Biomaterials 33 (2012) 4044e4058

Contents lists available at SciVerse ScienceDirect

Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials

Ectopic bone formation by 3D porous calcium phosphate-Ti6Al4V hybrids

produced by perfusion electrodeposition
Yoke Chin Chai a, b, c,1, Greet Kerckhofs c, g, 2, Scott J. Roberts a, c,1, Simon Van Bael c, d, e, 3, Evert Schepers c, f, 4,
Jozef Vleugels g, 5, Frank P. Luyten a, c,1, Jan Schrooten c, g, *
a
Laboratory for Skeletal Development and Joint Disorders, KU Leuven, O&N 1, Herestraat 49, Bus 813, 3000 Leuven, Belgium
b
Department of Biomedical Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia
c
Prometheus, Division of Skeletal Tissue Engineering, KU Leuven, O&N 1, Herestraat 49, Bus 813, 3000 Leuven, Belgium
d
Division of Biomechanics and Engineering Design, KU Leuven, Celestijnenlaan 300c, Bus 2419, 3001 Heverlee, Belgium
e
Department of Mechanical Engineering, Division of Production Engineering, Machine Design and Automation, KU Leuven, Celestijnenlaan 300b, 3001 Leuven, Belgium
f
Department of Prosthetics Dentistry, KU Leuven, Kapucijnenvoer 7, Blok A e Box 7001, 3000 Leuven, Belgium
g
Department of Metallurgy and Materials Engineering, KU Leuven, Kasteelpark Arenberg 44, Bus 2450, 3001 Leuven, Belgium

a r t i c l e i n f o a b s t r a c t

Article history: Successful clinical repair of non-healing skeletal defects requires the use of bone substitutes with robust
Received 9 December 2011 bone inductivity and excellent biomechanical stability. Thus, three-dimensionally functionalised porous
Accepted 9 February 2012 calcium phosphate-Ti6Al4V (CaPeTi) hybrids were produced by perfusion electrodeposition, and the
Available online 28 February 2012
in vitro and in vivo biological performances were evaluated using human periosteum derived cells
(hPDCs). By applying various current densities at the optimised deposition conditions, CaP coatings with
Keywords:
sub-micrometer to nano-scale porous crystalline structures and different ion dissolution kinetics were
Bone formation
deposited on the porous Ti6Al4V scaffolds. These distinctive physicochemical properties caused
Calcium phosphate
Osteoinduction
a significant impact on in vitro proliferation, osteogenic differentiation, and matrix mineralisation of
Electrodeposition hPDCs. This includes a potential role of hPDCs in mediating osteoclastogenesis for the resorption of CaP
Titanium scaffolds coatings, as indicated by a significant down-regulation of osteoprotegerin (OPG) gene expression and by
the histological observation of abundant multi-nucleated giant cells near to the coatings. By subcuta-
neous implantation, the produced hybrids induced ectopic bone formation, which was highly dependent
on the physicochemical properties of the CaP coating (including the Ca2þ dissolution kinetics and coating
surface topography), in a cell density-dependent manner. This study provided further insight on stem
cell-CaP biomaterial interactions, and the feasibility to produced bone reparative units that are predic-
tively osteoinductive in vivo by perfusion electrodeposition technology.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Despite the advances in biomaterials and tissue engineering,

clinical repair of critical-size or non-healing skeletal defects
* Corresponding author. Department of Metallurgy and Materials Engineering, remains challenging. This is due to the lack of essential biological
KU Leuven, Kasteelpark Arenberg 44, Bus 2450, 3001 Leuven, Belgium. and biomechanical entities at the deteriorated defect site, whereby
Tel.: þ3216321212; fax: þ3216341992. osteoprogenitor cells, extracellular matrix, signalling molecules/
E-mail addresses: yokechin.chai@med.kuleuven.be (Y.C. Chai), greet.kerckhofs@
growth factors, and mechanical stability of the fracture, are the four
mtm.kuleuven.be (G. Kerckhofs), Scott.Roberts@med.kuleuven.be (S.J. Roberts),
Simon.VanBael@mech.kuleuven.be (S. Van Bael), evert.schepers@med.kuleuven.be vital constituents that are necessary for an effective bone defect
(E. Schepers), jozef.vleugels@mtm.kuleuven.be (J. Vleugels), Frank.Luyten@ healing [1]. Other parameters, including vascularity, comorbidities,
uz.kuleuven.be (F.P. Luyten), Jan.Schrooten@mtm.kuleuven.be (J. Schrooten). and the physiological profile of the patient, are recently proposed to
1
Tel.: þ3216346169; fax: þ3216346200. have strong interactions with these constituents, and thus consti-
2
Tel.: þ3216321194; fax: þ3216321990.
3
Tel.: þ3216322772; fax: þ3216322480.
tuting a conceptual framework for enhancement of bone repair [2].
4
Tel.: þ3216332471; fax: þ3216332445. Potentially, this framework can be translated into a clinically-
5
Tel.: þ3216321244; fax: þ3216321270. relevant setting, through in vitro production of a robust bone

0142-9612/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biomaterials.2012.02.026
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
reparative unit that possesses both the necessary biological and
biomechanical characteristics. In this context, engineering a three-
dimensionally (3D) functionalised calcium phosphate-titanium
(CaPeTi) hybrid that combines both the superior mechanical
integrity and osteoinductivity, with or without the use of stem cells
technology, may appear as an ideal synthetic bone substitute that
could overcome certain limitations and clinical complications
related to the standard bone grafting procedures [3].
Indeed, titanium (Ti) and calcium phosphate (CaP)-based
biomaterials have respectively gained promising clinical track
records in orthopaedic and dental applications over the years,
either due to their well-known biocompatibility, biomechanical
strength, osteoconductivity or the increasing scientific evidences
on their osteoinductivity [4]. These have provoked strong inter-
ests in the research and development of various surface func-
tionalisation strategies to modify Ti-based biomaterials with CaP
layers, aiming to promote osseointegration and to induce bone
formation within a superior mechanical framework at the defect
site [5]. So far, most of the existing functionalisation methods
have failed to produce such an osteoinductive hybrid, with the
exception of biomimetic deposition of CaP coatings, where bone
formation was induced intramuscularly [6]. However, this tech-
nique has limited control over the deposition of biomimetic
apatite with desirable physicochemical properties, and the
process is relatively time consuming. Alternatively, electrodepo-
sition or electrolytic deposition appears to be a promising tech-
nique that can promptly functionalise three-dimensional (3D)
porous Ti structures with CaP layers and offers higher controlla-
bility and reproducibility on the coating properties [7]. Unfortu-
nately, the produced coating often lacks of osteoinductivity
in vivo, although higher osteogenicity has been reported using
in vitro culture systems [8]. This is mainly due to the lack of in-
depth understanding on the necessary material properties of
CaP that would effectively trigger osteogenesis in terms of the
in vitro stem cell-material and in vivo cell-host-material interac-
tions [9], as well as the absence of a comprehensive technological
knowledge on electrodepositing CaP layers with excellent bone
induction properties.
Previously, we reported on the integration of perfusion and
electrodeposition technology to deposit CaP layers onto additive
manufactured 3D porous Ti6Al4V scaffolds in a controllable and
reproducible manner [10]. In this study, we describe the use of
a six-channel perfusion electrodeposition system (6P-ELD) as
a laboratory up-scaling functionalisation tool, to produce three-
dimensionally (3D) functionalised porous CaPeTi hybrids with
different physicochemical properties. The in vitro biological
performance of the produced hybrid was assessed by analysis of
the in vitro proliferation and osteogenic differentiation (i.e. alka-
line phosphatase activity and osteogenic gene markers expression)
of human periosteum-derived cells (hPDCs). The in vivo ectopic
osteoinductivity was evaluated by seeding hPDCs onto the hybrids
followed by subcutaneous implantation in nude mice. After 8
weeks of implantation, samples were retrieved, characterised by
nanofocus X-ray computed tomography (nanoCT) and processed
for histological analysis for de novo bone formation. To correlate
the in vitro and in vivo biological outcomes, the physicochemical
properties of the deposited CaP coatings (including coating
distribution, surface morphology, dissolution kinetics and thick-
ness) before and after in vitro studies were analysed by scanning
electron microscopy (SEM), inductive-couple plasma-atomic
emission spectrophotometry (ICP-AES) and microfocus X-ray
computed tomography (mCT). These data are essential as feedback
for optimisation of the 6P-ELD production parameters, in order to
produce a bone reparative unit that is predictively osteoinductive
in vivo.
4045
2. Materials and methods
2.1. Production of three-dimensionally functionalised porous CaPeTi hybrids using
six-channel perfusion electrodeposition system (6P-ELD)
A six-channel perfusion electrodeposition (6P-ELD) chamber was fabricated for
lab-scale production of 3D porous CaPeTi hybrids, based on the single-channel
experimental prototype as reported previously (Fig. 1A) [10]. The six channels
were designed in a cylindrical array (Fig. 1B), and each channel (10 cm length)
consisted of a cathode (a 3-pin holder for Ti-scaffold) surrounded by platinum ring
anode (10 mm in height and diameter). A supersaturated calcium phosphate solu-
tion [11] was used as electrolyte and perfused through the 6 channels using a peri-
staltic pump (IsmatecÒ). Current densities (I) of 1.54, 5, 10 or 20 mA/cm2 were
applied on each of the six scaffolds in order to deposit CaP coatings with different
physicochemical properties onto the 3D porous Ti6Al4V scaffolds [Ti-scaffolds,
fabricated by selective laser melting (designed pore size ¼ 1000 mm)] [12], while
other parameters were fixed at optimum conditions: deposition time (t) ¼ 6 h,
electrolyte flow rate (f) ¼ 10 ml/min, and the process temperature (T) ¼ 50  C.
Scaffolds of 3 mm height  6 mm diameter (type A) were used for the in vivo ectopic
bone formation assays, and 10 mm height  6 mm diameter (type B) for in vitro
coating characterisation and cell culture experiments (Fig. 1E). The pH change of the
bulk electrolyte was recorded using a pH measurement kit (PicoÒ Technology, U.K.),
and the deposited CaP coating mass was calculated from the dry weight of the
scaffolds before and after electrodeposition (n ¼ 3) to represent deposition
efficiency.
2.2. Physicochemical characterisation of the deposited CaP coatings
2.2.1. Microfocus X-ray computed tomography (mCT) analysis
The distribution, morphology and thickness of the CaP coatings deposited on the
Ti6Al4V scaffolds were characterised using high-resolution mCT on a Skyscan 1172
system (Skyscan NV, Kontich, Belgium) at an isotropic voxel size of 4.5 mm3. The
samples were scanned using a source voltage and current of 60 kV and 167 mA, with
a filter of 0.5 mm Al and a rotation step of 0.3 over a total of 180 . The obtained
radiographic images were reconstructed using NRecon (Skyscan NV, Kontich, Bel-
gium) software and the coating thickness was measured using CTAn (Skyscan NV,
Kontich, Belgium) software. Briefly, the coating thickness at 20 different locations on
a reconstructed uCT image was measured, and in total 10 representative mCT images
were used for each produced hybrid. Then, the obtained values were subtracted by
the value measured on the plain Ti-scaffolds (due to partial volume effect (pv)), and
the average coating thickness was determine for each produced hybrid.
2.2.2. Dissolution test
The produced CaPeTi hybrids (n ¼ 3) were incubated in 15 ml phosphate
buffered solution (PBS, BiowhittakerÒ, without Ca2þ and Mg2þ ions) in free floating
condition on an orbital rotator at 10 rpm inside a 37  C incubator. The solutions were
refreshed at 3 and 6 h, and at 4, 7, 14 and 21 days, and the pH of the solutions were
recorded before storage at 80  C. The released [Ca2þ] and [PO3- 4 ] over 21 days were
measured by Inductive Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES)
(Varian 720-ES) at 393.366 nm (for Ca2þ) and 253.561 nm (for PO3 4 ) respectively.
After 21 days of incubation, the dried weights of the hybrids were recorded, and the
change in coating thickness were characterised by mCT as described above.
2.2.3. Scanning electron microscopy analysis (SEM)
The morphology of the CaP coatings before and after dissolution was assessed by
SEM coupled with an energy dispersive X-ray (EDAX) analysis (FEI XL30 FEG) at
10 kV. The samples were coated with gold to increase conductivity of the deposited
CaP coatings for better visualisation.
2.3. Characterisation of in vitro biological behaviours of hPDCs seeded on CaPeTi
hybrids
2.3.1. Isolation and expansion of hPDCs from human donors
Periosteal biopsies (10 mm  5 mm) were harvested from the medial side of the
proximal tibia of male and female of adolescent patients [6 donors, age 14.9  2.1]
during distraction osteogenesis. The periosteum was stripped from the tibia with
a periosteal lifter and transported in growth medium (GM) [DMEM þ GlutaMAXÔ-1,
Invitrogen] containing 10% foetal bovine serum (InvitrogenÔ), 1% antibiotic-
antimycotics solution and 1% sodium pyruvate. The biopsies were finely minced
and digested overnight at 37  C in 0.2% type IV collagenase (Invitrogen) and
subsequently centrifuged at 1300 rpm for 10 min to collect the deliberated perios-
teal cells. All collected cells were pooled together and seeded in T175 flask in GM.
Non-adherent cells were removed after 4 days by changing the medium and the
adherent cells were subsequently expanded in GM. Upon confluence, the cells were
harvested and resuspended in DMEM with 10% FBS and 10% DMSO, and stored in
liquid nitrogen. All procedures were approved by the ethical committee for Human
Medical Research (KU Leuven), and the patient informed consent forms were ob-
tained as described previously [13].
4046 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058

Fig. 1. Six-channel perfusion electrodeposition (6P-ELD) system and additive manufactured 3D porous Ti6Al4V scaffolds (Ti-scaffolds) by selective laser melting. (A & B) Photograph
of the 6P-ELD chamber and a schematic showing the 6 electrodeposition channels arranged in a cylindrical array, each containing an electrodeposition cell consisting of an anode
(platinum ring) and a cathode (Ti-scaffold) that were connected in parallel to a power supply (C & D). (E) 3D reconstructed images of the produced porous Ti-scaffolds used for
in vivo (type A) and in vitro (type B) studies.

2.3.2. Cell seeding efficiency, viability and proliferation of hPDCs polymerase chain reaction (PCR) as described previously [15] (n ¼ 3) (Table 1). For
The pooled hPDCs was thawed from liquid nitrogen and expanded in growth this study, 1 mg of the purified RNA for each sample was used for the synthesis of
medium. At confluence, cells were harvested and 35 ml of cell suspension containing complementary DNA (cDNA).
200,000 cells (50% of the scaffold filling volume) was drop seeded onto each hybrid
and incubated statically for 1 h to allow cell attachment. Then, the samples were
2.4. In vivo ectopic bone formation assay
transferred into 24-well plates for subsequent culture in GM. Cell seeding efficiency
(CSE) was calculated by quantifying the DNA content in the left over medium after
The CaPeTi hybrids produced at 5 and 10 mA/cm2 were selected for this assay.
seeding and normalised to the total DNA content of the initial cell seeding density
Before implantation, the samples were autoclaved and pre-wetted with GM for 2 h.
(n ¼ 3). As described previously, dynamic rotation seeding was performed to
To determine the optimum cell seeding density, cell suspensions containing 1, 2, 3,
compare the CSE with that of static drop seeding as well as the effect of pre-wetting
and 5 million hPDCs were drop seeded onto each hybrid (n ¼ 3) and incubated
samples with GM on CSE [14]. After 21 days of culture, the cell viability on hybrids
statically for 1 h to allow cell attachment prior to the dynamic rotation seeding
was evaluated qualitatively using a Live/DeadÒ cell viability kit (Invitrogen). Cell
method. The CSE was calculated by quantifying the DNA content of cells attached on
proliferation was assessed by measuring cellular metabolic activity over 21 days
the scaffolds and normalised to the DNA content of the initial seeding density. Based
using AlamarBlueÒ (Invitrogen) (10% in GM, n ¼ 3), and by quantifying the DNA
on the obtained CSE data, the CaPeTi hybrids seeded with 1 and 3 million hPDCs
content of each sample using a Quant-iTÔ dsDNA HS assay kit (Invitrogen) (n ¼ 3).
hPDCs seeded on uncoated Ti-scaffolds and cultured in growth medium or osteo-
genic medium (OM) [GM þ 100 nM dexamethasone þ 50 mg/ml ascorbic
acid þ 10 mM b-glycerophosphate] for the same culture duration were used as Table 1
negative and positive controls, respectively. Primers used for Sybr Green polymerase chain reaction (RT-PCR).

Gene Primer sequences

2.3.3. Physicochemical characterisation of cell-material interactions
During medium refreshment, the spent media were stored at 80  C till Ca2þ Forward Reverse
and PO3
4 dissolution analysis by ICP-AES. Before analysis, all collected samples were Runx2 50 -CGCATTCCTCATCC 50 -GCCTGGGGTCTGTAA
thawed to room temperature and diluted 1000x (for Ca2þ measurement) or 100x CAGTAT-30 TCTGA-30
(for PO3
4 ) in deionised water. After 21 days of in vitro culture, the change in coating Osterix 50 -AGTGACCTTTCAGC 50 -GGGAAAAGGGAGGG
thickness was characterised by mCT, whereas the morphology of cell growth on the CTCCAA-30 TAATCA-30
produced hybrids was visualised by SEM. Briefly, after fixation with 2.5% glutaral- Osteocalcin 50 -GTGCAGCCTTTGTG 50 -GCTCACACACCTCCC
dehyde and post-fixed with 2% osmium tetroxide, the samples were dehydrated, TCCAA-30 TCCT-30
dried with hexamethyldisilazine solution and finally sputter coated with gold for Osteopontin 50 -ACTGATTTTCCCAC 50 -TCAGGGTACTGGAT
SEM observation. GGACCT-30 GTCAGG-30
Col-1 50 -GACGAAGACATCCCA 50 -AGATCACGTCATCGC
2.3.4. Alkaline phosphatase (ALP) activity and osteogenic gene marker analysis CCAAT-30 ACAAC-30
At 21 days of culture in GM, the samples were harvested in 0.05% Triton X-100 BMP-2 50 -GTATCGCAGGCACTC 50 -TTTTCCCACTCGTTT
(in PBS) and clear cell lysates were obtained by 3 freeze-thaw cycles, followed by AGGTC-30 CTGGT-30
sonication and centrifugation at 13,000 rpm for 10 min. The ALP activity of each Osteoprotegerin 50 -GGGGACCACAATGAA 50 -AGCTGATGAGAGG
sample (n ¼ 3) was measured using a BluePhosÒ microwell phosphatase substrate CAACT-30 TTTCTTCG-30
system (Kirkegaard & Perry Laboratories) according to the manufacturer’s instruc- Matrix Gla Protein 50 -CAAGAGAGGATCCGA 50 -CGCTTCCTGAAGTA
tions. The expression levels of osteogenic gene markers for cell differentiation and GAACG-30 GCGATT-30
matrix mineralisation (i.e. Runx2, Osterix, Osteocalcin, Osteopontin, Collagen type-1, b-Actin 50 -CCCAGATCATGTTTGA 50 -CCTCGTAGATGGG
Matrix Gla Protein), bone morphogenetic protein-2 (BMP-2), and regulation of GACCT-30 CACAGT-30
osteoclastogenesis (Osteoprotegerin) were assessed quantitatively by the Sybr green
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058 4047

were implanted subcutaneously on the back in the cervical region of NMRI-nu mice
(female, 8 weeks) (n ¼ 3) [13]. Unseeded CaPeTi hybrids and uncoated Ti-scaffolds
were used as controls. After 8 weeks, the implants were retrieved and fixed in 4%
paraformaldehyde for 2 h before characterisation by nanofocus X-ray computed
tomography (nanoCT, Phoenix Nanotom SÒ, GE) for de novo bone formation (voxel
size ¼ (3.75 mm)3; voltage ¼ 80 kV; current ¼ 195 mA; filter ¼ 1 mm Al and 1 mm Cu;
exposure time ¼ 750 ms; frame averaging ¼ 1; scanning time per sample ¼ about
30 min). The obtained reconstructed images (n ¼ 1500e2000 per explant) were
analysed using CTAn (Skyscan NV, Kontich, Belgium) software, by applying a global
thresholding to calculate the volume of bone formed within each explant (n ¼ 3).
After thresholding, the percentage of bone formed within a produced hybrid was
determined by normalising the calculated bone volume to the volume of the
available space within the hybrid (excluding volume of the centre hollow and the
volume of the coating). For histological analysis, the explants were embedded in
poly-methyl methacrylate (PMMA) resins and tissue sections of w70 um were ob-
tained and stained with Stevenels’ Blue & Van Gieson’s Picrofuchsin, or Toluidine
Blue & Basic Fuchsin. All procedures were approved by the ethical committee for
animal research of the KU Leuven and the animals were housed according to the
guidelines of the university’s animalium.
2.5. Statistical analysis
Data are expressed as mean  standard deviation (S.D). Unpaired student t-test
(two-tailed) was performed to compare two independent groups, by establishing
the statistical significance at p < 0.05.
3. Results
3.1. Characterisation of the P-ELD process and surface morphology
of the deposited CaP coatings
Fig. 2A and B show the reduction of the pH of the bulk electrolyte
from the original neutral pH (¼ 7.0) to acidic conditions (constant
volume of 2 L were used) during the 6 h of 6P-ELD at different applied
current densities using type A and type B scaffolds. This reduction in
pH was dependent on the total applied current for the 6 channels (in
milliAmpere, mA), where an increase in current resulted in a faster
pH reduction rate. Due to the difference in surface area of the two
scaffold types, higher currents were needed for deposition of coat-
ings on type B scaffolds (with higher surface area) in order to achieve
a similar current density as for type A scaffold (with lower surface
area). After 6 h of 6P-ELD, the pH of the bulk electrode varied between
5 and 7 when type A scaffold was used, whereas the pH dropped
dramatically to highly acidic (w3.0) when a high current (174 mA)
was applied on the type B scaffold. Plotting the deposited coating
weights (normalised to the height of the scaffolds) against the

Fig. 2. Characterisation of the 6P-ELD process and surface morphology of the produced CaP-Ti6Al4V hybrids. (A & B) Reduction of pH of the bulk electrolyte during 6 h of 6P-ELD at
different applied current densities (1.54, 5, 10 and 20 mA/cm2) using type A (surface area ¼ 1.8 cm2) and type B (surface area ¼ 5.8 cm2) Ti-scaffolds. (C) Graphs showing the
deposited coating weight after 6 h of 6P-ELD at different applied current densities (1.54, 5, 10 and 20 mA/cm2) for both scaffold types. Mean  S.D. (n ¼ 3). (D) Electron micrographs
showing a different surface morphology of the deposited CaP coatings at different current densities on type B scaffolds, with nano- and sub-micrometer porous structures (insets).
At lower current densities (5 mA/cm2), the coatings were homogeneously deposited around the struts of the Ti-scaffolds without clogging of the pores. Cracks ([) were apparent
on coating produced at 5 mA/cm2. At higher current densities (10 mA/cm2), more pores were clogged by coatings with a sub-micrometer globular porous structure. SA, surface
area.
4048 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
applied current densities for both scaffold types (Fig. 2C), revealed
a linear relationship between the deposited coating weights and the
applied current densities up to 10 mA/cm2. A further increase in
current density (e.g. at 20 mA/cm2) was not beneficial as the
deposited coating weight was not increased for both scaffold types.
Scanning electron micrographs revealed homogeneous CaP
coating deposition around the struts of the Ti-scaffolds when 6P-
ELD was performed at low current densities (i.e. 1.54 and 5 mA/
cm2) (Fig. 2D (i & iii)). No clogging of the pores of scaffolds by the
deposited coatings was observed and coating cracks ([) became
apparent at current density of 5 mA/cm2 (see inset). At high current
densities (10 and 20 mA/cm2), thick CaP coatings were deposited
throughout the scaffolds and caused clogging of some scaffolds
pores (Fig. 2D (v & vii)). At high magnification, nano-scale porous
crystallites were observed on the surface of the coatings deposited
at low current densities, which became denser and thinner with an
increase in current density (Fig. 2D (ii & iv)). Meanwhile, sub-
micrometer globular porous structures were observed on the
surface of the thick coatings deposited at high current densities
(Fig. 2D (vi & viii)). At much higher magnification, nano-scale rod-
shape crystal structures were observed on these globular porous
structures (insets of Fig. 2D (vi & viii)).
3.2. mCT analysis of the distribution, morphology and thickness of
the deposited CaP coatings
Fig. 3A shows the physical appearance of the produced CaPeTi
hybrids as compared to plain Ti-scaffolds, where homogeneous
and thick coatings were deposited at low and high current densi-
ties, respectively. The deposited coatings (orange-red in pseudo-
colour) were found to be homogeneously distributed around each
strut (Ti, blue) of the scaffolds, as shown by the representative mCT
cross-section images at both the top and the centre of the produced
CaPeTi hybrids (Fig. 3B). At high magnification, the coatings were
found to be deposited along the peripheral of the struts at 1.54, 5
and 10 mA/cm2 with the presence of micropores (indicated by
white arrows in the insets), where the coating thickness and
uniformity were increased with increase in current density
(Fig. 3C). However, at 20 mA/cm2, the deposited coatings were less
uniform and scattered away from the struts. Partial volume effect
(revealed as yellow-orange thin structures on the surface of the
scaffold struts, indicated by D in Fig. 3C) was observed on mCT
images of the plain Ti-scaffolds. By subtracting the baseline due to
partial volume effect (pv) from the measured total coating thick-
ness (t) (Fig. 3D (i)), the thickness of the deposited coatings were
calculated and found to increase significantly from 15.4  1.25 mm
(at 1.54 mA/cm2) to 20.1  0.57 mm (at 5 mA/cm2) and
46.1  5.27 mm (at 10 mA/cm2) (Fig. 3D (ii)). The coating produced
at 20 mA/cm2 had highest thickness (55.4  9.98 mm), but was not
significantly higher than that produced at 10 mA/cm2. These
measured coating thicknesses were linearly correlated to the
coating weights when deposited at 5 and 10 mA/cm2, but not at
1.54 and 20 mA/cm2 (Fig. 3D (iii)).
3.3. In vitro dissolution of the deposited CaP coatings
By ICP-AES measurements, the coatings produced at 10 and
20 mA/cm2 had the highest Ca2þ dissolution over 21 days of
dynamic incubation in PBS, where the Ca2þ release from coating
deposited at 10 mA/cm2 became significantly higher than the
coating deposited at 20 mA/cm2 at 21 days (p < 0.05) (Fig. 4A (i)).
This was followed by the coating deposited at 5 mA/cm2, where the
released [Ca2þ] was significantly higher than that of coating
deposited at 1.54 mA/cm2 after 6 h of incubation in PBS. Interest-
ingly, the [PO3
4 ] in PBS containing CaP-Ti6Al4V hybrids were
measured to be ranged from 0.07 to 0.55 ppm, which were statis-
tically similar to the samples containing plain Ti-scaffolds (Fig. 4B
(ii)). The ion dissolution behaviour did not have a significant impact
on the pH of PBS, as statistically similar reduction in pH was
observed between samples containing the produced hybrids and
plain Ti-scaffolds over 21 days of incubation (Fig. 4A (iii)).
During in vitro culture, a significantly lower [Ca2þ] was found in
the medium of samples containing the hybrids when compared to
the plain Ti-scaffolds (Fig. 4B (i)). This reduction in [Ca2þ] was related
to the coating production current density. The coatings produced at
10 and 20 mA/cm2 resulted in the lowest [Ca2þ], followed by coat-
ings produced at 5 mA/cm2, whereas the highest [Ca2þ] was
measured for the coating produced at 1.54 mA/cm2. Interestingly,
the produced hybrids resulted in a significantly higher [PO3 4 ] in the
medium after 1 day (p < 0.05) and 5 days (only for coating deposited
at 10 mA/cm2) of in vitro culture, which became statistically similar
to that of plain Ti-scaffolds after 12 days, and were lower than that of
plain Ti-scaffolds for the subsequently culture time points (Fig. 4B
(ii)). No significant differences in [PO3 4 ] were found in medium
containing the hybrids produced at different current densities for
each time point, except at 1 day where the coating produced at
10 mA/cm2 released the highest [PO3 4 ] (p < 0.05).
This dissolution behaviour was supported by the coating weight
loss measurements, where coatings produced at higher current
densities had a higher weight loss (Fig. 4C). This coating weight loss
was not significantly different when incubated in PBS (ranged from
0.16 to 5.52%) or during in vitro cell culture (ranged from 0.74 to
4.01%), although the latter resulted in a higher weight loss by
coatings produced at 1.54 and 5 mA/cm2. Interestingly, a slight gain
in coating weight was observed by coating produced at 1.54 mA/
cm2 after dissolution test.
3.4. Surface morphology and change in coating thickness
distribution after in vitro dissolution and cell culture assays
SEM and EDAX analysis suggested the presence of NaCl (*) on the
surface of the plain Ti-scaffolds and on the coating deposited at
1.54 mA/cm2 (Fig. 5A (i & ii)) after 21 days of in vitro dissolution.
Interestingly, rod-shape crystals (indicated by white triangle) were
observed on the surface of the coating deposited at 5 mA/cm2 (inset
of Fig. 5A (iii)), but not on the surface of coatings deposited at 10 and
20 mA/cm2 (Fig. 5A (iv & v)). After 21 days of in vitro culture, SEM
revealed favourable growth of hPDCs on the surface of plain Ti-
scaffolds (the negative controls) and on coatings deposited at 1.54
and 5 mA/cm2 and to a much lower extent on coatings deposited at
10 and 20 mA/cm2 (Fig. 5A (viex)). Surprisingly, abundant fila-
mentous crystal structures and agglomerates (:, see insets of
Fig. 5A (viii & ix)) were observed on the surface of coatings produced
at 5 and 10 mA/cm2, which were found to contain high amounts of
carbon (C), oxygen (O), calcium (Ca) and phosphorus (P) by EDAX
analysis (insets of Fig. 5B (ii & iii)). These observations were com-
plemented by the mCT quantification of the coating thickness
distributions, where a reduction in main coating thickness was
observed on the coatings deposited at 5, 10 and 20 mA/cm2 after
in vitro dissolution, which were further decreased after in vitro cell
culture with a concomitant appearance of a peak around 60e70 mm
thickness ([) (Fig. 5B (iieiv)). No change in coating thickness
distribution was observed for the coatings deposited at 1.54 mA/cm2
after 21 days of dissolution test or in vitro culture (Fig. 5B (i)).
3.5. Cell seeding, viability and growth of hPDCs on the produced
CaPeTi hybrids
By static cell seeding onto non pre-wetted hybrids, a CSE of
about 65% was obtained for hybrids produced at 1.54 and 5 mA/
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058 4049

Fig. 3. mCT analysis of the deposited CaP coatings. (A) Photographs of the produced CaeTi hybrids, and (B) the representative mCT cross-section image (in pseudo-colour) at the top
and centre of scaffold showing homogenous CaP coating deposition (orange-red) around each strut (Ti, blue) of the scaffold under P-ELD at 10 mA/cm2 for 6 h (scale bar ¼ 1 mm).
(C) Comparison of coating distribution (*) around struts of the produced hybrids (scale bar ¼ 100 mm). Micropores (indicated by white arrows in insets) were observed within the
coatings deposited at 5, 10 and 20 mA/cm2 (D) (i & ii) The quantified thicknesses of the deposited coatings based on mCT image analysis after subtracting the baseline due to partial
volume effect (pv) calculated for the plain Ti-scaffold, and (iii) the correlation between the measured thicknesses and weights of coatings deposited at different current densities.
Mean  S.D. (n ¼ 3). Unpaired t-test (two-tailed):xp < 0.001. t, total coating thickness. (For interpretation of the references to colour in this figure legend, the reader is referred to the
web version of this article.)

cm2, which was statistically similar to the CSE of the negative
controls (cells seeded on plain Ti-scaffolds, GM) (Fig. 6A). A
significantly higher CSE (w75%, p < 0.001) was obtained with
hybrids produced at 10 and 20 mA/cm2. Although a higher CSE was
obtained on these hybrids, the cellular metabolic activity was
significantly lower (p < 0.001). By performing a dynamic seeding
onto pre-wetted hybrids, the CSE was increased to > 90%, including
the negative controls (Fig. 6B). In this case, the cellular metabolic
activities were increased 3e5 folds compared to static seeding.
Nevertheless, the cellular metabolic activities for cells seeded on
hybrids produced at 1.54 and 5 mA/cm2 were significantly higher
and those at 10 and 20 mA/cm2 were lower than that of the
negative controls.
By culturing cells seeded dynamically onto the pre-wetted
hybrids for up to 21 days, statistically the same metabolic
activities were measured on hybrids produced at 1.54 and 5 mA/
cm2 as on the negative controls (Fig. 6C). However, hybrids
produced at 10 and 20 mA/cm2 gave rise to significantly lower
metabolic activities. The positive controls (i.e. culturing cell-seed
plain Ti-scaffolds in osteogenic medium, OM) produced the high-
est metabolic activities over the 21 days. By DNA quantification,
a significantly higher DNA content was measured on hybrids
produced at 1.54 mA/cm2 and the positive control, and a signifi-
cantly lower DNA content on hybrids produced at 10 and 20 mA/
cm2 (Fig. 6D). These data were complemented by the qualitative
live-dead staining, where hybrids produced at 1.54 mA/cm2 resul-
ted in a similar cell viability as compared to the negative controls,
which was reduced on hybrids produced at 5, 10 and 20 mA/cm2
(Fig. 6E). High cell viability was demonstrated on the positive
controls.
4050 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058

Fig. 4. Characterisation of in vitro dissolution behaviour of the produced CaP coatings. (A) The dissolution of (i) Ca2þ and (ii) PO3
4 during 21 days of dynamic incubation in PBS and
the corresponding change in pH of the solutions. (B) The measured [Ca2þ] and [Pi] in medium during 21 days of in vitro culture of hPDCs on the produced hybrids. (C) The cor-
responding coating weight loss after incubation in PBS and in vitro cell culture. Mean  S.D. (n ¼ 3). Unpaired t-test (two-tailed): *p < 0.05, #p < 0.01, xp < 0.001.

3.6. ALP activity and the expression of osteogenic gene markers
expression by hPDCs seeded on the produced hybrids
Culturing cells seeded upon CaPeTi hybrids in GM for 21 days
caused a reduction in ALP activity, which was significantly reduced
for cells seeded on hybrids produced at 5 and 10 mA/cm2 (p < 0.05
and p < 0.01, respectively) as compared to the negative controls
(represented by GM) (Fig. 7). As expected, a significant increase in
ALP activity was induced by culturing cells seeded on plain
Ti-scaffolds in osteogenic medium (represented by OM).
Due to low cell viability on hybrids produced at 20 mA/cm2, this
condition was not assessed for gene expression analysis. Based on
the Sybr green PCR data, the Runx2 expression was found to be
significantly upregulated by cells seeded on hybrids produced at 5
and 10 mA/cm2 (p < 0.01 and p < 0.05, respectively) (Fig. 8 (i)).
Interestingly, Osterix (Osx) e a transcription factor for osteoblast
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058 4051

Fig. 5. Surface morphology and change in coating thickness distribution of the produced hybrids after 21 days of in vitro dissolution and cell culture studies. (A (i & ii)) After
dissolution test, SEM-EDAX analysis suggested the formation of NaCl (*) on the plain Ti-scaffolds and on CaP coatings produced at 1.54 mA cm2, whereas rod-shape crystals (white
triangle) were observed on the coating produced at 5 mA/cm2 (see inset of A(iii)). (A viex)) After culture, fibroblastic cells were observed on the coatings, together with the
formation of filamentous carbonated apatite-like CaP crystals and agglomerates (dark triangles, as suggested by EDAX spectra showing increased in carbon (C) and oxygen (O)
contents) on the coatings produced at 5, 10 and 20 mA/cm2 (see insets of A (viii & ix)). (B) mCT quantification showed a reduction in the frequency of the primary coating thickness
produced at 5, 10 and 20 mA/cm2 after dissolution, and a further reduction in the thickness after in vitro cell culture with the appearance of a peak in the region of higher coating
thickness (50 mm, indicated by the arrows). No change in coating thickness distribution for the coating deposited at 1.54 mA/cm2 under different conditions. C, cells; Ca, calcium;
Na, sodium; P, phosphorus; Cl, chloride; Ti, titanium; V, vanadium; Al, aluminium; Au, gold. Scale bar ¼ 20 mm.

differentiation, was upregulated by cells seeded on all produced (OCN) and osteopontin (OPN) were significantly down- (p < 0.01)
hybrids, where the expression level by cells seeded on hybrids and upregulated (p < 0.05) by cells seeded on hybrids produced at
produced at 1.54 mA/cm2 (p < 0.05) was significantly higher than 10 mA/cm2, respectively (Fig. 8 (iii & iv)). However, no significant
that of the negative controls (Fig. 8 (ii)). In addition, osteocalcin differences in Col-1 and BMP-2 transcripts were observed between
 4052
Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
Fig. 6. Cell seeding and growth of hPDCs on CaP-Ti6Al4V hybrids. (A & B) Cell seeding efficiencies and the corresponding metabolic activities of hPDCs seeded on non pre-wetted and pre-wetted hybrids using static and dynamic seeding
methods. (C) Proliferation of hPDCs seeded dynamically on pre-wetted scaffolds as suggested by the metabolic activity measurement and (D) the corresponding DNA content after 21 days of culture. (E) Merged fluorescence images
showing living (green) and dead (red) cells on different hybrids after 21 days of culture. Various intensities of orange background were due to the reflection of red fluorescence by the CaP coatings with different thickness or by the
mineralised matrix. Mean  S.D. (n ¼ 3). Unpaired t-test (two-tailed): *p < 0.05, #p < 0.01 and xp < 0.001. CSE, cell seeding efficiency; GM, cell seeded on plain Ti-scaffold and cultured in growth medium; OM, cell seeded on plain Ti-
scaffold and cultured in osteogenic medium. Scale bar ¼ 500 mm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
Fig. 7. ALP activity of hPDCs seeded on the produced CaPeTi hybrids after 21 days of
culture. A significant reduction in ALP activity was observed by cells seeded on the
coatings deposited at 5 and 10 mA/cm2 as compared to cells seeded on plain Ti-
scaffolds (represented by GM). Osteogenic medium (OM) induced a significant ALP
activity on cells seeded on plain Ti-scaffolds. Mean  S.D. (n ¼ 3). Unpaired t-test (two-
tailed): *p < 0.05, #p < 0.01.
cells seeded on the hybrids and the negative controls (Fig. 8 (v & vi).
Surprisingly, osteoprotegerin (OPG) e an osteoclastogenesis
inhibitory factor, was significantly down-regulated by cells seeded
on all produced hybrids, this expression was 3.4e6.7 fold lower
than that of the negative controls (Fig. 8 (vii)). Matrix Gla protein
(MGP), a calcification inhibitor, was also down-regulated by cells
seeded on all produced hybrids, although the expression levels
were not significantly different (Fig. 8 (viii)).
3.7. Optimisation of cell seeding efficiency (CSE) for in vivo ectopic
bone formation assay
Fig. 9A shows the CSE and the corresponding DNA content of
seeding 1, 2, 3 and 5 million hPDCs dynamically onto the hybrid
(type A scaffolds) produced at 5 and 10 mA/cm2. The results
showed that an increase in the cell seeding number decreased the
CSE on both scaffold lengths, where seeding 5 million cells gave rise
to a significantly lower CSE than seeding 1 million cells. However,
quantification of DNA content on the cell-seeded hybrids showed
that the DNA content increased significantly when >2 million cells
were seeded. Considering both CSE and DNA content data by
plotting (CSE  DNA) with the cell seeding numbers, it was clearly
shown that a cell seeding number of 3 million was the optimum for
both hybrids produced at 5 and 10 mA/cm2 (Fig. 9B). Fig. 9C shows
the metabolic activity per DNA of hybrids seeded with 3 million
hPDCs prior to implantation, where cells seeded on hybrids
produced at 5 mA/cm2 had a higher metabolic activity than those
produced at 10 mA/cm2. However, the metabolic activities were not
significantly different (p > 0.05). This was further illustrated by the
qualitative live-dead staining showing differences in cell viability
and distribution between both hybrids (Fig. 9D).
3.8. Ectopic bone formation of the hPDCs seeded CaPeTi hybrids
After 8 weeks of subcutaneous implantation, by nanoCT and
histological analysis, no bone formation was observed in any of the
unseeded CaPeTi hybrids and uncoated Ti-scaffolds (data no
shown). For hybrids seeded with 1 million hPDCs, histological
analysis showed ectopic bone formation within hybrids produced
4053
at 10 mA/cm2 after 8 weeks of subcutaneous implantation, where
small amounts of new bone spicules (B, with the presence of
osteocytes (:)) were observed adjacent to the CaP coating (Fig. 10A
(iii & iv)). However, no bone formation was observed within hybrids
produced at 5 mA/cm2. Instead abundant multi-nucleated giant
cells (indicated by white arrows) were found in close proximity to
the CaP coating (Fig. 10A (i & ii)). It was noted that the CaP coatings
remained apparent on the surface of the scaffold struts after
implantation, which did not stain with the histological stains. Soft
tissue (st) infiltration was observed within all explants, together
with the presence of adipose tissue (ft) which was mostly found
within the 2 mm centre hollow of the scaffolds. Surprisingly, by
seeding 3 million hPDCs, bone formation was observed throughout
the hybrids produced at 10 mA/cm2 (Fig. 10B (vii e xii). The newly
formed bone (B) was in direct contact [indicated by white arrow
head in Fig. 10B (viii)] with the surface of the struts. Interestingly,
some remnants of the CaP coating (*) were found to be surrounded
by the newly formed bone matrix and had no binding affinity to the
histological stains. By seeding 3 million hPDCs onto hybrids
produced at 5 mA/cm2, abundant multi-nucleated giant cells were
also found near to the coatings (Fig. 10B (ieiv)). These CaP coatings
had strong affinity to the histological stains (strongly stained in red
colour), and some bone spicules were observed especially at the
bottom part of the hybrids (Fig. 10B (vevi)). Of note, blood vessels
(bv) were observed within all explants as shown by Fig. 10B (iii).
Fig. 10C (i) shows a representative reconstructed nanoCT image
(in pseudo-colour) indicating bone formation (B, in orange colour)
within a hybrid produced at 10 mA/cm2. By applying a global
thresholding, the newly formed bone (B, in brown colour) was
segmented from the coating [indicated by D in the inset of Fig. 10C
(ii)] and the absolute volumes of bone and coating, as well as the
percentage of bone formed within the available internal volume of
a hybrid were quantified. The results showed that hybrids produced
at 10 mA/cm2 induced 11.67  0.96% (corresponding to
6.42  0.53 mm3 of bone) of bone formed within the available
internal hybrid volume, which was significantly higher than for
hybrids produced at 5 mA/cm2 (i.e. 7.00  1.34% or
3.92  0.65 mm3) (Fig. 10C (ii)).
4. Discussions
In this study, we have demonstrated the feasibility of depositing
CaP layers with different physicochemical properties onto the
surface of porous, complex 3D Ti6Al4V scaffolds, by performing 6P-
ELD at different current densities under the optimised deposition
conditions. This methodology offers controllability and reproduc-
ibility over the coating distribution, thickness, surface morphology
and ion dissolution kinetics. Two main issues were encountered
with the use of the developed 6P-ELD system. Firstly, pH shift of the
bulk electrolyte (the supersaturated calcium phosphate solution) to
acidic condition due to the high electric current used or prolonged
electrolysis reactions, and also a direct result of the consumption of
hydroxyl ion (OH) during CaP coating (i.e. hydroxyapatite) depo-
sition. This acidic condition may cause re-dissolution of the
deposited CaP coating during 6P-ELD, although an alkaline envi-
ronment was generated near to the electrodes that facilitate CaP
deposition. This may explain the levelling off of the deposited
coating weight when high current density (>20 mA/cm2) was
applied. Other factors such as excessive hydrogen gas production
and shielding of the electric field due to local partial clogging of
scaffold pores by the deposited thick coatings, may have also dis-
rupted further coating deposition. This resulted in non-linear
correlation between the measured thickness and weight of
coating produced at 20 mA/cm2, and indicated a controlled depo-
sition process was obtained within the range of 5e10 mA/cm2.
 4054
Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
Fig. 8. Quantitative PCR analysis of osteogenic gene markers expression by hPDCs seeded on the produced hybrids and cultured for 21 days. The y-axis represents the relative expression (2DCT) normalised to the expression level of the
housekeeping gene b-Actin. Runx2, runt-related transcription factor; Osx, osterix; OCN, osteocalcin; OPN, osteopontin; Col-1, collagen type-1; BMP-2, bone morphogenetic protein-2; OPG, osteoprotegerin; MGP, matrix gla protein.
Mean  S.D. (n ¼ 3).
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058 4055

Fig. 9. Optimisation of cell seeding efficiency (CSE) on hybrids produced at 5 and 10 mA/cm2 (Type B scaffolds) for in vivo ectopic bone formation assay. (A) Cell seeding efficiencies
and the corresponding DNA content of cell-seeded hybrids. (B) Graph showing three millions of cell seeding number was the optimal for both hybrids. (C) Metabolic activities per
DNA of three millions cells seeded dynamically on the hybrids and (D) the representative images after live-dead staining pre-implantation. Mean  S.D. (n ¼ 3). Unpaired t-test
(two-tailed): *p < 0.05, #p < 0.01 and xp < 0.001. CSE, cell seeding efficiency; Ti, Ti-scaffold; C, cells. Scale bar ¼ 1 mm.

Moreover, continuously perfusing an acidic electrolyte may alter
the physicochemical properties of the deposited coatings and thus
the subsequent biological responses. To minimise any negative
impact on the biological behaviour, thorough rinsing in deminer-
alised water was performed to remove any acidic residues that
might be present on the produced hybrids. Secondly, the produced
coatings are subject to delamination when the coating thickness
increases, which was determined by the applied current density
and deposition time [16]. Indeed, applying high current densities
(e.g. at 10 and 20 mA/cm2) produced thick coatings with superficial
layers that had relatively poor adhesion properties. In contrast,
coatings that were deposited at low current densities (e.g. at 1.54 or
5 mA/cm2) had better adhesion properties, but they were either
not osteoinductive or had a delayed onset of bone formation. This
4056 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058

Fig. 10. Ectopic bone forming capacity of the hPDC-seeded CaPeTi hybrids. (A) By seeding 1 million hPDCs, a small amount of bone spicules (B) were formed within the hybrids
produced at 10 mA/cm2 (iii & iv), whereas abundant multi-nucleated giant cells (indicated by white arrows) were observed within the hybrids produced at 5 mA/cm2 without any
bone formation (i & ii). (B) By seeding 3 million hPDCs, bone formation was observed throughout the hybrids produced at 10 mA/cm2 (viiexii) and some bone spicules were
observed especially at the bottom part of the hybrids produced at 5 mA/cm2 (vevi). Some CaP coatings might be converted to biological apatite (Apt) as shown by its’ high affinity to
the histological stain (i e iv). (C) (i) Representative reconstructed nanoCT image (in pseudo-colour) showing bone formation (B, in orange colour) within the hybrids produced at
10 mA/cm2 (ii) By applying a global thresholding, bone (B, in brown colour) was segmented from the coating (indicated by D in the inset), and (iii) quantified as the absolute bone
volume or percentage of bone formed in the available internal volume of a hybrid. Ti, scaffold strut; CaP, coating; Apt, apatite; st, soft tissue; ft, adipose tissue; bv, blood vessels;
mm3, cubic millimetre. Scale bar ¼ 100 mm. Mean þ S.D. (n ¼ 3). Unpaired t-test (two-tailed): *p < 0.05, #p < 0.01. Stained with Stevenels’ Blue & Van Gieson’s Picrofuchsin, or
Toluidine Blue & Basic Fuchsin. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
intimates that further fine-tuning the deposition parameters in
order to deposit CaP coatings with optimum material integrity and
osteoinductivity is required. For example, by performing 6P-ELD at
an intermediate current density (such as 7.5 mA/cm2) or through
the co-deposition of other materials (such as collagen fibres [17]
and chitosan [11]) as binders, an improvement in both the
coating property and the osteogenicity may be achieved.
Nevertheless, the key determinant/s of bone forming capacity
for an “electrodeposited” CaP coating on Ti-scaffold is/are currently
not well defined. Therefore, we carried out a series of analyses to
gain further insight into the in vitro biological interactions between
a clinically-relevant human MSC-like osteoprogenitor and the
produced hybrids with distinctive physicochemical properties, in
order to correlate to their in vivo ectopic bone forming capacities.
Interestingly, our current findings showed that hybrids with high-
est Ca2þ dissolution kinetic (produced at 10 mA/cm2) and
possessed a distinctive coating surface topography, have elicited
specific in vitro cellular events and carbonated apatite formation, as
well as induced ectopic bone formation in combination with
seeded hPDCs in a cell density-dependent manner. Whereas,
hybrids with intermediate physicochemical properties (i.e.
produced at 5 mA/cm2) induced a delayed onset of ectopic bone
formation, which also required the use of a higher dose of hPDCs.
This cell density-dependent ectopic bone forming capacity of the
produced hybrids may be explained by the relatively low cell
proliferation on the deposited coatings as shown by the in vitro
culture assays. Interestingly, all the deposited coatings were iden-
tified as carbonated hydroxyapatite with a Ca/P ratio of w1.42,
which did not change significantly in function of current density
applied as reported in our previous study [10]. In this study, hybrids
produced at 20 mA/cm2 were excluded for in vivo implantation due
to unfavourable cell growth, whereas hybrids produced at 1.54 mA/
cm2 were also excluded due to low Ca2þ release and thin coating
although favourable cell growth was observed.
In this study, we reported mCT as a non-invasive tool for 3D
characterisation of the coating thickness, as with microscopy only
2D measurements can be made. This also eliminates certain errors
due to sectioning such as loosening of the coating from the metallic
structure or cracks in the coating. However, we are well aware of
the artefacts (i.e. partial volume effect and scatter) inherently
present in the mCT images. Therefore, the coating thickness was
calculated by subtracting the assumed baseline due to these arte-
facts (i.e. by measuring the thickness of the edge effects in non-
coated Ti-scaffolds ¼ reference measurement) from the total
measured coating thickness, resulting in a potential underestima-
tion of the thickness. For thinner coating thickness (<15 mm), the
spatial image resolution was not sufficient and thus the coating
thickness might not be measured correctly. Hence, to assess the
accuracy of mCT as a non-invasive tool for precise quantification of
CaP coating thickness, further validation is required. This could be
achieved by comparing uCT data to reference measuring tech-
niques, such as scanning electron microscopy, or by using stan-
dardized phantoms with a known coating thickness.
It should be highlighted that the use of PBS for coating disso-
lution testing may not be appropriate (although it is a common
protocol besides using simulated body fluid), as the buffering effect
of PBS prevented the precise measurement of PO3 4 dissolution. This
resulted in lack of experimental data to interpret the potential
effect/s of released PO3
4 on in vitro and in vivo biological outcomes,
as PO34 was described to be an important signalling molecule in
osteoblast differentiation [18]. However, the release of Ca2þ was
measured using PBS, where the released level was specific for each
produced hybrid and the coating that had the highest Ca2þ release
induced in vivo ectopic bone formation. This further consolidated
the findings of several studies showing a high correlation between
4057
a specific Ca2þ dissolution kinetic and in vivo osteoinductivity
[19,20]. Perhaps, a continuous release of Ca2þ (at appropriate rate
and concentration) may provide long-term stimulation that initi-
ates essential cellular osteogenic events and thus ectopic bone
formation [21]. Surprisingly, this ion dissolution behaviour did not
occur within the in vitro culture system. In fact, reductions in Ca2þ
and PO3 4 concentrations in culture medium were measured,
possibly due to the formation of carbonated apatite on the surface
of the coatings (which was found to be a cell-mediated phenom-
enon by SEM and mCT quantification of coating thickness distribu-
tion after culture). The formation of biological apatite may occur
in vivo upon implantation of the hybrids, and in fact re-
precipitation of the released Ca2þ and PO3 4 to form biological
apatite was hypothesised to be one of the possible mechanisms of
action for CaP osteoinduction [22]. This phenomenon may be
related to the observation of CaP coatings that had high affinity to
the histological stains, where the deposited CaP coatings might
dissolve and convert to biological apatite.
The in vitro culture studies showed the benefits of combining
pre-wetting and dynamic seeding methods to improve cell seeding
efficiency and cell viability on the CaP-coated scaffolds. This may be
due to the adsorption of serum proteins onto the coatings, which
has been reported to be critical for cell attachment and survival on
CaP substrate [23], and/or an increase in cell-material contacts by
the dynamic seeding (thus increased cell attachments). However,
subsequent culturing in GM under static condition has led to a low
cell growth on thick coatings, i.e. produced at 10 and 20 mA/cm2,
which may be associated to the specific coating surface topography
and chemistries, or cell loss due to the detachment of loose surface
CaP layers during culture. Interestingly, the alkaline phosphatase
activities for cells seeded on hybrids produced at 5 and 10 mA/cm2
were significantly reduced, which was in contrast to other findings
[24,25]. This may represent a negative feedback mechanism of
hPDCs towards the Ca2þ and PO3 4 releases by the produced hybrids
for the regulation of matrix mineralisation [26]. Based on the gene
analysis data, osteogenic differentiation of hPDCs was observed on
the produced hybrids, as indicated by the upregulation of Runx2,
Osx and OPN. However, the effects were not conclusive. Most
encouragingly, culturing hPDCs on the produced hybrids induced
a significant down-regulation of osteoprotegerin (OPG) expression.
This may be a response to CaP coatings by hPDCs to initiate coating
resorption through the activation of osteoclastogenesis [27]. This
was supported by histological analysis, where abundant multi-
nucleated giant cells were found in close proximity to the CaP
coatings deposited at 5 mA/cm2 after 8 weeks of implantation, but
to a lower extent for coatings deposited at 10 mA/cm2. We believe
the osteoclastic resorption event may have had occurred earlier,
which gave rise to a faster onset of ectopic bone formation by the
hybrids produced at 10 mA/cm2. Essentially, the gene expression of
RankL needs to be assessed in order to determine the change in
RankL/OPG balance to confirm an active recruitment of osteoclasts
by hPDCs for CaP coating resorption [28]. Additionally, the
produced hybrids may promote mineralisation activity by hPDCs,
as indicated by the down-regulation of Matrix Gla protein (MGP,
a mineralisation inhibitor [29]) gene expression, although the effect
was not significant.
As indicated by qualitative histological analysis, the observed
ectopic bone induction by the hybrids was highly dependent on the
physicochemical properties of the coating and the cell seeding
density. Using a quantitative nanoCT imaging based approach, the
3D quantified bone volumes and percentages were under-
estimated, as the thresholding used was taken conservatively as it
excluded all structures close to the surface of Ti struts that could
still be regarded as coatings remnants. This underestimation was
clearly visible for the hybrids produced at 10 mA/cm2, as some of
4058 Y.C. Chai et al. / Biomaterials 33 (2012) 4044e4058
the newly formed bones were found in direct contact with the Ti-
substrate in the histological sections. This direct contact between
the newly formed bone and the Ti-substrate provided evidence on
the biocompatibility and potentially high osseointegration of the
produced hybrids with the host system. Of noted, abundant blood
vessels were observed within the explants either indicating
a minimal effect of the produced thick coatings (reduced pore size)
on host blood vessels in-growth, or the produced hybrids may have
an effect on new blood vessels formation. The latter effect can be
confirmed by the assessment of the angiogenic activity of the
produced hybrids, such as the detection of vascular endothelial
growth factor (VEGF) gene expression and protein secretion by
hPDCs in response to the deposited CaP coatings [30]. Finally, we
believe that the produced hybrids have great potential to be used
for clinical repair of skeletal defects, as the assessment of osteoin-
duction at ectopic sites has been reported to be of relevance to an
orthotopic implantation [31]. However, the ectopic bone formation
assay performed within this study was only a two-month end point
study. The potential use of the produced hybrids with a delayed
onset should not be jeopardised, as implantation for a longer time
point may give no difference in the bone forming capacity by the
different produced hybrids, which was frequently reported in
literature [4].
5. Conclusion
By using a 6-channel perfusion electrodeposition system (6P-
ELD), we have demonstrated the feasibility to produce 3D porous
CaPeTi hybrids with distinctive physicochemical properties that
have significant impact on the behaviour of in vitro human osteo-
progenitor cells and in vivo ectopic bone forming capacities. The
produced CaP coatings did not support ALP activity but may have
possibly supported osteoclastogenesis. We provided further
knowledge on the essential stem cell-CaP biomaterial interactions,
which are critical to fine-tuning the electrodeposition parameters
in the future, in order to produce in vitro bone reparative units with
predictive osteoinductivity and superior biomechanical properties
for a successful repair of skeletal defects.
Acknowledgement
The authors thank Carla Geeroms, Joop Van Deursen and Ing.
Johan Vanhulst for the excellent technical assistances on micro-
computer tomography analysis of bone formation and the devel-
opment of the six-channel perfusion electrodeposition system. This
work is funded by the KU Leuven IDO project 05/009 e QuEST, the
Stem Cell Institute of Leuven e KU Leuven, the Agency for Inno-
vation by Science and Technology in Flanders: IWT/OZM/090655,
and is part of Prometheus, the Leuven Research & Development
Division of Skeletal Tissue Engineering of the KU Leuven: www.
kuleuven.be/prometheus.
References
[1] Giannoudis PV, Einhorn TA, Marsh D. Fracture healing: the diamond concept.
Injury 2007;38(Suppl. 4):S3e6.
[2] Calori GM, Giannoudis PV. Enhancement of fracture healing with the diamond
concept: the role of the biological chamber. Injury 2011;42:1191e3.
[3] Beaman FD, Bancroft LW, Peterson JJ, Kransdorf MJ. Bone graft materials and
synthetic substitutes. Radiol Clin North Am 2006;44:451e61.
[4] Barradas AM, Yuan H, van Blitterswijk CA, Habibovic P. Osteoinductive
biomaterials: current knowledge of properties, experimental models and
biological mechanisms. Eur Cell Mater 2011;21:407e29. discussion 29.
[5] León B, Jansen J. Thin calcium phosphate coatings for medical implants. New
York: Springer; 2009.
[6] Habibovic P, Li J, van der Valk CM, Meijer G, Layrolle P, van Blitterswijk CA,
et al. Biological performance of uncoated and octacalcium phosphate-coated
Ti6Al4V. Biomaterials 2005;26:23e36.
[7] Lopez-Heredia MA, Weiss P, Layrolle P. An electrodeposition method of
calcium phosphate coatings on titanium alloy. J Mater Sci Mater Med 2007;18:
381e90.
[8] Lopez-Heredia MA, Sohier J, Gaillard C, Quillard S, Dorget M, Layrolle P. Rapid
prototyped porous titanium coated with calcium phosphate as a scaffold for
bone tissue engineering. Biomaterials 2008;29:2608e15.
[9] Chai YC, Roberts SJ, Desmet E, Kerckhofs G, van Gastel N, Geris L, et al.
Mechanisms of ectopic bone formation by human osteoprogenitor cells on
CaP biomaterial carriers. Biomaterials 2012;33:3127e42.
[10] Chai YC, Truscello S, Bael SV, Luyten FP, Vleugels J, Schrooten J. Perfusion
electrodeposition of calcium phosphate on additive manufactured titanium
scaffolds for bone engineering. Acta Biomater 2011;7:2310e9.
[11] Wang J, de Boer J, de Groot K. Preparation and characterization of electro-
deposited calcium phosphate/chitosan coating on Ti6Al4V plates. J Dent Res
2004;83:296e301.
[12] Van Bael S, Kerckhofs G, Moesen M, Pyka G, Schrooten J, Kruth JP. Micro-CT-
based improvement of geometrical and mechanical controllability of selective
laser melted Ti6Al4V porous structures. Mater Sci Eng A 2011;528:7423e31.
[13] Eyckmans J, Roberts SJ, Schrooten J, Luyten FP. A clinically relevant model of
osteoinduction: a process requiring calcium phosphate and BMP/Wnt sig-
nalling. J Cell Mol Med 2010;14:1845e56.
[14] Chai YC, Roberts SJ, Van Bael S, Chen Y, Luyten FP, Schrooten J. Multi-level
factorial analysis of Ca(2þ)/P(i) supplementation as bio-instructive media for
in vitro biomimetic engineering of three-dimensional osteogenic hybrids.
Tissue Eng Part C Methods 2012;18:90e103.
[15] Chai YC, Roberts SJ, Schrooten J, Luyten FP. Probing the osteoinductive effect
of calcium phosphate by using an in vitro biomimetic model. Tissue Eng Part A
2011;17:1083e97.
[16] Sharma S, Soni VP, Bellare JR. Electrophoretic deposition of nano-
biocomposites for orthopedic applications: influence of current density and
coating duration. J Mater Sci Mater Med 2009;20(Suppl. 1):S93e100.
[17] Miao S, Weng W, Li Z, Cheng K, Du P, Shen G, et al. Electrolytic deposition of
octacalcium phosphate/collagen composite coating on titanium alloy. J Mater
Sci Mater Med 2009;20:131e4.
[18] Beck Jr GR. Inorganic phosphate as a signaling molecule in osteoblast differ-
entiation. J Cell Biochem 2003;90:234e43.
[19] Yuan H, Fernandes H, Habibovic P, de Boer J, Barradas AM, de Ruiter A, et al.
Osteoinductive ceramics as a synthetic alternative to autologous bone graft-
ing. Proc Natl Acad Sci U S A 2010;107:13614e9.
[20] Roberts SJ, Geris L, Kerckhofs G, Desmet E, Schrooten J, Luyten FP. The
combined bone forming capacity of human periosteal derived cells and
calcium phosphates. Biomaterials 2011;32:4393e405.
[21] Nakamura S, Matsumoto T, Sasaki J, Egusa H, Lee KY, Nakano T, et al. Effect of
calcium ion concentrations on osteogenic differentiation and hematopoietic
stem cell niche-related protein expression in osteoblasts. Tissue Eng Part A
2010;16:2467e73.
[22] Habibovic P, Sees TM, van den Doel MA, van Blitterswijk CA, de Groot K.
Osteoinduction by biomaterialsephysicochemical and structural influences.
J Biomed Mater Res A 2006;77:747e62.
[23] Tsang EJ, Arakawa CK, Zuk PA, Wu BM. Osteoblast interactions within
a biomimetic apatite microenvironment. Ann Biomed Eng 2011;39:
1186e200.
[24] Wang J, de Boer J, de Groot K. Proliferation and differentiation of MC3T3-E1
cells on calcium phosphate/chitosan coatings. J Dent Res 2008;87:650e4.
[25] Richard D, Dumelie N, Benhayoune H, Bouthors S, Guillaume C, Lalun N, et al.
Behavior of human osteoblast-like cells in contact with electrodeposited
calcium phosphate coatings. J Biomed Mater Res B Appl Biomater 2006;79:
108e15.
[26] Chang YL, Stanford CM, Keller JC. Calcium and phosphate supplementation
promotes bone cell mineralization: implications for hydroxyapatite (HA)-
enhanced bone formation. J Biomed Mater Res 2000;52:270e8.
[27] Detsch R, Mayr H, Ziegler G. Formation of osteoclast-like cells on HA and TCP
ceramics. Acta Biomater 2008;4:139e48.
[28] Barbosa MA, Monteiro FJ, Correia R, Leon B. Quantitative analysis of osteo-
protegerin and RANKL expression in osteoblast grown on phosphate ceramics.
Key Eng Mater 2004;254-256:713e6.
[29] Roy ME, Nishimoto SK. Matrix Gla protein binding to hydroxyapatite is
dependent on the ionic environment: calcium enhances binding affinity but
phosphate and magnesium decrease affinity. Bone 2002;31:296e302.
[30] Barralet J, Gbureck U, Habibovic P, Vorndran E, Gerard C, Doillon CJ. Angio-
genesis in calcium phosphate scaffolds by inorganic copper ion release. Tissue
Eng Part A 2009;15:1601e9.
[31] Habibovic P, Yuan H, van den Doel M, Sees TM, van Blitterswijk CA, de
Groot K. Relevance of osteoinductive biomaterials in critical-sized orthotopic
defect. J Orthop Res 2006;24:867e76.