Porosity and Pore Size of B-tricalcium Phosphate Scaffold Can

Available online at www.sciencedirect.com

Acta Biomaterialia 4 (2008) 1904–1915

www.elsevier.com/locate/actabiomat

Porosity and pore size of b-tricalcium phosphate scaffold caninfluence protein production and osteogenic differentiation ofhuman mesenchymal stem cells: An in vitro and in vivo study

Philip Kasten a,*, Ingo Beyen a, Philipp Niemeyer b, Reto Luginbuhl c,Marc Bohner c, Wiltrud Richter a

a Orthopaedic University Hospital Heidelberg, Schlierbacher Landstr. 200a, 69118 Heidelberg, Germanyb Department for Orthopaedics and Traumatology, University of Freiburg, Germany

c Dr. h.c. Robert Mathys Foundation, CH-2544 Bettlach, Switzerland

Received 7 December 2007; received in revised form 8 May 2008; accepted 14 May 2008Available online 11 June 2008

Abstract

The interaction of stem cells and ceramics in bone regeneration is still poorly understood. The aim of this study was to examine theinfluence of the porosity (25%, 65% and 75%) of b-tricalcium phosphate (TCP) ceramics on osteogenic differentiation of mesenchymalstem cells (MSC) in vitro and in vivo. For the in vitro portion of the study, TCP scaffolds loaded with MSC were kept in osteogenicinduction medium for 21 days. For the in vivo portion of the study, scaffolds loaded with undifferentiated MSC were implanted subcu-taneously into SCID mice for 8 weeks and compared with similarly implanted controls that were not loaded with MSC. Measurements oftotal protein as well as specific alkaline phosphatase (ALP) activity were taken as indicators of growth/matrix production and osteogenicdifferentiation. An increase in the total protein concentration was noted from day 1 to day 21 on the in vitro TCP 65% and TCP 75%scaffolds (p < 0.05) with no such increase noted in the TCP 25% specimens. However, the specific alkaline phosphatase activity increasedfrom day 1 to day 21 in all three in vitro specimens (p < 0.02) and reached similar levels in each specimen by day 21. In vivo, ALP activityof cell-loaded TCP 65% ceramics was higher when compared with both the TCP 25% and TCP 75% specimens (p < 0.046), and higher inthe TCP 75% than TCP 25% specimens (p = 0.008). Histology revealed mineralization by human cells in the pores of the TCP ceramicscaffolds with a trend toward greater calcification in TCP 65% and 75%. In summary, a higher porosity of TCP scaffolds does not nec-essarily mean a higher ALP activity in vivo. The distribution and size of the pores, as well as the surface structure, might play an impor-tant role for osteogenic differentiation in vivo.� 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Porosity; Stem cell; Osteogenesis; Ceramic structure; Bone

1. Introduction

Regenerative medicine with the use of stem cells is a rap-idly emerging field in the treatment of bone defects [1–5].Mesenchymal stem cells (MSC) can easily be expanded tohigh cell numbers [6] and addition of MSC facilitates thehealing of bone defects [7]. Cell suspensions are difficultto apply as they hardly remain in bony defects and do

1742-7061/$ - see front matter � 2008 Acta Materialia Inc. Published by Else

doi:10.1016/j.actbio.2008.05.017

* Corresponding author. Tel.: +49 6221 96 5; fax: +49 6221 96 6347.E-mail address: [email protected] (P. Kasten).

not provide any biomechanical stability. Therefore, MSCare combined with biomaterials in a tissue engineeringapproach [8–11]. Calcium phosphate ceramics are usedclinically because they combine good stability with porosityand interconnectivity, and they are non-toxic during thedissolution and degradation process [11–13]. Moreover,they allow the adhesion and growth of MSC and osteo-blasts [14]. Among calcium phosphate ceramics, b-trical-cium phosphate (TCP), which dissolves in the presence ofacids released by cells such as osteoclasts or macrophages,is distinguished from hydroxyapatite (HA), which is hardly

vier Ltd. All rights reserved.

mailto:[email protected]

Table 1APhysical parameters of the biomaterials

Material Totalporosity(%)

Pore diameter

600–200 lm(%)

200–50 lm(%)

50–5 lm(%)

<5 lm(%)

TCP 25 block forms(Cerasorb�,Curasan)a

�25 ± 5 0 2 5 18

TCP 25 granules, 1000–2000 lm(Cerasorb�,Curasan)b

�59.3 ± 2.5 20 10 5 20

TCP 65 blockforms(Cerasorb M�,Curasan)

�65 ± 5 5 15 20 25

TCP 75 blockforms(RMS)c

�75 ± 0.4 54 0 0 21

These data were provided by the manufacturers.a The block forms are usually provided with macroporosity by drilling

interconnecting macropores with a diameter of 0.5–2.0 mm to reach anoverall porosity of 65–80%.

b The overall porosity of granulates is calculated by the intergranularporosity (measured by Hg porosimetry) and intergranular voids.

c Structure similar to that of chronOSTM.

Table 1BPhysical parameters of the biomaterials

Material Specificsurfacearea (m2 g�1)

Surface/apparentvolume(m2 cc�1)

Surface/scaffold(m2)

TCP 25 block forms(Cerasorb�, Curasan)a

0.12 0.28 0.0076

TCP 65 block forms(Cerasorb M�,Curasan)

0.18 0.2 0.0052

TCP 75 block forms(RMS)c

0.3 0.23 0.0068

P. Kasten et al. / Acta Biomaterialia 4 (2008) 1904–1915 1905

degradable at all [15]. In general, it seems favourable tohave cell-mediated degradation that proceeds at the samespeed as new bone forms, as this allows the formation ofbone with homogeneous elasticity and reduced fracturerisk [16]. Critical factors that may determine the successof a MSC/biomaterial construct for osteogenesis include

� the initial adhesion of the MSC;� survival of the MSC on the biomaterial;� cell proliferation after loading; and� the extent of osteogenic differentiation.

It is well described in previous studies that the porosityof the biomaterial plays a significant role in the success ofan MSC/biomaterial construct [17–19]. Although severalsolid TCP ceramics are on the market, there is sparse dataindicating what degree of porosity is most favourable foradhesion, proliferation and osteogenic differentiation ofMSC in vitro and in vivo [20,21].

To this end, we examined three different TCP blockmaterials that are often used clinically: (i) Cerasorb�, (ii)Cerasorb M� (both Curasan, Kleinostheim, Germany)and (iii) a TCP produced by the Dr. Robert Mathys Foun-dation (RMS, Bettlach, Switzerland) that is similar to chro-nOSTM (Synthes). Cerasorb� granules served as the positivein vivo control because of their well-known osteoconduc-tive properties. The porosity between the tested biomateri-als differed from 25% to 75% and pore size ranged from<10 to 600 lm. TCP Cerasorb� block material with alow porosity of 25% was included into the study as a neg-ative control since it is known that small pore size is notfavourable for bone formation in vivo [22]. Nevertheless,orthopaedic surgeons like to use this material because ofits strong mechanical properties (e.g. in open wedge osteot-omy of the acetabulum or the tibia). We examined adhe-sion, protein production and osteogenic differentiation ofMSC on these ceramics for 3 weeks in vitro and osteogenicdifferentiation of freshly loaded undifferentiated MSC com-posites in vivo for 8 weeks.

2. Material and methods

2.1. TCP scaffolds

Three different porous TCP block forms were used:Cerasorb� (TCP 25), Cerasorb M� (TCP 65) and theTCP by RMS (TCP 75). In comparing these constructs,there are significant differences regarding micro- andmacroporosity, while the specific surface area (SSA, inm2 g�1) is equally low in all of them (Tables 1A and 1B).In the in vivo assay, Cerasorb� granules with particlesranging from 1000 to 2000 lm and micropores of <5 lmwere used as a positive control. To illustrate the surfacethat is accessible to the cells, we calculated the surfaceper apparent volume (m2 cc�1 = SSA � density �(1-porosity)) and the total surface per scaffold (m2 =

weight of scaffold � SSA). Phase purity was high, at morethan 99% in all scaffolds.

The TCP for the Cerasorb� products was produced by asolid-state reaction from calcium carbonate and calciumhydrogen phosphate. Microporous Cerasorb� block formswere produced by milling after cold isostatic pressing and afinal sintering step. For producing Cerasorb M� blockforms with additional meso- and macropores, an organicporogen substance was added for sintering before pressing.The porogen substance burns away, leaving pores behind.The TCP by RMS was synthesized in an emulsion processand subsequent sintering [15]. All ceramic bodies weremachined in cuboids (5 � 3 � 2 mm) with a volume of30 mm3. The Cerasorb� granule samples had the sameweight as the block forms.

2.2. Structural and chemical analysis of the ceramics

The distribution of the pore sizes was analysed by mer-cury intrusion porosimetry (Pascal 140-240/440, Porotec,

1906 P. Kasten et al. / Acta Biomaterialia 4 (2008) 1904–1915

Hofheim, Germany). In addition, the scaffolds wereobserved by scanning electron microscopy (SEM;Cambridge S360) to determine the microporosity accordingto a protocol as published previously [23]. Microcomputer-ized tomography (l-CT) analysis was performed with alCT80 Micro-CT instrument (Scanco Medical, Switzer-land) using a high-resolution protocol (slice distance10 lm; in-plane pixel size 10 lm) to evaluate the porosityof the scaffolds. The microfocus of the X-ray source ofthe l-CT system had a spot size of 5 lm and a maximumvoltage of 70 kVp. Approximately 100 slices mm�1 wereacquired and three-dimensional (3-D) volumes recon-structed. The reconstructed 3-D images were analysed with3-D image analysis software (VGStudio Max 1.2.1, Vol-ume Graphics, Heidelberg, Germany; and IPL (Image Pro-cessing Language), Scanco Medical, Switzerland). Thesoftware programs allow for calculating the ‘‘trabecularthickness” and ‘‘trabecular separation”, which correspondin our experiment to the wall of the pores and the porediameter.

2.3. Isolation and cultivation of human mesenchymal stemcells

For the in vitro experiments, samples of human MSC(n = 5) were obtained from iliac bone marrow aspirates.This procedure was performed before autogenous bonegrafting was extracted from the iliac crest. Cells were alsorecovered from the femoral medullary canal after femoralosteotomy under general anesthesia. The age of the donors(three men, two women) ranged from 11 to 66 years(mean ± standard deviation (SD) = 28.2 ± 33.2). In detail,the patients were operated on for congenital hip dysplasia,spinal stenosis, wrist contracture and in two cases forosteoarthritis of the hip.

For the in vivo experiments, we pooled the MSC ofthree donors (one woman, three men, aged 11–24 years)from iliac crest bone marrow aspirates. We pooled theMSC to level out known inter-individual differences in pro-liferation and osteogenic differentiation [6,24]. The latterpatients were operated for Perthes–Legg–Calvee diseaseand hip dysplasia (�2), and the mean age was 15.6 ± 7.6.All donors were haematologically healthy Caucasians.Co-morbidities affecting bone metabolism were not pres-ent. All procedures were approved by the institutional eth-ics committee and all donors provided informed consent.MSC isolation and expansion were performed as reportedpreviously [6,24].

2.4. Loading of cells and 3-D culture of biocomposites

All ceramics were incubated for 12 h in 30 lg ml�1 fibro-nectin/PBS-solution (Calbiochem/Merk, Darmstadt) anddried for 10 min under a sterile flow at room temperature.For the in vitro experiments, the samples were then stati-cally loaded with 2 � 105 MSC in 10 ll of culture medium.The scaffolds for the in vitro experiments were cultured for

up to 3 weeks in an osteogenic induction medium contain-ing Dulbecco’s modified Eagle’s medium (DMEM; Gibco/Invitrogen, Karlsruhe, Germany), with the addition of 10%foetal calf serum, 0.1 lM dexamethasone (Sigma, Taufkir-chen, Germany), 0.17 mM ascorbic acid-2-phosphate(Sigma), 10 mM b-glycerolphosphate (Sigma) and 100IE ml�1 penicillin and 100 lg ml�1 streptomycin (Bio-chrom, Berlin, Germany). Samples were harvested for eval-uation on days 1, 7 and 21. All biomaterials were tested invitro with the MSC of five different patients.

For the in vivo experiments 1 � 106 undifferentiatedMSC in 15 ll of culture medium were applied to the scaf-folds. Ceramic scaffolds without MSC that served as con-trols were treated identically, but did not receive MSC.

2.5. Cell-loading efficacy and metabolic activity on day 1

To assess the cell-loading efficacy, the scaffold wasremoved from the well in which the loading of the scaffoldtook place. Two approaches were employed: in the first, welooked at the number of metabolically active cells on thebiomaterial; in the second, the cells in the well were mea-sured in the WST-1 test. Cell-loading efficacy was definedas the ratio of metabolic cells on the scaffold divided bythe number of cells on the scaffold and in the well. The cellvitality on day 1 was calculated by dividing the metabolicactivity of cells on the scaffold according to the WST-1 testdivided by the metabolic activity of cells that were applied,i.e. 2 � 105.

The reagent WST-1 (Roche Diagnostics GmbH, Mann-heim, Germany) was diluted in a ratio of 1:10 withDMEM. The osteogenic induction medium was removedfrom the wells and the matrices were transferred into a96-well plate; then 250 ll of diluted WST-1 reagent wasput either into each well where the cell loading was doneor added to the matrices or to a monolayer of 2 � 105

cells. During a 30 min incubation period at 37�C, WST-1 was converted into a coloured formazan salt by the met-abolic activity of the living cells. To quantify the amountof formazan, 100 ll of each sample was put into a sepa-rate 96-well plate and measured using an ELISA plate-reader at 450 nm. A standard curve was prepared in amonolayer with different cell numbers in the samemanner.

To account for the artefact produced by the absorptionof the formazan salt onto the empty scaffolds, a correctionfactor was calculated in triplicate for each biomaterial. Forthat purpose, 2 � 105 cells in a monolayer were incubatedwith the cell proliferation reagent WST-1. Then the solubleformazan salt was measured in the ELISA plate reader asdescribed above. Afterwards, each empty scaffold was incu-bated in the formazan-containing supernatant for 60 minat 37�C. A 100 ll quantity of each sample was measuredin a separate 96-well plate using an ELISA plate-readerat 450 nm. By dividing the initial optical density (OD) bythe OD after the incubation with the empty scaffolds, a cor-rection factor was obtained for each biomaterial: 18.59%


for Cerasorb�, 17.56% for Cerasorb M� (both Curasan,Kleinostheim, Germany) and 17.74% for TCP by RMS.

2.6. Protein production and osteogenic differentiation

The protein production was assessed by measuring theamount of protein on days 1, 7 and 21 using the MicroBCATM Protein Assay Kit (Fa.Pierce, Rockford, IL,USA), as described previously [24]. Osteogenic differentia-tion was assessed by measuring alkaline phosphatase(ALP) activity. The matrices were incubated with 1 ml0.1% Triton X-100 (Sigma–Aldrich Chemie GmbH,Taufkirchen, Germany) and homogenized mechanicallyto disintegrate the cells using a Polytron (dispersing andmixing technology, Model PT2000, Kinematica Inc., Cin-cinnati, OH, USA). The cell lysate was centrifuged (460g

for 10 min), the supernatant was recovered and the ALPenzyme activity was determined according to the followingprotocol: a 50 ll sample extract was incubated with 100 llof p-nitrophenyl-phosphate (100 mg capsule, Sigma 104�

Phosphatase Substrate, Sigma–Aldrich Chemie GmbH,Taufkirchen, Germany) and 50 ll of distilled water. Theconversion to p-nitrophenol was measured after 30 min at405/490 nm. In detail, one capsule was diluted with12.67 ml of an alkaline buffer (150 mM NaHCO3, 1 mMMgCl2; pH 10.3) to achieve a 30 mM p-nitrophenyl phos-phate solution. A standard curve was prepared with ap-nitrophenol standard solution (10 lmol ml�1; Sigma–Aldrich Chemie GmbH, Taufkirchen, Germany). For this,a 1:2 dilution series ranging from 25 to 0.195 lg ml�1

p-nitrophenol in 0.01% Triton X-100 was prepared andmeasured as described above. The amount of ALP wasdivided by the protein value to obtain the specific amountof ALP product.

2.7. Animals and surgical procedures

Three experimental groups (each n = 8) were tested: (i)Cerasorb�, (ii) Cerasorb M� and (iii) TCP by RMS. Afterloading the biomaterials with 1 � 106 undifferentiated cellsas described above two cell-constructs were subcutaneouslyimplanted into adult male SCID mice (Charles RiverDeutschland GmbH, Sulzfeld, Germany; C.B. 17-scid/leR-Crl). Two additional empty scaffolds were always insertedas controls. The total of implanted scaffolds per mouseconsequently was four with one exception: in the micewhich received Cerasorb�, not only Cerasorb� granulesas well as block forms were implanted as a positive control[22]. A 100 ll quantity of fibrin glue (Baxter DeutschlandGmbH, Unterschleißheim, Germany) was added to thegranules to form a clot and prevent their disassembly. Withthe animals under general intraperitoneal anesthesia andafter disinfection of the backs of the mice, two subcutane-ous pockets were created bluntly through two separate1 cm incisions on the back. After implantation of the scaf-folds, the wounds were closed with single interruptedsutures. The SCID mice, weighing about 30 g, were housed

in Macrolon cages size MII (Fa.Tecniplast DeutschlandGmbH, Hohenpeißenberg, Germany) and fed a specialmurine diet (V1185-300 ssniff M-Z, Fa. ssniff SpezialdiatenGmbH, Soest, Germany). To account for the variabilitycaused by the SCID mice, eight mice were used in eachgroup, in accordance with a pilot study and a biometricanalysis [24,25]. All but one from the Cerasorb� group of24 animals survived until the end of the experiment. At 8weeks the animals were sacrificed and the scaffolds har-vested. One cell/biomaterial construct and one empty con-trol were used for biochemical evaluation and a second wasused for histological evaluation.

2.8. Histology

For light microscopy, samples were fixed and embeddedin paraffin wax (Leica, Bensheim, Germany) as describedpreviously [24]. Sections were stained with toluidine blue(Waldeck GmbH & Co. Division Chroma, Munster, Ger-many). Three sequential sections 5 lm thick at a defineddepth of 0.75 mm were selected for evaluation. The blindedsamples were examined at a magnification of �100 by twoindependent observers to identify the type of tissue (bone-like, cartilage-like or connective tissue). Samples with tis-sues other than connective tissue were stained with alizarinred (Waldeck GmbH) for calcium deposits. The mineral-izations were graded as follows: 0 = no mineralization bythe cells in the region of interest (ROI); + = mineralizationaround <3 cells in ROI; ++ = mineralization around >3cells in the ROI. The ROIs were located in the area of vis-ible calcifications at the edges of the biomaterial.

A digoxigenin-labeled probe specific for the human Alurepetitive DNA element was prepared by polymerase chainreaction as described previously [24]. The following primerswere used: forward 50-CGAGGCGGGTGGATCATGAGGT-30, reverse 50-TTTTTTGAGACGGAGTCTCGC-30

[26]. Signals were detected by immunohistochemistry usinganti-digoxigenin alkaline phosphatase-conjugated Fab frag-ments (Roche) and NBT/BCIP (Roche) as substrates.

2.9. Statistical analysis

Data analysis was performed with SPSS for Windows15.0 (SPSS Inc., Chicago, IL, USA). Descriptive statisticsand non-parametric tests were employed because of smallsample sizes and distributional characteristics. Differencesbetween subgroups were analysed with Mann–WhitneyU-tests. The significance level was assigned at 5% (two-tailed) due to the explorative nature of the study.

3. Results

3.1. Structural and chemical analysis of the TCP scaffolds

A wider range of pore diameters was observed by mer-cury intrusion porosimetry in the TCP 75 as compared tothe other two ceramics (Fig. 1). The latter two had most

0%

20%

40%

60%

80%

100%

0 20 40 60 80 100 120

Pore diameter [microns]

Por

e vo

lum

e fr

actio

n [%

]

ß-TCP 75

ß-TCP 65

ß-TCP 25

Fig. 1. Mercury intrusion porosimetry demonstrating the pore sizedistribution in relation to the total pore volume. A wider range of porediameters was observed by mercury intrusion porosimetry in the TCP 75than in the other two ceramics.


of their porosity with a pore diameter below 10 lm. The‘‘steps” along the curve of the TCP 75 sample can beexplained either by breaking of the walls of isolated poresor by filling small interconnections between the pores.The absence of such steps in the curve of the two otherceramics, TCP 25 and 65, suggest that either the blocks

Fig. 2. (A) SEM pictures of TCP 25, (B) TCP 65 and (C) TCP 75, each atscaffolds. The surface seems to be smooth in the �2000 magnification of the Tmore irregular in shape.

were mechanically stable throughout the measurement orthe macropores were interconnected by only very smallpores (in the order of 6–7 lm). The SEM pictures (Fig. 2)support the latter explanation. The microstructures of theTCP 25 and TCP 65 samples were similarly smooth (Fig.2), while the TCP 75 sample had more rough particles.The l-CT analysis revealed that TCP 75 had a larger mac-ropore diameter than TCP 65, while their pore wall thick-nesses were similar (Table 2 and Fig. 3). The porosity wasmeasured as 56.22% for TCP 75 and 29.6% for TCP 65.Since pores below 5 lm could not be detected by l-CT,the values of total porosity were generally underestimated.The l-CT analysis of the pores of TCP 25 did not provideconclusive data because the pores were in the same sizerange as the voxel, i.e. too small to be detected properly.

3.2. Loading efficacy and cell vitality on day 1

Loading efficacy ranged from 71% for TCP 25 (SD20.8%) and TCP 65 (SD 26.5%) to 80% (SD 22.5%) forthe TCP 75 form blocks. There was no significant differenceregarding loading efficacy among the various TCP ceram-

�100, �500 and �2000 magnification, showing the microporosity of theCP 25 and TCP 65 scaffolds, whereas the surface of the TCP 75 samples is


ics. The mean values of metabolic activity per 2 � 105 cellson day 1 were higher in the groups with higher porosity,but this trend did not reach significance due to high stan-dard deviations. In detail, metabolic activity on day 1was 66.3 ± 43.6% on TCP 25, 99.3 ± 59.2% on TCP 65and 89.7 ± 48.4% on TCP 75.

3.3. Protein production in vitro and in vivo

Protein content increased significantly in vitro from day1 to day 21 on TCP 65 and TCP 75 and in the monolayer(p = 0.01, p = 0.048 and p = 0.01 respectively), but not on

Table 2l-CT analysis of ceramics

l-CT analysis TCP 65 (n = 3) TCP 75 (n = 3)

Porosity (%)a 29.6 ± 1.47 56.22 ± 1.88Macropore diameter (lm) 41.03 ± 2 135.6 ± 6.49Macropore wall thickness (lm) 97.56 ± 2.45 105.53 ± 4.33

a The l-CT cannot detect pores below 5 lm properly and consequentlythe porosity is underestimated.

Fig. 3. 3-D reconstructions of the l-CT slices illustrating the d

Fig. 4. Proliferation and matrix deposition in vitro assessed by measuring the pfrom day 1 to day 21 on TCP 65 and TCP 75 and in the monolayer, but not

TCP 25 (Fig. 4). Consistent with the higher porosity, therewere higher protein contents in the TCP 75 and TCP 65groups than in TCP 25 in vivo (both p = 0.001). Therewas no statistically significant difference between the TCP65 and TCP 75 groups in vivo (p = 0.059). The loadingwith undifferentiated cells did not increase the total amountof protein in vivo compared with empty scaffolds.

3.4. Osteogenic differentiation in vitro and in vivo

The specific ALP activity in vitro increased in all cell/biomaterial groups over 21 days (p < 0.02). However, therewere no significant differences between the groups (Fig. 5).These findings were in contrast to the in vivo ALP activity,which was highest in the TCP 65 + cells group, followed bythe TCP 75 + cells group (p = 0.046), while the TCP25 + cells group presented values close to those of the con-trol group (Fig. 6). There was higher ALP activity in theTCP 65 + cells groups and TCP 75 + cells groups (p<0.009) but not the TCP 25 + cells groups compared withthe empty scaffolds (Fig. 6).

istinct porosity, pore size and distribution of the scaffolds.

rotein content over 21 days. Protein content increased significantly in vitroon TCP 25. Mean values and SD are shown.


The histological sections revealed that a penetration ofcells into the inner parts of the scaffold occurred in the bio-materials with higher porosity (TCP 65 and 75) (Fig. 7).Semi-quantitative grading of histologic sections revealedmore mineralization in TCP 65 and TCP 75 cell-loaded

Specific ALP A

0

2

4

6

8

10

12

14

16

18

day1 day7 day21 day1 day7 day21 day1 day7 day21

-TCP 25 + cells -TCP 25 -TCP 65 + cells

[p-n

itro

ph

eno

l/ p

rote

in]

p=0.016

p=0.016

β β β

Fig. 5. ALP activity in vitro normalized to the protein content. The specific Awithout significant differences. Mean values and SD are shown.

Specific ALP

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

0.018

Cells Controls Cells

-TCP 25 -

[p-n

itro

ph

eno

l/ p

rote

in]

p=0.008

p

p=0.004

β β

Fig. 6. TCP ceramics with the higher porosity yielding higher ALP activity tha

explants than in the corresponding TCP 25 explants (Table3, Fig. 8A and B). The in situ hybridization confirmed thathuman cells were present in the block forms after 8 weeks(Fig. 8C). The positive control with the granules showedthe strongest mineralization, as expected (Fig. 8D and E).

ctivity in vitro

day1 day7 day21 day1 day7 day21 day1 day7 day21

-TCP 65 -TCP 75 +cells -TCP 75

mean value (n=5)

p=0.008

β β β

LP activity in vitro increased in all cell/biomaterial groups over 21 days

Activity in vivo

Controls Cells Controls

TCP 65 -TCP 75

mean value (n=8)

=0.009

p=0.001

p=0.046

β

n the one with the lower porosity in vivo. Mean values and SD are shown.

Fig. 7. Toluidine blue staining at �10 magnification of TCP 25 (A), TCP 65 (B) and TCP 75 (C) in vivo specimens, showing that there were fewer cellswithin the TCP 25 scaffold.

Table 3Mineralization after 8 weeks

TCP 25 blockforms

TCP 65 blockforms

TCP 75 blockforms

0 + ++ 0 + ++ 0 + ++

Empty 7/7 0/7 0/7 6/7 1/7 0/7 6/8 2/8 0/8With cells 7/7 0/7 0/7 2/7 5/7 0/7 2/8 6/8 0/8

0 = no mineralization around cells in region of interest (ROI); + = min-eralization around <3 cells in ROI; ++ = mineralization around >3 cellsin ROI.


4. Discussion

Our data demonstrate that porosity and pore size of dis-tinct TCP scaffolds influence not only protein productionin vitro and in vivo but also specific ALP activity, whichis an important marker for osteogenic differentiation.

Scaffolds for osteogenesis should mimic bone morphol-ogy, structure and function in order to optimize integrationinto surrounding tissue [21]. Generally, bone has a verycomplex structure. The cortical bone has a compact struc-ture with only 3–12% porosity. It contains a series of voids,e.g. haversian canals with a diameter of 100–200 lm [27].Trabecular bone has a porosity in the range of 50–90%and pore diameters close to 1 mm [28]. The complexity ofarchitecture and the variability of properties of bone tissue(e.g. porosity, pore size, mechanical properties, mineraliza-tion or mineral density, cell type and cytokines gradientfeatures), as well as differences in age, nutritional state,activity and disease status of individuals, establish a majorchallenge in fabricating scaffolds that meet the needs ofspecific repair sites in specific patients [21]. A key issue tocompensate for the complexity of bone tissue is to achievea fast replacement of the bone substitute with new maturebone. We therefore chose for our experiments the biode-gradable ceramic TCP, which is often used to fill bonedefects in orthopedic surgery.

4.1. Effect of pore size and porosity in vivo

A parameter that is independent of the material isporosity, or the percentage of void space in a solid. Pores

are necessary for bone tissue formation because they allowfor migration and proliferation of osteoblasts and mesen-chymal cells, as well as vascularization [29]. The minimumpore size required to regenerate mineralized bone is gener-ally considered to be 50–100 lm [17,30,31]. Smaller poresresult in ingrowth of unmineralized osteoid tissue, and evensmaller ones are penetrated only by fibrous tissue [30]. Arecent study examining different macropore diameters(150, 260, 510 and 1220 lm) and a total porosity of 75%of TCP in metaphyseal defects in sheep at 6 and 12 weeksrevealed that the biological response to implantation wasmarginally influenced by the pore size [20]. There was a bal-ance of factors; however, faster ceramic resorption andslower bone formation were found to occur in 510 lm poresamples compared to 150 and 260 lm pore samples.

Pore diameter is not the only important parameter in anexperimental setting; other parameters include the animalspecies, block size, implant chemistry, implant topography,resorption and interconnectivity. It was hypothesized thatin resorbable materials, pore density and interconnectiondensity are more important than pore size, contrary tounresorbable materials, in which the sizes and densitiesare equally important [32]. Also, according to a study byBohner and Baumgart [33], bone ingrowth should not beaffected by the pore diameter as long as the structure isfully interconnected and the pore interconnections have adiameter larger than 50 lm. In a non-fully interconnectedscaffold, bone ingrowth should generally be faster whenlarger macropores are present [29,34]. Although a few stud-ies show no effect of porosity on the amount of appositedbone [35,36], higher porosity is usually associated withgreater bone formation [37–40]. Although increased poros-ity and pore size facilitate bone ingrowth, the result is areduction in mechanical properties, since the structuralintegrity of the scaffold is compromised [41,42].

Our hypothesis was that a higher level of porosity wouldincrease osteogenic differentiation in vivo. However, in ourstudy TCP 65 yielded a higher ALP activity than TCP 75.This difference must be related to physical, morphologicalor chemical aspects. In terms of chemistry, both TCP 65and TCP 75 consisted of more than 99% TCP. Lookingat the architecture of the pores, several differences can be

Fig. 8. (A) TCP 65 in vivo specimen displaying an intense penetration by cells but no mineralization in the alizarin red/fast green staining (�20magnification). (B) A representative TCP 75 in vivo specimen with MSC showing mineralization (black arrow) in the alizarin red/fast green staining. (C)The ALU in situ hybridization of a serial section demonstrating the presence of human cells (black nuclei, arrow head) in the pores with mouse tissuearound the human cells (both �100 magnification). (D) The positive control of the in vivo TCP 25 granules showing mineralization in the �20 and (E) the�100 magnification.


noticed. The macropores of TCP 65 were smaller and moreirregular in shape, had on average a smaller diameter andhad fewer interconnections than those of TCP 75, asobserved in l-CT and histology (Figs. 3 and 7). Further-more, the microporosity differed qualitatively betweenTCP 65 and TCP 75: the surface structure of TCP 65seemed to be smoother than that of TCP 75 (Fig. 2). Thisis also reflected by the slightly lower specific surface area,surface per volume and surface per scaffold (Table 1B).In our opinion, a smoother surface at the �2000 magnifica-tion level seems to be more favourable for cell differentia-tion in vivo.

All our samples were resorbable and had a rather lowinterconnectivity, but had strong differences in pore sizeand total porosity: TCP 25 mainly had pores up to 2–10 lm and a relatively low total porosity of 25%. Accord-ing to the aforementioned studies, it is unlikely that bonecan form in solid blocks of TCP 25. The situation is com-pletely different, however, if granules are used. In this case,there is enough space between the granules to allow forbone formation. Indeed, biomaterials in the form of gran-ules demonstrated a higher rate of bone formation com-pared with solid scaffolds [22]. However, many surgeonsfavour solid biomaterials as they provide better handlingand greater mechanical stability. TCP 65 with a totalporosity of 65% has pores mainly ranging from 2 to100 lm, allowing bone formation to occur within the pores.TCP 75 has a higher total porosity of 75% and pores

mainly ranging from 200 to 600 lm, with some smallerthan 5 lm. Indeed, we could find mineralization producedby human MSC within larger pores and higher specificALP activity in the biomaterials with the larger pores ofthe TCP 65 and 75. However, there was no bone formationin the histological specimens of any ceramics. Donor-dependent heterogeneity, such as age [43–45] and disease[46], may influence the yield and proliferative capacity ofstem cells or their osteogenic potential [47,48]. The lackof bone formation may be attributed to a known donorvariability that cannot be completely ruled out by a preop-erative selection [24,47–49]. In addition, the time course ofimplantation may have been too short for bone formation.Furthermore, the model of implanting subcutaneously, i.e.in a poorly vascularized space, is not ideal for strong calci-fications [14,49].

One limitation of the study is that different donors wereused for the in vitro and in vivo approaches because of lim-ited cell numbers. This prevented any direct donor-to-donor comparison of the two settings. However, withinthe in vivo and in vitro approaches all scaffolds receivedthe same cell populations, and consequently a comparisonbetween the scaffolds within one setting was possible. Themean age in the in vitro group (28.6 years old) was slightlyhigher than in the in vivo group (15.6 years old). In con-trast to MSC data from rodents, where younger donorsshowed better osteogenic differentiation capacity thanolder donors [43], no such differences were observed for


human MCS [48,50,51]. This suggests that the slight differ-ence in donor age between the in vitro and in vivo groups isnot a limitation of this study.

Another issue is that in the current experimental designthe authors relied on the analysis of ALP activity as themain biochemical marker for osteogenic differentiation.However, the ALP activity can reliably be measured andcompared to previous studies [34,52–55]. ALP is a markerof the protein level that does not rise in uninduced MSC,but displays a strong induction during osteogenic differen-tiation [56]. The time curve of ALP activity with anincrease in the first weeks seemed suitable for the experi-mental design [55]. To further assess osteogenic differentia-tion in vivo with a separate parameter, we chose to usehistology [57]. However, other osteogenic markers, suchas osteonectin and osteopontin, could be included in thefuture studies.

One limitation of the study is that the ALP activity invivo was normalized to the total protein content. The lattermight be influenced by co-extraction of host tissue, espe-cially in the ceramics with the higher porosities. Indeed,the total protein content was significant higher in the scaf-folds with the higher porosities that might be derived byproliferation of human MSC or host tissue. The ALP activ-ity of the human MSC e.g. in the TCP 65, which was nor-malized to total protein content, was strong enough todisplay a significant difference to e.g. TCP 25 scaffolds,which had a lower total protein content. Therefore, theconclusions regarding the ALP activity in vivo are valid.Another option of normalization would have been to thescaffold. However, since the scaffolds are brittle, duringthe course of the experiments the edges sometimes brokeoff and the volume of the scaffold was thus reduced atthe time of retrieval. This reduced the ALP activity perscaffold and caused higher standard deviations.

4.2. The effect of porosity and pore size in vitro

There are few studies about the effect of porosity on pro-liferation and osteogenic differentiation in vitro. In onestudy, composites of apatite and collagen with pores rang-ing from 50 to 300 lm and porosities of 49–79% exhibitedno significant differences of MC3T3-E1 osteoblast prolifer-ation [58]. In another study, smaller pores (0.4 and 13 lm)in TiO2 films enhanced the proliferation of human cellstrypsinized from bone in contrast to larger pores (49 lm)[59]. Human MSC that were kept in dynamic spinner flaskcultivation exhibited a faster rate of osteogenic differentia-tion in coralline hydroxyapatite scaffolds with 200 lm poresize than with 500 lm pores, but proliferation was higher inthe 500 lm scaffolds [53]. In summary, it was proposed thatosteogenic differentiation in vitro was not affected by poresize, but was enhanced by a low porosity [21]. Our datacannot support this statement completely, since ALP activ-ity in our setting was independent of the total porosityranging from 25% to 75% and pore size. In our experimentthe porosity did not affect cell attachment, in agreement

with previous studies [60]. A higher porosity in our settingincreased the protein content of human MSC, since thepore space increased with porosity and facilitated transportof oxygen and nutrients. This affirms one study withhuman MSC [53], but is in contrast to another study withosteoblast cell lines [58].

5. Conclusion

In vitro porosity was beneficial for protein production,but did not influence osteogenic differentiation. In vivo,the higher porosities of 65% and 75% yielded higher ALPactivity than the 25% porosity. Comparing the two highestporosities, the TCP 75 allowed for lower ALP activity thanTCP 65. In summary, higher porosity does not necessarilymean higher ALP activity in vivo. The distribution and sizeof the pores, as well as the surface structure, might play animportant role for osteogenic differentiation in vivo. How-ever, the exact mechanisms have not been determined todate. Future studies will have to identify other relevantparameters of the scaffolds for osteogenic differentiation.

Acknowledgements

The work was performed in the Division of Experimen-tal Orthopaedics, Orthopaedic University Hospital Heidel-berg, Heidelberg, Germany. We thank Curasan AG,Kleinostheim, Germany and Dr. Robert Mathys Founda-tion, Bettlach, Switzerland for their support in providingthe biomaterials. Biopharm, Heidelberg, Germany andthe research fund of the Orthopaedic University Hospital,Heidelberg, Germany supported the study financially.

References

[1] Vacanti CA, Bonassar LJ. An overview of tissue engineered bone.Clin Orthop 1999(367 Suppl):S375–81.

[2] Laurencin CT, Ambrosio AM, Borden MD, Cooper Jr JA. Tissueengineering: orthopedic applications. Annu Rev Biomed Eng1999;1:19–46.

[3] Petite H, Viateau V, Bensaid W, Meunier A, de Pollak C, Bourgui-gnon M, et al. Tissue-engineered bone regeneration. Nat Biotechnol2000;18(9):959–63.

[4] Bruder SP, Fink DJ, Caplan AI. Mesenchymal stem cells in bonedevelopment, bone repair, and skeletal regeneration therapy. J CellBiochem 1994;56(3):283–94.

[5] Ohgushi H, Caplan AI. Stem cell technology and bioceramics: fromcell to gene engineering. J Biomed Mater Res 1999;48(6):913–27.

[6] Reyes M, Lund T, Lenvik T, Aguiar D, Koodie L, Verfaillie CM.Purification and ex vivo expansion of postnatal human marrowmesodermal progenitor cells. Blood 2001;98(9):2615–25.

[7] Hernigou P, Poignard A, Beaujean F, Rouard H. Percutaneousautologous bone-marrow grafting for nonunions. Influence of thenumber and concentration of progenitor cells. J Bone Joint Surg Am2005;87(7):1430–7.

[8] Bruder SP, Kraus KH, Goldberg VM, Kadiyala S. The effect ofimplants loaded with autologous mesenchymal stem cells on thehealing of canine segmental bone defects. J Bone Joint Surg Am1998;80(7):985–96.

[9] Kon E, Muraglia A, Corsi A, Bianco P, Marcacci M, Martin I, et al.Autologous bone marrow stromal cells loaded onto porous hydroxy-


apatite ceramic accelerate bone repair in critical-size defects of sheeplong bones. J Biomed Mater Res 2000;49(3):328–37.

[10] Puelacher WC, Vacanti JP, Ferraro NF, Schloo B, Vacanti CA.Femoral shaft reconstruction using tissue-engineered growth of bone.Int J Oral Maxillofac Surg 1996;25(3):223–8.

[11] Quarto R, Mastrogiacomo M, Cancedda R, Kutepov SM, Mukha-chev V, Lavroukov A, et al. Repair of large bone defects with the useof autologous bone marrow stromal cells. N Engl J Med2001;344(5):385–6.

[12] Marcacci M, Kon E, Zaffagnini S, Giardino R, Rocca M, Corsi A,et al. Reconstruction of extensive long-bone defects in sheep usingporous hydroxyapatite sponges. Calcif Tissue Int 1999;64(1):83–90.

[13] Rueger JM. Bone substitution materials. Current status and pros-pects. Orthopade 1998;27(2):72–9.

[14] Kasten P, Luginbuhl R, van Griensven M, Barkhausen T, Krettek C,Bohner M, et al. Comparison of human bone marrow stromal cellsseeded on calcium-deficient hydroxyapatite, b-tricalcium phos-phate and demineralized bone matrix. Biomaterials 2003;24(15):2593–603.

[15] Bohner M. Calcium orthophosphates in medicine: from ceramics tocalcium phosphate cements. Injury 2000;31(Suppl 4):37–47.

[16] Wiesmann HP, Joos U, Meyer U. Biological and biophysicalprinciples in extracorporal bone tissue engineering. Part II. Int JOral Maxillofac Surg 2004;33(6):523–30.

[17] Kuboki Y, Saito T, Murata M, Takita H, Mizuno M, Inoue M, et al.Two distinctive BMP-carriers induce zonal chondrogenesis andmembranous ossification, respectively; geometrical factors of matricesfor cell-differentiation. Connect Tissue Res 1995;32(1–4):219–26.

[18] Gauthier O, Bouler JM, Aguado E, Pilet P, Daculsi G. Macroporousbiphasic calcium phosphate ceramics: influence of macropore diam-eter and macroporosity percentage on bone ingrowth. Biomaterials1998;19(1–3):133–9.

[19] Kuhne JH, Bartl R, Frisch B, Hammer C, Jansson V, Zimmer M.Bone formation in coralline hydroxyapatite. Effects of pore sizestudied in rabbits. Acta Orthop Scand 1994;65(3):246–52.

[20] von Doernberg MC, von Rechenberg B, Bohner M, Grunenfelder S,van Lenthe GH, Muller R, et al. In vivo behavior of calciumphosphate scaffolds with four different pore sizes. Biomaterials2006;27(30):5186–98.

[21] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds andosteogenesis. Biomaterials 2005;26(27):5474–91.

[22] Mankani MH, Kuznetsov SA, Fowler B, Kingman A, Robey PG. Invivo bone formation by human bone marrow stromal cells: effect ofcarrier particle size and shape. Biotechnol Bioeng 2001;72(1):96–107.

[23] Bohner M, van Lenthe GH, Grunenfelder S, Hirsiger W, Evison R,Muller R. Synthesis and characterization of porous beta-tricalciumphosphate blocks. Biomaterials 2005;26(31):6099–105.

[24] Kasten P, Vogel J, Luginbuhl R, Niemeyer P, Tonak M, Lorenz H,et al. Ectopic bone formation associated with mesenchymal stem cellsin a resorbable Calcium deficient hydroxyapatite carrier. Biomaterials2005;26–29:5879–89.

[25] Kasten P, Luginbuhl R, Vogel J, Niemeyer P, Weiss S, van GriensvenM, et al. Induction of bone tissue on different biomaterials: an in vitroand a pilot in vivo study in the SCID mouse. Z Orthop Ihre Grenzgeb2004;142(4):467–75.

[26] Walker JA, Kilroy GE, Xing J, Shewale J, Sinha SK, Batzer MA.Human DNA quantitation using Alu element-based polymerasechain reaction. Anal Biochem 2003;315(1):122–8.

[27] Cooper DM, Matyas JR, Katzenberg MA, Hallgrimsson B. Com-parison of microcomputed tomographic and microradiographicmeasurements of cortical bone porosity. Calcif Tissue Int2004;74(5):437–47.

[28] Keaveny TM, Morgan EF, Niebur GL, Yeh OC. Biomechanics oftrabecular bone. Annu Rev Biomed Eng 2001;3:307–33.

[29] Kuboki Y, Takita H, Kobayashi D, Tsuruga E, Inoue M, Murata M,et al. BMP-induced osteogenesis on the surface of hydroxyapatitewith geometrically feasible and nonfeasible structures: topology ofosteogenesis. J Biomed Mater Res 1998;39(2):190–9.

[30] Hulbert SF, Young FA, Mathews RS, Klawitter JJ, Talbert CD,Stelling FH. Potential of ceramic materials as permanently implant-able skeletal prostheses. J Biomed Mater Res 1970;4(3):433–56.

[31] Itala AI, Ylanen HO, Ekholm C, Karlsson KH, Aro HT. Porediameter of more than 100 micron is not requisite for bone ingrowthin rabbits. J Biomed Mater Res 2001;58(6):679–83.

[32] Lu JX, Flautre B, Anselme K, Hardouin P, Gallur A, Descamps M,et al. Role of interconnections in porous bioceramics on bonerecolonization in vitro and in vivo. J Mater Sci Mater Med1999;10(2):111–20.

[33] Bohner M, Baumgart F. Theoretical model to determine the effects ofgeometrical factors on the resorption of calcium phosphate bonesubstitutes. Biomaterials 2004;25(17):3569–82.

[34] Tsuruga E, Takita H, Itoh H, Wakisaka Y, Kuboki Y. Pore size ofporous hydroxyapatite as the cell-substratum controls BMP-inducedosteogenesis. J Biochem (Tokyo) 1997;121(2):317–24.

[35] Kujala S, Ryhanen J, Danilov A, Tuukkanen J. Effect of porosity onthe osteointegration and bone ingrowth of a weight-bearing nickel–titanium bone graft substitute. Biomaterials 2003;24(25):4691–7.

[36] Fisher JP, Vehof JW, Dean D, van der Waerden JP, Holland TA,Mikos AG, et al. Soft and hard tissue response to photocrosslinkedpoly(propylene fumarate) scaffolds in a rabbit model. J Biomed MaterRes 2002;59(3):547–56.

[37] Lewandrowski KU, Gresser JD, Bondre S, Silva AE, Wise DL,Trantolo DJ. Developing porosity of poly(propylene glycol-co-fumaric acid) bone graft substitutes and the effect on osteointegration:a preliminary histology study in rats. J Biomater Sci Polym Ed2000;11(8):879–89.

[38] Chu TM, Orton DG, Hollister SJ, Feinberg SE, Halloran JW.Mechanical and in vivo performance of hydroxyapatite implants withcontrolled architectures. Biomaterials 2002;23(5):1283–93.

[39] Kruyt MC, de Bruijn JD, Wilson CE, Oner FC, van Blitterswijk CA,Verbout AJ, et al. Viable osteogenic cells are obligatory for tissue-engineered ectopic bone formation in goats. Tissue Eng 2003;9(2):327–36.

[40] Roy TD, Simon JL, Ricci JL, Rekow ED, Thompson VP, ParsonsJR. Performance of degradable composite bone repair products madevia three-dimensional fabrication techniques. J Biomed Mater Res A2003;66(2):283–91.

[41] Barralet JE, Grover L, Gaunt T, Wright AJ, Gibson IR. Preparationof macroporous calcium phosphate cement tissue engineering scaf-fold. Biomaterials 2002;23(15):3063–72.

[42] Zhang Y, Zhang M. Three-dimensional macroporous calcium phos-phate bioceramics with nested chitosan sponges for load-bearing boneimplants. J Biomed Mater Res 2002;61(1):1–8.

[43] Quarto R, Thomas D, Liang CT. Bone progenitor cell deficits and theage-associated decline in bone repair capacity. Calcif Tissue Int1995;56(2):123–9.

[44] Majors AK, Boehm CA, Nitto H, Midura RJ, Muschler GF.Characterization of human bone marrow stromal cells with respect toosteoblastic differentiation. J Orthop Res 1997;15(4):546–57.

[45] D’Ippolito G, Schiller PC, Ricordi C, Roos BA, Howard GA. Age-related osteogenic potential of mesenchymal stromal stem cells fromhuman vertebral bone marrow. J Bone Miner Res 1999;14(7):1115–22.

[46] Murphy JM, Dixon K, Beck S, Fabian D, Feldman A, Barry F.Reduced chondrogenic and adipogenic activity of mesenchymal stemcells from patients with advanced osteoarthritis. Arthritis Rheum2002;46(3):704–13.

[47] Phinney DG, Kopen G, Righter W, Webster S, Tremain N, ProckopDJ. Donor variation in the growth properties and osteogenicpotential of human marrow stromal cells. J Cell Biochem1999;75(3):424–36.

[48] Oreffo RO, Bennett A, Carr AJ, Triffitt JT. Patients with primaryosteoarthritis show no change with ageing in the number ofosteogenic precursors. Scand J Rheumatol 1998;27(6):415–24.

[49] Kasten P, Vogel J, Luginbuhl R, Niemeyer P, Weiss S, Schneider S,et al. Influence of platelet-rich plasma on osteogenic differentiation of


mesenchymal stem cells and ectopic bone formation in calciumphosphate ceramics. Cells Tissues Organs 2006;183(2):68–79.

[50] Leskela HV, Risteli J, Niskanen S, Koivunen J, Ivaska KK,Lehenkari P. Osteoblast recruitment from stem cells does not decreaseby age at late adulthood. Biochem Biophys Res Commun2003;311(4):1008–13.

[51] Roura S, Farre J, Soler-Botija C, Llach A, Hove-Madsen L, Cairo JJ,et al. Effect of aging on the pluripotential capacity of humanCD105(+) mesenchymal stem cells. Eur J Heart Fail 2006;8(6):555–63.

[52] Muraglia A, Martin I, Cancedda R, Quarto R. A nude mousemodel for human bone formation in unloaded conditions. Bone1998;22(5 Suppl):131S–4S.

[53] Mygind T, Stiehler M, Baatrup A, Li H, Zou X, Flyvbjerg A, et al.Mesenchymal stem cell ingrowth and differentiation on corallinehydroxyapatite scaffolds. Biomaterials 2007;28(6):1036–47.

[54] Coelho MJ, Fernandes MH. Human bone cell cultures in biocom-patibility testing. Part II: effect of ascorbic acid, beta-glycerophos-phate and dexamethasone on osteoblastic differentiation.Biomaterials 2000;21(11):1095–102.

[55] Plant A, Tobias JH. Characterisation of the temporal sequence ofosteoblast gene expression during estrogen-induced osteogenesis infemale mice. J Cell Biochem 2001;82(4):683–91.

[56] Winter A, Breit S, Parsch D, Benz K, Steck E, Hauner H, et al.Cartilage-like gene expression in differentiated human stem cellspheroids: a comparison of bone marrow-derived and adipose tissue-derived stromal cells. Arthritis Rheum 2003;48(2):418–29.

[57] Laurencin CT, Attawia MA, Lu LQ, Borden MD, Lu HH, GorumWJ, et al. Poly(lactide-co-glycolide)/hydroxyapatite delivery of BMP-2-producing cells: a regional gene therapy approach to boneregeneration. Biomaterials 2001;22(11):1271–7.

[58] Itoh M, Shimazu A, Hirata I, Yoshida Y, Shintani H, Okazaki M.Characterization of CO3Ap-collagen sponges using X-ray high-resolution microtomography. Biomaterials 2004;25(13):2577–83.

[59] Akin FA, Zreiqat H, Jordan S, Wijesundara MB, Hanley L.Preparation and analysis of macroporous TiO2 films on Ti surfacesfor bone-tissue implants. J Biomed Mater Res 2001;57(4):588–96.

[60] Takahashi Y, Tabata Y. Effect of the fiber diameter and porosity ofnon-woven PET fabrics on the osteogenic differentiation of mesen-chymal stem cells. J Biomater Sci Polym Ed 2004;15(1):41–57.

Porosity and Pore Size of B-tricalcium Phosphate Scaffold Can

Documents

Transcript of Porosity and Pore Size of B-tricalcium Phosphate Scaffold Can