Introduction
Bone tissue regeneration is a complex multiscale process in which the evolution of tissues depends on biochemical, physical, geometrical, and mechanical factors. These factors mutually influence the biophysical response of other factors at different spatial and time scales (
Wu et al. 2010). The mechanical properties of hard tissues such as bone arise from the complex hierarchical structure (
Murugan and Ramakrishna 2005). Hence, it is important to consider the hierarchy of bone in the design of biomaterials that can heal/replace damaged tissues or substitute as an implant in critical size bone defects in the body. Although biological grafts have traditionally been used in bridging the critical size defects in bone, the risk of donor site morbidity, long healing time, persistent pain in autografts (
Fernyhough et al. 1992;
Kim et al. 2009;
Kurz et al. 1989), and the associated concerns of graft rejection and transmission of diseases in allografts (
Cartmell and Dunn 2004;
Keating et al. 2005) affect the desirable defect healing outcome. To overcome these disadvantages, biomaterials such as the tissue-engineered polymer scaffold can be considered viable candidates for implant in critical size bone defects treatment. However, the material selection, structural integrity, and mechanical properties evolution during degradation and tissue regeneration are still major challenges in designing a complex hierarchical polymer scaffold for bone tissue regeneration. Use of predictive methodologies for the design of biomaterials in general, and scaffolds in particular, is limited; most studies in the literature are testing and experiment based. In recent years, efforts have been made toward a simulation basis for the design of optimized polymer scaffolds for bone tissue regeneration (
Katti et al. 2015a,
2017;
Raabe et al. 2007;
Sun et al. 2013). Recent studies have elucidated the emergence and importance of multiscale modeling of mechano-biological aspects (
Hellmich et al. 2012;
Hellmich and Katti 2015;
Shim et al. 2012) that can also be useful for novel biomaterials design.
Hydroxyapatite (HAP) is a mineral component in the bone, and for this reason, it is widely used as a synthetic biomaterial to enhance osteoconductivity, biocompatibility, and nonimmunogenicity (
Khanna et al. 2011;
Li et al. 2013;
Wang et al. 2007). However, due to its hard and brittle nature, it is often combined with aliphatic polyesters such as polycaprolactone (PCL) as an inorganic filler to improve the mechanical strength of the polymer-based scaffolds (
Lyons et al. 2014;
Roeder et al. 2008). PCL is a hydrophobic, semicrystalline polymer with excellent blend-compatibility, rheological, and viscoelastic properties (
Nair and Laurencin 2007;
Okada 2002;
Woodruff and Hutmacher 2010). Many PCL-based drug delivery devices have been approved by the U.S. Food and Drug Administration (FDA) because of PCL’s good biocompatibility and biodegradability (
Dash and Konkimalla 2012;
Kumari et al. 2010;
Torchilin 2001). Because of these properties, PCL has become a candidate for the fabrication of tissue-engineered scaffolds (
Hollister 2005;
Shor et al. 2007;
Yoshimoto et al. 2003). Experimental and molecular dynamics (MD) simulation studies performed in this group indicate that the mechanical properties of PCL composites are highly influenced by the presence of HAP (
Bhowmik et al. 2009;
Verma et al. 2006,
2010). Montmorillonite (MMT) nanoclay was first used by the Toyota research group in the 1990s to improve the mechanical properties of the nylon-6 nanoclay composite (
Okada et al. 1990). The addition of small amounts of modified nanoclay has shown significant improvement of the biocompatibility, the biodegradability, and the thermal and mechanical properties of polymer nanocomposites (PCN) (
Katti et al. 2008,
2006;
Okada et al. 1990;
Picard et al. 2007). Molecular dynamics studies have shown that the interactions between MMTclay and organic modifiers control the mechanical properties of PCN (
Sikdar et al. 2008a). Previous studies by this group have led to the development of the altered phase theory of PCNs (
Sikdar et al. 2008b). The altered phase model developed by this group revealed that the altered polymer has a higher elastic modulus compared with the unaltered polymer in the PCN (
Sikdar et al. 2008b). The authors have also used engineered nanoclay to successfully synthesize a biocomposite system for bone tissue engineering applications (
Katti et al. 2008).
Given the increase in biological and mechanical properties of the biomaterials resulting from the PCL, HAP, and modified nanoclay, the authors have synthesized a novel PCL/in situ HAP nanoclay biomaterial system for bone tissue engineering applications (
Ambre et al. 2015). The interlayer spacing and biocompatibility of MMT nanoclay were enhanced by modifying it with unnatural aminovaleric acid (
Katti et al. 2005,
2010). The modified MMTclay was used to prepare a biomineralized HAP (in situ HAPclay), using a biomimetic strategy (
Ambre et al. 2011). Finally, PCL was introduced to synthesize PCL/in situ HAP nanoclay films and scaffolds. The mechanical properties of nanocomposite films and scaffolds were increased compared with pristine PCL because of the presence of the in situ HAP nanoclay (
Ambre et al. 2011). The cell culture study on this system indicated a flat and spherical morphology of human Mesenchymal Stem Cells (hMSCs) and the formation of the mineralized extracellular matrix (ECM) similar to the human bone (
Ambre et al. 2015;
Katti et al. 2015b).
To understand the underlying interaction mechanisms and mechanical properties of our composite system, the authors have developed representative molecular models of organically modified MMTclay (OMMT), OMMT-HAP, and OMMT-HAP-PCL (
Katti et al. 2015a;
Sharma et al. 2015). Further, a steered molecular dynamics (SMD) study was performed to understand the mechanical properties of the OMMT-HAP-PCL nanocomposite scaffold (
Sharma et al. 2015). The mechanical response of OMMT-HAP-PCL at the molecular level showed two distinct regions in the stress–strain curve. The results of these molecular dynamics simulations described the contribution of molecular interactions among different constituents of HAP, the modifier, PCL, and MMTclay and provided a detailed insight into a complex biomaterial system for bone tissue engineering applications.
Besides the role of molecular interactions in the mechanical properties evaluation, structural integrity also plays an important role in determining the biological responses and performances in bone tissue formation in polymeric scaffolds. Therefore, it is paramount to examine the stress–strain characteristics in polymeric scaffolds at the macrostructural level. For this purpose, the finite element (FE) method is commonly accepted as a numerical tool to predict the mechanical behavior of scaffolds. Among different FE modeling approaches, the microcomputed tomography (μCT) based FE method is widely used to simulate a representative three-dimensional (3D) scaffold microstructure (
Jaecques et al. 2004;
Lacroix et al. 2006;
Lohfeld et al. 2012;
Sandino et al. 2008). The μCT–FE technique provides the ability to analyze quantitatively the mechanical behavior of highly porous and complex materials such as trabecular bone, polymeric scaffolds, and metallic and polymeric foams (
Jones et al. 2006;
Lacroix et al. 2009;
Müller and Rüegsegger 1995;
Saadatfar et al. 2005;
Singh et al. 2010;
Verhulp et al. 2008). The effects of pore interconnectivity, scaffold porosity, fluid velocity, and pressure on the stress–strain distribution in the porous scaffolds (
Lacroix et al. 2009;
Sandino et al. 2008) have also been described. The local stress concentration and deformation processes in cellular solids (ceramics, metals, and polymers) have been calculated (
Petit et al. 2013). Analysis correlating the biomechanical interactions between the porous bioactive HAP ceramics and bone tissue regeneration indicates a strong relationship between bone ingrowth and the overall stiffness of scaffolds (
Ren et al. 2012). Recently, the mechanical responses of polycaprolactone/tricalcium phosphate (PCL/TCP) composite scaffolds have been predicted for an optimizing scaffold design (
Lohfeld et al. 2015). MD and SMD simulations provide a framework to analyze the mechanical responses in a scaffolds material system at the molecular level quantitatively; similarly, the FE modeling approach allows the simulation of the stress–strain behavior in scaffolds at the macroscopic level. However, few attempts have been made to connect quantitatively the mechanical behavior obtained at the molecular level to the macroscopic mechanics for designing the scaffolds for bone tissue engineering applications.
Chemical degradation via hydrolysis or enzyme-catalyzed hydrolysis is the most common degradation mechanism in polymer scaffolds. The degradation of PCL-based scaffolds under accelerated (alkaline) and simulated physiological conditions indicates that the PCL composite scaffolds are degraded via surface degradation; for long periods, the degradation in simulated conditions appears similar to bulk degradation (
Lam et al. 2008). A subsequent in vivo degradation study revealed that for a short duration (6 months), the degradation mechanism is mainly through the surface and that for a longer period (2 years), it was bulk degradation (
Lam et al. 2009). These studies indicate that PCL composites exhibit similar degradation mechanisms in in vivo and in vitro conditions. Recently, it was reported that the addition of nanoparticles such as HAP could be used to modulate the degradation rate of PCL composite scaffolds (
Díaz et al. 2014). The addition of biomineralized HAP nanoclay to the PCL influences the degradation behavior of the scaffold (
Ambre et al. 2015). It has been reported that Young’s modulus of composite scaffolds was increased when cells were seeded on the scaffolds (
Perron et al. 2009). In addition to the in vitro and in vivo studies, several mathematical models have been developed that link the changes in molecular weight and mass loss to the degradation of polymeric scaffolds (
Bawolin et al. 2010;
Heljak et al. 2014;
Rothstein et al. 2009).
Several in vivo studies have reported PCL-nanocomposite scaffold degradation behavior (
Pektok et al. 2008;
Yeo et al. 2008). The authors have reported that mineralization in the nanoclay-PCL scaffolds mimics biology by the formation of intercellular vesicles in the PCL/in situ HAPclay composite films seeded with hMSCs cells (
Katti et al. 2015a) similar to the formation observed in biological systems. The transmission electron microscopy (TEM) imaging results showed highly crystalline mineral inside the vesicles, indicating that the composite films provide a favorable bone-mimetic environment for new bone formation. Thus, in addition to understanding the tissue regeneration and polymer degradation for bone tissue engineering applications, it is important to incorporate and investigate the effects of these processes on the mechanical behavior of polymeric scaffolds.
The main goal of this study is to develop a multiscale mechanics approach to designing a three-dimensional polymer scaffold for bone tissue regeneration. To achieve this, (1) a μCT-FE model of a PCL/in situ HAPclay scaffold was created using the 3D reconstructed μCT images and materials properties obtained from this group’s previous SMD simulations (
Sharma et al. 2015), the predicted FE model was verified, and the mechanical behavior was analyzed against the experimental compression tests’ results; (2) the effect of the accelerated degradation on the mechanical properties of scaffolds was obtained experimentally; (3) the response of human osteoblast (hOB) cells to scaffolds and the cells’ effect on the mechanical properties of seeded scaffolds were found experimentally; and (4) analytical models were developed for mechanical properties prediction in accelerated-degradation scaffolds and hOB cells–seeded scaffolds. A new scaffold degradation modeling approach based on damage mechanics principles was developed in this work. The authors used the mathematical formulation for the disturbed state concept (DSC), previously developed for porous materials, solids, and interfaces, to model the mechanical behavior of the scaffolds in this work. The DSC was introduced by Desai et al. to express the behavior of the disturbed/deformed material through the reference of its undisturbed/undeformed (intact) state (
Desai 1974). This concept is widely used because of its simple numerical formulation and ability to predict the behavior of a wide range of materials (
Katti and Desai 1995;
Kwak et al. 2013,
2014).
The mechanical compression tests were performed to study the stress–strain responses in the PCL/in situ HAPclay scaffolds. Scanning electron microscopy (SEM) was used to study the hOB cells’ behavior and to capture the scaffold wall porosity. A water-soluble Tetrazolium salt-1 (WST-1) assay was performed to assess the viability of the hOB cells on the scaffolds.
The final step in achieving the study goal was to develop an integrative FE model by using the predicted FE model and the analytical models of the PCL/in situ HAPclay scaffolds for critical size bone defects implants.
Materials and Methods
Materials
Sodium-montmorillonite clay (SWy-2; Crook County, Wyoming) was purchased from the Clay Minerals Repository at the University of Missouri, Columbia, Missouri. Polycaprolactone (average molecular weight ), 1,4 dioxane (anhydrous, 99.8%), and 5-aminovaleric acid were purchased from Sigma-Aldrich (St. Louis, Missouri). Sodium phosphate () and calcium chloride () were purchased from J. T. Baker (Phillipsburg, New Jersey) and EM Sciences (Hatfield, Pennsylvania), respectively. Human osteoblast cells (hFOB 1.19 cell line CRL 11372) were obtained from American Type Culture Collection (ATCC) (Manassas, Virginia). The cell culture medium used in the hOB cells culture was made of fetal bovine serum (FBS) from ATCC, a G418 antibiotic from JR Scientific (Woodland, California), and HyQ Dulbecco’s modified eagle’s medium (DMEM)-12 (1:1) from Hyclone (Logan, Utah).
Preparation of Polycaprolactone (PCL)/In Situ Hydroxyapatite (HAP) Clay Scaffolds
PCL/in situ HAPclay scaffolds were prepared following the method described in the group’s previous study (
Ambre et al. 2015). In summary, 3.6 g of PCL was added to 40 mL of 1,4 dioxane and stirred in a beaker until all the polymer was dissolved. A 0.4 g of in situ HAPclay (the preparation procedure is described in
Ambre et al. 2011) was sonicated in 16 mL of 1,4 dioxane. The sonicated suspension was added to the PCL solution, and the resulting solution was further stirred for another two hours at room temperature. The final solution was transferred into the polypropylene (PP) centrifuge tubes, and these tubes were frozen overnight in an isopropyl alcohol bath at
. These cylindrical frozen PCL composite samples were removed from the tubes and transferred into absolute ethanol (
) for the solvent extraction. The absolute ethanol solution was replaced every 24 h. After four days, the frozen PCL composite samples were removed from the solution and dried at room temperature. Finally, these scaffolds were cut into the size 13 mm long and 13 mm in diameter, to ensure the uniform tissue growth in the cell culture well plate. Samples of the same size were used for consistency in all the studies presented in this paper.
Scanning Electron Microscopy Characterization
A JEOL JSM-6490LV [Japan Electron Optics Laboratory (JOEL), Peabody, Massachusetts] scanning electron microscope was used to study the microstructure of the PCL/in situ HAPclay scaffolds. The SEM imaging of the adhesion of the hOB cells on the PCL/in situ HAPclay scaffolds was performed with the same SEM instrument. The seeded scaffolds were washed with phosphate buffer saline to remove the cell culture medium. For fixing the live hOB cells, the seeded scaffolds were first treated with glutaraldehyde (2.5%) and then dehydrated in an ethanol series (10% v/v, 30% v/v, 50% v/v, 70% v/v, and 100%). Then, hexamethyldisilazane was used to replace the dried 100% ethanol. Before their viewing in the SEM, all these samples were sputter coated with gold and mounted on an SEM sample stub.
Microcomputed Tomography (μCT) of PCL/In Situ HAPclay Scaffolds
Three samples of the PCL/in situ HAPclay scaffolds were scanned using a microcomputed tomography technique. The samples were attached to a glass rod using carbon tape and placed inside a GE Phoenix (General Electric, Boston, Massachusetts) v|tome|x s X-ray computed tomography system (μCT) equipped with an 180 kV high power nanofocus X-ray tube xs|180nf and a high contrast GE DXR250RT (General Electric, Boston, Massachusetts) flat panel detector. One thousand projections of the sample were acquired at 60 kV and 350 mA, using a molybdenum target. The detector timing was 1,000 ms and the total acquisition time was 1 h and 6 min. The sample magnification was 7.28× with the voxel size 24.38 μm. The acquired images were reconstructed into a volume data set using
datos|x 3D computer tomography software version 2.2. A Diconde image series was then acquired from the reconstructed volume. Finally, the image analysis software
Mimics was used for the three-dimensional reconstruction of the scaffold [Fig.
1(a)]. Cylinders of 1.5-mm height and 1.5-mm diameter from each sample were modeled.
Finite Element Modeling of PCL/In Situ HAPclay Scaffolds
To evaluate the mechanical properties of the PCL/in situ HAPclay scaffolds, the finite element (FE) models were created from the assembly of 3D STereoLithography (STL) geometric models of scaffolds. The small cylinders obtained after 3D reconstructions were then converted into a 3D tetrahedral mesh using
3-Matic software and used as FE models. The typical 3D FE model contains 134,929 nodes and 352,316 tetrahedral elements of approximately 10–25-μm length. The scaffold models were imported into
Mentat software for the FE analysis simulations. An unconfined compressive loading condition was simulated to determine the mechanical properties of the scaffolds. Parallel plates were attached to the top and bottom surfaces of the scaffold. For the loading condition, the axial load was uniformly applied on the upper plate of the model, whereas the lower plate was constrained in all directions. The material parameters used in the FE analysis simulations were obtained from this group’s previous SMD study [Fig.
1(b)] (
Sharma et al. 2015). The loading conditions for the scaffold are shown in Fig.
1(c). Finally, the stress–strain responses in the meshed scaffold were analyzed with
Marc 13.0 FE analysis software. All FE analysis simulations were run on 256 processors (32 nodes; eight processors per node) at the Center for Computationally Assisted Science and Technology (CCAST) clusters at North Dakota State University (Fargo, North Dakota) using
Marc 13.0 software. Figs.
2(a and b) show the SEM images of the scaffolds and the porosities and microporosities in the scaffold walls.
Accelerated Degradation Studies in 0.1M NaOH
The mechanical degradation of the PCL/in situ HAPclay scaffolds was studied in alkaline conditions (0.1M NaOH). Scaffold samples were ultraviolet (UV) sterilized for 45 min and then immersed in 70% alcohol overnight. The samples were then washed with phosphate buffer saline (PBS) to prepare them for degradation studies. The PBS-washed scaffold samples were transferred to a 0.1M NaOH solution in separate glass vials. The scaffold samples in glass vials were placed at 37°C and 5% for degradation. Finally, the scaffolds were removed from the glass vials after each degradation period (1, 5, 7, 14, and 18 days) and washed with deionized water and dried at room temperature conditions.
The compressive mechanical tests were carried out on undegraded (0 days, control) and degraded (1, 5, 7, 14, and 18 days) PCL/in situ HAPclay composite scaffold samples using a material testing servo mechanical test frame (MTS 858, MTS Systems, Eden Prairie, Minnesota). Each sample was placed between the flat and smooth platens to perform the tests. A of constant deformation rate was applied for up to 10% strain to each test sample. The load and corresponding displacement data were recorded. The load-displacement data were used to construct the stress–strain response curves for each sample.
Human Osteoblast (hOB) Cell Culture Studies
Cell culture experiments for cell viability and mechanical properties studies of the PCL/in situ HAPclay scaffolds were carried out using hOB cells. For this purpose, UV sterilized, 70% alcohol-immersed, and PBS-washed scaffolds samples were used. The PBS-washed scaffold samples were first immersed in a cell culture medium and kept at 37°C and 5% in the incubator. After 24 h, hOB cells were seeded on each sample and 1 mL of cell culture medium was then added. This was followed by incubation of cell-seeded scaffolds at 37°C and 5% for 4, 7, 18, and 28 days. The cell culture medium was changed every 3 days.
The compressive mechanical tests of PCL/in situ HAPclay scaffolds seeded with hOB cells were carried out in wet conditions using a compression testing machine (Mechanical test system SATEC model 22 EMF, Instron, Norwood, Massachusetts). For wet testing, the scaffolds were directly removed from the cell culture medium (at 37°C) after 4, 7, 18, and 28 days and tested at room temperature. The samples were immediately placed between the flat and smooth platens of the mechanical testing machine and remained under wet condition during the entire experiments. Deformation control loading was applied at a constant rate of
up to 10% strain. The load-displacement results were analyzed using
Bluehill software v. 2.5 and used for stress–strain calculations.
Cell Proliferation Study Using WST-1 Assay
An investigation of the viability of the hOB cells grown in the PCL/in situ HAPclay scaffolds was studied using WST-1 assay (Roche Applied Science, Mannheim, Germany). The cell viability analysis was performed on the hOB cells-seeded scaffolds for 4, 7, 18, and 28 days following the manufacturer protocol (
Cell Proliferation Reagent WST-1 2011). The hOB cells-seeded scaffolds were removed from the culture medium and washed with PBS. Then, the washed scaffolds were placed in a 96-well plate filled with culture medium and WST-1 reagent and incubated at 37°C and 5%
. After 4 h of incubation, the scaffolds were removed from the solution and the absorbance of formazan-formed solution, which is directly related to the number of live cells, was read at 450 nm using a microplate spectrophotometer (Bio-Rad, Benchmark Plus, Bio-Rad Laboratories, Hercules, California).
Integrative Finite Element Modeling of PCL/In Situ HAPclay Scaffold of Critical Size
An integrative FE modeling was performed to analyze the effect of the time-dependent process of accelerated degradation and the hOB cell culture on the mechanical behavior of the implant PCL/in situ HAPclay scaffold in critical size bone defects. The length and diameter of the FE model were 4 and 2.82 cm, respectively. The FE model was constructed using 1.5-mm tetrahedral elements. The element length was the same as the size of the FE model of the PCL/in situ HAPclay. The total numbers of nodes and elements of the FE model were 7,133 and 36,037, respectively. The bottom and top surfaces were attached to plates. The bottom plate was constrained in all three directions, and a uniform axial load in the -direction was applied on the top plate. The time-dependent material parameters used in the FE simulations were obtained from the integration of the stress–strain response of the predicted FE model of the PCL/in situ HAPclay scaffolds and the proposed degradation function. Finally, the stress–strain responses were reported for the FE simulations of the implant scaffold in critical size bone defects. The FE model was carried out using the Marc/Mentat 13.0.