Electrospun poly(l -lactide-co -acryloyl carbonate) fiber scaffolds with a mechanically stable crimp structure for ligament tissue engineering

Article abstract of DOI:10.1021/bm401813j

The aim of this study was to prepare a fibrous scaffold that possesses a crimped morphology using a photo-cross-linkable biodegradable copolymer. To obtain the crimped morphology, the polymer was first electrospun onto a rotating wire mandrel to obtain aligned straight fibers. Postprocessing by immersion in aqueous buffer at 37 °C generated a crimplike pattern in the fibers. It was reasoned that cross-linking the fibers following formation of the crimped structure would endow the scaffolds with a recoverable crimp pattern and mechanical properties similar to that of the collagen fibers in the anterior cruciate ligament (ACL). To achieve this aim, a trimethylene carbonate based monomer bearing an acrylate pendant group was synthesized and copolymerized with l-lactide. The copolymer was electrospun and photo-cross-linked yielding fibrous scaffolds possessing a substantial increase in tensile modulus and crimp stability compared to the uncross-linked fibrous scaffolds. The crimp-stabilized scaffolds also showed good cytocompatibility toward 3T3 fibroblasts, which attached and grew along the crimped fibers. These findings suggest that these cross-linked fiber scaffolds may be useful for the generation of cultured ligament tissue.

Full text of DOI:10.1021/bm401813j

Article
pubs.acs.org/Biomac
Electrospun Poly(L‑lactide-co-acryloyl carbonate) Fiber Scaffolds
With a Mechanically Stable Crimp Structure For Ligament Tissue
Engineering
Fei Chen, James W. S. Hayami, and Brian G. Amsden*
Department of Chemical Engineering and Human Mobility Research Centre, Queen’s University, Kingston, Canada
*
S Supporting Information
ABSTRACT: The aim of this study was to prepare a fibrous
scaffold that possesses a crimped morphology using a photo-
cross-linkable biodegradable copolymer. To obtain the
crimped morphology, the polymer was first electrospun onto
a rotating wire mandrel to obtain aligned straight fibers.
Postprocessing by immersion in aqueous buffer at 37 °C
generated a crimplike pattern in the fibers. It was reasoned that
cross-linking the fibers following formation of the crimped
structure would endow the scaffolds with a recoverable crimp
pattern and mechanical properties similar to that of the
collagen fibers in the anterior cruciate ligament (ACL). To
achieve this aim, a trimethylene carbonate based monomer
bearing an acrylate pendant group was synthesized and copolymerized with L-lactide. The copolymer was electrospun and photo-
cross-linked yielding fibrous scaffolds possessing a substantial increase in tensile modulus and crimp stability compared to the
uncross-linked fibrous scaffolds. The crimp-stabilized scaffolds also showed good cytocompatibility toward 3T3 fibroblasts, which
attached and grew along the crimped fibers. These findings suggest that these cross-linked fiber scaffolds may be useful for the
generation of cultured ligament tissue.
1
. INTRODUCTION
under in vivo conditions, the extracellular matrix formed is a
disorganized, hypercellular tissue having a different crimp
The anterior cruciate ligament (ACL) is the most frequently
injured ligament of the knee. The ACL does not heal effectively
when torn, likely as a result of the washing away of any blood
clot that begins to form by the synovial fluid that surrounds the
ligament. Without a clot to form a provisional matrix and
stabilize the torn ends of the ACL, cells cannot regenerate the
damaged tissue. Suture repair of the ACL has been ineffective
with over 90% of the patients treated in that manner failing to
heal. The afflicted knee joint cannot support the same loads,
and if the ligament is left untreated, subsequent chondral and
13
pattern than is present normally. As a result, the healing tissue
13
has reduced mechanical strength. Therefore, a clinically
acceptable engineered ACL tissue should possess biological
architecture and mechanical properties closely aligned with
native ACL to allow for physiologic loading after implantation.
In the human ACL, fibroblasts within the extracellular matrix
secrete collagen molecules that then self-assemble sequentially
1
1
4,15
2
,3
into microfibrils, fibrils, and fibers (1−32 μm in diameter).
The fibrils are cross-linked to each other, forming fiber bundles
1
4,15
4
called fascicles (100−250 μm in diameter).
The fibers are
meniscal injury, and early osteoarthritis, may result. For these
aligned along the long axis of the ligament and display a
reasons, surgical reconstruction of the ACL is often required.
The current dominant reconstruction strategy is to implant
autograft tendon tissue. Although generally successful, con-
version of the autograft to ligamentous tissue is a long process
and has the disadvantages of anterior knee pain, decreased
range of motion, donor site morbidity, and limited donor
tendon availability. The limitations of these existing ACL
replacement approaches have led to the investigation of tissue-
engineered ligament grafts.
However, to date no engineered ACL tissue has the
appropriate extracellular architecture or mechanical properties
following in vivo transplantation. This is due in large part to the
lack of scaffolds with adequate structure and mechanical
properties capable of supporting ligament development in vitro
and remodeling in vivo. During the remodeling stage of healing
14−16
wavelike pattern along their length called crimp.
This
16
crimp pattern repeats every 45−60 μm. Crimp plays a
1
4,15
biomechanical role at both the cellular and tissue level.
The
ACL is characterized by a bimodal stress−strain response with a
compliant response referred to as the toe region at low strains,
1
7,18
4,5
which changes to a stiff response at larger strains.
The
crimp pattern allows for 7 to 16% of creep before permanent
15
6
−12
deformation and ligament damage occurs. Larger loads then
elongate the fibers causing a gradual increase in tissue
1
5,19
stiffness.
The established standard mechanical properties
Received: December 10, 2013
Revised: April 2, 2014
Published: April 3, 2014
©
2014 American Chemical Society
1593
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
24a
Scheme 1. Four-Step Synthesis Process for the Preparation of the AC Monomer
a
Reaction conditions: (i) tris(hydroxymethyl)ethane, p-toluenesulfonic acid monohydrate, acetone, room temperature; (ii) acryloyl chloride,
triethylamine, tetrahydrofuran, 0 °C; (iii) HCl (1.0 M), methanol, room temperature; (iv) ethyl chloroformate, triethylamine, tetrahydrofuran, 0 °C.
Reproduced from ref 24. Copyright 2009 American Chemical Society.
Scheme 2. Ring-Opening Polymerization of LLA with the AC Monomer and Final Copolymer Structure
of the human ACL from young donors are a Young’s modulus
maintain a toe region and crimped pattern following dynamic
loading and for fibroblast adhesion and proliferation.
2
0
20
of 111 MPa, ultimate tensile strength of 38 MPa, and 12 N·
17
m for energy absorbed at failure.
We have recently developed an electrospinning technique
that generates nano/micro polymer fibers that are aligned.
These fibers spontaneously crimp when immersed in 37 °C
2
. MATERIALS AND METHODS
All materials were used as received unless otherwise noted. Stannous
-ethylhexanoate (95% purity), and Dulbecco’s Modified Eagle’s
2
21
phosphate-buffered saline. Scaffolds prepared from these
fibers also exhibit a toe region when pulled in uniaxial
Medium were obtained from Sigma-Aldrich, ON, Canada. L-lactide
was obtained from Purac, The Netherlands, and purified by
recrystallization from anhydrous toluene. The following were used in
the synthesis of the AC monomer: tris(hydroxymethyl)ethane and 2,2-
dimethoxypropane (ACROS Organics, New Jersey, U.S.A.), p-
toluenesulfonic acid monohydrate, acryloyl chloride, and ethyl
chloroformate (Sigma-Aldrich, Oakville, ON, Canada), acetone,
diethyl ether, potassium carbonate, tetrahydrofuran (THF), triethyl-
amine, methanol, and hydrochloric acid (Fisher Scientific, Canada).
THF was dried with 4 Å molecular sieves for 2 days before use.
2
1,22
tension.
We have also demonstrated that a crimp pattern
in scaffolds is beneficial in that fibroblasts seeded on such
scaffolds and cultured under dynamic stretching conditions
produced an extracellular matrix (ECM) approximating the
composition of native ACL and formed fascicle-like collagen
22
bundles, whereas fibroblasts seeded on linear fibers did not.
This response was reasoned to be due to the approximation of
the deformation conditions normally experienced by an ACL
fibroblast attached to the collagen fiber. While these results are
promising, the copolymer that has been used to date, poly(L-
lactide-co-D,L-lactide) did not possess adequate mechanical
properties when hydrated (Young’s modulus of ∼10 MPa), lost
its crimp pattern under long-term dynamic stimulation, and
2.1. Polymer Synthesis. Acryloyl carbonate (AC) was synthesized
via a four step process (Scheme 1) based on the synthesis described by
2
4
Chen et al. Once synthesized, the AC monomer was copolymerized
with L-lactide (LLA) at varying molar ratio via ring-opening
polymerization utilizing stannous 2-ethylhexanoate as a catalyst and
octan-1-ol as an initiator (Scheme 2). The copolymerization was
carried out in toluene at 110 °C for 24 h. The monomer to initiator
mole ratio was set at 500:1, while the initiator to catalyst mole ratio
was 300:1. The resulting copolymers were then purified by dissolving
in CH Cl and precipitating in cold methanol.
23
exhibited modest fibroblast attachment.
The primary objective of this study was to improve on this
scaffold design by altering the polymer composition. It was
reasoned that cross-linked fibrous scaffolds would maintain
fiber crimp structure upon dynamic loading and result in
significantly increased hydrated scaffold modulus. To produce a
cross-linkable polymer fiber, a trimethylene carbonate based
monomer bearing an acryloyl pendant group (5-methyl-2-oxo-
2
2
2.2. Polymer Characterization. The monomer compositions of
the poly(L-lactide-co-acryloyl carbonate) (P(LLA-AC)) copolymers
were determined from 1H NMR spectra obtained from a Bruker
2
4
Avance-400 MHz spectrometer. The synthesized copolymers were
dissolved in chloroform-d (Fluka, Canada) at 10 mg/mL and their
chemical shifts were measured relative to the methyl proton resonance
of an internal tetramethylsilane reference. The integration of chemical
shifts corresponding to ∼5.2 ppm (methine hydrogen of LLA groups),
and 5.6−6.4 ppm (intact vinyl protons of AC monomer) were used to
determine the monomer compositions. A Mettler Toledo DSC1
system was used to measure the thermal properties (glass transition
1
,3-dioxan-5-yl acrylate, or acryloyl carbonate for simplicity)
was synthesized (Scheme 1) and copolymerized with L-lactide
Scheme 2). The acryloyl carbonate (AC) monomer was
(
chosen as it can be easily copolymerized with other lactone
monomers and can be readily cross-linked through radical
polymerization initiated for example by UV irradiation.
Copolymers were synthesized containing different amounts of
AC monomer, and these copolymers were electrospun into
crimped polymer mats, which were subsequently photo-cross-
linked. The cross-linked scaffolds were assessed for mechanical
properties in uniaxial tension both initially and following in
vitro degradation under static conditions for their ability to
temperature (T ), melting point (T ), and heat of fusion (ΔH )) of
g
m
f
the synthesized copolymers. The samples were taken through two
heating cycles and one cooling cycle. The heating cycles ran from −80
to 180 °C, while the cooling cycle ran from 180 to −80 °C. A 10 °C/
min, heating rate with a 3 min hold time at each set point was utilized.
The T was measured from the second heating cycle, while Tm and
ΔH were measured from the first heating cycle. The minimum of the
endothermic curve in the first heating cycle was taken as T , while the
g
f
m
1
594
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
Figure 1. Schematic illustration of the preparation of the crimped and cross-linked fiber scaffolds. Top refers to the operating, or ambient,
temperature while T refers to the polymer glass transition temperature.
g
area bound by the endotherm normalized to mass of the polymer
2.4. Characterization of P(LLA-AC) Fibrous Scaffolds. Fiber
diameter and crimp parameters were measured from SEM images. The
fiber mats were air-dried, mounted on aluminum stubs, pulse sputter
coated (Anatech Hummer VI-A Sputter Coater, U.S.A.) and imaged
with a JEOL JSM840 SEM (Peabody, Japan) with HKL Flamenco
EBSD data acquisition software (version 5.0.6.0, U.S.A.). The
alignment of the fibers was manually quantified from the SEM images
using ImageJ, by measuring the orientation angles relative to the
horizontal for a minimum of 100 fibers. Fiber diameter (n = 100) and
crimp parameters (wavelength and amplitude, n = 20) were
determined from SEM images processed with Sigma Scan Pro that
was calibrated for each magnification with the in-image scale. The
mechanical properties of the crimped electrospun fiber mats were
measured in uniaxial tension at 37 °C using a micromechanical tester
sample was used to calculate ΔH . The crystallinity of the polymers
f
was calculated by dividing the measured ΔH value by the ΔH of
f
f
25
purely crystalline poly(L-lactide) (93.1 J/g ). The number-average
molecular weight (M ), weight-average molecular weight (M ), and
n
w
molar-mass dispersity (Đ ) were measured using a Viscotek 270max
M
Separation Module equipped with a refractive index (RI) detector. A
set of two porous PolyAnalytik columns with an exclusion limit
6
molecular weight of 20 × 10 g/mol were used in series. The
calibration curve based on the RI detector was constructed using
narrow molecular weight poly(styrene) standards ranging from 6910
to 844 000 g/mol. THF at 40 °C was used as the mobile phase at a
flow rate of 1 mL/min with the polymer sample dissolved in THF at
concentration of 10 mg/mL. Results were analyzed and fitted in Astra
v4.90.07 software (Wyatt Technology, U.S.A.).
(
Mach-1, Biomomentum, Canada), to determine the change in
material properties with respect to the change in AC content (3, 5,
, and 12 mol %) for dry scaffolds, and during in vitro degradation in
2
.3. Fabrication of Crimped and Photo-Cross-Linked Fibrous
8
Scaffolds. The copolymers were electrospun into fibers as described
21
PBS for scaffolds prepared with the copolymer containing 12 mol %
AC (0, 1, 3, and 5 weeks). For the week 0 value, the scaffolds were
immersed in PBS for 30 min to become hydrated prior to measuring
the mechanical properties. For these measurements, rectangular fiber
scaffolds were cut from the mandrel in the direction of the fiber
alignment and were transferred onto cardboard windows using double-
sided adhesive tape (Scotch, 3M Canada). The ends of the window
were mounted between the grips of the Mach-1, the sides of the
cardboard window cut with scissors, and the scaffolds pulled at a rate
of 1% strain/s until failure with load and displacement captured at 100
Hz. Tensile stress in the fibers was calculated as the measured force
normalized to the total cross sectional area of the fibers mounted in
previously. In this process, the polymers were dissolved in a 3:1
(
(
volume ratio) solution of dichloromethane/dimethylformamide
DCM/DMF) (Sigma-Aldrich Ltd., Canada) at a concentration
ranging from 20 to 35% (w/v). The electrospinning apparatus
consisted of a syringe pump (KD Scientific Inc., U.S.A.), high voltage
generator (Gamma Voltage Research, U.S.A.) and a rotating wire
mandrel (collection device) (Figure 1). The wire mandrel was
constructed using two circular end pieces of 5 cm diameter that were
connected by four equally spaced, 10 cm long connecting rods. The
wire mandrel was attached to a horizontally mounted in-line mixer
(
Barnant, Series 20, U.S.A.), which rotated at 1000 rpm (as measured
with a contact tachometer, Extech). The polymer solutions were
pumped at a flow rate of 0.03 mL/min through a 21 ga. blunt tip
needle (Becton Dickinson & Company, U.S.A.). The air gap between
the needle tip and ground plate was set at 15 cm, and a 1 kV/cm
positive electric field was applied between the air gap using the voltage
generator. Electrospinning was performed at ambient conditions
23
the sample, as described in Surrao et al. The applied strain was
measured as the deformation normalized to the gauge length of the
scaffold sample. The modulus of the crimped fiber scaffolds was
determined from a linear regression of the linear region of the stress−
strain curve. The equilibrium water uptake of the polymers was
determined by soaking the fibrous scaffolds in PBS for 24 h at 37 °C
(relative humidity 50−70% and 20−22 °C). The fibers collected on
(
n = 3). Following soaking, the scaffolds were dabbed dry with a
the wire mandrel were stored at room temperature in a vacuum
desiccator until required. The aligned fibers were cut from the mandrel
and immersed in phosphate buffered saline (PBS) at 37 °C for 30 min
to induce a crimped pattern and to remove trace amounts of residual
DMF. The resultant crimped fibers were then divided into two groups.
One group of fibers were photo-cross-linked for 5 min per side in their
wet state using a UV light source (EFOS 3000, EXFO Canada) at an
Kimwipe and wieghed, then dried in vacuo (15 mmHg, 20 °C) for 48
h and weighed again. The water content of the hydrated fibers is
reported as the difference in wet versus dry weight of the scaffolds
divided by the initial dry weight of the scaffolds.
To determine the ability of the crimp to be recovered for uncross-
linked and cross-linked fiber scaffolds prepared with P(LLA-AC) of
12% AC content, the initial gauge length of the fiber scaffolds
(measured from the exposed fibers between the edges of the mounting
frame) was used to set the gage distance of the micromechanical tester
−2
intensity and wavelength of 100 mW cm and 365 nm, respectively.
The second group of fibers was left uncross-linked as controls.
1
595
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
Table 1. Physical Properties of P(LLA-AC) Copolymers before and after Cross-Linking
uncross-linked
cross-linked
a
a
a
a
b
mol % AC in feed
product mol % AC
M (kDa)
n
ĐM
Tg (°C)
Tm (°C)
Xc (%)
Tg (°C)
Tm (°C)
Xc (%)
%w/w water
3
6
2
5
72.1
62.2
55.9
61.5
1.2
1.1
1.2
1.2
58
56
55
50
161
156
154
147
31
32
22
26
63
64
65
67
143
137
135
124
4
4
3
5
8 ± 5
11 ± 5
12 ± 6
9 ± 7
1
2
0
0
8
12
a
b
n = 2, standard deviations are ±1 °C. Hydrated water content, equilibrium value measured at 24 h for the electrospun scaffolds, n = 3.
40% were obtained. A representative 1H NMR spectrum
grips. The mounting frame was placed in the grips so that none of the
frame edges were exposed to ensure no slack was present in the fiber
mats prior to testing. The immediate load profile during testing also
indicated there was no slack in the mounted scaffolds (no preload was
applied). The scaffolds were then pulled to 20% strain at a strain rate
of 1%/s. The scaffolds were removed and placed in PBS at 37 °C for
obtained in CDCl of the synthesized AC monomer as well as
3
the peak assignments are provided in Supporting Information
Figure 1.
3
.2. Synthesis of P(LLA-AC). The aim of this study was to
provide a versatile approach to synthesizing photo-cross-
linkable polymers with similar mechanical properties to the
ACL. On the basis of our previous findings, wherein poly(L-
lactide-co-D,L-lactide) (80:20 L-lactide/D,L-lactide molar ratio)
fiber scaffolds exhibited modulus values of approximately 10
30 min to allow the crimp pattern to recover. The recovered scaffolds
were then remeasured for gauge length, which was used to calculate
the new 20% total strain and 1%/s strain rate before mounting the
scaffolds in the grips at the required zero strain gauge length. This
process was continued for a total of 10 pulls unless the scaffold broke.
The modulus was calculated from the initial linear slope of the stress
versus strain curve for each measurement, which occurred below 10%
strain, as described above.
Dynamic testing of crimped fiber scaffolds prepared with P(LLA-
AC) of 12% AC content was performed to determine the loading
profiles of the scaffolds under rapid loading and unloading conditions
at 37 °C. The scaffolds were mounted in the micromechanical tester as
described above and were pulled to 10 and 20% strain at 1 Hz for 10
cycles. The loading data were fitted and plotted together to show the
progression of the loading curves through the 10 cycles.
23
MPa in the hydrated state, L-lactide was chosen as the
predominant monomer and copolymerized with AC. The
resultant copolymers had a high molecular weight and
moderate molar-mass dispersity (1.2−1.5), as measured by
1
GPC (Table 1). Importantly, the H NMR spectra of the
copolymers clearly displayed chemical shifts at 5.6−6.4 ppm,
attributable to intact vinyl protons on the pendant acryloyl
group (Figure 2). The composition of the copolymers was
Sol content of cross-linked P(LLA-AC) fiber scaffolds was measured
by weight loss. The initial dry weights (dry mass initial) were measured
using a balance (AT-250 Balance, Mettler Instrument Corporation,
U.S.A.), the scaffolds were placed in glass vials containing 5 mL of
DCM, and the DCM was replaced every 2 h over a two day period. At
the end of day 2, the scaffolds were vacuum-dried overnight in an oven
set at 37 °C, and the dry mass (dry massfinal) measured. The sol
content of the scaffolds was calculated using
dry massinitial − dry massfinal
Sol content =
dry massinitial
2.5. Cell Culture. NIH 3T3 fibroblasts were maintained in
Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma-Aldrich,
Canada) supplemented with 10% fetal bovine serum (FBS Sigma-
Aldrich, Canada) and 1% antibiotic in a 37 °C and 5% CO incubator
2
between passage P1 and P4 before seeding on the fiber scaffolds. Discs
1
Figure 2. Representative H NMR spectrum of the P(LLA-AC)(12%
AC) copolymer.
(
8 mm dia.) were cut from the fiber scaffolds using a disposable biopsy
punch and sterilized by UV light exposure for 30 min per side in a
laminar flow hood and then placed in a 24 well culture plate. Cells4
were seeded onto one side of the scaffolds at a density of 2.5 × 10
cells/scaffold using a seeding volume of 50 μL. The cells were allowed
to attach for 4 h and then 1.95 mL of supplemented medium was
added. The medium was replaced after 2 days. The number of
metabolically active cells was measured using the MTT assay 24 and
determined by comparing the integrals of signals of the acryloyl
vinyl protons with those of PLLA methine protons at δ = 5.16
ppm. It was evident that both functionality and molecular
weight of copolymers of AC and LLA could be readily
controlled using the polymerization technique employed.
However, it must be noted that when a high feed ratio of AC
monomer (20%) was used in the copolymerization, only 12−
14% AC was present in the resulting copolymer (Table 1). The
relatively low AC incorporated into the copolymer structure is
believed to be due to a lower reactivity of the AC monomer
compared to LLA, likely as a result of the bulky pendant
24
7
2 h after seeding.
2.6. Statistics. Measured values are presented as the average ± the
standard deviation about the average with the number of replicate
samples indicated by n. Statistical differences were assessed using a
one-way Kruskal−Wallis or two-way Friedman analysis of variance
with a Boferroni posthoc test using p < 0.05 as a measure of
significance.
26
acryloyl group. A similar result was reported by Chen et al.
The number-average molecular weight (M ), molar-mass
dispersity (Đ ), and thermal properties of the synthesized
copolymers are also listed in Table 1. A single glass transition
3
. RESULTS AND DISCUSSION
.1. Synthesis of the AC monomer. The functional cyclic
carbonate monomer AC was synthesized in four steps
n
3
M
24
according to Chen et al. Moderate total yields of about
temperature (T ) was detected on the first and second heating
g
1
596
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
cycles for all the copolymers. The T , melting point (T ), and
electrospun P(LLA-AC) fibers were 0.9 ± 0.12 μm (Table 3).
The aligned fibers were effectively treated to possess a crimp
g
m
degree of crystallinity (X ) of the copolymers tended to
c
decrease with increasing AC monomer content due to AC
monomer insertion into the copolymer backbone resulting in
disruption of the regular close packing of molecular chains.
Table 3. Crimp Parameters and Average Diameter of the
P(LLA-AC) Fibers
a
Following cross-linking, the T tended to increase for all the
g
crimp parameters
copolymers and was roughly the same for each at approximately
fiber diameter
(μm)
wavelength
(μm)
amplitude
(μm)
65 °C. The melting point and degree of crystallinity also tended
copolymer
to decrease for each copolymer. The melting points of the
cross-linked samples again tended to decrease with increasing
AC content in the copolymer, while the degree of crystallinity
was substantially reduced and was approximately equivalent for
each network at ∼4%. These results are expected for a cross-
linked copolymer, as cross-linking restricts polymer chain
movement and interferes with the ability of the LLA blocks
P(LLA-AC)(5%AC)
P(LLA-AC)(8%AC)
0.9 ± 0.24
0.72 ± 0.12
0.85 ± 0.10
44 ± 13
49 ± 6
42 ± 9
28 ± 9
31 ± 5
22 ± 5
P(LLA-AC)(12%
AC)
a
n = 100 for fiber diameters and n = 20 for crimp parameters.
27
pattern (Figure 3), yielding crimp amplitudes and wavelengths
similar to those of the collagen fibers present in native ACL
along the backbone to align and crystallize.
Following incubation in PBS at 37 °C for 24 h, cross-linked
fiber scaffolds of the copolymers absorbed statistically
equivalent amounts of water (p = 0.12−0.58), circa 10 w/w
(
amplitude of 5−10 μm, and wavelength of 45−60 μm,
16
respectively ).
3
.4. Uniaxial Tensile Measurement of Fiber Scaffolds’
%
, indicating that notable hydration of the polymer occurs
Mechanical Properties. The Young’s moduli of the cross-
linked scaffolds, measured before immersion in water, are
presented in Figure 4. Although the scaffolds consisting of
before substantial degradation of the polymer begins and that
the extent of hydration is unaffected by the relatively low
amounts of AC monomer incorporated into the copolymer.
Further, the cross-linked polymer fibers had relatively low sol
contents (from 16 to 6%) that decreased as the AC monomer
content of the polymer increased (Table 2), indicating that the
cross-linking efficiency increased with an increase in available
pendant acryloyl groups.
Table 2. Sol Contents of the Cross-Linked P(LLA-AC)
a
copolymers
copolymer
sol content (%)
P(LLA-AC)(5%AC)
P(LLA-AC)(8%AC)
P(LLA-AC)(12%AC)
16.5 ± 0.2
11.3 ± 0.5
6.25 ± 0.2
a
Figure 4. Cross-linked scaffold Young’s modulus with respect to the
mol % AC monomer in the copolymer. (n = 4).
(
n = 3).
3
.3. Fabrication of Fibrous P(LLA-AC) Scaffolds. Defect
copolymers prepared with lower molar amounts of AC
monomer (3 and 5%) exhibited greater average moduli than
those with 8 and 12% AC, these measurements had greater
variability and there was no significant differences between the
Young’s moduli of the scaffolds in the dry state. Each dry
scaffold measured had a Young’s modulus greater than that of
free P(LLA-AC) fibers with smooth edges were obtained. The
diameter of the fibers (n = 100) was measured from SEM
images (Figure 3). The rotating mandrel used to collect the
fibers produced aligned microfibers; 85% of the fibers were
within ±10° from the horizontal. The average diameter of the
Figure 3. SEM images of electrospun P(LLA-AC) fibers before and after crimping (scale bar 200 μm).
1
597
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
the target value of 110 MPa. Given that the scaffolds prepared
with the copolymer possessing 12% AC monomer had the
lowest sol content and the least degree of variability in its
modulus values, these scaffolds were chosen for further
assessment.
Cross-linking the P(LLA-AC)(12%AC) scaffolds had an
obvious improvement on the modulus of the scaffolds (Figure
5). The modulus of the dry, cross-linked scaffolds (222 MPa ±
Figure 5. Stress−strain curves of dry P(LLA-AC)(12%AC) fiber
scaffolds before and after cross-linking and hydrated cross-linked fiber
scaffolds. The average modulus of the dry uncross-linked fiber scaffolds
was 86 ± 5 MPa and that of the dry cross-linked scaffolds was 222 ±
2
8 MPa. The average modulus of the hydrated scaffolds was 26 ± 1.4
MPa. (n = 3).
2
8) was almost two times higher than that of the uncross-linked
scaffolds (86 ± 5 MPa). Furthermore, the cross-linked scaffolds
displayed an obvious plateau region between the yield point
and the final break point. This plateau region is probably due to
an interconnected molecular entanglement structure in the
cross-linked sample. Once hydrated, the modulus of the fiber
scaffolds decreased to 26 ± 1.4 MPa, and a plateau region was
no longer observed. Instead, the fibrous scaffolds exhibited
rubbery behavior with a lower ultimate tensile strength and an
increased elongation at break. This behavior can be attributed
Figure 6. Stress−strain response of the hydrated and cross-linked fiber
scaffolds following dynamic loading. Conditions: 1 Hz for 10 cycles.
Top: up to 10% strain. Bottom: up to 20% strain. The lines represent
curve fits to the data. (n = 3).
28
23
to a plasticization effect due to absorbed water.
.5. Dynamic Mechanical Properties of Hydrated
8:12 P(LLA-AC) Scaffolds. The P(LLA-AC)(12%AC)
3
To further confirm the recoverability of the crimp pattern
following uniaxial extension induced by photo-cross-linking, the
change in the gauge length (which is a measure of the degree of
crimp unfolding under applied strain) and modulus of the
scaffolds were measured following a series of pull-relaxation
cycles. Differences between pulls and between cross-linked
versus uncross-linked were assessed using a two-way Friedman
analysis of variance. There was a slight, but not significant,
downward trend in modulus of the cross-linked scaffolds over
10 pull-relaxation cycles (Figure 7A), which remained at
approximately 30.1 ± 10 MPa. In contrast, a continuous and
significant decrease in modulus when compared to the initial
modulus for uncross-linked scaffolds was observed from pull-
relaxation cycles 3−6. Following pull 6, only one scaffold
remained unbroken. The modulus of this scaffold stayed
constant at approximately 10 MPa for the remaining pulls.
Moreover, the gauge length increased with each pull-relaxation
cycle for the uncross-linked scaffolds, whereas the gauge length
of cross-linked scaffolds remained constant (Figure 7B).
8
scaffolds were subjected to dynamic loading and a pull-
relaxation test following hydration in pH 7.4 buffer. The
hydrated condition was necessary to reflect the future
performance of the scaffolds within a dynamically stimulated
bioreactor. Repetitive crimp unfolding of the photo-cross-
linked P(LLA-AC) fibers over 10 loading cycles demonstrated a
further advantage of cross-linking the polymer following crimp
induction. The cross-linked scaffolds exhibited a nearly linear
initial loading profile (Figure 6), while a pronounced toe region
was observed for each following loading cycle. The change in
loading profile was due possibly to fiber breakage and/or
release of interfiber entanglements. Following the first loading
cycle, a reproducible loading profile was observed for the 10
cycles tested. The length of the toe region was dependent on
the ultimate strain applied; with an ultimate applied strain of
1
0%, the toe region was approximately 4%, while with an
ultimate applied strain of 20% the toe region was approximately
0%. The difference in the length of the toe region could be
1
The crimp pattern of the photo-cross-linked P(LLA-AC)
fibers after 10 pull-relaxation cycles was examined visually from
SEM images (Figure 8). These images clearly show that the
crimped and cross-linked fibers recovered their initial wave-
due to the difference in the overall strain applied coupled with
the time allowed for the fibers to relax and recover their initial
crimped pattern, or the differences in the strain rates applied.
1
598
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
Figure 7. Change in (A) modulus and (B) recoverable gauge length following a series of pull-relaxation cycles for both uncross-linked and cross-
linked hydrated P(LLA-AC)(12AC%) scaffolds. Conditions: 1 Hz for 5 cycles, up to 20% strain. Relaxation condition: immersed in 37 °C water for
30 min. (n = 3).
modulus of the cross-linked fiber scaffolds throughout the
degradation study time period. The copolymer exhibited a
slight decrease in T at 5 weeks when compared to the initial
g
Tg, indicative of the onset of polymer degradation. To
determine whether there were any changes in fiber structure
during the in vitro degradation, SEM images from each time
point were taken. Figure 9 shows the morphology of the
Figure 8. SEM images of P(LLA-AC)(12%AC) fibers following 10
cycles of pulling-relaxation: (A) cross-linked, (B) uncross-linked (scale
bar 200 μm).
length and amplitude after the scaffolds were allowed to rest in
an aqueous environment at 37 °C, while uncross-linked fibers
had lost a notable degree of their original crimp structure.
3
.6. Change in Properties of P(LLA-AC)(12%AC)
Scaffolds during in Vitro Degradation. The fiber scaffolds
used in this experiment possessed an initial modulus of
approximately 26 ± 1.4 MPa in the wet state (Table 4), which
Table 4. Mechanical and Thermal Properties of Cross-
Linked P(LLA-AC)(12%AC) Fibre Scaffolds Following in
Vitro Degradation
time
week)
modulus
(MPa)
Tg- first heating cycle Tg-second heating cycle
(
(°C)
(°C)
Figure 9. SEM micrographs of cross-linked P(LLA-AC)(12%AC) fiber
scaffolds incubated in 1× PBS at 37 °C showing that the crimp
structure in the fibers is retained over the period studied: (A) week 0,
(B) week 1, (C) week 3, and (D) week 5 (scale bar 200 μm).
0
1
3
5
26 ± 1.4
21 ± 4.0
18 ± 3.5
21 ± 5.2
41
40
38
37
55
55
53
54
a
n = 3 for modulus and n = 2 for thermal properties, for every time
point. The pooled standard deviation of the glass transition values is
1.
scaffolds of each degradation time points. There were no
obvious differences before and after incubation within the
aqueous medium, such as no evidence of surface pitting or
broken fibers, demonstrating that the crimped structure
parameters were unaffected by long-term incubation in an
aqueous environment under static conditions.
±
is consistent with the results reported above in Figure 7, and
significantly greater than that obtained previously (10 ± 4
2
3
MPa ) for scaffolds composed of crimped, but uncross-linked,
poly(L-lactide-co-D,L-lactide) fibers. The explanation for the
reduction in modulus of the fiber scaffolds upon hydration
being due to plasticization of the network by the absorption of
water provided above is supported by the observed difference
between the T of the hydrated sample (T from the first
3.7. Fibroblast Attachment and Proliferation on
Cross-Linked P(LLA-AC) Scaffolds. To determine whether
the AC content of the copolymer had any effect on cell
attachment and proliferation, 3T3 fibroblasts were seeded on
uncross-linked and cross-linked fiber scaffolds prepared using
copolymers with 3, 5, 8, and 12 mol % AC. While not relevant
for an assessment of the potential of the scaffolds for the
generation of ligament tissue, 3T3 fibroblasts are a standard cell
g
g
heating cycle) and that of the dry sample (T from the second
g
heating cycle) (Table 4). There was no significant drop in the
1
599
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
2
9
line used for the assessment of biomaterial cytotoxicity. The
MTT assay was used to assess cell number on the scaffolds,
based on a measurement of the metabolic activity of the cells at
linked versus uncross-linked scaffolds prepared with P(LLA-
AC) of the same AC content, regardless of the AC content of
the polymer, suggesting that the presence of the pendant
acrylate group on the polymer was not cytotoxic. Fewer
fibroblasts attached to the scaffolds prepared with P(LLA-AC)
2
4 and 72 h (Figure 10A). The data is normalized to the
metabolic activity of fibroblasts seeded onto TCPS at the same
(
3%) than on the scaffolds prepared with P(LLA-AC) of higher
seeding density in Figures 10B,C.
AC content (5−12% AC), for which there was no significant
difference with respect to AC content. It is possible that the
higher lactide content of the 3% AC polymers produced a
different protein deposition profile on these polymer fiber
surfaces with fewer fibroblast adhesion sites exposed for
effective attachment. At 72 h, there were no significant
differences between the cell numbers on any of the scaffolds
or the TCPS controls; however, the cell numbers on the
scaffolds tended to be lower than on the TCPS controls (Figure
1
0C). SEM images (Figure 11) confirmed that the 3T3
fibroblasts attached to the cross-linked fibersand were aligning
along the fiber long axes by 24 h and were proliferating.
Figure 11. SEM images of 3T3 fibroblasts on cross-linked P(LLA-
AC)(12% AC) scaffolds (A) 24 h and (B) 72 h after seeding (scale bar
1
00 (left) and 50 μm (right)).
CONCLUSIONS
■
Previous work by our group has demonstrated that scaffolds
possessing a crimped architecture similar to that of the native
anterior cruciate ligament are beneficial in terms of the rate and
composition of extracellular matrix produced by seeded
fibroblasts when grown under dynamic loading conditions in
a bioreactor. A limitation of the polymers used to date for
preparing these scaffolds is that this crimped structure is
gradually lost under dynamic mechanical loading. Copolymer-
izing an acryloyl carbonate monomer with L-lactide yielded a
copolymer with an acryloyl pendant group along its backbone
that was effectively utilized to cross-link electrospun microfibers
possessing an appropriate crimp structure. The cross-linked and
crimped fiber scaffolds had a Young’s modulus approximating
that of the modulus of human anterior cruciate ligaments.
Importantly, cross-linking provided a means of stabilizing the
fiber crimp structure following dynamic mechanical loading.
The crimp structure was also unaffected by static incubation in
PBS for up to 5 weeks; nevertheless, the stability of the crimp
needs to be assessed under dynamic conditions in buffer in
future studies. Finally, fibroblasts seeded onto these fibers
effectively attached and proliferated. Collectively, these results
demonstrate that this scaffolding approach has potential for the
in vitro generation of replacement ligaments.
Figure 10. (A) 3T3 fibroblast number following seeding on uncross-
linked (UX) cross-linked (X) fiber scaffolds prepared with copolymers
of varying mol % AC (number following X or UX), as assessed with
the MTT assay (top) and normalized to cell number on TCPS
controls at 24 h (middle) and 72 h (bottom) (n = 3, *: p < 0.05).
There was a significant difference between 24 and 72 h
absorbances for each scaffold (Figure 10A). This result
confirms that the 3T3 fibroblasts were able to attach and
proliferate on these scaffolds. After 24 h, except for scaffolds
prepared with P(LLA-AC)(12%), there were fewer fibroblasts
on the scaffolds than on the TCPS controls (Figure 10B).
There was no difference between the cell number on the cross-
ASSOCIATED CONTENT
■
*
S
Supporting Information
1
A representative H NMR spectrum of the synthesized AC
monomer is provided, confirming the structure as well as
1
600
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601

Biomacromolecules
Article
demonstrating it was free of impurities. This material is
available free of charge via the Internet at http://pubs.acs.org.
(25) Fischer, E. W.; Sterzel, H. J.; Wegner, G. Kolloid-Z. Z.Polym.
1
(
973, 251, 980−990.
26) Matsuo, J.; Aoki, K.; Sanda, F.; Endo, T. Macromolecules 1998,
3
1, 4432−4438.
AUTHOR INFORMATION
■
(27) Sperling, L. Introduction To Physical Polymer Science; John Wiley
and Sons: New York, 1992.
28) Labrecque, L. V.; Kumar, R. A.; Dave, V.; Gross, R. A.;
McCarthy, S. P. J. Appl. Polym. Sci. 1997, 66, 1507−1513.
29) von Recum, A. Handbook of Biomaterials Evaluation, 2nd ed.;
CRC Press: Boca Raton, FL, 1999; pp 275−289.
Notes
(
The authors declare no competing financial interest.
(
ACKNOWLEDGMENTS
■
Funding for this project was provided by the Canadian
Institutes of Health Research and the Natural Sciences and
Engineering Research Council of Canada (CREATE Bio-
interfaces Training Program, CREATE Training Program in
Bone and Joint Health Technologies).
REFERENCES
■
(
1
1) Murray, M. M.; Fleming, B. C. J. Orthop. Res. 2013, 31, 1501−
506.
(
2) Feagin, J. A.; Curl, W. W. Am. J. Sport Med. 1976, 4, 95−100.
(
3) Engebretsen, L.; Benum, P.; Fasting, O.; Mølster, A.; Strand, T.
Am. J. Sport Med. 1990, 18, 585−590.
(
4) Petrigliano, F. A.; McAllister, D. R.; Wu, B. M. Arthroscopy 2006,
2
(
(
2, 441−451.
5) Laurencin, C.; Freeman, J. Biomaterials 2005, 26, 7530−7536.
6) Barber, J. G.; Handorf, A. M.; Allee, T. J.; Li, W.-J. Tissue Eng.,
Part A 2013, 19, 1265−1274.
7) Little, D.; Guilak, F.; Ruch, D. S. Tissue Eng., Part A 2010, 16,
307−2319.
8) Kew, S. J.; Gwynne, J. H.; Enea, D.; Abu-Rub, M.; Pandit, A.;
(
2
(
Zeugolis, D.; Brooks, R. A.; Rushton, N.; Best, S. M.; Cameron, R. E.
Acta Biomater. 2011, 7, 3237−3247.
(
9) Hayami, J. W. S.; Surrao, D. C.; Waldman, S. D.; Amsden, B. G. J.
Biomed. Mater. Res., Part A 2010, 92, 1407−1420.
(
10) Vaquette, C.; Kahn, C.; Frochot, C.; Nouvel, C.; Six, J.-L.; De
Isla, N.; Luo, L.-H.; Cooper-White, J.; Rahouadj, R.; Wang, X. J.
Biomed. Mater. Res., Part A 2010, 94A, 1270−1282.
(
11) Cooper, J. A.; Sahota, J. S.; Gorum, W. J.; Carter, J.; Doty, S. B.;
Laurencin, C. T. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 3049−3054.
(12) Bourke, S.; Kohn, J.; Dunn, M. Tissue Eng. 2004, 10, 43−52.
(13) Frank, C.; Woo, S.; Amiel, D.; Harwood, F.; Gomez, M.;
Akeson, W. Am. J. Sports Med. 1983, 11, 379−389.
(
14) Frank, C.; Shrive, N.; Lo, I.; Hart, D.; Buckwalter, J.; Einhorn,
T.; Simon, S. In Orthopaedic Basic Science Foundation of Clinical
Practice; American Academy of Orthopaedic Surgeons: Rosemont, IL,
2
007.
15) Vunjak-Novakovic, G.; Altman, G.; Horan, R. Annu. Rev. Biomed
Eng. 2004, 6, 131−156.
16) Amiel, D.; Billings, E.; Akeson, W. H. In Knee ligaments structure,
(
(
function, injury, and repair; Daniel, D., Akeson, W. H., O’Connor, J. J.,
Eds.; Raven Press: New York, 1990; pp 77−94.
(
17) Woo, S.; Hollis, J. M.; Adams, D. J.; Lyon, R. M.; Takai, S. Am. J.
Sport Med. 1991, 19, 217−225.
18) Jones, R. S.; Nawana, N. S.; Pearcy, M. J.; Learmonth, D. J. A.;
Bickerstaff, D. R.; Costi, J. J.; Paterson, R. S. Clin. Biomech. 1995, 10,
39−344.
19) Woo, S.; Abramowitch, S.; Kilger, R.; Liang, R. J. Biomech. 2006,
9, 1.
20) Noyes, F. R.; Grood, E. S. J. Bone Joint Surg. Am. 1976, 58,
074−1082.
21) Surrao, D. C.; Hayami, J. W. S.; Waldman, S. D.; Amsden, B. G.
Biomacromolecules 2010, 11, 3624−3629.
22) Surrao, D. C.; Fan, J. C. Y.; Waldman, S. D.; Amsden, B. G. Acta
Biomater. 2012, 8, 3704−3713.
23) Surrao, D. C.; Waldman, S. D.; Amsden, B. G. Acta Biomater.
012, 8, 3997−4006.
24) Chen, W.; Yang, H.; Wang, R.; Cheng, R.; Meng, F.; Wei, W.;
Zhong, Z. Macromolecules 2010, 43, 201−207.
(
3
(
3
(
1
(
(
(
2
(
1
601
dx.doi.org/10.1021/bm401813j | Biomacromolecules 2014, 15, 1593−1601