ARTICLE IN PRESS
Biomaterials 28 (2007) 2677 2686
www.elsevier.com/locate/biomaterials
The influence of BMP-2 and its mode of delivery on the
osteoconductivity of implant surfaces during the early
phase of osseointegration
Yuelian Liua,b, Lukas Enggista, Alexander F. Kuffera, Daniel Buserc, Ernst B. Hunzikera,
a
ITI Research Institute for Dental and Skeletal Biology, University of Bern, Murtenstrasse 35, P.O. Box 54, 3010 Bern, Switzerland
b
Academic Centre for Dentistry Amsterdam (ACTA), Department of Oral Function, Section of Oral Implantology and Prosthetic Dentistry,
Louwesweg 1, 1066 EA Amsterdam, The Netherlands
c
Department of Oral Surgery and Stomatology, School of Dental Medicine, University of Bern, Freiburgstrasse 7, 3010 Bern, Switzerland
Received 23 November 2006; accepted 2 February 2007
Available online 12 February 2007
Abstract
Osteogenic agents, such as bone morphogenetic protein-2 (BMP-2), can stimulate the degradation as well as the formation of bone.
Hence, they could impair the osteoconductivity of functionalized implant surfaces.
We assessed the effects of BMP-2 and its mode of delivery on the osteoconductivity of dental implants with either a naked titanium
surface or a calcium-phosphate-coated one. The naked titanium surface bore adsorbed BMP-2, whilst the coated one bore incorporated,
adsorbed, or incorporated and adsorbed BMP-2. The implants were inserted into the maxillae of adult miniature pigs. The volume of
bone deposited within a defined   osteoconductive  (peri-implant) space, and bone coverage of the implant surface delimiting this space,
were estimated morphometrically 1 3 weeks later.
After 3 weeks, the volume of bone deposited within the osteoconductive space was highest for coated and uncoated implants bearing
no BMP-2, followed by coated implants bearing incorporated BMP-2; it was lowest for coated implants bearing only adsorbed BMP-2.
Bone-interface coverage was highest for coated implants bearing no BMP-2, followed by coated implants bearing either incorporated, or
incorporated and adsorbed BMP-2; it was lowest for uncoated implants bearing adsorbed BMP-2. Hence, the osteoconductivity of
implant surfaces can be significantly modulated by BMP-2 and its mode of delivery.
r 2007 Elsevier Ltd. All rights reserved.
Keywords: Biomimetic material; Bone morphogenetic protein (BMP); Bone; Bone healing; Osteoconduction
1. Introduction attempts have been made to improve the osteoconductivity
of implants, which are predominantly titanium-based [1].
Dental and orthopedic implants are widely used for These measures include surface roughening by sand- and
tooth and joint replacement. The surgical insertion corundum-blasting, and acid etching [2 4]. The osteophi-
techniques are safe, and the clinical success rates are high. licity of naturally oxidized titanium surfaces can also be
And the expectations of patients are high too. Success in heightened by binding fluoride or chloride ions thereto
dental and orthopedic implantology depends greatly on the [5,6], which enhance their affinity for serum- and bone-
biocompatibility, chemistry and microtopography of the matrix proteins. However, the greatest improvements in the
metal implant surface, since these properties contribute to osteoconductivity of implants have been achieved by
its osteoconductivity. If the efficacy of osteoconductivity is coating their surfaces with a layer of calcium phosphate
improved, implant osseointegration, and hence the re- [7]. And to render these layers osteoinductive, efforts have
establishment of functionality, are expedited. Numerous been made to functionalize them with osteogenic agents,
such as bone morphogenetic proteins (BMPs) [8,9].
However, since these agents can stimulate the degradation
Corresponding author. Tel.: [phone]; fax: [phone].
E-mail address: [e-mail] (E.B. Hunziker). as well as the formation of bone, they could impair the
0142-9612/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biomaterials.2007.02.003
ARTICLE IN PRESS
2678 Y. Liu et al. / Biomaterials 28 (2007) 2677 2686
osteoconductivity of the coated-implant surface. BMPs, for
A B
example, stimulate the recruitment, proliferation and
differentiation not only of osteoprogenitor cells but also
of osteoclasts at an early stage [10]. Hence, they may
promote the resorption of newly formed bone almost as
soon as it has been laid down [11]. Consequently, the net
volume of bone deposited may be lower in the presence
than in the absence of BMPs. The applied dose of the drug
is also critical, since an overdosage can trigger the
production of intrinsic BMP-inhibitors, such as noggins
[12].
To clarify the potential beneficial and adverse effects of
BMP-2 on the osteoconductivity of titanium implants, we
considered 2 basic types of surface: a sand-blasted, acid-
etched one and a sand-blasted, acid-etched one coated with
a layer of calcium phosphate. These 2 surfaces served as the
base-line controls. In the first category, BMP-2 was
adsorbed directly onto the metal surface; in the second, it
was either adsorbed onto, incorporated into, or incorpo-
rated into and adsorbed onto the calcium-phosphate
Fig. 1. (A) Photograph of the experimental dental implant and its
coating. The implants were inserted into the partially
insertion holder. (B) Two-dimensional representation of one of the
edentulous maxillae of adult Goettingen miniature pigs.
implant chambers [boxed area in (A)]. The region defined as the
The neoformation of bone both along the implant surface
osteoconductive space (see text) is indicated in black.
and up to a distance of 175 mm therefrom (the immediate
peri-implant space) was quantified 1 3 weeks after surgery,
175 mm therefrom within the inner half of each chamber (Fig. 1), during
viz., during the early phase of healing, when any
the early phase of healing (1 3 weeks).
interference with the osteoconductivity of the implants
would be apparent.
Our findings revealed the osteoconductivity of the
2.2. Experimental dental implant model
implants to be greatly influenced by BMP-2 and its mode
of delivery. The net volume of bone deposited within the The experimental dental implant is a cylindrical titanium bolt with a
length of 6 mm. It has a core diameter of 2.7 mm, which is flush with the
immediate vicinity of the coated implants (up to 175 mm
inner surface (or floor) of each bone chamber, and an outer diameter of
therefrom) could be increased by the presence of BMP-2
4.2 mm, which is flush with the point at which the sloping wall of each
during the first 2 postoperative weeks, but was decreased
bone chamber meets the outer surface of the bolt (Fig. 1). Hence, each of
by its presence during the third. Bone coverage of the
the 2 bone chambers has a depth of 0.75 mm [(4.2 2.7 mm)/2]. The
coated implants was lowered by the presence of BMP-2 at titanium implant has a standard sand-blasted, acid-etched surface.
each weekly juncture. The net volume of bone deposited
both directly upon and within the immediate vicinity of
2.3. Experimental animal model
uncoated implants was lowered by the presence of BMP-2
at each time point. In general, implant osteoconductivity
A total of 18 adult (2 4-year-old) Goettingen miniature pigs were used
was most severely compromised when BMP-2 was present
in this study. Six months prior to implant placement, 3 sinistral and 3
as an adsorbed depot (subject to a burst-release profile), dextral premolars were extracted from the maxilla of the animals under
conditions of general anesthesia, which was induced by Thiopental and
and less so when it was delivered via a slow-release system
maintained by Halothanes, both administered via an intratracheal tube.
(incorporated into a calcium-phosphate coating).
The animals were intravenously injected with benzylpenicillin [Duplocillin
LA (Intervet BV, NL): 12,000 IU/kg of body weight] at the time of surgery
and 3 days later, and with analgesics during the first 3 postoperative days.
2. Materials and Methods
Six months after tooth extraction, by which time the sockets had
completely healed, the implants were inserted under conditions of general
2.1. Experimental design
anesthesia (as described above). Six implants, 1 from each experimental
group, were inserted per animal using an essentially systematic random
The aim of this study was to assess the influence of BMP-2 and its mode protocol, but nevertheless ensuring that those bearing BMP-2 were not
of delivery on the osteoconductivity of purely sand-blasted and acid- placed next to those lacking this agent. Each implant was separated from
etched titanium surfaces, and of such surfaces that bore a coating of its neighbour(s) by a distance of at least 2 3 mm, to ensure that its healing
calcium phosphate. The experimental dental implants, whose bone activities were mutually exclusive.
chambers were designed to be large, were inserted at an orthotopic The implants were inserted using a standard surgical procedure [13].
maxillary site in adult Goettingen miniature pigs. Since we specifically Intra- and postoperatively, the animals were treated as during the tooth-
wished to address the issue of osteoconductivity, the quantification of extraction procedure. Those that were to be killed 1 week after surgery
newly formed bone was restricted to that covering the implant surface and were housed at the local veterinary hospital. Those that were destined for
to that laid down within its immediate vicinity, viz., up to a distance of sacrifice after 2 and 3 weeks were kept at a local farm.
ARTICLE IN PRESS
Y. Liu et al. / Biomaterials 28 (2007) 2677 2686 2679
The miniature pigs were killed in a generally anesthetized state by an coating during its deposition. It was introduced into the supersaturated
intravenous injection of potassium chloride, which induced cardiac arrest. solution of calcium phosphate at a concentration of 10 mg/ml.
The animal experiments were approved by the Bernese Animal
Protection Commission and were conducted in accordance with its
2.5.2. Adsorption of BMP-2
regulations.
BMP-2 was adsorbed directly onto the coated or uncoated surfaces of
the implants by applying 5 ml of a stock solution (1 mg/ml of sterilized
water) to each of the 2 chambers in turn. To ensure that the fluid was
2.4. Experimental groups
evenly distributed along the chamber surfaces during its application, the
implant was hand rotated around the central axis. The aqueous phase was
Six experimental groups were established (see Table 1). Groups 1 4
permitted to evaporate at ambient temperature.
consisted of titanium implants that were biomimetically coated with a
layer of calcium phosphate bearing either incorporated BMP-2 (group 1),
2.5.3. Incorporation and adsorption of BMP-2
adsorbed BMP-2 (group 2), incorporated and adsorbed BMP-2 (group 3),
In the hybrid group (group 3), BMP-2 was incorporated into the
or no BMP-2 (group 4: base-line control). Groups 5 and 6 consisted of
calcium-phosphate layer during the coating procedure, as indicated above.
naked (uncoated) titanium implants bearing either adsorbed BMP-2
An additional depot of the drug was then adsorbed directly onto these
(group 5) or no BMP-2 (group 6: base-line control).
coatings (10 mg per sample), likewise as indicated above.
The amount of BMP-2 incorporated into the coatings was determined
2.5. Biomimetic coating procedure and quantification of the
using the ELISA technique [16]. The amount of BMP-2 deposited on the
incorporated depot of BMP-2
coated or uncoated implants by adsorption (i.e., by evaporation of an
applied stock solution) was predetermined and therefore not estimated by
ELISA.
The biomimetic coating of implants with a layer of calcium phosphate
was conducted as previously described [14]. Briefly, the implants were
initially immersed in concentrated simulated body fluid (NaCl: 733.5 mM;
2.6. Tissue processing and sampling
CaCl2 2H2O: 12.5 mM; Na2HPO4 2H2O: 5 mM; NaCO3: 21 mM) under
high-nucleation conditions (i.e., in the presence of 7.5 mM MgCl2 6H2O),
The maxilla of each animal was bisected into its dextral and sinistral
to inhibit crystal growth, for 24 h at 37 1C.The thin (1 3-mm-thick)
halves using a Stryker orthopedic saw. The soft tissues were cut away and
amorphous calcium-phosphate layer thereby produced serves as a seeding
excess bone around the implants was removed using an Exact saw. The
substratum for the subsequent deposition of a crystalline coating, which
implants were then chemically fixed in 4% formaldehyde solution for 10
was prepared by immersing the implants in a supersaturated solution of
days at ambient temperature. After rinsing in tap water, they were
calcium phosphate (NaCl: 136 mM; CaCl2 2H2O: 4 mM; Na2HPO4 2H2O:
separated from each other using an Exact saw. The individual implants
2mM), buffered with TRIS (50 mM; pH 7.4), for 48 h at 37 1C. The entire
were dehydrated in ethanol and embedded in methylmethacrylate, as
coating procedure was conducted under sterile conditions. The mean
previously described [17]. Each implant was then sectioned according to a
thickness of the coatings deposited was measured in the light microscope
systematic random sampling protocol [18] using a Leica saw.
using 4 unimplanted, chemically fixed and plastic-embedded samples that
Each implant was sectioned parallel to its longitudinal axis. Relative to
were reserved for this purpose. The octacalcium phosphate microstructure
the central axis, the cutting direction was varied systematically in
of the coatings was confirmed by scanning electron microscopy and by
increments of 301 between the 6 implants of 1 experimental group. Once
Fourier-transform infrared spectroscopy, as previously described [15].
the cutting direction had been defined, each implant was sliced, at a fixed
interval of 500 mm, into 200-mm-thick parallel sections, with a random
2.5.1. Incorporation of BMP-2 start relative to the left-hand wall of the implant. This process yielded 8 10
Human recombinant BMP-2 (a generous gift from Wyeth, Cambridge, slices per implant. Every second slice was used for the quantitative analysis
Massachusetts, USA) was incorporated into the calcium-phosphate of bone and of the coating volume.
Table 1
Coating degradation and BMP-2 release
Group Absence/ Mode of Coating thickness (mm) [initial coating Amount of BMP-2 released (mg/implant), and
number presence of BMP-2 thicknessź54.7 (78.73) mm] BMP-2 (mg/ percentage of coating degraded with time
coating loading implant)
Inc. Ads. W. 1 W. 2 W. 3 W. 1 W. 2 W. 3
1 Coating + 22.2674.26 2.5871.36 1.0870.45 12.9571.80 7.6871.45 12.3371.77 12.6972.02
59.88% 95.39% 98.07 %
2 Coating + 45.5374.36 16.8275.45 13.3574.17 10 10  
17.06% 69.75% 76.04%
3 Coating + + 14.6372.46 6.5272.82 0.4670.2 22.9571.80 16.8171.56 20.2172.1 22.7571.82
73.72% 88.33% 99.17%
4 Coating 52.0071.83 27.7277.69 15.3173.14    
4.99% 49.91% 72.49%
5 No coating +    1010 
6 No coating       
Inc.źincorporated.
Ads.źadsorbed.
W.źweek.
Mean values7SEM are represented.
ARTICLE IN PRESS
2680 Y. Liu et al. / Biomaterials 28 (2007) 2677 2686
The saw-cuts were mounted on plexiglass holders, polished down to a
investigation revealed this layer to have a uniform
thickness of 100 mm, and surface stained with basic fuchsine, Toluidine
thickness of 54.7 (78.73) mm.
Blue O and McNeal s Tetrachrome [17].
New bone was deposited to varying degrees around all
implants in each experimental group by a process of
2.7. Light microscopy and histomorphometry
intramembranous growth (Fig. 2). Neither cartilage tissue
The sectioned specimens were examined in either a Vanox AH2, or a
Nikon Eclipse E100 light microscope.
A
The thickness of the coatings, and bone coverage of the coated or
uncoated implant surface delimiting the   osteoconductive  space (see
MB
below), were estimated using a grid of parallel vertical test lines, which
were separated by a fixed distance of 70 mm. The system of vertical lines
was orientated perpendicular to each implant-chamber surface in turn,
namely, first to the chamber floor and then to the chamber walls [14], for UB
the histomorphometric analysis at a final magnification of 200.
The volume densities of newly formed mineralized and unmineralized
UB
(osteoid) bone matrix deposited within the   osteoconductive  (peri-
Ti
implant) space were estimated separately using the point-counting
methodology described by Gundersen et al. [19] and by Cruz-Orive and
**
Weibel [20], at a final magnification of 115. The osteoconductive space
was defined as the inner half of each bone chamber up to an arbitrarily
fixed distance of 175 mm from the floor and the 2 side walls (see Fig. 1).
The central region of the inner-chamber half was not analyzed, since bone
B
would be deposited therein only as a consequence of considerable
osteoinductive activity. The outer-chamber half was likewise not analyzed,
since this region is dominated by native bone activity stemming from the
surrounding   cut  osseous surfaces.
**
2.8. Consumption and efficacy of BMP-2
AS
AS
On the basis of previously reported findings, BMP-2 is known to be
homogeneously incorporated into biomimetic calcium-phosphate coatings
[15]. The amount liberated is thus proportional to the amount of coating
degraded, and can be estimated from the thickness of the calcium- Ti
phosphate layer remaining at each time point.
Likewise on the basis of existing data, the depot of BMP-2 that was
superficially adsorbed onto coated or uncoated implant surfaces was
C
assumed to be completely liberated during the initial few days of
implantation, viz., before the first sampling time (1 week after surgery). MB
An efficacy index for BMP-2 was defined as the amount of newly formed
bone deposited within the osteoconductive space per microgram of BMP-2
MB
released from the implant during the first postoperative week.
MB
2.9. Statistical analysis
MB
All numerical data are presented as mean values together with the
Ti
standard error of the mean (SEM). Differences within the same group at
each sampling time were analyzed using the 1-way ANOVA test.
Differences between the various groups at a particular sampling time
were analyzed using the 2-way ANOVA test. The level of significance was
set at po0.05. SPSS software (Version 11.0.4) for an Apple Macintosh
Fig. 2. Light micrographs of 100-mm-thick sections through a bone
computer system was employed for the statistical analysis. Post-hoc
chamber of titanium implants after surface staining with basic fuchsine,
comparisons were made using Bonferroni corrections.
Toluidine Blue O and McNeal s Tetrachrome. (A) Titanium implant (Ti)
bearing a calcium-phosphate coating into which BMP-2 was incorporated,
1 week after surgery. At this stage, the coating (*) is being voraciously
3. Results
degraded by osteoclasts (arrows). Within the chamber, most of the newly
formed osseous tissue is as yet unmineralized (UB). A few seams of
At the time of sacrifice, all animals were in good health.
mineralized bone (MB) are nevertheless apparent. Barź500 mm. (B)
Histological inspection of the 108 implants inserted
High-magnification view of (A), illustrating serried osteoclasts (arrows)
revealed 1 of these to be misplaced, 1 to be placed too along the surface of the coating (*). The space (AS) between the coating
and the titanium implant (Ti) is an artefact of tissue processing.
deeply, and 3 to be associated with local inflammatory
Barź100 mm. (C) Titanium implant (Ti) bearing a calcium-phosphate
activity. These 5 implants were excluded from the analysis.
coating into which BMP-2 was incorporated, 3 weeks after surgery. By this
Histological analysis of the 4 coated and unimplanted
3-week juncture, the coating has been completely degraded by osteoclasts.
samples that were reserved to determine the mean thickness
All of the newly formed osseous tissue has undergone mineralization
of the calcium-phosphate layer at the onset of the (MB). Barź500 mm.
ARTICLE IN PRESS
Y. Liu et al. / Biomaterials 28 (2007) 2677 2686 2681
nor endochondral ossification was observed at any time highest in the absence of BMP-2. In the presence of the
point. growth factor, the highest total bone volumes were
associated with coated implants bearing an incorporated
depot. As at the 2-week juncture, the lowest total bone
3.1. Total bone deposited within the osteoconductive space
volumes were associated with coated implants bearing only
of implant chambers
adsorbed BMP-2.
The volume fractions of total bone deposited within the
osteoconductive space of the implant chambers in each 3.2. Proportions of mineralized and unmineralized bone
group are presented in Fig. 3. One week after the insertion matrix deposited within the osteoconductive space of implant
of implants, the total volume of bone deposited within the chambers
osteoconductive space was lowest in the 2 uncoated groups,
and lower in the presence than in the absence of BMP-2 The volume fractions of mineralized and unmineralized
(Fig. 3). The total volume of bone deposited within the bone matrix deposited within the osteoconductive space of
osteoconductive space of coated implants was higher, and implant chambers in each group are presented in Figs. 4
enhanced by the presence of BMP-2 (except in the hybrid and 5, respectively. At the 1-week juncture, the volume
group). By the second week, the total volume of bone had fraction of mineralized bone matrix was, not unexpectedly,
increased substantially in all groups. The bone volume low in all groups, and increased progressively during the
associated with uncoated implants was still lower in the second and third weeks. However, the relative relationships
presence than in the absence of BMP-2. The former value between the groups changed with time. Most notably, the
was comparable to that recorded for coated implants 1-week volume fraction of mineralized bone matrix was
bearing only adsorbed BMP-2, and the latter to that highest in association with coated implants bearing only
recorded for coated implants bearing no BMP-2. The total adsorbed BMP-2. But at the 2-week juncture, it was lower
volumes of bone associated with coated implants bearing in this group than in any of the other coated-implant
either incorporated BMP-2, or incorporated and adsorbed groups. And at 3 weeks, the value registered for coated
BMP-2, were comparable and the highest amongst all implants bearing only adsorbed BMP-2 was indeed the
groups at this 2-week juncture. lowest. At the 2-week juncture, the highest volume fraction
By the third week, the total volume of bone deposited of mineralized bone matrix was associated with coated
had increased further in all groups except the one with a implants bearing both incorporated and adsorbed BMP-2,
hybrid coating [incorporated and adsorbed BMP-2 (mean although the values recorded for uncoated implants
value unchanged)]. As at the 2-week juncture, the total bearing no BMP-2 and for coated ones bearing either no
bone volume associated with uncoated implants was higher BMP-2 or incorporated BMP-2 were not substantially
in the absence than in the presence of BMP-2. Indeed, the lower. At the 3-week juncture, the highest volume fractions
value in the former group was higher than in any of the of mineralized bone matrix were associated with coated
coated ones. Amongst the latter, the total bone volume was implants bearing no BMP-2 and with uncoated ones
Fig. 3. Volume fraction of total (mineralizied and unmineralized) bone deposited within the osteoconductive space of implants in each of the 6 groups, 1 3
weeks after surgery. Mean values (nź6 for each group at each time point) are represented together with the SEM. BMP inc: coated implants bearing
incorporated BMP-2; BMP inc+ads: coated implants bearing incorporated and adsorbed BMP-2; BMP ads: coated implants bearing adsorbed BMP-2;
BMP on metal: uncoated implants bearing adsorbed BMP-2; metal only: uncoated implants bearing no BMP-2.
ARTICLE IN PRESS
2682 Y. Liu et al. / Biomaterials 28 (2007) 2677 2686
Fig. 4. Volume fraction of mineralized bone deposited within the osteoconductive space of implants in each of the 6 groups, 1 3 weeks after surgery. Mean
values (nź6 for each group at each time point) are represented together with the SEM. Please see legend to Fig. 3 for an explanation of the abbreviations.
Fig. 5. Volume fraction of unmineralized bone deposited within the osteoconductive space of implants in each of the 6 groups, 1 3 weeks after surgery.
Mean values (nź6 for each group at each time point) are represented together with the SEM. Please see legend to Fig. 3 for an explanation of the
abbreviations.
bearing no BMP-2. At this 3-week juncture, the relation- incorporated BMP-2. The 3-week values for coated
ship observed between the 2 groups of coated implants implants bearing incorporated BMP-2 and for uncoated
bearing either incorporated BMP-2, or incorporated and implants bearing adsorbed BMP-2 were similar.
adsorbed BMP-2, at 2 weeks was reversed.
Interestingly, the volume fraction of unmineralized bone 3.3. Bone coverage of implant surfaces within the
matrix was highest in all groups at the 2-week juncture. At osteoconductive space
the 1-week juncture, the volume fraction of unmineralized
bone matrix associated with uncoated implants bearing Bone coverage of the coated or uncoated surface area of
adsorbed BMP-2 was very much lower than in any of the the osteoconductive space in each group is presented in
other groups. The highest value was recorded for coated Fig. 6.
implants bearing incorporated BMP-2. At the 3-week At each juncture, bone coverage of the implant surface
juncture, the lowest volume fraction of unmineralized bone was highest for coated implants bearing no BMP-2, and
matrix was associated with coated implants bearing no increased with time (Fig. 6). Likewise for coatings bearing
BMP-2. As at the 1- and 2-week junctures, the highest either adsorbed, incorporated, or incorporated and ad-
values were associated with coated implants bearing sorbed BMP-2, coverage of the implant surface with bone
ARTICLE IN PRESS
Y. Liu et al. / Biomaterials 28 (2007) 2677 2686 2683
Fig. 6. Bone coverage of the coated or uncoated implant surface in each of the 6 groups, 1 3 weeks after surgery. Bone deposited directly along the coated
or uncoated surface delimiting the osteoconductive space was expressed relative to the total area of the osteoconductive space to yield a percentage value.
Mean values (nź6 for each group at each time point) are represented together with the SEM. Please see legend to Fig. 3 for an explanation of the
abbreviations.
increased with time, although the relationship between 3.5. Efficacy of BMP-2
each of these groups differed at each juncture. In the
uncoated groups, coverage of the implant surface with Given that the incorporated depot of BMP-2 was
bone was lower in the presence than in the absence of homogeneously distributed throughout the coatings during
BMP-2 at each juncture. Indeed, the coverage was lower in their deposition, it is possible to calculate the amount that
the former group than in any of the others at all time was liberated together with their degraded volume during
points. each postoperative week. These values were determined on
the basis of the coating volumes remaining at each weekly
3.4. Degradation of coatings juncture. Knowing the initial loading dose of BMP-2 and
the coating volume into which it was incorporated (at time
Data relating to the degradation of implant coatings are zero), the amount of BMP-2 liberated during each
summarized in Table 1. Coatings bearing incorporated and postoperative week was calculated by subtracting the
adsorbed BMP-2 underwent the most extensive degrada- amount associated with the coating volume remaining at
tion during the 3-week monitoring period, followed by the end of any particular week from that associated with
those bearing only an incorporated depot of the drug. the coating volume present at the end of the preceding one
Coatings bearing no BMP-2 underwent the least degrada- (Table 1). The adsorbed depot of BMP-2 was assumed to
tion during the 3-week monitoring period, followed by be completely released during the first few days of
those bearing only an adsorbed depot of the drug. implantation (see Section 2). The amounts of BMP-2
During the first week of implantation, coatings bearing liberated during the first postoperative week were used to
incorporated and adsorbed BMP-2 were degraded by 74%, calculate the efficacy index for the drug during this period.
and those bearing only an incorporated depot of the drug This index was defined as the volume of bone formed
by 60%. Coatings bearing only adsorbed BMP-2 were within the osteoconductive space of the implant chambers
degraded by 17%, and those bearing no BMP-2 by 5%. per microgram of BMP-2 liberated during the first post-
By the end of the second week, coatings bearing operative week (Table 2). These data reveal the efficacy of
incorporated and adsorbed BMP-2 had been degraded by BMP-2 to be highest for coated implants that bore solely
88%, and those bearing only an incorporated depot of the an incorporated depot of the drug (0.19). It was almost 4-
drug by 95%. Coatings bearing only adsorbed BMP-2 had fold lower (0.05) when the incorporated depot of BMP-2
been degraded by 70%, and those bearing no BMP-2 by was supplemented with a superficially adsorbed one.
50%. However, for coatings that bore only an adsorbed depot
By the end of the third week, coatings bearing either of the drug, the efficacy index was 2.5-fold higher (0.13)
incorporated and adsorbed BMP-2, or only an incorpo- than for hybrid ones. But when BMP-2 was adsorbed
rated depot of the drug, had been almost completely directly onto the uncoated implant surface, the efficacy
degraded. Those bearing only adsorbed BMP-2 or no index was 10-fold lower (0.01) than when it was adsorbed
BMP-2 had been degraded by about 75%. onto the coated one.
ARTICLE IN PRESS
2684 Y. Liu et al. / Biomaterials 28 (2007) 2677 2686
Table 2
Efficacy index for BMP-2 within the osteoconductive space of implants during the first postoperative week
Group number Absence/presence of Mode of BMP-2 loading Efficacy index (mm3 of bone formed/mg of BMP-2
coating liberated)
Inc. Ads.
1 Coating + 0.19
2 Coating + 0.13
3 Coating + + 0.05
4 Coating 
5 No coating + 0.01
6 No coating 
Inc. źincorporated.
Ads.źadsorbed.
4. Discussion biomimetic calcium-phosphate coating. Uncoated titanium
implants bore adsorbed BMP-2, whereas coated ones bore
The osteoconductivity of titanium implant surfaces has either an adsorbed, an incorporated, or an incorporated
been markedly improved during the past 2 decades by and an adsorbed depot of the growth factor. During the
various physical and chemical means (see Section 1). And first 2 postoperative weeks, bone-formation activity within
more recently, attempts have been made to further the osteoconductive space was highest in association with
promote local bone-formation activity by conferring the coated implants bearing only incorporated BMP-2. The
implant surfaces with osteoinductive properties, namely, by effects elicited by other BMP-2-delivery modes during this
functionalizing them with an osteogenic agent, such as 2-week timespan were variable.
BMP-2 [21 23]. Thus far, these compounds have been When BMP-2 was available in both an incorporated and
generally adsorbed directly onto the implant surface an adsorbed form, the amount liberated during the first
[24 27], although attempts have been made either to bind postoperative week was greater than when it was available
them chemically thereto [28 30] or to incorporate them only in an incorporated form. This circumstance may
into a polymer carrier [31 34]. Since adsorbed osteogenic explain why the volume of bone formed close to the
molecules bind with low affinity to the metallic surface, implant surface during the initial week was lower in the
they are released rapidly within a physiological milieu, hybrid group: the high local concentration of BMP-2
thereby generating high local concentrations, which, by generated probably overstimulated the bone-resorption
stimulating the expression of noggins, can lead to their activity of osteoclasts. Furthermore, overstimulation of
unspecific inactivation within a few hours to maximally the osteoclasts probably had a catalytic effect on the
several days [8]. degradation rate of the coating, since, being a bone-matrix-
Adsorbed osteogenic molecules bind with higher affinity like material, it is subject to osteoclastic digestion [14].
to calcium-phosphate implant coatings than to the naked By the third postoperative week, the highest volumes of
metallic surface, and they are consequently released bone formed within the osteoconductive space were
more slowly. If these compounds are incorporated into, associated with the coated and uncoated implants that
rather than adsorbed onto, the crystal latticework which bore no BMP-2. Amongst the BMP-2-bearing implants,
has become possible only recently by virtue of the only coated ones carrying an incorporated depot of the
biomimetic coating procedure then they are liberated at drug were associated with bone volumes that approached
an even slower rate, as the inorganic layer undergoes these high levels. In the coated- and uncoated-implant
degradation [14]. groups with an adsorbed depot of the growth factor, and in
However, osteogenic agents recruit and activate not only the coated-implant group with an incorporated and an
osteoprogenitor cells but also bone-resorbing ones, namely, adsorbed depot of the drug, the values were clearly lower.
osteoclasts [35]. Since a fine balance probably exists Also these findings reflect the unspecific overstimulation of
between the opposing effects of an osteogenic agent, it is bone- (and of coating-) resorption activity by high local
necessary to optimize its dose, its mode of delivery, and the concentrations of BMP-2 at earlier junctures.
timing of its release to maximize not only the osteoinduc- As aforementioned, adsorbed osteogenic molecules bind
tive but also the osteoconductive response; otherwise, with lower affinity to a naked metallic surface than to a
osteoinductive activity could indirectly impair the osteo- calcium-phosphate-coated one (Liu et al., unpublished
conductive response along and near the implant surfaces. data). This circumstance is reflected in the net volumes of
In the present study, we compared the effects of BMP-2 bone deposited within the osteoconductive space of coated
and its mode of delivery on bone formation using 2 and uncoated implants bearing adsorbed BMP-2 during
established and well-characterized osteoconductive implant the first postoperative week, which were much lower in the
surfaces: a sand-blasted, acid-etched titanium one and a latter group than in the former. The adsorbed depot of
ARTICLE IN PRESS
Y. Liu et al. / Biomaterials 28 (2007) 2677 2686 2685
BMP-2 was released more rapidly from the uncoated formation activity to a degree that would compensate for
implants, thereby generating a high local concentration of the bone-degradation activity promoted at earlier junc-
the drug at an earlier juncture, which in turn stimulated tures. Hence, the bone-interface coverage occurring in the
bone-resorption activity at an earlier juncture, namely, presence of BMP-2 did not   catch up  with that occurring
during the first postoperative week. By the end of the in its absence. The same argument holds true for coated
second week, the bone volumes in these 2 groups were implants bearing only an adsorbed depot of the growth
similar. factor.
Since coating-incorporated BMP-2 molecules do not The osteoconductivity of implant surfaces is most
diffuse out of this delivery system to a significant degree severely impaired by BMP-2 when the drug is present as
(Liu et al., unpublished data), they are apparently liberated a superficially adsorbed depot. The   burst  release of
via a cell-mediated process, viz., during the osteoclastic adsorbed molecules generates a high local concentration of
degradation of the inorganic layer. In this respect, the the drug at a very early phase of healing, which does not
system is analogous to physiological bone remodeling, favour the expedition and augmentation of bone-formation
whereby growth factors are liberated from the bone matrix activity. To have the desired effects, BMP-2 must be
during its degradation. This appears to be the most efficient delivered in a more physiological-like manner, viz.,
way of inducing bone formation without overstimulating incorporated into the latticework of a bone-matrix-like
bone resorption. material (calcium-phosphate coating), wherefrom it can be
Bone-matrix formation is a rapid process, which is not liberated in a physiological-like manner, namely, by an
paralleled by the rate of mineralization in any mammalian osteoclast-mediated degradation of the layer within which
system. Our data bear this point out: unmineralized bone it is trapped. The facilitation of bone-formation activity by
matrix (osteoid) predominated over mineralized bone coating-incorporated BMP-2 was impaired by the presence
matrix during the first postoperative week; only during of an additional, superficially adsorbed depot of the drug.
the third week was its proportion significantly reduced. In the future, the dose of BMP-2 incorporated into
The deposition of osseous tissue close to the implant implant coatings must be tuned to establish the optimal
surface is important for the mechano-transduction of balance between high bone-formation and low bone-
loading forces, which will be compromised if bone growth degradation activities, without compromising the osteo-
in this region is patchy, viz., if the contact with bone conductive properties of the implant surface.
directly covering the implant surface is focal rather than
continuous. If only the bone-interface coverage is mea- 5. Conclusions
sured, this mechanical-force-transduction issue cannot be
addressed. For example, our data revealed the bone- The osteoconductivity of an implant surface can be
interface coverage to be highest for coated implants substantially modified by its functionalization with an
bearing no BMP-2 at each time point, but particularly at osteogenic agent, such as BMP-2, and by the mode of drug
the 2-week juncture, when it was 16% higher than in any of delivery. Osteoconductivity can be truly evaluated only if
the other groups. But this encouraging result was not the volume of bone deposited within the immediate vicinity
complimented by the findings relating to bone formation of the implant surface is determined in addition to the
within the osteoconductive space. At the 2-week juncture, bone-interface coverage. Indeed, the latter estimator may
the highest bone volumes within this region were associated be indicative only of a non-functional bony encapsulation
with coated implants bearing either incorporated, or of the implant, since in the absence of a continuous sheet of
incorporated and adsorbed BMP-2. Hence, bone-interface newly formed bone within the immediate vicinity of the
coverage is a parameter of limited value in assessing the implant surface, this bony covering cannot support load
anchorage requirements for optimal force transmission transmission. The osteoconductivity of functionalized
between the implant surface and the surrounding bony implant surfaces was most severely impaired when BMP-
tissue. Nevertheless, it assists in our reconstruction of the 2 was present as a superficially adsorbed depot upon
overall biological picture. That the bone-interface coverage calcium-phosphate-coated or uncoated implant surfaces,
was lower for coated implants in the presence than in the and least so when it was incorporated into a calcium-
absence of BMP-2 indirectly reflects the bone-resorption- phosphate coating. Using this latter drug-delivery system,
stimulating action of the drug at the early (1- and 2-week) the agent is liberated in a manner that most closely
junctures. By the third week, the bone-interface coverage resembles the release of growth factors from bone matrix
recorded for coated implants bearing either incorporated, during its osteoclast-mediated degradation under physio-
or incorporated and adsorbed BMP-2, closely approached logical conditions.
the value registered for coated implants bearing no BMP-2.
Similarly for uncoated implants, the bone-interface cover- Acknowledgements
age was consistently lower in the presence than in the
absence of BMP-2. And since BMP-2 was present only as The authors would like to express their gratitude to
an adsorbed depot, which is rapidly released, insufficient Wyeth (Cambridge, Massachusetts, USA), for providing
BMP-2 remained at the 3-week stage for stimulating bone- them with human recombinant BMP-2, and to Straumann
ARTICLE IN PRESS
2686 Y. Liu et al. / Biomaterials 28 (2007) 2677 2686
Institute AG (Basel, Switzerland) for supplying the [18] Gundersen HJ, Jensen EB. The efficiency of systematic sampling in
stereology and its prediction. J Microsc 1987;147(Pt 3):229 63.
implants. They would like to thank Simon Nussli, Karin
¨
[19] Gundersen HJ, Bendtsen TF, Korbo L, Marcussen N, Moller A.
Ludi, Nina Broggini and David Reist for their technical
¨
Some new, simple and efficient stereological methods and their use in
assistance, and the ESI-team of the Inselspital in Bern for
pathological research and diagnosis. APMIS 1988;96(5):379 94.
organizing the surgery.
[20] Cruz-Orive LM, Weibel ER. Recent stereological methods for
This work was supported by grants from the Swiss cell biology: a brief survey. Am J Physiol 1990;258(4 Pt 1):
L148 56.
National Science Foundation (to DB and EBH) and the
[21] Cochran DL, Schenk R, Buser D, Wozney JM, Jones AA.
ITI Foundation, Basel, Switzerland (to EBH and DB).
Recombinant human bone morphogenetic protein-2 stimulation of
bone formation around endosseous dental implants. J Periodontol
References
1999;70(2):139 50.
[22] Sykaras N, Woody RD, Lacopino AM, Triplett RG, Nunn ME.
[1] Schroeder A, Pohler O, Sutter F. Tissue reaction to an implant of a Osseointegration of dental implants complexed with rhBMP-2: a
titanium hollow cylinder with a titanium surface spray layer. SSO comparative histomorphometric and radiographic evaluation. Int J
Schweiz Monatsschr Zahnheilkd 1976;86(7):713 27. Oral Maxillofac Implants 2004;19(5):667 78.
[2] Gotfredsen K, Berglundh T, Lindhe J. Anchorage of titanium [23] Reddi AH, Cunningham NS. Initiation and promotion of bone
implants with different surface characteristics: an experimental study differentiation by bone morphogenetic proteins. J Bone Miner Res
in rabbits. Clin Implant Dent Relat Res 2000;2(3):120 8. 1993;8(Suppl. 2):S499 502.
[3] Albrektsson T, Wennerberg A. Oral implant surfaces: part 1 review [24] Kawai T, Mieki A, Ohno Y, Umemura M, Kataoka H, Kurita S,
focusing on topographic and chemical properties of different surfaces et al. Osteoinductive activity of composites of bone morphogenetic
and in vivo responses to them. Int J Prosthodont 2004;17(5):536 43. protein and pure titanium. Clin Orthop Relat Res 1993(290):
[4] Buser D, Schenk RK, Steinemann S, Fiorellini JP, Fox CH, Stich H. 296 305.
Influence of surface characteristics on bone integration of titanium [25] Ripamonti U, Yeates L, van den Heever B. Initiation of heterotopic
implants. A histomorphometric study in miniature pigs. J Biomed osteogenesis in primates after chromatographic adsorption of
Mater Res 1991;25(7):889 902. osteogenin, a bone morphogenetic protein, onto porous hydroxya-
[5] Cooper LF, Zhou Y, Takebe J, Guo J, Abron A, Holmen A, et al. patite. Biochem Biophys Res Commun 1993;193(2):509 17.
Fluoride modification effects on osteoblast behavior and bone [26] Noshi T, Yoshikawa T, Dohi Y, Ikeuchi M, Horiuchi K, Ichijima K,
formation at TiO2 grit-blasted cp. titanium endosseous implants. et al. Recombinant human bone morphogenetic protein-2 potentiates
Biomaterials 2006;27(6):926 36. the in vivo osteogenic ability of marrow/hydroxyapatite composites.
[6] Ellingsen JE, Johansson CB, Wennerberg A, Holmen A. Improved Artif Organs 2001;25(3):201 8.
retention and bone to-implant contact with fluoride-modified tita- [27] Hollinger JO, Schmitt JM, Buck DC, Shannon R, Joh SP, Zegzula
nium implants. Int J Oral Maxillofac Implants 2004;19(5):659 66. HD, et al. Recombinant human bone morphogenetic protein-2 and
[7] Wong M, Eulenberger J, Schenk R, Hunziker E. Effect of surface collagen for bone regeneration. J Biomed Mater Res
topology on the osseointegration of implant materials in trabecular 1998;43(4):356 64.
bone. J Biomed Mater Res 1995;29(12):1567 75. [28] Nanci A, Wuest JD, Peru L, Brunet P, Sharma V, Zalzal S, et al.
[8] Bagno A, Di Bello C. Surface treatments and roughness properties of Chemical modification of titanium surfaces for covalent attach-
Ti-based biomaterials. J Mater Sci Mater Med 2004;15(9):935 49. ment of biological molecules. J Biomed Mater Res 1998;40(2):
[9] Boyne P, Jones SD. Demonstration of the osseoinductive effect of 324 35.
bone morphogenetic protein within endosseous dental implants. [29] Park YJ, Kim KH, Lee JY, Ku Y, Lee SJ, Min BM, et al.
Implant Dent 2004;13(2):180 4. Immobilization of bone morphogenetic protein-2 on a nanofibrous
[10] Chen D, Zhao M, Mundy GR. Bone morphogenetic proteins. chitosan membrane for enhanced guided bone regeneration. Biotech-
Growth Factors 2004;22(4):233 41. nol Appl Biochem 2006;43(Pt 1):17 24.
[11] Liu Y, Huse RO, de Groot K, Buser D, Hunziker EB. Delivery mode [30] Puleo DA, Kissling RA, Sheu MS. A technique to immobilize
and efficacy of BMP-2 associated with implants. J Dent Res bioactive proteins, including bone morphogenetic protein-4 (BMP-4),
2007;86(1):84 9. on titanium alloy. Biomaterials 2002;23(9):2079 87.
[12] Rosen V. BMP and BMP inhibitors in bone. Ann NY Acad Sci [31] Heckman JD, Ehler W, Brooks BP, Aufdemorte TB, Lohmann CH,
2006;1068:19 25. Morgan T, et al. Bone morphogenetic protein but not transforming
[13] Buser D, Broggini N, Wieland M, Schenk RK, Denzer AJ, Cochran growth factor-beta enhances bone formation in canine diaphyseal
DL, et al. Enhanced bone apposition to a chemically modified SLA nonunions implanted with a biodegradable composite polymer. J
titanium surface. J Dent Res 2004;83(7):529 33. Bone Joint Surg Am 1999;81(12):1717 29.
[14] Liu Y, de Groot K, Hunziker EB. BMP-2 liberated from biomimetic [32] Agrawal CM, Pennick A, Wang X, Schenck RC. Porous-coated
implant coatings induces and sustains direct ossification in an ectopic titanium implant impregnated with a biodegradable protein delivery
rat model. Bone 2005;36(5):745 57. system. J Biomed Mater Res 1997;36(4):516 21.
[15] Liu Y, Hunziker EB, Randall NX, de Groot K, Layrolle P. Proteins [33] Itoh S, Matubara M, Kawauchi T, Nakamura H, Yukitake S,
incorporated into biomimetically prepared calcium phosphate coat- Ichinose S, et al. Enhancement of bone ingrowth in a titanium fiber
ings modulate their mechanical strength and dissolution rate. mesh implant by rhBMP-2 and hyaluronic acid. J Mater Sci Mater
Biomaterials 2003;24(1):65 70. Med 2001;12(7):575 81.
[16] Engvall E, Perlman P. Enzyme-linked immunosorbent assay (ELI- [34] Jones AA, Buser D, Schenk R, Wozney J, Cochran DL. The effect of
SA). Quantitative assay of immunoglobulin G. Immunochemistry rhBMP-2 around endosseous implants with and without membranes
1971;8(9):871 4. in the canine model. J Periodontol 2006;77(7):1184 93.
[17] Schenk RK, Olah AJ, Herrmann W. Preparation of calcified tissues [35] Cowan CM, Aalami OO, Shi YY, Chou YF, Mari C, Thomas R, et
for light microscopy. In: Dickson GR, editor. Methods of calcified al. Bone morphogenetic protein 2 and retinoic acid accelerate in vivo
tissue preparation. Amsterdam: Elsevier Science Publishers B.V.; bone formation, osteoclast recruitment, and bone turnover. Tissue
1984. p. 1 56. Eng 2005;11(3 4):645 58.