Osteogenic differentiation of marrow stromal cellscultured on nanoporous alumina surfaces

Ketul C. Popat,1,2 Kwan-Isara Chatvanichkul,2 George L. Barnes,3 Thomas Joseph Latempa Jr.,4

Craigs A. Grimes,4 Tejal A. Desai1,21Department of Physiology and Division of Bioengineering, University of California at San Francisco,San Francisco, California2Department of Biomedical Engineering, Boston University, Boston, Massachusetts3Department of Orthopedic Surgery, Boston University Medical Center, Boston, Massachusetts4Department of Electrical Engineering, Pennsylvania State University, University Park, Pennsylvania

Received 15 December 2005; revised 9 May 2006; accepted 28 July 2006Published online 6 November 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.31028

Abstract: A major goal in orthopedic biomaterialsresearch is to design implant surfaces, which will enhanceosseointegration in vivo. Several microscale as well asnanoscale architectures have been shown to significantlyaffect the functionality of bone cells i.e., osteoblasts. In thiswork, nanoporous alumina surfaces fabricated by a two-step anodization process were used. The nanostructure ofthese surfaces can be controlled by varying the voltageused for anodization process. Marrow stromal cells wereisolated from mice and seeded on nanoporous and amor-phous (control) alumina surfaces. Cell adhesion, prolifera-tion, and viability were investigated for up to 7 days ofculture. Furthermore, the cell functionality was investi-gated by calcein staining. The cells were provided withdifferentiation media after 7 days of culture. The alkaline

phosphatase (ALP) activity and matrix production werequantified using a colorimetric assay and X-ray photoelec-tron spectroscopy (XPS) for up to 3 weeks of culture (2weeks after providing differentiation media). Further, scan-ning electron microscopy (SEM) was used to investigateosteoblast morphology on these nanoporous surfaces. Overthe 3-week study, the nanoporous alumina surfaces dem-onstrated *45% increase in cell adhesion, proliferation,and viability, 35% increase in ALP activity, and 50%increase in matrix production when compared with thecontrol surfaces. � 2006 Wiley Periodicals, Inc. J BiomedMater Res 80A: 955–964, 2007

Key words: orthopedic biomaterials; marrow stromal cells;osseointegration; nanoporous alumina

INTRODUCTION

A major challenge in orthopedic biomaterialsresearch is the design of material surfaces that will (a)allow attachment of anchorage-dependent osteoproge-nitors and (b) promote in vitro and in vivo osteogenicdifferentiation of cells, thereby delivering a matureosteoblastic cell population capable of rapidly form-ing an interfacial layer with appropriate biomechani-cal properties.1,2 The important material related pa-rameters that influence events at bone-implant inter-face are surface composition, surface energy, surfaceroughness, and surface topography.3 Researchershave previously shown that altering the surface to-pography at a micro- or nanoscale can alter the differ-entiation of various cell types.4–7 On the basis of an

earlier work done by our group and others, webelieve that nanoscale topography also influencesfunctionality osteoblasts or bone-forming cells. Bonehas a varied arrangement of hierarchical structureconsisting of a defined macrostructure, microstruc-ture, and nanostructure. Thus, we believe that mate-rial surface topographies at a size scale, comparableto the hierarchical structure of bone, will induce a dif-ferential biological response in terms of elevated alka-line phosphatase (ALP) activity and enhanced matrixproduction.

Earlier work by Klawitter and Hulbert, using ox-ide ceramics, showed that a porous surface withpore diameter of *100 mm was needed for adequatebone ingrowth.8 It was thought that small pore sizesallowed incomplete mineralization of the infiltratingtissue. Subsequently, a work by Bobyn et al., usingmetallic implants, showed good bone ingrowth withpore sizes between 50 and 400 mm.9 Recently, how-ever, studies have revealed the possibility that muchsmaller pores may allow bone ingrowth when pre-

Correspondence to: T. A. Desai; e-mail: Tejal.desai@ucsf.eduContract grant sponsor: Boston University SPRInG

' 2006 Wiley Periodicals, Inc.

sented at high density within metal-oxide sub-strates.10,11 For example, porous Ca–P coatings onimplants have shown apposition of human bonegrowth within 2–3 weeks post surgery.12 Althoughthis work is encouraging, there are several problemsrelated to dissolution of the coatings over time, andcracking and separation from the metallic sub-strate.13,14 These studies point to the importance ofdeveloping more robust material surfaces withmicro- and nanoscale architecture that has potentialto incorporate osteoconductive or osteoinductive bio-molecules, which will then enhance the appositionof bone from the existing bone surfaces and stimu-late new bone formation. In a study, using aluminananoparticles of grain size 167 nm down to 24 nm,osteoblast adhesion increased by 50% as grain sizedecreased. Enhanced proliferation, ALP activity, andmatrix production were evident when comparedwith alumina with conventional (micron size) grainsizes. Several other recent studies10,11,15,16 also sup-port the notion that nanoscale topography directlyinfluences bone-cell behavior. Nanoporosity mayresult in improved cellular adhesion and thereforeenhanced matrix deposition on the surface. It couldalso be possible that these nanoporous structures,which are not able to allow cellular ingrowth due totheir size, will instead, be filled with collagen andbone matrix. In this case, a nanoporous surfacecould act as the framework in which osteoblasts willsynthesize new bone.17

In this study, bone marrow stromal cells (MSCs)are used to investigate the ability of nanostructuredsurfaces to enhance cell differentiation. Bone marrowextracts containing osteoprogenitors combined withvarious matrices have been demonstrated to acceler-ate and enhance bone formation within osseousdefects when compared with the matrix alone.18,19

The MSCs contain a pluripotent population of cellscapable of differentiating along multiple mesenchy-mal lineages (e.g., bone,20,21 ligament,22,23 adipose,24

cartilage,25 and muscle tissue26). Because tissue cul-ture techniques allow the isolation and ex vivoexpansion of this cell population from animals,27

these cells may represent an ideal osteogenic cellsource to be used to evaluate their interaction withnanostructured surfaces. The ability of MSCs toinduce bone formation in vivo is believed to be dueto the interaction of osteoprogenitors present withinthe cell populations with osteoinductive factors, suchas bone morphogenetic proteins and various growthfactors and cytokines, which cause them to differen-tiate into bone-forming cells i.e., osteoblasts,28–32

which will then eventually form bone matrix.Aluminum oxide (Al2O3), or alumina, is used as

our substrate material. As a biocompatible ceramic, itis currently used in orthopedic and dental implants.The oxide surface makes alumina chemically inert in

biological environments and hence it is a choice mate-rial for implant engineering applications.33 The bio-compatibility of metal oxides has already been provenas the materials have current clinical applications inorthopedic prostheses and dental implants.4 Our tech-nique of producing nanoporous alumina surfacesusing anodization provides ease of control over thesize of structure, with maintenance of mechanicalproperties that is not possible with nanophase materi-als or microstructured ceramics.34 Osteogenic differen-tiation on nanoporous alumina surfaces was examinedat the biochemical level to evaluate phenotypic res-ponses. MSCs adhesion after 1 day and proliferationafter 4 and 7 days of seeding cells on nanoporousalumina surfaces was investigated. Further, the cell vi-ability and the cytotoxicity due to the substrate wereinvestigated. Also, matrix deposited by the cells onthe nanoporous alumina surfaces was characterizedby measuring calcium and phosphorus surface con-centrations using X-ray photoelectron spectroscopy(XPS). The ALP activity was measured by a colorimet-ric assay. Scanning electron microscopy (SEM) wasused to evaluate the morphology of adhered cells innanoporous alumina surfaces. Similar analyses wereperformed on amorphous alumina surfaces (with nonanoarchitecture), which were used as control.

MATERIALS AND METHODS

Fabrication of nanoporous alumina surfaces

Pure aluminum sheets (2 cm � 2 cm � 0.5 mm, AlfaAesar) were electrochemically polished in a solution ofphosphoric and sulfuric acid (volume ratio 3:2, Sigma) at 3Amps and 1208C. This process was continued until thesheet appeared smooth and shiny. The sheets were thencleaned using an acetone ultrasonic bath. A polymer coat-ing (ethyl acetate–butyl acetate–dibutyl phthalate, Sigma)was applied on one side of the sheet to protect it fromanodization. The anodization was performed in a two-stepprocess using 0.3M oxalic acid (Sigma) at 60 V. Platinumsheet (Alfa Aesar) was used as cathode and aluminum asanode. The anodization temperature was kept constantthroughout the process by cooling the solution with an icebath. Between the first and second anodization, the sheetswere etched in a 4% (w/w) chromic acid (Sigma) and 8%(v/v) phosphoric acid mixture to remove the aluminalayer that had formed on the unprotected side of the sheet.A polymer was removed using an acetone and 10% NaOHsolution (Sigma) to expose the unoxidized aluminum. Thealuminum was etched using a solution of 10% (w/w) HCland 0.1M CuCl2 leaving only the nanoporous aluminaformed by anodization. The barrier layer present afteretching the aluminum was removed using 10% (v/v)phosphoric acid solution. Figure 1 shows the schematic ofanodization process and Figure 2 shows the SEM image ofthe nanoporous structure of alumina surfaces formed by

956 POPAT ET AL.

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

two-step anodization process at 60 V, which resulted inpore size of *72 nm.

Preparation of cell cultures

C57 BJ mice were used from Jackson Laboratories (BarHarbour ME) to isolate primary marrow stromal cells(MSCs). The animals were euthanized according to theIACUC approved protocol. Twelve animals were used foreach harvesting. Limbs were removed aseptically andplaced in cold PBS in 50-mL falcon tubes. Bones were dis-sected from soft tissues under a cell culture hood. Meta-physeal ends of the bones were removed to allow access tothe marrow cavity. Marrow cavity contents were flushedout using a 25-gauge needle attached to a 10-mL syringecontaining a-modified MEM (aMEM, Invitrogen) supple-mented with 10% FBS and 1% penn/strep. The flushedcell suspension was then filtered through a 70-mm nylonstrainer. Cells were then counted using a hemocytometer.Prior to cell isolation, nanoporous alumina surfaces wereplaced in wells of 6-well plates under ultraviolet lights ina biological hood for 24 h for presterilization. The surfaceswere then sterilized by immersing in 70% ethanol for30 min. MSCs were seeded at a density of 15–25 � 106 perwell in 6 well plates with sterile and clean nanoporoussurfaces. Higher seeding density was used to see signifi-cant changes in long-term phenotype behavior. Amor-phous alumina surfaces were used as controls. On day 4of culture, half of the media was removed and replacedwith fresh aMEM supplemented with 10% FBS and 1%penn/strep. On day 7 of culture, all media was removedand cells were fed with differentiation media. Differentia-tion media includes aMEM supplemented with 10% FBS,1% Penn/Strep, Dexamethasone (10�8M final concentra-tion), Ascorbic acid (50 mg/mL final), and b-glycerol phos-phate (8 mmol final). Media was changed every two days.

Cell adhesion and proliferation

Nanoporous alumina surfaces and amorphous aluminasurfaces were placed in wells of 6 well plates and weresterilized as mentioned earlier. MSCs were seeded at adensity of 15 � 106 cells/well. Cell adhesion was investi-gated after 1 day of seeding and proliferation was investi-gated after 4 and 7 days of seeding on surfaces. The boundcells were quantified by trypsinizing and counting using astandard hemocytometer.

Cell viability

The cell viability was investigated after 1 and 4 days ofseeding using commercially available MTT assay (Sigma).

Figure 1. Schematic of two-step anodization process for the fabrication of nanoporous alumina surfaces. [Color figurecan be viewed in the online issue, which is available at www.interscience.wiley.com.]

Figure 2. SEM image of nanoporous alumina surfaceshowing the porous structure (pore size 79 nm).

DIFFERENTIATION OF MSCs CULTURED ON NANOPOROUS ALUMINA SURFACES 957

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

The MTT method is simple, accurate, and yields reproduc-ible results. The key component is (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) or MTT. Solutionsof MTT, dissolved in medium or balanced salt solutionswithout phenol red, are yellowish in color. Mitochondrialdehydrogenases of viable cells cleave the tetrazolium ring,yielding purple formazan crystals. Standard procedureprovided with the MTT kit was followed. The resultingpurple solution was spectrophotometrically measured at570 nm using a Spetronix Instruments GENESYS 5. Anincrease or decrease in cell number results in a concomi-tant change in the amount of formazan formed, indicatingthe degree of cytotoxicity caused by the alumina surface.

Cell functionality

Calcein AM, which measures the intracellular esteraseactivity, was used to stain the live cells present on the sur-face. It is a fluorogenic esterase substrate that is hydro-lyzed intracellularly to a green fluorescent product; thus,green fluorescence is an indicator of live cells. The surfaceswith cells were incubated in a 2 mM solution of calceinAM for 30–45 min at room temperature. Calcein wasdetected using standard fluorescein filter set on a fluores-cence microscopy (Olympus BX60).

Alkaline phosphatase activity

ALP activity is an important parameter to access thenormal functionality of cells on a surface; hence, the activ-ity was measured up to 3 weeks of culture (i.e., 2 weeksafter providing differentiation media). The MSCs werewashed thrice with warm PBS and fixed with 2% parafor-maldehyde in 0.2 mol/L cacodylic buffer for 10 min atroom temperature. Following the fixation, surfaces withcells were washed extensively with deionized water. ALPsubstrate solution was made by adding 25 mL of alkalinebuffer solution (Sigma) to 25 mL of deionized water. Tothis, contents of 1 capsule of prepackaged ALP substrate(100 mg/capsule, Sigma) was dissolved. One milliliter ofthe ALP substrate solution was added per well andallowed to react for 30 min at 378C. Then, 0.5 mL of thesolution was removed and added to 0.5 mL of 0.2N NaOHto stop the reaction. Converted substrate is a direct mea-sure of alkaline phosphatase (ALP) activity and wasdetected photometrically at 410 nm. The ALP absorbencieswere normalized with total intracellular protein content.To release the intracellular protein, the adhered cells onthe NW surfaces were lysed in 2% Triton-X solution for30 min and the resulting lysate solution was used for theanalysis. The total protein content was determined using aBCA (bicinchoninic acid) assay kit (Pierce). The absorbanceat 562 nm was converted to protein concentration using analbumin standard curve.

Extracellular matrix

Calcium and phosphorus are the primary componentsof bone matrix. Once the cells start differentiating on the

surfaces, they will deposit bone matrix. To detect the pres-ence of calcium and phosphorus on surfaces, the sampleswere transferred from 6 well plates and were air-dried forXPS analysis. XPS is a surface-sensitive technique anddetects trace levels of elements present on the surface. Sur-vey spectra were collected from 0 to 1100 eV using aKratos AXIS Ultra Imaging X-ray Photoelectron Spectrome-ter with a monochromatic Al-Ka-X-ray small spot source(1486.6 eV) and multichannel detector with a pass energyof 160 eV. Data for percent atomic composition and atomicratios for deposited calcium and phosphorous on aluminasurfaces for up to three weeks of culture were calculatedfrom the survey scans using the manufacturer suppliedsoftware.

Cell morphology

Cell morphology on nanoporous alumina and amor-phous alumina surfaces were examined using SEM for upto three weeks of culture. Prior to imaging, the cells werefixed and dehydrated. Using a procedure modified fromCorning Life Sciences,35 the surfaces were rinsed twice inPBS and then soaked in the primary fixative of 3% glutar-aldehyde (Sigma), 0.1M of sodium cacodylate (Poly-sciences), and 0.1M sucrose (Sigma, ST. Louis MO) for45 min. The surfaces were subjected to two 5-min washeswith a buffer containing 0.1M sodium cacodylate and 0.1Msucrose. The cells were then dehydrated by replacing thebuffer with increasing concentrations of ethanol (35, 50, 70,95, and 100%) for 10 min each. Further, the cells weredried by replacing ethanol by hexamethyldisilazane(HMDS) (Polysciences) for 10 min. The HMDS wasremoved, and the surfaces were air-dried for 30 min. SEMimaging was conducted on the JEOL JSM 5910 SEM (Pea-body, MA) at voltages ranging from 10–20 kV after thesurfaces were sputter-coated in gold in the Cressington108 Sputter Coater (Cranberry Twp, PA). The sputter-coaterwas set at current of 20 mA and pressure of 0.05 mbarfor 20 s to deposit a 10-nm layer of gold.

Statistical analysis

Each experiment was reconfirmed at least thrice usingcells from different marrow stromal preparations. All theresults were analyzed using analysis of variance (ANOVA).Statistical significance was considered at p < 0.01.

RESULTS

Cell adhesion and proliferation

Cell adhesion was investigated by counting thetrypsinized cells 1 day after seeding them on nano-porous alumina surfaces. The results were comparedwith cells seeded on amorphous alumina surfaces.Figure 3 shows the cell count on nanoporous andamorphous alumina surfaces obtained by a hemocy-

958 POPAT ET AL.

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

tometer. Nanoporous alumina surfaces demonstratedhigher cell adhesion when compared with amor-phous alumina surfaces (p < 0.01).

Further, MSCs were allowed to proliferate on alu-mina surfaces for up to 7 days. The proliferationwas investigated by the similar trypsinization techni-ques described earlier in this section. After both 4and 7 days of culture, more cells were present onnanoporous alumina surface suggesting higher pro-liferation when compared with amorphous aluminasurface (p < 0.01).

Cell viability

The cell viability was investigated using commer-cially available MTT assay for the log phase of cellgrowth i.e., up to 4 days of culture on nanoporousand amorphous alumina surfaces. The absorbance ofthe solution directly relates to the mitochondrial ac-tivity, which is a direct measure of cell viability. Fig-ure 4 shows the absorbance measured for nanopo-rous and amorphous alumina surfaces after day 1and day 4 of seeding the cells. Results indicatehigher number of viable cells on nanoporous alu-mina surfaces when compared with those on amor-phous alumina surfaces (p < 0.01).

Cell functionality

Calcein AM was used to stain live cells on nano-porous and amorphous alumina surfaces for up to 4days of culture to investigate the cell functionality.The stained cells were visualized using a fluores-cence microscope. Figure 5 shows the images of live

cells on nanoporous and amorphous alumina surfa-ces after 1 and 4 days of culture.

ALP activity

MSCs were allowed to adhere and proliferate onalumina surfaces for up to 7 days after which differ-entiation media was provided. ALP activity (ALPabsorbance normalized with intracellular proteincontent) was measured for up to 2 weeks of cultureafter providing differentiation media using the color-imetric assay as described earlier (Fig. 6). The mea-sured absorbance directly relates to the concentrationof ALP. Cells cultured on nanoporous alumina surfa-ces show higher ALP activity when compared withamorphous alumina surfaces.

Extracellular matrix

XPS analysis was used to detect the presence ofextracellular matrix produced by MSCs on aluminasurfaces for a 2-week period of culture after differentia-tion media was provided. Survey scans were collectedto determine the presence of calcium and phosphorus,the major constituents of bone matrix, on the surface.Table I shows the ratio of elemental concentrations ofcalcium and phosphorus to aluminum (substratepeak). Ca and P ratios were higher for all weeks fornanoporous alumina surfaces when compared withother amorphous alumina surfaces suggesting thatmore extracellular matrix was deposited by osteoblastson these surfaces (p < 0.01).

Figure 4. Cell viability measured after day 1 and day 4 ofculture on amorphous and nanoporous alumina surfaceswith MTT assay; results indicate cells are more viable onnanoporous surfaces when compared with amorphous alu-mina surfaces; n ¼ 3, p < 0.01 for nanoporous aluminawhen compared with amorphous alumina. [Color figurecan be viewed in the online issue, which is available atwww.interscience.wiley.com.]

Figure 3. Cell adhesion (after 1 day) and proliferation (af-ter 4 and 7 days) on amorphous and nanoporous aluminasurfaces; results show an increased adhesion and prolifera-tion on nanoporous surfaces when compared with amor-phous alumina surfaces; n ¼ 3, p < 0.01 for nanoporousalumina when compared with amorphous alumina. [Colorfigure can be viewed in the online issue, which is availableat www.interscience.wiley.com.]

DIFFERENTIATION OF MSCs CULTURED ON NANOPOROUS ALUMINA SURFACES 959

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

Osteoblast morphology

One of the major goals of this study was to deter-mine the cell morphology on nanoporous aluminasurfaces up to a 3-week period of culture. SEMimages of cells fixed on the surface showed that thecells are proliferating after 1 week of culture. Cells onnanoporous alumina surfaces are present as clustersand show a spreading morphology when comparedwith those of amorphous alumina surfaces suggestingthat they are responding to the nanoarchitecture [Fig.7(a,b)]. Figure 7(c,d) shows SEM images of cells onamorphous and nanoporous alumina surfaces for upto 2 week of culture, respectively (1 week after pro-viding the differentiation media). Cells on nanopo-rous alumina surfaces are extending their processes,which indicate the differentiation stage when com-pared with those of amorphous alumina surfaces,which show a spherical morphology. After 3 weeksof culture (i.e. 2 weeks after providing differentiationmedia), proliferated cells have formed a network onthe surface by depositing the matrix, which is absentin cells on amorphous alumina surfaces [Fig. 7(e,f)].

DISCUSSION

The development of nanoporous platforms basedon novel metal-oxide films will provide insight into

cell-material interactions for designing theimproved implant surfaces. It is envisioned that theincorporation of such nanoarchitectures on implantsurfaces will further facilitate the culture and main-tenance of differentiated cell states, and providelong-term cell viability and functionality. The moti-vation to use nanostructured surfaces for bone bio-templating is physiologically driven. Human bone

Figure 6. Alkaline phosphatase (ALP) activity measuredafter 2 and 3 weeks of culture on amorphous and nanopo-rous alumina surfaces; ALP activity normalized with intra-cellular protein content; results indicate higher ALP activ-ity on nanoporous surfaces compared to amorphous alu-mina surfaces; n ¼ 3, p < 0.01 for nanoporous aluminawhen compared with amorphous alumina. [Color figurecan be viewed in the online issue, which is available atwww.interscience.wiley.com.]

Figure 5. 10� images of live cells stained with calcein on alumina surfaces; cells on amorphous surfaces are sphericalwhereas on nanoporous surfaces seem to be spreading. [Color figure can be viewed in the online issue, which is availableat www.interscience.wiley.com.]

960 POPAT ET AL.

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

consists of several levels of hierarchical structuralorganization35–37; the macrostructure, the micro-structure, and the nanostructure. The macrostruc-ture and the microstructure of bone are formed byseveral nanoscale constituents such as minerals, col-lagen, and noncollaganeous nanosized organics.Protein molecules such as phosphoproteins, osteo-pontin, sialoprotein, osteonectin, and osteocalcinpresent in the extracellular matrix of bone arenanostructured. Further, calcium phosphate, an im-portant constituent of bone matrix is composition-ally and structurally nanostructured. Thus, wehypothesized that controlled nanoarchitectures withsize scales comparable to that of bone nanostruc-ture will be able to promote osteoblast differentia-tion and matrix production, thus enhancingosseointegration.38

In this work, we have demonstrated that con-trolled nanoscale architecture can promote short-term in vitro osseointegration. MSCs were seeded onnanoporous and amorphous alumina surfaces andwere cultured for up to 3 weeks. Up to 1 week, theMSCs, the majority of which consists of osteoproge-nitors, were allowed to adhere and proliferate.Osteogenic differentiation was induced by providingmedia with ascorbic acid, dexamethasone, and b-glycerol phosphate. This will result in the formationof mature osteoblasts, which will then deposit bonematrix on the nanoporous alumina surface.

Our short-term results indicate a 45% increase incell adhesion on nanoporous alumina surfaceswhen compared with amorphous alumina surfaces.There was also an increase in proliferation up to40% on nanoporous surfaces (Fig. 3). Thus, nano-porosity or rather a well-defined nanoarchitectureon alumina surfaces seems to promote cell adhesionand proliferation. The adhesion and proliferationare much higher than those observed in traditionalmicroscale topography.8,9 The adhesion and prolifer-ation phases are extremely important since this willgovern the long-term functionality of these cells onthe surfaces. Apart from this, the cell viability dur-ing this initial period of cell-material interaction isextremely important. Although adhesion on somesurfaces may be higher, cells may not be viable andfunctional. Hence, we have investigated the cell via-

bility using standard MTT assay, which measuresthe intracellular mitochondrial activity. This assaycan be used for the log phase growth of the cellsand hence viability with this assay was investigatedfor up to 4 days of culture (before the cells reachconfluency). Results indicate that the cells are 40%more viable on nanoporous alumina surfaces whencompared with amorphous alumina surfaces after 4days of culture (Fig. 4). Further, calcein was used tostain live cells on surfaces. Results indicate that cellsare flattening and spreading on nanoporous aluminasurfaces. By day 4, images clearly indicate that thecells have started spreading. Alternatively, cells onamorphous alumina surfaces are still spherical inshape and show relatively lower degree of spreadingby 4 days (Fig. 5).

After 1 week of culture, differentiation mediawas provided to induce the cells to start differenti-ating into osteoblasts, produce marker proteins, andbegin to lay matrix on surfaces. The ALP activityand the extracellular matrix production are themost important indication of normal phenotypicbehavior of osteoblasts. The ALP activity was deter-mined using a colorimetirc assay and was found tobe 35% higher for cells on nanoporous alumina sur-faces than those on amorphous alumina surfaces(Fig. 6). Further, extracellular matrix deposited onsurfaces was characterized by determining the sur-face concentrations of calcium and phosphorous,the major constituents of bone matrix, using XPS.Results indicate that there is 50% increase in cal-cium and phosphorous surface concentration after 3weeks for cells on nanoporous alumina, suggestingmore matrix deposition than on amorphous alu-mina surfaces. The results are comparable to thoseobtained on nanophase ceramics.16,17 Further, SEMwas used to image cell morphology (Fig. 7). After 1week of culture, the cells have formed clusters onnanoporous alumina surfaces when compared withthose on amorphous alumina surfaces, which isnormal phenotype behavior of MSCs [Fig. 7(a,b)].As the cells differentiate into osteoblasts, they startdepositing matrix on the surface, which can beclearly seen from SEM images of cells on the nano-porous alumina surface. The cells on the amor-phous alumina surfaces are isolated and the

TABLE ICalcium and Phosphorous Concentrations on Nanoporous and Amorphous Alumina Surfaces Measured

Using XPS; n 5 3; p < 0.01 for Calcium and Phosphorous Concentrations for Week 2 and Week 3

Ca/Al P/Al

Nanoporous Amorphous alumina Nanoporous Amorphous alumina

Week 2 0.37 6 0.002 0.13 6 0.002 0.45 6 0.001 0.21 6 0.003Week 3 0.46 6 0.003 0.21 6 0.001 0.72 6 0.002 0.33 6 0.003

DIFFERENTIATION OF MSCs CULTURED ON NANOPOROUS ALUMINA SURFACES 961

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

spreading is less when compared with those onnanoporous alumina surfaces [Fig. 7(c,d)]. By theend of week 3, cells on nanoporous alumina surfacehave formed a network when compared with those

on amorphous alumina surface. This suggests thatthe precise nanoscale architecture promotes enhancedosteoblast functionality by increased ALP activityand matrix production.

Figure 7. SEM images of MSCs on amorphous [(a) week 1, (c) week 2, and (e) week 3] and nanoporous alumina surfaces[(b) week 1, (d) week 2, and (f) week 3]. Cells on nanoporous surfaces show higher degree of spreading and normal phe-notype behavior when compared with those on amorphous alumina surfaces.

962 POPAT ET AL.

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

CONCLUSION

Designing materials that elicit a favorable responsefor osteoblasts is important in orthopedic biomateri-als research. In this work, nanoporous alumina sur-faces, fabricated by a two-step anodization processwere used to investigate short-term and long-termperformance of MSCs. Nanoporous alumina surfacessupported higher cell adhesion, proliferation, and vi-ability up to 7 days of culture when compared withamorphous alumina surfaces. In addition, MSCswere cultured on these surfaces for a period of 3weeks to investigate the long-term effect of nano-architecture on their functionality and phenotypicbehavior. Cells cultured on nanoporous alumina sur-faces demonstrated higher ALP activity. Further-more, the calcium and phosphorous concentrationswere higher on nanoporous alumina surfaces suggest-ing that cells laid more matrix when compared withamorphous alumina surfaces. Thus, this researchshows that the bone-cell performance can be signifi-cantly improved using controlled nanoarchitectureand these surfaces have potential applications as coat-ings for orthopedic implants. To use these surfacesfor orthopedic implants, it is extremely important toinvestigate their in vivo performance and mechanicaland tribological properties to understand the failuremechanisms. Future studies will begin to continue toaddress these issues in hopes of improving the bone-implant interface.

References

1. Ishaug SL, Thomson RC, Mikos AG, Langer R. Biomaterials fororgan regeneration. In: Meyers RA, editor. Molecular Biologyand Biotechnology: A Comprehensive Desk Reference. NewYork: VCH; 1995. pp 86–93.

2. Ishaug SL, Crane GM, Miller MJ, Yasko AW, Yaszemski MJ,Mikos AG. Bone formation by three-dimensional stromal osteo-blast culture in biodegradable polymer scaffolds. J BiomedMater Res 1997;36:17–28.

3. Schwartz Z, Boyan BD. Underlying mechanisms at the bone-biomaterial interface. J Cell Biochem 1994;56:340–347.

4. Boyan BD, Hummert TW, Dean DD, Schwartz Z. Role of mate-rials surfaces in regulating bone and cartilage cell response.Biomaterials 1996;30:977–985.

5. Shackelford JF. Bioceramics—Current status and future trends.Bioceram Mater Sci Forum 1999;293:99–106.

6. Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920–926.

7. Chou L, Marek B, Wagner WR. Effects of hydroxyapatitescoating crystallinity on biosolubility, cell attachment, andproliferation in vitro. Biomaterials 1999;20:977–985.

8. Klawitter JJ, Hulbert SF. Application of porous ceramics forthe attachment of load bearing internal orthopedic applica-tions. J Biomed Mater Symp 1972;2:161–229.

9. Bobyn JD, Pilliar RM, Cameron HU, Weatherly GC. The opti-mum pore size for the fixation of porous-surfaced metalimplants by the ingrowth of bone. Clin Orthop Relat Res1980;150:263–270.

10. Leary-Swan EE, Popat KC, Desai TA. Fabrication and evalua-tion of nanoporous alumina membranes for osteoblast cul-ture. J Biomed Mater Res A 2005;72:288–295.

11. Popat KC, Leary-Swan EE, Desai TA. Peptide immobilizednanoporous alumina membranes for enhanced osteoblast ad-hesion. Biomaterials 2005;26:1969–1976.

12. Lee TM, Wang BC, Yang YC, Chang E. Comparison ofplasma-sprayed hydroxyapatite coatings and hydroxyapa-tite/tricalcium phosphate composite coatings: In vivo study.J Biomed Mater Res 2001;55:360–367.

13. Bauer TW, Taylor SK, Jiang M, Medendorp SV. An indirectcomparison of third-body wear in retrieved hydroxyapatite-coated, porous, and cemented femoral components. ClinOrthop Relat Res 1994;298:11–18.

14. Bloebaum RD, Beeks D, Dorr LD, Savory CG, DuPont JA,Hofmann AA. Complications with hydroxyapatites particu-late separation in total hip arthroplasty. Clin Orthop RelatRes 1994;298:19–26.

15. Schwartz Z, Martin JY, Dean DD, Simpson J, Cochran DL.Effect of titanium surface on chondrocytes proliferation,matrix production and differentiation depends on thestate of cell maturation. J Biomed Mater Res 1996;30:145–155.

16. Price RL, Haberstroh KM, Webster TJ. Enhanced functions ofosteoblasts on nanostructured surfaces of carbon and alu-mina. Med Biol Eng Comput 2003;41:372–375.

17. Webster TJ, Ergun C, Doremus RH, Seigel RW, Bizios R.Enhanced functions of osteoblasts on nanophase ceramics.Biomaterials 2000;21:1803–1810.

18. Tiedeman JJ, Garvin KL, Kile TA, Connolly JR. The role of acomposite, demineralized bone matrix and bone marrow inthe treatment of osseous defects. Orthopedics 1995;8:1153–1158.

19. Rougraff BT, Kling TJ. Treatment of active unicameral bonecysts with percutaneous injection of demineralized bone ma-trix and autogenous bone marrow. J Bone Joint Surg Am2002;84:921–929.

20. Haynesworth SE, Goshima J, Goldberg VM, Caplan AI. Char-acterization of cells with osteogenic potential from humanmarrow. Bone 1992;13:81–88.

21. Prockop DJ. Marrow stromal cells as stem cells for nonhema-topoietic tissues. Science 1997;276:71–74.

22. Caplan AI, Fink DJ, Goto T, Linton AE, Young RG, WakitaniS, Goldberg VM, Haynesworth SE. Mesenchymal stem cellsand tissue repair. In: Jackson DW, editor. The Anterior Cruci-ate Ligament: Current and Future Concepts. New York:Raven Press; 1993.

23. Altman GH, Horan RL, Martin I, Farhadi J, Stark PR, VollochV, Richmond JC, Vunjak-Novakovic G, Kaplan DL. Cell dif-ferentiation by mechanical stress. FASEB J 2002;16:270–272.

24. Beresford JN, Bennett JH, Devlin C, Leboy PS, Owen ME.Evidence for an inverse relationship between the differentia-tion of adipocytic and osteogenic cells in rat marrow stromalcell cultures. J Cell Sci 1992;102(Part 2):341–351.

25. Wakitani S, Goto T, Pineda SJ, Young RG, Mansour JM,Caplan AI, Goldberg VM. Mesenchymal cell-based repair oflarge, full-thickness defects of articular cartilage. J Bone JointSurg Am 1994;76:579–592.

26. Seshi B, Kumar S, Sellers D. Human bone marrow stromalcell: Coexpression of markers specific for multiple mesenchy-mal cell lineages. Blood Cells Mol Dis 2000;26:234–246.

27. Bianco P, Robey PG. Marrow stromal stem cells. J Clin Invest2000;105:1663–1668.

28. Harakas NK. Demineralized bone-matrix-induced osteogene-sis. Clin Orthop Relat Res 1984;188:239–251.

29. Urist MR. Bone formation by autoinduction. Science 1965;150:893–899.

30. Urist MR, DeLange RF, Finerman GA. Bone cell differentia-tion and growth factors. Science 1983;220:680–686.

DIFFERENTIATION OF MSCs CULTURED ON NANOPOROUS ALUMINA SURFACES 963

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a

31. Wang J, Glimcher MJ. Characterization of matrix-inducedosteogenesis in rat calvarial bone defects. II. Origins of boneforming cells. Calcif Tissue Int 1999;65:486–493.

32. Wang J, Yang R, Gerstenfeld LC, Glimcher MJ. Characteriza-tion of demineralized bone matrix-induced osteogenesis inrat calvarial bone defects. III. Gene and protein expression.Calcif Tissue Int 2000;67:314–320.

33. Lee SJ. Alumina. Encyclopedia of Biomaterials and Biomed-ical Engineering. New York: Marcel Dekker; 2004. pp 19–27.

34. Karlsson M, Palsgard E, Wilshaw PR, Di Silvio L. Initialin vitro interaction of osteoblasts with nano-porous alumina.Biomaterials 2003;24:3039–3046.

35. Corning Incorporated Life Sciences. Preparation of CostarTranswell inserts for scanning electron microscopy. Mar 92004. Available at: http://www.corning.com/Lifesciences/technical_information/techDocs/sem_transwell_protocol.pdf.

36. Mehta SS. Analysis of the mechanical properties of bone ma-terial using nondestructive ultrasound reflectometry, PhDDissertation, The University of Texas Southwestern MedicalCenter at Dallas, 1995.

37. Weiner S, Traub W. Bone structure: From angstroms tomicrons. FASEB J 1992;6:879–885.

38. Landis WJ. The strength of a calcified tissue depends in parton the molecular structure and organization of its constituentmineral crystals in their organic matrix. Bone 1995;16:533–544.

964 POPAT ET AL.

Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a