Available online at www.sciencedirect.

com

Acta Biomaterialia 4 (2008) 19041915 www.elsevier.com/locate/actabiomat

Porosity and pore size of b-tricalcium phosphate scaold can inuence protein production and osteogenic dierentiation of human mesenchymal stem cells: An in vitro and in vivo study
Philip Kasten a,*, Ingo Beyen a, Philipp Niemeyer b, Reto Luginbuhl c, c a Marc Bohner , Wiltrud Richter
a

Orthopaedic University Hospital Heidelberg, Schlierbacher Landstr. 200a, 69118 Heidelberg, Germany b 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 2008 Available 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 the inuence of the porosity (25%, 65% and 75%) of b-tricalcium phosphate (TCP) ceramics on osteogenic dierentiation of mesenchymal stem cells (MSC) in vitro and in vivo. For the in vitro portion of the study, TCP scaolds loaded with MSC were kept in osteogenic induction medium for 21 days. For the in vivo portion of the study, scaolds loaded with undierentiated MSC were implanted subcutaneously into SCID mice for 8 weeks and compared with similarly implanted controls that were not loaded with MSC. Measurements of total protein as well as specic alkaline phosphatase (ALP) activity were taken as indicators of growth/matrix production and osteogenic dierentiation. An increase in the total protein concentration was noted from day 1 to day 21 on the in vitro TCP 65% and TCP 75% scaolds (p < 0.05) with no such increase noted in the TCP 25% specimens. However, the specic alkaline phosphatase activity increased from 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 activity of cell-loaded TCP 65% ceramics was higher when compared with both the TCP 25% and TCP 75% specimens (p < 0.046), and higher in the TCP 75% than TCP 25% specimens (p = 0.008). Histology revealed mineralization by human cells in the pores of the TCP ceramic scaolds with a trend toward greater calcication in TCP 65% and 75%. In summary, a higher porosity of TCP scaolds does not necessarily mean a higher ALP activity in vivo. The distribution and size of the pores, as well as the surface structure, might play an important role for osteogenic dierentiation 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 rapidly emerging eld in the treatment of bone defects [15]. Mesenchymal stem cells (MSC) can easily be expanded to high cell numbers [6] and addition of MSC facilitates the healing of bone defects [7]. Cell suspensions are dicult to apply as they hardly remain in bony defects and do
*

Corresponding author. Tel.: +49 6221 96 5; fax: +49 6221 96 6347. E-mail address: Philip.Kasten@ok.uni-heidelberg.de (P. Kasten).

not provide any biomechanical stability. Therefore, MSC are combined with biomaterials in a tissue engineering approach [811]. Calcium phosphate ceramics are used clinically because they combine good stability with porosity and interconnectivity, and they are non-toxic during the dissolution and degradation process [1113]. Moreover, they allow the adhesion and growth of MSC and osteoblasts [14]. Among calcium phosphate ceramics, b-tricalcium phosphate (TCP), which dissolves in the presence of acids released by cells such as osteoclasts or macrophages, is distinguished from hydroxyapatite (HA), which is hardly

1742-7061/$ - see front matter 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2008.05.017

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

1905

degradable at all [15]. In general, it seems favourable to have cell-mediated degradation that proceeds at the same speed as new bone forms, as this allows the formation of bone with homogeneous elasticity and reduced fracture risk [16]. Critical factors that may determine the success of 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 dierentiation.

Table 1A Physical parameters of the biomaterials Material Total porosity (%) Pore diameter 600 200 lm (%) 0 200 50 lm (%) 2 50 5 lm (%) 5 <5 lm (%) 18

It is well described in previous studies that the porosity of the biomaterial plays a signicant role in the success of an MSC/biomaterial construct [1719]. Although several solid TCP ceramics are on the market, there is sparse data indicating what degree of porosity is most favourable for adhesion, proliferation and osteogenic dierentiation of MSC in vitro and in vivo [20,21]. To this end, we examined three dierent TCP block materials that are often used clinically: (i) Cerasorb, (ii) Cerasorb M (both Curasan, Kleinostheim, Germany) and (iii) a TCP produced by the Dr. Robert Mathys Foundation (RMS, Bettlach, Switzerland) that is similar to chronOSTM (Synthes). Cerasorb granules served as the positive in vivo control because of their well-known osteoconductive properties. The porosity between the tested biomaterials diered from 25% to 75% and pore size ranged from <10 to 600 lm. TCP Cerasorb block material with a low porosity of 25% was included into the study as a negative control since it is known that small pore size is not favourable for bone formation in vivo [22]. Nevertheless, orthopaedic surgeons like to use this material because of its strong mechanical properties (e.g. in open wedge osteotomy of the acetabulum or the tibia). We examined adhesion, protein production and osteogenic dierentiation of MSC on these ceramics for 3 weeks in vitro and osteogenic dierentiation of freshly loaded undierentiated MSC composites in vivo for 8 weeks.

TCP 25 block forms (Cerasorb, Curasan)a TCP 25 granules, 1000 2000 lm(Cerasorb, Curasan)b TCP 65 block forms(Cerasorb M, Curasan) TCP 75 block forms(RMS)c

$25 5

$59.3 2.5

20

10

20

$65 5

15

20

25

$75 0.4

54

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 0.52.0 mm to reach an overall porosity of 6580%. b The overall porosity of granulates is calculated by the intergranular porosity (measured by Hg porosimetry) and intergranular voids. c Structure similar to that of chronOSTM.

Table 1B Physical parameters of the biomaterials Material Specic surface area (m2 g1) 0.12 0.18 Surface/ apparent volume (m2 cc1) 0.28 0.2 Surface/ scaold (m2) 0.0076 0.0052

TCP 25 block forms (Cerasorb, Curasan)a TCP 65 block forms (Cerasorb M, Curasan) TCP 75 block forms (RMS)c

0.3

0.23

0.0068

2. Material and methods 2.1. TCP scaolds Three dierent porous TCP block forms were used: Cerasorb (TCP 25), Cerasorb M (TCP 65) and the TCP by RMS (TCP 75). In comparing these constructs, there are signicant dierences regarding micro- and macroporosity, while the specic surface area (SSA, in m2 g1) is equally low in all of them (Tables 1A and 1B). In the in vivo assay, Cerasorb granules with particles ranging from 1000 to 2000 lm and micropores of <5 lm were used as a positive control. To illustrate the surface that is accessible to the cells, we calculated the surface per apparent volume (m2 cc1 = SSA density (1-porosity)) and the total surface per scaold (m2 =

weight of scaold SSA). Phase purity was high, at more than 99% in all scaolds. The TCP for the Cerasorb products was produced by a solid-state reaction from calcium carbonate and calcium hydrogen phosphate. Microporous Cerasorb block forms were produced by milling after cold isostatic pressing and a nal sintering step. For producing Cerasorb M block forms with additional meso- and macropores, an organic porogen substance was added for sintering before pressing. The porogen substance burns away, leaving pores behind. The TCP by RMS was synthesized in an emulsion process and subsequent sintering [15]. All ceramic bodies were machined in cuboids (5 3 2 mm) with a volume of 30 mm3. The Cerasorb granule samples had the same weight as the block forms. 2.2. Structural and chemical analysis of the ceramics The distribution of the pore sizes was analysed by mercury intrusion porosimetry (Pascal 140-240/440, Porotec,

1906

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

Hofheim, Germany). In addition, the scaolds were observed by scanning electron microscopy (SEM; Cambridge S360) to determine the microporosity according to a protocol as published previously [23]. Microcomputerized tomography (l-CT) analysis was performed with a lCT80 Micro-CT instrument (Scanco Medical, Switzerland) using a high-resolution protocol (slice distance 10 lm; in-plane pixel size 10 lm) to evaluate the porosity of the scaolds. The microfocus of the X-ray source of the l-CT system had a spot size of 5 lm and a maximum voltage of 70 kVp. Approximately 100 slices mm1 were acquired and three-dimensional (3-D) volumes reconstructed. The reconstructed 3-D images were analysed with 3-D image analysis software (VGStudio Max 1.2.1, Volume Graphics, Heidelberg, Germany; and IPL (Image Processing Language), Scanco Medical, Switzerland). The software programs allow for calculating the trabecular thickness and trabecular separation, which correspond in our experiment to the wall of the pores and the pore diameter. 2.3. Isolation and cultivation of human mesenchymal stem cells For the in vitro experiments, samples of human MSC (n = 5) were obtained from iliac bone marrow aspirates. This procedure was performed before autogenous bone grafting was extracted from the iliac crest. Cells were also recovered from the femoral medullary canal after femoral osteotomy 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 for osteoarthritis of the hip. For the in vivo experiments, we pooled the MSC of three donors (one woman, three men, aged 1124 years) from iliac crest bone marrow aspirates. We pooled the MSC to level out known inter-individual dierences in proliferation and osteogenic dierentiation [6,24]. The latter patients were operated for PerthesLeggCalvee disease and hip dysplasia (2), and the mean age was 15.6 7.6. All donors were haematologically healthy Caucasians. Co-morbidities aecting bone metabolism were not present. All procedures were approved by the institutional ethics committee and all donors provided informed consent. MSC isolation and expansion were performed as reported previously [6,24]. 2.4. Loading of cells and 3-D culture of biocomposites All ceramics were incubated for 12 h in 30 lg ml1 bronectin/PBS-solution (Calbiochem/Merk, Darmstadt) and dried for 10 min under a sterile ow at room temperature. For the in vitro experiments, the samples were then statically loaded with 2 105 MSC in 10 ll of culture medium. The scaolds for the in vitro experiments were cultured for

up to 3 weeks in an osteogenic induction medium containing Dulbeccos modied Eagles medium (DMEM; Gibco/ Invitrogen, Karlsruhe, Germany), with the addition of 10% foetal calf serum, 0.1 lM dexamethasone (Sigma, Taufkirchen, Germany), 0.17 mM ascorbic acid-2-phosphate (Sigma), 10 mM b-glycerolphosphate (Sigma) and 100 IE ml1 penicillin and 100 lg ml1 streptomycin (Biochrom, Berlin, Germany). Samples were harvested for evaluation on days 1, 7 and 21. All biomaterials were tested in vitro with the MSC of ve dierent patients. For the in vivo experiments 1 106 undierentiated MSC in 15 ll of culture medium were applied to the scaffolds. Ceramic scaolds without MSC that served as controls were treated identically, but did not receive MSC. 2.5. Cell-loading ecacy and metabolic activity on day 1 To assess the cell-loading ecacy, the scaold was removed from the well in which the loading of the scaold took place. Two approaches were employed: in the rst, we looked at the number of metabolically active cells on the biomaterial; in the second, the cells in the well were measured in the WST-1 test. Cell-loading ecacy was dened as the ratio of metabolic cells on the scaold divided by the number of cells on the scaold and in the well. The cell vitality on day 1 was calculated by dividing the metabolic activity of cells on the scaold according to the WST-1 test divided by the metabolic activity of cells that were applied, i.e. 2 105. The reagent WST-1 (Roche Diagnostics GmbH, Mannheim, Germany) was diluted in a ratio of 1:10 with DMEM. The osteogenic induction medium was removed from the wells and the matrices were transferred into a 96-well plate; then 250 ll of diluted WST-1 reagent was put either into each well where the cell loading was done or added to the matrices or to a monolayer of 2 105 cells. During a 30 min incubation period at 37C, WST1 was converted into a coloured formazan salt by the metabolic activity of the living cells. To quantify the amount of formazan, 100 ll of each sample was put into a separate 96-well plate and measured using an ELISA platereader at 450 nm. A standard curve was prepared in a monolayer with dierent cell numbers in the same manner. To account for the artefact produced by the absorption of the formazan salt onto the empty scaolds, a correction factor was calculated in triplicate for each biomaterial. For that purpose, 2 105 cells in a monolayer were incubated with the cell proliferation reagent WST-1. Then the soluble formazan salt was measured in the ELISA plate reader as described above. Afterwards, each empty scaold was incubated in the formazan-containing supernatant for 60 min at 37C. A 100 ll quantity of each sample was measured in a separate 96-well plate using an ELISA plate-reader at 450 nm. By dividing the initial optical density (OD) by the OD after the incubation with the empty scaolds, a correction factor was obtained for each biomaterial: 18.59%

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

1907

for Cerasorb, 17.56% for Cerasorb M (both Curasan, Kleinostheim, Germany) and 17.74% for TCP by RMS. 2.6. Protein production and osteogenic dierentiation The protein production was assessed by measuring the amount of protein on days 1, 7 and 21 using the Micro BCATM Protein Assay Kit (Fa.Pierce, Rockford, IL, USA), as described previously [24]. Osteogenic dierentiation was assessed by measuring alkaline phosphatase (ALP) activity. The matrices were incubated with 1 ml 0.1% Triton X-100 (SigmaAldrich Chemie GmbH, Taufkirchen, Germany) and homogenized mechanically to disintegrate the cells using a Polytron (dispersing and mixing technology, Model PT2000, Kinematica Inc., Cincinnati, OH, USA). The cell lysate was centrifuged (460g for 10 min), the supernatant was recovered and the ALP enzyme activity was determined according to the following protocol: a 50 ll sample extract was incubated with 100 ll of p-nitrophenyl-phosphate (100 mg capsule, Sigma 104 Phosphatase Substrate, SigmaAldrich Chemie GmbH, Taufkirchen, Germany) and 50 ll of distilled water. The conversion to p-nitrophenol was measured after 30 min at 405/490 nm. In detail, one capsule was diluted with 12.67 ml of an alkaline buer (150 mM NaHCO3, 1 mM MgCl2; pH 10.3) to achieve a 30 mM p-nitrophenyl phosphate solution. A standard curve was prepared with a p-nitrophenol standard solution (10 lmol ml1; Sigma Aldrich Chemie GmbH, Taufkirchen, Germany). For this, a 1:2 dilution series ranging from 25 to 0.195 lg ml1 p-nitrophenol in 0.01% Triton X-100 was prepared and measured as described above. The amount of ALP was divided by the protein value to obtain the specic amount of 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. After loading the biomaterials with 1 106 undierentiated cells as described above two cell-constructs were subcutaneously implanted into adult male SCID mice (Charles River Deutschland GmbH, Sulzfeld, Germany; C.B. 17-scid/leRCrl). Two additional empty scaolds were always inserted as controls. The total of implanted scaolds per mouse consequently was four with one exception: in the mice which received Cerasorb, not only Cerasorb granules as well as block forms were implanted as a positive control [22]. A 100 ll quantity of brin glue (Baxter Deutschland GmbH, Unterschleiheim, Germany) was added to the granules to form a clot and prevent their disassembly. With the animals under general intraperitoneal anesthesia and after disinfection of the backs of the mice, two subcutaneous pockets were created bluntly through two separate 1 cm incisions on the back. After implantation of the scaffolds, the wounds were closed with single interrupted sutures. The SCID mice, weighing about 30 g, were housed

in Macrolon cages size MII (Fa.Tecniplast Deutschland GmbH, Hohenpeienberg, Germany) and fed a special murine diet (V1185-300 ssni M-Z, Fa. ssni Spezialdiaten GmbH, Soest, Germany). To account for the variability caused by the SCID mice, eight mice were used in each group, in accordance with a pilot study and a biometric analysis [24,25]. All but one from the Cerasorb group of 24 animals survived until the end of the experiment. At 8 weeks the animals were sacriced and the scaolds harvested. One cell/biomaterial construct and one empty control were used for biochemical evaluation and a second was used for histological evaluation. 2.8. Histology For light microscopy, samples were xed and embedded in paran wax (Leica, Bensheim, Germany) as described previously [24]. Sections were stained with toluidine blue (Waldeck GmbH & Co. Division Chroma, Munster, Ger many). Three sequential sections 5 lm thick at a dened depth of 0.75 mm were selected for evaluation. The blinded samples were examined at a magnication of 100 by two independent observers to identify the type of tissue (bonelike, cartilage-like or connective tissue). Samples with tissues other than connective tissue were stained with alizarin red (Waldeck GmbH) for calcium deposits. The mineralizations were graded as follows: 0 = no mineralization by the cells in the region of interest (ROI); + = mineralization around <3 cells in ROI; ++ = mineralization around >3 cells in the ROI. The ROIs were located in the area of visible calcications at the edges of the biomaterial. A digoxigenin-labeled probe specic for the human Alu repetitive DNA element was prepared by polymerase chain reaction as described previously [24]. The following primers were used: forward 50 -CGAGGCGGGTGGATCATGA GGT-30 , reverse 50 -TTTTTTGAGACGGAGTCTCGC-30 [26]. Signals were detected by immunohistochemistry using anti-digoxigenin alkaline phosphatase-conjugated Fab fragments (Roche) and NBT/BCIP (Roche) as substrates. 2.9. Statistical analysis Data analysis was performed with SPSS for Windows 15.0 (SPSS Inc., Chicago, IL, USA). Descriptive statistics and non-parametric tests were employed because of small sample sizes and distributional characteristics. Dierences between subgroups were analysed with MannWhitney U-tests. The signicance level was assigned at 5% (twotailed) due to the explorative nature of the study. 3. Results 3.1. Structural and chemical analysis of the TCP scaolds A wider range of pore diameters was observed by mercury intrusion porosimetry in the TCP 75 as compared to the other two ceramics (Fig. 1). The latter two had most

1908

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

Pore volume fraction [%]

100% 80% 60% 40% 20% 0% 0 20 40 60 80 100 120

-TCP 75

-TCP 65 -TCP 25

Pore diameter [microns]

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

were mechanically stable throughout the measurement or the macropores were interconnected by only very small pores (in the order of 67 lm). The SEM pictures (Fig. 2) support the latter explanation. The microstructures of the TCP 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 macropore diameter than TCP 65, while their pore wall thicknesses were similar (Table 2 and Fig. 3). The porosity was measured 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 provide conclusive data because the pores were in the same size range as the voxel, i.e. too small to be detected properly. 3.2. Loading ecacy and cell vitality on day 1 Loading ecacy ranged from 71% for TCP 25 (SD 20.8%) and TCP 65 (SD 26.5%) to 80% (SD 22.5%) for the TCP 75 form blocks. There was no signicant dierence regarding loading ecacy among the various TCP ceram-

of their porosity with a pore diameter below 10 lm. The steps along the curve of the TCP 75 sample can be explained either by breaking of the walls of isolated pores or by lling small interconnections between the pores. The absence of such steps in the curve of the two other ceramics, 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 at 100, 500 and 2000 magnication, showing the microporosity of the scaolds. The surface seems to be smooth in the 2000 magnication of the TCP 25 and TCP 65 scaolds, whereas the surface of the TCP 75 samples is more irregular in shape.

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

1909

ics. The mean values of metabolic activity per 2 105 cells on day 1 were higher in the groups with higher porosity, but this trend did not reach signicance due to high standard deviations. In detail, metabolic activity on day 1 was 66.3 43.6% on TCP 25, 99.3 59.2% on TCP 65 and 89.7 48.4% on TCP 75. 3.3. Protein production in vitro and in vivo

TCP 25 (Fig. 4). Consistent with the higher porosity, there were higher protein contents in the TCP 75 and TCP 65 groups than in TCP 25 in vivo (both p = 0.001). There was no statistically signicant dierence between the TCP 65 and TCP 75 groups in vivo (p = 0.059). The loading with undierentiated cells did not increase the total amount of protein in vivo compared with empty scaolds. 3.4. Osteogenic dierentiation in vitro and in vivo

Protein content increased signicantly in vitro from day 1 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 2 l-CT analysis of ceramics l-CT analysis Porosity (%) Macropore diameter (lm) Macropore wall thickness (lm)
a a

TCP 65 (n = 3) 29.6 1.47 41.03 2 97.56 2.45

TCP 75 (n = 3) 56.22 1.88 135.6 6.49 105.53 4.33

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

The specic ALP activity in vitro increased in all cell/ biomaterial groups over 21 days (p < 0.02). However, there were no signicant dierences between the groups (Fig. 5). These ndings were in contrast to the in vivo ALP activity, which was highest in the TCP 65 + cells group, followed by the TCP 75 + cells group (p = 0.046), while the TCP 25 + cells group presented values close to those of the control group (Fig. 6). There was higher ALP activity in the TCP 65 + cells groups and TCP 75 + cells groups (p< 0.009) but not the TCP 25 + cells groups compared with the empty scaolds (Fig. 6).

Fig. 3. 3-D reconstructions of the l-CT slices illustrating the distinct porosity, pore size and distribution of the scaolds.

Fig. 4. Proliferation and matrix deposition in vitro assessed by measuring the protein content over 21 days. Protein content increased signicantly in vitro from day 1 to day 21 on TCP 65 and TCP 75 and in the monolayer, but not on TCP 25. Mean values and SD are shown.

1910

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

The histological sections revealed that a penetration of cells into the inner parts of the scaold occurred in the biomaterials with higher porosity (TCP 65 and 75) (Fig. 7). Semi-quantitative grading of histologic sections revealed more mineralization in TCP 65 and TCP 75 cell-loaded

explants than in the corresponding TCP 25 explants (Table 3, Fig. 8A and B). The in situ hybridization conrmed that human cells were present in the block forms after 8 weeks (Fig. 8C). The positive control with the granules showed the strongest mineralization, as expected (Fig. 8D and E).

Specific ALP Activity in vitro

18 p=0.016 mean value (n=5) 16 p=0.016 14 p=0.008

[p-nitrophenol/ protein]

12

10

0 day1 day7 day21 day1 -TCP 25 + cells day7 day21 day1 day7 day21 day1 day7 day21 day1 -TCP 65 + cells -TCP 65 day7 day21 day1 day7 day21 -TCP 75 -TCP 25 -TCP 75 +cells

Fig. 5. ALP activity in vitro normalized to the protein content. The specic ALP activity in vitro increased in all cell/biomaterial groups over 21 days without signicant dierences. Mean values and SD are shown.

Specific ALP Activity in vivo

0.018
mean value (n=8)

p=0.008 0.016 p=0.004 0.014 p=0.046

0.012 [p-nitrophenol/ protein] p=0.009 0.010

0.008 p=0.001 0.006

0.004

0.002

0.000 Cells Controls Cells Controls Cells Controls

-TCP 25

-TCP 65

-TCP 75

Fig. 6. TCP ceramics with the higher porosity yielding higher ALP activity than the one with the lower porosity in vivo. Mean values and SD are shown.

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

1911

Fig. 7. Toluidine blue staining at 10 magnication of TCP 25 (A), TCP 65 (B) and TCP 75 (C) in vivo specimens, showing that there were fewer cells within the TCP 25 scaold.

Table 3 Mineralization after 8 weeks TCP 25 block forms 0 Empty With cells 7/7 7/7 + 0/7 0/7 ++ 0/7 0/7 TCP 65 block forms 0 6/7 2/7 + 1/7 5/7 ++ 0/7 0/7 TCP 75 block forms 0 6/8 2/8 + 2/8 6/8 ++ 0/8 0/8

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

4. Discussion Our data demonstrate that porosity and pore size of distinct TCP scaolds inuence not only protein production in vitro and in vivo but also specic ALP activity, which is an important marker for osteogenic dierentiation. Scaolds for osteogenesis should mimic bone morphology, structure and function in order to optimize integration into surrounding tissue [21]. Generally, bone has a very complex structure. The cortical bone has a compact structure with only 312% porosity. It contains a series of voids, e.g. haversian canals with a diameter of 100200 lm [27]. Trabecular bone has a porosity in the range of 5090% and pore diameters close to 1 mm [28]. The complexity of architecture and the variability of properties of bone tissue (e.g. porosity, pore size, mechanical properties, mineralization or mineral density, cell type and cytokines gradient features), as well as dierences in age, nutritional state, activity and disease status of individuals, establish a major challenge in fabricating scaolds that meet the needs of specic repair sites in specic patients [21]. A key issue to compensate for the complexity of bone tissue is to achieve a fast replacement of the bone substitute with new mature bone. We therefore chose for our experiments the biodegradable ceramic TCP, which is often used to ll bone defects in orthopedic surgery. 4.1. Eect of pore size and porosity in vivo A parameter that is independent of the material is porosity, or the percentage of void space in a solid. Pores

are necessary for bone tissue formation because they allow for migration and proliferation of osteoblasts and mesenchymal cells, as well as vascularization [29]. The minimum pore size required to regenerate mineralized bone is generally considered to be 50100 lm [17,30,31]. Smaller pores result in ingrowth of unmineralized osteoid tissue, and even smaller ones are penetrated only by brous tissue [30]. A recent study examining dierent 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 weeks revealed that the biological response to implantation was marginally inuenced by the pore size [20]. There was a balance of factors; however, faster ceramic resorption and slower bone formation were found to occur in 510 lm pore samples compared to 150 and 260 lm pore samples. Pore diameter is not the only important parameter in an experimental setting; other parameters include the animal species, block size, implant chemistry, implant topography, resorption and interconnectivity. It was hypothesized that in resorbable materials, pore density and interconnection density are more important than pore size, contrary to unresorbable materials, in which the sizes and densities are equally important [32]. Also, according to a study by Bohner and Baumgart [33], bone ingrowth should not be aected by the pore diameter as long as the structure is fully interconnected and the pore interconnections have a diameter larger than 50 lm. In a non-fully interconnected scaold, bone ingrowth should generally be faster when larger macropores are present [29,34]. Although a few studies show no eect of porosity on the amount of apposited bone [35,36], higher porosity is usually associated with greater bone formation [3740]. Although increased porosity and pore size facilitate bone ingrowth, the result is a reduction in mechanical properties, since the structural integrity of the scaold is compromised [41,42]. Our hypothesis was that a higher level of porosity would increase osteogenic dierentiation in vivo. However, in our study TCP 65 yielded a higher ALP activity than TCP 75. This dierence must be related to physical, morphological or chemical aspects. In terms of chemistry, both TCP 65 and TCP 75 consisted of more than 99% TCP. Looking at the architecture of the pores, several dierences can be

1912

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

Fig. 8. (A) TCP 65 in vivo specimen displaying an intense penetration by cells but no mineralization in the alizarin red/fast green staining (20 magnication). (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 tissue around the human cells (both 100 magnication). (D) The positive control of the in vivo TCP 25 granules showing mineralization in the 20 and (E) the 100 magnication.

noticed. The macropores of TCP 65 were smaller and more irregular in shape, had on average a smaller diameter and had fewer interconnections than those of TCP 75, as observed in l-CT and histology (Figs. 3 and 7). Furthermore, the microporosity diered qualitatively between TCP 65 and TCP 75: the surface structure of TCP 65 seemed to be smoother than that of TCP 75 (Fig. 2). This is also reected by the slightly lower specic surface area, surface per volume and surface per scaold (Table 1B). In our opinion, a smoother surface at the 2000 magnication level seems to be more favourable for cell dierentiation in vivo. All our samples were resorbable and had a rather low interconnectivity, but had strong dierences in pore size and total porosity: TCP 25 mainly had pores up to 2 10 lm and a relatively low total porosity of 25%. According to the aforementioned studies, it is unlikely that bone can form in solid blocks of TCP 25. The situation is completely dierent, however, if granules are used. In this case, there is enough space between the granules to allow for bone formation. Indeed, biomaterials in the form of granules demonstrated a higher rate of bone formation compared with solid scaolds [22]. However, many surgeons favour solid biomaterials as they provide better handling and greater mechanical stability. TCP 65 with a total porosity of 65% has pores mainly ranging from 2 to 100 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 smaller than 5 lm. Indeed, we could nd mineralization produced by human MSC within larger pores and higher specic ALP activity in the biomaterials with the larger pores of the TCP 65 and 75. However, there was no bone formation in the histological specimens of any ceramics. Donordependent heterogeneity, such as age [4345] and disease [46], may inuence the yield and proliferative capacity of stem cells or their osteogenic potential [47,48]. The lack of bone formation may be attributed to a known donor variability that cannot be completely ruled out by a preoperative selection [24,4749]. In addition, the time course of implantation 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 calcications [14,49]. One limitation of the study is that dierent donors were used for the in vitro and in vivo approaches because of limited cell numbers. This prevented any direct donor-todonor comparison of the two settings. However, within the in vivo and in vitro approaches all scaolds received the same cell populations, and consequently a comparison between the scaolds within one setting was possible. The mean age in the in vitro group (28.6 years old) was slightly higher than in the in vivo group (15.6 years old). In contrast to MSC data from rodents, where younger donors showed better osteogenic dierentiation capacity than older donors [43], no such dierences were observed for

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915

1913

human MCS [48,50,51]. This suggests that the slight dierence in donor age between the in vitro and in vivo groups is not a limitation of this study. Another issue is that in the current experimental design the authors relied on the analysis of ALP activity as the main biochemical marker for osteogenic dierentiation. However, the ALP activity can reliably be measured and compared to previous studies [34,5255]. ALP is a marker of the protein level that does not rise in uninduced MSC, but displays a strong induction during osteogenic dierentiation [56]. The time curve of ALP activity with an increase in the rst weeks seemed suitable for the experimental design [55]. To further assess osteogenic dierentiation in vivo with a separate parameter, we chose to use histology [57]. However, other osteogenic markers, such as osteonectin and osteopontin, could be included in the future studies. One limitation of the study is that the ALP activity in vivo was normalized to the total protein content. The latter might be inuenced by co-extraction of host tissue, especially in the ceramics with the higher porosities. Indeed, the total protein content was signicant higher in the scaffolds with the higher porosities that might be derived by proliferation of human MSC or host tissue. The ALP activity of the human MSC e.g. in the TCP 65, which was normalized to total protein content, was strong enough to display a signicant dierence to e.g. TCP 25 scaolds, which had a lower total protein content. Therefore, the conclusions regarding the ALP activity in vivo are valid. Another option of normalization would have been to the scaold. However, since the scaolds are brittle, during the course of the experiments the edges sometimes broke o and the volume of the scaold was thus reduced at the time of retrieval. This reduced the ALP activity per scaold and caused higher standard deviations. 4.2. The eect of porosity and pore size in vitro There are few studies about the eect of porosity on proliferation and osteogenic dierentiation in vitro. In one study, composites of apatite and collagen with pores ranging from 50 to 300 lm and porosities of 4979% exhibited no signicant dierences of MC3T3-E1 osteoblast proliferation [58]. In another study, smaller pores (0.4 and 13 lm) in TiO2 lms enhanced the proliferation of human cells trypsinized from bone in contrast to larger pores (49 lm) [59]. Human MSC that were kept in dynamic spinner ask cultivation exhibited a faster rate of osteogenic dierentiation in coralline hydroxyapatite scaolds with 200 lm pore size than with 500 lm pores, but proliferation was higher in the 500 lm scaolds [53]. In summary, it was proposed that osteogenic dierentiation in vitro was not aected by pore size, but was enhanced by a low porosity [21]. Our data cannot support this statement completely, since ALP activity in our setting was independent of the total porosity ranging from 25% to 75% and pore size. In our experiment the porosity did not aect cell attachment, in agreement

with previous studies [60]. A higher porosity in our setting increased the protein content of human MSC, since the pore space increased with porosity and facilitated transport of oxygen and nutrients. This arms one study with human MSC [53], but is in contrast to another study with osteoblast cell lines [58]. 5. Conclusion In vitro porosity was benecial for protein production, but did not inuence osteogenic dierentiation. In vivo, the higher porosities of 65% and 75% yielded higher ALP activity than the 25% porosity. Comparing the two highest porosities, the TCP 75 allowed for lower ALP activity than TCP 65. In summary, higher porosity does not necessarily mean higher ALP activity in vivo. The distribution and size of the pores, as well as the surface structure, might play an important role for osteogenic dierentiation in vivo. However, the exact mechanisms have not been determined to date. Future studies will have to identify other relevant parameters of the scaolds for osteogenic dierentiation. Acknowledgements The work was performed in the Division of Experimental Orthopaedics, Orthopaedic University Hospital Heidelberg, Heidelberg, Germany. We thank Curasan AG, Kleinostheim, Germany and Dr. Robert Mathys Foundation, Bettlach, Switzerland for their support in providing the biomaterials. Biopharm, Heidelberg, Germany and the research fund of the Orthopaedic University Hospital, Heidelberg, Germany supported the study nancially. References
[1] Vacanti CA, Bonassar LJ. An overview of tissue engineered bone. Clin Orthop 1999(367 Suppl):S37581. [2] Laurencin CT, Ambrosio AM, Borden MD, Cooper Jr JA. Tissue engineering: orthopedic applications. Annu Rev Biomed Eng 1999;1:1946. [3] Petite H, Viateau V, Bensaid W, Meunier A, de Pollak C, Bourguignon M, et al. Tissue-engineered bone regeneration. Nat Biotechnol 2000;18(9):95963. [4] Bruder SP, Fink DJ, Caplan AI. Mesenchymal stem cells in bone development, bone repair, and skeletal regeneration therapy. J Cell Biochem 1994;56(3):28394. [5] Ohgushi H, Caplan AI. Stem cell technology and bioceramics: from cell to gene engineering. J Biomed Mater Res 1999;48(6):91327. [6] Reyes M, Lund T, Lenvik T, Aguiar D, Koodie L, Verfaillie CM. Purication and ex vivo expansion of postnatal human marrow mesodermal progenitor cells. Blood 2001;98(9):261525. [7] Hernigou P, Poignard A, Beaujean F, Rouard H. Percutaneous autologous bone-marrow grafting for nonunions. Inuence of the number and concentration of progenitor cells. J Bone Joint Surg Am 2005;87(7):14307. [8] Bruder SP, Kraus KH, Goldberg VM, Kadiyala S. The eect of implants loaded with autologous mesenchymal stem cells on the healing of canine segmental bone defects. J Bone Joint Surg Am 1998;80(7):98596. [9] Kon E, Muraglia A, Corsi A, Bianco P, Marcacci M, Martin I, et al. Autologous bone marrow stromal cells loaded onto porous hydroxy-

1914

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915 [30] Hulbert SF, Young FA, Mathews RS, Klawitter JJ, Talbert CD, Stelling FH. Potential of ceramic materials as permanently implantable skeletal prostheses. J Biomed Mater Res 1970;4(3):43356. [31] Itala AI, Ylanen HO, Ekholm C, Karlsson KH, Aro HT. Pore diameter of more than 100 micron is not requisite for bone ingrowth in rabbits. J Biomed Mater Res 2001;58(6):67983. [32] Lu JX, Flautre B, Anselme K, Hardouin P, Gallur A, Descamps M, et al. Role of interconnections in porous bioceramics on bone recolonization in vitro and in vivo. J Mater Sci Mater Med 1999;10(2):11120. [33] Bohner M, Baumgart F. Theoretical model to determine the eects of geometrical factors on the resorption of calcium phosphate bone substitutes. Biomaterials 2004;25(17):356982. [34] Tsuruga E, Takita H, Itoh H, Wakisaka Y, Kuboki Y. Pore size of porous hydroxyapatite as the cell-substratum controls BMP-induced osteogenesis. J Biochem (Tokyo) 1997;121(2):31724. [35] Kujala S, Ryhanen J, Danilov A, Tuukkanen J. Eect of porosity on the osteointegration and bone ingrowth of a weight-bearing nickel titanium bone graft substitute. Biomaterials 2003;24(25):46917. [36] Fisher JP, Vehof JW, Dean D, van der Waerden JP, Holland TA, Mikos AG, et al. Soft and hard tissue response to photocrosslinked poly(propylene fumarate) scaolds in a rabbit model. J Biomed Mater Res 2002;59(3):54756. [37] Lewandrowski KU, Gresser JD, Bondre S, Silva AE, Wise DL, Trantolo DJ. Developing porosity of poly(propylene glycol-cofumaric acid) bone graft substitutes and the eect on osteointegration: a preliminary histology study in rats. J Biomater Sci Polym Ed 2000;11(8):87989. [38] Chu TM, Orton DG, Hollister SJ, Feinberg SE, Halloran JW. Mechanical and in vivo performance of hydroxyapatite implants with controlled architectures. Biomaterials 2002;23(5):128393. [39] Kruyt MC, de Bruijn JD, Wilson CE, Oner FC, van Blitterswijk CA, Verbout AJ, et al. Viable osteogenic cells are obligatory for tissueengineered ectopic bone formation in goats. Tissue Eng 2003;9(2): 32736. [40] Roy TD, Simon JL, Ricci JL, Rekow ED, Thompson VP, Parsons JR. Performance of degradable composite bone repair products made via three-dimensional fabrication techniques. J Biomed Mater Res A 2003;66(2):28391. [41] Barralet JE, Grover L, Gaunt T, Wright AJ, Gibson IR. Preparation of macroporous calcium phosphate cement tissue engineering scaffold. Biomaterials 2002;23(15):306372. [42] Zhang Y, Zhang M. Three-dimensional macroporous calcium phosphate bioceramics with nested chitosan sponges for load-bearing bone implants. J Biomed Mater Res 2002;61(1):18. [43] Quarto R, Thomas D, Liang CT. Bone progenitor cell decits and the age-associated decline in bone repair capacity. Calcif Tissue Int 1995;56(2):1239. [44] Majors AK, Boehm CA, Nitto H, Midura RJ, Muschler GF. Characterization of human bone marrow stromal cells with respect to osteoblastic dierentiation. J Orthop Res 1997;15(4):54657. [45] DIppolito G, Schiller PC, Ricordi C, Roos BA, Howard GA. Agerelated osteogenic potential of mesenchymal stromal stem cells from human vertebral bone marrow. J Bone Miner Res 1999;14(7): 111522. [46] Murphy JM, Dixon K, Beck S, Fabian D, Feldman A, Barry F. Reduced chondrogenic and adipogenic activity of mesenchymal stem cells from patients with advanced osteoarthritis. Arthritis Rheum 2002;46(3):70413. [47] Phinney DG, Kopen G, Righter W, Webster S, Tremain N, Prockop DJ. Donor variation in the growth properties and osteogenic potential of human marrow stromal cells. J Cell Biochem 1999;75(3):42436. [48] Oreo RO, Bennett A, Carr AJ, Tritt JT. Patients with primary osteoarthritis show no change with ageing in the number of osteogenic precursors. Scand J Rheumatol 1998;27(6):41524. [49] Kasten P, Vogel J, Luginbuhl R, Niemeyer P, Weiss S, Schneider S, et al. Inuence of platelet-rich plasma on osteogenic dierentiation of

[10]

[11]

[12]

[13] [14]

[15] [16]

[17]

[18]

[19]

[20]

[21] [22]

[23]

[24]

[25]

[26]

[27]

[28] [29]

apatite ceramic accelerate bone repair in critical-size defects of sheep long bones. J Biomed Mater Res 2000;49(3):32837. 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):2238. Quarto R, Mastrogiacomo M, Cancedda R, Kutepov SM, Mukhachev V, Lavroukov A, et al. Repair of large bone defects with the use of autologous bone marrow stromal cells. N Engl J Med 2001;344(5):3856. Marcacci M, Kon E, Zaagnini S, Giardino R, Rocca M, Corsi A, et al. Reconstruction of extensive long-bone defects in sheep using porous hydroxyapatite sponges. Calcif Tissue Int 1999;64(1):8390. Rueger JM. Bone substitution materials. Current status and prospects. Orthopade 1998;27(2):729. Kasten P, Luginbuhl R, van Griensven M, Barkhausen T, Krettek C, Bohner M, et al. Comparison of human bone marrow stromal cells seeded on calcium-decient hydroxyapatite, b-tricalcium phosphate and demineralized bone matrix. Biomaterials 2003;24(15): 2593603. Bohner M. Calcium orthophosphates in medicine: from ceramics to calcium phosphate cements. Injury 2000;31(Suppl 4):3747. Wiesmann HP, Joos U, Meyer U. Biological and biophysical principles in extracorporal bone tissue engineering. Part II. Int J Oral Maxillofac Surg 2004;33(6):52330. Kuboki Y, Saito T, Murata M, Takita H, Mizuno M, Inoue M, et al. Two distinctive BMP-carriers induce zonal chondrogenesis and membranous ossication, respectively; geometrical factors of matrices for cell-dierentiation. Connect Tissue Res 1995;32(14):21926. Gauthier O, Bouler JM, Aguado E, Pilet P, Daculsi G. Macroporous biphasic calcium phosphate ceramics: inuence of macropore diameter and macroporosity percentage on bone ingrowth. Biomaterials 1998;19(13):1339. Kuhne JH, Bartl R, Frisch B, Hammer C, Jansson V, Zimmer M. Bone formation in coralline hydroxyapatite. Eects of pore size studied in rabbits. Acta Orthop Scand 1994;65(3):24652. von Doernberg MC, von Rechenberg B, Bohner M, Grunenfelder S, van Lenthe GH, Muller R, et al. In vivo behavior of calcium phosphate scaolds with four dierent pore sizes. Biomaterials 2006;27(30):518698. Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaolds and osteogenesis. Biomaterials 2005;26(27):547491. Mankani MH, Kuznetsov SA, Fowler B, Kingman A, Robey PG. In vivo bone formation by human bone marrow stromal cells: eect of carrier particle size and shape. Biotechnol Bioeng 2001;72(1):96107. Bohner M, van Lenthe GH, Grunenfelder S, Hirsiger W, Evison R, Muller R. Synthesis and characterization of porous beta-tricalcium phosphate blocks. Biomaterials 2005;26(31):6099105. Kasten P, Vogel J, Luginbuhl R, Niemeyer P, Tonak M, Lorenz H, et al. Ectopic bone formation associated with mesenchymal stem cells in a resorbable Calcium decient hydroxyapatite carrier. Biomaterials 2005;2629:587989. Kasten P, Luginbuhl R, Vogel J, Niemeyer P, Weiss S, van Griensven M, et al. Induction of bone tissue on dierent biomaterials: an in vitro and a pilot in vivo study in the SCID mouse. Z Orthop Ihre Grenzgeb 2004;142(4):46775. Walker JA, Kilroy GE, Xing J, Shewale J, Sinha SK, Batzer MA. Human DNA quantitation using Alu element-based polymerase chain reaction. Anal Biochem 2003;315(1):1228. Cooper DM, Matyas JR, Katzenberg MA, Hallgrimsson B. Comparison of microcomputed tomographic and microradiographic measurements of cortical bone porosity. Calcif Tissue Int 2004;74(5):43747. Keaveny TM, Morgan EF, Niebur GL, Yeh OC. Biomechanics of trabecular bone. Annu Rev Biomed Eng 2001;3:30733. Kuboki Y, Takita H, Kobayashi D, Tsuruga E, Inoue M, Murata M, et al. BMP-induced osteogenesis on the surface of hydroxyapatite with geometrically feasible and nonfeasible structures: topology of osteogenesis. J Biomed Mater Res 1998;39(2):1909.

P. Kasten et al. / Acta Biomaterialia 4 (2008) 19041915 mesenchymal stem cells and ectopic bone formation in calcium phosphate ceramics. Cells Tissues Organs 2006;183(2):6879. Leskela HV, Risteli J, Niskanen S, Koivunen J, Ivaska KK, Lehenkari P. Osteoblast recruitment from stem cells does not decrease by age at late adulthood. Biochem Biophys Res Commun 2003;311(4):100813. Roura S, Farre J, Soler-Botija C, Llach A, Hove-Madsen L, Cairo JJ, et al. Eect of aging on the pluripotential capacity of human CD105(+) mesenchymal stem cells. Eur J Heart Fail 2006;8(6): 55563. Muraglia A, Martin I, Cancedda R, Quarto R. A nude mouse model for human bone formation in unloaded conditions. Bone 1998;22(5 Suppl):131S4S. Mygind T, Stiehler M, Baatrup A, Li H, Zou X, Flyvbjerg A, et al. Mesenchymal stem cell ingrowth and dierentiation on coralline hydroxyapatite scaolds. Biomaterials 2007;28(6):103647. Coelho MJ, Fernandes MH. Human bone cell cultures in biocompatibility testing. Part II: eect of ascorbic acid, beta-glycerophosphate and dexamethasone on osteoblastic dierentiation. Biomaterials 2000;21(11):1095102.

1915

[50]

[51]

[52]

[53]

[54]

[55] Plant A, Tobias JH. Characterisation of the temporal sequence of osteoblast gene expression during estrogen-induced osteogenesis in female mice. J Cell Biochem 2001;82(4):68391. [56] Winter A, Breit S, Parsch D, Benz K, Steck E, Hauner H, et al. Cartilage-like gene expression in dierentiated human stem cell spheroids: a comparison of bone marrow-derived and adipose tissuederived stromal cells. Arthritis Rheum 2003;48(2):41829. [57] Laurencin CT, Attawia MA, Lu LQ, Borden MD, Lu HH, Gorum WJ, et al. Poly(lactide-co-glycolide)/hydroxyapatite delivery of BMP2-producing cells: a regional gene therapy approach to bone regeneration. Biomaterials 2001;22(11):12717. [58] Itoh M, Shimazu A, Hirata I, Yoshida Y, Shintani H, Okazaki M. Characterization of CO3Ap-collagen sponges using X-ray highresolution microtomography. Biomaterials 2004;25(13):257783. [59] Akin FA, Zreiqat H, Jordan S, Wijesundara MB, Hanley L. Preparation and analysis of macroporous TiO2 lms on Ti surfaces for bone-tissue implants. J Biomed Mater Res 2001;57(4):58896. [60] Takahashi Y, Tabata Y. Eect of the ber diameter and porosity of non-woven PET fabrics on the osteogenic dierentiation of mesenchymal stem cells. J Biomater Sci Polym Ed 2004;15(1):4157.