Biomed Microdevices DOI 10.1007/s10544-013-9826-0

Reprogramming hMSCs morphology with silicon/porous silicon geometric micro-patterns M. D. Ynsa & Z. Y. Dang & M. Manso-Silvan & J. Song & S. Azimi & J. F. Wu & H. D. Liang & V. Torres-Costa & E. Punzon-Quijorna & M. B. H. Breese & J. P. Garcia-Ruiz

# Springer Science+Business Media New York 2013

Abstract Geometric micro-patterned surfaces of silicon combined with porous silicon (Si/PSi) have been manufactured to study the behaviour of human Mesenchymal Stem Cells (hMSCs). These micro-patterns consist of regular silicon hexagons surrounded by spaced columns of silicon equilateral triangles separated by PSi. The results show that, at an early culture stage, the hMSCs resemble quiescent cells on the central hexagons with centered nuclei and actin/β-catenin and a microtubules network denoting cell adhesion. After 2 days, hMSCs adapted their morphology and cytoskeleton proteins from cell-cell dominant interactions at the center of the hexagonal surface. This was followed by an intermediate zone with some external actin fibres/β-catenin interactions and an outer zone where the dominant interactions are cell-silicon. Cells move into silicon columns to divide, migrate and communicate. Furthermore, results show that Runx2 and vitamin D receptors, both specific transcription factors for skeleton-derived cells, are expressed in cells grown on micropatterned silicon under all observed circumstances. On the other hand, non-phenotypic M. D. Ynsa (*) : M. Manso-Silvan : V. Torres-Costa : E. Punzon-Quijorna Department of Applied Physics, Universidad Autónoma de Madrid, Campus de Cantoblanco, 28049 Madrid, Spain e-mail: [email protected] M. D. Ynsa : E. Punzon-Quijorna Centro de Micro-Análisis de Materiales (CMAM), Universidad Autónoma de Madrid, Campus de Cantoblanco Edif. 22, Faraday 3, 28049 Madrid, Spain Z. Y. Dang : J. Song : S. Azimi : J. F. Wu : H. D. Liang : M. B. H. Breese Centre for Ion Beam Applications (CIBA), Department of Physics, National University of Singapore, Singapore 117542, Singapore J. P. Garcia-Ruiz Molecular Biology Department, Universidad Autónoma de Madrid, Campus de Cantoblanco, 28049 Madrid, Spain

alterations are under cell growth and migration on Si/PSi substrates. The former consideration strongly supports the use of micro-patterned silicon surfaces to address pending questions about the mechanisms of human bone biogenesis/pathogenesis and the study of bone scaffolds. Keywords Human Mesenchymal Stem Cells . Silicon/porous silicon micro-pattern . Actin . ß-catenin . Runx2 receptor . Vitamin D receptor

1 Introduction Cell-cell and cell-extracellular matrix (ECM) adhesions are required in mechano-transduction of external physical cues into intracellular signaling, tissue development and biology. Adhesion molecules are engaged in intercellular bridges governing cell shape as well as translating the mechanical tension and orienting intercellular junctions. The contribution of cell-cell and cell-ECM adhesions is still poorly understood although it is fundamental in ECM; the dynamic scaffold is actively remodelled during morphogenesis, cell migration and cell polarization (Guilak et al. 2009; Rozario and DeSimone 2010). The interactions between progenitors and the material surface should be explored in a bench-top environment before recommending any further applications. Cells recognize ECM geometry to regulate intracellular architecture and provide spatial information for cell polarization (Thery et al. 2006). Interestingly, an elegant study (Tseng et al. 2012) showed that ECM affects the degree of stability of intercellular junction positioning and the magnitude of intraand intercellular forces. The heterogeneity of the spatial organization of ECM induces anisotropic distribution of mechanical constraints in cells at the time they adapt to minimize both intra- and intercellular forces. This reinforces the importance of specific scaffolds in cellular related therapies of specific

Biomed Microdevices

degenerative diseases. However the contribution of cell-cell, cell-ECM in a specific niche area is starting to be studied. Increasing data suggest that mesenchymal stem cells (MSCs) reside in a perivascular niche since they have a well-recognized utility for tissue engineering, bone and cartilage repair (Caplan and Dennis 2006). MSCs show multi-lineage potential, strong immunomodulatory activities and high proliferative capacities in 2-dimensional (2D) conventional tissue culture. However, the expansion of MSCs needs to be improved to avoid senescence or tumorigenic transition. 3D microspheres show a large surface for cells and their use in a reactor system resulted in an improved method for MSCs culture (King and Miller 2007; dos Santos et al. 2011). However, altering culture conditions from 2D to 3D resulted in major cellular cytoskeleton remodeling conditions that affected lineage commitment of stem cells (Tseng et al. 2011, 2012; Thery 2010; Xu et al. 2011). Thus, micropatterning methods are needed in biomedical research of biomaterials before their clinical use and to investigate cell physiology, multi-cellular morphogenesis or mechanical properties of micro-environmental impact. In this work we show the response of hMSCs to two dimensional (2D) micropatterned Si/PSi surfaces consisting of a central Si hexagon surrounded by columns of Si triangles of different sizes. These micro-patterns were precisely fabricated by, up to now, non-conventional lithography method for biological purposes. P-type bulk Si wafers were selectively irradiated with a high energy ion beam. The ion-induced defects locally reduce the free carrier density, hence increase the resistivity of the irradiated silicon regions. This reduces the rate of porous silicon (PSi) formation during subsequent electrochemical anodization. Higher fluence irradiation can act as an etch stop for PSi formation. For a review of recent work with this process, see Ref. (Azimi et al. 2012).

2 Material and methods 2.1 Preparation of PSi A p-type bulk Si wafer (525 μm thickness) with a resistivity of 0.4 Ω·cm was coated with hexamethyldisilazane (HMDS) by vapor priming to create adhesion between the sample and photoresist. A 6 μm thick photoresist, AZ9260, was spin coated on to the wafer and then pre-baked to 95 °C for 4 min. A chrome mask containing a pattern with a hexagon surrounded by columns of equilateral triangles with different sizes was fabricated by laser lithography and transferred to the photoresist by standard UV photolithography, resulting in a two-step (resist and silicon step) binary pattern on the photoresist. Samples were exposed to a 500 keV H2+(molecular hydrogen) ion beam which was uniformly distributed over an area of 2×2 cm2 using the nuclear microscope facility at the Centre for Ion Beam Applications (CIBA), National University of Singapore (Watt

et al. 2003) following the process described in Ref. (Mangaiyarkarasi et al. 2008). Molecular hydrogen was used rather than protons since, with a megavolt accelerator, better beam brightness is more easily attained at higher beam energies. 500 keV H2+ ions have a range of 4.0 and 2.4 μm in photoresist and silicon respectively, so the photoresist is thick enough to prevent irradiation induced-damage to the underlying silicon. A fluence of 1×1016 ions/cm2 was high enough to greatly reduce the subsequent rate of PSi formation in Si. In this way, no ioninduced defects were introduced at the un-exposed/un-irradiated areas and during subsequent electrochemical anodisation the PSi formation rate is not reduced in these areas compared to the irradiated regions. For a typical ion beam current of 500 nA the irradiation time for each sample was about 2 h. After irradiation, the photoresist layer was removed and the wafer was electrochemically anodized at a current density of 40 mA/cm2 in an electrolyte containing HF (48 %):water:ethanol in the ratio of 1:1:2. A Ga–In eutectic and copper wire was used to make an electrical contact to the unpolished back wafer surface and epoxy was used to protect the contact from the HF electrolyte. The resistivity of the p-type wafer and the anodization current density determine the pore size of the PSi regions as Lehmann has previously shown (Lehmann 2002). 2.2 Description of the patterns The Si/PSi patterns used to culture hMSCs consist of regular silicon hexagons surrounded on all six sides by spaced columns of silicon equilateral triangles, as shown schematically in Fig. 1a. The hexagons have side lengths of 5 mm and the triangle columns have a length of 2 mm. Three triangle sizes, with sides of 50, 25 and 10 μm, and two orientations, with a base or tip on the edges of the hexagon, were used, Fig. 1b shows the 50 μm triangles. All the triangles located on one edge have the same size and orientation. The Si triangle columns are separated by a 60 μm width of PSi regions (non-irradiated areas) in all cases (yellow regions in Fig. 1b). Figure 1c shows a low magnification SEM of the 50 μm Si/ PSi patterns, from which one observes a continuous Si surface between consecutive triangles. Figure 1d shows a higher magnification SEM of the porous surface adjacent to the triangles. The surface is pitted with pores of diameter of a few tens of nanometers. 2.3 Cell culture Two to four milliliters of human bone marrow samples from healthy donors were provided by Dr. Benjamin Fernández from the Fundación Jiménez-Diaz. The culture expansion of hMSCs was carried out as previously described (Lennon et al. 1996; Ogueta et al. 2002). Cells were plated and incubated with (LG)-DMEM plus10% FBS of selected batches to 80 % confluence. Cells were collected by treatment with 0.25 %

Biomed Microdevices Fig. 1 a Schematic of the Si/PSi patterns used to culture hMSCs. The hexagon has edges of 5 mm length and is surrounded by triangles forming columns of 2 mm length. b Optical image of Si triangle columns with size of 50 μm. The silicon triangle columns are separated by PSi regions (non-irradiated areas) and the distance between the triangle bases is 60 μm. c Low magnification SEM of the 50 μm Si/PSi patterns, d Higher magnification SEM of the porous surface adjacent to the triangles

Trypsin-EDTA and 15,000 were seeded on the central Si hexagon of Si/PSi micropatterns, and on control cover slips coated with 0.5 % gelatin (bovine skin, Sigma) in order to check the cell number and if the cells are correct. The micropatterned surfaces were UV irradiated, washed and equilibrated using 1 ml of PBS (Phosphate Buffered Saline), this process was repeated 4–6 times. 2.4 Immunofluorescence After cell culture, cells on Si/PSi and gelatin coated cover slips were processed as previously described (Javed et al. 2000; Romero-Prado et al. 2006). Briefly, for whole cell preparation, the cells were rinsed twice with ice-cold PBS and fixed in 3.7 % formaldehyde in PBS for 30 min at room temperature (RT), then washed in PBS. For cytoskeleton (CSK) preparations, the cells were incubated with 0.5% Triton in CSK buffer containing 10 mM pipes, pH 6.8, 3 mM MgCl2, 100 nM NaCl, 1 nM EGTA, 0.3 M sucrose, and 1 mM PMSF for 30 min on ice. For nuclear matrix preparations, DNA was digested. To this end, cell preparations were depleted of soluble proteinases as described above, and incubated with 1% NP-40, 0.5 % deoxycholate acid twice for 2 min. They were then washed with 0.3 mM imidazolein H2O, incubated at RT with DNase I (50 IU/ml) in CSK buffer, twice for 15 min, and cell preparations were treated with CSK containing 0.25 M (NH4)2SO4 and 2 M NaCl in CSK buffer for 5 min. The preparations were then fixed with 3.7 % formaldehyde. For the cytoskeleton,

nuclear matrix and NM-IF, the preparations were blocked with 1 % BSA-0.1 % Triton X-100 in PBS (PBSA) for 2 h at RT and exposed to different antibodies. Runx2, β-catenin and VDR antibodies (Santa Cruz Biotech, Europe) were used respectively at the concentration of 1:200, αtubulin antibody (Sigma, Spain) at the concentration of 1:2,000 and β-actin was detected Phalloidin-Alexa 488 or TRITC at the concentration of 1:100 and 1:500 (Microscopy Service of Molecular Biology Centre, CSIC-UAM, Spain). After primary antibody reactions, samples were washed four times with PBSA. Secondary antibodies were either Alexa 488 goat anti-rabbit or Alexa 594 goat anti-mouse (Molecular Probes, Eugene, OR). These antibodies were used at 1/500 dilution. When required, nuclei were stained with DAPI (CALBIOCHEM) for 5 min and then washed three times with PBS 0.1 % Triton X-100. Immune stains were observed with an inverted IX81 Olympus with a DP72 digital camera. Where required, preparations were observed with a confocal laser scanning microscope and multi-photon system LSM710 which was coupled to an inverted Zeiss Axio microscope.

3 Results 3.1 Si/PSi micropatterns analysed by hMSCs Micropatterned Si/PSi samples were used to study the adaptability of hMSCs to transitional surfaces. We first used a cytoskeletal study to identify the patterned features. It is well

Biomed Microdevices

documented that contrasts between properties of biomaterial surfaces determine cell/surface interactions (Lipski et al. 2007). The surface features were visualized by microphotography with DAPI overexposure. In this manner, a map was made to localize cell position within each micropatterned substrate. After this localization step, we analyzed cytoskeleton proteins, β-actin, β-catenin or α-tubulin on different areas of the Si/PSi micropatterns. The β-actin (green)/β-catenin (red) microphotography belonging to hMSCs on control gelatin can be observed in Fig. 2(a, a1), showing that it spreads in all directions with actin stress fibers stemming from the cell center. The actin structures frequently co-localize with β-catenin and combine in the observed yellow cellular edges. In the Si/PSi samples, three different situations can be distinguished: (i); the central Si hexagon (b), (ii); transition area from the hexagon to the Si/PSi columns (b1, b2) and, (iii); the triangle columns far from the hexagonal part (b3). In each of the three stages, the actin cytoskeleton presents a different shape, with an increasing tendency to appear axially compressed. These morphologies correspond closely to those described by Thery M. (2010) for MSCs in adhesion, proliferation, contractility and migration. In addition, we observed in our samples that when α-tubulin was determined in both, the central Si hexagon (c) and the Si/PSi columns (c1), microtubules followed the actin shape. Furthermore, cells behave differently at the different column sizes (Fig. 3). In the 50 μm triangle columns, cells adherence can be either transversal or the cells can align in columns with two or three rows of cells, as in Fig. 3a. hMSCs form lines and migrate on the 25 μm triangle columns. On the 10 μm columns, hMSCs present an enhanced migration state prioritizing a polarization along the column axis and suffering a reduced cellular volume. In addition, it is observed that there are cells with actin/ catenin co-localization around the cell edge. This fact is very interesting since β-catenin mediates many cell processes including the control of cell metabolism. In particular, catenin regulates the transcriptional factor FoxO1 in osteoblasts to control glucose homeostasis (Kode et al. 2012). These results show the value in using Si/PSi geometric micropatterns as supports for the evaluation of skeletal cells. On one hand, they mediate in hMSCs morphology, from an adhesive to a migration state in 1D surface columns. It is relevant to outline that previous works support that 1D migration can be extrapolated to migration in 3D scaffolds (MironMendoza et al. 2008; Friedl et al. 2012; Doyle et al. 2009). On the other hand, the Si/PSi micropatterns are compatible with the glucose homeostasis as in skeletal bone, which is fundamental for the regenerative capacity of skeletal cells. 3.2 hMSCs on Si/PSi micropatterns express Runx2 and VDR transcriptional factors We next assessed whether hMSCs growing on micropatterned Si/PSi express two transcriptional factors: Runx2 and the

Fig. 2 Effect of Si/PSi micropatterned surfaces on hMSCs cytoskeleton proteins by immunofluorescence. Evaluation of β-actin(green )/βcatenin(red)/nuclei(DAPI, in blue) on hMSCs on 0.5 % gelatin (a, a1) as control and on Si/PSi micropatterns; hexagonal-(b), hexagonal-columne proximal area (b1, b2), triangular columns far from the hexagonal part (b3). Evaluations of tubulin (green), hexagonal (c) and triangular columns far from the hexagonal part (c1). The scale bar represents 100 μm

receptor of vitamin D, which are involved in bone cell specification/differentiation (Romero-Prado et al. 2006; Montecino et al. 2008; Lieben and Carmeliet 2013). The results of this experiment are illustrated in Fig. 3. We recorded microphotographs of both DAPI and Runx2/VDR either on the central Si hexagon and on triangle columns of 10, 25 and 50 μm sizes. Results show that Runx2 (green labeling) is highly expressed in hMSCs and most of it remains associated

Biomed Microdevices Fig. 3 Evaluation of Runx2 (green) and VDR (red), both transcriptional factors involved in bone development and cell differentiation, on different regions of Si/PSi micropatterns. (h) Transcriptional nuclear factors on Si hexagonal area. Overexposed DAPI-nuclei image (blue left images) and Runx2/ VDR (right images) on: (a) 50 μm triangle columns, the top of this image corresponds to Si hexagonal area; (a/b) Columns with 50 μm (down) and 25 μm triangles (up); (b) 25 μm triangle columns; (b/c) Columns with 25 μm (up) and 10 μm triangles (down); (c) 10 μm triangle columns. The scale bar represents 100 μm

with proteins of CSK as reported for other transcription factors (Lai et al. 2011). In comparison, the receptor of VD can be observed mostly at the nuclei (red staining). As cells are growing, the interaction between both transcription factors was not expected (they do interact during differentiation to activate bone specific genes). Interestingly, the association of Runx2 with cytosolic proteins has not been reported. Results confirm the observation derived from cytoskeleton studies. 50 μm triangle columns collect groups of cells which interact

by cell-cell (Fig. 3a) as well as making connections between columns of the same or different sizes (Fig. 3a/b). On the other hand, results obtained with 25 and 10 μm columns show that cell adhesion is mostly individual with a significant reduction of their volume as a result of the induced stretching. In addition, cells from hexagonal areas and Si/PSi columns were observed in a multi-photon fluorescence microscope to analyze any interactions between Runx2 and VDR. Results are shown in Fig. 4a (cells on hexagon areas) and in Fig. 4b

Biomed Microdevices

encouraging in thinking about this biomaterial for new strategies to grow and direct cells towards tissue damaged regions.

4 Discussion This work, in agreement with the previous studies of our group (Munoz-Noval et al. 2012; Torres-Costa et al. 2012), reinforces that Si/PSi micropatterns are excellent for studies dealing with bone-marrow hMSC growth and migration. However, there are some discrepancies. The present study shows that hMSCs preferentially localize on Si columns independent of their size (Fig. 3), avoiding PSi areas, whereas in previous work the response of hMSCs depended on the Si/PSi pattern width. A possible reason to understand this discrepancy is the higher resistivity of the bulk Si wafer used in this study, 0.4 Ω·cm versus 0.05–0.1 Ω·cm used previously, where the porosity was around 65 %. There is a clear correlation between p-type wafer resistivity on the anodization current density and the porosity of the material (see Fig. 5), data taken from (Lehmann 2002). According to Fig. 5 the porosity of PSi regions of the present work is around 75 % for a current density of 40 mA/cm2 (see star). Furthermore, the wafer resistivity is within the regime which produces microoporous silicon (Lehmann 2002), consistent with a high photoluminescence intensity produced from it (Azimi et al. 2013). This porosity is higher than of 65 % in the former studies (Munoz-Noval et al. 2012; Torres-Costa et al. 2012) where mesoporous silicon was preferentially formed. On the other hand, neuroblastoma cells displayed different cell adhesion properties, depending on the pore size of the PSi substrate (Gentile et al. 2012; Whitehead et al. 2008; Fan et al.

Fig. 4 Runx2 and VDR from hMSCs grown on Si/PSi micropatterned surfaces. Runx2 (green), VDR (red), Nuclei (DAPI) in blue were determined on cells spread on hexagonal (a) and on columns area (b). The scale bar represents 10 μm

(Si/PSi column micropatterns). It is observed that Runx2 is associated with microtubules since the green color extends to the end of the cellular membrane. Multi-photon analysis of VDR shows that, in addition to its preferential nuclear localization, this transcription factor can be detected perinuclearly, the cytosolic part including membranous protrusions resembling dendritic filopodia. Interestingly, when cells adhere to Si/PSi columns, it can be observed that nuclei are off-center and aligned by transcriptional factors that reveals the orientation of the cell polarity. These results also show that hMSCs either adhere on hexagonal surfaces or migrate on Si/PSi micropatterns, remaining as bone cells, since they express the main transcription factors needed for bone. These results are

Fig. 5 Porosity as a function of doping density or resistivity for p-type (100) Si wafer anodized in 1:1 ethanoic HF at different current densities (as indicated). Data obtained from Lehmann 2002, Fig. 6.9, page 111

Biomed Microdevices

2011; Wang et al. 2012; Sun et al. 2007; Khung et al. 2008) and consequently, the hMSCs behavior may also be affected by the pore size. New experiments should be performed to confirm or reject this point. Furthermore, the greatest complexity of the Si/ PSi micropatterns used in this study gives us additional information about the behavior of hMSCs in Si/PSi substrates. Triangular micropatterns, although fabricated in other materials, have been previously been used to rectify random motions of motile cells (Mahmud et al. 2009). In that work, the triangles were connected through partial superposition and the cells moved preferentially from the triangle base to the vertex. Maximal directional movement was observed when the width of the superposed triangles was around 10 to 20 μm. However, in our micropatterns, the cells do not seem to have a preferential orientation. Even if the connections of the triangles built in our work are much smaller than 10 μm, the cells which initially seeded in the Si hexagonal part migrated to the Si triangle columns regardless of the triangle orientation. Although our results show that hMSCs prefer to adhere to Si rather than to a PSi substrate, the suitability of PSi as a biomaterial has been reported in the literature and the results have demonstrated that PSi films can promote cell adhesion and viability (Low et al. 2006; Bayliss et al. 2000). Thus, the biocompatibility of PSi could benefit the migration of the cells even in geometries less favorable to the cell migration in other materials. In order to better understand the cellular behavior in Si/PSi substrates, new geometric patterns should be tested and its fabrication using high energy ions is the best option due to its versatility and precision. The present results reveal that our Si/PSi patterns facilitate cell-cell interactions. Cells can also migrate, maintaining the expression of main bone transcriptional factors. Consequently, this work supports the use of Si/PSi for specific scaffold strategies that are needed in bone-aged degenerative diseases, trauma and/or genetic pathologies. Furthermore, considering recent results demonstrating the contribution of bone to the control of energy homeostasis, the numbers of pathologies associated with bone are extended to diabetes or metabolic syndrome (Kode et al. 2012). The understanding of cellular and molecular mechanisms by which a Si/PSi surface induces proliferation, migration, metabolic changes or orientation are of general importance. In this work, bone marrow-derived hMSCs showed that Si/PSi is a suitable material on which they can adhere and migrate using the cell growth conditioning medium. The results fit in well with those showing that silica nanoparticles influence cytoskeletal organization and function, or how neuroblastoma cells respond to surface topography of continuous porous silicon gradients (Lipski et al. 2007; Khung et al. 2008). In addition, this work shows that Si/PSi micropatterned surfaces are a substrate where human bone progenitor cells attach and make cell-cell and cell-Si/PSi adhesion, and can migrate in 1D. It is suggested that columns of Si/PSi behave as

microcarriers for hMSCs to act as an in vivo or in vitro bioreactor system to replace 2D-monolayer culture (Tseng et al. 2012; King and Miller 2007; Tseng et al. 2011). hMSCs, as described in mesoderm tissue development, change on moving only a few micrometers from a large, unconstrained area of several square millimeters to a constrained column of micrometer-size triangles. They undergo a significant transformation of their morphology, adapting cytoskeleton and associated proteins but without losing their expression of transcription factors Runx2, vitamin D receptor involved in bone genetic specification. There are many questions that need to be explored, such as whether cells can undergo osteoblastogenesis or adipogenesis or interact with bone niche cells, determine the extension of a directed micropattern or the mechanism involved when cells go from 2D to 1D. However, the location change of β-catenin is curious and could be interpreted as effecter of Wnt, insulin and FOXO signaling cascades of Drosophila and other eukaryotic organisms involved in energetic metabolism homeostasis (Accili and Arden 2004; Jin and Liu 2008; Ip et al. 2012a, b). Hopefully, we expect that hMSCs and micropatterned Si/PSi will help in different strategies for specific unresolved clinical problems such as immune suppression, diabetes or bone degenerative diseases in the near future.

5 Conclusion We have demonstrated that Si/PSi micropatterns, fabricated by a high energy proton beam, are excellent for bone-marrow hMSCs to grow, migrate and communicate. Non-phenotypic alterations occur during cell growth and migration on these substrates. Our results strongly support the use of Si/PSi micropatterned substrates to address pending questions about mechanisms of human bone biogenesis/pathogenesis and the analysis of bone scaffolds.

References D. Accili, K.C. Arden, Cell 117, 421 (2004) S. Azimi, J. Song, Z.Y. Dang, H.D. Liang, M.B.H. Breese, J. Micromech. Microeng. (2012). doi:10.1088/0960-1317/22/11/113001 S. Azimi, J. Song, Z.Y. Dang, M.B.H. Breese, J. Appl. Phys. 114, 053517 (2013) S.C. Bayliss, L.D. Buckberry, P.J. Harris, M. Tobin, J. Porous. Mater. 7, 191 (2000) A.I. Caplan, J.E. Dennis, J. Cell. Biochem. 98, 1076 (2006) F. dos Santos, P.Z. Andrade, G. Eibes, C.L. da Silva, J.M. Cabral, Methods Mol. Biol. 698, 189 (2011) A.D. Doyle, F.W. Wang, K. Matsumoto, K.M. Yamada, J. Cell Biol. 184, 481 (2009) D. Fan, G.R. Akkaraju, E.F. Couch, L.T. Canham, J.L. Coffer, Nanoscale 3, 354 (2011) P. Friedl, E. Sahai, S. Weiss, K.M. Yamada, Nat. Rev. Mol. Cell Biol. 13, 743 (2012)

Biomed Microdevices F. Gentile, R. La Rocca, G. Marinaro, A. Nicastri, A. Toma, F. Paonessa, G. Cojoc, C. Liberale, F. Benfenati, E. di Fabrizio, P. Decuzzi, ACS Appl. Mater. Interfaces 4, 2903 (2012) F. Guilak, D.M. Cohen, B.T. Estes, J.M. Gimble, W. Liedtke, C.S. Chen, Cell Stem Cell 5, 17 (2009) W. Ip, Y.T. Chiang, T. Jin, Cell Biosci (2012a). doi:10.1186/2045-3701-2-28 W. Ip, W. Shao, Y.T. Chiang, T. Jin, Am J Physiol Endocrinol Metab (2012b). doi:10.1152/ajpendo.00249.2012 A. Javed, B. Guo, S. Hiebert, J.Y. Choi, J. Green, S.C. Zhao, M.A. Osborne, S. Stifani, J.L. Stein, J.B. Lian, A.J. van Wijnen, G.S. Stein, J. Cell Sci. 113, 2221 (2000) T. Jin, L. Liu, Mol. Endocrinol. 22, 2383 (2008) Y.L. Khung, G. Barritt, N.H. Voelcker, Exp. Cell Res. 314, 789 (2008) J.A. King, W.M. Miller, Curr. Opin. Chem. Biol. 11, 394 (2007) A. Kode, I. Mosialou, B.C. Silva, S. Joshi, M. Ferron, M.T. Rached, S. Kousteni, J. Biol. Chem. 287, 8757 (2012) W.T. Lai, V. Krishnappa, D.G. Phinney, Stem Cells 29, 1102 (2011) V. Lehmann, Electrochemistry of Silicon: Instrumentation, Science, Materials and Applications (Wiley-VCH, New York, 2002) D.P. Lennon, S.E. Haynesworth, S.P. Bruder, N. Jaiswal, A.I. Caplan, In Vitro Cell Dev. Biol. 32, 602 (1996) L. Lieben, G. Carmeliet, Bone 54, 237 (2013) A. Lipski, N. Bornfeld, B. Jurklies, Retina 27, 864 (2007) S.P. Low, K.A. Williams, L.T. Canham, N.H. Voelcker, Biomaterials 27, 4538 (2006) G. Mahmud, C.J. Campbell, K.J.M. Bishop, Y.A. Komarova, O. Chaga, S. Soh, S. Huda, K. Kandere-Grzybowska, B.A. Grzybowski, Nat. Phys. 5, 606 (2009) D. Mangaiyarkarasi, Y.S. Ow, M.B.H. Breese, V.L.S. Fuh, E. Tang Xioasong, Opt. Express 16, 12757 (2008) M. Miron-Mendoza, J. Seemann, F. Grinnell, Mol. Biol. Cell 19, 2051 (2008) M. Montecino, G.S. Stein, J.L. Stein, J.B. Lian, A.J. van Wijnen, L. Carvallo, S. Marcellini, F. Cruzat, G. Arriagada, Crit. Rev. Eukaryot. Gene Expr. 18, 163 (2008)

A. Munoz-Noval, V. Sanchez-Vaquero, E. Punzon-Quijorna, V. Torres-Costa, D. Gallach-Perez, L. Gonzalez-Mendez, I. Montero, R.J. Martin-Palma, A. Climent-Font, J.P. GarciaRuiz, M. Manso-Silvan, J. Biomed. Mater. Res. A 100 , 1615 (2012) S. Ogueta, J. Munoz, E. Obregon, E. Delgado-Baeza, J.P. Garcia-Ruiz, Mol. Cell. Endocrinol. 190, 51 (2002) M. Romero-Prado, C. Blazquez, C. Rodriguez-Navas, J. Munoz, I. Guerrero, E. Delgado-Baeza, J.P. Garcia-Ruiz, J. Cell. Biochem. 98, 1457 (2006) T. Rozario, D.W. DeSimone, Dev. Biol. 341, 126 (2010) W. Sun, J.E. Puzas, T.-J. Sheu, X. Liu, P.M. Fauchet, Adv. Mater. 19, 921 (2007) M. Thery, J. Cell Sci. 123, 4201 (2010) M. Thery, V. Racine, M. Piel, A. Pepin, A. Dimitrov, Y. Chen, J.B. Sibarita, M. Bornens, Proc. Natl. Acad. Sci. U. S. A. 103, 19771 (2006) V. Torres-Costa, G. Martinez-Munoz, V. Sanchez-Vaquero, A. MunozNoval, L. Gonzalez-Mendez, E. Punzon-Quijorna, D. GallachPerez, M. Manso-Silvan, A. Climent-Font, J.P. Garcia-Ruiz, R.J. Martin-Palma, Int. J. Nanomedicine 7, 623 (2012) Q. Tseng, I. Wang, E. Duchemin-Pelletier, A. Azioune, N. Carpi, J. Gao, O. Filhol, M. Piel, M. Thery, M. Balland, Lab Chip 11 , 2231 (2011) Q. Tseng, E. Duchemin-Pelletier, A. Deshiere, M. Balland, H. Guillou, O. Filhol, M. Thery, Proc. Natl. Acad. Sci. U. S. A. 109, 1506 (2012) P.Y. Wang, L.R. Clements, H. Thissen, A. Jane, W.-B. Tsai, N.H. Voelcker, Adv. Funct. Mater. 22, 3414 (2012) F. Watt, J.A. Van Kan, I. Rajta, A.A. Bettiol, T.F. Choo, M.B.H. Breese, T. Osipowicz, Nucl. Inst. Methods Phys. Res. B 210, 14 (2003) M.A. Whitehead, D. Fan, P. Mukherjee, G.R. Akkaraju, L.T. Canham, J.L. Coffer, Tissue Engineering Part A 14, 195 (2008) J. Xu, C. Chen, X. Jiang, R. Xu, D. Tambe, X.J. Zhang, L.N. Liu, B. Lan, K.Y. Cai, L.H. Deng, Biochem. Biophys. Res. Commun. 415, 591 (2011)