Original Article

Colonization and Osteogenic Differentiation of DifferentStem Cell Sources on Electrospun Nanofiber Meshes

Yash M. Kolambkar, M.S.,1 Alexandra Peister, Ph.D.,2 Andrew K. Ekaputra, B.Appl.Sc.,3

Dietmar W. Hutmacher, Ph.D.,4,5 and Robert E. Guldberg, Ph.D.5

Numerous challenges remain in the successful clinical translation of cell-based therapies for musculoskeletaltissue repair, including the identification of an appropriate cell source and a viable cell delivery system. The aim ofthis study was to investigate the attachment, colonization, and osteogenic differentiation of two stem cell types,human mesenchymal stem cells (hMSCs) and human amniotic fluid stem (hAFS) cells, on electrospun nanofibermeshes. We demonstrate that nanofiber meshes are able to support these cell functions robustly, with both cell typesdemonstrating strong osteogenic potential. Differences in the kinetics of osteogenic differentiation were observedbetween hMSCs and hAFS cells, with the hAFS cells displaying a delayed alkaline phosphatase peak, but elevatedmineral deposition, compared to hMSCs. We also compared the cell behavior on nanofiber meshes to that on tissueculture plastic, and observed that there is delayed initial attachment and proliferation on meshes, but enhancedmineralization at a later time point. Finally, cell-seeded nanofiber meshes were found to be effective in colonizingthree-dimensional scaffolds in an in vitro system. This study provides support for the use of the nanofiber mesh as amodel surface for cell culture in vitro, and a cell delivery vehicle for the repair of bone defects in vivo.

Introduction

Bone is one of the few adult tissues with the capacityto regenerate. However, large, unstable, or infected

bone defects remain a challenging clinical problem.1 Tissueengineering strategies that deliver cells, growth factors, andgenetic material on scaffolds have demonstrated consider-able potential in developing bone graft substitutes.2,3 Deliv-ery of exogenous cells capable of forming bony tissue maybe especially important to repair bone defects in patientswith a limited endogenous progenitor cell supply, such asolder patients, smokers, or patients with certain diseases.4

The success of cell-based therapies for bone regeneration hasbeen limited, in part, by the inadequate availability of largequantities of osteogenic cells and an effective cell deliverysystem.

The identification of a cell source that may be easily har-vested, expanded to large numbers, and controllably differ-entiated may be tremendously beneficial clinically for thereconstruction of damaged tissues. Bone-marrow-derivedmesenchymal stem cells (MSCs) have demonstrated a strongpotential for differentiation into bone-forming cells, and have

been shown to promote repair of critically sized bone defectsin preclinical animal studies.5–7 These cells are well suited forautologous transplantation, making them a feasible cellsource for clinical deployment due to the lack of immuno-genic issues associated with this transplantation modality.However, MSCs are associated with reduced mineralizationcapacity in older donors and following expansion to achievetherapeutic cell numbers.8,9

Amniotic fluid stem (AFS) cells are c-Kit expressing cellsisolated from amniotic fluid that have demonstrated a highself-renewal capacity and the ability to differentiate into adiverse range of cell types, including those from the adipose,muscle, neuronal, cartilage, and bone lineages.10–13 Recently,our lab has demonstrated that these cells can produce robustmineralization in three-dimensional (3D) constructs in vitroand in vivo.14,15 Importantly, AFS cells have also shown alack of senescence through 250 population doublings anddisplay an absence of tumorigenicity in vivo.12 However,not much is known of their osteogenic potential comparedto MSCs. A critical step toward clinical translation is thequantitative comparison of the proliferation and bone-forming capacity of different cell sources.

1Wallace H. Coulter Department of Biomedical Engineering, Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Instituteof Technology, Atlanta, Georgia.

2Department of Biology, Morehouse College, Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology,Atlanta, Georgia.

3Division of Bioengineering, National University of Singapore, Singapore.4Institute of Health and Biomedical Innovation, Queensland University of Technology, Brisbane, Australia.5George W. Woodruff School of Mechanical Engineering, Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of

Technology, Atlanta, Georgia.

TISSUE ENGINEERING: Part AVolume 00, Number 00, 2010ª Mary Ann Liebert, Inc.DOI: 10.1089/ten.tea.2010.0004

1

The delivery of stem cells to the site of injury, througheither systemic introduction or local delivery, is anothercritical consideration for the success of cell-based therapies.Site-specific delivery has the advantage of being able to de-liver large numbers of cells directly to the required area. Intissue engineering strategies, this typically involves placingcells on a 3D scaffold, followed by implantation at the injurysite. However, the lack of initial vascularity at the center of a3D scaffold limits the transport of nutrients to, and wasteproducts from, the cells. This presents a very harsh envi-ronment that makes cell survival extremely difficult.16,17 Analternative is to deliver cells to the periphery of the defect viaa thin membrane or scaffold. This delivery strategy mayenhance cell survival by positioning the cells in proximity tothe surrounding highly vascularized tissues, and therebyproviding for nourishment and clearance of waste products.

Electrospinning has recently emerged as a technique tofabricate scaffolds for tissue engineering, with fiber diame-ters ranging from tens of nanometers to as large as 10mm.18–20

The nanofiber mesh obtained by this process is a uniquescaffold membrane that possesses structural features with asize scale similar to extracellular matrix (ECM) components,high porosity, and large surface-area-to-volume ratios. Theseproperties allow for enhanced cellular attachment andspreading,21,22 and therefore nanofiber meshes may serve asan effective delivery vehicle for cells to a defect site in vivo.However, it is important to first evaluate their efficacy insupporting cell function and as a cell delivery vehicle in vitro.Although a few studies have investigated the osteogenicdifferentiation of progenitor cells on nanofiber meshes, morethorough analyses are needed to characterize the differenti-ation and mineralization process, and to quantify mineraldeposition.23–25 In addition, nanofiber meshes may be uti-lized as an ECM-mimetic surface for evaluating cell behav-ior, and therefore serve as an improved in vitro cell culturesystem, compared to flat tissue culture plates.26

The aim of this study was to investigate the attachment,colonization, and osteogenic differentiation of human MSCs(hMSCs) and human AFS (hAFS) cells on electrospun na-nofiber meshes. We demonstrate that electrospun meshes areable to robustly support these functions for both cell types.Compared to tissue culture plastic, there is delayed initialattachment and proliferation, but enhanced mineralization ata later time point. Differences in the kinetics of osteogenicdifferentiation were observed between hMSCs and hAFScells. Cell-seeded nanofiber meshes were also effective incolonizing 3D scaffolds in an in vitro model. These resultsprovide support to further evaluate the nanofiber mesh as acell delivery vehicle for the repair of bone defects in vivo.

Experimental Procedures

Fabrication of nanofiber meshes

A polymer solution was made by dissolving 13% (w/v)poly (e-caprolactone) (PCL; Sigma-Aldrich, St. Louis, MO) ina 40:60 volume ratio of dichloromethane:dimethylformamide(Sigma-Aldrich). PCL pellets were added to the solventmixture, and gently stirred for 16–24 h. The polymer solutionwas loaded in a 3 mL syringe (Becton-Dickinson, FranklinLakes, NJ), and a 22-gauge blunt stainless steel needle ( JensenGlobal, Inc., Santa Barbara, CA) was attached to the syringeend. The syringe was mounted on a syringe pump (Harvard

Apparatus, Holliston, MA), and the pump was set to infuseat a rate of 0.75 mL/h. A flat, 6�6 inch copper plate(McMaster-Carr, Atlanta, GA) covered with aluminum foilwas used to collect the fibers, and placed at a distance of20 cm from the needle end. Fibers were electrospun for50 min at a voltage of 14 kV, supplied by a high-voltagepower supply (Gamma High Voltage Research, OrmondBeach, FL), to obtain a thin sheet of nanofiber mesh. To re-move any residual solvent, the meshes were placed in adesiccators for at least 1 day before further use.

Nanofiber mesh morphology

The morphology of the nanofiber meshes was examinedusing a scanning electron microscope. A small piece of thedry nanofiber mesh was cut and mounted on a metal stubusing double-sided adhesive tape. A thin layer of gold wasthen deposited on the mesh sample for 80 s using a sputtercoater (Quorum Technologies, East Granby, CT). The gold-coated sample was then viewed under a Hitachi S-800 FieldEmission scanning electron microscope (Hitachi HTA, Plea-santon, CA) with 10 kV accelerating voltage. The diametersof the fibers were quantified by analyzing the scanningelectron microscopy images (at 7000�magnification) using acustom MATLAB� (MATLAB� 7.0 R14; The MathWorks,Inc., Natick, MA) program. A total of at least 75 distinctfibers were measured from four randomly chosen locations.

Culture of AFS cells and MSCs

hAFS cells were kindly provided by Dr. Anthony Atalaand Dr. Shay Soker at the Wake Forest Institute forRegenerative Medicine (Winston-Salem, NC). The isolationmethod and culture protocols have been described previ-ously.10,12 Briefly, back-up human amniocentesis cultureswere harvested by trypsinization, and subjected to c-Kitimmunoselection. hAFS cells were subcultured routinely ata dilution of 1:4 to 1:8, and not permitted to expandbeyond 70% confluence. The hAFS cells were passaged inalpha-minimum essential medium supplemented with 16%fetal bovine serum (embryonic stem cell [ESC]-qualifiedFBS), 100 U/mL penicillin, 100 mg/mL streptomycin, 2 mML-glutamine (Invitrogen, Carlsbad, CA), 18% Chang B, and2% Chang C (Irvine Scientific, Santa Ana, CA). In all exper-iments, cells were used at passages 16–17.

hMSCs derived from the bone marrow were obtainedfrom the Tulane University Center for Gene Therapy (NewOrleans, LA) at passage 1. Cells were isolated using bonemarrow aspirates from the iliac crest of normal adult donorsas previously described.27 For expansion, these cells wereplated at a density of 50 cells/cm2, and cultured in the hMSCgrowth medium. The hMSC growth medium consisted ofalpha-minimum essential medium (Invitrogen) supple-mented with 16% FBS (Atlanta Biologicals, Atlanta, GA),100 U/mL penicillin, 100mg/mL streptomycin, and 2 mML-glutamine (Invitrogen). The cells were subcultured oncethey reached a confluency of *70%. Passage 2–3 hMSCswere then used for all experiments.

Cell culture on nanofiber fiber meshes

Square (15 mm) samples were cut from nanofiber meshsheets using scissors. Samples were placed in 24-well culture

2 KOLAMBKAR ET AL.

plates, submerged in 200 proof ethanol (Sigma-Aldrich), andsterilized by allowing the ethanol to evaporate overnight.After the samples had dried completely, they were pre-wetted with sterile 70% ethanol for 30 min. The 70% ethanolwas then aspirated, and sterile dead weights were placedaround the samples to prevent them from floating. The meshsamples were next rinsed three times with excess sterilephosphate-buffered saline (PBS; Mediatech, Inc., Manassas,VA). An 800mL volume of the medium was placed in eachwell containing the samples. The control groups received thehMSC growth medium, whereas the osteogenic groups werefurther supplemented with 10 nM dexamethasone, 6 mMb-glycerol phosphate, 50mg/mL ascorbic acid 2-phosphate,and 50 ng/mL L-thyroxine (Sigma-Aldrich). hMSCs andhAFS cells were then seeded onto nanofiber meshes in*200mL of the hMSC medium such that the density of cellswas 20,000 cells/cm2. Cells were also cultured in tissue cul-ture treated 24-well plates at the same density for compari-son. The medium was changed every 3–4 days, and theconstructs were cultured for up to 4 weeks.

Cell viability

On days 1, 7, 14, and 28, the viability of the cells in theconstructs was assessed using the Live/Dead� staining kit(Molecular Probes, Eugene, OR, Invitrogen). Harvestedconstructs were rinsed in PBS and incubated in 4mM calcein-AM and 4mM ethidium homodimer-1 for 45 min at roomtemperature. The samples were again rinsed in PBS, andimages were obtained on a Zeiss LSM 510 confocal micro-scope (Carl Zeiss, Thornwood, NY). Green fluorescence ofcalcein-AM was detected by using a 488-nm Argon ion laserand a band pass 505–550 filter. Red fluorescence of ethidiumhomodimer-1 was detected by using a 543-nm helium–neonlaser and a long pass 560 filter.

DNA content

Samples were harvested after 1, 7, 14, and 28 days toevaluate the construct cellularity, which was assessed bydetermining the DNA content. The cells were first lysedby freeze-thawing the constructs three times in 0.05% TritonX-100 (Sigma-Aldrich) with vigorous vortexing. To freezethe samples, they were placed in dry-ice-cooled methanol(Sigma-Aldrich) for 5 min. Samples were then thawed ina room-temperature water bath. The DNA amount in thelysate was quantified using the PicoGreen� dsDNA Quan-titation Kit (Molecular Probes), and standardized usingLambda DNA solutions of known concentrations. A workingsolution of the PicoGreen reagent was made following themanufacturer’s protocol, and incubated with experimentalsamples in the dark for 5 min at room temperature. Thefluorescence was measured on a plate reader (HTS 7000;Perkins-Elmer, Waltham, MA) at an excitation of 485 nm andemission of 535 nm. All samples were run in triplicate, andthe DNA content was normalized to the culture surface areaof the samples.

Alkaline phosphatase activity

To determine the osteogenic differentiation of the cellson nanofiber meshes, the alkaline phosphatase (ALP) ac-tivity assay was performed. In this assay, the release of

p-nitrophenol from p-nitrophenyl phosphate by the ALPenzyme is measured.28 The same lysate solution that wasused to determine DNA content was used for this purpose.The ALP substrate working solution was made by mixingequal parts of 20 mM p-nitrophenyl phosphate, 1.5 M2-amino-2-methyl-1-propanol (pH 10.25), and 10 mM MgCl2.The experimental samples were mixed with the freshly madesubstrate working solution, and incubated for 1 h at 378C.The reaction was stopped by adding 1 N NaOH, and theabsorbance was measured at 405 nm on a plate reader(PowerWave XS; Biotek, Winooski, VT). All samples wererun in triplicate and compared to p-nitrophenol standards.The ALP activity was normalized by the incubation time andthe amount of DNA obtained from the PicoGreen assay.

Calcium content

To quantify matrix mineralization, the calcium depositedby cells on nanofiber meshes after 28 days was determinedusing the Arsenazo III dye.29 Samples were vortexed with1 N acetic acid overnight to extract the calcium into solution.The extract was mixed with the Arsenazo III reagent (Diag-nostic Chemicals Limited, Oxford, CT) and incubated for 30 sat room temperature, and the absorbance read at 650 nm on aplate reader (PowerWave XS; Biotek). The samples wereassayed in triplicate and compared to calcium chloridestandards.

Calcein staining and quantification

For observation of the mineral deposited by the cells, thesamples were stained using calcein on day 28.30,31 Briefly, astock solution of 100mg/mL calcein (Sigma-Aldrich) in PBS(pH 7.4) was added to the medium on top of the samples,such that the final concentration of calcein was 10 mg/mL.The samples were incubated in the calcein solution for 4 h inthe incubator. After rinsing twice with PBS and fixing with10% neutral-buffered formalin (EMD Chemicals, Gibbstown,NJ), samples were rinsed with excess of deionized water. Thefluorescence of the samples was read on a fluorescence platereader (HTS 7000; Perkins-Elmer) at an excitation of 485 nmand emission of 535 nm. After this, the same samples wereimaged using an inverted microscope (Axio Observer.Z1;Carl Zeiss) and an FITC filter.

Fourier transform infrared spectroscopy

On day 28, constructs were also harvested for analyzingthe chemical composition of the mineral deposited on thenanofiber meshes. Samples were dehydrated in 100% ethanoland dried at 508C overnight. Acellular PCL nanofiber meshwas used as a negative control. After dehydration, thesamples were cut into small pieces, mixed with potassiumbromide (Sigma-Aldrich), and pressed into pellets using acustom built apparatus. Samples were analyzed with aNicolet Nexus 470 FTIR spectrometer (Thermo-Nicolet, Ma-dison, MI). Sixty-four scans were acquired at 4 cm�1 resolu-tion under nitrogen purge.

Cell delivery by nanofiber mesh in vitro

The ability of a cell-seeded nanofiber mesh to serve as acell delivery vehicle was studied using an in vitro model. AFScells were seeded on to nanofiber mesh samples (15�10 mm)

STEM CELL FUNCTION ON NANOFIBER MESHES 3

at a density of 200,000 cells/cm2. The cells were allowed toattach to the mesh overnight. On the following day, each cell-seeded mesh was wrapped around a cylindrical collagenscaffold (dry dimensions: 4 mm diameter and 9 mm length)aseptically, such that the cells were facing the scaffold (Fig.7A). The scaffolds were punched from a fibrous collagensheet (average pore size 61.7 mm, 93.7% pore volume; KenseyNash Corporation, Exton, PA). The mesh was held in posi-tion by placing two interrupted silk sutures through themesh and scaffold at the two ends of the scaffold. For com-parison, we also seeded 300,000 cells throughout collagenscaffolds by pipetting the cell suspension directly in thescaffolds. There was no nanofiber mesh in this control group.The constructs were statically cultured in the hAFS cellgrowth medium. After 2 weeks, the mesh was taken off,following which the mesh and scaffold were stained with theLive/Dead staining kit (Molecular Probes, Invitrogen) toobserve the cell migration into the scaffold. A confocal mi-croscope (Zeiss LSM 510; Carl Zeiss) was used to take serialimages to create 3D images.

Data analysis

Results are presented as mean� standard error of themean. Analysis of variance was performed on data, with pair-wise comparisons done using the Tukey multiple comparisonprocedure. A p-value of< 0.05 was considered significant.Residuals were used for diagnosing the appropriateness of themodel by analyzing the constancy of error variance and nor-mality of error terms.32 Wherever required, remedial measureswere taken by transforming the data according to the Box-Coxprocedure,33 or by using weighted least squares to make theerror variance constant and the error distribution normal.32,34

Minitab� 15 (Minitab, Inc., State College, PA) was used toperform the statistical analysis.

Results

Morphology of nanofiber meshes

PCL nanofiber meshes were electrospun on a flat collectorplate. The mesh formed a circular area of *8 cm diameter.

The thickness of the mesh was found to vary with location,with the central areas thicker than the edges. The thin meshsamples from the edges were discarded and not used for cellculture. Fibers appeared to be smooth and uniform, withminimal bead formation (Fig. 1A, B). The quantification ofthe fiber diameter using a custom MATLAB� programdemonstrated that the mean fiber diameter was 591 nm witha standard deviation of 199 nm. The fiber diameter histogramrevealed that most of the fibers were between 300 and900 nm, with the highest frequency occurring in the 500–600 nm range (Fig. 1C).

Viability and colonization of hMSCsand hAFS cells over time

hMSCs and hAFS cells were seeded on electrospun na-nofiber meshes and tissue culture plates, and cultured in theosteogenic medium for up to 28 days. The viability of thecells on the meshes was assessed on days 1, 7, 14, and 28 bythe Live/Dead staining kit. The live cells are stained green,whereas the dead cells appear red. At the same time points,DNA from the samples was extracted and quantified to es-timate the number of cells on the meshes, as well as in theculture wells. The Live/Dead images (Fig. 2) illustrate thatboth cell types attached to the nanofiber meshes by day 1and were able to spread out by day 7. During days 7–14 thenumber of cells increased considerably, and by day 28, thecells were confluent on the meshes. The viability of both celltypes on the meshes was found to be high, as seen by theextensive green stain, though a few dead cells were detected.No differences were observed in the viability and coloniza-tion between the two cell types.

The DNA quantification over the 4-week culture periodwas used to compare the colonization kinetics of the two celltypes on tissue culture plates and nanofiber meshes, re-spectively (Fig. 3). There was significant increase in DNAwith time for both cell types, on plates as well as meshes,indicating cellular proliferation. On plates, the number ofhAFS cells increased between days 1 and 7 but did notchange significantly after that, suggesting rapid initial pro-liferation and confluency around day 7 (Fig. 3A). On the

FIG. 1. Nanofiber mesh morphology. Nanofiber meshes were electrospun from a 13% (w/v) poly (e-caprolactone) (PCL)solution made in 40:60 dichloromethane:dimethylformamide. A scanning electron microscope was used to examine themorphology of the nanofibers. (A) Scanning electron microscopy image at low (1000�) magnification. (B) Scanning electronmicroscopy image at high (7000�) magnification. (C) The diameter of the fibers was quantified using a custom MATLAB�

program, and the diameter distribution was plotted on a histogram. The mean diameter of the fibers was found to be591� 199 nm.

4 KOLAMBKAR ET AL.

other hand, hMSCs increased in numbers between both days1–7 and days 14–28. However, the later increase in DNA isbecause the hMSCs lift off the plate after confluence aroundday 7 to form a pellet and then repopulate the plate. Thispelleting behavior was not seen with the hAFS cells for up to 28days, though the hAFS cells do ultimately lift off the plate. ThehMSC repopulation explains the increase in hMSC numberbetween days 14 and 28 and the differences seen betweenhMSCs and hAFS cells on days 14 and 28. There were alsosignificantly more hAFS cells than hMSCs on day 1, suggest-ing a higher initial attachment and/or proliferation rate.

The pelleting phenomenon does not occur on the nanofi-ber meshes, even at a later time points. There was a signifi-cant increase in DNA for both cell types between days 1–7,and an even higher increase between days 7 and 14, corre-sponding to the Live/Dead images (Fig. 3B). The number ofcells did not change significantly after that, suggesting con-fluency of the cells on the nanofiber meshes. There were nosignificant differences in the DNA between hMSCs andhAFS cells at any time point.

Figure 3C and D compare the colonization kinetics ofhAFS cells on the nanofiber meshes with that on the tissueculture plates. On day 1, there was significantly less DNA onthe mesh compared to plates, suggesting that not all cellsattach to the nanofibers within the first 24 h. The same trendwas seen on day 7. However, by day 14, there was no sig-nificant difference between meshes and plates. On day 28,there was again no significant difference, though the lines

crossed over. Although the hMSC pellet in the plates andrepopulate the culture surface by day 28, the amount of DNAwas not significantly different than that on the mesh.

Osteogenic differentiation of hMSCs and hAFScells: ALP activity

The osteogenic differentiation of the cells was first inves-tigated by analyzing the ALP activity of the cells (Fig. 4).ALP is a membrane-bound enzyme that hydrolyzes phos-phate esters, which results in inorganic phosphate beingavailable for incorporation into mineral deposits.35 Therewas significant increase in the ALP activity of both cell typeswith time, on plates as well on nanofiber meshes, suggestingosteogenic differentiation. On tissue culture plates, ALP ac-tivity peaked at day 7 for MSCs, whereas for AFS cells itincreased slowly but continuously up to day 28 (Fig. 4A).Interestingly, the maximum ALP activity of the hAFS cellswas greater than the hMSC peak. On the nanofiber meshes,hMSCs demonstrated a similar earlier rise in ALP activity onday 14 compared to day 28 for hAFS cells (Fig. 4B). The ALPactivity of the hMSCs was significantly greater than that ofhAFS cells on all time points other than day 1. The ALPresponse on meshes is delayed compared to that on the plate,as seen by the later increase in ALP activity. Interestingly, theALP activity of hMSCs on meshes was sustained longer thanthat observed on plates, with the maximum value on meshesat day 28 greater than that on plates at day 7 ( p< 0.05).

FIG. 2. Live/Dead staining. Human mesenchymal stem cells (hMSCs) and human amniotic fluid stem (hAFS) cells wereseeded on nanofiber meshes and cultured in the osteogenic medium. On days 1, 7, 14, and 28, the viability of the cells on themeshes was assessed by imaging the constructs with a confocal microscope after staining with the Live/Dead� stain. Green,live cells; red, dead cells. Images were taken at 10�magnification. Scale bar indicates 200mm and applies to all images. Bothcell types were able to attach, proliferate, and become confluent on the mesh. Color images available online at www.liebertonline.com/ten.

STEM CELL FUNCTION ON NANOFIBER MESHES 5

Osteogenic differentiation of hMSCs and hAFS cells:Matrix mineralization

The osteogenic differentiation of the cells was further in-vestigated by quantifying and staining the calcium depositsand by analyzing the chemical nature of the depositedmineral by Fourier transform infrared (FTIR) spectroscopy.An analysis of variance on the calcium deposited by the cellsrevealed that both cell type and the culture surface had asignificant effect on the calcium levels (Fig. 5A). Under os-teogenic stimulation, all groups demonstrated increasedcalcium deposition, compared to the growth medium, indi-cating that cells are able to differentiate to an osteoblasticphenotype on the surfaces. Calcium levels in the hMSCgrowth medium groups were negligible. hAFS cells depos-ited a higher amount of calcium than hMSCs on both platesand meshes. Also, both cell types deposited more calcium onmeshes compared to plates.

FTIR spectroscopy was used to characterize the composi-tion of the mineral that was deposited by the cells on thenanofiber meshes under osteogenic stimulation (Fig. 5B). Todistinguish the peaks associated with the mineral from thepeaks associated with the PCL mesh, an acellular piece ofPCL mesh was also scanned. The cellular samples displayedamide I/II peaks at 1655 and 1550 cm�1, a carbonate peak at870 cm�1, and a doublet split phosphate peak at 560 and

605 cm�1, which were not seen in the acellular mesh. Therewas also a peak at 1050 cm�1 in the cellular samples, but itoverlapped with a PCL mesh peak. These peaks are thesignature of a carbonate-containing, poorly crystalline hy-droxyapatite, the form of mineral that is found in nativebone. This suggests that both hMSCs and hAFS cells de-posited mineral that possessed the characteristic bands ofphysiological mineral.

Samples were stained with calcein to observe the presenceof calcium on the nanofiber meshes. The calcein stainingdemonstrated the presence of extensive calcium-containingmineral nodules, which were uniformly distributed on themeshes, as seen by the green fluorescence (Fig. 6A). This wasthe case in both the hMSC and hAFS cell osteogenic groups,whereas the growth medium groups stained minimally.Quantification of the fluorescence revealed that more mineralwas deposited by the hAFS cells compared to the hMSCs(Fig. 6B), thus supporting the calcium data in Figure 5.

Nanofiber mesh as a cell delivery vehicle

As a preliminary evaluation of the nanofiber mesh for celldelivery, AFS cells were seeded on nanofiber meshes andwrapped around cylindrical collagen scaffolds in vitro. Theconstructs were cultured for 2 weeks and then stained withthe Live/Dead staining kit. In the cell-seeded scaffold, we

FIG. 3. DNA content. To evaluate sample cellularity, the DNA content was determined after cell lysis using the PicoGreen�

reagent. Cells were cultured on (A) tissue culture plates and (B) nanofiber meshes. To compare the cellularity on nanofibermeshes with that on tissue culture plates, the data were plotted again for (C) hMSCs and (D) hAFS cells. DNA increased withtime, indicating cellular proliferation. p< 0.05 is considered significantly different (*significantly different from the other celltype at same time point; #significantly different from the previous time point for same cell type).

6 KOLAMBKAR ET AL.

observed that more cells were present at the periphery of thescaffold, even though cells were seeded throughout (Fig. 7A).We also detected a large number of dead cells in the interiorof the scaffold. When the cell-seeded mesh was wrappedaround a collagen scaffold, we found that numerous cellshad migrated off the mesh onto the collagen scaffold and hadhigh viability (Fig. 7C, D). The majority of cells were locatedclose to the peripheral surface. Interestingly, we noted thatthe top surface was completely covered with cells (Fig. 7D).This implies that the cells were able to migrate at least 2 mmfrom the edge, where the mesh was present, to the center ofthe top surface. A few cells were also seen in the interior ofthe scaffold, more than 500mm away from the peripheralsurface (Fig. 7D). The mesh was completely confluent withcells, indicating that only a subset of cells migrates from themesh onto the scaffold (Fig. 7E).

Discussion

In this study, we investigated the function of two kinds ofstem cells, adult hMSCs and fetal hAFS cells, on electrospunnanofiber meshes. Both cell types were able to attach, colo-nize, and undergo robust osteogenic differentiation on themeshes. This indicates that the nanofiber mesh is a scaffoldmembrane capable of supporting vital osteoprogenitor andosteoblast functions. Other groups have also reported theability of nanofiber meshes to promote differentiation of

FIG. 4. Alkaline phosphatase (ALP) activity. The osteo-genic differentiation of the cells was evaluated by measuringthe ALP activity of cell lysates on (A) tissues culture platesand (B) nanofiber meshes. ALP activity increased for bothcell types with time, suggesting osteogenic differentiation.p< 0.05 is considered significantly different (*significantlydifferent from the other cell type at same time point; #sig-nificantly different from a previous time point for samecell type; $significantly different from hMSC peak on plate atday 7).

FIG. 5. Calcium quantification and Fourier transform in-frared analysis. (A) The mineralization of the constructs wasassessed by measuring the calcium deposited by cells. Bothcell types deposited calcium in the osteogenic medium, in-dicating an osteoblast phenotype. p< 0.05 is considered sig-nificantly different (#significantly different from the growthmedium; *significantly different from other cell type on samesurface; $significantly different from plate with same celltype). (B) The chemical composition of the mineral was an-alyzed by Fourier transform infrared spectroscopy. The an-notated peaks are the signature of a carbonate-containing,poorly crystalline hydroxyapatite, indicative of physiologicmineral. The remaining peaks are due to the PCL nanofibermesh, as seen in the acellular mesh. Color images availableonline at www.liebertonline.com/ten.

STEM CELL FUNCTION ON NANOFIBER MESHES 7

osteoblasts36,37 and MSCs;23,38 however, a quantitative andmore thorough analysis of the matrix mineralization hasbeen missing. We utilized a sensitive calcium assay based onthe Arsenazo III dye to quantify the extent of matrix min-eralization.29,39 The mineral deposited on the nanofiber me-shes was confirmed to be biological in nature by FTIRspectroscopy, indicating that the process was cell mediated.Finally, we used calcein staining to observe and semi-quantify the mineral deposits on the mesh. Another advan-tage of the calcein stain is that it can be used for continuousmonitoring of the in vitro matrix mineralization process.30,31

Although acellular approaches to bone reconstructionusing scaffolds and osteogenic growth factors have shownmoderate clinical success, the delivery of osteogenic cellsmay be required for patients with a reduced local supply ofresponsive osteoprogenitor cells. For successful clinical

translation of cell-based bone defect repair, a cell sourceneeds to be identified that is readily available, propagatedeasily, has high osteogenic potential, and will be accepted bythe recipient immune system. Both hMSCs and hAFS cellspossess a number of these characteristics. MSCs have beenstudied extensively, especially for bone regeneration, andpreclinical studies have shown their ability to repair bonedefects in vivo.5,6 A number of human clinical trials haveshown variable, but encouraging results of hMSC therapy,including the treatment of graft-versus-host disease, myo-cardial infarction, osteogenesis imperfecta, and large bonedefects.40,41 However, hMSCs are known to progressivelylose their stem cell properties during expansion, limiting thetotal number of cells available for therapy, and limited via-bility following transplantation remains a significant chal-lenge.42 One of the objectives of this study was to comparethe osteogenic capacity of this widely used adult stem cellwith a more novel fetal stem cell source, the hAFS cells. AFScells have been demonstrated to be capable of extensive self-renewal, and therefore can be expanded to large numbersand still maintain their multipotency.10–13

The colonization kinetics of hMSCs and hAFS cells onnanofiber meshes was found to be similar, suggesting com-parable proliferation rates when seeded at a high density of20,000 cells/cm2. On tissue culture plates, there were morehMSCs than hAFS cells on day 28, but this was due to thepelleting and recolonization by the hMSCs. hAFS cellsdemonstrated a later rise in ALP activity than the hMSCs onboth plates and meshes, perhaps due to their primitive na-ture. ALP is a membrane-bound enzyme that plays an im-portant role in matrix mineralization.35 ALP activity isone of the earliest markers of osteogenic differentiationand rises as the osteoprogenitors commit to the osteoblastlineage. It peaks in the matrix maturation phase in prepa-ration of mineralization and decreases as mineralizationprogresses.43,44 However, ALP is expressed by other differ-entiated cells as well,45,46 and therefore it is important tosimultaneously analyze other osteogenic measures, such asmatrix mineralization. At 4 weeks, hAFS cells depositedsignificantly more mineral than hMSCs on both plates andmeshes, as demonstrated by calcium quantification and cal-cein staining. Thus, we observed that, while the rise in ALPactivity of the hAFS cells occurs later than in hMSCs, thehAFS cells mineralize more robustly after 4 weeks. This in-dicates that the kinetics of ALP activity and matrix miner-alization are differentially regulated for these two differentcell types. Our results demonstrate that hAFS cells have highosteogenic potential, even at the late passage numbers wehave used. In addition, unlike human ESCs, hAFS cells haveshown an absence of tumor formation in vivo.12 This suggeststhat the hAFS cells may be a feasible cell source for the repairof bone defects. They may be especially useful in the case ofpatients whose cells are not amenable for autologous trans-plantation due to disease or advanced age. Another advan-tage of the AFS cells is that they are suitable for convenientoff-the-shelf allogeneic cell delivery, as long as the majorhistocompatibility complex of donor and recipient are mat-ched. This would reduce the time and cost of delivering thecell therapy and may result in improved clinical acceptance.

Woo et al. have recently reported that a nanofibrousscaffold made by a modified solvent casting method resultedin improved expression of osteoblast phenotype versus a

FIG. 6. Calcein staining. Constructs were stained withcalcein to observe the presence of the mineral deposits. (A)The osteogenic samples stained with calcein, illustrating thatcells had deposited mineral uniformly on the nanofiber me-shes. Images were taken at 10�magnification. Scale bar in-dicates 200mm and applies to all images. (B) The calceinstaining was quantified by measuring the fluorescence usinga plate reader. The data revealed greater mineralization byAFS cells. p< 0.05 is considered significantly different (*sig-nificantly greater from the MSC osteogenic medium; #sig-nificantly greater from growth with same cell type). Colorimages available online at www.liebertonline.com/ten.

8 KOLAMBKAR ET AL.

solid-walled scaffold.47 We compared cell function on thenanofiber meshes with that on tissue culture plastic. Com-pared to tissue culture plastic, there is delayed initial at-tachment and proliferation on the meshes. However, by 2weeks, the cells on the meshes catch up with those on theplates, and there is no significant difference between thegroups. The cells, especially the hMSCs, did not lift off

the nanofiber mesh surface at high cell densities, as was seenon plates. This difference in cell attachment could be due tochanges in cell adherence, material, and topographic prop-erties or ECM deposited on the nanofiber mesh surface.Despite the initial lag in colonization on meshes compared tothat on plates, we observed enhanced mineralization onmeshes by both the cell types at 4 weeks. This suggests that

FIG. 7. Cell-seeded nanofiber meshes for in vitro delivery. (A) To investigate the use of nanofiber meshes for cell delivery,AFS cells were seeded on nanofiber meshes and wrapped around a three-dimensional (3D) collagen scaffold for 2 weeksin vitro. For comparison, cells were seeded throughout the scaffold. (B–E) Three-dimensional confocal images of the Live/Dead-stained scaffold and mesh. The projections of the 3D images are shown. The surface and top views are views of the 3Dimage looking from the top. The side view is the view of the 3D image looking from the side. The green box indicates the areaand view being analyzed. (B) The collagen scaffold with cells seeded throughout had more cells on the exterior withnumerous dead cells in the interior. (C) When a cell-seeded mesh was wrapped around the scaffold, cells migrated on to theperipheral surface of the scaffold and displayed high viability. (D) Top and cross-sectional surfaces of scaffold wrapped withcell-seeded mesh. Cells also colonized the top surface of the scaffold and migrated more than 500 mm into the scaffold fromthe mesh. (E) The seeded mesh was confluent with cells. Images were taken at 10�magnification. Scale bar indicates 200 mmand applies to all images. Color images available online at www.liebertonline.com/ten.

STEM CELL FUNCTION ON NANOFIBER MESHES 9

the ECM-mimetic morphology of the nanofibers provides anenvironment conducive for matrix maturation. In a recentarticle, Smith et al. demonstrated that the use of nanofibersresulted in a greater degree of ESC differentiation, comparedto films and tissue culture plates.48 This study also providessupport for the use of nanofiber meshes as an improvedin vitro cell culture model surface that better recapitulates thein vivo environment of cells.

Cell survival after delivery is a critical issue in the devel-opment of cell-based strategies, especially for thick tissuessuch as bone. The lack of initial vascularity in bone defectslimits the transport of nutrients to and waste products fromthe center of the defect. Therefore, if cells are seededthroughout a 3D scaffold and placed at the defect site, cellslocated at the center of the scaffold may not survive.16,17,49,50

Delivery of cells on the periphery of bone defects via a tissue-engineered periosteum may be an effective approach toenhance cell survival by the presence of a neighboring vas-culature. With time, as a continuous vasculature is estab-lished at the center, the cells may migrate toward the centerdue to an improved transport environment. Recently, Zhangreported that engraftment of bone morphogenetic protein-2producing MSCs using gelfoam wrapped around nonvitalallografts improved allograft incorporation and repair.51 Ourresults indicate that the electrospun nanofiber mesh pos-sesses characteristics suitable for supporting cell function. Inaddition, its design and thickness can be controlled to obtaina membrane suitable for creating a tissue-engineered peri-osteum.52 To begin preliminary investigations into cell de-livery, we asked following the question: Will cells migrate offthe mesh and populate a 3D scaffold in vitro? We observedthat the cells migrated off the mesh and colonized the scaf-fold within 2 weeks, traveling as far as 2 mm. In addition tothe migration, part of the colonization is probably due to cellproliferation. Interestingly, we noticed better viability of thecells in the scaffold when they were delivered on the meshcompared to when they were seeded uniformly in the scaf-folds. These results suggest that a cell-seeded nanofiber meshmay be an effective method to deliver cells to bone defectsand maintain high viability. Future work will determinewhether these effects are also observed in vivo.

In conclusion, we demonstrated that two types of stemcells, hMSCs and hAFS cells, are able to attach, colonize, andundergo robust osteogenic differentiation on electrospunnanofiber meshes. hAFS cells displayed a delayed ALP in-crease, but deposited significantly more mineral comparedto hMSCs. Cell-seeded nanofiber meshes were effective incolonizing 3D scaffolds in an in vitro model. These resultsindicate that the electrospun nanofiber mesh supportsosteoprogenitor cell function and may be useful as a mediumfor cell delivery for the repair of bone defects in vivo. Inaddition, this study provides support for the use of nanofibermeshes as a model surface for cell culture experiments.

Acknowledgments

This work was funded by NIH Grant AR051336. Collagensheets were generously provided by Kensey Nash Corpora-tion. The authors thank Dr. Charles Gersbach, Dr. JenniferPhillips, and Dr. Ge Zhao for technical assistance with thecalcium assay, FTIR, and the ALP assay, respectively. Theauthors would also like to thank Dr. Ayona Chatterjee for

assistance with statistical analysis and Vivek Mukhatyar forhelpful discussions.

Disclosure Statement

The authors do not have any competing financial interest.

References

1. Praemer, A., Furner, S., and Rice, D.P. MusculoskeletalConditions in the United States. Rosemont, IL: AmericanAcademy of Orthopaedic Surgeons, 1999.

2. Rose, F.R., and Oreffo, R.O. Bone tissue engineering: hope vshype. Biochem Biophys Res Commun 292, 1, 2002.

3. Salgado, A.J., Coutinho, O.P., and Reis, R.L. Bone tissueengineering: state of the art and future trends. MacromolBiosci 4, 743, 2004.

4. Bruder, S.P., and Fox, B.S. Tissue engineering of bone. Cellbased strategies. Clin Orthop Relat Res S68, 1999.

5. Bruder, S.P., Kraus, K.H., Goldberg, V.M., and Kadiyala, S.The effect of implants loaded with autologous mesenchymalstem cells on the healing of canine segmental bone defects.J Bone Joint Surg Am 80, 985, 1998.

6. Bruder, S.P., Kurth, A.A., Shea, M., Hayes, W.C., Jaiswal, N.,and Kadiyala, S. Bone regeneration by implantation of pu-rified, culture-expanded human mesenchymal stem cells.J Orthop Res 16, 155, 1998.

7. Tsuchida, H., Hashimoto, J., Crawford, E., Manske, P.,and Lou, J. Engineered allogeneic mesenchymal stem cellsrepair femoral segmental defect in rats. J Orthop Res 21, 44,2003.

8. McCulloch, C.A., Strugurescu, M., Hughes, F., Melcher,A.H., and Aubin, J.E. Osteogenic progenitor cells in rat bonemarrow stromal populations exhibit self-renewal in culture.Blood 77, 1906, 1991.

9. Quarto, R., Thomas, D., and Liang, C.T. Bone progenitor celldeficits and the age-associated decline in bone repair ca-pacity. Calcif Tissue Int 56, 123, 1995.

10. Kolambkar, Y.M., Peister, A., Soker, S., Atala, A., andGuldberg, R.E. Chondrogenic differentiation of amnioticfluid-derived stem cells. J Mol Histol 38, 405, 2007.

11. Siddiqui, M.M., and Atala, A. Amniotic fluid-derived plur-ipotential cells. In: Lanza, R., Gearhart, J., and Hogan, B. eds.Handbook of Stem Cells. Philadelphia: Academic Press,2004, p. 175.

12. De Coppi, P., Bartsch, G., Jr., Siddiqui, M.M., Xu, T., Santos,C.C., Perin, L., Mostoslavsky, G., Serre, A.C., Snyder, E.Y.,Yoo, J.J., Furth, M.E., Soker, S., and Atala, A. Isolation ofamniotic stem cell lines with potential for therapy. NatBiotechnol 25, 100, 2007.

13. Delo, D.M., De Coppi, P., Bartsch, G., Jr., and Atala, A.Amniotic fluid and placental stem cells. Methods Enzymol419, 426, 2006.

14. Peister, A., Porter, B.D., Kolambkar, Y.M., Hutmacher, D.W.,and Guldberg, R.E. Osteogenic differentiation of amnioticfluid stem cells. Biomed Mater Eng 18, 241, 2008.

15. Peister, A., Deutsch, E.R., Kolambkar, Y., Hutmacher, D.W.,and Guldberg, R. Amniotic fluid stem cells produce robustmineral deposits on biodegradable scaffolds. Tissue Eng PartA 15, 3129, 2009.

16. Thorrez, L., Shansky, J., Wang, L., Fast, L., Vandendriessche,T., Chuah, M., Mooney, D., and Vandenburgh, H. Growth,differentiation, transplantation and survival of human skel-etal myofibers on biodegradable scaffolds. Biomaterials 29,

75, 2008.

10 KOLAMBKAR ET AL.

17. Jager, M., Degistirici, O., Knipper, A., Fischer, J., Sager, M.,and Krauspe, R. Bone healing and migration of cord blood-derived stem cells into a critical size femoral defect afterxenotransplantation. J Bone Miner Res 22, 1224, 2007.

18. Pham, Q.P., Sharma, U., and Mikos, A.G. Electrospunpoly(epsilon-caprolactone) microfiber and multilayer nano-fiber/microfiber scaffolds: characterization of scaffolds andmeasurement of cellular infiltration. Biomacromolecules 7,

2796, 2006.19. Barnes, C.P., Sell, S.A., Boland, E.D., Simpson, D.G., and

Bowlin, G.L. Nanofiber technology: designing the nextgeneration of tissue engineering scaffolds. Adv Drug DelivRev 59, 1413, 2007.

20. Tan, S.H., Inai, R., Kotaki, M., and Ramakrishna, S. Sys-tematic parameter study for ultra-fine fiber fabrication viaelectrospinning process. Polymer 46, 6128, 2005.

21. Pham, Q.P., Sharma, U., and Mikos, A.G. Electrospinning ofpolymeric nanofibers for tissue engineering applications: areview. Tissue Eng 12, 1197, 2006.

22. Murugan, R., and Ramakrishna, S. Nano-featured scaffoldsfor tissue engineering: a review of spinning methodologies.Tissue Eng 12, 435, 2006.

23. Li, W.J., Tuli, R., Huang, X., Laquerriere, P., and Tuan, R.S.Multilineage differentiation of human mesenchymal stemcells in a three-dimensional nanofibrous scaffold. Biomater-ials 26, 5158, 2005.

24. Xin, X., Hussain, M., and Mao, J.J. Continuing differentiationof human mesenchymal stem cells and induced chondro-genic and osteogenic lineages in electrospun PLGA nanofi-ber scaffold. Biomaterials 28, 316, 2007.

25. Yoshimoto, H., Shin, Y.M., Terai, H., and Vacanti, J.P. Abiodegradable nanofiber scaffold by electrospinning and itspotential for bone tissue engineering. Biomaterials 24, 2077,2003.

26. Klinkhammer, K., Seiler, N., Grafahrend, D., Gerardo-Nava,J., Mey, J., Brook, G.A., Moller, M., Dalton, P.D., and Klee, D.Deposition of electrospun fibers on reactive substratesfor in vitro investigations. Tissue Eng Part C Methods 15,

77, 2009.27. Sekiya, I., Larson, B.L., Smith, J.R., Pochampally, R., Cui,

J.G., and Prockop, D.J. Expansion of human adult stem cellsfrom bone marrow stroma: conditions that maximize theyields of early progenitors and evaluate their quality. StemCells 20, 530, 2002.

28. Martin, J.Y., Dean, D.D., Cochran, D.L., Simpson, J., Boyan,B.D., and Schwartz, Z. Proliferation, differentiation, andprotein synthesis of human osteoblast-like cells (MG63)cultured on previously used titanium surfaces. Clin OralImplants Res 7, 27, 1996.

29. Leary, N.O., Pembroke, A., and Duggan, P.F. Single stablereagent (Arsenazo III) for optically robust measurement ofcalcium in serum and plasma. Clin Chem 38, 904, 1992.

30. Uchimura, E., Machida, H., Kotobuki, N., Kihara, T., Kita-mura, S., Ikeuchi, M., Hirose, M., Miyake, J., and Ohgushi,H. In-situ visualization and quantification of mineralizationof cultured osteogenetic cells. Calcif Tissue Int 73, 575, 2003.

31. Hale, L.V., Ma, Y.F., and Santerre, R.F. Semi-quantitativefluorescence analysis of calcein binding as a measurement ofin vitro mineralization. Calcif Tissue Int 67, 80, 2000.

32. Kutner, M.H., Nachtsheim, C.J., Neter, J., and Li, W. AppliedLinear Statistical Models, fifth edition. New York: McGraw-Hill, 2005.

33. Box, G.E.P., and Cox, D.R. An analysis of transformations.J R Stat Soc B 26, 211, 1964.

34. Wu, C.F.J., and Hamada, M. Experiments: Planning, Ana-lysis, and Parameter Design Optimization, first edition. NewYork: Wiley-Interscience, 2000.

35. Seibel, M.J. Molecular markers of bone turnover: biochemi-cal, technical and analytical aspects. Osteoporos Int 11

Suppl 6, S18, 2000.36. Venugopal, J., Low, S., Choon, A.T., Kumar, A.B., and Ra-

makrishna, S. Electrospun-modified nanofibrous scaffoldsfor the mineralization of osteoblast cells. J Biomed Mater ResA 85, 408, 2008.

37. Badami, A.S., Kreke, M.R., Thompson, M.S., Riffle, J.S., andGoldstein, A.S. Effect of fiber diameter on spreading,proliferation, and differentiation of osteoblastic cells onelectrospun poly(lactic acid) substrates. Biomaterials 27, 596,2006.

38. Shin, M., Yoshimoto, H., and Vacanti, J.P. In vivo bonetissue engineering using mesenchymal stem cells on anovel electrospun nanofibrous scaffold. Tissue Eng 10, 33,2004.

39. Gersbach, C.A., Le Doux, J.M., Guldberg, R.E., and Garcia,A.J. Inducible regulation of Runx2-stimulated osteogenesis.Gene Ther 13, 873, 2006.

40. Abdallah, B.M., and Kassem, M. The use of mesenchymal(skeletal) stem cells for treatment of degenerative diseases:current status and future perspectives. J Cell Physiol 218, 9,2009.

41. Quarto, R., Mastrogiacomo, M., Cancedda, R., Kutepov,S.M., Mukhachev, V., Lavroukov, A., Kon, E., andMarcacci, M. Repair of large bone defects with the use ofautologous bone marrow stromal cells. N Engl J Med 344,

385, 2001.42. Shi, S., Gronthos, S., Chen, S., Reddi, A., Counter, C.M.,

Robey, P.G., and Wang, C.Y. Bone formation byhuman postnatal bone marrow stromal stem cells isenhanced by telomerase expression. Nat Biotechnol 20,

587, 2002.43. Lian, J.B., and Stein, G.S. Development of the osteo-

blast phenotype: molecular mechanisms mediating oste-oblast growth and differentiation. Iowa Orthop J 15, 118,1995.

44. Aubin, J.E. Regulation of osteoblast formation and function.Rev Endocr Metab Disord 2, 81, 2001.

45. Hinnebusch, B.F., Siddique, A., Henderson, J.W., Malo, M.S.,Zhang, W., Athaide, C.P., Abedrapo, M.A., Chen, X., Yang,V.W., and Hodin, R.A. Enterocyte differentiation markerintestinal alkaline phosphatase is a target gene of thegut-enriched Kruppel-like factor. Am J Physiol GastrointestLiver Physiol 286, G23, 2004.

46. Matsumoto, H., Erickson, R.H., Gum, J.R., Yoshioka, M.,Gum, E., and Kim, Y.S. Biosynthesis of alkaline phosphataseduring differentiation of the human colon cancer cell lineCaco-2. Gastroenterology 98, 1199, 1990.

47. Woo, K.M., Jun, J.H., Chen, V.J., Seo, J., Baek, J.H., Ryoo,H.M., Kim, G.S., Somerman, M.J., and Ma, P.X. Nano-fibrous scaffolding promotes osteoblast differentiation andbiomineralization. Biomaterials 28, 335, 2007.

48. Smith, L.A., Liu, X., Hu, J., Wang, P., and Ma, P.X. En-hancing osteogenic differentiation of mouse embryonicstem cells by nanofibers. Tissue Eng Part A 15, 1855,2009.

49. Byers, B.A., Guldberg, R.E., and Garcia, A.J. Synergy be-tween genetic and tissue engineering: Runx2 overexpressionand in vitro construct development enhance in vivo miner-alization. Tissue Eng 10, 1757, 2004.

STEM CELL FUNCTION ON NANOFIBER MESHES 11

50. Cartmell, S., Huynh, K., Lin, A., Nagaraja, S., and Guldberg,R. Quantitative microcomputed tomography analysis ofmineralization within three-dimensional scaffolds in vitro.J Biomed Mater Res 69A, 97, 2004.

51. Zhang, X., Xie, C., Lin, A.S., Ito, H., Awad, H., Lieberman,J.R., Rubery, P.T., Schwarz, E.M., O’Keefe, R.J., andGuldberg, R.E. Periosteal progenitor cell fate in segmentalcortical bone graft transplantations: implications forfunctional tissue engineering. J Bone Miner Res 20, 2124,2005.

52. Ekaputra, A.K., Zhou, Y., Cool, S.M., and Hutmacher, D.W.Composite electrospun scaffolds for engineering tubularbone grafts. Tissue Eng Part A 15, 3779, 2009.

Address correspondence to:Robert E. Guldberg, Ph.D.

George W. Woodruff School of Mechanical EngineeringParker H. Petit Institute for Bioengineering and Bioscience

Georgia Institute of Technology315 Ferst Drive

Atlanta, GA 30332

E-mail: robert.guldberg@me.gatech.edu

Received: January 4, 2010Accepted: May 24, 2010

Online Publication Date: June 23, 2010

12 KOLAMBKAR ET AL.