Optimizing the osteogenic potential of adult stem cells for skeletal regeneration


  • Jung Yul Lim,

    1. Department of Engineering Mechanics, University of Nebraska, Lincoln, Nebraska 68588
    2. The Graduate School of Dentistry, Kyung Hee University, Seoul, Korea
    Search for more papers by this author
  • Alayna E. Loiselle,

    1. Division of Musculoskeletal Sciences, Department of Orthopaedics and Rehabilitation, Center for Biomedical Devices and Functional Tissue Engineering, College of Medicine, Pennsylvania State University, Hershey, Pennsylvania 17033
    Search for more papers by this author
  • Jeong Soon Lee,

    1. Department of Engineering Mechanics, University of Nebraska, Lincoln, Nebraska 68588
    Search for more papers by this author
  • Yue Zhang,

    1. Division of Musculoskeletal Sciences, Department of Orthopaedics and Rehabilitation, Center for Biomedical Devices and Functional Tissue Engineering, College of Medicine, Pennsylvania State University, Hershey, Pennsylvania 17033
    Search for more papers by this author
  • Joshua D. Salvi,

    1. Weill Cornell Medical College, Cornell University, New York, New York 10021
    Search for more papers by this author
  • Henry J. Donahue

    Corresponding author
    1. Division of Musculoskeletal Sciences, Department of Orthopaedics and Rehabilitation, Center for Biomedical Devices and Functional Tissue Engineering, College of Medicine, Pennsylvania State University, Hershey, Pennsylvania 17033
    • Division of Musculoskeletal Sciences, Department of Orthopaedics and Rehabilitation, Center for Biomedical Devices and Functional Tissue Engineering, College of Medicine, Pennsylvania State University, Hershey, Pennsylvania 17033. T: 717-531-4819; F: 717-531-7583.
    Search for more papers by this author

  • Jung Yul Lim and Alayna E. Loiselle contributed equally to this study.


Adult stem cells, including mesenchymal stem cells, display plasticity in that they can differentiate toward various lineages including bone cells, cartilage cells, fat cells, and other types of connective tissue cells. However, it is not clear what factors direct adult stem cell lineage commitment and terminal differentiation. Emerging evidence suggests that extracellular physical cues have the potential to control stem cell lineage specification. In this perspective article, we review recent findings on biomaterial surface and mechanical signal regulation of stem cell differentiation. Specifically, we focus on stem cell response to substrate nanoscale topography and fluid flow induced shear stress and how these physical factors may regulate stem cell osteoblastic differentiation in vitro. © 2011 Orthopaedic Research Society Published by Wiley Periodicals, Inc. J Orthop Res 29:1627–1633, 2011

Tissue engineering approaches utilize stem cells and scaffolds for in vitro cell seeding and growth followed by in vivo transplantation. These constructs offer alternatives to address limitations of using bone grafts. Autografts have limited availability and result in donor-site morbidity and allografts can potentially be immunologically rejected. Allografts undergo extensive processing to limit disease transmission resulting in decreased osteoinductive capacity, which results in delayed bone formation, and incomplete union. Regenerating traumatized skeletal tissue using tissue engineering strategies lessens these problems and is emerging as a promising approach to treat bone defects.

To accomplish successful bone tissue engineering, the control of bone-forming cells, scaffolding biomaterials, and chemical and physical extracellular signals is critical. As regards bone-forming cells, stem cells of varying origin, including adult mesenchymal stem cells (MSC), embryonic stem cells (ES), and induced pluripotent stem cells (iPS) have been examined for their osteogenic potential. When a stem cell divides, each new cell has the potential to either remain a stem cell (self-renewal) or become another type of cell with a more lineage-specific function. ES cells have near-perfect self-renewal capability and differentiate toward all three derivatives of the primary germ layer. However, the use of ES cells is limited by political and ethical issues. Adult stem cells, including MSC, hematopoietic, neural, epithelial, skin, and fat-derived stem cells, have received increased attention for tissue engineering, as the manipulation of adult stem cells is technically less challenging, cost effective, and raises fewer ethical concerns. More importantly, adult stem cells can differentiate into many, but not all, cell lineages, including bone, cartilage, fat, and muscle cells. However, the mechanisms controlling MSC lineage specification are incompletely understood. Also, adult stem cells display limitations to their use in tissue engineering. For instance, in vitro expansion of MSCs, a prerequisite for obtaining a sufficient number of cells for in vivo use, often leads to a loss of differentiation potential.1, 2 Considering the limitations of adult stem cells in self-renewal and differentiation potential, the exploitation of extracellular factors in regulating adult stem cell function and fate is of significant interest.

iPS cells are obtained by the transfection of stem cell-associated genes (Oct-3/4, SOX2, c-Myc, Klf4, and Nanog) into nonpluripotent cells, including fibroblasts, using either transfection through retroviruses or using integration-free approaches.3–6 While iPS cells show great promise in regenerative medicine, iPS technology is at an early stage in development. Therefore, the focus of this perspective will be optimizing MSC behavior.

While considerable information is available regarding the regulation of MSC osteogenic potential by growth factors, cytokines, and hormones, little is known about the optimal chemical and physical extracellular conditions for MSC osteoblastic differentiation. Biomaterial characteristics including chemistry, surface energy, topography, and 3D morphology can work together to affect not only short-term cell adhesion, spreading, migration, and proliferation, but also longer-term lineage specification and differentiation. The regulation of osteoblastic cell function via modulating biomaterial properties has been examined extensively for implant applications,7, 8 while directing stem cell fate through biomaterial cues has been less widely examined. This perspective will focus on the potential role of nanotopography in regulating osteoblastic differentiation of osteoprogenitor and stem cells.

Another promising approach to regulating cell function and fate, especially as regards tissue engineering of mechanically functional tissues (bone, cartilage, muscle, etc.), is the use of mechanical signals. Matrix deformation in vivo, in response to mechanical load, provides cells with complex mechanical milieus consisting of fluid flow, stretch, and electrokinetic effects. Considerable data indicate that fluid flow and tensile stretch significantly affect the function and fate of bone cells (osteoblasts and osteocytes).9–11 However, relatively little is known regarding the effect of mechanical signals on stem cell behavior, including differentiation.


Substrate Topography Effects

Cell sensing of, and response to, anisotropic topographies has long been observed, for example, cell contact guidance. Recently, whether cells align on nanoscale ridges and grooves and, if so, how small the topographic size can become, until the cell can no longer sense it, has been examined.12, 13 From these reports, it is clear that anisotropic topographies induce drastic morphological changes in cellular, skeletal, and focal adhesion structure and may induce changes in gene expression.

On the other hand, isotropic topographies filled with randomly or uniformly distributed topographic features do not align cells but affect collective cell behavior as a function of topographic scale, distribution, gap, and surface coverage. Lithographic techniques can produce topographies within a range of 0.2–1 µm (photolithography) or possibly 10–20 nm (e-beam lithography) and have been used to assess cell response to well-defined topographic surface modifications.14

Cell response to nanotopography has typically been demonstrated by using polymer demixed film substrates. In polymer demixing, spin-casting of an immiscible polymer blend solution produces thin films covered with randomly distributed nanotopographic features via phase separation. Our group showed that osteoblastic cell adhesion and bone cell differentiation were significantly enhanced on 11 nm high polystyrene/polybromostyrene (PS/PBrS) demixed islands relative to 85 nm high islands or PS flat control.15 We also observed that 3–29 nm nanopits or nanoislands, produced by poly(L-lactic acid) (PLLA) and PS demixing, induced greater osteoblast adhesion relative to PLLA flat control.16 For topographies produced by nanophase surface replication, to mimic the roughness of intact bone at the range of about 32 nm, osteoblastic cell adhesion and proliferation were increased.17 These data suggest that osteoblastic cells react to isotropic nanotopographies by exhibiting enhanced cell adhesion and bone cell differentiation potential.

Inducing or accelerating stem cell differentiation toward the osteogenic lineage has also been examined. Oh et al.18 provided evidence for the interesting concept that controlling only nanotopography can achieve two important goals in stem cell research: (1) control of cell proliferation without differentiation and (2) activation of differentiation toward a specific lineage. They showed that specific scale nanotubes (30 nm dia.) promoted adhesion of hMSCs without noticeable differentiation, whereas larger (100 nm dia.) nanotubes induce cytoskeletal stress formation and directed hMSC osteogenesis. It was also shown that the use of nanotopography is sufficient to stimulate hMSCs to produce bone mineral even in the absence of osteogenic media.19, 20 We observed that culture on specific scale (10–20 nm high) nanoislands accelerates hMSC lineage commitment toward osteogenesis, as seen by increased alkaline phosphatase (AP) activity and mineralization (Fig. 1). We also found that accelerated differentiation of hMSCs on specific nanotopographies could be detected as decreased expression of stem cell surface markers (Fig. 2). That hMSCs lose stem cell surface markers (SSEA-4, CD73, CD105) on 11 nm high nanoislands indicates that these cells are no longer stem cells (or already committed to a specific fate, in this case osteogenesis). These data suggest that specific nanoscale topography potentiates hMSC differentiation toward the osteoblastic lineage.

Figure 1.

hMSC differentiation toward the osteoblastic lineage is enhanced when cells are cultured on specific nanoscale topographies. (a) Nanoisland topographies with varying island heights (12, 21, and 45 nm) produced by PLLA/PS (70/30 w/w) demixing. AFM height images of nanoislands and flat control surfaces are shown. (b) hMSCs were cultured on test surfaces in growth or osteogenic differentiation media and AP activity quantified (n = 4). (c) Mineralized area. (n = 6). *p < 0.05, **p < 0.01 compared with flat PLLA; ANOVA with post hoc test.

Figure 2.

Stem cell surface marker expression in hMSCs on nanotopographies and flat control. Human MSCs were cultured on PS/PBrS films with growth or osteogenic media for 7 days, harvested, and tagged with antibodies for SSEA-4, CD73, CD90, and CD105. Flow cytometry was then utilized to determine the percent of the cell population with a positive stem cell surface marker. ##p < 0.01 compared among nanoislands; **p < 0.01 compared with flat control. ANOVA with post hoc test.

Many studies,15–17 including our own, have focused on z-axis nanotopographic scale in the context of the roughness parameter (Ra), which is defined as the arithmetic mean vertical deviation from the mean height line. The other two (x- and y-) dimensional parameters, for example, topographic feature area, gap, etc., are also expected to affect cell behavior depending on their size relative to that of the cells. If the size of these topographic features are comparable to, or larger than the cell size, cells will perceive the features as individual surfaces. If the feature area or gap is sufficiently small, cells will not be confined between the two features but will be affected at reduced size scale.

While emerging data suggest that cells respond to specific nanotopographies, the mechanism by which this occurs is not known. One clue comes from the finding of increased development and organization of cytoskeletal elements, especially actin stress fibers, in cells on nanotopographies relative to those on flat surfaces.15, 20 Actin fibers contribute to the elastic properties of cells and disrupting actin decreases the elastic modulus of cells.21 Thus, nanoscale topography can, through an effect on actin formation and organization, regulate cell compliance, and potentially actin-focal adhesion-associated integrin signaling. We showed that integrin-mediated signaling, including activation of focal adhesion kinase (FAK), is regulated by nanotopography. Phosphorylation of FAK (pY397) was increased in osteoblasts on PLLA/PS (50/50 w/w) demixed 14–29 nm deep nanopits relative to cells on 45 nm pits or flat control.22 This suggests that the local concentration of specific integrin-binding ligands may be altered by nanotopography. Recently, our data were adopted to examine the role of protein adsorption in nanotopography regulation of cells. Fibronectin (FN) adsorption, quantified by radiolabeling, was greater on 14 nm deep nanopits relative to 29 and 45 nm deep pits,23 suggesting a positive correlation between increased protein adsorption, ligand concentration, and integrin signaling. The effectiveness of nanotopography itself, relative to diffusible or bound ligands, in regulating cell function has not as yet been examined. In terms of the mechanism, a recent study using proteomics revealed that many of the proteins affected by nanotopographic culture are associated with extracellular signal-regulated kinase (ERK),24 potentially downstream of FAK.

Although the mechanism of nanotopographic regulation of cell function is not fully understood, the data strongly suggest that identifying optimal nanotopographies for bio-interfaces would be beneficial for regenerative medicine. Since bone cells in vivo experience unique nanotopographic milieus from collagen and hydroxyapatite, mimicking these in vitro would provide biomimetic conditions similar to what bone-forming cells experience in vivo.

While the short-term nanotopography effects on cells appear to be direct, the longer-term effects may not be direct. For example, nanotopography control of integrin signaling has been demonstrated. However, it is unclear whether nanotopography regulates mineralization. Whether the nanotopography effect is potentially masked by extracellular matrix (ECM) protein secretion by the cells, and if so, the duration of such masking, are yet to be determined. Also, if the nanotopographies were fabricated using degradable polymeric materials, this may influence the longer-term effects on cells. The PLLA used in Figures 1 and 2, however, had degradation times longer than that necessary for stem cell osteogenesis (mineralization on day 24), suggesting that the nanotopography effect persists at least long enough for stem cell fate determination and osteogenic differentiation. We predict this would also be the case in vivo, but as yet this has not been examined.

Mechanical Stimulation Effects

The potential of osteoprogenitor and stem cells to proliferate and differentiate toward bone cells is determined to a large extent by the ability of the cells to perceive and respond to their mechanical extracellular milieus. Bone cells are exposed to mechanical signals from substrate deformation and resultant fluid flow through the lacuna–canalicular network. These mechanical signals mediate the anabolic effect of load on bone. The physiologically relevant strain and shear stress levels to which bone cells are exposed are about 10,000 µstrain and 1–20 dyne/cm2.25 If the ability of mechanical signals to regulate bone and stem cell activity is diminished, the resultant decrease in bone formation and stem cell osteogenesis could result in bone loss. On the other hand, increasing the ability of physical signals to regulate bone cell activity is an attractive strategy to enhance bone formation in vivo and optimize skeletal tissue engineering protocols in vitro. Understanding the mechanism by which physical signals regulate bone cell activity would provide insights into the development of strategies to sensitize cells to the anabolic effects of mechanical load.

Recent studies showed that tensile stretching has the potential to promote stem cell osteogenesis, but there is a discrepancy in the level of stretching sufficient for inducing stimulatory effects. MSCs showed increased AP activity and core-binding factor α1 (Cbfa-1) mRNA expression under stretching at or below 5% strain, whereas higher strain (10%, 15%) decreased AP activity and the expression of other osteogenic markers.26 Mechanical stretch at 2–10% strain induced increased expression of several key osteogenic transcription factors and genes (Runx2, Ets-1, etc.).27, 28 The increase of Cbfa-1 in hMSCs under 3% stretch increased AP activity and mineralization, and this was more pronounced when hMSCs were stretched on FN and laminin than on collagen and vitronectin.29 Stretching of stem cells under more physiologically relevant conditions, for example, stretch within 3D collagen scaffolds, was found to increase osteoblastic differentiation of hMSCs.30, 31

It has long been known that bone cells, including osteocytes and osteoblasts, display an anabolic response to fluid flow induced shear stress. Our group has shown that fluid flow activates multiple signaling pathways including cytosolic Ca2+ mobilization, NF-κB activation, increased expression of osteopontin mRNA, increased PGE2 release and several other pathways.32–36 The relative importance of fluid flow activation in these different pathways in mediating the anabolic response to load is currently under intense investigation.

Some of the above-mentioned mechanisms of bone cell detection of fluid flow are also applicable to stem cell sensing of fluid flow. Exposure of hMSCs to fluid flow induces adenosine-5'-triphosphate (ATP) release, intracellular Ca2+ mobilization, phosphorylation of ERK1/2, and the activation of the calcium-sensitive protein phosphatase calcineurin.37, 38 Activation of these signaling pathways combined to induce a robust increase in cell proliferation in hMSCs. Additionally, transport phenomena contribute to fluid flow effects as we have shown that alterations in flow rate at a constant peak shear stress, for example, modulated chemotransport under the same shear stress, were associated with changes in fluid flow-induced hMSC proliferation.39 It is also important to note that contrary data on fluid flow effects on stem cell proliferation have been reported. For instance, flow stimulated osteoblastic maturation, but not proliferation, of bone marrow stromal cells, in a PGE2-dependent manner.40 Despite this discrepancy in cell proliferation, it has been reported that fluid flow promotes mineralized matrix deposition and osteogenic gene expression in hMSCs.41–43 Interestingly, murine C3H10T1/2 MSCs displayed up-regulation of Runx2, Sox9, and PPARγ mRNA in response to oscillatory fluid flow, suggesting that fluid flow has the potential to regulate transcription factors involved in multiple lineage pathways including osteogenesis, chondrogenesis, and adipogenesis.44 Taken together, abundant data suggest that fluid flow stimuli induce the up-regulation of many mechanosensitive signaling molecules in bone and stem cells thus promoting osteogenic differentiation of these cells. However, how these pathways interact and are integrated is poorly understood.

The relative effect of direct cell stretching versus flow-induced shear stress in affecting bone and stem cell behavior is not precisely known. However, our earlier studies showed that substrate deformation may play less role in bone cell mechanotransduction than does fluid flow.45 We found that cell straining at levels equivalent to those induced by routine physical activity (∼0.5%, 1 Hz) did not induce a significant cytosolic Ca2+ response or increase in osteopontin mRNA levels, while fluid flow at levels equivalent to those induced by routine physical activity (20 dyne/cm2, 1 Hz) caused significant increases in both markers.45

Substrate and Mechanical Co-Stimulation

Since mechanical signals have stimulatory effects on bone and stem cells, increasing cellular mechanosensitivity to mechanical cues is an attractive strategy to enhance skeletal tissue regeneration. One potential way to achieve this goal is to pursue the synergistic effects of cell culture substrate and mechanical signals. We recently proposed that mechanosensitivity of cells in response to fluid flow will be enhanced when cells are in contact with nanotopographic substrates.46, 47 We showed that osteoblastic cells display greater cellular elastic modulus, as determined by atomic force microscopy (AFM) indentation, when cultured on 11 and 38 nm high nanoislands relative to cells on flat surfaces.46 Furthermore, hMSCs cultured on specific scale (10–20 nm high) nanoislands displayed a greater cytosolic Ca2+ response to fluid flow induced shear stress than cells cultured on flat control and other (45–80 nm high) nanoislands.47 We speculate that this is a result of the increased elastic modulus of cells on nanotopographies. Thus, nanotopographic culture has the potential to enhance reactivity of stem cells to mechanical signals. An analogous approach has been attempted using osteoblasts cultured on microscale roughnesses. MG63 osteoblast-like cells cultured on titanium with µm scale roughness were more sensitive to fluid flow cues than cells cultured on flat titanium.48, 49 Taken together, these data suggest that the synergistic effects of surface topography, both at the nano and micro scale, may be beneficial in developing tissue engineering protocols for skeletal regeneration.

Little is known regarding the combined effects of nanotopography, fluid flow, and diffusible or bound ligands on osteoblastic differentiation. However, that culture on nanotopography sensitizes cells to flow induced shear stress47 suggests at least a synergy between fluid flow and nanotopgraphy. As regards diffusible ligands, flow induced differential diffusion of proteins that are contained in serum, and not yet adsorbed onto surfaces, may indeed affect cell behavior as suggested by our previous study.39 Less is known regarding the interaction of bound ligands with flow and topography. It is possible that adsorbed ECM proteins, that contain cell-adhesive, integrin-binding arginine–glycine–aspartic acid (RGD) sequences, would be washed away by fluid flow, negating the effects of cell differentiation. This is another area on which to focus future work.


Emerging data from several different groups suggest nanotopographic surfaces, alone or combined with mechanical signals, have a strong potential to optimize stem cells for skeletal regeneration. However, many issues remain unresolved. The mechanism by which nanotopographies affect cell behavior, while likely involving integrins and FAK-mediated signaling, is poorly understood. Emerging data suggest that nanotopography can affect signaling pathways that overlap with those induced by fluid flow induced shear stress including integrins, FAK, and ERK.24, 37 However, there is a clear need for integrated genomic and proteomic approaches to define how these signaling pathways overlap.

Information on signaling pathways and relevant bioactive molecules (e.g., RGD motifs) would be useful in developing biomolecules that act in synergy with nanotopographies and which could be used to functionalize nanotopographic surfaces to increase their osteogenic potential. Another gap in knowledge is the exact nanotopographic features that are more stimulatory of stem cell osteogenic potential. While data from our laboratory suggest an optimal height of 10–30 nm, the relative importance of height, spacing, uniformity, and geometry is only beginning to emerge. This information is critical to developing the most efficacious biomaterials.


This study was financially supported by Penn State Biomaterials and Bionanotechnology Summer Institute (BBSI), AO Foundation Research Grant (Lim, S-10-7L), UNL Layman award (Lim), and NIH AR054937-03 (Donahue).