directed osteogenic differentiation of human mesenchymal stem/precursor cells on silicate...

Directed osteogenic differentiation of human mesenchymal stem/precursor cells on silicate substituted calcium phosphate

Kate Cameron,1 Paul Travers,1 Chaman Chander,2 Tom Buckland,2 Charlie Campion,2

Brendon Noble3

1MRC Centre for Regenerative Medicine, SCRM Building, University of Edinburgh, 5 Little France Drive,

Edinburgh EH16 4UU, United Kingdom2ApaTech Ltd., 370 Centennial Avenue, Centennial Park, Elstree, Herts WD6 3TJ, United Kingdom3School of Science, Technology and Health, University Campus Suffolk, Waterfront Building, 19 Neptune Quay,

Ipswich IP4 1QJ, United Kingdom

Received 28 February 2012; revised 13 April 2012; accepted 3 May 2012

Published online 00 Month 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/jbm.a.34261

Abstract: Insufficient, underactive, or inappropriate osteo-

blast function results in serious clinical conditions such as

osteoporosis, osteogenesis imperfecta and fracture nonunion

and therefore the control of osteogenesis is a medical prior-

ity. In vitro mesenchymal stem cells (MSCs) can be directed

to form osteoblasts through the addition of soluble factors

such as b-glycerophosphate, ascorbic acid, and dexametha-

sone; however this is unlikely to be practical in the clinical

setting. An alternative approach would be to use a scaffold

or matrix engineered to provide cues for differentiation with-

out the need for soluble factors. Here we describe studies

using Silicate-substituted calcium phosphate (Si-CaP) and

unmodified hydroxyapatite (HA) to test whether these materi-

als are capable of promoting osteogenic differentiation of

MSCs in the absence of soluble factors. Si-CaP supported

attachment and proliferation of MSCs and induced osteogen-

esis to a greater extent than HA, as evidenced through upreg-

ulation of the osteoblast-related genes: Runx2 (1.2 fold),

Col1a1 (2 fold), Pth1r (1.5 fold), and Bglap (1.7 fold) Dmp1

(1.1 fold), respectively. Osteogenic-associated proteins, alka-

line phosphatase (1.4 fold), RUNX2, COL1A1, and BGLAP,

were also upregulated and there was an increased produc-

tion of mineralized bone matrix (1.75 fold), as detected by the

Von Kossa Assay. These data indicate that inorganic sub-

strates are capable of directing the differentiation programme

of stem cells in the absence of known chemical drivers and

therefore may provide the basis for bone repair in the clinical

setting. VC 2012 Wiley Periodicals, Inc. J Biomed Mater Res Part A:

00A:000–000, 2012.

Key Words: mesenchymal stem cells, biomaterial, osteogenic

differentiation, calcium phosphate

How to cite this article: Cameron K, Travers P, Chander C, Buckland T, Campion C, Noble B. 2012. Directed osteogenicdifferentiation of human mesenchymal stem/precursor cells on silicate substituted calcium phosphate. J Biomed Mater Res Part A2012:00A:000–000.

INTRODUCTION

Mesenchymal stem cells (MSCs) are well established as apotential source of cells for tissue engineering and cellulartherapy; as they can be autologously derived, readilyexpanded and directed to differentiate into multiple celltypes. In vivo, MSCs are thought to migrate to sites of damageand differentiate into the required cell type1; thus they play avital role in tissue repair. In vitro, MSCs can be directed alonga restricted set of lineages in the presence of soluble factors;for example, the addition of b-glycerophosphate, ascorbicacid and dexamethasone results in commitment to the osteo-blast lineage.2,3 This is a complex, highly regulated processinvolving the parallel and sequential expression of lineage-specific genes. An essential transcription factor controllingthis process is Runx2, which is believed to be both necessaryand sufficient to direct MSCs toward the osteoblast lineage.4,5

Runx2 acts by binding to a regulatory element found in the

promoter of all major osteoblast genes, including COL1A1(collagen type I)6 and BGLAP (osteocalcin).7 The proteinsassociated with these genes, amongst others, are secreted byosteoblasts and are incorporated into osteoid, the organicphase of bone.8 Alkaline phosphatase is produced, initially inparallel with osteoid production, then declining as matrixmineralization occurs,9 and calcium phosphate crystalsappear forming the mineral phase of bone.10 Osteoblasts thatare left behind during osteoid production are then encasedin the mineralized osteoid, cease proliferation and becometerminally differentiated osteocytes, known to express DMP1(dentin matrix acidic phosphoprotein 1).11

Bone formation is widely studied and is known to beactivated through a variety of soluble factors, includingbone morphogenic proteins and other supporting cyto-kines.12,13 However, the role that the local physical environ-ment plays in the process of osteogenic differentiation is

Correspondence to: K. Cameron; e-mail: [email protected]

VC 2012 WILEY PERIODICALS, INC. 1

less well understood and represents an emerging area ofresearch.

Ideally, bone replacement materials need to be porous,biodegradable scaffolds with osteogenic potential.14 Themineral component of bone has a structure similar to hy-droxyapatite (HA; Ca10 (PO4)6OH2) and this similarity formspart of the reasoning behind the use of synthetic HA as askeletal replacement and bone void filler in cases of boneinjury.15 Unlike HA, natural bone mineral incorporatesnumerous substitutions within the calcium phosphate lat-tice, such as carbonate ions and trace elements. These sub-stitutions appear to be crucial to the biological activity andsurface chemistry of bone.16 In particular, trace levels of sili-con in the form of silicate (SiO4�

4 ) are known to be essentialfor normal growth and development of bone.17,18

Silicate-substituted calcium phosphate-based materialsdemonstrate increased bone matrix production in vitro andpromote greater bone formation in vivo when comparedwith pure HA.19 It is not clear whether the increased boneformation associated with these materials is the result ofincreased osteoblast activity or increased numbers of osteo-blasts resulting from increased activation of stem/progeni-tor cells. Recent research has focused on differences inphysical and surface properties of bone scaffold materials,demonstrating that materials with different surface topogra-phies and chemical properties can significantly influenceprocesses such as cell adhesion,20 proliferation,21 and differ-entiation,22 suggesting that cells commit to a specific lineagedepending on the physical properties of the substrate towhich they adhere. This is the first study to examine andcompare the response of MSCs to hydroxyapatite and sili-cate substituted calcium phosphate in the absence of solublestimulators of differentiation. Using in vitro preparations ofhuman MSCs we are able to investigate whether a mecha-nism exists by which the increased bone formation associ-ated with Si-CaP compared with HA might, at least in partbe the result of increased production of osteoblasts fromtheir stem cell source rather than increased synthetic activ-ity alone. The aim of the present study was to ask whetherhydroxyapatite or a silicate-substituted hydroxyapatitematerial could promote osteogenic differentiation of MSCsin the absence of soluble osteogenic factors.

MATERIALS AND METHODS

Culture materialsAll calcium phosphate-based materials were provided byApaTech Ltd (Elstree, UK) in the form of dense sintereddiscs 1-mm thick and with 14 mm in diameter. Discs werecomposed of either HA or Si-CaP (0.8 wt % Si) with thesame physical structure. All experiments were carried outon discs in 24-well tissue culture plates (Corning LifeSciences). All reagents were from Sigma–Aldrich Ltd. (PooleDorset, UK) unless otherwise stated.

Isolation and cell cultureHuman femoral head tissue samples were obtained underinformed consent from donors undergoing orthopedic sur-gery for osteoarthritis. Human MSCs were isolated from

tissue samples and expanded using a method published pre-viously.23 In brief, whole bone marrow tissue was removedfrom femoral heads with a curette, triturated in sterile con-ditions and cultured in Dulbecco’s Modified Eagle medium(DMEM) supplemented with 10% fetal calf serum (FCS),L-glutamine (2 mM), penicillin (100 IU mL�1), and strepto-mycin (100 lg mL�1) (Sigma–Aldrich, Poole, UK), termedgrowth media. Whole bone marrow cells were maintainedfor 4 days at 37�C, 5% CO2, after which time nonadherentcells were removed and the medium changed. Media changewas repeated every 3 days until the adherent fibroblasticcells were confluent. Confluent cells were passaged bytreatment with 0.05% trypsin in phosphate buffered salinesolution (PBS; Sigma–Aldrich Ltd., Poole, UK) for use inexperiments or reseeded. All experiments were conductedusing cells up to passage 4.

Induction of osteogenic differentiationOsteogenic differentiation of cells was induced on tissueculture polystyrene as described previously2 by culture inosteogenic medium (OS: 50 lM ascorbic acid phosphate(Wako, Neuss, Germany), 10 mM b-glycerophosphate and100 nM dexamethasone) The cell cultures were maintainedfor up to 21 days and the culture medium was replacedthree times every 7 days.

Cell attachmentThree replicates of both HA and Si-CaP discs were placedinto 24-well tissue culture plates. Each well was seeded with2 � 104 MSCs in 100 lL of growth media. At 2, 4, 6, and 8 hafter seeding, substrates were washed three times with PBSand remaining cells fixed for 15 min with 500 lL of 4%paraformaldehyde (PFA). Fixed cell samples were incubatedwith 40, 6-diamidino-2-phenylindole (DAPI) for 10 min andthe total cell nuclei (equivalent to cell number) countedautomatically using image analysis software.24 The meannumber of cells attached per mm2 was then calculated.

Cell proliferationThree replicates of HA and Si-CaP discs were placed into24-well tissue culture plates and each well was seeded with2 � 104 MSCs in 100 lL of growth media. Eight hours afterseeding, 1 mL of standard growth medium was added tocells on all substrates, namely Si-CaP, HA, and plastic (PL); aseparate group of cells cultured on PL received OS media(PLþOS). Thereafter, media was replaced every 3 days. Cellswere washed three times in PBS, fixed for 15 min with 500lL of 4% PFA and stained with DAPI at the following time-points: days 1, 3, 5, 7, 14, and 21 after seeding. At eachtime-point, four images were taken in different randomlyselected substrate locations for each sample. Cell numberwas determined using image analysis software (Image J)24

and the mean number of cells per mm2 was calculated.

Alkaline phosphatase positive cell areaAlkaline phosphatase activity (ALP) was assessed by histo-chemical staining and quantified by determining the area ofALP positive (dark blue) cells. Cells were seeded and

2 CAMERON ET AL. NOVEL SUBSTRATE FOR OSTEOGENIC DIFFERENTIATION

cultured as described above. After 1, 7, 14, and 21 days ofseeding (experimental time-points), each well was washedthree times in PBS, fixed for 15 min with 500 lL 4% PFAand washed again three times with PBS. Alkaline phospha-tase histochemical staining was performed using a commer-cial analytical test kit (Leukocyte Alkaline Phosphatase Kit,Sigma–Aldrich, Poole, UK) following manufacturer’s instruc-tions. At each time-point, four images were taken in differ-ent random substrate locations for each sample and ALPexpression quantified by determining the area of ALP posi-tive (dark blue) cells using Image J.24

Mineralized extra cellular matrix identification(Von Kossa histochemistry)MSCs seeded onto HA or Si-CaP discs or on tissue culturepolystyrene were fixed and washed as for ALP detectionabove, and incubated in 5% silver nitrate solution beforebeing exposed to a 60-W light bulb at a distance of �5 cm for10 min. This allowed differential levels of staining of any min-eralized matrix produced the weakly staining substrates thatalso contain calcium phosphate. Following exposure, cellswere washed in distilled water and incubated in 5% sodiumthiosulphate solution for 5 min. Four images were taken indifferent substrate locations for each time point (1, 7, 14, and21 days) for each substrate. Areas of mineralized matrix(black deposits) were quantified using Image J.24

RNA extraction and RT-PCRTotal RNA was isolated from human MSCs cultured on Si-CaP,HA, PL, and PLþOS. On days 1, 7, 14, and 21 after seedingextraction procedures were undertaken using a commercialTrizol reagent (Invitrogen, Paisley, UK) and following instruc-tions described previously.25 Total RNA samples were tran-scribed to cDNA using Superscript III First strand SynthesisSuperMix for RT-PCR (Invitrogen). For the quantitative RT-PCR, 40 ng of template cDNA was added to each PCR reac-tion. Quantitative PCR was performed using an ABI 7500Fast Real-Time PCR system (Applied Biosystems., Warring-ton, UK). Pairs of specific primers (Applied Biosystems Ltd)were used for the following genes: beta-2-microglobulin(B2M; Hs99999907_m1), CD105 (Hs00164438_m1), runt-related transcription factor 2 (Runx2; Hs01047978_m1), col-lagen type I (Col1a1; Hs00164004_m1), osteocalcin (Bglap;Hs00609452_g1), parathyroid hormone receptor 1 (Pth1r;Hs00174895_m1), and dentin matrix acidic phosphoprotein1(Hs01009391_g1). Standard enzymes and cycling condi-tions for the ABI 7500 Fast System were used (initial denatu-ration step at 95�C for 20 s, followed by 40 amplificationcycles of 3-s duration at 95�C and then 30 sat 60�C). Dataanalysis was performed using 7500 Fast System SequenceDetection Software (version 1.3, Applied Biosystems). Levelsof expression of each gene of interest were normalized tob-2-microglobulin gene levels at each time point.

Immunocytochemical stainingRunx2. MSCs fixed for 15 min with 500 lL 4% PFA wereblocked with blocking buffer (10% goat serum, 1% BSA and0.1% Triton X-100) for 1 h at room temperature and then

incubated with Anti-human Runx2/CBFA1 antibody (R&D Sys-tems, Abingdon, UK, MAB2006; 1:200 dilution in antibody dil-uent- overnight at 4�C. Cells were washed three times withPBS (1 mL) on an orbital shaker at 40 rpm for 5 min. Cellswere incubated with anti-rat secondary antibody conjugatedwith Alexa Fluor 488 (Invitrogen; A11006; 1:400 dilution inantibody diluent) for 1 h in the dark and at room temperature.

Col1a1. MSCs fixed for 15 min with 500 lL 4% PFA wereblocked with Dako protein block (Dako, UK) for 1 h at roomtemperature and then incubated overnight at 4�C withmouse monoclonal anti-collagen type I antibody (Sigma–Aldrich; c2456; 1:200 dilution in antibody diluent). Cellswere washed three times with PBS (1 mL) on an orbitalshaker at 40 rpm for 5 min and then incubated with ananti-mouse secondary antibody conjugated with Alexa Fluor488 (Invitrogen; A11001; 1:1000 dilution in antibody dilu-ent) for 1 h at room temperature and protected from light.

Osteocalcin. Cells were fixed for 15 min with 500 lL 4%PFA, then blocked with blocking buffer for 45 min at roomtemperature and incubated in anti-human osteocalcin anti-body (R&D Systems; MAB1419; 1:200 dilution in PBS) over-night at 4�C. Cells were washed three times with PBS(1 mL) on an orbital shaker at 40 rpm for 5 min and thenincubated with anti-mouse secondary conjugated with AlexaFluor 488 (Invitrogen; A11001, 1:1000 dilution in PBS) atroom temperature for 1 h in the dark.

Image acquisition and analysisAll images were collected using a Zeiss Axio Observer.Z1microscope with an LD Plan-Neofluar 20�/0.4 Korr Ph 2objective lens (Carl Zeiss, Welwyn Garden City, UK). Themicroscope was coupled to a Zeiss AxioCamMR3 camera.The image acquisition and processing software used wasZeiss Axiovison Rel 4.8. and Axiovison Version 4.7.1.0,respectively. When visualizing immunofluorescent staining,Dako fluorescent imaging medium (Dako UK Ltd, Cam-bridgeshire, UK) was used. Alkaline phosphatase activityand area of mineralized matrix was quantified by an auto-mated image analysis macro created in image software(Image J), which detected staining as pixels with an inten-sity at least three fold greater than the background. Allimages were collected at room temperature.

Statistical analysisOne-way analysis of variance (ANOVA) was used to test fordifferences between group sets for attachment, growth, ALP,Von Kossa experiments, and for RT-PCR data. All analyseswere performed using Microscoft Excel 2007 and a 95%confidence interval or a significance level of p < 0.05 wasconsidered statistically significant.

RESULTS

To investigate whether undifferentiated multipotent stromalcells could be initiated to differentiate along the osteogeniclineage in the absence of soluble factors; we studied theirresponse to various substrates: hydroxyapatite (HA) and

ORIGINAL ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | MONTH 2012 VOL 00A, ISSUE 00 3

silicon substituted calcium phosphate (Si-CaP). Cells cul-tured on tissue culture polystyrene alone (PL) do not differ-entiate along the osteoblast lineage unless the culture issupplemented with known osteogenic factors, thus MSCscultured with osteogenic medium (PLþOS) were used aspositive controls. We examined separately the attachment,proliferation and differentiation of MSCs cultured on thesesubstrates.

Cell attachmentThe average number of cells attached to all substratesincreased with time, as shown in Figure 1(A). The numberof cells attached to tissue culture polystyrene (PL) wasgreater than that obtained for HA at all time-points. Thenumber of cells attached to Si-CaP discs was significantlylower than that on PL for time-points up to and including6 h. However, in the 6–8 h time period there was a markedincrease in the mean (6SD) number of cells attached toSi-CaP and by 8 h the number of cells attached to thissubstrate was comparable to that for PL (129 6 10.3 vs.119 6 8.6 cells mm�2; p ¼ 0.222). On HA, the rate ofattachment showed the same time dependence as for Si-CaP,with a rapid increase in the number of cells attached at 6and 8 h. Significantly more cells attached to Si-CaP than toHA at 2 h (32 6 1.3 vs. 18 6 0.8 cells mm�2; p < 0.05)and 8 h (129 6 10.3 vs. 74 6 2.3 cells mm�2; p < 0.05).

Cell proliferationA time-dependent increase in cell number for cultures onSi-CaP, HA, PL, and PLþOS was observed up to 14 days; thisrate of proliferation is shown in Figure 1(B). On day 1, celldensity was similar for all substrates, after which cells pro-liferated at different rates on each substrate. The greatestincrease in cell number over time was observed for cellscultured on PL without osteogenic media. The addition ofosteogenic factors led to a decrease in the growth rate, withcell density being significantly lower for PLþOS than for PLat all time-points after day 1. However, this difference onlyachieved statistical significance at days 7 and 14 (p ¼ 0.02

and p ¼ 0.03, respectively). Cells cultured on Si-CaP and HAshowed a similar growth curve to cells on PLþOS; cell num-bers on both Si-CaP and HA were lower than PL at all time-points but the difference between Si-CaP and HA did notachieve statistical significance (p > 0.19).

Alkaline phosphataseTissue nonspecific ALP is an enzyme expressed by cells dur-ing osteogenesis.26 To assess the degree of differentiationalong the osteogenic pathway, ALP activity was estimatedby measurement of the area of ALP positive staining cellsafter by histochemical staining at different time points inculture on different substrates, and in the presence of osteo-genic supplements. The ALP positive area was low on cellscultured on plastic alone (PL), but was strongly ‘induced’ inPLþOS cultures, as shown in Figure 2(A): upper panel. Incontrast to cells maintained on PL alone, cells cultured onSi-CaP and HA in the absence of osteogenic supplementsshowed strong staining. Corresponding images of DAPI-stained cell nuclei are also shown [Fig. 2(A): lower panel].Cell numbers and ALP staining were comparable for cellscultured on Si-CaP, HA, and PLþOS. The ALP activity on PLwas lower even though the cell numbers were similar. TheALP positive fraction was significantly increased on Si-CaPand HA compared with PL (42.9% 6 9.3% and 32.9% 66.7% vs. 6.3% 6 8.5% total substrate area, respectively;p < 0.01) [Fig. 2(B)]. There was no statistically significantdifference between the ALP activity of cells cultured onSi-CaP versus HA; there was also no significant differencebetween either of these versus the positive control, cellscultured on PLþOS.

MineralizationAreas of mineralization (stained in black by the von Kossareagent, indicating calcium deposits) are shown in Figure2(C) upper panel, with the corresponding DAPI image in thelower panel. Quantification of the stained area [Fig. 2(D)]showed that the extent of mineralization increased withtime for all substrates except for PL, where no changes

FIGURE 1. Mesenchymal stem cell attachment and proliferation on substrates. (A) Number of cells attached to substrates over time normalized

to surface area. The results represent the mean 6 SD of three donors with samples run in triplicate. (B) Cell proliferation over time on substrates

and on PL in OS media, normalized to surface area. The results represent the mean 6 SD of three donors, with five samples run per donor.

Abbreviations; Tissue culture polystyrene (PL), silicon substituted calcium phosphate (Si-CaP) and Hydroxyapatite (HA) and Tissue culture poly-

styrene with osteogenic supplements (PLþOS).


from baseline levels were observed. No significant increasein mineralized deposits on HA was detected until day 14. Atthis time-point, the area covered with mineralized matrixwas lower for HA when compared with Si-CaP (20.8% 68.2% vs. 36.5% 6 12.3% total substrate area). In all sub-strates the area covered with mineralized matrix continuedto increase up to the furthest time point at day 21. At thistime-point, the area of mineralization on Si-CaP and HA wassignificantly greater than on PL (48.9% 6 14.5% and42.5% 6 11.0% vs. 1.2% 6 0.2% total substrate area,respectively; p < 0.05 in both comparisons).

Gene expressionCulturing MSCs on different substrates for up to 21 daysresulted in differential expression of osteoblast specificgenes. To account for differences in cell numbers on thevarious substrates, the expression levels were normalized tothe house keeping gene b2 microglobulin and the normal-ized values for the expression of a range of MSCs and osteo-blast specific genes are shown in Figure 3.

Endoglin (CD105) is a cell surface receptor commonlyassociated with MSCs, and it decreases in expression asthese cells differentiate. CD105 expression decreased on

Si-CaP and PLþOS, but did not change significantly on HAand PL [Fig. 3(A)]. Expression of CD105 was significantlyreduced on Si-CaP and PLþOS on day 7 and remained lowerthan PL at all time points thereafter. Cells cultured on HAexpressed lower levels on day 1 than compared to all othersubstrates; but did not decrease over time, and remained atsimilar levels to PL on day 21.

Runt related transcription factor 2 (RUNX2) has a keyrole in osteoblast differentiation. It binds to the promotersof a number of osteoblast-specific genes including collagentype 1 and osteocalcin, inducing their expression. As can beseen from Figure 3(B), RUNX2 expression was upreguatedin cells cultured on Si-CaP, HA, and PLþOS. High levels ofexpression were observed much earlier in cells cultured onSi-CaP and HA compared with PLþOS. RUNX2 levels werehighest on day 1 for both Si-CaP and HA, where they were1.6- and 1.4-fold greater than PL, respectively. Cells onPLþOS demonstrated a high level of RUNX2 expressionmuch later, on day 14, where they reached level 1.8-foldgreater than PL.

Parathyroid hormone receptor 1 (PTH1R) expression isassociated with endochondral ossification and plays a vitalrole in calcium homeostasis in bone.27 PTH1R expression is

FIGURE 2. Alkaline phosphatase and Von Kossa staining on substrates. Abbreviations; Tissue culture polystyrene (PL), silicon substituted

calcium phosphate (Si-CaP) and Hydroxyapatite (HA) and Tissue culture polystyrene with osteogenic supplements (PLþOS). (A) Alkaline phos-

phatase staining of differentiating cells on substrates at day 21 (upper panels), lower panel shows nuclear stain (DAPI) of same area to show

cell density. (Scale bar represents 200 lm). (B) ALP activity of cells on substrates quantified by analyzing the area of ALP positive cells using

image analysis software and the mean area calculated per disk/well. The results represent the mean 6 SD of three donors with samples run in

triplicate. (C) Von Kossa staining of calcium deposits in mineralization on substrates at day 21(top panel), lower panel shows same fields with

nuclear stain (DAPI) to show cell density (Scale bar represents 200 lm). (D) Mineralization on substrates was measured by analyzing the area of

black deposits using image analysis software and the mean area covered was calculated per disk/well. The results represent the mean 6 SD of

three donors with samples run in triplicate. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

ORIGINAL ARTICLE


associated with a mature, cuboidal cell, and that isexpressed as osteoblastic cells differentiate.28 PTH1R wasupregulated on Si-CaP, HA, and PLþOS however, the timingof expression was different on each substrate [Fig. 3(C)].Initially cells on Si-CaP expressed high levels of PTH1R, fol-lowed by PLþOS, then HA. Cells on Si-CaP significantly

upregulated PTH1R expression on days 1 and 7: 1.8- and3.6-fold greater than PL. However, cultures on HA andPLþOS demonstrated the greatest levels of expression onday 21, 2.4- and 5.6-fold greater than PL respectively. Cellson PL did not show any increase in PTH1R expression over21 days.

FIGURE 3. Real time gene expression profile of MSCs on substrates. Data are mRNA levels of genes of interest and are represented as level of

expression normalized to beta-2 microglobulin and to day 0. Three donors were used, and each sample was run in duplicate. Abbreviations;

Tissue culture polystyrene (PL), silicon substituted calcium phosphate (Si-CaP) and Hydroxyapatite (HA) and Tissue culture polystyrene with

osteogenic supplements (PLþOS).


An early marker of osteoblast differentiation and areporter of osteoblast activity, Collagen type I (Col1a1) wasincreased in cells on Si-CaP, HA, and PLþOS [Fig. 3(D)].Col1a1 expression peaked at different time points on eachsubstrate, first on Si-CaP, then HA and much later PLþOS. Incells on Si-CaP, Col1a1 was highly upregulated on day 7, 3.9-fold greater than PL. On HA Col1a1 expression was highestat day 14, 2.5-fold greater than PL. Levels of Col1a1 in cul-tures on PLþOS increased from day 14 and but peaked lateron day 21, to 2.6-fold greater than PL. Expression of Col1a1was maintained at a low level in cells cultured on PL.

Osteocalcin (BGLAP) is a late marker of osteoblast differ-entiation, expressed at the onset of mineralization. Therewere no significant differences in BGLAP expression on anyof the substrates [Fig. 3(E)]. Cells cultured on Si-CaP, showedthe greatest expression of BGLAP at day 14, however thisincrease did not reach statistical significance (p ¼ 0.988).

Dentin matrix acidic phosphoprotein 1 is encoded by theDMP1 gene and thought to be restricted to mineralized tis-sue,29 where it is produced by terminally differentiatedosteoblasts.30 DMP1 expression was low in cells on all sub-strates on days 1–14; however, on day 21 expression greatlyincreased in cells on Si-CaP (5.5-fold), HA (5.1-fold), andPLþOS (7.8-fold). Cells on PL expressed low levels of DMP1peaking at 0.9-fold on day 21, but this was significantlylower than cells on Si-CaP (p ¼ 0.03), HA (p ¼ 0.03), andPLþOS (p ¼ 0.003).

ImmunocytochemistryRunt-related transcription factor 2 (RUNX2) was detectedvia immunohistochemistry in the cells cultured on PLþOS,Si-CaP, and HA [Fig. 4(A)]. Cells cultured on PLþOSexpressed nuclear RUNX2 on days 7 and 14. Cells grown onSi-CaP and on HA showed the same nuclear pattern ofRUNX2 expression on day 7. However, on days 14 and 21RUNX2 appeared to be present also in the cytoplasm,although RUNX2 protein expression was predominantly nu-clear (in �91 and 87% of cells cultured on Si-CaP and onHA, respectively). RUNX2 was not detected in cells from anyof the PL samples.

Collagen type I protein expression in cells cultured onthe various substrates on days 1, 7, 14, and 21 is also shown[Fig. 4(B)]. On cells grown on Si-CaP and HA, COL1a1 pro-tein was detected as early as day 1, with expression increas-ing in intensity until day 21. Collagen type 1 was detectedin cells cultured on PL and PLþOS at all time-points; how-ever, its expression was less intense and less widespreadcompared with that of cells grown in Si-CaP and HA.

Osteocalcin was first detected in cells cultured on HA(day 7). On days 14 and 21, osteocalcin expression intensitywas strong and widespread in cells grown on Si-CaP, HA,and PLþOS [Fig. 4(C)]. No osteocalcin was detected at anytime-point in cells cultured on PL.

DISCUSSION

Here we have sought to determine whether the osteogenicdifferentiation of undifferentiated mesenchymal stem cellsfrom human marrow can be engendered in the absence of

known chemical drivers such as dexamethasone. We havesuccessfully employed inorganic substrates as a driver todifferentiation and suggest that the in vivo directed differen-tiation of undifferentiated MSCs contributes to the bonehealing properties of this class of materials. The purpose ofsynthetic bone graft materials is to provide a structuralframework that encourages bone repair and where the

FIGURE 4. Immunofluorescence staining of (A) RUNX2, (B) COL1A1,

and (C) BGLAP on Pl with OS media, PL, Si-CaP, and HA with growth

media. RUNX2 (green), DAPI (blue). Scale bar represents 100 lm.

Abbreviations; Tissue culture polystyrene (PL), silicon substituted

calcium phosphate (Si-CaP) and hydroxyapatite (HA) and Tissue

culture polystyrene with osteogenic supplements (PLþOS). [Color

figure can be viewed in the online issue, which is available at

wileyonlinelibrary.com.]

ORIGINAL ARTICLE


synthetic matrix is eventually replaced with normal healthybone. The target for any synthetic bone graft is to mimic oreven outperform the repair profile and time course ofnatural bone healing. An ideal bone graft substitute shouldprovide an environment to encourage attachment, and pro-liferation of multipotent/progenitor cells, and ideally, directsdifferentiation along the osteoblast lineage. Several studieshave investigated the behavior of committed progenitorsgrown on calcium phosphate based biomaterials.31–35 How-ever, MSCs are more likely than differentiated osteoblasts tomigrate to the implant surface.36 This interaction has notyet been studied yet it is an important aspect of osseointe-gration and critical to improving the design and success ofbone graft substitutes. In the present study, we examinedattachment, proliferation and osteogenic differentiation ofMSCs on different substrates (PL, Si-CaP, and HA) comparedwith culture in the presence of known osteogenic inducers(PLþOS). MSCs differentiated into osteoblasts on all sub-strates; however, the calcium phosphate based substratesinitiated osteogenic differentiation in the absence of knownosteogenic supplements. Cell culture plastics are designedto allow maximal cell adhesion. As predicted this substrate(PL) generated the most rapid cell adhesion. Comparing sili-cate-substituted with unmodified HA, our results demon-strated a greater number of MSCs attached to Si-CaP at alltime points compared with HA. These findings are in agree-ment with a previous work that demonstrated increasedcell attachment rates to biomaterials containing trace levelsof Si ions.37 It has been postulated that changes in surfacecharge and the addition of Si to the HA matrix38–40 servesto increase protein adsorption that, in turn, enhances cellattachment.20,37,41 Once attached to the surface of thematrix, the degree of cell colonization is dependent on pro-liferation. The balance between proliferation and differentia-tion must be precisely regulated in order for a bone graft tosuccessfully integrate with the surrounding tissue. If MSCswere to differentiate into committed cells upon attachmentwithout any proliferation there may not be enough cells toform sufficient new bone without further recruitment fromexternal sources. This may be relevant in therapeutic appli-cations involving preseeded material grafts.42 Conversely, if

MSCs continually proliferate and do not differentiate therealso will be a failure of new bone formation. In both casesthis will lead to poor osseointegration and essentially, graftfailure.

As expected, MSCs cultured on PL showed simple expo-nential growth while cells cultured in differentiation media(PLþOS) doubled at a much slower rate, consistent with amodel in which MSCs cease proliferation as differentiationstarts. Cells on Si-CaP and HA also proliferated at a slowrate, with an approximately linear rate of increase overtime. One interpretation of this linear growth rate is that itreflects asymmetric division of stem cells, in which onecommitted cell and one daughter stem cell are produced ateach division. This may be ideal, as it allows differentiationof committed cells in addition to providing a source ofprogenitors for further differentiation when required.

The process of osteogenic differentiation is complex, andmarked by the changing expression of key genes and pro-teins as the cell moves down the differentiation pathway,shown in Figure 5, from a multipotent MSC to a terminallydifferentiated osteocyte.

Alkaline phosphatase (ALP) is an early marker of osteo-blast differentiation. ALP was not present in undifferenti-ated MSCs at day 0. ALP was detected in cells cultured onPLþOS but not in cells cultured on PL alone. ALP wasstrongly expressed in cells cultured on both Si-CaP and HA,suggesting both these substrates are capable of differentiat-ing MSCs to osteoblasts. Mineralization as visualized usingthe Von Kossa assay was seen in PLþOS (positive control)but not in PL (negative control). The area of mineralizationwas similar on Si-CaP and HA to that on PLþOS. This dem-onstrates that both calcium based materials initiated theformation of a mineralized matrix. Comparison of ALP andmineralization levels on Si-CaP and HA suggest Si-CaP ismore osteoinductive than HA.

To determine further differences between Si-CaP andHA, osteoblast-related proteins, and key transcription fac-tors were analyzed. RUNX2 is the most specific osteoblastassociated transcription factor, controlling osteoblast com-mitment, differentiation and matrix mineralization. It oper-ates by binding to the osteoblast specific region of all major

FIGURE 5. Osteoblast differentiation. Growth and differentiation of osteoblasts, schematic diagram of markers and stages of osteoblast differen-

tiation. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]


osteoblast-related genes and controls their expression.RUNX2 gene expression was significantly increased on day 1on both Si-CaP and HA compared to PL. RUNX2 protein wasdetected later than gene expression but was maintaineduntil day 21. On days 14 and 21 RUNX2 expressionappeared to be both cytoplasmic and nuclear on Si-CaP andHA. This cytoplasmic staining may indicate RUNX2 has al-ready played its role in differentiation and gene expression,and has shuttled into the cytoplasm in its inactive form.However, some studies suggest this subnuclear localizationmay be required for maximal transcriptional activity.43,44

CD105 or Endoglin is a commonly recognized marker ofbone marrow derived MSCs,1,45–47 and its expressiondecreases as cells become differentiated. Cells cultured onPL maintained CD105 expression, whereas cells on Si-CaPand PLþOS showed a decrease in expression of CD105, indi-cating these cells have begun to differentiate along theosteoblast lineage. This was confirmed by the increase inosteoblast related genes RUNX2, COL1A1, BGLAP, and PTH1R.Type I collagen is the most abundant protein in bone. Pro-duction of this type of collagen peaks as cells becomemature extracellular matrix producing osteoblasts.20 COL1A1was upregulated in cells cultured on Si-CaP, HA, and PLþOSconsistent with the differentiation of cells on the substratesalong the osteoblast lineage. The highest levels of COL1A1were seen on Si-CaP on day 7, earlier than on HA andPLþOS, which appeared to peak on days 14 and 21, respec-tively. Type I collagen protein levels increased over time onSi-CaP but did not parallel the level of translation and posttranslational processing required to produce stable collagen,detected by immunocytochemistry. Cells cultured on HA andPLþOS produced type I collagen in a similar manner, whichwas more closely related to gene expression.

Parathyroid hormone 1 receptor (PTH1R), is expressedat peak levels in osteoblastic cells that are involved in ma-trix mineralization, in vitro.48 PTH1R was also upregulatedearlier on Si-CaP on day 7 compared to HA and PLþOS.Indicating cells in Si-CaP were differentiating at an increaserate compared to cells on HA or PLþOS.

Osteocalcin or bone gamma-carboxyglutamic acid-containing protein (BGLAP) is a late stage osteoblast specificprotein, produced only by mature osteoblasts prior to theonset of mineralization.49 Protein production of BGLAPreflected gene expression on Si-CaP, HA, and PLþOS on day14. It is understood that BGLAP gene expression is downre-gulated as cells become mature osteoblasts, this was seenon day 21, but was not reflected in protein levels.

DMP1 expression was analyzed to determine whethercells were capable of terminal differentiation along theosteoblast lineage into osteocytes. Cells cultured on PLþOSexpressed DMP1 on day 21, indicating osteocytes were pres-ent. However in the absence of supplementation, only cellscultured on Si-CaP and HA expressed high levels of DMP1.

Results from gene expression and protein productiondata demonstrate cells were differentiating into osteoblastson Si-CaP, HA, and PLþOS. However, Si-CaP upregulated theosteoblast related genes RUNX2, COL1A1, and PTH1R earlierthan HA and PLþOS, suggesting that cells cultured on

Si-CaP were differentiating along the osteoblast lineage ear-lier than on HA or PLþOS. This elevated rate of differentia-tion may be due to an increase in rate of differentiation inthe same percentage of cells, or due to a greater percentageof cells differentiating. However, similar numbers of cellssuggest that differentiation is occurring at an increased rate(Figs. 2 and 4), Furthermore, this is also reflected in in vivostudies suggesting this response may be due to the tracelevels of silicon in Si-CaP.50

Previous studies have demonstrated the upregulation ofosteoblast-specific genes on calcium phosphate-based sub-strates. However, such studies have been carried out usingpartially or fully differentiated cells, or when additionallysupplementing the cell culture with known osteogenic sup-plements.21,49,51–54 It is important to note that MSCs differ-entiation towards the osteoblastic lineage in this study isthe result of the substrate properties and not the result ofsoluble osteogenic supplements.

Here we demonstrated MSCs cultured on Si-CaP and HA,in the absence of the known osteogenic driver dexametha-sone, differentiated along the osteogenic lineage as deter-mined by (i) increased ALP activity, (ii) downregulation theMSC associated gene CD105, (iii) up-regulation of early osteo-blast related genes, RUNX2, COL1A1 and late osteoblast genes,PTH1R and BGLAP, (iv) up-regulation of osteocyte-specificgene DMP1, (v) increased protein production of RUNX2,COL1A1, and BGLAP, (vi) and produced a mineralized matrixat a greater rate than cells cultured in the presence of osteo-genic supplementation on tissue culture polystyrene.

The potential implication of these findings for tissueengineering is significant; indicating these substrates maybe utilized to control stem cell after in vivo and in vitro,without the need for osteogenic supplementation. Further-more, the increased rate of differentiation seen on Si-CaPmay enable the development of novel substrates for osteo-genic differentiation of MSC, which will have an enormousimpact in regenerative medicine.

REFERENCES1. Chamberlain G, Fox J, Ashton B, Middleton J. Concise review:

Mesenchymal stem cells: Their phenotype, differentiation

capacity, immunological features, and potential for homing. Stem

Cells 2007;25:2739–2749.

2. Pittenger MF. Multilineage potential of adult human mesenchy-

mal stem cells. Science 1999;284:143–147.

3. Beresford J, Joyner C, Devlin C. The effects of dexamethasone and

1, 25-dihydroxyvitamin D3 on osteogenic differentiation of human

marrow stromal cells in vitro. Arch Oral Biol 1994;39:941–947.

4. Ziros PG, Basdra EK, Papavassiliou AG. Runx2: Of bone and

stretch. Int J Biochem Cell Biol 2008;40:1659–1663.

5. Shui C, Spelsberg TC, Riggs BL, Khosla S. Changes in Runx2/Cbfa1

expression and activity during osteoblastic differentiation of human

bonemarrow stromal cells. J BoneMiner Res 2003;18:213–221.

6. Komori T. Regulation of bone development and extracellular ma-

trix protein genes by RUNX2. Cell Tissue Res 2010;339:189–195.

7. Makita N, Suzuki M, Asami S, Takahata R, Kohzaki D, Kobayashi

S, Hakamazuka T, Hozumi N. Two of four alternatively spliced iso-

forms of RUNX2 control osteocalcin gene expression in human

osteoblast cells. Gene 2008;413:8–17.

8. Glimcher MJ. Recent studies of the mineral phase in bone and its

possible linkage to the organic matrix by protein-bound phos-

phate bonds. Philos Trans R Soc Lond B Biol Sci 1984;304:

479–508.

ORIGINAL ARTICLE


9. Stein GS, Lian JB, Owen TA. Relationship of cell growth to the

regulation of tissue-specific gene expression during osteoblast

differentiation. FASEB J 1990;4:3111–3123.

10. Zhao G, Raines A, Wieland M, Schwartz Z, Boyan B. Requirement

for both micron- and submicron scale structure for synergistic

responses of osteoblasts to substrate surface energy and topogra-

phy. Biomaterials 2007;28:2821–2829.

11. Dallas S. Dynamics of the transition from osteoblast to osteocyte.

Ann N Y Acad Sci 2010;1192:437–443.

12. Lavery K, Hawley S, Swain P, Rooney R, Falb D, Alaoui-Ismaili MH.

New insights into BMP-7 mediated osteoblastic differentiation of

primary human mesenchymal stem cells. Bone 2009;45:27–41.

13. Maegawa N, Kawamura K, Hirose M, Yajima H, Takakura Y,

Ohgushi H. Enhancement of osteoblastic differentiation of mesen-

chymal stromal cells cultured by selective combination of bone

morphogenetic protein-2 (BMP-2) and fibroblast growth factor-2

(FGF-2). J Tissue Eng Regen Med 2007;1:306–313.

14. Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds

and osteogenesis. Biomaterials 2005;26:5474–5491.

15. Dorozhkin SV. Bioceramics of calcium orthophosphates. Biomate-

rials 2010;31:1465–1485.

16. Pietak AM, Reid JW, Stott MJ, Sayer M. Silicon substitution in the

calcium phosphate bioceramics. Biomaterials 2007;28:4023–4032.

17. Schwarz K, Milne DB. Growth-promoting effects of silicon in rats.

Nature 1972;239:333–334.

18. Carlisle EM. Silicon: An essential element for the chick. Science

1972;178:619–621.

19. Hing K, Saeed S, Annaz B, Buckland T. Silicate substitution alters

the progression of bone apposition within porous hydroxyapatite

bone graft substitutes. Trans 50th A. Meeting Orthopaedic

Research Society, San Francisco, CA, vol. 29, poster 1041. Rose-

mont, IL: Orthopaedic Research Society, 2004.

20. Anselme K. Osteoblast adhesion on biomaterials. Biomaterials

2000;21:667–681.

21. Chen P-Y, Huang LLH, Hsieh H-J. Hyaluronan preserves the prolif-

eration and differentiation potentials of long-term cultured murine

adipose-derived stromal cells. Biochem Biophys Res Commun

2007;360:1–6.

22. Chai C, Leong KW. Biomaterials approach to expand and direct

differentiation of stem cells. Mol Ther 2007;15:467–480.

23. Tremoleda J, Forsyth N, Khan N. Bone tissue formation from

human embryonic stem cells in vivo. Cloning Stem Cells

2008;10:119–132.

24. Abr�amoff MD, Magalhaes PJ, Ram Sunanda J. Image processing

with ImageJ. Biophotonics 2004;11:36–42.

25. Sambrook J. A Single-step method for the simultaneous prepara-

tion of DNA, RNA, and protein from cells and tissues. Cold Spring

Harbor Protocols 2006; doi:10.1101/pdb.prot4056.

26. Malaval L, Modrowski D, Gupta AK, Aubin JE. Cellular expression

of bone-related proteins during in vitro osteogenesis in rat bone

marrow stromal cell cultures. J Cell Physiol 1994;158:555–572.

27. Alonso V, Ardura J, Wang B. A naturally occurring isoform inhibits

parathyroid hormone receptor trafficking and signaling. J Bone

Miner Res 2011;26:143–155.

28. McCauley LK, Koh AJ, Beecher CA, Cui Y, Rosol TJ, Franceschi

RT. PTH/PTHrP receptor is temporally regulated during osteoblast

differentiation and is associated with collagen synthesis. J Cell

Biochem 1996;61:638–647.

29. Kalajzic I, Braut A, Guo D, Jiang X, Kronenberg MS, Mina M, Har-

ris MA, Harris SE, Rowe DW. Dentin matrix protein 1 expression

during osteoblastic differentiation, generation of an osteocyte

GFP-transgene. Bone 2004;35:74–82.

30. Toyosawa S, Shintani S, Fujiwara T. Dentin matrix protein 1 is pre-

dominantly expressed in chicken and rat osteocytes but not in osteo-

blasts—Toyosawa—2001—Journal of Bone andMineral Research—

Wiley Online Library. J BoneMiner Res 2001;16:2017–2026.

31. Tortelli F, Pujic N, Liu Y, Laroche N, Vico L, Cancedda R. Osteo-

blast and osteoclast differentiation in an in vitro three-dimen-

sional model of bone. Tissue Eng A 2009;15:2373–2383.

32. Ramaswamy Y, Wu C, Zhou H, Zreiqat H. Biological response of

human bone cells to zinc-modified Ca-Si-based ceramics. Acta

Biomater 2008;4:1487–1497.

33. Langstaff S, Sayer M, Smith TJ, Pugh SM. Resorbable bioceram-

ics based on stabilized calcium phosphates. Part II: Evaluation of

biological response. Biomaterials 2001;22:135–150.

34. Sun H, Wu C, Dai K, Chang J, Tang T. Proliferation and osteoblas-

tic differentiation of human bone marrow-derived stromal cells on

akermanite-bioactive ceramics. Biomaterials 2006;27:5651–5657.

35. Annaz B, Hing KA, Kayser M, Buckland T, Di Silvio L. Porosity var-

iation in hydroxyapatite and osteoblast morphology: A scanning

electron microscopy study. J Microsc 2004;215 (Part 1):100–110.

36. Olivares-Navarrete R, Hyzy SL, Hutton DL, Erdman CP, Wieland

M, Boyan BD, Schwartz Z. Direct and indirect effects of micro-

structured titanium substrates on the induction of mesenchymal

stem cell differentiation towards the osteoblast lineage. Biomate-

rials 2010;31:2728–2735.

37. Zou S, Ireland D, Brooks RA, Rushton N, Best S. The effects of

silicate ions on human osteoblast adhesion, proliferation, and differ-

entiation. J Biomed Mater Res B Appl Biomater 2009;90:123–130.

38. Porter AE, Buckland T, Hing K, Best SM, Bonfield W. The structure

of the bond between bone and porous silicon-substituted hydroxy-

apatite bioceramic implants. J Biomed Mater Res 2006;78:25–33.

39. Guth K, Buckland T. Silicon dissolution from microporous silicon

substituted hydroxyapatite and its effect on osteoblast behavior.

KEM 2006;309-311:117–120.

40. Rashid N, Harding I, Hing KA. Effect of silicate substitution on the

surface charge of hydroxyapatite. Transactions—7th World Bio-

materials Congress, Sydney, 2004.

41. Ergun C, Webster TJ, Bizios R, Doremus RH. Hydroxylapatite with

substituted magnesium, zinc, cadmium, and yttrium. I. Structure

and microstructure. J Biomed Mater Res 2002;59:305–311.

42. Owen M. Marrow stromal stem cells. J Cell Sci Suppl 1988.

43. Pockwinse SM, Rajgopal A, Young DW, Mujeeb KA, Nickerson J,

Javed A, Redick S, Lian JB, van Wijnen AJ, Stein JL, Stein GS,

Doxsey SJ. Microtubule-dependent nuclear-cytoplasmic shuttling

of Runx2. J Cell Physiol 2005;206:354–362.

44. Kim S, Koga T, Isobe M, Kern BE, Yokochi T, Chin YE, Karsenty

G, Taniguchi T, Takayanagi H. Stat1 functions as a cytoplasmic at-

tenuator of Runx2 in the transcriptional program of osteoblast dif-

ferentiation. Genes Dev 2003;17:1979–1991.

45. Mareddy S, Crawford R, Brooke G, Xiao Y. Clonal isolation and

characterization of bone marrow stromal cells from patients with

osteoarthritis. Tissue Eng 2007;13:819–829.

46. Kolf CM, Cho E, Tuan RS. Mesenchymal stromal cells. Biology of

adult mesenchymal stem cells: Regulation of niche, self-renewal

and differentiation. Arthritis Res Ther 2007;9:204.

47. Yoshimura K, Shigeura T, Matsumoto D, Sato T, Takaki Y, Aiba-

Kojima E, Sato K, Inoue K, Nagase T, Koshima I, Gonda K. Characteri-

zation of freshly isolated and cultured cells derived from the fatty and

fluid portions of liposuction aspirates. J Cell Physiol 2006;208:64–76.

48. Koh AJ, Beecher CA, Rosol TJ, McCauley LK. 3‘‘,5-’’Cyclic adeno-

sine monophosphate activation in osteoblastic cells: Effects on

parathyroid hormone-1 receptors and osteoblastic differentiation

in vitro. Endocrinology 1999;140:3154–3162.

49. Marinucci L, Balloni S, Becchetti E, Bistoni G, Calvi EM, Lumare E,

Ederli F, Locci P. Effects of hydroxyapatite and Biostite on osteo-

genic induction of hMSC. Ann Biomed Eng 2010;38:640–648.

50. Hing KA, Wilson LF, Buckland T. Comparative performance of

three ceramic bone graft substitutes. Spine J 2007;7:475–490.

51. Kaur G, Valarmathi MT, Potts JD, Jabbari E, Sabo-Attwood T, Wang

Q. Regulation of osteogenic differentiation of rat bone marrow stro-

mal cells on 2Dnanorod substrates. Biomaterials 2010;31:1732–1741.

52. Sundelacruz S, Levin M, Kaplan DL. Membrane potential controls

adipogenic and osteogenic differentiation of mesenchymal stem

cells. PLoS ONE 2008;3:e3737.

53. GargiuloC, ThaoH, TuanH, ThuyT. In vitro culture anddifferentiation

of osteoblasts on coral scaffold fromhumanbonemarrowmesenchy-

mal stem cells. The Third International Conference on the Develop-

ment of Biomedical Engineering in Vietnam, January 11, 2010.

54. Schneider RK, Puellen A, Kramann R, Raupach K, Bornemann J,

Knuechel R, P�erez-Bouza A, Neuss S. The osteogenic differentia-

tion of adult bone marrow and perinatal umbilical mesenchymal

stem cells and matrix remodelling in three-dimensional collagen

scaffolds. Biomaterials 2010;31:467–480.


directed osteogenic differentiation of human mesenchymal stem/precursor cells on silicate...

Documents