Differentiation of mononuclear precursors into osteoclasts
`on the surface of Si-substituted hydroxyapatite
`
`C.M. Botelho,1,2 R.A. Brooks,3 G. Spence,3 I. McFarlane,4 M.A. Lopes,1,2 S.M. Best,5 J.D. Santos,1,2
`N. Rushton,3 W. Bonfield5
`1INEB - Instituto de Engenharia Biome´dica, Laborato´rio de Biomateriais, Rua do Campo Alegre, 823,
`4150 –180 Porto, Portugal
`2FEUP - Faculdade de Engenharia da Universidade do Porto, DEMM, Rua Dr. Roberto Frias,
`4200 – 465 Porto, Portugal
`3Orthopaedic Research Unit, Box 180, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, United Kingdom
`4Clinical Biochemistry, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, United Kingdom
`5Department of Materials Science and Metallurgy, University of Cambridge, Cambridge CB2 3QZ, United Kingdom
`
`Received 26 August 2005; revised 28 December 2005; accepted 10 January 2006
`Published online 31 May 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.30726
`
`In healthy bone, resorption and synthesis are in
`Abstract:
`perfect coordination. In previous studies we demonstrated
`that the incorporation of silicon into the hydroxyapatite
`(HA) lattice enhances the proliferation and differentiation of
`human osteoblasts. Therefore, the aim of this study was to
`demonstrate the effect of silicon-substituted HA (0.8 and 1.5
`wt % Si-HA) on the differentiation of mononuclear cells into
`osteoclasts, using two different starting cultures, peripheral
`blood mononuclear cells (PBMC) and monocytes expressing
`the CD14 antigen (CD14⫹). Through this study, it was pos-
`sible to demonstrate that Si-HA allows the differentiation of
`mononuclear cells into mature osteoclasts, independent of
`the starting culture, PBMC or CD14⫹. Most of the cells on
`the surface of the materials expressed osteoclastic markers:
`actin rings, several nuclei, positivity for tartrate-resistant
`
`acid phosphatase (TRAP), and vitronectin receptor. In the
`presence of osteoclasts, a higher release of calcium and
`phosphate into the medium from the 1.5 wt % Si-HA sub-
`strate was detected when compared to the HA substrate;
`therefore, these results indicate higher osteoclastic resorp-
`tive activity on the 1.5 wt % Si-HA surface. Si-HA can be
`resorbed by cellular mechanisms and have a stimulatory
`effect on osteoclasts, although the underlying mechanism is
`still poorly understood. © 2006 Wiley Periodicals, Inc.
`J Biomed Mater Res 78A: 709–720, 2006
`
`Key words: peripheral blood mononuclear cells; monocytes
`CD14 positive; differentiation; osteoclasts; silicon-substi-
`tuted hydroxyapatite
`
`INTRODUCTION
`
`The ideal bone graft for many applications would be
`a material that can both be resorbed and induce bone
`formation, and thereby being completely replaced by
`new bone. Bone is a dynamic organ because of its
`constant remodeling. This process can be divided into
`two main steps: resorption, where the “old” bone is
`resorbed by the osteoclasts, and synthesis, where the
`osteoblasts lay down new layers of bone.1
`In healthy bone, resorption and synthesis are in
`perfect coordination. Disruption to the balance be-
`
`tween resorption and formation can lead to osteopo-
`rosis, which is characterized by the loss of bone mass,2
`and osteopetrosis, resulting from a failure of oste-
`oclasts to resorb bone.3 The osteoclast is the only cell
`capable of resorbing mature bone. It is a tissue-specific
`macrophage polykaryon created by the differentiation
`and fusion of monocyte/macrophage precursor cells.4
`The resorption takes place under the ruffled border,
`where protons and proteases are secreted, leading to
`the formation of resorption lacunae.4
`In the last decade, there has been a breakthrough in
`the understanding of osteoclastogenesis, because of
`the discovery of the receptor activator of nuclear factor
`␬␤ligand (RANKL) in 1998. It was identified as an
`osteoblast-producing ligand, which promotes oste-
`oclast differentiation.4,5 RANK is a type I transmem-
`brane receptor of the TNF receptor superfamily that
`was identified in a dendritic cell cDNA library.6 Sev-
`eral studies showed that it is possible to generate
`MILLENIUM EXHIBIT 2050
`(cid:20)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`Baxter Healthcare Corp. et al. v. Millenium Biologix, LLC
`IPR2013-00582, -00590
`
`Correspondence to: R.A. Brooks; email: rb10003@cam.ac.uk
`Contract grant sponsor: Fundac¸a˜o para a Cieˆncia e Tecno-
`logia (FCT); contract grant numbers: SFRH/BD/6173 and
`POCTI/CTM/49238/2002
`
`© 2006 Wiley Periodicals, Inc.
`
`

`

`710
`
`BOTELHO ET AL.
`
`MATERIALS AND METHODS
`
`HA and Si-HA were prepared by a chemical precipitation
`route; the protocol is fully described elsewhere.16 –18 Sin-
`tered powders of the HA and Si-HA materials were pre-
`pared by compacting the as-precipitated powders and sin-
`tering it at 1300°C. The samples were sterilized by heat
`(180°C) for a period of 2 h.
`Two different osteoclast precursor cell populations were
`used: PBMC and CD14 positive monocytes (CD14⫹), as de-
`scribed later.
`
`resorbed by cellular mechanisms under physiologic
`conditions.13
`In previous studies, we demonstrated that the incor-
`poration of silicon into the HA lattice increases osteo-
`blast proliferation and differentiation,14 and Gibson et al.
`reported that the presence of this ion increases the met-
`abolic activity of osteoblast-like cells in vitro.15 Patel et al.
`showed an increase in bone apposition/ingrowth re-
`lated to the concentration of silicon incorporated into the
`HA lattice, although no osteoclast-like cells were ob-
`served on the surface or proximity of this material.13 The
`aim of this study was to assess the effect of silicon
`incorporation into HA on the differentiation of mononu-
`clear cells into mature osteoclasts, using two different
`precursor populations, PBMC and monocytes express-
`ing CD14, and to determine whether this material can be
`resorbed by osteoclasts.
`
`osteoclasts in vitro from peripheral mononuclear blood
`cells (PBMC), using RANKL and macrophage colony
`stimulating factor (MCSF), a polypeptide growth fac-
`tor.7,8 Quinn et al. suggested that osteoclasts are de-
`rived from monocytes that expressed the macrophage-
`associated antigen CD14 (CD14⫹).8 These results were
`supported by Massey et al., who showed that CD14⫹
`cells from PBMC showed enhanced osteoclastic bone
`resorption in co-cultures with UMR 106 rat osteoblast-
`like cells.9 The use of co-culture is related to the ability
`of osteoblast/stromal cells to produce the two hema-
`topoietic
`factors needed for osteoclastogenesis,
`RANKL and MCSF.4
`The use of soluble RANKL and MCSF allows the
`formation of mature osteoclasts in vitro directly from
`osteoclast precursors, eliminating the need for a co-
`culture system (PBMC and UMRS 106 rat osteoblast-
`like cells). The major advantage of this method is that
`only one cell type is used, and the final osteoclast
`culture is free of osteosarcoma cells.5 RANKL and
`MCSF induce gene expression for tartrate-resistant
`acid phosphatase (TRAP), cathepsin K, calcitonin re-
`␤
`ceptor, and ␣
`3-integrin (vitronectin receptor) that
`v
`are characteristic of mature osteoclasts.4 Therefore, in
`vitro osteoclasts can be identified by TRAP staining (a
`characteristic shared with macrophages), multinucle-
`arity, formation of an actin ring structure, a polar cell
`␤
`body during resorption, and by the presence of av
`3
`integrin.5,10 This integrin belongs to the superfamily of
`adhesion proteins, named as “vitronectin receptor,”
`because it binds to the extracellular matrix protein
`␤
`vitronectin.10 In osteoclasts, av
`3 is the predominant
`integrin, quantitatively and functionally, and it can
`mediate the recognition of several RGD-motif contain-
`ing bone-matrix proteins.10
`The PBMCs were isolated from a buffy coat, obtained
`Biomaterials are designed to fulfill a purpose; in the
`from the blood of healthy donors. Buffy coat cells were
`case of bone regeneration, they should stimulate the
`diluted in phosphate buffer solution (PBS) and transferred to
`proliferation and differentiation of osteoblast cells,
`tubes containing Histopaque-1077® (Sigma-Aldrich, Poole,
`UK), which allows the separation of erythrocytes, neutro-
`leading to the formation of “new bone.” They should
`phils, platelets, and PBMC by density, and the protocol is
`also be resorbed by a cellular mechanism to keep the
`fully described elsewhere.19 The end product is a mix cul-
`normal balance between bone formation and resorp-
`ture of monocytes, platelets, and lymphocytes.
`tion. It has been shown that it is possible to assess the
`The cells were suspended in ␣-MEM medium, with 10%
`resorbability of biomaterial by cellular mechanisms in
`(v/v) human AB serum, 1% (v/v) glutamine, and 30 ␮g/mL
`vitro.1,11,12 Schiling et al. reported that poly(methyl
`vitamin C (attachment medium). They were seeded at a
`methacrylate) cannot be resorbed by osteoclasts as no
`concentration of 2 ⫻ 106 cells/mL and allowed to attach for
`resorption pits were visualized, although these fea-
`2 h at 37°C, in a 5% CO2 atmosphere. After this period, the
`non-adherent cells were removed by washing and 250 ␮L of
`tures were present on the surface of calcium phos-
`attachment medium containing 25 ng/mL of MCSF and 30
`phate materials, indicating its resorbability by oste-
`ng/mL of RANKL was added. Experimental samples will be
`oclasts.1 Doi et al. reported that the incorporation of
`referred to as HA OC: 0.8 wt % Si-HA OC and 1.5 wt %
`carbonate ions increased osteoclast resorption in
`Si-HA OC. After 2 h of attachment, a set of samples (n ⫽ 3,
`vitro.11 Patel et al. reported similar results in vivo; in
`for each material) was stained with toluidine blue to visu-
`this case, the incorporation of carbonate ions into the
`alize the nucleus. The cells were visualized using a light
`hydroxyapatite (HA) lattice increased the number of
`microscope.
`multinucleated phagocytic cells
`resembling oste-
`The medium was changed every 3 days for a period of 21
`oclasts in close proximity to or directly on the surface
`days and analyzed for calcium (Ca2⫹) and phosphate
`3⫺) using automated colorimetric assays, calcium and
`of carbonate HA, indicating that this material was
`(PO4
`(cid:21)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Peripheral blood mononuclear cells
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`DIFFERENTIATION OF MONONUCLEAR PRECURSORS INTO OSTEOCLASTS
`
`711
`
`RESULTS
`
`Peripheral blood mononuclear cells
`
`(Dade Behring, Deerfield,
`phosphorous flex® reagents,
`USA) on a Dimension RxL Integrated Chemistry System.
`Two different controls were used: (a) the samples without
`cells were incubated only in attachment medium (referred to
`as HA, 0.8 wt % Si-HA and 1.5 wt % Si-HA) and (b) the
`samples were incubated with 2 ⫻ 106 cells/mL of PBMC, but
`only in attachment medium, without cytokines (designated
`as HA PBMC; 0.8 wt % Si-HA and 1.5 wt % Si-HA PBMC).
`Following 21 days of culture, the medium was removed,
`and the samples were washed with phosphate buffered
`saline solution, after which the cells were fixed for 5 min in
`4% paraformaldehyde and stained for TRAP.20 The sub-
`strate was naphthol AS-BI phosphate (Sigma-Aldrich, Poole,
`UK). Tartrate resistance was assessed in the presence of
`0.05M sodium tartrate.
`
`CD14ⴙ
`
`After 2 h of incubation, mononuclear cells attached
`at the surface of all materials. The cells attached to the
`surface of Si-HA seemed to form clusters (Fig. 1).
`On the HA surface, the cells were sparser than on
`Si-HA (Fig. 2). Although the difference in adherent
`cells between the materials did not reach statistical
`significance,
`these results support previous ones
`where a similar trend was observed regarding the
`adhesion of human osteoblasts; 0.8 wt % Si-HA
`promoted greater adhesion and proliferation of human
`osteoblasts compared to HA.14 Further research showed
`that 0.8 wt % Si-HA not only shows higher osteoblast
`adhesion when compared to HA but also when com-
`pared to 1.5 wt % Si-HA (data not shown).
`After a period of 21 days, the culture was terminated.
`The cells on the surface of the three materials showed
`typical features of osteoclasts, they were multinucleated
`(white arrow) and TRAP positive (Fig. 3).
`The release of calcium and phosphate into the me-
`dium was measured in three experimental conditions,
`to evaluate the resorptive ability of osteoclasts.
`The calcium release from HA to the medium was
`similar in all the three experimental conditions [Fig.
`4(a)]: HA in attachment medium, HA PBMC, and HA
`OC. Although a slight increase in the phosphate con-
`centrations was observed from the samples containing
`osteoclasts at day 19 [Fig. 4(b)], it was not statistically
`significant and, furthermore, the trend of phosphate
`release between the controls and the experimental
`samples is exactly the same. Therefore, it is not possi-
`ble to conclude that the slight increase in phosphate
`release is due to osteoclast activity.
`The behavior of the osteoclasts on the 0.8 wt %
`Si-HA surface differs from HA. In this case, at day 19,
`a slight increase in the calcium content of the medium
`with osteoclasts was detected, although it did not
`reach statistical significance [Fig. 5(a)]. While the in-
`crease in phosphate at day 19 reached statistical sig-
`nificance when compared to the values observed with
`the controls (0.8 wt % Si-HA – 1.10 mmol/L, p ⬍ 0.007;
`0.8 wt % Si-HA PBMC – 1.11 mmol/L, p ⬍ 0.045; and
`0.8 wt % Si-HA OC – 1.42 mmol/L) [Fig. 5(b)]. This
`could indicate that the osteoclasts observed on the
`surface of 0.8 wt % Si-HA are active and resorbing.
`The calcium concentration of the medium from the
`experimental sample 1.5 wt % Si-HA OC at day 19 was
`significantly higher when compared to the control
`samples [Fig. 6(a)] (1.5 wt % Si-HA – 1.88 mmol/L, p ⬍
`0.002; 1.5 wt % Si-HA PBMC – 1.89 mmol/L, p ⬍ 0.003;
`and 1.5 wt % Si-HA OC – 2.39 mmol/L). A similar
`All results were statistically evaluated by ANOVA, and
`result was observed for phosphate [Fig. 6(b)] (1.5 wt %
`posthoc testing used Bonferroni’s correction on SPSS statisti-
`Si-HA – 1.10 mmol/L, p ⬍ 0.001; 1.5 wt % Si-HA
`cal software. Significance was set at the 5% level, (p ⬍ 0.05).
`(cid:22)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`The first step to obtain a pure culture of CD14⫹ cells was
`the same as described earlier, but with the addition of a
`magnetic separation step. After centrifugation with His-
`topaque, the cells were labeled with magnetic microbeads®,
`presenting anti-CD14 antibodies and placed in a magnetic
`column (Miltenyi Biotec, Bisley, UK), where the CD14⫹ cells
`were retained by the column. Following removal of the
`magnetic source, the column was washed with a Macs Buff-
`er® (Miltenyi Biotec) and the CD14⫹ cells were collected.
`The cells were seeded at 2 ⫻ 106 cells/mL and incubated for
`a period of 2 h in an attachment medium at 37°C in a 5% CO2
`atmosphere. After this period, the non-adherent cells were
`washed off and 250 ␮L of attachment medium containing 25
`ng/mL of MCSF and 30 ng/mL of RANKL was added for a
`period of 21 days, (the experimental samples will be referred
`to as HA OC: 0.8 and 1.5 wt % Si-HA OC). Two different
`controls were used: (a) the cells were seeded at 2 ⫻ 106
`cells/mL with attachment medium but no cytokines were
`added (referred to as HA NC; 0.8 and 1.5 wt % Si-HA NC),
`(b) the cells were also seeded at 2 ⫻ 106 cells/mL with
`attachment medium plus 25 ng/mL of MCSF (samples des-
`ignated as HA MCSF; 0.8 and 1.5 wt % Si-HA MCSF).
`The medium was changed every 7 days for a period of 21
`calcium (Ca2⫹)
`days
`and analyzed for
`and phos-
`3⫺) using a colorimetric assay, as described ear-
`phate (PO4
`lier.
`After 21 days, the cells were fixed in 4% paraformalde-
`hyde/phosphate buffered saline solution, pH 7.4, at room
`temperature for 15 min. Using immunocytochemistry, the
`cells were stained with Phalloidin FITC (1 ␮g/mL), which
`binds to F-actin. The vitronectin receptor was identified
`␤
`using a primary mouse antibody against av
`3, and a second-
`ary antibody conjugated to TRITC to demonstrate vitronec-
`tin receptors (681 ␮g/mL). The nuclei were stained using
`4⬘,6-diamidino-2-phenylindole (20 ␮g/mL). The fluorescent
`stains were visualized using fluorescence microscopy (FM)
`and confocal laser scanning microscopy (CLSM).
`
`Statistical analysis
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`712
`
`BOTELHO ET AL.
`
`When comparing the samples with osteoclasts from
`the three materials at day 19, the increase in calcium
`and phosphate concentrations of the medium is asso-
`ciated with the increase in the silicon content of the
`sample. A statistical difference between HA and 1.5 wt
`% Si-HA (calcium p ⬍ 0.007; phosphate p ⬍ 0.001) was
`detected, although no statistically significant differ-
`ence was observed between 0.8 wt % Si-HA and HA or
`between the two compositions of Si-HA.
`
`CD14ⴙ
`
`A similar analysis protocol was followed with the
`CD14⫹ cultures, although in this case the osteoclast
`markers, actin rings, and vitronectin receptor were
`assessed in addition to TRAP staining and measure-
`ment of the calcium and phosphate release.
`After 21 days, most of the cells present at the surface
`the three materials were TRAP positive and
`of
`multinucleated. The localization of actin distribution
`using FITC-labeled phalloidin demonstrated the char-
`acteristic actin ring on the surface of HA and Si-HA.
`Very few cells differentiated on the control surfaces (NC
`or MCSF) or formed actin rings, as can be seen in the FM
`images (Fig. 7). On the other hand, most of the cells on
`the surface of the experimental samples presented actin
`rings, although some of them had discontinuities. They
`were positive for vitronectin (Fig. 8), and several nuclei
`were observed inside the actin ring [Fig. 8(d)].
`Similar results were obtained for Si-HA; the cells
`were multinucleated, presented actin rings, most of
`them complete, and they also presented podosomes
`rich in vitronectin. The cells were also positive for vitro-
`nectin and were multinucleated. On both compositions a
`large number of osteoclasts were present (Fig. 9).
`As mentioned earlier, osteoclasts are actively mi-
`grating cells, and they can form microfilaments called
`
`Figure 1. Distribution of adherent PBMC on (a) HA; (b) 0.8
`wt % Si-HA; and (c) 1.5 wt % Si-HA after 2 h of incubation. (i)
`The platelets and (ii) the nucleus of mononuclear cells were
`stained with toluidine blue. [Color figure can be viewed in the
`online issue, which is available at www.interscience.wiley.
`com.]
`
`PBMC – 1.11 mmol/L, p ⬍ 0.001; and 1.5 wt % Si-HA
`OC – 1.42 mmol/L). Once again this result could
`Figure 2. Number of adherent PBMC/mm2 after 2 h of
`indicate that the osteoclasts are active and resorbing.
`incubation.
`(cid:23)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`DIFFERENTIATION OF MONONUCLEAR PRECURSORS INTO OSTEOCLASTS
`
`713
`
`Figure 4. Calcium (a) and phosphate (b) concentration of
`the medium at different time points for the HA controls and
`the HA with osteoclasts (n ⫽ 3).
`
`Figure 3. Multinucleated, TRAP-positive cells on the sur-
`face of (a) HA; (b) 0.8 wt % Si-HA; and (c) 1.5 wt % Si-HA
`after 21 days of incubation. White arrow, nuclei. [Color
`figure can be viewed in the online issue, which is available
`at www.interscience.wiley.com.]
`
`Figure 5. Calcium (a) and phosphate (b) concentration of
`podosomes, which are rich in vitronectin receptor
`the medium at different time points for the 0.8 wt % Si-HA
`(␣
`␤
`3). The accumulation of these structures precedes
`controls and the 0.8 wt % Si-HA samples with osteoclasts.
`v
`(*Statistical difference p ⬍ 0.05, n ⫽ 3).
`the resorption phase. Podosomes were visible in more
`(cid:24)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`714
`
`BOTELHO ET AL.
`
`Figure 6. Calcium (a) and phosphate (b) concentration in
`the medium at different time points for the 1.5 wt % Si-HA
`controls and the 1.5 wt % Si-HA sample with osteoclasts.
`(*Statistical difference p ⬍ 0.05, n ⫽ 3).
`
`detail by CLSM (Fig. 10). This fact could indicate that
`the osteoclasts are entering the resorption phase. It can
`be seen in the CLSM images that some cells on the
`surface on Si-HA formed podosomes [Fig. 10(c,d), white
`arrows]. The localization of the microfilaments showing
`the actin ring is very important, because they reflect the
`osteoclast phases of migration and resorption.
`The results obtained regarding calcium and phos-
`phate release were similar to those described previ-
`ously for PBMCs. Calcium and phosphate concentra-
`tions were analyzed at every medium change, until
`the final day of culture.
`In this case, earlier phosphate release into the me-
`dium of HA OC was observed, as soon as day 8 [Fig.
`11(a)]. At day 21, a slight increase in phosphate con-
`centration in the medium containing osteoclasts was
`noted. A similar trend was observed for calcium (HA
`NC – 1.88 mmol/L; HA MCSF – 1.88 mmol/L; and
`HA OC – 1.97 mmol/L) [Fig. 11(b)]. However, none of
`these results reached a statistically significant differ-
`ence.
`A higher concentration of phosphate was observed
`at day 8, 15, and 21 from the samples containing 0.8 wt
`% Si-HA OC, but when compared to the controls these
`were not statistically significant [Fig. 12(a)]. Once
`again, a higher release of calcium was observed by the
`(cid:25)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Figure 7. Fluorescence microscopy showing few cells with
`actin rings (white arrow). (a) HA MCSF; (b) 0.8 wt % Si-HA
`MCSF; (c) 1.5 wt % Si-HA MCSF, after 21 days of incubation.
`[Color figure can be viewed in the online issue, which is
`available at www.interscience.wiley.com.]
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`DIFFERENTIATION OF MONONUCLEAR PRECURSORS INTO OSTEOCLASTS
`
`715
`
`Figure 8. Fluorescence microscopy image of the actin rings and vitronectin of multinucleated cells (osteoclasts), on the HA
`surface, after 21 days. (a) actin ring; (b) vitronectin; (c) colocalization of the vitronectin receptor within the cell and the surrounding
`actin ring; (d) nuclei, discontinuities (white arrow). [Color figure can be viewed in the online issue, which is available at
`www.interscience.wiley.com.]
`
`end of the culture period [Fig. 12(b)], but again it was
`not statistically significant.
`In the case of 1.5 wt % Si-HA, the release of phos-
`phate was significantly higher at day 15, when com-
`pared with the controls [Fig. 13(a)] (1.5 wt % Si-HA
`NC – 1.07 mmol/L, p ⬍ 0.014; 1.5 wt % Si-HA MCSF
`– 1.06 mmol/L, p ⬍ 0.011; and 1.5 wt % Si-HA OC –
`1.27 mmol/L). While a similar increase was observed
`at the same time point for calcium it did not reach
`statistical significance [Fig. 13(b)]. At day 21 the cal-
`cium and phosphate content in the 1.5 wt % Si-HA OC
`medium was significantly higher than the controls
`(calcium: 1.5 wt % Si-HA NC – 1.77 mmol/L, p ⬍
`0.001; 1.5 wt % Si-HA MCSF – 1.86 mmol/L, p ⬍ 0.001;
`and 1.5 wt % Si-HA OC – 2.58 mmol/L; phosphate, 1.5
`wt % Si-HA NC – 1.06 mmol/L, p ⬍ 0.001; 1.5 wt %
`Si-HA MCSF – 1.17 mmol/L, p ⬍ 0.001; 1.5 wt %
`Si-HA OC – 1.57 mmol/L). These results indicate that
`the osteoclasts are active and resorbing.
`
`Once again the calcium and phosphate concentra-
`tion in media from the experimental samples in-
`creased with the increase in silicon content. When
`comparing the three materials, the calcium and phos-
`phate content in the medium of 1.5 wt % Si-HA OC at
`day 21 was significantly higher than for HA OC and
`0.8 wt % Si-HA OC (calcium – HA OC, p ⬍ 0.001; 0.8
`wt % Si-HA OC, p ⬍ 0.001; phosphate – HA OC, p ⬍
`0.001; 0.8 wt % Si-HA OC, p ⬍ 0.015). No significant
`difference was detected when the results from HA and
`0.8 wt % Si-HA were compared.
`
`DISCUSSION
`
`Through this study, it was possible to demonstrate
`that HA and Si-HA allows the differentiation of PBMC
`and CD14⫹ selected mononuclear cells into mature
`(cid:26)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`716
`
`BOTELHO ET AL.
`
`Figure 9. A detailed view of the osteoclasts on the surface of 0.8 wt % Si-HA (a, b, and c) and 1.5 wt % Si-HA (d, e, and f)
`after 21 days. (a)and (d) actin rings; (b) and (e) vitronectin; (c) and (f) nuclei.
`
`tronectin receptors. Although some of these markers
`osteoclasts. Most of the cells on the surface of the
`are expressed by other cell types such as macrophages,
`materials expressed osteoclastic markers: actin rings,
`according to Monchau et al., the formation of an actin
`multinuclearity, expressed TRAP, and presented vi-
`(cid:27)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`DIFFERENTIATION OF MONONUCLEAR PRECURSORS INTO OSTEOCLASTS
`
`717
`
`Figure 10. Confocal laser scanning microscopy images of the actin rings surrounding vitronectin receptors within the
`multinucleated cells – osteoclasts, on the 0.8 wt % Si-HA (a, c) and 1.5 wt % Si-HA (b, d) surfaces, after 21 days. (a) and (b)
`actin rings; (c) (d) vitronectin (white arrow indicates podosomes). [Color figure can be viewed in the online issue, which is
`available at www.interscience.wiley.com.]
`
`ring is the best in vitro method to confirm the oste-
`through complex focal adhesion contacts containing
`oclastic phenotype.21 In the fluorescence images, it
`several proteins in a ring shape at the periphery of the
`cell.22,23 In the fluorescence images, it was possible to
`was possible to identify podosomes. It is through this
`observe this structure, the actin ring, in the experimen-
`structure that osteoclasts adhere to the surface, via an
`intramembranous integrin (␣
`␤
`tal samples. It is under this actin ring that the oste-
`3). These molecules act
`v
`oclast makes a tight contact with the surface, followed
`as an intermediate between the extracellular matrix
`by the secretion of protons and proteases, which will
`proteins and the intracellular cytoskeletal actin micro-
`dissolve the bone crystals or the ceramic samples, as
`filaments. The microfilaments bind to integrins
`(cid:28)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`718
`
`BOTELHO ET AL.
`
`Figure 11. Phosphate (a) and calcium (b) concentration of
`the medium at different time points for the HA controls and
`the HA sample with osteoclasts (n ⫽ 3).
`
`Figure 13. Phosphate (a) and calcium (b) concentration in
`the medium at different time points for the 1.5 wt % Si-HA
`controls and the 1.5 wt % Si-HA sample with osteoclasts.
`(*Statistical difference p ⬍ 0.05, n ⫽ 3).
`
`in this in vitro study.22,23 The resorption products
`are endocytosed from the ruffled border and then
`transcytosed and released at the functional secretory
`domain in the top of
`the osteoclast basal mem-
`brane.24,25 Some of the actin rings presented disconti-
`
`nuity, which could be due to the change in the oste-
`oclast phase, resorption to migration or a decrease in
`the cellular activity.21 When the osteoclast is changing
`from the non-resorbing to the resorbing stage a large
`reorganization takes place, and, in the first stage of the
`resorption cycle, actin and vinculin are distributed
`throughout the podosomes. In the next stage, these
`structures coalesce to a specific area of the osteoclast,
`and the actin and vinculin stains dissociate. These
`changes are important because they reflect the cells’
`interaction with the extracellular environment.26
`To ensure that the calcium and phosphate measured
`in the medium was related to the osteoclastic activity
`and not related to either acellular degradation or non-
`osteoclast cell activity, several controls were per-
`formed. Each ceramic was simultaneously exposed to
`culture medium alone (without cells) and to cells but
`no cytokines (first experiment). In the second experi-
`ment, a different control was added; the samples were
`incubated with cells plus MCSF, to verify that resorp-
`tion occurs only in the presence of osteoclasts. In this
`second control, only MCSF was added and no
`RANKL, to avoid the differentiation of osteoclasts
`precursors into mature osteoclasts. Several studies
`have demonstrated the importance of RANKL for the
`differentiation and activation of osteoclast precursors.
`Figure 12. Phosphate (a) and calcium (b) concentration in
`It has been demonstrated that RANKL is fundamental
`the medium at different time points for the 0.8 wt % Si-HA
`for osteoclastogenesis and that this phenomenon can
`controls and the 0.8 wt % Si-HA samples with osteoclasts.
`(cid:20)(cid:19)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`

`

`DIFFERENTIATION OF MONONUCLEAR PRECURSORS INTO OSTEOCLASTS
`
`719
`
`ceramic showed evidence of osteoclast resorption,
`while no evidence of cellular resorption was observed
`on the HA samples.12 Although Patel et al. did not
`reported the presence of osteoclast-like cells in vivo,13
`we were able to demonstrate that Si-HA material can
`be resorbed by cellular mechanisms and that silicon
`has a positive effect on osteoclast activity. Although
`the exact mechanism behind this is still unknown and
`further studies are required.
`The primary objective for synthetic bone substitute
`research is the development of an implantable mate-
`rial which combines initial rapid healing with the
`subsequent capability to be progressively resorbed by
`osteoclasts during normal continuous tissue remodel-
`ing.12 So, Si-HA seems to be a good candidate.
`
`inter-
`be affect by several factors, namely MCSF,
`lukin-1, transforming growth factor-␤, tumor necrosis
`factor-␣,
`interlukin-6, vitamin D3, and parathyroid
`hormone. Although all these cytokines are very im-
`portant, they cannot induce the differentiation of os-
`teoclast precursors alone.27–31 The ability of TNF-␣ to
`induce osteoclastogenesis is still controversial. It has
`been shown that TNF-␣ can induce osteoclastogenesis,
`but by the stimulation of osteoblast/stromal cells to
`produce RANKL,32 or, if this cytokine acts directly on
`osteoclast precursors, at
`least a small amount of
`RANKL must be present.32 RANKL has the ability to
`induce osteoclastogenesis in the presence of MCSF.28
`RANKL acts on osteoclast precursors through a mem-
`brane receptor, RANK (receptor activator of NF-␬␤)
`activating osteoclastogenesis, although if it binds to
`the decoy receptor osteoprotegerin osteoclastogenesis
`will be inhibited.28 It has been established that MCSF
`is involved in the proliferation and survival of oste-
`oclast progenitors and osteoclasts, but it does not stim-
`ulate their resorbing activity.7,33
`The analyses of culture medium from the experi-
`mental samples showed significant differences. 1.5 wt
`% Si-HA had higher concentrations of calcium and
`phosphate in the medium from the samples contain-
`ing osteoclasts when compared between the three ma-
`terials and the controls, with the exception of 1.5 wt %
`and 0.8 wt % Si-HA, using PBMC as a starting culture,
`which indicates that the osteoclasts were active and
`resorbing on 1.5 wt % Si-HA. Similar increases in
`calcium and phosphate in the medium were reported
`by Zhang et al., when he seeded mature osteoclasts on
`the surface of bioglass.34 In the case of 0.8 wt % Si-HA,
`a higher increase in phosphate was observed in both
`cultures. No significant difference was observed be-
`tween HA and 0.8 wt % Si-HA. The difference be-
`tween the results obtained in the two starting cultures,
`PBMC, and CD14⫹, could be due to the presence of a
`mixed population of cells in the case of PBMC, only a
`small fraction of which are CD14⫹.
`From these results, it seems that HA is likely to be
`less resorbable in vitro, although other studies have
`shown evidence of resorption on HA.35,36 Several fac-
`tors can influence osteoclastic resorption, such as sin-
`tering temperature, porosity, grain size, and experi-
`mental conditions.11,36
`It seems that there is an increased resorption from
`the Si-HA materials compared to HA. It is known that
`a stable collagenous matrix is important for osteoclast
`attachment, for the actin ring formation, and conse-
`quent resorption.32 It was previously demonstrated
`that Si-HA has a higher binding capacity to collagen
`and that its affinity increases with silicon content,
`which can lead to an increase in the osteoclast resorp-
`tion activity.37 Similar results were obtained by Lang-
`staff et al., where it was demonstrated that a silicon-
`stabilized calcium phosphate coating and bulk
`(cid:20)(cid:20)(cid:3)(cid:82)(cid:73)(cid:3)(cid:20)(cid:21)
`Journal of Biomedical Materials Research Part A DOI 10.1002/jbm.a
`
`1. Schilling A, Linhart W, Filke S, Gebauer M, Schinke T, Rueger
`J, Amling M. Resorbability of bone substitute biomaterials by
`human osteoclasts. Biomaterials 2004;25:3963–3972.
`2. Miller P, Baram D, Bilezkian J, Greenspan S, Lindsay R, Riggs
`B, Watts N. Practical clinical application of biochemical mark-
`ers of bone turnover: Consensus of an expert