Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
The effect of elastic biodegradable polyurethane electrospun nanofibers on the differentiation of mesenchymal stem cells Yi-Chia Kuo a , Shih-Chieh Hung b,c , Shan-hui Hsu a,∗ a b c
Institute of Polymer Science and Engineering, National Taiwan University, Taipei, Taiwan Institute of Clinical Medicine and Institute of Pharmacology, Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan Stem Cell Laboratory, Department of Medical Research & Education and Orthopaedics & Traumatology, Veterans General Hospital, Taipei, Taiwan
a r t i c l e
i n f o
Article history: Received 10 March 2014 Received in revised form 10 June 2014 Accepted 12 July 2014 Available online 21 July 2014 Keywords: Biodegradable Polyurethane (PU) Electrospinning Nanofibers Mesenchymal stem cells (MSCs)
a b s t r a c t Biodegradable polyurethane (PU) was synthesized based on using poly(-caprolactone) (PCL) as the soft segment. Fibers in different diameters (200–400 nm, 600–800 nm, and 1.4–1.6 m) were then made by electrospinning PU solution in N,N-dimethylacetamide and 2,2,2-trifluoroethanol. Human bone marrow derived mesenchymal stem cells (hMSCs) in the form of single dispersed cells or aggregates were seeded on the electrospun meshes for evaluation of cell behavior. Differentiation experiments showed that hMSC aggregates on electrospun fibers had greater differentiation capacities than single cells. Besides, nanofibers of 200–400 nm diameters significantly promoted the osteogenic and chondrogenic differentiation of hMSCs than fibers of the other diameters. The effect of substrate elasticity was further elucidated by comparing cell behaviors on the nanofibers of PCL-based PU and those of pure PCL. The more elastic PU nanofibers demonstrated more osteogenic and chondrogenic induction potential than PCL electrospun fibers. We suggested that the elastic nanofibers seeded with hMSC aggregates may be advantageous for cartilage and bone tissue engineering. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Mesenchymal stem cells (MSCs) are a promising cell source for tissue engineering applications mainly because of its multipotency to differentiate into various cell types such as osteoblasts [1], adipocytes [2], chondrocytes [3], myocytes [4] and neural cells [5]. Their fate is greatly influenced by both intrinsic and extrinsic signals, including cell–cell and cell–substrate interactions, the gradients of oxygen, nutrient, and protein concentrations, as well as the mechanical properties of the substrate [6]. Biological responses of stem cells could be quite different in two-dimensional (2D) and three-dimensional (3D) environments [7]. The traditional 2D culture method tends to separate single cells from their neighbors, which does not resemble the real in vivo situation. The 3D cell culture involves the generation of cellular aggregates or the growth of cells on 3D substrates (“scaffolds”), which may improve the cell–cell communication and differentiation potential of stem cells [8].
∗ Corresponding author at: Institute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan. Tel.: +886 2 33665313; fax: +886 2 33665237. E-mail address: [email protected] (S.-h. Hsu). http://dx.doi.org/10.1016/j.colsurfb.2014.07.017 0927-7765/© 2014 Elsevier B.V. All rights reserved.
Electrospinning, a fast and simple polymer processing technique, is not only capable of producing a wide variety of polymeric nanofibers, but also able to control the sizes, shapes, and porosity to achieve desired products [9]. Due to the above advantages, electrospinning is widely used in fabricating 3D scaffolds for regeneration of tissues such as blood vessel [10], bone [11], nerve [12], cartilage [13], skin [14], heart [15], and ligament [16]. Recent studies have shown that the adhesion, migration, proliferation, and differentiation of stem cells could be influenced by the topography [17], porosity [18], and mechanical properties of the substrate materials [19]. In particular, nanofibers that simulate the fibrillar structure of natural native extracellular matrix (ECM) can improve the differentiation efficiency of stem cells [20]. Electrospun substrates with high porosities or large pore sizes may have better proliferation capacities because cells can grow into the substrates [21]. Moreover, the osteogenesis of stem cells could be achieved by increasing the hardness of the matrix, while chondrogenesis could be induced by the softness of the substrate [22]. An ideal scaffold should mimic both the biomechanical properties and architecture of a natural tissue. Therefore, making nanofibrous scaffolds from elastic polymers such as polyurethane (PU) could be a potential approach for tissue engineering [23–25]. Cellular aggregates of MSCs have received much attention recently for their possible applications in tissue engineering
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
415
[26–29]. The behavior of MSC aggregates on electrospun 3D scaffolds remains unclear. There is no direct evidence that the differentiation capacity of aggregates after spreading on a fibrous matrix can still maintain their advantages over single (dispersed) cells on the same substrate. In this study, we first synthesized biodegradable polyurethane (PU) elastomers. Elastic fibers in nanoto micro-scale were then fabricated by electrospinning. The influence of topography and mechanical properties of different fibrous scaffolds on cell adhesion, proliferation, and differentiation potentials of MSC aggregates vs. single cells was investigated. Specifically, we sought to optimize the combination of electrospun fibers and MSC aggregates for cartilage and bone tissue engineering. 2. Materials and methods 2.1. PU synthesis The biodegradable PU used in this study was synthesized by a green process. Isophorone diisocyanate (IPDI, Evonik Degussa GmbH) was reacted with stirred (180 rpm) poly(-caprolactone) diol (PCL diol, molecular weight ∼2 kDa, Sigma) under nitrogen for 3 h at 75 ◦ C in the presence of 0.03% stannous octoate (Sn(Oct)2 , Alfa Aesar) as the catalyst. After prepolymerization, 2,2-bis(hydroxymethyl) propionic acid (DMPA, Sigma) and an appropriate amount (13 mL) of methyl ethyl ketone (MEK, J.T. Baker) were added to the reactor. After stirring for 1 h (180 rpm), the reaction temperature was then dropped to 50 ◦ C. Neutralization was completed by adding triethylamine (TEA, R.D.H) into the reactor. The neutralized prepolymer was then dispersed in deionized water (110 mL) and stirred vigorously at 1100 rpm before ethylenediamine (EDA, Tedia) in 110 mL of deionized water was further added with continuing stirring to complete the reaction. The stoichiometric ratio used for synthesis was optimized as IPDI/PCL diol/DMPA/EDA/TEA = 3.52:1:1:1.52:1 [30]. The solvent was removed to obtain the solid form of biodegradable PU. The molecular weight defined by gel permeation chromatography was ∼160 kDa. The weight percent of PCL segment in the final polymer was ∼65%. On the other hand, the pure PCL polymer was purchased (molecular weight 70–90 kDa, Sigma) and used as received. 2.2. Electrospinning process The setup of the electrospinning apparatus was constructed from a high-voltage power supplier (You-Shang Technical Corp., Taiwan), a syringe pump (KDS-100, KD Scientific, USA), a syringe with a stainless steel blunt-ended needle (20G, Terumo, Japan), and a plate type collector (Fig. 1(a)). The electrospinning was conducted in a custom-designed transparent acrylic chamber equipped with a hygrometer and a thermometer. The local humidity (40%) was controlled by a desiccant (silica gel). The chamber was placed in a humidity/temperature controlled room. In order to narrow down the distribution of fiber diameters, the mixed solvent system of a low-boiling point solvent 2,2,2-trifluoroethanol (TFE, Tedia; bp. 78 ◦ C) and a high-boiling point solvent N,N-dimethylacetamide (DMAc, Tedia; bp. 165 ◦ C) was used. PU was dissolved at the concentration of 6 wt% in the mixture of TFE and DMAc with different mixing ratios and stirred for overnight. The polymer solution was then poured into the syringe and drawn to the collecting plate by applying a high voltage. The flow rate, the applied voltage, and the distance between the nozzle and the collector could be adjusted. The temperature and the relative humidity were controlled at 21.5 ± 0.5 ◦ C and 40 ± 4%. After optimization, PU (5, 8, and 11 wt%) in a mixed solvent (wt% 60:40) was used to obtain electrospun membranes comprising of nanofibers (“ES1” groups), sub-microfibers (“ES2” groups), and microfibers (“ES3” groups). The relative humidity and temperature were controlled as mentioned. Other electrospinning parameters
Fig. 1. (a) The electrospinning apparatus. (b) The effect of mixed solvent (TFE/DMAc) ratio on the morphology of electrospun fibers.
such as the flow rate, the applied voltage, and the distance between nozzle and collector were varied in the ranges of 0.3–0.6 mL/h, 10–15 kV, and 15–32 cm, respectively. Electrospun meshes for cell studies were spun onto the PU-coated coverslip glass, while those for physico-chemical characterization were spun onto glass slides and then removed by immersion into ice water. The PU-coated coverslip glass was prepared by spin coating 3% solution in TFE onto 15 mm diameter coverslip glass (Assistant, Glaswarenfabrik Karl Hecht KG, Germany) at 2750 rpm for 20 s and dried in vacuum for 1 day, which also served as the control (“PU film” groups). PCL electrospun meshes (“PCL-ES” groups) were prepared for comparison. For the purpose, the solution of 10 wt% PCL in TFE was spun onto PCL-coated coverslip glass at a voltage of 20 kV, a flow rate of 0.3 mL/h, and a distance of 32 cm. The relative humidity and temperature were controlled as mentioned. The PCL-coated coverslip glass was prepared similarly to that of PU by 5% solution of PCL in TFE. All electrospun meshes were dried in vacuum for 2 days to remove residual solvent. 2.3. Characterization of electrospun fibers The electrospun meshes were examined by a scanning electron microscope (SEM, Hitachi S-4800) at an accelerating voltage
416
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
of 10 kV after sputter-coated with a 2 nm layer of carbon (Emitech K950X). The average fiber diameter was determined based on SEM images using the Image J software. The thickness without a supporting film was measured by a dial thickness gauge. The thickness of each mesh was kept as close as possible during the parameter optimization. The fiber density and porosity of the electrospun meshes were determined as follows. The volume and density of the samples were calculated based on their dimension and weight. The fiber density (, g/cm3 ) was estimated by dividing the weight of the meshes with the volume of the meshes. The porosity was then determined based on the equation porosity (%) = (1 − /0 ) × 100%, where 0 is the density of PU (1.167 g/cm3 ) or PCL (1.145 g/cm3 ). The water contact angle was measured by a goniometer (FTA-1000B, First Ten Angstrom Company, USA) at room temperature. Images of water drops (5 L) deposited on the surface of the membrane were used for water contact angle measurements. The mechanical properties of the meshes and PU films including the Young’s modulus, 100% modulus, tensile strength, and elongation were measured by a universal testing system (HT8504, Hung Ta) at a strain rate of 50 mm min−1 . All procedures followed the standard testing method (ASTM-D638). The meshes used for characterization were comprised only of electrospun meshes. The specimens were cut in standard sizes by a specially designed knife (parameters: overall length 38.2 mm, width of grip section 16.02 mm, gage length 23 mm, and gage width 4.67 mm; n = 7). 2.4. Cell culture Human bone marrow mesenchymal stem cells (hMSCs) were obtained from the Tulane Center for Preparation and Distribution of Adult Stem Cells. They were seeded in each well with a density of 5 × 104 cells per well, and cultured in the basal medium at 37 ◦ C with 5% CO2 for 3 days. Chitosan powder (molecular weight ∼416 kDa, 77.7% deacetylation, Fluka) was dissolved and stirred at room temperature in 1% aqueous acetic acid solution for 24 h to obtain a 1% chitosan solution. Chitosan membranes were made from casting 300 L of solution on 15 mm diameter coverslips and air-dried. The membranes were rinsed in 0.5 N NaOH for 1 min, washed extensively by phosphate buffered saline (PBS), sterilized with 75% ethanol for 1 h, washed five times with PBS, and placed into a 24-well tissue culture plate before use. The basal medium contained low glucose Dulbecco’s modified Eagle medium (DMEMLG) supplemented with 10% fetal bovine serum (FBS, Gibco), 1% penicillin-streptomycin (Invitrogen). After incubation for 3 days, hMSC aggregates were collected by aspiration and quantified using a CountessTM automated cell counter (Invitrogen). Cells seeded in the blank well (tissue culture polystyrene, TCPS) served as the control. 2.5. Cell culture on electrospun meshes All samples were sterilized under ultraviolet light for 3 h, and placed into 24-well tissue culture plates, where each well fit one coverslip glass. hMSC aggregates or single cells at a density of 4 × 104 cells in 1 mL culture medium were seeded on PU meshes (ES1, ES2, and ES3) and cultured in the basal medium for 1, 7, and 14 days, respectively. The medium was refreshed three times a week. To further compare PU and PCL electrospun meshes, hMSC aggregates (4 × 104 cells) were seeded on ES1 and PCL-ES meshes in the basal medium for 3 days. The medium was then replaced with the osteogenic induction medium or chondrogenic induction medium for 3 and 7 days to induce osteogenic or chondrogenic differentiation. The osteogenic induction medium contained 50 g/mL of
l-ascorbate-2-phosphate, 10−8 M dexamethasone, and 10 mM glycerophosphate. The chondrogenic induction medium contained 1% of insulin-transferrin-selenium (ITS) premix 100×, 50 g/mL of l-ascorbate-2-phosphate, 0.1 M of dexamethasone, 10 ng/mL of TGF-3, and 40 g/mL of l-proline. The induction medium was refreshed twice a week. The two control groups were hMSC aggregates seeded on “PU films” and hMSC single cells seeded on ES1 meshes (“ES1-SC” groups). The morphology of cells after seeding was examined by an inverted microscope (Leica, DMIRB). The cell number of aggregates was ensured by a parallel experiment where cell counting was performed after incubation in basal medium for 3 days. hMSC spheroids were collected by aspiration and trypsinized for cell counting using the automated cell counter.
2.6. Cell proliferation and gene expression Cells on different substrates were collected at the end of culture period, and the number was determined by the fluorometric assay of DNA. Briefly, samples were lyophilized and decomposed in 1.5 mL of papain solution, which contained 47.6 g/mL of papain (Sigma) in buffer solution with 55 mM of sodium citrate (Showa), 150 mM of sodium chloride (Sigma), 5 mM of cysteine HCL (Sigma), and 5 mM of Na2EDTA (Sigma), for 24 h at 60 ◦ C. Hoechst 33258 fluorescent dye solution (5 mL) which contained 0.1 g/mL of Hoechst 33258 (Sigma) in pH 7.4 buffer solution [10 mM Tris (Tedia), 1 mM Na2 EDTA (Sigma), and 0.1 mM sodium chloride (Sigma)] was then added to 0.5 mL of decomposed solution. The intensity of fluorescence (excitation wavelength 365 nm, emission wavelength 458 nm) was determined by a multidetection microplate reader (Molecular Devices SpectraMax M5, USA). The dye solution served as the background. To evaluate the osteogenic and chondrogenic gene expression, total RNA from collected samples was extracted by Trizol® reagent (Invitrogen, USA). cDNA was acquired via reverse transcription using RevertAidTM First Strand cDNA Synthesis kit (MBI Fermentas, St. Leon-Rot, Germany). Polymerase chain reaction (PCR) was used to amplify the gene expression of human runt-related transcription factor 2 (RUNX2) and human SRY (sex determining region Y)-box 9 (SOX9) with T100TM Thermal Cycler (BIO-RAD, USA) and estimated by capillary electrophoresis (HDA-GT12TM , eGene, USA). The gene expression analyses were carried out by measuring band intensity after electrophoresis. Results of gene expression of each sample were normalized to their own glyceraldehyde 3-phosphate dehydrogenase (GAPDH; housekeeping gene). The primer sequences are as follows: human GAPDH, 5 -CAAGGTCATCCATGACAACTTTG3 (forward) and 5 -GTCCACCACCCTGTTGCTGTAG-3 (reverse), human RUNX2, 5 -TTGCAGCCATAAGAGGGTAG-3 (forward) and 5 -GTCACTTTCTTGGAGCAGGA-3 (reverse), human SOX9, 5 -AAAAGGCAAGCAAAGGAGAT-3 (forward) and 5 -AAAAGGGATGGACAAAAAGG-3 (reverse). The cycling condition includes 35 cycles of 30 s at 95 ◦ C (denaturation), 30 s at 47 or 58 ◦ C (annealing), and 1 min at 72 ◦ C (extension) for SOX9, RUNX2, and GAPDH.
2.7. Statistical analysis Each experiment was performed independently for three times and with multiple samples. Numerical values were presented as mean ± standard deviation. Statistical differences between two groups were evaluated by unpaired two tailed t-tests. One way ANOVA with Tukey’s multiple comparison tests was used to determine the statistical differences of the results that involved more than two groups. p values less than 0.05 meant significant differences.
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
417
Table 1 Electrospinning parameters for producing meshes with three different fiber diameters and the corresponding fiber density and porosity (n = 7). All data are significantly different from each other, except the fiber diameter between ES1 and PCL-ES and the weight between ES1 and ES2. Meshesa
Solution conc. (wt%)
ES1 ES2 ES3 PCL-ES
5 8 11 10
a
Flow rate (L/min) 5 5 10 5
Voltage (kV)
Collector distance (cm)
Fiber diameter (m)
15 10 10 20
32 15 32 32
0.31 0.67 1.47 0.37
± ± ± ±
0.05 0.09 0.09 0.06
Weight (g)
Density of scaffold (g/cm3 ) 0.48 0.40 0.32 0.17
± ± ± ±
0.01 0.02 0.02 0.03
0.062 0.052 0.035 0.167
± ± ± ±
Porosity (%)
0.006 0.008 0.006 0.003
58.6 64.9 72.6 85.5
± ± ± ±
1.3 1.1 2.0 2.6
ES1, ES2, and ES3: PU fibers made from the mixed solvent of TFE and DMAc. PCL-ES: PCL fibers made from the solvent TFE.
Table 2 The tensile mechanical properties and water contact angle of the electrospun meshes at 25 ◦ C (n = 7). All data are significantly different from each other, except the Young’s modulus between ES2 and PCL-ES, Young’s modulus between ES3 and PCL-ES, 100% modulus between ES2 and ES3, strength between ES2 and ES3, and elongation between ES1 and ES2, and contact angle between ES2 and ES3. Samples
Young’s modulus (MPa)
PU film (control) ES1 ES2 ES3 PCL-ES
30.5 10.8 3.5 5.9 4.3
± ± ± ± ±
0.6 1.8 0.3 1.4 1.3
100% modulus (MPa) 7.1 4.2 1.9 1.7 2.6
± ± ± ± ±
0.3 0.7 0.2 0.3 0.4
Tensile strength (MPa) 37.4 8.5 4.5 4.0 3.1
± ± ± ± ±
1.9 1.7 0.3 0.6 0.2
Elongation (%) 582.3 315.4 327.8 444.4 107.0
± ± ± ± ±
24.0 40.7 11.2 46.5 12.7
Contact angle (◦ ) 77.7 91.4 107.8 110.5 127.7
± ± ± ± ±
0.7 7.0 7.0 3.0 2.9
Fig. 2. The morphology of PU electrospun meshes with different fiber diameters and that of PCL electrospun meshes. Abbreviation: ES1, PU electrospun meshes with 200–400 nm fiber diameter. ES2, PU electrospun meshes with 600–800 nm fiber diameter. ES3, PU electrospun meshes with 1.4–1.6 m fiber diameter. PCL-ES, PCL electrospun meshes with 200–400 nm fiber diameter.
418
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
Fig. 3. The adhesion and proliferation of aggregates and single cells on electrospun meshes after incubation in basal medium. (a) The schematics of the culture period (1–14 days). (b) The number of cells and (c) the seeding efficiency measured at 1 day. (e) The number of cells after incubation for 7 days. (f) The number of cells after incubation for 14 days. *p < 0.05; **p < 0.01.
3. Results
3.2. Characterization of electrospun meshes
3.1. Fibrous PU meshes by electrospinning
The fiber density and porosity of PU and PCL electrospun membranes are listed in Table 1. As the fiber diameter of PU meshes decreased, the density of the mesh increased from 0.32 ± 0.02 g/cm3 to 0.48 ± 0.02 g/cm3 and the porosity decreased from 72.6 ± 2.0% to 58.6 ± 1.3%. Among all meshes, PCL-ES had the lowest fiber density and the highest porosity. The contact angle and mechanical properties are listed in Table 2. As the fiber diameter of PU meshes decreased, the contact angle decreased from 110.5 ± 3.0◦ to 91.4 ± 7.0◦ . PCL-ES was the most hydrophobic (127.7 ± 2.9◦ ) among all. All PU meshes had outstanding elongation (>300%). Among them, ES1 meshes had the smallest elongation (315.4 ± 40.7%) and the largest Young’s modulus (10.8 ± 1.8 MPa), 100% modulus (4.2 ± 0.7 MPa), and tensile strength (8.5 ± 1.7 MPa). The modulus and strength of ES1 were about two times higher than those of ES2 and ES3 meshes. The control PU film had the lowest contact angle and highest mechanical properties. PCL-ES had the smallest elongation (∼100%).
Electrospun meshes with different topographies were made in this study to examine the proliferation and differentiation capacities of hMSCs on biodegradable and elastic PU materials. To produce electrospun fibers in nanometric scale, solvent selection is very important. The effects of mixed solvent of TFE and DMAc in different mixing ratios on the morphology of electrospun fibers are shown in Fig. 1(b). When the content of DMAc in the mixed solvent reached 50%, a part of the electrospun fibers stuck together. Such a tendency increased with the DMAc content. The optimized mixing ratio of TFE to DMAc was 3:2. Fibers of nano- to micro-scale could be obtained from PU solution of different concentrations and in specific electrospinning conditions. The thickness was about 0.4 mm for each mesh (without the supporting film). The dimension was (20 ± 0.1 mm) × (15 ± 0.1 mm) (where the sample size n was 7). The details of electrospun parameters of all the electrospun meshes are listed in Table 1 and the corresponding topographies are shown in Fig. 2. The average fiber diameter of ES1, ES2, and ES3 meshes was 0.31 ± 0.05 m, 0.67 ± 0.09 m, and 1.47 ± 0.09 m, respectively. SEM images showed that fiber diameter increased with the increased concentration of the polymer solution. In addition, PCL-ES nanofibers for comparison were produced from 10 wt% PCL solution in TFE and had a mean diameter of 0.37 ± 0.06 m.
3.3. The effect of fiber diameter on hMSC behavior Cells were distributed uniformly within the meshes. The number of hMSC single cells and aggregates on the electrospun fibers were tested at 1, 7, and 14 days in basal medium (Fig. 3(a)). When seeding hMSC single (dispersed) cells, the cell seeding efficiency increased with the increased fiber diameter. On the other hand,
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
419
Fig. 4. Gene expressions of aggregates and single cells on electrospun meshes after incubation in basal medium. (a) The schematics of the culture period (3–7 days). (b) RUNX2 and SOX9 expression before seeding on the electrospun meshes (0 day). (c) RUNX2 expression after incubation for 3 and 7 days. (d) SOX9 expression after incubation for 3 and 7 days. Results of gene expression of each sample were normalized to their own GAPDH. RUNX2, runt-related transcription factor 2; SOX9, SRY (sex determining region Y)-box 9. *p < 0.05; **p < 0.01.
hMSC aggregates had similar seeding efficiency on all PU electrospun meshes (Fig. 3(b) and (c)). Compared to single cells, hMSC aggregates showed better cell proliferative capacity on ES1 after incubation for 7 and 14 days (Fig. 3(d) and (e)). In addition, the cell proliferation of hMSC single cells slightly increased with the increased fiber diameter, while hMSC aggregates had similar proliferative capacity on ES2 and ES3 at 7 and 14 days. The gene expression of hMSC single cells and aggregates on the electrospun fibers were tested at 3 and 7 days in basal medium (Fig. 4(a)). SOX9 is an important chondrogenic marker gene, which encodes the transcription factor critical to chondrogenesis and cartilage development. RUNX2 is an important osteogenic marker gene, which encodes the transcription factor critical to osteogenesis
and bone development. The aggregates showed high SOX9 expression before seeding on the substrate (Fig. 4(b)). It also appeared that the SOX9 gene expression of aggregates on electrospun membranes increased with the incubation time, while that of single cells showed the opposite (Fig. 4(e) and (f)). After 7 days, aggregates had the highest SOX9 expression on ES1 meshes. The RUNX2 expression of hMSC single cells and aggregates were upregulated on all PU meshes at 3 days (Fig. 4(c)). At 7 days, hMSC aggregates showed better RUNX2 expression than single cells on all PU meshes (Fig. 4(d)). hMSC aggregates on ES1 had the highest RUNX2 expression at both 3 and 7 days. Overall, the data revealed that hMSC aggregates had better osteogenic and chondrogenic differentiation potential than single cells on electrospun meshes without the induction medium;
420
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
14 days (Fig. S2(a) and (b)). hMSC aggregates on ES1 meshes also revealed a significantly higher glycosaminoglycan (GAG, cartilage matrix component) content than those on TCPS after the chondrogenic induction for 14 days (Fig. S2(c)). ES1 appeared to be more favorable than PCL meshes regarding GAG secretion (p = 0.057, not reaching significance p < 0.05). On ES1 meshes, hMSC aggregates demonstrated significantly greater capacities for cartilage and bone matrix deposition than single cells. 4. Discussion
Fig. 5. The number of total cells after aggregates were seeded on ES1, PCL-ES, or PU film, incubated in basal medium for 3 days, and induced for 7 days. (a) The schematics of the culture period. (b) The cell number after induction. Abbreviation: ES1-SC, single cells seeded on ES1 and incubated in the same condition. *p < 0.05; **p < 0.01.
and ES1 meshes (i.e. 200–400 nm nanofibers) provided a better environmental cue for differentiation than the other fibers. 3.4. The effect of electrospun fibers on hMSC differentiation The number of hMSC single cells and aggregates on all samples were examined after incubation in basal medium for 3 days and further incubation in induction medium for 7 days (Fig. 5(a)). The cell number for hMSC aggregates seeded on the PU film was higher than that on ES1 and PCL-ES at the end of the culture period (Fig. 5(b)). Besides, the cell number was higher for hMSC aggregates seeded on ES1 meshes as compared to PCL-ES meshes. On the other hand, the cell number was higher for hMSC single cells (vs. aggregates) seeded on ES1. After osteogenic induction, hMSC aggregates on ES1 meshes showed the highest RUNX2 expression than the others (Fig. 6(a)). On the other hand, after chondrogenic induction for 7 days, there was no significant difference between hMSC aggregates on ES1, PCL-ES, and PU film, though it appeared that ES1 also had the best chondrogenic effect. The morphology of hMSCs on different substrates is shown in Supplemental data (Fig. S1). After incubation for 3 days, hMSC aggregates had already spread on PCL-ES and PU film, but some remained not spread (size between 80 and 180 m) on ES1 meshes. This tendency persisted until 7 days post induction. As shown in Supplemental data (Fig. S2), hMSC aggregates on ES1 meshes revealed a significantly higher calcium content (bone matrix component) than those on PCL meshes, PU flat films, or the blank well (TCPS) control after the osteogenic induction for
The high-boiling point solvent may evaporate incompletely during the electrospinning process, while the low-boiling point solvent may evaporate fast to generate fibers with large distribution of diameters. A mixed solvent system of TFE and DMAc, which appeared here for the first time, was used to overcome the above two problems frequently encountered. PU electrospun fibers of three different diameters could be achieved by adjusting the polymer concentration, the distance between nozzle and collector, the voltage, and the flow rate, all of which affected the fiber diameter. Electrospun PU fibers with decreasing fiber diameters showed an increase in hydrophilicity. This might be due to the fact that the opportunity of the hydrophilic functional groups exposed to the air would be higher when the fiber diameter was lower. Although the functional groups of PU are mainly considered as hydrophobic, PU also had many hydrophilic functional groups such as COOH, OH, COO, and NH. Besides, a few studies mentioned that electrospun fibers with thinner fiber diameter had pores of small sizes, which may increase the water absorption rate by capillary action and decrease the contact angle accordingly [31]. Moreover, the largerdiameter fibers could sometimes have big pores filled with air, which may increase the hydrophobicity [32]. The tensile properties of electrospun fibers including Young’s modulus, 100% modulus, and tensile strength were always higher in ES1 compared to the other two (ES2 and ES3). The alignment of nanofibers may influence the mechanical properties, i.e. highly alignment fibers normally show better mechanical properties. In the study, we did not control the alignment of the nanofibers; therefore, the nanofibers were randomly oriented, which were confirmed by our observation. On the other hand, the diameter and strength of a single fiber are both associated with the packing and orientation of the polymeric chains [33]. Our experimental result revealed that thinner nanofibers had greater strength. This may be attributed to the greater stretching deformation generated during single fiber formation, which may align the polymeric chains and subsequently lead to the greater strength of each single fiber [34]. Besides, fiber diameter can further affect fiber density and porosity; and in most cases, as the fiber diameter decreases, the fiber density increases in number, which also help enhance the mechanical properties [35]. The adhesion and proliferation of MSC aggregates and single cells in basal medium showed that the seeding efficiency of MSC aggregates were maintained in 70–90% on PU electrospun fibers measured at 1 day. MSC single cells showed only 30% seeding efficiency on ES1. It was believed that MSC aggregates produced more extracellular matrix such as fibronectin, which may promote cell adhesion on electrospun fibers [36,37]. The favorable cell adhesion may account for the higher seeding efficiency. In general, we found that fibers with smaller diameter promoted the differentiation while those with larger diameter enhanced proliferation. The enhanced cell proliferation may be associated with the cell ingrowth to pores on larger-diameter fibers [38]. The current results also indicated that MSC aggregates could be seeded without the help of plasma treatment or coating extracellular matrix and could well proliferate on the PU electrospun fibers. Although MSC aggregates tended to spread with the increased culture time on electrospun fibers, they had better differentiation
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
421
Fig. 6. (a) Gene expression for cells after induction for 7 days as indicated in Fig. 5(a). Abbreviation: ES1-basal, aggregates was seeded on ES1 and incubated in basal medium for 10 days. Results of gene expression of each sample were normalized to their own GAPDH. *p < 0.05; **p < 0.01. (b) The hypothetical mechanism for gene regulation on electrospun meshes.
potentials than single cells on electrospun fibers. It also seemed that MSC aggregates were more sensitive to the environmental (substrate) cue such as topography and hardness. After incubation for 7 days, MSC aggregates on ES1 showed higher RUNX2 expression than those on ES2 and ES3. ES1 had higher Young’s modulus and had the topography in nanoscale, both of which may contribute to enhance the osteogenic differentiation of stem cells [39,40]. In addition, MSC aggregates showed higher SOX9 expression on ES1 after incubation for 7 days. This might be attributed to the higher cell density and the biomimetic ECM-like structure of ES1 [41–43]. The behavior of cells on electrospun meshes may be related to their inherent properties before seeding. Prior to seeding, MSC aggregates showed higher SOX9 expression than single cells probably because of high cell density in the MSC aggregates. As the MSC aggregates spread on electrospun fibers, cell density may be decreased, which in turn favored cell–substrate interaction. After incubation for another 7 days, the expression of SOX9 obviously increased again. The hypothetical mechanism regarding how the MSC aggregates responded to the electrospun fibers is summarized in Fig. 6(b). Comparing the effect of fibers from different materials on cell behavior, it appeared that the hardness of substrate could significantly influence the osteogenic differentiation of MSCs. The biodegradable PU used in this study contained 65 wt% PCL segments in the structure. Nanofibers of PU (i.e. ES1) with higher modulus and elongation induced more osteogenic differentiation of MSCs than nanofibers of PCL (i.e. PCL-ES) having lower modulus and elongation. However, we could not tell if this was caused by chemistry,
hardness, or elasticity. Besides, the MSC aggregates spread rather slowly on ES1 meshes as noted in Supplemental data (Fig. S1). Some MSC aggregates remained intact even at 7 days post induction. This cell clustering effect may improve the cell–cell communication that is critical for the differentiation of stem cells. On the other hand, it was observed that the inclination for chondrogenic differentiation was all lower than that for osteogenic differentiation on all substrates. As the MSC aggregates gradually spread on the substrates, the 3D microenvironment which promoted the chondrogenic differentiation could no longer exist. As mentioned, some MSC aggregates remained intact on ES1 for a period, which may account for the greater chondrogenic differentiation on ES1. The reason behind the lower chondrogenic differentiation of MSCs on PCL-ES vs. ES1 may be the smaller amount of cells on PCL-ES (Fig. 5). MSC aggregates used in this study were shown to have better differentiation capacities upon induction. They may replace single cell suspension to be used in the cell-based therapy. However, the MSC aggregates are still susceptible to cell loss during administration. By seeding MSC aggregates onto the electrospun fibers, the combined constructs may help translate the better differentiation efficiency of the MSC aggregates. Our study demonstrates that even if the MSC aggregates may eventually spread and lose their morphology on the substrate, the differentiation potential remains to be better than single cells. In particular, the elastic PU electrospun nanofibers fabricated in this study promote the chondrogenic and osteogenic differentiation of MSC aggregates, which may have potential applications in cartilage and bone tissue engineering.
422
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
Acknowledgments This research was supported by the Program for Stem Cell and Regenerative Medicine Frontier Research (NSC102-2321-B-002033) sponsored by the National Science Council, Taiwan, R.O.C. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfb. 2014.07.017. References [1] Y. Mohammadi, M. Soleimani, M. Fallahi-Sichani, A. Gazme, V. Haddadi-Asl, E. Arefian, J. Kiani, R. Moradi, A. Atashi, N. Ahmadbeigi, Int. J. Artif. Organs 30 (2007) 204. [2] W.J. Li, R. Tuli, X. Huang, P. Laquerriere, R.S. Tuan, Biomaterials 26 (2005) 5158. [3] J. Liao, X. Guo, K.J. Grande-Allen, F.K. Kasper, A.G. Mikos, Biomaterials 31 (2010) 8911. [4] S. Sahoo, H. Ouyang, J.C. Goh, T.E. Tay, S.L. Toh, Tissue Eng. 12 (2006) 91. [5] K.S. Bae, J.B. Park, H.S. Kim, D.S. Kim, D.J. Park, S.J. Kang, Yonsei Med. J. 52 (2011) 401. [6] M.P. Lutolf, P.M. Gilbert, H.M. Blau, Nature 462 (2009) 433–441. [7] E. Cukierman, R. Pankov, D.R. Stevens, K.M. Yamada, Science 294 (2001) 1708. [8] J.E. Frith, B. Thomson, P.G. Genever, Tissue Eng. C: Methods 16 (2010) 735. [9] N. Bhardwaj, S.C. Kundu, Biotechnol. Adv. 28 (2010) 325. [10] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Biomaterials 25 (2004) 877. [11] N.G. Rim, S.J. Kim, Y.M. Shin, I. Jun, D.W. Lim, J.H. Park, H. Shin, Colloids Surf. B 91 (2012) 189. [12] E.M. Jeffries, Y. Wang, Biofabrication 5 (2013) 035015. [13] A. Fertala, W.B. Han, F.K. Ko, J. Biomed. Mater. Res. 17 (2001) 48. [14] J. Venugopal, S. Ramakrishna, Tissue Eng. 11 (2005) 847. [15] X. Zong, H. Bien, C.Y. Chung, L. Yin, D. Fang, B.S. Hsiao, B. Chu, E. Entcheva, Biomaterials 26 (2005) 5330. [16] C.H. Lee, H.J. Shin, I.H. Cho, Y.M. Kang, I.A. Kim, K.D. Park, J.W. Shin, Biomaterials 26 (2005) 1261.
[17] M. Chen, P.K. Patra, S.B. Warner, S. Bhowmick, Biophys. Rev. Lett. 1 (2006) 153. [18] A.S. Badami, M.R. Kreke, M.S. Thompson, J.S. Riffle, A.S. Goldstein, Biomaterials 27 (2006) 596. [19] Y. Yang, K. Kulangara, R.T. Lam, R. Dharmawan, K.W. Leong, ACS Nano 6 (2012) 8591. [20] C.T. Laurencin, A.M. Ambrosio, M.D. Borden, J.J. Cooper, Annu. Rev. Biomed. Eng. 1 (1996) 19. [21] J.B. Lee, S.I. Jeong, M.S. Bae, D.H. Yang, D.N. Heo, C.H. Kim, E. Alsberg, I.K. Kwon, Tissue Eng. Part A 17 (2011) 2695. [22] J.S. Park, J.S. Chu, A.D. Tsou, R. Diop, Z. Tang, A. Wang, S. Li, Biomaterials 32 (2011) 3921. [23] H.-W. Chien, J.-Y. Lai, W.-B. Tsai, Colloids Surf. B 116 (2014) 576. [24] R.O. Ritchie, M.J. Buehler, P. Hansma, Phys. Today 62 (2009) 41. [25] A.J. Engler, S. Sen, H.L. Sweeney, D.E. Discher, Cell 126 (2006) 677. [26] S.-h. Hsu, G.S. Huang, S.Y. Lin, F. Feng, T.T. Ho, Y.C. Liao, Tissue Eng. Part A 18 (2012) 67. [27] H.Y. Yeh, B.H. Liu, S.-h. Hsu, Biomaterials 33 (2012) 8943. [28] B.H. Liu, H.Y. Yeh, Y.C. Lin, M.H. Wang, D.C. Chen, B.H. Lee, S.-h. Hsu, Biores. Open Access 2 (2013) 28. [29] S.-h. Hsu, Y.H. Ni, Y.C. Lee, Macromol. Biosci. 13 (2013) 1100. [30] Taiwan patents (#I427091 and 102100076) and U.S. patent applications (13/317,651 Biodegradable and biocompatible waterborne polyurethane and 14/145,393 Biocompatible and biodegradable elastomer). [31] Y. Liu, N. Wu, Q. Wei, Y. Cai, A. Wei, J. Appl. Polym. Sci. 110 (2008) 3172. [32] N. Wang, K. Burugapalli, W. Song, J. Halls, F. Moussy, Y. Zheng, Y. Ma, Z. Wu, K. Li, J. Memb. Sci. 36 (2000) 75. [33] J. Yao, C.W.M. Bastiaansen, T. Peijs, Fibers 2 (2014) 158. [34] C.L. Pai, M.C. Boyce, G.C. Rutledge, Polymer 52 (2011) 2295. [35] I.K. Kwon, S. Kidoaki, T. Matsuda, Biomaterials 26 (2005) 3929. [36] C.L. Li, T. Tian, K.J. Nan, N. Zhao, Y.H. Guo, J. Cui, J. Wang, W.G. Zhang, Oncol. Rep. 20 (2008) 1465. [37] M.T. Santini, G. Rainaldi, P.L. Indovina, Crit. Rev. Oncol. Hematol. 36 (2000) 75. [38] J. Rnjak-Kovacina, S.G. Wise, Z. Li, P.K. Maitz, C.J. Young, Y. Wang, A.S. Weiss, Biomaterials 32 (2011) 6729. [39] K.J. Burg, S. Porter, J.F. Kellam, Biomaterials 21 (2000) 2347. [40] X.H. Chu, X.L. Shi, Z.Q. Feng, Z.Z. Gu, Y.T. Ding, Biotechnol. Lett. 31 (2009) 347. [41] M. Takagi, Y. Umetsu, M. Fujiwara, S. Wakitani, J. Biosci. Bioeng. 103 (2007) 98. [42] L. Zhang, P. Su, C. Xu, J. Yang, W. Yu, D. Huang, Biotechnol. Lett. 32 (2012) 1339. [43] J.M. Coburn, M. Gibson, S. Monagle, Z. Patterson, J.H. Elisseeff, Proc. Natl. Acad. Sci. U.S.A. 109 (2012) 10012.
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
The effect of elastic biodegradable polyurethane electrospun nanofibers on the differentiation of mesenchymal stem cells Yi-Chia Kuo a , Shih-Chieh Hung b,c , Shan-hui Hsu a,∗ a b c
Institute of Polymer Science and Engineering, National Taiwan University, Taipei, Taiwan Institute of Clinical Medicine and Institute of Pharmacology, Faculty of Medicine, National Yang-Ming University, Taipei, Taiwan Stem Cell Laboratory, Department of Medical Research & Education and Orthopaedics & Traumatology, Veterans General Hospital, Taipei, Taiwan
a r t i c l e
i n f o
Article history: Received 10 March 2014 Received in revised form 10 June 2014 Accepted 12 July 2014 Available online 21 July 2014 Keywords: Biodegradable Polyurethane (PU) Electrospinning Nanofibers Mesenchymal stem cells (MSCs)
a b s t r a c t Biodegradable polyurethane (PU) was synthesized based on using poly(-caprolactone) (PCL) as the soft segment. Fibers in different diameters (200–400 nm, 600–800 nm, and 1.4–1.6 m) were then made by electrospinning PU solution in N,N-dimethylacetamide and 2,2,2-trifluoroethanol. Human bone marrow derived mesenchymal stem cells (hMSCs) in the form of single dispersed cells or aggregates were seeded on the electrospun meshes for evaluation of cell behavior. Differentiation experiments showed that hMSC aggregates on electrospun fibers had greater differentiation capacities than single cells. Besides, nanofibers of 200–400 nm diameters significantly promoted the osteogenic and chondrogenic differentiation of hMSCs than fibers of the other diameters. The effect of substrate elasticity was further elucidated by comparing cell behaviors on the nanofibers of PCL-based PU and those of pure PCL. The more elastic PU nanofibers demonstrated more osteogenic and chondrogenic induction potential than PCL electrospun fibers. We suggested that the elastic nanofibers seeded with hMSC aggregates may be advantageous for cartilage and bone tissue engineering. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Mesenchymal stem cells (MSCs) are a promising cell source for tissue engineering applications mainly because of its multipotency to differentiate into various cell types such as osteoblasts [1], adipocytes [2], chondrocytes [3], myocytes [4] and neural cells [5]. Their fate is greatly influenced by both intrinsic and extrinsic signals, including cell–cell and cell–substrate interactions, the gradients of oxygen, nutrient, and protein concentrations, as well as the mechanical properties of the substrate [6]. Biological responses of stem cells could be quite different in two-dimensional (2D) and three-dimensional (3D) environments [7]. The traditional 2D culture method tends to separate single cells from their neighbors, which does not resemble the real in vivo situation. The 3D cell culture involves the generation of cellular aggregates or the growth of cells on 3D substrates (“scaffolds”), which may improve the cell–cell communication and differentiation potential of stem cells [8].
∗ Corresponding author at: Institute of Polymer Science and Engineering, National Taiwan University, Taipei 10617, Taiwan. Tel.: +886 2 33665313; fax: +886 2 33665237. E-mail address: [email protected] (S.-h. Hsu). http://dx.doi.org/10.1016/j.colsurfb.2014.07.017 0927-7765/© 2014 Elsevier B.V. All rights reserved.
Electrospinning, a fast and simple polymer processing technique, is not only capable of producing a wide variety of polymeric nanofibers, but also able to control the sizes, shapes, and porosity to achieve desired products [9]. Due to the above advantages, electrospinning is widely used in fabricating 3D scaffolds for regeneration of tissues such as blood vessel [10], bone [11], nerve [12], cartilage [13], skin [14], heart [15], and ligament [16]. Recent studies have shown that the adhesion, migration, proliferation, and differentiation of stem cells could be influenced by the topography [17], porosity [18], and mechanical properties of the substrate materials [19]. In particular, nanofibers that simulate the fibrillar structure of natural native extracellular matrix (ECM) can improve the differentiation efficiency of stem cells [20]. Electrospun substrates with high porosities or large pore sizes may have better proliferation capacities because cells can grow into the substrates [21]. Moreover, the osteogenesis of stem cells could be achieved by increasing the hardness of the matrix, while chondrogenesis could be induced by the softness of the substrate [22]. An ideal scaffold should mimic both the biomechanical properties and architecture of a natural tissue. Therefore, making nanofibrous scaffolds from elastic polymers such as polyurethane (PU) could be a potential approach for tissue engineering [23–25]. Cellular aggregates of MSCs have received much attention recently for their possible applications in tissue engineering
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
415
[26–29]. The behavior of MSC aggregates on electrospun 3D scaffolds remains unclear. There is no direct evidence that the differentiation capacity of aggregates after spreading on a fibrous matrix can still maintain their advantages over single (dispersed) cells on the same substrate. In this study, we first synthesized biodegradable polyurethane (PU) elastomers. Elastic fibers in nanoto micro-scale were then fabricated by electrospinning. The influence of topography and mechanical properties of different fibrous scaffolds on cell adhesion, proliferation, and differentiation potentials of MSC aggregates vs. single cells was investigated. Specifically, we sought to optimize the combination of electrospun fibers and MSC aggregates for cartilage and bone tissue engineering. 2. Materials and methods 2.1. PU synthesis The biodegradable PU used in this study was synthesized by a green process. Isophorone diisocyanate (IPDI, Evonik Degussa GmbH) was reacted with stirred (180 rpm) poly(-caprolactone) diol (PCL diol, molecular weight ∼2 kDa, Sigma) under nitrogen for 3 h at 75 ◦ C in the presence of 0.03% stannous octoate (Sn(Oct)2 , Alfa Aesar) as the catalyst. After prepolymerization, 2,2-bis(hydroxymethyl) propionic acid (DMPA, Sigma) and an appropriate amount (13 mL) of methyl ethyl ketone (MEK, J.T. Baker) were added to the reactor. After stirring for 1 h (180 rpm), the reaction temperature was then dropped to 50 ◦ C. Neutralization was completed by adding triethylamine (TEA, R.D.H) into the reactor. The neutralized prepolymer was then dispersed in deionized water (110 mL) and stirred vigorously at 1100 rpm before ethylenediamine (EDA, Tedia) in 110 mL of deionized water was further added with continuing stirring to complete the reaction. The stoichiometric ratio used for synthesis was optimized as IPDI/PCL diol/DMPA/EDA/TEA = 3.52:1:1:1.52:1 [30]. The solvent was removed to obtain the solid form of biodegradable PU. The molecular weight defined by gel permeation chromatography was ∼160 kDa. The weight percent of PCL segment in the final polymer was ∼65%. On the other hand, the pure PCL polymer was purchased (molecular weight 70–90 kDa, Sigma) and used as received. 2.2. Electrospinning process The setup of the electrospinning apparatus was constructed from a high-voltage power supplier (You-Shang Technical Corp., Taiwan), a syringe pump (KDS-100, KD Scientific, USA), a syringe with a stainless steel blunt-ended needle (20G, Terumo, Japan), and a plate type collector (Fig. 1(a)). The electrospinning was conducted in a custom-designed transparent acrylic chamber equipped with a hygrometer and a thermometer. The local humidity (40%) was controlled by a desiccant (silica gel). The chamber was placed in a humidity/temperature controlled room. In order to narrow down the distribution of fiber diameters, the mixed solvent system of a low-boiling point solvent 2,2,2-trifluoroethanol (TFE, Tedia; bp. 78 ◦ C) and a high-boiling point solvent N,N-dimethylacetamide (DMAc, Tedia; bp. 165 ◦ C) was used. PU was dissolved at the concentration of 6 wt% in the mixture of TFE and DMAc with different mixing ratios and stirred for overnight. The polymer solution was then poured into the syringe and drawn to the collecting plate by applying a high voltage. The flow rate, the applied voltage, and the distance between the nozzle and the collector could be adjusted. The temperature and the relative humidity were controlled at 21.5 ± 0.5 ◦ C and 40 ± 4%. After optimization, PU (5, 8, and 11 wt%) in a mixed solvent (wt% 60:40) was used to obtain electrospun membranes comprising of nanofibers (“ES1” groups), sub-microfibers (“ES2” groups), and microfibers (“ES3” groups). The relative humidity and temperature were controlled as mentioned. Other electrospinning parameters
Fig. 1. (a) The electrospinning apparatus. (b) The effect of mixed solvent (TFE/DMAc) ratio on the morphology of electrospun fibers.
such as the flow rate, the applied voltage, and the distance between nozzle and collector were varied in the ranges of 0.3–0.6 mL/h, 10–15 kV, and 15–32 cm, respectively. Electrospun meshes for cell studies were spun onto the PU-coated coverslip glass, while those for physico-chemical characterization were spun onto glass slides and then removed by immersion into ice water. The PU-coated coverslip glass was prepared by spin coating 3% solution in TFE onto 15 mm diameter coverslip glass (Assistant, Glaswarenfabrik Karl Hecht KG, Germany) at 2750 rpm for 20 s and dried in vacuum for 1 day, which also served as the control (“PU film” groups). PCL electrospun meshes (“PCL-ES” groups) were prepared for comparison. For the purpose, the solution of 10 wt% PCL in TFE was spun onto PCL-coated coverslip glass at a voltage of 20 kV, a flow rate of 0.3 mL/h, and a distance of 32 cm. The relative humidity and temperature were controlled as mentioned. The PCL-coated coverslip glass was prepared similarly to that of PU by 5% solution of PCL in TFE. All electrospun meshes were dried in vacuum for 2 days to remove residual solvent. 2.3. Characterization of electrospun fibers The electrospun meshes were examined by a scanning electron microscope (SEM, Hitachi S-4800) at an accelerating voltage
416
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
of 10 kV after sputter-coated with a 2 nm layer of carbon (Emitech K950X). The average fiber diameter was determined based on SEM images using the Image J software. The thickness without a supporting film was measured by a dial thickness gauge. The thickness of each mesh was kept as close as possible during the parameter optimization. The fiber density and porosity of the electrospun meshes were determined as follows. The volume and density of the samples were calculated based on their dimension and weight. The fiber density (, g/cm3 ) was estimated by dividing the weight of the meshes with the volume of the meshes. The porosity was then determined based on the equation porosity (%) = (1 − /0 ) × 100%, where 0 is the density of PU (1.167 g/cm3 ) or PCL (1.145 g/cm3 ). The water contact angle was measured by a goniometer (FTA-1000B, First Ten Angstrom Company, USA) at room temperature. Images of water drops (5 L) deposited on the surface of the membrane were used for water contact angle measurements. The mechanical properties of the meshes and PU films including the Young’s modulus, 100% modulus, tensile strength, and elongation were measured by a universal testing system (HT8504, Hung Ta) at a strain rate of 50 mm min−1 . All procedures followed the standard testing method (ASTM-D638). The meshes used for characterization were comprised only of electrospun meshes. The specimens were cut in standard sizes by a specially designed knife (parameters: overall length 38.2 mm, width of grip section 16.02 mm, gage length 23 mm, and gage width 4.67 mm; n = 7). 2.4. Cell culture Human bone marrow mesenchymal stem cells (hMSCs) were obtained from the Tulane Center for Preparation and Distribution of Adult Stem Cells. They were seeded in each well with a density of 5 × 104 cells per well, and cultured in the basal medium at 37 ◦ C with 5% CO2 for 3 days. Chitosan powder (molecular weight ∼416 kDa, 77.7% deacetylation, Fluka) was dissolved and stirred at room temperature in 1% aqueous acetic acid solution for 24 h to obtain a 1% chitosan solution. Chitosan membranes were made from casting 300 L of solution on 15 mm diameter coverslips and air-dried. The membranes were rinsed in 0.5 N NaOH for 1 min, washed extensively by phosphate buffered saline (PBS), sterilized with 75% ethanol for 1 h, washed five times with PBS, and placed into a 24-well tissue culture plate before use. The basal medium contained low glucose Dulbecco’s modified Eagle medium (DMEMLG) supplemented with 10% fetal bovine serum (FBS, Gibco), 1% penicillin-streptomycin (Invitrogen). After incubation for 3 days, hMSC aggregates were collected by aspiration and quantified using a CountessTM automated cell counter (Invitrogen). Cells seeded in the blank well (tissue culture polystyrene, TCPS) served as the control. 2.5. Cell culture on electrospun meshes All samples were sterilized under ultraviolet light for 3 h, and placed into 24-well tissue culture plates, where each well fit one coverslip glass. hMSC aggregates or single cells at a density of 4 × 104 cells in 1 mL culture medium were seeded on PU meshes (ES1, ES2, and ES3) and cultured in the basal medium for 1, 7, and 14 days, respectively. The medium was refreshed three times a week. To further compare PU and PCL electrospun meshes, hMSC aggregates (4 × 104 cells) were seeded on ES1 and PCL-ES meshes in the basal medium for 3 days. The medium was then replaced with the osteogenic induction medium or chondrogenic induction medium for 3 and 7 days to induce osteogenic or chondrogenic differentiation. The osteogenic induction medium contained 50 g/mL of
l-ascorbate-2-phosphate, 10−8 M dexamethasone, and 10 mM glycerophosphate. The chondrogenic induction medium contained 1% of insulin-transferrin-selenium (ITS) premix 100×, 50 g/mL of l-ascorbate-2-phosphate, 0.1 M of dexamethasone, 10 ng/mL of TGF-3, and 40 g/mL of l-proline. The induction medium was refreshed twice a week. The two control groups were hMSC aggregates seeded on “PU films” and hMSC single cells seeded on ES1 meshes (“ES1-SC” groups). The morphology of cells after seeding was examined by an inverted microscope (Leica, DMIRB). The cell number of aggregates was ensured by a parallel experiment where cell counting was performed after incubation in basal medium for 3 days. hMSC spheroids were collected by aspiration and trypsinized for cell counting using the automated cell counter.
2.6. Cell proliferation and gene expression Cells on different substrates were collected at the end of culture period, and the number was determined by the fluorometric assay of DNA. Briefly, samples were lyophilized and decomposed in 1.5 mL of papain solution, which contained 47.6 g/mL of papain (Sigma) in buffer solution with 55 mM of sodium citrate (Showa), 150 mM of sodium chloride (Sigma), 5 mM of cysteine HCL (Sigma), and 5 mM of Na2EDTA (Sigma), for 24 h at 60 ◦ C. Hoechst 33258 fluorescent dye solution (5 mL) which contained 0.1 g/mL of Hoechst 33258 (Sigma) in pH 7.4 buffer solution [10 mM Tris (Tedia), 1 mM Na2 EDTA (Sigma), and 0.1 mM sodium chloride (Sigma)] was then added to 0.5 mL of decomposed solution. The intensity of fluorescence (excitation wavelength 365 nm, emission wavelength 458 nm) was determined by a multidetection microplate reader (Molecular Devices SpectraMax M5, USA). The dye solution served as the background. To evaluate the osteogenic and chondrogenic gene expression, total RNA from collected samples was extracted by Trizol® reagent (Invitrogen, USA). cDNA was acquired via reverse transcription using RevertAidTM First Strand cDNA Synthesis kit (MBI Fermentas, St. Leon-Rot, Germany). Polymerase chain reaction (PCR) was used to amplify the gene expression of human runt-related transcription factor 2 (RUNX2) and human SRY (sex determining region Y)-box 9 (SOX9) with T100TM Thermal Cycler (BIO-RAD, USA) and estimated by capillary electrophoresis (HDA-GT12TM , eGene, USA). The gene expression analyses were carried out by measuring band intensity after electrophoresis. Results of gene expression of each sample were normalized to their own glyceraldehyde 3-phosphate dehydrogenase (GAPDH; housekeeping gene). The primer sequences are as follows: human GAPDH, 5 -CAAGGTCATCCATGACAACTTTG3 (forward) and 5 -GTCCACCACCCTGTTGCTGTAG-3 (reverse), human RUNX2, 5 -TTGCAGCCATAAGAGGGTAG-3 (forward) and 5 -GTCACTTTCTTGGAGCAGGA-3 (reverse), human SOX9, 5 -AAAAGGCAAGCAAAGGAGAT-3 (forward) and 5 -AAAAGGGATGGACAAAAAGG-3 (reverse). The cycling condition includes 35 cycles of 30 s at 95 ◦ C (denaturation), 30 s at 47 or 58 ◦ C (annealing), and 1 min at 72 ◦ C (extension) for SOX9, RUNX2, and GAPDH.
2.7. Statistical analysis Each experiment was performed independently for three times and with multiple samples. Numerical values were presented as mean ± standard deviation. Statistical differences between two groups were evaluated by unpaired two tailed t-tests. One way ANOVA with Tukey’s multiple comparison tests was used to determine the statistical differences of the results that involved more than two groups. p values less than 0.05 meant significant differences.
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
417
Table 1 Electrospinning parameters for producing meshes with three different fiber diameters and the corresponding fiber density and porosity (n = 7). All data are significantly different from each other, except the fiber diameter between ES1 and PCL-ES and the weight between ES1 and ES2. Meshesa
Solution conc. (wt%)
ES1 ES2 ES3 PCL-ES
5 8 11 10
a
Flow rate (L/min) 5 5 10 5
Voltage (kV)
Collector distance (cm)
Fiber diameter (m)
15 10 10 20
32 15 32 32
0.31 0.67 1.47 0.37
± ± ± ±
0.05 0.09 0.09 0.06
Weight (g)
Density of scaffold (g/cm3 ) 0.48 0.40 0.32 0.17
± ± ± ±
0.01 0.02 0.02 0.03
0.062 0.052 0.035 0.167
± ± ± ±
Porosity (%)
0.006 0.008 0.006 0.003
58.6 64.9 72.6 85.5
± ± ± ±
1.3 1.1 2.0 2.6
ES1, ES2, and ES3: PU fibers made from the mixed solvent of TFE and DMAc. PCL-ES: PCL fibers made from the solvent TFE.
Table 2 The tensile mechanical properties and water contact angle of the electrospun meshes at 25 ◦ C (n = 7). All data are significantly different from each other, except the Young’s modulus between ES2 and PCL-ES, Young’s modulus between ES3 and PCL-ES, 100% modulus between ES2 and ES3, strength between ES2 and ES3, and elongation between ES1 and ES2, and contact angle between ES2 and ES3. Samples
Young’s modulus (MPa)
PU film (control) ES1 ES2 ES3 PCL-ES
30.5 10.8 3.5 5.9 4.3
± ± ± ± ±
0.6 1.8 0.3 1.4 1.3
100% modulus (MPa) 7.1 4.2 1.9 1.7 2.6
± ± ± ± ±
0.3 0.7 0.2 0.3 0.4
Tensile strength (MPa) 37.4 8.5 4.5 4.0 3.1
± ± ± ± ±
1.9 1.7 0.3 0.6 0.2
Elongation (%) 582.3 315.4 327.8 444.4 107.0
± ± ± ± ±
24.0 40.7 11.2 46.5 12.7
Contact angle (◦ ) 77.7 91.4 107.8 110.5 127.7
± ± ± ± ±
0.7 7.0 7.0 3.0 2.9
Fig. 2. The morphology of PU electrospun meshes with different fiber diameters and that of PCL electrospun meshes. Abbreviation: ES1, PU electrospun meshes with 200–400 nm fiber diameter. ES2, PU electrospun meshes with 600–800 nm fiber diameter. ES3, PU electrospun meshes with 1.4–1.6 m fiber diameter. PCL-ES, PCL electrospun meshes with 200–400 nm fiber diameter.
418
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
Fig. 3. The adhesion and proliferation of aggregates and single cells on electrospun meshes after incubation in basal medium. (a) The schematics of the culture period (1–14 days). (b) The number of cells and (c) the seeding efficiency measured at 1 day. (e) The number of cells after incubation for 7 days. (f) The number of cells after incubation for 14 days. *p < 0.05; **p < 0.01.
3. Results
3.2. Characterization of electrospun meshes
3.1. Fibrous PU meshes by electrospinning
The fiber density and porosity of PU and PCL electrospun membranes are listed in Table 1. As the fiber diameter of PU meshes decreased, the density of the mesh increased from 0.32 ± 0.02 g/cm3 to 0.48 ± 0.02 g/cm3 and the porosity decreased from 72.6 ± 2.0% to 58.6 ± 1.3%. Among all meshes, PCL-ES had the lowest fiber density and the highest porosity. The contact angle and mechanical properties are listed in Table 2. As the fiber diameter of PU meshes decreased, the contact angle decreased from 110.5 ± 3.0◦ to 91.4 ± 7.0◦ . PCL-ES was the most hydrophobic (127.7 ± 2.9◦ ) among all. All PU meshes had outstanding elongation (>300%). Among them, ES1 meshes had the smallest elongation (315.4 ± 40.7%) and the largest Young’s modulus (10.8 ± 1.8 MPa), 100% modulus (4.2 ± 0.7 MPa), and tensile strength (8.5 ± 1.7 MPa). The modulus and strength of ES1 were about two times higher than those of ES2 and ES3 meshes. The control PU film had the lowest contact angle and highest mechanical properties. PCL-ES had the smallest elongation (∼100%).
Electrospun meshes with different topographies were made in this study to examine the proliferation and differentiation capacities of hMSCs on biodegradable and elastic PU materials. To produce electrospun fibers in nanometric scale, solvent selection is very important. The effects of mixed solvent of TFE and DMAc in different mixing ratios on the morphology of electrospun fibers are shown in Fig. 1(b). When the content of DMAc in the mixed solvent reached 50%, a part of the electrospun fibers stuck together. Such a tendency increased with the DMAc content. The optimized mixing ratio of TFE to DMAc was 3:2. Fibers of nano- to micro-scale could be obtained from PU solution of different concentrations and in specific electrospinning conditions. The thickness was about 0.4 mm for each mesh (without the supporting film). The dimension was (20 ± 0.1 mm) × (15 ± 0.1 mm) (where the sample size n was 7). The details of electrospun parameters of all the electrospun meshes are listed in Table 1 and the corresponding topographies are shown in Fig. 2. The average fiber diameter of ES1, ES2, and ES3 meshes was 0.31 ± 0.05 m, 0.67 ± 0.09 m, and 1.47 ± 0.09 m, respectively. SEM images showed that fiber diameter increased with the increased concentration of the polymer solution. In addition, PCL-ES nanofibers for comparison were produced from 10 wt% PCL solution in TFE and had a mean diameter of 0.37 ± 0.06 m.
3.3. The effect of fiber diameter on hMSC behavior Cells were distributed uniformly within the meshes. The number of hMSC single cells and aggregates on the electrospun fibers were tested at 1, 7, and 14 days in basal medium (Fig. 3(a)). When seeding hMSC single (dispersed) cells, the cell seeding efficiency increased with the increased fiber diameter. On the other hand,
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
419
Fig. 4. Gene expressions of aggregates and single cells on electrospun meshes after incubation in basal medium. (a) The schematics of the culture period (3–7 days). (b) RUNX2 and SOX9 expression before seeding on the electrospun meshes (0 day). (c) RUNX2 expression after incubation for 3 and 7 days. (d) SOX9 expression after incubation for 3 and 7 days. Results of gene expression of each sample were normalized to their own GAPDH. RUNX2, runt-related transcription factor 2; SOX9, SRY (sex determining region Y)-box 9. *p < 0.05; **p < 0.01.
hMSC aggregates had similar seeding efficiency on all PU electrospun meshes (Fig. 3(b) and (c)). Compared to single cells, hMSC aggregates showed better cell proliferative capacity on ES1 after incubation for 7 and 14 days (Fig. 3(d) and (e)). In addition, the cell proliferation of hMSC single cells slightly increased with the increased fiber diameter, while hMSC aggregates had similar proliferative capacity on ES2 and ES3 at 7 and 14 days. The gene expression of hMSC single cells and aggregates on the electrospun fibers were tested at 3 and 7 days in basal medium (Fig. 4(a)). SOX9 is an important chondrogenic marker gene, which encodes the transcription factor critical to chondrogenesis and cartilage development. RUNX2 is an important osteogenic marker gene, which encodes the transcription factor critical to osteogenesis
and bone development. The aggregates showed high SOX9 expression before seeding on the substrate (Fig. 4(b)). It also appeared that the SOX9 gene expression of aggregates on electrospun membranes increased with the incubation time, while that of single cells showed the opposite (Fig. 4(e) and (f)). After 7 days, aggregates had the highest SOX9 expression on ES1 meshes. The RUNX2 expression of hMSC single cells and aggregates were upregulated on all PU meshes at 3 days (Fig. 4(c)). At 7 days, hMSC aggregates showed better RUNX2 expression than single cells on all PU meshes (Fig. 4(d)). hMSC aggregates on ES1 had the highest RUNX2 expression at both 3 and 7 days. Overall, the data revealed that hMSC aggregates had better osteogenic and chondrogenic differentiation potential than single cells on electrospun meshes without the induction medium;
420
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
14 days (Fig. S2(a) and (b)). hMSC aggregates on ES1 meshes also revealed a significantly higher glycosaminoglycan (GAG, cartilage matrix component) content than those on TCPS after the chondrogenic induction for 14 days (Fig. S2(c)). ES1 appeared to be more favorable than PCL meshes regarding GAG secretion (p = 0.057, not reaching significance p < 0.05). On ES1 meshes, hMSC aggregates demonstrated significantly greater capacities for cartilage and bone matrix deposition than single cells. 4. Discussion
Fig. 5. The number of total cells after aggregates were seeded on ES1, PCL-ES, or PU film, incubated in basal medium for 3 days, and induced for 7 days. (a) The schematics of the culture period. (b) The cell number after induction. Abbreviation: ES1-SC, single cells seeded on ES1 and incubated in the same condition. *p < 0.05; **p < 0.01.
and ES1 meshes (i.e. 200–400 nm nanofibers) provided a better environmental cue for differentiation than the other fibers. 3.4. The effect of electrospun fibers on hMSC differentiation The number of hMSC single cells and aggregates on all samples were examined after incubation in basal medium for 3 days and further incubation in induction medium for 7 days (Fig. 5(a)). The cell number for hMSC aggregates seeded on the PU film was higher than that on ES1 and PCL-ES at the end of the culture period (Fig. 5(b)). Besides, the cell number was higher for hMSC aggregates seeded on ES1 meshes as compared to PCL-ES meshes. On the other hand, the cell number was higher for hMSC single cells (vs. aggregates) seeded on ES1. After osteogenic induction, hMSC aggregates on ES1 meshes showed the highest RUNX2 expression than the others (Fig. 6(a)). On the other hand, after chondrogenic induction for 7 days, there was no significant difference between hMSC aggregates on ES1, PCL-ES, and PU film, though it appeared that ES1 also had the best chondrogenic effect. The morphology of hMSCs on different substrates is shown in Supplemental data (Fig. S1). After incubation for 3 days, hMSC aggregates had already spread on PCL-ES and PU film, but some remained not spread (size between 80 and 180 m) on ES1 meshes. This tendency persisted until 7 days post induction. As shown in Supplemental data (Fig. S2), hMSC aggregates on ES1 meshes revealed a significantly higher calcium content (bone matrix component) than those on PCL meshes, PU flat films, or the blank well (TCPS) control after the osteogenic induction for
The high-boiling point solvent may evaporate incompletely during the electrospinning process, while the low-boiling point solvent may evaporate fast to generate fibers with large distribution of diameters. A mixed solvent system of TFE and DMAc, which appeared here for the first time, was used to overcome the above two problems frequently encountered. PU electrospun fibers of three different diameters could be achieved by adjusting the polymer concentration, the distance between nozzle and collector, the voltage, and the flow rate, all of which affected the fiber diameter. Electrospun PU fibers with decreasing fiber diameters showed an increase in hydrophilicity. This might be due to the fact that the opportunity of the hydrophilic functional groups exposed to the air would be higher when the fiber diameter was lower. Although the functional groups of PU are mainly considered as hydrophobic, PU also had many hydrophilic functional groups such as COOH, OH, COO, and NH. Besides, a few studies mentioned that electrospun fibers with thinner fiber diameter had pores of small sizes, which may increase the water absorption rate by capillary action and decrease the contact angle accordingly [31]. Moreover, the largerdiameter fibers could sometimes have big pores filled with air, which may increase the hydrophobicity [32]. The tensile properties of electrospun fibers including Young’s modulus, 100% modulus, and tensile strength were always higher in ES1 compared to the other two (ES2 and ES3). The alignment of nanofibers may influence the mechanical properties, i.e. highly alignment fibers normally show better mechanical properties. In the study, we did not control the alignment of the nanofibers; therefore, the nanofibers were randomly oriented, which were confirmed by our observation. On the other hand, the diameter and strength of a single fiber are both associated with the packing and orientation of the polymeric chains [33]. Our experimental result revealed that thinner nanofibers had greater strength. This may be attributed to the greater stretching deformation generated during single fiber formation, which may align the polymeric chains and subsequently lead to the greater strength of each single fiber [34]. Besides, fiber diameter can further affect fiber density and porosity; and in most cases, as the fiber diameter decreases, the fiber density increases in number, which also help enhance the mechanical properties [35]. The adhesion and proliferation of MSC aggregates and single cells in basal medium showed that the seeding efficiency of MSC aggregates were maintained in 70–90% on PU electrospun fibers measured at 1 day. MSC single cells showed only 30% seeding efficiency on ES1. It was believed that MSC aggregates produced more extracellular matrix such as fibronectin, which may promote cell adhesion on electrospun fibers [36,37]. The favorable cell adhesion may account for the higher seeding efficiency. In general, we found that fibers with smaller diameter promoted the differentiation while those with larger diameter enhanced proliferation. The enhanced cell proliferation may be associated with the cell ingrowth to pores on larger-diameter fibers [38]. The current results also indicated that MSC aggregates could be seeded without the help of plasma treatment or coating extracellular matrix and could well proliferate on the PU electrospun fibers. Although MSC aggregates tended to spread with the increased culture time on electrospun fibers, they had better differentiation
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
421
Fig. 6. (a) Gene expression for cells after induction for 7 days as indicated in Fig. 5(a). Abbreviation: ES1-basal, aggregates was seeded on ES1 and incubated in basal medium for 10 days. Results of gene expression of each sample were normalized to their own GAPDH. *p < 0.05; **p < 0.01. (b) The hypothetical mechanism for gene regulation on electrospun meshes.
potentials than single cells on electrospun fibers. It also seemed that MSC aggregates were more sensitive to the environmental (substrate) cue such as topography and hardness. After incubation for 7 days, MSC aggregates on ES1 showed higher RUNX2 expression than those on ES2 and ES3. ES1 had higher Young’s modulus and had the topography in nanoscale, both of which may contribute to enhance the osteogenic differentiation of stem cells [39,40]. In addition, MSC aggregates showed higher SOX9 expression on ES1 after incubation for 7 days. This might be attributed to the higher cell density and the biomimetic ECM-like structure of ES1 [41–43]. The behavior of cells on electrospun meshes may be related to their inherent properties before seeding. Prior to seeding, MSC aggregates showed higher SOX9 expression than single cells probably because of high cell density in the MSC aggregates. As the MSC aggregates spread on electrospun fibers, cell density may be decreased, which in turn favored cell–substrate interaction. After incubation for another 7 days, the expression of SOX9 obviously increased again. The hypothetical mechanism regarding how the MSC aggregates responded to the electrospun fibers is summarized in Fig. 6(b). Comparing the effect of fibers from different materials on cell behavior, it appeared that the hardness of substrate could significantly influence the osteogenic differentiation of MSCs. The biodegradable PU used in this study contained 65 wt% PCL segments in the structure. Nanofibers of PU (i.e. ES1) with higher modulus and elongation induced more osteogenic differentiation of MSCs than nanofibers of PCL (i.e. PCL-ES) having lower modulus and elongation. However, we could not tell if this was caused by chemistry,
hardness, or elasticity. Besides, the MSC aggregates spread rather slowly on ES1 meshes as noted in Supplemental data (Fig. S1). Some MSC aggregates remained intact even at 7 days post induction. This cell clustering effect may improve the cell–cell communication that is critical for the differentiation of stem cells. On the other hand, it was observed that the inclination for chondrogenic differentiation was all lower than that for osteogenic differentiation on all substrates. As the MSC aggregates gradually spread on the substrates, the 3D microenvironment which promoted the chondrogenic differentiation could no longer exist. As mentioned, some MSC aggregates remained intact on ES1 for a period, which may account for the greater chondrogenic differentiation on ES1. The reason behind the lower chondrogenic differentiation of MSCs on PCL-ES vs. ES1 may be the smaller amount of cells on PCL-ES (Fig. 5). MSC aggregates used in this study were shown to have better differentiation capacities upon induction. They may replace single cell suspension to be used in the cell-based therapy. However, the MSC aggregates are still susceptible to cell loss during administration. By seeding MSC aggregates onto the electrospun fibers, the combined constructs may help translate the better differentiation efficiency of the MSC aggregates. Our study demonstrates that even if the MSC aggregates may eventually spread and lose their morphology on the substrate, the differentiation potential remains to be better than single cells. In particular, the elastic PU electrospun nanofibers fabricated in this study promote the chondrogenic and osteogenic differentiation of MSC aggregates, which may have potential applications in cartilage and bone tissue engineering.
422
Y.-C. Kuo et al. / Colloids and Surfaces B: Biointerfaces 122 (2014) 414–422
Acknowledgments This research was supported by the Program for Stem Cell and Regenerative Medicine Frontier Research (NSC102-2321-B-002033) sponsored by the National Science Council, Taiwan, R.O.C. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.colsurfb. 2014.07.017. References [1] Y. Mohammadi, M. Soleimani, M. Fallahi-Sichani, A. Gazme, V. Haddadi-Asl, E. Arefian, J. Kiani, R. Moradi, A. Atashi, N. Ahmadbeigi, Int. J. Artif. Organs 30 (2007) 204. [2] W.J. Li, R. Tuli, X. Huang, P. Laquerriere, R.S. Tuan, Biomaterials 26 (2005) 5158. [3] J. Liao, X. Guo, K.J. Grande-Allen, F.K. Kasper, A.G. Mikos, Biomaterials 31 (2010) 8911. [4] S. Sahoo, H. Ouyang, J.C. Goh, T.E. Tay, S.L. Toh, Tissue Eng. 12 (2006) 91. [5] K.S. Bae, J.B. Park, H.S. Kim, D.S. Kim, D.J. Park, S.J. Kang, Yonsei Med. J. 52 (2011) 401. [6] M.P. Lutolf, P.M. Gilbert, H.M. Blau, Nature 462 (2009) 433–441. [7] E. Cukierman, R. Pankov, D.R. Stevens, K.M. Yamada, Science 294 (2001) 1708. [8] J.E. Frith, B. Thomson, P.G. Genever, Tissue Eng. C: Methods 16 (2010) 735. [9] N. Bhardwaj, S.C. Kundu, Biotechnol. Adv. 28 (2010) 325. [10] C.Y. Xu, R. Inai, M. Kotaki, S. Ramakrishna, Biomaterials 25 (2004) 877. [11] N.G. Rim, S.J. Kim, Y.M. Shin, I. Jun, D.W. Lim, J.H. Park, H. Shin, Colloids Surf. B 91 (2012) 189. [12] E.M. Jeffries, Y. Wang, Biofabrication 5 (2013) 035015. [13] A. Fertala, W.B. Han, F.K. Ko, J. Biomed. Mater. Res. 17 (2001) 48. [14] J. Venugopal, S. Ramakrishna, Tissue Eng. 11 (2005) 847. [15] X. Zong, H. Bien, C.Y. Chung, L. Yin, D. Fang, B.S. Hsiao, B. Chu, E. Entcheva, Biomaterials 26 (2005) 5330. [16] C.H. Lee, H.J. Shin, I.H. Cho, Y.M. Kang, I.A. Kim, K.D. Park, J.W. Shin, Biomaterials 26 (2005) 1261.
[17] M. Chen, P.K. Patra, S.B. Warner, S. Bhowmick, Biophys. Rev. Lett. 1 (2006) 153. [18] A.S. Badami, M.R. Kreke, M.S. Thompson, J.S. Riffle, A.S. Goldstein, Biomaterials 27 (2006) 596. [19] Y. Yang, K. Kulangara, R.T. Lam, R. Dharmawan, K.W. Leong, ACS Nano 6 (2012) 8591. [20] C.T. Laurencin, A.M. Ambrosio, M.D. Borden, J.J. Cooper, Annu. Rev. Biomed. Eng. 1 (1996) 19. [21] J.B. Lee, S.I. Jeong, M.S. Bae, D.H. Yang, D.N. Heo, C.H. Kim, E. Alsberg, I.K. Kwon, Tissue Eng. Part A 17 (2011) 2695. [22] J.S. Park, J.S. Chu, A.D. Tsou, R. Diop, Z. Tang, A. Wang, S. Li, Biomaterials 32 (2011) 3921. [23] H.-W. Chien, J.-Y. Lai, W.-B. Tsai, Colloids Surf. B 116 (2014) 576. [24] R.O. Ritchie, M.J. Buehler, P. Hansma, Phys. Today 62 (2009) 41. [25] A.J. Engler, S. Sen, H.L. Sweeney, D.E. Discher, Cell 126 (2006) 677. [26] S.-h. Hsu, G.S. Huang, S.Y. Lin, F. Feng, T.T. Ho, Y.C. Liao, Tissue Eng. Part A 18 (2012) 67. [27] H.Y. Yeh, B.H. Liu, S.-h. Hsu, Biomaterials 33 (2012) 8943. [28] B.H. Liu, H.Y. Yeh, Y.C. Lin, M.H. Wang, D.C. Chen, B.H. Lee, S.-h. Hsu, Biores. Open Access 2 (2013) 28. [29] S.-h. Hsu, Y.H. Ni, Y.C. Lee, Macromol. Biosci. 13 (2013) 1100. [30] Taiwan patents (#I427091 and 102100076) and U.S. patent applications (13/317,651 Biodegradable and biocompatible waterborne polyurethane and 14/145,393 Biocompatible and biodegradable elastomer). [31] Y. Liu, N. Wu, Q. Wei, Y. Cai, A. Wei, J. Appl. Polym. Sci. 110 (2008) 3172. [32] N. Wang, K. Burugapalli, W. Song, J. Halls, F. Moussy, Y. Zheng, Y. Ma, Z. Wu, K. Li, J. Memb. Sci. 36 (2000) 75. [33] J. Yao, C.W.M. Bastiaansen, T. Peijs, Fibers 2 (2014) 158. [34] C.L. Pai, M.C. Boyce, G.C. Rutledge, Polymer 52 (2011) 2295. [35] I.K. Kwon, S. Kidoaki, T. Matsuda, Biomaterials 26 (2005) 3929. [36] C.L. Li, T. Tian, K.J. Nan, N. Zhao, Y.H. Guo, J. Cui, J. Wang, W.G. Zhang, Oncol. Rep. 20 (2008) 1465. [37] M.T. Santini, G. Rainaldi, P.L. Indovina, Crit. Rev. Oncol. Hematol. 36 (2000) 75. [38] J. Rnjak-Kovacina, S.G. Wise, Z. Li, P.K. Maitz, C.J. Young, Y. Wang, A.S. Weiss, Biomaterials 32 (2011) 6729. [39] K.J. Burg, S. Porter, J.F. Kellam, Biomaterials 21 (2000) 2347. [40] X.H. Chu, X.L. Shi, Z.Q. Feng, Z.Z. Gu, Y.T. Ding, Biotechnol. Lett. 31 (2009) 347. [41] M. Takagi, Y. Umetsu, M. Fujiwara, S. Wakitani, J. Biosci. Bioeng. 103 (2007) 98. [42] L. Zhang, P. Su, C. Xu, J. Yang, W. Yu, D. Huang, Biotechnol. Lett. 32 (2012) 1339. [43] J.M. Coburn, M. Gibson, S. Monagle, Z. Patterson, J.H. Elisseeff, Proc. Natl. Acad. Sci. U.S.A. 109 (2012) 10012.
