Bioactive Glass Scaffolds
for Bone Regeneration
Julian R. Jones1, Eileen Gentleman1 and Julia Polak2
DOI: 10.2113/ GSELEMENTS .3.6.393
T
The treatment for advanced-stage
osteoarthritis at sites such as the
hip or knee joint is a total joint
replacement with a metal implant.
However, all orthopaedic implants
have a limited life span as they
lack three of the most critical characteristics of living tissues: (1) the
ability to self-repair, (2) the ability
to maintain a blood supply and (3)
the ability to modify in response to
stimuli such as mechanical load.
KEYWORDS: bioactivity, scaffold, bone regeneration, tissue engineering As life expectancy increases and
degenerative bone diseases become
more common, the need for an
INTRODUCTION
artificial alternative to an autograft will become even more
Regenerative medicine is the stimulation of the body’s own important. A paradigm shift from replacement to regeneration
mechanisms to restore diseased or damaged tissue to its of tissues may provide a solution (Hench and Polak 2002).
here is a need for new materials that can stimulate the body’s own
regenerative mechanisms and heal tissues. Porous templates (scaffolds)
are thought to be required for three-dimensional tissue growth. This
article discusses bone regeneration and the specifications of an ideal scaffold
and the materials that may be suitable. Bioactive glasses have high potential
as scaffold materials as they stimulate bone cells to produce new bone, they
are degradable in the body and they bond to bone. The two types of bioactive
glasses, their mechanisms for bioactivity and their potential for scaffold
production are reviewed. Examples of their current clinical use are highlighted.
original state and function. Bone and cartilage are tissues
that are often in need of regeneration due to trauma,
tumour removal, or more commonly, age-related diseases
such as osteoporosis and osteoarthritis. Osteoporosis is the
loss of bone strength and density due to an imbalance in
the normal bone regeneration cycle. In healthy bone, bonegenerating cells (osteoblasts) lay down new bone, while
bone-resorbing cells (osteoclasts) resorb old bone. In older
people with osteoporosis, the osteoblasts are lower in
number and less active than normal; osteoclasts, however,
continue their work, leading to a loss of bone mass.
Osteoarthritis is a disease that causes loss of the articular
cartilage that acts as a cushion between bones and is the
bearing surface in joints. Without an adequate cushioning
at the joint surface, bone rubs against bone, leading to
degeneration and fracture.
The gold standard in reconstructive surgery for damaged or
diseased bone is the autograft, which involves harvesting
the patient’s own tissue from a donor site and transplanting
it to the damaged site. Alternatives are homografts (transplantation from another patient) and xenografts (tissue
from a different species, e.g. freeze-dried bovine bone).
There are many limitations to these techniques. Autografts
have low availability and can cause death of healthy tissue
at the donor site. Homografts carry the risk of disease transmission and, furthermore, are in short supply. Xenografts
are in large supply but they have even greater risks of
immune rejection, in situ degeneration and disease transmission.
BONE REGENERATION
Materials used in regenerative medicine are often designed
to act as templates or scaffolds for tissue in three dimensions (3D) and then safely dissolve once they have performed their function, leaving the body to remodel the
tissue to its natural form. One strategy for this is tissue engineering, where cells are stimulated to grow tissue outside of
the body (in vitro) and then are implanted to repair or
restore the body after disease or degeneration. In skin-tissue
engineering, skin is grown on a scaffold, and the scaffold
dissolves in vitro so that only the new skin is implanted in
the patient. This is the only tissue-engineering process in
clinical use at the time of writing.
However, in the case of bone, the body must remodel new
bone tissue according to its local loading environment. In
bone-tissue engineering, therefore, a scaffold–tissue biocomposite is necessary to restore function. An alternative
strategy is in situ bone regeneration, where a load-bearing
scaffold that can stimulate bone growth would be
implanted directly into the defect without the addition of
cells prior to implantation. This scaffold would then dissolve at the same rate the body creates new bone at the site.
A diagram illustrating this process for a simplified defect in
the jawbone is depicted in FIGURE 1.
The general criteria for an ideal scaffold for bone regeneration are as follows (Jones et al. 2006a). The scaffold
Q acts as a template for bone growth in three dimensions;
1
Department of Materials, Imperial College London
South Kensington Campus, London SW7 2AZ, UK
E-mail: [email protected]; [email protected]
W resorbs at the same rate as the bone is repaired,
producing degradation products that are non-toxic
and that can be excreted easily by the body;
2
Tissue Engineering and Regenerative Medicine Centre
Roderic Hill Building, Imperial College London
South Kensington Campus, London SW7 2AZ, UK
E-mail: [email protected]
E is biocompatible (not toxic) and promotes cell adhesion
and activity, stimulating new bone growth (osteogenesis);
ELEMENTS, VOL. 3,
PP.
393–399
393
D ECEMBER 2007
R bonds to the host bone without the formation
of scar tissue, creating a stable interface;
At present, a scaffold that fulfils all of these criteria does not
exist. In fact, many issues still need to be clearly defined so
that researchers have specific goals to aim for. For example,
one problem with criterion 7 is that at the time of writing
no regulatory procedures for such a material are in place.
There are procedures and guidelines for drugs, polymers
that dissolve and permanent devices, but there is not yet a
regulatory system for devices that will stimulate tissue
growth and resorb over time.
T exhibits mechanical properties matching those
of the host bone after in vitro tissue culture;
Y is made from a processing technique that can produce
irregular shapes to match that of the bone defect;
U has the potential to be produced commercially and
sterilised to the required international standards
for clinical use.
MATERIAL SELECTION
There are three classes of biomedical materials. Initially it
was considered that implant materials should trigger as little reaction as possible from the body. Materials that are
compatible with the body and trigger little biological
response are considered bioinert. However, no material is
completely inert in the body; the body sees implanted
materials as foreign bodies and isolates them by encapsulating them in scar tissue. In many applications, it is advantageous to have a material safely dissolve or degrade in the
body after it has performed its function. These bioresorbable materials are usually synthetic polyesters, but can
be ceramics. When it was recognized that the lack of interface between the implant and the tissue being repaired is a
problem, materials were developed that stimulate a beneficial biological response from the body. These materials are
generally available in three forms: (1) a powder for packing
into defects, (2) monolithic (dense) discs or rods, and (3)
porous forms. Porous materials consist of either isolated
(closed) pores or an interconnected network where the
pores are connected by windows, also termed interconnects
or apertures. Full definitions of these terms can be found in
Cerruti and Sahai (2006). The ideal candidate material for
an ideal scaffold would combine controlled resorbability
and bioactivity.
Bioactive Materials
The word bioactive has several uses. In its most general use,
it means that the material stimulates an advantageous biological response from the body on implantation. The term
was coined by Larry Hench in 1971, when he and his colleagues at the University of Florida invented Bioglass®, the
first material to form a strong bond to bone (Hench et al.
1971). This discovery not only launched the field of bioactive glasses, but bioactive ceramics in general. Initially,
bioactivity referred to materials that could bond to bone,
but Bioglass was later found to stimulate new bone growth
and to bond to soft tissues. Researchers have also developed
other materials, usually polymers, that can release biological stimulants such as bone morphogenic proteins, which
can stimulate bone growth and, therefore, can also be considered bioactive. In this article, bioactivity will be considered as the ability to bond to bone and to stimulate bone
growth without drugs or biological agents incorporated
into the material.
BIOACTIVE GLASS
Diagram illustrating how a porous bioactive glass scaffold
could be used to regenerate a bone defect. The scaffold
is placed at the site of the injury or defect, where it releases ions that
stimulate cells of the native tissue to differentiate, migrate and remodel
the scaffold. Eventually, the scaffold completely dissolves, and the
defect is filled with the patient’s own bone.
FIGURE 1
ELEMENTS
Bioactive glasses are amorphous, silicate-based materials
that bond to bone and stimulate new bone growth while
dissolving over time, making them candidate materials for
tissue engineering. Bioactive glass 45S5 Bioglass® (46.1%
SiO2, 24.4% NaO, 26.9% CaO, 2.6% P2O5, in mol%) was the
first material seen to form an interfacial bond with host tissue, when it was implanted in rats (Hench et al. 1971). The
strength of the interfacial bond between Bioglass and cortical bone was equal to or greater than the strength of the
host bone (Weinstein et al. 1980). Bioglass particulate has
been in clinical use since 1993 as Perioglas®, used to fill
periodontal defects (USBiomaterials Corp., Alachua, Florida),
394
D ECEMBER 2007
and more recently as NovaBone® (NovaBone Corp., Alachua,
Florida), used in orthopaedic applications. Since the initial
discovery of Bioglass, other compositions and glass types
have also been found to be bioactive. There are two ways to
produce bioactive glass: the traditional melt-derived
approach and the sol–gel process, each yielding very different glasses. The interest in bioactive glasses has been
expanded since their initial discovery and now not only
focuses on bone bonding, but also on their osteogenic
potential and applications in tissue engineering.
Mechanism of Bioactivity
Large differences in rates of in vivo bone growth among
various formulations of bioactive glasses have been
observed (Oonishi et al. 1999, 2000), indicating that there
are two classes of bioactive materials. Bioglass bonds to
bone and stimulates both osteoconduction and osteoinduction (Wilson and Low 1992; Hench 1998). Osteoconduction
is the growth of bone along the implant surface where the
implant touches the host bone. Osteoinduction is the generation of new bone on the implant surface, and the surface
does not necessarily have to be in contact with the host
bone. Osteoinduction occurs when bone cells or their progenitors are recruited by the material and signals are
provided for it to produce bone (Hench and Polak 2002).
The bonding of bioactive glass to bone has been attributed
to the formation of a hydroxyl carbonate apatite (HCA)
layer on the glass surface in contact with body fluid. HCA is
similar to bone mineral and therefore forms a bond (Hench
et al. 1971). The HCA layer forms as a result of a sequence
of chemical reactions on the surface of the implant when it
is exposed to body fluid (Pantano et al. 1974; Hench 1998).
There are five proposed reaction stages that lead to the formation of a strong bond between bioactive glass and living
tissue. These involve the rapid release of soluble ionic
species (glass corrosion or dissolution) from the glass, ultimately leading to the formation of a high-surface-area
hydrated silica and polycrystalline HCA bilayer on the glass
surface.
Stage 1: Rapid exchange of Na+ and Ca2+ with H+ or H3O+
from solution, causing hydrolysis of the silica groups,
which creates silanols (Si–OH):
e.g. Si–O–Na+ + H+ → Si–OH + Na+(aq)
Ion exchange is diffusion-controlled, with a t1/2 dependence. The pH of the solution increases as a result of H+ ions
in the solution being replaced by cations.
Stage 2: Stage 1 increases the hydroxyl concentration of
the solution, which leads to the attack of the silica glass
network. Soluble silica is lost in the form of Si(OH)4 to the
solution, resulting from the breaking of Si–O–Si bonds and
the continued formation of Si–OH (silanols) at the glass–
solution interface:
Si–O–Si + H2O → Si–OH + OH–Si
This stage is an interface-controlled reaction depending linearly on time.
Stages 3 to 5: Condensation and repolymerisation of the
Si–OH groups is then thought to occur, leaving a silica-rich
layer on the surface, depleted in alkalis and alkali-earth
cations (stage 3). Ca2+ and PO43- groups then migrate to the
surface through the silica-rich layer and from the surrounding
fluid, forming a CaO–P2O5-rich film on top of the silica-rich
layer (stage 4). The CaO–P2O5 film crystallizes as it incorporates
OH- and CO32- anions from solution to form a mixed HCA
layer (stage 5).
ELEMENTS
This mechanism is based on corrosion mechanisms of
soda–lime–silica glass, although the corrosion process is
much more rapid in bioactive glass compositions. The
kinetics of the first two stages implies that stage 1 is the
rate-determining step for bone bonding. However, it may
be the atomic structure of the glass (network connectivity)
that truly determines bioactivity (Hill 1996). The slowest
step is the rate-determining one. So, when the glass first
starts to dissolve (stage 1), diffusion would be rate limiting;
once the system gets closer to equilibrium, however, stage 2
becomes the rate-determining step. If glass dissolution continues even after a surface layer of bone has formed, there
must be some steady-state solution concentrations that are
reached due to simultaneous dissolution of glass and diffusion of ions through the bone layer and precipitation of
HCA. At this point, the system proceeds through parallel
reactions (not sequential), so the fastest step, probably diffusion, will determine the overall reaction rate. These reactions are conceptually very similar to silicate weathering
reactions in the geochemical environment, and the relationship between silicate mineral or glass and their reactivity provides one area where geochemists and mineralogists
could potentially interact with biomaterials scientists.
The biological mechanisms of bonding that follow HCAlayer formation are thought to involve the adsorption of
growth factors, followed by the attachment, proliferation
and differentiation of osteoprogenitor cells (Hench and
Polak 2002). Osteoblasts (bone-growing cells) lay down
extracellular matrix (collagen matrix), which mineralises to
create a nanocomposite of mineral and collagen on the surface of the bioactive glass implant while the dissolution of
the glass continues over time (Ducheyne and Qiu 1999).
Osteoinduction versus Osteoconduction
Bioactive glasses have been found to bond more rapidly to
bone than bioactive ceramics such as synthetic hydroxylapatite
(sHA). Bioactive glasses have also been found to be osteoinductive, i.e. they stimulate new bone growth on the implant
away from the bone–implant interface. sHA is classified as
osteoconductive, i.e. it encourages bone to grow along the
implant at the bone–implant interface (Oonishi et al. 1999).
The reasons that bioactive glasses are osteoinductive and
that sHA is osteoconductive have long been linked to the
rate of formation of the HCA surface layer. A faster rate of
HCA formation is thought to cause more rapid bone bonding. However, this would only explain why Bioglass has
greater osteoconductivity than sHA, not why bioactive
glasses are osteoinductive. The mechanism for osteoinduction, therefore, must be more complicated.
Stimulating osteoinduction creates an ideal environment
for regenerating bone, so an understanding of how bioactive
glasses induce this response is essential. That is, what signals
do the osteogenic cells receive from a bioactive glass that
guide them to create new bone? As Bioglass degrades it releases
dissolved silicon (as silicic acid), calcium, sodium and phosphate species into solution. It is thought that a combination
of some of these ions triggers the cells to produce new bone;
especially critical are concentrations of soluble silicon and
calcium ions (Hench and Polak 2002). Molecular biology
studies have shown that seven families of genes involved in
osteogenesis are stimulated by bioactive glass dissolution
products, including insulin growth factor II (IGF-II), IGFbinding proteins, and proteases that cleave IGF-II from their
binding proteins (Xynos et al. 2001). IGFs are involved in the
osteoblast’s synthesis of collagen, an essential component
of bone. It is thought that bioactive glasses determine gene
expression by the rate and type of dissolution ions released.
The intracellular signalling pathways, however, remain
uncertain.
395
D ECEMBER 2007
It is important to ascertain which ions cause osteoinduction
via gene expression, and at what concentrations. The effect
has been seen to be concentration dependent (Xynos et al.
2001), with approximately 17 to 20 µg ml-1 of soluble Si
and 88 to 100 µg ml-1 of soluble Ca ions required. Sodium
ions are not thought to be beneficial to cells, and the phosphate content of the glass is not thought to affect gene
expression (Hench and Polak 2002) although it may be
needed in the body fluid for the extracellular matrix to mineralise and form HCA. Recent studies have shown that
release of phosphate from the glass is not required for extracellular matrix production and that bone cells can mineralise as long as phosphate is present in the solution (Jones
et al. 2007b). This observation has been supported by Reffitt
et al. (2003), who showed enhanced differentiation of
osteoblastic cell lines when exposed to soluble silica
(orthosilicic acid) and found that the collagen extracellular
matrix production increased in all cells treated with
orthosilicic acid. In fact, dietary silica supplements have
long been associated with increased bone mineral density
(Carlisle 1982; Jugdaohsingh et al. 2004). This has led to the
development of silicon-containing composites (Phan et al.
2003) and silicon-doped hydroxylapatite materials that
have recently shown enhanced bone bonding compared to
conventional sHA (Porter et al. 2004).
Melt-Derived Bioactive Glasses
The original Bioglass was produced by melt processing,
which involves melting high-purity oxides (SiO2, Na2CO3,
CaCO3, P2O5) in platinum crucibles in a furnace at 1370°C
(Hench et al. 1971). Bioglass can be poured into preheated
(350°C) moulds (e.g. graphite) to produce rods or as-cast
components. Bioglass particulate is made by pouring the
melt into water to quench, creating a frit. The frit is then
dried and ground to the desired particle size range.
The compositional range for bonding of bone to bioactive
glasses and glass ceramics is illustrated in FIGURE 2. The most
bioactive glasses lie in the middle (region S) of the
Na2O–CaO–SiO2 diagram (assuming constant 6 wt% P2O5)
(Hench 1998). Compositions that exhibit slower rates of
bonding lie between 52 and 60 wt% SiO2 in the glass.
Compositions with greater than 60 wt% SiO2 (region B) are
bioinert.
Adding multivalent cations, such as Al3+, Ti4+ and Ta5+, to
the glass shrinks the bone-bonding field (Greenspan and
Hench 1976). The scientific basis for the compositional
boundaries is associated with the dissolution rate of the
glasses. Andersson et al. (1990) modified the 45S5 composition and implanted glasses with 16 different compositions
in the SiO2–Na2O–CaO–P2O5–Al2O3–B2O3 system into rabbit tibiae. Bone bonding occurred only with glasses that
could form an HCA layer, when tested in physiologically
balanced ionic buffer solution in vitro. The presence of
greater than 1.5 wt% Al2O3 in the glass inhibited bone
bonding by slowing HCA formation rate, by stabilizing the
silica structure and reducing the dissolution rate to prevent
calcium phosphate build-up within the layer. Up to about
1.5 wt% Al2O3 can be included in the glass without destroying
its bioactivity. Glasses with compositions within regions A
and S in FIGURE 2 bonded to bone; glasses with compositions
outside the bioactive regions did not bond.
Melt-Derived Clinical Products
Despite all these developments, few bioactive glass products
are in current clinical use, especially when compared to the
less bioactive sHA. The first Bioglass device was approved in
the United States in 1985 and was used to treat conductive
hearing loss by replacing the bones (ossicles) of the middle
ear. The device was called the Bioglass® Ossicular
ELEMENTS
Reconstruction Prosthesis (MEP®) and was a solid, cast
Bioglass structure that acted to conduct sound from the
tympanic membrane to the cochlea. The advantage of the
MEP over other devices in use at the time was its ability to
bond with soft tissue (the tympanic membrane) as well as
bone tissue. A modification of the MEP design was made to
improve handling in surgery, and it is used clinically with
the trademark name DOUEK MED.
The first particulate material sold in the United States was
PerioGlas®, which was approved via the U.S. Food and Drug
Administration (FDA) 510[k] process in December 1993 and
produced by USBiomaterials (Alachua, Florida). In 1995,
PerioGlas obtained a CE Mark (European Union approval),
and marketing of the product began in Europe. The initial
indication for the product was to restore bone loss resulting
from periodontal disease (holes in the jawbone) (Karatzas et
al. 1999). The glass particles, in the size range 90–710 µm,
are generally mixed with some of the patient’s blood and
packed into the bone defect. During a ten-year clinical history, PerioGlas has demonstrated excellent results, with virtually no adverse reactions to the product; it is now sold in
over 35 countries. Bioactive glass particles in the size range
200–355 µm have also been shown to undergo continual
dissolution after the formation of the HCA layer in
mandible sites. Dissolution led to the formation of a hollow
shell of HCA in which new bone formed (Radin et al. 2000).
This was not observed in previous studies where pellets or
rods of Bioglass were used; therefore, the formation of the
hollow shell is likely dependent on particle size and specific
surface area. This led to the product Biogran®, which is
Bioglass granules marketed by Orthovita Corp.
Building on the marketing successes of PerioGlas, a Bioglass
particulate (90–710 µm size range) for orthopaedic bone
grafting was introduced into the European market in 1999
under the trade name NovaBone®. The product was cleared
for general orthopaedic bone grafting in non-load-bearing
sites in February 2000. To date, NovaBone is being sold in
the United States, Europe, China and a number of other
countries.
Bioglass particulate is also used for the treatment of dentinal
hypersensitivity (sensitive teeth). The Bioglass material used
in this application is a very fine particulate (<5 µm in diameter)
that is incorporated into toothpaste or applied to the tooth
surface around exposed root dentin via an aqueous vehicle.
When Bioglass particles are put in contact with dentin, they
adhere to the surface, rapidly form an HCA layer and
occlude exposed tubules, thereby relieving pain. Studies
have shown that small amounts of Bioglass particulate perform better than current therapies (Gillam et al. 2002).
Early in 2004, the FDA cleared two products for sale through
the 510[k] process, and product sales began in mid-2004 by
Novamin Corp. (Alachua, Florida).
No porous, melt-derived bioactive glass scaffolds are in
clinical use.
Sol–Gel-Derived Bioactive Glasses
For a melt-derived glass to bond to bone, the silica content
has to be 60 mol% or less. However, HCA layer formation
and bone bonding can be achieved for glasses with up to
90 mol% silica if the glass is derived by a sol–gel process
(Li et al. 1991). The first sol–gel bioactive glasses were developed in the early 1990s (Li et al. 1991; Pereira et al. 1994).
Glasses of the 58S (60 mol% SiO2, 36 mol% CaO, 4 mol%
P2O5) composition, very similar to the melt-derived compositions developed previously, were found to form the HCA
surface layer more rapidly than any melt-derived glass
(Sepulveda et al. 2002b).
396
D ECEMBER 2007
The sol–gel process involves the hydrolysis of alkoxide precursors to create a sol or colloidal liquid. In the case of silicatebased bioactive glasses, the silicate precursor is an alkoxide
such as tetraethyl orthosilicate (TEOS). If components other
than silica are required in the glass composition, they are
added to the sol either as other alkoxides or as salts. In the
case of 58S, phosphate and calcium are incorporated by
adding triethyl phosphate (TEP) and calcium nitrate
tetrahydrate, respectively. The sol can be considered as a
solution of silica species that can undergo polycondensation
to form the silica network of Si–O–Si bonds (Hench and
West 1990). A gel forms within three days at ambient temperature. Water and ethanol, which are by-products of the
condensation reaction, must be evaporated by using carefully
controlled low heating rates. The final step is to heat the dried
gel to at least 600°C in order to remove organic by-products
(Saravanapavan and Hench 2003).
Sol–gel glasses have a specific surface area, typically
~200 m2 g-1, about two orders of magnitude greater than
that of melt-derived glasses (Sepulveda et al. 2001). This is
because gel glasses contain a nanoporous network that is
inherent to the sol–gel process, whereas melt-derived
glasses are fully dense. The nanopores in sol–gel glasses are
usually in the range of 1–30 nm in diameter. The nanopore
size can be tailored during processing by controlling the pH
of the catalyst (Brinker and Scherer 1990), the nominal
composition (Arcos et al. 2002) and the final temperature
(Jones et al. 2006b). It is, however, difficult to produce large
crack-free monoliths (greater than 10 mm in thickness)
because the heating step to drive off water, organics and
nitrates can create capillary stresses that cause cracking.
Advantages of sol–gel bioactive glasses over melt-derived
glasses of similar composition are that they are generally
more bioactive and can remain bioactive with silica contents
of up to 80 mol%. The enhanced bioactivity is due to the
nanoporosity and enhanced surface area (Sepulveda et al.
2001), which cause increased rates of dissolution, accelerating
stages 1 and 2 of the bioactivity mechanism. Sol–gel glasses
can, therefore, be considered truly bioresorbable. Gel glasses
can also be bioactive while containing fewer components,
e.g. glasses composed of 70 mol% SiO2 and 30 mol% CaO
(70S30C) form an HCA layer as rapidly as the 58S glass
(Saravanapavan et al. 2003).
The reason that 70S30C glass can nucleate an HCA layer
even though it does not contain phosphate is that Si–OH
groups are thought to play a role in HCA layer nucleation.
These groups form during the bioactivity mechanism (glass
corrosion), but in sol–gel glasses there are several Si–OH
groups present in the unreacted glass that can quickly act as
nucleation sites. HCA layers can be nucleated on various
materials that have a high concentration of surface OH
groups when the materials are placed in supersaturated
solutions (Li et al. 1994). Apatite layers have even been
demonstrated to form on polymers (Miyazaki et al. 2003).
Sol–gel glasses inherently contain a substantial number of
OH groups in the glass network. The glass network is therefore not completely cross linked. Hence, several sol–gel
silica-based glasses have nanoporosity, often microporosity,
which causes the high surface area. The higher surface area
causes more rapid dissolution than with dense Bioglass,
even though sol–gel glasses have a higher silica content and
greater network connectivity.
A further advantage of sol–gel glasses is that their surfaces
can be modified by a variety of surface-chemistry methods,
e.g. with amine groups, which can make the surface
hydrophobic and attractive to specific proteins such as
laminin. Specially designed proteins can also be attached to
ELEMENTS
Composition diagram for the bioactivity of melt-derived
silicate glass. Region S is a region where bioactive glasses
not only bond to bone but also are osteoinductive and gene activating.
FIGURE 2
the material surface prior to implantation to obtain novel
bioactivity by delivering the proteins to the desired wound
site (Lenza et al. 2003).
Sol–Gel-Derived Clinical Products
Products involving sol–gel bioactive glasses have only just
started to appear for clinical use. Novabone Corp. recently
modified their product by adding sol–gel glass particles to
particles of the 58S composition. The aim is that the 58S
particles will dissolve more rapidly than the Bioglass particles, creating spaces that will encourage bone ingrowth
between the Bioglass particles. The new Novabone product
was the first product containing sol–gel bioactive glasses to
be approved (in 2005), which paves the way for bringing
sol–gel scaffolds to the clinic.
Novathera Ltd. (Cambridge, UK) has developed Theraglass®,
a wound-healing gel that incorporates particles with the
70S30C gel glass composition, modified by the addition of
2 mol% of silver ions. Low concentrations of silver ions
have been found to be bactericidal without killing useful
cells (Bellantone et al. 2002).
POROUS BIOACTIVE GLASS SCAFFOLDS
Perhaps the biggest benefit from using sol–gel glasses over
melt-derived glasses is that they have led to the development of porous scaffolds with interconnected macropores
suitable for tissue-engineering applications (Sepulveda et al.
2002a). Although Bioglass was invented in 1971, no successful porous scaffolds have been synthesised from it
because a sintering process is employed in all known methods for producing porous structures from glass powder.
Sintering requires glasses to be heated above their glass
transition temperature in order to initiate localised flow,
and Bioglass crystallizes immediately above its glass transition. Therefore, a sintered Bioglass scaffold is actually a
glass-ceramic or ceramic scaffold.
The sol–gel foaming process involves the foaming of the sol
with the aid of a surfactant by vigorous agitation as the viscosity rapidly increases (Sepulveda et al. 2002a; Jones et al.
2006b). On gelation, the spherical bubbles become permanent
in the gel, and as drainage occurs in the foam struts, the gel
shrinks and the bubbles merge. Circular apertures (interconnects) then open up at the point of contact between
397
D ECEMBER 2007
neighbouring bubbles, creating an interconnected network
similar to trabecular bone, a spongy highly porous form of
bone tissue (see Boskey 2007 this issue). This interconnected (i.e. permeable) pore structure is essential for creating
an effective scaffold because cells can migrate into the
pores, and tissue can then grow throughout the scaffold
template. The interconnect must be at least 100 µm in
diameter in order to create channels wide enough to allow
cell migration, bone ingrowth and blood vessel development.
FIGURE 3 shows an X-ray microcomputed tomography (µCT)
image of a bioactive glass foam scaffold. The image shows
that the macropores are well interconnected (the scaffold is
permeable). In fact, the pore structure is hierarchical
because the nanoporosity inherent to the sol–gel process is
maintained (Jones et al. 2007a). This nanoporosity is beneficial to cells as it mimics the hierarchical structure of natural tissues and, therefore, more closely simulates a
physiological environment that stimulates cell behaviour
than a surface without nanopores. Compressive strengths
of 2.4 MPa, while maintaining modal (peak of size distribution) interconnect diameters above 100 µm, have been
achieved by tailoring the nanoporosity during processing
(Jones et al. 2006b). The strength values are similar to those
of porous bone and clinically used porous hydroxylapatite
(Valentini et al. 2000) and are continually increasing as the
process is improved.
The compressive strengths of porous bioactive glass scaffolds may be suitable for low-load sites such as Hills-Sacks
lesions in the shoulder (bone defects in the ball of the
shoulder that are notoriously slow in healing) and in sites
that are primarily in compression only, such as fused spinal
vertebrae. However, these scaffolds are brittle and would
not survive in a dynamic high-load environment such as
the hip. It is likely that the only way to produce a porous
scaffold with the bone-bonding and cell-stimulating properties of a bioactive glass and the toughness of a composite
would be to create an inorganic–organic nanocomposite by
incorporating biodegradable polymers into the sol–gel
process (Pereira et al. 2005; Vallet-Regí et al. 2006). A scaffold with this composition would mimic the structure of
natural bone, which is a composite of brittle hydroxylapatite and tough collagen. Similarly to fibreglass, an inorganic–organic nanocomposite would yield a material
incorporating the desired properties from each component.
Cell-response studies on bioactive glass foam scaffolds have
found that primary human osteoblasts lay down mineralised immature bone tissue, without additional signalling
species such as dexamethasone and ß-glycerophosphate.
This occurs in scaffolds of both the 58S composition
(Gough et al. 2004) and the 70S30C composition (Jones et
al. 2007b), which indicates that phosphate is not required
in the glass composition for bone matrix production and
mineralisation to occur. This is thought to be due to the
stimulation of the bone cells by the silicon and calcium
ions to lay down matrix and to the subsequent mineralisation
of the matrix; however these mechanisms are not yet clear.
Whatever material is used as a scaffold, it is vital to be able
to characterise the pore networks of scaffolds in 3D to
ensure that the scaffold has the potential to allow 3D bone
growth throughout. For in vivo tests or clinical trials, it is
imperative to know the exact structure of the pore network
before and after implantation, so a non-destructive technique is required for imaging and quantification of the scaffold.
Many authors have used µCT images to display the pore
networks of their scaffolds, but little has been done to
quantify the images. Now, novel methods of 3D image
analysis have been developed to quantify the structure by
ELEMENTS
X-ray microcomputed tomography (µCT) image of a typical bioactive glass scaffold produced by the sol–gel foaming process, with streak lines showing calculated paths of fluid flow.
IMAGE COURTESY OF GOWSIHAN POOLOGASUNDARAMPILLAI
FIGURE 3
applying combinations of computer algorithms (Jones et al.
2007a). The µCT data can also be input into finite-element
models to predict mechanical properties and permeability
as a function of specific pore networks. In FIGURE 3, streak
lines represent the flow path of a fluid for a pressure applied
to the top surface of the bioactive glass scaffold. The flow
path is calculated from the resultant flow vectors determined by solving Stokes equations at the local scale using
the 3D geometry of the scaffold obtained via µCT. For this
study, the permeability was calculated using a program
code previously developed to study water flow in reservoir
rocks (Anguy et al. 1996). These models will allow optimisation of scaffold architecture and cell culture (bioreactor)
conditions for optimal tissue growth.
CONCLUSIONS
Since the invention of Bioglass, the first bioactive ceramic,
several bioactive ceramics and glass ceramics have been
developed. Bioactive glasses have been found to release ions
that stimulate bone cells at the genetic level, causing
osteoinduction. It is this biological mechanism that must
be fully understood if bioactive materials are to be fully
optimised. Porous scaffolds can then be developed that will
take these mechanisms into account. In this way, materials
can be optimised from the atomic to the macro level with
respect to cell response.
ACKNOWLEDGMENTS
The authors would like to thank Mr. Gowsihan
Poologasundarampillai, Dr. Robert Atwood and Professor
Peter Lee of Imperial College London for their collaboration
on 3D image analysis and Professor Dominique Bernard
(Institute of Chemistry and Condensed Matter of Bordeaux)
for his assistance with the flow calculations. Dr. Jones is a
Royal Academy of Engineering and EPSRC Research Fellow.
.
398
D ECEMBER 2007
REFERENCES
Andersson OH, Liu GZ, Karlsson KH, Niemi
L, Miettinen J, Juhanoja J (1990) In vivo
behaviour of glasses in the SiO2-Na2OCaO-P2O5-Al2O3-B2O3 system. Journal
of Materials Science: Materials in
Medicine 1: 219-227
of the Royal Society A-Mathematical
Physical and Engineering Science 364:
263-281
Jones JR, Ehrenfried LM, Hench LL (2006b)
Optimising bioactive glass scaffolds for
bone tissue engineering. Biomaterials 27:
964-973
Anguy Y, Bernard D, Ehrlich R (1996)
Towards realistic flow modelling.
Creation and evaluation of twodimensional simulated porous media:
An image analysis approach. Surveys in
Geophysics 17: 265-287
Jones JR, Poologasundarampillai G, Atwood
RC, Bernard D, Lee PD (2007a) Nondestructive quantitative 3D analysis for
the optimisation of tissue scaffolds.
Biomaterials 28: 1404-1413
Arcos D, Greenspan DC, Vallet-Regí M (2002)
Influence of the stabilization temperature
on textural and structural features and
ion release in SiO2-CaO-P2O5 sol-gel glasses.
Chemistry of Materials 14: 1515-1522
Jones JR, Tsigkou O, Coates EE, Stevens
MM, Polak JM, Hench LL (2007b)
Extracellular matrix formation and
mineralization on a phosphate-free
porous bioactive glass scaffold using
primary human osteoblast (HOB) cells.
Biomaterials 28: 1653-1663
Bellantone M, Williams HD, Hench LL (2002)
Broad-spectrum bactericidal activity of
Ag2O-doped bioactive glass. Antimicrobial
Agents and Chemotherapy 46: 1940-1945
Boskey AL (2007) Mineralization of bones
and teeth. Elements 3: 385-391
Brinker J, Scherer GW (1990) Sol-Gel
Science: The Physics and Chemistry of
Sol-Gel Processing. Academic Press, San
Diego, 912 pp
Carlisle EM (1982) The nutritional
essentiality of silicon. Nutrition Reviews
40: 193-198
Cerruti M, Sahai N (2006) Silicate biomaterials
for orthopaedic and dental implants. In:
Sahai N, Schoonen MAA (eds) Medical
Mineralogy and Geochemistry, Reviews
in Mineralogy & Geochemistry 64,
pp 283-313
Ducheyne P, Qiu Q (1999) Bioactive ceramics:
the effect of surface reactivity on bone
formation and bone cell function.
Biomaterials 20: 2287-2303
Gillam DG, Tang JY, Mordan NJ, Newman
HN (2002) The effects of a novel Bioglass®
dentifrice on dentine sensitivity: a scanning
electron microscopy investigation. Journal
of Oral Rehabilitation 29: 305-313
Gough JE, Jones JR, Hench LL (2004) Nodule
formation and mineralisation of human
primary osteoblasts cultured on a porous
bioactive glass scaffold. Biomaterials 25:
2039-2046
Greenspan DC, Hench LL (1976) Chemical
and mechanical behavior of Bioglass-coated
alumina. Journal of Biomedical Materials
Research 10: 503-509
Hench LL (1998) Bioceramics. Journal of the
American Ceramic Society 81: 1705-1728
Hench LL, Polak JM (2002) Third-generation
biomedical materials. Science 295: 10141017
Hench LL, West JK (1990) The sol–gel process.
Chemical Reviews 90: 33-72
Hench LL, Splinter RJ, Allen WC, Greenlee
TK (1971) Bonding mechanisms at the
interface of ceramic prosthetic materials.
Journal of Biomedical Materials Research
2: 117-141
Hill R (1996) An alternative view of the
degradation of bioglass. Journal of Materials
Science Letters 15: 1122-1125
Jones JR, Lee PD, Hench LL (2006a)
Hierarchical porous materials for tissue
engineering. Philosophical Transactions
ELEMENTS
Jugdaohsingh R, Tucker KL, Qiao N, Cupples
LA, Kiel DP, Powell JJ (2004) Dietary
silicon intake is positively associated with
bone mineral density in men and
premenopausal women of the Framingham
Offspring cohort. Journal of Bone and
Mineral Research 19: 297-307
Karatzas S, Zavras A, Greenspan D, Amar S
(1999) Histologic observations of
periodontal wound healing after treatment
with PerioGlas in nonhuman primates.
International Journal of Periodontics and
Restorative Dentistry 19: 489-499
Lenza RFS, Jones JR, Vasconcelos WL, Hench
LL (2003) In vitro release kinetics of proteins
from bioactive foams. Journal of Biomedical
Materials Research 67A: 121-129
Li R, Clark AE, Hench LL (1991) An
investigation of bioactive glass powders
by sol-gel processing. Journal of Applied
Biomaterials 2: 231-239
Li PJ, Ohtsuki C, Kokubo T, Nakanishi K,
Soga N, Degroot K (1994) The role of
hydrated silica, titania, and alumina in
inducing apatite on implants. Journal of
Biomedical Materials Research 28: 7-15
Miyazaki T, Ohtsuki C, Akioka Y, Tanihara
M, Nakao J, Sakaguchi Y, Konagaya S
(2003) Apatite deposition on polyamide
films containing carboxyl group in a
biomimetic solution. Journal of Materials
Science: Materials in Medicine 14: 569-574
Oonishi H, Hench LL, Wilson J, Sugihara F,
Tsuji E, Kushitani S, Iwaki H (1999)
Comparative bone growth behavior in
granules of bioceramic materials of various
sizes. Journal of Biomedical Materials
Research 44: 31-43
Oonishi H, Hench LL, Wilson J, Sugihara F,
Tsuji E, Matsuura M, Kin S, Yamamoto T,
Mizokawa S (2000) Quantitative
comparison of bone growth behavior in
granules of Bioglass® A-W glass-ceramic
and hydroxyapatite. Journal of Biomedical
Materials Research 51: 37-46
Pantano CG, Clark AE, Hench LL (1974)
Multilayer corrosion films on Bioglass
surfaces. Journal of the American
Ceramic Society 57: 412-413
Pereira MM, Clark AE, Hench LL (1994)
Calcium-phosphate formation on sol-gelderived bioactive glasses in vitro. Journal of
Biomedical Materials Research 28: 693-698
Pereira MM, Jones JR, Orefice RL, Hench LL
(2005) Preparation of bioactive glasspolyvinyl alcohol hybrid foams by the
sol-gel method. Journal of Materials Science:
Materials in Medicine 16: 1045-1050
399
Phan PV, Grzanna M, Chu J, Polotsky A,
El-Ghannam A, Van Heerden D,
Hungerford DS, Frondoza CG (2003) The
effect of silica-containing calciumphosphate particles on human
osteoblasts in vitro. Journal of Biomedical
Materials Research 67A: 1001-1008
Porter AE, Patel N, Skepper JN, Best SM,
Bonfield W (2004) Effect of sintered
silicate-substituted hydroxyapatite on
remodelling processes at the bone–implant
interface. Biomaterials 25: 3303-3314
Radin S, Ducheyne P, Falaize S, Hammond
A, (2000) In vitro transformation of
bioactive glass granules into Ca-P shells.
Journal of Biomedical Materials Research
49: 264-272
Reffitt DM, Ogston N, Jugdaohsingh R,
Cheung HFJ, Evans BAJ, Thompson RPH,
Powell JJ, Hampson GN (2003) Orthosilicic
acid stimulates collagen type 1 synthesis
and osteoblastic differentiation in human
osteoblast-like cells in vitro. Bone 32:
127-135
Saravanapavan P, Hench LL (2003)
Mesoporous calcium silicate glasses. I.
Synthesis. Journal of Non-Crystalline Solids
318: 1-13
Saravanapavan P, Jones JR, Pryce RS, Hench
LL (2003) Bioactivity of gel-glass powders
in the CaO-SiO2 system: A comparison
with ternary (CaO-P2O5-SiO2) and
quaternary glasses (SiO2-CaO-P2O5-Na2O).
Journal of Biomedical Materials Research
66A: 110-119
Sepulveda P, Jones JR, Hench LL (2001)
Characterization of melt-derived 45S5 and
sol-gel-derived 58S bioactive glasses.
Journal of Biomedical Materials Research
58: 734-740
Sepulveda P, Jones JR, Hench LL (2002a)
Bioactive sol-gel foams for tissue repair.
Journal of Biomedical Materials Research
59: 340-348
Sepulveda P, Jones JR, Hench LL (2002b)
In vitro dissolution of melt-derived 45S5
and sol-gel derived 58S bioactive glasses.
Journal of Biomedical Materials Research
61: 301-311
Valentini P, Abensur D, Wenz B, Peetz M,
Schenk R (2000) Sinus grafting with porous
bone mineral (Bio-Oss) for implant
placement: A 5-year study on 15 patients.
International Journal of Periodontics and
Restorative Dentistry 20: 245-254
Vallet-Regí M, Salinas AJ, Arcos D (2006)
From the bioactive glasses to the star gels.
Journal of Materials Science: Materials in
Medicine 17: 1011-1017
Weinstein AM, Klawitter JJ, Cook SD (1980)
Implant-bone interface characteristics of
Bioglass dental implants. Journal of
Biomedical Materials Research 14: 23-29
Wilson J, Low SB (1992) Bioactive ceramics
for periodontal treatment: Comparative
studies in the patus monkey. Journal of
Applied Biomaterials 3: 123-129
Xynos ID, Edgar AJ, Buttery LDK, Hench LL,
Polak JM (2001) Gene-expression profiling
of human osteoblasts following treatment
with the ionic products of Bioglass® 45S5
dissolution. Journal of Biomedical
Materials Research 55A: 151-157 .
D ECEMBER 2007
ELEMENTS
400
D ECEMBER 2007