Article

In vitro biocompatibility of novel

biphasic calcium phosphate-mullite
composites

Journal of Biomaterials Applications

27(5) 497509
! The Author(s) 2011
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0885328211412206
jba.sagepub.com

Shekhar Nath, Sushma Kalmodia and Bikramjit Basu

Abstract
In designing new calcium phosphate (CaP)-based composites, the improvement in physical properties (strength, toughness) without compromising the biocompatibility aspect is essential. In a recent study, it has been demonstrated that
significant improvement in compressive strength as well as modest enhancement in toughness is achievable in biphasic
calcium phosphate (BCP)-based composites with mullite addition (up to 30 wt%). Herein, we report the results of the
in vitro cell adhesion, cell proliferation, alkaline phosphatase (ALP) activity, and osteocalcin (OC) production for a series
of BCP-mullite (up to 30 wt%) composites. Mouse fibroblast (L929) cell lines were used to examine in vitro cell adhesion
and cell proliferation; while osteoblast-like (osteosarcoma, MG63) cells were used for in vitro osteoblastic function study
by ALP and OC expression. Much emphasis has been provided to discuss the cell viability and proliferation as well as
osteoblastic differentiation marker on the investigated biocomposites in relation to the characteristics of the phase
assemblage. On the basis of various observations using multiple biochemical assays, it has been suggested that BCPmullite composites would be a candidate material for orthopedic applications.
Keywords
BCP-mullite, composite, cell adhesion, MTT, ALP, osteocalcin

Introduction
In the search of ideal bioactive bone implant materials,
substantial eorts have been invested to develop
hydroxyapatite (HA) or CaP-based composites with
various reinforcements.14 As a synthetic analogue of
calcied tissues of vertebrate, HA is a candidate material for bone implant applications. Also, bone matrix
can directly bind to HA, which is a necessary prerequisite for implant osseointegration.5 However, its low
strength and fracture toughness have reduced the eld
of possible applications only to those, where the
implant will be subjected to very low stress.68
Another possible application of HA could be a coating
on metallic implants. However, recent clinical studies9,10 revealed that HA coating did not produce any
benet, when implanted for long period. Keeping these
information in mind, it is therefore needed to improve
the performance of CaP (HA, tricalcium phosphate
TCP]), which could be used as bulk. In search of such
materials, researchers used mixture of HA-TCP materials along with second phase reinforcement to develop

composites with desirable properties. To this end, the

physical property enhancement without any compromise on biocompatibility aspect necessitates the use of
optimal amount of reinforcement as well as tailoring
the processing parameters.
For this purpose, in the present study, dierent
amounts (1030 wt%) of mullite (3Al2O3.2SiO2) were
mixed with HA and the powder mixtures were sintered
under various optimal conditions. Mullite is a solid
solution of alumina (Al2O3) and silica (SiO2). Mullite
is chemically inert and mainly used as refractory material in high temperature furnaces.11 It has lower density

Laboratory for Biomaterials, Department Materials Science and

Engineering, Indian Institute of Technology Kanpur, Kanpur 208016, UP,
India
Corresponding author:
Bikramjit Basu, Laboratory for Biomaterials, Department of Materials
Science and Engineering, Indian Institute of Technology Kanpur, Kanpur
208016, UP, India.
Email: bikram@iitk.ac.in

498
(3.05 g/cc) than Al2O3 (3.95 g/cc) and ZrO2 (6.1 g/
cc). It has good combination of structural properties,
like high hardness of 15 GPa, high elastic modulus of
240 GPa, and moderate fracture toughness of
3 MPa m0.5.12 When mullite was mixed (up to
30 wt%) with HA, very good combinations of mechanical properties were measured (E-modulus: 70 GPa,
hardness: 45 GPa, and fracture toughness: 1.5 MPa
m0.5).13 This toughness value was 2.5 times higher
than that of pure monolithic HA (0.6 MPa m0.5,
measured by SEVNB technique). The compressive
strength of the developed composites was 230 MPa,
which is by far better than pure HA (50 MPa).13
From microstructural phase assemblage point of
view, these composites contain a mixture of HA and
TCP and therefore they are called as biphasic calcium
phosphate (BCP)-mullite composites.1416 Hence, it is
likely that such composite could be an excellent alternative of HA, provided that the composite shows
required biocompatibility.
In this backdrop, an important part of research is to
characterize the in vitro properties of new biomaterials.
Despite various conicts in results of in vitro and in vivo
tests, the in vitro assays are considered the primary biocompatibility screening tests for a wide variety of
implant materials.17 For example, the response of osteoblastic cells to a thin lm of poorly crystalline calcium
phosphate apatite (PCA) crystals was examined in vitro.
The osteoblasts were reported to exhibit high cellular
activity, such as adhesion, proliferation etc. In fact, the
cells were attached more rapidly to PCA thin lm than
to reference dishes.18 In another study, HA was used as
an additive for zirconia-alumina nanocomposite. The
addition of HA was reported to increase the biocompatibility of the nanocomposites signicantly, as evident from the results of the in vitro tests using MG63
osteoblast-like cells.5 In some cases, the HA-based
composite materials showed better in vitro biocompatibility than pure HA. For example, Boanini et al.19
studied the interaction of osteoblast-like cells on nanocomposite of HA with aspartic acid and glutamic acid.
Their results revealed that the nanocomposite of
HA possessed better cell proliferation, alkaline phosphatase (ALP) activity, and osteocalcin (OC) gene
expression than pure HA. Shu et al.20 studied the role
of HA on the dierentiation and growth of MC3T3-E1
osteoblasts cells. Their results indicated that HA
enhanced osteoblast dierentiation while suppressing
cell growth. In contrast, Licht et al.21 reported that
due to the phagocytosis of HA particles, osteoblast
exhibited reduced cell growth and ALP activity.
In another study, Ogata et al.22 compared the osteoblast response to HA with HA/soluble CaP (SCaP)
composites. Their results revealed that HA/CaP

Journal of Biomaterials Applications 27(5)

showed greater ability in osteogenesis than HA by
increasing collagen synthesis and calcication of the
extracellular matrix.
Reviewing literature, it is evident that HA, TCP, and
other CaP phases are biocompatible. However, there is
no published data available on the biocompatibility of
mullite ceramics. Therefore, the biocompatibility of
BCP-mullite composite materials is needed to be examined before any biomedical application can be proposed. In this paper, we report the cell adhesion,
proliferation, and dierentiation behavior using
mouse broblast (L929) as well as MG63 cell lines.
The sintered HA was used as baseline material in all
the in vitro experiments.

Materials and experimental procedures

Synthesis of materials
HA powder was synthesized in-house using commercially available chemicals, such as calcium oxide
(CaO) and phosphoric acid (H3PO4), following a wellestablished suspensionprecipitation route.23,24 Phase
pure mullite (3Al2O3.2SiO2) powder was procured commercially (KCM Corporation, Japan). As a rst step of
sample preparation, the mixing of HA and mullite powders (1030 wt% mullite) was carried out in a ball mill
for 16 h. Following this, the powder was subsequently
pressed to obtain pellets of 5 mm diameter. The sintering temperature was optimized based on the densication and mechanical properties results. The sintering of
the composite pellets was carried out at 1350 C for 2 h,
while sintering of pure HA was performed at 1200 C
for 2 h, both in conventional pressureless sintering furnace. After the sintering, the diameter of the sample
was around 4 mm. The sintered composites are designated by their initial mullite content (BCP  M means
BCP x wt% mullite and likewise), irrespective of the
phases present in the sintered materials. Therefore,
throughout the text, such designation is followed for
the composites.

Material characterization
X-ray diraction (Rich-Seifert, 2000 D) patterns were
acquired from the sintered ceramics to identify the different phases present. Based on the X-ray peak intensity of the characteristic phases, the qualitative presence
of various phases is reported in this paper. Scanning
electron microscopy (SEM; model JSM-6330 F,
Philips, The Netherlands) was performed on the polished and etched surface of the sintered composite to
investigate various microstructural features.

Nath et al.
Surface topography is an important parameter
which, inuences the cell adhesion and, in general, the
biocompatibility property of materials. In order to
characterize the ner scale topography of the investigated materials, smoothly polished and thermally
etched surfaces were observed under atomic force
microscope (AFM; Molecular Imaging, Pico-SPM I,
USA) using contact mode. A Si3N4 cantilever with a
three-sided pyramidal single crystal Si3N4 tip with apex
angle of 20 , a tip radius of curvature of 10 nm, and a
normal stiness of  0.6 N/m were employed.

Cell culture experiment. L929 and MG63 cell lines

obtained from CCMB, Hyderabad (India), and
ATCC (USA), respectably were preserved in an LN2
container. Prior to seeding the cells on biomaterials
surfaces, the cells were revived. Following this, cells
were cultured in Dulbeccos modied Eagles medium
(DMEM; Sigma Aldrich), supplemented with 10% fetal
bovine serum (Sigma Aldrich) and 1% penicillin/streptomycin cocktail (Sigma Aldrich). The culture plate
with the cell lines (gelatin-coated) was incubated for
further proliferation and growth in a CO2 incubator
(Thermo, USA) operated under conditions of 5%
CO2, 90% humidity, and 37 C temperature. The
medium was being replaced every 2 d interval and the
numbers of cells in conuent monolayer was approximately 5  105/mL, in 35-mm culture plate. The conuent monolayer was detached from the culture plate
using 0.50% trypsin and 0.20% EDTA solution
(Sigma Aldrich).

Cell adhesion test. As described in the previous subsection (section Cell culture experiment), L929 cells
were cultured. The samples used for cell adhesion
experiment were pure HA, BCP10M, BCP20M,
BCP30M, and control glass disc. All samples were sterilized in steam autoclave (121 C, 15 lb pressure for
15 min) and, subsequently, the cells were seeded on
the samples at approximate density of 5  105/mL. In
a separate experiment, the cell seeding density of L929
was reduced to 1  105/mL and the eect of cell seeding
density on cell adhesion was studied. The seeded test
samples were incubated in a CO2 incubator with the
standard culture condition, 5% CO2, 37 C temperature, and 90% humidity. The culture medium was
being aspirated after 2-d interval and fresh culture
medium was being added into each wells. After the
stipulated time period (1 and 3 d), the samples were
washed twice with phosphate buer saline (PBS;
1  PBS, pH 7.4) and then xed by 2% glutaraldehyde
in PBS. The cells, adhered on the material surfaces,
were dehydrated using a series of ethanol solutions
(30%, 50%, 70%, 95%, 100%) for 10 min twice and

499
then further dried using critical point drier (CPD;
Quramtech, UK). The dried samples were sputtercoated (Vacuum Tech, Bangalore, India) with gold
and examined under SEM. For AFM observation, the
dried samples were directly used, without gold coating.
The experiment was repeated for at least three times
and the representative results are presented in the present article.

MTT assay
L929 and MG63 cells were cultured following the previously described cell culture method. The cell proliferation was investigated on pure HA, BCP-mullite
composites (4 mm diameter, 4 mm height), and control
glass disc. At rst, autoclaved samples were placed in
the 4-well plate and then washed with PBS. Following
this, 5  104 cells/mL were seeded on each sample.
Subsequently, the culture plate was incubated for 2 d
in the CO2 incubator. After the incubation period, the
medium was aspirated and samples were washed
twice with PBS. Then, 200 mL of fresh DMEM was
added (without phenol red) into each well followed
by 10 mL reconstitute MTT (3(4,5-dimethylthiazol-2
yl)-2-5 diphenyltetrazolium bromide: SIGMA, USA,
Cat No.M5655; 5 mg/mL in DMEM culture medium
and serum) per 100 mL of DMEM was added in
each well and the plate was incubated for 68 h. In
the meantime, the culture plate was viewed under the
phase contrast microscope (Nikon, Eclipse 80i, Japan)
to check for the formation of purple formazane
crystal. After the incubation, samples were removed
from the wells and placed in a new culture plate.
Thereafter, 200 mL of dimethyl sulfoxide (DMSO)
was added into each well, including control. The optical density of the solution was measured at 540 nm
using ELISA automated microplate reader (Bio-Tek,
EL  800).

ALP activities
ALP is a widely recognized biochemical marker to
assess osteoblast activity on biomaterial substrate. It
is known that ALP plays a major role in skeletal mineralization. This enzyme is bound to the membrane of
osteoblasts and functions to enhance osteogenesis by
degrading pyrophosphates. For the ALP assay, the
autoclaved samples were placed in the 4-well plates
and MG63 cells were seeded approximately at a density
of 5  104 cells/mL. Three replicates of each composition were selected for this experiment. The culture protocol and conditions were similar to that described in
the earlier section. After 2 d of culture, ascorbic
acid and vitamin D3 were added to activate the

500
phenotypic expression of dierentiated MG63 cell line.
The concentration of 1,25-dihydroxy vitamin D3 solution has signicant eect on the expression of dierentiated MG63 cells. In an important study, Robert and
Gideon25 reported that the concentration of 10 810 7
M, 1,25-dihydroxy vitamin D3 inhibited growth and
elevated ALP as well as total cell protein. However,
lower D3 concentrations (10 10 and 10 9 M) reduced
ALP activity. Therefore, in the present study, the concentration of vitamin D3 was kept at 10 810 7 M for
better ALP activity.
The ALP activity experiments were conducted after
3 rd and 7th d of culture. When the cells were lysed
followed by centrifuged. Therefore, the supernatant
was used as the sample for the ALP activity. The subsequent steps were followed as per the commercial kit
protocol (ALP, code no. 25904, Span Diagonstics Ltd.,
Surat, India) in a 96-well culture plate. At the last step
of the experiment, optical density was measured by
ELISA reader at 405 nm.

Osteocalcin. It is known that OC, the most abundant

noncollageneous protein of the bone extracellular
matrix, is biologically synthesized by osteoblast-like
cells. In the present study, MG63 cells were cultured
for the OC assay, following the method described in
an earlier section. After subconuent, cells were trypsinized and seeded on sterilized pure HA and BCP-mullite (all samples were of 4 mm diameter) in a 4 well
culture plate at a density of 5  104 cells/mL. After
day 2 and 5, culture medium was replaced by dierentiating medium that contain vitamin D3 (10 8 M nal
concentration), ascorbic acid (50 mg/mL nal), and
glycerol phosphate (108 mg/mL) to enhance osteogenic
activity. OC production was tested at 7th d of culture and the sample preparation was similar to ALP
assay. In brief, three strips of coated wells were taken
from the strips supplied with the OC kit (HOSTEASIA KAP1381, Biosource Europe S.A., Belgium).
Thereafter, 25 mL of each calibrator (human serum
with protease inhibitors and benzamidin: supplied
with kit), control (human serum with protease inhibitors, benzamidin and thymol: supplied with OC kit)
and samples (Pure HA, BCP10M, BCP20M, and
BCP30M) were transferred into the appropriate wells
and subsequently 100 mL of anti-OST-HRP conjugate
was added to all the wells, followed by 2 h shaking.
Following this, the liquid from each well was aspirated
and washed by 400 mL of wash solution. Then, the solution from each well was aspirated and 100 mL of chromogenic solution was added into each well. Thereafter,
the well plate was incubated in a shaker at room temperature. Finally, the optical density was measured at
450 nm.

Journal of Biomaterials Applications 27(5)

Results
Phase assemblages and microstructure. In Table 1,
the quantitative analysis of XRD results, obtained from
the polished surface of the investigated materials, are
presented. Also, a typical microstructure of BCP30M
composite is provided in Figure 1. The larger grain
matrix was identied as TCP/HA phase and the grain
boundary phases were mullite, alumina, CaO, and gehlenite. More details of the microstructural characterization can be found elsewhere.15,16 From Table 1 it
should be clear that all the HA-based composites
after sintering predominantly contained TCP. In contrast, single-phase HA without any sign of dissociation
to any TCP polymorph could be obtained as baseline
HA. The addition of mullite as well as higher sintering
temperature, in combination, caused dissociation of
HA to TCP phases and it is clear that all the mullitecontaining samples were essentially BCP composites.
In Fig. 2, AFM images show some representative
surface topography (three-dimensional view). Figure
2(a) shows the equiaxed grains of HA of variable
sizes with good grain boundary adhesion. Figure 2(b)
shows the AFM image captured from BCP10M sample,
sintered at 1350 C for 2 h. Here, the ner grains of
reaction products were present at the boundary region
of larger HA/TCP grains. Similarly, Figure 2(c) shows
a coarse CaP-grain surrounded by the reaction products, that is, calciumalumino silicate phase. Figure
2(d) presents a high magnication AFM image of
BCP30M sample and the grain boundary area was
almost covered by the reaction products. In all the
investigated samples, the grain boundary phases
appeared to be well bonded with the matrix phase,
that is, HA/TCP. Also, the crystals of grain boundary
phases were much smaller (less than 300 nm) compared
to that of the matrix phase.

Cell adhesion. Figure 3 shows the cell adhesion behavior of BCP20M (Figure 3(a) and (b)), BCP30M
(Figure 3(c) and (d)), and control (Figure 3(e) and
(f)) samples, after 1 d of culture. Here, the cell seeding
density was 5  105 cells/mL. In all cases, cell division,
proliferation and cellcell contacts were evident.
Clearly, as far as the cell adhesion was concerned, various composites, despite compositional dierence/phase
assemblage, exhibited comparable behavior. Also, the
cell adhesion for investigated ceramics was comparable
with that of control sample. Beside major observation
of cell spreading, a transient signal of cell migration
was also observed through structural changes in the
form of cell protrusion.
More detailed observations of cellmaterial interaction are made with SEM and AFM. Figure 4 shows the
results of similar cell adhesion experiments, but the

Nath et al.

501

Table 1. The starting powder composition and sintering conditions are mentioned as well as the sample designation of each sample
used in the present investigation.
Sample designation

Starting composition

Pressureless sintering
conditions

HA
BCP10M

Pure HA
HA with 10 wt% mullite

1200 C, 2 h
1350 C, 2 h

BCP20M

HA with 20 wt% mullite

1350 C, 2 h

BCP30M

HA with 30 wt% mullite

1350 C, 2 h

Phase assemblages (after sintering)

HA-ss
a-TCP-ss, HA-ss, b-TCP-w, mullite-ww,
CaO-ww
a-TCP-s, b-TCP-ss, HA-ww, mullite-w,
gehlenite-ww, CaO-ww, alumina-ww
b-TCP-ss, HA-ww, mullite-s, gehleniteww, CaO-ww, alumina-ww

The phases present in the sintered ceramics are summarized based on the XRD peak intensities.
Ss: very strong, s: strong, ww: very weak, w: weak.

longer distance than that was observed in the previous

case (when cell seeding density was 5  105/mL) on the
materials surface. Figure 4(i) shows an AFM image of a
lapodia attached on pure HA sample. The branching
of lapodia is clearly visible. This type of branching
provided good anchorage with the material surface.

MTT assay. It is known that MTT reagent directly

Figure 1. Representative microstructure of thermally etched

BCP30M ceramic. Note the uniform presence of grain boundary
phases around each grain.

culture time was more (3 d) and the cell seeding density

was approximately 1  105/mL. SEM images reveal
adhesion of L929 cells on BCP10M (Figure 4(c) and
(d)), BCP20M (Figure 4(e) and (f)), BCP30M
(Figure 4(g) and (h)) samples. The cell adhesion on
baseline HA sample can be seen in Figure 4(a) and
(b). From the images, some interesting observations
can be made, for example, Figure 4(e) reveals how the
cells were connected to each other by lapodia extension on BCP20M sample. In this case, some observable
dierences could be found, when the cell seeding density was lower. The cells were larger in size with an
approximate dimension of 4050 mm. The major dierence could be found in cell morphology. In case of cell
density being 1  105/mL, most of the cells were spread
on the surface and intended to enhance the cell-material
contacts. The lopodia of the cells were extended over

reacts with the mitochondria (mitochondrial dehydrogenase) of living cells. Therefore, the reduction of MTT
will be more if more numbers of metabolically active
cells are present. In this context, MTT is widely
regarded as one of the quantitative assays to determine
the cytotoxicity of the materials, detecting the cell viability on the sample surface. The measured optical density, as recorded with ELISA plate reader, is directly
proportional to the number of viable cells in the culture
medium.
Figures 5(a) and (b) plot the MTT assay results
obtained using MG63 and L929 cells, respectively. In
both the plots, the results are compared in reference to
pure HA. Figure 5(a) shows that in all the mullite-containing composites, the numbers of metabolically active
cells were comparable to pure HA. BCP10M, BCP20M,
and BCP30M possessed 103%, 113%, and 101% cell
viability, respectively, compared to pure HA. Similar to
these results, Figure 5(b) shows the result of MTT assay
using L929 cells. Here also, the result obtained with HA
was considered as baseline observation (control).
Again, the MTT reduction rate data revealed that
both BCP10M and BCP20M, in average, possessed
similar or slightly better cell viability than pure HA.
Owing to the fact that error bars overlap among various MTT datasets, no statistical signicance could be
conrmed in terms of cell viability.

ALP activities. ALP activity is considered as a phenotypic marker of dierentiated cell. The ALP produced
by metabolically active MG63 cells, after 3 and 7 d of

502

Journal of Biomaterials Applications 27(5)

Figure 2. AFM images showing the surface topography of the various investigated materials: (a) pure HA sintered at 1200 C for 2 h
and the composites: (b) BCP10M, (c) BCP20M, and (d) BCP30M samples, all sintered at 1350 C for 2 h. GB represents grain
boundary.

experiments, is quantitatively plotted in Figure 6. After

3 d of culture, the maximum ALP activity was measured with pure HA and for composites (1030 wt%
mullite) the ALP activity was very near to pure HAp.
After 7 d of culture, the ALP expression of the cultured
cells was signicantly higher in all mullite-containing
composite with respect to pure HA. However, almost
no dierence in terms of ALP expression was found
among various mullite-containing composites.

Osteocalcin. OC is a later stage marker of bone cell

dierentiation. Declercq et al.26 described calcication

as a predictor of bone mineralization capacity of biomaterials in osteoblastic cell cultures. It is normally

accepted that the OC production measurement is
important to substantiate the use of investigated ceramics as bone replacement materials. As part of the present study, the results of the OC assay, are plotted in
Figure 7. The OC production after 7 d of culture on
control, pure HA, and BCP10M showed signicantly
lower values in comparison with BCP20M and
BCP30M samples. However, no dierence in OC production could be noticed between BCP20M and
BCP30M samples.

Nath et al.

503

Figure 3. Selected SEM images illustrating the adhesion of L 929 cells on various material surfaces BCP20M (a, b), BCP30 M (c, d),
and negative control sample (e, f) after in vitro culture for 1 d.

504

Journal of Biomaterials Applications 27(5)

Figure 4. SEM images of L929 cells adhered on pure HA (a, b), BCP10M (c, d), BCP20M (e, f), and BCP30M samples (g, h). AFM
image revealing extensive filopodia extension on pure HA sample is also shown (i). Results were obtained after 3 d of culture. The
seeded cell density was 1  105 cells/mL.

Discussion
This study demonstrates that the biological response of
BCP (HA and TCP) ceramics containing 20 and
30 wt% mullite is comparable (as per cell adhesion,
MTT) or better (ALP and OC results) than pure HA.
It is indeed an important result that the presence of
mullite did not aect the cellular viability property on
BCP-mullite composites, when compared to sintered
HA monolith. In the following, results will be interpreted in terms of the substrate compositional dierences or phase assemblage. Such interpretation will be
helpful to realize the following aspects: (a) In vitro cytocompatibility property on the basis of microstructural

phase assemblages; (b) prediction of osteoconduction

property based on biocompatibility with MG 63 cell
line; and (c) bone-cell functionality and phenotypic
marker expression based on ALP and OC expression
assay.

Materials
and
microstructure
effect
on
cytocompatibility. In recent years, the mixture of HA
and TCP ceramics has been reported to be better than
the monolithic forms of the either (HA or TCP) due to
the controlled resorbability of BCP ceramics. In an
interesting work, Arinzeh et al.27 studied the bone formation capabilities of BCP ceramics using human

Nath et al.

505

3.00

Days of culture
3.00
7.00

120
2.50

110
Mean optical density (405 nm)

Metabolically active cells (% Pure HA)

(a) 130

100
90
80
70
60

2.00

1.50

1.00

50
BCP10M

BCP20M

BCP30M

0.50

0.00

120

Pure HA

110

BCP10M
BCP20M
Samples

BCP30M

Figure 6. ALP activity of MG63 cells on pure HA, BCP10M,

BCP20M, and BCP30M ceramics, after 3 and 7 d of culture in
osteogenic medium.

100
90
80

60

70
60

50

50
BCP10M

BCP20M
Samples

BCP30M

Figure 5. MTT assay results showing the relative number of

metabolically active (a) MG63 cells (b) L929 adhered on pure HA,
BCP10M, BCP20M, and BCP30M samples. In (a) and (b), the
results were compared with pure HA. The results are shown
after 2 d of culture.

Osteocalcin (ng/mL)

Metabolically active cells (% Pure HA)

(b) 130

40
30
20
10

mesenchymal stem cells. The results revealed that HA :

TCP ratio of 20:80 was the optimum combination for
better bone formation, whereas pure HA and TCP
phase had much less eect on bone formation. A
closer look at Table 1 further reveals that while aTCP dominated in BCP10M material, b-TCP was the
major phase in BCP20M and BCP30M composites.
The additional presence of CaO, Al2O3, and gehlenite,
all in minor amounts (much weaker X-ray peak intensity), was also noticed. From the examples mentioned
in introduction part, it is quite clear that pure HA and
TCP are undoubtedly biocompatible material.
However, the presence of other calcium-alumino-silicate phases may also have an inuence on the biocompatibility of this material. In some earlier reported
results,28,29 it was mentioned that the biocompatibility
of the materials mainly depended on the leaching of

0
Control

Pure HA

BCP10M
Samples

BCP20M

BCP30M

Figure 7. OC expression of cultured MG63 cells on various

ceramics samples after 7 d of culture. The results were compared
with control solution supplied with the kit.

ions. The adhered cells mainly proliferate and dierentiate due to the activation of surface ions. In the present
case, the possible leaching ions would be Ca, P, Si, and
Al. The eect of Ca and P ions need not be discussed
here, as these are already known marker for the biomineralization.30 The eects of other ions, therefore,
need to be discussed with cited literature. For example,
silicon-containing materials were already investigated
as silicon-substituted HA, Si3N4, etc. Thain et al.31

506
described the in vitro biocompatibility of silicon-substituted HA thin lm and reported that the presence of Si
did not induce any toxic eect. In another study, Kue
et al.32 showed enhanced cell proliferation and OC production by human osteoblast-like MG63 cells on silicon
nitride ceramic discs. Their results revealed that Si3N4
was a nontoxic biocompatible ceramic, which could be
used as a potential biomaterial. This proves that the
presence of Si in the composite, should not ideally degrade the biocompatibility. As far as the Alleaching is concerned, a study by Ku et al. on Ti-6Al4 V alloy suggested that the release kinetics of Al-ions
could play a major role in inuencing the osteoblast
behavior.28
Figure 1 shows some residual porosity mainly at
intragranular and few at intergranular regions. In general, porosity degrades mechanical properties, but
microporosity has specic advantages during the initial
period of implantation. It should be mentioned here
that Hornez et al.33 reported that both mesoporosity
(1050 mm) and microporosity (110 mm) in HA essentially stimulated signicant cell growth of MC3T3-E1
osteoblast cells. The cell viability and cell functions
were better for microporous samples. The presence of
micropores would therefore allow better anchorage
with bone-forming cells and thus improved the mechanical attachment of these materials during initial period
of implantation.34
Another parameter that is important in the context
of cell adhesion was the surface roughness. In the present case, any minor dierence in cell viability could be
attributed to dierence in surface roughness. The subgrains are clearly visible in the AFM images, while
those are not distinct in SEM image (Figure 1). From
Figure 2, it should also be clear that the presence of
large fraction of grain boundary phase increased the
local surface roughness at the nanoscale and this
should enhance the cell adhesion property in mullitecontaining BCP composites.
In addition, eorts were made to study the adhesion
and expansion behavior of single cell in isolation and in
contact with other cells on individual material surface.
This is important to know how a single cell expands or
attaches to material surface. The initial attachment of
cells is mediated by integrins, which induces dramatic
cytoskeletal changes leading to cell spreading, development of focal adhesion complexes, cell migration, etc.
In fact, these are morphological signs, which are usually
described as overall cell morphology changes when a
cell recognizes the adsorbed adhesive proteins (and
their conformation) on a given material surface. The
cell adhesion and expansion of single cell on various
material surfaces can be seen in Figure 4(b, d, f, h).
Among various mammalian cell types, the broblast
is one of the least-dierentiated cell lines. When

Journal of Biomaterials Applications 27(5)

broblast cells were seeded at lower density, it
showed higher motility on the material surface. One
single cell can migrate and construct cellcell interaction with a recognizable polarity of movement.35 AFM
images, presented in Figure 4(i), reveal the evidences of
cell migration via the extension of cell lopodium. On
the other hand, lamellipodia extends the movement
toward the direction of cell travel and at the same
time adhere on the materials surface.35 When the cell
density was low, the cellcell interaction was also lower
and, hence, the cells interact more with material surface
by spreading themselves on the surface (Figure 4(e)).
The unidirectional long lopodia extension from the
adhered cells often formed well-developed Y or star
junction, as can be seen in Figure 4(e). In such a situation, cells were suitable to attach to the substrate and
grow in size as a part of the cell cycle in presence of
some growth factors.35 In contrast, high cell density
inhibits the growth of normal cells. In a cell crowding
surface, cell growth is inhibited by cell-to-cell contact
and as a result, it reduces spreading. The above discussion corroborates well with the dierence in cell size
and cell spreading, when cell seeding density was
varied (Figures 3 and 4).
The attachment of substrate-dependent cells, such as
broblasts and osteoblast, to a substratum (e.g. biomaterial) is a synchronized process involving cytoskeleton
reorganization, cell spreading, and formation of focal
contact.36,37 The cell adhesion behavior of pure HA
and composite samples showed presence of cytoplasmic
extension, lopodia (Figure 4(a-h)). From Figure 4(a)
(h), it can be said qualitatively that the lopodia was
more visible in case of pure HA (Figure 4(a) and (b)),
BCP20M (Figure 4(e) and (f)), and BCP30M
(Figure 4(g) and (h)) samples. However, the lopodia
was less visible and the cells surface had minimum
adhesion on BCP10M sample (Figure 4(c) and (d)).
BCP20M and BCP30M samples mainly contained
b-TCP (Table 1), whereas BCP10M mainly contained
a-TCP. Therefore, the dierence in cytocompatibility
behavior could be attributed to the dierence in phase
assemblage.

Osteoconduction
and
biochemical
markers
of bone turnover. As mentioned earlier, the aspects
of bone cell dierentiation and functionality could be
explained on the basis of ALP and OC assay results.
The synthesis of these biochemical markers of bone cell
increases with the increasing expression of osteoblasts
and decreases with the maturation of osteoblasts.38 It is
well known that OC and ALP are the phenotypic markers of the late and early stages of dierentiation of
osteoblast-like cells, respectively. An increased specic
activity of ALP in a bone cell essentially reects a shift
toward a more dierentiated state.

Nath et al.
Recalling ALP results after 7 d, it is clear that all the
mullite-containing composites had comparable expression of early stage osteoblast dierentiation marker,
which was much higher than baseline HA. This
means that osteoblast-like cells were in a better functionally dierentiated state in contact with the BCPmullite composites. Furthermore, the extent of OC
expression in case of BCP20M and BCP30M was
much higher than both baseline HA and BCP10M.
The above observations therefore conrmed that
BCP20M composite could exhibit the best combination
of OC and ALP expression.
Our results need to be explained in the perspective of
earlier literature reports. In several earlier research
reports, it was mentioned that among CaP-based materials, b-TCP containing composites showed better dierentiation and in vivo bone formation/mineralization
capability. For example, Shiratori et al.39 studied the
bone-forming ability of b-TCP, when implanted in
bone defects of rat femur. Their results revealed that
b-TCP was an appropriate material for the treatment
of bone defects. In another study, Matsuno et al.40 compared the in vitro (ALP) as well as in vivo behavior of bTCP/collagen sponge composite with only collagen
sponge (CS). In the in vitro experiments, human mesenchymal stem cell lines were used and for in vivo experiment, the samples were placed under the back skin of
nude mice for a time period of up to 12 weeks. Their
results revealed that the composite containing b-TCP
showed better osteogenic properties and had ability to
promote bone formation. In an interesting study,
Arinzeh et al.27 tried to optimize the optimum HA/bTCP ratio for better stem cellinduced bone formation.
For this purpose, they selected various ratios of HA : bTCP, that is, 100:1, 76:24, 63:37, 56:44, 20:80, and 1:100.
In both in vitro (OC) and in vivo experiments (mouse), it
was revealed that HA: b-TCP ratio of 20:80 was the best
combination for new bone formation in bone remodeling process. It can be further noted that this combination
showed better results compared to monolithic form of
HA and b-TCP. In fact, all other combination of HA
and b-TCP showed improved results compared to pure
HA and b-TCP. At the molecular level, HA phase provided the required binding sites, whereas TCP, due to its
higher dissolution properties, increased the concentration of Ca and P locally. All these led to the cascading of
dierential gene expression and positively increased the
cell dierentiation.41,42
In the present case, Table 1 clearly shows that both
BCP20M and BCP30M predominantly contained bTCP phase, while BCP10M contained more a-TCP
phase. On the basis of the above information, it
should therefore be clear that in view of the dominant
presence of b-TCP, both BCP20M and BCP30M exhibited better expression of osteblastic phenotypic marker

507
than BCP10M. The present investigation also reconrmed that single-phase HA had inferior expression
of dierentiated bone cell than BCP microstructure
containing varying ratio of b-TCP. Additionally, it
can be stated that the additional presence of gehlenite,
alumina, or CaO did not have any inhibitory eect as
far as the osteoconduction property of BCP20M and
BCP30M was concerned. In addition to the dierences
in phase assemblage, the dierence in surface roughness
at nanoscale due to uniform presence of grain boundary phase could also explain the observed dierence in
osteogenic property in the present case.
In summary, it is clearly understood from MTT data
and cell adhesion tests that HA and BCP-mullite substrates supported comparable number of viable cells.
This, in combination with all the above observations,
suggests that BCP-mullite biocomposites should be
used as a substrate for bone-forming cells.

Conclusions
a. The principal nding of the present study is that the
pressureless sintered BCP-mullite composites favorably supported cell adhesion of L929 cells in vitro.
Also, such observations were independent of mullite content (up to 30 wt%). The morphology and
motility of adhered cells was dependent on cell seeding density.
b. As far as the quantication of the metabolically
active cells is concerned (cell viability), MTT
assay results with broblast and osteoblast-like
cells did not reveal any statistically signicant difference in BCP-mullite composites in comparison
with pure HA. This conrmed that mullite and
additional presence of various phases (alumina,
CaO, gehlenite) did not have any toxic eect
in vitro and the presence of mullite did not degrade
cell viability with respect to pure HA.
c. The combination of ALP activity and OC expression results indicates that BCP-based composite
substrates with 20% or 30% mullite supported
superior osteoconduction than baseline monolithic
HA ceramic. The presence of predominant b-TCP
phase was found to be suitable for osteoblastic
function and phenotypic expression. The presence
of additional phases, like gehlenite or alumina, was
not found to have any inhibitory eect in vitro.
Based on all the in vitro analysis and observations,
in combination, BCP-biocomposites containing
20% or 30% mullite can be considered as suitable
substrate for bone cell adhesion and proliferation.
d. AFM study in combination with SEM analysis
revealed the uniform presence of grain boundary
reaction phases, which increased the surface

508

Journal of Biomaterials Applications 27(5)

roughness at the nanoscale. The increased surface
roughness property could explain better cell adhesion/proliferation property of BCP-mullite composites, compared to single-phase HA.

Acknowledgements
The authors wish to thank Department of Biotechnology
(DBT), Government of India, for the nancial help. Thanks
to Dr Lakshmi Nair of CCMB, Hyderabad, India, for providing us L929 cells. Thanks are also due to Dr S. Ganesh and
S. Mittal of BSBE Department, IIT Kanpur, India, for their
help during initial period of cell culture experiments.

References
1. Nath S and Basu B. Designing materials for hard tissue
replacements. J Kor Ceram Soc 2008; 45: 129.
2. Nath S, Bodhak S and Basu B. HDPE-Al2O3-HAp composites for biomedical applications: processing and characterization. J Biomed Mater Res B Appl Biomaterials
2009; 88B: 111.
3. Nath S, Tripathi R and Basu B. Understanding phase
stability, microstructure development and biocompatibility in calcium phosphate-titania composites, synthesized
from hydroxyapatite and titanium powder mix. Mater Sci
Eng C 2009; 29: 97107.
4. Nath S, Kalmodia S and Basu B. Sintering, microstructure and in vitro properties of HA-10 wt % Ag composites. J Mat Sci Mater Med 2010; 21: 12731287.
5. Kong YM, Bae CJ, Lee SH, Kim HW and Kim HE.
Improvement in biocompatibility of ZrO2Al2O3 nanocomposite by addition of HA. Biomaterials 2005; 26:
509517.
6. Adolfson E, Henning PA and Hermannson L. Phase
analysis and thermal stability of hot isostatically pressed
zirconia-hydroxyapatite composites. J Am Ceram Soc
2000; 83: 27982802.
7. Clifford A and Hill R. Apatite-mullite glass ceramics.
J Non-crys Sol 1996; 196: 346351.
8. Kim HW, Kong YM, Koh YH and Kim HE. Pressureless
sintering and mechanical and biological properties of
flour-hydroxyapatite composites with zirconia. J Am
Ceram Soc 2003; 86: 20192026.
9. Camazzola D, Hammond T, Gandhi R and Davey JR. A
randomized trial of hydroxyapatite-coated femoral stems
in total hip arthroplasty a 13-year follow-up. J Arthrop
2009; 24: 3337.
10. Gandhi R, Davey JR and Mahomed NN.
Hydroxyapatite coated femoral stems in primary total
hip arthroplasty. J Arthrop 2009; 24: 3842.
11. Schneider H, Okada K and Pask J. Mullite and mullite
ceramics. New York: Wiley, 1994.
12. Kanzaki S, Tabata H, Kumazawa T and Ohta S.
Sintering and mechanical properties of stoichiometric
mullite. J Am Ceram Soc 1984; 68: C6C7.
13. Nath S, Dubey A and Basu B. Mechanical properties
evaluation of novel calcium phosphate-mullite composites. J Biomat Appl, doi: 10.1177/0885328210393292.

14. Daculsi G, Laboux O, Malard O and Weiss P. Current

state of the art of biphasic calcium phosphate bioceramics. J Mater Sci Mater Med 2003; 14: 195200.
15. Nath S, Biswas K, Wang K, Bordia RK and Basu B.
Sintering, phase stability, and properties of calcium phosphate-mullite composites. J Am Ceram Soc 2010; 93:
16391649.
16. Nath S, Biswas K and Basu B. Phase stability and microstructure development in hydroxyapatite-mullite system.
Scripta Mater 2008; 58: 10541057.
17. Schottler FJ. Tissue culture methods for determining biocompatibility. ASTM STP 1983; 810: 1924.
18. Hong JY, Kim YJ, Lee H W, Lee WK, Ko JS and
Kim HM. Osteoblastic cell response to thin film
of poorly crystalline calcium phosphate apatite
formed at low temperatures. Biomaterials 2003; 24:
29772984.
19. Boanini E, Torricelli P, Gazzano M, Giardino R and
Bigi A. Nanocomposites of hydroxyapatite with aspartic
acid and glutamic acid and their interaction with osteoblast-like cells. Biomaterials 2006; 27: 44284433.
20. Shu R, McMullen R, Baumann MJ and McCabe LR.
Hydroxyapatite accelerates differentiation and suppresses
growth of MC3T3-E1 osteoblasts. J Biomed Mater Res A
2003; 67: 11961204.
21. Licht BA, Gregoire M, Orly I and Menanteau J. Cellular
activity of osteoblasts in the presence of hydroxyapatite:
an in vitro experiment. Biomaterials 1991; 12: 752756.
22. Ogata K, Imazato S, Ehara A, Ebisu S, Kinomoto Y,
Nakano T, et al. Comparison of osteoblast responses to
hydroxyapatite and hydroxyapatite/soluble calcium
phosphate composites. J Biomed Mater Res A 2005; 72:
127135.
23. Rathje W. Zur Kentnis de Phosphate: I. Uber hydroxyapatite. Bodenk Pflernah 1939; 12: 121128.
24. Santos MH, Oliveira M, Souza LPF, Mansurand HS and
Vasconcelos WL. Synthesis control and characterization
of hydroxyapatite prepared by wet precipitation process.
Mater Res 2004; 7: 625630.
25. Robert JM and Gideon AR. The Effect of 1,25(OH)zD3
on alkaline phosphatase in osteoblastic osteosarcoma
cell. J Bio Chem 1982; 257: 33623365.
26. Declercq HA, Verbeeck RMH, Ridder LIFJMD,
Schacht EH and Cornelissen MJ. Calcification as an indicator of osteoinductive capacity of biomaterials in osteoblastic cell cultures. Biomaterials 2005; 26: 49644974.
27. Arinzeh TL, Tran T, Mcalary J and Daculsi G. A comparative study of biphasic calcium phosphate ceramics
for human mesenchymal stem-cell-induced bone formation. Biomaterials 2005; 26: 36313638.
28. Ku CH, Pioletti DP, Browne M and Gregson PJ. Effect
of different Ti6Al4V surface treatments on osteoblasts
behavior. Biomaterials 2002; 23: 14471454.
29. Es-Souni M and Brandies HF. Assessing the biocompatibility of NiTi shape memory alloys used for medical
applications. Anal Bioanal Chem 2005; 381: 557567.
30. Koshihara M, Masuyama M, Uehara M and Suzuki K.
Effect of dietary calcium: phosphorus ratio on bone mineralization and intestinal calcium absorption in ovariectomized rats. BioFactors 2004; 22: 3942.

Nath et al.
31. Thian ES, Huang J, Best SM, Barber ZH, Brooks RA,
Rushton N, et al. The response of osteoblasts to nanocrystalline silicon-substituted hydroxyapatite thin films.
Biomaterials 2006; 27: 26922698.
32. Kue R, Sohrabi A, Nagle D, Frondoza C and
Hungerford D. Enhanced proliferation and osteocalcin
production by human osteoblast-like MG63 cells on silicon nitride ceramic discs. Biomaterials 1999; 20:
11951201.
33. Hornez JC, Chai F, Monchau F, Blanchemain N,
Descampsand M and Hildebrand HF. Biological and
physico-chemical assessment of hydroxyapatite (HA)
with different porosity. Biomol Eng 2007; 24: 505509.
34. Annaz B, Hing KA, Kayser M, Buckland T and
Silvio LD. Porosity variation in hydroxyapatite and osteoblast morphology: a scanning electron microscopy
study. J Microscopy 2004; 215: 100110.
35. Freshney RI. Biology of cultured cells, culture of animal
cells: a manual of basic technique, 5th ed. New York: John
Wiley & Sons, Inc, 2005, p.3142.
36. Misra A, Lim RPZ, Wu Z and Thanabalu T. N-WASP
plays a critical role in fibroblast adhesion and spreading.
Biochem Biophy Res Comm 2007; 364: 908912.
37. West KA, Zhang H and Brown MC. The LD4 motif of
paxillin regulates cell spreading and motility through an

509

38.

39.

40.

41.

42.

interaction with paxillin kinase linker (PKL). J Cell Biol

2001; 154: 161176.
Chang X, Hou ZM, Yasuaki S, Tomoo T and Akira Y.
Quantitative study of alkaline phosphatase and osteocalcin in the process of new bone. West China J Stomat
2005; 23: 424426.
Shiratori K, Matsuzaka K, Koike Y, Murakami S,
Shimono M and Inoue T. Bone formation in b-tricalcium
phosphate-filled bone defects of the rat femur: morphometric analysis and expression of bone related protein
mRNA. Biomed Res 2005; 26: 5159.
Matsuno T, Hashimoto Y, Nakahara T, Kuremoto K,
Nakamura T and Satoh T. b-TCP/Collagen sponge
composite enhances the osteogenic differentiation of
mesenchymal stem cells. J Hard Tissue Bio 2005; 14:
189191.
Piattelli A, Scarano A and Mangano C. Clinical and histologic aspects of biphasic calcium phosphate ceramic
(BCP) used in connection with implant placement.
Biomaterials 1996; 17: 17671770.
Uoshima MH, Ishikawa I, Kinoshita A, Weng HT and
Oda S. Clinical and histological observation of replacement of biphasic calcium phosphate by bone tissue in
monkeys. Int J Periodont Rest Dent 1995; 15: 204213.