Synthesis of highly porous crosslinked elastin hydrogels and their interaction
with fibroblasts in vitro
Nasim Annabia, Suzanne M. Mithieuxb, Elizabeth A. Boughtonc, Andrew J. Ruysc, Anthony S. Weissb,
aSchool of Chemical and Biomolecular Engineering, University of Sydney, Sydney, NSW 2006, Australia
bSchool of Molecular and Microbial Biosciences, University of Sydney, Sydney, NSW 2006, Australia
cSchool of Aerospace, Mechanical and Mechatronic Engineering, University of Sydney, Sydney, NSW 2006, Australia
a r t i c l e i n f o
Received 3 April 2009
Accepted 10 May 2009
Available online 4 June 2009
High pressure CO2
a b s t r a c t
In this study the feasibility of using high pressure CO2 to produce porous a-elastin hydrogels was
investigated. a-Elastin was chemically crosslinked with hexamethylene diisocyanate that can react with
various functional groups in elastin such as lysine, cysteine, and histidine. High pressure CO2substan-
tially affected the characteristics of the fabricated hydrogels. The pore size of the hydrogels was
enhanced 20-fold when the pressure was increased from 1 bar to 60 bar. The swelling ratio of the
samples fabricated by high pressure CO2was also higher than the gels produced under atmospheric
pressure. The compression modulus of a-elastin hydrogels was increased as the applied strain magnitude
was modified from 40% to 80%. The compression modulus of hydrogels produced under high pressure
CO2was 3-fold lower than the gels formed at atmospheric conditions due to the increased porosity of the
gels produced by high pressure CO2. The fabrication of large pores within the 3D structures of these
hydrogels substantially promoted cellular penetration and growth throughout the matrices. The highly
porous a-elastin hydrogel structures fabricated in this study have potential for applications in tissue
? 2009 Elsevier Ltd. All rights reserved.
Porous, three-dimensional (3D) scaffolds have been used
extensively as biomaterials in tissue engineering since they mayact
as an analogue of the extracellular matrix and provide a physical
support structure for cellular growth . Scaffold mechanical
properties and microstructure including porosity, mean pore size,
interconnectivity, and surface area may influence, sometimes
significantly, cell adhesion and infiltration [2–6]. Many natural
hydrogels containing collagen, hyaluronic acid (HA), fibrin, alginate,
agarose, dextran, chitosan, and elastin-like polypeptides (ELPs) are
attractive materials that have been used as scaffolds due to their
similarities with the extracellular matrix, excellent biological
performance, and inherent cellular interaction capabilities [7,8].
Elastin is an insoluble extracellular matrix protein that provides
tissues with the properties of elasticityand resilience. Due tothe
presence of crosslinks in elastin, the protein is highly insoluble and
therefore difficult to process into new biomaterials. Consequently,
soluble forms of elastin including tropoelastin , a-elastin
[11,12], and elastin-like polypeptides (ELPs)  are frequently used
to form crosslinked hydrogels.
Various crosslinking methods including chemical [10–24],
enzymatic [14,25], physical [26–28], and g-irradiation [29–31] have
been used to fabricate elastin-based hydrogels. The reactions may
be carried out in an aqueous or organic phase depending on the
type of crosslinker and desired properties of hydrogel. McMillan
et al.  showed that crosslinking of an ELP in an organic solvent,
where the ELP molecules exhibit no inverse phase transition,
resulted in the formation of hydrogels that were more uniform and
homogenous compared with those crosslinked in an aqueous
solution. This process commonly comprises two steps: coacerva-
tion and crosslinking . The dissimilarity observed between
these hydrogels fabricated in aqueous and organic solutions
presumably resulted from differences in interactions between the
protein and the solvent during the crosslinking reaction . In this
study homogenous and non-porous hydrogels were obtained when
the protein was chemically crosslinked in an organic solvent in the
absence of coacervation . In general, lack of cellular growth into
* Corresponding author. Tel.: þ61 2 9351 4794; fax: þ61 2 9351 2854.
E-mail address: email@example.com (F. Dehghani).
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/biomaterials
0142-9612/$ – see front matter ? 2009 Elsevier Ltd. All rights reserved.
Biomaterials 30 (2009) 4550–4557
the 3D structures of ELP hydrogels due to the presence of small
pores or porosity gradient is an issue associated with the current
methods used for ELP hydrogel formation.
Dense gas carbon dioxide (CO2) has been used widely as a gas
foaming agent to induce porosity in the structure of amorphous or
semi-crystalline hydrophobic polymers such as poly(lactic acid)
caprolactone (PCL) [32–36]. A dense gas is a fluid at above or close
to its critical temperature and pressure with the properties inter-
mediate to those of gases and liquids. Dense gas CO2with a low
critical temperature (Tc: 31?C) is attractive for biomaterial pro-
cessing because it is inert, non-toxic, and non-flammable [32,33].
Porous hydrogels have been fabricated using supercritical CO2–
wateremulsion templated techniques [37–39] and also using dense
gas CO2as a medium for crosslinking reactions [7,12,40,41]. The
supercritical CO2–water emulsion templated methodology has
been successfully employed to produce highly porous crosslinked
structures of different polymers such as dextran , chitosan ,
alginate , polyvinyl alcohol (PVA), and blended PVA/poly
ethylene glycol (PEG) . A 3D structure of highly interconnected
and thin-walled porous dextran hydrogel with the average pore
size less than 26 mm was formed using a surfactant to stabilise the
CO2–water emulsion . Lee et al. synthesised a biodegradable
and inexpensive poly(vinyl acetate)-based surfactant that can be
used in this process to promote the product biocompatibility .
Further research may be required to enhance the pore sizes of these
hydrogels for tissue engineering applications.
We have shown previously that highly porous a-elastin hydro-
gels can be formed by crosslinking a-elastin with glutaraldehyde
(GA) in an aqueous solution using high pressure CO2. These
hydrogels were formed rapidly, in less than an hour, with pore
structures resembling natural elastin . Fibroblast cells prolif-
erated in the 3D structures as a result of CO2induced channels
within the structure of the a-elastin hydrogels . The low
mechanical properties of the hydrogels were attributed to the low
number of lysine residues (less than 1%) in a-elastin that were
available for crosslinking with GA.
The objective of this study was to fabricate a-elastin hydrogels
with increased porosity and mechanical properties using hexam-
ethylene diisocyanate (HMDI) as a crosslinking agent. Hexam-
ethylene diisocyanate is a bifunctional molecule with lower
cytotoxicity than GA . In general, isocyanates may react with
nucleophilic functional groups such as amines, alcohols, and
protonated acids. Hexamethylene diisocyanate, in particular, reacts
with the side chains of lysine, cysteine, and histidine, and to a lesser
amino acids available in the a-elastin structure to increase the
crosslinking density and mechanical properties of fabricated
hydrogels. Dimethyl sulfoxide (DMSO) was used as a solvent to
fabricate HMDI crosslinked hydrogels due to the low solubility of
HMDI in aqueous solution. Our research objective was to create an
enhanced 3D pore network throughout a-elastin hydrogels, during
the crosslinking process, by dissolving high pressure CO2 into
a DMSO solution, where the dissolved CO2would be released on
depressurisation. The pore morphology, swelling ratio, and
were compared with those produced at atmospheric conditions. In
vitro studies were also conducted to assess the cellular growth and
proliferation in the 3D structures of fabricated hydrogels.
acid)(PLGA), and poly-
2. Materials and methods
a-Elastin extractedfrombovine ligament was obtained fromElastin Products Co.
(Missouri, USA). Dimethyl sulfoxide (DMSO) and hexamethylene diisocyanate
(HMDI) were purchased from Sigma. Food grade carbon dioxide (99.99% purity) was
supplied by BOC. GM3348 fibroblast cell line was obtained from the Coriell Cell
Repository. Cells were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM)
supplemented with 10% v/v fetal bovine serum (FBS), penicillin and streptomycin.
All tissue culture reagents were obtained from Sigma.
2.2. Hydrogel formation
2.2.1. Hydrogel fabrication at atmospheric pressure
In each experiment 100 mg/mla-elastin in DMSO was mixed with HMDI and the
solution was immediately pipetted into a glass Lab-Tek chamber slide. The sample
was allowed to react for 18 h at room temperature under nitrogen. The crosslinked
hydrogels were then swelled in MilliQ water and gently agitated on a shaker for 2 h
to remove residual DMSO. The media were exchanged with fresh MilliQ water every
15 min with shaking. The hydrogels were then stored in PBS.
A preliminary set of experiments were conducted to determine the required
amount of HMDI for the hydrogel fabrication. a-Elastin solutions at 100 mg/ml were
mixed with various concentrations of HMDI ranging between 0.25% (v/v) and 10%
(v/v). The solutions were then pipetted into Lab-Tek chamber slides and placed in
a chamber connected to a nitrogen line for 18 h at room temperature (25?C). The
results demonstrated that hydrogels were formed at HMDI concentrations above 1%
(v/v). At low concentrations of HMDI, dissolved a-elastin was partially crosslinked,
while at high concentrations of HMDI (10%) the hydrogel was very rigid and non-
elastic. Consequently, in this study, 2% (v/v) and 5% (v/v) HMDI were used toproduce
2.2.2. Dense gas hydrogel formation
The experimental apparatus used to fabricate a-elastin hydrogels using dense
gas CO2involved a high-pressure vessel coupled with a CO2source and a high-
pressure pump. This apparatus has been previously described . An a-elastin
solution containing HMDI was injected into a custom-made Teflon mould placed
inside the high-pressure vessel. After the vessel was sealed and approached thermal
equilibrium at 25?C, the system was pressurised with CO2to 60 bar, isolated and
maintained at these conditions for a set period of time. The system was then dep-
ressurised and samples were collected. Then the crosslinked structures were
swelled in MilliQ water and gentlyagitated on a shaker for 2 h to remove the residue
of DMSO. The hydrogels were stored in PBS for further analysis.
2.3. Scanning electron microscopy (SEM)
The SEM images of samples were obtained using a Philips XL30 scanning elec-
tron microscope (15 kV) to determine the pore characteristics of the fabricated
hydrogels and to examine cellular infiltration and adhesion. Lyophilised a-elastin
hydrogels were mounted on aluminium stubs using conductive carbon paint, then
gold coated prior to SEM analysis.
Cell-seeded hydrogels were fixed with 2% (v/v) GA in 0.1 M Na-cacodylate buffer
with 0.1 M sucrose for 1 h at 37?C. Samples underwent post-fixation with 1%
osmium in 0.1 M Na-cacodylate for 1 h and were then dehydrated in ethanol solu-
tions at 70%, 80%, 90% and 3 times 100% for 10 min each. For drying, the samples
were immersed for 3 min in 100% hexamethyldisilazane (HMDS) then transferred to
a desiccator for 25 min to avoid water contamination. Finally they were mounted on
stubs and sputter coated with 10 nm gold.
2.4. Mechanical characterisation: compressive properties
Uniaxial compression tests were performed in an unconfined state using a Bose
ELF3400 mechanical tester with a 50 N load cell. The testing procedure was done in
accordance with previously documented hydrogel mechanical testing reports [43–
45]. Prior to mechanical testing, the hydrogels produced at high pressure CO2and
atmospheric condition were swelled for 2 h in PBS. The thickness (3 ?0.1 mm) and
diameter (12.5 ?0.7 mm) of each sample were measured using a digital calliper
prior to mechanical testing. The compressive properties of the samples were tested
in the hydrated state, in PBS, at room temperature. Compression (mm) and load (N)
were recorded using Wintest software at a cross speed of 30 mm/s and different
strain levels ranging from 40% to 80%. At each strain level, the samples were cycli-
cally preconditioned for 7 cycles to minimize artefact interference. The hydrogels
were subsequently subjected to another loading and unloading cycle (8th cycle)
where compression (mm) and load (N) were collected. At each strain level, the
compressive modulus for the 8th cycle was obtained as the tangent slope of the
stress–strain curve. In addition, at each strain level, the energy loss based on the 8th
compression cycle was computed. Three specimens were tested for each sample
type (sample produced at high pressure CO2or atmospheric pressure) at each strain
2.5. Swelling property
The swelling behaviours of the HMDI crosslinked hydrogels produced at high
pressure CO2 and atmospheric conditions were evaluated at room temperature
(25?C) in three different solvent environments: MilliQ water, phosphate-buffered
N. Annabi et al. / Biomaterials 30 (2009) 4550–45574551
dry. The samples were then swelled in 10 ml of solvent (MilliQ water, PBS or DMSO)
for 24 h. For each solvent, at least three samples were placed in the media overnight.
The excess liquid was then removed from the swelled samples and the swelling ratio
was calculated based on the ratio of the increase in mass to that of the dry sample.
2.6. In vitro cell culture
The ability of human skin fibroblast cells (GM3348) to grow into the 3D hydrogel
structure was assessed. Following crosslinking, hydrogels were transferred into
a 48-well plate and washed twice with ethanol to sterilise the materials. The
hydrogels were then washed at least twice with culture media to remove any
residual ethanol and equilibrated in culture media (DMEM, 10% FBS, pen-strep) at
37?C overnight. The cells were then seeded ontothe hydrogels at 1.6?105cells/well
and compared with unseeded hydrogel in an adjacent well. Cells were cultured in
a CO2incubator for 3 days at 37?C, after which the hydrogels were fixed to assess
cell proliferation and infiltration using light microscopy and SEM analysis.
2.7. Light microscopy analysis on histological samples
The growth of the cells in fabricated hydrogels was confirmed using light
microscopy analysis after fixing, sectioning, and staining cross-sections of cell-
seeded scaffolds. The hydrogels containing cells were fixed by soaking in 10%
formalin overnight. The scaffolds were then immersed in 70% ethanol. The samples
were processed on an automated tissue processor on a 6 h cycle to paraffin through
a graded series of ethanol, and xylene. They were embedded in paraffin wax and
5 mm sections were taken and collected onto glass slides and dried. The slides were
then deparaffinised, rehydrated, stained using a standard haematoxylin and eosin
staining procedure, dehydrated, cleared in xylene, and mounted in DPX. The cross-
sections were examined using a light microscope (Olympus DP50) connected to the
3. Results and discussion
Inthis studythefeasibilityof usinghighpressureCO2toproduce
HMDI crosslinked a-elastin hydrogels with increased mechanical
properties and pore sizes was assessed. The effects of reaction time
and crosslinker concentration on the characteristics of the hydrogel
produced at high pressure CO2 were investigated. Preliminary
experiments demonstrated that when 2% (v/v) HMDI was used, the
hydrogels were not formed at reaction times below 2 h. A partially
crosslinked film of a-elastin was observed in the mould when the
system was depressurised after 1 h. However, when the crosslinker
concentration was increased to 5% (v/v), a hydrogel with desirable
pore size and elasticity was obtained within 1 h. As the reaction
time was further increased from 1 to 2 h, the 5% (v/v) HMDI
crosslinked hydrogel became rigid and less elastic.
3.1. Pore structure of the a-elastin hydrogel
The structures of hydrogels including porosity, pore sizes, and
interconnectivity have a substantial influence on penetration,
adhesion, and growth of the cells within the 3D structures. The
macrostructures of 2% (v/v) HMDI crosslinked a-elastin hydrogels
produced at high pressure CO2 and atmospheric conditions in
lyophilised and hydrated states are shown in Fig. 1. Rigid a-
elastin hydrogels were formed at both high pressure CO2 and
atmospheric conditions as shown in Fig. 1. The hydrogels were
easily handled and kept their macrostructures after swelling. This
was due to the high degree of crosslinking within the structures
of HMDI crosslinked hydrogels. As shown in Fig. 1, increased
porosity in the hydrogels fabricated at high pressure was clearly
In this study, SEM analysis was used to characterise the pore
morphologyof fabricateda-elastin hydrogels. As shown in Fig. 2c–f,
the hydrogels fabricated at 60 bar CO2pressure were highly porous.
Comparison of SEM images of a-elastin hydrogels produced under
high pressure CO2and atmospheric conditions indicated that high
pressure CO2increased the pore size of the fabricated hydrogels as
shown in Fig. 2. Equivalent circle diameter (ECD) of the pores was
calculated using Image J software. The average pore size of hydro-
gels fabricated by using 2% (v/v) HMDI increased from 3.9?0.8 mm
to 79.8?54.8 mm when pressure was increased from 1 bar to
60 bar. In hydrogels fabricated by the dense gas CO2, 52.6% of the
pores were above 80 mm in diameter which makes these hydrogels
suitable for cellular growth through the 3D structures. The pres-
ence of large channels in the cross-sections of samples fabricated at
high pressure CO2(Fig. 2d) could facilitate cellular penetration and
proliferation into the 3D structures. These large channels were not
present in hydrogels fabricated at atmospheric conditions as shown
in Fig. 2b. The generation of large channels is a unique feature of
dense gas CO2 that facilitated the penetration and growth of
fibroblast cells in 3D structures of GA crosslinked a-elastin hydro-
gels reported in our previous study .
The formation of large pores in the 3D structures of HMDI
crosslinked hydrogels produced at high pressure CO2may be due
to the high solubility of CO2in DMSO at operating temperature
(25?C) and pressure (60 bar). Phase behaviour studies on binary
mixtures of CO2 and DMSO indicated that solubility of CO2 in
DMSO is a function of temperature and pressure . Kordikowski
et al. reported that increasing the temperature decreased the CO2
solubility in DMSO. The solubility of CO2 in DMSO was also
enhanced by increasing the pressure at constant temperature .
The molar fractions of CO2increased from 0.1 to 0.9 when pres-
sure raised from 10 bar to 60 bar at 25?C . The extent of
miscibility is reflected in the volume expansions of the liquid
phase; no further expansion takes place with the occurrence of
liquid–liquid immiscibility. The volume expansions for the binary
mixture of CO2and DMSO increased from 10% to 820% at 25?C
when pressure was increased from 10 bar to 60 bar . However,
using a higher temperature (30?C) the volume expansions of the
mixture increased from 6.5% to 180% when pressure was
enhanced from 10 bar to 60 bar . Consequently, in this study
the operating conditions of 25?C and 60 bar were used to achieve
higher solubility of CO2 in DMSO. a-Elastin solution in DMSO
could be expanded by CO2at 60 bar and 25?C as the crosslinking
Fig. 1. HMDI crosslinked a-elastin hydrogel produced at 60 bar CO2(a), and atmospheric conditions (b) before and after hydration in water at 25?C.
N. Annabi et al. / Biomaterials 30 (2009) 4550–4557 4552
reaction and gelation took place. The presence of the crosslinker
in the a-elastin solution is likely to limit the expansion of a-
elastin solution by CO2.
Using 5% (v/v) HMDI, the number of the large pores on the top
and also in the cross-section of the sample was reduced as
compared to the 2% (v/v) HMDI crosslinked hydrogels (Fig. 2 e and
f). However, using a higher concentration of HMDI (5% (v/v)), the
hydrogels were more rigid than 2% (v/v) HMDI crosslinked hydro-
gels. Increasing the crosslinker concentration resulted in an
increase in the degree of crosslinking through the hydrogel
matrices, which limited the ability of the a-elastin solution to
expand in CO2and led to the formation of smaller pores.
3.2. Mechanical characterisation: compressive properties
The compressive mechanical properties of HMDI crosslinked
samples produced at high pressure CO2and atmospheric condi-
tions are shown in Fig. 3. The compression modulus of both
samples, produced at high pressure and atmospheric conditions,
increased as the applied strain magnitude was increased. The
compression modulus of the sample produced at atmospheric
conditions ranged from 11.25?0.4 kPa to 18.8?3.4 kPa at 40% and
80% strain, respectively (Fig. 3c). The stress-strain curve became
nonlinear at strain levels above 40% as shown in Fig. 3b. This
indicated that plastic deformation occurred in the hydrogels
produced at atmospheric conditions at strains greater than 40%.
However, for the samples fabricated at high pressure CO2, the
stress–strain curve was still linear at 60% strain (Fig. 3a). Therefore,
the samples produced at high pressure CO2were expected to be
more elastic than hydrogels produced at atmospheric conditions.
The compression modulus of the HMDI crosslinked hydrogel
produced by high pressure CO2 ranged from 3.99?0.5 kPa to
8.62?1.7 kPa at 40% and 80% strain, respectively (Fig. 3c). The
compression modulus of the hydrogels produced at high pressure
CO2was generallylower than the samples produced at atmospheric
conditions. This means that the samples produced at atmospheric
pressure were stiffer than those produced using high pressure CO2.
This was expected due to the increased porosity of the samples
fabricated at high pressure CO2.
The compression properties of the fabricated materials were
comparable to those reported in the literature for a number of
biomaterials. Srokowski et al. found that the compression modulus
of a non-porous ELP gel was between 6.26 kPa and 215 kPa over
a strain range of 20–80% . However, for a porous foam-like ELP
Fig. 2. SEM images of a-elastin hydrogels fabricated at atmospheric pressure using 2% HMDI (a, b), 60 bar CO2pressure using 2% HMDI (c, d), 60 bar CO2pressure using 5% HMDI
N. Annabi et al. / Biomaterials 30 (2009) 4550–4557 4553
the compression modulus was reported between 4.48 kPa and
146.50 kPa . The compression strength of a flexible poly-
urethane foam increased from 2.4 kPa to 11.7 kPa for strain levels
between 20% and 60% as reported by Lutter et al. . Thomas et al.
fabricated a copolymer gel of polyvinyl alcohol (PVA)/polyvinyl
pyrrolidine (PVP) for endoscopic replacement of the nucleus pul-
posus of a lumbar intervertebral disk . The compression
modulus of fabricated PVA/PVP was reported between 4 kPa and
5.5 kPa for 15% and 25% in strain level . The compression
modulus of a poly(ethylene glycol) diacrylate (PEGDA) used for
lamina propria regeneration ranged between w30 kPa and 100 kPa
depending on the PEGDA concentration . The compression
modulus of porousnatural–synthetic
produced by impregnation monomers such as butylmethacrylate
(BMA) into porous alginate using emulsion templating and super-
39.9?4.2 kPa by Partap et al. . The mean compressive modulus
of the nucleus pulposus of the lumbar intervertebral disc was
reported to be 5.4 kPa  which is in the range of the compressive
modulus of fabricated a-elastin hydrogel using high pressure CO2.
As the intervertebral disc consists of approximately 1.7% elastin
, the HMDI crosslinked hydrogels at high pressure CO2offer
unique promise for an orthopaedic application.
High resilience of elastin provides the ability to deform revers-
ibly without loss of energy . Energy loss is proportional to
hysteresis. Using 60% strain level, the energy loss for the hydrogels
fabricated at high pressure CO2was 6.03?1.14%. However, for the
gels produced at atmospheric conditions the energy loss was
11.72?4.02%, demonstrating greater hysteresis for the samples
produced at atmospheric pressure. As shown in Fig. 3d, the energy
loss of both hydrogels fabricated at high pressure CO2and atmo-
spheric conditions increased with increasing strain levels. The
energy loss for samples produced at high pressure CO2increased
from 1.43?0.86% to 13.16?1.93%, when the strain level increased
from 40% to 80% as shown in Fig. 3d. For the hydrogels produced at
14.67?3.31% when the strain level increased from 40% to 80%.
The energy loss of the fabricated a-elastin hydrogel was either
lower or comparable with those reported for fabricated ELPs using
various crosslinkers. Veith et al. found that the energy loss for
a genipin crosslinked elastin-based recombinant polypeptide was
9.7?6.2% . However, using pyrroloquinoline quinone (PQQ) as
a crosslinking agent, energy loss increased to 20.4 ?1.6% indicating
greater hysteresis in PQQ crosslinked sheets . Srokowski et al.
also reported a relatively constant energy loss of 51.3 ?10.1% at
different strain levels for an ELP . The energy loss of aorta
elastin purified from porcine tissue was reported to be 23?2% by
Bellingham et al. and for two different engineered ELPs 20?7% and
3.3. Swelling properties
The swelling behaviour of the fabricated hydrogels is shown in
Fig. 4. Generally, the swelling ratio of the samples exposed to high
pressure CO2was higher than the hydrogels fabricated at atmo-
spheric condition as indicated in Fig. 4. Both hydrogels produced at
high pressure CO2 and atmospheric pressure swelled more in
DMSO than in water and PBS. Hydrogels produced at 60 bar CO2
40% 60% 80%
Strain level (mm)
Strain level (mm)
Energy Loss %
Fig. 3. Unconfined compressive behaviour of HMDI crosslinked hydrogel. Cyclic stress–strain data for the sample produced at high pressure CO2(a), and atmospheric condition (b).
Compressive modulus (c) and energy loss (d) at each strain level for both hydrogels produced at high pressure CO2and atmospheric conditions.
N. Annabi et al. / Biomaterials 30 (2009) 4550–45574554
pressure absorbed 6.81?0.46, 10.9?2.46, and 18.65?0.88 g
liquid/g proteinwhen they were hydrated in PBS, water, and DMSO,
respectively. However, the gels formed at atmospheric pressure
absorbed 4.79?0.15, 9.45?0.25, and 9.82?1.97 g liquid/g protein
when they were swelled in PBS, water, and DMSO, respectively. The
exact mechanism of an increase in the swelling ratio of HMDI
crosslinked hydrogels in DMSO is unknown . DMSO can
destabilize the secondary structure of proteins and peptides by
interacting with the polypeptide backbone . Lillie and Gosline
reported that higher swelling in DMSO may be attributed to an
increase in the molecular mobility of the elastin network through
plasticisation . In the DMSO system the a-elastin hydrogel may
be plasticised by DMSO, resulting in an increase in segment length
between crosslinks. The higher swelling ratio of the sample
produced at high pressure CO2was due to the presence of larger
pores through their structures compared to the hydrogel formed at
The swelling ratio of HMDI crosslinked a-elastin hydrogels
produced at high pressure CO2was either considerably greater or
comparable with other crosslinked ELP hydrogels using diisocya-
nate crosslinker [21,43]. The swelling behaviour of a-elastin
hydrogels produced at high pressure CO2was comparable with
a lysine diisocyanate (LDI) crosslinked ELPs with the swelling
capacity of approximately 6 g liquid/g protein in both PBS and
water . However, the swelling ratio of gels produced at high
pressure CO2was 10-fold higher than an HMDI crosslinked ELPs
reported by Nowatzki et al. (0.37 g H2O/g protein at 4?C) . This
may be due to the increased porosity within the structures of
fabricated hydrogels or lower degree of crosslinking.
The swelling ratio of HMDI crosslinked a-elastin using 60 bar
CO2pressure was lower than the GA crosslinked a-elastin hydrogel
fabricated at 60 bar CO2, which was 33.2?0.8 g H2O/g protein as
reported in our previous study . This may due to the higher
degree of crosslinking through the structures of HMDI crosslinked
a-elastin hydrogel as the swelling ratio of hydrogels is correlated to
the degree of crosslinking. Generally, the hydrogels with a high
degree of crosslinking exhibit low swelling ratio.
The swelling behaviour of the HMDI crosslinked hydrogel was
correlated with the compressive modulus. In general, the samples
with greater swelling have lower compressive moduli as indicated
in Table 1. This phenomenon has also been reported by Srokowski
et al., where a higher compressive modulus corresponded to lower
level of swelling for LDI crosslinked hydrogels . Vieth et al. also
found that the elastic modulus of genipin crosslinked ELPs was
irreversibly correlated with the swelling ratio .
3.4. In vitro fibroblast cell proliferation using elastin hydrogels
Cellular growth and proliferation in a-elastin hydrogels were
examined by light microscopy and SEM analysis to demonstrate
the feasibility of using the processed material for soft tissue
engineering applications. The light microscopy images of adherent
fibroblast cells cultured on hydrogel produced at 60 bar CO2are
shown in Fig. 5. Haematoxylin and eosin were used for staining, as
a result the cells appear as dark grey and the a-elastin scaffolds as
light grey. As shown in Fig. 5b and c fibroblast cells were able to
grow into the 3D structures of a-elastin due to the presence of
large pores induced by high pressure CO2. However, cells were
only able to form a monolayer on the surface of the hydrogel
fabricated at atmospheric conditions as indicated in Fig. 5a, due to
the presence of small pores on the surface. The SEM images in
Fig. 6 corroborated the cell proliferation into the 3D structure of
hydrogel fabricated at high pressure CO2. As shown in Fig. 6, cells
were able to colonise at the top surface (Fig. 6b and c) and also
into the 3D structure (Fig. 6d–f) of a-elastin hydrogels produced
under high pressure CO2due to the presence of large pores within
Fig. 5. Images of fibroblast cells cultured on hydrogel produced at atmospheric pressure (a) and high pressure CO2(b and c).
Swelling ratio and compressive modulus of hydrogels fabricated at high pressure
CO2and atmospheric conditions.
Crosslinked hydrogelSwelling ratio
(g liquid/g protein)
aApplied strain magnitude: 40%.
bApplied strain magnitude: 80%.
Swelling Ratio (g liquid / g protein)
Fig. 4. Swelling behaviour of hydrogels produced at high pressure CO2and atmo-
spheric conditions in PBS, water, and DMSO.
N. Annabi et al. / Biomaterials 30 (2009) 4550–45574555
This study demonstrated the feasibility of fabricating elastin
hydrogels with enhanced mechanical properties and pore sizes
using dense gas CO2 and HMDI crosslinker. High pressure CO2
significantly increased the pore sizes of fabricated hydrogels due to
the high solubility and diffusion of CO2in DMSO. Hydrogels fabri-
cated using dense gas CO2were more elastic compared with those
fabricated at atmospheric pressure. The compressive modulus of
the sample produced at high pressure CO2was lower than the gel
fabricated at atmospheric conditions due to the presence of larger
pores within the structures of crosslinked material produced under
high pressure CO2. The fabrication of these large pores within the
3D structures substantially promoted fibroblast infiltration and
growth throughout the matrices. The fabricated a-elastin hydrogels
display promising characteristics that may be suitable for soft
tissue repair and orthopaedic applications such as spinal nucleus
The authors acknowledge financial support from Sydnovate, the
Australian Research Council (AW), and Merck Pty Ltd.
Figures with essential colour discrimination. Fig. 5 in this article
may be difficult to interpret in black and white. The full colour
images can be found in the on-line version, at doi:10.1016/j.
 Liu C, Xia Z, Czernuszka JT. Design and development of three-dimen-
sional scaffolds for tissue engineering. Chem Eng Res Des 2007;85(A7):
 Harley BA, Leung JH, Silva ECCM, Gibson LJ. Mechanical characterization
 O’Brien FJ, Harley BA, Yannas IV, Gibson LJ. The effect of pore size on cell
adhesion in collagen–GAG scaffolds. Biomaterials 2005;26(4):433–41.
 Zeltinger J, Sherwood JK, Graham DA, Mueller R, Griffith LG. Effect of pore size
and void fraction on cellular adhesion, proliferation, and matrix deposition.
Tissue Eng 2001;7(5):557–72.
 Nehrer S, Breinan HA, Ramappa A, Young G, Shortkroff S, Louie LK, et al. Matrix
collagen type and pore size influence behavior of seeded canine chondrocytes.
 Wake MC, Patrick Jr CW, Mikos AG. Pore morphology effects on the fibro-
vascular tissue growth in porous polymer substrates. Cell Transplant
 Temtem M, Casimiro T, Mano JF, Aguiar-Ricardo A. Green synthesis of
a temperature sensitive hydrogel. Green Chem 2007;9(1):75–9.
Fig. 6. SEM images of fibroblast cells attached to a-elastin hydrogel fabricated at atmospheric conditions (a), and high pressure CO2(b–f). Top surfaces (a–c), internal surfaces of
hydrogel obtained by cross sectioning the sample (d–f).
N. Annabi et al. / Biomaterials 30 (2009) 4550–45574556
 Malafaya PB, Silva GA, Reis RL. Natural-origin polymers as carriers and scaf- Download full-text
folds for biomolecules and cell delivery in tissue engineering applications. Adv
Drug Deliv Rev 2007;59(4–5):207–33.
 Mithieux SM, Weiss AS. Elastin. Adv Protein Chem 2005;70:437–61.
 Mithieux SM, Rasko JEJ, Weiss AS. Synthetic elastin hydrogels derived from
massive elastic assemblies of self-organized human protein monomers.
 Leach JB, Wolinsky JB, Stone PJ, Wong JY. Crosslinked alpha-elastin biomate-
rials: towards a processable elastin
 Annabi N, Mithieux SM, Weiss AS, Dehghani F. The fabrication of elastin-based
hydrogels using high pressure CO2. Biomaterials 2009;30(1):1–7.
 Lim DW, Nettles DL, Setton LA, Chilkoti A. In situ cross-linking of elastin-like
 Mithieux SM. Synthetic elastin: construction and properties of cross-linked
human tropoelastin. Ph.D. dissertation. Sydney, NSW: University of Sydney;
 Welsh ER, Tirrell DA. Engineering the extracellular matrix: a novel approach to
polymeric biomaterials. I. Control of the physical properties of artificial protein
matrices designed to support adhesion of vascular endothelial cells. Bio-
 Martino M, Tamburro AM. Chemical synthesis of cross-linked poly(KGGVG), an
elastin-like biopolymer. Biopolymers 2001;59(1):29–37.
 McMillan RA, Caran KL, Apkarian RP, Conticello VP. High-resolution topo-
graphic imaging of environmentally responsive, elastin-mimetic hydrogels.
for control of cell and tissue behavior. Macromolecules 2003;36(5):1553–8.
 McMillan RA, Conticello VP. Synthesis and characterization of elastin–mimetic
protein gels derived from a well-defined polypeptide precursor. Macromole-
 Bellingham CM, Lillie MA, Gosline JM, Wright GM, Starcher BC, Bailey AJ, et al.
Recombinant human elastin polypeptides self-assemble into biomaterials
with elastin-like properties. Biopolymers 2003;70(4):445–55.
 Nowatzki PJ, Tirrell DA. Physical properties of artificial extracellular matrix
protein films prepared by isocyanate crosslinking. Biomaterials 2003;
 Trabbic-Carlson K, Setton LA, Chilkoti A. Swelling and mechanical behaviors of
 Lim DW, Nettles DL, Setton LA, Chilkoti A. Rapid cross-linking of elastin-like
polypeptides with (hydroxymethyl)phosphines in aqueous solution. Bio-
 Vieth S, Bellingham CM, Keeley FW, Hodge SM, Rousseau D. Microstructural
and tensile properties of elastin-based polypeptides crosslinked with genipin
and pyrroloquinoline quinone. Biopolymers 2007;85(3):199–206.
 McHale Melissa K, Setton Lori A, Chilkoti A. Synthesis and in vitro evaluation
of enzymatically cross-linked elastin-like polypeptide gels for cartilaginous
tissue repair. Tissue Eng 2005;11(11–12):1768–79.
 Mithieux Suzanne M, Tu Y, Korkmaz E, Braet F, Weiss Anthony S. In situ
polymerization of tropoelastin in the absence of chemical cross-linking.
 Nagapudi K, Brinkman WT, Thomas BS, Park JO, Srinivasarao M, Wright E, et al.
Viscoelastic and mechanical behavior of recombinant protein elastomers.
 Nagapudi K, Brinkman WT, Leisen J, Thomas BS, Wright ER, Haller C, et al.
Protein-based thermoplastic elastomers. Macromolecules 2005;38(2):345–54.
 Lee J, Macosko CW, Urry DW. Phase transition and elasticity of protein-based
hydrogels. J Biomater Sci Polym Ed 2001;12(2):229–42.
 Lee J, Macosko CW, Urry DW. Mechanical properties of cross-linked synthetic
elastomeric polypentapeptides. Macromolecules 2001;34(17):5968–74.
 Lee J, Macosko CW, Urry DW. Swelling behavior of g-irradiation cross-linked
 Quirk RA, France RM, Shakesheff KM, Howdle SM. Supercritical fluid tech-
nologies and tissue engineering scaffolds. Curr Opin Solid State Mater Sci
mimetic scaffold. Acta Biomater
 Tai H, Popov VK, Shakesheff KM, Howdle SM. Putting the fizz into chemistry:
applications of supercritical carbon dioxide in tissue engineering, drug
delivery and synthesis of novel block copolymers. Biochem Soc Trans
 Barry JJA, Silva MMCG, Popov VK, Shakesheff KM, Howdle SM. Supercritical
carbon dioxide: putting the fizz into biomaterials. Philos Trans R Soc London
Ser A 2006;364(1838):249–61.
 Cansell F, Aymonier C, Loppinet-Serani A. Review on materials science
and supercritical fluids. Curr Opin Solid State Mater Sci 2003;7(4–5):
 Barry JJA, Gidda HS, Scotchford CA, Howdle SM. Porous methacrylate scaf-
folds: supercritical fluid fabrication and in vitro chondrocyte responses.
 Partap S, Rehman I, Jones JR, Darr JA. Supercritical carbon dioxide in water
emulsion-templated synthesis of porous calcium alginate hydrogels. Adv
 Palocci C, Barbetta A, La Grotta A, Dentini M. Porous biomaterials obtained
using supercritical CO2–water emulsions. Langmuir 2007;23(15):8243–51.
 Lee J-Y, Tan B, Cooper AI. CO2-in-water emulsion-templated poly(vinyl
alcohol) hydrogels using poly(vinyl acetate)-based surfactants. Macromole-
 Cooper AI, Holmes AB. Synthesis of molded monolithic porous polymers using
supercritical carbon dioxide as the porogenic solvent. Adv Mater 1999;11(15):
 Cooper AI, Wood CD, Holmes AB. Synthesis of well-defined macroporous
polymer monoliths by sol–gel polymerization in supercritical CO2. Ind Eng
Chem Res 2000;39(12):4741–4.
 Van Luyn MJA, Van Wachem PB, Damink LO, Dijkstra PJ, Feijen J,
Nieuwenhuis P. Relations between in vitro cytotoxicity and crosslinked dermal
sheep collagens. J Biomed Mater Res 1992;26(8):1091–110.
 Srokowski EM, Woodhouse KA. Development and characterisation of novel
cross-linked bio-elastomeric materials. J Biomater Sci Polym Ed 2008;19(6):
 Stammen JA, Williams S, Ku DN, Guldberg RE. Mechanical properties of
a novel PVA hydrogel in shear and unconfined compression. Biomaterials
 Joshi A, Fussell G, Thomas J, Hsuan A, Lowman A, Karduna A, et al. Functional
compressive mechanics of a PVA/PVP nucleus pulposus replacement. Bioma-
 Kordikowski A, Schenk AP, Van Nielen RM, Peters CJ. Volume expansions and
vapor–liquid equilibria of binary mixtures of a variety of polar solvents and
certain near-critical solvents. J Supercrit Fluids 1995;8(3):205–16.
 Lutter HD, Hinz W, Decker W, Leppkes R, Reich E, Peters R, et al., inventors.
Block polyoxyethylene–polyoxypropylene polyols for the production of flex-
ible polyurethane foams with low compression hardness. US Patent No.
 Thomas J, Gomes K, Lowman A, Marcolongo M. The effect of dehydration
history on PVA/PVP hydrogels for nucleus pulposus replacement. J Biomed
Mater Res B Appl Biomater 2004;69(2):135–40.
 Liao H, Munoz-Pinto D, Qu X, Hou Y, Grunlan Melissa A, Hahn Mariah S.
Influence of hydrogel mechanical properties and mesh size on vocal fold
fibroblast extracellular matrix production and phenotype. Acta Biomater
 Partap S, Hebb AK, Rehman I, Darr JA. Formation of porous natural–synthetic
polymer composites using emulsion templating and supercritical fluid assis-
ted impregnation. Polym Bull 2007;58(5–6):849–60.
 Umehara S, Tadano S, Abumi K, Katagiri K, Kaneda K, Ukai T. Effects of
degeneration on the elastic modulus distribution in the lumbar intervertebral
disc. Spine 1996;21(7):811–9. discussion 820.
 Mikawa Y, Hamagami H, Shikata J, Yamamuro T. Elastin in the human
intervertebral disk: a histological and biochemical study comparing it with
elastin in the human yellow ligament. Arch Orthop Trauma Surg 1986;105(6):
 Gosline J, Lillie M, Carrington E, Guerette P, Ortlepp C, Savage K. Elastic
proteins: biological roles and mechanical properties. Philos Trans R Soc Lond B
Biol Sci 2002;357(1418):121–32.
 Lillie MA, Gosline JM. Swelling and viscoelastic properties of osmotically
stressed elastin. Biopolymers 1996;39(5):641–52.
N. Annabi et al. / Biomaterials 30 (2009) 4550–45574557