Preparation and Cell Compatibility Evaluation of
Chitosan/Collagen Composite Scaffolds Using
Amino Acids as Crosslinking Bridges
Sung-Pei Tsai,1 Chien-Yang Hsieh,1 Chung-Yu Hsieh,1 Da-Ming Wang,1
Lynn Ling-Huei Huang,2 Juin-Yih Lai,3 Hsyue-Jen Hsieh1
1Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan
2Institute of Biotechnology, National Cheng Kung University, Tainan 701, Taiwan
3Research and Development Center for Membrane Technology, Chung Yuan University, Chung-Li 320, Taiwan
Received 13 March 2006; accepted 16 December 2006
DOI 10.1002/app.26157
Published online 26 April 2007 in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT:
In this study, a novel freeze-gelation method
amino acids to the scaffolds. Cell compatibility was demon-
instead of the conventional freeze-drying method was used
strated by the in vitro culture of human skin ?broblasts on
to fabricate porous chitosan/collagen-based composite scaf-
the scaffolds. The ?broblasts attached and proliferated well
folds for skin-related tissue engineering applications. To
on the chitosan/collagen composite scaffolds, especially the
improve the performance of chitosan/collagen composite
one with glutamic acid molecules as crosslinking bridges,
scaffolds, we added 1-ethyl-3-(3-dimethylaminopropyl)-car-
whereas cells did not grow on the chitosan scaffolds. Our
bodiimide (EDC) and amino acids (including alanine, gly-
results suggest that the collagen-modi?ed chitosan scaffolds
cine, and glutamic acid) in the fabrication procedure of the
with glutamic acid molecules as crosslinking bridges are
composite scaffolds, in which amino acid molecules act as
very promising biomaterials for skin-related tissue engineer-
crosslinking bridges to enhance the EDC-mediated crosslink-
ing applications because of their enhanced tensile strength
ing. This novel combination enhanced the tensile strength of
and improved cell compatibility with skin ?broblasts. Ó 2007
the scaffolds from 0.70 N/g for uncrosslinked scaffolds to 2.2
Wiley Periodicals, Inc. J Appl Polym Sci 105: 1774–1785, 2007
N/g for crosslinked ones; the crosslinked scaffolds also
exhibited slower degradation rates. The hydrophilicity of the
Key words: amino acid; biocompatibility; biomaterials; chi-
scaffolds was also signi?cantly enhanced by the addition of
tosan; crosslinking
INTRODUCTION
N-acetylglucosamine. The term, degree of deacetyla-
tion, represents the molar ratio of glucosamine units
Skin substitutes such as xenografts, allografts, and
to all repeating units on the chitosan molecule.3
autografts have been employed for therapeutic treat-
Because of the presence of amino groups, the chitosan
ments of severe burns and nonhealing wounds. Tissue
polymer is positively charged and solubilized by pro-
engineering applies both life science and engineering
tonation at environmental pH values of <6. Therefore,
principles to develop functional substitutes for dam-
the solubility of chitosan is in?uenced by its degree of
aged tissues.1,2 Engineered skin products have the
deacetylation. Usually a 1–3% acetic acid solution is
potential to assist wound healing and reduce scarring
used as a solvent for chitosan. The presence of amino
after skin cancer surgery. To ?nd suitable biomaterials
groups, due to their chemical reactivity, also provides
for arti?cial skin development, a combination of chito-
the possibility for the chemical modi?cation of chito-
san, collagen, and amino acids was employed to pre-
san. Chitosan can be enzymatically degraded by chiti-
pare composite scaffolds in this study.
nase, chitosanase, and pectinase. It can also be
Chitosan-based biomaterials are receiving increased
degraded by lysozyme in vivo. Its mechanical proper-
attention in tissue engineering-related applications.
ties can be improved by crosslinking.4 The degrada-
Chitosan is a polysaccharide derived from the N-
tion rate of chitosan is relatively slow, but this can be
deacetylation of chitin. The molecular structure of chi-
modi?ed by the degree of deacetylation. A higher
tosan is a copolymer comprised of glucosamine and
degree of deacetylation produces a lower rate of de-
gradation.5,6 Because of its antiseptic, biocompatible,
Correspondence to: H.-J. Hsieh (hjhsieh@ntu.edu.tw).
and degradable properties, chitosan has been widely
Contract grant sponsor: National Science Council, Taiwan;
used in various biomedical applications,7,8 including
contract grant number: NSC93-2214-E-002-034.
wound dressings,9 drug delivery,10 and tissue engi-
Contract grant sponsor: Ministry of Education, Taiwan.
neering.11–14
Collagen scaffolds have also been widely used as a
Journal of Applied Polymer Science, Vol. 105, 1774–1785 (2007)
V
C 2007 Wiley Periodicals, Inc.
dermal equivalent to induce ?broblast in?ltration and

---

SCAFFOLDS WITH AMINO ACIDS AS CROSSLINKING BRIDGES
1775
dermal regeneration.9 Collagen is a major component
Our previous research utilized polyglutamic acid to
of the human extracellular matrix and connective tis-
modify chitosan scaffolds, and the results were prom-
sues such as dermis, bone, cartilage, tendons, liga-
ising.31
ments, and basement membrane. Because of its excel-
In the present study, we prepared composite scaf-
lent biocompatibility, low antigenicity, and high avail-
folds containing chitosan as a framework and collagen
ability, collagen is widely used as a biomaterial.
as a cell-recognizing component plus amino acids as
Various forms of collagen materials, such as ?lms,
crosslinking bridges, and the proposed crosslinking
sponges, gels, have been developed for medical appli-
scheme is shown in Figure 1. Additionally, a novel
cations. However, collagen scaffolds generally de-
freeze-gelation method32 instead of the conventional
grade quickly in vivo, so some treatments, such as glu-
freeze-drying method was used to fabricate porous
taraldehyde crosslinking,15 thermal dehydration,16
composite scaffolds. This method saves time and
and UV radiation,17 are used to increase the mechani-
energy, and is suitable for fabricating large-sized scaf-
cal strength of collagen by introducing intra- and
folds.32 The structure, mechanical strength and elon-
intermolecular linkages as well as to decrease its de-
gation, hydrophilicity of the scaffolds, and the ability
gradation rate. Even though collagen is a major com-
of ?broblasts to attach to and proliferate on these com-
ponent of skin, it is relatively expensive.
posite scaffolds were examined.
Scaffolds composed of collagen and chitosan may
create an appropriate environment for the regenera-
tion of skin. Our previous studies have indicated that
EXPERIMENTAL
skin ?broblasts are not very compatible with chito-
Materials
san.18,19 The addition of collagen should facilitate the
attachment and proliferation of skin ?broblasts. Chito-
Chitosan (with a degree of deacetylation of 90% and a
san/collagen composite scaffolds have been fabri-
molecular weight of $ 300,000) was purchased from
cated by many researchers.20–24 Nevertheless, some
Kiotek (Taipei, Taiwan). EDC, NHS, phosphate-buf-
studies did not use crosslinking agents, and thus the
fered saline (PBS), alanine, glycine, glutamic acid, so-
scaffolds, especially collagen ones, quickly degraded.
dium dodecyl sulfate (SDS), Tergitol (type NP-40),
A few studies used glutaraldehyde as a crosslinking
and phenylmethanesulfonyl ?uoride (PMSF) were
reagent, and some studies mentioned that glutaralde-
purchased from Sigma-Aldrich (St. Louis, MO). Colla-
hyde did not have harmful effect to cells,25,26 but some
gen was puri?ed from fresh pig skin by pepsin diges-
studies reported that glutaraldehyde was toxic even at
tion and an acetic acid dissolution method.33 The pu-
a concentration of 3.0 ppm.17,27 To improve the cell
rity of the type I collagen was con?rmed by SDS-
compatibility of chitosan and the biological stability of
PAGE analysis with Coomassie brilliant blue staining.
collagen, this study utilized the crosslinking reagent,
Distilled and deionized water was used throughout
1-ethyl-3-(3-dimethylamino-propyl)-carbodiimide/N-
this study.
hydroxyl-succinimide (EDC/NHS),28,29 and amino
Minimum essential medium (MEM) was purchased
acids including alanine, glycine, and glutamic acid as
from Sigma-Aldrich, and fetal bovine serum (FBS) and
crosslinking bridges to fabricate novel composite scaf-
penicillin–streptomycin (10,000 U/mL) from Biologi-
folds (Fig. 1). EDC is a zero-length crosslinking agent
cal Industries (Kibbute, Israel). Trypsin-EDTA was
used to conjugate carboxyl to amino groups. NHS can
purchased from Biochrom AG (Berlin, Germany), and
improve the ef?ciency of EDC coupling reactions. Ala-
sodium pyruvate was purchased from GIBCO Invitro-
nine has a nonpolar, hydrophobic side chain [Fig.
gen (Grand Island, NY). Tissue culture ?asks and 12-
1(a)]. Glycine is the simplest amino acid and is nonpo-
and 24-well plates were obtained from Corning (Big
lar [Fig. 1(b)]. The number ratio of carboxyl to amino
Flats, NY). Cells used for the cell compatibility study
groups in alanine or glycine is 1 (COOH/NH
were WS1 human embryonic skin ?broblasts (ATCC
2 ¼ 1 : 1).
Alanine differs from glycine in that a hydrophobic
CRL-1502; American Type Culture Collection, Mana-
methyl group rather than hydrogen atoms is attached
ssas, VA).
to its a-carbon. Glutamic acid has an acidic carboxyl
group on its side chain [Fig. 1(c)]. The number ratio of
Preparation of scaffolds
carboxyl to amino groups in glutamic acid is 2
(COOH/NH2 ¼ 2 : 1).30 Amino acids with positively
Porous scaffolds were prepared by the freeze-gelation
charged side chains like lysine were not used in this
method.32 Brie?y, chitosan was dissolved in a 0.2M
study because chitosan provides a large amount of
acetic acid aqueous solution to form a 3.5 wt % chito-
amino groups and also many positive charges. Thus
san polymer solution. Afterwards, EDC, NHS, amino
for increasing the possibility of crosslinking in the
acid, and collagen acetic acid solutions were added to
EDC/NHS system, we should increase the amount of
form a viscous polymer solution, which was continu-
carboxyl groups, and that led to the use of alanine,
ally stirred at 48C for 6 h. The polymer solution
glycine, and especially glutamic acid in this study.
was centrifuged for 15 min at 3000 Â g to remove the
Journal of Applied Polymer Science DOI 10.1002/app

---

1776
TSAI ET AL.
Figure 1
Structural formula of amino acids utilized as crosslinking bridges. (a) Alanine; (b) glycine; (c) glutamic acid; (d)
structure of chitosan; (e) proposed crosslinking diagram of chitosan, collagen, and amino acids. A.A.: amino acid.
insoluble impurities, and then the polymer solution
fuged for 15 min at 3000 Â g. The ?lm solution was
was poured into a square stainless steel plate (150 Â
then poured into dishes, and dried in an oven for 24 h
150 mm2) with a specially made mold and frozen at
with the temperature maintained at 408C. The dehy-
À808C for 6 h. The frozen chitosan solution was
drated ?lms were then immersed in a 3M NaOH solu-
immersed in a precooled NaOH aqueous solution for
tion for 12 h, followed by rinsing with ethanol and a
24 h, and gelation occurred below the freezing point
PBS solution. At the end, the ?lms were washed with
of the polymer solution. Subsequently the scaffolds
distilled water six times and then stored at 48C for fur-
were washed by 95% ethanol and PBS buffer. Finally
ther
cell
compatibility
assays.
For
comparison
the scaffolds were kept in a moist condition at 48C
between the freeze-gelation method and lyophiliza-
until further experiments were carried out.
tion, we also made lyophilized samples. Brie?y, the
To observe the ?uorescence staining of cells, nonpo-
polymer solution was poured into the same square
rous ?lms were made from the same polymer solu-
stainless steel plate (described in the preceding para-
tion. Each solution was prepared in a 0.2M acetic acid
graph) and then frozen at À808C for 6 h. Finally, the
solution, continually stirred for 12 h, and then centri-
frozen solutions were lyophilized for 72 h in a freeze-
Journal of Applied Polymer Science DOI 10.1002/app

---

SCAFFOLDS WITH AMINO ACIDS AS CROSSLINKING BRIDGES
1777
TABLE I
Compositions and Abbreviations of the Five Types of Scaffold Solutions
Scaffold
Chitosan
Collagen
Alanine
Glycine
Glutamic
solution
(wt %)
(wt %)
(mM)
(mM)
acid (mM)
Total (mM)
chi
3.5 (190, 0)a
–
–
–
–
(190, 0)
chi-col-non
3.5 (190, 0)
0.0188 (1.8, 1.8)
–
–
–
(191.8, 1.8)
chi-col-ala
3.5 (190, 0)
0.0188 (1.8, 1.8)
150 (150, 150)
–
–
(341.8, 151.8)
chi-col-gly
3.5 (190, 0)
0.0188 (1.8, 1.8)
–
150 (150, 150)
–
(341.8, 151.8)
chi-col-glu
3.5 (190, 0)
0.0188 (1.8, 1.8)
–
–
150 (150, 300)
(341.8, 301.8)
The concentrations of acetic acid, EDC, and NHS were 0.2M, 21.5 mM, and 10.75 mM, respectively, in all scaffold solutions.
According to the average molecular weight of the chitosan repeating units being 165.2, the amino group concentration of chi-
tosan in the scaffold solution was about 190 mM. The amino and carboxyl group concentrations of collagen in the scaffold so-
lution were about 1.8 mM.
a Values within parentheses are the concentrations of amino (NH2) and carboxyl (COOH) groups, respectively.
drier (Heto LyoLab 3000, Allerød, Denmark) to obtain
?xed with 2.5% glutaraldehyde in a 0.1M PBS solu-
lyophilized samples.
tion. Samples were dehydrated through a graded se-
We divided the scaffold solutions into ?ve groups
ries of ethanol, supercritical carbon dioxide-dried, and
according to their compositions. Three different amino
then gold-coated.
acids [alanine (Ala), glycine (Gly), and glutamic acid
(Glu)] were used in this study. The weight percentage of
chitosan in the scaffold solution was 3.5%. Chitosan is
Determination of water uptake
comprised of two kinds of repeating units, glucosamine
(M
Hydrophilicity is an important characteristic property
w: 161) and N-acetyl-glucosamine (Mw: 203). The
degree of deacetylation of chitosan we used was 90%, so
of biomaterials. To determine the hydrophilicity of the
the average molecular weight of the repeating units was
porous scaffolds, the bulk water absorption of the
165.2. By using the value of the molecular weight, the
scaffolds was determined to reveal their hydrophilic
number of amino groups in the chitosan solution could
behavior. To determine the water uptake,34 scaffolds
be calculated. The amount of amino acid (150 mM) used
were immersed in PBS at 378C to obtain the change in
was 70% of the amount of the amino groups on the chito-
water uptake with respect to time. Six samples were
san in the solution. The compositions of all ?ve types of
measured for each type of scaffold. The percentage of
scaffold solutions are summarized in Table I.
water uptake was calculated using the following equa-
tion:
Thermal properties analysis by differential
Percentage of water uptake
scanning calorimetry
¼ ðWwet À WdryÞ=Wdry  100%;
Differential scanning calorimetry (DSC) was employed
to investigate the endothermic peak temperature shift
where Wdry and Wwet are the weights of the scaffolds
among the scaffolds. Scaffolds were cut into small
before and after immersion in PBS, respectively.
pieces (about 2 mm3), and then four to six small pieces,
weighing about 5 mg, were placed in an aluminum
pan. The aluminum pan was pressed to seal it, and thus
Analysis of scaffold tensile strength and elongation
the porous scaffolds were crushed to ensure complete
contact with the aluminum pan. Analysis was carried
The tensile strength and elongation to fracture of the
out in a differential scanning calorimeter (DSC2010; TA
porous scaffolds were examined using a universal
Instruments, New Castle, DE) from 50 to 2508C at a con-
testing machine (model LRX; LLOYD, Paoli, PA) at a
trolled heating rate of 58C/min. Three samples from
constant speed of 10 mm/min with a preload of 0.5 N.
each composition were measured.
Samples were cut into a typical dog-bone shape simi-
lar to the American Society for Testing Material
(ASTM) D 4762-04 standard, and then each sample
Analysis of the scaffold structure by scanning
weight was measured. The measured maximum ten-
electron microscopy
sile strength (load to fracture) was normalized by
The surface and cross-section morphologies of the po-
weight for each scaffold. Before measuring the tensile
rous scaffolds were examined by scanning electron
strength and elongation, all test specimens were
microscopy (SEM; Hitachi JSM-6300, Tokyo, Japan).
immersed in PBS for 30 min at room temperature. At
Samples were placed on a Cu mount and coated using
least eight samples were measured for each composi-
a gold-coating apparatus. In vitro cell samples were
tion of scaffold.
Journal of Applied Polymer Science DOI 10.1002/app

---

1778
TSAI ET AL.
In vitro enzymatic degradation studies
removed, and then cultured cells were washed with
PBS three times. Lysis buffer was introduced to the
In vitro biodegradation test of the scaffolds with and
cells with a reaction time of 10 min. Lysis buffer con-
without the addition of amino acids as crosslinking
tained 0.25 wt % sodium deoxycholate, 1 wt % NP-40,
bridges was performed by lysozyme digestion. The
0.2 wt % SDS, and 4 mM PMSF. The resulting cell
scaffolds were immersed in PBS (pH 7.4) containing
lysate was recovered for the protein concentration
0.5 mg/mL lysozyme (L6876; Sigma-Aldrich) at 378C
assay using the Pierce BCA protein assay kit (Rock-
on an orbital shaker with speed of 100 rpm for 1, 2, 3,
ford, IL). At least three samples were measured for
and 4 weeks. At the end of each period, the scaffolds
each group.
were dried at 508C for 12 h, and then the weights were
measured. Five samples were measured for each type
of scaffold. The remaining weight for the scaffolds is
Statistical analysis
de?ned by the following equation:
Signi?cant differences in the mechanical properties
Remaining weight ð%Þ ¼ W=W
and cell compatibility were determined utilizing a
i  100%;
one-way analysis of variance (ANOVA test) and New-
where W
man–Keuls post hoc comparisons. The independent
i is the initial dry weight and W is the dry
weight after enzymatic degradation.
variable was the composition. Signi?cant differences
in the water uptake and in vitro enzymatic degrada-
tion were determined utilizing a two-way ANOVA
test and Newman–Keuls post hoc comparisons. The in-
Determination of cell compatibility
dependent variables were composition and time. Dif-
WS1 human embryonic skin ?broblasts were cultured
ferences were considered to be statistically signi?cant
in the MEM supplemented with 10% FBS at 378C in a
at P < 0.05. All statistical analyses were undertaken
5% CO2 incubator. Scaffolds were sterilized with a
using the SPSS statistical software (Chicago, IL).
70% ethanol solution and ultraviolet light prior to cell
culture. WS1 skin ?broblasts were seeded onto 12-
well plates at a concentration of 1 Â 104 cells/well.
RESULTS AND DISCUSSION
Cell morphology and attachment were monitored
Thermal properties of chitosan/collagen
using an image analysis system connected to an Olym-
composite scaffolds
pus (Tokyo, Japan) IX70 inverted phase-contrast
The thermal properties of the scaffolds were analyzed
microscope. Images were captured with a high-resolu-
by DSC as shown in Figure 2. There was an endother-
tion digital camera, and processed using image analy-
mic peak around 958C for the chitosan/collagen (chi-
sis software. For ?uorescence microscopy, the culture
col-non) scaffold, 1208C for the chi-col-ala scaffold,
medium was removed and then PBS supplemented
1228C for the chi-col-glu scaffold, and 1288C the for
with a ?uorescein diacetate (FDA) and propidium
iodide (PI) working solution was added and incu-
bated for 10 min at 378C in the dark. FDA is a nonpo-
lar and non?uorescent compound that is able to enter
cells and is hydrolyzed to ?uorescein and acetate by
nonspeci?c esterases in the cytoplasm.35 Fluorescein is
retained by the cell if the plasma membrane is intact.
Fluorescein can be monitored by excitation at 495 nm
and emission at 535 nm (green ?uorescence). PI inter-
calates into double-stranded nucleic acids and can be
monitored by excitation at 530 nm and emission at
610 nm (red ?uorescence).36 PI cannot pass through
intact cell membranes and is used to stain nucleic
acids in dead or dying cells with injured membranes.
The ?uorescence images were obtained using a digital
camera mounted on an Olympus inverted microscope
with ?uorescence accessories.
To determine the cell numbers on various surfaces,
a standard curve of cell number versus the respective
protein concentration was ?rst established. Then it
Figure 2
Differential scanning calorimetric (DSC) curves
from chitosan/collagen composite scaffolds with or without
was possible to indirectly determine the cell number
the addition of different amino acids as crosslinking bridges.
by measuring the protein concentration of a cell lysate.
See Table I for the composition of each scaffold. The heating
To carry out the measurement, the medium was
rate was 58C/min.
Journal of Applied Polymer Science DOI 10.1002/app

---

SCAFFOLDS WITH AMINO ACIDS AS CROSSLINKING BRIDGES
1779
Figure 3
Scanning electron microscopic (SEM) images (magni?cation, 500Â) showing the surface morphology of chitosan/
collagen composite scaffolds with or without the addition of amino acids. (a) chi-col-non scaffold; (b) chi-col-ala scaffold;
(c) chi-col-gly scaffold; (d) chi-col-glu scaffold. Scale bar ¼ 50 mm.
chi-col-gly scaffold. The scaffolds with amino acids as
SEM analysis of porous scaffolds
crosslinking bridges showed higher peak tempera-
The surface morphology and interior cross-sectional
tures than did the scaffold without amino acid addi-
structure of the chitosan/collagen composite scaffold
tion. Each composite scaffold had only one obvious
endothermic peak, indicating that there was no notice-
were examined by SEM. The surface pore structures
able phase separation in the composite scaffolds.
of the scaffolds are shown in Figure 3. All scaffold sur-
Shanmugasundaram et al.37 reported that chitosan/
faces were porous, demonstrating that the use of the
collagen has a transition band between 63 and 1688C.
freeze-gelation method could construct double-sided
The endothermic peak could result from loss of bound
porous scaffolds. In contrast, a dense layer at the sur-
water retained within the scaffold.38 Zhang et al.
face is produced by the traditional freeze-drying pro-
reported that the peak temperature for the chi-col-non
cedure.40
composite material shifts to a higher temperature
Figure 4 displays the cross-sectional structures of
because of the rigidity of the collagen and chitosan
the chitosan/collagen composite scaffolds. The entire
molecular chains,39 and our results demonstrate that
scaffold shows an interconnected porous structure
the addition of amino acids to the chi-col-non scaf-
that is con?rmed by the surface and cross-sectional
folds caused the peak temperature to shift to a higher
SEM images (Figs. 3 and 4). The pores are continuous
temperature. Our data thus suggest that covalent
and uniformly distributed. These results demonstrate
bonds are formed among the chitosan, collagen, and
that the use of the freeze-gelation method resulted in a
amino acids. The presence of covalent bonds could
uniformly distributed pore structure even with the
change the microstructure and porosity of the scaf-
addition of amino acids. We also observed that using
fold, and that may affect the amount of bound water
different amino acids as crosslinking bridges pro-
retained within the scaffold.
duced pores with different sizes in the composite
Journal of Applied Polymer Science DOI 10.1002/app

---

1780
TSAI ET AL.
Figure 4
Scanning electron microscopic (SEM) images (magni?cation, 100Â) showing the cross-sectional morphology of chi-
tosan/collagen composite scaffolds with or without the addition of amino acids. (a) chi-col-non scaffold; (b) chi-col-ala scaf-
fold; (c) chi-col-gly scaffold; (d) chi-col-glu scaffold. Scale bar ¼ 100 mm.
scaffold. From Figure 4, we observed that the chi-col-
for the chi-col-non and 108% for the chi scaffolds. The
non and chi-col-gly scaffolds had comparatively
post hoc Newman–Keuls test showed that the composi-
larger pores than did the chi-col-ala and chi-col-glu
tion (independent variable) had four signi?cantly dif-
scaffolds, but most pores were in the range of 50–250
ferent groups: chi and chi-col-non, chi-col-ala, chi-col-
mm. The pore size was determined by direct measure-
gly, and chi-col-glu. Thus, we concluded that the addi-
ment of 10 randomly chosen areas on the SEM images.
tion of amino acids to chitosan/collagen composite
Scaffolds used for skin tissue engineering should have
scaffolds signi?cantly increased the water uptake abil-
pore sizes of from 100 to 200 mm.41 Thus, our chito-
ity and made them more hydrophilic. In addition to
san/collagen composite scaffolds with amino acids as
the above post hoc Newman–Keuls test, the effect of
crosslinking bridges had appropriate pore sizes for
time on water uptake was analyzed, and we found
skin tissue engineering applications.
that after 20 min, the water uptake values of the chi-
col-gly, chi-col-ala, and chi-col-glu scaffolds was sig-
ni?cantly higher than those of the chi and chi-col-non
Water uptake of the scaffolds
scaffolds. The water uptake percentages of the chi and
Figure 5 shows the water uptake percentages of the
chi-col-non scaffolds had similar results compared
chitosan and chitosan/collagen composite scaffolds
with previous research.42 With the addition of amino
with or without the addition of amino acids. The
acids, the water uptake of the scaffolds increased by
water uptake was comparatively low for the chi and
more than 2 to 2.5 times. This might have been due to
chi-col-non scaffolds and signi?cantly increased for
the hydrophilic nature of the amino acids. For tissue
the chi-col-gly, chi-col-ala, and chi-col-glu scaffolds.
engineering applications, it is very important for
The ?nal water uptake reached 371% for the chi-col-
nutrients to be able to in?ltrate into porous scaffolds;
ala, 473% for the chi-col-gly, and 337% for the chi-col-
for this reason, a hydrophilic scaffold is more desira-
glu scaffolds; however, water uptake was only 135%
ble.26 Previous research also showed that a scaffold
Journal of Applied Polymer Science DOI 10.1002/app

---

SCAFFOLDS WITH AMINO ACIDS AS CROSSLINKING BRIDGES
1781
groups to amino groups signi?cantly affects the nor-
malized maximum tensile strength of scaffolds.
According to the molecular structure, the amino
groups on the chitosan and carboxyl and amino
groups on the amino acids and collagen all provide
possibilities for crosslinking (Fig. 1). Our group also
tried to use the conventional freeze-drying method to
fabricate porous chi-col-glu scaffolds (FD), and the
normalized maximum tensile strength of the FD scaf-
folds was around 2.3 N/g [Fig. 6(a)], similar to the
normalized maximum tensile strength of scaffolds
prepared by the freeze-gelation method. The statistical
tests showed that there was no signi?cant difference
in the normalized maximum tensile strength between
chi-col-glu and FD scaffolds [Fig. 6(a)].
Figure 5
Water uptake of chitosan and chitosan/collagen
composite scaffolds with or without the addition of amino
acids. (~) chi-col-gly, (!) chi-col-ala, (n) chi-col-glu, (*)
chi-col-non, and (l) chi. Error bars represent the mean 6 SD
for n ¼ 6; *P < 0.05 compared with the chi scaffold.
with water uptake of about 350% can be used in arti?-
cial skin application,43 and our prepared scaffolds also
fall in that range. This suggests that the addition of
amino acids enhances the hydrophilicity of the scaf-
folds and thus increases the percentage of water
uptake.
Tensile strength and elongation of porous scaffolds
Tensile strength and elongation of the scaffolds are
important for their application in tissue engineering.44
Soft scaffolds are suitable for brain cells, and stiff scaf-
folds are suitable for skin cells.45 Because the weights
of scaffolds of different compositions varied slightly,
we decided to report the load to fracture normalized
by the weight (normalized maximum tensile strength).
Figure 6 displays the normalized maximum tensile
strength and maximum tensile elongation of the chito-
san/collagen scaffolds with different amino acids as
crosslinking bridges. The normalized maximum ten-
sile strength of the chitosan/collagen scaffolds with-
out the addition of amino acids was low (0.7 N/g), but
with the addition of amino acids as crosslinking
bridges, the normalized maximum tensile strength of
the scaffolds increased by more than 1.5–2 times [Fig.
Figure 6
(a) Normalized maximum tensile strength of scaf-
6(a)]. We found that by adding alanine or glycine to
folds with or without the addition of different amino acids
the scaffolds, the normalized maximum tensile
as crosslinking bridges. The normalized maximum tensile
strength is de?ned as the load to fracture (maximum load)
strength was 1.9 N/g (chi-col-ala) and 1.8 N/g (chi-
divided by the weight of scaffold. The freeze-dried (FD)
col-gly). Furthermore, by adding glutamic acid to the
group contains chi-col-glu scaffolds fabricated by the con-
chitosan/collagen scaffold, the normalized maximum
ventional freeze-drying method. *P < 0.05 versus chi-col-
tensile strength reached 2.2 N/g. These results indi-
non scaffold. (b) Maximum tensile elongation (strain at max-
cated that adding amino acids to the chitosan/colla-
imum load) of composite scaffolds with or without the addi-
tion of different amino acids as crosslinking bridges. The FD
gen scaffolds signi?cantly increased the normalized
group contains chi-col-glu scaffolds fabricated by the con-
maximum tensile strength compared with chi-col-non
ventional freeze-drying method. Error bars represent the
scaffolds. Our data imply that the ratio of carboxyl
mean 6 SD for n ‡ 8. *P < 0.05 versus chi-col-non scaffold.
Journal of Applied Polymer Science DOI 10.1002/app

---

1782
TSAI ET AL.
Crosslinking using EDC/NHS consumes one car-
highest and lowest groups were within the range of
boxyl and one amino group each time. The concentra-
one standard deviation. Therefore, statistically, there
tion of amino groups ([NH2]) was 341.8 mM, includ-
was no signi?cant difference in the elongation proper-
ing [NH2] in chitosan of 190 mM plus that of the
ties of these scaffolds. Because the FD group contained
amino acids of 150 mM and that in collagen of 1.8
scaffolds fabricated by the traditional freeze-drying
mM, whereas the concentration of carboxyl groups
method, Figure 6(b) also demonstrates that the elonga-
([COOH]) was 151.8 mM in the chi-col-ala and chi-col-
tion properties of scaffolds prepared by the freeze-ge-
gly systems in which the amino acids had the number
lation method and those prepared by the freeze-dry-
ratio of COOH/NH2 of 1 (Table I). With the COOH/
ing method were similar.
NH2 ¼ 1 amino acid system, in the case where the car-
Our data also indicate that the tensile strength of
boxyl group on the amino acid ?rst forms a covalent
composite chitosan/collagen scaffolds with the addi-
bond with the amino group on the chitosan, the amino
tion of amino acids is signi?cantly higher than that of
acid has only one free amino group, and is thus unable
scaffolds prepared in a previous study,46 and thereby
to form covalent bonds with amino groups on the chi-
these composite scaffolds are suitable for skin tissue
tosan to continue the crosslinking reaction. With the
engineering applications.47
COOH/NH2 ¼ 2 amino acid system (such as chi-col-
glu with [NH2] ¼ 341.8 mM and [COOH] ¼ 301.8
mM) (Table I), one carboxyl group on the amino acid
In vitro enzymatic degradation
can ?rst be used to form a covalent bond with chito-
san, and the amino acid still has one free carboxyl
The in vitro enzymatic degradation of the scaffolds by
group that can form a covalent bond with another
lysozyme is shown in Figure 7. It can be seen that the
amino group. Because in the COOH/NH
remaining weight for the chi and chi-col-non scaffolds
2 ¼ 2 system,
the amino acid has a higher probability of continuing
was about 77% after 4 weeks of degradation. In con-
the crosslinking reaction, scaffolds prepared using the
trast, the remaining weights for the chi-col-ala, chi-
COOH/NH
col-gly, and chi-col-glu scaffolds were about 85% after
2
¼ 2 amino acid system would be
expected to have increased values for the normalized
4 weeks of degradation. The post hoc Newman–Keuls
maximum tensile strength.
test showed that the chi-col-non scaffold degraded
Although there are twice as many carboxyl groups
signi?cantly more slowly than the chi scaffold in
on glutamic acid than on alanine and glycine, the nor-
weeks 1 and 2, but exhibited no difference with the chi
malized maximum tensile strength of chi-col-glu scaf-
scaffold in weeks 3 and 4. In addition, the percent
fold, however, was only 15–20% higher. One possible
weights remaining of the chi-col-ala, chi-col-gly, and
reason may be that the EDC/NHS reagent mediates
chi-col-glu scaffolds were signi?cantly higher than
the crosslinking between the carboxyl and amino
those for the chi and chi-col-non scaffolds from weeks
groups, whether they are on chitosan, collagen, or
amino acids. Because the numbers of chitosan mole-
cules and amino acid molecules far exceeded those of
collagen
molecules,
most
crosslinking
occurred
between the amino acid and chitosan, and between
amino acids. However, other crosslinking reactions
still existed, and thus it is dif?cult to quantitatively
predict the increment in the normalized tensile
strength of scaffolds when using glutamic acid.
Both alanine and glycine belong to the COOH/NH2
¼ 1 system. The difference between alanine and gly-
cine is that one hydrogen atom of glycine is replaced
by a methyl group to become alanine. The normalized
maximum tensile strengths of the two systems were
similar. This indicates that the number ratio of car-
boxyl to amino groups signi?cantly in?uences the nor-
malized maximum tensile strength. However, the sub-
stituted group, like the methyl group, which cannot
react with EDC/NHS, does not signi?cantly in?uence
Figure 7
Percent weight remaining of chitosan/collagen
the normalized maximum tensile strength of the scaf-
composite scaffolds subjected to in vitro enzymatic degrada-
folds.
tion. The percent weights remaining of the chi-col-ala, chi-
col-gly, and chi-col-glu scaffolds were signi?cantly higher
The elongation data of chitosan/collagen composite
than those of the chi and chi-col-non scaffolds from weeks 1
scaffolds with or without the addition of amino acids
to 4. Error bars represent the mean 6 SD for n ¼ 5. *P < 0.05
as crosslinking bridges are shown in Figure 6(b). The
compared with the chi and chi-col-non scaffolds.
Journal of Applied Polymer Science DOI 10.1002/app

---

SCAFFOLDS WITH AMINO ACIDS AS CROSSLINKING BRIDGES
1783
Figure 8
Morphology of WS1 skin ?broblasts on various surfaces visualized by ?uorescence staining. (a) chi, day 3; (b) chi,
day 7; (c) chi-col-non, day 3; (d) chi-col-non, day 7; (e) chi-col-ala, day 3; (f) chi-col-ala, day 7; (g) chi-col-gly, day 3; (h) chi-col-
gly, day 7; (i) chi-col-glu, day 3; (j) chi-col-glu, day 7. Scale bar ¼ 200 mm.
Journal of Applied Polymer Science DOI 10.1002/app

---