Biomaterials

Science
View Article Online
PAPER View Journal | View Issue

Evaluation of absorbable hemostatic agents of

Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.

Cite this: Biomater. Sci., 2018, 6,

polyelectrolyte complexes using carboxymethyl
3332 starch and chitosan oligosaccharide both in vitro
and in vivo
Xingtao Chen, a Yonggang Yan,*a Hong Li,a Xuehong Wang,b Songchao Tang,b
Quan Li,c Jie Weib and Jiacan Su *c

Received 6th June 2018,
Accepted 2nd August 2018
DOI: 10.1039/c8bm00628h
rsc.li/biomaterials-science
Absorbable hemostatic agents with a high hemostatic efficacy play an important role in surgical and
severely traumatic hemostasis. In the present study, by applying polyelectrolyte assembly, polyelectrolyte
complexes (PECs), using carboxymethyl starch (CMS) and chitosan oligosaccharide (COS), with control-
lable physicochemical properties were prepared and optimized for hemorrhage control. Particle size, zeta
potential, morphology and water absorption of the PECs with different CMS/COS ratios were systemati-
cally evaluated. The results of in vitro degradation in PBS suggested that CMS/COS PECs were degradable
and their degradation rates, which decreased with the increase of the COS content, were suitable for
absorbable hemostatic agents. The in vivo hemostatic efficacy of the PECs with 10 wt% COS content
(PEC 10), which was evaluated in a rabbit hepatic hemorrhage model, was better than CMS but decreased
with the increase of the COS content. The plasma coagulation evaluation revealed that the PECs could
significantly activate and accelerate the coagulation cascade through both the intrinsic and extrinsic path-
ways but could not directly affect the common pathway. CMS/COS PECs also showed antimicrobial
activity against S. aureus, which enhanced with the increase of the COS content, but failed against E. coli.
Moreover, PEC 10 displayed excellent cytocompatibility with MC3T3-L1 and good tissue compatibility in a
rabbit liver model. These findings not only suggest that CMS/COS PECs with a suitable COS content were
promising absorbable hemostatic agents for internal use but they are also useful to understand the under-
lying principles for designing PEC based hemostatic agents.

1. Introduction by oxidized cellulose (e.g., Surgicel® and Oxycel®) limits its

Absorbable hemostatic agents have been widely developed for
surgical and severely traumatic bleeding, because in some
cases control of bleeding by pressure, ligature, and other con-
ventional procedures is ineffective or impractical.1 The
approved absorbable hemostatic agents, including collagen,
oxidized cellulose, chitosan and starch, have been investigated
and are commercially available.2,3 However, some issues with
regard to biological safety, potential infections, hemostatic
efficacy and biodegradation rate limit their wide use, and none
of these agents can be considered as the ideal absorbable
hemostatic agents for surgery. For instance, the low pH caused
a
College of Physical Science and Technology, Sichuan University, Chengdu 610064,
China. E-mail: yan_yonggang@vip.163.com
b
Key Laboratory for Ultrafine Materials of Ministry of Education, East China
University of Science and Technology, 200237 Shanghai, China
c
Department of Orthopaedics Trauma, Changhai Hospital, Second Military Medical
University, Shanghai 200433, China. E-mail: jiacansu@126.com
application in sensitive tissues such as nervous and cardiac
systems;4 chitosan (e.g., HemCon® and Gelox®) suffers from
its animal origin which may increase the risk of disease or
viral infection;5,6 starch microspheres (e.g., Arista®) are insuffi-
cient to stop severe bleeding and have no antimicrobial
activity.7–9 Therefore, it is necessary to find a more effective
agent with better hemostatic efficacy as well as biocompatibil-
ity, nonantigenicity, biodegradability and antimicrobial
activity.
Carboxymethyl starch (CMS) is a derivative of starch in
which some of the hydroxyl groups are modified into carboxy-
methyl groups. Owing to carboxymethylation, CMS is an
anionic polysaccharide and can be dissolved in aqueous solu-
tion at room temperature. It has been used as a biomaterial in
wound healing, regenerative medicine and drug delivery due
to its good biocompatibility, solubility and biodegradabil-
ity.10,11 Compared to starch, CMS is more biodegradable and
hydrophilic,12 which is favorable for application as an absorb-
able hemostatic agent. In addition, previous studies suggested

3332 | Biomater. Sci., 2018, 6, 3332–3344 This journal is © The Royal Society of Chemistry 2018

Biomaterials Science
that a negatively charged carboxyl group may directly activate
the coagulation cascade through the intrinsic pathway leading
to the accelerated formation of thrombin and fibrin clot, thus
being beneficial for hemostasis.13,14 Moreover, carboxyl in
CMS can be modified or incorporated with an active agent by
chemical bonding or electrostatic interaction to improve its
hemostatic efficacy (e.g., Ca2+)15 or endow with multifunction-
ality (e.g., Cu2+ and Ag+ for antimicrobial activity).16,17
Chitosan is a natural cationic polysaccharide derived by
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
partial deacetylation of chitin. The hemostatic activity and
safety of chitosan have been confirmed by many studies and
its FDA approved products are available on the market for
external use, but its application as an absorbable surgical
hemostatic agent is limited due to poor solubility at physio-
logical pH, insufficient antihemorrhagic effect, potential risk
inducing inflammatory response as well as intestinal and per-
itoneal adhesion and low degradation rate in vivo.5,6,18 One fre-
quently used strategy to overcome these drawbacks is the
chemical modification of the primary amines of chitosan.
Water-soluble chitosan derivatives (e.g., chitosan-catechol and
carboxymethyl chitosan) with enhanced hemostatic perform-
ance and biodegradability have been developed and used for
surgical hemostasis.19–23 An alternative strategy is the
reduction of the molecular weight of chitosan. As the degraded
products of chitosan, chitosan oligosaccharides (COS) have
been produced and their biological activities including anti-
inflammatory and antimicrobial activities have been investi-
gated in some research studies.24,25 In comparison to chitosan,
COS have improved biodegradability and solubility in aqueous
solutions due to a smaller molecular weight and showed
hemostatic and antimicrobial activities if the molecular weight
is controlled appropriately,26,27 thus may be being more suit-
able as an absorbable hemostatic agent.
Polyelectrolyte complexes (PECs) may be formed from CMS
and COS and held together by ion pairing interactions
between negatively charged carboxyl groups on CMS and posi-
tively charged amino groups on COS. The formation of these
polyvalent interactions, which is driven by the entropic release
of counter ions and water molecules, was termed polyelectro-
lyte assembly.28,29 It was reasonable to hypothesize that, by
applying polyelectrolyte assembly, PECs using negatively
charged CMS and positively charged COS could combine the
two polysaccharides that individually have hemostatic effects
and are more suitable as novel absorbable hemostatic agents
with improved hemostatic efficacy, biocompatibility, tunable
physicochemical properties and antimicrobial activity.
Furthermore, some research indicated that forming PECs
could improve tissue compatibility and cytotoxicity of the poly-
electrolyte via affecting its surface charge and charge
density.30,31 In the present study, COS was synthesized by oxi-
dative degradation using hydrogen peroxide and PECs with
various ratios of CMS to COS prepared by simply adding COS
solution to CMS solution with an emulsifying machine fol-
lowed by lyophilization. Since the physicochemical properties
of PECs strongly depend on parameters such as the molar
ratio, the influences of the CMS/COS ratio on water absorption,
This journal is © The Royal Society of Chemistry 2018
View Article Online
Paper
degradability, hemostatic efficacy, plasma coagulation, anti-
microbial activity and cytotoxicity of CMS/COS PECs were sys-
tematically evaluated and the potential for internal use as
hemostatic agents was studied in a rabbit hepatic hemorrhage
model.
2. Materials and methods
2.1 Synthesis of COS
COS were prepared by oxidative degradation of chitosan using
hydrogen peroxide according to previous studies with some
modifications.32 Low molecular weight chitosan (18 g)
(Aladdin Reagent, Shanghai) was suspended in 300 mL of a
hydrogen peroxide solution (6 wt%) under stirring for 1 h at
60 °C. After centrifugation, the supernatant containing COS
was collected and COS were precipitated by adding excess
ethanol to the supernatant. The COS precipitates were col-
lected by centrifugation, washed twice with ethanol, and then
dried under vacuum at 40 °C. The viscosity of the COS was
measured in a solution of NaCl (0.20 M) and CH3COOH (0.10
M) at 30 °C using an Ubbelohde viscometer and the molecular
weight was determined according to the classic Mark–
Houwink equation.33
2.2 Preparation and characterization of CMS/COS PECs
Aqueous stock solutions (2% w/v) of CMS (Mw = 1.5 × 105,
cassava source, degree of substitution (DS) = 0.4, Aladdin
Reagent, Shanghai) and COS were prepared by separately dis-
solving CMS and COS in distilled water with magnetic stirring.
The PEC solutions were prepared by mixing the two solutions
with a predetermined mass ratio (CMS to COS of 9 : 1, 4 : 1 and
7 : 3, abbreviated as PEC 10, PEC 20 and PEC 30, respectively)
using an emulsifying machine (JRJ300-S; Shanghai Specimen
Model Factory, Shanghai) at 4000 rpm for 40 min. The
acquired PEC gels were collected after centrifugation at 4000
rpm for 10 min and then washed three times with distilled
water before lyophilization.
Freeze-dried PEC granules were characterized by Fourier
transform infrared spectrometry (FTIR) (Nicolet 6700; Thermo
Fisher Scientific, Waltham, MA, USA). In order to determine
the particle size and zeta potential of the PECs in phosphate
buffer saline (PBS, pH = 7.4), the as-prepared PECs without
drying as well as CMS were adjusted at 7.4 by PBS and ana-
lyzed with a laser particle size analyzer (JL-1177; Chengdu
Jingxin Powder Analyzer Instruments Co., Chengdu, China) for
particle size distribution and laser Doppler electrophoresis
(Zetasizer Nano, Malvern Instruments, USA) for zeta potential.
The span value (SP) was used to evaluate size distribution and
calculated according to the formula,
SP ¼ ðD90  D10 Þ=D50
where D90, D10, and D50 are volume size diameters at 90, 10,
and 50% of the cumulative volume, respectively. A smaller SP
indicated narrower size distribution. The morphology of PEC
granules and lyophilized gel prepared by mixing PEC granules
Biomater. Sci., 2018, 6, 3332–3344 | 3333

Paper
and PBS solutions until reaching the equilibrium swelling
state was observed through scanning electron microscopy
(SEM) (JSM-6360LV, JEOL, Japan) after coating with gold in a
vacuum.
2.3 Water absorption and in vitro degradation
The water absorption capacity of the PECs was evaluated using
simulated body fluid (SBF) solution with ion concentrations
approximately equal to those of human blood plasma. Briefly,
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
PEC granules were vacuum dried at 50 °C overnight to elimin-
ate the residual water completely and then weighed (recorded
as W1) prior to immersing. Dried samples were dipped into a
centrifuge tube with 30 mL SBF for 10 min. Afterwards, the
samples were collected by air pump filtration until no fluid
dropped into the flask and then the weight was measured for
each sample (recorded as W2). The water absorption of each
sample was calculated according to the formula,
A% ¼ ½ðW 2  W 1 Þ=W 1   100%:
In vitro degradation behaviors were investigated by immer-
sing the samples into PBS at a weight to volume ratio of
0.4 g : 5 mL in plastic tubes, followed by incubating at 37 °C in
a shaker at 100 rpm for up to 21 days, with the solution being
refreshed every three days. At predetermined time intervals
(3, 7, 14, 21 days), the sample was collected after being centri-
fuged at 4000 rpm for 5 min, gently rinsed with deionized
water, and vacuum dried at 50 °C until a constant weight. The
weight of each sample before (W1) and after (W2) immersion
was accurately measured and the weight loss of each sample
was calculated according to the equation,
D% ¼ ½ðW 1  W 2 Þ=W 1   100%:
The pH values of the supernatants were tested with a pH
meter (PS-25; Shanghai Leici Chuangyi Apparatus and
Instrument Co., Shanghai, China).
2.4 Whole blood clotting
The whole blood clotting experiments were performed to
measure the blood-clotting efficiency of PECs according to pre-
vious studies with some modifications.3,5,34 Fresh blood was
obtained from New Zealand white rabbits and mixed with a
one-tenth volume of 3.8% (w/v) sodium citrate. The citrated
blood (300 µL) was dropped into a 15 mL tube containing 0.3 g
of hemostatic agent followed by vortex mixing the powders
with the blood well. Subsequently, CaCl2 solution (30 µL, 0.2
M) was added to the tube in order to initiate blood clotting,
and then incubated at 37 °C for various times (1, 2, 3, 4,
5 min). At each time point, deionized water (10 mL) was care-
fully added to each tube without disturbing the clot so as to
hemolyze the red blood cells (RBCs) which were not trapped in
the clot. The absorbance of hemoglobin in uncoagulated
blood was measured at 540 nm with a microplate reader. The
blood clot formed after 3 min incubation was collected, fixed
in 2.5% glutaraldehyde solution, dehydrated in a graded series
of ethanol and dried in a fume hood at room temperature for
SEM analysis. Commercial gelatin sponge (GS) was gently
3334 | Biomater. Sci., 2018, 6, 3332–3344
View Article Online
Biomaterials Science
ground into powders and tested by the same method for
comparison.
2.5 Plasma coagulation
To explore the effect of CMS/COS PECs on plasmatic coagu-
lation, various coagulation assays were performed according to
previous studies using a semi-automatic coagulation analyzer
(Biomerieux, France) including the plasma recalcification time
(PRT), activated partial thromboplastin time (APTT), prothrom-
bin time (PT) and thrombin time (TT).7,9,17,35 The platelet poor
plasma (PPP) and platelet rich plasma (PRP) were prepared by
the centrifugation of citrated rabbit blood at 3000 rpm for
10 min and at 1000 rpm for 5 min, respectively. For the PRT
test, the sample was mixed and incubated with 100 µL of PRP
at 37 °C for 2 min. Subsequently, the mixture was recalcified
with 100 µL of pre-incubated CaCl2 solution (25 mM) and the
PRT was recorded as the clotting time. Prior to APTT, PT and
TT tests, the samples were mixed and incubated with PPP at
37 °C for 2 min. The influence of hemostatic agents on APTT
was determined by mixing 100 µL of the mixture with 100 µL
of the APTT reagent, incubating for 4 min, and adding 100 µL
of CaCl2 solution (25 mM), in sequence. PT or TT under the
influence of samples was tested by mixing 100 µL of the
mixture with 200 µL of the PT or TT reagent, respectively. All
the reagents were preincubated at 37 °C for 2 min prior to
using. The same plasma without the addition of the hemo-
static agent was tested by the same method as the control.
2.6 Antimicrobial activity
The antimicrobial activity of the PECs against Escherichia coli
(E. coli, Gram-negative, ATCC 8739) and Staphylococcus aureus
(S. aureus, Gram-positive ATCC 6538) was investigated by a
modified colony counting method.17,36 Sterilized samples were
incubated with the bacteria suspension, which was adjusted to
the turbidity of the 0.5 McFarland standard followed by
dilution with sterile normal saline solution (0.85%, w/v) result-
ing in a concentration of about 5log CFU mL−1, at 37 °C up to
24 h in the vials with constant shaking. The suspension
(0.1 ml) obtained after a fixed incubating time (8, 16, 24 h) was
diluted, and then seeded on Luria–Bertani agar. After incu-
bation at 37 °C for additional 24 h, the resulting bacterial colo-
nies on the plates were counted. The growth inhibition of bac-
teria was expressed using the difference between the number
of colonies from bacteria with samples and that in the vial as
the control.
2.7 Cytotoxicity
Cytotoxicity was assayed using MC3T3-L1 fibroblasts by the
Cell Counting KIT-8 (CCK-8). The cells were cultured in a
24-well plate for the cell viability test and a confocal dish for
the observation of cell morphology, both in a 5% CO2 incuba-
tor at 37 °C. The culture medium was Dulbecco’s Modified
Eagle’s Medium (DMEM) containing the samples and sup-
plemented with 10% fetal bovine serum and 2% antibiotics
(200 mg mL−1 penicillin and 200 mg mL−1 streptomycin).
After 1, 4 and 7 days of incubation, the medium was replaced
This journal is © The Royal Society of Chemistry 2018

Biomaterials Science
by a culture medium containing 10% CCK-8 solution and the
cells were incubated under the same conditions for 3 h. The
absorbance of the culture medium was measured at 450 nm
and the cell viability was presented as the viability of cells rela-
tive to negative control (without the samples) at each time
point. At the same time point, the confocal dishes were
washed three times with PBS and the cultured cells were fixed
in 4% paraformaldehyde for 20 min for cell morphology obser-
vation. Prior to confocal laser scanning microscopy (CLSM)
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
(Nikon A1R, Japan) analysis, the fixed cells were stained with
rhodamine phalloidin (Cytoskeleton Inc., America) for 40 min
and 4′,6-diamidino-2-phenylindole (DAPI) (Beyotime Institute
of Biotechnology, China) for 10 min, sequentially.
2.8 Evaluation of hemostatic efficacy and tissue compatibility
in vivo
The hemostatic efficacy and tissue compatibility were evalu-
ated in a model of rabbit liver injury. With the approval of the
ethical committee of the Shanghai University of Traditional
Chinese Medicine, all animal experiments complied with the
ARRIVE guidelines37 and were carried out strictly in accord-
ance with the “Guide for the Care and Use of Laboratory
Animals” by the 2003 National Research Council. Thirty New
Zealand white rabbits (weight: 3–4 kg) were randomly assigned
to five groups (CMS, PEC 10, PEC 20, PEC 30 and GS group,
n = 6 per group) and were fasted 24 h before experiments. All
samples were vacuum dried at 50 °C overnight and then
gamma-ray sterilized (Heming Co. Ltd, Shanghai) for 2 h
before use.
Rabbits were anesthetized and cleaned. The abdomen was
opened to expose the liver, and then an “X” shaped incision
(approximately 15 × 15 mm with 10 mm depth) was created on
the left lateral hepatic lobe by a scalpel. Free bleeding was
allowed for 10 s, the hemostatic sample (1 g) was applied to
the bleeding site subsequently with manual compression for
10 s, and no additional hemostatic technique was used. The
hemostatic time was recorded when no blood flow was
observed and hemostasis was achieved (only the hemostasis
that could maintain 10 s was recorded), and then the abdomi-
nal cavity was sutured. Postoperative rabbits of CMS, PEC 10
and PEC 30 treatment groups were used for tissue compatibil-
ity evaluation and were individually housed, allowing free
access to water and diet, and other rabbits were euthanized. At
1 and 4 weeks after experiments, rabbits for tissue compatibil-
ity evaluation were euthanized and the injured sites of liver
were photographed and sampled, fixed in 10% buffered forma-
lin, dehydrated in grading ethanol, embedded in paraffin,
then serial sectioned at 5 mm and finally stained with hema-
toxylin and eosin (H&E) according to the protocol for micro-
scopic observation.
2.9 Statistical analysis
All samples were run in triplicate and the data points were
expressed as mean ± standard deviation (SD). Statistical ana-
lysis was performed using one-way ANOVA with the Bonferroni
This journal is © The Royal Society of Chemistry 2018
View Article Online
Paper
post hoc test, where p < 0.05 was considered statistically
significant.
3. Results
3.1 Characterization of CMS/COS PECs
In order to identify the interactions involved in CMS/COS
PECs, FTIR analysis was performed and the spectra are shown
in Fig. 1. In the spectrum of CMS, two strong peaks at 1606
and 1427 cm−1 corresponded to the asymmetric and sym-
metric stretching vibration of –COO− groups and a broad band
at 3447 cm−1 was ascribable to the stretching vibration of –OH
groups. COS exhibited a broad peak at 1629 cm−1 attributed to
N–H bending vibration, a sharp peak at 1384 cm−1 assigned to
the stretching vibration of C–N, and a broad band at
3409 cm−1 corresponding to O–H and N–H stretching
vibrations. Compared with the spectrum of CMS and COS, the
peaks corresponding to the asymmetric stretching vibration of
–COO− groups and bending vibration of N–H of the PECs
coupled and blue-shifted to 1603 cm−1 in PEC 10, 1600 cm−1
in PEC 20 and 1596 cm−1 in PEC 30, together with blue-shifts
of symmetric stretching vibration of –COO− groups to
1424 cm−1 in PEC 10, 1422 cm−1 in PEC 20 and 1421 cm−1 in
PEC 30, implying that ionic interaction occurred between CMS
and COS and enhanced with the increase of the COS
content.38,39 In addition, the band corresponding to O–H and
N–H stretching vibrations blue-shifted to 3419 cm−1 in PEC 10,
3416 cm−1 in PEC 20 and 3403 cm−1 in PEC 30, suggesting
that both ionic interaction and hydrogen bonding might be
involved in the PECs.
According to the viscosity of COS along with the Mark–
Houwink equation ([μ] = 1.8 × 10−3 M0.93), the molecular
weight of COS was ∼5 kDa. Table 1 depicts that the mean par-
ticle size and zeta potential ( pH = 7.4) of CMS was 11.2 μm
and −22.7 mV, respectively, which was similar to that of PEC
10 (12.4 μm and −20.6 mV, respectively). However, when the
Fig. 1 FTIR spectra of COS, CMS and CMS/COS PECs.
Biomater. Sci., 2018, 6, 3332–3344 | 3335
 View Article Online

Paper Biomaterials Science

Table 1 Size distribution, zeta potential and water absorption of CMS and CMS/COS PECs

Samples Mean particle size (μm) Span value Zeta potential (mV) Water absorption (%)

CMS 11.2 ± 1.5 1.55 ± 0.09 −22.7 ± 1.1 —

PEC 10 12.4 ± 1.1 1.54 ± 0.08 −20.6 ± 1.5 660 ± 34
PEC 20 25.1 ± 2.2 1.59 ± 0.15 −11.2 ± 1.3 471 ± 24
PEC 30 50.1 ± 3.8 1.65 ± 0.11 +4.8 ± 0.3 272 ± 16
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.

COS content was increased from 10% to 30%, the mean par-
ticle size and the zeta potential increased significantly (from
12.4 to 50.1 μm and from −20.6 to +4.8 mV, respectively),
suggesting that negatively charged CMS and positively charged
COS were aggregated into polyelectrolyte complexes with
varied ratios. The span of each sample was relatively large
(1.54–1.65) and no significance was found between them, indi-
cating a large and similar size distribution width of the
materials. The water absorption (Table 1) of CMS/COS PECs in
SBF was also strongly dependent on the ratio of CMS to COS.
CMS was miscible with SBF at any proportion, while the
absorption ratio of the PECs dropped remarkably with the
increase in the COS content. More specifically, the absorption
ratio of PEC 10 (660%) was much higher than that of PEC 20
(471%) and that of PEC 30 (272%). The SEM images of freeze-
dried PEC granules and gels are shown in Fig. 2. Some pores
and porous structures were found in the granules of PEC 10
and PEC 20, whereas those of PEC 30 showed fewer pores and
more solid and compacted morphology. The micrographs of
the lyophilized gels (Fig. 2B, D and F) depicted that all the PEC

Fig. 2 SEM micrographs of PEC 10 (A, B), PEC 20 (C, D) and PEC 30 (E, F) in the form of xerogel granules (A, C, E) and lyophilized gels prepared by
mixing PEC granules and PBS solutions (B, D, F) (scale bar: 10 μm for (A, C, E) and 50 μm for (B, D, F)).

3336 | Biomater. Sci., 2018, 6, 3332–3344 This journal is © The Royal Society of Chemistry 2018

Biomaterials Science
granules could coalesce and form integrated porous gel
matrixes when equilibrium absorbed PBS and, with an
increase of the COS content, smaller pores and less intercon-
nected porous structures were found, which was in agreement
with the results of water absorption.
3.2 Degradation in PBS
The degradation profiles of CMS and CMS/COS PECs in PBS,
which were expressed as the weight loss ratio as a function of
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
the incubation time, are presented in Fig. 3(A). CMS powders
were completely dissolved in PBS within one day and forming
PECs with COS significantly reduced their degradation rate. All
the PECs showed a rapid weight loss rate during the first week,
followed by gradually reduced degradation rate until the end
of incubation time. The degradation rates were highly depen-
dent on the ratio of CMS to COS and reduced significantly
with a higher COS content in the PECs. After 21 days of incu-
bation, the weight loss of PEC 10 was 92%, while that of PEC
20 and PEC 30 was 77% and 65%, respectively. The changes of
pH values in the PBS are shown in Fig. 3(B). The pH values of
all the PECs dropped to approximately 7.0 within 3 days and
Fig. 3 Weight loss (A) and pH change (B) of PBS after immersion of CMS and CMS/COS PECs.
Fig. 4 SEM observations of the stable clot formed after 3 min treatment with CMS (A), PEC 10 (B), PEC 20 (C) and PEC 30 (D) (scale bar: 10 μm); (E)
whole blood clotting efficiency.
This journal is © The Royal Society of Chemistry 2018
View Article Online
Paper
recovered slowly to 7.4 after 21 days. No significant difference
in pH values was observed among the PECs with different
ratios of CMS to COS. The results suggested that the degra-
dation rate of the PECs could be regulated by changing the
ratio of CMS to COS without affecting the pH values.
3.3 Whole blood clotting
The blood clotting experiments were performed to evaluate
blood clotting efficiency of the PECs and study how they inter-
acted with red blood cells (RBCs). As illustrated in Fig. 4(E),
the absorbance at 540 nm corresponded to the amount of red
blood cells that were not trapped in the clot at each time
point, which means the samples with lower absorbance have
higher efficiency to form clots and trap RBCs. All the absor-
bance profiles showed a dramatic reduction within 3 min
because of proceeding of blood clotting, and then became flat
when relatively stable blood clots were formed. The highest
blood clotting efficiency (lowest absorbance at 540 nm) was
found with both CMS and PEC 10, indicating that forming
PECs with a small amount of COS would not affect the blood
clotting capacity. However, when the COS content in the PECs
Biomater. Sci., 2018, 6, 3332–3344 | 3337

Paper
increased to 20% and 30%, the absorbance increased signifi-
cantly meaning reduced blood clotting capacity. SEM obser-
vations of the clot formed after 3 min treatment with CMS and
the PECs are shown in Fig. 4(A–D). The disc-shaped RBCs
adhered to all the materials. More RBCs were found with
regard to CMS and PEC 10 compared with other PECs, which
was in agreement with the results of the OD value measure-
ments. It was interesting to note that the RBCs adhered to the
PECs seemed to expand and lose their typical biconcave mor-
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
phology, which may be attributed to the PEC gels tightly
adhered to the surface of the RBSs. Particularly, in the PEC
30 group the RBSs were completely enwrapped by PEC 30 gels
and became a part of the gel matrix.
3.4 PRT, APTT, PT, and TT
Effects of the PECs on plasma coagulation were investigated by
PRT, APTT, PT and TT measurements. As shown in Fig. 5(A, B
and C), the PRT, APTT and PT values were significantly
reduced by CMS and CMS/COS PECs compared with the
control (100%, without any agent). CMS and PEC 10 got the
lowest values without a significant difference, while increasing
the COS content in the PECs (to 20% and 30%) led to pro-
longed PRT, APTT and PT, indicating impaired effects on accel-
erating plasma coagulation through the intrinsic pathway (PRT
and APTT) and extrinsic pathway (PT). As for TT shown in
Fig. 5(D), no significant difference between these materials
and the control was observed, which indicated that CMS and
the PECs did not directly affect the common pathway.
Fig. 5 Effects of CMS and CMS/COS PECs on plasma recalcification time (PRT) (A), activated partial thromboplastin time (APTT) (B), prothrombin
time (PT) (C) and thrombin time (TT) (D). *p < 0.05 compared with control. **p < 0.05 compared with PEC 10.
3338 | Biomater. Sci., 2018, 6, 3332–3344
View Article Online
Biomaterials Science
3.5 Antimicrobial activity
Antimicrobial activities of CMS and CMS/COS PECs against
S. aureus are shown in Fig. 6. CMS showed no antimicrobial
activity, while the PECs exhibited different degrees of bacter-
iostasis depending on the ratio of CMS to COS. With an
increase in the COS content in the PECs, the antibacterial
rate increased correspondingly at each time point. In
addition, the antibacterial rate of PEC 20 and PEC 30
increased with the incubation time reaching 51% and 83% at
24 h, respectively, which could be due to the sustained
release of COS from the PECs, while that of PEC 10 increased
within 16 h and exhibited a slight reduction at 24 h which
could be attributed to insufficient COS. The digital images of
the resulting S. aureus colonies after incubation with CMS
and the PECs (Fig. 6A–D) are also presented, which were con-
sistent with the results of Fig. 6E. Neither CMS nor CMS/COS
PECs showed significant antimicrobial activity against E. coli
(data not shown).
3.6 Cytotoxicity
Fig. 7(A) shows the cell viabilities of MC3T3-L1 fibroblasts
incubated with CMS or CMS/COS PECs, which were deter-
mined by CCK-8 assays and normalized with the negative
control (without any agent). PEC 30 exerted mild effects on
MC3T3-L1 fibroblasts and the cell viability dropped from 95%
to 86% when the culture time increased from 1 to 7 days. The
relative cell viabilities retained slightly higher than 100% for
CMS, PEC 10 and PEC 20 at each time point and no evident
This journal is © The Royal Society of Chemistry 2018
 View Article Online

Biomaterials Science Paper

Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.

Fig. 6 Antimicrobial activity of CMS and CMS/COS PECs against S. aureus. The resulting bacterial colonies incubated with CMS (A), PEC 10 (B), PEC
20 (C) and PEC 30 (D) for 24 h. (E) Antibacterial ratio against S. aureus at 8, 16 and 24 h.

Fig. 7 The viability (A) of MC3T3-L1 fibroblasts incubated with CMS and CMS/COS PECs and CLSM images of the cells for PEC 10 (B), PEC 20 (C),
PEC 30 (D) after 7 days of cultivation (scale bar: 50 μm).

cytotoxic effect was observed, indicating good cytocompatibil- blast characteristics with a bipolar structure were found for
ity. The cell morphology for CMS/COS PECs at day 7 was PEC 10 and PEC 20 groups, while the cell morphology for the
observed by CLSM and is shown in Fig. 7(B–D). Typical fibro- PEC 30 group was more spherical and not very well spread

This journal is © The Royal Society of Chemistry 2018 Biomater. Sci., 2018, 6, 3332–3344 | 3339

Paper
which could be attributed to the mild effect of PEC 30 on the
cells.
3.7 Hemostatic efficacy and tissue compatibility in a rabbit
hepatic hemorrhage model
Fig. 8(A–D) illustrate the hemostasis model of rabbit hepatic
hemorrhage. After applying CMS/COS PECs on the injured
liver (Fig. 8A), a sticky gel matrix (Fig. 8C) which further trans-
formed into hemostasis clots (Fig. 8D) was observed and
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
hemostasis was achieved rapidly. The hemostatic time
(Fig. 8E) of CMS and CMS/COS PECs was significantly shorter
than that of commercial GS (121 s). More specifically, the
hemostatic time of PEC 10 (48 s) was significantly shorter than
CMS (61 s), PEC 20 (67 s) and PEC 30 (82 s).
Fig. 9 shows the typical gross views and H&E staining of
liver defects treated with PEC 10, PEC 30 and CMS after 1 and
4 weeks, respectively. For the PEC 10 group, extremely mild
Fig. 8 Hemostasis in a rabbit hepatic hemorrhage model. (A) Incision and profuse bleeding of the left lobe of the liver. (B) Treatment with PEC 10
on the bleeding site. (C) Sticky gel matrix and (D) subsequent hemostasis clots created by PEC 10. (E) Corresponding hemostatic time. *p < 0.05
compared with GS. **p < 0.05 compared with CMS, PEC 20 or PEC 30.
3340 | Biomater. Sci., 2018, 6, 3332–3344
View Article Online
Biomaterials Science
inflammatory response was found with some residuals of PEC
10 in the 1st week (Fig. 9a and g) and disappeared with its
complete degradation in the 4th week (Fig. 9d and j), indicat-
ing good tissue compatibility. The incision was surrounded by
a small amount of fibrous tissues in the 1st week (Fig. 9a), and
after 4 weeks, no fibrous tissue was observed around the
incision (Fig. 9d). The CMS group (Fig. 9c, i, f and l) showed a
similar tissue response compared to the PEC 10 group with
slightly worse wound healing performance, while acute inflam-
mation was induced and adhesion between the two liver lobes
mediated by plenty of fibrous tissues occurred for the PEC
30 group during the 1st week (Fig. 9b and h) and then sub-
sided in the 4th week (Fig. 9e and k). For the PEC 30 group, it
is worth noting that owing to the relatively low degradation
rate and persistent release of COS with PEC disassembly,
inflammatory cells were still observed and the tissue response
remained after 4 weeks (Fig. 9k).
This journal is © The Royal Society of Chemistry 2018

Biomaterials Science
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
Fig. 9 Gross views (a–f ) and H&E staining (g–l, ×100, scale bar: 100 μm) of liver defects treated with PEC 10 (a, d, g, j), PEC 30 (b, e, h, k) and CMS
(c, f, i, l) after 1 (a–c, g–i) and 4 (d–f, j–l) weeks.
4. Discussion
Starch based hemostatic agents have been investigated by
many studies, and their hemostatic efficacy mainly results
from their function as a molecular filter by separating serum
from the cellular constituents such as platelets and erythro-
cytes. They can absorb much of the water portion of the blood
rapidly leading to concentrated blood solids, which then form
a gel matrix. The gel matrix slows blood flow and serves to
enhance blood clotting.7,8,40 Therefore, water absorption
capacity is the key factor responsible for the hemostatic
efficacy of starch based hemostatic agent. In the present study,
both CMS and COS could be completely dissolved in water,
while CMS/COS PECs were insoluble due to electrostatic inter-
actions between CMS and COS which were supported by the
results of FTIR analysis (Fig. 1). Owing to the excellent hydro-
This journal is © The Royal Society of Chemistry 2018
View Article Online
Paper
philicity of CMS and COS, CMS/COS PECs, especially PEC 10,
were also highly hydrophilic and had excellent water absorp-
tion capacity which were essential for hemostasis application.
Generally, the structure and properties of PECs strongly
depend on parameters such as molar ratio, pH and ionic
strength. Highly hydrophilic PEC microgels with small particle
sizes, high water content and large negative or positive charge
values are obtained with an excess of polyanionic or polycatio-
nic charges, respectively. In contrast, insoluble PEC precipi-
tates with large particle sizes, low water content and near zero
total charges are derived from equal charge molar ratios of
polyelectrolytes.41–43 In our study, the water absorption, par-
ticle size, zeta potential and degradation rate of CMS/COS
PECs were highly dependent on the ratio of CMS to COS,
which can be explained by enhanced electrostatic interactions
with the increase of the COS content (supported by FTIR
Biomater. Sci., 2018, 6, 3332–3344 | 3341

Paper
results). Therefore, the performance of CMS/COS PECs on
hemostasis could also be strongly influenced by the CMS/COS
ratio as confirmed by the results of hemostatic efficacy. In
addition, the results of in vitro degradation in PBS suggested
that CMS/COS PECs were degradable and their degradation
rates decreased with the increase of the COS content (Fig. 3).
Furthermore, their degradation rates were suitable when used
as absorbable hemostatic agents, while rapidly dissolved CMS
may easily enter a blood vessel leading to general or systemic
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
emboli and thrombosis.13
A previous study suggested that material-induced blood
plasma coagulation is a surface-mediated event, which is
related to both the surface area for immobilizing participants
of surface-dependent clotting reactions and the surface pro-
perties for activating the intrinsic pathway of the blood clotting
cascade of hemostatic agents.13,44 More recently, material-
induced rapid blood coagulation was also achieved by chito-
san–catechol via interacting with plasma proteins.23 For hemo-
static application, PEC 10 particles with a smaller particle size
(12.4 μm) and larger negative charge value (−20.6 mV) were
more propitious to contact activation of blood leading to accel-
erated plasma coagulation demonstrated by the results of
whole blood clotting (Fig. 4), with respect to PEC 20 and PEC
30.
Since blood plasma coagulation is the key phase during
hemostasis, effects of CMS and CMS/COS PECs on plasma
coagulation including PRT, APTT, PT and TT were determined
(Fig. 5) in order to further understand the hemostatic mecha-
nisms. PRT and APTT are both evaluation for the intrinsic
coagulation pathway, while the influence of platelets is
involved for the PRT and not involved for the APTT.45 Both the
PRT and APTT were significantly shortened for CMS and the
PECs compared with the control, and the increase of the COS
content in the PECs resulted in a prolonged PRT and APTT,
which along with the results of particle size and zeta potential
were in agreement with the principles of material-induced
contact activation. Interestingly, although previous research
reported that the hemostatic activity of chitosan and COS is
related to platelet activation, our study suggested that the
plasma coagulation process was mainly activated and acceler-
ated by negatively charged CMS rather than positively charged
COS since both the PRT and APTT of PEC 10 were significantly
lower than those of PEC 30. A similar trend was observed for
PT, which is evaluation for the extrinsic coagulation pathway,
indicating that CMS could also accelerate the extrinsic coagu-
lation pathway. As for TT, which is evaluation for the common
coagulation pathway, all the samples did not show significant
impact on TT. Based on this evidence, it could be concluded
that CMS could accelerate both the intrinsic and extrinsic
coagulation pathways and not directly affect the common
pathway, while the effect of COS on plasma coagulation was
not that dramatic.
Typical hemostasis refers to the process of halting the
bleeding from damaged blood vessels and proceeds in several
sequential phases, including constriction of blood vessels,
activation of a coagulation cascade and formation of a blood
3342 | Biomater. Sci., 2018, 6, 3332–3344
View Article Online
Biomaterials Science
clot. So the efforts to accelerate any of the above phases can be
helpful in achieving a high hemostatic efficacy.35,46 The hemo-
static efficacy of CMS and CMS/COS PECs was evaluated
in vitro by whole blood clotting assay (Fig. 4) and in vivo in a
rabbit hepatic hemorrhage model (Fig. 8) in consideration of
the fact that the liver is the most representative organ due to
its abundant blood supply and sensitivity to traumatic hemor-
rhaging. The highest hemostatic efficacy both in vitro and
in vivo was achieved by PEC 10, which was consistent with the
results of water absorption (highest), particle size (smallest)
and zeta potential (most negatively charged). Interestingly, no
significant difference was found between CMS and PEC 10 in
the whole blood clotting assay, whereas the hemostatic time
for PEC 10 in vivo was significantly shorter than CMS. Our
hypothesis is that CMS and PEC 10 showed similar surface
properties (Table 1, Fig. 4A and B) for material-induced blood
plasma coagulation leading to similar results of blood clotting
assay (Fig. 4E);1,13,14 in the case of in vivo hemostasis, highly
hydrophilic but insoluble PEC 10 could form a sticky gel
matrix to adhere to the injured site and occlude the blood
vessel rapidly (Fig. 8), while soluble CMS dissolved in blood
and couldn’t occlude the blood vessel until fibrin clots were
formed.19,23
In summary, the hemostatic process of CMS/COS PECs
could be depicted as follows. CMS/COS PECs could absorb
much of the water portion of the blood rapidly leading to a
sticky gel matrix which could tightly adhere and occlude the
injured site to stop bleeding efficiently. Subsequently, the gel
matrix could accelerate the coagulation process by providing
negatively charged and hydrophilic surfaces which are
regarded as key factors to activate the intrinsic pathway of the
coagulation cascade. Ultimately, stable fibrin clots were
formed to reinforce the primary occlusion and hemostasis was
achieved.
Since infection is often induced by trauma and bleeding,
hemostatic agents with antimicrobial activity can be
helpful.5,17,47 The antimicrobial activity of COS has been con-
firmed by previous studies, and can be explained by altered
permeability of the microbial cell membrane when positively
charged COS bind to the bacterial cell wall.24,48 CMS/COS
PECs also showed antimicrobial activity against S. aureus
(Fig. 6) but failed against E. coli. Unsurprisingly, the anti-
microbial activity against S. aureus enhanced with the increase
of the COS content, suggesting that the antimicrobial activity
derived from COS and the contribution of CMS were
unapparent.
In addition to hemostatic efficacy and antimicrobial
activity, another important issue was biological safety for an
absorbable hemostatic agent. The cytotoxicities of CMS and
CMS/COS PECs were evaluated with MC3T3-L1 fibroblasts
using CCK-8. The obtained data revealed that only PEC 30
showed a mild effect on the proliferation and morphology of
MC3T3-L1 (Fig. 7), which might be due to the excessive
amount of COS released to the cell culture medium resulting
in too high cation concentration for cells to grow. Besides, no
evident cytotoxic effect was observed for CMS, PEC 10 and
This journal is © The Royal Society of Chemistry 2018

Biomaterials Science
PEC 20, indicating good cell compatibility. In accordance
with the cytotoxicity results of CMS/COS PECs, PEC 10
showed good tissue compatibility with extremely mild inflam-
matory response which weakened over time, while the inten-
sity of the inflammatory response increased remarkably with
the increase of the COS content (PEC 30). The differences
between PEC 10 and PEC 30 can be explained by different
contents of COS, which can cause inflammatory tissue
responses and tissue adhesion,3,18 and different degradation
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
rates. PEC 10 with a low COS content and fast degradation
rate resulted in a slight and fast release of COS leading to
extremely mild inflammation weakened and eventually dis-
appeared, whereas PEC 30 with a high COS content and low
degradation rate resulted in a plentiful and long-term release
of COS leading to acute inflammation and adhesion which
were still observed after 4 weeks. Interestingly, in consider-
ation of the good cytocompatibility and tissue compatibility
of CMS confirmed by this research, these in vitro and in vivo
results also suggested that a large amount of water-soluble
COS could cause mild cytotoxicity and induce intense inflam-
mation and adhesion despite of the accepted good biocom-
patibility of chitosan.
5. Conclusion
Three CMS/COS PECs with different ratios of CMS to COS
(9/1, 4/1, 7/3) have been developed as absorbable hemostatic
agents for surgery and their hemostatic efficacy, hemostatic
mechanism and biocompatibility were evaluated. The results
demonstrated that PEC 10 displayed superior hemostatic
efficacy but reduced with the increase of the COS content.
The results of whole blood clotting, plasma coagulation and
hemostasis in a rabbit hepatic hemorrhage model revealed
that the possible hemostasis mechanism of CMS/COS PECs
was attributed to (1) physical absorption of the fluid com-
ponent of blood, (2) occlusion of the bleeding sites by
forming a sticky gel matrix, and (3) activation and accelera-
tion of the coagulation cascade by negatively charged and
hydrophilic surfaces. To our knowledge, these findings
provide the first evaluation and optimization of CMS/COS
PECs for hemostatic applications and are useful to under-
stand the hemostasis mechanism and underlying principles
for designing PEC based hemostatic agents. The degradation
rates, which decreased with the increase of the COS content,
were also suitable for absorbable hemostatic agents. In
addition, PEC 10 displayed excellent cytocompatibility with
MC3T3-L1 and good tissue compatibility with a rabbit liver.
CMS/COS PECs also showed antibacterial activities against
S. aureus, which were enhanced with the increase of the COS
content. Thus, it is reasonable to conclude that CMS/COS
PEC with a suitable COS content (e.g., PEC 10) has a great
potential for internal use as an absorbable hemostatic agent
and the above findings will open a new door for designing a
host of PEC based hemostatic agents with great potential in
clinical applications.
This journal is © The Royal Society of Chemistry 2018
View Article Online
Paper
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (51772194, 51502340 and 81771990) and
Key Medical Program of Science and Technology Development
of Shanghai (17441900600 and 17441902000).
References
1 T. A. Ostomel, Q. H. Shi and G. D. Stucky, J. Am. Chem.
Soc., 2006, 128, 8384–8385.
2 M. Gabay, Am. J. Health-Syst. Pharm., 2006, 63, 1244–1253.
3 S. Pourshahrestani, E. Zeimaran, N. A. Kadri, N. Gargiulo,
H. M. Jindal, S. V. Naveen, S. D. Sekaran, T. Kamarul and
M. R. Towler, ACS Appl. Mater. Interfaces, 2017, 9, 31381–
31392.
4 S. Ohta, T. Nishiyama, M. Sakoda, K. Machioka, M. Fuke,
S. Ichimura, F. Inagaki, A. Shimizu, K. Hasegawa,
N. Kokudo, M. Kaneko, Y. Yatomi and T. Ito, J. Biosci.
Bioeng., 2015, 119, 718–723.
5 S. Y. Ong, J. Wu, S. M. Moochhala, M. H. Tan and J. Lu,
Biomaterials, 2008, 29, 4323–4332.
6 R. L. Gu, W. Z. Sun, H. Zhou, Z. N. Wu, Z. Y. Meng,
X. X. Zhu, Q. Tang, J. Dong and G. F. Dou, Biomaterials,
2010, 31, 1270–1277.
7 F. P. Chen, X. Y. Cao, X. L. Chen, J. Wei and C. S. Liu,
J. Mater. Chem. B, 2015, 3, 4017–4026.
8 K. Bjorses, L. Faxalv, C. Montan, K. Wildt-Persson, P. Fyhr,
J. Holst and T. L. Lindahl, Acta Biomater., 2011, 7, 2558–
2565.
9 Y. Hou, Y. Xia, Y. K. Pan, S. C. Tang, X. F. Sun, Y. Xie,
H. Guo and J. Wei, Mater. Sci. Eng., C, 2017, 76, 340–349.
10 E. Assaad and M. A. Mateescu, Int. J. Pharm., 2010, 394, 75–84.
11 N. Masina, Y. E. Choonara, P. Kumar, L. C. du Toit,
M. Govender, S. Indermun and V. Pillay, Carbohydr. Polym.,
2017, 157, 1226–1236.
12 L. P. Massicotte, W. E. Baille and M. A. Mateescu,
Int. J. Pharm., 2008, 356, 212–223.
13 X. Yang, W. Liu, N. Li, M. Wang, B. Liang, I. Ullah,
A. L. Neve, Y. Feng, H. Chen and C. C. Shi, Biomater. Sci.,
2017, 5, 2357–2368.
14 T. A. Ostomel, Q. H. Shi, P. K. Stoimenov and G. D. Stucky,
Langmuir, 2007, 23, 11233–11238.
15 Y. X. Chen and G. Y. Wang, Colloids Surf., A, 2006, 289, 75–
83.
16 A. Bari, N. Bloise, S. Fiorilli, G. Novajra, M. Vallet-Regí,
G. Bruni, A. Torres-Pardo, J. M. González-Calbet, L. Visai
and C. Vitale-Brovarone, Acta Biomater., 2017, 55, 493–504.
17 C. L. Dai, Y. Yuan, C. S. Liu, J. Wei, H. Hong, X. S. Li and
X. H. Pan, Biomaterials, 2009, 30, 5364–5375.
Biomater. Sci., 2018, 6, 3332–3344 | 3343

Paper
18 X. F. Huang, Y. F. Sun, J. Y. Nie, W. T. Lu, L. Yang,
Z. L. Zhang, H. P. Yin, Z. K. Wang and Q. L. Hu, Int. J. Biol.
Macromol., 2015, 75, 322–329.
19 J. H. Ryu, Y. Lee, W. H. Kong, T. G. Kim, T. G. Park and
H. Lee, Biomacromolecules, 2011, 12, 2653–2659.
20 K. Kim, J. H. Ryu, D. Y. Lee and H. Lee, Biomater. Sci.,
2013, 1, 783–790.
21 J. M. Lee, J. H. Ryu, E. A. Kim, S. Jo, B.-S. Kim, H. Lee and
G.-I. Im, Biomaterials, 2015, 39, 173–181.
Published on 18 September 2018. Downloaded by Kings College London on 1/21/2019 2:26:23 AM.
22 Q. Xia, Z. Liu, C. Wang, Z. Zhang, S. Xu and C. C. Han,
Biomacromolecules, 2015, 16, 3083–3092.
23 M. Shin, S. G. Park, B. C. Oh, K. Kim, S. JO, M. S. Lee,
S. S. Oh, S. H. Hong, E. C. Shin, K. S. Kim, S. W. Kang and
H. Lee, Nat. Mater., 2017, 16, 147–152.
24 P. Zou, X. Yang, J. Wang, Y. F. Li, H. L. Yu, Y. X. Zhang and
G. Y. Liu, Food Chem., 2016, 190, 1174–1181.
25 H. K. No, N. Y. Park, S. H. Lee and S. P. Meyers, Int. J. Food
Microbiol., 2002, 74, 65–72.
26 J. Y. Kim, J. K. Lee, T. S. Lee and W. H. Park, Int. J. Biol.
Macromol., 2003, 32, 23–27.
27 B. C. Lee, M. S. Kim, S. H. Choi, K. Y. Kim and T. S. Kim,
Int. J. Mol. Med., 2009, 24, 327–333.
28 C. B. Bucur, Z. Sui and J. B. Schlenoff, J. Am. Chem. Soc.,
2006, 128, 13690–13691.
29 P. Schaaf and J. B. Schlenoff, Adv. Mater., 2015, 27, 2420–
2432.
30 M. Nagahata, R. Nakaoka, A. Teramoto, K. Abe and
T. Tsuchiya, Biomaterials, 2005, 26, 5138–5144.
31 J. S. Martinez, T. C. S. Keller and J. B. Schlenoff,
Biomacromolecules, 2011, 12, 4063–4070.
32 Y. Lu, D. L. Slomberg and M. H. Schoenfisch, Biomaterials,
2014, 35, 1716–1724.
33 G. G. Maghami and G. A. F. Roberts, Makromol. Chem.,
1988, 189, 195–200.
3344 | Biomater. Sci., 2018, 6, 3332–3344
View Article Online
Biomaterials Science
34 B. K. Gu, S. J. Park, M. S. Kim, C. M. Kang, J. I. Kim and
C. H. Kim, Carbohydr. Polym., 2013, 97, 65–73.
35 C. L. Dai, C. S. Liu, J. Wei, H. Hong and Q. H. Zhao,
Biomaterials, 2010, 31, 7620–7630.
36 H. Wang, M. Li, J. M. Hu, C. H. Wang, S. S. Xu and
C. C. Han, Biomacromolecules, 2013, 14, 954–961.
37 C. Kilkenny, W. J. Browne, I. C. Cuthill, M. Emerson and
D. G. Altman, Osteoarthritis Cartilage, 2012, 20,
256–260.
38 Q. Zhao, J. Qian, Q. An, C. Gao, Z. Gui and H. Jin, J. Membr.
Sci., 2009, 333, 68–78.
39 M. S. Belluzo, L. F. Medina, A. M. Cortizo and
M. S. Cortizoa, Ultrason. Sonochem., 2016, 30, 1–8.
40 P. Rajagopal and N. Hakim, Transplant. Proc., 2011, 43,
424–426.
41 H. V. Saether, H. K. Holme, G. Maurstad, O. Smidsrød and
B. T. Stokke, Carbohydr. Polym., 2008, 74, 813–821.
42 D. L. Cerf, A. S. Pepin, P. M. Niang, M. Cristea,
C. Karakasyan-Dia and L. Picton, Carbohydr. Polym., 2014,
113, 217–224.
43 E. Spruijt, A. H. Westphalm, J. W. Borst, M. A. C. Stuart and
J. Gucht, Macromolecules, 2010, 43, 6476–6484.
44 Z. Guo, K. M. Bussard, K. Chatterjee, R. Miller,
E. A. Vogler and C. A. Siedlecki, Biomaterials, 2006, 27, 796–
806.
45 C. Q. Xu, Y. J. Zeng and H. Gregersen, Med. Eng. Phys.,
2002, 24, 587–593.
46 T. A. Ostomel, Q. Shi, C. K. Tsung, H. J. Liang and
G. D. Stucky, Small, 2006, 2, 1261–1265.
47 S. Pourshahrestani, E. Zeimaran, N. A. Kadri, N. Gargiulo,
S. Samuel, S. V. Naveen, T. Kamarul and M. R. Towler,
J. Mater. Chem. B, 2016, 4, 71–86.
48 C. Qin, Y. Zhang, W. Liu, L. Xu, Y. Yang and Z. Zhou, Fish
Shellfish Immunol., 2014, 40, 267–274.
This journal is © The Royal Society of Chemistry 2018