Article pubs.acs.org/Biomac
Nanofibers of Elastin and Hydrophilic Segmented Polyurethane Solution Blends Show Enhanced Mechanical Properties through Intermolecular Protein−Polymer H Bonding Markus Heiny and V. Prasad Shastri* Institute for Macromolecular Chemistry, Hermann Staudinger Haus, University of Freiburg, Stefan-Meier Str. 31, 79104 Freiburg, Germany BIOSS Centre for Biological Signalling Studies, University of Freiburg, Schänzlestr. 18, 79104 Freiburg, Germany S Supporting Information *
ABSTRACT: Combining mechanical properties with enhanced cell interaction is highly desirable in a biomaterial. In this study, a new paradigm for enhancing the mechanical properties of segmented polyurethanes (SPUs) through solution blending with a biopolymer is presented. This noncovalent approach is based on the premise that molecular level blending of SPUs rich in hydrogen bonding (H bonding) domains with a biopolymer capable of H bonding will promote H-bond bridges between the components, leading to molecular annealing and modification of the physicochemical properties of the SPU. We demonstrate that by solution-blending solubilized elastin with a triblock copolymer-derived SPU, a 5-fold increase in tensile modulus of electrospun constructs of the SPU can be achieved, with concomitant enhancement in human endothelial cell attachment. Spectroscopic and calorimetric analysis confirm the role of H bonding in the enhancement, thus providing the impetus to further explore blending with biopolymers as a means of improving the property profiles of synthetic polymeric biomaterials.
■
INTRODUCTION
and mammalian origin, proteins such as collagen, as well as synthetic peptide-amphiphiles have been explored extensively for cell and tissue contacting applications.5−7 Elastin, an abundant extracellular matrix protein that is responsible for the elastic behavior of several tissues including skin and blood vessels has gained prominence of late8,9 and has been explored as dressings for burn or chronic wounds10,11 and in scaffolds for tissue regeneration.12 The solubilized form of elastin is believed to enhance elastin formation in vivo13,14 and can have a positive impact on wound healing. Combining the advantages of both synthetic polymers and biological materials therefore is an increasingly important strategy to create biomaterials that can be specifically tailored for a biomedical application. When two materials are blended, intermolecular interactions can result in unexpected outcomes in materials properties. For
In recent years, there has been an increased interest in functional materials that combine elastic properties with biodegradability. Such materials can be utilized in various healthcare-related applications including topical wound dressings and vascular grafts.1,2 In the context of materials for replacement or mimicking of tissues, elastic properties are important and in many instances, such as for vascular grafts, a requirement. Therefore, much effort has been placed on the synthesis of polymers that show good elastic behavior. Among biocompatible polymers, segmented polyurethanes (SPUs) derived from degradable building blocks are particularly attractive for cardiovascular applications as elasticity, blood biocompatibility,3 and biodegradation can be combined in a single system. Therefore, much effort has focused on combining elastic properties with biodegradation, for example “bio-rubber”, a biodegradable elastic polymer derived from glycerol and sebacic acid.4 In addition to synthetic polymers, many biopolymers including polysaccharides of both marine © XXXX American Chemical Society
Received: December 13, 2015 Revised: March 1, 2016
A
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules
Poly(ethylene glycol) (PEG, average mol wt 1450 g mol−1) and 1,4butanediol (99%) were purchased from Sigma-Aldrich, Germany and dried in vacuo prior to use. Hexamethylene diisocyanate (HMDI, ≥99.0%), tin(II) 2-ethylhexanoate (SnOct2, 95%), 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, ≥99.0%), and glutaraldehyde solution (grade I, 25%) were purchased from Sigma-Aldrich, Germany and used as received. DMSO (p.a.) was purchased from Merck, Germany and used as received. Toluene (p.a., Carl Roth, Germany) was distilled from sodium (Sigma-Aldrich, Germany) and stored over 3 Å molecular sieves. Elastin (from bovine neck ligament, Sigma-Aldrich, Germany) was solubilized according to standard procedures19 by stirring in 1 M KOH (Merck, Germany) in EtOH/H2O 4:1 (v/v) for 48 h at 37 °C and subsequently neutralized with acetic acid and dialyzed against H2O. Tetrahydrofuran (Carl Roth, Germany), diethyl ether, and methanol (VWR, Germany) were used without further treatment. Cell culturing supplements and viability assays consisted of Vasculife EnGS Basal Medium (Lifeline, CellTech), penicillin− streptomycin−amphotericin B (Pan Biotech), RPMI 1640 (Life Technologies), and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT, Sigma-Aldrich). Circular Dichroism. Circular dichroism (CD) spectra were obtained using a J-810 spectropolarimeter (Jasco) equipped with a Pelletier temperature cell PFD-425S (Jasco). Samples were prepared as solutions in HFIP at concentrations of 50 μg mL−1 and measured in quartz cuvettes (path length: 10 mm) at a temperature of 4 °C under a N2 atmosphere. Each spectrum is the average of three samples. Differential Scanning Calorimetry (DSC). Thermal transitions were determined on a Pyris 1 differential scanning calorimeter (PerkinElmer). Polymer samples (∼10 mg) were sealed in aluminum pans and measured under a nitrogen atmosphere. Indium and cyclohexene were used for temperature calibration. Samples were subjected to a heat−cool−heat cycle ranging from −50 to 180 °C with a heating/cooling rate of 10 °C min−1. Second heating cycle was used for data interpretation with Pyris Manager Version 8.0.0.0172 (PerkinElmer). Fourier Transform Infrared (FTIR) Spectroscopy. Infrared (IR)spectra were recorded on a Vector 22 spectrometer (Bruker; using OPUS software version 5.5) equipped with a LADTGS detector (12.000−360 cm−1) using 30 scans/sample under nitrogen purging. Samples were prepared by casting a thin film of the polymer/blend from HFIP on a freshly polished KBr crystal window previously cleaned with reagent-grade acetone and by purging with pressurized air to remove any dust particles. Gel Permeation Chromatography (GPC). Analyses of molecular weight and polydispersity of the synthesized polymers were carried out applying a 1200 Series GPC-SEC analysis system consisting of pump G1310A, RI detector G1362A, and autosampler G1329A (Agilent
example, lipophilic interactions15 and hydrogen bonding (H bonding)16 have been exploited extensively in drug delivery to alter the transport properties of drugs across biological barriers.17,18 In the context of advanced biomaterials, exploitation of molecular-level interactions between synthetic and biological polymers to enhance mechanical properties of a biomaterial composite remains a fairly unexplored realm. We theorized that synthetic polymers featuring domains with extensive H-bonding capacity could form intermolecular bridges with the protein backbone (Scheme 1) and thereby Scheme 1. Potential H-Bonding Interactions between Synthetic SPUs (above) and Elastin Segments (below) in Hydrophilic (light blue) and Hydrophobic (light red) Segments of SPU/Elastin Blends
influence the material properties of a synthetic polymer− protein blend. To test this hypothesis we processed solution blends of SPUs consisting of hydrophilic poly(ethylene glycol) (PEG) and hydrophobic poly(L-lactic acid) (PLLA) domains (Scheme 2) with solubilized elastin into nanofibers using electrospinning and made the surprising finding that the tensile modulus of blend nanofibers exceeds that of both the SPU and elastin by nearly an order of magnitude. Using spectroscopy and thermal analysis we show that this enhancement is afforded by H-bonding interactions between the PLA segment in the SPU and elastin. Additionally, molecular blends of SPU and elastin show enhanced interaction with human endothelial cells, thus making them suitable for cardiovascular applications.
■
EXPERIMENTAL SECTION
Materials. (3S)-cis-3,6-Dimethyl-1,4-dioxane-2,5-dione (L-lactide, 98%) was purchased from Sigma-Aldrich, Germany and recrystallized from ethyl acetate (>99.5%, Carl Roth, Germany) prior to use.
Scheme 2. Synthetic Scheme and Structure of Synthesized SPUs
B
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules Technologies), which was equipped with SDVB columns (5 μm, 100 Å, 1000 Å, 10 000 Å) (Polymer Standard Service). Samples were prepared at a concentration of 4 mg mL−1 in THF, filtered (0.45 μm filter), and eluted using THF at a flow rate of 1 mL min−1 at 30 °C. Calibration was done using polystyrene EasiVial PS-H Tri-Pack with a nominal range of 162−6 000 000 g mol−1 (Agilent Technologies). Elugrams were analyzed using WinGPC Unity, Build 5403 (Polymer Standard Solutions) according to the manufacturer’s guidelines and ISO 13885. Proton Nuclear Magnetic Resonance (1H NMR) Spectroscopy. Samples were measured on an ARX 300 MHz spectrometer (Bruker) at 25 °C using CDCl3 as a solvent. Chemical shift in ppm was referred to the residual solvent peak (δ 7.26 for CHCl3). Data analysis was carried out using TopSpin 3.0 software (Bruker). Scanning Electron Microscopy. All scanning electron microscopy (SEM) micrographs (both fiber samples and samples with adhered cells) were recorded after gold sputter coating on a Quanta 250 FEG microscope (FEI) with an accelerating voltage of 20 kV in high vacuum. Average fiber diameters and diameter distributions of electrospinning samples were determined using ImageJ (NIH) by measuring the diameters of 70 random fibers on SEM micrographs taken at three different random positions per sample. Thermogravimetric Analysis (TGA). Weight loss of polymer and blend samples was recorded with samples of 5 to 10 mg in a temperature range from 30 to 650 °C under an air flow of 20 mL min−1 using a TGA 4000 device (PerkinElmer). Uniaxial Tensile Testing. Tensile tests were carried out on a Zwick Z005 universal testing machine (Zwick Roell) at a uniaxial test speed of 10 mm min−1 at 25 °C with test specimen of uniform thicknesses of 150 μm cut into dimensions of 5 × 25 mm2. Viscosity Measurements. Solution viscosities of the electrospinning solutions were determined on an MCR 301 rheometer (Anton Paar) equipped with a cone−plate geometry (25 mm diameter, 2° cone angle) using RheoPlus software. The measurements were carried out at 20 °C as rotation frequency sweeps in the range of 0.001−1000 s−1. Polymer Synthesis. In a typical procedure, predried PEG (2.90 g, 2.00 mmol) and L-lactide (10−40 equiv, 20.0−80.0 mmol) were stirred under vacuum for 4 h. Anhydrous toluene (40 mL) was added and the reaction mixture was heated to 90 °C. After complete dissolution of the reagents, SnOct2 (1 wt % of L-lactide) was added and the reaction mixture was stirred at 90 °C for 12 h. The flask was cooled to 75 °C and HMDI (2 equiv, 673 mg, 4.00 mmol) was added and the reaction mixture was stirred for an additional 2.5 h, following which the temperature was raised to 90 °C and 1,4-butanediol (1 equiv, 181 mg, 2.00 mmol) was added. The temperature was kept at 90 °C for 30 min, followed by stirring at room temperature for 12 h, after which the resulting polymer was precipitated in diethyl ether, washed with diethyl ether and methanol, and dried in vacuo at room temperature to constant weight. Polymers are referred to as SPUxL, with x being the equivalent of L-lactide applied in the polymerization. Solution Blending. Solution blends were realized by dissolving SPU30L and solubilized elastin in the respective composition in HFIP under stirring at room temperature until homogeneous. Blends are denominated by stating the composition (in wt %) after PUE; thus PUE8020 stands for a blend consisting of 80 wt % SPU30L and 20 wt % solubilized elastin. Electrospinning. Nonwoven fiber mats were obtained by electrospinning (ES) of SPU or PUE blends. The polymer or blend solutions were loaded into a plastic syringe (B. Braun Melsungen AG) with a high voltage supply (Gamma High Voltage Research) connected to its blunt 21G stainless-steel canula. A continuous flow of the polymer/blend solution (feeding rate) was realized by means of a syringe pump (Cole-Parmer). Fibers were collected on a grounded metal sheet target. Solvent, concentration c of the polymer solution, feeding rate Q of the polymer solution, electrical potential E, and the distance d between needle tip and target were subject to optimizations. Specific spinning conditions were as follows. For SPU30L from THF: c = 20 wt %, E = 10 kV, Q = 1.25 mL h−1, d = 20 cm; for SPU30L from toluene: c = 17.5 wt %, E = 16 kV, Q = 0.75 mL h−1, d = 15 cm; for
PUE8020 blend from HFIP: c = 10 wt %, E = 12.4 kV, Q = 0.40 mL h−1, d = 13 cm; for PUE7030 blend: c = 10 wt %, E = 10 kV, Q = 0.40 mL h−1, d = 13 cm. Glutaraldehyde Cross-Linking. PUE scaffolds were cross-linked by immersing them into glutaraldehyde vapor produced from an excess of fresh 25% glutaraldehyde solution in a sealed glass desiccator for 4 h at room temperature. The scaffolds were subsequently stored in vacuo at room temperature for at least 48 h to remove remaining glutaraldehyde. Cell Culture Experiments. Human umbilical vein endothelial cells (HUVECs) were cultured in Vasculife EnGS Basal Medium supplemented with 1% penicillin−streptomycin−amphotericin B in a humidified incubator at 5% CO2/37 °C. For MTT assays, cells were seeded at a density of 10 000 cells/well in standard TC plates (96 wells) and coincubated with EtOH-sterilized and PBS-washed fiber mats of SPU30L and SPU40L (cut into squares of 50 × 50 mm2). The culture medium was removed after the indicated time span (1, 2, or 4 days), and the cells were incubated with RPMI 1640 containing 0.25 mg mL−1 of MTT for 3 h. The resultant Formazan crystals were dissolved in DMSO, and the absorbance at 550 nm was measured on a Synergy HT plate reader (BioTek). For adhesion tests, HUVECs were seeded at a density of 50 000 cells/well on EtOH-sterilized and PBSwashed fiber mats that were kept immersed in the wells of standard TC plates (24 wells) by means of sterile stainless-steel rings. After 3 h, the culturing medium was removed, followed by thoroughly washing the scaffolds with PBS. The samples were either used for MTT assays (same procedure as previously stated) or prepared for SEM microscopy by dehydration of the adhered cells in a series of graduated ethanol solutions (25, 50, 75, 90, and 100% (20 min each)) and drying at room temperature.
■
RESULTS AND DISCUSSION SPU Synthesis. To explore our hypothesis, we decided to use segmented polyurethanes (SPUs) based on three considerations: (1) The simple synthesis approach allows for incorporation of segments with well-defined length (molecular weight) and physicochemical characteristics along the polymer backbone for interaction with proteins. (2) SPUs have a history of use in cardiovascular applications and therefore any advancement made would have a direct relevance. (3) The polymer backbone architecture can be tailored to possess Hbonding possibilities as part of both hydrophobic and hydrophilic domains to maximize interactions between the elastin backbone and the synthetic SPU (cf. Scheme 1). We consequently chose to base our SPU design on PLLA−PEG− PLLA triblock copolymers20−22 as they offer distinct structural features that might be exploited in promoting molecular level interactions between the SPU and elastin, namely, juxtaposition of the helical structure of the PLLA next to the extensive Hbonding capacity of both the urethane and PEG domains. Additionally, the degradability of PLLA23 and the hydrophilicity and perceived low protein adsorption of PEG24 offer further advantages with regards to in vivo use. A straightforward one-pot procedure to synthesize SPUs bearing these structural motifs was developed based on classical PU synthesis. In the first step, PLLA−PEG−PLLA triblock macrodiols were synthesized by the SnOct2-catalyzed polymerization of L-lactide using the hydroxyl groups of short-chainlength polyethylene glycol (MW of 1450 g mol−1; n = 33) as initiating moieties, which were then converted in the second step into isocyanate-terminated prepolymers using 1,6-hexamethylene diisocyanate, followed by a chain extension step using 1,4-butanediol to increase the molecular weight (Scheme 2). By varying the PEG/lactide feed ratio in the monomer mixture, a series of SPUs containing PEG segments of fixed chain length and varied molecular weights of the PLLA blocks were realized C
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules Table 1. Composition and Properties of the Synthesized SPUs
a
polymer
feed ratio PEG/lactide
Mn (g mol−1)a
Mw (g mol−1)a
PDIa
PEG/lactide ratio in polymerb
Tg (°C)c
Tm,1 (° C)c
Tm,2 (° C)c
SPU10L SPU20L SPU30L SPU40L
1:10 1:20 1:30 1:40
30 000 38 000 43 000 69 000
56 000 84 000 86 000 162 000
2.1 2.2 1.9 2.3
1:9 1:21 1:30 1:41
−28 1 10 20
75 86 126
127 137
Determined by GPC. bDetermined by 1H NMR spectroscopy. cDetermined by DSC.
at high yields (≥90%). The progression of polymerization and the final chemical composition of the SPUs were confirmed using 1H NMR spectroscopy (Figure S1 in the Supporting Information) and the well-defined composition of the constituting macrodiol segments was further verified by GPC (Table S1 in the Supporting Information). A close correspondence between prescribed PEG/lactide ratio and experimental values was observed (Table 1), and in all ratios SPUs with molecular weights suitable for processing by ES (based on the literature data on chain entanglement) were obtained. SPUs of all lactide block lengths showed low glasstransition temperatures (Tg) below 20 °C. Melting transitions (Tm) were only observed in the SPUs with lactide block lengths higher than 20 units. The lower transition in SPU20L and SPU30L cannot be attributed to the melting of pure PEG domains because for the molecular weight of 1450 g mol−1 employed here a lower melting temperature of ∼45 °C should be expected. We assume that more complex cooperative crystallization and recrystallization processes25 might be at play, but further efforts have to be undertaken to understand the contributing factors. For SPU30L and SPU40L, distinct melting temperatures in the range of 127−137 °C could be detected and attributed to the melting of the PLLA domains. (DSC thermograms can be found in Figure S2 in the Supporting Information.) SPU and Elastin Fragments Exhibit Molecular Level Interactions. SPU30L was chosen as the model polymer system for further elucidation, as this composition exhibited crystallization of lactide domains (as previously discussed) and additionally showed good fiber formation using ES (as discussed later). Because the SPU contains optically active domains in the PLLA segment, any changes to either the secondary structure of elastin or PLLA would be a good first indicator of potential molecular level interactions between the two blend components. To investigate this we applied CD spectroscopy, a technique commonly used to study secondary structure of optically active macromolecules such as proteins26,27 and polysaccharides.7 CD spectra of SPU30L containing elastin in weight ratios ranging from 10 to 40 wt % were recorded using 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) as the solvent. The choice of HFIP was based on two considerations: (1) HFIP was the only solvent capable of solubilizing both elastin and SPU30L in concentrations sufficiently high for fiber formation using ES and (2) HFIP has been known to promote favorable protein−protein interactions.28 (Both aspects are discussed later in some detail in the Electrospinning section). Because UV absorption spectra of all blend compositions showed absorption maxima around 203 nm (for the solubilized elastin) and below (cf. Figure S3 in the Supporting Information), we confine the interpretation of the obtained CD spectra to above this wavelength. The spectrum of the pure solubilized elastin showed a negative ellipticity with a minimum around λ = 220 nm (Figure 1A). Because of the fact that the CD spectrum of solubilized
Figure 1. (A) CD spectra of PUE blends and blend components. (B) Comparison of theoretically expected and observed ellipticity at λ = 212 nm.
elastin in HFIP has not been reported in the literature, one cannot draw definitive conclusion on the secondary structure of the protein fragments in that solvent. Nevertheless, on the basis of CD data of respective compounds in aqueous systems29,30 in combination with the known fact that HFIP stabilizes helical conformations in proteins,31 we can assume that the solubilized elastin exhibits a predominantly unordered structure with a few helical domains. In contrast, the CD spectra of SPU30L showed a maximum at λ = 210 nm. This observed positive Cotton effect can most likely be attributed to the n-π* transition of the PLLA ester groups and strongly indicates an α-helical conformation of the PLLA domains of the SPU;32−34 however, the lack of detailed understanding of PLLA behavior in HFIP hampers a conclusive characterization of the polymer secondary structure. The CD spectrum of two components in solution is a proportional sum of the two underlying spectra. On the basis of this theoretical consideration, we compared the theoretically expected ellipticities at a representative wavelength of 212 nm (calculated from the ellipticities of the pure blend components) to the experimentally observed values at that wavelength (Figure 1B). Interestingly, although one might ascribe, to a certain extent, the expected linearity to the observed ellipticities, especially in the case of the PEU8020 blend, a D
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules
elastin showed its major denaturation transitions in a temperature range between 250 and 400 °C, which is close to the degradation behavior reported for native elastin.37 The pure SPU30L on the contrary experienced its major decomposition at an onset temperature of ∼250 °C (Figure 2A).
clear deviation from the theory could be observed. The particular behavior of this blend composition manifested itself in the melting point depression as well and is discussed in some detail later under thermal behavior. In summary, the studied CD spectra of the solution blends provided first evidence of molecular level interactions between SPU chains and elastin fragments in solution. SPU and Elastin Fragments Show H-Bonding in Solid State. To further gain insights into the molecular basis of the interactions observed between SPU and elastin in solution, we studied cast films of the blends using IR spectroscopy. This method can provide evidence of the presence of H-bonding interactions that can be discerned via frequency shifts and broadening of the absorption bands of the functional groups contributing to the H bonds.35 The resultant effect can most clearly be observed for the N−H and the CO stretching vibrations resulting from the urethane and ester groups. The N−H absorption band of the pure SPU30L consisted of a broad signal centered at 3360 cm−1, and the CO stretching vibration was located at 1768 cm−1. Upon blending with elastin these bands were shifted by 60 and 10 cm−1, respectively (Table 2, IR spectra can be found in Figure S4 in the Table 2. Infrared Absorption Bands in SPU/Elastin Molecular Hybrids system
N−H stretching (cm−1)
CO stretching (cm−1)
SPU30L PUE9010 PUE8020 PUE7030 PUE6040 sol. elastin
3360 3293 3290 3302 3295 3305
1768 1760 1758 1760 1759 1659
Supporting Information). In the case of the N−H band an exclusive attribution of the observed shift to H-bonding interactions proved difficult due to overlaying effects resulting from contributions of spectral components of the elastin content in the blends (Figure S3B); however, between 1725 and 1800 cm−1, where the SPU shows strong absorption, the elastin molecules exhibit no absorption. Hence the frequency shift in the CO band can be unequivocally attributed to the SPU, and a clear role for H bonds in the observed shifts can be inferred (Figure S3C). All in all, the observed frequency shifts are consistent with literature reports on the effect of H bonds on IR absorption behavior36 and provide strong evidence of a prominent role for H bonds in the interaction of SPU with solubilized elastin, as hypothesized. Blending of SPU with Elastin Yields Materials with Altered Thermal Properties. Having confirmed our hypothesis regarding molecular level interactions between the SPU and solubilized elastin, we were interested in knowing as to what extent these interactions manifest themselves in macroscopic properties of the materials. In this context, thermal decomposition behavior and melting behavior are readily explored properties on which the H-bonding interactions between the two blend components ought to exert a detectable influence. We therefore subjected the blends and their virgin components to TGA and DSC. As far as thermal decomposition is concerned, proteins are known to undergo denaturation upon thermal treatment. Next to the initial weight loss occurring between 50 and 80 °C, which can be ascribed to the removal of water, the solubilized
Figure 2. Analysis of the thermal properties of the blend materials: (A) Representative TGA profiles of blend components and PUE8020 blend (remaining TGA profiles can be found in Figure S5 in the Supporting Information). (B) Representative 1st derivatives of the above TGA profiles. (C) Difference between experimental and theoretical inflection points as a function of the elastin content in the blends.
To assess the envisioned impact of molecular level interactions on the thermal decomposition, we examined the decomposition behavior of the blend materials by means of an analysis of the detected minima in the first derivative of the weight loss profiles. Such a minimum in the first derivative of a TGA curve, the inflection point Ti, represents the temperature where the highest rate of weight loss occurs. As a first step, we calculated theoretical weight-loss profiles for the blends based on the assumption that a mixture of two materials showing no cooperative interactions between the components would experience simple additive behavior of the underlying curves (Figure 2A,B and Figure S5 in the Supporting Information). We then compared the (theoretically expected) inflection points of these curves (Ti,theor) to the experimentally E
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules
obtained by CD where the largest deviation from the theoretically expected ellipticity was observed in this blend composition as well. Because circular dichroism is impacted by the electronic environment of the optically active species, this suggests that the blend components in the PUE8020 have very strong molecular-level interactions. We theorize that this unique outcome in the case of the PUE8020 might represent an optimum in terms of configuration and chemical composition of the interacting species. Electrospinning of SPU and SPU-Elastin Blends. In regards to exploiting blends of synthetic polymers and proteins in biomedical and other applications such as membranes for filtration, ES provides a unique platform to generate nonwoven micro- and nanofibrous 2D and 3D structures.38,39 Additionally, this technique allows for leveraging the molecular interactions observed between SPU and elastin, as polymer fibers can be processed directly from solutions and the elongation and thinning of the polymer fibers during jet thinning can promote polymer−polymer interaction and chain alignment.40,41 Both factors can further enhance potential cooperative interactions between the two blend components and might therefore provide for a optimal translation of the observed molecularlevel interactions into macroscopic material properties. To assess the general possibility to process SPUs into fibers using ES, we carried out preparatory ES experiments with THF and toluene as common solvents. Under optimized conditions, smooth fibers with average fiber diameters in the range of several micrometers for the THF systems (3−5 μm) and below 1 μm (0.8−1 μm) for the toluene systems were obtained; however, most promising results were achieved with the SPU30L system, and for this polymer, average fiber diameters were 3 μm and 800 nm for THF and toluene, respectively (Figure 4A,B). Because it is evident that fiber formation from pure SPUs is strongly dependent on solvent properties, it is reasonable to assume that during the ES of SPU/elastin solutions a faster evaporation of the solvent can promote more thermodynamically unstable interactions such as the theorized H bonding between the SPU backbone and elastin. Good solubility of both blend components in the solvent is therefore essential to ensure mixing of the two phases during all stages of ES. Furthermore, molecular interactions between two macromolecules in solution is governed by their solvation state, and in this regard, HFIP, a highly volatile solvent, was found to be
determined values of the blend materials (Ti,exp). The actual values Ti,exp of the blend materials were without exception higher than the calculated Ti,theor. Depending on the elastin content, differences (Ti,exp − Ti,theor) of up to 37 °C were observed (Figure 2C). This shift of the major thermal transition toward higher temperatures is a noticeable deviation from the behavior expected for noninteracting blend components and provides another direct evidence of cooperative interactions between the SPU and the solubilized elastin. Furthermore, this is the first example to show that the thermal properties of a synthetic polymer have been enhanced by incorporation of a protein. To detect changes in the thermal transitions of the materials, we subjected the blends and their virgin components to DSC measurements. Interaction in the form of H bonds between two polymeric blend components is known to result in a depression of the melting temperature.35 Consistent with the hypothesized presence of strong interactions between the two blend components, upon blending with elastin, we were able to detect a decrease of 8 and 5 °C in Tm,1 and Tm,2, respectively, of the SPU30L (Figure 3). In the case of the PUE8020 blend, a
Figure 3. DSC thermograms of blends and blend components.
slightly larger shift in the melting temperatures (11 and 8 °C for Tm,1 and Tm,2, respectively) was observed (cf. dashed vertical line in Figure 3), and this is in accordance with the results
Figure 4. Representative SEM micrographs and fiber diameter frequency distributions of electrospun SPUs/SPU-elastin blends: (A) SPU30L electrospun from THF, (B) SPU30L electrospun from toluene, (C) PUE8020 blend electrospun from HFIP, (D) PUE7030 blend electrospun from HFIP, and (E) freestanding electrospun fibrous mat and tubular scaffold prepared by ES of PUE8020 blend. F
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules capable of solubilizing both the SPU and elastin in concentrations suitable for ES. HFIP has been utilized extensively for ES of both natural and synthetic polymers42,43 and has been shown to induce and preserve helical conformations of proteins such as collagen31 and promote the hierarchical organization of collagen during ES.28 On the basis of these critical considerations, HFIP was chosen for ES of the SPU/elastin solution blends. Solutions of SPU30L with solubilized elastin content of up to 40 wt % could be electrospun into uniform fiber mats by optimizing process parameters. Smooth fibers with lower average diameters as compared with the pure SPU fibers of ∼300 nm with narrow frequency distributions were obtained (Figure 4C,D). The full preservation of the fiber formation capacity of the blend solutions, despite the fact that the solution viscosity was considerably decreased by the addition of the low-molecular weight protein fragments (from 2.5 Pa s for the pure SPU spinning solution to below 1 Pa s for the highly blended solutions), is a further indication of molecular interactions between the two macromolecular components. As a proof of concept, both freestanding scaffolds and small-diameter tubular constructs could be realized (Figure 4E), and these configurations could be valuable for applications ranging from wound dressing and templated skin regeneration to vascular grafting and vascular tissue engineering. Solubilized Elastin Enhances Mechanical Properties of SPU. To explore whether the observed molecular level interactions of the blend components could indeed be translated into macroscopic mechanical properties of the fabricated nanofibrous devices, we determined the properties of the freestanding fiber mats concerning uniaxial tension. Incorporation of solubilized elastin resulted in an up to 5-fold increase in the tensile modulus depending on the elastin content in comparison with native SPU30L, with no substantial changes in elongation at break, which was ∼100% (Figure 5A). Additional increase in tensile modulus to 30 MPa was achieved by glutaraldehyde cross-linking. While the cross-linking as expected increased brittleness of the materials, as reflected in the observed loss in the elongation at break, especially above an amount of incorporated elastin of ∼20 wt % (red shaded region in Figure 5B), the tensile strengths of the materials, however, remained unaffected at ∼1 MPa and showed no dependence on either elastin incorporation or cross-linking (Figure 5C, stress− strain curves can be found in Figure S6 in the Supporting Information). The observed evolution of the mechanical properties can be attributed to the molecular-level interactions between the synthetic polymer and the protein fragments because neither the pure SPU nor the pure solubilized elastin show comparable behavior. As a matter of fact, the solubilized elastin could under no conditions be electrospun into fibers (let alone free-standing fiber mats) and therefore possessed no mechanical integrity in its pure state. H bonds are dynamic in the sense that they can break and reform upon mechanical load. This design concept leads to materials that combine high ductility and structural stability and can be found in natural structures such as collagen and spider silk but has also been exploited for synthetic polymer fibers.44 Therefore, in the non-cross-linked state, these H-bond interactions contribute to an increased resistance to mechanical deformation, as reflected in the increased tensile modulus, while still maintaining the ductility of the material as confirmed by negligible differences in the elongations at break. However, upon cross-linking, the dynamic component is
Figure 5. (A,B) Tensile moduli and elongations at break of SPU30L/ elastin hybrid scaffolds as a function of elastin content for non-crosslinked (A) and cross-linked (B) scaffolds. (C) Tensile strengths for non-cross-linked and cross-linked scaffolds.
diminished as H bonds are consolidated by covalent crosslinks between the blend components. This results in the formation of a 3D cross-linked network to a certain extent, and this reasoning is consistent with the observed loss in elongation at break, indicating an increase in brittleness in the materials. Blend Scaffolds Show Enhanced Endothelial Cell Attachment. Because one conceivable application of the SPU/elastin hybrid nanofibrous meshes is in cardiovascular repair, we explored the potential of these materials to support the attachment of HUVECs. A common concern with biomaterials processed using solvents is the leaching of residual solvent, leading to undesirable biological response in vivo.45 Therefore, to ensure that the SPU/elastin blend nanofibrous mats were not cytotoxic, the proliferation of HUVECs on tissue culture plastic (TCPS) was assessed in the presence of G
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules
electron microscopy of the substrates 48 h after seeding (Figure 6C).
nanofibrous meshes derived from SPU30L and SPU40L for a period of 4 days (Figure 6A) using the MTT assay. In the MTT
■
CONCLUSIONS In this study it was theorized that materials with new properties could be realized by blending a synthetic polymer with a biopolymer if such blends could interact at a molecular level. Solution blends of soluble elastin with SPUs possessing extensive H-bonding capabilities were prepared by varying the weight fraction of elastin and then processed into freestanding fibrous mats using ES to explore this paradigm. CD and IR spectroscopy confirmed the presence of strong H-bond interactions between the SPU and elastin. The presence of H bonding gave rise to unexpected changes in both thermal and mechanical behavior of the blends. Tensile testing revealed that a 5-fold increase in tensile modulus over the native SPU could be achieved through the incorporation of elastin. The cytocompatibility of the nanofibrous SPU/elastin substrates was assessed using HUVECs. SPU/elastin substrates, in addition to being well tolerated by HUVECs, interestingly showed an elastin-dependent enhancement in the attachment of HUVECs. In summary, SPU/elastin nanofibrous mats prepared by electrospinning of solution blends of SPU and soluble elastin fulfill many of the prerequisites for the engineering of cardiovascular prosthesis such as vascular grafts and heart valves.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biomac.5b01681. 1 H NMR spectra of macrodiols and SPUs; GPC data of macrodiols; DSC thermograms of SPUs; and IR spectra, UV absorption spectra, TGA profiles, and stress−strain curves of blend materials. (PDF)
■
AUTHOR INFORMATION
Corresponding Author
Figure 6. (A) MTT assay results of coincubation of SPU30L and SPU40L polymer fiber meshes with HUVECs. (B) MTT assay results of adhesion tests (3 h) of HUVECs on SPU30L and SPU30L/elastin scaffolds. The result for tissue-culture polystyrene (TCPS) is added as a reference point. (C) SEM micrograph of HUVECs (red pseudo color) seeded on PUE8020 scaffold, 48 h postseeding.
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Vincent Ahmadi for the acquisition of SEM images and Melika Sarem for technical help with circular dichroism. This work was funded by the Excellence Initiative of the German Federal and State Governments (Grant EXC 294, Centre for Biological Signalling Studies) and the University of Freiburg.
assay the accumulation of Formazan dye requires mitochondrial esterase activity and therefore is indicative of metabolically active cells.46 In general, proliferation of HUVECs in the presence of polymer meshes showed identical behavior within the reliability limits of the MTT assay.47 Encouraged by these findings, we evaluated the adhesion behavior of HUVECs on SPU/elastin hybrid nanofibrous substrates derived from SPU30L with elastin contents ranging from 10−40 wt %. After 3 h, a systematic increase in Formazan accumulation was observed with increasing elastin content, with a 2-fold increase in substrates with 40 wt % elastin (Figure 6B). These preliminary results are suggestive of an increase in HUVEC attachment, most likely mediated by cell-interactive sites present in elastin, and indirectly imply that the elastin in the nanofibrous scaffolds is accessible to the cells.9 The robust attachment of HUVECs was further confirmed by scanning
■
REFERENCES
(1) Stitzel, J.; Liu, J.; Lee, S. J.; Komura, M.; Berry, J.; Soker, S.; Lim, G.; van Dyke, M.; Czerw, R.; Yoo, J. J.; Atala, A. Biomaterials 2006, 27 (7), 1088−1094. (2) Najafabadi, A. H.; Tamayol, A.; Annabi, N.; Ochoa, M.; Mostafalu, P.; Akbari, M.; Nikkhah, M.; Rahimi, R.; Dokmeci, M. R.; Sonkusale, S.; Ziaie, B.; Khademhosseini, A. Adv. Mater. (Weinheim, Ger.) 2014, 26 (33), 5823−5830. (3) Yilgör, I.; Yilgör, E.; Wilkes, G. L. Polymer 2015, 58, A1−A36. (4) Wang, Y.; Ameer, G. A.; Sheppard, B. J.; Langer, R. Nat. Biotechnol. 2002, 20 (6), 602−606.
H
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX
Article
Biomacromolecules (5) Francis Suh, J.-K.; Matthew, H. W. Biomaterials 2000, 21 (24), 2589−2598. (6) Lee, K. Y.; Mooney, D. J. Prog. Polym. Sci. 2012, 37 (1), 106−126. (7) Forget, A.; Christensen, J.; Lüdeke, S.; Kohler, E.; Tobias, S.; Matloubi, M.; Thomann, R.; Shastri, V. P. Proc. Natl. Acad. Sci. U. S. A. 2013, 110 (32), 12887−12892. (8) Rnjak-Kovacina, J.; Wise, S. G.; Li, Z.; Maitz, P. K. M.; Young, C. J.; Wang, Y.; Weiss, A. S. Biomaterials 2011, 32 (28), 6729−6736. (9) Almine, J. F.; Bax, D. V.; Mithieux, S. M.; Nivison-Smith, L.; Rnjak, J.; Waterhouse, A.; Wise, S. G.; Weiss, A. S. Chem. Soc. Rev. 2010, 39 (9), 3371−3379. (10) Vasconcelos, A.; Cavaco-Paulo, A. Appl. Microbiol. Biotechnol. 2011, 90 (2), 445−460. (11) Vasconcelos, A.; Gomes, A. C.; Cavaco-Paulo, A. Acta Biomater. 2012, 8 (8), 3049−3060. (12) Yeo, G. C.; Aghaei-Ghareh-Bolagh, B.; Brackenreg, E. P.; Hiob, M. A.; Lee, P.; Weiss, A. S. Adv. Healthcare Mater. 2015, 4 (16), 2530− 2556. (13) Duca, L.; Floquet, N.; Alix, A. J. P.; Haye, B.; Debelle, L. Critical reviews in oncology/hematology 2004, 49 (3), 235−244. (14) Daamen, W. F.; Nillesen, S. T. M.; Wismans, R. G.; Reinhardt, D. P.; Hafmans, T.; Veerkamp, J. H.; van Kuppevelt, T. H. Tissue Eng., Part A 2008, 14 (3), 349−360. (15) Paulsson, M.; Edsman, K. J. Colloid Interface Sci. 2002, 248 (1), 194−200. (16) Kim, B.-S.; Park, S. W.; Hammond, P. T. ACS Nano 2008, 2 (2), 386−392. (17) Pino, C. J.; Gutterman, J. U.; Vonwil, D.; Mitragotri, S.; Shastri, V. P. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (52), 21283−21288. (18) Lee, P. J.; Langer, R.; Shastri, V. P. Pharm. Res. 2003, 20 (2), 264−269. (19) Robert, L.; Poullain, N. Bull. Soc. Chim. Biol. 1963, 45 (12), 1317−1326. (20) Riley, T.; Stolnik, S.; Heald, C. R.; Xiong, C. D.; Garnett, M. C.; Illum, L.; Davis, S. S.; Purkiss, S. C.; Barlow, R. J.; Gellert, P. R. Langmuir 2001, 17 (11), 3168−3174. (21) Luu, Y. K.; Kim, K.; Hsiao, B. S.; Chu, B.; Hadjiargyrou, M. J. Controlled Release 2003, 89 (2), 341−353. (22) Cohn, D.; Hotovely Salomon, A. Biomaterials 2005, 26 (15), 2297−2305. (23) Lasprilla, A. J. R.; Martinez, G. A. R.; Lunelli, B. H.; Jardini, A. L.; Filho, R. M. Biotechnol. Adv. 2012, 30 (1), 321−328. (24) Zhang, M.; Desai, T.; Ferrari, M. Biomaterials 1998, 19 (10), 953−960. (25) Ling, X.; Spruiell, J. E. J. Polym. Sci., Part B: Polym. Phys. 2006, 44 (22), 3200−3214. (26) Sarem, M.; Lüdeke, S. MRS Bull. 2015, 40 (06), 490−498. (27) Kelly, S. M.; Jess, T. J.; Price, N. C. Biochim. Biophys. Acta, Proteins Proteomics 2005, 1751 (2), 119−139. (28) Matthews, J. A.; Wnek, G. E.; Simpson, D. G.; Bowlin, G. L. Biomacromolecules 2002, 3 (2), 232−238. (29) Scelsi, A.; Bochicchio, B.; Smith, A.; Saiani, A.; Pepe, A. RSC Adv. 2015, 5 (115), 95007−95013. (30) Debelle, L.; Alix, A. J. P.; Jacob, M.-P.; Huvenne, J.-P.; Berjot, M.; Sombret, B.; Legrand, P. J. Biol. Chem. 1995, 270 (44), 26099− 26103. (31) Walgers, R.; Lee, T. C.; Cammers-Goodwin, A. J. Am. Chem. Soc. 1998, 120 (20), 5073−5079. (32) Matsuo, S.; Iwakura, Y. Makromol. Chem. 1972, 152 (1), 203− 215. (33) Nisha, S. K.; Asha, S. K. J. Mater. Chem. C 2014, 2 (11), 2051. (34) Chen, H.; Xue, Q.; Li, Z.; Sun, L.; Zhang, Q. Polymer 2011, 52 (2), 400−408. (35) He, Y.; Zhu, B.; Inoue, Y. Prog. Polym. Sci. 2004, 29 (10), 1021− 1051. (36) Kaminski, A. M.; Urban, M. W. J. Coat. Technol. 1997, 69 (10), 113−121.
(37) Samouillan, V.; Dandurand-Lods, J.; Lamure, A.; Maurel, E.; Lacabanne, C.; Gerosa, G.; Venturini, A.; Casarotto, D.; Gherardini, L.; Spina, M. J. Biomed. Mater. Res. 1999, 46 (4), 531−538. (38) Greiner, A.; Wendorff, J. H. Angew. Chem., Int. Ed. 2007, 46 (30), 5670−5703. (39) Yang, F.; Murugan, R.; Wang, S.; Ramakrishna, S. Biomaterials 2005, 26 (15), 2603−2610. (40) Jaeger, R.; Schönherr, H.; Vancso, G. J. Macromolecules 1996, 29 (23), 7634−7636. (41) Shenoy, S. L.; Bates, W. D.; Frisch, H. L.; Wnek, G. E. Polymer 2005, 46 (10), 3372−3384. (42) Telemeco, T. A.; Ayres, C.; Bowlin, G. L.; Wnek, G. E.; Boland, E. D.; Cohen, N.; Baumgarten, C. M.; Mathews, J.; Simpson, D. G. Acta Biomater. 2005, 1 (4), 377−385. (43) Rho, K. S.; Jeong, L.; Lee, G.; Seo, B.-M.; Park, Y. J.; Hong, S.D.; Roh, S.; Cho, J. J.; Park, W. H.; Min, B.-M. Biomaterials 2006, 27 (8), 1452−1461. (44) Beese, A. M.; Sarkar, S.; Nair, A.; Naraghi, M.; An, Z.; Moravsky, A.; Loutfy, R. O.; Buehler, M. J.; Nguyen, S. T.; Espinosa, H. D. ACS Nano 2013, 7 (4), 3434−3446. (45) Standard Practice for Reporting and Assessment of Residues on Single Use Implants; ASTM F2847-10; ASTM International: West Conshohocken, PA, 2010. www.astm.org. (46) Mosmann, T. J. Immunol. Methods 1983, 65 (1−2), 55−63. (47) Wang, P.; Henning, S. M.; Heber, D. PLoS One 2010, 5 (4), e10202.
I
DOI: 10.1021/acs.biomac.5b01681 Biomacromolecules XXXX, XXX, XXX−XXX