lable at ScienceDirect

ARTICLE IN PRESS

Biomaterials xxx (2009) 1–11

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

A biaxial rotating bioreactor for the culture of fetal mesenchymal stem cellsfor bone tissue engineering

Zhi-Yong Zhang a,b, Swee Hin Teoh b,c, Woon-Shin Chong d, Toon-Tien Foo d, Yhee-Cheng Chng d,Mahesh Choolani e, Jerry Chan e,*

a Graduate Program in Bioengineering (GPBE), National University of Singapore, Singaporeb Centre for Biomedical Materials Applications and Technology (BIOMAT), Department of Mechanical Engineering, Faculty of Engineering, National University of Singapore, Singaporec National University of Singapore Tissue Engineering Programme (NUSTEP), National University of Singapore, Singapored Bioengineering Laboratory, Technology Centre for Life Sciences, Singapore Polytechnic, Singaporee Experimental Fetal Medicine Group, Department of Obstetrics and Gynaecology, Yong Loo Lin School of Medicine, National University of Singaporeand National University Hospital, Singapore

a r t i c l e i n f o

Article history:Received 2 November 2008Accepted 13 January 2009Available online xxx

Keywords:BioreactorFetal mesenchymal stem cellBone tissue engineeringPolycaprolactoneNOD/SCID mice

* Corresponding author. Experimental Fetal MedObstetrics and Gynaecology, Yong Loo Lin School of MeSingapore, 5 Lower Kent Ridge Road, Singapore 1190þ65 6779 4753.

E-mail address: jerrychan@nus.edu.sg (J. Chan).

0142-9612/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.biomaterials.2009.01.028

Please cite this article in press as: Zhang Z-Yengineering, Biomaterials (2009), doi:10.101

a b s t r a c t

The generation of effective tissue engineered bone grafts requires efficient exchange of nutrients andmechanical stimulus. Bioreactors provide a manner in which this can be achieved. We have recentlydeveloped a biaxial rotating bioreactor with efficient fluidics through in-silico modeling. Here weinvestigated its performance for generation of highly osteogenic bone graft using polycaprolactone–tricalcium phosphate (PCL–TCP) scaffolds seeded with human fetal mesenchymal stem cell (hfMSC).hfMSC scaffolds were cultured in either bioreactor or static cultures, with assessment of cellular viability,proliferation and osteogenic differentiation in vitro and also after transplantation into immunodeficientmice. Compared to static culture, bioreactor-cultured hfMSC scaffolds reached cellular confluence earlier(day 7 vs. day 28), with greater cellularity (2�, p< 0.01), and maintained high cellular viability in thecore, which was 2000 mm from the surface. In addition, bioreactor culture was associated with greaterosteogenic induction, ALP expression (1.5� p< 0.01), calcium deposition (5.5�, p< 0.001) and bonynodule formation on SEM, and in-vivo ectopic bone formation in immunodeficient mice (3.2�, p< 0.001)compared with static-cultured scaffolds. The use of biaxial bioreactor here allowed the maintenance ofcellular viability beyond the limits of conventional diffusion, with increased proliferation and osteogenicdifferentiation both in vitro and in vivo, suggesting its utility for bone tissue engineering applications.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Bone tissue engineering provides a promising approach toaddress the significant drawbacks of existing bone grafts, whichinclude: firstly, the limited availability of bone tissue and donor-sitemorbidity associated with autografts [1,2]; secondly, the significantrisk of disease transmission and immune reaction arising from theuse of allografts [3]; and lastly, the lack of remodeling and subse-quent fatigue-associated graft failure with the use synthetic grafts[4]. In bone tissue engineering, biodegradable porous scaffolds areseeded with osteogenic cell types to develop an in-vitro-maturedengineered cellular bone graft, which can be fashioned into

icine Group, Department ofdicine, National University of

74. Tel.: þ65 6772 2672; fax:

All rights reserved.

et al., A biaxial rotating biore6/j.biomaterials.2009.01.028

different shapes and sizes, stimulate bone healing and remodelaccordingly [5].

Three dimensional (3D) scaffolds provide the necessary supportfor cells to attach, proliferate and differentiate, and define theoverall shape of the tissue engineered transplant [6]. Scaffoldsmade of polycaprolactone (PCL) have recently been shown topossess favorable properties for load bearing bone tissue engi-neering application compared with other materials such as PLGA[7]. They have a slower degradation speed with structural degra-dation kinetics of only 7% over a six month period in vivo [8],maintaining a sustained period of mechanical support. In additionthe slow degradation kinetics would result in a reduced risk ofacidosis from the rapid accumulation of acidic by-products, sug-gesting its utility for bone tissue engineering applications.

Mesenchymal stem cells (MSCs) are increasingly being used asa cellular source for osteogenic tissue engineering applications, dueto their ease of isolation and well defined osteogenic differentiationpathways [9,10]. However, the use of human adult MSC has been

actor for the culture of fetal mesenchymal stem cells for bone tissue

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–112

ARTICLE IN PRESS

handicapped by their low frequencies, slow proliferation time andgenerally limited proliferation capacity [11]. More recently, humanfetal MSC (hfMSC) has been characterized, with significantly higherproliferation capacity and reduced immunogenicity whencompared to their adult counterparts [12–15]. In a head-to-headcomparison, we have showed that hfMSC demonstrates significantproliferative and osteogenic advantages over MSC types derivedfrom the umbilical cord, adult adipose tissue or adult bone marrowin 3D culture systems both in vitro and in vivo [16].

The simple loading of osteogenic cell sources to suitable scaf-folds to generate a bone graft however has largely been limited bythe challenge of maintaining cellular viability at the core wherelarger grafts are concerned. This is due to the limits of diffusion ofboth nutrients and waste products, generally accepted to be around1–200 mm [17,18]. In addition, homogenous cellular seeding andextracellular matrix distribution become inefficient where largergrafts are concerned [19,20]. One way of overcoming these limi-tations is with the use of bioreactors. Bioreactors enable improvedflow perfusion of cellular scaffolds, which in turn improves masstransfer efficiencies. In addition, they provide a mechanical stim-ulation which triggers the mechano-transduction signalingpathway important for osteogenic differentiation in MSC cell types[19,21,22].

While most of these commercially available bioreactors areuniaxial in design, we have recently shown that a biaxial designwhich rotates simultaneously in two independent orthogonal axesresulted in improved fluidics over an uniaxial design through in-silico simulations [23]. Here we investigated the use of this biaxialbioreactor system for the in-vitro maturation of high-porosity PCL/TCP scaffolds seeded with hfMSC to generate bone tissue engi-neered grafts. We report that the incorporation of a biaxial biore-actor maturation step resulted in enhanced homogenous cellularproliferation, ECM distribution and osteogenic differentiation, andmaintained high cellular viability in the core of the scaffolds.

2. Materials and methods

2.1. Samples, animals and ethics

Fetal tissue collection was approved by the Domain Specific Review Board ofNational University Hospital, Singapore in compliance with international guidelinesregarding the use of fetal tissue for research [24]. Pregnant women gave separatewritten consent for the clinical procedure and for the use of fetal tissue for researchpurposes. And fetal tissues were collected from fetuses after clinically indicatedtermination of pregnancy. Fetal gestational age was determined by crown-rumplength measurement. Two fetal samples at 10þ6 and 14þ2 (weeksþ days) gestationswere utilized for this study.

Eight weeks old immunodeficient male NOD/SCID mice were acquired throughCharles Rivers, Australia, and all procedures were approved by the Institutional AnimalCare and Use Committee (IACUC) at National University of Singapore. All materialsused were purchased from Sigma–Aldrich (Singapore) unless otherwise stated.

2.2. Isolation and culture of hfMSCs

Bone marrow derived hfMSCs were isolated as previously described [25]. Briefly,single-cell suspensions of fetal bone marrow were prepared by flushing the bonemarrow cells out of the humeri and femurs using a 22-gauge needle into Dulbecco’smodified Eagle’s medium (DMEM, Sigma, USA)–Glutamax (GIBCO, USA) supple-mented with 10% fetal bovine serum (FBS), 50 U/ml penicillin, and 50 mg/mlstreptomycin (GIBCO, USA), which will be referred to as D10 medium, and plated on100 mm dishes at 106 mono-nuclear cells/ml in the D10 medium. Medium waschanged every 2–3 days and non adherent cells were removed. Cells were trypsi-nized and replated at 104/cm2 at sub-confluence. hfMSCs at Passage 3 were used forboth characterization and the experiments described here. Both samples yieldedsimilar results, and data presented here are representative of the experiments.

2.3. Characterization of hfMSC

2.3.1. ImmunophenotypehfMSCs were characterized by immunocytochemistry for CD14, CD34, CD45,

CD31, von Willebrand factor (vWF) (Abcam, USA), CD105 (SH2), CD73 (SH3, SH4)

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

(Abcam, USA), vimentin, laminin, CD29 (Chemicon, USA), CD44 (BD, USA), CD 106,CD 90 (Chemicon, USA), HLA I, HLA II (Dako, USA),Oct-4 (Abcam, USA) and Nanog(Abcam, USA) while flow-cytometry was used to screen for Stro-1 (Chemicon, USA)as previously described [16,25].

2.3.2. Tri-lineage mesenchymal differentiationIn order to assay the tri-lineage differentiation of the hfMSC, standard osteo-

genic, adipogenic and chondrogenic differentiation protocols were utilized and theresults were confirmed with von Kossa staining for osteogenic differentiation, OilRed staining for adipogenic differentiation and Safranin O staining for chondrogenicdifferentiation, as previous reported [16,26].

2.4. Scaffold manufacturing and surface treatment

We have previously described the generation of a high-porosity bioactive PCL–tricalcium phosphate (TCP) composite scaffold which would allow for rapid vascu-larization and maintenance of the structural integrity through a honeycombstructure, while allowing customisation of the scaffold to fit any particular shapeand size [26]. PCL–TCP 3D bioactive scaffold specimens were fabricated using thefused deposition modeling techniques as previously described [27]. A lay-downpattern of 0/60/120� was utilized to give a honeycomb-like pattern of triangularpores with a porosity of 70% and average pore size of 0.523 mm. A scaffold size of6� 6� 4 mm was used in these experiments (Fig. 3A).

The scaffolds were surface treated in 5 M NaOH for 3 h to enhance their hydro-philicity. After rinsing 3 times with PBS, they were sterilized in 70% ethanol for 24 h,and rinsed twice in PBS, and then transferred into an incubator at 37 �C for 24 h fordrying.

2.5. Seeding hfMSC to PCL–TCP scaffold

hfMSCs (Passage 3) were seeded onto the porous scaffolds by adding 50 ml of cellsuspension media with 5�105 cells to each scaffold (5�104/ml) in a dropwisemanner, placed in 24-well culture plates, and incubated for 3 h in an incubator.Thereafter, an additional 2 ml of D10 medium was slowly added to each well andhfMSC cellular scaffolds were incubated in a humidified atmosphere at 37 �C and 5%CO2 for one week with D10 medium changed 3 times a week to allow for cellattachment to scaffolds (Fig. 1A).

2.6. Biaxial rotating bioreactor culture vs. static culture

After one week of static culture, scaffolds were randomly divided into twogroups: (A) a bioreactor group where 45 scaffolds were transferred to the biaxialrotating bioreactor loaded with a total of 500 ml of osteogenic inductive mediumwhich fills the vessel and the reservoir. This medium was changed once a week fora total of four weeks, with 11.1 ml of medium per scaffold per week; and (B) thestatic culture group where scaffolds were transferred to new 24-well plates with3.5 ml of osteogenic inductive medium per well which was changed three timesa week for four weeks, with 10.5 ml of medium made available to each scaffoldper week (Fig. 1A). No major change of pH (Phenol Red) occurred in the culturemedium during the experiments to suggest the depletion of nutrients from eithergroup.

This biaxial rotation bioreactor system consists of a spherical vessel forculture (volume 500 ml), where the cellular-scaffold constructs are anchored tothe lip of bioreactor by pins, and a medium reservoir (500 ml), which allows thecontinuously replenishing medium and the real time monitoring and control ofoxygen, pH, and temperature. The spherical vessel and reservoir are connectedby tubings to form a perfusion system with medium flow circulating betweeneach other (as indicated in red arrow) (Fig. 1B). The entire bioreactor was placedin an incubator with humidified atmosphere at 37 �C and 5% CO2. Gaseousexchange was enabled through a special membrane incorporated into thespherical vessel. The spherical vessel was programmed to rotate in twoperpendicular axes (X and Z), with both axial rotation set to 5 rpm. This settingachieves a perfusion flow rate of 3.8 ml/min (Fig. 1B), allowing a completechange of medium every 132 min. A rotational speed of 5 rpm was chosen as itresulted in the greatest cellular proliferation in preliminary experimentsinvestigating rotational speeds between 3 and 15 rpm.

2.7. Cellular adhesion, viability and proliferation of hfMSC cellular scaffolds

The morphology of the cell in 3D culture, cellular adhesion and extracellularmatrix (ECM) production were examined daily by phase contrast light microscope(PCLM) over 28 days. Scaffolds were examined in two perpendicular panels, inplanar (top) and side view points (Fig. 3B).

The qualitative analysis of cell viability in 3D was performed via fluoresceindiacetate/propidium iodide (FDA/PI) staining, where FDA stains viable cells green,and PI stains necrotic and apoptotic cell nuclei red. Scaffolds were bisected in half toexpose the centre of the scaffold to achieve a core view of the scaffold (Fig. 4A),stained with FDA/PI as previously described [16], and viewed under a confocal laser

actor for the culture of fetal mesenchymal stem cells for bone tissue

A

B

Pre-culture for 1 week

“Static” vs. “Bioreactor”(4 wks)

In vitro study

Loading cells to scaffolds:

“Static” vs. “Bioreactor” (2 wks)

In vivo study

In vivo implantation(12 wks)

Week 0

Week 2

Week 4

Week 16

Experimental design

Fig. 1. Experimental and bioreactor design: A) Schematic representation of the study design; and B) Biaxial bioreactor design: The bioreactor system consisted of a spherical culturevessel (volume of 500 ml) connected to the medium reservoir through tubings through which a perfusion flow is generated (flow direction indicated by red arrows). The sphericalvessel sits on an articulator which allows rotation in two perpendicular axes (X and Z).

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–11 3

ARTICLE IN PRESS

microscope (Olympus, FV300 Fluoview, Japan). Cellular scaffolds were examined inboth planar view and side view on days 14 and 28 (Fig. 4A).

The total cell number in the 3D cellular scaffold on days 0 (day of transfer tothe bioreactor or static culture conditions, as illustrated by Fig. 1A) 14, 28 (n¼ 3)was estimated by quantifying the dsDNA content of each scaffold using a Pico-Green dsDNA Quantification Kit (Molecular Probes, USA). The total DNA wasextracted from each cellular scaffold by incubating the cellular scaffolds in 0.4 mlenzymatic cocktail (consisting of 0.1% collagenase A (Roche) with 0.1% Trypsinmixed in PBS) at 37 �C for 2 h, with vortex every 30 min then followed by threecycles of freeze and thaw; and assayed by following the manufacturer’s instruc-tion. The proliferation of the hfMSC inside 3D scaffold was interpreted by thechanges of dsDNA amount.

2.8. Comparison of osteogenic differentiation and mineralization in 3D scaffoldculture

2.8.1. ALP activity assayThe intracellular ALP activities of hfMSC cellular scaffolds under two culture

conditions were compared on days 0, 7, 14, 28 (n¼ 3, triplicates). Cell lysates weretested for ALP activity using SensoLyte� pNPP Alkaline Phosphatase Assay Kit (AnaS-pec USA) and the ALP activities were normalized to the total protein content deter-mined using the Bradford assay (Bio-Rad Laboratories, US) as previously described [16].

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

2.8.2. von Kossa staininghfMSC cellular scaffolds on days 14 and 28 were stained with von Kossa staining.

Briefly samples were gently rinsed twice with PBS then fixed in 10% formalin for 1 h,and washed in dH2O. Finally, they were stained with freshly made 2% silver nitrate indH2O (w/v) for 10 min in the dark and exposed to an incandescent lamp (100 W) for30 min.

2.8.3. Calcium content assayThe calcium content of the hfMSC cellular scaffolds on days 7, 14 and 28 (n¼ 3)

was assayed as previously described [16]. Briefly, the calcium deposition is dissolvedin 0.4 ml 0.5 N acetic acid and determined by a colorimetric assay using calciumassay kit (BioAssay Systems, USA). And control cell-free empty scaffolds cultured asabove (n¼ 3 per time point) were used as a negative control to offset the elution ofcalcium from the tricalcium phosphate component in the scaffold (27.5� 4.3 at day7; 29.7� 5.1 at day 14; 38.8� 6.4 at day 28 mg/acellular scaffold).

2.8.4. Scanning electron microscope (SEM) and energy dispersive X-rayspectrometer (EDX) analysis

After micro-CT scanning, the same 3 scaffolds were dehydrated, gold sputtered,viewed under the SEM and element component of crystal-like structure inside thesamples was analyzed by EDX as previously described [16].

actor for the culture of fetal mesenchymal stem cells for bone tissue

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–114

ARTICLE IN PRESS

2.9. Comparison of ectopic bone formation after subcutaneous implantation

2.9.1. Cellular-scaffold constructs’ preparationhfMSCs were seeded onto PCL–TCP scaffolds pre-cultured in D10 medium for

one week and pre-differentiated in osteogenic differentiation medium under eitherbioreactor culture or static culture condition for two weeks before implantation (asillustrated in Fig. 1A).

2.9.2. Surgical procedureAfter inducing general anesthesia, a midline longitudinal skin incision was made

on the dorsal surface of each mouse, and subcutaneous pockets created, into whichthe scaffolds were inserted. The skin was closed with interrupted 6-0 vicryl sutures.After three months, animals were euthanized, and the implants retrieved forhistological and micro-CT analysis (Fig. 1A).

2.9.3. HistologyCellular-scaffold constructs (n¼ 3) from each group were embedded in OCT

medium (Tissue-Tek, USA), and sectioned at 10 mm thickness and laid onto poly-L-lysine slides. Sections were stained with Masson Trichrome and von Kossa (coun-terstained with nuclear fast red) to ascertain evidence of bone formation andmineralization respectively.

2.9.4. Human to mouse chimerismHuman-specific lamins A/C immunostaining was used to investigate chimerism

of human cells in murine tissue as previously described [25,28]. Briefly, sectionsfrom each sample were blocked with 5% normal goat serum for 2 h, and left to reactwith monoclonal mouse anti-human Lamin A/C antibody (1:100, Vectorlabs, UK)overnight at 4 �C; sections were then incubated with goat anti-mouse secondaryantibodies (1:100 Alexaflour 488, Invitrogen, UK) for 1 h, and counterstained withpropidium iodide (PI). Images were visualized through confocal microscopy asdescribed above. The number of human and murine cells within scaffolds (n¼ 3)was enumerated manually on six low powered fields (LPFs) to calculate the rate ofchimerism of human cells. A mean of 217 cells was counted for each LPF (range 157–313) of each specimen.

2.9.5. Micro-CT analysisAfter cellular scaffolds (n¼ 3) have been harvested and fixed in 10% neutral

buffered formalin, they were placed in the sample holder and scanned through 180�

with a rotation step of 1� at a spatial resolution of 35 mm. An averaging of 5 anda 1 mm aluminum filter were used during the scanning. The scan files werereconstructed at a step size of 1 using a modified Feldkamp algorithm as provided bySkyscan (Belgium). The reconstructed data were loaded onto the 3D modelingsoftware, VGstudio (Volume Graphics GmbH, Germany) to stack the 2D image intoa 3D model for quantitative histomorphometric analysis. Cellular scaffolds wereimaged together with transplanted acellular scaffolds, un-transplanted emptyscaffold and bone tissue. The threshold value between un-transplanted emptyscaffold and known bone tissue was used for delineating ectopic bone from thescaffolds through visual and histograph analysis, and a value of 195 was chosen.Three acellular scaffold implants were utilized as negative controls to cancel out theinfluence of the TCP component within the scaffold on mineralization.

2.10. Statistical analyses

Parametric data have been represented as mean� SD, and compared usingeither two-way ANOVA or Student t-test. A value of p< 0.05 was taken as significant.

3. Results

3.1. Characterization of hfMSC

Culture expanded hfMSC grew as plastic-adherent spindle-shaped cells (Fig. 2A), with 45.1% of hfMSC being positive for theosteoprogenitor marker Stro-1. (Fig. 2B). hfMSC revealed animmunophenotype which was negative for haemopoietic andendothelial markers CD14, CD34, CD45, CD31, vWF and HLA II andpositive for mesenchymal markers CD105 (SH2), CD73 (SH3, SH4);intracellular marker vimentin and laminin; cell adhesion moleculesCD29, CD44, CD 106, CD 90; and HLA I as previously reported[25,28]. Furthermore, a small proportion of hfMSC was positive forthe embryonic stem cell markers Oct-4 (25.6�1.4% positive) andNanog (34.7�2.1% positive) (Fig. 2C). Under permissive cultureconditions, they underwent osteogenic, adipogenic and chondro-genic differentiation (Fig. 2D), confirming their bona-fide MSCphenotype.

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

3.2. Bioreactor culture resulted in homogenous cellular distribution,rapid proliferation and maintenance of cellular viability at thescaffold core

Under bioreactor culture, hfMSC proliferated rapidly, saturatingall the spaces within the scaffolds and reached confluence in 7 days,as seen by light microscopy (Fig. 3C–F), which was confirmedthrough measurement of total double-stranded DNA content(Picogreen dsDNA assay, Fig. 3S). In contrast, static-cultured scaf-folds reached confluence within the scaffolds only after 28 days ofculture (Fig. 3G–J, S), and achieved a lower final cell number at 28days compared to the bioreactor treated scaffolds (0.6�, p< 0.001,Fig. 3S).

Bioreactor-cultured scaffolds demonstrated homogenouscellular distribution at the surface (planar view, Fig. 4A) as well asthe internal core (core view, Fig. 4A) of the scaffolds at days 7 and14, which was evident with both light microscopy and FDA/PIstaining (Figs. 3E and F and 4B and C). While bioreactor-culturedscaffolds maintained good cellular viability at the core of the scaf-fold after 28 days of culture (Fig. 4D and E), we observed massivecellular death in core of the scaffold in scaffolds under static cultureconditions (Fig. 4I), despite good viabilities earlier in culture (day14, Fig. 4F and G).

3.3. Bioreactor culture enhanced osteogenic differentiation andmineralization

We observed the deposition of extracellular calcium crystalsappearing in bioreactor treated scaffolds by day 14 (Fig. 3L), whichwas only noticeable at day 28 in static-cultured scaffolds (Fig. 3R).This observation was confirmed through the darker von Kossastaining of the bioreactor cultured over static-cultured scaffolds atboth days 14 and 28 time points (Fig. 5A).

Scanning electron microscopy (SEM) revealed higher levels ofcrystal-like extracellular matrix (ECM) deposition in bioreactorcultured vs. static-cultured scaffolds (Fig. 5B). These deposits werecalcium phosphate salts as shown through element componentanalysis (EDX) to consist mainly of P, Ca and O elements (Fig. 5C).

Supporting this observation, we found that bioreactor treatedscaffolds expressed significantly higher levels of ALP activity thanthe static culture scaffolds from day 7 through day 28 (1.5�,p< 0.001, Fig. 5D), and higher calcium deposition (4.9� at day 14,p< 0.001, and 5.7� at day 28, p< 0.001, Fig. 5E).

hfMSC assumed an osteoblastic morphology after bioreactorculture (day 28, Fig. 4D and E), appearing ovoid in shape asobserved on FDA/PI staining, in contrast to static-cultured scaffolds,where hfMSC maintained a typical spindle-shaped morphology(Fig. 4H and I).

3.4. Bioreactor treatment promoted higher ectopic bone formationin vivo

Three months after subcutaneous implantation in immunode-ficient mice, scaffolds from both bioreactor and static culture werefound to be well integrated into the surrounding host tissues, withno evidence of any tumour formation (n¼ 18) (Fig. 6A and B).Bioreactor-cultured scaffolds demonstrated higher human:mousechimerism rates than static-cultured scaffolds (78.5�14.6% vs.57.6� 8.3%, p¼ 0.02) (Fig. 6C–G). In addition, bioreactor-culturedscaffolds generated more woven (Masson’s Trichrome staining,Fig. 6H and I), and calcified bone (von Kossa, Fig. 6J and K). In linewith these observations, micro-CT analyses demonstrated moreectopic bone formation in bioreactor over static-cultured scaffolds(3.2�, p< 0.001) (Fig. 6L and M).

actor for the culture of fetal mesenchymal stem cells for bone tissue

X 100

Basic morphology Immunophenotype by FACS A B

Stro-1: 45.12%

D Tri-lineages differentiation

CImmunophenotype by ICC

Fig. 2. Characterization of hfMSC. A) hfMSC grew as spindle-shaped plastic-adherent cells; B) with a significant proportion being positive for the osteoprogenitor marker Stro-1(FACS: 45.1% positive); C) hfMSC were non-haemopoietic, non-endothelial and expressed typical markers associated with MSC cell types (nuclei were counterstained with PI, scalebar¼ 20 mm). In addition, hfMSCs were positive for the embryonic stem cell markers Oct-4 and Nanog (green nuclear stain). D). Under permissive induction medium, hfMSCunderwent osteogenic (black extracellular crystals on von Kossa staining), adipogenic (red intracytoplasmic Oil Red O staining) and chondrogenic (red Safranin O staining)differentiation.

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–11 5

ARTICLE IN PRESS

4. Discussion

Bioreactors have been developed in order to address the need tocircumvent the limitations of mass transfer in thick scaffoldcultures, and in the case of bone tissue engineering, to apply suit-able mechano-transduction forces which in turn facilitate osteo-genic differentiation through specific signaling pathways [19,22]. Inthis study, we demonstrated that biaxial bioreactor maturedhfMSC/PCL–TCP scaffolds resulted in significantly higher cellular

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

proliferation, homogenous cellular distribution and in-vitro and in-vivo osteogenic differentiation than those cultured in staticcondition. In addition, biaxial bioreactor-cultured scaffolds retainedhigh cellular viabilities in the core of the scaffold not achieved withstatic cultures.

Currently, several bioreactors such as spinner flasks, perfusionbioreactors and rotating wall vessel (RWV) bioreactors have beeninvestigated for bone tissue engineering applications, but havebeen beset by various limitations. The use of spinner flasks resulted

actor for the culture of fetal mesenchymal stem cells for bone tissue

C

G

D

H

E

I

F

J

K

O

L

P

M

Q

N

R

A B

Planar view

Side view

S. Picogreen Assay

Fig. 3. Cell adhesion and proliferation of hfMSC cultured in the PCL–TCP scaffolds. A) High-porosity PCL–TCP scaffolds measuring 6 mm� 6 mm� 4 mm were seeded with hfMSCfor these experiments; B) Scaffolds were viewed from the planar (top) and side profile under phase contrast light microscopy through the culture. C–R) Scaffolds which had beencultured for one week were transferred to either bioreactor or static culture conditions over 28 days. Biaxial bioreactor-cultured scaffolds achieved confluence of all the availablespace within the scaffolds by day 7 (D–F), with the appearance of extracellular crystals from day 14 of culture (red arrow, L), which increased in amount by day 28, limiting thepassage of light through the scaffold (K–N). In contrast, static-cultured scaffolds reached confluence only after 28 days of culture, with evidence of mineralization (red arrow, R)appearing only at day 28 (G–J & O–R). S) Quantification of dsDNA content in cellular scaffolds by Picogreen assay supported the significantly higher cellular proliferation rates(p< 0.001), achieving a confluence by day 7, and achieving a higher final cell content within the scaffolds at all time points.

ARTICLE IN PRESS

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreactor for the culture of fetal mesenchymal stem cells for bone tissueengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

Fig. 4. Cellular viability studies (FDA/PI staining): A) FDA/PI was used to stain for live and dead cells respectively, through confocal microscopy imaging of the planar (top) view, andafter bisecting the scaffold into two in the middle, achieving a view of the scaffold’s centre (core view). B–I) Biaxial bioreactor-cultured scaffolds demonstrated confluence of thescaffold surface and interior, with homogenous cellular distribution at an earlier time point (B and C) then static-cultured scaffolds (F and G) (D14 shown here). By day 28, hfMSC inbioreactor-cultured scaffolds assumed an osteoblast morphology, appearing ovoid in shape (D and E), while retaining high cellular viabilities in the core of the scaffolds (E). Incontrast, hfMSC in static-cultured scaffolds remained spindle shaped (H), with poor cellular viabilities in the core of the scaffolds (I). (All images here are confocal z-stack images,constructed from 44 horizontal image sections with 300 mm in depth. Mag. 100�.)

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–11 7

ARTICLE IN PRESS

in improved fluid flow through a simple design [29–31], althoughthe turbulence generated can be detrimental for seeded cells andnewly laid down ECM [21,32]. Perfusion bioreactors, which havebeen shown to enhance MSC proliferation and osteogenic differ-entiation in scaffold constructs [33–35], have been limited by non-homogenous cellular distributions, with cells at the frontal zonesbeing washed away by the oncoming perfusion flow with higherflow rates [36]. Rotating wall vessel (RWV) bioreactors [37–39],which generate low shear forces and 3D high mass transfercapacity, are prone to similar problems of non-homogenouscellular growth and ECM deposition [21,33,40]. In addition, duringthe free floating culture, collision between the scaffolds and thebioreactor walls can induce cellular damage and disrupt cellularattachment and matrix deposition on the scaffolds [21,33,40].

Consequently, the biaxial rotating bioreactor used here has beendesigned to address the deficiencies found in current bioreactordesigns: firstly, a perfusion system was included to allow theexchange between the vessel and reservoir, allowing maximal masstransfer with consequential low shear stress, without the washoutassociated with RWV bioreactors; secondly, biaxial rotation of theculture vessel improved upon uniaxial rotating designs, leading tomore homogenous cellular and ECM distribution of the scaffold, aspreviously predicted from in silico [23], and; lastly, cellular scaf-folds were secured by pins and were not kept in free suspension,avoiding the risk of scaffold collisions with the vessel walls.

By day 28 of static culture, we found cellular necrosis within thecore of scaffolds, which is 2000 mm (2 mm) away from the surface.

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

This coincided with the achievement of cellular confluence withinthe scaffold in static culture, and hence the need for nutrients andoxygen to reach the core through diffusion alone. The cellularnecrosis observed here is largely predictable as the limits of masstransfer of nutrients are generally held to be around 200 mm [19]. Incontrast, biaxial bioreactor-cultured scaffolds, which achievedcellular confluence by day 7 of culture, resulted in high degrees ofcellular viability at the core of the scaffolds from 7 through 28 daysdespite the high cellularity and mineralization, with consequentialhigher final cellular numbers and osteogenic differentiation andmineralization. This performance was better than the experiencewith spinner flasks, where Sutherland et al. found a central necroticcore in 1 mm spheroid, which is 500 mm from the surface [41].

Culturing scaffolds in bioreactor resulted in higher cellularitywhich was obvious through both light microscopy and dsDNAcontent measurements, indicating the beneficial effects ofimproved mass transport. In addition, shear stresses found withinthe biaxial bioreactor can provide key mechanical stimulationwhich has been implicated in the promotion of cellular prolifera-tion of osteoprogenitor cell types [22,42].

In the bioreactor culture, shear forces would have been appliedto both the surface and the interior of the scaffold during the firstweek of culture, in line with in-silico fluid dynamic modelingsimulations [23]. At the end of one week, however, cellularconfluence of the scaffold would prevent any further convectionfrom taking place within the scaffold itself. The maintenance ofcellular viability at the scaffold core, which is 2000 mm from the

actor for the culture of fetal mesenchymal stem cells for bone tissue

Fig. 5. Osteogenic differentiation and mineralization of cellular scaffold. A) von Kossa staining of scaffolds in bioreactor culture demonstrated more robust osteogenic differentiationand mineralization than static culture, as shown in much darker of von kossa staining at days 14 and 28. B) Scanning electron microscopy (SEM) images of cellular scaffolds at day 28of culture demonstrating higher mineralization in the scaffolds grown in bioreactor cultures. C) EDX analysis of the element components revealed the mineralised nodules ascalcium phosphate salts, consisting of P, Ca and O elements. D) Cellular scaffolds cultured in biaxial bioreactor expressed higher level of ALP activity than those in static culture fromday 7 (*p< 0.05, **p< 0.01, ***p< 0.001), and finally, E) analysis of calcium deposition in the scaffolds revealed significantly higher calcium deposition in bioreactor-culturedscaffolds than static-cultured scaffolds from days 14 to 28. ***p< 0.001.

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–118

ARTICLE IN PRESS

surface, would depend on the improvement of mass transfer ofnutrients afforded by convection forces at the surface of the scaffold[19].

Shear stress generated by the biaxial bioreactor can triggermechano-transduction signaling pathways, which in turn up-regulates production of cyclic adenosine monophosphate (cAMP)[43], transforming growth factor beta1 (TGF-b1) [44] and nitric

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

oxide [45], all of which have been implicated in bone tissue repairmechanisms. The improved osteogenic differentiation and miner-alization of hfMSC cellular scaffolds in a dynamic bioreactor envi-ronment over a static culture system are in keeping with thegeneral observation that shear forces stimulate osteogenicprogramming in MSC [4,33,46–49]. In addition, we have observedthat biaxial bioreactor-cultured hfMSC assumed an ovoid

actor for the culture of fetal mesenchymal stem cells for bone tissue

Fig. 6. In-vivo ectopic bone formation A and B) Implanted scaffolds from both groups integrated into the surrounding tissues with no evidence of tumour formation. C and G)Immunostaining for human-specific lamins A/C (green nuclei, counterstained with propidium iodide) showed a higher human:mouse chimerism rate in bioreactor-cultured vs.static-cultured scaffolds (78.5�14.6% vs. 57.6� 8.3%, p< 0.05). (S indicates scaffold). H–K) Bioreactor-cultured scaffold showed more woven and mature bone formation andmineralization by Masson’s Trichrome and von Kossa staining respectively (scale bar 100 mm) L and M) This was confirmed by micro-CT analysis of the new ectopic bone formationaround the cellular scaffolds, with 3.2-fold more ectopic bone formed in bioreactor-cultured scaffolds compared with static-cultured scaffolds (*p< 0.05, ***p< 0.001).

ARTICLE IN PRESS

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreactor for the culture of fetal mesenchymal stem cells for bone tissueengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–1110

ARTICLE IN PRESS

morphology similar to those of mature osteoblasts, suggestinga higher degree of differentiation being achieved compared tostatic-cultured hfMSC.

The improved in-vitro performance of biaxial bioreactor-cultured scaffolds over static-cultured scaffolds was accompaniedby higher human cell chimerism rates than static-cultured scaffoldsafter implantation into NOD/SCID mice. Additionally, the higherectopic bone formation in the bioreactor-cultured group confirmedMendes et al.’s work, showing better in-vivo osteogenic perfor-mance with the use of pre-differentiated scaffolds for trans-plantation [50,51].

5. Conclusion

The use of this biaxial rotating bioreactor should allow shorterin-vitro maturation culture time to be achieved, along withimproved cellular and ECM distribution, and more efficient osteo-genic induction and mineralization of bone tissue engineered graftscompared to standard static culture systems. More importantly, theuse of our biaxial bioreactor resulted in a ten-fold improvement inmass transfer in thicker grafts, which will have important impli-cation for their eventual clinical application in bone tissueengineering.

Acknowledgement

The authors would like to thank these people for their help inthis project: Lay Geok Tan from Department of Obstetrics &Gynaecology; Mark Chong, Eddy Lee from Graduate Programme inBioengineering; Erin Teo from Department of Mechanical Engi-neering and Harmeet Singh from NUS Graduate School for Inte-grative Sciences and Engineering; National University of Singapore;Dr. Oi Wah Liew, Jenny Chong, Siew Hui Lee, Cui Xia Ang, JingyingChong, from Singapore Polytechnics.

This work is supported by the Cross Faculty Grant of NUS, Grant# R-174-000-107-123, NMRC Grant NMRC/0974/2005, NMRC/1179/2008, and National Healthcare Group SIG Grant 06013 and 08031,and funding from the Clinician Scientist Unit, NLAM, NUS. JCreceived salary support from Exxon-Mobil-NUS Fellowship.

Appendix

Figures with essential colour discrimination. Certain parts of themajority of the figures in this article are difficult to interpret inblack and white. The full colour images can be found in the on-lineversion, at doi:10.1016/j.biomaterials.2009.01.028.

Appendix. Supplementary data

Supplementary data associated with this article can be found inthe online version, at doi:10.1016/j.biomaterials.2009.01.028.

References

[1] Banwart JC, Asher MA, Hassanein RS. Iliac crest bone graft harvest donor sitemorbidity: a statistical evaluation. Spine 1995;20:1055–60.

[2] Goulet JA, Senunas LE, DeSilva GL, Greenfield ML. Autogenous iliac crest bonegraft: complications and functional assessment. Clin Orthop Relat Res1997;339:76–81.

[3] Parikh SN. Bone graft substitutes: past, present, future. J Postgrad Med2002;48:142–8.

[4] Datta N, Pham QP, Sharma U, Sikavitsas VI, Jansen JA, Mikos AG. In vitrogenerated extracellular matrix and fluid shear stress synergistically enhance3D osteoblastic differentiation. Proc Natl Acad Sci U S A 2006;103:2488–93.

[5] Service RF. Tissue engineers build new bone. Science 2000;289:1498–500.[6] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920–6.

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

[7] Sung HJ, Meredith C, Johnson C, Galis ZS. The effect of scaffold degradationrate on three-dimensional cell growth and angiogenesis. Biomaterials2004;25:5735–42.

[8] Lam CX, Hutmacher DW, Schantz JT, Woodruff MA, Teoh SH. Evaluation ofpolycaprolactone scaffold degradation for 6 months in vitro and in vivo.J Biomed Mater Res A 2008. July 21 (Epub).

[9] Ballas CB, Zielske SP, Gerson SL. Adult bone marrow stem cells for cell and genetherapies: implications for greater use. J Cell Biochem Suppl 2002;38:20–8.

[10] Mauney JR, Volloch V, Kaplan DL. Role of adult mesenchymal stem cells inbone tissue engineering applications: current status and future prospects.Tissue Eng 2005;11:787–802.

[11] Bruder SP, Jaiswal N, Haynesworth SE. Growth kinetics, self-renewal, and theosteogenic potential of purified human mesenchymal stem cells duringextensive subcultivation and following cryopreservation. J Cell Biochem1997;64:278–94.

[12] Campagnoli C, Roberts IA, Kumar S, Bennett PR, Bellantuono I, Fisk NM.Identification of mesenchymal stem/progenitor cells in human first-trimesterfetal blood, liver, and bone marrow. Blood 2001;98:2396–402.

[13] Gotherstrom C, Ringden O, Tammik C, Zetterberg E, Westgren M, Le Blanc K.Immunologic properties of human fetal mesenchymal stem cells. Am J ObstetGynecol 2004;190:239–45.

[14] Chan J, O’ Donoghue K, deFuente J, Roberts IA, Kumar S, Morgan JE, et al.Human fetal mesenchymal stem cells as vehicles for gene delivery. Stem Cells2005;23:93–102.

[15] Guillot PV, Gotherstrom C, Chan J, Kurata H, Fisk NM. Human first-trimesterfetal MSC express pluripotency markers and grow faster and have longertelomeres than adult MSC. Stem Cells 2007;25:646–54.

[16] Zhang ZY, Teoh SH, Chong MS, Schantz JT, Fisk NM, Choolani MA, et al.Superior osteogenic capacity for bone tissue engineering of fetal compared toperinatal and adult mesenchymal stem cells. Stem Cells 2008. Oct 2 (Epub).

[17] Ishaug SL, Crane GM, Miller MJ, Yasko AW, Yaszemski MJ, Mikos AG. Boneformation by three-dimensional stromal osteoblast culture in biodegradablepolymer scaffolds. J Biomed Mater Res 1997;36:17–28.

[18] Martin I, Obradovic B, Freed LE, Vunjak-Novakovic G. Method for quantitativeanalysis of glycosaminoglycan distribution in cultured natural and engineeredcartilage. Ann Biomed Eng 1999;27:656–62.

[19] Martin I, Wendt D, Heberer M. The role of bioreactors in tissue engineering.Trends Biotechnol 2004;22:80–6.

[20] Holy CE, Shoichet MS, Davies JE. Engineering three-dimensional bone tissue invitro using biodegradable scaffolds: investigating initial cell-seeding densityand culture period. J Biomed Mater Res 2000;51:376–82.

[21] Chen HC, Hu YC. Bioreactors for tissue engineering. Biotechnol Lett2006;28:1415–23.

[22] Bilodeau K, Mantovani D. Bioreactors for tissue engineering: focus onmechanical constraints. A comparative review. Tissue Eng 2006;12:2367–83.

[23] Singh H, Teoh SH, Low HT, Hutmacher DW. Flow modelling within a scaffoldunder the influence of uni-axial and bi-axial bioreactor rotation. J Biotechnol2005;119:181–96.

[24] Polkinghorne J. Review of the guidance on the research use of fetuses and fetalmaterial. London: HMSO; 1989. CM 762.

[25] Chan J, Waddington SN, O’Donoghue K, Kurata H, Guillot PV, Gotherstrom C,et al. Widespread distribution and muscle differentiation of human fetalmesenchymal stem cells after intrauterine transplantation in dystrophic mdxmouse. Stem Cells 2007;25:875–84.

[26] Hutmacher DW, Schantz T, Zein I, Ng KW, Teoh SH, Tan KC. Mechanical prop-erties and cell cultural response of polycaprolactone scaffolds designed andfabricated via fused deposition modeling. J Biomed Mater Res 2001;55:203–16.

[27] Rai B, Teoh SH, Ho KH, Hutmacher DW, Cao T, Chen F, et al. The effect ofrhBMP-2 on canine osteoblasts seeded onto 3D bioactive polycaprolactonescaffolds. Biomaterials 2004;25:5499–506.

[28] Chan J, O’ Donoghue K, Gavina M, Torrente Y, Kennea N, Mehmet H, et al.Galectin-1 induces skeletal muscle differentiation in human fetal mesenchymalstem cells and increases muscle regeneration. Stem Cells 2006;24:1879–91.

[29] Mygind T, Stiehler M, Baatrup A, Li H, Zou X, Flyvbjerg A, et al. Mesenchymalstem cell ingrowth and differentiation on coralline hydroxyapatite scaffolds.Biomaterials 2007;28:1036–47.

[30] Song K, Liu T, Cui Z, Li X, Ma X. Three-dimensional fabrication of engineeredbone with human bio-derived bone scaffolds in a rotating wall vessel biore-actor. J Biomed Mater Res A 2008;86:323–32.

[31] Stiehler M, Bunger C, Baatrup A, Lind M, Kassem M, Mygind T. Effect of dynamic3-D culture on proliferation, distribution, and osteogenic differentiation ofhuman mesenchymal stem cells. J Biomed Mater Res A 2008. April 22 (Epub).

[32] Sikavitsas VI, Bancroft GN, Mikos AG. Formation of three-dimensional cell/polymer constructs for bone tissue engineering in a spinner flask anda rotating wall vessel bioreactor. J Biomed Mater Res 2002;62:136–48.

[33] Sikavitsas VI, Bancroft GN, Holtorf HL, Jansen JA, Mikos AG. Mineralized matrixdeposition by marrow stromal osteoblasts in 3D perfusion culture increaseswith increasing fluid shear forces. Proc Natl Acad Sci U S A 2003;100:14683–8.

[34] van DJ, Bancroft GN, Sikavitsas VI, Spauwen PH, Jansen JA, Mikos AG. Flowperfusion culture of marrow stromal osteoblasts in titanium fiber mesh.J Biomed Mater Res A 2003;64:235–41.

[35] Gomes ME, Bossano CM, Johnston CM, Reis RL, Mikos AG. In vitro localizationof bone growth factors in constructs of biodegradable scaffolds seeded withmarrow stromal cells and cultured in a flow perfusion bioreactor. Tissue Eng2006;12:177–88.

actor for the culture of fetal mesenchymal stem cells for bone tissue

Z.-Y. Zhang et al. / Biomaterials xxx (2009) 1–11 11

ARTICLE IN PRESS

[36] Singh H, Ang ES, Lim TT, Hutmacher DW. Flow modeling in a novel non-perfusion conical bioreactor. Biotechnol Bioeng 2007;97:1291–9.

[37] Yu X, Botchwey EA, Levine EM, Pollack SR, Laurencin CT. Bioreactor-basedbone tissue engineering: the influence of dynamic flow on osteoblastphenotypic expression and matrix mineralization. Proc Natl Acad Sci U S A2004;101:11203–8.

[38] Granet C, Laroche N, Vico L, Alexandre C, Lafage-Proust MH. Rotating-wallvessels, promising bioreactors for osteoblastic cell culture: comparison withother 3D conditions. Med Biol Eng Comput 1998;36:513–9.

[39] Molnar G, Schroedl NA, Gonda SR, Hartzell CR. Skeletal muscle satellite cellscultured in simulated microgravity. In Vitro Cell Dev Biol Anim1997;33:386–91.

[40] Goldstein AS, Juarez TM, Helmke CD, Gustin MC, Mikos AG. Effect of convec-tion on osteoblastic cell growth and function in biodegradable polymer foamscaffolds. Biomaterials 2001;22:1279–88.

[41] Sutherland RM, Sordat B, Bamat J, Gabbert H, Bourrat B, Mueller-Klieser W.Oxygenation and differentiation in multicellular spheroids of human coloncarcinoma. Cancer Res 1986;46:5320–9.

[42] Ignatius A, Blessing H, Liedert A, Schmidt C, Neidlinger-Wilke C, Kaspar D,et al. Tissue engineering of bone: effects of mechanical strain on osteoblasticcells in type I collagen matrices. Biomaterials 2005;26:311–8.

[43] Reich KM, Gay CV, Frangos JA. Fluid shear stress as a mediator of osteoblastcyclic adenosine monophosphate production. J Cell Physiol 1990;143:100–4.

Please cite this article in press as: Zhang Z-Y et al., A biaxial rotating bioreengineering, Biomaterials (2009), doi:10.1016/j.biomaterials.2009.01.028

[44] Sakai K, Mohtai M, Iwamoto Y. Fluid shear stress increases transforminggrowth factor beta 1 expression in human osteoblast-like cells: modulation bycation channel blockades. Calcif Tissue Int 1998;63:515–20.

[45] Johnson DL, McAllister TN, Frangos JA. Fluid flow stimulates rapid andcontinuous release of nitric oxide in osteoblasts. Am J Physiol 1996;271:205–8.

[46] Gomes ME, Sikavitsas VI, Behravesh E, Reis RL, Mikos AG. Effect of flowperfusion on the osteogenic differentiation of bone marrow stromal cellscultured on starch-based three-dimensional scaffolds. J Biomed Mater Res A2003;67:87–95.

[47] Holtorf HL, Jansen JA, Mikos AG. Flow perfusion culture induces the osteo-blastic differentiation of marrow stroma cell-scaffold constructs in theabsence of dexamethasone. J Biomed Mater Res A 2005;72:326–34.

[48] Kurpinski K, Chu J, Hashi C, Li S. Anisotropic mechanosensing by mesenchymalstem cells. Proc Natl Acad Sci U S A 2006;103:16095–100.

[49] Park JS, Chu JS, Cheng C, Chen F, Chen D, Li S. Differential effects of equiaxialand uniaxial strain on mesenchymal stem cells. Biotechnol Bioeng2004;88:359–68.

[50] Mendes SC, Tibbe JM, Veenhof M, Bakker K, Both S, Platenburg PP, et al. Bonetissue-engineered implants using human bone marrow stromal cells: effect ofculture conditions and donor age. Tissue Eng 2002;8:911–20.

[51] Mendes SC, Tibbe JM, Veenhof M, Both S, Oner FC, van Blitterswijk CA, et al.Relation between in vitro and in vivo osteogenic potential of cultured humanbone marrow stromal cells. J Mater Sci Mater Med 2004;15:1123–8.

actor for the culture of fetal mesenchymal stem cells for bone tissue