University of ConnecticutOpenCommons@UConn
Doctoral Dissertations University of Connecticut Graduate School
1-12-2018
Fabrication and Characterization of Collagen-Apatite Composites for Bone Tissue EngineeringChangmin HuUniversity of Connecticut - Storrs, [email protected]
Follow this and additional works at: https://opencommons.uconn.edu/dissertations
Recommended CitationHu, Changmin, "Fabrication and Characterization of Collagen-Apatite Composites for Bone Tissue Engineering" (2018). DoctoralDissertations. 1695.https://opencommons.uconn.edu/dissertations/1695
Fabrication and Characterization of Collagen-Apatite Composites
for Bone Tissue Engineering
Changmin Hu, PhD
University of Connecticut, 2018
Abstract
Due to the high demand of scaffolds for treating bone fractures every year,
biomimetic coatings and scaffolds resembling the composition and structure of natural
bone are of keen interest of bone tissue engineers. In this research, a series of
biomimetic coatings and scaffolds have been developed, including biomimetic apatite
coatings, biomimetic collagen-apatite composite coatings, intrafibrillar calcified
collagen fibrils, intrafibrillar silicified collagen fibrils, and intrafibrillar calcified
collagen scaffolds.
First, a crack-free apatite coating was deposited on the surface-treated Ti6Al4V
substrate using a biomimetic process, in which the surface-treated titanium substrate
was immersed in a modified simulated body fluid (m-SBF). Dual-beam focused ion
beam/scanning electron microscopy (FIB/SEM) was used to characterize the
cross-sectional microstructures of the ceramic coating. Cross-sectional SEM images
and TEM images revealed that crack-free apatite coatings have been achieved by
adjusting the pH of the m-SBF. The coating formed demonstrates a gradient porous
structure, which is dense close to the substrate and then it becomes more and more
porous towards the surface of the coating.
Second, a biomimetic collagen-apatite composite coating was also successfully
Changmin Hu-University of Connecticut, 2018
formed on the surface-treated Ti6Al4V alloy using collagen-containing m-SBF.
During the composite coating formation, collagen and apatite co-precipitate to form
the composite coatings. Dual-beam FIB/SEM was used to characterize the
cross-sectional microstructures of the composite coatings. The result indicates that the
cross-section of the collagen-apatite composite coating also exhibits a gradient porous
structure, and that collagen-apatite composite coating is thinner and less porous than
the pure apatite coating.
Third, inspired by the structure of natural bone, collagen fibrils with intrafibrillar
calcification were prepared for bone tissue engineering. Poly(acrylic acid) (PAA) was
used as a sequestration analog of non-collagenous proteins (NCPs) for stabilizing
amorphous calcium phosphate (PAA-ACP) nanoprecursors to infiltrate into collagen
fibrils. Additionally, sodium tripolyphosphate (TPP) was applied as a templating
analog of NCPs to regulate the orderly deposition of minerals within the gap zone of
collagen fibrils. The effect of PAA concentration on the intrafibrillar mineralization of
reconstituted collagen fibrils was also investigated.
Fourth, silicon is an essential element contributing significantly to the health of
bone, and great efforts have been made to produce silica-containing biomaterials.
Poly (allylamine) hydrochloride (PAH) was used as an analog of zwitterionic proteins
to stabilize silicic acid (SA) to produce fluidic silica (PAH-SA) nanoprecursors, and
TPP was used as a templating analog of zwitterionic proteins. Silicified collagen
fibrils with core-shell, twisted and banded structures were produced by controlling the
zwitterionic proteins analogs. Intrafibrillar silicified collagen fibrils were produced
Changmin Hu-University of Connecticut, 2018
using PAH-SA as fluidic silica nanoprecursors and TPP as the templating analog to
modulate the deposition of silica within the gap zone of collagen fibrils.
Fifth, with the success in reproducing intrafibrillar mineralized collagen fibrils,
biomimetic collagen/apatite scaffolds consisting of intrafibrillar mineralized collagen
fibrils were prepared via a bottom-up approach, resembling the composition and
structure of natural bone from the nanoscale to the macroscale. At the macro-scale,
unidirectional macro-pores are aligned across the entire scaffold, and at the
micro-scale, each layer is comprised of aligned and well stacked lamella. Finally, at
the nano-scale, apatite minerals deposit within the gap zone of collagen fibrils and
orient along the long axis of these fibrils. The biomimetic collagen-apatite scaffolds
demonstrate a good biocompatibility in vitro, indicating a great potential to be used
for bone tissue engineering applications.
In summary, a novel methodology has been developed for the preparation and
characterization of biomimetic coatings and scaffolds with a hierarchical structure,
demonstrating a great potential for bone tissue engineering.
i
Fabrication and Characterization of Collagen-Apatite Composites
for Bone Tissue Engineering
Changmin Hu
B.S., Jinan University, 2010
M.S., Shanghai Jiao Tong University, 2013
A Dissertation
Submitted in Partial Fulfillment of the
Requirements for the Degree of
Doctor of Philosophy
at the
University of Connecticut
2018
ii
Copyright by
Changmin Hu
2018
iii
APPROVAL PAGE
Doctor of Philosophy Dissertation
Fabrication and Characterization of Collagen-Apatite Composites
for Bone Tissue Engineering
Presented by
Changmin Hu, B.S., M.S.
Major Advisor ______________________________________________________
Mei Wei
Associate Advisor ___________________________________________________
Mark Aindow
Associate Advisor ___________________________________________________
Yu Lei
University of Connecticut
2018
iv
Acknowledgement
First and foremost, my special and greatest thanks go to my advisor Prof. Mei
Wei for her invaluable guidance and support though my Ph.D. study and research.
Prof. Wei is a great advisor, she not only teaches me how to do research, but also
helps me with my future development. She is very supportive when I have a chance to
develop my career in industry. She cares about my research, study, life, and future
development. I would want to express my greatest appreciation to her.
I would also like to thank my associate advisors: Prof. Mark Aindow and Prof. Yu
Lei for their valuable comments, suggestions and help for my research projects. I
would want to thank my committee members: Prof. Seok-woo Lee and Prof.
Montgomery Shaw for their support of my work. I am also grateful for my previous
committee members: Prof. Harold D. Brody and Prof. Luyi Sun for their suggestions
and comments for my research. I would also want to thank our collaborators Prof.
David Rowe and Dr. Liping Wang from the University of Connecticut Health Center
(UCHC) for all their help.
I would like to thank Dr. Lichun Zhang and Dr. Roger Ristau for their technical
assistance on electron microscopy. I would also like to thank my previous and current
group members: Dr. Zengmin Xia, Dr. Erica Kramer, Dr. Max Villa, Dr. Michael Zilm,
Mr. Jonathan Russo, Dr. Stephanie Bendtsen, Mr. Drew Clearfield, Mr. Bryant
Heimbach, and Mr. Le Yu. Especially, I would want to thank Zengmin Xia. She is so
nice and helpful. She not only helped me with my research, but also helped me to
adapt to the new environment.
v
Last but not least, I would like to thank my parents and my sister for their
unconditioned support of all my choices.
vi
Table of Contents
1. Introduction ........................................................................................................... 1
1.1 Background ...................................................................................................... 1
1.2 Structures and composition of natural bone ...................................................... 2
1.3 Intrafibrillar mineralization .............................................................................. 3
1.4 Implant coating ................................................................................................ 5
1.5 Dual beam focused ion beam/scanning electron microscopy (FIB/SEM) .......... 5
2. Cross-sectional microstructure study of apatite coating on titanium alloy............. 19
2.1 Introduction.................................................................................................... 20
2.2 Materials and methods.................................................................................... 23
2.2.1 Biomimetic HA coating preparation ......................................................... 23
2.2.2 Microstructural characterization ............................................................... 24
2.3 Results and discussion .................................................................................... 26
2.4 Conclusions.................................................................................................... 39
3. Cross-sectional microstructure study of collagen-apatite composite coating on
titanium alloy .......................................................................................................... 47
3.1 Introduction.................................................................................................... 48
3.2 Materials and methods.................................................................................... 50
3.2.1 Biomimetic Col-HA coating formation..................................................... 50
3.2.2 Materials characterization ........................................................................ 52
3.3 Results and discussion .................................................................................... 54
3.4 Conclusions.................................................................................................... 66
4. Fabrication of intrafibrillar calcified collagen fibrils ............................................ 74
4.1 Introduction.................................................................................................... 75
vii
4.2 Materials and methods.................................................................................... 78
4.2.1 Materials .................................................................................................. 78
4.2.2 Preparation of intrafibrillar and extrafibrillar mineralized collagen fibrils 78
4.2.3 Characterization ....................................................................................... 79
4.3 Results ........................................................................................................... 80
4.4 Discussion ...................................................................................................... 84
4.5 Conclusions.................................................................................................... 89
5. Intrafibrillar silicification of collagen fibrils ........................................................ 97
5.1 Introduction.................................................................................................... 98
5.2 Materials and methods...................................................................................100
5.2.1 Materials .................................................................................................100
5.2.2 Preparation of hydrolyzed silicic acid ......................................................100
5.2.3 Preparation of PAH-stabilized silica (PAH-SA) precursors ......................100
5.2.4 Controlled biomimetic silicification of collagen fibrils via an OCSS
approach ..........................................................................................................101
5.2.5 Preparation of silicified collagen fibrils via a two-step approach .............102
5.2.6 Materials characterization .......................................................................103
5.2.7 In vitro cell culture and proliferation .......................................................104
5.2.8 Statistical analysis ...................................................................................105
5.3 Results and discussion ...................................................................................105
5.3.1 Col-SA fibrils ..........................................................................................105
5.3.2 Col-PAH-SA fibrils .................................................................................107
5.3.3 Col-inSA fibrils .......................................................................................109
5.3.4 Silicified collagen fibrils via a two-step approach ................................... 112
5.3.5 In vitro cell culture .................................................................................. 115
viii
5.4 Conclusions................................................................................................... 116
6. Intrafibrillar calcified collagen scaffolds .............................................................125
6.1 Introduction...................................................................................................127
6.2 Materials and methods...................................................................................131
6.2.1 Materials .................................................................................................131
6.2.2 Fabrication of extrafibrillar mineralized collagen-apatite (E-Col-Ap)
scaffolds ..........................................................................................................131
6.2.3 Fabrication of intrafibrillar and extrafibrillar mineralized collagen apatite
(IE-Col-Ap) scaffolds ......................................................................................132
6.2.4 Materials characterization .......................................................................133
6.2.5 In vitro cell culture ..................................................................................134
6.3 Results and discussions .................................................................................135
6.4 Conclusions...................................................................................................146
7. Conclusions and future work ..............................................................................159
7.1 Conclusions...................................................................................................159
7.2 Future work ...................................................................................................161
Appendice A: List of publications...........................................................................164
ix
List of Figures
Chapter 1
Figure 1.1. Hierarchical structures of natural bone: starting from sub-nanostructure of
collagen molecules and tropocollagen triple helix, to nanostructure of ordered HA
crystals, and to sub-microstructure of collagen fibrils and macrostructure (right) (a).
The arrangement of HA crystals with collagen fibrils.
Figure 1.2. The specimen is tilted towards the ion beam (left), a cross-section is
milled that can be imaged with the electron beam (right).
Chapter 2
Figure 2.1. SE SEM images of HA coatings formed at different initial pH values:
HA-6.4 (A), HA-6.5 (B), HA-6.6 (C), and HA-6.7 (D). The black arrows in (B)
indicate cracks formed on the surface of the coating.
Figure 2.2. XRD spectra of HA coatings on Ti-6Al-4V substrates after 24 h of
immersion in m-SBF.
Figure 2.3. Cross-sectional SE SEM images of HA coatings on Ti-6Al-4V substrates:
HA-6.5 (A) and HA-6.6 (B). White arrow in image (A) is a crack induced by the ion
beam of PFIB and black arrow in image (A) is a crack between the HA-6.5 coating
and the Ti-6Al-4V substrate.
Figure 2.4. Gas-assisted deposition of Pt layer at 8 kV and 30 kV ion beam voltage in
PFIB.
Figure 2.5. SE SEM images of HA-6.6 coating on Ti-6Al-4V substrate from
cross-sections prepared using GFIB (A, B) and PFIB (C, D). The cross-sections were
x
polished using CCS pattern (A, C) or Rec pattern (B, D). Cracks indicated by white
arrows in HA coating in image (C) were induced by the CCS process in PFIB.
Figure 2.6. An extraction sequence of a thin lamella from the bulk in GFIB (A-H).
The thickness of the TEM specimen was confirmed by STEM in GFIB (I).
Figure 2.7. The thin lamella from the site of interest before extraction (A) and after
mounted onto copper omni grids (B) in PFIB.
Figure 2.8. Data obtained from the TEM specimens prepared by GFIB: BF TEM
image (A) and HAADF STEM image (B). The inset to (A) is the SAED pattern from
the coating. C-F: Compositional maps obtained from the region in (B), showing the
distribution of Ca (C), P (D), Ti (E), and Ga (F).
Figure 2.9. Data obtained from the TEM specimens prepared by PFIB: BF TEM
image (A) and HAADF STEM image (B). C-F: Compositional maps obtained from
the region in (B), showing the distribution of Ca (C), P (D), Ti (E), and Xe (F).
Chapter 3
Figure 3.1. Illustration of the formation process of biomimetic Col-HA composite
coatings on the surface-treated Ti-6Al-4V substrates: Surface morphology of the
Ti-6Al-4V substrate after polishing using silicon carbide abrasive paper (A) and SLA
treatment (B), and coating formation by immersing surface-treated Ti-6Al-4V
substrate in the collagen-containing m-SBF solution (C).
Figure 3.2. SE SEM images of Col-HA composite coatings on Ti-6Al-4V substrates.
Col-HA coated Ti-6Al-4V alloy was titled 52° (C).
Figure 3.3. SE SEM images of the surface morphologies of HA (A) and Col-HA (B)
xi
coatings on Ti-6Al-4V substrates.
Figure 3.4. XRD (A) and ATR-FTIR (B) spectra of HA and Col-HA coated Ti-6Al-4V
substrates.
Figure 3.5. SE SEM images of the cross-sections of Col-HA composite coatings on
Ti-6Al-4V substrates prepared using GFIB (A1, A2, A3) and PFIB (B1, B2, B3). The
cross-sections were polished using Rec pattern (A1, B1), CCS pattern (A2, B2), and
RCS pattern (A3, B3). White arrows in B1, B2, and B3 indicate the cracks between Pt
layer and Col-HA coatings.
Figure 3.6. SE SEM images of the cross-sections of Col-HA (A) and HA (B) coatings
on the Ti-6Al-4V substrates.
Figure 3.7. TEM sample preparation sequence from Col-HA coated Ti-6Al-4V alloy
using GFIB (A-K). The TEM sample was confirmed to be electron transparent by
STEM in GFIB (L).
Figure 3.8. TEM sample preparation sequence of Col-HA coated Ti-6Al-4V alloy
using PFIB (A-K). Col-HA coated Ti-6Al-4V sample was mounted on TEM grid
holders for preparing a TEM sample using PFIB (L).
Figure 3.9. Characterization of the TEM specimens prepared by GFIB: BFTEM (A)
and HAADF STEM image (B). C-F: Compositional maps obtained from the region in
(B), showing the distribution of Ca (C), C (D), Ti (E), and Ga (F).
Figure 3.10. Characterization of the TEM specimens prepared by PFIB: BFTEM (A)
and HAADF STEM image (B). C-F: Compositional maps obtained from the region in
(B), showing the distribution of Ca (C), C (D), Ti (E), and Xe (F).
xii
Chapter 4
Figure 4.1. TEM images of unstained mineralized collagen fibrils with different PAA
concentrations: (A) 5.6 mg/mL, (B) 2 mg/mL (insert: SAED result), (C) 0.8 mg/mL
and (D) 0.4 mg/mL.
Figure 4.2. TEM image (A) and SEM image (B) of mineralized collagen fibrils
without PAA. SAED result for mineralized collagen fibrils without PAA (Insert, A).
White arrow in (B): HA minerals on the surface. The cracks on collagen fibrils were
induced by high voltage employed during SEM operation.
Figure 4.3. FESEM images of mineralized collagen fibril bundles with different PAA
concentrations (white arrows indicate D-banding): (A) 5.6 mg/mL, (B) 2 mg/mL, (C)
0.8 mg/mL (the inset is the enlarged image of the spot indicated by the white arrow),
and D) 0.4 mg/mL (white dotted arrow and white arrow heads show minerals on the
collagen surface), minerals aligned on the surface of collagen fibrils with an almost 67
nm spacing distance (white dotted arrow and thin arrows in inset image). The inset is
the enlarged image of the spot indicated by the white dotted arrow.
Figure 4.4. TEM images of collagen solution containing PAA-ACP nanoprecursor
droplets on copper grid. (A) Low magnification TEM image showing PAA-ACP
nanoprecursor droplets that were formed in the close vicinity of the collagen fibrils.
White arrow shows the collagen fibrils that have been removed. (B) & (C) High
magnification images of PAA-ACP nanoprecursor droplets. (C) One electron dense
xiii
PAA-ACP nanoprecursor droplet highlighted in the white circle consisting of many
amorphous calcium phosphate nanoparticles.
Figure 4.5. Schematic illustration of the intrafibrillar mineralization process induced
by PAA as a sequestration analog and TPP as a templating analog.
Chapter 5
Figure 5.1. TEM images (A1 and A2), FESEM images (B1 and B2) of Col-SA fibrils
with different SA/Col ratios: 3:1 (A1, B1), and 4:1 (A2, B2). ATR-FTIR spectra (C)
of Col-SA fibrils (Stars indicate the peaks for Si-O-Si, and triangle shows the peak for
Si-OH). Schematic illustration (D) of the process to form Col-SA fibrils, and the TEM
image shows Col-SA fibrils with a SA/Col ratio of 3:1.
Figure 5.2. TEM images of Col-PAH-SA fibrils with different SA/Col ratios: 1:1 (A),
4: 3 (B), and 2:1 (C). ATR-FTIR spectra (D) of Col-PAH-SA fibrils (Stars indicate the
peaks for Si-O-Si, triangle shows the peak for Si-OH, and arrow indicates amine
hydrogen bonding). Schematic illustration (E) of the process to form Col-PAH-SA
fibrils, and the TEM image shows Col-PAH-SA fibrils with a SA/Col ratio of 4:3.
Figure 5.3. TEM images (A1, B1, and C1) and FESEM images (A2, B2, and C2) of
Col-inSA fibrils with different SA/Col ratios: 1:1 (A1, A2), 4:3 (B1, B2), and 2:1 (C1,
C2).
Figure 5.4. ATR-FTIR spectra (A) and TGA spectra (B) of Col-inSA fibrils with
different SA/Col ratios. (Stars indicate the peaks for Si-O-Si, triangle shows the peak
for Si-OH, and arrow indicates amine hydrogen bonding.)
Figure 5.5. The initial stage of PAH-SA precursors interacted with collagen fibrils in
xiv
the presence of TPP as a regulator. Low (A) and high (B) magnification of the fluidic
silica precursors infiltrated into the collagen fibrils during the collagen self-assembly
process. White arrows in B: Vague D-banding can be observed.
Figure 5.6. Schematic illustration of the intrafibrillar mineralization process of
Col-inSA fibrils. TEM image shows Col-inSA fibrils with a SA/Col ratio of 4:3.
Figure 5.7. Col-SA fibrils produced using either a two-step process (T-Col-SA (A1),
T-Col-PAH-SA (B1), and T-Col-inSA (C1)) or a one-step approach (Col-SA (A2),
Col-PAH-SA (B2), and Col-inSA (C2)).
Figure 5.8. Proliferation of MC3T3-E1 cells on mineralized collagen membranes:
Collagen-apatite (Col-Ap, control), Col-SA (SA/Col=4:3), Col-PAH-SA
(SA/Col=4:3), and Col-inSA (SA/Col=4:3) over the course of 14 days. Data is
presented as mean ± standard error. Statistically significant differences (p) between
various groups were calculated using two-way RM ANOVA analysis of variance, and
*: p<0.05, **: p<0.01, ***: p<0.001.
Chapter 6
Figure 6.1. TEM images of a small piece of IE-Col-Ap (A) and E-Col-Ap (B)
hydrogel fabricated using the two-temperature process. Inset images in A and B:
SAED results. The mineralized collagen fibrils were unstained.
Figure 6.2. SEM images of IE-Col-Ap (A, A1, A2) and E-Col-Ap (B, B1, B2)
scaffolds. Double headed arrows in A2 and B2 indicate the freezing direction of
self-compressed hydrogels. Black circle in B2 shows the randomly dispersed minerals
on the surface.
xv
Figure 6.3. Freeze-dried IE-Col-Ap (A) and E-Col-Ap (B) scaffolds were ground into
small pieces in liquid nitrogen, and one of these small pieces were observed using
TEM. Inset images in A and B: SAED results. White arrows in A and B indicate the
extrafibrillar minerals.
Figure 6.4. XRD spectra of IE-Col-Ap (A) and E-Col-Ap (B) scaffolds.
Figure 6.5. FT-IR spectra of IE-Col-Ap (A) and E-Col-Ap (B) scaffolds.
Figure 6.6. TGA of IE-Col-Ap (A) and E-Col-Ap (B) scaffolds.
Figure 6.7. Proliferation of MC3T3-E1 cells on collagen, E-Col-Ap, and IE-Col-Ap
scaffolds over the course of 7 days. Data is represented as the mean ± standard error.
# denotes statistical significance (p < 0.05) between collagen scaffolds and
experimental scaffolds.
Chapter 7
Figure 7.1. TEM images of a small piece of hydrogel composed of intrafibrillar apatite (A),
silica (B), and apatite/silica (C) mineralized collagen fibrils. The mineralized collagen fibrils
were unstained.
1
1. Introduction
1.1 Background
In the United States, approximately one million bone fractures are treated every
year. [1, 2, 3] Various materials have been used, such as autografts, allografts, and
artificial grafts, including metals, polymers and ceramics. Autografts have long been
considered as the gold standard for bone grafting procedures. However, the amount of
bone that can be harvested is limited, and there is a high potential for donor site
morbidity. Allografts are alternatives to autografts, but there are concerns of disease
transmission and immuno-rejection. [4, 5, 6, 7] Artificial grafts can overcome the
resource problem of autografts and disease transmission issues of allografts; however,
many bone grafts are not bioactive. A promising strategy to address these problems is
bone tissue engineering, where a combination of supporting scaffolds, cells and/or
biological signals is used to regenerate defected bone. [4, 8, 9]
Supporting scaffolds play an important role in bone tissue engineering, and the
ideal supporting scaffolds for bone regeneration should have the following properties:
(1) biocompatibility, that does not induce immunogenic responses; (2)
biodegradability, that degrades at a suitable rate to match the regeneration rate of new
tissue; (3) osteoconductive, that facilitates the cell attachment and proliferation; (4)
osteoinductive, that stimulates the osteoprogenitor cells to differentiate into osteoblast
cells and induces new bone formation. [1]
Based on these criteria, bioinspired or biomimetic scaffolds have been prepared,
which have demonstrated great promises and attracted great attention from the
2
biomaterial community. [10, 11] Polymer scaffolds or metal implants were coated
with hydroxyapatite, the main inorganic composition of bone, to render these
materials with excellent bioactivity. [12, 13] Pek et al. produced the collagen-apatite
nanocomposite foams by mixing collagen fibers with apatite nanocrystals, mimicking
the composition of natural bone. [14] Xia et al. developed a biomimetic
collagen-apatite scaffold with multi-lamellar structure in order to mimic the
composition and structure of natural bone. [15]
1.2 Structures and composition of natural bone
Natural bone is a hierarchical composite mainly composed of apatite and collagen.
Type I collagen accounts for 95% of the organic material in bone, and has good
resorbability and high affinity to cells. [16, 17] Apatite constitutes about 65 wt% of
natural bone, exhibiting biocompatibility, osteoconductivity, and bone-bonding ability.
[3, 18] Additionally, silicon has been found in active bone growth area, such as the
osteoid of the bone of young rats, [19] indicating osteoinductive property of silicon
for new bone formation and stimulative for neovascularization without supplements
of growth factors. [20, 21]
3
Figure 1.1. Hierarchical structures of natural bone: starting from sub-nanostructure of
collagen molecules and tropocollagen triple helix, to nanostructure of ordered HA
crystals, and to sub-microstructure of collagen fibrils and macrostructure (a). The
arrangement of HA crystals with collagen fibrils (b). [22]
As shown in Figure 1.1, natural bone is a complex hierarchical composite having
multi-scale microstructure. Apatite nanocrystals deposit within the gap zone of
collagen fibrils (intrafibrillar mineralization) at an early stage of mineralization and
then on the surface of the collagen fibrils (extrafibrillar mineralization) at a later stage.
[23, 24, 25, 26, 27, 28] Bundles and arrays of mineralized collagen fibrils further
arrange into a multilayered structure rotating across concentric lamellae, which pack
densely into compact bone or loosely into spongy bone. [29, 30, 31]
1.3 Intrafibrillar mineralization
In order to mimic the composition and structure of natural bone, extensive
researches have been done to replicate the intrafibrillar mineralization process in
nature. [32, 33, 34] Non-collagenous proteins (NCPs) are found to play an important
4
role in regulating the intrafibrillar mineralization of collagen fibrils. [35, 36] Their
functions are to sequester calcium and phosphate ions to form liquid amorphous
calcium phosphate nanoprecursors, and then template the nucleation and growth of
the nanocrystals at the gap zone of the collagen fibrils. [35, 37, 38] According to
Kerschnitzki et al., abundant vesicles containing small mineral particles were found in
the lumen of blood vessels, in the bone-forming tissue between the blood vessels and
the forming bone surfaces, and in the cells associated with bone formation, proving
the important role of nanoprecursors in bone formation. [39, 40] As NCPs play an
important role in regulating intrafibrillar mineralization, polyanionic polymers such as
PAMAM-COOH dendrimer, poly (aspartic acid) and poly (acrylic acid) have been
used as the sequestration analog of the NCPs. [32, 41, 42] Liu et al. employed poly
(acrylic acid) as the sequestration analog and sodium tripolyphosphate as the
templating analog to induce collagen intrafibrillar mineralization, and proved that the
intrafibrillar mineralized collagen fibers have improved nanomechanics and
cytocompatibility. [43] Based on the success of apatite intrafibrillar mineralization,
silica was also deposited within the gap zone of collagen fibrils using poly (allylamine
hydrochloride) or choline chloride as the analogs of zwitterionic proteins. [44] Niu et
al. produced intrafibrillar silicified collagen scaffolds by incubating collagen scaffolds
in the choline-stabilized silica precursor, and the intrafibrillar silicified collagen
scaffolds promoted the osteocalcin expression and calcium deposition after
transplantation in vivo. [45]
5
1.4 Implant coating
Titanium and its alloy (Ti-6Al-4V) have been widely used as orthopedic and
dental implants because of their good biocompatibility, good corrosion resistance and
mechanical strength. [46, 47] However, titanium implants do not have a good bone
bonding and integration ability, resulting in implant failures. [48] Thus, coating
hydroxyapatite, the main inorganic composition of natural bone, on titanium has been
widely applied to impart the bioactivity and osteoconductivity to titanium implants
and improve the bone-bonding ability of the implants. [49, 50] After implanting, a
bone-like apatite layer forms on the bioactive coating, acting as an intermediate layer
connecting new bone and implants. [51] A lot of approaches have been used to form a
hydroxyapatite coating on the titanium implant, such as electrodeposition, [52] cold
spray, [53] electrostatic spraying, [54] plasma spray, [55] sol-gel, [56] pulse laser
deposition [57] and biomimetic approach. [58] The biomimetic process closely
mimics apatite formation in vivo, which significantly improves the biological
performance of the biomaterial. This coating technique has attracted extensive
attention because it is conducted at a condition close to body temperature (37 °C) and
pH (7.4), and it is able to co-precipitate with biologically active molecules without
compromising their bioactivities. [59]
1.5 Dual beam focused ion beam/scanning electron microscopy (FIB/SEM)
Focused ion beam (FIB) was developed in the middle 1970’s mainly for
semiconductor applications. The basic principal is to sputter atoms from the materials
using accelerated heavy ions. [60] FIB is able to mill into the bulk of a specimen,
6
exposing the hidden internal microstructure without causing significant damages. As a
result, it has been used to prepare TEM samples, cross-sections, and metal
depositions.
Dual-beam focused ion beam/scanning electron microscopy (FIB/SEM) was
introduced to include an FIB column for material etching and an SEM column for
imaging to facilitate simultaneous milling, imaging, serial sectioning, and many other
functionalities (Figure 1.2). [61] Both ion beam and electron beam are focused at the
same point, enabling the electron column to monitor the process while the ion beam is
millings the specimen. The dual-beam FIB/SEMs not only have all the capabilities as
single beam instruments, but also allow a more precise cross-sectioning of
heterogeneous, fragile, and/or porous materials and structures, 3-D imaging and
preparation of TEM specimens.
The majority of FIB/SEMs are equipped with a Ga liquid metal ion source on the
FIB column (GFIB); however, they are limited by the low material removal rate. [62]
This problem has been addressed by the introduction of a new generation of
dual-beam FIB/SEM, which are equipped with plasma ion source on the FIB column
(PFIB). [63, 64] The PFIB offers milling rates up to two orders of magnitude higher
than those of GFIB. [65, 66]
7
Figure 1.2. The specimen is tilted towards the ion beam (left), and a cross-section is
milled to be imaged by the electron beam (right). [67]
8
References
[1] Z. Xia, M. Wei. Biomimetic fabrication of collagen-apatite scaffolds for bone
tissue regeneration. Journal of Biomaterials and Tissue Engineering. 2013;
369-384.
[2] L. Meinel, V. Karageorgiou, R. Fajardo, B. Snyder, V. Shinde-Patil, L. Zichner,
D. Kaplan, R. Langer, G. Vunjak-Novakovic. Bone tissue engineering using
human mesenchymal stem cells: effects of scaffold material and medium flow.
Annals of Biomedical Engineering. 2004; 32:112-122.
[3] A. A. Al‐Munajjed, N. A. Plunkett, J. P. Gleeson, T. Weber, C. Jungreuthmayer,
T. Levingstone, J. Hammer, F. J. O'Brien. Development of a biomimetic
collagen‐hydroxyapatite scaffold for bone tissue engineering using a SBF
immersion technique. Journal of Biomedical Materials Research Part B: Applied
Biomaterials. 2009; 90:584-591.
[4] J. R. Porter, T. T. Ruckh, K. C. Popat. Bone tissue engineering: a review in bone
biomimetics and drug delivery strategies. Biotechnology Progress. 2009;
25:1539-1560.
[5] C. Szpalski, M. Barbaro, F. Sagebin, S. M. Warren. Bone tissue engineering:
current strategies and techniques—part II: cell types. Tissue Engineering Part B:
Reviews. 2012; 18:258-269.
[6] J. O. Smith, A. Aarvold, E. R. Tayton, D. G. Dunlop, R. O. Oreffo. Skeletal tissue
regeneration: current approaches, challenges, and novel reconstructive strategies
9
for an aging population. Tissue Engineering Part B: Reviews. 2011; 17:307-320.
[7] S. Liao, F. Cui, W. Zhang, Q. Feng. Hierarchically biomimetic bone scaffold
materials: nano‐HA/collagen/PLA composite. Journal of Biomedical Materials
Research Part B: Applied Biomaterials. 2004; 69:158-165.
[8] L. H. Nguyen, N. Annabi, M. Nikkhah, H. Bae, L. L. Binan, S. Park, Y. Kang, Y.
Yang, A. Khademhosseini. Vascularized bone tissue engineering: approaches for
potential improvement. Tissue Engineering Part B: Reviews. 2012; 18:363-382.
[9] I. Marcos-Campos, D. Marolt, P. Petridis, S. Bhumiratana, D. Schmidt, G.
Vunjak-Novakovic. Bone scaffold architecture modulates the development of
mineralized bone matrix by human embryonic stem cells. Biomaterials. 2012;
33:8329-8342.
[10] X. Su, K. Sun, F. Cui, W. Landis. Organization of apatite crystals in human
woven bone. Bone. 2003; 32:150-162.
[11] T. Dvir, B. P. Timko, D. S. Kohane, R. Langer. Nanotechnological strategies for
engineering complex tissues. Nature Nanotechnology. 2011; 6:13-22.
[12] S. S. Kim, M. S. Park, S. J. Gwak, C. Y. Choi, B. S. Kim. Accelerated bonelike
apatite growth on porous polymer/ceramic composite scaffolds in vitro. Tissue
Engineering. 2006; 12:2997-3006.
[13] H. Qu, M. Wei. Improvement of bonding strength between biomimetic apatite
coating and substrate. Journal of Biomedical Materials Research Part B: Applied
Biomaterials. 2008; 84:436-443.
10
[14] Y. Pek, S. Gao, M. M. Arshad, K. J. Leck, J. Y. Ying. Porous collagen-apatite
nanocomposite foams as bone regeneration scaffolds. Biomaterials. 2008;
29:4300-4305.
[15] Z. Xia, M. Villa, M. Wei. A biomimetic collagen–apatite scaffold with a
multi-level lamellar structure for bone tissue engineering. Journal of Materials
Chemistry B. 2014; 2:1998-2007.
[16] Z. Xia, X. Yu, X. Jiang, H. D. Brody, D. W. Rowe, M. Wei. Fabrication and
characterization of biomimetic collagen–apatite scaffolds with tunable structures
for bone tissue engineering. Acta Biomaterialia. 2013; 9:7308-7319.
[17] J. M. Holzwarth, P. X. Ma. Biomimetic nanofibrous scaffolds for bone tissue
engineering. Biomaterials. 2011; 32:9622-9629.
[18] S. Yunoki, H. Sugiura, T. Ikoma, E. Kondo, K. Yasuda, J. Tanaka. Effects of
increased collagen-matrix density on the mechanical properties and in vivo
absorbability of hydroxyapatite–collagen composites as artificial bone materials.
Biomedical Materials. 2011; 6:015012.
[19] E. Zhang, C. Zou, S. Zeng. Preparation and characterization of silicon-substituted
hydroxyapatite coating by a biomimetic process on titanium substrate. Surface
and Coatings Technology. 2009; 203:1075-1080.
[20] A. J. Mieszawska, N. Fourligas, I. Georgakoudi, N. M. Ouhib, D. J. Belton, C. C.
Perry, D. L. Kaplan. Osteoinductive silk–silica composite biomaterials for bone
regeneration. Biomaterials. 2010; 2010:8902-8910.
11
[21] W. Zhai, H. Lu, L. Chen, X. Lin, Y. Huang, K. Dai, K. Naoki, G. Chen, J. Chang.
Silicate bioceramics induce angiogenesis during bone regeneration. Acta
Biomaterialia. 2012; 8:341-349.
[22] B. Basu. Natural Bone and Tooth: Structure and Properties. Biomaterials for
Musculoskeletal Regeneration. 2016; 45-85.
[23] X. Su, K. Sun, F. Cui, W. Landis. Organization of apatite crystals in human
woven bone. Bone. 2003; 32:150-162.
[24] L. Chen, R. Jacquet, E. Lowder, W. J. Landis. Refinement of collagen–mineral
interaction: A possible role for osteocalcin in apatite crystal nucleation, growth
and development. Bone. 2015; 71:7-16.
[25] Y. Wang, T. Azaïs, M. Robin, A. Vallée, C. Catania, P. Legriel, G.
Pehau-Arnaudet, F. Babonneau, M.-M. Giraud-Guille, N. Nassif. The
predominant role of collagen in the nucleation, growth, structure and orientation
of bone apatite. Nature materials. 2012; 11:724-733.
[26] X. Wang, F. Cui, J. Ge, Y. Wang. Hierarchical structural comparisons of bones
from wild-type and liliput dtc232 gene-mutated Zebrafish. Journal of structural
biology. 2004; 145:236-245.
[27] F. Nudelman, P. H. Bomans, A. George, N. A. Sommerdijk. The role of the
amorphous phase on the biomimetic mineralization of collagen. Faraday
Discussions. 2012; 159:357-370.
[28] F. Nudelman, K. Pieterse, A. George, P. H. Bomans, H. Friedrich, L. J. Brylka, P.
12
A. Hilbers, G. de With, N. A. Sommerdijk. The role of collagen in bone apatite
formation in the presence of hydroxyapatite nucleation inhibitors. Nature
Materials. 2010; 9:1004-1009.
[29] W. Landis, M. M. Song, A. Leith, L. McEwen, B. McEwen. Mineral and organic
matrix interaction in normally calcifying tendon visualized in three dimensions
by high-voltage electron microscopic tomography and graphic image
reconstruction. Journal of Structural Biology. 1993; 110:39-54.
[30] T. T. Thula, D. E. Rodriguez, M. H. Lee, L. Pendi, J. Podschun, L. B. Gower. In
vitro mineralization of dense collagen substrates: a biomimetic approach toward
the development of bone-graft materials. Acta Biomaterialia. 2011; 7:3158-3169.
[31] W. Wagermaier, H. Gupta, A. Gourrier, M. Burghammer, P. Roschger, P. Fratzl.
Spiral twisting of fiber orientation inside bone lamellae. Biointerphases. 2006;
1:1-5.
[32] J. Li, J. Yang, J. Li, L. Chen, K. Liang, W. Wu, X. Chen, J. Li. Bioinspired
intrafibrillar mineralization of human dentine by PAMAM dendrimer.
Biomaterials. 2013; 34:6738-6747.
[33] S. S. Jee, L. Culver, Y. Li, E. P. Douglas, L. B. Gower. Biomimetic
mineralization of collagen via an enzyme-aided PILP process. Journal of Crystal
Growth. 2010; 312:1249-1256.
[34] S. S. Jee, R. K. Kasinath, E. DiMasi, Y. Y. Kim, L. Gower. Oriented
hydroxyapatite in turkey tendon mineralized via the polymer-induced
13
liquid-precursor (PILP) process. CrystEngComm. 2011; 13:2077-2083.
[35] S. Gajjeraman, K. Narayanan, J. Hao, C. Qin, A. George. Matrix macromolecules
in hard tissues control the nucleation and hierarchical assembly of
hydroxyapatite. Journal of Biological Chemistry. 2007; 282:1193-1204.
[36] Y. K. Kim, L. Gu, T. E. Bryan, J. R. Kim, L. Chen, Y. Liu, J. C. Yoon, L. Breschi,
D. H. Pashley, F. R. Tay. Mineralisation of reconstituted collagen using
polyvinylphosphonic acid/polyacrylic acid templating matrix protein analogues
in the presence of calcium, phosphate and hydroxyl ions. Biomaterials. 2010;
31:6618-6627.
[37] G. He, S. Gajjeraman, D. Schultz, D. Cookson, C. Qin, W. T. Butler, J. Hao, A.
George. Spatially and temporally controlled biomineralization is facilitated by
interaction between self-assembled dentin matrix protein 1 and calcium
phosphate nuclei in solution. Biochemistry. 2005;44:16140-16148.
[38] G. He, T. Dahl, A. Vei, A. George. Nucleation of apatite crystals in vitro by
self-assembled dentin matrix protein 1. Nature Materials. 2003; 2:552-558.
[39] M. Kerschnitzki, A. Akiva, A. B. Shohamb, N. Koifman, E. Shimoni, K. Rechav,
A. A. Arraf, T. M. Schultheiss, Y. Talmon, E. Zelzer, S. Weiner, L. Addadi.
Transport of membrane-bound mineral particles in blood vessels during chicken
embryonic bone development. Bone. 2016; 83:65-72.
[40] M. Kerschnitzki, A. Akiva, A. B. Shohamb, Y. Asscher, W. Wagermaier, P. Fratzl,
L. Addad, S. Weiner. Bone mineralization pathways during the rapid growth of
14
embryonic chicken long bones. Journal of Structural Biology. 2016; 195:82-92.
[41] S. S. Jee, T. T. Thula, L. B. Gower. Development of bone-like composites via the
polymer-induced liquid-precursor (PILP) process. Part 1: Influence of polymer
molecular weight. Acta Biomaterialia. 2010; 6:3676-3686.
[42] Y. Liu, Y. K. Kim, L. Dai, N. Li, S. O. Khan, D. H. Pashley, F. R. Tay.
Hierarchical and non-hierarchical mineralisation of collagen. Biomaterials. 2011;
2011:1291-1300.
[43] Y. Liu, S. Liu, D. Luo, Z. Xue, X. Yang, L. Gu, Y. Zhou, T. Wang. Hierarchically
staggered nanostructure of mineralized collagen as a bone-grafting scaffold.
Advanced Materials. 2016; 28:8740-8748.
[44] L. Niu, K. Jiao, Y. Qi, C. K. Yiu, H. Ryou, D. D. Arola, J. Chen, L. Breschi, D. H.
Pashley, F. R. Tay. Infiltration of silica inside fibrillar collagen. Angewandte
Chemie. 2011; 123:11892-11895.
[45] L. Niu, J. Sun, Q. Li, K. Jiao, L. Shen, D. Wu, F. Tay, J. Chen.
Intrafibrillar-silicified collagen scaffolds enhance the osteogenic capacity of
human dental pulp stem cells. Journal of Dentistry. 2014; 42:839-849.
[46] S. B. Goodman, Z. Yao, M. Keeney, F. Yang. The future of biologic coatings for
orthopaedic implants. Biomaterials. 2013; 34:3174-3183.
[47] Z. Xia, X. Yu. Biomimetic collagen/apatite coating formation on Ti6Al4V
substrates. Journal of Biomedical Materials Research Part B: Applied
Biomaterials. 2012; 100:871-881.
15
[48] L. M. Bjursten, L. Rasmusson, S. Oh, G. C. Smith, K. S. Brammer, S. Jin.
Titanium dioxide nanotubes enhance bone bonding in vivo. Journal of
Biomedical Materials Research Part A. 2010; 92:1218-1224.
[49] E. Sandrini, C. Giordano, V. Busini, E. Signorelli, A. Cigada. Apatite formation
and cellular response of a novel bioactive titanium. Journal of Materials Science:
Materials in Medicine. 2007; 18:1225-1237.
[50] T. Kokubo, H. Kim, M. Kawashita, T. Nakamura. Bioactive metals: preparation
and properties. Journal of Materials Science: Materials in Medicine. 2004;
15:99-107.
[51] J. Forsgren, F. Svahn, T. Jarmar, H. Engqvist. Formation and adhesion of
biomimetic hydroxyapatite deposited on titanium substrates. Acta Biomaterialia.
2007; 3:980-984.
[52] S. H. Chen, W. L. Tsai, P. C. Chen, A. Fang, W. C. Say. Influence of applied
voltages on mechanical properties and in-vitro performance of electroplated
hydroxyapatite coatings on pure titanium. Journal of the Electrochemical Society.
2016; 163:D305-D308.
[53] E. Kergourlay, D. Grossin, N. Cinca, C. Josse, S. Dosta, G. Bertrand, I. Garcia, J.
M. Guilemany, C. Rey. First cold spraying of carbonated biomimetic
nanocrystalline apatite on Ti6Al4V: physical-chemical, microstructural, and
preliminary mechanical characterizations. Advanced Engineering Materials.
2016; 18:496-500.
16
[54] P. J. D. Manders, J. G. C. Wolke, J. A. Jansen. Bone response adjacent to calcium
phosphate electrostatic spray deposition coated implants: an experimental study
in goats. Clinical Oral Implants Research. 2006; 17:548-553.
[55] D. Chen, E. H. Jordan, M. Gell, M. Wei. Apatite formation on alkaline treated
dense tio2 coatings deposited using the solution precursor plasma spray process.
Acta Biomaterialia. 2008; 4:553-559.
[56] W. Zhang, W. Liu, Y. Liu, C. Wang. Tribological behaviors of single and dual
sol-gel ceramic films on Ti6Al4V. Ceramics International. 2009; 35:1513-1520.
[57] O. Blind, L. H. Klein, B. Dailey, L. Jordan. Characterization of hydroxyapatite
films obtained by pulsed laser deposition on ti and ti6al4v substrates. Dental
Materials. 2005; 21:1017-1024.
[58] H. Teng, C. Yang, J. Lin, Y. Huang, F. Lu. A simple method to functionalize the
surface of plasma electrolytic oxidation produced tio2 coatings for growing
hydroxyapatite. Electrochimica Acta. 2016; 193:216-224.
[59] X. Yu, H. Qu, D. A. Knecht, M. Wei. Incorporation of bovine serum albumin into
biomimetic coatings on titanium with high loading efficacy and its release
behavior. Journal of Materials Science: Materials in Medicine. 2009; 20:287-294.
[60] R. Wirth. Focused Ion Beam (FIB) combined with SEM and TEM: Advanced
analytical tools for studies of chemical composition, microstructure and crystal
structure in geomaterials on a nanometre scale. Chemical Geology. 2009;
261:217-229.
17
[61] A. J. Bushby, K. M. P'ng, R. D. Young, C. Pinali, C. Knupp, A. J. Quantock.
Imaging three-dimensional tissue architectures by focused ion beam scanning
electron microscopy. Nature Protocols. 2011; 6:845-858.
[62] R. Young, C. Rue, S. Randolph, C. Chandler, G. Franz, R. Schampers, A.
Klumpp, L. Kwakman. A comparison of xenon plasma FIB technology with
conventional gallium LMIS FIB: imaging, milling, and gas-assisted applications.
Microscopy and Microanalysis. 2011; 17:652-653.
[63] N. Smith, W. Skoczylas, S. Kellogg, D. Kinion, P. Tesch, O. Sutherland, A.
Aanesland, R. Boswell. High brightness inductively coupled plasma source for
high current focused ion beam applications. Journal of Vacuum Science &
Technology B. 2006; 24:2902-2906.
[64] N. S. Smith, J. A. Notte, A. V. Steele. Advances in source technology for focused
ion beam instruments. MRS Bulletin. 2014; 39:329-335.
[65] T. Burnett, R. Kelley, B. Winiarski, L. Contreras, M. Daly, A. Gholinia, M.
Burke, P. Withers. Large volume serial section tomography by Xe Plasma FIB
dual beam microscopy. Ultramicroscopy. 2016; 161:119-129.
[66] M. N. Kelly, K. Glowinski, N. T. Nuhfer, G. S. Rohrer. The five parameter grain
boundary character distribution of α-Ti determined from three-dimensional
orientation data. Acta Materialia. 2016; 111:22-30.
[67] D. Stokes, F. Morrissey, B. Lich. A new approach to studying biological and soft
materials using focused ion beam scanning electron microscopy (FIB SEM).
18
Journal of Physics: Conference Series. 2006; 26:50.
19
2. Cross-sectional microstructure study of apatite coating on titanium
alloy
Abstract
Biomimetic hydroxyapatite (HA) coating has been used to render titanium
implants with excellent osteoconductivity and osseointegration in orthopedic and
dental applications. However, to date it has not been possible to reveal the fine details
of the coating structure or the coating/substrate interface. In this study, a crack-free
biomimetic HA coating has been successfully formed on sand-blasted, acid-etched
Ti-6Al-4V discs by simply immersing the discs in a modified simulated body fluid
(m-SBF). Dual-beam FIB/SEMs with either gallium ion source (GFIB) or xenon
plasma ion source (PFIB) were used to prepare the cross-sections and thin
transmission electron microscopy (TEM) specimens of the coating, and the
capabilities of GFIB and PFIB were compared. Both the cross-sectional SEM and
TEM images confirmed that a well bonded biomimetic HA coating has been formed
on the Ti-6Al-4V substrate. The coating exhibits a gradient structure with a dense
microstructure adjacent to the titanium substrate to facilitate strong bonding, while
there is a porous structure at the surface, which is beneficial to bone cell attachment
and subsequent new bone integration.
Keywords: Biomimetic HA coating; Cross-section; TEM specimen; Dual-beam
FIB/SEM; Gradient structure.
20
2.1 Introduction
The titanium alloy Ti-6Al-4V (Ti – 6% Al – 4% V (wt.%)) has been used
extensively in orthopedic and dental implants because of its good biocompatibility,
superior mechanical properties, and good corrosion resistance. [1, 2] However,
Ti-6Al-4V implants do not have a good bone-integration ability. [3] Thus, it has
become common practice to coat the Ti-6Al-4V implant with bioactive
hydroxyapatite (HA) to impart good bioactivity, to induce osteoconductivity, and to
promote osseointegration. [4] A variety of different deposition techniques has been
used to produce HA coatings on titanium alloy substrates. These include: plasma
spraying, sol-gel coating, electrophoretic deposition, and biomimetic immersion. [5-9]
The biomimetic coating approach has attracted extensive attention due to its mild
conditions (low temperature and neutral pH), ability to produce coatings with good
bonding strength, and capability to incorporate bioactive molecules during the coating
process. [2, 10-15] The surface morphology of the biomimetic coating has been
studied extensively using scanning electron microscopy (SEM). [16-18] However,
there are very few publications that describe the fine details of the internal
microstructure for these biomimetic coatings, or of the interface between the coating
and the alloy substrate.
Cross-sectional studies of the coating microstructure and the coating/substrate
interface are essential for investigations of crack formation, coating dissolution, ion
diffusion and distribution, and interfacial bonding. [19] For SEM studies the samples
are typically produced by fracturing or by cutting, grinding and polishing the coated
21
substrate. Both approaches have severe limitations. Fracture surfaces often have
complex topographies that complicate the interpretation of structural detail, and there
can be problems associated with plastic deformation in the substrate leading to
interfacial delamination. Cutting, grinding and polishing to produce
metallographic-type cross sections often induces significant mechanical damage and
limits the useful detail that can be extracted from such samples. [20-22] More serious
complications arise if one attempts to study such coatings by transmission electron
microscopy (TEM). These studies require extremely thin (typically <100nm) samples
free of mechanical damage or other artefacts. Traditional approaches for TEM
specimen preparation such as ultramicrotomy, or dimple grinding followed by ion
beam milling, are simply not feasible for HA-coated Ti alloy substrates. The
mechanical stresses introduced by these approaches would cause significant structural
damage both to the coating and to the coating/substrate interface. [23-25]
Focused ion beam (FIB) microscopy is a well-established technique for both
imaging and modifying samples on the micro- or nano-scale. The ability of FIB to
remove material from specific locations without inducing significant ion beam
damage makes it an ideal approach to prepare cross-sections of various materials and,
in particular, to produce site-specific cross-sectional TEM specimens. [26] Most
modern instruments include both an FIB column and an SEM column to facilitate
simultaneous milling and imaging, serial section tomography and many other
functionalities. [23] These dual-beam FIB/SEMs are capable of precise
cross-sectioning of heterogeneous, fragile, and/or porous materials and structures, and
22
preparation of TEM specimens. The vast majority of FIB/SEMs are equipped with a
Ga liquid metal ion source on the FIB column. While the latest generation of Ga FIB
(GFIB) columns is capable of operating over a wide range of probe sizes and beam
currents, there are restrictions on the material removal rate that can be achieved; this
limits the volume that can be interrogated in a reasonable milling time. [27] This
problem has been addressed by the introduction of a new generation of dual-beam
FIB/SEMs which utilize an inductively coupled plasma ion source with a heavy inert
gas feed (typically Xe) on the ion column. [28, 29] These plasma FIB (PFIB) columns
offer milling rates up to two orders of magnitude higher than those that can be
achieved using GFIB. [30, 31] They also offer advantages for materials that undergo
undesirable chemical interactions with the Ga ions during GFIB milling. However, to
date there have been very few studies that consider the relative merits of the GFIB
and PFIB approaches for specific complex materials systems.
Here we present a study of the microstructural details in a crack-free
biomimetic HA coating deposited onto a Ti-6Al-4V substrate by immersing it in a
modified simulated body fluid (m-SBF) at body temperature and pH. Conventional
microstructural studies on such coatings have been reported previously. [22] In the
current study we have concentrated on the use of dual-beam FIB/SEM techniques to
obtain high quality cross-sectional specimens for SEM and TEM experiments. These
specimens were prepared using both GFIB and PFIB instruments to reveal the relative
advantages of these approaches for such coated samples.
23
2.2 Materials and methods
2.2.1 Biomimetic HA coating preparation
Ti-6Al-4V discs with a thickness of 2 mm and a diameter of 15 mm were surface
treated as described previously. [2, 22] The Ti-6Al-4V discs were sandblasted with 46
grit SiC, followed by acid etching with a mixed solution of HCl and HNO3 and then
with a hot acid mixture of HCl and H2SO4. The acid-treated discs were then exposed
to UV irradiation for 8 h. Biomimetic HA coating was produced by immersing the
surface-treated Ti-6Al-4V in m-SBF. The ion concentrations of m-SBF and human
blood plasma are listed in Table 2.1. The initial pH of the m-SBF solution was
adjusted to 6.4, 6.5, 6.6, and 6.7 using HCl and HEPES
(4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), and the coatings formed at
different initial pH values were labeled as HA-6.4, HA-6.5, HA-6.6, and HA-6.7,
respectively. The coating formation was carried out by soaking the Ti-6Al-4V discs in
the m-SBF at a temperature of 37 °C for 24 h. The HA coated Ti-6Al-4V discs were
then washed gently with deionized water and air-dried overnight.
24
Table 2.1 Ion concentrations (mM) of human blood plasma and m-SBF.
Ion Human blood plasma (mM) m-SBF (mM)
Na+ 142.0 109.9
K+ 5.0 6.0
Mg2+ 1.5 1.5
Ca2+ 2.5 7.9
Cl- 103.0 111.2
HCO3- 27.0 17.5
HPO42- 1.0 3.0
SO42- 0.5 0
2.2.2 Microstructural characterization
The phase(s) present in the coatings were evaluated by X-ray diffraction (XRD)
using a Bruker AXS D2 Phaser diffractometer with CuKα radiation. Data were
acquired at 2θ values of 5° to 60° using a step size of 0.02° and a scan rate of
0.5 °/min.
Observations of surface morphology and FIB sectioning were performed in two
FEI Helios dual-beam FIB/SEMs: a Nanolab 460F1 GFIB and a Xenon PFIB in the
UConn/Thermo Fisher Scientific Center for Advanced Microscopy and Materials
Analysis (CAMMA). Both instruments are equipped with EasyLift nano-manipulators,
and the GFIB also has a flip-stage and a scanning transmission electron microscopy
(STEM) detector for improved final thinning of TEM specimens. To reduce charging
effects during SEM imaging and FIB milling, a thin conducting layer of Au was
sputtered onto the HA coating surfaces using a Polaron E5100 coating unit.
Secondary electron (SE) SEM images of the HA coating surface morphology
25
were obtained using the Everhart-Thornley and in-lens detectors on the electron
columns. All sectioning was performed with the coating surface perpendicular to the
optic axis of the ion column (at 52˚ to the electron column). Prior to cross-sectional
milling, a 3-µm-thick Pt layer was deposited onto the chosen region of the HA coating
surface to protect the coating from ion beam damage. This Pt deposition was
performed in two steps. Firstly, a 1-µm-thick Pt layer was deposited using a 5 kV
electron beam (E-beam) to crack the organometallic Pt precursor. A further
2-µm-thick Pt layer was then deposited using the ion beam (I-beam). [32] The ion
column accelerating voltages used in this process were 30 kV for the GFIB and 8 kV
for the PFIB.
For cross-sectional SEM studies a stepped, wedge-shape was first milled out
using a regular cross-section (RCS) pattern, so that the microstructure could be
observed on the vertical wedge wall using the electron column. The RCS pattern is a
multi-pass scan method used to mill the initial wedge. In this approach, the beam is
scanned through the entire pattern and then repeats for a set number of passes. For the
GFIB sectioning study, the wedge was 20 µm long, 10 µm wide, and 5 µm deep,
giving a wedge angle of ≈ 45˚. In the PFIB study, larger wedges with the same angle
were produced. In these cases, the dimensions were 60 µm long, 12 µm wide and 6
µm deep. After the wedge was milled, the vertical wedge wall was polished using an
ion beam at a reduced beam current to eliminate curtaining and other milling artefacts.
Two different pre-set milling patterns were used for the process, namely cleaning
cross-section (CCS) pattern, and rectangle (Rec) pattern. The CCS pattern is
26
processed line by line with each line containing a set number of passes, and the Rec
pattern repeatedly scans the beam over a rectangular array of equally spaced
positions.
For TEM studies, specimens were produced in a similar manner, but a second
wedge was milled on the other side of the protective Pt strap so that the vertical
wedge walls defined a pre-thinned FIB lamella. The lamella was then attached to the
EasyLift W probe, cut free from the surrounding material and transferred to a Cu
omni-grid. Final thinning of the lamellae to electron transparency was performed on
the grid using the ion beam with iteratively decreased ion currents to minimize any
residual ion damage.
The TEM specimens were examined using an FEI Talos F200X S/TEM operating
at an accelerating voltage of 200 kV. This instrument is equipped with a Super-X
silicon drift detector (SDD) energy dispersive X-ray spectrometry (EDXS) system,
which allows for a rapid acquisition of spectrum images for elemental mapping.
2.3 Results and discussion
Figure 2.1 is a series of SE SEM images showing the surface morphologies of the
HA coatings produced by immersing surface-treated Ti-6Al-4V substrates in an
m-SBF solution at different initial pH values for 24 h. When the initial pH was 6.4, a
thin, crack-free dense mineral coating was formed (HA-6.4, Figure 2.1A). As the
coating is very thin, the morphology of the underlying Ti-6Al-4V substrate can be
observed clearly. When the initial pH was 6.5, a smooth coating with a few minor
cracks due to dehydration shrinkage was formed, as indicated by the black arrows in
27
Figure 2.1B (HA-6.5). [33] A crack-free coating was formed when the initial pH was
6.6 (HA-6.6, Figure 2.1C). In contrast, the coating produced at an initial pH of 6.7
exhibits mudflat cracking with globules of HA scattered on the surface (HA-6.7,
Figure 2.1D). It has been observed previously that the HA-6.5, HA-6.6, and HA-6.7
coatings exhibit a porous and rough coating surface, which has been discovered to be
beneficial to cell attachment and the initiation of bone formation and integration. [34]
Figure 2.1. SE SEM images of HA coatings formed at different initial pH values:
HA-6.4 (A), HA-6.5 (B), HA-6.6 (C), and HA-6.7 (D). The black arrows in (B)
indicate cracks formed on the surface.
Examples of the corresponding XRD patterns are shown in Figure 2.2. These data
indicate that an HA coating can be formed at all initial pH values selected. However,
very weak HA peaks together with strong peaks from the hexagonal-close-packed
(HCP) α phase of the Ti-6Al-4V substrate were observed for the HA-6.4 coating. This
28
indicates that a very thin layer of HA coating was formed on the Ti-6Al-4V substrate,
which is in agreement with the SEM observation in Figure 2.1A. Strong peaks for HA
at 2θ=26° ((002) plane) and 31-34° (overlap of (211), (112), and (300) planes) were
observed for HA-6.5, HA-6.6, and HA-6.7 coatings, indicating that a thick HA
coating has been formed on the Ti-6Al-4V substrate. However, the broad diffraction
peak around 31-34° suggests that the HA coatings were not fully crystalline. A
preferred c-axis orientation of the HA crystals can be inferred from the high intensity
of the (002) peak compared to the (112) peak. [35] Diffraction peaks from the HCP α
phase of the Ti-6Al-4V substrate were present in all of the XRD patterns.
Figure 2.2. XRD spectra of HA coatings on Ti-6Al-4V substrates after 24 h of
29
immersion in m-SBF.
Based on the SEM and XRD data, it was found that the best-quality coatings
were HA-6.5 and HA-6.6 coatings. As such, these two coatings were selected for
more detailed cross-sectional investigation using FIB techniques. As shown in Figure
2.3, HA-6.5 and HA-6.6 coatings exhibit gradient structures: the HA coating adjacent
to the substrate was dense, and it became more and more porous towards the coating
surface. It is the first time that we clearly observed a gradient, porous cross-sectional
microstructure of the coating. When cross-sections produced are observed by the
conventional method (fracturing/grinding), only a dense structure was observed, but
the porous gradient structure was completely ruined by the sample preparation
process. [22, 36] Moreover, it has been shown that the pore size of HA-6.5 coating at
the outermost surface was larger than that of the HA-6.6 coating. Also, the HA-6.5
coating was thicker than HA-6.6 coating during the same immersion time (24 h),
which is in agreement with our previous observations.[12]
Figure 2.3. Cross-sectional SE SEM images of HA coatings on Ti-6Al-4V substrates:
HA-6.5 (A) and HA-6.6 (B). White arrow in image (A) is a crack induced by the ion
beam of PFIB and black arrow in image (A) is a crack between the HA-6.5 coating
30
and the Ti-6Al-4V substrate.
The formation of the HA coating can be expressed by Equations (1) and (2). As
discussed in our previous work, the pH of the m-SBF solution is determined by the
initial pH of the solution and the amount of HCO3- in the solution. The pH profile of
the m-SBF during the coating process has been investigated in details as reported
previously. [12, 22] Briefly, at the initial stage of coating formation, the pH of the
solution increases due to the decomposition of HCO3-, as shown in Equation (2).
Simultaneously, Ca2+ and HPO42- ions are adsorbed onto the substrate surface to form
HA nuclei, as expressed in Equation (1). Previous research reported that a high
concentration of HCO3- ions in m-SBF resulted in a dense apatite coating. [2, 22, 37]
Thus, in the initial stage of coating formation, a lower pH leads to slower coating
formation, resulting in a dense coating adjacent to the substrate. With the
decomposition of HCO3-, the increase in pH and the decrease in the HCO3
-
concentration leads to the formation of a porous HA coating toward the coating
surface. When the initial pH of m-SBF was 6.5, more HCO3- ions were needed to
bring up the pH than for an initial pH=6.6, thus less HCO3- ions remained,
contributing to the larger pore size observed toward the surface of the HA coating.
5Ca2++3HPO42-+4OH-→Ca5(PO4)3OH+3H2O (1)
HCO3-→CO2+OH- (2)
As shown in Figure 2.3A, there is a large crack between the HA-6.5 coating and
the Ti-6Al-4V substrate (black arrow in Figure 2.3A); this is consistent with the
cracks observed on the coating surface, as indicated in Figure 2.1B. Cracks at the
31
surface are due to the dehydration shrinkage, which significantly reduces the
interfacial bonding strength. [21] Thus, the cracks on the surface indicate weak
bonding between the HA-6.5 coating and the Ti-6Al-4V substrate. However, no
cracks were observed between the HA-6.6 coating and Ti-6Al-4V substrate, indicating
a strongly bonded HA coating was formed (Figure 2.3B). This might be because
HA-6.6 coating is thinner and denser compared to HA-6.5 coating. Because a
well-bonded HA coating formed at an initial pH=6.6, HA-6.6 coating system was
selected for further cross-sectional microscopy investigations.
In principle, cross-sections of the coating can provide critical information
regarding the interface as well as about the coating itself. However, due to the
significant differences between the stiffness and density of the coating and substrate,
it becomes extremely difficult to create the cross-sectional samples without damaging
the microstructure both at the interface and within the coating itself. FIB instruments
are powerful tools for studies of interfaces between materials with such different
properties, but there are certain drawbacks and instrumental artefacts. A common
artefact produced during FIB milling is the “curtaining” effect in which fluctuations in
the ion beam current, position and/or milling efficiency (due to sample
inhomogeneities) lead to undulations on the surface of an ostensibly planar FIB cut.
This effect can significantly degrade the resolution of images obtained from FIB-cut
cross-sections. Ideally, cross-sectional FIB cuts should be free of such curtaining
effects. After the initial wedge milling using the RCS pattern, the microstructure of
the cross-section was obscured by the curtaining (Figure 2.4). Then CCS and Rec
32
patterns were applied to obtain cross-sections with reduced curtaining in both GFIB
and PFIB, and the efficiencies of these two approaches were compared.
Figure 2.4. Gas-assisted deposition of Pt layer at 8 kV and 30 kV ion beam voltage in
PFIB.
Figure 2.5. SE SEM images of HA-6.6 coating on Ti-6Al-4V substrate from
cross-sections prepared using GFIB (A, B) and PFIB (C, D). The cross-sections were
polished using CCS pattern (A, C) or Rec pattern (B, D). Cracks indicated by white
33
arrows in HA coating in image (C) were induced by the CCS process in PFIB.
Figure 2.5 shows the cross-sectional images of HA-6.6 coated Ti-6Al-4V
substrates cleaned by Rec and CCS patterns in GFIB and PFIB. First, a 3-µm-thick Pt
layer was deposited onto the area of interest by gas-assisted Pt deposition to protect
the HA coating surface, allowing the features close to or at the surface to be analyzed
without damage. [38] In GFIB, an ion beam with accelerating voltage of 30 kV was
used to deposit the Pt-protective layer, however, in PFIB a dense Pt layer was
obtained at 8 kV. Increasing the ion beam accelerating voltage to 30 kV in PFIB led to
the formation of a porous Pt layer (Figure 2.4). In GFIB, wedges 20 µm in length, 10
µm in width and 5 µm in depth were milled using RCS, and then cleaned up using the
CCS and Rec patterns as shown in Figures 2.5A and 2.5B, respectively. In PFIB,
wedges 60 µm in length, 12 µm in width and 6 µm in depth were milled using RCS.
The images obtained from after cleaning up using CCS and Rec patterns are shown in
Figures 2.5C and 2.5D, respectively. Even though much larger wedges were milled
using PFIB, the total time for milling and the subsequent clean-up was less than those
used in GFIB. As shown in Figure 2.5, in GFIB, the Rec pattern proved to be more
suitable for the production of curtaining-free cross-sections than the CCS pattern at
the same ion beam current. However, the Rec pattern is as efficient as the CCS pattern
in creating a reasonable cross-section using PFIB. When CCS was applied to clean up
PFIB cross-sections, cracks were formed in the HA coating (white arrows in Figure
2.5C). Overall, the detailed microstructure of the HA coating and the interface
between the coating and the substrate have been well preserved by both GFIB and
34
PFIB, but GFIB produces a cross-section with higher quality than PFIB, which may
be due to the smaller ion beam spot size achievable in GFIB. [27]
Figure 2.6. An extraction sequence of a thin lamella from the bulk in GFIB (A-H).
The thickness of the TEM specimen was confirmed by STEM in GFIB (I).
Figure 2.7. The thin lamella from the site of interest before extraction (A) and after
mounted onto copper omni grids (B) in PFIB.
TEM experiments were performed to study the interface between the HA-6.6
coating and the Ti-6Al-4V substrate on the nano-scale. The TEM specimens were
35
prepared as thin lamellae from the sites of interest following the extraction sequence
presented in Figure 2.6. Firstly, a 3-µm-thick Pt strap was deposited onto the area of
interest (Figure 2.6A); this served as a protective layer for the HA coating during the
milling process. Wedge-shaped trenches were then milled out on either side of the Pt
strap (Figure 2.6B), defining a lamella that was then thinned further in preparation for
lift-out (Figures 2.6C and 2.7A). Due to the larger beam size and higher beam current
in the PFIB, some of the lamellae sustained significant damage during thinning (data
not shown), as indicated by the cracks in the HA coating in Figure 2.5C. As a result,
for PFIB a thicker lamella was prepared (~4 µm; Figure 2.7) than for GFIB (~900 nm,
Figures 2.6C and 2.6H). In both cases, the lamella was cut free at the bottom and on
one side, the W EasyLift probe was attached to the Pt strap by depositing additional Pt
at the point of contact, the lamella was cut free on the other side and was then lifted
out (Figure 2.6D). The lamella was then transferred to a copper omni grid (Figure
2.6E), attached to the grid at two corners by depositing Pt in these locations (Figure
2.6F), and then cut free from the W EasyLift probe (Figure 2.6G). The TEM lamella
was finally thinned to electron transparency using the ion beam with iteratively
decreasing ion currents. In GFIB, STEM imaging at an electron column accelerating
voltage of 30 kV was used to confirm that the specimen was thin enough for TEM
analysis (Figure 2.6I).
36
Figure 2.8. Data obtained from the TEM specimens prepared by GFIB: BF TEM
image (A) and HAADF STEM image (B). The inset to (A) is the SAED pattern from
the coating. C-F: Compositional maps obtained from the region in (B), showing the
distribution of Ca (C), P (D), Ti (E), and Ga (F).
37
Figure 2.9. Data obtained from the TEM specimens prepared by PFIB: BF TEM
image (A) and HAADF STEM image (B). C-F: Compositional maps obtained from
the region in (B), showing the distribution of Ca (C), P (D), Ti (E), and Xe (F).
38
The nano-scale microstructure of the HA coatings on Ti-6Al-4V was investigated
by S/TEM analysis of the TEM specimens prepared by both GFIB and PFIB, and
examples of the data are shown in Figures 2.8 and 2.9, respectively. In each case,
these include a bright field (BF) TEM image, a high angle annular dark field
(HAADF) STEM image, and compositional maps for Ca, P, Ti and the appropriate
milling ion (Ga and Xe, respectively). The maps were obtained from spectrum
imaging experiments in which full EDXS spectra were acquired at each pixel, and
then standard-less quantification was performed at each point. Thus, Figures 2.8C-F
and 2.9C-F are compositional maps (rather than X-ray maps) and the intensities in
these maps are proportional to the measured concentrations of the various species.
The BF TEM and HAADF STEM images showed that the thickness of the
coating was about 1.5-2.5 µm, and no structural defects were observed between the
HA coating and the Ti-6Al-4V substrate, indicating that a strong bond has been
formed at the interface. These images also reveal the details of the gradient structure
in the HA coating, which is consistent with the SEM observations in Figures 2.3 and
2.5. A dense coating is formed adjacent to the substrate to facilitate the strong bonding
between the coating and the substrate, and there is a highly porous surface that could
contribute to coating dissolution and cell attachment. [33] The preferential crystallite
alignment was also observed by STEM (Figure 2.8). Previous research reported that
the HA crystal dimensions decreased with increasing HCO3- ion concentration, and
that the crystals were more equi-axed and less needle-like at a higher concentrations.
[39] In the initial stage of coating formation, a smaller crystal size and denser coating
39
was formed adjacent to the substrate due to a higher HCO3- ion concentration and a
lower pH. [2, 22, 35, 39] As the immersion time increases, the HCO3- ions decompose
to bring up the pH, thus the dimensions of the HA crystal increased, and crystallites
elongated preferentially perpendicular to the surface, which is in accordance with the
XRD results (Figure 2.2). [35, 40] The corresponding mineral phase of the coating
was evaluated by selected area electron diffraction (SAED) coupled to TEM, and the
inset pattern in Figure 2.8A indicates that the HA coating is poly-crystalline, which
resembles the apatite structure in natural bone. [33] The compositional maps show
that the distributions of Ca and P in the coating are uniform, and thus there is no
change in composition associated with the structural gradient. These maps also reveal
that there is some implantation of Ga and Xe ion in the TEM specimens prepared by
GFIB and PFIB, respectively (Figures 2.8F and 2.9F). Such implantation is typical in
FIB analyses,[41] and does not appear to have caused any significant changes in the
structure of the coatings. We note, however, that the data from the TEM specimen
produced by PFIB (Figure 2.9) appears to be at slightly lower resolution and shows
somewhat more residual curtaining than the specimen produced by GFIB (Figure 2.8).
This is consistent with there being low-level damage and artefacts due to the coarser,
more powerful Xe ion beam in the PFIB.
2.4 Conclusions
A crack-free biomimetic HA coating was formed by immersing a surface-treated
Ti-6Al-4V substrate in an m-SBF solution at an initial pH of 6.6. Both GFIB and
PFIB were used to create curtaining-free cross-sections and TEM specimens of the
40
coating, and the capabilities of these two tools were compared. Although GFIB and
PFIB share a lot of similarities, there are also significant differences, including the
optimal voltage of ion-beam-induced Pt deposition, the TEM sample preparation
process, and the efficiency and the quality of the cross-sectional cuts produced.
Particularly, the GFIB is more beneficial for preparing the high-quality final finish of
the sample, while the PFIB efficient for creating a large cut of the sample. Also, both
SEM and TEM observations confirmed that a well-bonded HA coating has been
formed on the Ti-6Al-4V substrate. The coating demonstrates a unique gradient
structure, with a dense layer adjacent to the substrate to facilitate a strong bonding,
and a porous structure at the surface to allow for bone cell attachment.
41
References
[1] S. B. Goodman, Z. Yao, M. Keeney, F. Yang. The future of biologic coatings for
orthopaedic implants. Biomaterials. 2013; 34:3174-3183.
[2] Z. Xia, X. Yu. Biomimetic collagen/apatite coating formation on Ti6Al4V
substrates. Journal of Biomedical Materials Research Part B: Applied Biomaterials.
2012; 100:871-881.
[3] L. M. Bjursten, L. Rasmusson, S. Oh, G. C. Smith, K. S. Brammer, S. Jin.
Titanium dioxide nanotubes enhance bone bonding in vivo. Journal of Biomedical
Materials Research Part A. 2010; 92:1218-1224.
[4] E. Sandrini, C. Giordano, V. Busini, E. Signorelli, A. Cigada. Apatite formation
and cellular response of a novel bioactive titanium. Journal of Materials Science:
Materials in Medicine. 2007; 18:1225-1237.
[5] P. Yu, F. Lu, W. Zhu, D. Wang, X. Zhu, G. Tan, et al. Bio-inspired citrate
functionalized apatite coating on rapid prototyped titanium scaffold. Applied Surface
Science. 2014; 313:947-953.
[6] Y. Huang, Q. Ding, S. Han, Y. Yan, X. Pang. Characterisation, corrosion resistance
and in vitro bioactivity of manganese-doped hydroxyapatite films electrodeposited on
titanium. Journal of Materials Science: Materials in Medicine. 2013; 24:1853-1864.
[7] C. Aparicio, D. Rodriguez, F. Gil. Variation of roughness and adhesion strength of
deposited apatite layers on titanium dental implants. Materials Science and
Engineering: C. 2011; 31:320-324.
[8] D. Chen, E. H. Jordan, M. Gell, M. Wei. Apatite formation on alkaline-treated
42
dense TiO 2 coatings deposited using the solution precursor plasma spray process.
Acta Biomaterialia. 2008; 4:553-559.
[9] H. W. Kim, Y. H. Koh, L. H. Li, S. Lee, H. E. Kim. Hydroxyapatite coating on
titanium substrate with titania buffer layer processed by sol–gel method. Biomaterials.
2004; 25:2533-2538.
[10] L. N. Luong, K. M. McFalls, D. H. Kohn. Gene delivery via DNA incorporation
within a biomimetic apatite coating. Biomaterials. 2009; 30:6996-7004.
[11] G. Ciobanu, O. Ciobanu. Investigation on the effect of collagen and vitamins on
biomimetic hydroxyapatite coating formation on titanium surfaces. Materials Science
and Engineering: C. 2013; 33:1683-1688.
[12] H. Qu, M. Wei. The effect of temperature and initial pH on biomimetic apatite
coating. Journal of Biomedical Materials Research Part B: Applied Biomaterials. 2008;
87:204-212.
[13] X. Yu, H. Qu, D. A. Knecht, M. Wei. Incorporation of bovine serum albumin into
biomimetic coatings on titanium with high loading efficacy and its release behavior.
Journal of Materials Science: Materials in Medicine. 2009; 20:287-294.
[14] X. Yu, M. Wei. Preparation and evaluation of parathyroid hormone incorporated
CaP coating via a biomimetic method. Journal of Biomedical Materials Research Part
B: Applied Biomaterials. 2011; 97:345-354.
[15] X. Yu, L. Wang, X. Jiang, D. Rowe, M. Wei. Biomimetic CaP coating
incorporated with parathyroid hormone improves the osseointegration of titanium
implant. Journal of Materials Science: Materials in Medicine. 2012; 23:2177-2186.
43
[16] S. E. Bae, J. Choi, Y. K. Joung, K. Park, D. K. Han. Controlled release of bone
morphogenetic protein (BMP)-2 from nanocomplex incorporated on
hydroxyapatite-formed titanium surface. Journal of Controlled Release. 2012;
160:676-684.
[17] J. Faure, A. Balamurugan, H. Benhayoune, P. Torres, G. Balossier, J. Ferreira.
Morphological and chemical characterisation of biomimetic bone like apatite
formation on alkali treated Ti6Al4V titanium alloy. Materials Science and
Engineering: C. 2009; 29:1252-1257.
[18] A. W. Nijhuis, M. R. Nejadnik, F. Nudelman, X. F. Walboomers, J. te Riet, P.
Habibovic, et al. Enzymatic pH control for biomimetic deposition of calcium
phosphate coatings. Acta Biomaterialia. 2014; 10:931-939.
[19] S. Jalota, S. Bhaduri, S. B. Bhaduri, A. C. Tas. A protocol to develop crack-free
biomimetic coatings on Ti6Al4V substrates. Journal of Materials Research. 2007;
22:1593-1600.
[20] C. Auclair‐Daigle, M. Bureau, J. G. Legoux, L. H. Yahia. Bioactive
hydroxyapatite coatings on polymer composites for orthopedic implants. Journal of
Biomedical Materials Research Part A. 2005; 73:398-408.
[21] E. Zhang, C. Zou, S. Zeng. Preparation and characterization of silicon-substituted
hydroxyapatite coating by a biomimetic process on titanium substrate. Surface and
Coatings Technology. 2009; 203:1075-1080.
[22] H. Qu, M. Wei. Improvement of bonding strength between biomimetic apatite
coating and substrate. Journal of Biomedical Materials Research Part B: Applied
44
Biomaterials. 2008; 84:436-443.
[23] A. J. Bushby, K. M. P'ng, R. D. Young, C. Pinali, C. Knupp, A. J. Quantock.
Imaging three-dimensional tissue architectures by focused ion beam scanning electron
microscopy. Nature Protocol. 2011; 6:845-858.
[24] E. Coutinho, T. Jarmar, F. Svahn, A. Neves, B. Verlinden, B. Van Meerbeek, et al.
Ultrastructural characterization of tooth–biomaterial interfaces prepared with broad
and focused ion beams. Dental Materials. 2009; 25:1325-1337.
[25] M. Corazza, S. B. Simonsen, H. Gnaegi, K. T. Thydén, F. C. Krebs, S. A.
Gevorgyan. Comparison of ultramicrotomy and focused-ion-beam for the preparation
of TEM and STEM cross section of organic solar cells. Applied Surface Science. 2016;
389:462-468.
[26] M. Schaffer, B. Schaffer, Q. Ramasse. Sample preparation for atomic-resolution
STEM at low voltages by FIB. Ultramicroscopy. 2012; 114:62-71.
[27] R. Young, C. Rue, S. Randolph, C. Chandler, G. Franz, R. Schampers, et al. A
comparison of xenon plasma FIB technology with conventional gallium LMIS FIB:
imaging, milling, and gas-assisted applications. Microscopy and Microanalysis. 2011;
17:652-653.
[28] N. Smith, W. Skoczylas, S. Kellogg, D. Kinion, P. Tesch, O. Sutherland, et al.
High brightness inductively coupled plasma source for high current focused ion beam
applications. Journal of Vacuum Science & Technology B. 2006; 24:2902-2906.
[29] N. S. Smith, J. A. Notte, A. V. Steele. Advances in source technology for focused
ion beam instruments. MRS Bulletin. 2014; 39:329-335.
45
[30] T. Burnett, R. Kelley, B. Winiarski, L. Contreras, M. Daly, A. Gholinia, et al.
Large volume serial section tomography by Xe Plasma FIB dual beam microscopy.
Ultramicroscopy. 2016; 161:119-129.
[31] M. N. Kelly, K. Glowinski, N. T. Nuhfer, G. S. Rohrer. The five parameter grain
boundary character distribution of α-Ti determined from three-dimensional orientation
data. Acta Materialia. 2016; 111:22-30.
[32] H. Yu, Y. Sun, S. P. Alpay, M. Aindow. Solidification microstructures in Ag3Sn–
Cu3Sn pseudo-binary alloys. Journal of Materials Science. 2016; 51:6474-6487.
[33] E. Zhang, C. Zou, G. Yu. Surface microstructure and cell biocompatibility of
silicon-substituted hydroxyapatite coating on titanium substrate prepared by a
biomimetic process. Materials Science and Engineering: C. 2009; 29:298-305.
[34] Y. Huang, S. Han, X. Pang, Q. Ding, Y. Yan. Electrodeposition of porous
hydroxyapatite/calcium silicate composite coating on titanium for biomedical
applications. Applied Surface Science. 2013; 271:299-302.
[35] L. Müller, E. Conforto, D. Caillard, F. A. Müller. Biomimetic apatite
coatings—Carbonate substitution and preferred growth orientation. Biomolecular
Engineering. 2007; 24:462-466.
[36] D. O. Costa, B. A. Allo, R. Klassen, J. L. Hutter, S. J. Dixon, A. S. Rizkalla.
Control of surface topography in biomimetic calcium phosphate coatings. Langmuir.
2012; 28:3871-3880.
[37] F. Barrere, P. Layrolle, C. Van Blitterswijk, K. De Groot. Biomimetic calcium
phosphate coatings on Ti6Al4V: a crystal growth study of octacalcium phosphate and
46
inhibition by Mg 2+ and HCO 3−. Bone. 1999; 25:107S-111S.
[38] A. Thorfve, A. Palmquist, K. Grandfield. Three-dimensional analytical
techniques for evaluation of osseointegrated titanium implants. Materials Science and
Technology. 2014.
[39] N. Blumenthal, F. Betts, A. Posner. Effect of carbonate and biological
macromolecules on formation and properties of hydroxyapatite. Calcified Tissue
Research. 1975; 18:81-90.
[40] B. Feng, X. Chu, J. Chen, J. Wang, X. Lu, J. Weng. Hydroxyapatite coating on
titanium surface with titania nanotube layer and its bond strength to substrate. Journal
of Porous Materials. 2010; 17:453-458.
[41] R. Wirth. Focused Ion Beam (FIB) combined with SEM and TEM: Advanced
analytical tools for studies of chemical composition, microstructure and crystal
structure in geomaterials on a nanometre scale. Chemical Geology. 2009;
261:217-229.
47
3. Cross-sectional microstructure study of collagen-apatite composite
coating on titanium alloy
Abstract
A biomimetic bone-like collagen-hydroxyapatite (Col-HA) composite coating
was formed on a surface-treated Ti-6Al-4V alloy substrate via simultaneous collagen
self-assembly and hydroxyapatite nucleation. The coating process has been carried
out by immersing sand-blasted, acid-etched and UV irradiated Ti-6Al-4V alloy in type
I collagen-containing modified simulated body fluid (m-SBF). The surface
morphology and phase composition of the coating were characterized using scanning
electron microscopy (SEM), X-ray diffraction (XRD), and Fourier transformation
infrared spectroscopy (FTIR). More importantly, dual-beam FIB/SEMs with either
gallium ion source (GFIB) or xenon plasma ion source (PFIB) were used to
investigate the cross-sectional features of the biomimetic Col-HA composite coating
in great details. As a result, the cross-sectional images and thin transmission electron
microscopy (TEM) specimens were successfully obtained from the composite coating.
Both the cross-sectional SEM and TEM results confirmed that the Col-HA coating
had a similar microstructure to pure hydroxyapatite coating that consists of a gradient
coating at the cross-section, with a dense layer adjacent to the interface between the
titanium substrate and the coating, while a porous structure at the coating surface.
Keywords: Collagen-hydroxyapatite coating, Dual Beam FIB, TEM sample,
cross-section, microstructure
48
3.1 Introduction
Ti-6Al-4V (Ti-6%Al-4%V (wt.%)) alloy has been widely used as load-bearing
implants due to its excellent mechanical properties, corrosion resistance and
biocompatibility. [1, 2] However, there is still concern of the poor bone-implant
integration due to the formation of a fibrous tissue layer surrounding the Ti-6Al-4V
implants. [3] Thus, a layer of hydroxyapatite (HA) coating is usually deposited on the
surface of the Ti-6Al-4V implants to improve their bioactivity, osteoconductivity and
osteointegration. [4-6] HA, the main component of natural bone, exhibits excellent
biocompatibility, osteoconductivity and bone-bonding ability. [7] A variety of
deposition techniques, such as plasma spray process, hot isostatic pressing,
electro-deposition, sol-gel coating, and biomimetic method have been applied to
successfully deposit HA coatings on the Ti-6Al-4V implants. [5, 8-13] Collagen, the
most abundant protein in bone, is effective in accelerating cellular adhesion and
promoting cell proliferation. [14, 15] Thus, a composite coating comprised of both
collagen and HA will better mimic the composition of natural bone and thereby
improve the bone-implant integration.
Collagen/HA (Col-HA) composite coating has been deposited onto the surface of
Ti-6Al-4V substrates via electrolytic deposition, spin coating, and biomimetic
immersion over the past few years. [16-18] Due to the sensitivity of collagen, there
are limited methods available to deposit Col-HA composite coatings. For example,
plasma spraying and thermal spray, which are two widely used techniques to deposit
HA coatings, are impractical to deposit collagen composite coating due to the high
49
operation temperatures. [3] Biomimetic coating method, on the other hand, is a
promising method due to its low cost, easy operation, low temperature, and the ability
to incorporate biomolecules such as proteins and collagens. [7, 17, 19, 20] Thus,
Col-HA composite coatings have been successfully deposited on the Ti-6Al-4V
substrates. [7, 17, 18, 21]
The reliability of the coating and the bonding strength between coating and the
substrate are very important for the success of the implantation. [3] Because most of
the implant failures result from the debonding between the coating and the substrate,
as well as the dissolution of the coating. [22, 23] Cross-sectional studies of the
microstructure of the coating and the interface between the coating and the substrate
are essential to investigate the coating stability and bonding strength. [5, 24] However,
traditional techniques for producing cross-sections, such as coating fracturing or
coating cutting, grinding and polishing, suffer from losing microstructure details or
resulting in the debonding of the coating due to the plastic deformation. [5, 7, 25]
Meanwhile, focused ion beam (FIB) microscopy has been widely used as a
modern tool for structural microanalysis and micro-machining of materials. The most
important application of FIB is to prepare cross-sectional milling and transmission
electron microscopy (TEM) sample, since the microstructure of the sample is
maintained without damage. [26] The most advanced FIB instrument incorporates
both an FIB column and a scanning electron microscopy (SEM) column in a single
system, allowing simultaneous milling and imaging, serial sectioning and many other
functionalities. [27] The dual beam FIB/SEM is capable of preparing cross-sections
50
and TEM specimens made of heterogeneous, fragile, and porous materials and
structures. [5] The most widely used FIB/SEM is equipped with gallium liquid ion
source (GFIB), however, is limited with slow removal rate. [28] Thus, a dual beam
FIB/SEM system with plasma ion source (PFIB), which offers two orders of
magnitude higher milling rate than that of GFIB, was used in the current study. [29,
30] In our previous work, a comparison study has been conducted to investigate the
microstructure of the biomimetic HA coating on Ti-6Al-4V substrates using dual
beam FIB/SEM. The capabilities of GFIB and PFIB were also compared. [5] However,
due to ion bombardment and beam heating effects of FIB, the cross-sectioning the
sample using FIB is challenging when the sample involves polymer, which has a low
heating conductivity. [31]
In the present work, a biomimetic Col-HA coating was deposited on the
surface-treated Ti-6Al-4V alloy through simultaneous collagen self-assembly and HA
crystal nucleation. Dual-beam FIB/SEM techniques were used to produce high quality
cross-sections and TEM samples to study the cross-sectional microstructures of the
composite coating and the interface between the coating and the substrate. Finally, the
microstructure of the composite coating was compared with pure HA coating.
Additionally, a comparison between GFIB and PFIB in characterization of Col-HA
composite coatings was conducted.
3.2 Materials and methods
3.2.1 Biomimetic Col-HA coating formation
Ti-6Al-4V discs with a thickness of 2 mm and a diameter of 15 mm were
51
polished using 240 grit silicon carbide abrasive paper (LECO Corporation, USA), and
then treated by sandblasting and acid etching (SLA) techniques, followed by UV
irradiation for 8 h. SLA is a widely used surface treatment technique for Ti-based
implants. [5] Briefly, Ti-6Al-4V discs were sandblasted with 46 grit SiC, and then
acid etched by a mixed acid solution of HCl and HNO3, followed with a hot acid
mixture of HCl and H2SO4 for 5 min. After the surface treatment, the Ti-6Al-4V discs
were immersed in the m-SBF containing collagen (0.05 g/L) to produce biomimetic
Col-HA coatings. Type I collagen was extracted from rat tails and purified based on
the protocol reported by Rajan et al.. [32] The m-SBF was prepared as reported
previously (Table 3.1), [5] and the initial pH of the solution was adjusted to 6.75 by
using NaOH and HEPES (4-(2-hydroxyethyl)-1-piperazinneethanesulfonic acid). [7]
During the coating process, urea with a concentration of 0.5 mol/L was added to the
m-SBF containing collagen solution to slow down the collagen polymerization
process. [7] The coating formation was carried out by soaking the surface-treated
Ti-6Al-4V discs in the collagen-containing m-SBF solution at a temperature of 37 °C
for 24 h, as shown in Figure 3.1. As a control, the HA coating was formed by
immersing the surface treated Ti-6Al-4V substrates in the m-SBF solution with an
initial pH of 6.6 at 37 °C for 24 h. Such formed Col-HA and HA coated Ti-6Al-4V
discs were then washed gently with deionized water and dried in air for 24 h.
52
Table 3.1. Ion concentration (mM) of human blood plasma and m-SBF.
Ion Human blood plasma (mM) m-SBF (mM)
Na+ 142.0 109.9
K+ 5.0 6.2
Mg2+ 1.5 1.5
Ca2+ 2.5 7.9
Cl- 103.0 111.2
HCO3- 27.0 17.5
HPO42- 1.0 3.1
SO42- 0.5 0
3.2.2 Materials characterization
The phases of the Col-HA and HA coated Ti-6Al-4V alloys were evaluated by
X-ray diffraction (XRD) using a Bruker AXS D2 Phaser Diffractometer with CuKα
radiation. The diffraction data was collected using a step size of 0.02° and a scan
speed of 0.5 ° /min, and the 2θ range is 5° to 60°. The Col-HA and HA coatings were
scratched from the Ti-6Al-4V alloys, and the coatings were then characterized using
attenuated total reflection-Fourier transformation infrared spectroscopy (ATR-FTIR).
A Nicolet Magna 560 FTIR spectrophotometer (Artisan Technology Group®, IL) with
an ATR setup was used to collect the infrared spectra in the range of 4,000-400 cm-1.
The surface morphology and surface roughness of the Ti-6Al-4V substrates,
Col-HA coatings and HA coatings on the substrate were observed using FEI Helios
dual-beam FIB/SEMs: a Nanolab 460F1 GFIB and a Xenon PFIB in the
UConn/Thermo Fisher Scientific Center for Advanced Microscopy and Materials
53
Analysis (CAMMA). The milling and observation of the cross-sections of the Col-HA
coating on Ti-6Al-4V alloy were also performed in the Helios dual-beam FIB/SEMs.
Both GFIB and PFIB are equipped with EasyLift nano-manipulators to facilitate TEM
sample preparation, and the GFIB also equipped with a flip-stage and a scanning
transmission electron microscopy (STEM) detector for improved final TEM sample
thinning process. Before SEM imaging and FIB milling, both the Col-HA and HA
coatings were sputter coated with a conducting layer of Au using a Polaron E5100
coating instrument.
The applications of GFIB and PFIB to prepare the cross-sections and TEM
samples have been described previously, [5] and the detailed process is as following:
SEM images were obtained using through-lens detector (TLD) or Everhart-Thornley
detector (ETD) in the electron columns. All the sectioning process was performed
with the coating surface normal to the ion column, and 38° to the electron column.
Firstly, a 2-µm-thick Pt strap was deposited on the coating surface at the site of
interest, with the aim to protect the coating from ion beam damage. The Pt layer was
deposited using either electron beam or ion beam to crack organometallic Pt precursor.
[33] The ion beam voltages used in the Pt deposition were 30 kV in GFIB and 8 kV in
the PFIB.
Secondly, a stepped, wedge-shape was milled out using regular cross-section
(RCS) pattern, so that the microstructure of the coating and the interface between the
coating and the substrate could be exposed on the vertical wedge wall, and be
observed by SEM column. In the GFIB, the size of the wedge was 20 µm in length,
54
10 µm in width and 5 µm in depth, and in the PFIB, a large wedge was milled, and the
dimensions were 100 µm in length, 10 µm in width and 5 µm in depth.
Thirdly, after the initial wedge was milled, the vertical wedge wall was polished
by GFIB and PFIB using an ion beam with reduced accelerating current to reduce the
curtaining effects and other milling artefacts. Three different pre-set milling patterns
were used for the polishing process: RCS pattern, cleaning cross-sectional (CCS)
pattern, and rectangle (Rec) pattern.
For the TEM sample preparation, 3-µm-thick Pt strap was deposited at the site of
interest, and then two wedges were milled on both sides of the Pt strap, producing a
pre-thinned FIB lamella. A U-shape was cut on the vertical wedge wall of the
pre-thinned lamella, leaving 2 small pieces on the surface still attached to the bulky
substrate. Then this lamella was welded to the EasyLift W probe, cut totally free from
surrounding material and transferred to a Cu omni-grid. Finally, the lamella was
further thinned to electron transparency on the grid using reduced ion current to
minimize ion damage.
The TEM samples were examined using an FEI Talos F200X S/TEM operating at
an accelerating voltage of 200 kV in the UConn/Thermo Fisher Scientific CAMMA.
This instrument is equipped with a Super-X silicon drift detector (SDD) energy
dispersive X-ray spectrometry (EDS) system, allowing for a rapid acquisition of
spectrum images for elemental mapping.
3.3 Results and discussion
Collagen-HA (Col-HA) composite coatings have been obtained by a biomimetic
55
coating process, in which Col-HA coatings were deposited on the Ti-6Al-4V
substrates by immersing them in the collagen molecules containing m-SBF at 37 °C
and initial pH of 6.75 for 24 h (Figure 3.1). Figure 3.1 shows the process of the
formation of biomimetic Col-HA coatings on the surface-treated Ti-6Al-4V substrates.
Ti-6Al-4V substrates were polished using a 240 grit silicon carbide abrasive paper,
and the scratches were observed due to the silicon carbide abrasive paper (Figure
3.1A). Then the polished substrates were sand-blasted with 46 grit silicon carbide and
acid treated (SLA treatment, Figure 3.1B). The acid treatment produced a uniform
rough surface, which consists of sharp peaks at the microscale level, resulting in
higher surface areas. [35, 36] The acid treated Ti-6Al-4V substrate was then irradiated
using UV light at a wavelength of 254.8 nm for 8 h, resulting in a photocatalytic
chemical reaction. The surface chemistry alteration by photochemical reaction
decontaminates hydrocarbon and increase hydrophilicity by converting Ti4+ sites to
Ti3+ sites. [7, 35, 37] Thus abundant Ti-OH or Ti-O¯ groups are formed on the surface
of the substrate when in contact with SBF solution, imparting the surface with
negative charge and high surface energy. The negative charge and high surface area of
the pre-treated surface will attract and absorb Ca2+ to the surface, contributing to the
precipitation and deposition of coating on the Ti-6Al-4V substrate. [25] Then the
Ti-6Al-4V substrates with a bioactive surface were immersed in the
collagen-containing m-SBF (Figure 3.1C).
56
Figure 3.1. Illustration of the formation process of biomimetic Col-HA composite
coatings on the surface-treated Ti-6Al-4V substrates: Surface morphology of the
Ti-6Al-4V substrate after polishing using silicon carbide abrasive paper (A) and SLA
treatment (B), and coating formation by immersing surface-treated Ti-6Al-4V
substrate in the collagen-containing m-SBF solution (C).
Figure 3.2 shows the surface morphologies of the Col-HA coatings on the
Ti-6Al-4V substrates which is produced as described above. After 24 h of immersion
in the collagen-containing m-SBF, a crack-free homogeneous coating was formed on
the surface of the treated Ti-6Al-4V substrates at 37 °C with an initial pH of 6.75
(Figures 3.2A and 3.2B). The coating demonstrates a porous and rough surface with
numerous globules, and the roughness of the surface was clearly observed when the
Col-HA coated Ti-6Al-4V substrate was tilted 52° in the Dual Beam FIB/SEM
(Figure 3.2C). Compared with the HA coatings on the Ti-6Al-4V substrates, the
globule size of Col-HA coating was significantly smaller than that of HA coating
(Figure 3.3). In our previous research, a crack-free HA coating was formed at an
initial pH of 6.6 (Figure 3.3A). [5] In contrast, higher initial pH was required to
induce Col-HA composite coating, which is in accordance with the report by Xia et al..
[7]
57
Figure 3.2. SE SEM images of Col-HA composite coatings on Ti-6Al-4V substrates.
Col-HA coated Ti-6Al-4V alloy was titled 52° (C).
Figure 3.3. SE SEM images of the surface morphologies of HA (A) and Col-HA (B)
coatings on Ti-6Al-4V substrates.
Figure 3.4 shows the XRD spectra and ATR-FTIR patterns of HA and Col-HA
coatings formed after 24 h of immersion. Characteristic peaks of HA phase at 2θ=26.7°
((002) plane), 31.7-34.3° (overlap of (211), (112), and (300) planes), and 53.8° ((004)
plane) were observed for Col-HA and HA coatings. The XRD spectra indicate that the
crystalline HA presents in both the Col-HA and HA coatings. Further, two additional
peaks at 28.9° ((102) plane) and 50.2° ((213) plane) were only observed in the HA
coating. [38, 39] However, the broad diffraction peak around 31.7-34.3° indicates that
the HA phase in both the Col-HA and HA coatings are poorly crystallized. Weaker HA
peaks together with strong peaks from the hexagonal phase of the Ti-6Al-4V substrate
were observed for the Col-HA coatings. In comparison, the intensity of the Ti peaks
58
detected in the HA coating are much lower (Figure 3.4A). This indicates that a thinner
Col-HA coating was formed on the Ti-6Al-4V substrate compares with the HA
coating. Diffraction peaks for the hexagonal phase of the Ti-6Al-4V substrate were
observed in all of the XRD spectra.
Figure 3.4. XRD (A) and ATR-FTIR (B) spectra of HA and Col-HA coated
Ti-6Al-4V substrates.
HA and Col-HA coatings on the Ti-6Al-4V substrates were scratched from the
substrates and ATR-FTIR was also applied to characterize the coatings. Typical
characteristic peaks of the HA phase were shown in Figure 3.4B. The P=O stretching
mode (~1019.5 cm-1) and P-O bending mode (~597.8 cm-1 and ~560.6 cm-1) indicate
the phosphate ions, and the peak at 873.8 cm-1 is due to the joint contribution from
carbonate and HPO42- ions. [40] The peaks at 1448.0 cm-1 and 1414.5 cm-1 were
assigned to the CO32- groups. The absorption peak at 1638.4 cm-1 was attributed to the
stretching vibration of amide I, which is the characteristics of collagen, indicating the
presence of collagen in the coating. [20, 41]
59
Figure 3.5. SE SEM images of the cross-sections of Col-HA composite coatings on
Ti-6Al-4V substrates prepared using GFIB (A1, A2, A3) and PFIB (B1, B2, B3). The
cross-sections were polished using Rec pattern (A1, B1), CCS pattern (A2, B2), and
RCS pattern (A3, B3). White arrows in B1, B2, and B3 indicate the cracks between Pt
layer and Col-HA coatings.
FIB was then used to mill the Col-HA coated Ti-6Al-4V substrate to study the
details of cross-sectional microstructure of the coating and the interface between the
coating and the substrate. As described in our previous work, [5] a 2-µm-thick dense
Pt protective layer was deposited onto the surface of Col-HA coating at the area of
interest by gas-assisted Pt deposition, protecting the surface of the coating from ion
damage. [42] The ion beam accelerating voltages were 30 kV for GFIB and 8 kV for
PFIB. Then RCS pattern was used to mill a stepped, wedge-shape trench, exposing
the cross-sections of the Col-HA coated Ti-6Al-4V substrates and the interface
between the coatings and substrates. By applying RCS pattern, which is a multi-pass
scan method, the beam scanned through the entire pattern and then repeats for four
passes. Then Rec, CCS, and RCS patterns with reduced current (40 pA for GFIB and
60
24 pA for PFIB) were applied to reduce the “curtaining” effect, which is one of the
main drawbacks and artefacts of FIB milling. The curtaining effect is a result from the
fluctuations of the ion beam current and position, and the inhomogeneity of sample,
[5] which significantly reduces the resolution of the cross-sectional images after the
initial wedge cutting. Thus, eliminating curtaining effect is usually followed after the
initial wedge cutting using RCS pattern. Rec pattern is a pre-set milling pattern, in
which beam repeatedly scans over a rectangular array of equally spaced positions. The
CCS pattern is processed line by line with each line containing a set number of passes.
Figure 3.5 shows the cross-sectional images of Col-HA coated Ti-6Al-4V
substrates polished by Rec, CCS and RCS patterns using a reduced ion beam current
of 40 pA in GFIB (Figure 3.5 (A1-A3)) and an even lower ion beam current of 24 pA
in PFIB (Figure 3.5 (B1-B3)). With stepwise polish of Rec, CCS and RCS patterns in
reduced ion beam current, the curtaining effect on the cross-sectional images of
Col-HA coatings can be attenuated in both GFIB and PFIB, disclosing more details of
the coating cross-sections. Whereas it is also obvious that even low ion beam current
can cause curtaining effect on the cross-sectional images of Col-HA coating. This
issue is predominant in PFIB mainly owing to its larger ion beam spot size compared
to that of GFIB. The cracks between Pt protective layer and Col-coating (marked by
white arrows in Figure 3.5 (B1-B3)) were generated by ion beam of PFIB. [5]
Nevertheless, the cross-sectional details of Col-HA coating have been preserved after
a serial of milling in both GFIB and PFIB, indicating their capability of revealing
inner coating microstructures.
61
Figure 3.6. SE SEM images of the cross-sections of Col-HA (A) and HA (B) coatings
on the Ti-6Al-4V substrates.
In a previous study, our research group for the first time managed to clearly
observe a gradient, porous HA coating using dual beam GFIB and PFIB. [5] In this
work, a similar characteristic was observed in Col-HA coatings. Figure 3.6 shows a
cross-sectional comparison of Col-HA and HA coatings. Both coatings demonstrate a
relatively dense structure adjacent to the interface of the coating and the substrate,
whereas they become more and more porous towards the coating surfaces. The
thicknesses of Col-HA and HA coatings were 1-1.5 μm and 1.5-2.5 μm, respectively.
The cross-sectional SEM images also show that Col-HA coating exhibits a less porous
structure compared to the HA coating possibly owing to the fact that collagen
molecule polymerization occurred simultaneously with the HA nucleation. During this
process, the collagen molecules penetrated into the coating, and therefore filled some
of the pores in the coating, resulting in a less porous structure.
62
Figure 3.7. TEM sample preparation sequence from Col-HA coated Ti-6Al-4V alloy
using GFIB (A-K). The TEM sample was confirmed to be electron transparency by
STEM in GFIB (L).
Figure 3.8. TEM sample preparation sequence of Col-HA coated Ti-6Al-4V alloy
using PFIB (A-K). Col-HA coated Ti-6Al-4V sample was mounted on TEM grid
holder for preparing a TEM sample using PFIB (L).
To study the coating inner structure and the interface between coating and
63
substrate at a nanoscale, TEM samples were prepared from Col-HA coating using
both GFIB and PFIB. A representative extraction sequence of TEM sample
preparation in GFIB is shown in Figure 3.7. Briefly, a 3-μm-thick Pt protective layer
was firstly deposited onto the area of interest (Figure 3.7B). Then two wedge-shaped
trenches were milled out along with either edge of the Pt stripe and further thinned
and cleaned (Figures 3.7C and 3.7D). The sample stage was then tilted to allow a 52°
angle between ion beam and coating cross-section in order to completely cut through
the TEM lamella (Figure 3.7E). The W Easylift probe was then moved adjacent to the
top of the TEM lamella, and an extra Pt layer was deposited to connect the probe and
TEM lamella (Figures 3.7F and 3.7G). After being completely free from the pristine
sample, the TEM lamella was lifted out and welded onto a copper omni-grid (Figures
3.7H and 3.7I). The W Easylift probe was then removed after cutting free from the
TEM lamella (Figure 3.7J), which was finally thinned using ion beam to be electron
transparent (Figures 3.7K and 3.7L). A similar process was conducted on the same
Col-HA sample in PFIB to achieve a TEM lamella (Figure 3.8).
The nanoscale microstructure of the TEM specimens fabricated using GFIB and
PFIB were investigated using S/TEM and EDS. In each case, the representative
images including a bright field (BF) TEM image, a high angle annular dark field
(HAADF) STEM image, and EDS mappings for Ca, C, Ti and the milling ion (Ga and
Xe, respectively) are shown in Figures 3.9 and 3.10, respectively. The BF and
HAADF images confirm that the thickness of the Col-HA coating is 1-1.5 μm, and it
consists of a gradient structure with a more porous layer on the surface, and a denser
64
structure towards the interface of the coating and the substrate. By comparing the Ca,
C, and Ti maps, it is clear that a uniform, defect-free interface between the substrate
and the coating was formed. There was very limited amount of milling ions (Ga or Xe)
detected within the coating that underneath the Pt protective layer. The porous Pt layer
on the Col-HA coating was due to the deposition voltage (30 kV) chosen in PFIB
(Figure 3.10B), which is in accordance with our previous work. [5] There are many
studies reported that a porous surface topography is beneficial to cell adhesion,
proliferation as well as drug loading and delivery, [43-45] while a dense interface
could facilitate a strong bonding between the coating and the substrate, which is an
important parameter for in vivo surgical handling. [7, 46] Moreover, our previous
result suggested that with the incorporation of collagen protein molecules in
hydroxyapatite coating could significantly enhance the biocompatibility of the coating.
[7] Therefore, the Col-HA coated Ti-6Al-4V could be used as a hard tissue
replacement material, and the disclosure of the cross-sectional details in this work
provides an insight on the great potential of utilizing Col-HA composite coating in
clinical applications.
65
Figure 3.9. Characterization of the TEM specimens prepared by GFIB: BFTEM (A)
and HAADF STEM image (B). C-F: Compositional maps obtained from the region in
(B), showing the distribution of Ca (C), C (D), Ti (E), and Ga (F).
Figure 3.10. Characterization of the TEM specimens prepared by PFIB: BFTEM (A)
66
and HAADF STEM image (B). C-F: Compositional maps obtained from the region in
(B), showing the distribution of Ca (C), C (D), Ti (E), and Xe (F).
3.4 Conclusions
A biomimetic collagen-containing hydroxyapatite composite coating was
successfully fabricated on the surface of Ti-6Al-4V by simply immersing the
surface-treated Ti-6Al-4V substrates into a collagen-containing m-SBF. The
cross-sectional microstructure of the Col-HA coating has been mainly investigated by
a dual-beam FIB/SEMs with either gallium (GFIB) or xenon plasma (PFIB) ion
source and the capabilities of these two systems were compared. Both techniques
were able to acquire cross-sectional images and obtain TEM samples from the
Col-HA coating with good quality. It was revealed that the Col-HA coating consists of
a dense layer adjacent to the interface between the coating and the substrate, while a
porous structure towards the surface. Compared to GFIB, PFIB is more time efficient
even at a lower ion beam voltage and current. However, GFIB process induces less
damage to coating and the curtaining effect could be better attenuated.
67
References
[1] S. B. Goodman, Z. Yao, M. Keeney, F. Yang. The future of biologic coatings for
orthopaedic implants. Biomaterials. 2013; 34:3174-3183.
[2] L. T. de Jonge, S. C. Leeuwenburgh, J. J. van den Beucken, J. te Riet, W. F.
Daamen, J. G. Wolke, D. Scharnweber, J. A. Jansen. The osteogenic effect of
electrosprayed nanoscale collagen/calcium phosphate coatings on titanium.
Biomaterials. 2010; 31:2461-2469.
[3] E. Mohseni, E. Zalnezhad, A. R. Bushroa. Comparative investigation on the
adhesion of hydroxyapatite coating on Ti–6Al–4V implant: A review paper.
International Journal of Adhesion and Adhesives. 2014; 48:238-257.
[4] E. Sandrini, C. Giordano, V. Busini, E. Signorelli, A. Cigada. Apatite formation
and cellular response of a novel bioactive titanium. Journal of Materials Science:
Materials in Medicine. 2007; 18:1225-1237.
[5] C. Hu, M. Aindow, M. Wei. Focused ion beam sectioning studies of biomimetic
hydroxyapatite coatings on Ti-6Al-4V substrates. Surface and Coatings Technology.
2017; 313:255-262.
[6] H. Daugaard, B. Elmengaard, J.E. Bechtold, T. Jensen, K. Soballe. The effect on
bone growth enhancement of implant coatings with hydroxyapatite and collagen
deposited electrochemically and by plasma spray. Journal of Biomedical Materials
Research Part A. 2010; 92:913-921.
[7] Z. Xia, X. Yu. Biomimetic collagen/apatite coating formation on Ti6Al4V
substrates. Journal of Biomedical Materials Research Part B: Applied Biomaterials.
68
2012; 100:871-881.
[8] P. Yu, F. Lu, W. Zhu, D. Wang, X. Zhu, G. Tan, X. Wang, Y. Zhang, L. Li, C. Ning.
Bio-inspired citrate functionalized apatite coating on rapid prototyped titanium
scaffold. Applied Surface Science. 2014; 313:947-953.
[9] Y. Huang, Q. Ding, S. Han, Y. Yan, X. Pang. Characterisation, corrosion resistance
and in vitro bioactivity of manganese-doped hydroxyapatite films electrodeposited on
titanium. Journal of Materials Science: Materials in Medicine. 2013; 24:1853-1864.
[10] C. Aparicio, D. Rodriguez, F. Gil. Variation of roughness and adhesion strength
of deposited apatite layers on titanium dental implants. Materials Science and
Engineering: C. 2011; 31:320-324.
[11] D. Chen, E. H. Jordan, M. Gell, M. Wei. Apatite formation on alkaline-treated
dense TiO 2 coatings deposited using the solution precursor plasma spray process.
Acta Biomaterialia. 2008; 4:553-559.
[12] H. W. Kim, Y. H. Koh, L. H. Li, S. Lee, H. E. Kim. Hydroxyapatite coating on
titanium substrate with titania buffer layer processed by sol–gel method. Biomaterials.
2004; 25:2533-2538.
[13] M. Auger, B. Savoini, A. Muñoz, T. Leguey, M. Monge, R. Pareja, J. Victoria.
Mechanical characteristics of porous hydroxyapatite/oxide composites produced by
post-sintering hot isostatic pressing. Ceramics International. 2009; 35:2373-2380.
[14] Z. Xia, X. Yu, X. Jiang, H.D. Brody, D.W. Rowe, M. Wei. Fabrication and
characterization of biomimetic collagen–apatite scaffolds with tunable structures for
bone tissue engineering. Acta Biomaterialia. 2013; 9:7308-7319.
69
[15] J. M. Holzwarth, P. X. Ma. Biomimetic nanofibrous scaffolds for bone tissue
engineering. Biomaterials. 2011; 32:9622-9629.
[16] S. Manara, F. Paolucci, B. Palazzo, M. Marcaccio, E. Foresti, G. Tosi, S.
Sabbatini, P. Sabatino, G. Altankov, N. Roveri. Electrochemically-assisted deposition
of biomimetic hydroxyapatite–collagen coatings on titanium plate. Inorganica
Chimica Acta. 2008; 361:1634-1645.
[17] S. H. Teng, E. J. Lee, C. S. Park, W. Y. Choi, D. S. Shin, H. E. Kim. Bioactive
nanocomposite coatings of collagen/hydroxyapatite on titanium substrates. Journal of
Materials Science: Materials in Medicine. 2008; 19:2453-2461.
[18] K. Hu, X. J. Yang, Y. L. Cai, Z. D. Cui, Q. Wei. Preparation of bone-like
composite coating using a modified simulated body fluid with high Ca and P
concentrations. Surface and Coatings Technology. 2006; 201:1902-1906.
[19] H. Qu, M. Wei. The effect of temperature and initial pH on biomimetic apatite
coating. Journal of Biomedical Materials Research Part B: Applied Biomaterials. 2008;
87:204-212.
[20] H. Qu, Z. Xia, D. A. Knecht, M. Wei. Synthesis of dense collagen/apatite
composites using a biomimetic method. Journal of American Ceramic Society. 2008;
91:3211-3215.
[21] Y. Fan, K. Duan, R. Wang. A composite coating by electrolysis-induced collagen
self-assembly and calcium phosphate mineralization. Biomaterials. 2005;
26:1623-1632.
[22] W. Tong, Z. Yang, X. Zhang, A. Yang, J. Feng, Y. Cao, J. Chen. Studies on
70
diffusion maximum in x‐ray diffraction patterns of plasma‐sprayed hydroxyapatite
coatings. Journal of Biomedical Materials Research Part A. 1998; 40:407-413.
[23] F. Garcia, J. Arias, B. Mayor, J. Pou, I. Rehman, J. Knowles, S. Best, B. Leon, M.
Pérez‐Amor, W. Bonfield. Effect of heat treatment on pulsed laser deposited
amorphous calcium phosphate coatings. Journal of Biomedical Materials Research
Part A. 1998; 43:69-76.
[24] S. Jalota, S. Bhaduri, S.B. Bhaduri, A. C. Tas. A protocol to develop crack-free
biomimetic coatings on Ti6Al4V substrates. Journal of Materials Research. 2007;
22:1593-1600.
[25] E. Zhang, C. Zou, S. Zeng. Preparation and characterization of silicon-substituted
hydroxyapatite coating by a biomimetic process on titanium substrate. Surface and
Coatings Technology. 2009; 203:1075-1080.
[26] M. Iliescu, V. Nelea, J. Werckmann, I. Mihailescu. Transmission electron
microscopy investigation of pulsed-laser deposited hydroxylapatite thin films
prepared by tripod and focused ion beam techniques. Surface and Coatings
Technology. 2004; 187:131-140.
[27] M. Groeber, B. Haley, M. Uchic, D. Dimiduk, S. Ghosh. 3D reconstruction and
characterization of polycrystalline microstructures using a FIB–SEM system.
Materials Characterization. 2006; 57:259-273.
[28] R. Young, C. Rue, S. Randolph, C. Chandler, G. Franz, R. Schampers, A.
Klumpp, L. Kwakman. A comparison of xenon plasma FIB technology with
conventional gallium LMIS FIB: imaging, milling, and gas-assisted applications.
71
Microscopy and Microanalysis. 2011; 17:652-653.
[29] N. Smith, W. Skoczylas, S. Kellogg, D. Kinion, P. Tesch, O. Sutherland, A.
Aanesland, R. Boswell. High brightness inductively coupled plasma source for high
current focused ion beam applications. Journal of Vacuum Science & Technology B.
2006; 24:2902-2906.
[30] T. Burnett, R. Kelley, B. Winiarski, L. Contreras, M. Daly, A. Gholinia, M. Burke,
P. Withers. Large volume serial section tomography by Xe Plasma FIB dual beam
microscopy. Ultramicroscopy. 2016; 161:119-129.
[31] S. Kim, M.J. Park, N.P. Balsara, G. Liu, A.M. Minor. Minimization of focused
ion beam damage in nanostructured polymer thin films. Ultramicroscopy. 2011;
111:191-199.
[32] N. Rajan, J. Habermehl, M. F. Coté, C. J. Doillon, D. Mantovani. Preparation of
ready-to-use, storable and reconstituted type I collagen from rat tail tendon for tissue
engineering applications. Nature Protocol. 2006; 1:2753-2758.
[33] P. Munroe. The application of focused ion beam microscopy in the material
sciences. Materials characterization. 2009; 60:2-13.
[34] H. Qu, M. Wei. Improvement of bonding strength between biomimetic apatite
coating and substrate. Journal of Biomedical Materials Research Part B: Applied
Biomaterials. 2008; 84:436-443.
[35] H. Aita, W. Att, T. Ueno, M. Yamada, N. Hori, F. Iwasa, N. Tsukimura, T. Ogawa.
Ultraviolet light-mediated photofunctionalization of titanium to promote human
mesenchymal stem cell migration, attachment, proliferation and differentiation. Acta
72
Biomaterialia. 2009; 5:3247-3257.
[36] A. Nagaoka, K. I. Yokoyama, J. I. Sakai. Evaluation of hydrogen absorption
behaviour during acid etching for surface modification of commercial pure Ti, Ti–
6Al–4V and Ni–Ti superelastic alloys. Corrosion Science. 2010; 52:1130-1138.
[37] W. Att, N. Hori, F. Iwasa, M. Yamada, T. Ueno, T. Ogawa. The effect of
UV-photofunctionalization on the time-related bioactivity of titanium and chromium–
cobalt alloys. Biomaterials. 2009; 30:4268-4276.
[38] R.K. Brundavanam, G.E.J. Poinern, D. Fawcett. Modelling the crystal structure
of a 30 nm sized particle based hydroxyapatite powder synthesised under the
influence of ultrasound irradiation from X-ray powder diffraction data. American
Journal of Materials Science. 2013; 3:84-90.
[39] C. Hu, M. Zilm, M. Wei. Fabrication of intrafibrillar and extrafibrillar
mineralized collagen/apatite scaffolds with a hierarchical structure. Journal of
Biomedical Materials Research Part A. 2016; 104:1153-1161.
[40] X. Hu, H. Shen, K. Shuai, E. Zhang, Y. Bai, Y. Cheng, X. Xiong, S. Wang, J.
Fang, S. Wei. Surface bioactivity modification of titanium by CO 2 plasma treatment
and induction of hydroxyapatite: in vitro and in vivo studies. Applied Surface Science.
2011; 257:1813-1823.
[41] Z. Xia, M. Wei. Biomimetic fabrication of collagen-apatite scaffolds for bone
tissue regeneration. Journal of Biomaterials and Tissue Engineering. 2013; 3:369-384.
[42] E. Solla, B. Rodríguez-González, H. Aguiar, C. Rodríguez-Valencia, J. Serra, P.
González. Revealing the nanostructure of calcium phosphate coatings using
73
HRTEM/FIB techniques. Materials Characterization. 2016; 122:148-153.
[43] L. Yu, S. Qian, Y. Qiao, X. Liu. Multifunctional Mn-containing titania coatings
with enhanced corrosion resistance, osteogenesis and antibacterial activity. Journal of
Materials Chemistry B. 2014; 2:5397-5408.
[44] H. Hu, W. Zhang, Y. Qiao, X. Jiang, X. Liu, C. Ding. Antibacterial activity and
increased bone marrow stem cell functions of Zn-incorporated TiO2 coatings on
titanium. Acta Biomaterialia. 2012; 8:904-915.
[45] H. W. Kim, J.C. Knowles, H. E. Kim. Hydroxyapatite/poly (ε-caprolactone)
composite coatings on hydroxyapatite porous bone scaffold for drug delivery.
Biomaterials. 2004; 25:1279-1287.
[46] G. Hu, L. Zeng, H. Du, X. Fu, X. Jin, M. Deng, Y. Zhao, X. Liu. The formation
mechanism and bio-corrosion properties of Ag/HA composite conversion coating on
the extruded Mg-2Zn-1Mn-0.5 Ca alloy for bone implant application. Surface and
Coatings Technology. 2017; 325:127-135.
[47] M. Stigter, J. Bezemer, K. De Groot, P. Layrolle. Incorporation of different
antibiotics into carbonated hydroxyapatite coatings on titanium implants. release and
antibiotic efficacy. Journal of Controlled Release, 2004; 99:127-137.
74
4. Fabrication of intrafibrillar calcified collagen fibrils
Abstract
Bone is an organic-inorganic hierarchical biocomposite. Its basic building block
is mineralized collagen fibrils with both intrafibrillar and extrafibrillar mineralization,
which is believed to be regulated by non-collagenous proteins (NCPs) with
polyanionic domains. In this study, collagen fibrils with both intrafibrillar and
extrafibrillar mineralization were successfully prepared and the mechanism of
biomineralization was proposed. To achieve intrafibrillar mineralization, polyacrylic
acid (PAA) was used to sequester calcium and phosphate ions to form fluidic
PAA-amorphous calcium phosphate (PAA-ACP) nanoprecursors. At the presence of
sodium tripolyphosphate (TPP), PAA-ACP nanoprecursors were modulated to orderly
deposit within the gap zone of collagen fibrils. The effect of PAA concentration on
the intrafibrillar and extrafibrillar mineralization of reconstituted collagen fibrils was
investigated. It was found that with the decrease of PAA concentration, the inhibitory
effect of PAA on mineralization and the stability of ACP nanoprecursors decreased.
As a result, more minerals were deposited both within and on the surface of the
collagen fibrils.
Keywords: Intrafibrillar mineralization; extrafibrillar mineralization;
non-collagenous proteins; amorphous calcium phosphate nanoprecursors; hierarchical
structure
75
4.1 Introduction
Bone is an outstanding organic-inorganic nanocomposite, which is made of
mineralized collagen fibrils. It consists of seven hierarchical levels of organization,
[1] ranging from the nanometer scale to the millimeter scale. Nanoscale
hydroxyapatite (HA) crystals are embedded within the interstices of the collagen
fibrils and aligned along the long axis of collagen fibrils, which are in turn present in
bundles or arrays. [2, 3] These aligned mineralized collagen fibrils are arranged into a
multilayered cylindrical structure rotating across the concentric lamellae and forming
osteons. [4] Finally, the osteons are packed densely into compact bone or loosely into
a spongy bone. [5, 6] Since mineralized collagen fibrils are the primary unit of the
complex bone structure, it is important to understand the mineralization mechanism of
collagen fibrils in order to replicate biomineralization of collagen fibrils in vitro.
Biomineralization of collagen fibrils includes oriented HA crystals within and on
the surface of self-assembled collagen fibrils. Triple helical type I collagen molecules
self-assemble into collagen fibrils in a quarter-staggered manner, which leads to 67
nm periodicity with alternating gap and overlap zones. [7, 8] During the
mineralization of collagen fibrils, HA crystals first form within the gap zone, and then
spread through the fibrils, leading to arrays of HA nanocrystals embedded within and
oriented along the collagen fibrils, namely intrafibrillar mineralization. [9-11]
Following intrafibrillar mineralization, HA crystals grow on the surface of the
collagen fibrils, leading to extrafibrillar mineralization. [9] It has been reported that
intrafibrillar mineralization of collagen fibrils attributes to the high mineral content in
76
bone and the improvement of the mechanical and biological properties of the collagen
matrix. [5, 6, 12-14] Meanwhile, extrafibrillar mineralization of collagen fibrils plays
an important role in improvement of mineral content and mechanical properties. [13]
The HA minerals in the bone contribute to better biocompatibility, osteoconductivity
and osteointegration. [15, 16]
Substantial research have been done to deposit HA crystals onto the surface of
collagen fibrils (extrafibrillar mineralization) and within collagen fibrils (intrafibrillar
mineralization), [17-19] and it has been widely accepted that collagen matrix cannot
induce intrafibrillar mineralization by itself. Anionic non-collagenous proteins
(NCPs) are found to play an important role in modulating HA crystal nucleation and
the alignment of apatite deposited within collagen fibrils. [20-23] Owing to the
limited availability and high cost of NCPs, polyanionic electrolytes, such as
polyacrylic acid (PAA) and polyaspartic acid (PAsP), are often used to mimic the
functional groups of NCPs. Their functions are to sequester calcium and phosphate
ions to form fluid amorphous calcium phosphate nanoprecursors (polymer induced
liquid precursors, PILP). [21, 24, 25] It has been identified recently that bone
mineralization initially nucleates from an amorphous phase and then grows into
crystals. [15, 26-29] The advantages of using fluid amorphous precursors are that they
are moldable and infiltrate easily into collagen fibrils. [10, 27] However, polyanionic
macromolecules induced amorphous precursors alone could not reproduce highly
ordered intrafibrillar mineralization. Inorganic polyphosphate (sodium
tripolyphosphate, TPP) is normally used to replicate the template function of
77
phosphoproteins, which modulate HA deposition due to its excellent binding capacity
for calcium ions and its strong affinity to collagen fibrils. [14, 24, 30]
With considerable progresses in unveiling mechanisms of biomineralization of
collagen fibrils in vivo and in vitro replicating the process, most of the recent research
has been focused on the biomineralization of 3-D constructs through a PILP process.
For example, collagen scaffolds, turkey tendon, eggshell membrane and
demineralized human dentine have been biomineralized with intrafibrillar and
extrafibrillar minerals via polyanionic macromolecules-induced liquid precursors. [19,
22, 29, 31-34] Dai et al. have proved that the presence of polyaspartic acid mimicking
NCPs is essential for inducing intrafibrillar mineralization. [35] Meanwhile, Jee et al.
have revealed that oriented HA crystals in turkey tendon can be formed via the PILP
process, in which polyanionic polyaspartic acid is used to mimic the role of NCPs to
facilitate the formation of fluid amorphous nanoprecursors. [19] Li et al. have also
used PAA to mimic NCPs to stabilize amorphous calcium phosphate precursors to
mineralize eggshell membrane with intrafibrillar mineralization. [34]
Although many studies have been conducted to investigate intrafibrillar
mineralization via the PILP process to mimic the structural organization in natural
bone or teeth, little work has been done to investigate the concentration of polyanionic
polymers on the formation of intrafibrillar or extrafibrillar minerals, and there is no
report on the bottom-up process to produce intrafibrillar mineralized collagen fibrils.
In this study, we actively investigated the effect of polyanionic polymer
78
concentrations on the intrafibrillar mineralization of collagen fibrils using a bottom-up
approach in order to develop better understanding of the mineralization process in
natural bone.
4.2 Materials and methods
4.2.1 Materials
Type I collagen was extracted from rat tails, purified based on the protocol
reported by Rajan et al.. [36] Polyacrylic acid (PAA, Mw 2000 Da) and sodium
tripolyphosphate (TPP) were purchased from Sigma-Aldrich. All chemicals were of
analytical chemical grade.
4.2.2 Preparation of intrafibrillar and extrafibrillar mineralized collagen fibrils
Based on our previously established protocol for biomimetic collagen apatite
scaffolds, [37] intrafibrillar and extrafibrilar mineralized collagen fibrils were
fabricated at different PAA concentrations. Briefly, PAA, the sequestration analog,
was added to a modified simulated body fluid (m-SBF) to form stabilized amorphous
calcium phosphate (ACP) nanoprecursors (PAA-ACP). The m-SBF was prepared as
reported previously: NaOH, K2HPO4·3H2O, MgCl2·6H2O, CaCl2 were added in
deionized water (DIW) in the order as they are listed, and the ion concentrations were
adjusted according to Table 4.1. [38-40] Concentrations of PAA in PAA-ACP
nanoprecursors were 0, 0.4, 0.8, 2, and 5.6 mg/mL, and the pH of each PAA-ACP
nanoprecursor was adjusted to 7.2 by addition of HEPES
(4-(2-hydroxyapatiteethyl)-1-piperazineethanesulfonic acid) and NaOH. Type I
79
collagen stock solution (4.5 mg/mL) was diluted to a given concentration (0.2 mg/mL)
using PAA-ACP nanoprecursors. TPP (1.2 wt%), the templating analog, was added to
the collagen solution to template the hierarchical arrangement of hydroxyapatite
within collagen fibrils. The ion concentrations of all chemicals in the suspensions are
listed in Table 4.1. All of the reactions were taken under 4 ℃, and then transferred to
a water bath at 37 ℃ for 24 hr.
Table 4.1. Ion concentrations of suspension and body plasma
Ion Human blood plasma (mM) Suspension (mM)
Na+ 142.0 439.0
K+ 5.0 25.0
Mg2+ 1.5 6.0
Ca2+ 2.5 31.7
Cl- 103.0 444.6
HCO3- 27.0 70
HPO42- 1.0 12.5
SO42- 0.5 0
4.2.3 Characterization
PAA-ACP nanoprecursors with a PAA concentration of 0.4 mg/mL were added
to the collagen solution, and TPP (1.2 wt%) was then added to template minerals
deposition within collagen fibrils. The pH of the solution was adjusted to 7.2 using
HEPES and NaOH. After pH adjustment, the solution was dropped onto Formvar
carbon coated copper grid, rinsed with DIW, dried under room temperature, and then
observed using transmission electron microscopy (TEM, JEOL JEM-2010) at an
80
accelerating voltage of 80 kV. Five TEM images were analyzed using Image J
software to achieve the average size of ACP nanoprecursors and amorphous calcium
phosphate nanoparticles.
In order to examine the mineralization of collagen fibrils, intrafibrillar and
extrafibrillar mineralized collagen fibrils using PAA-ACP nanoprecursors at different
PAA concentrations were deposited on copper grid and observed using TEM. The
corresponding mineral phase was evaluated by selected area electron diffraction
(SAED) coupled to TEM. The surface morphology of the mineralized collagen fibrils
using PAA-ACP nanoprecursors at different PAA concentrations was also imaged
using field emission scanning electron microscopy (FESEM, JEOL JSM-6335F,
Japan).
4.3 Results
To investigate the effect of PAA concentration in PAA-ACP nanoprecursors on
the biomineralization of collagen fibrils, different PAA concentrations (0, 0.4, 0.8, 2,
5.6 mg/mL) were used to induce PAA-ACP nanoprecursor formation. PAA-ACP
nanoprecursors were added to the type I collagen stock solution and TPP was used to
modulate mineral arrangement in the collagen fibrils. Figure 4.1 shows the TEM
images of unstained collagen fibrils with different PAA concentrations in PAA-ACP
nanoprecursors. An increasingly distinct banding pattern (D-banding) in
electron-dense nucleation sites can be observed in collagen fibrils with decreasing
PAA concentration, implying that the mineral content formed within the gap zone of
81
collagen fibrils increased with the reduction of PAA concentration (Figure 4.1). At a
PAA concentration of 5.6 mg/mL, vague D-periods are observed (Figure 4.1A),
indicating electron-dense minerals deposited, poorly reproducing the D-banding
pattern of naturally mineralized collagen fibrils normally observed in bone. With the
decrease of PAA concentration, the intrafibrillar mineralization becomes more
obvious and homogeneous (Figure 4.1A-D), which is attribute to the higher mineral
content formed within the collagen fibrils. The inserted SAED result of the
mineralized collagen fibrils formed at the PAA concentration of 2 mg/mL (Figure
4.1B) does not exhibit a ring pattern, which indicates that the minerals formed within
the collagen fibrils are either amorphous or nanocrystallined. The control is the
sample with no PAA in the collagen-containing calcium phosphate solution. As such,
no D-banding is observed, suggesting that minerals are not deposited in specific
nucleating sites within collagen fibrils. Both SEM and TEM results reveal that HA
minerals (white arrow, Figure 4.2B) only formed on the surface of collagen fibrils
(Figure 4.2). SAED result (Figure 4.2A, insert) exhibits rings of typical low
crystalline apatite. These results collectively suggest that ordered mineralization
within collagen fibrils cannot be achieved without the presence of NCPs analogs.
82
Figure 4.1. TEM images of unstained mineralized collagen fibrils with different PAA
concentrations: (A) 5.6 mg/mL, (B) 2 mg/mL (insert: SAED result), (C) 0.8 mg/mL
and (D) 0.4 mg/ mL.
83
Figure 4.2. TEM image (A) and SEM image (B) of mineralized collagen fibrils
without PAA. SAED result for mineralized collagen fibrils without PAA (Insert, A).
White arrow in (B): HA minerals on the surface. The cracks on collagen fibrils were
induced by high voltage employed during SEM operation.
Figure 4.3 shows FESEM images of mineralized collagen fibrils with different
PAA concentrations. Vague D-banding is observed at a PAA concentration of 5.6
mg/mL (Figure 4.3A, white arrow). This suggests that few minerals have deposited
within collage fibrils. The D-banding becomes more distinguished with the decrease
of the PAA concentration, indicating more minerals are incorporated into the collagen
fibrils (Figures 4.3A-D), which is in accordance with the TEM result (Figure 4.1).
White dotted arrow and white arrow heads in Figure 4.3D show minerals formed on
the surface of collagen fibrils (extrafibrillar mineralization) with low PAA
concentration in PAA-ACP nanoprecursors. The inset image in Figure 4.3D is the
enlarged image indicated by white dotted arrow, and white thin arrows show minerals
aligned on the surface of collagen fibrils with an almost 67 nm spacing distance. This
suggests that this specific site is the entering sites for calcium phosphate nanoclusters
(Figure 4.3D). Moreover, intrafibrillar and extrafibrillar mineralized collagen fibrils
are arranged into bundles or fused to form collagen fibers (Figure 4.3).
84
Figure 4.3. FESEM images of mineralized collagen fibril bundles with different PAA
concentrations (white arrows indicate D-banding): (A) 5.6 mg/mL, (B) 2 mg/mL, (C)
0.8 mg/mL (the insert is the enlarged image of the spot indicated by the white arrow),
and D) 0.4 mg/mL (white dotted arrow and white arrow heads show minerals on the
collagen surface), minerals aligned on the surface of collagen fibrils with an almost 67
nm spacing distance (white dotted arrow and thin arrows in insert image). The insert
is the enlarged image of the spot indicated by the white dotted arrow.
4.4 Discussion
Since mineralized collagen fibrils are the primary unit of the complex bone
structure, investigation of the intrafibrillar and extrafibrillar mineralization of collagen
fibrils has become essential. It is widely accepted that biomineralization is well
regulated by NCPs, such as DMP 1, [41] osteonectin, osteocalcin and bone
85
sialoprotein. [8] Two functional motifs in NCPs have been found to play an important
role in regulation of nucleation, dimension and order of HA deposition within natural
bone. First, acidic sequestrating motif of NCPs binds to calcium to form
calcium-bound organic materials, and thereby inhibits the precipitation of calcium
phosphate. This in turn forms stabilized amorphous calcium phosphate
nanopresursors. [42] PAA was used in our study to replicate the sequestration
functional groups in NCPs. The –COO- groups of PAA attract calcium ions from
solution, which further recruit phosphate ions into the fluid droplets to form
supersaturated amorphous calcium phosphate nanoprecursors (Figures 4.4 and 4.5).
[35] Electron dense nanoprecursors were observed in Figure 4.4, and amorphous
calcium phosphate nanoparticles were also shown, indicating nanophase separation
has formed in the PAA-ACP nanoprecursors. The mean size of nanoprecursors is
37.61±6.46 nm. Amorphous calcium phosphate nanoparticles were observed in the
PAA-ACP nanoprecursor droplets with a mean size of 6.70±1.04 nm measured by
Image J. The small size of ACP nanoprecursors and ACP nanoparticles makes it
possible for PAA-ACP to infiltrate into collagen fibrils. It has been reported that
matrix vesicles (MVs) play a crucial role in bone mineralization. During bone
mineralization, MVs with a diameter of 40-200 nm are secreted during osteoid
formation, and observed at sites where mineralized tissue is synthesized de novo. [42]
As a result, MVs are considered the initial mineralization site. MV membranes
enriched with acidic phospholipids have high affinities for calcium cations, which can
induce the formation of stable calcium phosphate-phospholipid complex. Obviously,
86
each polymer stabilized amorphous calcium phosphate nanoprecursor has a similar
function as an MV, so PAA-ACP is regarded as an analog to MV. [42, 43]
Besides, templating motif of NCPs regulates the alignment of HA within collagen
fibrils through its excellent binding capacity for calcium ions and collagen fibrils. [22]
Due to the low cost and availability of the chemical, sodium tripolyphosphate (TPP)
has been used as an analog to NCPs in templating the HA deposition within collagen
fibrils in the current study. [24] Highly negatively charged TPP accumulates in the
gap zone of collagen fibrils dominated by highly positive net charge through
electrostatic attraction. [8, 11, 24] The high binding capacity of TPP to calcium and
collagen makes it possible to act as an ionic bridge to attract PAA-ACP
nanoprecursors to enter the gap zone of collagen fibrils (Figure 4.5). [24] PAA-ACP
nanoprecursors were found to accumulate in the close vicinity of collagen fibrils and
in contact with collagen fibrils, demonstrating the attraction of PAA-ACP to the
collagen fibrils (Figure 4.4A). Once the PAA-ACP nanoprecursors penetrate the gap
zone of collagen fibrils via capillary action and electrostatic attraction, the desorption
of calcium from PAA is expected because TPP is a much stronger pentanionic
chelating agent to ACP comparing to the carboxylate groups of PAA. [44] When a
high concentration of TPP (2%) is added to PAA-ACP nanoprecursors, white
precipitates are formed in the solution (data not shown). This also suggests that TPP
has attracted calcium ions from PAA-ACP and induced the aggregation of ACP in the
gap zone of collagen fibrils (see Figure 4.4A). The free polyanionic macromolecules
would exit the gap zone and re-enter the collagen and PAA containing m-SBF
87
solution to continuously recruit free calcium and phosphate ions to form PAA-ACP
nanoprecursors and re-infiltrate the fibrils to form ACP minerals. [35] These ACP
minerals are then crystalized under the influence of the amino acid side chains of
collagen fibrils. [33, 45, 46] A large fraction of charged amino acid side chains in the
gap zone helps to recruit and bind calcium and phosphate ions, and the resulting
calcium-phosphate prenucleation cluster become crystallized HA with time. [8, 34,
46] Thus the gap zone of collagen fibrils serves as the entering (as inferred from the
inset image in Figure 4.3D) and mineralization sites, which forms the D-banding of
collagen fibrils (Figure 4.5). This is in agreement with the finding by Nudelman et al.,
who reported that the gap region is the infiltrating and nucleating site for the
intrafibrillar mineraliation of collagen fibrils. [11]
Figure 4.4. TEM images of collagen solution containing PAA-ACP nanoprecursor
droplets on copper grid. (A) Low magnification TEM image showing PAA-ACP
nanoprecursor droplets that were formed in the close vicinity of the collagen fibrils.
White arrow shows the collagen fibrils that have been removed. (B) & (C) High
magnification images of PAA-ACP nanoprecursor droplets. (C) One electron dense
88
PAA-ACP nanoprecursor droplet highlighted in the white circle consisting of many
amorphous calcium phosphate nanoparticles.
Fibrils with intrafibrillar and extrafibrillar mineralization are fabricated using
PAA as a sequestration analog to induce amorphous calcium phosphate
nanoprecursors and TPP as a templating analog, and the mechanisms involved have
been illustrated in Figure 4.5. [47, 48] The rate of collagen mineralization is
dependent on the concentration of PAA. The lower the concentration of PAA, the
weaker the inhibition of PAA for ACP nanoprecursor precipitation, the faster the
mineralization and crystallization process, and the more minerals deposited. [10, 49,
50] When the PAA concentration is high, vague D-banding has been observed,
implying that only a small number of minerals are formed in the gap zone of collagen
fibrils (Figures 4.1A and 4.2A). This is due to that the ACP nanoprecursors are highly
stabilized by PAA (Figures 4.1 and 4.3). With the decreasing concentration of PAA,
PAA-ACP precursors became less stable and crystallization process could be quick
enough to form both intra- and extra-minerals, which can be proved by the increase of
the mineral amount and formation of extrafibrillar mineralization (Figures 4.1, 4.2 and
4.3).
89
Figure 4.5. Schematic illustration of the intrafibrillar mineralization process induced
by PAA as a sequestration analog and TPP as a templating analog.
4.5 Conclusions
Self-assembly of collagen fibrils and intrafibrillar and extrafibrillar
mineralization were achieved using PAA as sequestration analog and TPP as a
templating analog. The effect of the PAA concentration on intrafibrillar and
extrafibrillar mineralization was realized by adjusting the concentration of NCPs
analog during the biomineralization process. Thus, this study demonstrates a
promising approach to produce intrafibrillar mineralized collagen fibrils using a
bottom-up approach.
90
References
[1] S. Weiner, H. D. Wagner. The material bone: structure-mechanical function
relations. Annual Review of Materials Science. 1998; 28:271-298.
[2] W. Landis, M. Song, A. Leith, L. McEwen, B. McEwen. Mineral and organic
matrix interaction in normally calcifying tendon visualized in three dimensions by
high-voltage electron microscopic tomography and graphic image reconstruction.
Journal of Structural Biology. 1993; 110:39-54.
[3] T. T. Thula, D. E. Rodriguez, M. H. Lee, L. Pendi, J. Podschun, L. B. Gower. In
vitro mineralization of dense collagen substrates: a biomimetic approach toward the
development of bone-graft materials. Acta Biomaterialia. 2011; 7:3158-3169.
[4] W. Wagermaier, H. Gupta, A. Gourrier, M. Burghammer, P. Roschger, P. Fratzl.
Spiral twisting of fiber orientation inside bone lamellae. Biointerphases. 2006; 1:1-5.
[5] I. Jäger, P. Fratzl. Mineralized collagen fibrils: a mechanical model with a
staggered arrangement of mineral particles. Biophysical Journal. 2000; 79:1737-1746.
[6] M. J. Olszta, X. Cheng, S. S. Jee, R. Kumar, Y. Y. Kim, M. J. Kaufman, et al. Bone
structure and formation: a new perspective. Materials Science and Engineering: R:
Reports. 2007; 58:77-116.
[7] F. Z. Cui, Y. Li, J. Ge. Self-assembly of mineralized collagen composites.
Materials Science and Engineering: R: Reports. 2007; 57:1-27.
[8] Q. Wang, X. Wang, L. Tian, Z. Cheng, F. Cui. In situ remineralizaiton of partially
demineralized human dentine mediated by a biomimetic non-collagen peptide. Soft
Matter. 2011; 7:9673-9680.
91
[9] X. Wang, F. Cui, J. Ge, Y. Wang. Hierarchical structural comparisons of bones
from wild-type and liliput dtc232 gene-mutated Zebrafish. Journal of Structural
Biology. 2004; 145:236-245.
[10] F. Nudelman, P. H. Bomans, A. George, N. A. Sommerdijk. The role of the
amorphous phase on the biomimetic mineralization of collagen. Faraday Discuss.
2012; 159:357-370.
[11] F. Nudelman, K. Pieterse, A. George, P. H. Bomans, H. Friedrich, L. J. Brylka, et
al. The role of collagen in bone apatite formation in the presence of hydroxyapatite
nucleation inhibitors. Nature materials. 2010; 9:1004-1009.
[12] R. B. Martin, D. B. Burr, N. A. Sharkey. Skeletal tissue mechanics. 1998.
[13] A. K. Nair, A. Gautieri, M. J. Buehler. Role of intrafibrillar collagen
mineralization in defining the compressive properties of nascent bone.
Biomacromolecules. 2014; 15:2494-2500.
[14] Y. Liu, D. Luo, X. X. Kou, X. D. Wang, F. R. Tay, Y. L. Sha, et al. Hierarchical
intrafibrillar nanocarbonated apatite assembly improves the nanomechanics and
cytocompatibility of mineralized collagen. Advanced Functional Materials. 2013;
23:1404-1411.
[15] Z. Xia, M. Wei. Biomimetic fabrication of collagen-apatite scaffolds for bone
tissue regeneration. Journal of Biomaterials and Tissue Engineering. 2013; 3:369-384.
[16] A. J. W. Johnson, B. A. Herschler. A review of the mechanical behavior of CaP
and CaP/polymer composites for applications in bone replacement and repair. Acta
Biomaterialia. 2011; 7:16-30.
92
[17] Z. Xia, M. Villa, M. Wei. A biomimetic collagen–apatite scaffold with a
multi-level lamellar structure for bone tissue engineering. Journal of Materials
Chemistry B. 2014; 2:1998-2007.
[18] S. S. Jee, L. Culver, Y. Li, E. P. Douglas, L. B. Gower. Biomimetic
mineralization of collagen via an enzyme-aided PILP process. Journal of Crystal
Growth. 2010; 312:1249-1256.
[19] S. S. Jee, R. K. Kasinath, E. DiMasi, Y. Y. Kim, L. Gower. Oriented
hydroxyapatite in turkey tendon mineralized via the polymer-induced liquid-precursor
(PILP) process. CrystEngComm. 2011; 13:2077-2083.
[20] S. Gajjeraman, K. Narayanan, J. Hao, C. Qin, A. George. Matrix macromolecules
in hard tissues control the nucleation and hierarchical assembly of hydroxyapatite.
The Journal of Biological Chemistry. 2007; 282:1193-1204.
[21] S. S. Jee, T. T. Thula, L. B. Gower. Development of bone-like composites via the
polymer-induced liquid-precursor (PILP) process. Part 1: Influence of polymer
molecular weight. Acta Biomaterialia. 2010; 6:3676-3686.
[22] Y. K. Kim, L. Gu, T. E. Bryan, J. R. Kim, L. Chen, Y. Liu, et al. Mineralisation of
reconstituted collagen using polyvinylphosphonic acid/polyacrylic acid templating
matrix protein analogues in the presence of calcium, phosphate and hydroxyl ions.
Biomaterials. 2010; 31:6618-6627.
[23] H. A. Goldberg, K. J. Warner, M. C. Li, G. K. Hunter. Binding of bone
sialoprotein, osteopontin and synthetic polypeptides to hydroxyapatite. Connective
Tissue Research. 2001; 42:25-37.
93
[24] Y. Liu, Y. K. Kim, L. Dai, N. Li, S. O. Khan, D. H. Pashley, et al. Hierarchical
and non-hierarchical mineralisation of collagen. Biomaterials. 2011; 32:1291-1300.
[25] R. Q. Song, H. Colfen, A. W. Xu, J. Hartmann, M. Antonietti.
Polyelectrolyte-directed nanoparticle aggregation: systematic morphogenesis of
calcium carbonate by nonclassical crystallization. Acs Nano. 2009; 3:1966-1978.
[26] L. Addadi, D. Joester, F. Nudelman, S. Weiner. Mollusk shell formation: a source
of new concepts for understanding biomineralization processes. Chemistry-A
European Journal. 2006; 12:980-987.
[27] J. Mahamid, A. Sharir, L. Addadi, S. Weiner. Amorphous calcium phosphate is a
major component of the forming fin bones of zebrafish: Indications for an amorphous
precursor phase. Proceedings of the National Academy of Sciences. 2008;
105:12748-12753.
[28] W. J. Landis, F. H. Silver. Mineral deposition in the extracellular matrices of
vertebrate tissues: identification of possible apatite nucleation sites on type I collagen.
Cells, Tissues, Organs. 2008; 189:20.
[29] J. Mahamid, B. Aichmayer, E. Shimoni, R. Ziblat, C. Li, S. Siegel, et al.
Mapping amorphous calcium phosphate transformation into crystalline mineral from
the cell to the bone in zebrafish fin rays. Proceedings of the National Academy of
Sciences. 2010; 107:6316-6321.
[30] Y. Liu, N. Li, Y. p. Qi, L. Dai, T. E. Bryan, J. Mao, et al. Intrafibrillar collagen
mineralization produced by biomimetic hierarchical nanoapatite assembly. Advanced
Materials. 2011; 23:975-980.
94
[31] B. Antebi, X. Cheng, J. N. Harris, L. B. Gower, X. D. Chen, J. Ling. Biomimetic
collagen–hydroxyapatite composite fabricated via a novel perfusion-flow
mineralization technique. Tissue Engineering Part C: Methods. 2013; 19:487-496.
[32] J. Li, J. Yang, J. Li, L. Chen, K. Liang, W. Wu, et al. Bioinspired intrafibrillar
mineralization of human dentine by PAMAM dendrimer. Biomaterials. 2013;
34:6738-6747.
[33] B. Marelli, C. E. Ghezzi, A. Alessandrino, J. E. Barralet, G. Freddi, S. N. Nazhat.
Silk fibroin derived polypeptide-induced biomineralization of collagen. Biomaterials.
2012; 33:102-108.
[34] N. Li, L. Niu, Y. Qi, C. K. Yiu, H. Ryou, D. D. Arola, et al. Subtleties of
biomineralisation revealed by manipulation of the eggshell membrane. Biomaterials.
2011; 32:8743-8752.
[35] L. Dai, E. P. Douglas, L. B. Gower. Compositional analysis of a polymer-induced
liquid-precursor (PILP) amorphous CaCO 3 phase. Journal of Non-Crystalline Solids.
2008; 354:1845-1854.
[36] N. Rajan, J. Habermehl, M. F. Coté, C. J. Doillon, D. Mantovani. Preparation of
ready-to-use, storable and reconstituted type I collagen from rat tail tendon for tissue
engineering applications. Nature Protocol. 2006; 1:2753-2758.
[37] Z. Xia, X. Yu, X. Jiang, H. D. Brody, D. W. Rowe, M. Wei. Fabrication and
characterization of biomimetic collagen–apatite scaffolds with tunable structures for
bone tissue engineering. Acta Biomaterialia. 2013; 9:7308-7319.
[38] Z. Xia, X. Yu. Biomimetic collagen/apatite coating formation on Ti6Al4V
95
substrates. Journal of Biomedical Materials Research Part B: Applied Biomaterials.
2012; 100:871-881.
[39] H. Qu, Z. Xia, D. A. Knecht, M. Wei. Synthesis of dense collagen/apatite
composites using a biomimetic method. Journal of American Ceramic Society. 2008;
91:3211-3215.
[40] X. Yu, L. Wang, F. Peng, X. Jiang, Z. Xia, J. Huang, et al. The effect of fresh
bone marrow cells on reconstruction of mouse calvarial defect combined with
calvarial osteoprogenitor cells and collagen–apatite scaffold. Journal of Tissue
Engineering and Regeneration Medicine. 2013; 7:974-983.
[41] G. He, S. Gajjeraman, D. Schultz, D. Cookson, C. Qin, W. T. Butler, et al.
Spatially and temporally controlled biomineralization is facilitated by interaction
between self-assembled dentin matrix protein 1 and calcium phosphate nuclei in
solution. Biochemistry. 2005; 44:16140-16148.
[42] H. Ozawa, K. Hoshi, N. Amizuka. Current concepts of bone biomineralization.
Journal of Oral Bioscience. 2008; 50:1-14.
[43] K. Hoshi, S. Ejiri, H. Ozawa. Organic components of crystal sheaths in bones.
Journal of Electron Microscopy (Tokyo). 2001; 50:33-40.
[44] T. Saito, A. Arsenault, M. Yamauchi, Y. Kuboki, M. Crenshaw. Mineral induction
by immobilized phosphoproteins. Bone. 1997; 21:305-311.
[45] B. Marelli, C. E. Ghezzi, Y. L. Zhang, I. Rouiller, J. E. Barralet, S. N. Nazhat.
Fibril formation pH controls intrafibrillar collagen biomineralization in vitro and in
vivo. Biomaterials. 2015; 37:252-259.
96
[46] Z. Xu, Y. Yang, W. Zhao, Z. Wang, W. J. Landis, Q. Cui, et al. Molecular
mechanisms for intrafibrillar collagen mineralization in skeletal tissues. Biomaterials.
2015; 39:59-66.
[47] G. Boivin, P. Meunier. The degree of mineralization of bone tissue measured by
computerized quantitative contact microradiography. Calcified Tissue International.
2002; 70:503-511.
[48] D. Toroian, J. E. Lim, P. A. Price. The Size Exclusion Characteristics of Type I
Collagen Implications for the Role of Noncollagenous Bone Constitutes in
Mineralization. Journal Biological Chemistry. 2007; 282:22437-22447.
[49] F. R. Tay, D. H. Pashley. Guided tissue remineralisation of partially
demineralised human dentine. Biomaterials. 2008; 29:1127-1137.
[50] C. Holt, S. Lenton, T. Nylander, E. S. Sørensen, S. C. Teixeira. Mineralisation of
soft and hard tissues and the stability of biofluids. Journal of Structural Biololy. 2014;
185:383-396.
97
5. Intrafibrillar silicification of collagen fibrils
Abstract
Inspired by the silicification process in nature, intrafibrillar silicified collagen
fibrils were successfully fabricated using a biomimetic one-step collagen
self-assembly/silicification approach, in which collagen self-assembly and
intrafibrillar silicification occurred simultaneously. To the best of our knowledge, it is
the first time that intrafibrillar silicified collagen fibrils were formed via a one-step
simultaneous approach, where collagen fibrils served as a templating matrix, and
poly(allylamine) hydrochloride and sodium tripolyphosphate acted as the respective
positive and negative analogs of zwitterionic proteins. By tailoring zwitterionic
proteins analogs, silicified collagen fibrils with different microstructures, including
core-shell, twisted and banded structures, were achieved. The intrafibrillar silicified
collagen fibrils demonstrated significant improvement of osteoblasts proliferation
compared with apatite mineralized collagen fibrils. These findings open a new avenue
for preparation of silicon-containing hierarchical biocomposites for biomedical needs.
Keywords: Biomimetic; intrafibrillar silicification; hierarchical structure; zwitterionic
proteins; one-step approach
98
5.1 Introduction
Silicon is an essential element contributing significantly to the health of bone and
cartilage. [1] It has been discovered that silicon-containing organic-inorganic
biocomposites with unique hierarchical structures exist in numerous natural biological
systems, such as bone, fish scales, diatom frustules, and spicules of glass sponges.
[2-5] For example, the long spicules of Hyalonema sieboldi demonstrate a perfect
hierarchical structure, remarkable optical properties, and excellent durability and
flexibility. [6] More importantly, silicon-containing biomaterials have been found to
be osteoinductive for new bone formation and stimulative for neovascularization
without supplements of growth factors. [7, 8] Due to the unique properties of the
silicon-containing biomaterials, great enthusiasm has been generated and extensive
effort has been made to replicate these materials to be used in different biomedical
applications. [6, 9, 10]
In recent years, effort has been directed to develop an in-depth understanding of
the biosilicification mechanism so as to replicate the process in vitro and thereby
fabricate organic-inorganic hybrid biomaterials with a hierarchical structure.
Chemical analyses of glass sponge and diatom have shown that fibrillar collagen can
serve as templates for silicification, while the highly zwitterionic proteins, such as
silaffins, silicateins, and long-chain polyamines, are actively involved in the interplay
with collagen and the mediation of silicification process. [6, 11] The zwitterionic
proteins exhibit both positive charges introduced by the quaternary ammonium groups
and negative charges by the phosphorylated serine residues. [12] The zwitterionic
99
nature of these proteins enables them to form aggregates, which controls the
silicification process and silica morphology. [13] Polyamines have been found to
induce silica precipitation in vitro in the presence of polyanions, where polyamines
and polyanions act as positive and negative analogs of zwitterionic proteins,
respectively. [3] Despite all the successes in fabrication of silicon-containing
biomaterials, an in-depth understanding of the mechanism of biomimetic silicification
process has not yet been achieved.
A variety of silicified collagen hybrids have been produced, such as
silica-collagen xerogels, and silicified collagen hydrogels with different silicon
sources (silica and silicates). [14-18] Typically, silica is formed on the surface of
collagen fibrils (extrafibrillar silicification), resulting in a weak bond between silica
and collagen. [14, 15, 19] However, the recent successes in intrafibrillar calcification
of collagen fibrils have shed light on the understanding of intrafibrillar silicification
(forming silica within collagen fibrils), as intrafibrillar calcification and silicification
of collagen fibrils may share a similar mechanism. In our previous work, intrafibrillar
calcified collagen fibrils have been achieved by a one-step collagen self-assembly and
calcification approach. [20, 21] In the present study, we also propose to produce
intrafibrillar silicified collagen fibrils via a one-step collagen
self-assembly/silicification (OCSS) approach, where collagen self-assembly and
silicification occur simultaneously with the aid of positive and negative analogs of
zwitterionic proteins. To test this hypothesis, a systemic study was conducted to
investigate the effect of zwitterionic proteins analogs on the OCSS approach, and an
100
in vitro cell culture study was also conducted to test the biocompatibility of the
intrafibrillar silicified collagen fibrils.
5.2 Materials and methods
5.2.1 Materials
Type I collagen was extracted from rat tails, purified based on the protocol
reported by Rajan et al.. [22] Poly(allylamine) hydrochloride (PAH, Mw = 15,000 Da)
and sodium tripolyphosphate (TPP) were purchased from Sigma-Aldrich. All
chemicals were of analytical chemical grade.
5.2.2 Preparation of hydrolyzed silicic acid
A 3 wt% hydrolyzed silicic acid stock solution was prepared as followed:
Silbond® 40 (40% hydrolyzed tetraethyl orthosilicate (TEOS); Silbond Corp., Weston,
MI, USA), absolute ethanol, deionized water (DIW) and 37% HCl were mixed at a
molar ratio of 1.03:199:6.11:0.03 for 1 hour at room temperature to hydrolyze TEOS
into orthosilicic acid and its oligomers. [23]
5.2.3 Preparation of PAH-stabilized silica (PAH-SA) precursors
PAH-stabilized silica (PAH-SA) precursors were prepared by the following
procedure: A solution with a PAH concentration of 0.267 mM was mixed with 3 wt%
silicic acid at a 1:1 volume ratio to obtain 1.5 wt% silicic acid stabilized by PAH.
After stirring for 1 min, the mixture was centrifuged at 3000 RPM and 25 °C for 10
min. The PAH-stabilized silica (PAH-SA) precursors were then collected. [24]
101
5.2.4 Controlled biomimetic silicification of collagen fibrils via an OCSS approach
Buffer solution was used to avoid sudden pH changes and prepared based on our
established protocol, [20, 25] where 184.8 mM NaCl, 6.3 mM K2HPO4·3H2O and
100.7 mM HEPES (4-(2-hydroxyethyl)piperazine-1-ethane-sulfonic acid) were added
to DIW in the order listed above.
*Collagen-silica (Col-SA) fibrils: Type I collagen stock solution (4.5 mg/mL)
was diluted to a given concentration (0.3 mg/mL) using buffer solution at 4 °C. Then
different amounts of silicic acid solution (3 wt%) were added dropwise to the diluted
collagen solution immediately with vigorous stirring. The amount of silicic acid added
was controlled at a SA/Col ratio of 1:2, 1:1, 2:1, 3:1 and 4:1. After mixing the diluted
collagen solution with the silicic acid solution and stirring for 5 min, the mixtures (pH
= 6.4) were incubated at 25 °C for 10 min, and then incubated in a water bath at 37 °C
for 24 h.
*Collagen-poly(allylamine) hydrochloride-silica (Col-PAH-SA) fibrils: Type I
collagen stock solution (4.5 mg/mL) was diluted to a given concentration (0.3 mg/mL)
using the buffer solution at 4 °C. Then different amounts of PAH-SA precursors were
added dropwise into the diluted collagen solution immediately with vigorous stirring.
The amount of PAH-SA precursors was controlled at a SA/Col ratio of 1:1, 4:3 and
2:1. Higher ratios of SA/Col were not used since precipitates formed rather quickly
when the ratios of SA/Col were 3:1 and 4:1. After mixing the diluted collagen
solution with PAH-SA precursors and stirring for 5 min, Col-PAH-SA mixtures (pH =
6.4) were incubated at 25 °C for 10 min, and then transferred to a water bath at 37 °C
102
for 24 h.
*Intrafibrillar silicified Collagen (Col-inSA) fibrils: Type I collagen stock
solution (4.5 mg/mL) was diluted to a given concentration (0.3 mg/mL) using the
buffer solution at 4 °C. TPP (1.2 wt%) was added to this diluted collagen solution and
stirred for 5 min. Different amounts of PAH-SA precursors were then added dropwise
into this collagen-TPP solution with vigorous stirring. The amount of PAH-SA
precursors was controlled at a SA/Col ratio of 1:1, 4:3 and 2:1. After stirring for 5 min,
Collagen containing mixtures (pH = 6.8) were incubated at 25 °C for 10 min, and then
transferred to a water bath at 37 °C for 24 h.
5.2.5 Preparation of silicified collagen fibrils via a two-step approach
Firstly, collagen fibrils without mineralization were produced by the following
procedure: type I collagen stock solution (4.5 mg/mL) was diluted to a given
concentration (0.3 mg/mL) using the buffer solution at 4 °C, followed by incubation
at 25 °C for 10 min and then in a water bath at 37 °C for another 24 h to precipitate
collagen fibrils. Secondly, the collagen fibrils without mineralization were silicified
by incubating in silica precursors, which was described below. *Col-SA fibrils via a
two-step approach (T-Col-SA): Silicic acid was added dropwise to the collagen fibrils
at a SA/Col ratio of 4:1, and then incubated at 37 °C for another 24 h. *Col-PAH-SA
fibrils via a two-step approach (T-Col-PAH-SA): PAH-SA precursors were added to
the self-assembled collagen fibrils at a SA/Col ratio of 4:3, and then incubated at
37 °C for 24 h. *Col-inSA fibrils via a two-step approach (T-Col-inSA): TPP was
dissolved in the self-assembled collagen fibrils at a concentration of 1.2 wt%, and
103
PAH-SA precursors (SA/Col = 4:3) were added after 5 min of stirring. The mixed
solution was then incubated at 37 °C for 24 h.
5.2.6 Materials characterization
Transmission electron microscopy (TEM): Silicified collagen fibrils and collagen
fibrils without mineralization were dropped onto Formvar carbon coated copper grids,
rinsed with DIW and then dried at room temperature. TEM (JEOL, JEM-2010) was
used to examine the silicification of collagen fibrils at an accelerating voltage of 80
kV. All samples were examined without staining.
Field emission scanning electron microscopy (FESEM): FESEM (JEOL
JSM-6335F, Japan) was used to characterize the surface morphology of the silicified
collagen fibrils. Different silicified collagen fibrils were sputter-coated with gold and
examined using FESEM operating at 8-10 kV.
Thermogravimetric analysis (TGA): TGA was performed using a Q500 TGA
analyzer (TGAQ-500, TA Instrument, New Castle, DE, USA). About 20 mg
freeze-dried silicified collagen fibrils were placed in a platinum pan and heated from
30 to 900 °C at a rate of 10 °C/min in air to determine the amount of minerals in the
silicified collagen fibrils.
Attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR):
The freeze dried silicified collagen fibrils were characterized using ATR-FTIR. A
Nicolet Magna 560 FT-IR spectrophotometer (Artisan Technology Group ®, Illinois,
USA) with an ATR setup was used to collect infrared spectra in the range 4,000-400
cm-1 at a 4 cm-1 resolution, 32 scans.
104
5.2.7 In vitro cell culture and proliferation
Mouse calvaria 3T3-E1 (MC3T3-E1, ATCC, USA) cell line, an osteoblast
precursor cell line, was used to determine the cytocompatibility of various samples.
MC3T3-E1 cells were plated in a culture flask (FALCON, USA) containing 10 mL of
α-minimum essential medium (α-MEM; Gibco, Invitrogen, Inc., USA), which was
supplemented with 10% fetal bovine serum (FBS; Gibco, Invitrogen, Inc., USA) and
1% penicillin–streptomycin (Gibco, Invitrogen, Inc., USA). In a humidified
atmosphere of 5% CO2 under 37 °C that maintained by an incubator (NAPCO, USA),
cells were further passaged when reaching 80-90% confluence. The fibrous
membranes of 5.5 mm in diameter were collected and cross-linked with 1 wt% EDC
(1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide). Collagen-apatite membrane with
an apatite content of about 67 wt% was produced using a co-precipitation approach,
which serves as a control. The cross-linked fibrous membranes were sterilized and
placed in a 96-well plate.
To observe cell proliferation and viability using alamarBlue assay, cells were
seeded on various samples (four replicates) in 96-well plates at a cell density of 8,000
cells per well in 100 μL cell culture medium. The cells were cultured for 14 days and
the cell culture medium was refreshed every 2 days. After incubating for 1 day,
samples that attached with cells were transferred to a new 96-well plate to leave
behind the cells escaped from samples. Phenol-red-free cell culture medium with 10%
alamarBlue (Thermo Scientific, USA) was added to each well. After culturing for
another 4 h, 100 μL of the culture solution from each well was transferred to a new
105
96-well plate to measure the absorbance at wavelengths of 570 nm and 600 nm by a
mircoplate reader (BioTek, USA). The samples were then rinsed against PBS twice to
remove residue alamarBlue, and 100 μL fresh cell culture medium was added to each
well and refreshed every other day. After culturing for 4, 7, 10, and 14 days, cell
viabilities were investigated according to the procedure descripted above. Percentage
of reduction of alamarBlue was calculated according to the instruction provided by
the assay.
5.2.8 Statistical analysis
Statistically significant differences (p) between various groups were measured
using two-way RM ANOVA analysis of variance that carried out by GraphPad Prism
5.
5.3 Results and discussion
5.3.1 Col-SA fibrils
Silicified collagen fibrils with a core-shell structure are successfully produced
using a one-step approach. When silicic acid is mixed with collagen molecules with a
SA/Col ratio of 3:1 or 4:1, a thick layer of dense silica is coated on the collagen fibrils,
forming a core-shell structured Col-SA hybrid (Figures 5.1A1, 5.1A2, 5.1B1 and
5.1B2). This result is in accordance with the reports by Eglin et al. and Ono et al.,
where hollow silica fibrils were produced with silica coated onto the surface of
templating collagen fibrils. [26, 27] Some Col-SA fibrils also demonstrate a twisted
structure, which is due to the presence of negatively charged silanol groups of silica
106
coating on the surface of Col-SA hybrid (Figures 5.1B1, 5.1B2, and 5.1C). [9, 28]
When the SA/Col ratio is low (1:2, 1:1, or 2:1), silicified collagen fibrils with
core-shell structures are not formed (Data not shown). When the SA/Col ratio is 3:1,
the distance between two adjacent protrusions is 66.9±5.8 nm (n = 30, Figures 5.1A1
and 5.1B1), which is identical to the distance of the gap zone present in collagen
fibrils (67 nm), indicating the templating function of collagen fibrils. Negatively
charged silicic acid preferentially accumulates at sites of highly positively charged
gap zone of collagen fibrils, leading to the formation of protrusions with a periodicity
of about 67 nm (Figure 5.1D). [15, 29] When the SA/Col ratio is 4:1, the shell
thickness of the Col-SA core-shell structure increases, and the periodicity of the
distinguished protrusions increases to 82.7±21.0 nm (n = 40, Figure 5.1A2 and 5.1B2).
This is due to that with the increase of the SA/Col ratio, more and more silica
precipitates on the surface of collagen fibrils, resulting in an increase in shell
thickness and a decrease in electrostatic attraction between silicic acid and the gap
zone of collagen fibrils. As a result, the surface of the coating was smoothened, and
some adjacent protrusions fused together, leading to a longer distance between two
adjacent protrusions (Figures 5.1A2 and 5.1B2).
107
Figure 5.1. TEM images (A1 and A2), FESEM images (B1 and B2) of Col-SA fibrils
with different SA/Col ratios: 3:1 (A1, B1), 4:1 (A2, B2). ATR-FTIR spectra (C) of
Col-SA fibrils (Stars indicate the peaks for Si-O-Si, and triangle shows the peak for
Si-OH). Schematic illustration (D) of the process to form Col-SA fibrils, and the TEM
image shows Col-SA fibrils with a SA/Col ratio of 3:1.
5.3.2 Col-PAH-SA fibrils
By controlling the silica precursors, twisted silicified collagen fibrils with
bandings are produced (Figure 5.2). A long-chain polyamine (poly(allylamine)
hydrochloride (PAH, Mw = 15,000 Da)) was used to stabilize silica by the formation
of hydrogen bonds with silanol groups of silicic acid, which reduces the availability of
silicic acid from forming polysilicic acid, and thereby leads to the formation of fluidic
PAH-SA nanoprecursors. [4] Rope-like, twisted silicified collagen fibrils are formed
via a one-step approach by mixing PAH-SA precursors and collagen molecules
108
(Figures 5.2A-D). This result is in agreement with the reports by Desimone et al. and
Heinemann et al., [9, 28] where twisted fibrils were formed due to the presence of
negatively charged silanol groups of silica (Figure 5.2D). The average distances of
two adjacent bandings for Col-PAH-SA fibrils are 143.3±12.1 nm, 145.8±7.5 nm, and
145.0±11.8 nm for SA/Col ratio of 1:1, 4:3 and 2:1, respectively (Figures 5.2A-C, n =
30). All the banding distances are significantly larger than that of the typical
D-banding of collagen fibrils (67 nm). This may be due to that the amine groups of
PAH in PAH-SA precursors interact with carboxyl groups of collagen molecules
during the self-assembly of collagen fibrils, where PAH is inserted between two
adjacent collagen molecules as a space linker, leading to a longer distance between
adjacent gap zones (Figure 5.2E). In PAH-SA precursors, silica is stabilized by PAH
through ionic bonding and/or hydrogen bonding. [4] Subsequently, PAH-SA
precursors infiltrate into collagen fibrils along with PAH during the collagen fibril
formation process. PAH is then separated from silicic acid, leading to the precipitation
of silica at the C-terminus of collagen molecules, where the potential energy is the
lowest. [29] This explains why the banding in Figure 5.2 is larger than that of the
typical collagen D-banding because of the longer distance between adjacent gap zones
of the fibrils and the precipitation of electron-dense minerals at the C-terminus of
collagen molecules. The FTIR result also confirms the attachment of PAH onto
collagen fibrils (Figure 5.2D). The 528 cm-1 peak of FT-IR is assigned to the amine
hydrogen bonding of PAH, indicating the attachment of PAH onto the silicified
collagen fibrils (Figure 5.2D).
109
Figure 5.2. TEM images of Col-PAH-SA fibrils with different SA/Col ratios: 1:1 (A),
4: 3 (B), and 2:1 (C). ATR-FTIR spectra (D) of Col-PAH-SA fibrils (Stars indicate the
peaks for Si-O-Si, triangle shows the peak for Si-OH, and arrow indicates amine
hydrogen bonding). Schematic illustration (E) of the process to form Col-PAH-SA
fibrils, and the TEM image shows Col-PAH-SA fibrils with a SA/Col ratio of 4:3.
5.3.3 Col-inSA fibrils
As shown in Figure 5.3, intrafibrillar silicified collagen fibrils were produced
using PAH and TPP as the positive and negative analogs of zwitterionic proteins,
respectively. Compared to Col-PAH-SA fibrils, the morphology of collagen fibrils in
the Col-inSA group exhibits a normal structure instead of twisted fibrils (Figure 5.3).
[20] This may be due to the competition between TPP and PAH to interact with
collagen molecules during the collagen self-assembly process. As a penta-anion, TPP
110
has a stronger electrostatic attraction to the gap zone of collagen fibrils than PAH,
leading to prohibition of further bonding of collagen molecules with PAH. Instead,
fluidic PAH-SA precursors infiltrate into the collagen fibrils and precipitate within the
gap zone (Figure 5.5). Clear D-bandings are observed using TEM due to
electron-dense silica precipitating within specific sites of collagen fibrils (Figures 5.3
and 5.4). As shown in Figure 5.3, the D-banding becomes more and more obvious
with the increase of SA/Col ratios, and silica nanoparticles are also formed on the
surface of collagen fibrils. Similar to silaffins and long-chain polyamines in diatoms,
PAH and TPP have been used as respective positive and negative analogs of highly
zwitterionic proteins to promote silica formation in vitro. [30] In details, PAH mimics
the positively charged groups of lysine residues in silaffins, while TPP mimics the
negatively charged groups introduced by phosphorylation of serine residues. [3, 31]
Similar to the mechanism of calcium phosphate intrafibrillar mineralization, TPP
accumulates in the gap zone of collagen fibrils due to electrostatic attraction between
the highly negatively charged TPP and the positively charged gap zone of collagen
fibrils. [20, 29] TPP further attracts fluidic PAH-SA nanoprecursors to infiltrate into
gap zones of collagen fibrils via electrostatic attraction and capillary action (Figures
5.5 and 5.6). [4, 32, 33] PAH promotes silica precipitation in the presence of TPP in
the gap zone of collagen fibrils, leading to intrafibrillar silicification of collagen
fibrils (Figure 5.6). [31, 34]
111
Figure 5.3. TEM images (A1, B1, and C1) and FESEM images (A2, B2, and C2) of
Col-inSA fibrils with different SA/Col ratios: 1:1 (A1, A2), 4:3 (B1, B2), and 2:1 (C1,
C2).
Figure 5.4. ATR-FTIR spectra (A) and TGA spectra (B) of Col-inSA fibrils with
different SA/Col ratios. (Stars indicate the peaks for Si-O-Si, triangle shows the peak
for Si-OH, and arrow indicates amine hydrogen bonding.)
112
Figure 5.5. The initial stage of PAH-SA precursors interacted with collagen fibrils in
the presence of TPP as a regulator. Low (A) and high (B) magnification of the fluidic
silica precursors infiltrated into the collagen fibrils during the collagen self-assembly
process. White arrows in B: Vague D-banding can be observed.
Figure 5.6. Schematic illustration of the intrafibrillar mineralization process of
Col-inSA fibrils. TEM image shows Col-inSA fibrils with a SA/Col ratio of 4:3.
5.3.4 Silicified collagen fibrils via a two-step approach
Moreover, silicification of collagen fibrils via a one-step approach was compared
with a two-step approach. In a two-step approach, collagen fibrils were first
self-assembled, and then silica precursors were added to the self-assembled collagen
113
fibril suspension to enable the silicification of the precipitated collagen fibrils.
Compared T-Col-SA fibrils with Col-SA fibrils at the same SA/Col ratio
(SA/Col=4:1), silica particles rather than a continuous coating are deposited on the
surface of collagen fibrils (Figures 5.7A1 and 5.7A2). Silica particles may be
absorbed onto the collagen fibrils due to electrostatic attraction, or through hydrogen
bonding between silanol groups of silica and amine or hydroxyl groups of collagen. [6]
In the T-Col-PAH-SA fibrils, silica particles are only formed on the surface of
collagen fibrils, while twisted silicified collagen fibrils are formed in Col-PAH-SA
fibrils using one-step approach (Figures 5.7B1 and 5.7B2). The comparison between
the T-Col-PAH-SA and the Col-PAH-SA fibrils indicates that in the one-step approach
PAH and silica are well integrated into the collagen self-assembly process and that the
silicification and self-assembly of collagen fibrils occur simultaneously. In contrast,
intrafibrillar silicification was observed in the Col-inSA fibrils formed via both the
two-step and the one-step processes (Figures 5.7C1 and 5.7C2). In both the two-step
and the one-step approaches, highly negatively charged TPP accumulates in the
positively charged gap zone of collagen fibrils through electrostatic attraction, which
further attracts fluidic PAH-SA precursors to infiltrate into collagen fibrils. [4, 29, 31]
PAH and TPP function as the positive and negative analogs of zwitterionic proteins,
respectively, which mediate the biosilicification process in nature systems. Similarly,
the accumulation of PAH-SA and TPP in the gap zone of collagen fibrils enables the
PAH-SA precursors to aggregate and form silica within collagen fibrils. [3, 13]
Two kinds of polyanions were used: phosphate ions and TPP, but they have
114
different functions. TPP inhibits the incorporation of PAH during the collagen
self-assembly process (Figures 5.7B2 and 5.7C2), and it also functions as a regulator
to guide precise deposition of silica within collagen fibrils (Figures 5.5B, 5.7B1 and
5.7C1). Phosphate ions, being one kind of kosmotrope anions, were used to favor
collagen self-assembly. [16] The comparison between T-Col-PAH-SA and T-Col-inSA
fibrils indicates that, even though both phosphate ions and TPP induce phase
separation of PAH-SA precursors, [34, 35] phosphate ions do not function as a
regulator to attract PAH-SA to infiltrate into collagen fibrils (Figure 5.7B1), but TPP
does (Figure 5.7C1).
Figure 5.7. Col-SA fibrils produced using either a two-step process (T-Col-SA (A1),
T-Col-PAH-SA (B1), and T-Col-inSA (C1)) or a one-step approach (Col-SA (A2),
Col-PAH-SA (B2), and Col-inSA (C2)).
115
5.3.5 In vitro cell culture
Different mineralized collagen fibrous membranes were tested for their ability to
support cell growth in vitro. An osteoblastic cell line MC3T3-E1 was used.
Collagen-apatite (Col-Ap) matrices have been widely used as scaffolds for bone tissue
engineering, thus the Col-Ap fibrous membrane was used as the control. [36-42] Cell
proliferation and viability on different fibrous membranes were quantified using
alamarBlue assay. The Col-inSA group supports a higher proliferation rate of
MC3T3-E1 cells, especially at the late stage of incubation, compared with Col-Ap,
Col-SA and Col-PAH-SA (Figure 5.8). This result indicates that the intrafibrillar
silicified collagen fibrous membrane has a better biological response to osteoblast
cells compared with Col-Ap. It is well-established that silicon-induced cell responses
are dose-dependent. [43, 44] As shown in the TEM and SEM images of Col-inSA
fibrils (Figure 5.3), silica is wrapped up by collagen fibrils, demonstrating dense and
obvious D-banding structures, allowing gradual release of silicon. [45] Therefore,
intrafibrillar silicified collagen fibril can serve as a carrier and gradually release
silicon ions from the collagen fibrils to positively impact cell activities instead of
burst release causing negative cell responses. [46]
116
Figure 5.8. Proliferation of MC3T3-E1 cells on mineralized collagen membranes:
Collagen-apatite (Col-Ap, control), Col-SA (SA/Col=4:3), Col-PAH-SA
(SA/Col=4:3), and Col-inSA (SA/Col=4:3) over the course of 14 days. Data is
presented as mean ± standard error. Statistically significant differences (p) between
various groups were measured using two-way RM ANOVA analysis of variance, and
*: p<0.05, **: p<0.01, ***: p<0.001.
5.4 Conclusions
We have presented a one-step collagen self-assembly/silicification approach in
preparation of intrafibrillar silicified collagen fibrils. Highly negatively charged TPP
first accumulates in the positively charged gap zone of collagen fibrils due to
electrostatic attraction, which further attracts positively charged fluidic PAH-SA
precursors to enter the collagen fibrils. As the positive analog of the zwitterionic
proteins in silicification, PAH promotes silica precipitation in the presence of negative
analog, TPP. Thus the accumulation of PAH-SA and TPP in the gap zone of collagen
fibrils facilitates the formation of silica within gap zone of collagen fibrils, resulting
117
in intrafibrillar silicification. Moreover, the structure of the silicified collagen fibrils
can be manipulated by varying silica precursors, polyanions and SA/Col ratios to
produce silicified collagen fibrils with a core-shell, twisted, or banded structure. The
intrafibrillar silicified collagen fibrils possess better cell compatibility compared with
the collagen-apatite fibrils. Thus, this approach provides a facile method to produce
intrafibrillar silicified collagen fibrils for biomedical applications.
118
References
[1] C. C. Perry, T. Keeling-Tucker. Biosilicification: the role of the organic matrix in
structure control. JBIC Journal of Biological Inorganic Chemistry. 2000; 5:537-550.
[2] E. Beniash. Biominerals—hierarchical nanocomposites: the example of bone.
Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology. 2011;
3:47-69.
[3] S. Wenzl, R. Hett, P. Richthammer, M. Sumper. Silacidins: highly acidic
phosphopeptides from diatom shells assist in silica precipitation in vitro. Angewandte
Chemie. 2008; 120:1753-1756.
[4] L.-n. Niu, K. Jiao, H. Ryou, A. Diogenes, C. K. Yiu, A. Mazzoni, et al.
Biomimetic silicification of demineralized hierarchical collagenous tissues.
Biomacromolecules. 2013; 14:1661-1668.
[5] A. Woesz, J. C. Weaver, M. Kazanci, Y. Dauphin, J. Aizenberg, D. E. Morse, et al.
Micromechanical properties of biological silica in skeletons of deep-sea sponges.
Journal of Materials Research. 2006; 21:2068-2078.
[6] H. Ehrlich, R. Deutzmann, E. Brunner, E. Cappellini, H. Koon, C. Solazzo, et al.
Mineralization of the metre-long biosilica structures of glass sponges is templated on
hydroxylated collagen. Nature Chemistry. 2010; 2:1084-1088.
[7] A. J. Mieszawska, N. Fourligas, I. Georgakoudi, N. M. Ouhib, D. J. Belton, C. C.
Perry, et al. Osteoinductive silk–silica composite biomaterials for bone regeneration.
Biomaterials. 2010; 31:8902-8910.
[8] W. Zhai, H. Lu, L. Chen, X. Lin, Y. Huang, K. Dai, et al. Silicate bioceramics
119
induce angiogenesis during bone regeneration. Acta Biomaterialia. 2012; 8:341-349.
[9] M. F. Desimone, C. Hélary, G. Mosser, M. M. Giraud-Guille, J. Livage, T. Coradin.
Fibroblast encapsulation in hybrid silica–collagen hydrogels. Journal of Materials
Chemistry. 2010; 20:666-668.
[10] T. Delclos, C. Aimé, E. Pouget, A. Brizard, I. Huc, M. H. Delville, et al.
Individualized silica nanohelices and nanotubes: Tuning inorganic nanostructures
using lipidic self-assemblies. Nano Letters. 2008; 8:1929-1935.
[11] S. Heinemann, H. Ehrlich, T. Douglas, C. Heinemann, H. Worch, W. Schatton, et
al. Ultrastructural studies on the collagen of the marine sponge Chondrosia reniformis
Nardo. Biomacromolecules. 2007; 8:3452-3457.
[12] N. Kröger, S. Lorenz, E. Brunner, M. Sumper. Self-assembly of highly
phosphorylated silaffins and their function in biosilica morphogenesis. Science. 2002;
298:584-586.
[13] C. C. Lechner, C. F. Becker. Silaffins in Silica Biomineralization and Biomimetic
Silica Precipitation. Marine Drugs. 2015; 13:5297-5333.
[14] M. F. Desimone, C. Hélary, I. B. Rietveld, I. Bataille, G. Mosser, M. M.
Giraud-Guille, et al. Silica–collagen bionanocomposites as three-dimensional
scaffolds for fibroblast immobilization. Acta Biomaterialia. 2010; 6:3998-4004.
[15] S. Heinemann, C. Heinemann, R. Bernhardt, A. Reinstorf, B. Nies, M. Meyer, et
al. Bioactive silica–collagen composite xerogels modified by calcium phosphate
phases with adjustable mechanical properties for bone replacement. Acta
Biomaterialia. 2009; 5:1979-1990.
120
[16] S. Quignard, G. J. Copello, C. Aimé, I. Bataille, C. Hélary, M. F. Desimone, et al.
Influence of Silicification on the Structural and Biological Properties of Buffer‐
Mediated Collagen Hydrogels. Advanced Engineering Materials. 2012; 14:B51-B55.
[17] S. Heinemann, C. Heinemann, H. Ehrlich, M. Meyer, H. Baltzer, H. Worch, et al.
A novel biomimetic hybrid material made of silicified collagen: perspectives for bone
replacement. Advanced Engineering Materials. 2007; 9:1061-1608.
[18] S. Heinemann, H. Ehrlich, C. Knieb, T. Hanke. Biomimetically inspired hybrid
materials based on silicified collagen. International Journal of Materials Research.
2007; 98:603-608.
[19] K. Spinde, M. Kammer, K. Freyer, H. Ehrlich, J. N. Vournakis, E. Brunner.
Biomimetic silicification of fibrous chitin from diatoms. Chemistry of Materials. 2011;
23:2973-2978.
[20] L. Zhang. C. Hu, Mei Wei. Development of biomimetic scaffold woth both
intrafibrilar and extrafibrillar mineralization. ACS Biomaterials Science &
Engineering. 2015; 1:669-676.
[21] C. Hu, M. Zilm, M. Wei. Fabrication of intrafibrillar and extrafibrillar
mineralized collagen/apatite scaffolds with a hierarchical structure. Journal of
Biomedical Materials Research Part A. 2016; 104:1153-1161.
[22] N. Rajan, J. Habermehl, M. F. Coté, C. J. Doillon, D. Mantovani. Preparation of
ready-to-use, storable and reconstituted type I collagen from rat tail tendon for tissue
engineering applications. Nature Protocol. 2006; 1:2753-2758.
[23] L. Niu, K. Jiao, Y. Qi, C. K. Yiu, H. Ryou, D. D. Arola, et al. Infiltration of silica
121
inside fibrillar collagen. Angewandte Chemie. 2011; 123:11892-11895.
[24] L. Niu, K. Jiao, H. Ryou, C. K. Yiu, J. Chen, L. Breschi, et al. Multiphase
intrafibrillar mineralization of collagen. Angewandte Chemie. 2013; 125:5874-5878.
[25] D. A. Cisneros, J. Friedrichs, A. Taubenberger, C. M. Franz, D. J. Muller.
Creating ultrathin nanoscopic collagen matrices for biological and biotechnological
applications. Small. 2007; 3:956-963.
[26] Y. Ono, Y. Kanekiyo, K. Inoue, J. Hojo, M. Nango, S. Shinkai. Preparation of
Novel Hollow Fiber Silica Using Collagen Fibers as a Template. Chemistry Letters.
1999:475-476.
[27] D. Eglin, G. Mosser, M. M. Giraud-Guille, J. Livage, T. Coradin. Type I collagen,
a versatile liquid crystal biological template for silica structuration from nano-to
microscopic scales. Soft Matter. 2005; 1:129-131.
[28] S. Heinemann, C. Heinemann, M. Jager, J. Neunzehn, H. Wiesmann, T. Hanke.
Effect of silica and hydroxyapatite mineralization on the mechanical properties and
the biocompatibility of nanocomposite collagen scaffolds. ACS Applied Materials &
Interfaces. 2011; 3:4323-4331.
[29] F. Nudelman, K. Pieterse, A. George, P. H. Bomans, H. Friedrich, L. J. Brylka, et
al. The role of collagen in bone apatite formation in the presence of hydroxyapatite
nucleation inhibitors. Nature Materials. 2010; 9:1004-1009.
[30] N. Kröger, R. Deutzmann, M. Sumper. Polycationic peptides from diatom
biosilica that direct silica nanosphere formation. Science. 1999; 286:1129-1132.
[31] N. Li, L. Niu, Y. Qi, C. K. Yiu, H. Ryou, D. D. Arola, et al. Subtleties of
122
biomineralisation revealed by manipulation of the eggshell membrane. Biomaterials.
2011; 32:8743-8752.
[32] Y. Liu, Y. K. Kim, L. Dai, N. Li, S. O. Khan, D. H. Pashley, et al. Hierarchical
and non-hierarchical mineralisation of collagen. Biomaterials. 2011; 32:1291-1300.
[33] Y. K. Kim, L.-s. Gu, T. E. Bryan, J. R. Kim, L. Chen, Y. Liu, et al. Mineralisation
of reconstituted collagen using polyvinylphosphonic acid/polyacrylic acid templating
matrix protein analogues in the presence of calcium, phosphate and hydroxyl ions.
Biomaterials. 2010; 31:6618-6627.
[34] E. Brunner, K. Lutz, M. Sumper. Biomimetic synthesis of silica nanospheres
depends on the aggregation and phase separation of polyamines in aqueous solution.
Phys Chem Chem Phys. 2004; 6:854-857.
[35] J. Fothergill, M. Li, S. A. Davis, J. A. Cunningham, S. Mann.
Nanoparticle-Based Membrane Assembly and Silicification in Coacervate
Microdroplets as a Route to Complex Colloidosomes. Langmuir. 2014; 30:14591-6.
[36] Z. Xia, M. Wei. Biomimetic fabrication of collagen-apatite scaffolds for bone
tissue regeneration. Journal of Biomaterials and Tissue Engineering. 2013; 3:369-384.
[37] Z. Xia, X. Yu, X. Jiang, H. D. Brody, D. W. Rowe, M. Wei. Fabrication and
characterization of biomimetic collagen–apatite scaffolds with tunable structures for
bone tissue engineering. Acta Biomaterialia. 2013; 9:7308-7319.
[38] Y. Pek, S. Gao, M. M. Arshad, K. J. Leck, J. Y. Ying. Porous collagen-apatite
nanocomposite foams as bone regeneration scaffolds. Biomaterials. 2008;
29:4300-4305.
123
[39] M. M. Villa, L. Wang, J. Huang, D. W. Rowe, M. Wei. Bone tissue engineering
with a collagen–hydroxyapatite scaffold and culture expanded bone marrow stromal
cells. Journal of Biomedical Materials Research Part B: Applied Biomaterials. 2015;
103:243-253.
[40] X. Yu, Z. Xia, L. Wang, F. Peng, X. Jiang, J. Huang, et al. Controlling the
structural organization of regenerated bone by tailoring tissue engineering scaffold
architecture. Journal of Materials Chemistry. 2012; 22:9721-9730.
[41] X. Yu, L. Wang, F. Peng, X. Jiang, Z. Xia, J. Huang, et al. The effect of fresh
bone marrow cells on reconstruction of mouse calvarial defect combined with
calvarial osteoprogenitor cells and collagen–apatite scaffold. Journal of Tissue
Engineering and Regenerative Medicine. 2013; 7:974-983.
[42] M. M. Villa, L. Wang, J. Huang, D. W. Rowe, M. Wei. Improving the
permeability of lyophilized collagen–hydroxyapatite scaffolds for cell‐based bone
regeneration with a gelatin porogen. Journal of Biomedical Materials Research Part B:
Applied Biomaterials. 2015; 104:1580-1590.
[43] H. A. Santos, J. Riikonen, J. Salonen, E. Mäkilä, T. Heikkilä, T. Laaksonen, et al.
In vitro cytotoxicity of porous silicon microparticles: effect of the particle
concentration, surface chemistry and size. Acta Biomaterialia. 2010; 6:2721-2731.
[44] E. A. Bortoluzzi, L. Niu, C. D. Palani, A. R. El-Awady, B. D. Hammond, D. Pei,
et al. Cytotoxicity and osteogenic potential of silicate calcium cements as potential
protective materials for pulpal revascularization. Dental Materials. 2015;
31:1510-1522.
124
[45] L. Niu, K. Jiao, Y. Qi, S. Nikonov, C. K. Yiu, D. D. Arola, et al. Intrafibrillar
silicification of collagen scaffolds for sustained release of stem cell homing
chemokine in hard tissue regeneration. The FASEB Journal. 2012; 26:4517-4529.
[46] S. Bhattacharjee, I. M. Rietjens, M. P. Singh, T. M. Atkins, T. K. Purkait, Z. Xu,
et al. Cytotoxicity of surface-functionalized silicon and germanium nanoparticles: the
dominant role of surface charges. Nanoscale. 2013; 5:4870-4883.
125
6. Intrafibrillar calcified collagen scaffolds
Abstract
A biomimetic collagen-hydroxyapatite (Col-HA) scaffold resembling the
composition and structure of natural bone from the nano- to macro-scale has been
successfully prepared for bone tissue engineering. We have developed a bottom-up
approach to fabricate hierarchically biomimetic Col-HA scaffolds with both
intrafibrillar and extrafibrillar mineralization. To achieve intrafibrillar mineralization,
polyacrylic acid (PAA) was used as a sequestrating analog of non-collagenous
proteins (NCPs) to form a fluidic amorphous calcium phosphate (ACP) nanoprecursor
through attraction of calcium and phosphate ions. Sodium tripolyphosphate (TPP) was
used as a templating analog to regulate orderly deposition of HA within collagen
fibrils. Both X-ray diffraction and Fourier transform infrared spectroscopy suggest
that the mineral phase was HA. Field emission scanning electron microscopy,
transmission electron microscopy and selected area electron diffraction indicate that
hierarchical collagen-HA scaffolds were formed with both intrafibrillar and
extrafibrillar mineralization. Biomimetic Col-HA scaffolds with both intrafibrillar and
extrafibrillar mineralization (IE-Col-Ap) were compared to Col-HA scaffolds with
extrafibrillar mineralization only (E-Col-Ap) as well as pure collagen scaffolds in
vitro for cellular proliferation using MC3T3-E1 cells. Pure collagen scaffolds had the
highest rate of proliferation, while there was no statistically significant difference
between IE-Col-Ap and E-Col-Ap scaffolds. This approach renders a promising
126
Col-HA scaffold, which bears high resemblance to natural bone in terms of
composition and structure.
Keywords: Intrafibrillar mineralization; bottom-up approach; hierarchical scaffolds;
amorphous calcium phosphate nanoprecursors
127
6.1 Introduction
In the United States, approximately one million bone fractures are treated
every year. [1-3] Autografts have long been considered as the gold standard for
bone grafting procedures. However, the amount of bone that can be harvested is
limited and there is a high potential for donor site morbidity. Allografts are
alternatives to autografts, but there are concerns of disease transmission. [4-7]
Thus a promising strategy to address these problems is bone tissue engineering,
where a combination of a supporting scaffold, cells and stimuli is used to
regenerate bone. [4, 8, 9]
There are many different types of scaffold materials, but biomimetic
scaffolds have attracted great attention from researchers. [10, 11] Natural bone is
a composite of hydroxyapatite (HA) and collagen fibrils. HA constitutes 65 wt% of
natural bone, exhibiting biocompatibility, osteoconductivity, and bone-bonding ability.
[3, 12] Collagen, the most abundant protein in bone, has good resorbability and high
affinity to cells. [13, 14] Inspired by the composition of natural bone, scaffolds for
bone repair consisting of collagen and apatite have attracted extensive attention of
researchers in recent years. [10, 11]
Recently, great efforts have been made to produce collagen-apatite composites
resembling the composition and structure of natural bone from the nano- to the
macro-scale. The conventional strategies to produce bone substitutes include coating a
collagen scaffold with biomimetic apatite, physically mixing apatite nanocrystals with
collagen fibrils, and biomimetic co-precipitation of collagen and apatite in which
128
apatite is deposited on the surface of collagen fibrils. [3, 12, 15-17] Collagen/apatite
scaffolds with improved mechanical properties were achieved by either coating
collagen scaffolds with HA or physically mixing pre-synthesized apatite with type I
collagen. [3, 18] Recently, our research group employed a self-assembly approach to
form extrafibrillar mineralized collagen fibrils in a collagen-containing modified
simulated body fluid (m-SBF). [13, 16] The materials characterization confirmed that
the scaffold has a composition closely mimicking that of natural bone, and our in vitro
cell culture and in vivo animal tests suggested that this structure has excellent
osteoconductivity and is suitable for bone repair. [1, 15, 16, 19] Although these
biomimetic collagen/apatite scaffolds successfully mimic the composition of natural
bone, they do not recapitulate the structure of bone at all levels, especially at the
nanostructural level. [20]
In natural bone, HA nanocrystals deposit within the gap zone of collagen fibrils
(intrafibrillar mineralization) at the early stage of mineralization and then on the
surface of collagen fibrils (extrafibrillar mineralization) at a later stage. [10, 21-25]
Bundles and arrays of mineralized collagen fibrils further arrange into a multilayered
structure rotating across concentric lamellae, which pack densely into compact bone
or loosely into spongy bone. [26-28] Both intrafibrillar and extrafibrillar
mineralization are essential for the functions of natural bone. Intrafibrillar and
extrafibrillar mineralization of collagen fibrils have been reported to contribute to the
high mineral content in bone and the improvement of mechanical and biological
properties in collagen matrix. [29-33] Thus, biomimetic mineralization must
129
recapitulate both intrafibrillar and extrafibrillar mineralization of collagen fibrils to
mimic natural bone.
Extensive research has been conducted on the mechanism of intrafibrillar
mineralization, and it has been proposed that non-collagenous proteins (NCPs) play
an important role in the mineralization process. Chen et al. investigated the effect of
NCPs (osteocalcin (OC) and bone sialoprotein (BSP)) on intrafibrillar and
extrafibrillar mineralization of collagen fibers. [21] Both BSP and OC are present on
the surface of collagen fibrils, but only OC is present within the intrafibrillar space of
collagen fibrils. [21] BSP is too large to be accommodated by the gap zone of
collagen fibrils due to the size exclusion rule for NCPs - molecules larger than a 40
kDa protein are completely excluded from collagen fibrils while those smaller than 6
kDa can freely diffuse into all spaces within collagen fibrils. [34] Moreover, OC
would potentially be located at the boundary between the gap and the overlap region
of collagen fibers, further confirming the observations by Nudelman et al..[25]
Nudelman et al. proposed that the net charges of each band play crucial roles in the
deposition and maturation of HA. Rodriguez et al. further proved that osteopontin, the
most abundant NCP in bone, can induce intrafibrillar mineralization of collagen and
modulate osteoclastic activity in vitro. [35]
With a better understanding of the mechanism for intrafibrillar mineralization of
collagen, biomimetic collagen/apatite scaffolds mimicking the natural structure of
bone from nano- to macro-scale have been investigated. Collagen scaffolds, turkey
tendons, eggshell membranes and demineralized human dentines have been subjected
130
to either intrafibrillar or extrafibrillar mineralization via a polymer-induced
liquid-precursor (PILP) process. [36-42] In the PILP process, an electrolyte is used to
form a fluidic amorphous calcium phosphate (ACP) precursor. Polyanionic
electrolytes, such as poly (aspartic acid) (PAsP), poly (acrylic acid) (PAA) and poly
(vinyl phosphonic acid) (PVPA), are often used to mimic the functional groups of
NCPs due to the limited availability and high cost of NCPs. [43-45] Collagen tapes
were immersed in PAsP-stabilized ACP nanoprecursors, and hydroxyapatite initially
precipitated within collagen fibrils, and then a mineral coating formed on the surface
of the collagen tape. [43, 44] In a separate study, human demineralized dentine was
incubated in ACP precursors, which were stabilized by poly(amidoamine) (PAMAM),
another candidate to mimic NCPs. [40] Intrafibrillar HA mineralization was achieved,
which mimicked the arrangement of natural minerals in human dentine at the
nanometer scale. Thus, a bioinspired intrafibrillar mineralization process can be
achieved via employing polyanionic electrolytes, which are analogs of NCPs. [36-40]
However, a two-step process is normally required to produce a collagen/HA
scaffold with intrafibrillar mineralization. Typically, a collagen scaffold is produced,
and then incubated in a precursor solution to induce mineral infiltration into collagen
fibrils. There are no reports on a bottom-up approach to fabricate scaffolds with a
bone-like composition and hierarchical structure. In the bottom-up approach,
materials assemble from the nanoscale, such as molecules, ions and atoms, to form a
macroscopic structure. [46] In this study, we employed a bottom-up approach to
produce intrafibrillar and extrafibrillar mineralized collagen-apatite scaffolds from the
131
collagen molecules and ACP nanoprecursors instead of the aforementioned two-step
process. The structure and biological performance of the intrafibrillar and
extrafibrillar mineralized collagen-apatite scaffolds were compared with that of
extrafibrillar mineralized scaffolds produced via our previous co-precipitation
method.
6.2 Materials and methods
6.2.1 Materials
Type I collagen was extracted from rat tails, purified based on the protocol
reported by Rajan et al.. [47] Polyacrylic acid (PAA, Mw 5000 Da) and sodium
tripolyphosphate (TPP) were purchased from Sigma-Aldrich. All chemicals were of
analytical chemical grade.
6.2.2 Fabrication of extrafibrillar mineralized collagen-apatite (E-Col-Ap) scaffolds
Based on our previously established protocol, E-Col-Ap scaffolds were produced
by the following procedure: A 6X simulated body fluid, called modified simulated
body fluid (m-SBF) was prepared as reported previously: [48-50] NaCl,
K2HPO4·3H2O, MgCl2·6H2O, CaCl2 were added in deionized water (DIW) in the
order as they are listed, and the ion concentrations were adjusted according to Table
6.1. A collagen stock solution (4.5 mg/mL) was diluted to 2.2 mg/mL by the addition
of m-SBF. The pH of the collagen containing m-SBF solution was adjusted to 7.2 by
addition of HEPES (4-(2-hydroxyapatiteethyl)-1-piperazineethanesulfonic acid),
NaHCO3 and NaOH. A two-temperature process was used to form a mineralized
132
collagen hydrogel: The solution was incubated at 25 °C for 1 h and then transferred to
a water bath at 37 °C for 24 hr. Subsequently, the collagen-apatite hydrogel was
allowed to undergo unconfined self-gravity driven compression (self-compression) at
room temperature for 20 min. The resulting self-compressed gel was transferred to a
custom-made mold and subjected to unidirectional freezing in a freeze-drier (Free
Zone®, Labconco, USA) precooled to -25 °C. The freeze-dried scaffolds were
subsequently cross-linked with 1 wt% N-(3-dimethylaminopropyl)-N’
-ethylcarbodiimide hydrochloride (EDC) for 24 h. Cross-linked scaffolds were then
rinsed thoroughly in DIW, followed by rinsing with a 5% glycine solution to react
with residual active carboxylic acid groups. [51, 52] The scaffolds were then rinsed
again with DIW, and freeze-dried for a second time to completely remove all the
moisture within the scaffold while maintaining the structure of the scaffold.
6.2.3 Fabrication of intrafibrillar and extrafibrillar mineralized collagen apatite (IE-Col-Ap)
scaffolds
Based on our previously established protocol for the biomimetic intrafibrillar
mineralized collagen fibrils, [16, 53] IE-Col-Ap scaffolds with a lamellar structure
were fabricated via the following procedure: PAA, the sequestration analog, was
added to a m-SBF solution to form stabilized ACP nanoprecursors (PAA-ACP). The
m-SBF was prepared as described above. [48-50] The concentration of PAA in
PAA-ACP nanoprecursor was 800 ug/mL. Type I collagen stock solution (4.5 mg/mL)
was diluted to a given concentration (2.2 mg/mL) using PAA-ACP nanoprecursors.
TPP (1.2 wt%), the templating analog, was added to the collagen stock solution to
133
template the orderly arrangement of hydroxyapatite within collagen fibrils. The pH of
the solution was then adjusted to 7.2 by addition of HEPES, NaHCO3 and NaOH. All
of the reactions were taken place at 4 °C, and then transferred to a water bath at 37 °C
for 24 h. The self-compression process, unidirectional freeze-drying and cross-linking
process were conducted following the same procedure as described above.
6.2.4 Materials characterization
A small piece of collagen-apatite hydrogel (IE-Col-Ap and E-Col-Ap hydrogel)
prepared by the two-temperature process was placed on a copper grid and observed
using transmission electron microscopy (TEM) at an accelerating voltage of 80 kV
and selected area electron diffraction (SAED) coupled to TEM (TEM, JEOL
JEM-2010). [54] The cross-linked dry scaffold (IE-Col-Ap and E-Col-Ap scaffold)
was grinded into tiny pieces in liquid nitrogen, dispersed in ethanol, deposited onto a
copper grid, dried in air, and then observed using both TEM and SAED coupled to the
TEM. The surface morphology of the freeze-dried collagen-apatite scaffold was
imaged using field emission scanning electron microscopy (FESEM, JEOL
JSM-6335F, Japan).
Freeze dried IE-Col-Ap and E-Col-Ap scaffolds were characterized using
attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR). A
Nicolet Magna 560 FT-IR spectrophotometer (Artisan Technology Group ®, Illinois,
USA) with an ATR setup was used to collect infrared spectra in the range of
4,000-400 cm-1 at a 4 cm-1 resolution, 32 scans. X-ray diffraction (BRUKER AXS D2
Phaser) was also performed on the mineralized collagen scaffolds from 10° to 50° at a
134
step size of 0.02° and a scan rate of 0.5 °/min with CuKα radiation.
Thermogravimetric analysis (TGAQ-500, TA Instrument, USA) was performed from
30 to 900 °C with a heating rate of 10 °C/min in air to determine the mineral content
in the collagen composite scaffolds.
6.2.5 In vitro cell culture
The ability of E-Col-Ap, and IE-Col-Ap scaffolds to support cell proliferation
was monitored through the reduction of a 3-(4,5-dimethylthiazol
-2-yl)-2,5-diphenyltetrazolium (MTT) dye. A pre-osteoblastic cell line, mouse
calvaria 3T3-E1 (MC3T3-E1), was used for the study. Cells were grown in alpha
modified eagles medium (α-MEM) supplemented with 10% fetal bovine serum and 1%
penicillin-streptomycin at 37°C under an atmosphere of 5% CO2. The medium was
changed every other day until the culture reached 90% confluence, at which point
cells were harvested and seeded onto pure collagen, E-Col-Ap, and IE-Col-Ap
scaffolds at 7,500 cells per scaffold (n ≥ 4). Scaffolds were cylindrical with a diameter
of 3.5 mm and a height of 1 mm.
For the proliferation study the medium was replaced every other day.
Proliferation was assessed by replacing the medium at 1, 3, and 7 days with a working
solution of α-MEM supplemented with 10% MTT solution (5 mg/ml). The treated
scaffolds were incubated for 4 hours at 37 °C. The medium was removed and
dimethyl sulfoxide was added to dissolve the formed formazan crystals. Aliquots of
150 μl were measured using a microplate reader (Biotek) at a wavelength of 560 nm.
All data are represented as the mean ± standard error. Statistical analysis was
135
conducted using a student t-test with two tails, and differences were considered
significant if p < 0.05.
6.3 Results and discussions
Figure 6.1 shows the TEM images of mineralized collagen fibrils in the
IE-Col-Ap and E-Col-Ap hydrogels. At the initial stage of IE-Col-Ap scaffold
preparation, a hydrogel consisting of intrafibrillar and extrafibrillar mineralized
collagen fibrils was formed using a two-temperature process (Figure 6.1A). In Figure
6.1A, a clear D-banding pattern is observed in the unstained collagen fibrils,
indicating electron-dense minerals have deposited within specific sites of collagen
fibrils, reproducing the D-banding pattern of naturally mineralized collagen fibrils
normally observed in bone. [22] The SAED result of intrafibrillar minerals does not
demonstrate a ring pattern, but only a diffraction spot (Figure 6.1A, inset), indicating
that the minerals deposited are nanocrystalline. This specific alignment of minerals
within collagen fibrils was regulated by TPP, a templating analog of NCPs. [37]
Highly negatively charged TPP accumulates within positively charged gap zones of
collagen fibrils through electrostatic attraction, which further attracts fluidic
PAA-ACP nanoprecursors to enter the gap zone of collagen fibrils via electrostatic
attraction. [25, 44, 53, 55] Thus, TPP functions as an ionic bridge between collagen
fibrils and PAA-ACP nanoprecursors due to the high affinity of TPP to collagen and
calcium. [53] A large fraction of the charged functional groups in the gap zone recruit
and bind with calcium and phosphate ions from the PAA-ACP nanoprecursors to
precipitate electron-dense minerals within the gap zone, which results in the observed
136
D-banding of collagen fibrils. [42, 55, 56] In extrafibrillar mineralized collagen fibrils,
D-banding is not observed. Instead, minerals are only formed on the surface of
collagen fibrils (Figure 6.1B). The SAED result of extrafibrillar mineralized collagen
fibrils exhibits a ring pattern of typical low crystalline hydroxyapatite (Figure 6.1B,
inset).
Figure 6.1. TEM images of a small piece of IE-Col-Ap (A) and E-Col-Ap (B)
hydrogel fabricated using the two-temperature process. Inset images in A and B:
SAED results. The mineralized collagen fibrils were unstained.
Then IE-Col-Ap and E-Col-Ap hydrogels were self-compressed and frozen
radially to form the hierarchical scaffold as shown in Figure 6.2. Both IE-Col-Ap and
E-Col-Ap scaffolds exhibit a multi-lamellar structure from the macrostructure to the
microstructure level. At the macrostructure level, unidirectional macro-pores are
aligned across the entire scaffold (Figures 6.2A and 6.2B). At the microstructure level,
each layer is comprised of aligned and well stacked lamellae with a thickness of
137
several hundred nanometers (Figures 6.2A1 and 6.2B1). In each lamella, mineralized
collagen fibers are partially aligned along the freezing direction of the scaffold, as
indicated by double headed arrows shown in Figures 6.2A2 and 6.2B2. In the
extrafibrillar mineralized scaffold, minerals are observed to be randomly dispersed on
the surface of collagen fibrils (Figure 6.2B2, black circle). The shear stress during the
gravity self-compression and unidirectional freezing processes leads to the alignment
of macro-pores, a multi-lamellar structure, and partially aligned mineralized collagen
fibrils of the Col-HA scaffold (Figure 6.2). [16]
138
Figure 6.2. SEM images of IE-Col-Ap (A, A1, A2) and E-Col-Ap (B, B1, B2)
scaffolds. Double headed arrows in A2 and B2 indicate the freezing direction of
self-compressed hydrogels. Black circle in B2 shows the randomly dispersed minerals
on the surface.
Freeze-dried IE-Col-Ap and E-Col-Ap scaffolds were grinded into tiny pieces,
and observed using TEM (Figure 6.3). At the nanostructure level of the IE-Col-Ap
139
scaffold, minerals are arranged within collagen fibrils and a clear D-banding pattern is
observed (Figure 6.3A). Minerals are also observed on the surface of the collagen
fibrils, as indicated by white arrow in Figures 6.3A. As demonstrated by SAED (inset
of Figure 6.3A), the typical ring pattern for HA is observed, indicating that the
minerals formed within collagen fibrils are HA. An arc-shape diffraction pattern of
the (002) growth face is observed in Figure 6.3A, implying that the mineral phase is
oriented along the long axis direction of collagen fibrils. [22, 57] In contrast, the
minerals within the E-Col-Ap scaffold are randomly dispersed on the surface of
collagen fibrils (Figures 6.2B2 and 6.3B). The ring pattern observed from SAED is
consistent with the typical ring pattern for HA which implies that the minerals formed
are HA (inset of Figure 6.3B).
Figure 6.3. Freeze-dried IE-Col-Ap (A) and E-Col-Ap (B) scaffolds were ground into
small pieces in liquid nitrogen, and one of these small pieces were observed using
TEM. Inset images in A and B: SAED results. White arrows in A and B indicate the
extrafibrillar minerals.
140
Figure 6.4 shows the XRD spectra of IE-Col-Ap and E-Col-Ap scaffolds.
IE-Col-Ap and E-Col-Ap scaffolds exhibit similar XRD spectra (Figures 6.4A and
6.4B). The broad peak at about 20° (2θ) is attributed to collagen. The peak at 26.2°
(2θ) corresponds to the (002) plane of HA. The peak at 29.4° is assigned to the (210)
plane. The broad peak from 30.7-35° is an overlap of three major reflections (112),
(211) and (300). The peak at about 39.6° corresponds to (310) plane of HA. The broad
peaks of 30.7-35° indicate that the apatite obtained is at the nanosized level and
poorly crystalline. [16]
Figure 6.4. XRD spectra of IE-Col-Ap (A) and E-Col-Ap (B) scaffolds.
Figure 6.5 shows the FT-IR spectra obtained from IE-Col-Ap and E-Col-Ap
scaffolds. Both IE-Col-Ap and E-Col-Ap scaffolds show similar spectra. The broad
peak around 3400 cm-1 is associated with the absorption of water in the scaffolds.
Both spectra show typical peaks of collagen: the peak around 1660 cm-1 is attributed
to the C=O stretching vibration (amide I), while those at 1549 cm-1 and 1188 cm-1
arise from angular deformation of N-H bond (amide II) and vibration of C-N bond
(amide III). All of these peaks are typical major peaks for collagen. Peaks between
141
500-600 cm-1 are assigned to anti-symmetric bending vibration of the PO43- groups,
and peaks at 950 cm-1 and 1036 cm-1 are identified as the stretching vibration of PO43-
groups. Moreover, the peaks at 1410 cm-1 and 875 cm-1 are typical vibration bands of
carbonate apatite, which is similar to the mineral phase in natural bone. [58]
Figure 6.5. FT-IR spectra of IE-Col-Ap (A) and E-Col-Ap (B) scaffolds.
Figure 6.6 shows TGA measurements of E-Col-Ap and IE-Col-Ap scaffolds.
Both E-Col-Ap and IE-Col-Ap scaffolds have a mineral content of approximately 30
wt%. In the E-Col-Ap scaffolds, the first 7 % of weight loss occurs between
30-150 °C (peak at 71.2 °C), which corresponds to the evaporation of physisorbed
water on the scaffold. Decomposition of collagen molecules occurs between
200-650 °C (peaks at 308.9 °C and 544.5 °C), leaving a residual 30 wt% which are
the minerals obtained (Figure 6.6B). In the IE-Col-Ap scaffolds, the initial weight loss
(10 wt%) occurs between 30-150 °C (peak at 50 °C), and the decomposition of
collagen molecules occurs between 150-800 °C (peaks at 227.7 °C, 306.6 °C and
142
576.2 °C), which yields approximately 30.6 wt% of minerals (Figure 6.6A). From
Figure 6.6 the decomposition of collagen molecules in E-Col-Ap scaffolds started at a
higher temperature than that of IE-Col-Ap scaffolds. This may be due to the
extrafibrillar HA minerals existing on the surface of collagen fibrils, which in turn
protects the collagen fibrils from decomposition. However, a higher temperature was
needed to fully decompose the collagen in the IE-Col-Ap scaffolds, which may be due
to the association of intrafibrillar minerals with collagen fibrils.
Figure 6.6. TGA of IE-Col-Ap (A) and E-Col-Ap (B) scaffolds.
Compared to E-Col-Ap scaffolds fabricated in our previous study using a
co-precipitation method, similar composition and crystallinity of HA minerals were
obtained. However, oriented HA minerals were formed within collagen fibrils via this
bottom-up approach, which closely mimics the organization of mineralized collagen
fibrils in natural bone.
Substantial research has been performed on the mechanism of intrafibrillar
mineralization of collagen fibrils. [25, 43, 44, 56, 59] Two functional groups in NCPs
143
have been found to play vital roles in regulation of nucleation, dimension, and order
of HA deposition within natural bone. Acidic sequestrating motifs of NCPs bind to
calcium and inhibit the precipitation of calcium phosphate, forming stabilized-ACP
nanoprecursors. [60] The templating groups of NCPs regulate the deposition of HA
within gap zone of collagen fibrils through its excellent binding capacity for calcium
and its strong affinity for collagen fibrils. [37] Fluidic ACP nanoprecursors infiltrate
and are retained within the gap zone of collagen fibrils through electrostatic attraction,
capillary action and size exclusion. [61] Then the infiltrated ACP minerals mature into
HA within collagen fibrils.
However, it is still not clear why the polymer-induced liquid precursor droplets
are more likely to infiltrate collagen fibrils instead of precipitating on the surface of
collagen fibrils. If only size exclusion and electrostatic attraction played a role in
mineralization, then more minerals should have deposited on the surface of collagen
fibrils since there is more room on the surface for nucleation than within the collagen
fibers. The electrostatic attraction is presumably the same, since charged groups are
located on both the interior and exterior of collagen fibers. [35, 56] Thus, more
minerals should have deposited on the surface of collagen fibrils. However, this was
not observed in our study. Herein, we propose that there are other driving forces
playing crucial roles in directing the deposition and maturation of HA minerals within
collagen fibrils: the stereochemical match between the interior of collagen fibrils and
ACP precursors. On one hand, the charged functional groups of collagen molecules
are located and oriented spatially within collagen fibrils to increase the likelihood of
144
calcium and phosphate ions binding to the fibrils, leading to the nucleation of ACP
within collagen fibrils. This has been confirmed by Xu et al. by applying all-atom
Hamiltonian replica exchange molecular dynamics simulation. [56] On the other hand,
calcium ions induce conformation changes of many proteins. [21, 62] Bound calcium
ions to dentin matrix protein 1 (DMP 1) induce a conformation change resulting in
self-assembly into elongated dimers. [63] Osteocalcin also undergoes a conformation
change when calcium binds. The conformation change resembles the interatomic
locations of calcium in the HA lattice. [21, 64] Thus, conformation changes of
proteins occur after binding with calcium to form ACP nanoprecursors, which leads to
the re-arrangement of calcium and phosphate ions into a structure resembling HA.
This re-arrangement of calcium and phosphate ions decreases the activation energy
for the transformation of ACP into HA. Moreover, the conformation change helps
NCPs interact specifically and spatially with certain domains of collagen sequences,
which leads to the spatial deposition of minerals within collagen fibrils. [65] Thus, we
propose that HA minerals form within collagen fibrils due to two factors: (1) The
stereochemical match of NCPs with the interior functional groups of collagen fibrils,
and (2) the conformation change of NCPs decreases the activation energy of
nucleation within collagen fibrils and regulates HA orientation. [65, 66] Meanwhile,
the work by other research groups also proposed the stereochemical template function
of collagen fibrils, which is conducive to nucleation, growth and development of
apatite crystals within collagen fibrils. [67, 68] A similar mechanism has been
proposed that specific molecular recognition between DMP1 and apatite surface aids
145
in the controlled growth of HA crystals with oriented crystallography. [63] Thus,
PAA-ACP nanoprecursors penetrate the gap zone of collagen fibrils via capillary
action and electrostatic attraction. Subsequently, both spatial interactions between
collagen molecules and ACP nanoprecursors along with preferential association of
NCPs on certain HA crystallographic planes play important roles in the orderly
alignment of HA within collagen fibrils. [65, 69]
Figure 6.7. Proliferation of MC3T3-E1 cells on collagen, E-Col-Ap, and IE-Col-Ap
scaffolds over the course of 7 days. Data is represented as the mean ± standard error.
# denotes statistical significance (p < 0.05) between collagen scaffolds and
experimental scaffolds.
The biocompatibility of the scaffolds was assessed through the proliferation of
MC3T3-E1 cells on E-Col-Ap and IE-Col-Ap scaffolds compared to cell proliferation
on pure collagen scaffolds which were used as a control, and the results are depicted
in Figure 6.7. All scaffolds supported cell growth over the course of 7 days. Cell
146
proliferation was the greatest on collagen scaffolds and statistically significant from
mineralized scaffolds. The difference in mineralization strategies did not result in a
statistically significant difference in proliferation between E-Col-Ap and IE-Col-Ap.
Despite the lower rate of cell proliferation on E-Col-Ap and IE-Col-Ap scaffolds
compared to collagen scaffolds, the mineralized scaffolds are a suitable material for
bone tissue engineering. The in vitro performance of a material is not always
indicative of its in vivo performance. Calcium phosphate materials are well known for
their excellent biocompatibility and in vivo performance, but these materials have
been shown to have inhibitory effects on osteoblasts in vitro. [70-73] Lyons et al.
observed inhibited cell proliferation on a biomimetic collagen-calcium phosphate
scaffold compared to a collagen-glycosaminoglycan scaffold in vitro but observed a
better response in vivo with the collagen-calcium phosphate scaffold. [73] Minardi et
al. also observed inhibited cell proliferation of hBM-MSC on a collagen HA
composite in vitro compared to a pure collagen scaffold. [71] The presence of the
biomimetic apatite phase was found to support the expression of osteogenic genes
better than the pure collagen scaffold. [74] E-Col-Ap scaffolds fabricated with the
described technique have previously been shown to have excellent in vivo
performance. [15] Future studies to compare the in vivo performance of IE-Col-Ap to
E-Col-Ap scaffolds will be conducted to determine the role of intrafibrillar
mineralization on the scaffolds bone forming ability.
6.4 Conclusions
A biomimetic Col-HA scaffold mimicking the composition and structure of
147
natural bone has been successfully prepared via a bottom-up approach. Compared
with E-Col-Ap scaffolds prepared using the co-precipitation approach in our previous
study, IE-Col-Ap scaffolds showed similar multi-lamellar structure and mineral
composition, which were confirmed by FESEM, XRD and FT-IR examinations.
However, the TEM and SAED results indicated that HA minerals in IE-Col-Ap
scaffolds were deposited within collagen fibrils and oriented along the long axis of
these fibrils, which bears higher resemblance to natural bone than the scaffold
fabricated using the co-precipitation approach. The biomimetic Col-HA scaffolds
demonstrated good biocompatibility in vitro with comparable cellular proliferation for
both E-Col-Ap and IE-Col-Ap scaffolds. This biomimetic IE-Col-AP scaffold can be
a promising biomaterial for bone tissue engineering applications.
148
References:
[1] Z. Xia, M. Wei. Biomimetic fabrication of collagen-apatite scaffolds for bone
tissue regeneration. Journal of Biomaterials and Tissue Engineering. 2013; 3:369-384.
[2] L. Meinel, V. Karageorgiou, R. Fajardo, B. Snyder, V. Shinde-Patil, L. Zichner, et
al. Bone tissue engineering using human mesenchymal stem cells: effects of scaffold
material and medium flow. Annal of Biomedical Engineering. 2004; 32:112-122.
[3] A. A. Al‐Munajjed, N. A. Plunkett, J. P. Gleeson, T. Weber, C. Jungreuthmayer, T.
Levingstone, et al. Development of a biomimetic collagen‐hydroxyapatite scaffold for
bone tissue engineering using a SBF immersion technique. Journal of Biomedical
Materials Research Part B: Applied Biomaterials. 2009; 90:584-591.
[4] J. R. Porter, T. T. Ruckh, K. C. Popat. Bone tissue engineering: a review in bone
biomimetics and drug delivery strategies. Biotechnology Progress. 2009;
25:1539-1560.
[5] C. Szpalski, M. Barbaro, F. Sagebin, S. M. Warren. Bone tissue engineering:
current strategies and techniques—part II: cell types. Tissue Engineering Part B:
Reviews. 2012; 18:258-269.
[6] J. O. Smith, A. Aarvold, E. R. Tayton, D. G. Dunlop, R. O. Oreffo. Skeletal tissue
regeneration: current approaches, challenges, and novel reconstructive strategies for
an aging population. Tissue Engineering Part B: Reviews. 2011; 17:307-320.
[7] S. Liao, F. Cui, W. Zhang, Q. Feng. Hierarchically biomimetic bone scaffold
materials: nano‐HA/collagen/PLA composite. Journal of Biomedical Materials
Research Part B: Applied Biomaterials. 2004; 69:158-165.
149
[8] L. H. Nguyen, N. Annabi, M. Nikkhah, H. Bae, L. Binan, S. Park, et al.
Vascularized bone tissue engineering: approaches for potential improvement. Tissue
Engineering Part B: Reviews. 2012; 18:363-382.
[9] I. Marcos-Campos, D. Marolt, P. Petridis, S. Bhumiratana, D. Schmidt, G.
Vunjak-Novakovic. Bone scaffold architecture modulates the development of
mineralized bone matrix by human embryonic stem cells. Biomaterials. 2012;
33:8329-8342.
[10] X. Su, K. Sun, F. Cui, W. Landis. Organization of apatite crystals in human
woven bone. Bone. 2003; 32:150-162.
[11] T. Dvir, B. P. Timko, D. S. Kohane, R. Langer. Nanotechnological strategies for
engineering complex tissues. Nature nanotechnology. 2011; 6:13-22.
[12] S. Yunoki, H. Sugiura, T. Ikoma, E. Kondo, K. Yasuda, J. Tanaka. Effects of
increased collagen-matrix density on the mechanical properties and in vivo
absorbability of hydroxyapatite–collagen composites as artificial bone materials.
Biomedical Materials. 2011; 6:015012.
[13] Z. Xia, X. Yu, X. Jiang, H. D. Brody, D. W. Rowe, M. Wei. Fabrication and
characterization of biomimetic collagen–apatite scaffolds with tunable structures for
bone tissue engineering. Acta Biomaterialia. 2013; 9:7308-7319.
[14] J. M. Holzwarth, P. X. Ma. Biomimetic nanofibrous scaffolds for bone tissue
engineering. Biomaterials. 2011; 32:9622-9629.
[15] X. Yu, Z. Xia, L. Wang, F. Peng, X. Jiang, J. Huang, et al. Controlling the
structural organization of regenerated bone by tailoring tissue engineering scaffold
150
architecture. Journal of Materials Chemistry. 2012; 22:9721-9730.
[16] Z. Xia, M. Villa, M. Wei. A biomimetic collagen–apatite scaffold with a
multi-level lamellar structure for bone tissue engineering. Journal of Materials
Chemistry B. 2014; 2:1998-2007.
[17] H. S. Yang, W. G. La, S. H. Bhang, T. J. Lee, M. Lee, B. S. Kim. Apatite-coated
collagen scaffold for bone morphogenetic protein-2 delivery. Tissue Engineering Part
A. 2011; 17:2153-2164.
[18] Y. Pek, S. Gao, M. M. Arshad, K. J. Leck, J. Y. Ying. Porous collagen-apatite
nanocomposite foams as bone regeneration scaffolds. Biomaterials. 2008;
29:4300-4305.
[19] M. M. Villa, L. Wang, D. W. Rowe, M. Wei. Effects of Cell-Attachment and
Extracellular Matrix on Bone Formation In Vivo in Collagen-Hydroxyapatite
Scaffolds. PLoS One. 2014; 9:e109568.
[20] S. Weiner, H. D. Wagner. The material bone: structure-mechanical function
relations. Annual Review of Materials Science. 1998; 28:271-298.
[21] L. Chen, R. Jacquet, E. Lowder, W. J. Landis. Refinement of collagen–mineral
interaction: A possible role for osteocalcin in apatite crystal nucleation, growth and
development. Bone. 2015; 71:7-16.
[22] Y. Wang, T. Azaïs, M. Robin, A. Vallée, C. Catania, P. Legriel, et al. The
predominant role of collagen in the nucleation, growth, structure and orientation of
bone apatite. Nature materials. 2012; 11:724-733.
[23] X. Wang, F. Cui, J. Ge, Y. Wang. Hierarchical structural comparisons of bones
151
from wild-type and liliput dtc232 gene-mutated Zebrafish. Journal of Structural
Biology. 2004; 145:236-245.
[24] F. Nudelman, P. H. Bomans, A. George, N. A. Sommerdijk. The role of the
amorphous phase on the biomimetic mineralization of collagen. Faraday Discuss.
2012; 159:357-370.
[25] F. Nudelman, K. Pieterse, A. George, P. H. Bomans, H. Friedrich, L. J. Brylka, et
al. The role of collagen in bone apatite formation in the presence of hydroxyapatite
nucleation inhibitors. Nature Materials. 2010; 9:1004-1009.
[26] W. Landis, M. Song, A. Leith, L. McEwen, B. McEwen. Mineral and organic
matrix interaction in normally calcifying tendon visualized in three dimensions by
high-voltage electron microscopic tomography and graphic image reconstruction.
Journal of Structural Biology. 1993; 110:39-54.
[27] T. T. Thula, D. E. Rodriguez, M. H. Lee, L. Pendi, J. Podschun, L. B. Gower. In
vitro mineralization of dense collagen substrates: a biomimetic approach toward the
development of bone-graft materials. Acta Biomaterialia. 2011; 7:3158-3169.
[28] W. Wagermaier, H. Gupta, A. Gourrier, M. Burghammer, P. Roschger, P. Fratzl.
Spiral twisting of fiber orientation inside bone lamellae. Biointerphases. 2006; 1:1-5.
[29] I. Jäger, P. Fratzl. Mineralized collagen fibrils: a mechanical model with a
staggered arrangement of mineral particles. Biophysical Journal. 2000; 79:1737-1746.
[30] M. J. Olszta, X. Cheng, S. S. Jee, R. Kumar, Y. Y. Kim, M. J. Kaufman, et al.
Bone structure and formation: a new perspective. Materials Science and Engineering:
R: Reports. 2007; 58:77-116.
152
[31] A. K. Nair, A. Gautieri, M. J. Buehler. Role of intrafibrillar collagen
mineralization in defining the compressive properties of nascent bone.
Biomacromolecules. 2014; 15:2494-2500.
[32] Y. Liu, D. Luo, X. X. Kou, X. D. Wang, F. R. Tay, Y. L. Sha, et al. Hierarchical
intrafibrillar nanocarbonated apatite assembly improves the nanomechanics and
cytocompatibility of mineralized collagen. Advanced Functional Materials. 2013;
23:1404-1411.
[33] R. B. Martin, D. B. Burr, N. A. Sharkey. Skeletal tissue mechanics. 1998.
[34] D. Toroian, J. E. Lim, P. A. Price. The Size Exclusion Characteristics of Type I
Collagen Implications for the Role of Noncollagenous Bone Constitutes in
Mineralization. Journal of Biological Chemistry. 2007; 282:22437-22447.
[35] D. E. Rodriguez, T. Thula-Mata, E. J. Toro, Y. W. Yeh, C. Holt, L. S. Holliday, et
al. Multifunctional role of osteopontin in directing intrafibrillar mineralization of
collagen and activation of osteoclasts. Acta Biomaterialia. 2014; 10:494-507.
[36] S. S. Jee, R. K. Kasinath, E. DiMasi, Y. Y. Kim, L. Gower. Oriented
hydroxyapatite in turkey tendon mineralized via the polymer-induced liquid-precursor
(PILP) process. CrystEngComm. 2011; 13:2077-2083.
[37] Y. K. Kim, L. Gu, T. E. Bryan, J. R. Kim, L. Chen, Y. Liu, et al. Mineralisation of
reconstituted collagen using polyvinylphosphonic acid/polyacrylic acid templating
matrix protein analogues in the presence of calcium, phosphate and hydroxyl ions.
Biomaterials. 2010; 31:6618-6627.
[38] J. Mahamid, B. Aichmayer, E. Shimoni, R. Ziblat, C. Li, S. Siegel, et al.
153
Mapping amorphous calcium phosphate transformation into crystalline mineral from
the cell to the bone in zebrafish fin rays. Proceedings of the National Academy of
Sciences. 2010; 107:6316-6321.
[39] B. Antebi, X. Cheng, J. N. Harris, L. B. Gower, X. D. Chen, J. Ling. Biomimetic
collagen–hydroxyapatite composite fabricated via a novel perfusion-flow
mineralization technique. Tissue Engineering Part C: Methods. 2013; 19:487-496.
[40] J. Li, J. Yang, J. Li, L. Chen, K. Liang, W. Wu, et al. Bioinspired intrafibrillar
mineralization of human dentine by PAMAM dendrimer. Biomaterials. 2013;
34:6738-6747.
[41] B. Marelli, C. E. Ghezzi, A. Alessandrino, J. E. Barralet, G. Freddi, S. N. Nazhat.
Silk fibroin derived polypeptide-induced biomineralization of collagen. Biomaterials.
2012; 33:102-108.
[42] N. Li, L. Niu, Y. Qi, C. K. Yiu, H. Ryou, D. D. Arola, et al. Subtleties of
biomineralisation revealed by manipulation of the eggshell membrane. Biomaterials.
2011; 32:8743-8752.
[43] S. S. Jee, T. T. Thula, L. B. Gower. Development of bone-like composites via the
polymer-induced liquid-precursor (PILP) process. Part 1: Influence of polymer
molecular weight. Acta Biomaterialia. 2010; 6:3676-3686.
[44] Y. Liu, Y. K. Kim, L. Dai, N. Li, S. O. Khan, D. H. Pashley, et al. Hierarchical
and non-hierarchical mineralisation of collagen. Biomaterials. 2011; 32:1291-1300.
[45] R. Q. Song, H. Colfen, A. W. Xu, J. Hartmann, M. Antonietti.
Polyelectrolyte-directed nanoparticle aggregation: systematic morphogenesis of
154
calcium carbonate by nonclassical crystallization. Acs Nano. 2009; 3:1966-1978.
[46] Y. Liu, S. Mai, N. Li, C. K. Yiu, J. Mao, D. H. Pashley, et al. Differences between
top-down and bottom-up approaches in mineralizing thick, partially demineralized
collagen scaffolds. Acta biomaterialia. 2011; 7:1742-1751.
[47] N. Rajan, J. Habermehl, M. F. Coté, C. J. Doillon, D. Mantovani. Preparation of
ready-to-use, storable and reconstituted type I collagen from rat tail tendon for tissue
engineering applications. Nature Protocol. 2006; 1:2753-2758.
[48] Z. Xia, X. Yu. Biomimetic collagen/apatite coating formation on Ti6Al4V
substrates. Journal of Biomedical Materials Research Part B: Applied Biomaterials.
2012; 100:871-881.
[49] H. Qu, Z. Xia, D. A. Knecht, M. Wei. Synthesis of dense collagen/apatite
composites using a biomimetic method. Journal of American Ceramic Society. 2008;
91:3211-3215.
[50] X. Yu, L. Wang, F. Peng, X. Jiang, Z. Xia, J. Huang, et al. The effect of fresh
bone marrow cells on reconstruction of mouse calvarial defect combined with
calvarial osteoprogenitor cells and collagen–apatite scaffold. Journal of Tissue
Engineering and Regenerative Medicine. 2013; 7:974-983.
[51] M. Gelinsky, P. Welzel, P. Simon, A. Bernhardt, U. König. Porous
three-dimensional scaffolds made of mineralised collagen: preparation and properties
of a biomimetic nanocomposite material for tissue engineering of bone. Chemical
Engineering Journal. 2008; 137:84-96.
[52] X. Duan, C. McLaughlin, M. Griffith, H. Sheardown. Biofunctionalization of
155
collagen for improved biological response: scaffolds for corneal tissue engineering.
Biomaterials. 2007; 28:78-88.
[53] L. Zhang. C. Hu, Mei Wei. Development of biomimetic scaffold woth both
intrafibrilar and extrafibrillar mineralization. ACS Biomaterials Science &
Engineering. 2015; 1:669-676.
[54] C. Hu, W. Cui. Microscopy of Biopolymer Systems. Handbook of
Biopolymer-Based Materials: From Blends and Composites to Gels and Complex
Networks. 2013; 611-644.
[55] Q. Wang, X. Wang, L. Tian, Z. Cheng, F. Cui. In situ remineralizaiton of partially
demineralized human dentine mediated by a biomimetic non-collagen peptide. Soft
Matter. 2011; 7:9673-9680.
[56] Z. Xu, Y. Yang, W. Zhao, Z. Wang, W. J. Landis, Q. Cui, et al. Molecular
mechanisms for intrafibrillar collagen mineralization in skeletal tissues. Biomaterials.
2015; 39:59-66.
[57] N. Nassif, F. Gobeaux, J. Seto, E. Belamie, P. Davidson, P. Panine, et al.
Self-assembled collagen− apatite matrix with bone-like hierarchy. Chemistry of
Materials. 2010; 22:3307-3309.
[58] G. Piga, A. Santos-Cubedo, A. Brunetti, M. Piccinini, A. Malgosa, E. Napolitano,
et al. A multi-technique approach by XRD, XRF, FT-IR to characterize the diagenesis
of dinosaur bones from Spain. Palaeogeography, Palaeoclimatology, Palaeoecology.
2011; 310:92-107.
[59] F. R. Tay, D. H. Pashley. Guided tissue remineralisation of partially
156
demineralised human dentine. Biomaterials. 2008; 29:1127-1137.
[60] H. Ozawa, K. Hoshi, N. Amizuka. Current concepts of bone biomineralization.
Journal of Oral Bioscience. 2008; 50:1-14.
[61] L. Gu, Y. K. Kim, Y. Liu, K. Takahashi, S. Arun, C. E. Wimmer, et al.
Immobilization of a phosphonated analog of matrix phosphoproteins within
cross-linked collagen as a templating mechanism for biomimetic mineralization. Acta
Biomaterialia. 2011; 7:268-277.
[62] S. R. Shi, R. J. Cote, D. Hawes, S. Thu, Y. Shi, L. L. Young, et al.
Calcium-induced modification of protein conformation demonstrated by
immunohistochemistry: What is the signal? J Histochem Cytochem. 1999;
47:463-469.
[63] G. He, S. Gajjeraman, D. Schultz, D. Cookson, C. Qin, W. T. Butler, et al.
Spatially and temporally controlled biomineralization is facilitated by interaction
between self-assembled dentin matrix protein 1 and calcium phosphate nuclei in
solution. Biochemistry. 2005; 44:16140-16148.
[64] P. V. Hauschka, S. A. Carr. Calcium-dependent. alpha.-helical structure in
osteocalcin. Biochemistry. 1982; 21:2538-2547.
[65] E. Beniash. Biominerals—hierarchical nanocomposites: the example of bone.
Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology. 2011;
3:47-69.
[66] G. He, T. Dahl, A. Veis, A. George. Nucleation of apatite crystals in vitro by
self-assembled dentin matrix protein 1. Nature Materials. 2003; 2:552-558.
157
[67] W. Stetler-Stevenson, A. Veis. Bovine dentin phosphophoryn: calcium ion
binding properties of a high molecular weight preparation. Calcified Tissue
International. 1987; 40:97-102.
[68] W. J. Landis, F. H. Silver. Mineral deposition in the extracellular matrices of
vertebrate tissues: identification of possible apatite nucleation sites on type I collagen.
Cells Tissues Organs. 2009; 189:20-24.
[69] S. S. Jee, L. Culver, Y. Li, E. P. Douglas, L. B. Gower. Biomimetic
mineralization of collagen via an enzyme-aided PILP process. Journal of Crystal
Growth. 2010; 312:1249-1256.
[70] Y. F. Chou, W. Huang, J. C. Dunn, T. A. Miller, B. M. Wu. The effect of
biomimetic apatite structure on osteoblast viability, proliferation, and gene expression.
Biomaterials. 2005; 26:285-295.
[71] J. B. Nebe, L. Müller, F. Lüthen, A. Ewald, C. Bergemann, E. Conforto, et al.
Osteoblast response to biomimetically altered titanium surfaces. Acta Biomaterialia.
2008; 4:1985-1995.
[72] R. O. Oreffo, F. C. Driessens, J. A. Planell, J. T. Triffitt. Growth and
differentiation of human bone marrow osteoprogenitors on novel calcium phosphate
cements. Biomaterials. 1998; 19:1845-1854.
[73] F. G. Lyons, A. A. Al-Munajjed, S. M. Kieran, M. E. Toner, C. M. Murphy, G. P.
Duffy, et al. The healing of bony defects by cell-free collagen-based scaffolds
compared to stem cell-seeded tissue engineered constructs. Biomaterials. 2010;3
1:9232-9243.
158
[74] S. Minardi, B. Corradetti, F. Taraballi, M. Sandri, J. Van Eps, F. Cabrera, et al.
Evaluation of the osteoinductive potential of a bio-inspired scaffold mimicking the
osteogenic niche, for bone augmentation. Biomaterials. 2015; 62:128-137.
159
7. Conclusions and future work
7.1 Conclusions
The objective of this study was to prepare and characterize biomimetic
collagen-apatite composite coatings and scaffolds for enhanced bone repair and
regeneration. In Chapter 2, a crack-free apatite coating was produced using a
biomimetic approach. Biomimetic apatite was deposited on the surface-treated
Ti-6Al-4V substrates by soaking them in the modified simulated body fluid (m-SBF)
for 24 hours. The apatite morphology could be adjusted by controlling the pH of the
m-SBF solution. Dual beam FIB/SEM was used to mill the cross-sections and prepare
TEM foils. Both cross-sectional SEM images and TEM observations confirmed that a
well-bonded HA coating has been formed on the Ti-6Al-4V substrate, and that the
biomimetic apatite coating possessed a unique gradient structure, with a dense
structure adjacent to the substrate to facilitate a strong bonding, and a porous structure
at the surface to allow bone cell attachment.
Furthermore, building on the success of preparation of biomimetic apatite
coatings, biomimetic collagen-apatite composite coatings were successfully fabricated
on the Ti-6Al-4V substrates to better mimic the composition of natural bone. The
coating was also formed by simply immersing the surface-treated Ti-6Al-4V substrate
into a collagen-containing m-SBF. The cross-sectional microstructure of the
collagen-apatite coating has been investigated using dual beam FIB/SEM. It has been
revealed that the collagen-apatite coating also consists of a gradient structure with a
160
dense a dense layer adjacent to the interface and a porous structure towards the
surface. Compared with apatite coating, collagen-apatite composite coating formed is
thinner and less porous than the apatite coating.
The cross-sectional microstructure of both the apatite coating and the
collagen-apatite composite coating has been investigated using dual beam FIB/SEM
with either gallium (GFIB) or xenon plasma (PFIB) ion source, and the capabilities of
these two dual beam FIB/SEM systems were also compared. Although GFIB and
PFIB share a lot of similarities, there are also significant differences, including the
optimal voltage of ion-beam-induced Pt deposition, the TEM sample preparation
process, and the efficiency and the quality of the cross-sectional cuts produced.
Particularly, the GFIB is more beneficial for preparing the high-quality final finish of
the sample, while the PFIB is more efficient for creating a large cut in the sample.
Inspired by the composition and structures of natural bone, intrafibrillar
calcification of collagen fibrils have been successfully produced using poly (acrylic
acid) as a sequestration analog and sodium tripolyphosphate as a templating analog.
First, the effect of the sequestration analog on the intrafibrillar mineralization has
been investigated. Based on the success of intrafibrillar calcification of collagen
fibrils, intrafibrillar silicification of collagen fibrils has been successfully produced by
a one-step collagen self-assembly/silicification approach, in which both the collagen
fibril self-assembly and the intrafibrillar silicification occurred simultaneously. As the
positive analog of the zwitterionic proteins in silicification, poly (allylamine
hydrochloride) promotes silica precipitation in the presence of a negative analog, such
161
as sodium tripolyphosphate. The structure of the silicified collagen fibrils can be
manipulated by varying silica precursors, polyanions and silica/collagen ratios to
produce silicified collagen fibrils with a core-shell, twisted, or banded structure.
Further, the mechanism of the collagen silicification has also been explored. The
intrafibrillar silicified collagen fibrils possess better cell compatibility compared with
the collagen-apatite fibrils.
A collagen-apatite scaffold consisting of intrafibrillar mineralized collagen fibrils
has been successfully prepared via a bottom-up approach, closely mimicking the
composition and structure of natural bone. Such prepared scaffolds demonstrate a
hierarchical structure at nano-, micro- and macro-levels, bearing striking resemblance
to natural bone. The results indicated that at the nanoscale level, the apatite minerals
deposited within the gap zone of collagen fibrils and oriented along the long axis of
these fibrils; At the microscale level, the aligned thin multi-lamella formed a thick
layer; At the macroscale level, the scaffolds consisted of interconnected macropores
throughout the entire scaffold. This biomimetic collagen-apatite scaffold with high
resemblance to natural bone can be a promising biomaterial for future bone tissue
engineering applications.
7.2 Future work
Great progresses have been made to produce intrafibrillar calcification and
silicification of collagen fibrils, but there are still many challenges remaining, for
examples: (1) the apatite mineral content in the intrafibrillar mineralized
collagen-apatite scaffolds is much lower than that in the natural bone; (2) the lack of
162
osteoinductivity of the scaffolds.
Silicon-containing biomaterials have been found to be osteoinductive for new
bone formation and stimulative for neovascularization. Intrafibrillar silification of
collagen fibrils have been successfully developed in our study. In order to impart the
osteoinductivity to the collagen-apatite scaffolds, a scaffold with both intrafibrillar
apatite and silica have been produced. Figure 7.1 shows the TEM images of
intrafibrillar mineralized collagen fibrils in the collagen-apatite, collagen-silica and
collagen-apatite-silica hydrogels. At the initial stage of scaffold preparation, a
hydrogel consisting of apatite and/or silica mineralized collagen fibrils was formed
using a two-temperature process (Figure 7.1). In Figure 7.1, clear D-banding patternd
are observed in the unstained collagen fibrils, indicating electron-dense minerals have
deposited within the gap zone of collagen fibrils, reproducing the D-banding pattern
of naturally mineralized collagen fibrils normally observed in bone. As discussed
above, sodium tripolyphosphate was applied as the templating analog to direct the
deposit of calcium phosphate (Figure 7.1A) and silica (Figure 7.1B) within gap zone
of collagen fibrils, and then composite minerals of calcium phosphate and silica were
deposited within the collagen fibrils (Figure 7.1C).
Further characterization needs to be conducted to prove the application of the
intrafibrillar collagen-apatite-silica composite scaffolds in bone tissue engineering. In
vitro cell culture studies are essential for initial assessing the biocompatibility of the
material, because they help predict how the material will interact with cells in vivo.
Cell culture studies will be conducted using MC3T3-E1 cells, an osteoblast precursor.
163
Intrafibrillar silica, apatite and apatite/silica mineralized collagen scaffolds will be
prepared to assess the viability and proliferation of cells, which will be monitored
through the use of MTT dye.
Figure 7.1. TEM images of a small piece of hydrogel composed of intrafibrillar
apatite (A), silica (B) and apatite/silica (C) mineralized collagen fibrils. The
mineralized collagen fibrils were unstained.
164
Appendice A: List of publications
Publications
1. Changmin Hu, Le Yu, Mei Wei*. Sectioning studies of biomimetic
collagen-hydroxyapatite coatings on Ti-6Al-4V substrates using focused ion beam.
Paper in preparation.
2. Changmin Hu, Le Yu, Mei Wei*. Intrafibrillar silicification of collagen fibrils
through a one-step collagen self-assembly/silicification approach, RSC Advances.
2017: 55, 34624-34632.
3. Changmin Hu, Mark Aindow, Mei Wei*. Focused ion beam sectioning studies of
biomimetic hydroxyapatite coatings on Ti-6Al-4V substrates, Surface and Coatings
Technology. 2017: 313, 255-262.
4. Changmin Hu, Michael Zilm, Mei Wei*. Fabrication of intrafibrillar and
extrafibrillar mineralized collagen/apatite scaffolds with a hierarchical structure,
Journal of Biomedical Materials Research Part A. 2016: 5, 1153-1161.
5. Changmin Hu, Lichun Zhang, Mei Wei*. Development of biomimetic scaffolds
with both intrafibrillar and extrafibrillar mineralization, ACS Biomaterials Science &
Engineering. 2015: 8, 669-676.
Conference presentations
1. Changmin Hu, Mei Wei*. Sectioning biomimetic collagen-hydroxyapatite coatings
on Ti-6Al-4V substrates using focused ion beam. Society For Biomaterials 2017
Annual Meeting & Exposition. USA, MN, 2017.
2. Changmin Hu, Mark Aindow, Mei Wei*. A comparison of using plasma FIB and
165
Ga FIB to study porous hydroxyapatite coating on Ti6Al4V alloy. 2016 MRS Fall
Meeting & Exhibit, USA, MA, 2016.
3. Changmin Hu, Mei Wei*. Biomimetic silicification of collagen fibrils through a
one-step collagen self-assembly/silicification approach. 10th World Biomaterials
Congress, Canada, Montreal, 2016.
4. Changmin Hu, Lichun Zhang, Mei Wei*. Development of biomimetic scaffolds
with intrafibrillar and extrafibrillar mineralization. Society For Biomaterials 2015
Annual Meeting & Exposition. USA, NC, 2015.