Full text of DOI:10.1359/jbmr.2001.16.2.231

JOURNAL OF BONE AND MINERAL RESEARCH
Volume 16, Number 2, 2001
© 2001 American Society for Bone and Mineral Research
Bioactive Glass Stimulates In Vitro Osteoblast
Differentiation and Creates a Favorable Template for Bone
Tissue Formation
C. LOTY,1,2 J.M. SAUTIER,1 M.T. TAN,3 M. OBOEUF,1 E. JALLOT,4 H. BOULEKBACHE,5
D. GREENSPAN,3 and N. FOREST1
ABSTRACT
In this study, we have investigated the behavior of fetal rat osteoblasts cultured on bioactive glasses with 55
wt% silica content (55S) and on a bioinert glass (60S) used either in the form of granules or in the form of
disks. In the presence of Bioglass granules (55 wt% silica content), phase contrast microscopy permitted
step-by-step visualization of the formation of bone nodules in contact with the particles. Ultrastructural
observations of undecalcified sections revealed the presence of an electron-dense layer composed of needle-
shaped crystals at the periphery of the material that seemed to act as a nucleating surface for biological
crystals. Furthermore, energy dispersive X-ray (EDX) analysis and electron diffraction patterns showed that
this interface contains calcium (Ca) and phosphorus (P) and was highly crystalline. When rat bone cells were
cultured on 55S disks, scanning electron microscopic (SEM) observations revealed that cells attached, spread
to all substrata, and formed multilayered nodular structures by day 10 in culture. Furthermore, cytoenzy-
matic localization of alkaline phosphatase (ALP) and immunolabeling with bone sialoprotein antibody
revealed a positive staining for the bone nodules formed in cultures on 55S. In addition, the specific activity
of ALP determined biochemically was significantly higher in 55S cultures than in the controls. SEM
observations of the material surfaces after scraping off the cell layers showed that mineralized bone nodules
remained attached on 55S surfaces but not on 60S. X-ray microanalysis indicated the presence of Ca and P
in this bone tissue. The 55S/bone interfaces also were analyzed on transverse sections. The interfacial analysis
showed a firm bone bonding to the 55S surface through an intervening apatite layer, confirmed by the X-ray
mappings. All these results indicate the importance of the surface composition in supporting differentiation of
osteogenic cells and the subsequent apposition of bone matrix allowing a strong bond of the bioactive materials
to bone. (J Bone Miner Res 2001;16:231–239)
Key words: bioactive glasses, in vitro osteogenesis, mineralization, bone bonding, interfaces
1Laboratoire de Biologie-Odontologie, Faculte´ de Chirurgie Dentaire, Institut Biome´dical des Cordeliers, Universite´ Paris 7, Paris,
France.
2This study represents part of a dissertation to be submitted in partial fulfillment of the requirements for the degree of “Diploˆme de
Doctorat” by Christine Loty.
3US Biomaterials Corporation, Alachua, Florida, USA.
4Laboratoire de Microscopie Electronique, INSERM U314, Universite´ de Reims, Reims, France.
5Laboratoire de Biologie du De´veloppement, Universite´ Paris 7, Paris, France.
231

232
LOTY ET AL.
TABLE 1. BIOGLASS COMPOSITIONS (IN MOL%)
INTRODUCTION
SiO2
Na2O
CaO
P2O5
ISSUE ENGINEERING can be conceptualized as the use of
materials to promote new tissue formation and it in-
55S
60S
55.1
60.1
20.1
17.7
22.2
19.6
2.6
2.6
T
volves interactions of cells with the material. It has been
established that the essential requirement for implantable
materials to bond to living bone is the formation of a
biologically active apatite layer on their surface.(1) For
example, bioactive materials, such as calcium phosphate
ceramics, bioactive glasses, and glass ceramics have been
shown to form a strong chemical bond with bone. These
so-called bioactive materials may influence attachment pro-
liferation and differentiation of cells and the subsequent
integration in a host tissue. Reactions occurring at the sur-
face of bioactive glasses lead to the formation of a silica gel
layer and the subsequent crystallization of hydroxycarbon-
ated apatite (HCA).(2,3) Hench and Ethridge(4) defined key
compositional features that allowed a direct bond with bone
tissue: SiO2 content must be less than 60 mol%, have high
Na2O and CaO concentrations, and have a high Ca/P2O5
ratio. Such compositions developed a crystalline HCA when
implanted in the body that mediates a bond to both hard and
soft tissues. Glasses between 53 and 56 mol% of SiO2
formed an HCA layer at a slower rate and bond to bone but
not to soft tissues. Additionally, bioactive materials are
capable of releasing ions, which may affect cellular re-
sponses. For example, Matsuda and Davies(5) reported in
vitro a better osteoblast expression on bioactive glass than
on nonreactive glass. Similarly, Vrouwenvelder et al.(6–8)
observed a stimulatory effect of 45S5 Bioglass (U.S. Bio-
materials Corp., Alachua, FL, USA) on cultured rat osteo-
blasts in terms of cellular proliferation and differentiation.
Also, these studies have provided valuable information on
the effect of bioactive glasses on cell proliferation and
extracellular matrix formation; they have not explored the
effects of such bioactive surfaces on matrix biomineraliza-
tion. Compositions similar to 55% composition used in this
work have been studied for various bone grafting
applications.(9–11) Although 55S bioactive glass has been
shown to bond to bone but not to soft tissue,(12) the biolog-
ical mechanisms concerning the interfacial reactions be-
tween the glass and the cells remain poorly understood. For
this reason, this work was undertaken to examine the be-
havior of fetal rat osteoblasts cultured in the presence of a
bioactive glass with 55 wt% silica content (55S Bioglass)
and the interfacial interactions between the material and
bone cells. We have used an osteoblast culture system of
isolated fetal rat calvaria cells that forms nodular structures
with the characteristics of woven bone.(13) The aim of the
present study was to investigate the supposed stimulatory
effect of rat bone cells cultured in the presence of 55S
Bioglass. The second objective of this work was to examine
bone matrix formation and mineralization in contact with
the material. Immunocytochemical and biochemical param-
eters showed that 55S Bioglass promoted differentiation of
fetal rat osteoblasts. Furthermore, analytical scanning elec-
tron microscopy (SEM) and transmission electron micros-
copy (TEM) showed that the bioactive glass created a
MATERIALS AND METHODS
Materials
The materials used in this study were bioactive glasses
with 55% silica and a bioinert glass composition with
60 wt% silica by weight, respectively (U.S. Biomaterials
Corp., Alachua, FL, USA; Table 1). Both disks and granules
were used. Disks were 20 mm in diameter and 2 mm in
thickness. The size of the granules used in this study was
710–790 ␮m. Before exposing the bioactive glasses to cell
cultures, all samples were cleaned ultrasonically in acetone
and sterilized by dry heat at 180°C for 2 h in a furnace.
Glasses with 60% silica by weight were used as control.
Cell culture method
Osteoblasts were isolated enzymatically from the calvaria
of 21-day-old fetal Sprague–Dawley rats, as previously
described by Nefussi et al.(13) Briefly, calvaria were asep-
tically dissected and fragments were incubated in
phosphate-buffered saline solution (PBS) with 0.25% col-
lagenase (type I; Sigma, St. Louis, MO, USA) for 2 h at
37°C. Then the cells, dissociated from the bone fragments,
were plated at 4 ϫ 104 cells/cm2 either directly onto 55S
and 60S disks or in culture dishes in the presence of loosely
distributed granules (size, 710–90 ␮m) at a concentration of
1 mg/ml of culture medium. The culture medium used is
Dulbecco’s modified Eagle’s medium (DMEM; Gibco,
Grand Island, NY, USA) supplemented with 10% fetal calf
serum (FCS; Biosis, Philadelphia, PA, USA), 10 mM
␤-glycerophosphate (Sigma), 50 ␮g/ml of ascorbic acid,
and 50 U/ml of penicillin-streptomycin (Gibco). Cells were
cultured up to 23 days in a humidified atmosphere of 5% in
air at 37°C. Culture media were changed 24 h after seeding
and at 48-h intervals thereafter.
TEM
On day 15, cultures in the presence of Bioglass particles
were fixed in situ in Karnovsky solution (4% paraformal-
dehyde and 1% glutaraldehyde) for 1 h and rinsed in 0.2 M
sodium cacodylate buffer (pH, 7.4). Bone cell cultures were
then postfixed in 1% osmium tetroxide diluted in 0.2 M
cacodylate buffer. Cells were then dehydrated in a graded
series of alcohol, embedded in Epon-Araldite, and incu-
bated for 1 day at 60°C. Semithin sections were cut through
cell layer and biomaterial granules with a diamond knife,
mounted on glass slides, and stained with toluidine blue.
Semithin sections were performed for orientation purpose
and then ultrathin sections were collected on copper grids
template for bone formation and allowed contact osteogen- and stained with 2.5% uranyl acetate in absolute ethanol for
esis. 4 minutes and lead citrate for 2 minutes. All sections were

IN VITRO BONE FORMATION ON BIOACTIVE GLASSES
233
examined using a Philips CM-12 transmission electron mi- with a rubber policeman. Cell extracts were sonicated be-
croscope.
fore enzyme assay to dissociate extracellular matrix and
liberate membranous alkaline phosphatase (ALP).
Estimation of protein content was carried out using the
Pierce BCA Protein Assay Kit (Pierce Chemicals, Rock-
ford, IL, USA). The specific activity of ALP was assayed in
the cell layers as the release of p-nitrophenol from
p-nitrophenolphosphate. The optical density was read at 410
nm in a spectrophotometer (Beckman 25; Beckman Coulter,
Fulerton, CA, USA) and the enzyme activity was expressed
as units per milligram of total proteins.
SEM
Cell cultures on 55S and 60S disks were fixed on day 1
and day 10 of culture in Karnovsky fixative solution (4%
paraformaldehyde and 1% glutaraldehyde) for 1 h and
rinsed three times with 0.2 M sodium cacodylate buffer (pH,
7.4). Then, samples were postfixed with 1% osmium tetrox-
ide and dehydrated in a graded series of alcohol and in
amyl-acetate before critical point drying. Specimens were
mounted on copper stubs with silver paints, coated with
30-nm gold in a Polaron sputtering apparatus (Kiln Ferm,
Milton Kayes, UK), and examined on a Jeol JSM-35 scan-
Energy dispersive X-ray analysis
Bone cell cultures with Bioglass particles: After 15 days
ning electron microscope (Jeol France, Croisysurfeine, in culture, samples were embedded in Epon 812. Ultrathin
France) at 15 kV.
sections were made, collected on copper grids, and stained
with 2.5% uranyl acetate in absolute ethanol for 4 minutes
and lead citrate for 2 minutes. Specimens were coated with
a conductive layer of carbon (Х10 nm) in a sputter coater.
Sections were observed with a TEM (Philips CM 30) oper-
ating at a voltage of 50–300 kV. Energy dispersive X-ray
(EDX) measurements were performed with an EDAX ana-
lyzer (SiLi detector; Oxford Instruments, Oxford, UK).
EDX spectra are collected on elements with X-ray energy
under 6 kV. Electron diffraction analysis was performed in
micro-micro diffraction mode.
Cytoenzymatic localization of alkaline phosphatase
Cultured cells were fixed in situ on day 10 at room
temperature in 2% citrate water solution for 30 s and dried.
The cells were then exposed for 5 minutes to a solution
containing naphthol AS-MX phosphate as substrate and fast
violet B salt as coupler (Sigma). The cultures were incu-
bated and observed without counterstaining. As a control,
cultures were incubated in the absence of the substrate.
Bone cell cultures on bioactive glass disks: The extrac-
tion buffer (0.1 M sodium carbonate bicarbonate, 1 M
MgCl2, and 0.2% NP-40) used for ALP activity assay
completely removed rat bone cells with their organic matrix,
and the resulting disks were prepared for SEM examina-
tions. Each disk was washed ultrasonically in acetone and
then in 70% alcohol, rinsed with distilled water, and dried
before observation. A number of disks also were embedded
in methyl methacrylate resin (Technovit 7200; Kulzer,
Francheville, France), sectioned with a diamond saw
(EXAKT; Microm, Francheville, France) to examine the
bone/material interface. Specimens were mounted on cop-
per stubs with silver paints and coated with carbon. Obser-
vations were performed with a scanning electron micro-
scope (JSM-840A) connected to an EDX microanalyzer
(EDX II; Link Analytical, Oxford Instruments).
Immunocytochemistry
Cells cultured on Bioglass disks were fixed on day 10
with 4% paraformaldehyde for 15 minutes and cell mem-
branes were permeabilized with 0.1% triton X-100 in PBS.
After three washes of PBS, rabbit anti-rat immunoglobulin
G (IgG) bone sialoprotein (dilution, 1/50; a gift from Dr
L.W. Fisher, National Institute for Dental Research, Na-
tional Institutes of Health [NIH], Bethesda, MD, USA) was
applied for 1 h at 37°C. Subsequently, cultures were rinsed
three times with PBS, incubated for 30 minutes with the
second antibody, fluorescein isothiocyanate (FITC)–labeled
goat anti-rabbit IgG (dilution, 1/40; Southern Biotechnol-
ogy Associated, Inc., Birmingham, AL, USA). Primary
antibodies were omitted in negative controls. After washes
in PBS, the bioactive glass disks and control coverslips were
mounted with glycerol and observed under an epifluores-
cence microscope (Leitz Orthoplan, Wetzlar, Germany).
Photomicrographs were taken on TMAX 400 ASA film
(Eastman Kodak Corp., Rochester, NY, USA).
RESULTS
Cultures with 55S Bioglass granules
Bone nodule formation: Use of phase contrast micros-
copy allowed for in vitro analysis of the morphological
changes associated with formation of bone nodules on Bio-
glass granules. During the first days in culture, the cells
Protein synthesis and alkaline phosphatase–specific
activity assay
Before biochemical assay, osteoblastic cultures on the proliferated and reached confluence by day 4 of culture.
two substrata were prepared at different times (4 h, days 3, Thereafter, multilayers of cells were visible around the glass
5, 8, 12, 17, and 22), washed with DMEM (0% FCS) on ice, granules and formed a refringent collar that made observa-
and incubated with sodium-carbonate-bicarbonate buffer tions of the cell outline impossible (Fig. 1A). On day 7, a
(0.1 M NaHCO3-Na2CO3, pH 10.2). Samples were stored at differentiated area was noted at the vicinity of the glass and
Ϫ80°C. For all assays (in triplicate), bone cell cultures were a refringent material could be observed in the center of this
unfrozen and incubated in an extraction buffer (0.1 M area in contact with the granule (Fig. 1B). On day 12 of
sodium carbonate bicarbonate, pH 10.2, 1 M MgCl2, and culture, a change was noted by the observation that the
0.2% NP-40) for 10 minutes and removed of their substrate refringent material became more opaque (Fig. 1C). Further-

234
LOTY ET AL.
FIG. 2. (A) Semithin section of rat bone cell cultures on day 15 with
55S bioactive glass particles (magnification ϫ100). Two bone nodules
are located under the 55S particle (BG, bioactive glass; arrows, bone
nodules). (B) The same semithin section at a higher magnification
showing the overall morphology of bone nodules and their relation-
ships with the 55S particle (magnification ϫ200).
Histology: Thin sections processed perpendicularly to the
cell layer (fixed on day 15) of 55S granules show bone
nodules directly in contact to the material (Fig. 2A). Cross-
sections through the dense glass matrix produced a break-
down of the material, which looked like a hollow zone. At
a higher magnification, these bone nodules appeared to be
formed of multilayered polygonal cells surrounding a min-
eralized zone (Fig. 2B).
TEM observation of undecalcified sections revealed the
presence of a dense extracellular matrix in contact with the
material (Fig. 3A). The matrix was composed of densely
packed collagen fibers either cut in cross-sections or in
longitudinal sections with their characteristic banding pat-
tern. An electron-dense layer is located at the periphery of
the 55S Bioglass. This layer was composed of two zones:
the inner portion was granular in appearance whereas the
external part is composed of needle-shaped crystals. Fur-
thermore, collagen fibers were in close contact with the
external portion of the glass, which seemed to act as a
nucleating surface for biological crystals. In addition, min-
eralized foci were visible in the osteoid matrix that devel-
oped independently of the material and were composed of
needle-shaped crystals similar to that observed at the surface
of the glass. Transmission electron microscopy (TEM) obser-
vations revealed an electron-dense layer, composed of needle-
shaped crystals at the periphery of the material. EDX spectrum
on this electron-dense layer shows the presence of P, Pb, U,
Ca, and Cu elements. Cu is caused by the copper grid. Lead
and uranium are a consequence of lead citrate and uranyl
FIG. 1. Observations in phase contrast microscopy of rat bone cell
culture with 55S bioactive glass particles at different times of culture.
(A) Day 4 in culture. Rat bone cells proliferated and were in confluence
around a 55S granule (magnification ϫ200). (B) Day 7 in culture. A
55S bioactive particle was surrounded by differentiated areas com-
posed of multilayers of cells with a refringent matrix (magnification
ϫ200). (C) Day 12 in culture. Bone nodules completely surrounded acetate coloration. X-ray microanalysis revealed an accu-
55S Bioglass (magnification ϫ200).
mulation of Ca and P but no Si (Fig. 3C), contrary to the
inner portion (Fig. 3B), which is amorphous (Fig. 3D). The
electron diffraction patterns obtained on electron-dense
needle-shaped areas revealed a crystalline form (Fig. 3E).
more, the bone nodule outline became well defined, clearly
distinguishing this region from the background cell layer.
The time sequence of the morphological changes noted in
the 55S composition were similar to those observed with the
bioinert 60S Glass (data not shown).
Cultures on 55S and 60S glass disks
Osteoblast differentiation and bone nodule formation:
SEM observations 24 h after seeding show that osteoblasts

IN VITRO BONE FORMATION ON BIOACTIVE GLASSES
235
FIG. 3. (A) TEM observation of ultrathin sections of the bone cell
culture with 55S granules on day 15 (BG, bioactive glass; magnifica-
tion ϫ20,000). A dense collagenous matrix is present at the periphery
of the material and biological mineralized foci are visible inside the
FIG. 4. (A) SEM observation of a single rat bone cell attached to the
surface of a 55S disk, 24 h after seeding (bar ϭ 10 ␮m). This cell
anchored to the material by the meaning of numerous lamellipodia and
exhibited a standoff morphology. (B) SEM image of the cell layer at
matrix (white arrows). An electron-dense layer composed of two zones day 10 of culture on 55S disk. Note the presence of nodular structures
surrounds the material. The interior of this electron-dense layer is
granular (asterisk) whereas the external part is composed of needle-
shaped crystals (black arrow). (B) EDX microanalysis of the inner part
of the electron-dense layer located at the periphery of the material.
X-ray spectrum reveals the presence of Si, P, and Ca. (C) EDX
microanalysis of the outer part of the electron-dense layer located at the
periphery of the material. X-ray spectrum reveals the presence of P and
Ca but no Si. (D) Electronic diffraction pattern of the inner layer at the
periphery of the material indicated that this part is poorly crystalline.
(E) Electronic diffraction pattern of the outer layer at the periphery of
the material indicated that this external part is crystalline.
composed of multilayered polygonal cells (bar ϭ 1 mm).
were anchored to the 55S bioactive glass disk by means of
numerous filopodia (Fig. 4A). SEM observations of rat
calvaria cell cultures at day 10 showed that rat bone cells
had proliferated and formed multilayered nodular structures
spread into a confluent cell layer on the 55S surface
(Fig. 4B).
FIG. 5. Cytoenzymatic localization of ALP activity is positive in
bone nodule on day 10 of culture (bar ϭ 100 ␮m).
Cytoenzymatic localization showed concentration of ALP
in the bone nodules that appear as a dark region (Fig. 5).
Furthermore, immunolabeling with bone sialoprotein anti-
ALP activity measured by enzyme assay gradually in-
body revealed a specific staining for the bone nodules in creased after 3 days of cultures (Fig. 7). By day 8 in culture,
55S cultures (Figs. 6A and 6B). It is noteworthy that osteo- a difference in ALP activity is seen between 55S and 60S
blast differentiation also occurred in 60S cultures and that compositions, and by day 12 this difference is 50% greater
bone nodules developed (data not shown).
for the 55S composition.

236
LOTY ET AL.
FIG. 6. (A) Immunolocalization of bone sialoprotein in culture on
55S bioactive glass showed a specific staining in the bone nodule on
day 10 of culture (bar ϭ 100 ␮m). (B) The same bone nodule observed
in phase contrast microscopy (bar ϭ 100 ␮m).
FIG. 8. SEM observations of 55S disks after elimination of the cell
layer and the organic matrix, after 22 days of culture. (A) SEM
examination revealed the presence of irregular nodular structures that
remained attached to the 55S surface (magnification ϫ90). (B) At a
higher magnification, this bone nodule appeared as a spongy structure
with numerous osteocytes lacunae (magnification ϫ450). (C) The same
nodule at a higher magnification (magnification ϫ1500).
FIG. 7. Time course of ALP-specific activity during 22 days of
culture on 55S disks and 60S glass control disks.
Bone bonding: The extraction buffer used for enzyme
assay completely eliminated the cells and the organic matrix
from the disks, and the resulting surfaces were observed
under an SEM. Examination of disks after 22 days of culture the disks showed evidence of the presence of mineralized
revealed the presence of mineralized nodular structures at- bone nodules in contact with the bioactive glass (Fig. 8A).
tached to the surface of 55S disks (Fig. 8) but not on 60S Furthermore, the surface of the glass is covered with the
(data not shown). Low-magnification SEM micrographs of reactive surface and exhibited cracks due to the SEM prep-

IN VITRO BONE FORMATION ON BIOACTIVE GLASSES
237
silica content 55S Bioglass promoted differentiation of fetal
rat osteoblasts. Furthermore, the glass created a template for
bone formation and allowed contact osteogenesis.
Immunocytochemical and biochemical data clearly indi-
cated a positive effect of 55S Bioglass on osteogenic dif-
ferentiation. Our findings agree with those of Vrouwen-
velder et al.(6–8) who observed a stimulatory effect of 45S5
Bioglass on cultured rat osteoblasts. In the same way, Mat-
suda and Davies(5) showed a better cellular colonization and
extracellular matrix production on bioactive glasses versus
nonbioactive glasses in calvaria organ cultures. Cells are
sensitive to the physicochemical characteristics of the ma-
terials with which they interact. This is the case particularly
for bioactive glasses that release ions and create on their
surface a microenvironment that could positively influence
the behavior of the cells. The reaction kinetics of bioactive
glasses has been studied extensively after incubation in
simulated body fluids. Glasses with a higher level of bio-
activity undergo surface reactions very rapidly, for example,
within 3 h for 45S5 Bioglass, and lead to the formation of
an HCA layer on its surface.(14,15) The more gradual devel-
opment of a crystalline HCA on 55S Bioglass results in
interactions of the cells with the reactive interface over a
longer period of time. Furthermore, the ion exchange cre-
ated an alkaline pH at the surface of the Bioglass that could
encourage osteoblast differentiation.
This phenomenon of ion exchange that also occurred in
the culture medium could explain the promotion of the
osteoblast phenotype in 55S Bioglass cultures. A recent
study provides evidence that this hypothesis is valid. This
study showed that soluble extracts of bioactive glasses
enhanced bone mineralization in organ cultures.(16) Like-
wise, it was shown that the process of bone formation
depended on an optimal alkaline pH in the extracellular
milieu surrounding the osteoblasts.(17) On the other hand,
others have shown that the products of glass corrosion
elevated the pH of the culture medium to a value that
adversely affects osteoblast activity. To prevent pH shifts,
these authors immersed the material in Tris buffer before
the cultures.(18,19) This discrepancy may be related to dif-
ferences in the composition of the glasses and their surface
reaction kinetics.
FIG. 9. EDX mapping of bone nodule observed in Fig. 8 by SEM
with secondary electrons (SE) showed the presence of P and Ca in the
nodule but no Si (bar ϭ 100 ␮m).
arations. At a higher magnification, a mineralized bone
nodule attached to 55S appeared as a spongy structure with
numerous lacunae, corresponding to osteocyte cell pro-
cesses that were very similar to trabecular bone (Figs. 8B
and 8C). EDX mapping of the surface of 55S indicated the
presence of Ca and P (Fig. 9) with a little silicon (Table 2).
The mineralized bone nodule contained calcium and phos-
phorous (Fig. 9) with a Ca/P ratio similar to the glass
surface layer, but no silicon (Table 2).
The SEM (secondary and backscattered electrons) and
X-ray microanalysis of the sectioned disks examined the
interface between the mineralized bone matrix and bioactive
glass. Figure 10A shows a section of a 55S disk on which rat
osteoblasts have been cultured 22 days and where the cells
have been removed as explained previously. The SEM
micrograph shows continuity between the ceramic and a
bonelike structure that seems to be mediated by an interme-
diate layer. These findings were confirmed using backscat-
tered electron microscopy where the mineralized bone nod-
ule was clearly visible with osteocyte lacunae inside the
mineralized matrix (Fig. 10B). X-ray microanalytical map-
pings of the bone mineralized nodule/55S Bioglass interface
indicated the distribution of Ca and P throughout the mate-
rial and the bone tissue (Fig. 10C). Si was present in the
glass bulk but in larger amounts beneath the Ca-P–rich
layer.
A biological explanation of the stimulatory effect of an
alkaline pH on osteoblast differentiation concerned the reg-
ulation of gap junction intercellular communications.
Yamaguchi et al.(20) have shown that the gap junction cou-
pling was increased at an alkaline pH in MC3T3-E1 osteo-
blastic cells. Gap junction communication seemed to play a
key role during cellular condensations that proceeded in
vivo as well as in vitro osteoblast differentiation.(21) The
increased pH created by the reaction of the 55S glass may
be enhancing gap junction communication, thereby increas-
ing osteoblast differentiation.
DISCUSSION
Osteoblast differentiation
Authors have reported that 55S Bioglass releases soluble
Bioactive materials have been defined based on a specific silicon immediately on exposure of the glass to an in vitro
biological response elicited when implanted into bone. This solution or to body fluids.(4,22,23) For example, silicon was
concept is based on control of the surface chemistry of the known to ensure a structural role in the formation of cross-
material. Reactions occurring at the surface of bioactive links between collagen and proteoglycans during bone
glasses lead to the formation of a silica gel layer and the growth.(24) A potential metabolic function served by soluble
subsequent crystallization of HCA.(2,3) The results of the silicon was shown in an elegant work by Keeting et al.,(25)
present study showed that a bioactive glass with 55 wt% which showed that zeolite-A present in soluble silicon stim-

238
LOTY ET AL.
TABLE 2. ELEMENT ATOMIC PERCENTAGE OF THE INTERFACE BONE/55S
Si K
Ca K
P K
Na K
O K
Bone nodule on 55S
Apatite layer on 55S
Si-rich layer on 55S
55S bioactive glass
0.2
2.0
21.0
13.0
16.2
15.7
2.0
8.9
8.6
1.2
1.4
Ϫ0.0
Ϫ0.3
0.1
74.7
74.0
75.7
68.9
7.1
9.0
namic surface chemistry that provides an extracellular stim-
ulus to the cell and an extracellular environment that is
compatible with adsorption of biologically active mole-
cules.
Matrix mineralization and contact osteogenesis
Another important finding reported in this study confirms
in vitro the capacity of mineralized bone tissue to form
directly on 55S bioactive glass and to bond with it. Bioac-
tive implants have differing rates of bonding depending on
their bulk composition. For example, glass compositions
with more than 60% SiO2 do not bond to bone. The present
study confirmed this data and showed evidence of bone
apposition to the surface of the bioactive glass but not on the
60S bioinert surface. This finding is surprising because
osteoblast differentiation and bone nodule formation oc-
curred normally on 60S glass. Hence, the bioactive surface
not only encouraged osteoblast differentiation but also al-
lowed contact osteogenesis. In contrast, mineralized bone
nodules were not present on the 60S surface after crapping-
off the cell layers. It is possible that bone nodules formed on
60S developed at a distance from the glass surface, inside
the multicellular layers. Another explanation is that the
mineralized bone nodules were in contact (but not bond)
with the bioinert surface and have been removed by the
mechanical action of the rubber policeman and the ultra-
sonic cleaning of the disks. This probably is because of the
fact that the bioinert 60S surface does not develop a well
HCA layer to which cells can attach. SEM observations of
mineralized bone nodules at the bone/55S Bioglass inter-
face, and analysis of calcium, phosphorus, and silicon evi-
denced the ability of the bioactive glass to bond with bone.
Bioactive glasses formed a hydrated silica-gel layer within
minutes of exposure to body fluids.(22) In addition, soluble
silicon accelerated the precipitation of an amorphous cal-
cium phosphate, and silanol, soluble organic silicon, pro-
vided a heterogeneous mechanism for hydroxyl carbonate
apatite.(26) In our study, the bioactive surface was confirmed
to be composed of calcium and phosphorous and was highly
crystallized. Furthermore, collagen fibrils coalesced with
this layer, which seemed to act as a nucleating surface for
biological crystals. Such observations suggested that at
FIG. 10. SEM observations and EDX analysis of the interface be-
tween bone nodules and 55S sectioned disks after elimination of the
cell layer on day 22 of culture. (A) SEM image by secondary electrons
(R, resin; B, bone nodule; BG, bioactive glass; bar ϭ 10 ␮m). (B) The
same zone observed by backscattered electron microscopy (black star,
Ca-P–rich layer; black asterisk, Si-rich layer; white asterisks, osteocyte
lacunae). (C) EDX mapping of the interface. Results showed the
presence of Ca and P in the bone nodule, a Ca-P–rich layer under the
nodule, and a Si-rich layer under the Ca-P–rich layer.
ulated human osteoblast proliferation and differentiation by early stages of osteogenesis, calcified foci might serve to
regulating transforming growth factor ␤ (TGF-␤) produc- bridge to collagen fibers and thereby create a continuous
tion. In our study, the negatively charged silica gel that mineralized bone tissue and allow fusion with the surface
formed at the surface of the bioglasses before crystallization layer of the glass. Hench and Paschall(1) first reported the
could have adsorbed adhesion molecules and growth factors bonding of bioactive glasses by interdigitation of collagen
and subsequently promoted osteoblast differentiation. Fi- fibers with the material surface. Last, we have previously
nally, we can hypothesize that the stimulatory effect of 55S shown in a TEM study that AW glass-ceramic bonded to
Bioglass on osteoblast differentiation is caused by a dy- bone formed in vitro through the formation of a Ca-P–rich

IN VITRO BONE FORMATION ON BIOACTIVE GLASSES
239
layer.(27) Furthermore, we showed that biomineralization
can be initiated on the bioactive layer, which acts as a
In: Sedel L, Rey C (eds.) Bioceramics, vol. 10, Elsevier
Science, Oxford, UK, pp. 429–432.
12. Hench LL, Wilson J 1998 Bioactive glasses: Present and
future. In: LeGeros RZ, LeGeros JP (eds.) Bioceramics, vol
11. World Scientific, New York, NY, USA, pp. 31–36.
13. Nefussi JR, Boy-Lefevre ML, Boulekbache H, Forest N 1985
Mineralization in vitro of matrix formed by osteoblasts iso-
lated by collagenase digestion. Differentiation 29:60–168.
14. Clark AE, Pantano CG, Hench LL 1976 Auger spectroscopic
analysis of bioglass corrosion films. J Am Ceram Soc 59:37–39.
15. Kim CY, Clark AE, Hench LL 1988 Early stages of calcium-
phosphate layer formation on bioglasses. J Non Crystalline
Solids 113:195–202.
16. Maroothynaden J, Zang X, Hench LL, Revell PA 1998 Effect
of bioglass and loading environment on whole bone organs
cultured in vitro. In: LeGeros RZ, LeGeros JP (eds.) Bioce-
ramics, vol 11. World Scientific, New York, NY, USA, pp.
501–504.
nucleating surface for biological crystals and then serves as
a template for the organization of the matrix. Richard et
al.(28) showed, by TEM and electron diffraction, that cal-
cium phosphate implants encourage an epitaxial biological
crystal deposition, which supports the data of the present
investigation.
In conclusion, there appears to be a better osteoblast
expression on 55S bioactive glass compared with bioinert
60S glass as shown by enzyme cytochemical, immunocy-
tochemical, and biochemical parameters. Furthermore, min-
eralization can be initiated in vitro with 55S bioactive glass
serving as a nucleating surface and a template for the
organization of the matrix and for generating bone tissue.
17. Green J 1994 Cytosolic pH regulation in osteoblasts. Miner
Electrolyte Metab 20:16–30.
18. El Ghannam A, Ducheyne P, Shapiro IM 1995 Bioactive
material template for in vitro synthesis of bone. J Biomed
Mater Res 29:359–370.
19. El Ghannam A, Ducheyne P, Shapiro IM 1997 Formation of
surface reaction products on bioactive glass and their effects
on the expression of the osteoblastic phenotype and the dep-
osition of mineralized extracellular matrix. Biomaterials 18:
295–303.
20. Yamaguchi DT, Huang JT, Ma D 1995 Regulation of gap
junction intercellular communication by pH in MC3T3–E1
osteoblastic cells. J Bone Miner Res 10:1891–1899.
21. Hall BK, Miyake T 1995 Divide, accumulate, differentiate:
Cell condensation in skeletal development revisited. Int J Dev
Biol 39:881–893.
22. Ogino M, Ohuchi F, Hench LL 1980 Compositional depen-
dence of the formation of calcium phosphate films on bioglass.
J Biomed Mater Res 14:55–64.
23. Hench LL, Wilson J 1984 Surface-active biomaterials. Science
226:630–636.
24. Hott M, de Pollak C, Modrowski D, Marie PJ 1993 Short-term
effects of organic silicon on trabecular bone in mature ovari-
ectomized rats. Calcif Tissue Int 53:174–179.
25. Keeting PE, Oursler MJ, Wiegand KE, Bonde SK, Spelsberg
TC, Riggs BL 1992 Zeolite A increases proliferation, differ-
entiation and transforming growth factor beta production in
normal adult human osteoblast-like cells in vitro. J Bone Miner
Res 7:1281–1289.
ACKNOWLEDGMENTS
The authors are grateful for SEM investigations per-
formed by the “Center Inter-Universitaire de Microscopie
Electronique” Jussieu, Universite´s Paris VI et Paris VII,
CNRS, 7, quai Saint Bernard, 75252 Paris Cedex 05,
France. We thank F. Meury for expert technical assistance.
This work was supported by a grant from CANAM—
Assistance Publique—Hoˆpitaux de Paris and by INSERM
and CNRS (n°4M109D). This work also was supported by
the U.S. Biomaterials Corp., Alachua, FL, USA.
REFERENCES
1. Hench LL, Paschall HA 1973 Direct chemical bonding be-
tween bioactive glass-ceramic materials and bone. J Biomed
Mater Res Symp 4:25–42.
2. Clark AE, Pantano CG, Hench LL 1976 Auger spectroscopic
analysis of bioglass corrosion films. J Am Ceram Soc 59:37–39.
3. Ogino M, Ohuchi F, Hench LL 1980 Compositional depen-
dence of the formation of calcium phosphate films on bioglass.
J Biomed Mater Res 14:55–64.
4. Hench LL, Ethridge EC 1982 Biomaterials: An interfacial
approach. Biophysics and Bioengineering Series No. 4, Aca-
demic Press, London, UK, pp. 221–254.
5. Matsuda T, Davies JE 1987 The in vitro response of osteo-
blasts to bioactive glass. Biomaterials 8:275–284.
6. Vrouwenvelder WCA, Groot GC, de Groot K 1992 Behavior
of fetal rat osteoblasts cultured in vitro on bioactive glass and
nonreactive glass. Biomaterials 13:382–392.
7. Vrouwenvelder WCA, Groot GC, de Groot K 1993 Histolog-
ical and biochemical evaluation of osteoblasts cultured on
bioactive glass, hydroxyapatite, titanium alloy, and stainless
steel. J Biomed Mater Res 27:465–475.
8. Vrouwenvelder WCA, Groot GC, de Groot K 1994 Better
histology and biochemistry for osteoblasts cultured on
titanium-doped bioactive glass: 45S5 compared with iron,
titanium, fluoride and boron-containing bioactive glasses. Bio-
materials 15:7–106.
9. Larmas E, Sewon L, Loustarinen T, Kangasniemi I, Yli-Urpo
A 1995 Bioactive glass in periodontal defects. Initial findings
of soft tissue and osseous repair. In: Wilson J, Hench LL,
Greenspan G (eds.) Bioceramics, vol. 8. Elsevier Science,
Oxford, UK, pp. 279–284.
26. Damen JJ, ten Cate JM 1989 The effect of silicic acid on
calcium phosphate precipitation. J Dent Res 68:1355–1359.
27. Sautier JM, Kokubo T, Ohtsuki T, Nefussi JR, Boulekbache H,
Oboeuf M, Loty S, Loty C, Forest N 1994 Bioactive glass-
ceramic containing crystalline apatite and wollastonite initiates
biomineralization in bone cell culture. Calcif Tissue Int 55:
458–466.
28. Richard M, Aguado E, Cottrel M, Daculsi G 1998 Ultrastruc-
tural and electron diffraction of the bone-ceramic interfacial
zone in coral and biphasic CaP implants. Calcif Tissue Int
62:437–442.
Address reprint requests to:
Dr. Christine Loty
Laboratoire de Biologie-Odontologie
Institut Biome´dical des Cordeliers
15–21, rue de l’Ecole de Me´decine
F-75270 Paris Cedex 06, France
10. Suominenn E, Kinnunen J 1996 Bioactive glass granules and
plates in the reconstruction of defects of facial bones. Scand J
Plast Reconstr Surg Hand Surg 30:281–289.
11. Aitasalo K, Sounpaa J, Peltola M, Yli-Urpo A 1997 Behavior Received in original form December 3, 1999; in revised form May
of bioactive glass (S53P4) in human frontal sinus obliteration. 25, 2000; accepted July 18, 2000.