Home
Search
Collections
Journals
About
Contact us
My IOPscience
Additive manufacturing of collagen scaffolds by three-dimensional plotting of highly viscous dispersions
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2016 Biofabrication 8 015015 (http://iopscience.iop.org/1758-5090/8/1/015015) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 165.123.34.86 This content was downloaded on 07/03/2016 at 11:51
Please note that terms and conditions apply.
Biofabrication 8 (2016) 015015
doi:10.1088/1758-5090/8/1/015015
PAPER
RECEIVED
11 September 2015
Additive manufacturing of collagen scaffolds by three-dimensional plotting of highly viscous dispersions
REVISED
10 January 2016 ACCEPTED FOR PUBLICATION
29 January 2016 PUBLISHED
Anja Lode1,4, Michael Meyer2,4, Sophie Brüggemeier1, Birgit Paul1, Hagen Baltzer2, Michaela Schröpfer2, Claudia Winkelmann3, Frank Sonntag3 and Michael Gelinsky1 1
22 February 2016 2 3 4
Centre for Translational Bone, Joint and Soft Tissue Research, University Hospital Carl Gustav Carus and Faculty of Medicine of Technische Universität Dresden, Germany Research Institute of Leather and Plastic Sheeting (FILK) Freiberg, Germany Fraunhofer Institute of Material and Beam Technology (IWS) Dresden, Germany These authors contributed equally to this work.
E-mail: [email protected] Keywords: additive manufacturing, rapid prototyping, extrusion, collagen, freeze-drying
Abstract Additive manufacturing (AM) allows the free form fabrication of three-dimensional (3D) structures with distinct external geometry, fitting into a patient-specific defect, and defined internal pore architecture. However, fabrication of predesigned collagen scaffolds using AM-based technologies is challenging due to the low viscosity of collagen solutions, gels or dispersions commonly used for scaffold preparation. In the present study, we have developed a straightforward method which is based on 3D plotting of a highly viscous, high density collagen dispersion. The swollen state of the collagen fibrils at pH 4 enabled the homogenous extrusion of the material, the deposition of uniform strands and finally the construction of 3D scaffolds. Stabilization of the plotted structures was achieved by freeze-drying and chemical crosslinking with the carbodiimide EDC. The scaffolds exhibited high shape and dimensional fidelity and a hierarchical porosity consisting of macropores generated by strand deposition as well as an interconnected microporosity within the strands as result of the freezedrying process. Cultivation of human mesenchymal stromal cells on the scaffolds, with and without adipogenic or osteogenic stimulation, revealed their cytocompatibility and potential applicability for adipose and bone tissue engineering.
1. Introduction Collagen type I is one of the most common structural elements of various connective tissues. It is characterized by a hierarchical composition: tropocollagen molecules, the nanoscale subunits consisting of a triple helix formed by three polypeptide chains, assembled into microfibrils and finally into fibers which are arranged in a highly organized manner. As the most abundant extracellular matrix (ECM) protein, collagen provides not only structure and strength but also specifically interacts with other biomolecules, contains specific sequence motifs mediating cell adhesion and guides cell function. Moreover, cells can remodel collagenous matrices by digestion through secreted collagenases and synthesis of endogenous collagen [1–3].
© 2016 IOP Publishing Ltd
Due to its excellent biocompatibility, biodegradability and function as ECM, collagen is widely used as biomaterial for current medical applications and for the development of tissue engineering (TE) products and regenerative approaches [4–10]. Collagen-based biomaterials mimicking the natural ECM structure can be produced by decellularization of various mammalian tissues [11]. A number of such biological scaffolds (e.g. derived from skin) has been successfully used in preclinical studies and clinical applications [12]. Moreover, collagen can be extracted from tissues of mammalian origin (mainly bovine and porcine) which are rich in the fibrous protein such as skin and tendon, but also marine organisms are increasingly important sources [13]. Extracted collagen is mostly soluble; the collagen fibrils can be reconstituted in vitro [14]. The physical form of the collagen material
Biofabrication 8 (2016) 015015
A Lode et al
can be adapted to diverse applications: injectable hydrogels create three-dimensional (3D) microenvironments for culture and differentiation of cells, micro-/nanospheres serve as delivery vehicles, membranes and films are used to guide tissue regeneration and porous 3D scaffolds are suitable for TE approaches [3, 14]. TE-based therapies are a promising strategy for patient-specific treatment in regenerative medicine. Suitable scaffolds or implants should have an outer shape perfectly restoring the defect and an inner pore structure providing sufficient stability and allowing cell ingrowth and vascularization. Both, shaping for specific applications and control of pore architecture, is challenging and hard to achieve with conventional molding techniques. Therefore, in the last years additive manufacturing (AM) technologies have attained great importance and attention for the development of implants and TE scaffolds [15–18]. The principle of AM is the ‘free form fabrication’ of 3D structures through layer-by-layer construction in a computer aided design/computer aided manufacturing (CAD/ CAM) process [19]. In the present study, collagen scaffolds of predesigned architecture were fabricated by using 3D plotting, an extrusion technology-based AM technique developed by Landers and Mülhaupt [20]. Pasty or highly viscous materials are extruded through a dosing needle and deposited as strands on a building platform to construct 3D structures. A broad spectrum of biomaterials can be processed with this technique. Depending on the type of material, suspensions, solutions, slurries or melts are applied. In addition, by using multi-channel dosing systems, more than one biomaterial can be plotted in the same fabrication process allowing the realization of more complex, multiphasic scaffold designs [21]. Fabrication of predesigned collagen scaffolds using AM methods is especially challenging due to the low viscosity of collagen solutions. Up to now, only very few studies were conducted and two strategies have been reported so far. Firstly, AM was used to fabricate sacrificial negative molds with specific size and shape in which the low-viscous collagen material was then cast (indirect method) [22, 23]. Secondly, Ahn and coworkers developed a direct cryogenic plotting method which allows the deposition of a low-viscosity collagen solution at temperatures below 0 °C [24]. In contrast, we applied in the present study a highly viscous collagen dispersion consisting of insoluble collagen fibrils and fibers for the fabrication of 3D scaffolds by direct plotting at room temperature. The plotted collagen scaffolds were characterized with respect to their pore structure and porosity, dimensional fidelity (swelling and shrinkage) and mechanical properties as well as in cell culture experiments with human mesenchymal stromal cells (hMSC). 2
2. Materials and methods 2.1. Preparation and characterization of the collagen dispersion The preparation technology of the collagen dispersion aimed to manufacture plottable preparations from porcine collagen which should be endotoxin-free and can be stored and sterilized in dry state. Porcine hides were obtained from a local abattoir and were split to remove the fat on a Turner splitting machine 537-VI which had been adapted to split wet porcine hides. Remaining skins were washed with deionized water and all solutions used for the following processing steps were prepared with deionized water. The skins were limed with saturated calcium hydroxide solution (Ph. Eur; Th. Geyer, Renningen, Germany) and 3% (m/v) sodium sulfide (60% scales, technical quality; Th. Geyer) overnight to remove hair. The float was separated and the resulting pelts were washed and treated once more with lime overnight. After removing the float and washing the purified material again, the alkaline pH was first adjusted to 8.5 with a 3% (m/ v) solution of ammonium chloride (Ph. Eur; Th. Geyer). Then the material was treated with 3% (v/v) hydrogen peroxide (stock solution 30%, Ph. Eur; Th. Geyer) for 3 h and acidified to pH 4 by addition of hydrochloric acid (stock solution 32%, p.a.; Carl Roth, Germany). The swollen material was minced with a meat chopper (Nagütec SW114, Germany) followed by several grinding steps in a colloid mill (in house construction) under addition of ice, prepared from deionized water. The dispersion was neutralized with 0.1 M sodium hydroxide solution (prepared from granules, p.a.; Carl Roth), concentrated by centrifugation, dialyzed against purified water (regenerated cellulose, MWCO 12–14 kDa; Serva) and finally freeze dried. Prior to freeze-drying, the endotoxin level in the neutral dialyzed dispersion was measured with the Pyrochrome® Kinetic Chromogenic Endotoxin Testing kit (Associates of Cape Cod, USA) according to the manufacturer’s instructions. For medical products the limit is 0.5 EU ml−1. The concentration of the collagen dispersion before plotting was analyzed by measurement of dry matter content and ash content according to DIN EN ISO 4047 and 4684, respectively. Hydroxyproline content was measured and the collagen content was calculated using the factor 7.46 according to Reich [25]. To evaluate the purity, the amino acid composition was measured according to Ph. Eur.2.2.56. The freeze-dried collagen could be stored over months and sterilized by gamma irradiation, which could be shown by repeated preparations from the same stock. Immediately prior to plotting, the dry collagen was redispersed either in phosphate buffered saline (PBS; Gibco, Life Technologies, Darmstadt, Germany) or in 0.2 M hydrochloric acid in a mass ratio of 1: 6 (collagen: dispersant) and thoroughly mixed to yield a neutral and an acidic dispersion, respectively.
Biofabrication 8 (2016) 015015
A Lode et al
Viscosity of the collagen dispersion was measured with a plate rheometer MCR 501 (Anton Paar GmbH, Germany) using the plate to plate measuring system PP 25. The gap distance was set to 1 mm and the temperature to 10 °C. This low temperature was necessary to prevent evaporation at the edge of the plates which would lead to an increase of the concentration of the collagen dispersion and therefore would affect the measurement. The collagen dispersion was applied between the plates and the system was tempered for 5 min. First, the viscosity of the dispersion was measured in the oscillating modus with a shear rate in the range between 0.01 and 100 s−1 to characterize the flow behavior. Then, different preparations were compared at the same shear rate of 0.1 s−1. The water binding capacity was measured by centrifugation of 80 g dispersion at 2650 RCF, the supernatant was separated, and the mass was weighed again.
2.2. Fabrication of collagen scaffolds: 3D plotting and post-processing The 3D plotting system used in this study was the BioScaffolder 2.1 (GeSiM, Großerkmannsdorf, Germany). The collagen dispersion was extruded through dosing needles (Globaco GmbH, Roedermark, Germany) with an inner diameter of 610 μm with a dosing pressure of 1.0 bar and a plotting speed of 18 mm min−1. Collagen strands were deposited layerby-layer in a 0°0°/90°90° or a 0°0°/45°45°/90°90°/ 135°135° lay-down-pattern with a strand distance of 1.6 mm in air as plotting environment. The number of layers varied between 6 and 33 and is indicated in the respective sections and figure legends. After plotting, the collagen scaffolds were frozen at −20 °C and subsequently freeze-dried. Crosslinking of collagen was carried out by incubation of the scaffolds for 1.5 h in 1% (m/v) EDC (1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide; Fluka, Germany) dissolved in 80% (v/v) ethanol. After thorough rinsing of the scaffolds in distilled water, 1% (w/v) glycine solution and again in water, a final freeze-drying step was conducted. Prior to cell culture, the scaffolds were sterilized by gamma irradiation at 25 kGy.
2.3. Characterization of the plotted collagen scaffolds 2.3.1. Macro- and microstructure The scaffolds (0°0°/90°90° configuration, 10 layers) were imaged by stereo light microscopy using a Leica M205 C equipped with a DFC295 camera (Leica, Germany). Macro- and microporosity were investigated by scanning electron microscopy (SEM). Scaffold samples were mounted on aluminum pins, coated with gold, and analyzed using a Philips XL 30/ESEM with field emission gun (FEG), operated in SEM mode. 3
2.3.2. Porosity and dimensional fidelity The real volume of the plotted collagen scaffolds (0° 0°/90°90° configuration, 26 layers; n=12) was evaluated by helium-pycnometry (Ultrapyc 1200e, Quantachrome Instruments, Boynton Beach, USA). Porosity was then calculated by the ratio of real to apparent density. Dimensional fidelity was evaluated by measurement of length and height of the scaffolds (0°0°/90°90° configuration, 10 layers; n=5) at various processing steps and comparison to the predefined dimensions. In addition, thickness and spacing of the strands were determined by analyzing the microscopic pictures of scaffolds (both configurations) with ImageJ 1.44p (National Institutes of Health, Bethesda, Maryland, USA). 2.3.3. Mechanical testing Compressive mechanical properties of collagen scaffolds with an edge length of 10 mm in each direction (26 layers with 6 parallel strands each were plotted on top of each other in a 0°0°/90°90° configuration) were obtained by uniaxial compression using a Z010 equipped with a 100 N load cell (Zwick, Germany). Prior to testing, the samples (n=12) were incubated in simulated body fluid [26] for 24 h and then investigated in wet state at room temperature. Static compression was performed with a velocity of 1% s−1. Compressive modulus was evaluated at the linear slope. Compressive yield stress was set at the slope alteration of the compressive curve. 2.4. Cell culture experiments 2.4.1. Human mesenchymal stromal cells (hMSC) Human MSC, isolated from bone marrow aspirates of healthy donors, were kindly provided by Professor M Bornhäuser and co-workers (Medical Clinic I, University Hospital Carl Gustav Carus Dresden). The application of hMSC for the in vitro experiments was approved by the ethics commission of the Faculty of Medicine of Technische Universität Dresden. The cells were expanded at 37 °C and 5% CO2 in Dulbecco’s modified eagle’s medium low glucose (DMEM; Gibco, Life Technologies) supplemented with 9% fetal calf serum (Biochrom, Berlin, Germany), 100 U ml−1 penicillin and 100 μg ml−1 streptomycin (Gibco, Life Technologies). Cells in passage 4 were used for the experiments. 2.4.2. Cultivation of hMSC in plotted collagen scaffolds The produced scaffolds (8×8 mm2, 6 layers with 6 parallel strands each were plotted in a 0°0°/90°90° configuration) were incubated in cell culture medium for 24 h and thereafter seeded at a density of 2×105 cells per scaffold. Adipogenic differentiation of hMSC was stimulated by addition of 5 μg ml−1 insulin, 1 μM dexamethasone, 60 μM indomethacin and 0.5 mM 3-isobutyl-1-methylxanthine (all from Sigma-Aldrich, Taufenkirchen, Germany); osteogenic differentiation of
Biofabrication 8 (2016) 015015
A Lode et al
Table 1. Primers for RT-PCR. Gene
Primer sequences
FABP4 [29]
for: 5′-ATGCTTTTGTAGGTACCTGG-3′ rev: 5′-CTCTCTCATAAACTCTCGTG-3′ for: 5′-AAGACCACTCCCACTCCTTTG-3′ rev: 5′-GTCAGCGGACTCTGGATTCA-3′ for: 5′-AAACAGCATCAGCGTTCCCATC-3′ rev: 5′-AGTGTTGGCAGCAAATTCCG-3′ for: 5′-CTG AAG ACA CAG CTG AGG AC-3′ rev: 5′-CTG GTG AAT GTG TGT AAG AC-3′ for: 5′-GGACTTCGAGCAAGAGATGG-3′ rev: 5′-AGCACTGTGTTGGCGTACAG-3′ for: 5′-ACC ATT CCC ACG TCT TCA CAT TTG-3′ rev: 5′-ATT CTC TCG TTC ACC GCC CAC-3′ for: 5′-AAT GAA AAC GAA GAA AGC GAA G-3′ rev: 5′-ATC ATA GCC ATC GTA GCC TTG T-3′ for: 5′-GGT GAA GGT CGG AGT CAA CGG-3′ rev: 5′-GGT CAT GAG TCC TTC CAC GAT-3′
PPARγ [29] PLIN [30] LPL [29] β-actin ALP [31] BSP II [31] GAPDH
hMSC was induced by addition of 100 nM dexamethasone, 3.5 mM β-glycerophosphate and 0.05 mM ascorbic acid 2-phosphate (all from Sigma-Aldrich) to the cell culture medium. Induction of osteogenic and adipogenic differentiation was started one day after seeding. Cell-seeded scaffolds were cultivated over 21 days. 2.5. Characterization of cell-seeded collagen scaffolds 2.5.1. Analysis of cell adhesion, proliferation and viability 24 h after seeding as well as after 21 days of cultivation, the number of cells attached and grown on the scaffolds was evaluated by DNA quantification and measurement of the lactate dehydrogenase (LDH) activity as previously described [27]. In brief, samples were washed with PBS, frozen, homogenized using a Precellys24 apparatus (Peqlab, Erlangen, Germany) and incubated in lysis buffer (1% (m/v) Triton X-100 in PBS). DNA content was determined with the Quant-iTTM PicoGreen® dsDNA reagent (Life Technologies) according to manufacturer’s instructions and correlated with the cell number using a calibration line. LDH activity was measured with the CytoTox 96® Non-Radioactive Cytotoxicity Assay (Promega, Madison, USA) according to manufacturer’s instructions and correlated with the cell number. Cell morphology and distribution were analyzed by SEM and confocal laser scanning microscopy (cLSM). Samples, cultured for 1 and 21 days, were washed with PBS and fixed with 3.7% formaldehyde/ PBS. For SEM, fixed samples were washed with distilled water and dehydrated using a gradation series of ethanol/distilled water solutions. Critical point drying was performed with a CPD 030 apparatus (BAL-TEC, Liechtenstein). Dried samples were coated with gold and imaged using a Philips XL 30/ESEM with FEG, operating in SEM mode. Samples for cLSM were prepared according to a previously published
4
Tannealing
Amplicon size
52 °C
387 bp
55 °C
554 bp
55 °C
118 bp
52 °C
505 bp
55 °C
234 bp
55 °C
162 bp
55 °C
450 bp
55 °C
520 bp
protocol [28]. In brief, the samples were incubated with AlexaFluor 488® phalloidin (Life Technologies) and DAPI (Sigma-Aldrich) to stain cytoskeletons and nuclei. Confocal LSM was performed using a Zeiss (Germany) cLSM 510, located in the MTZ imaging facility of Technische Universität Dresden, with excitation wavelengths of 488 nm (phalloidin) and 405 nm (DAPI). 2.5.2. Analysis of hMSC differentiation along the adipogenic and osteogenic lineage Gene expression of adipogenic and osteogenic markers was analyzed by reverse transcriptase PCR (RTPCR). RNA was isolated from the cell-seeded scaffolds using the peqGOLD MicroSpin Total RNA Kit (Peqlab, Erlangen, Germany) according to manufacturer’s instructions. During the procedure, lysates of five replicates were pooled. 200 ng of total RNA were reverse transcribed into cDNA in a 20 μl reaction mixture containing 100 U of Superscript II (Life Technologies), 0.5 mM dNTPs (Peqlab), 12.5 ng μl−1 random hexamers (Eurofins MWG Operon, Ebersberg, Germany) and 40 U RNase OUT (Life Technologies). For PCR experiments, 1 μl of cDNA was amplified in a 20 μl reaction mixture containing 1.5 U HotTaq-Polymerase (Peqlab), 0.2 mM dNTPs, 1.5 mM MgCl2 and specific primer pairs (1 μM of each primer). Sequences and annealing temperatures of the primers (Eurofins MWG Operon) as well as the amplicon sizes are summarized in table 1. The PCR products were visualized using a 2% agarose gel stained with Gel Red (Biotium, Hayward, CA, USA). Furthermore, adipogenic differentiation was assessed by measurement of glycerol 3-phosphate dehydrogenase (GPDH) activity as well as by cLSM analysis of Nile red stained samples as described previously [32]. GPDH activity was determined according to the method described by Wise and Green [33]. In brief, samples were taken on day 1 and 21 of culture,
Biofabrication 8 (2016) 015015
A Lode et al
washed with PBS and frozen at −80 °C. After thawing, they were incubated in lysis buffer containing 50 mM Tris, 1 mM EDTA and 1 mM β-mercaptoethanol (all from Sigma-Aldrich) followed by sonication. 50 μl of each cell lysate were mixed with 150 μl substrate solution consisting of 0.2 mM dihydroxyacetone phosphate and 0.12 mM NADH in 100 mM triethanolamine/2.5 mM EDTA (all from Sigma-Aldrich). Enzyme kinetics was determined by measurement of absorbance at 340 nm. The enzyme activity of GPDH was detected as decrease of the NADH concentration whereby 1 mU corresponds to the oxidation of 1 nmol NADH min−1. In order to relate GPDH activity to the cell number, the DNA content of the same lysates was determined as described above. For microscopic detection of cells with characteristic fat vacuoles, samples were taken on day 21 of culture, washed with PBS, fixed with 3.7% formaldehyde/PBS and incubated with Nile red (Sigma-Aldrich) and DAPI to stain fat droplets and nuclei. Microscopic analysis was performed applying a Zeiss cLSM 510 with excitation wavelengths of 561 nm (Nile red) and 405 nm (DAPI). Osteogenic differentiation was further verified by measurement of alkaline phosphatase (ALP) activity and by alizarin red staining of deposited mineral. ALP activity was determined as described [27]. In brief, samples were taken on day 1 and 21 of culture, washed with PBS, frozen, homogenized and incubated in lysis buffer (1% (m/v) Triton X-100 in PBS). 20 μl of each cell lysate were mixed with 100 μl substrate solution (1 mg ml−1 p-nitrophenylphosphate (Sigma-Aldrich) in 0.1 M diethanolamine/1% (m/v) Triton X-100/ 1 mM MgCl2, pH 9.8). After incubation at 37 °C for 30 min, the enzymatic reaction was stopped by addition of 1 M NaOH; p-nitrophenolate (pNp) formation was quantified by absorbance measurement at 405 nm. The amount of pNp produced by the cell lysate was calculated using a p-nitrophenol calibration line and related to the cell number in each sample (calculated from DNA content which was determined as described above) to express specific ALP activity (μmol pNp/30 min /106 cells). For alizarin red staining, samples were taken on day 21 of culture, washed with PBS, fixed with 3.7% formaldehyde/PBS and incubated in alizarin red staining solution (Sigma-Aldrich) for 2 min followed by incubation in distilled water. Microscopic analysis was performed applying the Leica M205 C equipped with a DFC295 camera.
2.6. Statistical analysis One-way analysis of variance was used to evaluate statistical significance; post-hoc analysis using the Tukey-method was used to determine multiple comparisons (Origin 8.5.0G, OriginLab, USA). Significant differences were assumed at p
Search
Collections
Journals
About
Contact us
My IOPscience
Additive manufacturing of collagen scaffolds by three-dimensional plotting of highly viscous dispersions
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2016 Biofabrication 8 015015 (http://iopscience.iop.org/1758-5090/8/1/015015) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 165.123.34.86 This content was downloaded on 07/03/2016 at 11:51
Please note that terms and conditions apply.
Biofabrication 8 (2016) 015015
doi:10.1088/1758-5090/8/1/015015
PAPER
RECEIVED
11 September 2015
Additive manufacturing of collagen scaffolds by three-dimensional plotting of highly viscous dispersions
REVISED
10 January 2016 ACCEPTED FOR PUBLICATION
29 January 2016 PUBLISHED
Anja Lode1,4, Michael Meyer2,4, Sophie Brüggemeier1, Birgit Paul1, Hagen Baltzer2, Michaela Schröpfer2, Claudia Winkelmann3, Frank Sonntag3 and Michael Gelinsky1 1
22 February 2016 2 3 4
Centre for Translational Bone, Joint and Soft Tissue Research, University Hospital Carl Gustav Carus and Faculty of Medicine of Technische Universität Dresden, Germany Research Institute of Leather and Plastic Sheeting (FILK) Freiberg, Germany Fraunhofer Institute of Material and Beam Technology (IWS) Dresden, Germany These authors contributed equally to this work.
E-mail: [email protected] Keywords: additive manufacturing, rapid prototyping, extrusion, collagen, freeze-drying
Abstract Additive manufacturing (AM) allows the free form fabrication of three-dimensional (3D) structures with distinct external geometry, fitting into a patient-specific defect, and defined internal pore architecture. However, fabrication of predesigned collagen scaffolds using AM-based technologies is challenging due to the low viscosity of collagen solutions, gels or dispersions commonly used for scaffold preparation. In the present study, we have developed a straightforward method which is based on 3D plotting of a highly viscous, high density collagen dispersion. The swollen state of the collagen fibrils at pH 4 enabled the homogenous extrusion of the material, the deposition of uniform strands and finally the construction of 3D scaffolds. Stabilization of the plotted structures was achieved by freeze-drying and chemical crosslinking with the carbodiimide EDC. The scaffolds exhibited high shape and dimensional fidelity and a hierarchical porosity consisting of macropores generated by strand deposition as well as an interconnected microporosity within the strands as result of the freezedrying process. Cultivation of human mesenchymal stromal cells on the scaffolds, with and without adipogenic or osteogenic stimulation, revealed their cytocompatibility and potential applicability for adipose and bone tissue engineering.
1. Introduction Collagen type I is one of the most common structural elements of various connective tissues. It is characterized by a hierarchical composition: tropocollagen molecules, the nanoscale subunits consisting of a triple helix formed by three polypeptide chains, assembled into microfibrils and finally into fibers which are arranged in a highly organized manner. As the most abundant extracellular matrix (ECM) protein, collagen provides not only structure and strength but also specifically interacts with other biomolecules, contains specific sequence motifs mediating cell adhesion and guides cell function. Moreover, cells can remodel collagenous matrices by digestion through secreted collagenases and synthesis of endogenous collagen [1–3].
© 2016 IOP Publishing Ltd
Due to its excellent biocompatibility, biodegradability and function as ECM, collagen is widely used as biomaterial for current medical applications and for the development of tissue engineering (TE) products and regenerative approaches [4–10]. Collagen-based biomaterials mimicking the natural ECM structure can be produced by decellularization of various mammalian tissues [11]. A number of such biological scaffolds (e.g. derived from skin) has been successfully used in preclinical studies and clinical applications [12]. Moreover, collagen can be extracted from tissues of mammalian origin (mainly bovine and porcine) which are rich in the fibrous protein such as skin and tendon, but also marine organisms are increasingly important sources [13]. Extracted collagen is mostly soluble; the collagen fibrils can be reconstituted in vitro [14]. The physical form of the collagen material
Biofabrication 8 (2016) 015015
A Lode et al
can be adapted to diverse applications: injectable hydrogels create three-dimensional (3D) microenvironments for culture and differentiation of cells, micro-/nanospheres serve as delivery vehicles, membranes and films are used to guide tissue regeneration and porous 3D scaffolds are suitable for TE approaches [3, 14]. TE-based therapies are a promising strategy for patient-specific treatment in regenerative medicine. Suitable scaffolds or implants should have an outer shape perfectly restoring the defect and an inner pore structure providing sufficient stability and allowing cell ingrowth and vascularization. Both, shaping for specific applications and control of pore architecture, is challenging and hard to achieve with conventional molding techniques. Therefore, in the last years additive manufacturing (AM) technologies have attained great importance and attention for the development of implants and TE scaffolds [15–18]. The principle of AM is the ‘free form fabrication’ of 3D structures through layer-by-layer construction in a computer aided design/computer aided manufacturing (CAD/ CAM) process [19]. In the present study, collagen scaffolds of predesigned architecture were fabricated by using 3D plotting, an extrusion technology-based AM technique developed by Landers and Mülhaupt [20]. Pasty or highly viscous materials are extruded through a dosing needle and deposited as strands on a building platform to construct 3D structures. A broad spectrum of biomaterials can be processed with this technique. Depending on the type of material, suspensions, solutions, slurries or melts are applied. In addition, by using multi-channel dosing systems, more than one biomaterial can be plotted in the same fabrication process allowing the realization of more complex, multiphasic scaffold designs [21]. Fabrication of predesigned collagen scaffolds using AM methods is especially challenging due to the low viscosity of collagen solutions. Up to now, only very few studies were conducted and two strategies have been reported so far. Firstly, AM was used to fabricate sacrificial negative molds with specific size and shape in which the low-viscous collagen material was then cast (indirect method) [22, 23]. Secondly, Ahn and coworkers developed a direct cryogenic plotting method which allows the deposition of a low-viscosity collagen solution at temperatures below 0 °C [24]. In contrast, we applied in the present study a highly viscous collagen dispersion consisting of insoluble collagen fibrils and fibers for the fabrication of 3D scaffolds by direct plotting at room temperature. The plotted collagen scaffolds were characterized with respect to their pore structure and porosity, dimensional fidelity (swelling and shrinkage) and mechanical properties as well as in cell culture experiments with human mesenchymal stromal cells (hMSC). 2
2. Materials and methods 2.1. Preparation and characterization of the collagen dispersion The preparation technology of the collagen dispersion aimed to manufacture plottable preparations from porcine collagen which should be endotoxin-free and can be stored and sterilized in dry state. Porcine hides were obtained from a local abattoir and were split to remove the fat on a Turner splitting machine 537-VI which had been adapted to split wet porcine hides. Remaining skins were washed with deionized water and all solutions used for the following processing steps were prepared with deionized water. The skins were limed with saturated calcium hydroxide solution (Ph. Eur; Th. Geyer, Renningen, Germany) and 3% (m/v) sodium sulfide (60% scales, technical quality; Th. Geyer) overnight to remove hair. The float was separated and the resulting pelts were washed and treated once more with lime overnight. After removing the float and washing the purified material again, the alkaline pH was first adjusted to 8.5 with a 3% (m/ v) solution of ammonium chloride (Ph. Eur; Th. Geyer). Then the material was treated with 3% (v/v) hydrogen peroxide (stock solution 30%, Ph. Eur; Th. Geyer) for 3 h and acidified to pH 4 by addition of hydrochloric acid (stock solution 32%, p.a.; Carl Roth, Germany). The swollen material was minced with a meat chopper (Nagütec SW114, Germany) followed by several grinding steps in a colloid mill (in house construction) under addition of ice, prepared from deionized water. The dispersion was neutralized with 0.1 M sodium hydroxide solution (prepared from granules, p.a.; Carl Roth), concentrated by centrifugation, dialyzed against purified water (regenerated cellulose, MWCO 12–14 kDa; Serva) and finally freeze dried. Prior to freeze-drying, the endotoxin level in the neutral dialyzed dispersion was measured with the Pyrochrome® Kinetic Chromogenic Endotoxin Testing kit (Associates of Cape Cod, USA) according to the manufacturer’s instructions. For medical products the limit is 0.5 EU ml−1. The concentration of the collagen dispersion before plotting was analyzed by measurement of dry matter content and ash content according to DIN EN ISO 4047 and 4684, respectively. Hydroxyproline content was measured and the collagen content was calculated using the factor 7.46 according to Reich [25]. To evaluate the purity, the amino acid composition was measured according to Ph. Eur.2.2.56. The freeze-dried collagen could be stored over months and sterilized by gamma irradiation, which could be shown by repeated preparations from the same stock. Immediately prior to plotting, the dry collagen was redispersed either in phosphate buffered saline (PBS; Gibco, Life Technologies, Darmstadt, Germany) or in 0.2 M hydrochloric acid in a mass ratio of 1: 6 (collagen: dispersant) and thoroughly mixed to yield a neutral and an acidic dispersion, respectively.
Biofabrication 8 (2016) 015015
A Lode et al
Viscosity of the collagen dispersion was measured with a plate rheometer MCR 501 (Anton Paar GmbH, Germany) using the plate to plate measuring system PP 25. The gap distance was set to 1 mm and the temperature to 10 °C. This low temperature was necessary to prevent evaporation at the edge of the plates which would lead to an increase of the concentration of the collagen dispersion and therefore would affect the measurement. The collagen dispersion was applied between the plates and the system was tempered for 5 min. First, the viscosity of the dispersion was measured in the oscillating modus with a shear rate in the range between 0.01 and 100 s−1 to characterize the flow behavior. Then, different preparations were compared at the same shear rate of 0.1 s−1. The water binding capacity was measured by centrifugation of 80 g dispersion at 2650 RCF, the supernatant was separated, and the mass was weighed again.
2.2. Fabrication of collagen scaffolds: 3D plotting and post-processing The 3D plotting system used in this study was the BioScaffolder 2.1 (GeSiM, Großerkmannsdorf, Germany). The collagen dispersion was extruded through dosing needles (Globaco GmbH, Roedermark, Germany) with an inner diameter of 610 μm with a dosing pressure of 1.0 bar and a plotting speed of 18 mm min−1. Collagen strands were deposited layerby-layer in a 0°0°/90°90° or a 0°0°/45°45°/90°90°/ 135°135° lay-down-pattern with a strand distance of 1.6 mm in air as plotting environment. The number of layers varied between 6 and 33 and is indicated in the respective sections and figure legends. After plotting, the collagen scaffolds were frozen at −20 °C and subsequently freeze-dried. Crosslinking of collagen was carried out by incubation of the scaffolds for 1.5 h in 1% (m/v) EDC (1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide; Fluka, Germany) dissolved in 80% (v/v) ethanol. After thorough rinsing of the scaffolds in distilled water, 1% (w/v) glycine solution and again in water, a final freeze-drying step was conducted. Prior to cell culture, the scaffolds were sterilized by gamma irradiation at 25 kGy.
2.3. Characterization of the plotted collagen scaffolds 2.3.1. Macro- and microstructure The scaffolds (0°0°/90°90° configuration, 10 layers) were imaged by stereo light microscopy using a Leica M205 C equipped with a DFC295 camera (Leica, Germany). Macro- and microporosity were investigated by scanning electron microscopy (SEM). Scaffold samples were mounted on aluminum pins, coated with gold, and analyzed using a Philips XL 30/ESEM with field emission gun (FEG), operated in SEM mode. 3
2.3.2. Porosity and dimensional fidelity The real volume of the plotted collagen scaffolds (0° 0°/90°90° configuration, 26 layers; n=12) was evaluated by helium-pycnometry (Ultrapyc 1200e, Quantachrome Instruments, Boynton Beach, USA). Porosity was then calculated by the ratio of real to apparent density. Dimensional fidelity was evaluated by measurement of length and height of the scaffolds (0°0°/90°90° configuration, 10 layers; n=5) at various processing steps and comparison to the predefined dimensions. In addition, thickness and spacing of the strands were determined by analyzing the microscopic pictures of scaffolds (both configurations) with ImageJ 1.44p (National Institutes of Health, Bethesda, Maryland, USA). 2.3.3. Mechanical testing Compressive mechanical properties of collagen scaffolds with an edge length of 10 mm in each direction (26 layers with 6 parallel strands each were plotted on top of each other in a 0°0°/90°90° configuration) were obtained by uniaxial compression using a Z010 equipped with a 100 N load cell (Zwick, Germany). Prior to testing, the samples (n=12) were incubated in simulated body fluid [26] for 24 h and then investigated in wet state at room temperature. Static compression was performed with a velocity of 1% s−1. Compressive modulus was evaluated at the linear slope. Compressive yield stress was set at the slope alteration of the compressive curve. 2.4. Cell culture experiments 2.4.1. Human mesenchymal stromal cells (hMSC) Human MSC, isolated from bone marrow aspirates of healthy donors, were kindly provided by Professor M Bornhäuser and co-workers (Medical Clinic I, University Hospital Carl Gustav Carus Dresden). The application of hMSC for the in vitro experiments was approved by the ethics commission of the Faculty of Medicine of Technische Universität Dresden. The cells were expanded at 37 °C and 5% CO2 in Dulbecco’s modified eagle’s medium low glucose (DMEM; Gibco, Life Technologies) supplemented with 9% fetal calf serum (Biochrom, Berlin, Germany), 100 U ml−1 penicillin and 100 μg ml−1 streptomycin (Gibco, Life Technologies). Cells in passage 4 were used for the experiments. 2.4.2. Cultivation of hMSC in plotted collagen scaffolds The produced scaffolds (8×8 mm2, 6 layers with 6 parallel strands each were plotted in a 0°0°/90°90° configuration) were incubated in cell culture medium for 24 h and thereafter seeded at a density of 2×105 cells per scaffold. Adipogenic differentiation of hMSC was stimulated by addition of 5 μg ml−1 insulin, 1 μM dexamethasone, 60 μM indomethacin and 0.5 mM 3-isobutyl-1-methylxanthine (all from Sigma-Aldrich, Taufenkirchen, Germany); osteogenic differentiation of
Biofabrication 8 (2016) 015015
A Lode et al
Table 1. Primers for RT-PCR. Gene
Primer sequences
FABP4 [29]
for: 5′-ATGCTTTTGTAGGTACCTGG-3′ rev: 5′-CTCTCTCATAAACTCTCGTG-3′ for: 5′-AAGACCACTCCCACTCCTTTG-3′ rev: 5′-GTCAGCGGACTCTGGATTCA-3′ for: 5′-AAACAGCATCAGCGTTCCCATC-3′ rev: 5′-AGTGTTGGCAGCAAATTCCG-3′ for: 5′-CTG AAG ACA CAG CTG AGG AC-3′ rev: 5′-CTG GTG AAT GTG TGT AAG AC-3′ for: 5′-GGACTTCGAGCAAGAGATGG-3′ rev: 5′-AGCACTGTGTTGGCGTACAG-3′ for: 5′-ACC ATT CCC ACG TCT TCA CAT TTG-3′ rev: 5′-ATT CTC TCG TTC ACC GCC CAC-3′ for: 5′-AAT GAA AAC GAA GAA AGC GAA G-3′ rev: 5′-ATC ATA GCC ATC GTA GCC TTG T-3′ for: 5′-GGT GAA GGT CGG AGT CAA CGG-3′ rev: 5′-GGT CAT GAG TCC TTC CAC GAT-3′
PPARγ [29] PLIN [30] LPL [29] β-actin ALP [31] BSP II [31] GAPDH
hMSC was induced by addition of 100 nM dexamethasone, 3.5 mM β-glycerophosphate and 0.05 mM ascorbic acid 2-phosphate (all from Sigma-Aldrich) to the cell culture medium. Induction of osteogenic and adipogenic differentiation was started one day after seeding. Cell-seeded scaffolds were cultivated over 21 days. 2.5. Characterization of cell-seeded collagen scaffolds 2.5.1. Analysis of cell adhesion, proliferation and viability 24 h after seeding as well as after 21 days of cultivation, the number of cells attached and grown on the scaffolds was evaluated by DNA quantification and measurement of the lactate dehydrogenase (LDH) activity as previously described [27]. In brief, samples were washed with PBS, frozen, homogenized using a Precellys24 apparatus (Peqlab, Erlangen, Germany) and incubated in lysis buffer (1% (m/v) Triton X-100 in PBS). DNA content was determined with the Quant-iTTM PicoGreen® dsDNA reagent (Life Technologies) according to manufacturer’s instructions and correlated with the cell number using a calibration line. LDH activity was measured with the CytoTox 96® Non-Radioactive Cytotoxicity Assay (Promega, Madison, USA) according to manufacturer’s instructions and correlated with the cell number. Cell morphology and distribution were analyzed by SEM and confocal laser scanning microscopy (cLSM). Samples, cultured for 1 and 21 days, were washed with PBS and fixed with 3.7% formaldehyde/ PBS. For SEM, fixed samples were washed with distilled water and dehydrated using a gradation series of ethanol/distilled water solutions. Critical point drying was performed with a CPD 030 apparatus (BAL-TEC, Liechtenstein). Dried samples were coated with gold and imaged using a Philips XL 30/ESEM with FEG, operating in SEM mode. Samples for cLSM were prepared according to a previously published
4
Tannealing
Amplicon size
52 °C
387 bp
55 °C
554 bp
55 °C
118 bp
52 °C
505 bp
55 °C
234 bp
55 °C
162 bp
55 °C
450 bp
55 °C
520 bp
protocol [28]. In brief, the samples were incubated with AlexaFluor 488® phalloidin (Life Technologies) and DAPI (Sigma-Aldrich) to stain cytoskeletons and nuclei. Confocal LSM was performed using a Zeiss (Germany) cLSM 510, located in the MTZ imaging facility of Technische Universität Dresden, with excitation wavelengths of 488 nm (phalloidin) and 405 nm (DAPI). 2.5.2. Analysis of hMSC differentiation along the adipogenic and osteogenic lineage Gene expression of adipogenic and osteogenic markers was analyzed by reverse transcriptase PCR (RTPCR). RNA was isolated from the cell-seeded scaffolds using the peqGOLD MicroSpin Total RNA Kit (Peqlab, Erlangen, Germany) according to manufacturer’s instructions. During the procedure, lysates of five replicates were pooled. 200 ng of total RNA were reverse transcribed into cDNA in a 20 μl reaction mixture containing 100 U of Superscript II (Life Technologies), 0.5 mM dNTPs (Peqlab), 12.5 ng μl−1 random hexamers (Eurofins MWG Operon, Ebersberg, Germany) and 40 U RNase OUT (Life Technologies). For PCR experiments, 1 μl of cDNA was amplified in a 20 μl reaction mixture containing 1.5 U HotTaq-Polymerase (Peqlab), 0.2 mM dNTPs, 1.5 mM MgCl2 and specific primer pairs (1 μM of each primer). Sequences and annealing temperatures of the primers (Eurofins MWG Operon) as well as the amplicon sizes are summarized in table 1. The PCR products were visualized using a 2% agarose gel stained with Gel Red (Biotium, Hayward, CA, USA). Furthermore, adipogenic differentiation was assessed by measurement of glycerol 3-phosphate dehydrogenase (GPDH) activity as well as by cLSM analysis of Nile red stained samples as described previously [32]. GPDH activity was determined according to the method described by Wise and Green [33]. In brief, samples were taken on day 1 and 21 of culture,
Biofabrication 8 (2016) 015015
A Lode et al
washed with PBS and frozen at −80 °C. After thawing, they were incubated in lysis buffer containing 50 mM Tris, 1 mM EDTA and 1 mM β-mercaptoethanol (all from Sigma-Aldrich) followed by sonication. 50 μl of each cell lysate were mixed with 150 μl substrate solution consisting of 0.2 mM dihydroxyacetone phosphate and 0.12 mM NADH in 100 mM triethanolamine/2.5 mM EDTA (all from Sigma-Aldrich). Enzyme kinetics was determined by measurement of absorbance at 340 nm. The enzyme activity of GPDH was detected as decrease of the NADH concentration whereby 1 mU corresponds to the oxidation of 1 nmol NADH min−1. In order to relate GPDH activity to the cell number, the DNA content of the same lysates was determined as described above. For microscopic detection of cells with characteristic fat vacuoles, samples were taken on day 21 of culture, washed with PBS, fixed with 3.7% formaldehyde/PBS and incubated with Nile red (Sigma-Aldrich) and DAPI to stain fat droplets and nuclei. Microscopic analysis was performed applying a Zeiss cLSM 510 with excitation wavelengths of 561 nm (Nile red) and 405 nm (DAPI). Osteogenic differentiation was further verified by measurement of alkaline phosphatase (ALP) activity and by alizarin red staining of deposited mineral. ALP activity was determined as described [27]. In brief, samples were taken on day 1 and 21 of culture, washed with PBS, frozen, homogenized and incubated in lysis buffer (1% (m/v) Triton X-100 in PBS). 20 μl of each cell lysate were mixed with 100 μl substrate solution (1 mg ml−1 p-nitrophenylphosphate (Sigma-Aldrich) in 0.1 M diethanolamine/1% (m/v) Triton X-100/ 1 mM MgCl2, pH 9.8). After incubation at 37 °C for 30 min, the enzymatic reaction was stopped by addition of 1 M NaOH; p-nitrophenolate (pNp) formation was quantified by absorbance measurement at 405 nm. The amount of pNp produced by the cell lysate was calculated using a p-nitrophenol calibration line and related to the cell number in each sample (calculated from DNA content which was determined as described above) to express specific ALP activity (μmol pNp/30 min /106 cells). For alizarin red staining, samples were taken on day 21 of culture, washed with PBS, fixed with 3.7% formaldehyde/PBS and incubated in alizarin red staining solution (Sigma-Aldrich) for 2 min followed by incubation in distilled water. Microscopic analysis was performed applying the Leica M205 C equipped with a DFC295 camera.
2.6. Statistical analysis One-way analysis of variance was used to evaluate statistical significance; post-hoc analysis using the Tukey-method was used to determine multiple comparisons (Origin 8.5.0G, OriginLab, USA). Significant differences were assumed at p
