journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

3D braid scaffolds for regeneration

of articular cartilage
Hyunchul Ahna, Kyoung Ju Kimb, Sook Young Parkc, Jeong Eun Huhd,
Hyun Jeong Kimc, Woong-Ryeol Yua,n
a
Department of Materials Science and Engineering and Research Institute of Advanced Materials, Seoul National
University, 1 Gwanak-ro, Gwanak-gu, Seoul 151-744, Republic of Korea
b
Automotive Material Development Group 1, Cheil Industry, 332-2 Gocheon-dong, Uiwang-si, Gyeonggi-do 437-711,
Republic of Korea
c
Department of Dental Anesthesiology and Dental Research Institute, School of Dentistry, Seoul National University,
Seoul 110-768, Republic of Korea
d
Oriental Medicine Research Center for Bone & Joint Disease, Kyung Hee University, 149 Sangil-dong, Gangdong-gu,
Seoul 134-727, Republic of Korea

art i cle i nfo

ab st rac t

Article history:

Regenerating articular cartilage in vivo from cultured chondrocytes requires that the cells

Received 18 October 2013

be cultured and implanted within a biocompatible, biodegradable scaffold. Such scaffolds

Received in revised form

must be mechanically stable; otherwise chondrocytes would not be supported and patients

7 January 2014

would experience severe pain. Here we report a new 3D braid scaffold that matches the

Accepted 8 January 2014

anisotropic (gradient) mechanical properties of natural articular cartilage and is permissive

Available online 28 January 2014

to cell cultivation. To design an optimal structure, the scaffold unit cell was mathemati-

Keywords:

cally modeled and imported into nite element analysis. Based on this analysis, a 3D braid

3D braid scaffold

structure with gradient axial yarn distribution was designed and manufactured using a

Axial yarn gradient

custom-built braiding machine. The mechanical properties of the 3D braid scaffold were

Unit cell

evaluated and compared with simulated results, demonstrating that a multi-scale

Finite element method

approach consisting of unit cell modeling and continuum analysis facilitates design of

Articular cartilage

scaffolds that meet the requirements for mechanical compatibility with tissues.

1.

Introduction

Articular cartilage is a soft tissue sustaining the pressure at

joints between the hard ends of bones, and is thereby frequently damaged by wear. Articular cartilage consists mainly of
collagen bers and proteoglycan matrix, which is a type of berreinforced composite (Mow and Guo, 2003). Fig. 1 shows a
n

Corresponding author. Tel.: 82 2 880 9096; fax: 82 2 885 9671.

E-mail addresses: woongryu@snu.ac.kr,
woongryu@gmail.com (W.-R. Yu).
1751-6161/$ - see front matter & 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jmbbm.2014.01.004

& 2014 Elsevier Ltd. All rights reserved.

schematic diagram illustrating typical articular cartilage. Within

articular cartilage, chondrocytes maintain the tissue structure
by synthesizing and modulating collagen bers and proteoglycan. Chondrocytes may be cultivated outside the body and
implanted within articial articular cartilage (e.g., scaffolds) to
regenerate the articular cartilage (Buschmann et al., 1992). To
provide a supportive environment for chondrocytes to produce

38

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

articular cartilage and survive, the mechanical properties of

scaffold must be as close as possible to those of natural
articular cartilage (Cohen et al., 1999). The mechanical properties of articular cartilage are anisotropic and heterogeneous.
Furthermore, due to its gradient structure (Fig. 1), the mechanical properties of articular cartilage also vary along a gradient
(Akizuki et al., 1986; Buckwalter and Mankin, 1998; Cohen et al.,
1998), making the development of scaffolds for regenerating
articular cartilage challenging.
Woven composite scaffolds have been developed that can
readily mimic the mechanical properties of natural articular
cartilage. Woven composite scaffolds with uniform in-plane
structure and isotropic mechanical properties have been
proposed as a cartilage scaffold (Hutmacher, 2000; Temenoff
and Mikos, 2000). Because their isotropic mechanical properties did not match those of anisotropic articular cartilage, the
woven composite scaffolds could only be implanted into a
limited, local region of the articular cartilage (Woodeld et al.,
2004). Later, an anisotropic three-dimensional woven composite scaffold was developed that can mimic natural human
cartilage; however, the structural variation of this scaffold did
not match the layer-like structure of cartilage through thickness (Moutos et al., 2007). Recently, multi-layered composite
scaffolds with aligned and randomly distributed ber structures have been developed (McCullen et al., 2012). Even though
their structure mimics native articular cartilage, creating
scaffolds with gradient properties is only possible using quite
complex manufacturing processes, necessitating the adoption
of new technologies. To address the need for straightforward
methods of manufacturing multi-layered scaffolds, this study
introduces 3D braid composites as a feasible scaffold for
articular cartilage due to the ability to control the mechanical
properties by adjusting the processing parameters.
Braid composites, a type of 3D textile composite, are highperformance structural materials with high impact resistance, high failure strength, and long fatigue life, and have
been used in the high-end automobile, aerospace, and sporting goods industries (Tate et al., 2006). Several studies have
used unit cell modeling to numerically predict the elastic
properties of 3D braid composites (Zuorong et al., 1999).
Moreover, the mechanical analysis of braid composites using
FE approach has been well developed (Miravete et al., 2006;
Yu and Cui, 2007). In addition to braiding yarns, axial yarns
have also been introduced into 3D braid composites. As the
axial yarns are additional straight yarns introduced between

the braiding yarns, they improve in-plane mechanical stiffness and enhance shape stability, damage tolerance, and
fatigue resistance by increasing the ber volume fractions
(Karbhari and Wang, 2007; Potluri et al., 2003). In summary,
braiding is a well-developed, established technology allowing
the virtual design of anisotropic 3D braid scaffolds with
gradient ber volume fraction in a specic direction that
may then be manufactured by varying processing parameters
such as braiding angle, yarn density, and number of layers.
The structure of braided scaffolds closely matches the ber
structure within human organs; they have thus been examined as potential ligament replacement materials (Freeman
et al., 2007; Ide et al., 2001).
This study aimed to develop biocompatible 3D braid
scaffolds that can degrade completely following the regeneration of articular cartilage (Athanasiou et al., 1991; Dai
et al., 2010). For this purpose, a new unit cell model was
developed to predict the mechanical performance of braid
composites and optimize their internal structure in virtual
space. Using the unit cell model and nite element analysis,
an optimal structure that mimics the gradient mechanical
properties of natural articular cartilage is proposed.

2.
Methods: designing new 3D braid scaffolds
and experimental
In addition to functioning physiologically, scaffolds intended
to support regeneration of articular cartilage should be
mechanically stable; otherwise, their malfunction causes
unbearable pain. Producing mechanically stable scaffolds
requires careful design, for which mechanical modeling of
scaffolds is well suited. Here, a unit-cell approach; i.e., mathematical modeling of unit cell geometry and its nite element
analysis using the actual material properties, was employed to
predict the mechanical properties of 3D braid scaffolds.

2.1.
Mathematical modeling of unit cell geometry in 3D
braid scaffold
The unit-cell of 3D braid scaffolds can be modeled mathematically, provided that the running paths of individual
yarns inside the braid are determined by the manufacturing
process (in this case, circular braiding) (Fig. 2). The yarn
carriers on concentric layers in the braiding bed move

Fig. 1 Schematic diagram of the knee joint and articular cartilage structure (Buckwalter and Mankin, 1998; Cohen et al., 1998).
Left and right schemes of the articular cartilage represent variable distribution and density (from high at the surface area to
low at the bottom) of chondrocytes and brous structure, i.e., its gradient structure.

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

39

Fig. 2 Schematic diagram of the braiding process (a) and motion of the yarn carriers: (b) overall, (c) rst, (d) second, (e) third,
(f) fourth steps in the formation of each unit cell. The blue (b) and orange (b) and green ((c)(f)) arrows represent the movement
of the yarn carriers. The red and green circles represent axial and braiding yarn carriers, respectively; solid and dotted box in
(b) represent the unit cell after the rst and second steps and after the third and fourth steps, respectively. The unit cell
dimensions are described by w, t and h in (b). (For interpretation of the references to color in this gure legend, the reader is
referred to the web version of this article.)
Table 1 Example braiding parameters and target composite
dimensions.
Braiding parameters

Target braid dimension

Layers of braiding
Braiding yarn carriers per layer (n)
Axial yarn carriers per layer
Braiding angle ()
Inner radius (Ri)
Outer radius (Ro)
Tube thickness (T)

circumferentially; i.e., clockwise and counter-clockwise for

each carrier in the odd and even layers, respectively (Fig. 2(c)).
The carriers then move radially; i.e., inward and outward for
each carrier in the odd and even columns, respectively (Fig. 2
(d)). After these two steps, the carriers move in the opposite
direction to the previous rst and second steps, respectively
(Fig. 2(e) and (f)), completing a total of four steps. Considering
these carrier motions, the repeating unit of the resulting
braid can be identied schematically as shown in Fig. 2(b).
Two assumptions were made for the mathematical description of the unit cell: the braiding yarns in the unit cell are
straight and their cross-sections are elliptical with major and
minor radii of a and b, respectively. The geometrical description was then completed by mathematically expressing the
actual running paths of individual yarns. This description
requires the unit cell size of the preform braid (width (w) 
thickness (t)  height (h)), which can be determined using the
braiding parameters and the physical dimensions of the target
braid composite (Table 1). The axial yarn thickness does not
affect the size of the unit cell. The width of the unit cell was
determined from the number of braiding yarns per layer, while

4
72
72
301
16.48 mm
17.91 mm
1.43 mm

its thickness was calculated from the target composite thickness. The height of the unit-cell was calculated using the
braiding angle, as follows:
w

Ri Ro
T
w=2
 4; t
 4; h
n
N1
tan

where N and n are the number of braiding layers and yarn

carriers per layer, respectively. , Ri , and R0 are representing
the braiding angle and the inner and outer radius, respectively.
The meanings of characters in Eq. (1) are provided in
Tables 1 and 2. Yarns were assumed to be elliptical. The radii
of the ellipses were determined from the ber volume
fraction. The ber volume fraction (v) in the unit cell can be
calculated using the volume of the axial (Va) and braiding (Vb)
yarns:
v%

Va Vb
 100
Vt

where Vt (wth/2) is the total volume of the unit cell. The

volumes of the axial and braiding yarns are given by:
Va 8abl; Vb 8abh0

40

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

Table 2 Unit cell geometry determined from the braiding parameters.

Dimensions
t (thickness)
w (width)
h (height)
a (major semi-axis yarn length)
b (minor semi-axis yarn length)

1.1440 mm
6.0022 mm
5.1980 mm
0.5020 mm
0.0661 mm
R h q
Axial yarn curve parameters determining axial yarn length (l 0 1 z'2 dy; z A sin By C D)
A
B
C
D

0.1375
1.1589
0.6775
0.1375

where l represents the length of the axial yarn in the unit cell
q
andh0 w=22 t=22 h2 . From the ber volume fraction
of the target composite, the product of the two radii (a, b) can
be determined using Eq. (2). The radius of the short axis (b) is
assumed to be 0.96t/8 to prevent interpenetration between
yarns (see Table 2 for detailed expression of each variable).

2.2.

Finite element analysis of the simulated unit cell

The Young's modulus of a 3D braid scaffold can be calculated

by nite element analysis (FEA) of the unit cell model
described in Section 2.1. The unit cell was geometrically
constructed using the TexGen software, and was imported
into commercial FEA software (ABAQUS/standard).
To calculate the mechanical properties of the unit cell using
FEA, the constitutive equation of each yarn in the unit cell was
required, for which the yarn and matrix were assumed to be
transversely isotropic and isotropic, respectively. Note that the
ber and resin, which are constituent materials of the yarn and
matrix, were assumed isotropic. The yarn properties were
considered in the axial and transverse direction. The axial
direction represented the ber direction (longitudinal) in the
yarn (3-direction), while the transverse directions are represented by 1 and 2-directions. The stiffness matrix of the
transverse isotropic materials is expressed by Eq. (4).
1 2
1=Et
11
C 6
B
B 22 C 6 tt =Et
C 6
B
B 33 C 6 ta =Et
C 6
B
C6
B
0
B 12 C 6
C 6
B
B 13 C 6
0
A 4
@
0
23
0

tt =Et
1=Et

at =Ea
at =Ea

ta =Et

1=Ea

1=Gtt

0
0

0
0

0
0

1=Gta
0

0
0

0
0

1
s11
7B
C
7B s22 C
7B
C
B s33 C
0 7
7B
C
7B
C
0 7B s12 C
7B
C
B s13 C
0 7
5@
A
1=Gta
s23
0
0

30

where s11, s22, s33, 11, 22, 33, are the normal stress and strain in
the material directions, respectively, and s12, s13, s23, 12, 13, 23
are shear stress and strain.
The basic mechanical properties of the yarns (coefcient
in Eq. (4)) were then calculated using the following mixture
rule (Clyne, 1990):
Ea Vf Ef Vm Em ; Et

Ef Em
Vf Em Vm Ef

Gf Gm
Et
; Gtt
Vf Gm Vm Gf
21 vtt
vat
vta
Et
Et

; vta Vf vf Vm vm ; vtt 1 vta 

Ea
Et
Ea
3K
Gat Gta

Kf Km
Ef
Em
; Km
; Kf
V f Km V m Kf
31 2vf
31 2vm
Ef
Em
; Gm
Gf
21 vf
21 vm
K

where Ea and Et are the axial and transverse yarn moduli,

respectively, and Ef and Em are the ber and matrix elastic
moduli, respectively. Gf and Gm are the ber and matrix shear
moduli and Gat and Gtt are the axial-transverse and transversetransverse moduli, respectively. vat and vtt are the axialtransverse and transverse-transverse Poisson's ratio, respectively. K, Kf, and Km are bulk moduli of composite, ber and
matrix, respectively. Elastic moduli were determined experimentally as shown in Table 3. Because the agarose gel used as
matrix resin of the composite is nearly incompressible, its
Poisson's ratio should be 0.5 but for numerical stability 0.49 or
similar value (e.g., 0.49999) has been used frequently. In this
research, however, 0.3 was inputted for Poisson's ratio of
agarose gel to improve the numerical stability. The following
simulation justied this assumption. With 0.49999 inputted for
Poisson's ratio of the matrix, Young's modulus of the composites was calculated and compared with the results from the
assumption of Poisson's ratio of 0.3, showing very small
difference (0.02%) between the two results. Computational cost,
however, was increased more than 10%. Therefore the Poisson's
ratio of the agar gel was assumed to match with that of the ber
(0.3) for numerical stability in the nite element analysis. The
ber volume fraction in the unit cell can be controlled during
the fabrication processes. The ber volume fraction of the yarns
was determined to be  70% from experimental unit cell
dimensions and yarn density. The elastic moduli in the axial
and transverse directions of the yarns were then calculated
using Eq. (5).
To obtain the Young's modulus, the unit cell was elongated to have a strain of 0.005 in the axial direction under
periodic boundary conditions. The reaction force at the xed
side was then converted into the modulus value.

2.3.

Experimental

2.3.1.

Mechanical test

To verify the mathematical modeling in Section 2.1, braided

composites were fabricated from ultra-high-molecular-weight
polyethylene (UHMWPE) bers (Dong-il Industrial Co., Korea)

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

Table 3 Elastic constants and volume fraction of

cartilage-mimic composite components.

PLGA
yarn
Agarose
gel

Young's
modulus (MPa)

Poisson's
ratio (v)

Volume
fraction (%)

48,000

0.3

30

0.3

70

0.235

and epoxy resin (YD-128 with curing agent, G-A0432, both

from Kukdo Chemical Co., LTD., Korea). Three-dimensional
braid preforms were manufactured using a four-step process
with a custom-built circular braiding machine. Braid composites were then fabricated using compression molding by
immersing braid preforms in epoxy resin and removing the
void in the resin by vacuum. Based on the melting point of
UHMWPE, the molding temperature was set to 90 1C for 30 min
at 1 MPa. Following epoxy resin mounting, braid composites
were cut into sections of 9/8 the unit cell size in width using a
diamond saw to observe the upper half of the unit cell. Crosssections were polished using sand paper and aluminum oxide
powder and observed by optical microscopy and computed
tomography (CT).
To validate the numerical analysis (Section 2.2), the tensile
and compressive strengths of UHMWPE preforms were measured using universal test machines (Model 5565 and 8801,
Instron, USA) at strain rates of 0.002 and 0.004 s  1, respectively. For the articular cartilage scaffold, 3D braid preform
composites were manufactured from biocompatible, biodegradable poly(lactic-co-glycolic acid) (PLGA, L:G ratio 10:90,
Samyang, Korea) bers and agarose gel matrix, also by
compression molding. For the tensile test, rectangular samples in 15  2  150 mm (width  thickness  height) were
used. The specimens were pulled in the height direction with
the clamping length of 30 mm. A preload of 5 N was applied
to the specimens for stabilizing composite ber structure. For
the compression test, rectangular samples in 15  2  15 mm
(width  thickness  height) were tested with free-end. A preload of 0.3 N was applied to the specimens for stabilizing
composite ber structure.
The tensile and compressive behaviors of these composites were then characterized and compared with the numerical solution calculated using the method in Section 2.2.
Biodegradability was an important criterion in selecting these
materials because engineered tissue scaffolds should be
completely cleared after bearing mechanical load during
cartilage regeneration. The composite was fabricated using
unit-cell model dimensions.

2.3.2.

Cell cultivation on the scaffold

To cultivate chondrocytes, the scaffold must be hydraulically

permeable (Buckwalter and Mankin, 1998). Braid scaffolds
have been proven to be mechanically robust enough to
sustain the load between the knees; however, their capacity
for supporting cells has not yet been examined. This section
describes tests of 3D braid scaffolds' compatibility with

41

chondrocytes, a necessary prerequisite for their development

as scaffolds for regenerating articular cartilage.
Primary chondrocytes were obtained from rabbits (2-weekold New Zealand white). Cartilage samples were washed
twice in phosphate-buffered saline (PBS), chopped into small
pieces, and incubated in in PBS with 0.2% collagenase type II
(Sigma Chemical Co., St. Louis, MO, USA) at 37 1C water bath
for 30 min. Then, fetal bovine serum (FBS) was added for
inhibition the collagenase, and dissociated single cells were
collected by centrifuging at 1300 rpm for 5 min. The incubation and centrifuging steps were repeated 4 times. Isolated
cartilage cells were cultured in Dulbecco's modied Eagle's
medium (DMEM) supplemented with 10% (v/v) FBS and 1%
(v/v) antibiotic/antimycotic solution in a humidied air with
carbon dioxide concentration of 5% at 37 1C. After harvest
using trypsin/EDTA, cells were seeded directly on each
sterilized PLGA 3D braid scaffolds (2.5  105 cells/cm2) and
cultured for 1, 4, 7, and 14 days. The culture medium was
changed every 3 days. The cells cultured on scaffolds were
xed in 2.5% glutaraldehyde after washing with PBS and postxed with 1% (w/v) OsO4. Then, samples were dehydrated in
a stepwise ethanol gradient, dried, coated with platinum, and
observed with a scanning electron microscope (SEM, SUPRA
55VP, Carl Zeiss, Germany) for measuring the morphological
features of cells with scaffolds. The samples were tested
using deoxyribonucleic acid (DNA) and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays to
characterize the cell proliferation. For MTT assay, the samples were incubated in serum free alpha modication of
eagle's minimum essential media (-MEM) supplemented
with 0.5 g MTT at 37 1C for 4 h and the purple formazan was
extracted by 0.04 M HCl in 2-propanol. The extracted solution
was measured using a UV/VIS spectrophotometer at 570 nm.
Total 44 scaffolds were prepared for SEM observation (one
sample per day), MTT assay (ve samples per day) and DNA
assay (ve samples per day) tests. The DNA amount of cells
cultured on scaffolds was isolated following the kit protocol
(QIAGEN, Valencia, CA) and the concentration was measured
by nanodrop (ND-1000, Wilmington, USA). The data were
statistically analyzed by independent Student's t-test using
Origin ver. 8.0 (Origin Lab, Northampton, MA). nnnpo0.001 as
compared with 1 day; #po0.5; ##po0.05; and ###po0.001 as
compared with the previous day.

3.

Results and discussion

3.1.
Validation and numerical analysis of the model
unit cell
Unit cell geometry was constructed using TexGen as shown
in Fig. 3 and compared with experimental observations
(Fig. 4). Cross-sections of braid composites revealed that axial
yarns were not perfectly elliptical due to yarn lament
compaction; however, the positions of the two yarns in the
braid composites were similar to those in the model (Fig. 4(a)).
Comparison of modeled shapes with CT images of 3D
UHMWPE preforms with and without axial yarns (Fig. 4(b))
demonstrates that unit cell modeling (Section 2.1) is an

42

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

Fig. 3 Geometrical model of a 3D braid preform unit cell with axial yarns.

Fig. 4 Cross sections of 3D braid composites. (a) SEM, and (b) micro-CT image of 3D braid preform, (i) with and (ii) without
axial yarns.

accurate mathematical description of the geometry of 3D

braid preforms and composites.
The Young's moduli of UHMWPE 3D braid preforms with
and without axial yarns were calculated by numerical analysis. To investigate the effect of axial yarns on the tensile
properties of 3D braid preforms, four yarn volume fractions
were analyzed. Fig. 5 shows unit cells modeled using various
axial yarn densities: 1.78% (two axial yarns), 3.58% (four), and
7.14% (eight) (Fig. 5(b)). The total ber volumes of the unit
cells were set to be 56.39% and 49.25% for 3D braid preforms
with and without axial yarns, respectively. Comparison of
experimental and simulated results (Fig. 6) demonstrated
that this modeling and numerical analysis approach

accurately predicted the small deformation of 3D braid ;preforms. The axial yarn is the main determinant of the Young's
modulus of the 3D braid preform, offering exibility in designing scaffolds suitable for regenerating articular cartilage.

3.2.

Newly designed 3D braid scaffold

The mechanical properties of native articular cartilage are

anisotropic, especially in the depth (thickness) direction. To
mimic such anisotropic properties, the number of axial yarns
should vary in the direction of the braid; i.e., the ber fraction
should vary layer by layer. To mimic natural articular cartilage, a 3D braid preform with four layers was manufactured.

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

43

Fig. 7 Optimal composite scaffold unit-cell structure.

Fig. 5 Mathematical modeling of braid preforms with axial

yarns. (a) Possible axial yarn locations (red circles), (b) two
axial yarns (1.78% ber volume), (c) four axial yarns (3.58%
ber volume) (d) eight axial yarns (7.14% ber volume). (For
interpretation of the references to color in this gure legend,
the reader is referred to the web version of this article.)

Fig. 6 Correlation of experimental moduli with predicted

values for varying axial yarn contents.

To investigate the effect of gradient axial yarns on the

mechanical properties of 3D braid scaffolds, the gradient and
overall tensile properties of the four-layer preform were
calculated by numerical analysis using a commercial nite
element code (ABAQUS). An initial unit cell was constructed
by rst setting the ber volume fractions to values compatible
with the hydraulic permeability required for physiological
function of chondrocytes (i.e., braiding yarn: 30%, axial yarn:
04%) (Buckwalter and Mankin, 1998). For ease of manufacturing and allow an anisotropic in-plane modulus (larger
than 1 MPa in both x and y directions), the braiding angle of
the unit cell was set to 301. The mechanical properties of this
optimized unit cell structure (Fig. 7) were calculated and
compared with those of native articular cartilage (Table 4).

These calculations suggest that 3D braid preforms with

gradient axial yarn fraction are a suitable mimic of native
articular cartilage. Next, the feasibility of manufacturing 3D
braid preforms with a varying number of axial yarns in each
layer was assessed by producing the designed preform and
observing its internal structure (Fig. 8). As shown in Fig. 8, the
composite scaffold with four layers was successfully manufactured. Each layer shows different amount of axial yarns
contained in it. Layer 1 (Fig. 8(b) and (c)) represent top layer
with highest amount of axial yarns. As the layer number
increased, the amount of the axial yarns was reduced, so
layer 4 contained no axial yarns. Micro-CT revealed properly
formed gradient structures in axial yarns (pixel size set to
6 m using SKYSCAN 1173.)
Finally, the tensile and compressive behaviors of the 3D
braid scaffolds were assessed. Fig. 9 compares the experimental stressstrain curves of the scaffolds with FEA predictions (described in Section 2.2). The predicted tensile stress
strain curve was well matched to experimental observations
up to 4% strain. While the predicted stressstrain curve is
linear, the experimental appeared to stiffen at higher strains,
implying that small deformation compacts and jams the
yarn, which was not predicted by the simulation. Slight
overestimation of the stress in the simulation was due to
the manufacturing process, in which yarns swelled, reducing
the load-carrying capacity somewhat. In contrast, the compressive behavior of these 3D braid preforms was almost
linear up to  12%. The simulation results agreed well with
experiments, implying that this design is adequate for developing various scaffolds requiring mechanical stability. Since
the mechanical properties of these scaffolds are similar to
those of natural articular cartilage, they can be used as a
scaffold to regenerate articular cartilage.

3.3.

Cell cultivation results

The morphologies of cultured cells were visualized by SEM at

1, 4, 7 and 14 days after cell seeding (Fig. 10). At 1 day,
only few cells were observed (Fig. 10(A-a)). However, from
4 day observation, chondrocytes has started to proliferate
and a greater number of cells are visible under the SEM
(Fig. 10(A-b)). After 7 days, the growth of chondrocytes start to
spreads over individual bers (Fig. 10(A-c)) and by 14 days the
gaps between bers are lled with chondrocytes, forming a
sheet like structure (Fig. 10(A-d)). The white arrows on Fig. 10
(A) indicate chondrocytes. These qualitative observations
support the conclusion that 3D braid scaffolds are an appropriate substrate for cell cultivation. Cellular proliferation and

44

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

Table 4 Elastic constants of engineered 3D braid scaffolds and native articular cartilage (Akizuki et al., 1986; Elliott et al.,
1999; Khalsa and Eisenberg, 1997; Setton et al., 1993).
Compressive modulus (MPa)

Young's modulus (MPa)

Lower (V0)

Native cartilage
Engineered scaffold (predicted)

0.40.8
0.581

Middle 1 (V1)

135 (continuous distribution)

1.407
10.494

Middle 2 (V2)

Upper (V4)

19.797

37.925

Fig. 8 Micro-CT images of gradient braided scaffold. (a) 3D braid preform with four-layer axial yarn gradient; (b) trans-axial
view of the 3D braid preform; (c) cross-sections revealing gradient variation in axial yarn content. Red lines mark boundaries
between layers; red arrows mark the axial yarns. (For interpretation of the references to color in this gure legend, the reader
is referred to the web version of this article.)

3.0

2.0
1.5
1.0
0.5
0.0

Experimental
Numerical

5e+4

Stress (Pa)

Stress ( Mpa)

6e+4

Experimental
Numerical

2.5

4e+4
3e+4
2e+4
1e+4

10

12

Strain (%)

-2

-4

-6
Strain (%)

-8

-10

Fig. 9 Stressstrain curves of PLGA/agarose 3D braid composite scaffolds. (a) Tensile behavior and (b) compressive behavior.

metabolic activities were also investigated quantitatively

using MTT (Fig. 10(B-a)) and total DNA content (Fig. 10(B-b))
assays. The number of cells and viability increased signicantly compared with 1 day (marked by asterisks) or previous

day of each time point (marked by sharps) in both MTT and

DNA graphs. The steady increase of cell quantity indicates
and reinforce that the 3D braid scaffolds are suitable for
chondrocyte cultivation.

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

45

Fig. 10 Cellular morphology and viability results of primary chondrocytes cultured on PLGA 3D braid scaffolds. (A) SEM
images of cells at 1 (A-a), 4 (A-b), 7 (A-c) and 14 (A-d) days after cell seeding. White arrows in picture indicate the
chondrocytes. (B) MTT assay (metabolic rate, (B-a)) and DNA content (proliferation rate, (B-b)). nnnpo0.001 versus 1 day; #po0.5;
##
po0.05; and ###po0.001 versus the previous day by independent Student's t-test.

4.

Conclusions

To develop biocompatible 3D braid scaffolds that can sustain

the loads placed on human joints, support chondrocytes to
allow regeneration of articular cartilage, and degrade to allow
complete clearance following regeneration, a design strategy
utilizing multi-scale modeling concept was employed: unitcell modeling of the target structure, nite element analysis
to predict its mechanical properties, and experimental validation. The results demonstrated that the 3D braid structure
provides an environment suitable for chondrocyte cultivation; specically, a multi-layered 3D braid scaffold with axial
yarns mimics the anisotropic and gradient mechanical properties of natural articular cartilage. An optimal scaffold was
designed based on numerical analysis and manufactured
using a braiding machine. Comparison of its mechanical
properties with predicted results revealed reasonable agreement, conrming the validity of this prediction approach.

Acknowledgement
The authors would like to thank the Korea Science and
Engineering Foundation (KOSEF) for sponsoring this research
through the SRC/ERC program of MOST/KOSEF (R11-2005-065)
and also funded by the Korea government Ministry of Knowledge Economy under the project 10038479.

r e f e r e nc e s

Akizuki, S., Mow, V.C., Muller, F., Pita, J.C., Howell, D.S., Manicourt,
D.H., 1986. Tensile properties of human knee joint cartilage: I.
Influence of ionic conditions, weight bearing, and fibrillation
on the tensile modulus. J. Orthop. Res. 4, 379392.
Athanasiou, K.A., Rosenwasser, M.P., Buckwalter, J.A., Malinin, T.I.,
Mow, V.C., 1991. Interspecies comparisons of in situ intrinsic
mechanical properties of distal femoral cartilage. J. Orthop. Res.
9, 330340.

46

journal of the mechanical behavior of biomedical materials 34 (2014) 37 46

Buckwalter, J.A., Mankin, H.J., 1998. Articular cartilage: tissue

design and chondrocyte-matrix interactions. Instr. Course
Lect. 47, 477486.
Buschmann, M.D., Gluzband, Y.A., Grodzinsky, A.J., Kimura, J.H.,
Hunziker, E.B., 1992. Chondrocytes in agarose culture
synthesize a mechanically functional extracellular matrix.
J. Orthop. Res. 10, 745758.
Clyne, T.W., 1990. A compressibility-based derivation of simple
expressions for the transverse Poissons ratio and shear
modulus of an aligned long fibre composite. J. Mater. Sci. Lett.
9, 336339.
Cohen, N.P., Foster, R.J., Mow, V.C., 1998. Composition and
dynamics of articular cartilage: structure, function, and
maintaining healthy state. J. Orthop. Sports. Phys. Ther. 28,
203215.
Cohen, Z.A., McCarthy, D.M., Kwak, S.D., Legrand, P., Fogarasi, F.,
Ciaccio, E.J., Ateshian, G.A., 1999. Knee cartilage topography,
thickness, and contact areas from MRI: in-vitro calibration
and in-vivo measurements. Osteoarthr. Cartil. 7, 95109.
Dai, W., Kawazoe, N., Lin, X., Dong, J., Chen, G., 2010. The
influence of structural design of PLGA/collagen hybrid
scaffolds in cartilage tissue engineering. Biomaterials 31,
21412152.
Elliott, D.M., Guilak, F., Vail, T.P., Wang, J.Y., Setton, L.A., 1999.
Tensile properties of articular cartilage are altered by
meniscectomy in a canine model of osteoarthritis. J. Orthop.
Res. 17, 503508.
Freeman, J.W., Woods, M.D., Laurencin, C.T., 2007. Tissue
engineering of the anterior cruciate ligament using a braid
twist scaffold design. J. Biomech. 40, 20292036.
Hutmacher, D.W., 2000. Scaffolds in tissue engineering bone and
cartilage. Biomaterials 21, 25292543.
Ide, A., Sakane, M., Chen, G., Shimojo, H., Ushida, T., Tateishi, T.,
Wadano, Y., Miyanaga, Y., 2001. Collagen hybridization with
poly(-lactic acid) braid promotes ligament cell migration.
Mater. Sci. Eng., C 17, 9599.
Karbhari, V.M., Wang, Q., 2007. Influence of triaxial braid denier
on ribbon-based fiber reinforced dental composites. Dent.
Mater. 23, 969976.

Khalsa, P.S., Eisenberg, S.R., 1997. Compressive behavior of

articular cartilage is not completely explained by proteoglycan
osmotic pressure. J. Biomech. 30, 589594.
McCullen, S.D., Autefage, H., Callanan, A., Gentleman, E., Stevens,
M.M., 2012. Anisotropic fibrous scaffolds for articular cartilage
regeneration. Tissue Eng. Part A 18, 20732083.
Miravete, A., Bielsa, J.M., Chiminelli, A., Cuartero, J., Serrano, S.,
Tolosana, N., de Villoria, R.G., 2006. 3D mesomechanical
analysis of three-axial braided composite materials. Compos.
Sci. Technol. 66, 29542964.
Moutos, F.T., Freed, L.E., Guilak, F., 2007. A biomimetic threedimensional woven composite scaffold for functional tissue
engineering of cartilage. Nat. Mater. 6, 162167.
Mow, V.C., Guo, X.E., 2003. Mechano-electrochemical properties of
articular cartilage: their inhomogeneities and anisotropies.
Annu. Rev. Biomed. Eng. 4, 175209.
Potluri, P., Rawal, A., Rivaldi, M., Porat, I., 2003. Geometrical
modelling and control of a triaxial braiding machine for
producing 3D preforms. Composites Part A 34, 481492.
Setton, L.A., Zhu, W., Mow, V.C., 1993. The biphasic
poroviscoelastic behavior of articular cartilage: role of the
surface zone in governing the compressive behavior. J.
Biomech. 26, 581592.
Tate, J.S., Kelkar, A.D., Whitcomb, J.D., 2006. Effect of braid angle
on fatigue performance of biaxial braided composites. Int. J.
Fatigue 28, 12391247.
Temenoff, J.S., Mikos, A.G., 2000. Review: tissue engineering for
regeneration of articular cartilage. Biomaterials 21, 431440.
TexGen, http://texgen.sourceforge.net/index.php/Main_Page.
Woodfield, T.B.F., Malda, J., de Wijn, J., Peters, F., Riesle, J., van
Blitterswijk, C.A., 2004. Design of porous scaffolds for cartilage
tissue engineering using a three-dimensional fiber-deposition
technique. Biomaterials 25, 41494161.
Yu, X.G., Cui, J.Z., 2007. The prediction on mechanical properties
of 4-step braided composites via two-scale method. Compos.
Sci. Technol. 67, 471480.
Zuorong, C., Dechao, Z., Lu, M., Ye, L., 1999. Evaluation of elastic
properties of 3-D (4-step) regular braided composites by a
homogenisation method. Compos. Struct. 47, 477482.