- Email: [email protected]
Characterization of a biodegradable electrospun polyurethane nanofiber scaffold: Mechanical properties and cytotoxicity
Acta Biomaterialia 6 (2010) 3847–3855
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Characterization of a biodegradable electrospun polyurethane nanofiber scaffold: Mechanical properties and cytotoxicity Masoud Yeganegi a,b, Rita A. Kandel a,b,c, J. Paul Santerre b,d,* a
CIHR – Bioengineering of Skeletal Tissues Team, Mount Sinai Hospital, Toronto, Canada M5G 1X5 Institute of Biomaterials and Biomedical Engineering, Department of Materials Science and Engineering, University of Toronto, Toronto, Canada M5S 3G9 c Department of Pathobiology and Laboratory Medicine, Mt. Sinai Hospital, University of Toronto, Toronto, Canada M5G 1X5 d Faculty of Dentistry, University of Toronto, Toronto, Canada M5G 1G6 b
a r t i c l e
i n f o
Article history: Received 1 March 2010 Received in revised form 18 April 2010 Accepted 6 May 2010 Available online 11 May 2010 Keywords: Annulus fibrosus Biodegradation Polyurethane Tissue engineering Nanofiber
a b s t r a c t The current study analyzes the biodegradation of a polycarbonate polyurethane scaffold intended for the growth of a tissue-engineered annulus fibrosus (AF) disc component. Electrospun scaffolds with random and aligned nanofiber configurations were fabricated using a biodegradable polycarbonate urethane with and without an anionic surface modifier (anionic dihydroxyl oligomer), and the mechanical behavior of the scaffolds was examined during a 4 week biodegradation study. Both the tensile strength and initial modulus of aligned scaffolds (r = 14 ± 1 MPa, E = 46 ± 3 MPa) were found to be higher than those of random fiber scaffolds (r = 1.9 ± 0.4 MPa, E = 2.1 ± 0.2 MPa) prior to degradation. Following initial wetting of the scaffold, the initial modulus of the aligned samples showed a significant decrease (dry: 46 ± 3 MPa; pre-wetted: 9 ± 1 MPa, p < 0.001). The modulus remained relatively constant during the remainder of the 4 week incubation period (aligned at 4 weeks: 8.0 ± 0.3 MPa). The tensile strength for aligned fiber scaffolds was affected in the same manner. Similar changes were not observed for the initial modulus of the random scaffold configuration. Biodegradation of the scaffold in the presence of cholesterol esterase (a monocyte derived enzyme) yielded a 0.5 mg week–1 weight loss. The soluble and non-soluble degradation products were found to be non-toxic to bovine AF cells grown in vitro. The consistent rate of material degradation along with stable mechanical properties comparable to those of native AF tissue and the absence of cytotoxic effects make this polymer a suitable biomaterial candidate for further investigation into its use for tissue-engineering annulus fibrosus. Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Back pain is ranked as the most prevalent chronic disease for people under the age of 60 [1]. Degenerative disc disease may contribute to the pathogenesis of lower back pain and involves the progressive degeneration of the intervertebral disc (IVD). The intact disc is necessary to support compressive and bending stresses while providing flexibility to the spine [2]. Currently, existing treatments include discectomy, the use of a prosthetic substitute or the fusion of adjacent vertebrae, none of which are optimal. Spinal fusion may be effective in some cases, but a number of patients can develop further degeneration, due to reduced flexibility and increased stress in adjacent segments [3–6]. Complications can also occur in patients undergoing disc replacement with synthetic substrates. Dislocations, slippage, wear and mechanical failure,
* Corresponding author at: Institute of Biomaterials and Biomedical Engineering, Department of Materials Science and Engineering, University of Toronto, Toronto, Canada M5S 3G9. Tel.: +1 416 979 4903x4341. E-mail address: [email protected] (J.P. Santerre).
although rare, have been reported [7]. The formation of wear debris can induce an inflammatory response mediated by various cytokines, leading to pain, osteolysis, fibrous tissue formation, prosthetic loosening and pain [7,8]. The human vertebral column is made up of 26 vertebral bodies that provide support to the body and protect the spinal cord. The intervertebral discs lying between the vertebrae provide flexibility and help dissipate mechanical loads and shocks that would otherwise damage the vertebral column [9]. The intervertebral discs are composed of the annulus fibrosus (AF), a fibrocartilaginous tissue that surrounds the gelatinous inner nucleus pulposus. The hyaline cartilage endplates found at each end represent the anatomical limits of the disc and contribute to the interface of the disc and bone. The mechanical properties of intervertebral discs are complex and the literature on this topic shows significant variability [10–13], perhaps in part due to the diversity of experimental conditions and mechanical models used to measure material properties. Studies indicate that the annulus fibrosus exhibits both matrix viscoelastic (flow-independent) and biphasic viscoelastic
1742-7061/$ - see front matter Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2010.05.003
3848
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
(flow-dependent) behavior [13]. The elastic modulus of lamellae for the annulus fibrosus has been found to vary with radial position in the disc, with values ranging from 5 ± 4 MPa for posterior inner AF to 10 ± 6 MPa for anterior inner AF and 49 ± 32 MPa for anterior outer AF [14]. The ultimate stress of lamellae for the annulus fibrosus also varies radially from 0.9 ± 0.3 to 3.3 ± 1.3 MPa for anterior outer AF [14]. Thus the forces experienced by the AF tissue can be quite high and therefore an intact AF is critical to proper disc function. One alternative strategy for the treatment of intervertebral disc degeneration is the replacement of the diseased tissue by a tissueengineered substitute [15–17]. Tissue engineering of the annulus fibrosus is particularly challenging due to the complex structure of the tissue. The AF, which is responsible for withstanding circumferential tensile forces and to a lesser extent compressive forces [14], consists of concentric lamella, made up of collagen fibers and oriented at approximately 60° to the vertical. The alignment of these fibers alternates between successive lamellae and is of great importance to the functional nature of the annulus fibrosus [18]. Efforts to produce AF tissue in vitro have involved various polymeric scaffolds including poly-DL-lactic acid (PDLLA)/45S5 BioglassÒ films, polyglycolic acid (PGA), collagen/hyaluronan, collagen/glycosaminoglycans (GAGs), atelocollagen and alginate scaffolds [19–24]. However, there are various issues that have yet to be overcome. Nucleus pulposus (NP) and AF cells cultured in alginate scaffolds have been found to adopt similar phenotypes rather than their respective native phenotypes [25]. Both cell types displayed an NP-like phenotype with rounded chondrocyte-like morphology, higher than normal levels of GAG and type II collagen vs. type I collagen. Similar problems have been encountered in collagen-hyaloronan scaffolds [23]. Polyglycolic-based polymers create acidic by-products throughout their biodegradation, which can significantly alter cell behavior and tissue production, and possibly cause cell death [26,27]. While the use of PDLLA/45S5 Bioglass has produced AF tissue with predominantly type I collagen, typical of native AF tissue, there exists an inherent challenge in reproducing the aligned structural arrangement of annulus tissue [28]. A potential alternative to the above materials for tissue engineering of an IVD consist of polycarbonate urethane polymers, owing to their relative biocompatibility [29,30], the possibility of tailoring their biodegradability and the inherent reproducible nature of the materials [29–32]. Recent work has shown that a degradable polycarbonate urethane (PU) previously reported in Ref. [32] could be processed, with the addition of anionic dihydroxyl oligomers (ADO), which enhance cell attachment [33], using electrospinning applications to fabricate aligned fibers. These structures mimic the aligned nature of the AF, and thus provide an appropriate scaffold to support the growth of such a tissue. Specifically, such topographical properties have been shown to influence cell alignment [17]. While previous studies have used random nanofiber polymeric scaffolds for IVD tissue generation [34], recent efforts have been aimed at utilizing the capacity of aligned electrospun scaffolds to generate aligned AF tissue. Nerurkar et al. [35–37] have assessed the mechanics of aligned poly(e-caprolactone) scaffolds and their ability to induce alignment in AF cells and tissue. The high surface to volume ratio of these scaffolds is also expected to favor cell attachment and retention of cell phenotype [33,38]. The purpose of this study was to characterize a biodegradable electrospun PU nanofiber biomaterial and determine its suitability for use in repairing the AF. Given the mechanical function of AF, it was important to assess the mechanical properties of the proposed polycarbonate-based urethane scaffolds containing the ADO additives which had not previously been reported on. Along with an emphasis on the mechanical consequences of the materials and the scaffold’s inherent biodegradable nature, this study was also
designed to determine the potential cytotoxic effects of the PU and the ADO’s biodegradation products on AF cells, as the latter materials represent a new class of scaffold polymers that are quite distinct in terms of balanced pH degradation by-products, from the classical degradable acidic by-product-producing polyesters, such as PDLLA, PGA and polycaprolactone-based polymers. 2. Materials and methods 2.1. Materials Polyurethanes have been used in biomedical devices since the 1960s. Efforts in the 1980s and 1990s had previously been directed at producing biostable polyurethanes; however, in the past decade, the focus has shifted to developing bioactive/biocompatible and biodegradable polyurethanes for the purpose of tissue engineering or regeneration [30,39]. In this study, PU was synthesized, as has been described [40], with hexane diisocyanate:polycarbonate diol:butane diol in a molar ratio of 3:2:1. In addition, an ADO additive was synthesized through the reaction of lysine diisocyanate, polytetramethylene oxide and hydroxyethyl methacrylate (HEMA), followed by subsequent hydrolysis of ester groups on the lysine and HEMA moieties [33]. ADO has been previously shown to increase AF cellular adhesion and collagen production, mediated by protein adsorption to the surface [33], taking advantage of both the polar and ionic function of the material. In more recent work, other groups have also recognized the importance of using charged materials to generate nanofiber scaffold systems, where increased tissue production has been observed [41]. Electrospinning was employed to fabricate either random or aligned nanofiber scaffolds using the base polymer (PU) with or without ADO (0.5 wt.%). The polymer was dissolved in 1,1,1,3,3,3hexafluora-2-propanol at a concentration of 16 wt.%. The scaffold was then electrospun by injecting the polymer solution from a metallic syringe at a rate of 0.5 ml h–1 onto the collecting surface. An 18 kV difference was applied across the syringe and the collecting surface. In the case of aligned scaffolds, the collecting surface consisted of a cylindrical aluminum mandrel rotating at 1250 rpm, such that the surface of the mandrel moved at 10 m s–1 [33,42]. For the random scaffold, a stationary aluminum plate was used [33]. Relative humidity (<30%) and temperature (approximately 25 °C) were controlled to minimize adverse effects on the scaffold quality and reproducibility. The resulting electrospun films were approximately 106 ± 5 and 550 ± 50 lm thick for aligned and random scaffolds, respectively. The differences in thickness between the two types of membranes were inherent to the individual methods of generating aligned vs. random scaffolds, since the random scaffolds were more fragile and required a thicker film in order to cover the entire plate surface with a mechanically stable membrane. The fibers ranged from 200 to 400 nm in diameter as measured by scanning electron microscopy (Fig. 1). 2.2. Biodegradation study Cholesterol esterase (CE) has been shown to be present in monocytes as they differentiate into macrophages and has been reported to degrade PUs [43,44]. CE (C3766, Sigma Aldrich, St. Louis, MO) was adjusted to 10 U ml–1, with 1 U defined as generating 1 nmol min–1 of p-nitrophenol from p-nitrophenylbutyrate (as measured colorimetrically by the DU 800 Beckman Coulter Spectrophotometer) [43]. This enzyme concentration was selected based on previous CE dose response studies, and was chosen to ensure that adequate degradation would occur throughout the 4 week study [45]. The half-life of CE in the presence of the polymer was approximately 12 h (Fig. 2); thus, throughout the 4 week
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
3849
Fig. 1. Scanning electron microscopy images of aligned (a) and random (b) electrospun polycarbonate urethane nanofiber scaffolds. (Solution: 16% w/v PU; injection rate: 0.5 ml h–1; potential difference: 18 kV.)
biodegradation study, appropriate volumes of a concentrated enzyme solution (100 U ml–1) were added daily to adjust the CE activity to a final concentration of 10 U ml–1 in a phosphate buffer solution (PBS, pH 7.0, 37 °C) containing the samples. The solutions containing degradation products were spun down at 3000 RCF to isolate the non-soluble particulate. Aliquots of the non-soluble degradation products in suspension were placed directly onto formvar-coated grids and imaged using transmission electron microscopy (TEM; FEI Tecnai 20, Hillsboro, Oregon) to study particulate size and morphology. Following the separation of soluble and non-soluble degradation products, cytotoxicity studies were performed on each group.
went tensile testing prior to and during the 4 week biodegradation period to assess any ongoing changes in mechanical properties. The measured dimensions at each time point were used to account for changes in cross-sectional area for the modulus and strength calculations. Each polymer sample was carefully immobilized on either end, using metallic clamps, and the tensile strength was evaluated using a mechanical testing device (InstronÒ model 8501) under a tensile strain of 10 mm min–1 to breaking point with a 50 N load cell. The initial modulus and the ultimate stress were used as measures of intrinsic mechanical properties throughout the degradation period. 2.4. Differential scanning calorimetry
2.3. Assessment of mass loss and mechanical properties Fiber membranes studied for physical property changes after biodegradation were initially prepared by cutting dried scaffolds into 6 30 mm pieces (with thicknesses of 106 ± 5 and 550 ± 50 lm for aligned and random scaffolds respectively) and weighed (Ohaus Explorer analytical balance) prior to placing in the enzymatic solution. The differences in thickness between the aligned and random scaffolds stem from the fact that, following the fabrication process, thinner random polymers are difficult to remove from the deposition surface due to their inferior mechanical properties, and thus thicker polymers were required to prevent polymer deformation prior to mechanical testing. The scaffolds were collected weekly, washed in dH2O, lyophilized and weighed to determine the average mass loss. The thickness and dimensions of the scaffolds were measured prior to and following biodegradation. Scaffolds under-
Differential scanning calorimetry (DSC) was performed to assess the crystallinity of scaffold polymers and changes in their microstructure due to the pre-wetting and drying process. The samples were pre-wet in phosphate buffer in the absence of CE for 1 week, washed in distilled water and lyophilized. The dried samples were then compared to as-made controls to understand the effects of the wetting and drying processes on the scaffolds. A small section of each sample, weighing approximately 2–3 mg, was cooled using liquid nitrogen and the thermograms were recorded between 100 and 200 °C at a heating rate of 15 °C min–1 (DSC was performed using the TA Instruments differential scanning calorimeter (model 2910) at the Brockhouse Institute for Material Research, McMaster University, Hamilton, Ontario, Canada). Data from two consecutive heating cycles (the first of which is reported here) was used to analyze possible crystalline state changes. 2.5. Cytotoxicity study
Fig. 2. Determination of CE half-life. It was determined that the half-life of CE in the presence of the aligned PU scaffolds was approximately 12 h. Thus CE was added daily to adjust the enzyme activity (n = 3).
To evaluate if the degradation products were cytotoxic, bovine annulus fibrous cells were chosen as a model for human lumbar AF [46–48]. Briefly, to obtain AF cells, intervertebral discs from bovine caudal spines (6–9 months old) were dissected out aseptically. Five discs from one spine were combined to obtain sufficient cells for an experiment. The outer annulus was separated from the disc and minced into small pieces. To isolate the cells, the tissues underwent serial digestion with 0.5% protease (Sigma, St. Louis, MO) for 1 h at 37 °C, followed by 0.25% collagenase A (Roche, Laval, Quebec, Canada) overnight at 37 °C. The cell suspension was filtered through a sterile mesh (pore size: 70 lm) and resuspended in Ham’s F12 medium supplemented with 5% fetal bovine serum. AF cells were placed in monolayer culture (seeding density of 1 105 cm–2) and grown for 48 h in Ham’s F12 medium containing 5% fetal bovine serum. The non-soluble degradation products were added to the tissue culture medium at concentrations ranging from
3850
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
0.001 to 0.1% w/v. Soluble degradation products (in PBS) were added to the medium at various concentrations ranging from 20% to 100% (relative to concentrations recovered from the biodegradation study). Cytotoxicity of the degradation products was assessed using the MTT viability assay and Live/Dead fluorescent imaging. In the MTT assay, mitochondrial dehydrogenase of viable cells cleave the tetrazolium ring of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to yield purple formazan crystals, which are insoluble in aqueous solutions. The resulting crystals were dissolved in methoxyethanol (acidified with HCl to pH 3) and the solution was measured spectrophotometrically at a wavelength of 570 nm (Thermo Scientific model Multiskan Ex photometer). Hydrogen peroxide (0.01%) was used as a negative control to induce cell death [49]. In the case of non-soluble degradation products, feeding medium was used as positive control for the MTT assay. For each soluble degradation product solution, corresponding concentrations of PBS and medium were used as a positive control and the latter was used to calculate the relative percent viability. In the Live/Dead assay (L-3224, Invitrogen, Burlington, ON), Calcein AM is enzymatically converted to its fluorescent variant, which in cells with intact membranes is concentrated within the cytoplasm. Ethidium homodimer-1 enters cells with damaged membranes and binds to nucleic acids and becomes fluorescent. ImageJ (v. 1.4g) was used to find regions of interest (ROIs) by isolating local maxima, which targeted the intensified Calcein AM observed around the nucleus of viable cells. These ROIs were then manually examined to count the viable cells. This latter procedure was similarly used to detect dead cells stained by Ethidium homodimer-1. On average, nearly 600 cells were counted for each data point. 2.6. Statistical analysis All test groups contained at least quadruplicate repeats and each experiment was repeated at least three times. The results from all experiments were combined and expressed as the mean ± standard error of the mean (SEM). The data were analyzed using one-way analysis of variance and all pair-wise comparisons between groups were conducted using the Scheffe post hoc test. Significance was assigned at p-values < 0.05. 3. Results and discussion When comparing sample mass for scaffolds before and after degradation, it was observed that degraded samples lost an average of 2.0 mg per sample for all groups by week 4. The rate of biodegradation was 0.56 ± 0.05 mg week–1 and was relatively linear over the 4 weeks (r2 = 0.86) (Fig. 3). An interesting observation was the fact that degradation appeared to be primarily surfacemediated since both the thicker random fiber scaffold and thinner aligned scaffold lost similar amounts of material despite the difference in their initial mass, as attributed to differences in thickness. These findings possibly reflect enzyme exposure limitations imposed by the scaffold. It was observed in Fig. 2 that the scaffold influenced the activity of CE, and therefore hydrolytic activity could be limited by the exposure of the enzyme to the outer surface of the samples, whether in a random or aligned configuration. Also, the diffusion of oligomeric degradation products from the scaffold and enzyme within the scaffold may be controlling the degradation mechanism. It should be noted that previous studies on the base polyurethane without ADO in the form of films, using CE, show that the degradation was primarily surface-mediated and it would therefore not be anticipated to see a sudden collapse of the bulk fiber scaffold early on for the non-loaded incubated materials [50,51]. The fact that no difference was observed for aligned
Fig. 3. Cumulative absolute mass loss during biodegradation. Scaffolds were incubated in 100 U ml–1 CE over 4 weeks. Data are reported as mean ± SEM (n = 6). () Absolute mass loss was found to increase significantly at every week (p < 0.05) for all groups, while no statistical differences were observed between the scaffold groups within each week.
nanofibers with ADO vs. no ADO is the first reported evidence that the additive did not appear to influence the biostability of the material. The aligned samples experienced a significant decrease (from 45.6 to 8.9 MPa, p < 0.05) in the initial modulus following wetting and drying processes, prior to degradation (Fig. 4a). In addition, a significant drop in tensile strength (from 13.8 to 6.6 MPa, p < 0.05) was also observed in aligned scaffolds (Fig. 4b). There was no observed difference between PU and PU + 0.5%ADO samples, suggesting that ADO did not appear to have any plasticizing or hardening effects on the mechanical behavior of the scaffolds. In order to further study the structural transformations introduced into the materials during the pre-wetting and drying process, samples were analyzed using DSC. The thermal transition temperatures were relatively similar for both PU and PU + 0.5%ADO scaffolds (Fig. 5a and b). The Tv hard-segment melting transition temperature remained relatively unchanged near 93 °C across all samples and was unaffected by the pre-wetting process. This temperature corresponds to the crystalline segments containing polar urethane groups, which are anticipated to be the most stable regions of the polymer due to the potential of extensive hydrogen bonding within this domain [40]. Likewise, the Tg values for polycarbonate near 36 °C remained relatively unchanged. Perhaps the most striking difference between the as-made controls and the pre-wet samples was the shift in the soft-segment melting transition temperature (Tiii and Tiv in Fig. 5) from approximately 51 °C for Tiii to 68 °C for Tiv. The shift of this polycarbonate (PCN) polymer phase towards higher temperatures suggests a loss of PCN phase purity (pure polycarbonate has a melt transition temperature of 45 °C [40]) and an increase in phase mixing of hard segment with the soft segment content. More evidence that the pre-wetting process disrupted the crystalline state of the soft segment PCN phase was observed near 33 °C (Tii, Fig. 5). The appearance of the onset of the soft-segment melt phase for the pre-wet sample is distant from the endotherm at 67 °C. Such a broad transition is characteristic of a heterogeneous phase. These transformations may be disrupting the organization of the crystalline matrix for the soft segments and are believed to be contributing in part to the softening of the oriented material post exposure to aqueous medium and the resulting deterioration in its bulk mechanical properties. In the case of random scaffolds, no significant drop in mechanical properties were observed with exposure to water, however the initial mechanical properties of this random structure are much lower than those of the oriented fiber. This suggests that the
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
3851
Fig. 4. (a) Elastic modulus and (b) tensile strength of the electrospun polyurethane nanofiber scaffolds following the pre-wetting (for 1 week in pH 7.0 PBS at 37 °C) and drying process, comparing non-ADO vs. ADO, as well as aligned vs. random scaffolds. Data are reported as mean ± SEM (n = 6). *,**Statistically significant differences at p < 0.05.
Fig. 5. DSC for as-made and pre-wet/dried samples for (a) aligned PU, (b) aligned PU + 0.5%ADO and (c) random PU + 0.5%ADO. Ti is the glass transition temperature for the polycarbonate soft segment, Tii indicates the onset of the soft-segment melt phase, Tiii and Tiv are soft-segment melting transition temperatures, and Tv indicates the hardsegment melting transition temperature.
entanglements (Fig. 1) of the overlapping fibers rather than the oriented structure of the polymer chains were dominant in terms of defining the mechanical properties of the scaffold. While DSC results for random scaffolds show a slightly lower shift in the soft-segment melt temperature (a shift of 12 °C as compared to 17 °C), which may partially account for some differences in the effects of buffer uptake, it is not apparent that these could rationalize such dramatic differences in performance. Aligned scaffolds are highly dependent on the strength of the fiber in the direction of orientation, whereas in the case of the random scaffolds the properties are more anisotropic and more dependent on the physical assembly of the fibers, yielding equal resistance in either direction of the scaffold. Furthermore, the latter are less prone to stress-induced changes in mechanical properties that may occur as a result of the electrospining process, whereby solidifying fibers could have polymer chains within the fiber undergo orientation after they are taken up by the rotating mandrel. Such stress-induced character in
polyurethanes has been recognized as affecting material morphology and chemistry and may contribute to the varying responses of the aligned vs. the random scaffolds (Fig. 4b) [30]. After initial exposure to the aqueous medium (pre-wet condition), the elastic modulus of the scaffolds (aligned: 8.0 ± 0.4 MPa; random: 2.2 ± 0.2 MPa) showed no significant changes when stress force was normalized to the cross-sectional area of the sample at the degradation time points throughout the remaining weeks of the biodegradation study, despite showing a degradation rate of 0.56 ± 0.05 mg week–1 (Figs. 3 and 6a). This is further evidence of a surface-mediated biodegradation mechanism, since the observed mass losses, although significant for both types of scaffolds (aligned vs. random) and particularly for the thinner aligned membranes (30% mass loss), are not accompanied by a loss in modulus. However, there was a significant drop in the ultimate stress in week 1 for both random and aligned samples, with the drop for aligned being more substantial. Throughout the final 3 weeks of
3852
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
Fig. 6. (a) Initial modulus and (b) tensile strength of the electrospun polyurethane nanofiber scaffolds over 4 weeks of biodegradation in CE (100 U ml–1) at 37 °C in pH 7.0 PBS. Data are reported as mean ± SEM (n = 6). Aligned scaffolds showed significantly higher modulus than random scaffolds at all time points. Ultimate stress of aligned polymers decreased in the first week of degradation, but remained stable thereafter. *Statistically significant difference at p < 0.05. Reported values for native AF tissue are included for Ref. [14] and represent samples having the long axis of the specimen aligned with the circumference of the AF.
the biodegradation process, the tensile strength of the aligned scaffolds continued to drop, although not significantly, with the values being approximately 2.3 MPa, thereby remaining comparable to those of different segments of native AF tissue [14] (Fig. 6b). ADO did not induce any significant changes to the degradation behavior (Fig 3) or mechanical properties of the scaffolds (Fig 6). The bulk mechanical properties of random scaffolds were unaffected throughout the final 3 weeks of degradation, suggesting that the physical entanglements of the random fibers were influencing material performance in this period more so than the mass loss. On the other hand, the initial drop in the ultimate stress for the aligned scaffolds after week 1, followed by its relatively slower deterioration during the final 3 weeks, suggests two different phases of degradation. In the first phase there is immediate hydrolysis of polymer chains on the outside surface, which affects the strength of the material. This initial chain cleavage results in some loss of soluble products; however, substantial chain cleavage within the polymer fibers will not occur until the residual non-solubilized cleaved chains at the surface are degraded to the point of being soluble enough to expose fresh surface for the enzymes to attack. Hence, the degradation of physical properties in this latter phase will be controlled somewhat by the rate of surface polymer chain solubilization and not necessarily by monomer cleavage. Following the initial breakdown of the polymer chains within week 1, residual oligomeric degradation products are suspected to contribute to the existing diffusion barrier to bulk degradation [45], supporting the stabilization of mechanical properties in the following 3 weeks. The reported literature values for the elastic modulus of layers of native human AF are 5 ± 4 MPa for posterior inner AF, 20 ± 12 MPa for posterior outer AF, 10 ± 6 MPa for anterior inner AF and 49 ± 32 MPa for anterior outer AF, with the test samples having the long axis of each specimen aligned with the circumference of the AF [14]. More recent studies have observed elastic moduli of 17.4 ± 14.3 and 5.6 ± 4.7 MPa for outer and inner AF, respectively [52]. The ultimate stress for layers of AF similarly varies radially from 0.9 ± 0.3 MPa for posterior inner AF to 1.1 ± 0.3 MPa for posterior outer AF, 0.9 ± 0.7 MPa for anterior inner AF and 3.3 ± 1.3 MPa for anterior outer AF [14]. In this study, CE was used to degrade the polymer samples. Although this provides an indication of what may occur in vivo as CE is known to be secreted by macrophages [43,44], the latter also secrete other proteases as well as oxidative agents, which would introduce oxidative degradation in addition to the hydrolytic degradation catalyzed by CE. Furthermore, the specific nature
of enzymes released by AF cells would most likely differ from those of macrophages. However, the use of this enzyme does demonstrate biosensitivity of the membranes with respect to changes in the physical character the scaffold. Similar results were observed for both ADO and non-ADO surfaces implying that the degradation was not affected from the presence of the anionic oligomers at the surface. Further in vivo studies are now required to fully assess the biodegradation behavior. In the current study, the degradation products were spun down and separated into soluble and non-soluble products. The non-soluble products, imaged by TEM, varied in shape and size (Fig. 7). The thickness of the fiber fragment was found to be approximately on average 400 nm, which is comparable to that of the electrospun nanofibers (Fig. 1). Different concentration gradients of the soluble and non-soluble degradation products were applied to bovine AF
Fig. 7. TEM of a non-soluble degradation product.
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
3853
Fig. 8. An MTT assay was used to evaluate potential cytotoxicity of various concentrations of (a) non-soluble and (b) soluble degradation products on bovine AF cells. The experiment was repeated four times (n = 8 per condition). Data are expressed as mean ± SEM. H2O2 (0.01% w/v) was used as a negative control for viability.
Fig. 9. Effect of PU degradation products on cell viability was determined by incubating AF cells for 24 h with various concentrations of (a) non-soluble and (b) soluble degradation products. A Live/Dead assay was used to assess cell viability. The number of dead cells were counted and expressed as percent of total number of cells. The experiment was repeated four times (n = 8 per condition) and data expressed as mean ± SEM. H2O2 was used as a negative control for viability.
Fig. 10. Representative images of a Live/Dead assay of AF cells treated with (a) untreated negative control (medium with carrier); (b) H2O2-treated positive control; (c) 0.1% w/v non-soluble degradation products; (d) 100 vol.% soluble degradation products.
3854
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
cells grown in monolayer culture. The MTT assay indicated no significant cytotoxic effects from either non-soluble or soluble degradation products, from the scaffolds made with PU or PU + ADO (Fig. 8a and b). Although the MTT assay is reported to measure cytotoxicity, it is more specifically an indicator of metabolic activity. Thus, a Live/Dead assay was performed concurrently with the MTT assay to confirm the findings. The Live/Dead assay showed that the material’s degradation products did not induce cytotoxicity (Figs. 9 and 10). These results are in agreement with our hypothesis that degradation of the PU would result in production of non-toxic by-products, based on previous studies of the virgin PU [30]. This work has concentrated on investigating possible effects of degradation products on cell viability and morphology. Future studies are required to explore additional indicators of cell function, such as the effect of degradation fragments on cell proliferation and extracellular matrix synthesis. 4. Conclusion The findings in this study have shown that electrospun aligned scaffolds produced superior mechanical properties compared to random scaffolds, confirming that this formulation is an appropriate scaffold for engineering AF tissue. An important consideration in the design of such scaffolds is the issue of structural changes due to water uptake into the material, particularly as it relates to the rate of degradation. It was established that exposure to aqueous medium disrupted the material’s intermolecular structure and resulted in a reduction of its mechanical properties. DSC data showed the appearance of a disrupted soft segment crystalline phase with changes in the degree of phase mixing of hard and soft segments as well as the PCN crystal state within the polymer. The degradation of the polymer by CE provided a useful model for biodegradation, yielding a controlled and consistent mass loss rate. Of particular importance is that the degradation of the materials, which resulted in mass losses as high as 30% over 4 weeks, did not result in a significant deterioration of mechanical properties, indicating a surface degradation process rather than bulk material breakdown. The degradation products did not cause any acute cytotoxicity in vitro. The ADO additive was found to have no significant effects on mechanical properties, degradation behavior or cytotoxic character. Additional mechanical studies to explore the changes in mechanical properties of the aligned polymer in the presence of AF tissue grown on the PU substrate will provide a further understanding of the potential use of this polymeric material for AF tissue engineering. In summary, the electrospun aligned PU nanofiber polymer appears to be a suitable biomaterial candidate for use in the formation of tissue-engineered AF. Acknowledgements The funding for this work was provided by a University of Toronto Fellowship Award, an Ontario Graduate Scholarship in Science and Technology (OGSST), in addition to a Natural Sciences and Engineering Research Council/Canadian Institute of Health Research (NSERC–CIHR) Collaborative Health Research Program (CHRP) Grant (312882) and a CIHR operating grant (MOP86723). We would like to thank Mr. Harry Bojarski and Ryding-Regency Meat Packer for providing the tissues. Appendix. Figures with essential color discrimination Certain figure in this article, particularly Figure 10, is difficult to interpret in black and white. The full color images can be found in the on-line version, at doi:10.1016/j.actbio.2010.05.003.
References [1] Rapoport J, Jacobs P, Bell NR, Klarenbach S. Refining the measurement of the economic burden of chronic diseases in Canada. Chronic Dis Can 2004;25(1): 13–21. [2] Urban JP, Roberts S. Degeneration of the intervertebral disc. Arthritis Res Ther 2003;5(3):120–30. [3] Lopez-Espina CG, Amirouche F, Havalad V. Multilevel cervical fusion and its effect on disc degeneration and osteophyte formation. Spine 2006;31(9): 972–8. [4] Javedan SP, Dickman CA. Cause of adjacent-segment disease after spinal fusion. Lancet 1999;354(9178):530–1. [5] Boden SD. Overview of the biology of lumbar spine fusion and principles for selecting a bone graft substitute. Spine 2002;27(16 Suppl. 1):S26–31. [6] Huang RC, Sandhu HS. The current status of lumbar total disc replacement. Orthop Clin North Am 2004;35(1):33–42. [7] Anderson PA, Rouleau JP. Intervertebral disc arthroplasty. Spine 2004;29(23): 2779–86. [8] Anderson JM. Inflammatory response to implants. ASAIO Trans 1988;34(2): 101–7. [9] Bogduk N. Clinical anatomy of the lumbar spine and sacrum. 4th ed. Edinburgh: Elsevier Churchill Livingstone; 2005. [10] Riches PE, Dhillon N, Lotz J, Woods AW, McNally DS. The internal mechanics of the intervertebral disc under cyclic loading. J Biomech 2002;35(9):1263–71. [11] Alkalay R. The material and mechanical properties of the healthy and degenerated intervertebral disc. In: Integrated biomaterials science. New York, USA: Springer; 2002. p. 403–24. [12] Baer AE, Laursen TA, Guilak F, Setton LA. The micromechanical environment of intervertebral disc cells determined by a finite deformation, anisotropic, and biphasic finite element model. J Biomech Eng 2003;125(1):1–11. [13] Wu HC, Yao RF. Mechanical behavior of the human annulus fibrosus. J Biomech 1976;9(1):1–7. [14] Ebara S, Iatridis JC, Setton LA, Foster RJ, Mow VC, Weidenbaum M. Tensile properties of nondegenerate human lumbar anulus fibrosus. Spine 1996;21(4):452–61. [15] Chang G, Kim HJ, Kaplan D, Vunjak-Novakovic G, Kandel RA. Porous silk scaffolds can be used for tissue engineering annulus fibrosus. Eur Spine J 2007;16(11):1848–57. [16] O’Halloran DM, Pandit AS. Tissue-engineering approach to regenerating the intervertebral disc. Tissue Eng 2007;13(8):1927–54. [17] Johnson WE, Wootton A, El HA, Eisenstein SM, Curtis AS, Roberts S. Topographical guidance of intervertebral disc cell growth in vitro: towards the development of tissue repair strategies for the anulus fibrosus. Eur Spine J 2006;15(15):S389–96. [18] Marchand F, Ahmed AM. Investigation of the laminate structure of lumbar disc anulus fibrosus. Spine 1990;15(5):402–10. [19] Sato M et al. An atelocollagen honeycomb-shaped scaffold with a membrane seal (ACHMS-scaffold) for the culture of annulus fibrosus cells from an intervertebral disc. J Biomed Mater Res A 2003;64(2):248–56. [20] Thonar E, An H, Masuda K. Compartmentalization of the matrix formed by nucleus pulposus and annulus fibrosus cells in alginate gel. Biochem Soc Trans 2002;30(Pt 6):874–8. [21] Wilda H, Gough JE. In vitro studies of annulus fibrosus disc cell attachment, differentiation and matrix production on PDLLA/45S5 Bioglass composite films. Biomaterials 2006;27(30):5220–9. [22] Rong Y, Sugumaran G, Silbert JE, Spector M. Proteoglycans synthesized by canine intervertebral disc cells grown in a type I collagen–glycosaminoglycan matrix. Tissue Eng 2002;8(6):1037–47. [23] Alini M, Li W, Markovic P, Aebi M, Spiro RC, Roughley PJ. The potential and limitations of a cell-seeded collagen/hyaluronan scaffold to engineer an intervertebral disc-like matrix. Spine 2003;28(5):446–54. [24] Mizuno H, Roy AK, Vacanti CA, Kojima K, Ueda M, Bonassar LJ. Tissueengineered composites of anulus fibrosus and nucleus pulposus for intervertebral disc replacement. Spine 2004;29(12):1290–7. [25] Chou AI, Reza AT, Nicoll SB. Distinct intervertebral disc cell populations adopt similar phenotypes in three-dimensional culture. Tissue Eng A 2008;14(12): 2079–87. [26] Ishihara H, Urban JP. Effects of low oxygen concentrations and metabolic inhibitors on proteoglycan and protein synthesis rates in the intervertebral disc. J Orthop Res 1999;17(6):829–35. [27] Li HY, Chang J. PH-compensation effect of bioactive inorganic fillers on the degradation of PLGA. Compos Sci Technol 2005;65(14):2226–32. [28] Helen W, Gough JE. Cell viability, proliferation and extracellular matrix production of human annulus fibrosus cells cultured within PDLLA/Bioglass composite foam scaffolds in vitro. Acta Biomater 2008;4(2):230–43. [29] Mathur AB, Collier TO, Kao WJ, Wiggins M, Schubert MA, Hiltner A, et al. In vivo biocompatibility and biostability of modified polyurethanes. J Biomed Mater Res 1997;36(2):246–57. [30] Santerre JP, Woodhouse K, Laroche G, Labow RS. Understanding the biodegradation of polyurethanes: from classical implants to tissue engineering materials. Biomaterials 2005;26(35):7457–70. [31] van Minnen B, van Leeuwen MB, Stegenga B, Zuidema J, Hissink CE, van Kooten TG, et al. Short-term in vitro and in vivo biocompatibility of a biodegradable polyurethane foam based on 1,4-butanediisocyanate. J Mater Sci Mater Med 2005;16(3):221–7.
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855 [32] Tang YW, Labow RS, Santerre JP. Enzyme-induced biodegradation of polycarbonate polyurethanes: dependence on hard-segment concentration. J Biomed Mater Res 2001;56(4):516–28. [33] Yang L, Kandel RA, Chang G, Santerre JP. Polar surface chemistry of nanofibrous polyurethane scaffold affects annulus fibrosus cell attachment and early matrix accumulation. J Biomed Mater Res A 2009;91(4): 1089–99. [34] Nesti LJ et al. Intervertebral disc tissue engineering using a novel hyaluronic acid-nanofibrous scaffold (HANFS) amalgam. Tissue Eng A 2008;14(9): 1527–37. [35] Nerurkar NL, Mauck RL, Elliott DM. ISSLS prize winner: integrating theoretical and experimental methods for functional tissue engineering of the annulus fibrosus. Spine (Phila, Pa, 1976) 2008;33(25):2691–701. [36] Nerurkar NL, Baker BM, Sen S, Wible EE, Elliott DM, Mauck RL. Nanofibrous biologic laminates replicate the form and function of the annulus fibrosus. Nat Mater 2009;8(12):986–92. [37] Nerurkar NL, Elliott DM, Mauck RL. Mechanics of oriented electrospun nanofibrous scaffolds for annulus fibrosus tissue engineering. J Orthop Res 2007;25(8):1018–28. [38] Thapa A, Miller DC, Webster TJ, Haberstroh KM. Nano-structured polymers enhance bladder smooth muscle cell function. Biomaterials 2003;24(17): 2915–26. [39] Guelcher SA. Biodegradable polyurethanes: synthesis and applications in regenerative medicine. Tissue Eng B Rev 2008;14(1):3–17. [40] Tang YW, Labow RS, Santerre JP. Enzyme-induced biodegradation of polycarbonate-polyurethanes: dependence on hard-segment chemistry. J Biomed Mater Res 2001;57(4):597–611. [41] Gruber HE, Hoelscher G, Ingram JA, Hanley Jr EN. Culture of human anulus fibrosus cells on polyamide nanofibers: extracellular matrix production. Spine (Phila, Pa, 1976) 2009;34(1):4–9.
3855
[42] Stankus JJ, Guan J, Wagner WR. Fabrication of biodegradable elastomeric scaffolds with sub-micron morphologies. J Biomed Mater Res A 2004;70(4): 603–14. [43] Labow RS, Meek E, Santerre JP. Synthesis of cholesterol esterase by monocytederived macrophages: a potential role in the biodegradation of poly(urethane)s. J Biomater Appl 1999;13(3):187–205. [44] Labow RS, Sa D, Matheson LA, Santerre JP. Polycarbonate-urethane hard segment type influences esterase substrate specificity for human-macrophage-mediated biodegradation. J Biomater Sci Polym Ed 2005;16(9):1167–77. [45] Tang YW, Labow RS, Santerre JP. Enzyme induced biodegradation of polycarbonate-polyurethanes: dose dependence effect of cholesterol esterase. Biomaterials 2003;24(12):2003–11. [46] Demers CN, Antoniou J, Mwale F. Value and limitations of using the bovine tail as a model for the human lumbar spine. Spine 2004;29(24):2793–9. [47] Cotterill PC, Kostuik JP, D’Angelo G, Fernie GR, Maki BE. An anatomical comparison of the human and bovine thoracolumbar spine. J Orthop Res 1986;4(3):298–303. [48] Wilke HJ, Krischak S, Claes L. Biomechanical comparison of calf and human spines. J Orthop Res 1996;14(3):500–3. [49] Hoffmann ME, Meneghini R. Action of hydrogen peroxide on human fibroblast in culture. Photochem Photobiol 1979;30(1):151–5. [50] Christenson EM, Patel S, Anderson JM, Hiltner A. Enzymatic degradation of poly(ether urethane) and poly(carbonate urethane) by cholesterol esterase. Biomaterials 2006;27(21):3920–6. [51] Skarja GA, Woodhouse KA. In vitro degradation and erosion of degradable, segmented polyurethanes containing an amino acid-based chain extender. J Biomater Sci Polym Ed 2001;12(8):851–73. [52] Elliott DM, Setton LA. Anisotropic and inhomogeneous tensile behavior of the human anulus fibrosus: experimental measurement and material model predictions. J Biomech Eng 2001;123(3):256–63.
Contents lists available at ScienceDirect
Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat
Characterization of a biodegradable electrospun polyurethane nanofiber scaffold: Mechanical properties and cytotoxicity Masoud Yeganegi a,b, Rita A. Kandel a,b,c, J. Paul Santerre b,d,* a
CIHR – Bioengineering of Skeletal Tissues Team, Mount Sinai Hospital, Toronto, Canada M5G 1X5 Institute of Biomaterials and Biomedical Engineering, Department of Materials Science and Engineering, University of Toronto, Toronto, Canada M5S 3G9 c Department of Pathobiology and Laboratory Medicine, Mt. Sinai Hospital, University of Toronto, Toronto, Canada M5G 1X5 d Faculty of Dentistry, University of Toronto, Toronto, Canada M5G 1G6 b
a r t i c l e
i n f o
Article history: Received 1 March 2010 Received in revised form 18 April 2010 Accepted 6 May 2010 Available online 11 May 2010 Keywords: Annulus fibrosus Biodegradation Polyurethane Tissue engineering Nanofiber
a b s t r a c t The current study analyzes the biodegradation of a polycarbonate polyurethane scaffold intended for the growth of a tissue-engineered annulus fibrosus (AF) disc component. Electrospun scaffolds with random and aligned nanofiber configurations were fabricated using a biodegradable polycarbonate urethane with and without an anionic surface modifier (anionic dihydroxyl oligomer), and the mechanical behavior of the scaffolds was examined during a 4 week biodegradation study. Both the tensile strength and initial modulus of aligned scaffolds (r = 14 ± 1 MPa, E = 46 ± 3 MPa) were found to be higher than those of random fiber scaffolds (r = 1.9 ± 0.4 MPa, E = 2.1 ± 0.2 MPa) prior to degradation. Following initial wetting of the scaffold, the initial modulus of the aligned samples showed a significant decrease (dry: 46 ± 3 MPa; pre-wetted: 9 ± 1 MPa, p < 0.001). The modulus remained relatively constant during the remainder of the 4 week incubation period (aligned at 4 weeks: 8.0 ± 0.3 MPa). The tensile strength for aligned fiber scaffolds was affected in the same manner. Similar changes were not observed for the initial modulus of the random scaffold configuration. Biodegradation of the scaffold in the presence of cholesterol esterase (a monocyte derived enzyme) yielded a 0.5 mg week–1 weight loss. The soluble and non-soluble degradation products were found to be non-toxic to bovine AF cells grown in vitro. The consistent rate of material degradation along with stable mechanical properties comparable to those of native AF tissue and the absence of cytotoxic effects make this polymer a suitable biomaterial candidate for further investigation into its use for tissue-engineering annulus fibrosus. Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction Back pain is ranked as the most prevalent chronic disease for people under the age of 60 [1]. Degenerative disc disease may contribute to the pathogenesis of lower back pain and involves the progressive degeneration of the intervertebral disc (IVD). The intact disc is necessary to support compressive and bending stresses while providing flexibility to the spine [2]. Currently, existing treatments include discectomy, the use of a prosthetic substitute or the fusion of adjacent vertebrae, none of which are optimal. Spinal fusion may be effective in some cases, but a number of patients can develop further degeneration, due to reduced flexibility and increased stress in adjacent segments [3–6]. Complications can also occur in patients undergoing disc replacement with synthetic substrates. Dislocations, slippage, wear and mechanical failure,
* Corresponding author at: Institute of Biomaterials and Biomedical Engineering, Department of Materials Science and Engineering, University of Toronto, Toronto, Canada M5S 3G9. Tel.: +1 416 979 4903x4341. E-mail address: [email protected] (J.P. Santerre).
although rare, have been reported [7]. The formation of wear debris can induce an inflammatory response mediated by various cytokines, leading to pain, osteolysis, fibrous tissue formation, prosthetic loosening and pain [7,8]. The human vertebral column is made up of 26 vertebral bodies that provide support to the body and protect the spinal cord. The intervertebral discs lying between the vertebrae provide flexibility and help dissipate mechanical loads and shocks that would otherwise damage the vertebral column [9]. The intervertebral discs are composed of the annulus fibrosus (AF), a fibrocartilaginous tissue that surrounds the gelatinous inner nucleus pulposus. The hyaline cartilage endplates found at each end represent the anatomical limits of the disc and contribute to the interface of the disc and bone. The mechanical properties of intervertebral discs are complex and the literature on this topic shows significant variability [10–13], perhaps in part due to the diversity of experimental conditions and mechanical models used to measure material properties. Studies indicate that the annulus fibrosus exhibits both matrix viscoelastic (flow-independent) and biphasic viscoelastic
1742-7061/$ - see front matter Ó 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2010.05.003
3848
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
(flow-dependent) behavior [13]. The elastic modulus of lamellae for the annulus fibrosus has been found to vary with radial position in the disc, with values ranging from 5 ± 4 MPa for posterior inner AF to 10 ± 6 MPa for anterior inner AF and 49 ± 32 MPa for anterior outer AF [14]. The ultimate stress of lamellae for the annulus fibrosus also varies radially from 0.9 ± 0.3 to 3.3 ± 1.3 MPa for anterior outer AF [14]. Thus the forces experienced by the AF tissue can be quite high and therefore an intact AF is critical to proper disc function. One alternative strategy for the treatment of intervertebral disc degeneration is the replacement of the diseased tissue by a tissueengineered substitute [15–17]. Tissue engineering of the annulus fibrosus is particularly challenging due to the complex structure of the tissue. The AF, which is responsible for withstanding circumferential tensile forces and to a lesser extent compressive forces [14], consists of concentric lamella, made up of collagen fibers and oriented at approximately 60° to the vertical. The alignment of these fibers alternates between successive lamellae and is of great importance to the functional nature of the annulus fibrosus [18]. Efforts to produce AF tissue in vitro have involved various polymeric scaffolds including poly-DL-lactic acid (PDLLA)/45S5 BioglassÒ films, polyglycolic acid (PGA), collagen/hyaluronan, collagen/glycosaminoglycans (GAGs), atelocollagen and alginate scaffolds [19–24]. However, there are various issues that have yet to be overcome. Nucleus pulposus (NP) and AF cells cultured in alginate scaffolds have been found to adopt similar phenotypes rather than their respective native phenotypes [25]. Both cell types displayed an NP-like phenotype with rounded chondrocyte-like morphology, higher than normal levels of GAG and type II collagen vs. type I collagen. Similar problems have been encountered in collagen-hyaloronan scaffolds [23]. Polyglycolic-based polymers create acidic by-products throughout their biodegradation, which can significantly alter cell behavior and tissue production, and possibly cause cell death [26,27]. While the use of PDLLA/45S5 Bioglass has produced AF tissue with predominantly type I collagen, typical of native AF tissue, there exists an inherent challenge in reproducing the aligned structural arrangement of annulus tissue [28]. A potential alternative to the above materials for tissue engineering of an IVD consist of polycarbonate urethane polymers, owing to their relative biocompatibility [29,30], the possibility of tailoring their biodegradability and the inherent reproducible nature of the materials [29–32]. Recent work has shown that a degradable polycarbonate urethane (PU) previously reported in Ref. [32] could be processed, with the addition of anionic dihydroxyl oligomers (ADO), which enhance cell attachment [33], using electrospinning applications to fabricate aligned fibers. These structures mimic the aligned nature of the AF, and thus provide an appropriate scaffold to support the growth of such a tissue. Specifically, such topographical properties have been shown to influence cell alignment [17]. While previous studies have used random nanofiber polymeric scaffolds for IVD tissue generation [34], recent efforts have been aimed at utilizing the capacity of aligned electrospun scaffolds to generate aligned AF tissue. Nerurkar et al. [35–37] have assessed the mechanics of aligned poly(e-caprolactone) scaffolds and their ability to induce alignment in AF cells and tissue. The high surface to volume ratio of these scaffolds is also expected to favor cell attachment and retention of cell phenotype [33,38]. The purpose of this study was to characterize a biodegradable electrospun PU nanofiber biomaterial and determine its suitability for use in repairing the AF. Given the mechanical function of AF, it was important to assess the mechanical properties of the proposed polycarbonate-based urethane scaffolds containing the ADO additives which had not previously been reported on. Along with an emphasis on the mechanical consequences of the materials and the scaffold’s inherent biodegradable nature, this study was also
designed to determine the potential cytotoxic effects of the PU and the ADO’s biodegradation products on AF cells, as the latter materials represent a new class of scaffold polymers that are quite distinct in terms of balanced pH degradation by-products, from the classical degradable acidic by-product-producing polyesters, such as PDLLA, PGA and polycaprolactone-based polymers. 2. Materials and methods 2.1. Materials Polyurethanes have been used in biomedical devices since the 1960s. Efforts in the 1980s and 1990s had previously been directed at producing biostable polyurethanes; however, in the past decade, the focus has shifted to developing bioactive/biocompatible and biodegradable polyurethanes for the purpose of tissue engineering or regeneration [30,39]. In this study, PU was synthesized, as has been described [40], with hexane diisocyanate:polycarbonate diol:butane diol in a molar ratio of 3:2:1. In addition, an ADO additive was synthesized through the reaction of lysine diisocyanate, polytetramethylene oxide and hydroxyethyl methacrylate (HEMA), followed by subsequent hydrolysis of ester groups on the lysine and HEMA moieties [33]. ADO has been previously shown to increase AF cellular adhesion and collagen production, mediated by protein adsorption to the surface [33], taking advantage of both the polar and ionic function of the material. In more recent work, other groups have also recognized the importance of using charged materials to generate nanofiber scaffold systems, where increased tissue production has been observed [41]. Electrospinning was employed to fabricate either random or aligned nanofiber scaffolds using the base polymer (PU) with or without ADO (0.5 wt.%). The polymer was dissolved in 1,1,1,3,3,3hexafluora-2-propanol at a concentration of 16 wt.%. The scaffold was then electrospun by injecting the polymer solution from a metallic syringe at a rate of 0.5 ml h–1 onto the collecting surface. An 18 kV difference was applied across the syringe and the collecting surface. In the case of aligned scaffolds, the collecting surface consisted of a cylindrical aluminum mandrel rotating at 1250 rpm, such that the surface of the mandrel moved at 10 m s–1 [33,42]. For the random scaffold, a stationary aluminum plate was used [33]. Relative humidity (<30%) and temperature (approximately 25 °C) were controlled to minimize adverse effects on the scaffold quality and reproducibility. The resulting electrospun films were approximately 106 ± 5 and 550 ± 50 lm thick for aligned and random scaffolds, respectively. The differences in thickness between the two types of membranes were inherent to the individual methods of generating aligned vs. random scaffolds, since the random scaffolds were more fragile and required a thicker film in order to cover the entire plate surface with a mechanically stable membrane. The fibers ranged from 200 to 400 nm in diameter as measured by scanning electron microscopy (Fig. 1). 2.2. Biodegradation study Cholesterol esterase (CE) has been shown to be present in monocytes as they differentiate into macrophages and has been reported to degrade PUs [43,44]. CE (C3766, Sigma Aldrich, St. Louis, MO) was adjusted to 10 U ml–1, with 1 U defined as generating 1 nmol min–1 of p-nitrophenol from p-nitrophenylbutyrate (as measured colorimetrically by the DU 800 Beckman Coulter Spectrophotometer) [43]. This enzyme concentration was selected based on previous CE dose response studies, and was chosen to ensure that adequate degradation would occur throughout the 4 week study [45]. The half-life of CE in the presence of the polymer was approximately 12 h (Fig. 2); thus, throughout the 4 week
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
3849
Fig. 1. Scanning electron microscopy images of aligned (a) and random (b) electrospun polycarbonate urethane nanofiber scaffolds. (Solution: 16% w/v PU; injection rate: 0.5 ml h–1; potential difference: 18 kV.)
biodegradation study, appropriate volumes of a concentrated enzyme solution (100 U ml–1) were added daily to adjust the CE activity to a final concentration of 10 U ml–1 in a phosphate buffer solution (PBS, pH 7.0, 37 °C) containing the samples. The solutions containing degradation products were spun down at 3000 RCF to isolate the non-soluble particulate. Aliquots of the non-soluble degradation products in suspension were placed directly onto formvar-coated grids and imaged using transmission electron microscopy (TEM; FEI Tecnai 20, Hillsboro, Oregon) to study particulate size and morphology. Following the separation of soluble and non-soluble degradation products, cytotoxicity studies were performed on each group.
went tensile testing prior to and during the 4 week biodegradation period to assess any ongoing changes in mechanical properties. The measured dimensions at each time point were used to account for changes in cross-sectional area for the modulus and strength calculations. Each polymer sample was carefully immobilized on either end, using metallic clamps, and the tensile strength was evaluated using a mechanical testing device (InstronÒ model 8501) under a tensile strain of 10 mm min–1 to breaking point with a 50 N load cell. The initial modulus and the ultimate stress were used as measures of intrinsic mechanical properties throughout the degradation period. 2.4. Differential scanning calorimetry
2.3. Assessment of mass loss and mechanical properties Fiber membranes studied for physical property changes after biodegradation were initially prepared by cutting dried scaffolds into 6 30 mm pieces (with thicknesses of 106 ± 5 and 550 ± 50 lm for aligned and random scaffolds respectively) and weighed (Ohaus Explorer analytical balance) prior to placing in the enzymatic solution. The differences in thickness between the aligned and random scaffolds stem from the fact that, following the fabrication process, thinner random polymers are difficult to remove from the deposition surface due to their inferior mechanical properties, and thus thicker polymers were required to prevent polymer deformation prior to mechanical testing. The scaffolds were collected weekly, washed in dH2O, lyophilized and weighed to determine the average mass loss. The thickness and dimensions of the scaffolds were measured prior to and following biodegradation. Scaffolds under-
Differential scanning calorimetry (DSC) was performed to assess the crystallinity of scaffold polymers and changes in their microstructure due to the pre-wetting and drying process. The samples were pre-wet in phosphate buffer in the absence of CE for 1 week, washed in distilled water and lyophilized. The dried samples were then compared to as-made controls to understand the effects of the wetting and drying processes on the scaffolds. A small section of each sample, weighing approximately 2–3 mg, was cooled using liquid nitrogen and the thermograms were recorded between 100 and 200 °C at a heating rate of 15 °C min–1 (DSC was performed using the TA Instruments differential scanning calorimeter (model 2910) at the Brockhouse Institute for Material Research, McMaster University, Hamilton, Ontario, Canada). Data from two consecutive heating cycles (the first of which is reported here) was used to analyze possible crystalline state changes. 2.5. Cytotoxicity study
Fig. 2. Determination of CE half-life. It was determined that the half-life of CE in the presence of the aligned PU scaffolds was approximately 12 h. Thus CE was added daily to adjust the enzyme activity (n = 3).
To evaluate if the degradation products were cytotoxic, bovine annulus fibrous cells were chosen as a model for human lumbar AF [46–48]. Briefly, to obtain AF cells, intervertebral discs from bovine caudal spines (6–9 months old) were dissected out aseptically. Five discs from one spine were combined to obtain sufficient cells for an experiment. The outer annulus was separated from the disc and minced into small pieces. To isolate the cells, the tissues underwent serial digestion with 0.5% protease (Sigma, St. Louis, MO) for 1 h at 37 °C, followed by 0.25% collagenase A (Roche, Laval, Quebec, Canada) overnight at 37 °C. The cell suspension was filtered through a sterile mesh (pore size: 70 lm) and resuspended in Ham’s F12 medium supplemented with 5% fetal bovine serum. AF cells were placed in monolayer culture (seeding density of 1 105 cm–2) and grown for 48 h in Ham’s F12 medium containing 5% fetal bovine serum. The non-soluble degradation products were added to the tissue culture medium at concentrations ranging from
3850
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
0.001 to 0.1% w/v. Soluble degradation products (in PBS) were added to the medium at various concentrations ranging from 20% to 100% (relative to concentrations recovered from the biodegradation study). Cytotoxicity of the degradation products was assessed using the MTT viability assay and Live/Dead fluorescent imaging. In the MTT assay, mitochondrial dehydrogenase of viable cells cleave the tetrazolium ring of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to yield purple formazan crystals, which are insoluble in aqueous solutions. The resulting crystals were dissolved in methoxyethanol (acidified with HCl to pH 3) and the solution was measured spectrophotometrically at a wavelength of 570 nm (Thermo Scientific model Multiskan Ex photometer). Hydrogen peroxide (0.01%) was used as a negative control to induce cell death [49]. In the case of non-soluble degradation products, feeding medium was used as positive control for the MTT assay. For each soluble degradation product solution, corresponding concentrations of PBS and medium were used as a positive control and the latter was used to calculate the relative percent viability. In the Live/Dead assay (L-3224, Invitrogen, Burlington, ON), Calcein AM is enzymatically converted to its fluorescent variant, which in cells with intact membranes is concentrated within the cytoplasm. Ethidium homodimer-1 enters cells with damaged membranes and binds to nucleic acids and becomes fluorescent. ImageJ (v. 1.4g) was used to find regions of interest (ROIs) by isolating local maxima, which targeted the intensified Calcein AM observed around the nucleus of viable cells. These ROIs were then manually examined to count the viable cells. This latter procedure was similarly used to detect dead cells stained by Ethidium homodimer-1. On average, nearly 600 cells were counted for each data point. 2.6. Statistical analysis All test groups contained at least quadruplicate repeats and each experiment was repeated at least three times. The results from all experiments were combined and expressed as the mean ± standard error of the mean (SEM). The data were analyzed using one-way analysis of variance and all pair-wise comparisons between groups were conducted using the Scheffe post hoc test. Significance was assigned at p-values < 0.05. 3. Results and discussion When comparing sample mass for scaffolds before and after degradation, it was observed that degraded samples lost an average of 2.0 mg per sample for all groups by week 4. The rate of biodegradation was 0.56 ± 0.05 mg week–1 and was relatively linear over the 4 weeks (r2 = 0.86) (Fig. 3). An interesting observation was the fact that degradation appeared to be primarily surfacemediated since both the thicker random fiber scaffold and thinner aligned scaffold lost similar amounts of material despite the difference in their initial mass, as attributed to differences in thickness. These findings possibly reflect enzyme exposure limitations imposed by the scaffold. It was observed in Fig. 2 that the scaffold influenced the activity of CE, and therefore hydrolytic activity could be limited by the exposure of the enzyme to the outer surface of the samples, whether in a random or aligned configuration. Also, the diffusion of oligomeric degradation products from the scaffold and enzyme within the scaffold may be controlling the degradation mechanism. It should be noted that previous studies on the base polyurethane without ADO in the form of films, using CE, show that the degradation was primarily surface-mediated and it would therefore not be anticipated to see a sudden collapse of the bulk fiber scaffold early on for the non-loaded incubated materials [50,51]. The fact that no difference was observed for aligned
Fig. 3. Cumulative absolute mass loss during biodegradation. Scaffolds were incubated in 100 U ml–1 CE over 4 weeks. Data are reported as mean ± SEM (n = 6). () Absolute mass loss was found to increase significantly at every week (p < 0.05) for all groups, while no statistical differences were observed between the scaffold groups within each week.
nanofibers with ADO vs. no ADO is the first reported evidence that the additive did not appear to influence the biostability of the material. The aligned samples experienced a significant decrease (from 45.6 to 8.9 MPa, p < 0.05) in the initial modulus following wetting and drying processes, prior to degradation (Fig. 4a). In addition, a significant drop in tensile strength (from 13.8 to 6.6 MPa, p < 0.05) was also observed in aligned scaffolds (Fig. 4b). There was no observed difference between PU and PU + 0.5%ADO samples, suggesting that ADO did not appear to have any plasticizing or hardening effects on the mechanical behavior of the scaffolds. In order to further study the structural transformations introduced into the materials during the pre-wetting and drying process, samples were analyzed using DSC. The thermal transition temperatures were relatively similar for both PU and PU + 0.5%ADO scaffolds (Fig. 5a and b). The Tv hard-segment melting transition temperature remained relatively unchanged near 93 °C across all samples and was unaffected by the pre-wetting process. This temperature corresponds to the crystalline segments containing polar urethane groups, which are anticipated to be the most stable regions of the polymer due to the potential of extensive hydrogen bonding within this domain [40]. Likewise, the Tg values for polycarbonate near 36 °C remained relatively unchanged. Perhaps the most striking difference between the as-made controls and the pre-wet samples was the shift in the soft-segment melting transition temperature (Tiii and Tiv in Fig. 5) from approximately 51 °C for Tiii to 68 °C for Tiv. The shift of this polycarbonate (PCN) polymer phase towards higher temperatures suggests a loss of PCN phase purity (pure polycarbonate has a melt transition temperature of 45 °C [40]) and an increase in phase mixing of hard segment with the soft segment content. More evidence that the pre-wetting process disrupted the crystalline state of the soft segment PCN phase was observed near 33 °C (Tii, Fig. 5). The appearance of the onset of the soft-segment melt phase for the pre-wet sample is distant from the endotherm at 67 °C. Such a broad transition is characteristic of a heterogeneous phase. These transformations may be disrupting the organization of the crystalline matrix for the soft segments and are believed to be contributing in part to the softening of the oriented material post exposure to aqueous medium and the resulting deterioration in its bulk mechanical properties. In the case of random scaffolds, no significant drop in mechanical properties were observed with exposure to water, however the initial mechanical properties of this random structure are much lower than those of the oriented fiber. This suggests that the
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
3851
Fig. 4. (a) Elastic modulus and (b) tensile strength of the electrospun polyurethane nanofiber scaffolds following the pre-wetting (for 1 week in pH 7.0 PBS at 37 °C) and drying process, comparing non-ADO vs. ADO, as well as aligned vs. random scaffolds. Data are reported as mean ± SEM (n = 6). *,**Statistically significant differences at p < 0.05.
Fig. 5. DSC for as-made and pre-wet/dried samples for (a) aligned PU, (b) aligned PU + 0.5%ADO and (c) random PU + 0.5%ADO. Ti is the glass transition temperature for the polycarbonate soft segment, Tii indicates the onset of the soft-segment melt phase, Tiii and Tiv are soft-segment melting transition temperatures, and Tv indicates the hardsegment melting transition temperature.
entanglements (Fig. 1) of the overlapping fibers rather than the oriented structure of the polymer chains were dominant in terms of defining the mechanical properties of the scaffold. While DSC results for random scaffolds show a slightly lower shift in the soft-segment melt temperature (a shift of 12 °C as compared to 17 °C), which may partially account for some differences in the effects of buffer uptake, it is not apparent that these could rationalize such dramatic differences in performance. Aligned scaffolds are highly dependent on the strength of the fiber in the direction of orientation, whereas in the case of the random scaffolds the properties are more anisotropic and more dependent on the physical assembly of the fibers, yielding equal resistance in either direction of the scaffold. Furthermore, the latter are less prone to stress-induced changes in mechanical properties that may occur as a result of the electrospining process, whereby solidifying fibers could have polymer chains within the fiber undergo orientation after they are taken up by the rotating mandrel. Such stress-induced character in
polyurethanes has been recognized as affecting material morphology and chemistry and may contribute to the varying responses of the aligned vs. the random scaffolds (Fig. 4b) [30]. After initial exposure to the aqueous medium (pre-wet condition), the elastic modulus of the scaffolds (aligned: 8.0 ± 0.4 MPa; random: 2.2 ± 0.2 MPa) showed no significant changes when stress force was normalized to the cross-sectional area of the sample at the degradation time points throughout the remaining weeks of the biodegradation study, despite showing a degradation rate of 0.56 ± 0.05 mg week–1 (Figs. 3 and 6a). This is further evidence of a surface-mediated biodegradation mechanism, since the observed mass losses, although significant for both types of scaffolds (aligned vs. random) and particularly for the thinner aligned membranes (30% mass loss), are not accompanied by a loss in modulus. However, there was a significant drop in the ultimate stress in week 1 for both random and aligned samples, with the drop for aligned being more substantial. Throughout the final 3 weeks of
3852
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
Fig. 6. (a) Initial modulus and (b) tensile strength of the electrospun polyurethane nanofiber scaffolds over 4 weeks of biodegradation in CE (100 U ml–1) at 37 °C in pH 7.0 PBS. Data are reported as mean ± SEM (n = 6). Aligned scaffolds showed significantly higher modulus than random scaffolds at all time points. Ultimate stress of aligned polymers decreased in the first week of degradation, but remained stable thereafter. *Statistically significant difference at p < 0.05. Reported values for native AF tissue are included for Ref. [14] and represent samples having the long axis of the specimen aligned with the circumference of the AF.
the biodegradation process, the tensile strength of the aligned scaffolds continued to drop, although not significantly, with the values being approximately 2.3 MPa, thereby remaining comparable to those of different segments of native AF tissue [14] (Fig. 6b). ADO did not induce any significant changes to the degradation behavior (Fig 3) or mechanical properties of the scaffolds (Fig 6). The bulk mechanical properties of random scaffolds were unaffected throughout the final 3 weeks of degradation, suggesting that the physical entanglements of the random fibers were influencing material performance in this period more so than the mass loss. On the other hand, the initial drop in the ultimate stress for the aligned scaffolds after week 1, followed by its relatively slower deterioration during the final 3 weeks, suggests two different phases of degradation. In the first phase there is immediate hydrolysis of polymer chains on the outside surface, which affects the strength of the material. This initial chain cleavage results in some loss of soluble products; however, substantial chain cleavage within the polymer fibers will not occur until the residual non-solubilized cleaved chains at the surface are degraded to the point of being soluble enough to expose fresh surface for the enzymes to attack. Hence, the degradation of physical properties in this latter phase will be controlled somewhat by the rate of surface polymer chain solubilization and not necessarily by monomer cleavage. Following the initial breakdown of the polymer chains within week 1, residual oligomeric degradation products are suspected to contribute to the existing diffusion barrier to bulk degradation [45], supporting the stabilization of mechanical properties in the following 3 weeks. The reported literature values for the elastic modulus of layers of native human AF are 5 ± 4 MPa for posterior inner AF, 20 ± 12 MPa for posterior outer AF, 10 ± 6 MPa for anterior inner AF and 49 ± 32 MPa for anterior outer AF, with the test samples having the long axis of each specimen aligned with the circumference of the AF [14]. More recent studies have observed elastic moduli of 17.4 ± 14.3 and 5.6 ± 4.7 MPa for outer and inner AF, respectively [52]. The ultimate stress for layers of AF similarly varies radially from 0.9 ± 0.3 MPa for posterior inner AF to 1.1 ± 0.3 MPa for posterior outer AF, 0.9 ± 0.7 MPa for anterior inner AF and 3.3 ± 1.3 MPa for anterior outer AF [14]. In this study, CE was used to degrade the polymer samples. Although this provides an indication of what may occur in vivo as CE is known to be secreted by macrophages [43,44], the latter also secrete other proteases as well as oxidative agents, which would introduce oxidative degradation in addition to the hydrolytic degradation catalyzed by CE. Furthermore, the specific nature
of enzymes released by AF cells would most likely differ from those of macrophages. However, the use of this enzyme does demonstrate biosensitivity of the membranes with respect to changes in the physical character the scaffold. Similar results were observed for both ADO and non-ADO surfaces implying that the degradation was not affected from the presence of the anionic oligomers at the surface. Further in vivo studies are now required to fully assess the biodegradation behavior. In the current study, the degradation products were spun down and separated into soluble and non-soluble products. The non-soluble products, imaged by TEM, varied in shape and size (Fig. 7). The thickness of the fiber fragment was found to be approximately on average 400 nm, which is comparable to that of the electrospun nanofibers (Fig. 1). Different concentration gradients of the soluble and non-soluble degradation products were applied to bovine AF
Fig. 7. TEM of a non-soluble degradation product.
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
3853
Fig. 8. An MTT assay was used to evaluate potential cytotoxicity of various concentrations of (a) non-soluble and (b) soluble degradation products on bovine AF cells. The experiment was repeated four times (n = 8 per condition). Data are expressed as mean ± SEM. H2O2 (0.01% w/v) was used as a negative control for viability.
Fig. 9. Effect of PU degradation products on cell viability was determined by incubating AF cells for 24 h with various concentrations of (a) non-soluble and (b) soluble degradation products. A Live/Dead assay was used to assess cell viability. The number of dead cells were counted and expressed as percent of total number of cells. The experiment was repeated four times (n = 8 per condition) and data expressed as mean ± SEM. H2O2 was used as a negative control for viability.
Fig. 10. Representative images of a Live/Dead assay of AF cells treated with (a) untreated negative control (medium with carrier); (b) H2O2-treated positive control; (c) 0.1% w/v non-soluble degradation products; (d) 100 vol.% soluble degradation products.
3854
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855
cells grown in monolayer culture. The MTT assay indicated no significant cytotoxic effects from either non-soluble or soluble degradation products, from the scaffolds made with PU or PU + ADO (Fig. 8a and b). Although the MTT assay is reported to measure cytotoxicity, it is more specifically an indicator of metabolic activity. Thus, a Live/Dead assay was performed concurrently with the MTT assay to confirm the findings. The Live/Dead assay showed that the material’s degradation products did not induce cytotoxicity (Figs. 9 and 10). These results are in agreement with our hypothesis that degradation of the PU would result in production of non-toxic by-products, based on previous studies of the virgin PU [30]. This work has concentrated on investigating possible effects of degradation products on cell viability and morphology. Future studies are required to explore additional indicators of cell function, such as the effect of degradation fragments on cell proliferation and extracellular matrix synthesis. 4. Conclusion The findings in this study have shown that electrospun aligned scaffolds produced superior mechanical properties compared to random scaffolds, confirming that this formulation is an appropriate scaffold for engineering AF tissue. An important consideration in the design of such scaffolds is the issue of structural changes due to water uptake into the material, particularly as it relates to the rate of degradation. It was established that exposure to aqueous medium disrupted the material’s intermolecular structure and resulted in a reduction of its mechanical properties. DSC data showed the appearance of a disrupted soft segment crystalline phase with changes in the degree of phase mixing of hard and soft segments as well as the PCN crystal state within the polymer. The degradation of the polymer by CE provided a useful model for biodegradation, yielding a controlled and consistent mass loss rate. Of particular importance is that the degradation of the materials, which resulted in mass losses as high as 30% over 4 weeks, did not result in a significant deterioration of mechanical properties, indicating a surface degradation process rather than bulk material breakdown. The degradation products did not cause any acute cytotoxicity in vitro. The ADO additive was found to have no significant effects on mechanical properties, degradation behavior or cytotoxic character. Additional mechanical studies to explore the changes in mechanical properties of the aligned polymer in the presence of AF tissue grown on the PU substrate will provide a further understanding of the potential use of this polymeric material for AF tissue engineering. In summary, the electrospun aligned PU nanofiber polymer appears to be a suitable biomaterial candidate for use in the formation of tissue-engineered AF. Acknowledgements The funding for this work was provided by a University of Toronto Fellowship Award, an Ontario Graduate Scholarship in Science and Technology (OGSST), in addition to a Natural Sciences and Engineering Research Council/Canadian Institute of Health Research (NSERC–CIHR) Collaborative Health Research Program (CHRP) Grant (312882) and a CIHR operating grant (MOP86723). We would like to thank Mr. Harry Bojarski and Ryding-Regency Meat Packer for providing the tissues. Appendix. Figures with essential color discrimination Certain figure in this article, particularly Figure 10, is difficult to interpret in black and white. The full color images can be found in the on-line version, at doi:10.1016/j.actbio.2010.05.003.
References [1] Rapoport J, Jacobs P, Bell NR, Klarenbach S. Refining the measurement of the economic burden of chronic diseases in Canada. Chronic Dis Can 2004;25(1): 13–21. [2] Urban JP, Roberts S. Degeneration of the intervertebral disc. Arthritis Res Ther 2003;5(3):120–30. [3] Lopez-Espina CG, Amirouche F, Havalad V. Multilevel cervical fusion and its effect on disc degeneration and osteophyte formation. Spine 2006;31(9): 972–8. [4] Javedan SP, Dickman CA. Cause of adjacent-segment disease after spinal fusion. Lancet 1999;354(9178):530–1. [5] Boden SD. Overview of the biology of lumbar spine fusion and principles for selecting a bone graft substitute. Spine 2002;27(16 Suppl. 1):S26–31. [6] Huang RC, Sandhu HS. The current status of lumbar total disc replacement. Orthop Clin North Am 2004;35(1):33–42. [7] Anderson PA, Rouleau JP. Intervertebral disc arthroplasty. Spine 2004;29(23): 2779–86. [8] Anderson JM. Inflammatory response to implants. ASAIO Trans 1988;34(2): 101–7. [9] Bogduk N. Clinical anatomy of the lumbar spine and sacrum. 4th ed. Edinburgh: Elsevier Churchill Livingstone; 2005. [10] Riches PE, Dhillon N, Lotz J, Woods AW, McNally DS. The internal mechanics of the intervertebral disc under cyclic loading. J Biomech 2002;35(9):1263–71. [11] Alkalay R. The material and mechanical properties of the healthy and degenerated intervertebral disc. In: Integrated biomaterials science. New York, USA: Springer; 2002. p. 403–24. [12] Baer AE, Laursen TA, Guilak F, Setton LA. The micromechanical environment of intervertebral disc cells determined by a finite deformation, anisotropic, and biphasic finite element model. J Biomech Eng 2003;125(1):1–11. [13] Wu HC, Yao RF. Mechanical behavior of the human annulus fibrosus. J Biomech 1976;9(1):1–7. [14] Ebara S, Iatridis JC, Setton LA, Foster RJ, Mow VC, Weidenbaum M. Tensile properties of nondegenerate human lumbar anulus fibrosus. Spine 1996;21(4):452–61. [15] Chang G, Kim HJ, Kaplan D, Vunjak-Novakovic G, Kandel RA. Porous silk scaffolds can be used for tissue engineering annulus fibrosus. Eur Spine J 2007;16(11):1848–57. [16] O’Halloran DM, Pandit AS. Tissue-engineering approach to regenerating the intervertebral disc. Tissue Eng 2007;13(8):1927–54. [17] Johnson WE, Wootton A, El HA, Eisenstein SM, Curtis AS, Roberts S. Topographical guidance of intervertebral disc cell growth in vitro: towards the development of tissue repair strategies for the anulus fibrosus. Eur Spine J 2006;15(15):S389–96. [18] Marchand F, Ahmed AM. Investigation of the laminate structure of lumbar disc anulus fibrosus. Spine 1990;15(5):402–10. [19] Sato M et al. An atelocollagen honeycomb-shaped scaffold with a membrane seal (ACHMS-scaffold) for the culture of annulus fibrosus cells from an intervertebral disc. J Biomed Mater Res A 2003;64(2):248–56. [20] Thonar E, An H, Masuda K. Compartmentalization of the matrix formed by nucleus pulposus and annulus fibrosus cells in alginate gel. Biochem Soc Trans 2002;30(Pt 6):874–8. [21] Wilda H, Gough JE. In vitro studies of annulus fibrosus disc cell attachment, differentiation and matrix production on PDLLA/45S5 Bioglass composite films. Biomaterials 2006;27(30):5220–9. [22] Rong Y, Sugumaran G, Silbert JE, Spector M. Proteoglycans synthesized by canine intervertebral disc cells grown in a type I collagen–glycosaminoglycan matrix. Tissue Eng 2002;8(6):1037–47. [23] Alini M, Li W, Markovic P, Aebi M, Spiro RC, Roughley PJ. The potential and limitations of a cell-seeded collagen/hyaluronan scaffold to engineer an intervertebral disc-like matrix. Spine 2003;28(5):446–54. [24] Mizuno H, Roy AK, Vacanti CA, Kojima K, Ueda M, Bonassar LJ. Tissueengineered composites of anulus fibrosus and nucleus pulposus for intervertebral disc replacement. Spine 2004;29(12):1290–7. [25] Chou AI, Reza AT, Nicoll SB. Distinct intervertebral disc cell populations adopt similar phenotypes in three-dimensional culture. Tissue Eng A 2008;14(12): 2079–87. [26] Ishihara H, Urban JP. Effects of low oxygen concentrations and metabolic inhibitors on proteoglycan and protein synthesis rates in the intervertebral disc. J Orthop Res 1999;17(6):829–35. [27] Li HY, Chang J. PH-compensation effect of bioactive inorganic fillers on the degradation of PLGA. Compos Sci Technol 2005;65(14):2226–32. [28] Helen W, Gough JE. Cell viability, proliferation and extracellular matrix production of human annulus fibrosus cells cultured within PDLLA/Bioglass composite foam scaffolds in vitro. Acta Biomater 2008;4(2):230–43. [29] Mathur AB, Collier TO, Kao WJ, Wiggins M, Schubert MA, Hiltner A, et al. In vivo biocompatibility and biostability of modified polyurethanes. J Biomed Mater Res 1997;36(2):246–57. [30] Santerre JP, Woodhouse K, Laroche G, Labow RS. Understanding the biodegradation of polyurethanes: from classical implants to tissue engineering materials. Biomaterials 2005;26(35):7457–70. [31] van Minnen B, van Leeuwen MB, Stegenga B, Zuidema J, Hissink CE, van Kooten TG, et al. Short-term in vitro and in vivo biocompatibility of a biodegradable polyurethane foam based on 1,4-butanediisocyanate. J Mater Sci Mater Med 2005;16(3):221–7.
M. Yeganegi et al. / Acta Biomaterialia 6 (2010) 3847–3855 [32] Tang YW, Labow RS, Santerre JP. Enzyme-induced biodegradation of polycarbonate polyurethanes: dependence on hard-segment concentration. J Biomed Mater Res 2001;56(4):516–28. [33] Yang L, Kandel RA, Chang G, Santerre JP. Polar surface chemistry of nanofibrous polyurethane scaffold affects annulus fibrosus cell attachment and early matrix accumulation. J Biomed Mater Res A 2009;91(4): 1089–99. [34] Nesti LJ et al. Intervertebral disc tissue engineering using a novel hyaluronic acid-nanofibrous scaffold (HANFS) amalgam. Tissue Eng A 2008;14(9): 1527–37. [35] Nerurkar NL, Mauck RL, Elliott DM. ISSLS prize winner: integrating theoretical and experimental methods for functional tissue engineering of the annulus fibrosus. Spine (Phila, Pa, 1976) 2008;33(25):2691–701. [36] Nerurkar NL, Baker BM, Sen S, Wible EE, Elliott DM, Mauck RL. Nanofibrous biologic laminates replicate the form and function of the annulus fibrosus. Nat Mater 2009;8(12):986–92. [37] Nerurkar NL, Elliott DM, Mauck RL. Mechanics of oriented electrospun nanofibrous scaffolds for annulus fibrosus tissue engineering. J Orthop Res 2007;25(8):1018–28. [38] Thapa A, Miller DC, Webster TJ, Haberstroh KM. Nano-structured polymers enhance bladder smooth muscle cell function. Biomaterials 2003;24(17): 2915–26. [39] Guelcher SA. Biodegradable polyurethanes: synthesis and applications in regenerative medicine. Tissue Eng B Rev 2008;14(1):3–17. [40] Tang YW, Labow RS, Santerre JP. Enzyme-induced biodegradation of polycarbonate-polyurethanes: dependence on hard-segment chemistry. J Biomed Mater Res 2001;57(4):597–611. [41] Gruber HE, Hoelscher G, Ingram JA, Hanley Jr EN. Culture of human anulus fibrosus cells on polyamide nanofibers: extracellular matrix production. Spine (Phila, Pa, 1976) 2009;34(1):4–9.
3855
[42] Stankus JJ, Guan J, Wagner WR. Fabrication of biodegradable elastomeric scaffolds with sub-micron morphologies. J Biomed Mater Res A 2004;70(4): 603–14. [43] Labow RS, Meek E, Santerre JP. Synthesis of cholesterol esterase by monocytederived macrophages: a potential role in the biodegradation of poly(urethane)s. J Biomater Appl 1999;13(3):187–205. [44] Labow RS, Sa D, Matheson LA, Santerre JP. Polycarbonate-urethane hard segment type influences esterase substrate specificity for human-macrophage-mediated biodegradation. J Biomater Sci Polym Ed 2005;16(9):1167–77. [45] Tang YW, Labow RS, Santerre JP. Enzyme induced biodegradation of polycarbonate-polyurethanes: dose dependence effect of cholesterol esterase. Biomaterials 2003;24(12):2003–11. [46] Demers CN, Antoniou J, Mwale F. Value and limitations of using the bovine tail as a model for the human lumbar spine. Spine 2004;29(24):2793–9. [47] Cotterill PC, Kostuik JP, D’Angelo G, Fernie GR, Maki BE. An anatomical comparison of the human and bovine thoracolumbar spine. J Orthop Res 1986;4(3):298–303. [48] Wilke HJ, Krischak S, Claes L. Biomechanical comparison of calf and human spines. J Orthop Res 1996;14(3):500–3. [49] Hoffmann ME, Meneghini R. Action of hydrogen peroxide on human fibroblast in culture. Photochem Photobiol 1979;30(1):151–5. [50] Christenson EM, Patel S, Anderson JM, Hiltner A. Enzymatic degradation of poly(ether urethane) and poly(carbonate urethane) by cholesterol esterase. Biomaterials 2006;27(21):3920–6. [51] Skarja GA, Woodhouse KA. In vitro degradation and erosion of degradable, segmented polyurethanes containing an amino acid-based chain extender. J Biomater Sci Polym Ed 2001;12(8):851–73. [52] Elliott DM, Setton LA. Anisotropic and inhomogeneous tensile behavior of the human anulus fibrosus: experimental measurement and material model predictions. J Biomech Eng 2001;123(3):256–63.