Send Orders of Reprints at
[email protected] Current Pharmaceutical Design, 2013, 19, 000-000
1
Osteoinductive Biomaterial Geometries for Bone Regenerative Engineering Tugba Ozdemir, Andrew M. Higgins and Justin L. Brown* Department of Bioengineering, The Pennsylvania State University, Hallowell Building, University Park, PA 16802 Abstract: Worldwide, more than 2.2 million patients undergo bone graft procedures annually. In each of these procedures an interface is formed between the host tissue and the graft material. Synthetic implants exhibit an interface with the host tissue and the formation of a homogenous interface consisting of bone and void of intervening soft tissue is desired (osseointegration); recent developments have highlighted the benefit of incorporating nanostructures at that interface. Autograft and allograft bone are frequently used for bone loss, nonunion fractures, and spinal fusions; however, both are plagued with complications either due to supply or inadequate graft properties. In contrast to bone tissue engineering, which uses a top-down approach to repair bone defects, bone regenerative engineering uses a bottomup approach focused on strategies incorporating stem cells, biomaterials, and growth factors alone or in combination to generate or regenerate bone tissue. Early constructs developed for bone regenerative engineering utilized polymeric microstructures, presenting surface features with characteristic dimensions similar to that of a cell (1 m – 1000 m). These microstructures were typically biodegradable and demonstrated an excellent ability to match the mechanics of native bone tissue. They were also osteoconductive—capable of promoting osteoblast growth. On the other hand, the osteoinductive abilities of these microstructures were lacking. Osteoinduction, or the ability to promote the progression of a preosteoblastic cell to a mature osteoblast, historically was achieved in two ways: via the addition of nanoscale ceramics to the microstructures or via an external stimulus such as the addition of bone morphogenetic proteins (BMPs). More recent developments in bone regenerative engineering have utilized polymeric nanostructures (less than 1 m) with characteristic dimensions an order of magnitude or more less than that of a cell to stimulate and drive an osteoinductive response in the absence of growth factors. Despite strong literature evidence supporting the nanostructures’ ability to be both osteoconductive and osteoinductive, there is still disparity regarding how nanostructures regulate the progression towards an osteoblastic phenotype. This review will explore unique micro- and nano-architectures, how they initiate osteoinductive signals through pathways similar to BMPs, and how these unique geometries can be translated to the clinic.
Keywords: Osteoinduction, biomaterials, substrate surface geometries. BIOMATERIALS IN BONE REGENERATION: Biomaterials have shaped the field of tissue engineering throughout last quarter century. Clinically, they have been used to replace missing tissue architecture (hip implants), provide mechanical support (spinal grafts), transport nutrients, and deliver growth factors to the extracellular microenvironment in a controlled manner (growth factor functionalized hyrdogels for stem cell engineering). In the laboratory settings, biomaterials are used in fundamental research to understand the underlying principles of cell behavior, such as the use of extracellular matrix mimicking biomaterial constructs to investigate the importance of cell-matrix interactions. Although the term biomaterial was developed around the principle of passive structural scaffolds there is a growing trend in biomaterials to design scaffold structures capable of inducing desired biological outcomes without the use of expensive growth factors or hormones. Future generations of biomaterials promise minimally invasive and cheaper solutions towards tissue regeneration [1]. Recently, trends within tissue engineering have resulted in the evolution of the field of regenerative engineering, which takes a fundamental look at how cells interact with materials and how those interactions can be used to achieve regeneration of the native tissue [2]. Bone regenerative engineering has grown rapidly and has produced numerous clinical products [3]. Procedures involving bone grafts are performed on over 2.2 million patients globally at a combined cost of more than $2.5 billion annually [4]. With a quickly growing market and increased demand of regenerative biomaterials due to an aging population, it is crucial to improve current efforts towards more affordable and increasingly biocompatible graft materials. Currently there are four types available: autografts (self*Address correspondence to this author at the Department of Bioengineering, The Pennsylvania State University, Hallowell Building, University Park, PA 16802; Tel/Fax: +1 (814) 865-5190 E-mail:
[email protected] 1381-6128/13 $58.00+.00
donated tissue), allografts (tissue donated from another individual of the same species), xenografts (tissue donated from an individual of another species), and synthetic grafts made from biocompatible materials. Regardless of the graft type, the graft should exhibit the following characteristics that have shown to be essential to successfully regenerate bone [5]: • osteoconductivity, an ability to facilitate the migration and proliferation of osteoblasts and progenitor cells • osteoinductivity, an ability to induce progenitor cells to differentiate down osteogenic lineages • osteointegrativity, an ability to incorporate with the surrounding native tissue • osteogenicity, an ability to form new bone by osteoblasts seeded within the construct Currently these procedures largely require donor material in the form of autografts and/or allografts. Autografts are the preferred material and exhibit all four ideal graft properties. However, limited availability and donor-site morbidity as high as 25% deter their use [3, 4]. Allografts and xenografts are plentiful and exhibit all characteristics of an ideal graft with the exception of osteogenicity. Risks of immune response and disease transmission, such as HIV (1: 1,500,000), hepatitis B (1: 100,000), and hepatitis C (1: 60,000), are major drawbacks to these materials [4]. Development of a synthetic biomaterial for bone graft scaffolds is advantageous since it alleviates the disadvantages present with auto- and allografts. Synthetic grafts currently only account for 10% of the bone graft market and only exhibit two of the characteristics of an ideal bone graft: osteointegrativity and osteoconductivity [6, 7]. The success of future synthetic grafts hinges on the ability to gain osteoinductive and osteogenic properties. Current efforts to generate an osteoinductive response in synthetic grafts largely focus on the addition of growth factors and nanoscale ceramics to the graft. However, there are emerging ef© 2013 Bentham Science Publishers
2 Current Pharmaceutical Design, 2013, Vol. 19, No. 00
forts to study how surface geometry can induce osteogenic differentiation of mesenchymal stem cells (MSCs). This review will discuss clinical problems facing musculoskeletal medicine, specific geometric cues that can affect the molecular mechanisms necessary for inducing BMP-like osteogenic differentiation, and current in vivo translations of certain osteoinductive geometries. Bone Diseases and Fracture Healing Bone diseases can progress over multiple years and have a significant impact on the patient’s quality of life. The primary types of bone disease are: osteoporosis, osteonecrosis, osteogenesis imperfect, bone tumors in which the bone tissue regresses, and osteopetrosis—overly hardened bone tissue which, if severe enough, can result in bone fractures and regression of surrounding bone [3]. In addition to fractures and bone loss due to trauma, regenerative engineering strategies have the potential to generate solutions for bone diseases requiring replacement of bone tissue. Most bone fractures are closed through indirect fracture healing. Indirect fracture healing consists of inflammation which leads to callus formation through a combination of intramembranous and endochondral ossification. The callus is eventually remodeled to become healthy bone tissue. Small fractures, which do not destroy the structural and mechanical integrity of the bone, can heal completely by indirect fracture healing. On the other hand, nonunion fractures often require surgery and extensive post-operative rehabilitation to facilitate the fracture healing process [8]. Even with combined pharmaceutical treatments and physical therapy, nonunion bone fractures may not heal completely and can be prone to breakage. Currently, autologuous bone grafting is the primary method used to treat nonunion fractures. Autografts are chosen as the preferred material because their matrix is rich with osteoprogenitor cells, bone matrix proteins, and bone inducing growth factors. These autografts are typically harvested from a small portion of trabecular bone from the donor’s hip and are subsequently implanted into the fractured region. Along with autografts, freeze dried allografts, xenografts, and bioactive glass are other clinical treatment strategies used to treat bone fractures. Currently, no biomaterial construct is capable of providing all the characteristics present in autologous bone grafts; however, significant progress has been made in understanding why autografts are successful and what design criteria should be considered in order to create a successful biomaterial scaffold. Firstly, autografts have the structure of native bone, which facilitates MSC adhesion, spreading, and proliferation. Secondly, autografts have the necessary hormones and growth factors at the optimal concentrations necessary to promote key intracellular signaling mechanisms that facilitate the induction of the MSCs down osteogenic lineages. Thirdly, autografts already contain extracellular matrix proteins and mineralized nodules (calcium phosphate and hydroxyapatite) that supply necessary signals for bone modeling and remodeling. Despite autologous bone being an excellent bone grafting material, the necessity of donor-site surgical complications and donor-site morbidity has motivated researchers to seek alternatives to autografts. Properties of Osteoinductive Materials Regardless of their origin, an osteoinductive material should successfully accomplish three individual steps to regenerate new bone tissue (Fig. 1) [6]. 1. MSC recruitment 2. MSC differentiation 3. Bone formation 1. MSC Recruitment In healthy adults, bones are constantly resorbed and regenerated as a normal part of homeostasis. The day-to-day maintenance of bone tissue requires MSCs. MSCs can be isolated from bone marrow, and the initially fibroblast-like colonies of MSCs are capable
Ozdemir et al.
of differentiation into adipocytes, chondrocytes, osteoblasts, and myoblasts as reported by Friedenstein forty years ago [9]. Relative to the other cells types present in bone marrow, MSCs are identified based on having the ability to adhere to a polystyrene dish, having a number of defined phenotypic markers, and being capable of differentiating into multiple tissue types. Specifically, the criteria for determining MSC by the International Society of Cellular Therapy are [10] ; • Adhere to polystyrene tissue culture dishes, • Stain positive for CD105, CD73 and CD90, • Stain negative for CD34, CD45,CD11a, CD19, and HLA-DR (MHC class II receptor ), • Differentiate into adipocytes, osteocytes, and chondrocytes under regular in vitro culture conditions. MSCs play a critical role in initiating the bone regeneration process through recruitment and migration from the surrounding periosteum and endosteum to sites of remodeling. Subsequent osteogenic differentiation of the MSCs leads to the development of new bone tissue. The exact mechanism that regulates the MSC migration to the site of remodeling is not well understood [11]. However, recent findings in leukocyte recruitment in fracture healing have shed light on MSC homing [12, 13]. During fracture healing, several growth factors, including bone morphogenetic proteins (BMPs) and platelet derived growth factors (PDGFs), have demonstrated an ability to rapidly stimulate MSC migration to the site of injury [13]. BMPs are pleitropic morphogens that induce several events including chemotaxis, regulation of differentiation, growth, and angiogenesis both endogenously and exogenously. Several in vitro and in vivo studies have shown that BMPs are able increase proliferation of MSCs and initiate differentiation into osteoblasts [14]. 2. MSC Differentiation into Osteoblasts The process of bone formation via differentiation of MSCs has numerous similarities with the bone formation that occurs during embryogenesis. Both organogenesis and fracture healing consist of callus formation through the combination of intramembranous and endochondral ossification. Fracture healing does have some differences from organogenesis: inflammation, scarcity of pluripotent stem cells, and elevation of the mechanical forces experienced by the healing tissue. There is a diverse set of signal transduction mechanisms that take place throughout fracture healing, and it is the interplay between different sets of signaling mechanisms that determine the success of fracture healing. Osteoinduction begins with the increased expression and DNA binding of the CBF -1 (Runx2) transcription factor [6]. Although necessary for osteogenic transformation of MSCs, this transcription factor is not sufficient for stimulating osteoblastic phenotype progression. There is a cascade of molecular events involving the increased activity of Wnt/ -catenin, MAPK, Indian hedgehog (Ihh) that regulates the temporal expression of bone matrix proteins (BMPs, ColIaI, ALP, PTH-R1, OSC, BSP, OCN, OPN) and osteogenic transcripton factors (Osx, Dlx5, Msx2) that promote osteoblastic phenotype progression and lead to the development of mature bone tissue. Among the aforementioned proteins, BMPs have demonstrated the most success in the induction of bone growth both in vivo and in vitro [15-19]. When injected into mice, BMPs are shown to rapidly induce endochondral ossification [16]. The process occurs through the crosstalk of BMPs and Ihh, resulting in the increased expression of Runx2. Earlier studies demonstrated that although osteogenesis occurs in BMP deficient mice (i.e. collar bone formation and mineralization of cartilage), no bone marrow cavity, trabecular bone, or cortical bone is formed [18]. BMPs are also potent activators of the transcription factors Dlx5 and Dlx3 that induce Osterix, another key transcription factor that regulates the expression of bone matrix
Osteoinductive Biomaterial Geometries for Bone Regenerative Engineering
Current Pharmaceutical Design, 2013, Vol. 19, No. 00
3
Fig. (1). The steps involved in osteoinduction.
proteins independent of Runx2. As seen in (Fig. 2) the activation of Dlx5 via BMP-2 occurs through a MAPK activation dependent mechanism [20]. The study suggests that there are two unique stages of BMP-2 activity: an early stage and a maturation stage. The early stage involves direct induction of Runx2 expression by BMP-2, whereas the maturation stage contains a Runx2 positive feedback mechanism that increases BMP-2 expression that subsequently leads to Dlx activation [21]. MAPK signaling is clearly an important element of BMP-mediated osteoinduction. Recent results from the Brown laboratory at the Pennsylvania State University have demonstrated the ability to activate unique MAPK signaling cascades through modulation of the characteristic dimension of a scaffold.
Fig. (2). BMP induced osteogenic transcripton regulation. Red arrows indicate the early stage Runx2 activation. Purple arrows indicate later stage positive feedback mechanism of Dlx5/3 activation and further DNA binding of Osterix.
Osteoblasts are contact dependent cells, and it is expected that some mechanism exists where adhesion to a surface communicates a change to intracellular BMP signaling. At this point in time, the
connection between adhesion-dependent mechanisms and BMPmediated osteoinduction is not well understood. A recent study has shown that BMPs induce cytoskeletal tension through RhoA activation, which ultimately leads to enhanced bone formation [19]. The design and investigation of new scaffold architectures that increase cytoskeletal tension may mimic the proposed mechanism of BMPinduced bone formation through RhoA. If successful, these scaffold architectures could decrease the need of using expensive growth factors. 3. Ectopic Bone Formation MSCs have the capacity to differentiate into bone, cartilage, adipose, neural and muscle tissue in vivo; therefore, a scaffold designed to induce bone formation must ensure that it delivers only the signals that promote bone formation in MSCs. Formation of ectopic bone by an implant is a frequently employed technique used to evaluate directed differentiation of MSCs down osteogenic lineages. There are two ways to investigate osteoinduction that results in ectopic bone formation: implant the scaffold subcutaneously and implant the scaffold into intramuscular defects [22]. Animal models evaluating the scaffold mediated formation of ectopic bone both subcutaneously and in intramuscular defects have been successfully conducted in mice [23], rats [24], dogs [25], and sheep [26]. The typical procedure involves the in vivo transplantation of the graft into the animal and subsequent investigation of bone formation over the next several weeks. Models evaluating ectopic bone formation have shown that BMP-2 and PDGF appear to be the most successful candidates of inducing bone formation as compared to other growth factors [6]. The combination of BMP with porous ceramics, chitosan hydrogels, or demineralized bone have all resulted in ectopic bone formation [6]. MECHANISMS DEPENDENT ON SURFACE TOPOGRAPHY: Influential work by Paul Weiss and Beatrice Garber in 1992 laid the foundation for how cell shape changes due to the physical structure of its substrate. The effects of substrate shape on mesenchymal cell growth and differentiation processes were also studied. Based upon their experiments, they were able to demonstrate cytosolic concentration changes as a cell elongates when it interacts with an extracellular fibrous matrix. The physical processes responsible for change in cell shape due to certain structural stimulations were also explained [27]. Currently, there are two main theories to elucidate these changes: topography sensing and mechanotransduction.
4 Current Pharmaceutical Design, 2013, Vol. 19, No. 00
Changes in Adhesion Patterns Through Topography Sensing There are several mechanisms that cells can use to sense the underlying surface [28]. The first mechanism employs adhesion receptors, specifically integrin molecules. Integrins form adhesive interactions with the underlying surface and generate clusters that eventually recruit a large number, over 150, of structural and signaling proteins. These clusters, commonly known as focal adhesion complexes, are capable of modulating intracellular signaling cascades [29]. Many vital processes, such as proliferation, differentiation, polarity, and motility, are also associated with integrins [3032]. There is considerable evidence on how micro- and nanostructured surfaces can affect integrin- induced FAK activity [33-35]. A second mechanism by which a cell can sense a surface is through an adhesion is related to cytoskeletal tension that results in membrane deformations in both the cell and nuclear membranes. In this mechanism, actin stress fibers within the cytosol deform the nuclear membrane. The stretching of the nuclear membrane results in the opening of nucleopores. When more nucleopores are fully opened, there is an increase in the transport of mRNA molecules into the cytosol, which increases protein translation. Similarly, the cytoskeletal tension can result in the stretching of the cell membrane. When the cell membrane stretches in response to certain surface geometries, membrane-bound calcium ion channels open and allow the transport of ions into the cytosol. This results in positive feedback with myosin light-chain kinase that results in further cellular contractility. Finally, cytoskeletal remodeling can also result in compression of the cell membrane, forming the basis of a fourth mechanism that leads to chromosomal translocation, epigenetic DNA, and nucleolus changes [28]. In addition to integrin-induced focal adhesion formation and cytoskeletal reorganization, growing evidence demonstrates that contact guidance is a fundamental mechanism by which cells sense their environment. Recently, Dalby et al. showed that cells induce signaling activity via receptors for activated C-kinase (RACK) and MAPK on grooved surfaces. When the groove spacing was reduced there was a decrease in adhesive contacts [36-38]. Microtubules have been reported to be the cytoskeletal component responsible for initiating contact guidance [38]. A final mechanism by which cells potentially sense certain topographies is via curvature sensing mechanisms. One hypothesis of curvature sensing is via the vesicle trafficking protein arfaptin. Arfaptin has an either-or binding relationship between the small GTPase Rac-1 and a curved membrane. Thus a sufficiently curved membrane with a high affinity for arfaptin can preferentially bind arfaptin, therefore causing arfaptin to release Rac-1. An increase in Rac-1 signaling may result from the increased soluble Rac-1 available [30]. There are also numerous other vesicle trafficking proteins capable of generating vesicles, so despite a current lack of evidence for the arfaptin mechanism, vesicle trafficking proteins may ultimately prove to be suitable geometry sensors. Mechanotransduction Translation of physical processes driven by surface geometry— such as the hydrodynamic and elastic forces generated by the cytoskeletal reorganization—into biomolecular interactions must exist. Considering that the medium contained in a cell is a nonhomogeneous mixture of proteins, any change in the cell’s physical processes would change the local concentration of these proteins and initiate signaling. Like topography sensing there are several hypotheses to explain the exact mechanism by which surface geometry regulates these biomolecular processes within a cell that govern cell growth, differentiation, death, and migration [39]. One widely accepted hypothesis is direct/self-induced mechanotransduction, which was proposed by Dalby et al. [40, 41]. They hypothesized that the changes in nuclear morphology and chromosomal locations are a result of the surface topography and that this mechanotransduction leads to transcriptional activation of specific genes.
Ozdemir et al.
This model integrates well with the model proposed by Ingber et al. [42]. Ingber suggested that tension-induced integration of actin and tubulin filaments activates local focal adhesion kinase (FAK) phosphorylation and signal transduction towards the nucleus via microtubules [42, 43]. Once a cell adheres to a surface with a compatible shape, the cell’s integrin molecules begin forming adhesive bonds and, depending on the strength of those bonds, integrin molecules begin clustering around one another to initiate the formation of focal adhesion complexes. Focal adhesion complexes are composed of structural and signaling proteins tied to polymerized actin filaments within the cytosol. These actin filaments generate enough tension to ensure the structural integrity of the cell when it is attached to a surface. Since a focal adhesion complex is a direct interaction between a cell and an interface, logic dictates that Ingber’s hypothesis regarding topography sensing is made possible via these focal adhesion complexes. Alternatively, Dalby et al. proposed that the coordinated stretching of actin stress fibers with focal adhesion complexes changes the position and proximity of certain signaling proteins. Cytoskeletal remodeling through focal adhesion complex formation thus leads to altered signaling cascade activity [39]. OSTEOINDUCTIVITY OF COMMON BIOMATERIAL ARCHITECTURES: The ability to sense and respond to physicochemical cues allows cells to adapt to a variety of different environments. Cells have demonstrated an ability to differentially respond to surface geometry, stiffness, chemistry, wettability, energy, and charge. In biomaterial scaffolds, multiple variables are often dependent on one another. This complicates the process of explaining and examining how each parameter alters the cellular response. On the other hand, unraveling the physicochemical mechanisms is necessary if osteoinductive scaffolds are to be designed without requiring the use of growth factors. Studies mimicking the geometry of the native extracellular microenvironment, such as pillars, spheres, and fibers, have been in use for more than a half-century and have shown significant success over biomaterials with flat surfaces [44-46]. The our group postulates that in order to improve the design of regenerative materials, how the target tissues characteristic niche, composed of unique extracellular matrix architectures and biochemical factors, influences intracellular protein localization and signaling must be better understood. Once understood the response to this niche can be translated into scaffold designs so that the cell’s response to the scaffold will mimic that of the natural architecture and biochemical factors. These bioanalogous architectures are necessary if biomimetic scaffolds are to become effective clinical tools. For instance, bone is a composite structure with type I collagen fiber bundles of 20-200 nm in diameter with embedded hydroxyapatite crystals [45, 46]. Instead of attempting to replicate the geometry explicitly, replicating the natural structure with cell adhesion motifs, such as RGD peptide sequences, could promote cell polarity by aligning the cells, modify cell motility, and induce contact guidance. This could also be used to manipulate MSC differentiation. In addition to inducing MSC differentiation, having an understanding of the effects of surface topography on fracture healing and bone diseases would be beneficial since the underlying extracellular matrix geometry has diverged from the norm in these situations [47]. The effects of biomaterial surface roughness and geometries such as islands, posts, ridges, fibers, and curved surfaces are being studied in an attempt achieve the ideal bioanalogous architecture. Surface Roughness and Islands Modifying the surface of titanium (Ti) implants were some of the first attempts to increase osteoinduction in an effort to improve bone healing [48-50]. Many independent studies have reported that sandblasted flat titanium surfaces had significantly different initial cell adhesion, spreading, differentiation, and expression of os-
Osteoinductive Biomaterial Geometries for Bone Regenerative Engineering
teogenic genes when compared to smooth Ti implants [50, 51]. Despite the ability of microroughened surfaces to promote osteoblast differentiation, the process is rudimentary and does little to isolate the specific element that is responsible for the improved cell response. Recent developments in micro- and nanofabrication have allowed researchers to investigate specific topographical parameters of characteristic geometries such as: posts, tubes, spheres, fibers and ridges. These fabrication techniques allow for better control over the surface geometry of the synthetic extracellular matrix geometry [46, 47, 52]. The effects of some of these bioanalogous osteoinductive surfaces will be discussed in later sections. A more controlled method of creating rough surfaces with nanotopographic structures is polymer demixing [36, 53]. Using randomly spread nanoislands of polystyrene and polybromostyrene Lim et al. showed cells exhibit increased alkaline phosphatase activity when grown on nanoislands as compared to flat surfaces. Alkaline phosphatase activity is indicative of an osteoinductive response. When experimenting with the size of the nanoislands, they discovered that smaller islands (11 nm) promoted increased cell spreading, well-developed actin stress fiber formation, and focal adhesion maturation when compared to islands with larger diameters (38 and 85 nm) [54, 55]. This data suggests that surface topographies with convex features exhibiting diameters on the order of 10 nm could direct either the differentiation of MSCs or the phenotype progression of osteoblasts, or both. Posts and Ridges An alternative to creating raised features or surface roughness is using posts and ridges. In 2010, Chavez et al. showed that cubicleshaped nanopores (~2nm) cut into the surface of a material demonstrated an improved capability of promoting cell adhesion and survival than cylindrical porous films. This suggests that cells are capable of sensing structures as small as 2nm in diameter [56]. Although the translation and mechanistic explanation of these results are lacking, the atomic level structural patterns underlying the nanopores could provide the basis for the next generation of scaffold designs. Hollow nanotubes are an additional example of an implant surface coating that presents a high aspect ratio,. Studied by Bauer et al. the nanotubes were oriented perpendicularly to the surface, such that the MSCs interacted with the cross-sectional face of the nanotubes and not along the length of the nanotubes. Results of their study showed that MSCs preferentially spread and differentiated on titanium oxide (TiO2) nanotubes smaller than 15 nm in diameter. Their proposed explanation for this phenomenon was that the 15 nm nanotubes were small enough to provide the necessary proximity of integrin receptors (10 nm) so that they could form clusters at the focal adhesion points. In the 100 nm nanotubes, the integrins were unable to be within 10 nm of one another, and were unable to achieve integrin clustering. Osteogenic genes, along with other adhesion proteins, were markedly increased on the 15 nm diameter TiO2 nanotubes [57]. However, Bauer did observe interesting phenomena on 100 nm nanotubes: an increase in the chondroblastic markers, such as type II collagen, and an increase in the cell-cell adhesive parameters, such as N-cadherin. Bauer and his colleagues further postulated that on BMP-2 coated 100 nm nanotubes, cells were unable to form adhesive contacts with the surface but were able to form cell-cell contacts. Through these cell-cell contacts, the cells were able to form a three-dimensional network that allowed them to differentiate towards a chodrogenic phenotype [58]. From a developmental standpoint, this is the logical result. MSCs will differentiate into chondrocytes in soft, gel-like three-dimensional environments; similarly, the 100 nm nanotubes provide such a limited surface area for the formation of focal adhesions that the cells had to bind to one another for support. Further testing found that these results were independent of material properties. BMP-2 coatings on
Current Pharmaceutical Design, 2013, Vol. 19, No. 00
5
small (15 nm) and large (100 nm) nanotubes promoted differentiation of MSCs into osteoblasts and chondrocytes respectively. Another geometry currently being used is aligned posts and ridges on a planar surface (Table 1) [59]. Lamers et al. generated a high throughput nanoridge array to study the effects of ridge aspect ratio on osteoblasts. Interestingly, they found that osteoblasts had the greatest increase in their motility at an aspect ratio of 1: 3. The authors explained this increase in motility based on the required spacing of integrin molecules in order to form focal adhesions (50 m pore diameter) calcium phosphate (CaP) and silicate-substituted calcium phosphate (SiCaP) implants with varying porosities were implanted intramuscularly into sheep. Twelve weeks after implantation, the implants were excised and evaluated. Both scaffold types exhibited increased bone formation with increasing porosity; however, the SiCaP scaffolds had greater bone formation than the CaP scaffolds at 30% porosity, the largest porosity tested, indicating a possible increase in osteoinduction. Chuenjitkuntaworn et al. used solvent casting and particulate leaching to create polycaprolactone/hydroxyapatite composite scaffolds with an average pore diameter of 480-500 m [79]. The implants were inserted into a calvarial defect in a mouse for six weeks. Once excised, the composite scaffold showed greater bone formation as compared to the neat polycaprolactone scaffold and control. In vitro testing demonstrated that the composite had increased levels of type I collagen and osteocalcin mRNA. Pore size has been shown to be of particular importance in regulating osteoinductivity [26]. Results of studies conducted by Huan et al. have suggested that osteoinductivity can be increased with microporostiy—pores
