Ann Anat 191 (2009) 136—144 Mesenchymal stem cell differentiation on microstructured poly (methyl methacrylate) substrates Elisabeth Engel a,b, , Elena Martı ´nez c , Chris A. Mills c , Miriam Funes c , Josep A. Planell a,b , Josep Samitier c,d a Bio/Non-Bio Interactions for Regenerative Medicine Group, Institute for Bioengineering of Catalonia (IBEC), Josep Samitier 1-5, 08028 Barcelona, Spain b Biomaterials, Biomechanics and Tissue Engineering Group, Department of Materials Science and Metallurgy, Universitat Polite`cnica de Catalunya, Avda. Diagonal 647, 08028 Barcelona, Spain c Nanobioengineering Group, Institute for Bioengineering of Catalonia (IBEC), Josep Samitier 1-5, 08028 Barcelona, Spain d Department of Electronics, University of Barcelona, c/ Martı ´i Franque`s 1, 08028 Barcelona, Spain Received 1 February 2008; received in revised form 4 July 2008; accepted 8 July 2008 KEYWORDS Mesenchymal stem cells; Osteoblasts; Topography; Microstructures Summary Recent studies on 2D substrates have revealed the importance of surface properties in affecting cell behaviour. In particular, surface topography appears to influence and direct cell migration. The development of new technologies of hot embossing and micro-imprinting has made it possible to study cell interactions with controlled micro features and to determine how these features can affect cell behaviour. Several studies have been carried out on the effect of microstructures on cell adhesion, cell guidance and cell proliferation. However, there is still a lack of knowledge on how these features affect mesenchymal stem cell differentiation. This study was designed to evaluate whether highly controlled microstructures on PMMA could induce rMSC differentiation into an osteogenic lineage. Structured PMMA was seeded with rMSC and cell number; cell morphology and cell differentiation were evaluated. Results confirm that microstructures not only affect cell proliferation and alignment but also have a synergistic effect with osteogenic medium on rMSC differentiation into mature osteoblasts. & 2008 Elsevier GmbH. All rights reserved. ARTICLE IN PRESS www.elsevier.de/aanat 0940-9602/$ - see front matter & 2008 Elsevier GmbH. All rights reserved. doi:10.1016/j.aanat.2008.07.013 Corresponding author at: Biomaterials, Biomechanics and Tissue Engineering Group, Department of Materials Science and Metallurgy, Universitat Polite `cnica de Catalunya, Avda. Diagonal 647, 08028 Barcelona, Spain. E-mail address: [email protected] (E. Engel).
Elisabeth Engela,b,�, Elena Martınezc, Chris A. Millsc, Miriam Funesc,Josep A. Planella,b, Josep Samitierc,d
aBio/Non-Bio Interactions for Regenerative Medicine Group, Institute for Bioengineering of Catalonia (IBEC),Josep Samitier 1-5, 08028 Barcelona, SpainbBiomaterials, Biomechanics and Tissue Engineering Group, Department of Materials Science and Metallurgy,Universitat Politecnica de Catalunya, Avda. Diagonal 647, 08028 Barcelona, SpaincNanobioengineering Group, Institute for Bioengineering of Catalonia (IBEC), Josep Samitier 1-5,08028 Barcelona, SpaindDepartment of Electronics, University of Barcelona, c/ Martı i Franques 1, 08028 Barcelona, Spain
Received 1 February 2008; received in revised form 4 July 2008; accepted 8 July 2008
ing author at: Biomativersitat Politecnica deess: Elisabeth.Engel@up
SummaryRecent studies on 2D substrates have revealed the importance of surface propertiesin affecting cell behaviour. In particular, surface topography appears to influenceand direct cell migration. The development of new technologies of hot embossingand micro-imprinting has made it possible to study cell interactions with controlledmicro features and to determine how these features can affect cell behaviour.Several studies have been carried out on the effect of microstructures on celladhesion, cell guidance and cell proliferation. However, there is still a lack ofknowledge on how these features affect mesenchymal stem cell differentiation. Thisstudy was designed to evaluate whether highly controlled microstructures on PMMAcould induce rMSC differentiation into an osteogenic lineage. Structured PMMA wasseeded with rMSC and cell number; cell morphology and cell differentiation wereevaluated. Results confirm that microstructures not only affect cell proliferation andalignment but also have a synergistic effect with osteogenic medium on rMSCdifferentiation into mature osteoblasts.& 2008 Elsevier GmbH. All rights reserved.
Elsevier GmbH. All rights reserved.
erials, Biomechanics and Tissue Engineering Group, Department of Materials Science andCatalunya, Avda. Diagonal 647, 08028 Barcelona, Spain.c.edu (E. Engel).
Research on stem cells is advancing the body ofknowledge on how an organism develops from asingle cell and how healthy cells replace damagedcells in adult organisms. This promising area ofscience is also leading researchers to examine thepossibility of cell-based therapies for treatment ofdisease, which is often referred to as regenerativeor reparative medicine.
Research using embryonic stem cells is growing ininterest due to their potential medical applica-tions. Despite this potential, the ethical issuessurrounding these cells have created intense con-troversy. This concern has promoted in-depthresearch into adult stem cells. Adult stem cellsare undifferentiated cells found in differentiatedtissue that can renew themselves and differentiate(within certain limitations) to give rise to all thespecialized cell types of the tissue from which theyoriginate. Such adult stem cells have been dis-covered and characterized in a multitude oftissues, which suggests that autologous clinicalimplantation or genetically engineered stem cellsfor protein or drug delivery could be used withoutrisk of immunorejection (The National Institute ofHealth resource for stem cell research, 2007).
In adult stem cell research, one of the main fieldsis the differentiation of stem cells into an osteo-genic lineage for subsequent use in bone tissueengineering and reconstructive medicine. Interestin this field is related to the increased incidence ofosteodegenerative diseases (i.e. osteoporosis andosteoarthritis) in the rapidly aging populations ofdeveloped countries. Bone is a mineralized tissuethat confers multiple mechanical and metabolicfunctions to the skeleton. Bone contains twodistinct cell types: osteoblasts and osteoclasts.Osteoblasts are bone-forming cells of mesenchymalorigin. Mesenchymal stem cells (MSCs) have thepotential to differentiate into osteoblasts, chon-drocytes, adipocytes, fibroblasts, marrow stromaand other tissues of mesenchymal origin. TheseMSCs can be obtained from many sources (blood,adipose tissue, periosteum, muscle, dermis, etc.),but bone marrow aspirate is the most commonsupply. Due to the variety of cell types and therecently discovered plasticity of adult stem cells(stem cells from one tissue may give rise to celltypes of a completely different tissue), there is aneed for well-defined and efficient protocols fordirecting the differentiation of stem cells into theosteogenic lineage.
Cell behaviour depends on interactions with theenvironment. Consequently, the contact betweencells and implantable materials will determine the
success or failure of an implantable material or amedical device. It is well known that the cellresponse is affected by the physicochemical para-meters of the biomaterial surface, such as surfaceenergy, surface charges or chemical composition.Topography is one of the most crucial physical cuesfor cells. Microtopography influences cell adhesion,proliferation and differentiation (Zinger et al.,2005; Boyan et al., 2002; Aparicio et al., 2002).More recently, it has become clear that nano-topography also guides cell behaviour (Dalbyet al., 2007). Chemical cues in the form ofvarious biomolecules, such as adhesive proteins(i.e. fibronectin and laminin), also influence cellbehaviour in a crucial manner (Lan et al., 2005).
Here we evaluate the effect of various surfacemodifications with the aim of directing rat MSCs todifferentiate into osteoblast lineage.
Materials and methods
Polymer fabrication
Poly (methyl methacrylate) (PMMA, 125 mm thick)was purchased from Goodfellow (UK) and used asreceived. A nanoimprint lithography apparatus(Obducat, AB, Sweden) was used to fabricatefree-standing PMMA samples with surface featuresthat had lateral dimensions ranging from 50 to2 mm, following a previously reported hot emboss-ing procedure (Mills et al., 2007). Briefly, thistechnique uses a microstructured silicon mould,covered by a grown silicon oxide layer, which ispressed into the bulk PMMA polymer at a tempera-ture of 130 1C and pressure of 30 bars. The mouldshad been previously treated with a fluoroalkylsi-lane, (trichloro(tridecafluoro-octyl)-silane, to pre-vent adhesion so that they could be reused afterdemoulding. The structural dimensions in thepolymer replica were measured by an interfero-metric microscope (WYKO NT1100, Veeco, USA).After the nanoimprinting procedure, and prior tocell culturing, the transparent, free-standing PMMAsamples were sterilized by ethanol.
Cell isolation and culture
Bone marrow stromal cells obtained from thefemora of female Wistar rats aged 10–12 weekswere cultured as follows. Rats were killed and thefemora were removed and placed in culturemedium. The articular ends of each femur wereexcised at the epiphysis and the marrow expelledby injecting a stream of medium into the medullary
ARTICLE IN PRESS
E. Engel et al.138
cavity, using a syringe fitted with a 20-gaugeneedle. The expelled marrow was suspended in10ml medium (advanced modified Eagle medium+15% foetal bovine serum+1% penicillin/streptomy-cin+1% glutamine) and heparin (10 u/ml), anddispersed by repeated aspiration and expulsionthrough the needle. Cells obtained from eightanimals were seeded on T-75 tissue culture flasks.After 24 h, non-adhered cells were removed bymedium replacement. The medium was changedevery 2 or 3 days. Once the cells were at 80% ofconfluence, they were split at a ratio of 1:2. Cellsfrom passages 2–5 were used for all experiments.Cells from passage 2 were characterized forosteoblast lineage. Medium supplemented withb-glycerophosphate (10mM), dexamethasone(10�8 M) and ascorbic acid (50 mg/ml) was used asthe osteogenic medium (OM).
Alkaline phosphatase and von Kossa staining wereused to analyse the osteogenic potential of therMSC. Cells seeded on 24-well plates at a density of103 cells/cm2 were assayed on days 14 and 21 todetect alkaline phosphatase activity and miner-alization nodules. For this purpose, cells were fixedin 10% formalin and a solution of naphthol, AS-MXphosphate (70476 Sigma) and fast blue RR dissolvedin Milli-Q water was poured into the wells, inaccordance with the manufacturer’s instructions(Sigma Kit #85). Cells were incubated in the darkfor 30min and rinsed in water. Cells were stainedusing the von Kossa method to detect calcium-phosphate-containing matrices. A solution of silvernitrate (2%) in water was added to the wells, whichwere then incubated for 15min in the dark andrinsed twice in water. Finally, the samples wereexposed to white light (15min) and mounted inglycerine jelly. The samples were imaged using amicroscope (Nikon Eclipse E600).
Cell morphology, proliferation anddifferentiation
rMSC from passage 4 were used to assay cellmorphology, proliferation and differentiation onthe different PMMA surfaces. Cell culture plates ofpolystyrene (PS) were used as a control.
Cell morphologyCell morphology was analysed by confocal micro-
scopy and scanning electron microscopy. Immunos-taining was performed for the confocal microscopy.Cells were fixed (3% paraformaldehyde in phos-phate buffer 0.1M and sucrose 60mM) and thenpermeabilised for 10min in 1% Triton X-100 solu-tion. Blockage was completed using 10mM
PBS/20mM glycine/1% BSA for 20min. Afterwards,primary antibody for actin was added and thesamples were incubated for 1 h at 37 1C in a humidatmosphere followed by a final incubation withsecondary antibodies (goat anti-rabbit Alexa Fluor568), phalloidin and Hoechst for nuclei staining inblocking solution at 37 1C for 1 h. Dried sampleswere mounted in Mowiol+antifade and imagedusing a confocal microscope.
The preparation for scanning electronic micro-scopy (SEM) was performed as follows. Cells werefixed in 2.5% glutaraldehyde solution, post-fixed in1% OsO4 solution and dehydrated at room tem-perature via immersion in increasing concentra-tions of ethanol. Finally, critical point drying wasundertaken in liquid CO2.
A time-lapse video recording was performed for4 h, taking pictures every 10min.
Cell proliferationCell proliferation was determined at 10 days of
culture using the cell proliferation reagent WST-1(Roche, Germany). Briefly, the overall activity ofmitochondrial dehydrogenases in the sample wasmeasured by adding 20 ml of reagent to each welland quantifying the formazan produced in amicroplate ELISA reader at 450 and 600 nm asreference wavelengths. Measurements were com-pleted in triplicate.
Cell differentiationQuantification of alkaline phosphatase activity
(Sigma Diagnostic Kit #104) and osteocalcin(MetraTM, DPC Dipesa, Barcelona) was carried outon PS samples, on flat PMMA and on structuredPMMA (round and square structures) according tothe manufacturer’s instructions. rMSC at a densityof 104/well in a 24-well plate were seeded for 7days. The medium was replaced twice a week.After this period, the medium was replaced by OMin half of the samples and with regular medium inthe other half, in order to evaluate the effect ofthe differentiation factors on the structures.Results were correlated to the number of cells.
Results
Cell isolation and culture
rMSC from bone marrow were isolated andpassaged twice to evaluate their potential tobecome osteoprogenitors and mature osteoblasts.rMSC proliferated for 11 days. The cell populationincreased 5-fold. After 14 days of culture, 75% of
the total cell population was ALP positive(Figure 1A). After 21 days of culture, mineralizationnodules were seen (Figure 1B). Very few cells withadipocyte morphology were seen at the beginningof the cultures, and hardly any after 21 days.
Cell morphology, proliferation anddifferentiation
Cell morphologyrMSC acquired different morphologies, according
to the structures on the PMMA (Figure 2). Figure 3shows confocal images of rMSC on square and roundstructures, cultured for 3 days and stained forvinculin to visualize focal contacts, actin filamentsand nuclei. Cells were more stressed on the squaresmall structure. They spread out but were starshaped. Several filopodia could be seen in thestar shape. However, in some cases, cells had aspindle, needle-like appearance. Closer examina-
Figure 1. rMSC cultured for 14 days (A) and 21 days (B). Positwells after 14 days. After 21 days, staining of mineralizationmineralization nodules.
Figure 2. Interferometry images of the microstructured PMheight and (B) the square structure of 2mm width and 1 mm
tion showed that the filopodia were alreadyinteracting, mostly with the microstructures. Eachof these filopodia was attached to one structure atthe end (Figure 3), as shown by the SEM pictures(Figure 4). In addition, a certain orientation ofthe cells was observed on the small structure(Figures 3D and 4B).
The rounded structures were wide enough for thecells to stay inside and even divide a couple oftimes (Figure 3A). Cells were spread to the limit ofthe rounded structure. When rMSC were not insidethe wells, they had a very well-organized cytoske-leton, with a lot of vinculin expression, whichindicates that there were many focal contacts(Figure 3B). There were fewer focal contacts andless organized actin filaments on the roundedstructures than on the small structure (Figure 3Cand D).
The time-lapse video (Figure 5) recorded on theround structures for 4 h showed that MSC tended topenetrate the structures. However, not all of them
ive alkaline phosphatase staining could be observed in allnodules could be seen. Figure 2(B) shows a zoom of the
MA: (A) The rounded structure of 50 mm width and 1 mmheight.
ARTICLE IN PRESS
Figure 3. Immunohistochemical staining of actin (red), vinculin (green) and cell nuclei (blue): (A) rMSC on roundstructures. Three cells are located inside the structures and are connected to one another among the structures. (B)rMSC on round structures are well spread, with many stress fibres and a high number of focal contacts (vinculin). (C)rMSC on square structures have two types of morphologies: one more elongated, with no spreading at all, and anotherthat is star-shaped. (D) rMSC on square structures show stress fibres but fewer focal contacts, due to the lowerspreading. Notice the alignment of the cytoskeleton.
Figure 4. SEM images of rounded structures (A), square structures (B) and flat PMMA. rMSC on round structures can alsoget around these features. On square (small) structures, cells are aligned and several prolongations or filopodia can beobserved that attach preferentially to the structures. On flat PMMA, the cells are randomly oriented and severalmorphologies can be observed, depending on the cell activity.
E. Engel et al.140
entered, some just bypassed the structures. WhenMSCs were already inside the round features theydid not come out.
Cell proliferationResults of cell proliferation are shown in
Figure 6. rMSC proliferated for 10 days on the flat
ARTICLE IN PRESS
Figure 5. Pictures taken by time-lapse video recording show how cells get inside the round structures (pointed witharrows) and how the ones that were already inside these structures do not get out (pointed by arrow heads).
Figure 6. rMSC proliferation after 10 days in culture.Cells were cultured on all surfaces, including PS, flatPMMA and structured PMMA, either with or without OM.
Figure 7. rMSC differentiation. Cells cultured for 11 daysin the presence or absence of OM: (A) Alkaline phospha-tase activity correlated with cell number. (B) Osteocalcinamount correlated with cell number.
and structured PMMA. Cells on flat PMMA prolifer-ated extensively, at the same levels as on PS withno OM. When factors for osteoblast differentiationwere added, proliferation decreased, both on flatPMMA and on PS. Cells on the structured PMMAshowed greater proliferation after 10 days on thesamples without the OM, compared to flat PMMAand PS. The behaviour when the OM was added wassimilar to that on flat PMMA and PS.
ARTICLE IN PRESS
E. Engel et al.142
Cell differentiationIn the absence of OM, cells on flat PMMA
substrates showed the same levels of alkalinephosphatase activity as on PS. When OM was added,alkaline phosphatase activity increased on flatPMMA, even though this result was not significant.In the absence of OM, levels of AP activity werelower on microstructured substrates than on flatPMMA and PS. This difference was enhanced for therounded structures. In the presence of OM, rMSC onthe square structures showed the same levels of APactivity as PS (Figure 7A).
In terms of osteocalcin (Figure 7B), low levels ofthis osteoblastic late marker were observed in theabsence of OM in all samples. On structures, theselevels were half of those obtained for flat PMMA andPS. When OM was added, levels of osteocalcinincreased 5-fold for the flat PMMA and PS and 6-foldfor the structured PMMA. There were no differ-ences among the samples.
Discussion
MSCs are usually grown on PS treated withpolylysine, to enhance cell adhesion and cellgrowth. The stability of these cultures on other2D synthetic substrates has not been studied indepth. The most common polymers used for MSCcultures are natural polymers, such as collagen(mostly Type I), fibrin, alginates, hyaluronic acid,etc. However, synthetic polymers allow bettercontrol of physicochemical properties than naturalpolymers. In addition, the use of synthetic poly-mers reduces the risk of potential biohazardouscomplications associated with animal polymers(Cancedda et al., 2003).
PMMA is a highly hydrophilic substrate that hasshown good biocompatibility with cells and is beingused in contact lenses. It is a common polymer,which is used for micro- and nanoimprintingprocesses. It is an amorphous, thermoplastic poly-mer with excellent optical transparency and a glasstransition temperature �105 1C (Mills et al., 2007).
Several studies have shown that cells react totopographical features, which affect cell adhesion,morphology and proliferation (Wilkinson et al.,2002). The work done by the Boyan’s group (Batzeret al., 1998) on titanium has had a great impact onthe dental implant field. Their results show that arange of roughness of around 4 mm induces osteo-blast differentiation. However, the establishmentof a relationship between stem cell differentiationand microstructures has not been well studied.Only a few works (Zinger et al., 2005; Dalby et al.,
2007; Popat et al., 2008) have studied the effect ofnano- and microstructures on cell differentiation.
Cell differentiation is of major relevance intissue engineering applications. The possibility ofcreating topography that is able to induce MSCdifferentiation would revolutionise the field ofbone implants. In this sense, the results of thisstudy support the idea that certain surface patternscan induce differentiation.
The cell proliferation results illustrate that rMSCtends to increase on patterned PMMA surfaces, ineither large or small structures. These results werenot significant, which could be due to the hetero-geneity of rMSC (Derubeis and Cancedda, 2004).There was less cell proliferation on all substrateswhen differentiation was induced by OM. Never-theless, there were still more cells on the struc-tured PMMA than on the flat substrate.
When analysing the activity of AP and osteocalcinas early and late markers for osteoblast differen-tiation in the absence of OM, both markers showedlow levels on all surfaces, as expected. Levels onthe structured surfaces were lower than those onPS and flat PMMA. When inducing osteogenicdifferentiation, a greater amount of both markerswas quantified on PS and flat PMMA. The mostinteresting result was that levels of AP activity andosteocalcin secretion on structured surfaces wereequivalent to those encountered on flat surfaces(PS and PMMA). This result was more marked forosteocalcin, where no differences among thesurfaces were detected. This means that thestructures act synergistically with the osteogenicfactors present in the medium. These results are inconcordance with the results of the Boyan’s group(Batzer et al., 1998) on titanium, and with thesynergistic effect of roughness and 1a, 25-(OH)2D3,a well-known differentiation factor in osteoblastscultures.
Several studies indicate that chemical surfacemodifications have the greatest impact on osteo-blast behaviour (Lan et al., 2005; Keselowsky et al.,2005). However, topographical modifications alsopromote chemical modifications. The wettability ofa surface is known to be affected by its topography(Navarro et al., 2006). Previous results (Mills et al.,2007) on PMMA contact-angle measurements haveshown that the embossed, nonstructured PMMAsurface is originally hydrophilic but becomeshydrophobic when patterned with microstructures.The surface also becomes increasingly hydrophobicas the dimensions of the structures are reduced.These facts may explain why the cells are thinnerand more elongated, producing a more sphericalcross section, than those on the nonstructuredsurface, where the cell is forced to interact with a
flat surface. Cells seem to prefer hydrophilicsurfaces. We have observed this trend in adhesion(data not shown) to the structured surfaces (themost hydrophobic) during the first 24 h. However,the number of cells was equivalent after 10 days ofculture. This is due to the different proteinadsorption associated with hydrophobicity andhydrophilicity, which affects the initial adhesion.However, after 2 or 3 days, cells have producedtheir own extracellular matrix and have developeda microenvironment that is feasible for cell growth.
The differential reaction to the two types ofstructures is in accordance with the cell morphol-ogy. As seen in Figure 3, cells on the structuresshowed stress fibres. These were more evident inthe small structure and in large structures whencells were in contact with the feature. It has alsobeen reported that cell shape regulates the switchin lineage commitment of human MSCs (hMSC), bymodulating endogenous RhoA activity. When hMSCare allowed to adhere, flatten and spread, theyundergo osteogenesis, whereas unspread, roundcells become adipocytes (McBeath et al., 2004).Apparently, changes in cell spreading alter RhoA-mediated cytoskeletal contractility, focal adhesionassembly and downstream integrin signaling. Incontrast, we could not observe adipocytes in ourcultures. This could be due to the different types ofstructures and the cell origin.
In conclusion, our results confirm the relevanceof surface topography to MSC behaviour. Structuresenhanced MSC proliferation and gave good levels ofdifferentiation in the presence of OM. Surfacetopography is one of the main parameters used tocontrol the design of medical devices and implants.Moreover, the design of microtextured tissueengineering matrices for specific tissues can beused to facilitate experiments on cultured cells invitro and to provide more physiological conditionsunder which to maintain cells.
Acknowledgements
This study was carried out in the context of theCellPROM project, funded by the European Com-munity as contract no. NMP4-CT-2004-500039. Theinformation contained in this paper reflects theauthors’ views only. The authors are grateful tothe Spanish Ministry of Science and Education (MEC)for support provided through project TEC2004-06514-C03, and for the provision of grants by theRamon y Cajal programme (EM). Drs. C. Moormannand T. Wahlbrink from AMO Gmb (Germany) aregratefully acknowledged for providing the mouldcontaining the microstructures.
References
Aparicio, C., Gil, F.J., Planell, J.A., 2002. Humanosteoblasts proliferation and differentiation on grit-blasted and bioactive titanium for dental applications.J. Mater. Sci.: Mater. Med. 13, 1105–1111.
Batzer, R., Kieswetter, K., Liu, Y., Cochran, D.L.,Szmuckler-Moncler, S., Dean, D.D., Schwartz, Z.,1998. Titanium surface roughness alters responsive-ness of MG63 osteoblast-like cells to 1a, 25-(OH)2D3. J.Biomed. Mater. Res. 39, 77–85.
Cancedda, R., Dozin, Beatrice, Giannoni, P., Quarto, R.,2003. Tissue engineering and cell therapy of cartilageand bone. Matrix Biol. 22, 81–91.
Dalby, M., Gadegaard, N., Tare, R., Andar, A., Riehle,M.O., Herzyk, P., Wilkinson, C.D.V., Oreffo, R.O.C.,2007. The control of human mesenchymal celldifferentiation using nanoscale symmetry and disor-der. Nat. Mater., 997–1003.
Derubeis, A.R., Cancedda, R., 2004. Bone marrowstromal cells (BMSCs) in bone engineering: limitationsand recent advances. Ann. Biomed. Eng. 32 (1),160–165.
Mills, C.A., Fernandez, J.G., Martinez, E., Funes, M.,Engel, E., Errachid, A., Planell, J.A., Samitier, J.,2007. Directional alignment of MG63 cells on polymersurfaces containing point microstructures. Small 3 (5),871–879.
Navarro, M., Aparicio, C., Charles-Harris, M., Ginebra,M.P., Engel, E., Planell, J.A., 2006. Development of abiodegradable composite scaffold for bone tissueengineering: physicochemical, topographical, me-chanical, degradation, and biological properties.Adv. Polym. Sci. 200, 209–231.
Popat, K.C., Chatvanichkul, K.I., Barnes, G.L., LatempaJr., T.J., Grimes, C.A., Desai, T.A., 2008. Osteogenicdifferentiation of marrow stromal cells cultured onnanoporous alumina surfaces. J. Biomed. Mater.Res. A.
The National Institutes of Health resource for stem cellresearch, 2007. /http://stemcells.nih.gov/info/glos-sary.aspS.
Wilkinson, C.D.W., Riehle, M., Wood, M., Gallagher, J.,Curtis, A.S.G., 2002. The use of materials patternedon a nano- and micro-metric scale in cellularengineering. Mater. Sci. Eng. C – Biomimetic Supra-mol. Syst. 19, 263–269.
Zinger, O., Zhao, G., Schwartz, Z., Simpson, J.,Wieland, M., Landolt, D., Boyan, B., 2005. Differ-ential regulation of osteoblasts by substratemicrostructural features. Biomaterials 26 (14),1837–1847.
