A new approach to mineralization of biocompatible hydrogel scaffolds: An efficient process toward 3-dimensional bonelike composites

Article abstract of DOI:10.1021/ja028559h

As a first step toward the design and fabrication of biomimetic bonelike composite materials, we have developed a template-driven nucleation and mineral growth process for the high-affinity integration of hydroxyapatite with a poly(2-hydroxyethyl methacrylate) (pHEMA) hydrogel scaffold. A mineralization technique was developed that exposes carboxylate groups on the surface of cross-linked pHEMA, promoting high-affinity nucleation and growth of calcium phosphate on the surface, along with extensive calcification of the hydrogel interior. Robust surface mineral layers a few microns thick were obtained. The same mineralization technique, when applied to a hydrogel that is less prone to surface hydrolysis, led to distinctly different mineralization patterns, in terms of both the extent of mineralization and the crystallinity of the apatite grown on the hydrogel surface. This template-driven mineralization technique provides an efficient approach toward bonelike composites with high mineral-hydrogel interfacial adhesion strength.

Full text of DOI:10.1021/ja028559h

Published on Web 01/08/2003
A New Approach to Mineralization of Biocompatible Hydrogel
Scaffolds: An Efficient Process toward 3-Dimensional
Bonelike Composites
Jie Song,†,‡ Eduardo Saiz, and Carolyn R. Bertozzi*
†
,†,‡,§,|
Contribution from the Materials Sciences DiVision, Lawrence Berkeley National Laboratory,
Departments of Chemistry and Molecular and Cell Biology, and Howard Hughes Medical
Institute, UniVersity of California, Berkeley, California 94720
Received September 16, 2002 ; E-mail: bertozzi@cchem.berkeley.edu
Abstract: As a first step toward the design and fabrication of biomimetic bonelike composite materials, we
have developed a template-driven nucleation and mineral growth process for the high-affinity integration
of hydroxyapatite with a poly(2-hydroxyethyl methacrylate) (pHEMA) hydrogel scaffold. A mineralization
technique was developed that exposes carboxylate groups on the surface of cross-linked pHEMA, promoting
high-affinity nucleation and growth of calcium phosphate on the surface, along with extensive calcification
of the hydrogel interior. Robust surface mineral layers a few microns thick were obtained. The same
mineralization technique, when applied to a hydrogel that is less prone to surface hydrolysis, led to distinctly
different mineralization patterns, in terms of both the extent of mineralization and the crystallinity of the
apatite grown on the hydrogel surface. This template-driven mineralization technique provides an efficient
approach toward bonelike composites with high mineral-hydrogel interfacial adhesion strength.
Introduction
outcomes of these implants are not satisfactory. First, these
materials exhibit serious mechanical property mismatches with
Bone is a complex tissue that serves many essential functions
in the body. It protects organs, provides support and site-of-
muscle attachment, generates blood cells, and helps maintain
essential ion levels. Structurally, natural bone is a composite
of collagen, a protein-based hydrogel template, and inorganic
dahilite (carbonated apatite) crystals. The unusual combination
of a hard inorganic material and an underlying elastic hydrogel
network endows native bone with unique mechanical properties,
such as low stiffness, resistance to tensile and compressive
natural bone tissues, which can cause stress shielding and lead
to bone resorption when the material has a higher Young’s
3
modulus than bone. Very often, revision surgery will be
4
required to follow up the initial implantation. A second major
limitation of traditional bone implants is the lack of interaction
between these implants and their tissue environment. These
materials typically do not bear any functionalities that encourage
communication with their cellular environment and, therefore,
can only be categorized as bio-inert, far from being bioactive.5
These “static” implants are not capable of effectively triggering
the healing cascade upon surgical implantation, therefore
limiting the potential for tissue attachment and in-growth.
The development of a new generation of bonelike composite
materials with improved mechanical properties and enhanced
biocompatibility calls for a biomimetic synthetic approach using
natural bone as a guide. It is suggested that in natural bone
synthesis, the biomineralization process starts with the formation
1
forces, and high fracture toughness. Throughout the cavities
of bone, there are bone cells and a myriad of soluble factors
and extracellular matrix components that are constantly involved
2
with the bone formation and remodeling process. The unique
biological functions and mechanical properties of bone are
appealing to materials scientists, engineers, and clinicians for a
variety of applications, yet the complex nature of bone’s
structure has hindered real biomimetic design of artificial
bonelike materials for a broad range of applications including
the treatment of bone defects.
Current artificial materials used in the fabrication of ortho-
pedic implants, metals, ceramics, or polymers, were originally
developed for nonbiological applications. Although they could
provide an immediate solution for many patients, the long-term
6
of poorly crystalline calcium apatites (preceded by possible
transient amorphous calcium phosphates), which are then
modified through crystalline phase transitions to form the more
7
stable mature bone apatites with increased crystallinity. During
these crystalline phase transitions, acidic extracellular matrix
proteins that are attached to the collagen scaffold play important
†
Materials Sciences Division, Lawrence Berkeley National Laboratory.
Department of Chemistry, University of California, Berkeley.
Department of Molecular and Cell Biology, University of California,
‡
(3) Black, J. Biological Performance of Materials: Fundamentals of Biocom-
patibility, 3rd, rev. and expanded ed.; Marcel Dekker: New York, 1999.
§
(
4) National Materials AdVisory Board Newsletter, 1997.
Berkeley.
(
5) Willmann, G. AdV. Eng. Mater. 1999, 1, 95-105.
|
Howard Hughes Medical Institute, University of California, Berkeley.
(6) Buckwalter, J. A.; Glimcher, M. J.; Cooper, R. R.; Recker, R. J. Bone
Joint Surg.-Am. Vol. 1995, 77A, 1256-1275.
(7) Mann, S. Biomineralization: Principles and Concepts in Bioinorganic
Materials Chemistry; Oxford University Press: New York, 2001.
(
1) Weiner, S.; Wagner, H. D. Annu. ReV. Mater. Sci. 1998, 28, 271-298.
2) Sikavitsas, V. I.; Temenoff, J. S.; Mikos, A. G. Biomaterials 2001, 22,
(
2
581-2593.
J. AM. CHEM. SOC. 2003, 125, 1236-1243
1
236
9
10.1021/ja028559h CCC: $25.00 © 2003 American Chemical Society

3-D Bonelike Composites
A R T I C L E S
templating and inhibitory roles.8,9 Presumably, the acidic groups
serve as binding sites for calcium ions and align them in an
materials that integrate organic scaffolds and HA, and demon-
strate the level of integration of natural bone, have not yet been
achieved.
1
0,11
orientation that matches the apatite crystal lattice.
The
detailed mechanism of the formation and remodeling of natural
bone has long been a debate and is still subject to intensive
investigations. However, it is clear that template-driven biom-
ineralization, regulated by a number of extracellular matrix
components and the participation of bone cells, plays an
important role in formation of the highly integrated composite
structure of bone. This is the critical feature that the biomimetic
synthesis of artificial bone should emulate. We hypothesize that
this can be realized by the generation of functional polymer
scaffolds displaying surface ligands that mimic critical extra-
cellular matrix components known to direct template-driven
Hydrogel polymers are particularly appealing candidates for
the design of highly functional tissue engineering scaffolds.
27,28
The intrinsic elasticity and water retention ability of synthetic
hydrogels resemble those of natural hydrogels, such as collagen
matrices that are prevalent as structural scaffolds in various
27
connective tissues including bone. The porosity of synthetic
hydrogels may be controlled by various techniques including
29,30
31
solvent casting/particulate leaching,
phase separation, gas
32
33
34
foaming, solvent evaporation, freeze-drying, and blending
with non-crosslinkable linear polymers to afford a range of
35
mechanical properties. Another important feature of hydrogels
is that they can be assembled in 3-dimensional form displaying
multiple functional domains through copolymerization of dif-
ferent monomers. The polymerization chemistry is water
compatible, allowing incorporation of polar ligands such as
anionic peptides that mimic the acidic matrix proteins regulating
biomineralization, or to stimulate or assist cell adhesion,
proliferation, migration, and differentiation.1
2-15
Ideally, such
3
-dimensional scaffolds should also possess proper physical
properties, such as desired porosity and water retention ability,
to allow both pre-implantation cell seeding and post-implantation
tissue ingrowth.
mineral growth, and biological epitopes such as the tripeptide
There has been considerable effort to mimic bone by the
mineralization of polymer films with hydroxyapatite (HA) (Ca10-
36-38
RGD
that promote cellular adhesion.
Poly(2-hydroxyethyl methacrylate), or pHEMA, is one of the
(PO4)6(OH)2), the major inorganic component of natural bone.
3
9
most well-studied synthetic hydrogel polymers. With its high
biocompatibility, pHEMA and its functionalized copolymers
have become some of the most widely used synthetic hydrogels
in tissue engineering. Applications include ophthalmic devices
This is usually attempted through the time-consuming incubation
of substrates with simulated body fluid (SBF).16 This leads to
slow growth of crystalline or amorphous biominerals that exhibit
poor adhesion and lack a structural relationship with the
4
0,41
42
substrate.1
6-18
(e.g., contact lens),
cartilage replacements, bonding agents
Efforts aimed at improving the interface of HA
43-45
1
9,20
in dental resins and bone cements, and various drug delivery
One of the major challenges for its application as
with polymers include the use of silane coupling agents,
zirconyl salts, polyacids,
4
6,47
2
1
22,23
24,25
vehicles.
and isocyanates.
Recently,
a 3-dimensional scaffold of artificial bonelike materials, how-
ever, is to realize a high-affinity integration of inorganic minerals
with the organic pHEMA-based scaffold. This will be required
to achieve a composite material with unique mechanical
chemical treatment of biodegradable poly(lactide-co-glycolide)
films with aqueous base has been shown to facilitate the growth
of crystalline carbonate apatite on the surface.26 Both the
morphology and the thickness of the resulting crystalline apatite
layer, however, suggest that it will suffer from inadequate
interfacial adhesion (between the mineral and the polymer
substrate) and poor mechanical properties. Overall, composite
(
27) Peppas, N. A. Hydrogels in Medicine and Pharmacy; CRC Press: Boca
Raton, 1986.
(28) Lee, K. Y.; Mooney, D. J. Chem. ReV. 2001, 101, 1869-1879.
(
29) Holy, C. E.; Shoichet, M. S.; Davies, J. E. J. Biomed. Mater. Res. 2000,
51, 376-382.
(
8) Raj, P. A.; Johnson, M.; Levine, J. M.; Nancollas, H. G. J. Biol. Chem.
(30) Mooney, D. J.; Cima, L.; Langer, R.; Johnson, L.; Hansen, L. K.; Ingber,
D. E.; Vacanti, J. P. Mater. Res. Soc. Symp. Proc. 1992, 252, 345-352.
(31) Klawitter, J. J.; Hulbert, S. F. J. Biomed. Mater. Res. Symp. 1971, 2, 161-
229.
1
992, 267, 5968-5976.
(
9) Clark, R. H.; Campbell, A. A.; Klumb, L. A.; Long, C. J.; Stayton, P. S.
Calcif. Tissue Int. 1999, 64, 516-521.
(
(
10) Addadi, L.; Weiner, S. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 4110-
(32) Harris, L. D.; Kim, B.-S.; Mooney, D. J. J. Biomed. Mater. Res. 1998, 42,
396-402.
(33) Mikos, A. G.; Sarakinos, G.; Leite, S. M.; Vacanti, J. P.; Langer, R.
Biomaterials 1993, 14, 323-330.
(34) Whang, K.; Healy, K. E. In Methods of Tissue Engineering; Atala, A.,
Lanza, R. P., Eds.; Academic Press: San Diego, San Francisco, New York,
Boston, London, Sydney, Tokyo, 2002; pp 697-704.
(35) Brauker, J. H.; Carr-Brendel, V. E.; Martinson, L. A.; Crudele, J.; Johnston,
W. D.; Johnson, R. C. J. Biomed. Mater. Res. 1995, 29, 1517-1524.
(36) Massia, S. P.; Hubbell, J. A. Ann. N. Y. Acad. Sci. 1990, 589, 261-270.
(37) Barrera, D. A.; Zylstra, E.; Lansbury, P. T.; Langer, R. J. Am. Chem. Soc.
1993, 115, 11 010-11 011.
(38) Ratner, B. D. J. Mol. Recognit. 1996, 9, 617-625.
(39) Kost, J.; Langer, R. In Hydrogels in Medicine and Pharmacy; Peppas, N.,
Ed.; CRC Press: Boca Raton, 1987; Vol. III, p 95.
(40) Phillips, A. J.; Stone, J. Contact Lenses; Butterworth & Co.: London,
Boston, 1989.
4
114.
11) George, A.; Bannon, L.; Sabsay, B.; Dillon, J. W.; Malone, J.; Veis, A.;
Jenkins, N. A.; Gillbert, D. J.; Copeland, N. G. J. Biol. Chem. 1996, 271,
3
2 869-32 873.
(
(
(
(
12) Shea, L. D.; Smiley, E.; Bonadio, J.; Mooney, D. J. Nat. Biotechnol. 1999,
7, 551-554.
13) Lee, K. Y.; Peters, M. C.; Anderson, K. W.; Mooney, D. J. Nature 2000,
08, 998-1000.
14) Mann, B. K.; Schmedlen, R. H.; West, J. L. Biomaterials 2001, 22, 439-
44.
15) Luo, Y.; Dalton, P. D.; Shoichet, M. S. Chem. Mater. 2001, 13, 4087-
1
4
4
4
093.
(
(
(
16) Kokubo, T. Acta Mater. 1998, 46, 2519-2527.
17) Rhee, S.-H.; Tanaka, J. Biomaterials 1999, 20, 2155-2160.
18) Saiz, E.; Goldman, M.; Gomez-Vega, J. M.; Tomsia, A. P.; Marshall, G.
W.; Marshall, S. J. Biomaterials 2002, 23, 3749-3756.
(
(
19) Labella, R.; Braden, M.; Deb, S. Biomaterials 1994, 15, 1197-1200.
20) Dupraz, A. M. P.; de Wijn, J. R.; vander Meer, S. A. T.; de Groot, K. J.
Biomed. Mater. Res. 1996, 30, 231-238.
(41) Kidane, A.; Sxabocsik, J. M.; Park, K. Biomaterials 1998, 19, 2051-2055.
(42) Oxley, H. R.; Corkhill, P. H.; Fitton, J. H.; Tighe, B. J. Biomaterials 1993,
14, 1064-1072.
(
(
21) Misra, D. N. J. Dent. Res. 1985, 12, 1405-1408.
(43) Vermeiden, J. P. W.; Rejda, B. B.; Peelen, J. G. J.; de Groot, K. In
EValuation of Biomaterials; Winter, G. D., Leray, J. L., de Groot, K., Eds.;
Wiley: New York, 1980; pp 405-411.
22) Bradt, J.-H.; Mertig, M.; Teresiak, A.; Pompe, W. Chem. Mater. 1999, 11,
2
694-2701.
(
(
(
(
23) Liu, Q.; de Wijn, J. R.; Bakker, D.; van Toledo, M.; van Blitterswijk, C.
(44) Yang, J. M.; You, J. W.; Chen, H. L.; Shih, C. H. J. Biomed. Mater. Res.
1996, 33, 83-88.
A. J. Mater. Sci.-Mater. Med. 1998, 9, 23-30.
24) Liu, Q.; de Wijn, J. R.; van Blitterswijk, C. A. J. Biomed. Mater. Res.
(45) Prati, C.; Mongiorgi, R.; Valdre, G.; Montanary, G. Clin. Mater. 1991, 8,
137-143.
1
998, 40, 257-263.
25) Liu, Q.; de Wijn, J. R.; van Blitterswijk, C. A. J. Biomed. Mater. Res.
998, 40, 358-364.
26) Murphy, W. L.; Mooney, D. J. J. Am. Chem. Soc. 2002, 124, 1910-1917.
(46) Lu, S.; Anseth, K. S. J. Controlled Release 1999, 57, 291-300.
(47) Sefton, M. V.; May, M. H.; Lahooti, S.; Babensee, J. E. J. Controlled
Release 2000, 65, 173-186.
1
J. AM. CHEM. SOC.
9
VOL. 125, NO. 5, 2003 1237

A R T I C L E S
Song et al.
Scheme 1. Urea-Mediated Solution Mineralization of Hydroxyapatite (HA) onto pHEMA Hydrogel Scaffoldsa
a
Thermo-decomposition of urea produces a gradual increase in pH, resulting in the hydrolysis of surface 2-hydroxyethyl esters (A) and the precipitation
of HA from the aqueous solution. The in situ generated surface carboxylates strongly interact with calcium ions (B) and facilitate the heterogeneous nucleation
and 2-dimensional growth of a high-affinity calcium-phosphate (CP) layer on the pHEMA surface (C). Prolonged mineralization allows for the growth of
a thicker CP layer that covers the entire hydrogel surface as shown in D. A proposed urea-mediated, pH-dependent HA nucleation and growth behavior from
the solution in qualitatively depicted by the red dotted curve shown in E (curve 1, the solubility of HA, was taken from reference 50), guiding the chemical
and physical transformation of the pHEMA hydrogel to a highly integrated Gel-CP composite.
properties, biocompatibility and osteophilic interaction with its
natural bone environment.
water content (EWC). All subsequent experiments were carried
out with gels cross-linked by 2% EGDMA.
Here, we report the development of a rapid and effective
mineralization method that leads to high affinity integration of
HA with a pHEMA hydrogel scaffold. This work serves as an
important first step toward the development of hydrogel-based
biomimetic composite materials. The correlation between the
surface properties of hydrogel substrates, the extent of miner-
alization, the strength of mineral adhesion (at the gel-mineral
interface), and the mineral crystallinity was also investigated.
This robust mineralization technique, when combined with
further incorporation of functional domains of increasing
molecular complexity through copolymerization with HEMA,
could lead to new generations of composite materials with
enhanced tissue-implant interactions. Such a “bottom-up” ap-
proach will facilitate our understanding of structure-function
relationships in template-driven biomineralization, and help us
derive a set of rules to guide future rational design of composite
materials.
Calcium apatites are known to promote bone apposition and
differentiation of mesenchymal cells to osteoblasts.49 In this
work, we chose to use synthetic HA in the fabrication of
hydrogel-based bonelike composite materials. HA has limited
solubility in water at neutral and basic pH but is highly soluble
5
0
at acidic pH. On the basis of this property, we employed a
urea-mediated solution precipitation technique that has been
previously used in the preparation of composite ceramic
5
1,52
powders.
In an adapted protocol, a segment of pHEMA
hydrogel was soaked in an acidic solution (pH 2.5-3) of HA
containing a high concentration of urea (2 M). As depicted in
Scheme 1, upon gradual heating (without stirring) from room
temperature to 90-95 °C (within 2 h), urea started to decompose
and the pH slowly increased (around pH 8). Under these
conditions, some hydrolysis of the 2-hydroxyethyl esters oc-
curred at the hydrogel surface, promoting heterogeneous
nucleation and 2-dimensional growth of calcium phosphate (CP)
(
Scheme 1, A-C). By allowing mineralization to proceed for a
Results
longer time after reaching 95 °C (12 h total), a CP layer several
microns thick formed over the entire hydrogel surface (Scheme
2
-Hydroxyethyl methacrylamide (HEMAm) was synthesized
1
D).
There are several notable features of this procedure. First,
increasing pH and temperature during the process promotes the
through direct coupling of ethanolamine with methacryloyl
chloride under slightly basic (pH 8) conditions. A standard
radical polymerization protocol was used for the preparation
of pHEMA and pHEMAm.48 The cross-linker ethylene glycol
dimethacrylate (EGDMA) was used at 2%, 5%, and 10% (by
weight) to afford pHEMA gels with varied degrees of cross-
linking. These gels were found to have same (40%) equilibrium
(
(
(
(
49) Darimont, G. L.; Cloots, R.; Heinen, E.; Seidel, L.; Legrand, R. Biomaterials
2002, 23, 2569-2575.
50) Nancollas, G. H.; Zhang, J. In Hydroxyapatite and Related Materials;
Brown, P. W., Constantnz, B., Eds.; CRC: Boca Raton, 1994; p pp73.
51) Blendell, J. E.; Bowen, H. K.; Coble, R. L. Am. Ceram. Soc. Bull. 1984,
63, 797-801.
52) De Jonghe, L. C.; He, Y. In Ceramic Microstructures: Control at the Atomic
LeVel; Tomsia, A. P., Glaeser, A., Eds.; Plenum Press: New York, 1998;
pp 559-565.
(
48) Chilkoti, A.; Lopez, G. P.; Ratner, B. D. Macromolecules 1993, 26, 4825-
4
832.
1238 J. AM. CHEM. SOC.
9
VOL. 125, NO. 5, 2003

3-D Bonelike Composites
A R T I C L E S
hydrolysis of the ethyl ester side chains of pHEMA and leads
to the in-situ generation of an acidic surface and a partially acidic
interior that has high affinity for calcium ions. Simultaneously,
the mineral concentration in the solution undergoes a dramatic
change (as the pH rose from 3 to 8) as depicted by the red
dotted curve shown in Scheme 1E. Curve 1, adapted from the
work of Nancollas et al., depicts the general solubility of
hydroxyapatite.5 Curves 2 and 3 represent qualitative depictions
of typical nucleation behavior of calcium phosphates derived
0
5
1,52
from basic nucleation theory reported in the literature.
The
subtle differences in solubility and precipitation behavior of
various calcium phosphates are not reflected in this scheme.
The high affinity between the calcium ions and the exposed
carboxylate groups at the gel surface translates into a low
interfacial energy between the hydrogel and calcium phosphate,
and consequently, a low energy barrier for the heterogeneous
nucleation of mineral on the hydrogel surface (as shown by
curve 2 in Scheme 1E). Second, the thermo-decomposition of
urea allows a homogeneous variation of pH across the solution,
avoiding a sudden local pH change that is commonly observed
with strong base-induced precipitation.52 Third, any homoge-
neous precipitates that fall on the hydrogel and remain loosely
attached to the composite surface can be easily washed away
during the workup rinsing. Reproducible urea-mediated min-
eralization can be achieved by closely monitoring the amount
of urea used (affecting the final pH), the heating rate (affecting
mineral nucleation and growth rate) and the duration of the
process (affecting the extent of mineral coverage on the hydrogel
surface). We have shown that heating rates between 0.2 °C/
min and 0.5 °C/min will promote the formation of a CP layer
that uniformly covers the gel surface.
Figure 1. Morphology, crystallinity, and mineral-gel interfacial affinity
of calcium phosphate apatite layer grown on the surface of pHEMA after
2 h. (A) SEM showing 2-dimensional circular outward growth of calcium
apatite from multiple nucleation sites on the acidic surface of pHEMA. (B)
SEM showing the merge of circular mineral layers and the full coverage of
the hydrogel surface with calcium apatite. (C) SEM showing an indent
formed on the surface of mineralized pHEMA using a Vickers microindenter
with a load of 5 N. The calcium phosphate layer did not delaminate. (D)
SEM-associated EDS area analysis of the mineral layer shown in micrograph
B, confirming the chemical composition and Ca/P ratio (1.6 ( 0.1) that is
typical for HA. Synthetic HA was used to calibrate the determination of
the Ca/P ratio. (E) X-ray diffraction patterns of the pHEMA composite (a)
and unmineralized pHEMA gel (b). The lack of diffraction peaks corre-
sponding to crystalline HA suggests that an amorphous or nanocrystalline
layer was formed on the pHEMA surface.
hydrogel composite. No delamination of the mineral layer was
observed by SEM even after Vickers indentations with loads
of 5-15 N (Figure 1C) which is an indication of good adhesion.
The calibrated energy dispersive spectroscopy (EDS) area
analysis performed on the mineral surface of the composite
revealed a Ca/P ratio (1.6(0.1) similar to that of synthetic HA
(Figure 1D).
The hydrolysis of 2-hydroxyethyl esters during the thermo-
decomposition of urea was expected to lead to an increase of
surface hydrophilicity, which was confirmed by contact angle
and EWC measurements. In a mock experiment, a segment of
pHEMA gel was thermally treated as described above in a
solution containing the same concentration of urea, without the
presence of HA. Diiodomethane, a hydrophobic solvent that is
known to from stable droplets on hydrophilic materials without
Most of the flakelike crystal apatite coatings obtained by
mineralization in SBF on bioactive glasses, polymer scaffolds
or collagen films are not robust and tend to delaminate easily
1
6-18,55
upon drying.
We attempted microindentation analysis
53
noticeable penetration and contact angle hysteresis, was used
for measuring contact angles against water on both the treated
and untreated pHEMA gels. The contact angle of a di-
iodomethane droplet on the gel surface was found to increase
from 129° to 142° upon the urea-mediated thermal treatment
for 2 h. The observed decrease in surface wettability by a
hydrophobic solvent is consistent with the postulated in-situ
generation of polar surface carboxylates during the urea-
mediated mineralization. The hydrolysis also led to a slight
increase (2-3%) in the EWC of the gel.
with the apatite layers formed on bioglass using SBF mineral-
1
8
ization. The minerals crumbled with even very low loadings
and are therefore considered not amendable for such analysis.
Recently, a modified mineralization approach, involving the use
of very high concentrations of SBF and careful control of
-
HCO3 concentrations, was used to bind nanosized apatite
crystals to commercial metal implants with improved strength.5
However, no mechanical tests that could characterize the
interfacial adhesion between these apatites and the metal
substrates were reported.
6,57
X-ray diffraction (XRD) analysis performed on the mineral-
ized pHEMA composite indicated that the calcium phosphate
layer was either nanocrystalline or amorphous. No typical
reflections for crystalline HA were observed, with only a few
broad reflections (Figure 1E, a) similar to those observed in
the hydrogel prior to mineralization (Figure 1E, b). This suggests
The strong affinity between calcium and the in situ generated
acidic surface of pHEMA led to the 2-dimensional outward
growth of calcium phosphate from individual nucleation sites
(Figure 1A). After 2 h, the circular mineral layers merged and
covered the entire surface (Figure 1B). The adhesion strength
of the apatite layer to the gel surface was studied by microin-
dentation analysis54 performed on the surface of the mineral-
(
55) Du, C.; Cui, F. Z.; Zhang, W.; Feng, Q. L.; Zhu, X. D.; de Groot, K. J.
Biomed. Mater. Res. 2000, 50, 518-527.
(
53) Timmons, C. O.; Zisman, W. A. J. Colloid Interface Sci. 1966, 22, 165-
71.
54) Gomez-Vega, J. M.; Saiz, E.; Tomsia, A. P. J. Biomed. Mater. Res. 1999,
6, 549-559.
(56) Barrere, F.; van Blitterswijk, C. A.; de Groot, K.; Layrolle, P. Biomaterials
2002, 23, 1921-1930.
1
(
(57) Barrere, F.; van Blitterswijk, C. A.; de Groot, K.; Layrolle, P. Biomaterials
2002, 23, 2211-2220.
4
J. AM. CHEM. SOC.
9
VOL. 125, NO. 5, 2003 1239

A R T I C L E S
Song et al.
These mineral spheres are composed of platelike crystallites
(insert of Figure 2B), a typical morphology observed with the
crystalline apatite grown on bioactive glasses polymer substrates
16-18,26
or collagen films using SBF mineralization.
EDS analysis
performed on the spherical apatite aggregates and XRD
performed on the composite material confirmed the expected
Ca/P ratio (1.6(0.1) and typical reflections, (002) and (112),
1
8
for crystalline HA (Figure 2C & 2D). It is worth noting,
however, that HA crystals formed by adventitious precipitation
often adopt a similar preferential orientation. This suggests that
the observation of a preferential alignment of the apatitic crystal
lattice in a composite material does not necessarily reflect a
specific interaction with the underlying substrate. The cross-
section examination of the composite material revealed that there
were significant degrees of calcification inside the hydrogel as
well, although the degree of phosphate incorporation was limited
(
Figure 2E & 2F). The extensive calcification of the hydrogel
interior may be promoted by the partial hydrolysis of the
-hydroxyethyl ester side chains inside the pHEMA gel. The
2
growth of calcium apatites inside the pHEMA scaffold is limited
by both the space (note that no special technique was applied
to promote the formation of a porous hydrogel scaffold in the
current study) and the concentration of free anions achieved
inside the already partially anionic hydrogel.
To better understand the relationship between surface chem-
istry of the substrate and the mineralization pattern of the
composite material, we applied the same mineralization tech-
nique to poly(2-hydroxyethyl methacrylamide) (pHEMAm), a
hydrogel that is not prone to side chain hydrolysis under the
mineralization conditions. Indeed, an entirely different surface
mineral pattern was obtained with pHEMAm gels as shown in
Figure 3. The hydrogel was patterned with flowerlike minerals,
with much less extensive surface coverage even after 12 h of
mineralization. The apatite grown on pHEMAm was crystalline
as suggested by both dark field optical image (Figure 3A) and
XRD of the composite material (Figure 3B), with major
reflections matching with those of crystalline HA. The relative
intensities of these reflections suggest a preferential alignment
along (002), with the c-axis perpendicular to the substrate. SEM
micrographs revealed further details of the mineral pattern,
showing an upward growth of the bundles of whiskers away
from the gel surface (Figure 3C & 3D). This, along with the
relatively low surface mineral coverage, is consistent with the
decreased affinity between calcium apatite and the neutral
hydrogel surface of pHEMAm. EDS analysis performed on the
mineral bundles again revealed a Ca/P ratio (1.6(0.1) matching
HA (insert of Figure 3D). X-ray elemental mapping of Ca and
P (Figure 3E & 3F) showed that the mineral patterns were
composed of uniform calcium phosphate apatite. There was no
calcification or phosphate incorporation detected at the interior,
or on the dark hydrogel surface positions devoid of flowerlike
mineral patterns, in agreement with the low-calcium binding
nature of pHEMAm. Clearly, surface chemistry plays a very
important role in determining both the extent and the pattern
of the mineralization (2-D vs 3-D mineral growth). Tightly
bound HA prefers to spread on the surface, forming a circularly
grown mineral layer that eventually covers the entire gel surface.
By contrast, more loosely bound HA, as in the case of
pHEMAm, can grow more readily in three dimensions.
Figure 2. Extended mineralization of pHEMA via a urea-mediated process
(
12 h): morphology, crystallinity and thickness of calcium apatite layer
grown on the hydrogel surface and the calcification of the hydrogel interior.
A) SEM showing the side view of a cross-section of the pHEMA-apatite
(
composite after extended mineralization. The sample stage was tilted 45°.
Note the micron scale thickness of the mineral layer and the fine integration
at the mineral-gel interface. (B) SEM showing fully mineralized surface of
pHEMA after extended mineralization, along with the growth of spherical
clusters of crystalline HA from nucleation sites located at the center of
circular mineral layers. The inset shows an expanded view of one spherical
cluster of HA crystallites. Note the platelike morphology of the calcium
apatite crystallites that is commonly observed with crystalline apatites grown
on various substrates during SBF mineralization. (C) Calibrated EDS area
analysis of the surface mineral layer shown in micrograph B, confirming
the chemical composition and Ca/P ratio (1.6 ( 0.1) that is typical of HA.
Synthetic HA was used to calibrate the quantification of Ca/P ratio. Same
result was obtained on the selected area analysis performed on plate-like
HA shown in the inset of micrograph B. (D) X-ray diffraction pattern of
pHEMA-HA composite obtained by the urea-mediated mineralization
process after 12 hours (the composite shown in Fig. 2B). Peaks corre-
sponding to crystalline HA were detected, suggesting the crystalline nature
of the spherical HA grown on top of the calcium apatite coating. Note the
preferential alignment along the c-axis. (E) SEM showing the hydrogel
interior of the pHEMA-apatite composite. (F) SEM-associated EDS area
analysis of the cross-section of pHEMA-apatite composite shown in
micrograph E, suggesting significant calcification throughout the hydrogel
interior.
that although there is high affinity binding between calcium
ions and the in-situ generated surface carboxylates, the spacing,
order and/or alignment of these surface anionic ligands do not
promote the epitaxial growth of large calcium apatite crystallites
along any particular orientation.
By extending mineralization time to 12 h, mineral coatings
with thicknesses up to several microns were obtained, with good
integration at the mineral-gel interface as shown in a side view
image of the composite (Figure 2A). Once the surface was fully
covered with the amorphous apatite layer, the growth of HA
crystals forming spherical aggregates on top of initial nucleation
sites (centers of circular apatite rings) was observed (Figure 2B).
1240 J. AM. CHEM. SOC.
9
VOL. 125, NO. 5, 2003

3-D Bonelike Composites
A R T I C L E S
properties of 3-dimensional hydrogel scaffolds, and opens the
door for future application of pHEMA in the design of
functionalized bone replacements.
As represented in Scheme 1, the formation of the robust CP
layer on the hydrolyzed surface of pHEMA gel is driven by
the lower interfacial energy between the carboxylate rich gel
surface and the mineral. Once the pH and temperature reach an
equilibrium and the gel surface is fully covered with an initial
nanocrystalline or amorphous CP layer of a critical thickness,
the growth of CP composed of large platelet crystallites that
are easily detectable by XRD becomes energetically favorable.
Maintaining the final pH of the mineralization solution at around
8
prevents competing homogeneous precipitation from the
medium.
The distinctly different mineralization patterns observed on
pHEMA and pHEMAm gels using the same mineralization
procedure indicates that the chemical nature of the hydrogel
dictates the affinity and extent of mineralization. Other surfaces,
such as bioglass have been shown to mineralize with crystalline
1
6,18
CP.
However, crystalline apatite growth is not necessarily
a result of template-driven mineralization. Instead, it could occur
even with poor substrate-mineral interaction resulting from a
random crystallization process. Existing mineralization tech-
niques applied to a vast range of substrates using SBF often
result in the formation of similar crystalline apatite coatings
composed of either platelike or flakelike apatite crystals.1
The formation of such loosely covered crystalline mineral layers
indicates limited substrate-mineral interaction. That is, such
mineral nucleation and growth are not highly surface dependent
or strictly template driven. The precipitation is instead promoted
by the subtle local pH changes and/or local ion saturation
induced by the substrate. Similarly, in this study, the growth of
crystalline calcium apatite spheres after saturation of the initial
mineral coating is surface independent. Our results suggest that
a true template driven biomineralization process requires the
establishment of direct and extensive mineral-substrate contact,
as we observed between pHEMA and HA. High affinity
2-dimensional mineral growth at the substrate-mineral interface
reflects such extensive interactions.
6,18,26
Figure 3. Mineralization of pHEMAm via the urea-mediated process after
h: morphology and crystallinity of hydroxyapatite (HA) grown on the
2
hydrogel surface. (A) Dark field optical microscopy image of the mineralized
pHEMAm gel. The bright reflection of the flowerlike minerals suggests
their crystalline nature. (B) X-ray diffraction pattern of the pHEMAm-
mineral composite obtained by the urea mediated mineralization process.
Hydroxyapatite peaks can be observed, suggesting the crystalline nature of
the minealization. Note the preferential alignment along the c-axis. (C) SEM
showing unique surface mineral patterns of pHEMAm-HA composite. (D)
SEM showing an expanded view of the tip of one “petal” of the flower-
like mineral shown in micrograph C. Note the upward growth of crystal
bundles away from the dark hydrogel surface. The inset shows calibrated
EDS area analysis of the surface of the composite material, confirming the
chemical composition and Ca/P ratio (1.6 ( 0.1) of the surface mineral
that is typical of HA. The EDS analysis performed on the un-mineralized
hydrogel surface did not detect any Ca or P signal (data not shown). (E)
X-ray elemental mapping of Ca with the same sample area shown in
micrograph C. Note that there is no calcification of the dark hydrogel
background outside of the flowerlike mineral pattern. (F) X-ray elemental
mapping of P with the same sample area shown in micrograph C. Note
that there is no phosphate incorporation in the dark hydrogel background
outside of the flower-like mineral pattern.
In natural bone synthesis, it has been suggested that the
biomineralization process starts with the formation of transient
amorphous calcium phosphates (although the direct detection
of these transient precursors is difficult using static techniques
5
8,59
such as XRD and solid-state NMR
apatites.
) and poorly crystalline
5
9,60
These precursors then undergo several crystalline
Discussion
phase transitions, such as brushite, octacalcium phosphate, and
tricalcium phosphate, before the more stable bone apatites with
The mineralization method we describe here takes advantage
of the dramatically different solubilities of HA in acidic and
basic aqueous media and the chemically labile nature of ester
groups of pHEMA in basic media. We employed thermo-
decomposition of urea as a facile pH modulator to generate an
anionic surface and partially acidic interior of the pHEMA gel.
This initiated the heterogeneous nucleation and high-affinity
growth of calcium apatite on the gel surface and extensive
calcification inside the gel. Overall, this method provides a fast
and convenient approach for producing robust pHEMA-calcium
apatite composite materials with high quality interfacial integra-
tion between the mineral and the polymer substrate. Further-
more, this approach provides a foundation for integrating high-
affinity template-driven biomineralization with the versatile
7
higher crystallinity finally form. This complex process is
mediated by the dissolution and saturation of mineral ions at
the mineral-substrate interface, as well as bone matrix proteins.
Two different paradigms can be envisioned that mimic natural
bone synthesis in the fabrication of bone-like composite
materials. One could start with a highly ordered molecular
template to promote template-driven mineralization of crystalline
61
apatites. Indeed, this approach has been pursued in the context
(
58) Roberts, J. E.; Heughebaert, M.; Heughebaert, J. C.; Bonar, L. C.; Glimcher,
M. J.; Griffin, R. G. Calcif. Tissue Int. 1991, 49, 378-382.
(59) Roberts, J. E.; Bonar, L. C.; Griffin, R. G.; Glimcher, M. J. Calcif. Tissue
Int. 1992, 50, 42-48.
(
60) Kim, H. M.; Rey, C.; Glimcher, M. J. Calcif. Tissue Int. 1996, 59, 58-63.
(61) Hartgerink, J. D.; Beniash, E.; Stupp, S. I. Science 2001, 294, 1684-1688.
J. AM. CHEM. SOC. VOL. 125, NO. 5, 2003 1241
9

A R T I C L E S
Song et al.
of nanotube composites, using supramolecular aggregates of
peptide-amphiphiles bearing biomimetically designed mineral
nucleating ligands. Alternatively, one could start with composite
materials with properly adhered amorphous or nanocrystalline
osteophilic mineral compositions and encourage nature’s re-
modeling pathway to eventually take over and engineer the
initial mineral phase. It is this latter approach that may be
enabled by the method presented herein.
reduced pressure prior to use. In a typical procedure, 500 mg
of hydrogel monomer was combined with 10 µL of ethylene
glycol dimethacrylate, 100 µL of Milli-Q water and 150 µL of
ethylene glycol. To this mixture was added 50 µL each of an
aqueous solution of sodium metabisulfite (150 mg/mL) and
ammonium persulfate (400 mg/mL). The well-mixed viscous
solution was then poured into a glass chamber made by
microscope slides and allowed to stand at room temperature
overnight. The gels (5.5 cm × 1.5 cm × 1 mm) were then
soaked in Milli-Q water for 2-3 d, with daily exchange of
freshwater, to ensure the complete removal of unreacted
monomers before they were used for mineralization and further
physical characterizations.
Equilibrium Water Content (EWC) Measurements. The
EWC at room temperature is defined as the ratio of the weight
of water absorbed by a dry hydrogel to the weight of the fully
hydrated hydrogel. The amount of water absorbed by the
hydrogel is determined from the weight of a freeze-dried gel
(Wd) and the weight of the corresponding hydrated gel (Wh)
according to the following equation
In addition to strong adhesion at the gel-mineral interface,
the composites generated here have a HA layer with a structure
and thickness that are ideal for bone implant applications.
Analysis of calcium phosphate coatings on titanium implants
has shown that resorption of the coating occurs mostly in the
less organized apatite region and stops where the coating has
higher crystallinity.49 Thus, amorphous or nanocrystalline HA
coatings as we have achieved should promote resorption and
bone integration. In addition, earlier studies suggest that a thin
layer of HA with thickness on the order of 1-7 µm, as we
have achieved, provides a sufficient HA resorption time frame
to allow a progressive bone contact with the implant substrate
and is, therefore, ideal for inducing integration of the material
into natural bone.49 By contrast, classical plasma spray tech-
niques applied to metal implants6 produce HA coatings over
EWC (%) ) [(W - W )/W ] × 100
h
d
h
2
Contact Angle Measurements. The contact angles of di-
iodomethane droplets against water on hydrogels were measured
using the sessile drop method. A 3-5 µL droplet of di-
iodomethane was placed on the surface of a segment of hydrogel
submerged in water. The static contact angles were measured
with a Goiniometer from both sides of the droplet within 10 s
after depositing the drop, and the values were averaged. The
contact angle was found to be 129° and 142° for the prehydro-
lyzed pHEMA and posthydrolyzed pHEMA, respectively.
Mineralization of Hydrogels with the Urea-Mediated
Process. HA (2.95 g) was suspended into 200 mL of Milli-Q
water with stirring, and 2 M HCl was added sequentially until
all the HA suspension was dissolved at a final pH of 2.5-3.
Urea (24 g) was then dissolved into the solution to reach a
concentration of 2 M. Each hydrogel strip was then immersed
into 50 mL of the acidic stock HA solution containing urea.
The solution was slowly heated without stirring to 90-95 °C
5
0 µm thick. The favorable properties of the hydrogel-calcium
apatite composite obtained using the approach described here
may potentially maximize the chance for initiating in vivo
remodeling cascades and subsequent positive tissue-implant
integration. More in vitro and in vivo studies will have to be
performed before the real value of composites formed using
this approach can be estimated. In addition, the porosity of the
hydrogel scaffold has not yet been tuned, although many
available techniques (e.g., solvent casting/particulate leaching,
gas foaming and freeze-drying)2
9,30,32,34
can be applied to this
effort. Such modifications could further enhance the degree of
mineralization at the interior of the composite material and allow
deeper tissue ingrowth. These applications and extensions of
the method are presently under investigation.
Experimental Protocols
Synthesis. 2-Hydroxyethyl methacrylamide (HEMAm). To
(within 2 h) and maintained at that temperature overnight when
2
3
0 mL of ice-cold methanolic solution of ethanolamine (2 mL,
3 mmol) was slowly added 3.26 mL (34 mmol) of methacryloyl
necessary. The final pH was around 8.
Structural Characterizations. Mineralized hydrogel strips
were repeatedly washed in water to remove loosely attached
minerals and soluble ions before they were freeze-dried for
further structural analyses and mechanical characterization. The
surface microstructures and crystallinity of the materials grown
on the surface and inside of the hydrogel were analyzed by
scanning electron microscopy (SEM) with associated energy
dispersive spectroscopy (EDS) and X-ray powder diffraction
chloride (diluted in 20 mL of THF). Potassium hydroxide (1
M, aqueous) was added to maintain the solution pH at 8-9
throughout the reaction. The mixture was warmed to room
temperature over 2 h and stirred for another 2 h before it was
quenched by the addition of hydrochloric acid to a final pH of
5. The product was concentrated and redissolved in cold ethanol
to precipitate the potassium chloride salt. After silica gel flash
chromatography purification (chloroform: methanol/9:1), the
product (Rf 0.5) was isolated in 95% yield. H NMR (500 MHz,
CD3OD): δ 7.10 (1H, b), 5.55 (1H, s), 5.16 (1H, s), 3.51 (2H,
t, J ) 5.5 Hz), 3.25 (2H, q, J ) 5.0 Hz), 1.76 (3H, s);
(XRD).
1
SEM-EDS. All SEM micrographs of freeze-dried hydrogels
and hydrogel-mineral composites were obtained with a ISI-DS
3OC dual stage SEM with associated EDS. Samples were
1
3
C
1
NMR (125 MHz, CD3OD): δ 169.31, 139.06, 119.82, 60.71,
coated with either Au or Pt on a BAL-TEC, SCD 050 sputter
coater to achieve optimal imaging results, or coated with carbon
for EDS analysis. The imaging and analysis of composite
materials were performed at 15 kV, and those of hydrogels were
performed under reduced voltage (8-12 kV). The determination
of Ca/P ratios of all composite materials were based on
calibration using a standard synthetic HA sample.
+
+
4
1
1.96, 18.09; HRMS FAB (NBA): C6O2NH12 [M+H] , Calcd,
30.0868. Found 130.0868.
Hydrogel Preparation. 2-Hydroxyethyl methacrylate (HEMA)
was purchased from Aldrich and purified via distillation under
(62) de Groot, K.; Geesink, R.; Klein, C. J. Biomed. Mater. Res. 1987, 21, 1375-
1
381.
1242 J. AM. CHEM. SOC.
9
VOL. 125, NO. 5, 2003

3-D Bonelike Composites
A R T I C L E S
XRD. The presence and overall orientation of crystalline
phases in the precipitated mineral layers were evaluated by XRD
with a Siemens D500 instrument using Cu KR radiation.
Evaluation of Mineral-Hydrogel Interfacial Adhesion. To
evaluate the adherence of the mineral layers attached to pHEMA
hydrogels, the relative crack resistance was qualitatively evalu-
ated by indentation. The indentation test was performed using
a Vickers indentor (Micromet, Buehler, Ltd., USA) that has a
diamond pyramidal tip with a 136° angle between its faces.
Loads from 5 to 15 Newtons were applied for 20 s for each
measurement. After indentation, the samples where analyzed
by SEM to check for delamination. Lack of delamination is an
indication of strong adhesion between the mineral layer and
the hydrogel substrate.
Acknowledgment. We thank Dr. Catherine Klapperich,
Howard C. Hang, and Justin S. Cisar for critical reading of the
manuscript. This work was supported by the Laboratory Directed
Research and Development Program of Lawrence Berkeley
National Laboratory under the Department of Energy Contract
No. DE-AC03-76SF00098.
JA028559H
J. AM. CHEM. SOC.
9
VOL. 125, NO. 5, 2003 1243