RESEARCH
Hydroxyapatite as a Carrier for Bone
Morphogenetic Protein
Ramin Rohanizadeh*
Kimberly Chung
Bone morphogenetic proteins (BMPs) can induce the formation of new bone in numerous
orthopedic and dental applications in which loss of bone is the main issue. The combination
of BMP with a biomaterial that can carry and deliver proteins has been demonstrated to
maximize the therapeutic effects of BMPs. However, no ideal candidate with optimal
characteristics as a carrier has emerged for clinical use of BMPs. Hydroxyapatite (HA) is a
potential BMP carrier with its osteoconductive properties and desirable characteristics as a
bone graft biomaterial. In this study, 3 different methods to load BMP into HA materials were
characterized and compared based on the BMP uptake and release profile. BMP was loaded
into HA in 3 ways: (1) incorporation of BMP during HA precipitation, (2) HA immersion in
BMP solution, and (3) BMP incorporation during dicalcium phosphate dihydrate (DCPD)
conversion to HA. The size of HA crystals decreased when BMP was loaded during HA
precipitation and HA immersion in BMP solution; however, it did not change when BMP was
loaded during DCPD-to-HA conversion. The highest BMP uptake was achieved using the
immersion method followed by HA precipitation, and the lowest via DCPD conversion. It is
interesting to note that BMP loading during HA precipitation resulted in sustained and
prolonged BMP release compared with the 2 other BMP loading methods. In conclusion, BMP
incorporation during HA precipitation revealed itself to be the best loading method.
Key Words: hydroxyapatite, bone morphogenetic protein, drug delivery, carrier
INTRODUCTION
growth factors are able to lead and induce
bone formation, even in nonskeletal sites
lthough bone is one of the few
(eg, muscle). Biomaterials used in orthopedic
regenerative tissues in the
applications, such as ceramics, metals, and
human body, its regenerative
polymers, are only osteoconductive in nature
ability is limited.1 Some bone
and generally are not able to induce bone
A
defects undergo incomplete
formation.4 The combination of an osteo-
fracture healing (nonunion fractures), or the
conductive biomaterial with osteoinductive
defect size is beyond the body s healing
molecules overcomes the limitations of
capacity (critical size defects).1 3 In such
synthetic bone graft biomaterials. One such
cases, further intervention is required, typi-
hybrid material involves incorporating bone
cally of a surgical nature, to replace the
morphogenetic proteins (BMPs) into a bio-
damaged bone with a bone graft material.
degradable carrier. Once this hybrid material
Osteoinductive materials such as bone
is implanted, it will accelerate bone forma-
tion and regeneration in surrounding defec-
Advanced Drug Delivery Group, Faculty of Pharmacy,
University of Sydney, Sydney, Australia. tive bone tissues.
* Corresponding author, e-mail: ramin.rohanizadeh@
BMPs are part of the transforming growth
sydney.edu.au
DOI: 10.1563/AAID-JOI-D-10-00005 factor-b cytokine family.2 BMPs have a range
Journal of Oral Implantology 659
Hydroxyapatite as a Carrier for BMP
of functions, including fetal organ develop- structure for cell infiltration and ingrowth.
ment and postnatal kidney, neuronal, and Finally, the overall manipulation process that
bone development.5 Of greatest interest are includes loading BMP onto a carrier and its
its osteoinductive abilities, which allow BMP release must retain the biological activity of
signaling to induce postnatal bone forma- BMP.6,11,12 A range of BMP carriers of an
tion in nonskeletal sites.6,7 Of the many organic and an inorganic nature have been
isoforms, BMP-2 and -7 show the greatest investigated.2,13 These carriers have showed
in vivo potential for osteogenesis.8,9 They varying levels of success, but the major
recruit undifferentiated mesenchymal stem components of human bone collagen and
cells to the defective site that differentiate hydroxyapatite are preferred BMP carriers.2
into osteoblasts, which form new bone6,7; Hydroxyapatite (HA), Ca10(PO4)6(OH)2,
this is known as intramembranous bone has been used as a carrier for antibiotics,
formation.10 Another bone formation path- analgesics, and anticancer agents for the
way, endochondral formation, occurs when skeletal system.15,16 It is a preferred bioma-
chondroblasts (cartilage cells) are created terial in orthopedic and dental applications
from stem cells, modulating the bone shape; because of its chemically similar structure
they are then calcified and replaced by to the inorganic component of bone and
bone.10 Both pathways can be induced by other hard tissues.2,17 Although more soluble
BMPs and are highly dependent on the phases of calcium phosphate exist (eg,
properties of the BMP carrier and the isoform octacalcium phosphate, dicalcium phos-
of BMP.10,11 The main interest behind using phate dihydrate), HA has shown the greatest
BMP in dental and orthopedic applications is potential in bone tissue engineering.3 Syn-
its potential for reconstructing critical size thetic HA may be created through several
bone defects and nonunion fractures.3 To techniques and in a range of forms and
date, areas for which BMP have been studied shapes, such as powder, blocks, discs, and
include spinal fusion, long bone trauma, granules.12 HA is capable of directly bonding
ligament reconstruction, and orthopedic, to bone, and emerging evidence suggests
craniomaxillofacial, and periodontal dis- osteoinductivity with HA materials of certain
ease.3,12 Several BMP delivery systems are pore sizes and characteristics.18,19
available, ranging from the gene therapy BMP adsorption onto HA can enhance
approach to local implantation and systemic interfacial strength and contact between
administration.2 However, most research has the HA implant and surrounding bone,
focused on localized delivery systems, par- thereby promoting greater bone regenera-
ticularly in search of the ideal BMP carrier. tion around the implant than is seen with HA
The carrier for BMP should enhance the alone.3,17 On the other hand, incorporating
activity of the protein by maintaining a BMP into an HA porous carrier, even in small
certain BMP concentration in the defective doses, may increase the strength of the
area for a sufficient time to allow new bone porous carrier.20,21 BMPs adsorb on HA
formation.3,6,12 14 Good affinity should exist crystals via binding between functional
between BMP and its carrier to maintain sus- groups COO2, OH, and NH2 of BMP and
tained and prolonged BMP release. The carrier the calcium site of HA.22 One BMP molecule
should be easily sterilized and biodegradable may adsorb on HA, or up to 3 molecules may
and should have no immunogenicity.12,14 An adsorb cooperatively at once.22 It has been
ideal carrier not only is a mechanism for drug shown that the right side of BMP-2 is most
delivery but preferably should support bone prone to adsorption, with maximal adsorp-
growth by having an appropriate porous tion occurring in acidic conditions.23 Acidic
660 Vol. XXXVII/No. Six/2011
Rohanizadeh and Chung
conditions result in higher BMP/HA affinity study was to investigate the optimal method
because below the isoelectric point of BMP of loading BMP into an HA carrier. It has
(pH 7.9), BMP is positively charged and HA been shown that BMP-2 remains stable at
has a negative surface charge.23 temperatures up to 70uC,27 making plausible
BMP adsorption is thought to be limited to the BMP incorporation during HA wet
the macropores of HA, but there is a tendency precipitation synthesis. The present study
for a rapid initial burst release of BMP once a characterized and compared 3 methods of
BMP/HA composite is implanted in the loading BMP-2 into HA: (1) BMP incorpora-
body.23 25 Transient high BMP concentrations tion during HA formation in a calcifying
in the bone defect can lead to adverse effects solution, (2) immersion of HA powder in a
by activating osteoclast cells, which in turn BMP solution, and (3) BMP incorporation
resorb bone and promote inhibitory proteins
during dicalcium phosphate dihydrate
of BMP, thus limiting the osteogenic ability of
(DCPD) conversion to HA. Comparison of
the implant.24 Most studies have loaded BMP
these methods was based on the extent to
to HA via coating of the surface of HA materials
which BMP could be loaded into the HA
by BMP. This can be achieved by immersion of
carrier, the in vitro release profile of BMP
HA materials in a BMP solution because it can
from the carrier, and the influence of BMP on
be done under physiologic conditions, reduc- the physicochemical and morphologic prop-
ing the possibility of BMP denaturation. In
erties of HA crystals in the carrier.
contrast, incorporating BMP into HA materials
results in BMP adsorption onto both macro-
MATERIALS AND METHODS
pores and micro-pores of materials with
predominant attachment to calcium sites.
Recombinant human bone morphogenetic
Hence, BMP incorporation into HA is thought
protein-2 produced by Escherichia coli was
to prolong the retention profile because BMP
purchasedfromGenScript (Piscataway, NJ) asa
should be released as HA is dissolved.13 Liu
sterile lyophilized powder that was reconsti-
et al24 recently demonstrated that incorpora-
tuted in 5 mM acetic acid. Sodium hydroxide,
tion of BMP into biomimetic calcium phos-
acetic acid, calcium chloride, sodium phos-
phate revealed osteoinductive properties in
phate monobasic monohydrate, and sodium
rats and was fivefold more efficacious in
phosphate dibasic anhydrous were supplied
osteoinductivity compared with implants sur-
by MP Biomedicals (Solon, Ohio). Water used
face coated with BMP.24
throughout the project was purified by reverse
Many animal studies have been conduct-
osmosis (Milli-Q, Millipore, Billerica, Mass).
ed on HA/BMP hybrid materials. These trials
Optimization of HA synthesis for BMP
used HA materials immersed for a time in a
incorporation into HA powder
BMP solution and convincingly demonstrated
HA as a suitable carrier for BMP.3,7,17,25,26 A 10 mL solution of 2 M calcium chloride and
Studies involving BMP-2 with rats demon- a 5 mL solution of 1.2 M sodium phosphate
strated that the HA/BMP hybrid showed dibasic anhydrous were prepared to obtain
complete union bone repair and more a calcium (Ca)/phosphate (P) ratio of 1.67,
extensive bone formation compared with equal to that of stoichiometric HA. The
HA alone or autografts.3,7,25 calcium chloride solution was slowly added
Despite confirmation of the osteogenic to the phosphate solution while stirring
potential of BMP and the suitability of HA as using a magnetic stirrer. The pH of the
its carrier within animal trials, the clinical use solution was adjusted to 8, and the solu-
of BMP remains contentious. The aim of this tion was heated to one of the following
Journal of Oral Implantology 661
Hydroxyapatite as a Carrier for BMP
TABLE 1
Product of calcium and phosphate precipitation at different pH values and temperatures as
determined by XRD*
Temperature, uC
pH 37 40 43 45 47 50
7.5 DCPD + HA DCPD + HA HA HA HA HA
8HA HA HA
HA HA HA
*DCPD refers to dicalcium phosphate dehydrate; HA, hydroxyapatite.
temperatures: 37uC, 40uC, 43uC, 45uC, 47uC, conversion was allowed to proceed at the
and 50uC. During heating, the solution was following temperatures: 37uC, 40uC, 43uC,
covered to prevent evaporation. Stirring 45uC, 47uC, and 50uC in an oven with shaker.
continued for 2 hours to mature the After 24 hours, the solution was removed,
precipitated HA crystals. The mixture was filtered, and dried at 40uC to retain the
subsequently filtered quickly using a filter- powder. The entire process was repeated at
funnel and flask vacuum; this was followed a pH of 8.5 and 9 at the same temperatures.
by rinsing once with 25 mL of warmed (50uC) The reaction was also investigated at 50uC
water. The filter paper and the powder were
and at 60uC at pH 9, 9.5, 10, and 11. A 0.2 M
then removed from the funnel and were
solution of sodium hydroxide at 70uC was also
placed immediately in an oven at 40uCtobe
trialed. Similarly, the DCPD-to-HA conversion
dried overnight. The HA fabrication process
condition with the least risk of denaturing
was repeated at a pH of 7.5 over the same
BMP was selected for BMP incorporation.
temperature range and conditions. The
Methods to load BMP into HA powder
synthesis condition that resulted in precipi-
tation of crystalline HA and that had the
N Incorporation during HA precipitation: Based
lowest risk of causing BMP denaturation
on the results of optimization conditions for
(temperature and pH close to 37uC and 7.4,
HA synthesis (Table 1), 37uC and pH 8 were
respectively, ie, physiologic conditions) was
the selected parameters for this method.
chosen for BMP incorporation.
BMP-2 solution was added to 5 mL unheat-
Preparation of DCPD and optimization ed calcium chloride solution, which then
of DCPD-to-HA conversion for BMP
was slowly poured into the phosphate
incorporation into HA powder
solution (as described earlier) to obtain a
concentration of 120 mg/mL of BMP in the
Two equimolar (0.8 M) 15 mL solutions of
calcium chloride and sodium phosphate diba- 15 mL solution used for HA preparation. The
weight ratio between obtained HA and the
sic anhydrous were prepared. Calcium chloride
BMP available in the solution was approxi-
was added to the phosphate solution at room
mately 1000:1; the same ratio was used
temperature, and the pH was adjusted to 5.
across all other BMP loading methods.
The solution was covered and was left to be
stirred at room temperature for 2 hours before N Immersion of HA powder in a BMP solution:
filtering and rinsing with 25 mL of room The second method of loading BMP into
temperature water. The DCPD powder was HA involved immersing HA powder in a
dried overnight at room temperature. solution containing BMP. A total of 700 mg
Once dried, 1 g of the DCPD powder was of preformed HA was immersed at 37uC in
placed in 8.25 mL of 0.1 M sodium hydroxide a 5.8 mL solution containing 120 mg/mL
(dissolved as a buffered solution). The pH of BMP-2, yielding the same HA/BMP weight
the solution was adjusted to 8, and hydrolytic ratio and BMP concentration in solution as
662 Vol. XXXVII/No. Six/2011
Rohanizadeh and Chung
TABLE 2
Product of hydrolytic conversion of DCPD at different pH values and temperatures as
determined by XRD*
Temperature, uC
pH 37 40 43 45 47 50 60
8 DCPD DCPD DCPD DCPD DCPD DCPD 
8.5 DCPD DCPD DCPD DCPD DCPD DCPD 
9      DCPD 
9.5      DCPD 
10      DCPD DCPD
11      DCPD DCPD
* refers to a condition not studied by clinical trial; DCPD, dicalcium phosphate dihydrate.
the previous method. After 24 hours, the 14. The removed supernatant was replaced
solution was filtered and was rinsed with with 850 mL of fresh PBS at each time point.
25 mL warmed water (50uC). The powder The experiment was set up for all 3 BMP-
was allowed to dry overnight at 37uC. loading methods using 5 samples per method
N Incorporation during DCPD-to-HA conver- and time point. As a control, HA without BMP
sion: None of the conditions tested to was also incubated in PBS at 37uC with
convert DCPD to HA was successful (Table 2); identical time points for supernatant with-
therefore based on previous works of Le- drawal. All collected supernatant was stored at
Geros,28 0.2 M sodium hydroxide solution at 220uC until the protein concentration was
70uC was used for this method. A total of measured using 2 different protein assays.
120 mg/mL of BMP was added to the 8.25 mL The Lavapep total protein fluorescence
of 0.2 M sodium hydroxide solution (70uC). assay (Fluorotechnics, Sydney, Australia) and
The HA/BMP weight ratio and the BMP the Quantikine (R&D Systems Inc, Minneapo-
concentration in solution were adjusted as lis, Minn) BMP-2 enzyme-linked immunosor-
in the previously mentioned methods. The bent assay (ELISA) were used to determine
conditions of DCPD-to-HA conversion were total protein uptake into HA and its release
described earlier. profile in vitro. The microplate reader used for
both protein kits was the POLARstar OPTIMA
BMP release profile
in conjunction with FLUOstar OPTIMA soft-
The amount of BMP incorporated into the HA
ware (BMG Labtech, Ortenberg, Germany).
powder was determined by subtracting the
Because calcium ions released from HA
BMP concentration remaining in the calcifying
powder may interfere with the measurement
or immersion solution at the end of the
of BMP concentration in the total protein
loading process (before filtration) from the
assay, 2 g of HA (without BMP) was soaked in
initial protein concentration of 120 mg/mL. A
16.7 mL of water for 14 days. The supernatant
1.2 mM solution of sodium phosphate mono- was removed and was used in making the
basic monohydrate adjusted to pH 7.4 was the
BMP stock solution and in the serial dilutions
phosphate buffered solution (PBS) used for the
to make standards for the total protein assay.
in vitro BMP release assay. The rate of BMP
Physicochemical properties of HA (before
release was determined by incubating 120 mg
and after BMP loading)
BMP-loaded HA in 1.2 mL of PBS in an
Eppendorf tube at 37uC in a shaker oven. In X-ray diffraction (XRD) (Siemens D5000 X-ray
all, 850 mL of the supernatant was withdrawn diffractometer, Siemens Healthcare Diagnostics,
and frozen at 12 hours and at days 1, 3, 7, and Deerfield, Ill) was used to determine the crystal
Journal of Oral Implantology 663
Hydroxyapatite as a Carrier for BMP
FIGURE 1. The X-ray diffraction (XRD) spectra of (A) hydroxyapatite (HA) precipitates at pH 8 and 37uC; (B)
dicalcium phosphate dihydrate (DCPD) precipitates at pH 5 and room temperature; and (C) DCPD
converted to HA at pH 13 and 70uC.
structure of the obtained materials before and to 8, across all tested temperatures, the
after BMP loading. The parameters used were materials obtained had only HA crystal
40 kV and 30 mA, with a scan range between structure. Table 1 summarizes the products
10 and 40 degrees 2h and a step size of obtained from precipitation in calcium and
0.02 degrees with 2 seconds per step. Scanning phosphate solutions. For the second loading
electron microscopy (SEM) determined the method (DCPD-to-HA conversion), because
crystal morphology of the synthesized materials the conditions shown in Table 2 did not
and the effects of BMP incorporation on HA
elicit conversion of DCPD crystals to HA, the
crystal size and shape. The Zeiss Ultra Plus (Carl
DCPD powder was soaked in 0.2 M NaOH
Zeiss, G�ttingen, Germany) electron microscope
(pH 13) at 70uC to be hydrolyzed to HA.28
with a working distance of 4.4 mm and 20 kV
Physicochemical properties of HA (before
was usedinthis study.
and after BMP loading)
The typical XRD spectrum of HA precipitates
RESULTS
obtained from solution at pH 8 and 37uC is
Optimization of HA synthesis for
shown in Figure 1A. Diffraction peaks in this
BMP loading
spectrum are identified as those of HA. Major
peaks at 2h 5 26 degrees, 32 degrees, and
XRD analysis of powders obtained from the
calcium and phosphate precipitation reac- 34 degrees indicate HA lattice planes of
(002), (211), and (300), respectively. Fig-
tions demonstrated that a mixture of DCPD
and HA formed at pH 7.5 and at tempera- ure 1B and C shows the typical XRD spectra
tures of 37uC and 40uC (Table 1). At temper- of DCPD and HA converted from DCPD,
respectively. Figure 1B shows major peaks
atures above 43uC, only HA crystals were
precipitated at pH 7.5. By increasing the pH at 2h 5 11.7 degrees, 21 degrees, and
664 Vol. XXXVII/No. Six/2011
Rohanizadeh and Chung
FIGURE 2. The X-ray diffraction (XRD) pattern of crystals after bone morphogenetic protein (BMP) uptake
via (A) incorporation during hydroxyapatite (HA) precipitation; (B) HA immersion in BMP solution; and
(C) incorporation during BMP dicalcium phosphate dihydrate (DCPD)-to-HA conversion.
29 degrees, corresponding to DCPD lattice were larger, agglomerated, and more elongat-
planes of (020), (121), and (112) respectively. ed (Figure 3C). The micrograph of DCPD
Figure 1C identifies the typical HA peaks (Figure 3B) shows large (micron-sized) plate-
after DCPD-to-HA conversion; no DCPD like crystals. The micrographs in Figure 3D
peaks were observed after the conversion. through G reveal morphologic changes in HA
Following BMP incorporation via (1) crystals following BMP loading to HA powder.
incorporation during HA precipitation, (2) The BMP incorporated during HA precipita-
HA immersion in BMP solution, and (3) tion reduced the size of the crystals but
incorporation during DCPD conversion to retained the HA plate-like shape (Figure 3A vs
HA, the XRD spectra (Figure 2A through C) Figure 3D). After HA powder was immersed in
revealed that HA remained the only mineral the BMP solution, in most areas, HA crystals
compound present, with major peaks at 2h had a plate-like shape (Figure 3E); however, in
5 26 degrees and 32 degrees. The XRD some areas, crystals appeared to be very small
spectra showed a much higher level of noise and agglomerated compared with those
after BMP incorporation (Figure 2A through before BMP loading (Figure 3F: magnified of
C). No detectable alterations in the broad- boxed area in Figure 3E). BMP incorporated
ness of these peaks of the XRD spectra were during DCPD-to-HA conversion resulted in
observed following BMP loading. little variation in the morphology of HA
SEM micrographs in Figure 3A through G crystals compared with those formed without
show crystal morphology of different groups the presence of BMP (Figure 3C vs Figure 3G).
before and after BMP loading. The HA crystals
Release profile of BMP
synthesized from wet precipitation were nano-
sized and mostly had a plate-like shape The extent of BMP uptake into HA powder
(Figure 3A), whereas those formed from DCPD was measured using ELISA and total protein
conversion were more needle-shaped and assays (summarized in Table 3). Based on the
Journal of Oral Implantology 665
Hydroxyapatite as a Carrier for BMP
FIGURE 3. Scanning electron micrograph (SEM) images of (A) hydroxyapatite (HA) precipitates at pH 8
and 37uC; (B) dicalcium phosphate dihydrate (DCPD) precipitates at pH 5 and at room temperature; (C)
DCPD converted to HA at pH 13 and 70uC; (D) HA crystals after bone morphogenetic protein (BMP)
incorporation during HA precipitation; (E) HA crystals after immersion in a BMP solution; (F) Figure 3E
boxed area magnified (HA crystals); and (G) HA crystals after BMP incorporation during DCPD-to-
HA conversion.
666 Vol. XXXVII/No. Six/2011
Rohanizadeh and Chung
TABLE 3
Percentage of BMP uptake into HA measured by total protein and ELISA assays*
BMP Loading Incorporation During HA Immersion in Incorporation During
Method HA Precipitation BMP Solution DCPD Conversion to HA
Assay Total Protein ELISA Total Protein ELISA Total Protein ELISA
% BMP loaded 54 6 3.5 72 6 3.7 94.6 6 0.5Ą 96.2 6 0.5 50.5 6 2.4 96.7 6 0.5
*BMP refers to bone morphogenetic protein; DCPD, dicalcium phosphate dehydrate; ELISA, enzyme-
linked immunosorbent assay; HA, hydroxyapatite.
Ą
Significantly higher than the other 2 BMP loading methods (n 5 5; ANOVA P , .001; post hoc
Tukey pairwise).
total protein assay, the highest percentage and 619.3 6 139.8 ng/mL, respectively, for
of BMP uptake into HA was achieved with the HA precipitation, HA immersion, and DCPD-
immersion technique followed by incorpora- to-HA conversion loading methods. The
tion during HA precipitation, then DCPD-to- percentage of BMP released (based on the
HA conversion. When measured by total percentage of BMP uptake) after 14 days was
protein assay, BMP uptake during the HA 22.9 6 4.3, 0.43 6 0.18, and 1.23 6 0.28,
immersion method was significantly higher respectively, for each of the methods previ-
than that obtained using other loading ously mentioned (Figure 5). All 3 BMP-
methods (Table 3). The ELISA assay demon- loading methods revealed a rapid increase
strated significantly lower BMP uptake during in BMP release during the first 12 hours
HA precipitation and comparable percentage when compared with control. However, BMP
(around 96%) when BMP was loaded by released for samples prepared using the
immersion and the DCPD-to-HA conversion immersion method did not significantly
method. No significant differences were seen increase after 12 hours, unlike the 2 other
when the amount of BMP in solution was loading methods, which showed further BMP
measured before and after the filtration release after 12 hours. Loading BMP into HA
process, indicating that BMP adsorption on powder during DCPD conversion showed a
filter materials and funnel was negligible. secondary burst of BMP release at the 7 day
Using the ELISA assay, the release profiles time point with minimal increase thereon,
of BMP were determined for all 3 BMP- whereas for samples in which BMP was
loading methods and are shown in Figure 4. loaded during HA precipitation, the BMP
The total amount of BMP released after
14 days was 1028 6 181.6, 413.6 6 183.5,
FIGURE 5. Percentage of bone morphogenetic
FIGURE 4. Release profile of bone morphogenetic protein (BMP) released from hydroxyapatite (HA)
protein (BMP) from hydroxyapatite (HA) powders powders, based on the extent of BMP uptake for
prepared using 3 BMP loading methods. each respective BMP loading method.
Journal of Oral Implantology 667
Hydroxyapatite as a Carrier for BMP
release profile continued to increase over the reaction details the hydrolysis of DCPD to
14 day period. HA34:
10 CaHPO4 . 2 H2O?
DISCUSSION
Ca10�PO4�6�OH�2z18 H2Oz12 Hzz4PO3{
4
Precipitation of calcium phosphates from
A range of temperatures (60uC 140uC) and
homogenous solutions requires the satura- pH values (6 14) have been investigated
tion point to be reached, followed by
to facilitate this hydrolysis reaction.35 It was
subsequent nucleation and crystal growth.
seen that conditions at pH greater than 9
HA crystals are obtained in human physio- produced more elongated HA crystals. The
logic conditions, alkaline conditions, and
condition selected in this study 0.2 M
temperatures above 37uC; HA crystallinity is
NaOH (pH 13) and 70uC are of an extreme
increased by increasing the temperature in
nature but were used here because the
the wet precipitation method.29,30 Table 1
threshold for DCPD transformation was not
reflects these observations by showing that
determined (Table 2). At these conditions,
increasing the acidity and decreasing the
the HA produced showed different crystal
temperature led to the precipitation of a
morphology with elongated rod particles (as
mixture of DCPD and HA. However, at a pH
previously reported) but with reduced width
of 8, only HA and no DCPD was formed. The
in comparison with HA crystals obtained
XRD pattern (Figure 1A) showed that the HA
through direct wet precipitation (Figure 3C
obtained at a pH of 8 and at 37uC was poorly
vs Figure 3A).
crystalline, demonstrated by broad peaks in
BMP incorporation in HA has not been
the diffraction spectra. The significance of
specifically investigated in terms of the
poorly crystalline HA is correlated with its
physicochemical effects of BMP on HA.
increased solubility, which may facilitate
However, the presence of other bone matrix
BMP release once incorporated.29
proteins, bovine serum albumin, and amy-
SEM micrograph (Figure 3A) showed that lase has retarded HA crystal growth when
HA crystals were nano-sized (500 3 100 nm) present during precipitation of HA.35 38 SEM
plate shapes. The HA powder obtained had a images in Figure 3D through G show smaller
tendency to agglomerate, which has been HA crystals upon BMP uptake compared
previously reported in wet precipitation of with HA without BMP, supporting what has
HA.31,32 Smaller HA crystals are desirable for been previously found with other proteins.
their expected increased bioactivity property Incorporation of BMP during HA precipita-
and their higher surface area for protein tion (Figure 3D) reveals uniform rod needle-
adsorption.32 In contrast to HA, DCPD pre- like crystal morphology (approximately 200
cipitation from calcium and phosphate solu- 3 10 nm) compared with HA precipitated in
tions is predominant at a pH less than 6.5 the absence of BMP. The reduction in HA
and at room temperature.33 Figure 1B shows crystal size in BMP loaded samples might be
highly crystalline DCPD precipitates as seen due to binding of BMP to calcium sites in HA,
by sharp peaks in the XRD spectra when thereby inhibiting further crystal growth.
formed under acidic conditions (pH 5) and at Samples prepared via immersion of HA
room temperature. Figure 3B shows macro- powder in BMP solution (Figure 3E) revealed
sized (up to 10 mm 3 3 mm) plate-like crystals less uniform crystal morphology after BMP
typical of DCPD crystal morphology. In uptake compared with those prepared by
alkaline conditions, DCPD undergoes hydro- other loading methods. In this group, most
lytic transformation to HA. The following crystals have a plate-like shape with crystals
668 Vol. XXXVII/No. Six/2011
Rohanizadeh and Chung
of similar size to those obtained via BMP BMP adsorption to the containers used for
incorporation during HA precipitation. Figure 3F BMP loading was not accounted for.
(magnified boxed area in Figure 3E) shows an The amount of BMP uptake into HA via the
area of dissolution/reprecipitation of HA crystals 3 methods incorporation during HA precipi-
during the 24 hour immersion period in the tation, incorporation during DCPD-to-HA con-
BMP solution. Dissolution of HA crystals results version, and immersion of HA in BMP solu-
in a local increase in calcium and phosphate tion varied significantly based on the method
ions concentrations, leading to reprecipitation of measuring protein in solution used (ELISA vs
of very small HA crystals (50 3 100 nm) off the total protein assays). The total protein assay
primary HA crystals.38 These secondary crystals determines the state of BMP as total protein
showed different morphology and appeared (both denatured and undenatured protein
smaller and more agglomerated than the molecules), whereas the ELISA assay determines
primary HA crystals. the amount of BMP in the biologically active
Unlike the other BMP-loading methods, conformation (undenatured molecules). Table 3
BMP incorporated into HA powders during shows that the most extensive BMP loading
DCPD-to-HA conversion resulted in little occurred during HA immersion in the BMP
change in crystal morphology (Figure 3G). solution, demonstrated by both ELISA and total
The XRD spectra from all the methods in protein assays. This may have happened as a
the presence of BMP (Figure 2) show that result of the longer duration of BMP loading in
the XRD spectra with a much higher level of the immersion method (24 hours) compared
noise compared with those prepared with- with the 2 hour interval for BMP incorporation
out BMP (Figure 1) indicate the presence of during HA precipitation. Moreover, HA crystals
organic matter (such as BMP). The high level immersed in the BMP solution experienced a
of noise in XRD spectra made it difficult to secondary HA crystal growth, allowing for BMP
precisely determine the changes in peak adsorption onto the newly formed HA crystals,
broadening; therefore, for all 3 groups, no increasing the overall amount of BMP loaded.
detectable changes in HA crystallinity were Based on the total protein assay, BMP uptake
observed after BMP loading. during HA precipitation was the next highest
percentage. Unexpectedly, the DCPD-to-HA
BMP uptake
conversion loading method resulted in the
The amount of BMP uptake into HA powders lowest percentage, despite having a 24 hour
was determined via the protein assays seen loading interval. HA formed via hydrolysis of
in Table 3. However, assumptions underlined DCPD is known to be calcium deficient,34
the calculated percentage of BMP uptake. which may reduce the availability of BMP
Foremost is the BMP concentration in binding sites to calcium, thereby reducing
calcifying or immersion solutions, which the extent of BMP loading during DCPD-
was taken to imply that the remaining BMP to-HA. Moreover, the conversion reaction
was loaded into the HA powder. BMP involves DCPD dissolution and reprecipitation
adsorption to filter paper, by sampling of HA from calcium and phosphate ions.
before and after filtering, in each method However, precipitation of new HA crystals
was considered and revealed little adsorp- does not take the allotted 24 hour period,
tion of BMP onto filter materials. However, thereby reducing the time BMP had to bind to
the washing step in every method may have calcium sites. More important, the HA precip-
reduced the extent of the actual amount of itates obtained via DCPD conversion were
loaded BMP by removal of BMP, which was largest in comparison with those obtained by
loosely attached or incorporated. Finally, the 2 other tested BMP loading methods.
Journal of Oral Implantology 669
Hydroxyapatite as a Carrier for BMP
Therefore, the decrease in total surface area of unable to detect the very low concentration
HA crystals for BMP adsorption may be of BMP released from HA powder. BMP
another factor in reduced BMP uptake in the release from the samples prepared using
DCPD-to-HA conversion method. HA precipitation and DCPD conversion was
It should be noted that the total protein greater than in those prepared using HA
assay did not differentiate between dena- immersion in BMP solution. BMP uptake
tured and undenatured BMP molecules, during HA precipitation showed the best
whereas the ELISA assay detected BMP only release profile, with the most BMP (22.9%)
in the biologically active conformation state. released in conjunction with a sustained and
The ELISA assay is unable to measure prolonged release profile. The low release of
denatured BMP molecules in the remaining BMP (total of 0.43%) from samples prepared
calcifying or immersion solution; therefore a by immersion in BMP solution could be due
higher percentage of BMP was presumed to to BMP detachment during washing caused
be loaded into HA when the ELISA assay was by loose adsorption of BMP on HA crystals.
used, which is inaccurate. For this reason, BMP uptake during DCPD-to-HA conversion
across all 3 BMP-loading methods, the showed a biphasic release profile with 2
percentages of BMP uptake were elevated stages of burst release, within the first
when measured by ELISA assay rather than 12 hours and 7 days. For this group, the
the total protein assay (Table 3). Because the BMP release profile may indicate BMP initially
total protein assay accounted for denatured released from the HA crystals located on the
and active BMP, it gave a more accurate surface of HA agglomerates; after a week,
depiction of BMP uptake into HA powder. BMP was released from crystals located in
the inner regions of HA agglomerates.
The ELISA assay showed that BMP uptake
However, the release profiles for all loading
for the DCPD-to-HA conversion method was
methods showed an initial burst of BMP
significantly greater than the respective value
release within the first 12 hours. It should be
obtained using the total protein assay. This
noted that these release profiles reflect only
may be explained by the conditions used to
the amount of undenatured protein (detect-
load BMP during DCPD-to-HA conversion. It
ed by ELISA assay); the actual amount of
has been reported that after 8 hours of heat
total released BMP (both denatured and
treatment (70uC), the activity of BMP was
undenatured) is probably higher.
significantly decreased.27 Conversion of DCPD
powder to HA occurred at 70uC over 24 hours, Based on results obtained in this study,
increasing the percentage of denatured BMP further optimization of BMP loading during
molecules and subsequently reducing their the HA precipitation method should be
detection in remaining solution via the ELISA considered. Enhanced understanding of
assay. This contrasts with BMP loading during BMP in HA adsorption would develop with
HA precipitation at 37uC and pH 8, wherein variation of the stage in which BMP can be
the secondary structure of BMP is better added during the HA precipitation process
maintained, allowing more accurate BMP (eg, in the calcium solution vs in the
measurement in solution with ELISA and phosphate solution vs during HA crystal
thereby BMP uptake in HA powder. maturation). Upon optimization, the biolog-
ical osteogenic potential of these implants
BMP release profile
would have to be evaluated through in vivo
Unlike the ELISA assay, which is capable of trials. The addition of other bone augment-
BMP-2 detection in the range of picogram, ing drugs may be considered with HA as a
the total protein assay used in this study was potential carrier.
670 Vol. XXXVII/No. Six/2011
Rohanizadeh and Chung
ACKNOWLEDGMENT
Overall, these findings show the suitabil-
ity of HA as a BMP carrier. This study
This work was supported by a research grant
highlights the necessity of investigating the
from American Academy of Implant Dentistry
protein loading method because it directly
Research Foundation.
affects many properties of the carrier and its
ability to take up and release protein.
REFERENCES
1. Bringhurst FR, Demay MB, Krane SM, Kronenberg
CONCLUSION
HM. Bone and mineral metabolism in health and
disease. In: Fauci A, Braunwald E, Kasper D, et al, eds.
The potential for BMP in bone tissue Harrison s Principles of Internal Medicine. 17th ed. New
York, NY: McGraw-Hill; 2007.
engineering remains limited so long as its
2. Kirker-Head CA. Potential applications and de-
delivery is hampered. This study has dem- livery strategies for bone morphogenetic proteins. Adv
Drug Delivery Rev. 2000;43:65 92.
onstrated different methods of combining
3. Alam MI, Asahina I, Ohmamiuda K, Takahashi K,
BMP, and that a single carrier, HA, can
Yokota S, Enomoto S. Evaluation of ceramics composed of
different hydroxyapatite to tricalcium phosphate ratios as
produce great variation in protein uptake
carriers for rhBMP-2. Biomaterials. 2001;22:1643 1651.
and release, also affecting carrier physico-
4. Vaccaro AR. The role of the osteoconductive
scaffold in synthetic bone graft. Orthopedics. 2002;25:
chemical and morphologic properties. Based
S571 S578.
on parameters for HA synthesis closest to
5. Chen D, Zhao M, Mundy GR. Bone morphoge-
physiologic conditions, BMP incorporation netic proteins. Growth Factors. 2004;22:233 241.
6. Winn SR, Uludag H, Hollinger JO. Sustained release
reduced HA crystal size during HA precipita-
emphasizing recombinant human bone morphogenetic
tion (pH 8, 37uC) and the immersion method.
protein-2. Adv Drug Delivery Rev. 1998;31:303 318.
7. Noshi T, Yoshikawa T, Ikeuchi M, et al. Enhance-
However, protein loading during DCPD
ment of the in vivo osteogenic potential of marrow/
conversion to HA (pH 13, 70uC) showed no
hydroxyapatite composites by bovine bone morpho-
genetic protein. J Biomed Mater Res. 2000;52:621 630.
significant alteration in HA crystals.
8. Groeneveld EHJ, Burger EH. Bone morphogenetic
The negligible amount of BMP released
proteins in human bone regeneration. Eur J Endocrinol.
2000;142:9 21.
through the immersion method reflects the
9. Zhao JY, Shinkai M, Takezawa T, Ohba S, Chung
need for future studies to optimize methods
U, Nagamune T. Bone regeneration using collagen type
of BMP loading, because this is the predom- I vitrigel with bone morphogenetic protein-2. J Biosci
Bioeng. 2009;107:318 323.
inant BMP loading method used in past and
10. Zou XN, Zou LJ, Foldager C, Bendtsen M, Feng
current trials. BMP incorporation during HA
WZ, B�nger CE. Different mechanisms of spinal fusion
using equine bone protein extract, rhBMP-2 and
precipitation produced the best release
autograft during the process of anterior lumbar
profile, despite not acquiring the highest
interbody fusion. Biomaterials. 2009;30:991 1004.
11. Tsuruga E, Takita H, Itoh H, Wakisaka Y, Kuboki Y.
amount of BMP loaded. Its BMP release
Pore size of porous hydroxyapatite as the cell-substratum
profile could be described as slow, sustained,
controls BMP-induced osteogenesis. J Biochem. 1997;121:
317 324.
and prolonged, as desired in BMP clinical
12. Bessa PC, Casal M, Reis RL. Bone morphoge-
applications.
netic proteins in tissue engineering: the road from
laboratory to clinic, part II (BMP delivery). J Tissue Eng
Regen Med. 2008;2:81 96.
13. Seeherman H, Wozney JM. Delivery of bone
ABBREVIATIONS
morphogenetic proteins for orthopedic tissue regener-
ation. Cytokine Growth Factor Rev. 2005;16:329 345.
BMP: bone morphogenetic protein
14. Friess W, Uludag H, Foskett S, Biron R. Bone
DCPD: dicalcium phosphate dihydrate
regeneration with recombinant human bone morpho-
genetic protein-2 (rhBMP-2) using absorbable collagen
HA: hydroxyapatite
sponges (ACS): influence of processing on ACS
ELISA: enzyme-linked immunosorbent assay
characteristics and formulation. Pharm Dev Technol.
1999;4:387 396.
PBS: phosphate buffered solution
15. Thamaraiselvi TV, Prabakaran K, Rajeswari S.
SEM: scanning electron microscopy
Synthesis of hydroxyapatite that mimic bone mineral-
XRD: X-ray diffraction ogy. Trends Biomater Artif Organs. 2006;19:81 83.
Journal of Oral Implantology 671
Hydroxyapatite as a Carrier for BMP
16. Kamitakahara M, Ohtsuki C, Miyazaki T. Review 27. Ohta H, Wakitani S, Tensho K, et al. The effects
paper: behavior of ceramic biomaterials derived from of heat on the biological activity of recombinant
tricalcium phosphate in physiological condition. J human bone morphogenetic protein-2. J Bone Miner
Biomed Mater Res. 2008;23:197 212. Metab. 2005;23:420 425.
17. Freilich M, Patel CM, Wei M, et al. Growth of 28. LeGeros RZ. Calcium phosphates in oral biology
new bone guided by implants in a murine calvarial and medicine. In: Myers H, ed. Monographs in Oral
model. Bone. 2008;43:781 788. Medicine. Vol 15. Karger: San Francisco, Calif; 1991.
18. Lin LW, Chow K, Leng Y. Study of hydroxyap- 29. Matsumoto T, Okazaki M, Inoue M, et al.
atite osteoinductivity with an osteogenic differentiation Hydroxyapatite particles as a controlled release carrier
of mesenchymal stem cells. J Biomed Mater Res. of protein. Biomaterials. 2004;25:3807 3812.
2009;89A:326 335. 30. Prado Da Silva MH, Lima JHC, Soares GA, et al.
19. Kikawa T, Kashimoto O, Imaizumi H, Kokubun S, Transformation of monetite to hydroxyapatite in bioac-
Suzuki O. Intramembranous bone tissue response to tive coatings on titanium. Surf Coat Technol. 2001;137:
biodegradable octacalcium phosphate implant. Acta 270 276.
Biomater. 2009;5:1756 1766. 31. Saeri MR, Afshar A, Ghorbani M, Ehsani N,
20. Ono I, Tateshita T, Inoue M, Kuboki Y. In vivo Sorrell CC. The wet precipitation process of hydroxy-
strength enhancement of hydroxyapatite combined apatite. Mater Lett. 2003;57:4064 4069.
with rhBMP-2. J Bone Miner Metab. 1998;16:81 87. 32. Mobasherpour I, Heshajin MS, Kazemzadeh A,
21. Liu Y, De Groot K, Hunziker EB. BMP-2 liberated Zakeri M. Synthesis of nanocrystalline hydroxyapatite
from biomimetic implant coatings induces and sustains by using precipitation method. J Alloys Comp. 2007;430:
direct ossification in an ectopic rat model. Bone. 330 333.
2005;36:745 757. 33. Arifuzzaman SM, Rohani S. Experimental study
22. Dong XL, Wang Q, Wu T, Pan HH. Understand- of brushite precipitation. J Cryst Growth. 2004;267:
ing adsorption-desorption dynamics of BMP-2 on 624 634.
hydroxyapatite (001) surface. Biophys J. 2007; 93: 34. Shih WJ, Chen YF, Wang MC, Hon MH. Crystal
750 759. growth and morphology of the nano-sized hydroxyap-
23. Boix T, Gomez-Morales J, Torrent-Burgues J, atite powders synthesized from CaHPO4.2H2O and
Monfort A, Puigdomenech P, Rodriguez-Clemente R. CaCO3 by hydrolysis method. J Cryst Growth. 2004;270:
Adsorption of recombinant human bone morphoge- 211 218.
netic protein rhBMP-2m onto hydroxyapatite. J Inorg 35. Stulajterov� R, Medveckż L. Effect of calcium
Biochem. 2005;99:1043 1050. ions on transformation brushite to hydroxyapatite in
24. Liu Y, Huse RO, De Groot K, Buser D, Hunziker aqueous solutions. Colloid Surf A. 2008;316:104 109.
EB. Delivery mode and efficacy of BMP-2 in association 36. Leonor IB, Alves CM, Azevedo HS, Reis RL.
with implants. J Dent Res. 2007;86:84 89. Effects of protein incorporation on calcium phosphate
25. Boden SD, Martin GJ, Morone MA, Ugbo JL, coating. Mater Sci Eng C. 2009;29:913 918.
Moskovitz PA. Posterolateral lumbar intertransverse 37. Azevedo HS, Leonor IB, Alves CM, Reis RL.
process spine arthrodesis with recombinant human Incorporation of proteins and enzymes at different
bone morphogenetic protein 2/hydroxyapatite-trical- stages of the preparation of calcium phosphate
cium phosphate after laminectomy in the nonhuman coatings on a degradable substrate by a biomimetic
primate. Spine. 1999;24:1179 1185. methodology. Mater Sci Eng C. 2005;25:169 179.
26. Sachse A, Wagner A, Keller M, et al. Osteointe- 38. Rohanizadeh R, Padrines M, Bouler JM, Cou-
gration of hydroxyapatite-titanium implants coated chourel D, Fortun Y, Daculsi G. Apatite precipitation
with nonglycosylated recombinant human bone mor- after incubation of biphasic calcium-phosphate ceramic
phogenetic protein-2 (BMP-2) in aged sheep. Bone. in various solutions: influence of seed species and
2005;37:699 710. proteins. J Biomed Mater Res. 1988;42:530 539.
672 Vol. XXXVII/No. Six/2011