Osteoid-Mimicking Dense Collagen/Chitosan Hybrid Gels Florencia Chicatun,† Claudio E. Pedraza,‡ Chiara E. Ghezzi,† Benedetto Marelli,† Mari T. Kaartinen,‡ Marc D. McKee,‡,§ and Showan N. Nazhat*,† †
Department of Mining and Materials Engineering, Faculty of Engineering, ‡Faculty of Dentistry, Department of Anatomy and Cell Biology, Faculty of Medicine, McGill University, Montreal, Quebec, Canada H3A 2B2
bS Supporting Information ABSTRACT: Bone extracellular matrix (ECM) is a 3D network, composed of collagen type I and a number of other macromolecules, including glycosaminoglycans (GAGs), which stimulate signaling pathways that regulate osteoblast growth and diﬀerentiation. To model the ECM of bone for tissue regenerative approaches, dense collagen/chitosan (Coll/CTS) hybrid hydrogels were developed using diﬀerent proportions of CTS to mimic GAG components of the ECM. MC3T3-E1 mouse calvaria preosteoblasts were seeded within plastically compressed Coll/CTS hydrogels with solid content approaching that of native bone osteoid. Dense, cellular Coll/CTS hybrids were maintained for up to 8 weeks under either basal or osteogenic conditions. Higher CTS content signiﬁcantly increased gel resistance to collagenase degradation. The incorporation of CTS to collagen gels decreased the apparent tensile modulus from 1.82 to 0.33 MPa. In contrast, the compressive modulus of Coll/ CTS hybrids increased in direct proportion to CTS content exhibiting an increase from 23.50 to 55.25 kPa. CTS incorporation also led to an increase in scaﬀold resistance to cell-induced contraction. MC3T3-E1 viability, proliferation, and matrix remodeling capability (via matrix metalloproteinase expression) were maintained. Alkaline phosphatase activity was increased up to two-fold, and quantiﬁcation of phosphate mineral deposition was signiﬁcantly increased with CTS incorporation. Thus, dense Coll/CTS scaﬀolds provide osteoid-like models for the study of osteoblast diﬀerentiation and bone tissue engineering.
’ INTRODUCTION Bone tissue engineering (BTE) is an alternative regenerative approach that may overcome host rejection and sourcing problems associated with allografts and autografts, respectively.1 BTE is based on a 3D biomimetic scaﬀold that models the extracellular matrix (ECM) of bone by providing favorable environmental cues that support osteoblast attachment, proliferation, and diﬀerentiation. Reconstituted nanoﬁbrillar collagen gel scaﬀolds are biocompatible, biodegradable, and exhibit low antigenicity.2 However, these hydrogels are highly hydrated (>99% ﬂuid) and therefore lack control in structure and exhibit poor mechanical properties for tissue replacement applications. Plastic compression (PC) has recently been developed by Brown et al.3 to increase rapidly the collagen ﬁbrillar density (>10 wt %) by removing the excess ﬂuid that results from casting. This controllable process not only enables the production of collagen scaﬀolds with increased biomechanical properties but also allows for the organization of seeded cells in 3D by providing the space for their growth and diﬀerentiation without compromising their viability.4 The application of PC enabled the development of tissue-equivalent matrices that have been demonstrated to inﬂuence osteoblastic diﬀerentiation by the expression of markers such as alkaline phosphatase (ALP), bone sialoprotein, and osteopontin as well as collagen matrix remodeling (MMP-13).5 8 Along with abundant collagen type I, bone ECM is also composed of adhesive and mineral-binding proteins, polysaccharides such as glycosaminoglycans (GAGs), and proteoglycans, constituting 66, 33, and 1 wt % water, collagen, and proteoglycans, r 2011 American Chemical Society
respectively.9 Fibrillar proteins are known to provide mechanical and structural integrity to the ECM and regulate cell attachment and spreading.10 GAGs are unbranched disaccharides composed of N-acetylglucosamine or N-acetylgalactosamine that are responsible for growth factor binding, interactions with cellular receptors and adhesion proteins, and water molecular retention that confers tissue resistance to compression.11 Chitosan, a polysaccharide produced via the partial deacetylation of chitin, also possesses N-acetylglucosamine and glucosamine in its structure, and therefore, at least conceptually, can be used as a substitute for GAGs. Therefore, the incorporation of CTS into collagen scaﬀolds better mimics the native composition of bone osteoid for osteoblast diﬀerentiation in hydrogel scaﬀolds for BTE. However, whereas collagen/GAG and collagen/chitosan (Coll/CTS) matrices have been developed as tissue replacements, these scaﬀolds commonly rely on cell seeding postfabrication as a consequence of freezedrying process.12,13 Whereas direct cell seeding within highly hydrated gels has been reported,14 16 these scaﬀolds do not fully mimic the osteoid in terms of solid volume fraction and mechanical properties and have limitations associated with extensive cellinduced contraction. In an eﬀort to overcome these limitations, this study reports on dense Coll/CTS hybrid hydrogel scaﬀolds produced by PC to Received: April 18, 2011 Revised: June 7, 2011 Published: June 10, 2011 2946
dx.doi.org/10.1021/bm200528z | Biomacromolecules 2011, 12, 2946–2956
Biomacromolecules mimic bone osteoid, where cell-seeding is part of the processing route. The eﬀect of CTS incorporation on modulating MC3T3E1 osteoblastic cell-based remodeling of the scaﬀolds was related to their morphological and mechanical properties. In addition, osteoblastic diﬀerentiation and matrix mineralization were evaluated for up to 8 weeks. In this regard, it is expected that dense Coll/CTS hydrogels reﬂect natural bone ECM macromolecular components and qualities, oﬀering an in vitro model for constructing tissue-like structures with biomimetic function.
’ MATERIALS AND METHODS Preparation and Characterization of Acellular Scaffolds. Sterile, rat-tail tendon-derived type I collagen (2.11 mg/mL of protein in 0.6% acetic acid, First Link) and ultrapure chitosan powder (79.8% deacetylated, molecular weight 328 kDa, Ultrasan, BioSyntech) were combined to prepare Coll/CTS hybrids. Coll/CTS scaffolds of relative compositions of 2:1 and 1:1 (w/w) were prepared and compared with collagen gel scaffolds. CTS (13.5 and 27 mg for 2:1 and 1:1, respectively) was dissolved in acetic acid (2.2 mL/0.1 M) at 4 °C and stirred overnight. Coll/CTS solutions were prepared by the addition of 12.8 mL collagen solution under gentle stirring on ice. Coll/CTS selfassembly was achieved by mixing the solution with X10 minimum essential medium (MEM) at a ratio of 4:1 and neutralized with 5 M NaOH. Gels were cast in either rectangular molds (43 50 7 mm3) or four-well plates (A = 2 cm2) by pouring 15 mL/mold or 0.9 mL/well, respectively, and allowing to set at 37 °C in a 5% CO2 incubator for 30 min. After setting, gels were removed from the mold, and dense scaﬀolds were produced by a standardized PC protocol.3 In brief, highly hydrated gels were placed on a stack of blotting paper, nylon, and metal meshes to allow for capillary ﬂuid ﬂow in combination with an unconﬁned compressive stress of 0.5 kN/m2 for 5 min to remove the excess casting ﬂuid. The eﬃcacy of processing in generating dense scaﬀolds was assessed gravimetrically by measuring the weight loss in the gels as a function of compression time (n = 3). In brief, as-prepared gels were weighed for their initial weight, and the weight loss attributable to PC was measured at 1 min intervals. The percentage weight loss and solid content in compressed scaﬀolds were also conﬁrmed by mass measurements after freezedrying (BenchTop K freeze-dryer, VirTis, n = 3). These data were used to calculate the increase solid weight % due to PC. Morphological and Structural Characterization. Scaffolds were morphologically assessed by scanning electron microscopy (FEGSEM Hitachi S-4700 microscope). Gels were fixed overnight at 4 °C in a solution of 4% paraformaldehyde and 2% glutaraldehyde in 0.1 M sodium cacodylate buffer. Fixation was followed by dehydration over 15 min in a series of graded ethanol solutions and critical-point drying. SEM was carried out on Au/Pd sputter-coated samples. Attenuated total reﬂectance Fourier-transform infrared spectroscopy (ATR-FTIR, model Spectrum 400, PerkinElmer Instruments) equipped with a ZnSe diamond-coated ATR crystal was used to analyze scaﬀold composition. ATR-FTIR was performed in transmittance mode at 4 cm 1 resolution over a range of 4000 650 cm 1 (16 scans). Spectra were baseline-corrected for comparison (Spectrum software, Perkin-Elmer). Mechanical Analyses. Quasi-static tensile and compressive mechanical tests were performed on as-prepared acellular dense hydrogel scaffolds using a Bose ElectroForce BioDynamic instrument equipped with a 20 N load cell. All tests were carried out in displacement control at 0.01 mm/sec. Measurements were conducted at room temperature while maintaining constant sample hydration using drops of distilled water (dH2O). Tensile tests were performed on cylindrical-shaped specimens (n = 4) prepared by rolling compressed rectangular sheets along their long axis. Spiral constructs were used because they facilitated better handling and gripping when hydrated as compared with sheets of between
100 and 300 μm in thickness. Specimen dimensions (lengths of 10 ( 2 mm and diameters of 1.5 ( 0.1, 1.8 ( 0.1, and 2.4 ( 0.2 mm for collagen, Coll/CTS 2:1 and 1:1, respectively) were measured using a digital calliper and confirmed by optical microscopy. Test specimens were gripped in small metal screw tension chucks in between silicon carbide paper.17,18 The break strength and strain values were defined as the stress and strain, respectively, at the corresponding maximum load. The average apparent modulus was computed from the slope of the linear phase subsequent to the toe region. Compression testing was performed on cylindrical specimens prepared by rolling compressed sheets along their short axis (n = 6 8) of 2.8 ( 0.2 mm in height and 2.6 ( 0.1, 3.5 ( 0.1, and 4.6 ( 0.2 mm in diameter for collagen and Coll/CTS 2:1 and 1:1, respectively. Tests were carried out using two parallel nonporous platens up to 80% strain. The compressive modulus of all samples was calculated from the slope of the initial linear region (