendsMiscibilityStability

A B S T R A C T

A protein-polymer blend system based on silkworm silk fibroin (SF) and polylactic acid (PLA) was systematicallyinvestigated to understand the interaction and miscibility of proteins and synthetic biocompatible polymers inthe macro- and micro-meter scales, which can dramatically control the cell responses and enzyme biode-gradation on the biomaterial interface. Silk fibroin, a semicrystalline protein with beta-sheet crystals, providescontrollable crystal content and biodegradability; while noncrystallizable PDLLA provides hydrophobicity andthermal stability in the system. Differential scanning calorimetry (DSC) combined with scanning electron mi-croscope (SEM) showed that the morphology of the blend films was uniform on a macroscopic scale, yet withtunable micro-phase patterns at different mixing ratios. Fourier transform infrared analysis (FTIR) revealed thatstructures of the blend system, such as beta-sheet crystal content, gradually changed with the mixing ratios. Allblended samples have better stability than pure SF and PLA samples as evidenced by thermogravimetric analysis.Protease XIV enzymatic study showed that the biodegradability of the blend samples varied with their blendingratios and microscale morphologies. Significantly, the topology of the micro-phase patterns on the blends canpromote cell attachment and manipulate the cell growth and proliferation. This study provided a useful platformfor understanding the fabrication strategies of protein-synthetic polymer composites that have direct biomedicaland green chemistry applications.

1. Introduction

Tissue engineering aims to restore, regenerate or replace functionaltissues compromised by injuries or diseases [1,2] In recent years, arti-ficial tissues have become promising in clinics, which have been usedfor in-vivo cornea and cartilage repairing, as well as skin and musclereplacements [3,4]. However, due to the complexity of the tissue ex-tracellular matrix (ECM) and the different biological environments inthe body, it is still difficult to finely control the properties of the arti-ficial ECM. This has become a critical issue, but few studies have fo-cused on how to use physical methods to effectively and efficientlycontrol the attachment, proliferation and differentiation of target cellsin new extracellular matrices. In addition, no biomaterial system so farthat can cover the biophysical properties of different tissues rangingfrom thin skin to highly vascularized hard tissue (e.g. bones), in order

to meet the specific requirements of various clinic applications [5].Therefore, the fabrication of multifunctional, biodegradable blendsprovides a useful direction for the field of biomaterial and bioengi-neering [6,7], as the properties of biomaterials can become tunable bysimply changing the proportions of components in the same materialsystem.

Numerous studies have illustrated that flexible blend materialsbased on proteins and synthetic polymers are promising. Since thebackbones and side chains of these two types of macromolecules aresignificantly different, while also have the ability to form different in-teractions at the molecular level. For example, the selectivity, specifi-city, precise chemical structure, diverse functionalities of protein bio-molecules are critical in ECM, while the synthetic polymer componentsare highly resistant to corrosion and their stability or processability arealso necessary for biological materials [6,8].

https://doi.org/10.1016/j.msec.2019.109890Received 28 March 2019; Received in revised form 8 June 2019; Accepted 12 June 2019

⁎ Corresponding author at: Department of Physics and Astronomy, Rowan University, Glassboro, NJ 08028, USA.E-mail address: [email protected] (X. Hu).

Materials Science & Engineering C 104 (2019) 109890

Available online 15 June 20190928-4931/ © 2019 Elsevier B.V. All rights reserved.

T

http://www.sciencedirect.com/science/journal/09284931

https://www.elsevier.com/locate/msec

https://doi.org/10.1016/j.msec.2019.109890


mailto:[email protected]


http://crossmark.crossref.org/dialog/?doi=10.1016/j.msec.2019.109890&domain=pdf

Here, we blended crystallizable SF with noncrystallizable polylacticacid at different ratios to better understand the interactions betweenproteins and synthetic biocompatible polymers. Silk fibroin (SF) wasderived from cocoons of domestic silkworm, Bombyx mori, which is atypical natural biopolymer with unique properties [7,8], such as re-newability, mechanical flexibility, non-toxicity, as well as tunablebiodegradability and perfect biocompatibility [9–13]. In addition, it hastwo major phase conformations, beta-sheet crystalline phase and non-crystalline phase including random coils, alpha-helix and beta-turns etc.[7–14]. The fibroin molecular chains are composed of a complex of twocomponents: a large protein fibroin (MW~350 kDa) linked to a secondsmall protein (MW~25 kDa) via disulfide bonds. Moreover, the naturalSF has very high tensile strength and toughness, even superior to Nylonand Kevlar materials [7,9,15].

However, the physical limitations of biomolecules, such as theirsensitivity to temperature, pH, organic solvents, and degradation, ofteninhibit their practical applications. For example, the regenerated SFmaterials often become rigid and brittle when fully dried and are not beflexible after crystallization in the organic solvents (e.g. methanoltreatments), heat treatment or mechanical deformation. These crystal-lization procedures make the fibroin's conformations transformed fromrandom coil/alpha-helix to beta–sheet crystalline [16,17]. Conse-quently, once the crystal content of SF is increased, its elasticity de-creased significantly. Therefore, many blending methods have beenexplored in order to enhance the physical properties of regenerated SFmaterials [18–22]. For example, Cai et al. [23] blended chitosan withsilk fibroin during electrospinning, and found that the mechanicalproperty of composite membranes was significantly enhanced. Ma-cLeod and Rosei [11] fabricated an protein-based inverse opal by in-serting colloidal poly(methyl methacrylate) (PMMA) into the silk fi-broin network, which can be used as resorbable flexible biosensors tosystematically target and destroy unwanted tissues in anticancertherapies. We have [24] also produced a new class of biomedical ma-terials by blending semi-crystalline silk protein with tropoelastin atdifferent ratios, and revealed that the porous blends at micro-/nano-scales could support human mesenchymal stem cells attachment andproliferation with good mechanical elasticity.

In contrast, poly(lactic acid) (PLA) is a typical synthetic biode-gradable thermoplastic polymer. Because lactic acid has two en-antiomeric forms, L-lactide and D-lactide, semicrystalline poly(L-lactide)(PLLA) and poly(D-lactide) (PDLA) materials can be directly derivedfrom L-lactide and D-lactide monomers, respectively. While non-crystalline poly(D,L-lactic acid) (PDLLA) would come from a mixture ofL-lactide and D-lactide monomers, with stable physical properties atdifferent conditions due to lack of crystal structures. Generally, all threekinds of poly(lactic acid) are called PLA. [25–27]. In the previous SF/PLA composites studies, several manufacturing methods have beenused, such as injection molding [25,28], solvent casting, salt leaching[29,30] and electrospinning [17,28–30] etc. Stoppato et al. [31] stu-died the scaffolds of poly(D,L-lactic acid) blended with physicallychopped silk fibers, and demonstrated that the blend increased thestiffness of sponge, and promoted endothelial cells growth and thevascularization in vivo. Ho et al. [25] mixed chopped silk powders andpolylactic acid (PLA) by using injection molding process, which showedthat the coefficient of linear thermal expansions of the composite ma-terial was reduced by 28% than the pure PLA, as well as the loss factorof composites. However, most of these methods did not dissolve andmix silk and PLA at the molecular level [25,28]. Therefore, the mole-cules of SF and PLA have not fully interacted and form new phases/structures in an equilibrium solution system. Some of solvent mixingmethods only investigated the lower content of PLA in the blend system(< 10%) [17,28–32], which did not provide a broad picture of PLA-silkmolecular interactions at different mixing ratios.

In this study, we blended SF with PLA (noncrystalline PDLLA) sys-tematically in solvent at different mixing ratios to fabricate SF-PLAblend films. The ratios of SF are from 0% to 100% at 10–20% intervals.

Besides, the molecules interaction between SF and PLA was in-vestigated by DSC and FTIR techniques. Furthermore, the thermal sta-bility of blends was evaluated by thermo-gravimetric (TG) analysis andthe morphology of blends was observed by SEM. In addition, the en-zymatic degradation and cell response on SF/PLA blends were alsorevealed. This study not only provided comprehensive phase informa-tion and structural properties of SF/PLA blend films, but also theore-tically proved that a fully thermodynamically miscible blending systemof protein-synthetic polymer is achievable, which will help us to un-derstand how to control the interactions between the two components,and generate homogeneous functional microstructures in various forms(e.g. films, gels, microspheres, nanofibers, sponges), which has a widerange of applications in tissue regeneration, tissue fixation, woundclosure, wound dressings and drug delivery systems [7,33–36].

2. Experimental

2.1. Materials

Bombyx mori raw silk cocoons were purchased from China. Poly(D,L-lactide) (PDLLA, referred as ‘PLA’ in this study) powders, with endgroup of acid terminated-free carboxylic acid, molecular weight (Mw)of 18,000–24,000, and Tg of 48–50 °C, were purchased from Sigma-Aldrich Co., Ltd., USA. Dichloromethane (DCM) and Formic acid werebought from Pharmco-Aaper Co., Ltd., USA. All reagents were of ana-lytical grade.

2.2. Preparation of SF/PLA films

Cocoons of Bombyx mori silkworm silk were boiled for 30min in anaqueous solution of 0.02M NaHCO3 and rinsed thoroughly with Milli-Qwater to remove the glue-like sericin proteins [37,38]. The extractedsilk proteins were dried and dissolved in a 9.3M LiBr solution at 60 °Cfor 4–6 h at a concentration of 20 wt%. The silk solution was dialyzedagainst Milli-Q water for at least 2 days using a dialysis cassette (PierceSnake Skin MWCO 3500; Thermo Fisher Scientific, Waltham, MA).After centrifugation and filtration to remove insoluble residues, a 6 wt%silk fibroin aqueous solution was obtained. The silk solution was frozenat −20 °C for 24 h, and then freeze dried to become solid silk foams.The SF solution (25%) was prepared by dissolving SF solid scaffolds intoformic acid.

The PLA was first slowly dissolved into DCM at ambient tempera-ture to form a 25.0 wt% PLA solution, and then slowly mixed with the25wt% silk formic acid solution using a glass pipette to avoid proteinaggregation during mixing. The mixed solution was shaken by abenchmixer (BV1000 Vorter mixer, Benchmark Scientific Inc. USA) for2min, then waited for 20 mins to stabilize the solution. The final so-lutions obtained were based on a mass ratio of Silk fibroin: PLA=90:10(SP90), 70:30 (SP70), 50:50 (SP50), 30:70 (SP30), 10:90 (SP10), withpure silk (SP100) and pure PLA (SP0) used as controls. Finally, thesolutions were cast onto Teflon dishes to form films individually, fol-lowed by solvent evaporation at ambient temperature for 2 days andfurther dried thoroughly in a vacuum oven for 48 h. The SF/PLA filmswere about 50–70 μm thick after the drying process (SupplementalFig. 1).

2.3. Experimental methods

2.3.1. Fourier transform infrared spectroscopy (FTIR)FTIR analysis of silk/PLA films was performed using a FTIR spec-

trometer (Tensor 27, Bruker Co. Ltd., Germany), equipped with adeuterated triglycine sulfate detector and a multiple reflection, hor-izontal MIRacle ATR attachment (using a Ge crystal, from Pike Tech.).The instrument was continuously purged by nitrogen gas to eliminatethe spectral contributions of atmospheric water vapor. For each mea-surement, 128 scans were co-added with resolution 4 cm−1, with the

F. Wang, et al. Materials Science & Engineering C 104 (2019) 109890

2

wavenumbers ranged from 400 to 4000 cm−1. Fourier self-deconvolu-tion (FSD) of the IR spectra covering the Amide I region(1595–1705 cm−1) was performed by using the Opus 5.0 software.Deconvolution was performed using Lorentzian line shape with a half-bandwidth of 25 cm−1 and a noise reduction factor of 0.3, and the self-deconvoluted spectra were then curve fitted by subsequent Gaussianpeaks. FSD is a common signal-processing tool that allows deconvolu-tion of overlapping bands [15]. Using a high pass filter, the broad andindistinct Amide I bands (C]O stretching bonds in protein backbones)can be narrowed synthetically to provide a deconvoluted spectrum withbetter peak resolution. The deconvoluted Amide I spectra were area-normalized, and the relative areas of the single bands were used todetermine the fraction of the secondary structures in the blends.

2.3.2. Differential scanning calorimetry (DSC)The dried silk/PLA blend films (each about 5mg) were encapsulated

in Al pans and heated in a TA Instruments DSC (Q100, TA InstrumentsCo. Ltd., USA), with purged dry nitrogen gas flow (50mL·min−1), andequipped with a refrigerated cooling system. The instrument was cali-brated with indium and sapphire for temperature and heat flow.Standard mode DSC measurements were performed at a heating rate of2 °C·min−1. Temperature-modulated differential scanning calorimetry(TMDSC) measurements were also performed at a heating rate of2 °C·min−1 with a modulation period of 60 s and temperature amplitudeof 0.318 °C. Aluminum and sapphire reference standards were used forcalibration of the heat capacity. All aluminum sample pans were pairedwith a same weight.

In TMDSC, the “reversing heat capacity”, which represents the re-versible heat effect within the temperature range of the modulation, canbe measured [37]. The modulation of TS (t) with amplitude ATs andperiod p (ω=2π/p) can be obtained by using [39]:

= + < > + −T t T q C K A ωt ε( ) [ / ] sin( )S S Ts0 (1)

where ɛ is the phase shift related to the internal reference frequency, Kis the Newton's law constant, CS is the heat capacity of the sample ca-lorimeter,< q>is the underlying heating rate, and T0 is the startingtemperature. So the apparent reversing heat capacity CP can be ex-pressed by Eq. (2):

= < > < >C A A ω K ω[ / ] ( )ϕp Ts (2)

= +K ω τ ω( ) 1 2 2 (3)

where<Aɸ>is the amplitude of the heat-flow-rate in modulationcycle; <ATs> is the modulation amplitude of the temperature (TS)with the frequency ω; and K(ω) is a calibration factor with τ being acorrection value used at the given conditions of the measurement,which was determined from the sapphire calibration scans.

2.3.3. Thermogravimetric analysisThermal gravimetric analysis (Hi-Res TGA 2950, TA Instruments

Co., USA) was used to measure changes in the weight of SF/PLDAsamples with increasing temperature. The weights of all the sampleswere about 5mg, and all measurements were carried out under a ni-trogen atmosphere. The specimens were heated up to 700 °C at aheating rate of 10 °Cmin−1. The curves and calculations of the wholeprocess were used by the TA analysis software. The first derivative ofthe weight functions in each sample, which revealed degradation ratesand middle degradation temperatures of each film were also obtained.

2.3.4. Mechanical analysisStatic tensile experiments were performed on a dynamic mechanical

analyzer (DMA, Diamond, Perkins-Elmer Instruments Co., USA) withapparatus under SS pattern for all SF/PLDA samples. The shapes of allfilm samples were rectangular: 5.0 mm×6.0mm×1.1mm(length×width× thickness). The force was loaded from 5mN to4000mN at 50mNmin−1 at 20 °C. Stress-strain curves were recorded

after the tests. Five tests (n=5) were performed on each type of sampleto calculate the average mechanical parameters, and the standard de-viations are shown in Table 3.

2.3.5. Scanning electron microscopy (SEM)The surface morphologies of silk/PLA samples were observed with

SEM (LEO 1530 VP, ZEISS Corporation, Japan). Samples were fixed onthe SEM holder by conducting tape, and were coated with platinum by20 s plasma deposition (Desk II, Denton Vacuum Inc., USA).

2.3.6. Enzymatic degradationThe seven groups of films samples were incubated at 37 °C in 10mL

solutions containing 2 U/mL protease XIV (Sigma-Aldrich, St. Louis,MO) in phosphate buffered saline (PBS) at pH 7.4 [9]. Solutions werereplaced daily and the samples were rinsed at designated time points indistilled water and were dehydrated for weight measurements. Controlsamples were also studied under the same condition but without en-zyme. The remaining mass percentage of each film was estimated bydividing remaining dry weight of the sample by its initial dry weight[40–44].

2.3.7. Cell attachment and viabilityNIH/3T3 fibroblast cells, a kind gift from Dr. Dimitri Pestov (Rowan

University), were grown in Dulbecco's modified Eagle's medium(HyClone, with 4.00mM L-Glutamine and 4500mg/L Glucose) supple-mented with 10% fetal bovine serum (Life Technologies Inc.) and100 U/mL Penicillin-Streptomycin (Life Technologies Inc.). Cell culturewas carried out according to standard protocols [45], where all sampleswere coated on cell culture glass plates for accurate cell observation andmeasurements. Blank glass surface was used as the control group. Equalnumber of cells (4× 103) per cm2 was seeded on different SF/PLAblend films. 6 h after seeding, the cells were incubated with freshgrowth medium containing 5 μg/mL of Hoechst 33342 (Life Technol-ogies Inc.) for 15min, changed into fresh culture media and allowed togrow. Images of the cells were acquired at 6 and 48 h after seedingusing Nikon Eclipse Ti microscope equipped with epifluorescence. After48 h, the cells were washed once using Hank's Balanced Salt Solution(Life Technologies Inc.), fixed with 4% paraformaldehyde (ElectronMicroscopy Sciences) for 5min and then viewed by Zeiss Observer Z1microscope. NIH ImageJ software was used to quantify cell numbers.

3. Results and discussion

3.1. FTIR analysis

To understand the interactions between the two components in theSF/PLA films, FTIR analysis was first performed to measure the struc-tures of these blends. Fig. 1 shows the FTIR absorbance spectra of SF/PLA samples (SP100, SP90, SP70, SP50, SP30, SP10, and SP0) for thewavenumber region of 900–2000 cm−1. Generally, for the silk protein,the IR spectral region within 1700–1500 cm−1 is assigned to the pep-tide backbone of Amide I (1700–1600 cm−1) and Amide II(1600–1500 cm−1) absorptions [42]. The Amide I region mainly comesfrom the C]O stretching vibration (> 80%) [37], which directly re-veals the secondary structure of the protein backbone. The region1600–1640 cm−1 is related to the intermolecular beta-sheet bands,which increases during silk crystallization. The region between1640 cm−1 and 1660 cm−1 are dominated by vibrations from randomcoils and alpha-helices. The remaining parts of the spectra(1660–1690 cm−1) are mainly from beta-turn structures, with somesmall bands from other structures, such as a possible beta-sheet peakbetween 1690 and 1705 cm−1. The Amide II bands are mainly from theNeH in-plane bending vibrations and the out-of-phase combination ofthe CeN stretching in protein backbone [37]. Amide II regions oftenmix with the vibrational modes of protein side chain groups for solidsamples, which can be used to analyze the change of the micro-


3

environment and tertiary conformation of the proteins [24]. For PLA,the significant IR absorption bands (Fig. 1, SP0) at 1750 cm−1,1452 cm−1, 1183 cm−1 and 1082 cm−1 are associated with C]Ostretching of carbonyl group, CeH deformation vibration, CeOeCstretching and CeO antisymmetric stretching [17,27], respectively.

In our study, the absorption bands of pure SF film (Fig. 1a, SP100)observed at 1645 cm−1 (amide I), 1525 cm−1 (amide II) and1236 cm−1 (amide III) were assigned to random-coil dominated Silk Istructure [17,18]. By comparing the normalized FTIR curves of SF/PLAblends (Fig. 1a), it was found that the intensity of FTIR spectra at1750 cm−1 and 1082 cm−1 gradually decreased with the increase of SFcontent. On the other hand, the spectra intensity at 1645 cm−1,1525 cm−1 and 1452 cm−1 gradually increased with the increase of SFcontent. Besides, after blending SF with PLA, many absorption peakswere found shifted or changed. Specifically, the peaks at 1645 cm−1

and 1525 cm−1 became wider, and additional small peaks appeared at1625 cm−1, 1540 cm−1 and 1514 cm−1 (Fig. 1b), with the peak of1236 cm−1 shifted to 1267 cm−1. These findings indicate that newconformation (e.g. beta-sheet crystals dominated structure in silk at1625 cm−1, 1540 cm−1 and 1267 cm−1) has formed after the blending[18,24]. It also implies that the molecular interactions between SF and

PLA can lead to the conformation change of protein from Silk I to Silk IIgradually. Zhu et al. [17] studied the composites of SF-PLA at 98:2 to90:10 mixing ratios, and pointed out that the hydroxyls of amino acids,such as Ser, Asp and Glu, on the SF chains can have strong interactionswith the carbonyls of the PLA chains, which promoted the formation ofintermolecular hydrogen bonds between the two molecules.

To quantify the percentage of the secondary structures in the SF-PLAsamples, a deconvolution method based on curve fitting spectra afterFourier-Self-Deconvolution was performed on the Amide I region(1595–1705 cm−1) [24]. Fig. 1c shows FTIR spectra Amide I and IIregion of all samples treated in methanol for 20mins. Fig. 1d shows FSDAmide I spectra of the pure SF (SP100) with fitted vibrational bands (inred dot line). The peak position and their related secondary structuresfor silk proteins can be assigned from reference [37], as side chains (S),beta-sheets (B), random coil (R), alpha helix (A), and turns (T). Table 1summarizes the percentage of different secondary structures for eachuntreated and methanol treated SF-PLA samples.

For the untreated original blends, pure silk protein (SP100) had only8.76% beta-sheets, with 86.79% random coils and/or alpha helix, and2.55% turns. In the untreated blends, with an increase of PLA content,the amount of beta-sheet structures gradually increased to 33.50% in

Fig. 1. FTIR absorbance spectra of theSilk/PLA samples, for (a) Untreatedsample spectra in 900–2000 cm−1, (b)1200–1750 cm−1 (with normalizationof the Amide I intensity to better un-derstand their peak shifts), and for (c)20min MeOH treated sample spectra inAmide I and II regions. (d) A curvefitting example of FSD Amide I spectra(sample SP70). The fitted peaks areshown by dashed lines, and assigned asside chains (S), β-sheets (B), randomcoils (R), α-helix (A), and turns (T).

Table 1Percentage of secondary structures in SF/PLA blends.

SF/PDLA Silk fraction in sample β-sheet (B) in silk α-helix & random coils (A+R) in silk Turns (T) in silk Side chains (S) in silk Silk amorphous in sample(%)

SP100 100 8.76/59.23a 86.79 2.55 1.90 40.77SP90 90 20.49/52.55a 69.56 7.90 2.11 42.31SP70 70 23.50/46.65a 61.61 12.89 2.00 37.34SP50 50 25.52/45.18a 59.08 13.60 1.80 27.41SP30 30 28.60/41.20a 55.33 14.37 1.73 7.64SP10 10 33.50/36.6a 43.80 21.00 1.70 6.34

*All calculated secondary structure fractions have a same unit (wt%) with a± 2wt% error bar;a Beta-sheet percentages after 20min methanol treatment.


4

the SP10 sample, while random coils and alpha helix decreased to43.80%, and turns increased to 21.00%. The side chains did not changesignificantly (from 2.11% in sample SP90 to 1.7% in sample SP10).After 20min methanol treatment, the beta-sheet crystal peak (around1620 cm−1) increased significantly with the increase of silk in the SF-PLA blends, as seen in Fig. 1c. In the treated pure silk films, 59.23%beta-sheet crystal was formed. With the increase of PLA content, beta-sheet crystals reduced to 36.6% in the treated sample SP10, but is stillhigher than the untreated SP10 samples (33.5%) (see Table 1). Thebeta-sheet crystals in the blends are significantly important to promotethe formation of stable protein-polymer networks between the silk andPLA molecules [24]. The results from the IR spectrum studies provedthat SF and PLA have strong molecular interactions, which lead to newconformation after the blending.

3.2. DSC analysis

In general, polymer interactions include van der Waals forces, hy-drogen bonds, ionic bonds, and disulfide bonds between the specificdomains of polymers. From the view of polymer-protein interactions,the free energy of mixing is the key factor to govern the miscibility ofblends [46–48]. A negative enthalpic contribution can occur when theinteraction between the polymer-protein components becomes sig-nificant, which will increase the miscibility of the blend system [46,47].Therefore, a homogeneous macrophase in the polymer-protein blendcan be obtained if specific favorable interactions exist between eachcomponent in the mixture [46]. The clearest experimental evidence of amiscible blend on macroscopic scale (above 100 μm) is the occurrenceof a single glass transition temperature (Tg) for the protein-polymersystem, which is normally located between the Tgs of their two in-dividual components [48,49]. DSC is one of the most important tech-nologies to acquire evidence for the miscibility of polymer-proteinblends by investigating their glass transition temperature (Tg) regions[8,50,51].

For a binary component system, the interaction of SF with PLA canbe viewed as the interaction between a “solvent” and a “solute” basedon the Flory-Huggins's lattice model [46,47]. The variation of Tgs in themixture of polymer-protein blend systems can be predicted by Foxequation [50], Kwei equation [51] or Gordon-Taylo equation [52],which are listed below:

= +T W T W TFox equation: 1/ ( / ) ( / )g 1 g1 2 g2 (4)

= + + +T W T W T W W W WKwei equation: (( k )/( k )) qg 1 g1 2 g2 1 2 1 2 (5)

− = + +T W T W T W WGordon Taylo equation: ( k )/( k )g 1 g1 2 g2 1 2 (6)

where Tg is the glass transition temperature of the final blend; Tg1 andTg2 are the glass transition temperatures of their individual compo-nents, respectively; W1 and W2 are the weight percentages of these twocomponents, respectively; In Kwei and Gordon-Taylo equations, k is aparameter representing the strength of molecular interaction betweenblend components. Kwei equation has an advanced format than theGordon-Taylo equation. The first term on the right side of Kwei Eq. (5)is identical to the Gordon-Taylor Eq. (6), while the last term on the rightside of Kwei Eq. (5) represents addtional effect of interaction, such asthe hydrogen bonding in the binary system. Specifically, q is a para-meter for the specific intermolecular interactions in the mixture, ac-counting for the effects of rearrangements between the molecules. Thevalues of q can be either positive or negative [51]. We used standardDSC and temperature modulated DSC (TMDSC) to examine the thermalproperties of the SF/PLA blend films in order to assess the miscibility ofSF/PLA mixtures. Besides, we used Fox Eq. (4) and Kwei Eq. (5) todescribe the Tg-component relationships/interactions. The subscript ‘1’and ‘2’ in the equations stand for pure SF (SP100) and pure PLA (SP0),respectively.

Fig. 2a shows the standard DSC curves of SF/PLA film samples with

different mixing ratios. The SP100, SP90 and SP70 samples showed anobvious endothermic dehydration peak between 25 °C and 200 °C, re-spectively. All other blend samples did not display a clear dehydrationpeak during this range. Since the pure SF film is hydrophilic, it un-dergoes a dehydration process during heating, but as the SF contentdecreases, the dehydration peak gradually disappears. Previously [53],we have studied the silk-bound water films and found that the boundsolvent molecules in the film could act as a plasticizer in the silk films.For the pure silk sample, following a glass transition at 178 °C (Tg, ofSP100, in Fig. 2a), an irregular and asymmetric endothermic peak startsto appear from 220 °C (onset temperature), which can be assigned tothe degradation of the pure SF film. As for the pure PLA sample (SP0,Fig. 2a), following a glass transition at 48 °C (Tg), the degradation oc-curs at 250 °C (onset temperature). The heat flow curves between thetemperature ranges 240 °C~350 °C in Fig. 2a demonstrate decomposi-tion endothermic peaks from different blend samples. Analysis of thesedecomposition peaks revealed that as the PLA content increased, thedecomposition peak shifted to a higher temperature, from 246.75 °C(peak center) of pure SF (SP100) to 281 °C of pure PLA (SP0). All SF/PLA blend samples had higher degradation temperatures than pure silksamples, indicating that the SF/PLA samples have higher thermal sta-bility than pure silk proteins after mixing.

To investigate the Tg behavior of SF/PLA samples in detail, TMDSC,a technique that can eliminate the non-reversing thermal phenomena of

Fig. 2. (a) Standard DSC scans of SF/PLA blend films. The samples were heatedat 2 °C·min−1 from −30 °C to 400 °C. The endothermic direction is downward.(b) Reversing heat capacities of the silk/PLA samples measured by TMDSC witha 2 °C·min−1 heating rate, a modulation period of 60 s and a temperature am-plitude of 0.318 °C from −30 °C to 400 °C. Tg, Tonset, Tp stand for the glasstransition, solvent releasing onset temperature, and first degradation peaktemperature, respectively.


5

the sample (e.g., water evaporation process, or physical aging/de-gradation process), was used to measure the reversible thermal prop-erties of the SF/PLA samples. Fig. 2b shows the reversing heat capacityof the samples from −30 °C to 250 °C. The glass transition temperature(Tg) was determined from the midpoint of the step in the reversing heatcapacity. With the increasing of SF content, the glass transition tem-peratures (Tg) of the blends moved gradually from 48 °C (pure PLA) to178 °C (pure SF). At the same time, no individual glass transition fromthe SF or PLA component was observed for all samples. This indicatesthat the formed SF/PLA blend system is completely thermodynamicallymiscible at different ratios of the two components.

Table 2 summarizes the glass transition temperatures of SF/PLAsamples experimentally measured by TMDSC, or theoretically predictedby Fox and Kwei equations, respectively. Fig. 3 also shows the fittingcurves of the glass transition temperature of SF/PLA blends accordingto the results of TMDSC (square, solid line), Fox (circle, dash line) andKwei (Triangle, dot-dash line) Equations (k= 1.00, q=−150), re-spectively. Three fitted curves are expressed in the equations below:

= +

=

T W RThe fitting curve from DSC data: 37.393 8.123exp( /35.194)

0.995

g

= +

=

T W RThe fitting curve from Fox method: 42.176 7.452exp( /34.591)

0.999

g

= + =

T

W R

The fitting curve from Kwei method:

28.780 15.504exp( /43.909) 0.998

g

These fitting curves, from DSC method, Fox equation and Kweiequation all showed high fitting correlation coefficients that are close to0.99. The data from DSC data was almost coinciding with that calcu-lated from Fox equation when silk content is dominated in the blend.Kwei curve is also very close to the DSC curve when PLA content isdominated. The large negative value of q (q=−150) indicates a strongintermolecular attraction occurred between SF and PLA molecules afterblending [49,51,52]. Therefore, the results from Fox and Kwei equa-tions proved theoretically that a stable homogeneous system withoutmacrophase separation of silk and PLA components has been achieved.

3.3. TG analysis

Thermogravimetric analysis was used to further understand thethermal stability and thermal degradation behavior of the blend films.Fig. 4a shows the weight percentage change of samples during a heatingscan from 25 °C to 450 °C. During the initial heating between roomtemperature and 200 °C, all samples have a mass loss due to the eva-poration of the bound solvent molecules or water [11,43]. Pure PLAfilm had almost no mass loss, 0.92%, before it began to fully decom-pose. With the addition of silk protein, the percentage of mass loss in-creased when heated from room temperature to 200 °C, which are

Table 2Measured or calculated glass transition temperatures (Tg) of SF/PLA blend filmsat different mixing ratios.

SF/PLA sample SP100 SP90 SP70 SP50 SP30 SP10 SP0

SF/PLA (W1/W2) 100/0 90/10 70/30 50/50 30/70 10/90 0/100Tg-DSC/°C 178.00 138.23 102.10 70.20 51.74 48.47 48.00Tg-Fox/°C 178.00 140.06 98.21 75.59 61.47 51.79 48.00Tg-Kwei/°C 178.00 151.50 107.50 75.50 55.50 47.50 48.00

*Tg-DSC represents the Tg values from DSC measurement. Tg-Fox, Tg-Kwei arecalculated from Fox equation and Kwei equation, with k=1.0, q=−150[51,54], respectively. W1 and W2 represent the weight of SF and PLA, respec-tively.

Fig. 3. Glass transition temperature vs. SF weight content (WSF) for all SF/PLAblend films: (a) solid line is the best fit of the blend films data (square) usingTMDSC; (b) dash line is the best fit of the blend films data (dot) using FoxEquation (Eq. (1)); (c) short dashed dot line is the best fit of the blend films data(triangle) using Kwei Equation (Eq. (2)), with k= 1.0 and q=−150.

Fig. 4. (a) Mass percentage change of the SF/PLA blend films by TGA duringheating from room temperature to 450 °C at 10 °Cmin−1. (b) First derivative ofthe mass percentage curves. (c) The mass loss percentage of SF vs. the peaktemperature in DTG curve of Fig. 4b at which the sample gets the maximumthermal decomposition rate.


6

4.0 wt% for SP10, 4.2 wt% for SP30, 8.4 wt% for SP50, 13.1 wt% forSP70, and 10.45 wt% for SP90, 18.7 wt% for the pure SF (SP100) at250 °C. Above 250 °C, strong thermal degradation begins to occur. Allsamples showed one degradation stage during decomposition tem-perature range of 200 °C to 400 °C. Once the initial decompositiontemperature (~270 °C) is reached, the SP0 sample shows the fastestmass loss. Fasting degradation may be beneficial for biological cellscaffold applications [36,42]. However, SP100 samples (pure SF films)always lost mass (about 10.20%) during the heating before 105 °C. Twopeaks were shown in the first derivative of mass loss among 25 °C to200 °C (Fig. 4b). When the temperature reached to 400 °C, the SP100film still has 26.7% sample left. Interestingly, for SP90 and SP70 blendsamples, their decomposition curves are similar to SP100 due to theirhigh silk fibroin content. For the SP50 blend film, the mass loss of thissample (about 2.3% at 105 °C) is less than these of SP90 and SP70(about 3.6% and 5.7%), and its decomposition rate is faster after 105 °C.Thus, the composition of the blend will significantly affect sample'sthermal degradation rate and its degradation onset temperature. Afterthe start of degradation, all mixed samples quickly lost about 24% to98% of the mass, which may be related to the decomposition of sidechain groups and molecular bonds in the SF/PLA films [17]. Finally, theresidue amount at 400 °C decreased as the PLA content increases. Forexample, the SP0 film has about 1.7% residues at 400 °C, and about76.5% for the SP90 sample.

The degradation peak temperature (Tp) in DTG curve represents themaximum thermal decomposition rate of sample. Clearly, SP0 has thehighest thermal decomposition rate among all samples. Then the ratedecreases with the increase of silk content. Besides, the degradationpeak temperature of the blend samples was different from the pure SFor pure PLA curves. Fig. 4c shows the Tp values of all samples in theDTG curves. Above 400 °C, all samples degraded uniformly and nopeaks appeared in the first derivative curves.

In summary, the TG results showed that the bound solvent contentin SF/PLA blend films were less than the pure SF film, and their thermalstability varied with the change of SF contents [7]. Furthermore, the TGcurves also illustrated that the SF and PLA components were well mixedin the film systems and did not follow the individual thermal profiles ofSP100 and SP0.

3.4. Mechanical property

Stress-strain curves of different silk/PLA blends are shown in Fig. 5,and their mechanical parameters such as elastic modulus and elonga-tion ratio were then calculated (Table 3). Elastic modulus is an im-portant parameter that can demonstrate the object's resistance

capability to the elastic deformation, and elongation ratio may reflectthe bonding strength between molecules in blending samples [31]. It isinteresting to find that the addition of silk into the materials increasedthe elastic modulus of samples, from 0.17MPa (SP10) to 0.44Mpa(SP90), while pure silk (SP100) has the highest modulus about0.69MPa. And their elongation ratios also increased with the increaseof silk content in the blends when silk content is< 50% (PLA domi-nated blend samples), from 4.9% (SP10) to 26.4% (SP50). However, theresults also showed that the silk dominated samples of SP70, SP90 andSP100 have shorter elongation ratios than that of SP50 while theirtensile strengths and elastic modulus still follow the general trend.Therefore, it may be implied that when silk and PLA were mixed in aclose ratio (such as 50%–50% in SP50), the interactions between theirmolecules are strongest, which improved the elasticity (elongationratio) of the blend sample. It also needs to be noted that the PLA (SP0)control sample was very rigid and brittle which yielded 0.3% de-formation before broken, and its elastic modulus was about 34MPa (50times of all other silk blended samples), which is similar to the PLAsamples with the same molecular weight reported in literatures[27–29]. This also indicated that adding a small amount of silk cansignificantly increase the softness of PLA samples, which might havemany potential applications in the future.

3.5. Morphology of SF/PLA films

SEM was then used to investigate the cross-section morphology ofSF and PLA blends with different mixing ratios (Fig. 6). It shows thatthe morphology of all blend films is uniform on a macroscopic scale(Fig. 6A, 20 μmbar), but with interesting micro-phase patterns (Fig. 6B,200 nm bar). For pure PLA sample (SP0), a uniform, relatively smoothtopology was observed (Fig. 6Aa, SP0). However, for the pure silk fi-broin film (SP100), an irregular small fibril structure was observed(Fig. 6Ae, SP100), which is typical for regenerated silk materials[17,24,25,28]. By comparing the images of blend films with differentratios in Fig. 6A (20 μmbar), when SF content was below 50%, a net-work formed inside the SF/PLA blend was observed. The result illu-strated that the bridging effect between PLA and silk chains is able toprevent crack propagation [28] and improve the physical properties ofpure silk or PLA materials, which provide the blends with suitablemechanical properties [17]. Besides, as the PLA content increases, thecompactness of the film also increased. This result was also attributed tothe increased charge density of the mixture surface [40].

Furthermore, SEM images with a 200 nm bar (Fig. 6B, 200 nm bar)showed that some agglomerates at nano-scale can be observed when thePLA content reached 10wt% (SP90) which made the micro surface ofthe mixed membrane rough, as Zhu et al. [17] reported. As the silkcontent increased, spherical (SP30), insular (SP50) and linear (SP70)micro-phase patterns were observed. These micro-scale structures canprovide additional surface properties to the hybrid film, such as im-proved cell attachment and proliferation [41,42]. Simultaneously, theyalso affected the biodegradability of blended materials [43,44], whichwill be discussed in the enzyme biodegradation section below. Fig. 6Cshowed a schematic of the SF-PLA interaction mechanisms based on the

Fig. 5. Representative tensile stress-strain curves for the Silk-PLA films SP100,SP90, SP70, SP50, SP30, SP10 with SP0, measured at 20 °C.

Table 3Elastic modulus and elongation ratios of silk-PLA films (sample size n=5).

Sample Elastic modulus(MPa)

Elongation(%)

SP0 33.53 ± 3.1 0.24 ± 1.2SP10 0.17 ± 0.25 4.9 ± 2.4SP30 0.13 ± 0.33 14.9 ± 2.2SP50 0.21 ± 0.26 26.4 ± 5.7SP70 0.30 ± 0.21 6.3 ± 2.6SP90 0.44 ± 0.82 5.9 ± 1.3SP100 0.69 ± 0.64 15.6 ± 4.6


7

SEM images.

3.6. Enzyme degradation

Silk fibroin is an enzymatically biodegradable polymer [44]. Manyimportant factors, such as structure, morphology, molecular weight,hydrophobic and hydrophilic properties, as well as additives, syntheticmethods and environmental conditions, can directly influence the

accessibility of the material to the enzymes that catalyze the polymerchain cleavage [44]. Therefore, in vitro degradations of the SF/PLA bio-composite films incubated in water and protease XIV solution was ob-served (Fig. 7). To mimic the same sterilization conditions in cell stu-dies, all films were pre-treated in a 70% ethanol solution before thedegradation study. In the aqueous system (Fig. 7a), it was observed thatall samples exhibited very low degradation and their weight remainedfairly stable over the time [55]. It also can be seen that the weight loss

Fig. 6. SEM images of SF/PLA blend films within (A) 20 μmbar and (B) 200 nm bar. (C) Schematic of the SF-PLA interaction mechanisms based on the SEM images.


8

obtained is a function of the amount of SF present in the blend. Duringthe initial 20 h, small biodegradation (sample weight loss) was found.The degradation then slowed down and stopped around 90 h. Theamount of pure SF sample (SP100) degraded in the water is about 17%[20,56].

Similar behavior was observed in the enzyme solution system(Fig. 7b). All samples initially lost weight rapidly before 70 h, and thenthe degradation rate decreased. The pure PLA film (SP0) exhibited veryslow degradation [55], and the weight lost is only about 7% after 120 hof exposure in the protease XIV solution [55,57]. While pure SF film(SP100) demonstrated a much higher weight loss rate in protease XIVsolution [56]: after incubation at 37 °C for 20 h, the weight loss of thesilk film was about 20% [58]; at 120 h, the weight loss is approximately50%, which is consistent with previous results [56,58]. Generally, asthe SF content increases, the in-vitro biodegradation of the SF/PLAfilms increases. However, SP30, SP50 and SP70 samples showed higherbiodegradation at 120 h than other blend samples, which may be re-lated to their unique micro-phase separation patterns reported in SEMsection. Vasconcelos et al. [20] prepared blend films of silk fibroin andkeratin from formic acid, and pointed out that proteins do not followthe additives rule, but are able to establish intermolecular interactionsduring mixing, and the solvent could induce the crystallization of beta-sheets in SF chains and made the amino acids more accessible fortrypsin hydrolysis. In addition, Yoon Sung Nam and Park [59] found

that the interfacial energy at the blends surface is the key to the for-mation of a phase separating system, which then affects distributionand release behavior of carbonic anhydrase II. Mi et al. [60] fabricatedthe chitin/poly(D,L-lactide-co-glycolide) blend microspheres, which isphase-separated as a biodegradable drug-delivery system. They foundthat 50/50 ratio mixture degraded faster than PLA and indicated thatthe amorphous structure of the mixture allowed rapid water penetra-tion. Therefore, it is reasonable to conclude that the micro-phase se-paration patterns promote rapid decomposition of the blends in theenzymatic solution in our study. In addition, it was found in Fig. 6 thatthe tightness tendency of the micro-patterns is spherical (SP30) <insular (SP50) < linear (SP70), wherein SP30 has the loosest structurein the film, and it also has the smallest β-sheet content (41.20%) aftermethanol treatment when compared to the samples SP50 (45.18%),SP70 (46.65%) and SP90 (52.55%) (Table 1). These two factors maycause the SP30 sample to be almost completely degraded in the pre-sence of protease XIV. All of these studies have demonstrated thatbiodegradation of blend materials is a very complex process. In thiscase, the mixed structure, the surface morphology of blends, and themolecular interactions between the silk protein and the PLA polymerchain could jointly influence the material's accessibility to the proteaseXIV enzyme. Since the samples SP30, SP50 and SP70 showed thestrongest micro-phase separation patterns in the SEM, it is reasonable tohave higher biodegradation of these samples in the enzyme solution,which could be an important parameter for controlling the biomedicalproperties of the material in the future.

3.7. Cell interaction

Mouse fibroblast NIH/3 T3 cells were then utilized to understandthe cell viability on different SF/PLA blends. After seeding an equalnumber of cells on samples for 6 h and 48 h, the cells were stained andquantified in Fig. 8 and Fig. 9. Compared with the glass control surface,the morphology of the cells on all films remained normal at 48 h(Supplemental Fig. 2), which indicates the NIH/3T3 cells can pro-liferate on all SF/PLA films with long time. After 6 h, cell can attach toall samples successful, and compared with the control, the cell densityon the 70%, 50%, 30% silk blends were slightly higher than that of thecontrol surface (Fig. 8). In addition, the PLA (SP0) did not hinder thegrowth of cells at the initial stage. It has been reported that porousbiomaterial surface patterns can affect initial cell attachment [61–63].Therefore, we can conclude that the pure silk (SP100) and SP90 sur-faces are the least porous in the group (see SEM images), and addingsuitable amount of PLA can promote initial cell attachment by in-troducing a porous topology.

Fig. 7. In vitro enzyme degradation profiles of SF/PLA films. Remaining massof SP films in (a) distilled water and (b) protease XIV solution are demonstrated.(N=4, statistically significant difference p < 0.01).

Fig. 8. The quantification of NIH/3T3 fibroblast cell density at 6 and 48 h, onthe surfaces of different silk/PLA blend samples using a glass surface as thecontrol. Scale bar: 200 μm, p < 0.05.


9

Cell proliferation on different silk/PLA blends was also assessed at48 h (Figs. 8, 9). Compared to the control, cell growth and proliferationwere significantly improved on silk dominated samples (SP100, SP90,and SP70). When the weight percentage of PLA reaches 50% (SP50),the cell density decreased to 6188 ± 396 cells per mm2 but still com-parable to the control (p=0.7256). However, after the content of PLAreached 30wt%, the cell densities were declined drastically (SP30,3348 ± 310 cells per mm2; SP10, 3224 ± 342 cells per mm2; SP0,4010 ± 611 cells per mm2). Thus, the results indicate that a suitablemixture of SF/PLA mixtures can support and promote cell growth andproliferation, while excess PLA (PLA > 50%) may hinder it. However,compared to the pure PLA (SP0), similar to the 6-h control, there was aslight increase in cell growth and proliferation at 48 h, indicating thatPLA does not significantly promote or hinder cell growth and pro-liferation (SP10, SP0). Kim et al. [64] used thermally induced phaseseparation to form a PLLA scaffold and pointed out that the surface ofscaffold is the key to hold enough number of cells to induce cell–cellinteraction. Many studies [60,65–67] also report that hydrophilicity,surface roughness and phase separation patterns may affect cell

attachment and growth. Therefore, for the blending system in ourstudy, different phase separation patterns could have different effectson cell growth and proliferation, similar to their impact on the enzy-matic degradation rates. For example, SP70 blend sample showed thelinear pattern which the cell growth and proliferation was significantlyhigher than both insular SP50 and spherical SP30, while the enzymaticdegradation rate of the SP70 sample was also lower than that of theSP50 and SP30 samples. Due to changes in the physical and biochem-ical properties of the silk/PLA blend, the blend matrix can be used tofinely control cell attachment and proliferation at the interface of thebiomaterial.

3.8. Mechanism of protein-polymer interaction and miscibility

Miscibility between protein (silk fibroin) and synthetic polymer(polylactic acid) is an important property in the study of polymer blendsystems. According the classical thermodynamics and kinetics ofpolymer blending theory [68,69], three main types of miscibility basedon polymer blends can be found. One is that in the fully miscible blend,

Fig. 9. The morphology of NIH/3T3 cells on different SF/PLA blend biomaterials (SP100, SP90, SP70, SP50, SP30 and SP10 SP0) at 48 h. Scale bar: 25 μm.


10

the components are homogenously mixed at or below the nanoscale.Both polymers completely dissolve at each other at the molecular level,which leads to improved properties. The second is a miscible blend thatexhibits some degree of homogeneity in which one polymer can bepartially dissolved into another polymer at molecular level, and theminor polymer chain is evenly distributed within the major polymernetwork. Thus, if the two polymers are 50–50 mixed, there could be twotypes of microscale domains in the blend, dominated by either of thetwo polymer components. If the two types of polymer chains do nothave strong interactions (e.g., hydrogen bonds), strong phase separa-tion properties between these microdomains (e.g., two glass transitionsin DSC) can be thermodynamically revealed. Otherwise, strong inter-actions between the polymer chains will overcome the effects of mi-crophase separation and maintain a thermodynamically misciblesystem on a macroscopic scale. In this case, a uniform morphology atmacroscale can still be displayed, and a satisfactory homogeneousproperty can be provided. The third case is due to the major phaseseparation in the blend, the two polymers are completely immiscible,indicating that the properties of the blend are distorted compared to theneat polymer. Accordingly, both miscible and partially miscible blendscould be compatible (thermodynamically miscible) and can produceuniform properties on a macroscopic scale, whereas immiscible blendsare incompatible with each other with inhomogeneous properties fromdifferent parts of the material [69]. Therefore, the miscibility of theblend is directly related to the interactions between the blend compo-nents and resulting morphology, as well as their physical and biologicalproperties. The interaction parameters between polymers are also re-lated to the blend composition, as reported by Zhong and Al-Saigh [70].

In our DSC glass transition studies, the single glass transition ob-served from each blend ratio indicates that the blend is thermo-dynamically miscible at all ratios. However, the downward shape of thefitted curve in Fig. 3 also indicates that the molecules from the twocomponents bind to each other less strongly than themselves, so the Tgsof the blend is slightly lower than expected (linear fit line). The nega-tive values of the fitted parameters (q=−150) calculated by the Kweiequation show the intermolecular interactions between the rearrange-ments of molecules in the SF/PLA polymer blend system [71,72]. Theamide group (R-NH) on the silk fibroin interacts with the carbonylgroup (C]O) on the polylactic acid molecular chain to form a hydrogenbond. When PLA and SF are mixed, intramolecular hydrogen bond insilk fibroin can be broken, the amide structure redistributed, and the

carbonyl in the polylactic acid molecule stretched, in which the alphacarbon atom is more positive and the oxygen atom becomes more ne-gative which has three lone pair electrons in the molecule of lactic acid.This structure is that the conducive oxygen atom in the PLA molecularchain forms an intermolecular hydrogen bond with the -NH in the silkfibroin peptide. The FTIR, DSC and TG analysis results proved thishypothesis. After 20min of methanol treatment, the percentage of β-sheets (showing strong hydrogen bonds in the protein) in the SF/PLAsample decreased from 59.23% to 36.6%, and the PLA content in-creased from 0 to 90%. Meanwhile, the degradation temperature de-creases, and the degradation rate increases as the PLA content in-creases. As the silk protein content increases, interfacial cross-linkingalso significantly increases the modulus of the blend [71]. Moreover,from the view of a molecular scale, silk protein is a copolymer with abackbone comprising of alternating hydrophilic and hydrophobic do-mains, which can facilitate various polar–polar, hydrogen bond andhydrophobic–hydrophobic/hydrophilic interactions [71]. Hence, thebinding between hydrophobic silk molecular domains and hydrophobicPLA molecules can also be via the hydrophobic-hydrophobic interac-tions. Compared to the physical property measurements, cell pro-liferation test as well as the enzyme biodegradation test were moresensitive to the microphase separation patterns of polymer, as demon-strated by SEM [60]. With the silk content increasing, spherical (SP30),insular (SP50) and linear (SP70) micro-phase patterns appeared(Fig. 6). The SP30, SP50 and SP70 samples displayed higher enzymebiodegradation rates among other samples, with distinguished cellgrowth and proliferation after 48 h. Hence, the silk fibroin content ofthe blend matrix alters the surface morphology of the silk/PLA blend atthe material interface, thereby finely controlling cell attachment andproliferation at the interface of the biomaterial, as described above.Base on this discussion, it can be seen that the SF/PLA mixture is athermodynamically miscible blend with uniform tunable properties ona macroscopic scale, while with different microscopic patterns to con-trol various biological responses (Fig. 10). This balance between mac-roscopic and microscale material properties is critical in protein-syn-thetic polymer blend studies because most protein structures may formsimilar hydrogen bonds with synthetic polymers, yet it is still difficult tofind a pair of fully miscible protein-synthetic polymer at nanoscale dueto the significant difference between the protein and synthetic polymerchain backbones.

4. Conclusions

SF/PLA physical blends represent a new class of polymer-proteincomposite biomaterials. Synthetic polymer PLA confers stability andhydrophobicity to silk protein. While semicrystalline silk protein con-fers biodegradability, promoting cell interactions and providing phy-sical cross-linkers by varying beta-sheet crystal content to PLA. Thesebiomaterial blends can support and tune the attachment and pro-liferation of NIH/3T3 mouse fibroblast cells. As evidenced by DSC,TMDSC, SEM and TG studies, these protein-polymer blending systemsare stable and fully miscible in the macroscale with unique micro-phasepatterns. FTIR analysis demonstrated the conformation of silk proteinschanged gradually from Silk I to Silk II due to the molecular interac-tions with PLA polymers. All blended samples have improved stability,while their biodegradability is associated with their blend ratios andmicro-phase patterns. This study not only prove that the physicalblending method is feasible to provide polymer-protein blends fullmiscibility and tunable physical and biological properties, but alsoprovide experimental and theoretical foundations to help evaluate thephase miscibility information and the tunable properties. In addition, itwill also help for tailor-making novel silk-based composites with en-hanced properties that have broad applications in biomedical scienceand engineering.

Fig. 10. The interaction and miscibility mechanism of silk-PLA biocompositefilms. The silk protein content in the blend matrix can be used to control thephysiochemical and biological properties of the blend at the material interface.


11

Acknowledgements

This study was supported by the Rowan University Start-up andSeed Grants, NSF-MRI Program (DMR-1338014), NSF BiomaterialsProgram (DMR-1809541) and NSF Materials Eng. and Processing pro-gram (CMMI-1561966). FW is supported by the Nanjing NormalUniversity Scholarship for Overseas Studies Foundation of China(2013–2014), the college of Natural Science Foundation of JiangsuProvince (15KJB150018) as well as Nanjing Laboratory PlatformFoundation (1640703064).

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.msec.2019.109890.

References

[1] L. Tian, S.C. George, Biomaterials to prevascularize engineered tissues, J.Cardiovasc. Transl. Res. 4 (5) (2011) 685.

[2] J.S. Temenoff, A.G. Mikos, Tissue engineering for regeneration of articular carti-lage, Biomaterials 21 (5) (2000) 431–440.

[3] A. Shah, J. Brugnano, S. Sun, A. Vase, E.J.P.r. Orwin, The development of a tissue-engineered cornea: biomaterials and culture methods, Pediatr. Res. 63 (5) (2008)535.

[4] A.H. Huang, M.J. Farrell, R.L. Mauck, Mechanics and mechanobiology of me-senchymal stem cell-based engineered cartilage, J. Biomech. 43 (1) (2010)128–136.

[5] M. Nomi, A. Atala, P. De Coppi, S. Soker, Principals of neovascularization for tissueengineering, Mol. Asp. Med. 23 (6) (2002) 463–483.

[6] J.Y. Shu, B. Panganiban, T. Xu, Peptide-polymer conjugates: from fundamentalscience to application, Regenerative Biomaterials 64 (2013) 631–657.

[7] J.G. Hardy, T.R. Scheibel, Composite materials based on silk proteins, Prog. Polym.Sci. 35 (9) (2010) 1093–1115.

[8] X. Hu, P. Cebe, A.S. Weiss, F. Omenetto, D.L. Kaplan, Protein-based compositematerials, Mater. Today 15 (5) (2012) 208–215.

[9] P. Cebe, X. Hu, D.L. Kaplan, E. Zhuravlev, A. Wurm, D. Arbeiter, C. Schick, Beatingthe heat-fast scanning melts silk beta sheet crystals, Sci. Rep. 3 (2013) 1130.

[10] S. Mazzi, E. Zulker, J. Buchicchio, B. Anderson, X. Hu, Calorimetry, comparativethermal analysis of Eri, Mori, Muga, and Tussar silk cocoons and fibroin fibers, J.Therm. Anal. Calorim. 116 (3) (2014) 1337–1343.

[11] J. MacLeod, F. Rosei, Photonic crystals: sustainable sensors from silk, Nat. Mater. 12(2) (2013) 98–100.

[12] K.J. Koski, P. Akhenblit, K. McKiernan, J.L. Yarger, Non-invasive determination ofthe complete elastic moduli of spider silks, Nat. Mater. 12 (3) (2013) 262–267.

[13] Y.J. Yang, Y.S. Choi, D. Jung, B.R. Park, W.B. Hwang, H.W. Kim, H.J. Cha,Production of a novel silk-like protein from sea anemone and fabrication of wet-spun and electrospun marine-derived silk fibers, NPG Asia Materials 5 (6) (2013)e50 1–7.

[14] T.D. Sutherland, Y.Y. Peng, H.E. Trueman, S. Weisman, S. Okada, A.A. Walker,A. Sriskantha, J.F. White, M.G. Huson, J.A. Werkmeister, V. Glattauer,V. Stoichevska, S.T. Mudie, V.S. Haritos, J.A.M. Ramshaw, A new class of animalcollagen masquerading as an insect silk, Sci. Rep. 3 (2013) 2864.

[15] D.N. Rockwood, R.C. Preda, T. Yücel, X.Q. Wang, M.L. Lovett, D.L. Kaplan,Materials fabrication from Bombyx mori silk fibroin, Nat. Protoc. 6 (10) (2011)1612–1631.

[16] E.M. Pritchard, P.B. Dennis, F. Omenetto, R.R. Naik, D.L. Kaplan, Physical andchemical aspects of stabilization of compounds in silk, Biopolymers 97 (6) (2012)479–498.

[17] H.L. Zhu, X.X. Feng, H.P. Zhang, Y.H. Guo, J.Z. Zhang, J.Y. Chen, Structuralcharacteristics and properties of silk fibroin/poly (lactic acid) blend films, J.Biomater. Sci. Polym. Ed. 20 (9) (2009) 1259–1274.

[18] Q. Ma, B. Mao, P. Cebe, Chain confinement in electrospun nanocomposites: usingthermal analysis to investigate polymer–filler interactions, Polymer 52 (14) (2011)3190–3200.

[19] E. Steven, W.R. Saleh, V. Lebedev, S.F. Acquah, V. Laukhin, R.G. Alamo, J.S. Brooks,Carbon nanotubes on a spider silk scaffold, Nat. Commun. 4 (2013) 2435.

[20] A. Vasconcelos, G. Freddi, A. Cavaco-Paulo, Biodegradable materials based on silkfibroin and keratin, Biomacromolecules 9 (4) (2008) 1299–1305.

[21] Z. Fan, F. Zhang, T. Liu, B.Q. Zuo, Effect of hyaluronan molecular weight onstructure and biocompatibility of silk fibroin/hyaluronan scaffolds, Int. J. Biol.Macromol. 65 (2014) 516–523.

[22] E.M. Pritchard, D.L. Kaplan, Silk fibroin biomaterials for controlled release drugdelivery, Expert Opin. Drug Deliv. 8 (6) (2011) 797–811.

[23] Z.X. Cai, X.M. Mo, K.H. Zhang, L.P. Fan, A.L. Yin, C.L. He, H.S. Wang, Fabrication ofchitosan/silk fibroin composite nanofibers for wound-dressing applications, Int. J.Mol. Sci. 11 (9) (2010) 3529–3539.

[24] X. Hu, X.L. Wang, J. Rnjak, A.S. Weiss, D.L. Kaplan, Biomaterials derived fromsilk–tropoelastin protein systems, Biomaterials 31 (32) (2010) 8121–8131.

[25] M.P. Ho, H. Wang, K.T. Lau, J. Leng, Effect of silk fiber to the mechanical and

thermal properties of its biodegradable composites, J. Appl. Polym. Sci. 127 (4)(2013) 2389–2396.

[26] J. Qin, Y. Jiang, J. Fu, Y. Wan, R. Yang, W. Gao, H. Wang, Evaluation of drugrelease property and blood compatibility of aspirin-loaded electrospun PLA/RSFcomposite nanofibers, Iran. Polym. J. 22 (10) (2013) 729–737.

[27] H.T. Zou, C.H. Yi, L.X. Wang, H.T. Liu, W.L. Xu, Thermal degradation of poly (lacticacid) measured by thermogravimetry coupled to Fourier transform infrared spec-troscopy, J. Therm. Anal. Calorim. 97 (3) (2009) 929–935.

[28] H.Y. Cheung, K.T. Lau, X.M. Tao, D. Hui, A potential material for tissue engineering:silkworm silk/PLA biocomposite, Compos. Part B 39 (6) (2008) 1026–1033.

[29] G.B. Yin, Y.Z. Zhang, J.L. Wu, S.D. Wang, D.B. Shi, Z.H. Dong, Study on the elec-trospun PLA/silk fibroin-gelatin composite nanofibrous scaffold for tissue en-gineering, J. Fiber Bioeng. Inf. 2 (3) (2009) 182–188.

[30] C.J. Liao, C.F. Chen, J.H. Chen, S.F. Chiang, Y.J. Lin, K.Y. Chang, Fabrication ofporous biodegradable polymer scaffolds using a solvent merging/particulateleaching method, J. Biomed. Mater. Res. 59 (4) (2002) 676–681.

[31] M. Stoppato, H.Y. Stevens, E. Carletti, C. Migliaresi, A. Motta, R.E. Guldberg, Effectsof silk fibroin fiber incorporation on mechanical properties, endothelial cell colo-nization and vascularization of PDLLA scaffolds, Biomaterials 34 (19) (2013)4573–4581.

[32] H. Kwon, L. Sun, D.M. Cairns, R.S. Rainbow, R.C. Preda, D.L. Kaplan, L. Zeng, Theinfluence of scaffold material on chondrocytes under inflammatory conditions, ActaBiomater. 9 (5) (2013) 6563–6575.

[33] K. Numata, Y. Doi, Dual biosyntheses of poly [(R)-3-hydroxybutyric acid] and silkprotein for the fabrication of biofunctional bioplastic, Polym. J. 43 (7) (2011) 642.

[34] L. Meinel, D.L. Kaplan, Silk constructs for delivery of musculoskeletal therapeutics,Adv. Drug Deliv. Rev. 64 (12) (2012) 1111–1122.

[35] W.S. Muller, L.A. Samuelson, S.A. Fossey, D.L. Kaplan, Formation and character-ization of Langmuir silk films, Langmuir 9 (7) (1993) 1857–1861.

[36] F. Wang, G. Guo, Q. Ma, M. Gu, X. Wu, S. Sheng, X. Wang, Calorimetry, in-vestigation on the thermo-mechanical properties and thermal stability of polylacticacid tissue engineering scaffold material, J. Therm. Anal. Calorim. 113 (3) (2013)1113–1121.

[37] X. Hu, D. Kaplan, P. Cebe, Determining beta-sheet crystallinity in fibrous proteinsby thermal analysis and infrared spectroscopy, Macromolecules 39 (18) (2006)6161–6170.

[38] H.J. Jin, D.L. Kaplan, Mechanism of silk processing in insects and spiders, Nature424 (6952) (2003) 1057–1061.

[39] M. Pyda, B. Wunderlich, Reversing and nonreversing heat capacity of poly (lacticacid) in the glass transition region by TMDSC, Macromolecules 38 (25) (2005)10472–10479.

[40] S. Essa, J.M. Rabanel, P. Hildgen, Characterization of rhodamine loaded PEG-g-PLAnanoparticles (NPs): effect of poly (ethylene glycol) grafting density, Int. J. Pharm.411 (1–2) (2011) 178–187.

[41] C. Huang, M. Moosmann, J. Jin, T. Heiler, S. Walheim, T. Schimmel, Polymer blendlithography: a versatile method to fabricate nanopatterned self-assembled mono-layers, Beilstein J. Nanotechnol. 3 (1) (2012) 620–628.

[42] R.M. Rasal, A.V. Janorkar, D.E. Hirt, Poly (lactic acid) modifications, Prog. Polym.Sci. 35 (3) (2010) 338–356.

[43] V.S. Bisaria, T.K. Ghose, M. Technology, Biodegradation of cellulosic materials:substrates, microorganisms, enzymes and products, Enzym. Microb. Technol. 3 (2)(1981) 90–104.

[44] S.K. Mohan, T. Srivastava, Microbial Deterioration and Degradation of PolymericMaterials, vol. 2(4), (2010), pp. 210–215.

[45] V. Venkataraman, T. Duda, S. Ravichandran, R.K. Sharma, Neurocalcin δ modula-tion of ROS-GC1, a new model of Ca2+ signaling, Biochemistry 47 (25) (2008)6590–6601.

[46] P.J. Flory, Am, Chem. Sac. 63 (1941) 3091 Principles of Polymer Chemistry, CornellUniversity Press, Ithaca, New York, USA, 1953, pp. P70–P75.

[47] Y. Miwa, K. Usami, K. Yamamoto, M. Sakaguchi, M. Sakai, S. Shimada, Direct de-tection of effective glass transitions in miscible polymer blends by temperature-modulated differential scanning calorimetry, Macromolecules 38 (6) (2005)2355–2361.

[48] F. Wang, L. Zhou, J. Zhou, X. Gu, Y. Feng, Calorimetry, characterization of antic-ancer hypocrellin A encapsulated with silica nanoparticles, J. Therm. Anal. Calorim.102 (1) (2010) 69–74.

[49] H. Chen, M. Pyda, P. Cebe, Non-isothermal crystallization of PET/PLA blends,Thermochim. Acta 492 (1–2) (2009) 61–66.

[50] T.G. Fox Jr., P.J. Flory, Second-order transition temperatures and related propertiesof polystyrene. I. Influence of molecular weight, J. Appl. Phys. 21 (6) (1950)581–591.

[51] T. Kwei, E.M. Pearce, J.R. Pennacchia, M.R. Charton, Correlation between the glasstransition temperatures of polymer mixtures and intermolecular force parameters,Macromolecules 20 (5) (1987) 1174–1176.

[52] M. Gordon, J.S. Taylor, Ideal copolymers and the second-order transitions of syn-thetic rubbers. I. Non-crystalline copolymers, J. Appl. Chem. 2 (9) (1953) 493–500.

[53] X. Hu, D. Kaplan, P. Cebe, Effect of water on the thermal properties of silk fibroin,Thermochim. Acta 461 (1–2) (2007) 137–144.

[54] D. Liu, Q. Shen, H. Ding, L. Zhong, Influence of the components ratio on cellulose/poly (ethylene glycol) blend related T g and a model, J. Cell. Sci. Technol. 14 (2)(2006) 40–45.

[55] V. Vichaibun, M. Chulavatnatol, A new assay for the enzymatic degradation ofpolylactic acid, ScienceAsia 29 (2003) 297–300.

[56] E.M. Pritchard, X. Hu, V. Finley, C.K. Kuo, D.L. Kaplan, Effect of silk protein pro-cessing on drug delivery from silk films, Macromol. Biosci. 13 (3) (2013) 311–320.

[57] S.H. Lee, I.Y. Kim, W.S. Song, Biodegradation of polylactic acid (PLA) fibers using


12



http://refhub.elsevier.com/S0928-4931(19)31182-8/rf0005




















































































































































different enzymes, Macromol. Res. 22 (6) (2014) 657–663.[58] Q. Lu, X. Hu, X. Wang, J.A. Kluge, S. Lu, P. Cebe, D.L. Kaplan, Water-insoluble silk

films with silk I structure, Acta Biomater. 6 (4) (2010) 1380–1387.[59] Y.S. Nam, T.G. Park, Biodegradable polymeric microcellular foams by modified

thermally induced phase separation method, Biomaterials 20 (19) (1999)1783–1790.

[60] F.L. Mi, Y.M. Lin, Y.B. Wu, S.S. Shyu, Y.-H. Tsai, Chitin/PLGA blend microspheres asa biodegradable drug-delivery system: phase-separation, degradation and releasebehavior, Biomaterials 23 (15) (2002) 3257–3267.

[61] X. Hu, Q. Lu, L. Sun, P. Cebe, X. Wang, X. Zhang, D.L. Kaplan, Biomaterials fromultrasonication-induced silk fibroin−hyaluronic acid hydrogels,Biomacromolecules 11 (11) (2010) 3178–3188.

[62] B. Chehroudi, T. Gould, D.M. Brunette, Titanium-coated micromachined grooves ofdifferent dimensions affect epithelial and connective-tissue cells differently in vivo,J. Biomed. Mater. Res. 24 (9) (1990) 1203–1219.

[63] Q. Cai, J. Yang, J. Bei, S. Wang, A novel porous cells scaffold made of poly-lactide–dextran blend by combining phase-separation and particle-leaching tech-niques, Biomaterials 23 (23) (2002) 4483–4492.

[64] H.D. Kim, E.H. Bae, I.C. Kwon, R.R. Pal, J.D. Nam, D.S. Lee, Effect of PEG–PLLAdiblock copolymer on macroporous PLLA scaffolds by thermally induced phaseseparation, Biomaterials 25 (12) (2004) 2319–2329.

[65] X. Chen, W. Wu, Z. Guo, J. Xin, J. Li, Controlled insulin release from glucose-

sensitive self-assembled multilayer films based on 21-arm star polymer,Biomaterials 32 (6) (2011) 1759–1766.

[66] X. Wang, T. Yucel, Q. Lu, X. Hu, D.L. Kaplan, Silk nanospheres and microspheresfrom silk/PVA blend films for drug delivery, Biomaterials 31 (6) (2009) 1025–1035.

[67] B. Kong, J.S. Choi, S. Jeon, I.S. Choi, The control of cell adhesion and detachmenton thin films of thermoresponsive poly[(N-isopropylacrylamide)-r-((3-(methacry-loylamino) propyl)-dimethyl (3-sulfopropyl) ammonium hydroxide)], Biomaterials30 (29) (2009) 5514–5522.

[68] H.Y. Yu, F. Wang, Q.C. Liu, Q.Y. Ma, Z.G. Gui, Structure and kinetics of thermaldecomposition mechanism of novel silk fibroin films, Acta Phys. -Chim. Sin. 33 (2)(2017) 344–355.

[69] K. Hamad, M. Kaseem, M. Ayyoob, J. Joo, F. Deri, Polylactic acid blends: the futureof green, light and tough, Prog. Polym. Sci. 85 (2018) 83–127.

[70] B. Zhong, Z.Y. Al-Saigh, Characterization of biodegradable polymers by inverse gaschromatography. III. Blends of amylopectin and poly(l-lactide), J. Appl. Polym. Sci.123 (2012) 2616–2627.

[71] F. Wang, S.S.J. Aravind, H. Wu, J. Forys, V. Venkataraman, K. Ramanujachary,X. Hu, Tunable green graphene-silk biomaterials: mechanism of protein-based na-nocomposites, Mater. Sci. Eng. C 79 (2017) 728–739.

[72] F. Wang, P. Chandler, R. Oszust, E. Sowell, Z. Graham, W. Ardito, X. Hu, Thermaland structural analysis of silk–polyvinyl acetate blends, J. Therm. Anal. Calorim.127 (2017) 923–929.


13













































materials science & engineering c

Documents