Journal of Controlled Release 69 (2000) 225–236

www.elsevier.com / locate / jconrel

Self-assembled hydrogel nanoparticles from curdlan derivatives:

characterization, anti-cancer drug release and interaction with a
hepatoma cell line (HepG2)
Kun Na, Keun-Hong Park, Sung Won Kim, You Han Bae*
Center for Biomaterial and Biotechnology, Department of Materials Science and Engineering,
Kwangju Institute of Science and Technology, 1 Oryong-Dong, Puk-gu, Kwangju 500 -712, South Korea

Received 30 November 1999; accepted 24 February 2000

Abstract

Self-assembled hydrogel nanoparticles were synthesized from carboxymethylated (CM)-curdlan, substituted with a
sulfonylurea (SU) as a hydrophobic moiety for self-assembly. The degree of SU substitution was 2.4, 5.6, or 7.2 SU groups
per hundred anhydroglucose units of curdlan. The physicochemical properties of the self-assembled hydrogel nanoparticles
(DS 2.4, DS 5.6, and DS 7.2) in aqueous media were characterized by dynamic light scattering, transmission electron
microscopy, and fluorescence spectroscopy. The mean diameter of all samples was less than 300 nm with a unimodal size
distribution. The critical aggregation concentrations (CAC) of self-assembled hydrogel nanoparticles in distilled water were
4.2310 22 , 3.1310 22 and 1.9310 22 mg / ml for DS 2.4, 5.6 and 7.2, respectively. The loading and release of all-trans
retinoic acid (ATRA) was studied. The ATRA loading efficiencies and loading contents of CM-curdlan / SU nanoparticles
increased as the degree of SU substitution increased. The ATRA release rate was controlled by the degree of substitution and
drug-loading. For specific interaction with a hepatic carcinoma cell line (HepG2), CM-curdlan was additionally conjugated
with lactobionic acid (LBA; galactose moiety) (5.5 LBA molecules per hundred glucose units). HepG2 was strongly
luminated by ligand–receptor interactions with fluorescence-labeled LBA / CM-curdlan / SU hydrogel nanoparticles. The
luminescence was not observed for other control cases. It is concluded that LBA / CM-curdlan / SU hydrogel nanoparticles are
a useful drug carrier for the treatment of liver cancer, because of the potential immunological enhancement activities of
CM-curdlan in the body, the ligand–receptor mediated specific interactions, and the controlled release of the anti-cancer
drug.  2000 Elsevier Science B.V. All rights reserved.

Keywords: Self-assembled hydrogel nanoparticle; Curdlan; Ligand–receptor interactions; HepG2

1. Introduction nanoparticles of polysaccharides have been con-

ducted because of their potential biomedical and
Many studies of self-assembled hydrogel pharmaceutical application. The self-assembled hy-
drogel nanoparticles are composed of an inner core
*Corresponding author. Tel.: 182-62-970-2305; fax: 182-62- of hydrophobic segments and an outer shell of
970-2304. hydrophilic segments. This core-shell structure of a
E-mail address: yhbae@kjist.ac.kr (Y.H. Bae). self-assembled amphiphilic polymer is a potential

0168-3659 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved.
PII: S0168-3659( 00 )00256-X
226 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
delivery system because the inner core may contain
hydrophobic bioactive agents.
Sunamoto and co-workers have investigated
cholesterol-bearing pullulan, which forms monodis-
perse self-aggregates by intra- and / or intermolecular
association in a diluted aqueous solution [1,2]. The
cholesterol groups provide noncovalent cross-linking
points by aggregation, resulting in hydrogel
nanoparticles with polycore structures. This micro-
scopic structure increased the thermal stability of
proteins that were complexed within self-aggregates.
Jeong et al. studied the self-aggregates of hydro-
phobically modified chitosan in water, and reported
that, while critical aggregation concentration values
of self-aggregates decreased with increasing degree
of substitution (DS), the micro-viscosity of the
interior of self-aggregates was not significantly af-
fected by the DS of the hydrophobic group, pH, or
the ionic strength of the medium [3,4].
In this study, we have used CM-curdlan deriva-
tives for the preparation of self-assembled hydrogel
nanoparticles. Curdlan, one of the microbial polysac-
charides, is a naturally occurring linear (triple helix)
polysaccharide composed of 1,3-b-linked D-glucose
units produced by a strain of Alcaligenes faecalis
(Fig. 1a). It potentially has an inhibitory effect
against AIDS virus infection and blood anti-coagul-
ant activity, as well as low toxicity in vitro and in
vivo [5–7]. Especially, carboxymethylated curdlan
(CM-curdlan) is known to have anti-tumor activity
[8–10]. Although the mechanism of its anti-tumor
action is not clearly understood, CM-curdlan may
not act on tumor cells directly. Ohya et al. reported
Fig. 1. Chemical structures of a repeat unit of curdlan (a) and
sulfonylurea (b).
the immunological enhancement activity of the
muramyl dipeptide analogue (GADP) / CM-curdlan
conjugate [11] and suggested that the enhancement
was due not only to giving polymeric character to
GADP but also to the hybridization of GADP with
the immunologically active polysaccharide, curdlan.
Due to its anti-tumor activity, CM-curdlan may act
as a new anti-tumor drug delivery carrier.
CM-curdlan was hydrophobically modified with a
carboxylated sulfonylurea derivative (SU) (see Fig.
1b for its chemical structure). This SU derivative has
a strong tendency to aggregate in water (neutral pH)
due to the interactions between phenyl groups, hexyl
groups, and hydrogen bonding. The molecule has a
relatively long rigid structure, which may endow an
interesting feature as a hydrophobization moiety of
polysaccharides. The compound is also thought to be
safe in the body because similar derivatives have
been used as oral drugs for diabetic patients during
last two decades [12].
Nano-size particles can be delivered to specific
sites in the body by size-dependent passive targeting
or by active targeting using receptor-mediated inter-
actions [13–16]. For a potential liver cancer target-
ing system, we attempted to develop a receptor-
mediated delivery system using galactosylated self-
assembled hydrogel nanoparticles. Galactosylated
proteins and other macromolecules are taken up by
asialoglycoprotein receptor-mediated endocytosis in
vivo; the hepatocyte-specific targeting of drugs and
genes has already been widely investigated using
these carriers [17,18]. Hashida and co-workers have
reported that carboxymethyl-dextran, conjugated
with galactose and mannose for hepatic targeting,
selectively delivered up to 80% of the dose to
hepatocytes [19,20].
All-trans retinoic acid (ATRA) was selected as a
model anti-cancer drug. It is an intermediate in the
physiologic metabolic pathway of retinol (vitamin A)
and plays essential roles in the regulation of cell
differentiation and in the growth of a wide variety of
cell types including keratinocytes, melanoma cells,
human neuroblastoma cells, and human
promyelocytic leukemia cells. ATRA has shown
potential for cancer therapy and for the prevention of
various cancers. It has been shown to suppress
carcinogenesis in various epithelial tissues [21,22].
In this work, self-assembled hydrogel nanoparti-
 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236 227

Fig. 2. Synthesis schemes of CM-curdlan / SU (a) and LBA / CM-curdlan / SU (b).

cles were prepared by conjugation of SU to CM- 2. Materials and methods

curdlan and their physicochemical characteristics
were examined by dynamic light scattering (DLS),
transmission electron microscopy (TEM) and fluo-
rescence spectroscopy. ATRA release from self-as-
sembled nanoparticles was also tested in an attempt
to develop a controlled delivery system for ATRA.
For the potential targeting of liver cancer, CM-
curdlan / SU was conjugated with lactobionic acid
(LBA) and then tested for the degree of interaction
between self-assembled nanoparticles and a liver
cancer cell line (HepG2) in vitro.
2.1. Materials
Curdlan (Mw 68,000) was purchased from Wako,
Japan. All-trans retinoic acid (ATRA), lactobionic
acid (LBA: 4-O-b-D-galactopyranosyl-D-gluconic
acid), ethylenediamine, dicyclohexyl carbodiimide
(DCC) and 4-dimethylaminopyridine (DMAP) were
obtained from Sigma, USA. Pyrene as a fluorescence
probe was purchased from Aldrich, USA.
Terephthalic acid, N-hydroxysuccinimide (Su-OH),
228 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
allylamine, 4-(2-aminoethyl) benzenesulfonamide,
acrylic acid, and N-vinyl-2-pyrrolidone were pur-
chased from ACROS (Janssen Pharmaceuticalaan,
Belgium). N-[4-[2-[(4-carboxyphenyl (amino) ethyl)]
phenyl] sulfonyl]-N9-cyclohexylurea (sulfonylurea;
SU) was synthesized according to the method of
Hwang et al. [23]. All other chemicals used were
reagent grade.
2.2. Carboxymethylation of curdlan ( CM-curdlan)
Curdlan was carboxymethylated by the method
described by Sasaki [8]. Briefly, a suspension of 3 g
of curdlan in 80 ml of isopropyl alcohol was stirred
at room temperature for 30 min. Then, 8 ml of a 30%
sodium hydroxide solution was added in 1-ml por-
tions at 15-min intervals and stirring was continued
at room temperature for 90 min. Next, 3.6 g of
monochloroacetic acid were added over three sepa-
rate intervals of 10 min each and the mixture was
stirred at 508C for 5 h. The product was collected by
centrifugation, dissolved in 100 ml of water, and
neutralized with acetic acid. The neutralized solution
was mixed with 90 ml of methanol and the resulting
precipitate was collected by centrifugation. It was
washed with a mixture of 400 ml of 80% aqueous
methanol and 400 ml of diethyl ether and then
lyophilized, to give 2.5 g of carboxymethylated
curdlan. The degree of substitution was 0.65 units of
carboxymethyl group per one glucose unit of cur-
dlan, determined by 1 H-NMR.
2.3. Synthesis of CM-curdlan /SU derivatives and
LBA /CM-curdlan /SU derivative
SU was coupled to CM-curdlan by DCC and
DMAP-mediated ester formation. The carboxyl
groups of SU (1 g) in dried DMSO were activated by
addition of DCC (600 mg) and DMAP (400 mg)
(Fig. 2a). Different amounts of activated SU (100–
200 mg) were added in 50 ml of dried DMSO
containing 1 g of CM-curdlan and reacted for 24 h at
room temperature. After 24 h, the reactant mixture
was filtrated and dialyzed against distilled water for
3 days using a dialysis tube (molecular cut-off
12,000). Distilled water was exchanged at intervals
of 3–6 h. Precipitated CM-curdlan / SU was retrieved
by filtration and washed thoroughly with diethyl
ether and distilled water, then lyophilized. To re-
move the non-reactant completely, the lyophilized
sample was again dissolved in DMSO and dialyzed
against distilled water. This process was repeated
three times. The degree of substitution (DS), defined
as the number of sulfonylurea (SU) groups per
hundred anhydroglucose units of CM-curdlan, was
determined by 1 H-NMR and 13 C-NMR. The re-
sulting DS of three CM-curdlan / SU derivatives was
2.4, 5.6, and 7.2, respectively.
LBA / CM-curdlan / SU derivatives were synthes-
ized as follows: 1 g of CM-curdlan / SU (DS 5.6) in
dry DMSO (50 ml) was added to ethylenediamine
(100 mg) in the presence of DCC, to afford amino-
ethylcarbamylmethyl curdlan. The galactose moiety
was introduced by treating LBA (300 mg) in dry
DMSO (50 ml) with aminoethylcarbamylmethyl
curdlan (1.0 g) at room temperature for 24 h. The
degree of substitution of LBA was determined by
1
H-NMR to be 5.5 of LBA per hundred glucose units
(Fig. 2b).
2.4. Preparation of self-assembled nanoparticles
CM-curdlan derivatives were suspended in dis-
tilled water (pH 7.4), gently shaking at 378C for 48
h, followed by sonication using a probe-type sonifier
(Sonic & Materials Inc., VCX 400, USA) at 30 W for
2 min. To protect the solution from heat build-up
during sonication, a pulse function was used (pulse
on, 5.0 s; pulse off, 2.0 s). The solution of self-
assembled nanoparticles was then filtrated through a
0.45-mm filter and stored at room temperature.
2.5. Measurement of fluorescence spectroscopy
( pyrene)
Stock solutions of pyrene (6.0310 22 M) were
prepared in acetone and stored at 58C until used. For
the measurement of steady-state fluorescence spectra,
the pyrene solution in acetone was added to distilled
water to give a pyrene concentration of 12.0310 27
M. The solution was then distilled under vacuum at
608C for 1 h to remove acetone from the solution.
The acetone-free pyrene solution was mixed together
with solutions of self-assembled nanoparticles of
which the concentration ranged from 1310 24 to 1.0
mg / ml. The final concentration of pyrene in each
 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
sample solution was 6.0310 27 M, which is nearly
equal to its solubility in water at 228C.
2.6. Dynamic light scattering ( DLS) measurements
The dynamic light scattering experiment was
performed with an argon ion laser system (DLS,
Malvern Instruments, Series 4700) tuned at a wave-
length of 488 nm. Each sample was filtrated through
a 0.45-mm filter directly into a precleaned 10-mm-
diameter cylindrical cell. The intensity autocorrela-
tion was measured at a scattering angle (u ) of 908 at
258C. The sample concentration was kept at 0.3
mg / ml. DLS measurements were performed over a
time period sufficient to reach equilibrium, with no
evidence of change in size with time.
2.7. Transmission electron microscopy ( TEM)
observation
A drop of the self-assembled nanoparticles in
water was placed on a carbon-coated copper grid and
then dried under vacuum at 308C. Samples were
observed without any staining by a transmission
electron microscope (TEM, Jeol, JEM-2000 FX II,
Japan) at 80 kV, by a method similar to that used by
Jeong et al. [24].
2.8. Drug loading and release rate measurements
CM-curdlan / SU derivative (50 mg) was dissolved
in 10 ml of DMSO and 10 mg of ATRA were added.
The solution was stirred at room temperature and
dialyzed against distilled water for 3 days using a
dialysis tube (molecule cut-off 12,000) (distilled
water was exchanged at intervals of 3 h during the
first 24 h). The solution was filtered with a 0.45-mm
filter to remove precipitated polymer and drug and
then freeze-dried to obtain the product. For the
measurement of drug-loading content, a freeze-dried
sample was suspended in dichloromethane, vigorous-
ly stirred for 2 h and then sonicated for 3 min. The
resulting solution was centrifuged at 20,0003g for
30 min and the supernatant was analyzed using a UV
spectrophotometer at 365 nm.
The release rate measurements in vitro were
carried out as follows: 10 mg of drug-loaded CM-
curdlan / SU nanoparticles were suspended in 1 ml of
229
phosphate buffered saline (PBS, pH 7.4), in a
dialysis tube. The tube was then introduced into 15
ml of PBS and stirred at 100 rev. / min at 378C. At
predetermined sampling times, the whole medium
was removed and replaced by fresh PBS to maintain
sink conditions. The amount of ATRA in the solu-
tion was detected by UV at 365 nm. Because ATRA
is light sensitive, all procedures were carried out in
the dark.
2.9. Cell culture
HepG2 and Chang cells were cultured in Dulbec-
co’s modified Eagle’s medium (DMEM, 25 mmol / l
glucose) equilibrated with 5% CO 2 and 95% air at
378C. The medium was supplemented with 10% fetal
calf serum (FCS), 50 mg / l streptomycin and 75
mg / l penicillin sulphate.
2.10. RITC labeling of polymers
Each polymer was labeled with rhodamine B
isothiocyanate (RITC). To a solution of the polymer
(500 mg) in 5 ml of DMSO, 50 mg of RITC and 15
mg of dilaurated dibutyltin were added and allowed
to stand for 2 h at 908C. The cooled mixture was
then added to an excess of ethanol to precipitate the
fluorescence labeled polymer. The labeled polymer
was dissolved in 30 ml water and dialyzed against
1000 ml of alkaline water (pH 8.0 with NaOH) for 1
day, and then against distilled water for 3 days with
three separate exchanges of water. The RITC-labeled
polymers were obtained by freeze-drying.
2.11. Fluorescence intensity analysis ( RITC)
HepG2 and Chang cells were plated in 24-well
tissue culture dishes and cultured for 3 days. The
suspension of RITC-labeled polymer was added and
incubated at 378C for 1 h. The HepG2 and Chang
cells were washed three times with HEPES-balanced
Krebs-Ringer bicarbonate buffer and analysed by a
fluorescence microplate reader (FL600, Bio-Tek,
USA).
2.12. Confocal laser microscopic analysis
Confluent states of HepG2 and Chang cells were
230 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
detached by treatment with phosphate-buffered saline
(PBS) containing 0.5 mmol / l EDTA for 5 min. The
detached cells were preincubated in HEPES-balanced
Krebs-Ringer bicarbonate buffer (pH 7.4, 3.3 mM of
glucose) under 5% CO 2 and 95% air at 378C for 20
min, followed by incubation in HEPES-balanced
Krebs-Ringer bicarbonate buffer containing RITC-
labeled polymer at 378C for 1 h.
The cells (HepG2 and Chang) were then washed
with HEPES-balanced Krebs-Ringer bicarbonate buf-
fer three times and placed on a cover glass. The
cover glass was gently placed on a slide glass. Then,
the cells were observed through a confocal laser
microscope (Leika, Heidelburg, Germany).
3. Results and discussion
3.1. Properties of self-assembled hydrogel
nanoparticles of CM-curdlan /SU
Three samples of DS, which contained 2.4, 5.6,
and 7.2 SU groups per hundred anhydroglucose units
of curdlan, were prepared. The size and size dis-
tribution of self-assembled hydrogel nanoparticles in
aqueous media were measured by DLS. Fig. 3 shows
the particle size and distribution of DS 5.6 in
Fig. 3. Particle size distribution of DS 5.6 self-assembled hydro-
gel nanoparticles by dynamic light scattering.
aqueous media. Swollen DS 5.6 hydrogel particles at
pH 7.4 have a mean diameter of 209 nm with a
unimodal size distribution (6129 nm). The mor-
phological structure of DS 5.6 observed by TEM is
shown in Fig. 4. DS 5.6 is spherical with a diameter
ranging from about 20 to 50 nm (dried size).
The aggregation behavior of CM-curdlan / SU in
aqueous media was monitored by fluorometry in the
presence of pyrene as a fluorescence probe. The
fluorescence excitation spectra of DS 5.6 at various
concentrations of the polymer in the presence of
6.0310 27 M pyrene are displayed in Fig. 5a. At low
concentrations of the amphiphilic polymer, changes
in the total fluorescence intensity and the shift of the
(0,0) band at 334 nm in distilled water are negli-
gible. As the concentration of the amphiphilic poly-
mer increases, however, an increase in the total
fluorescence intensity and a red shift of the (0,0)
band are clearly observed. Since pyrene in a polar
environment shows only a weak fluorescence intensi-
Fig. 4. Transmission electron microscopy (TEM) photograph of
DS 5.6 self-assembled hydrogel nanoparticles (scale bar5100
nm).
 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236 231

Fig. 5. (a) Excitation spectra of pyrene (6.0310 27 M) in distilled water in the presence of DS 5.6 as a function of polymer concentration
(emission wavelength was 390 nm) and (b) plot of intensity ratio I337 /I334 from excitation spectra vs. log C of DS 5.6 self-assemblies.

ty, the increase in the total fluorescence intensity
with the addition of DS 5.6 indicates that the probe
is transferred from aqueous media to the less polar
microdomains of the interior of self-assembled
nanoparticles [4]. The (0,0) band for pyrene at 334
nm shifts to 337 nm on addition of DS 5.6. The
critical aggregation concentration (CAC), which is
the threshold concentration of self-aggregate forma-
tion by intra- and / or intermolecular association, was
determined from the change of the intensity ratio
(I337 /I334 ) of the pyrene in the presence of am-
phiphilic polymer. Fig. 5b shows the intensity ratio
(I337 /I334 ) of DS 5.6 as a function of amphiphile
concentration. The CAC values were determined
from the crossover point at low concentrations.
These excitation results are supported by the emis-
sion spectra of the samples. Fig. 6a shows the
fluorescence emission spectra of pyrene incorporated
into self-aggregates of DS 5.6 at pH 7.4. When the
pyrene coexists with DS 5.6, the total emission
intensity increases, especially the intensity of the first
highest vibrational band at 373 nm (I1 ), which begins

Fig. 6. (a) Emission spectra of pyrene (6.0310 27 M) in distilled water in the presence of DS 5.6 as a function of polymer concentration
(excitation wavelength was 336 nm) and (b) plot of intensity ratio I1 /I3 from emission spectra vs. log C of DS 5.6 self-assemblies.
232 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236

to increase dramatically at a certain concentration of

the amphiphilic polymer. This concentration is the
critical aggregation concentration (CAC) and can be
determined by measuring the intensity ratio (I1 /I3 )
(Fig. 6b). At low concentrations of amphiphilic
polymer, the I1 /I3 values are close to the value
(1.84) for pyrene in water, and decrease linearly with
the addition of amphiphilic polymer above the CAC.
The CAC is determined by the interception of two
straight lines [3]. The CAC values from the emission
data are consistent with those from the excitation
data.
The mean diameters of DS 2.4, 5.6 and 7.2 were
2366140, 2096129 and 1796112 nm, respectively.
This result indicates that the particle size of the
self-assembled hydrogel nanoparticle decreased with
increasing DS; evidently the increase of the DS
enhances the chances of hydrophobic interactions
between SU groups bound to CM-curdlan, resulting Fig. 7. Particle size distribution of ATRA-loaded DS 5.6 self-
in the formation of stronger hydrophobic cores [3]. assembled hydrogel nanoparticles by dynamic light scattering.
The CAC values of DS 2.4, 5.6 and 7.2 are 4.233
10 22 , 3.09310 22 and 1.89310 22 mg / ml, respec-
tively. The increase in hydrophobicity resulting from in a dialysis tube containing PBS buffer (pH 7.4).
the introduction of hydrophobic groups reduced the The total amounts of drug released from DS 2.4, 5.6
CAC values of CM-curdlan derivatives. Thus the and 7.2 in 12 h were 49, 30 and 24%, respectively
size and CAC of self-assembled hydrogel nanoparti- (Fig. 8). ATRA was released with first-order-like
cles can be controlled by the amount of DS. kinetics from the self-assembled nanoparticles; an
increased drug-loading and DS of self-assembled
3.2. ATRA loading and release test nanoparticles resulted in a slower release of ATRA.
This finding is consistent with the results of other
The mean diameter of drug-loaded DS 5.6 researchers [24,25]. Langer et al. [25] used DSC to
nanoparticles at pH 7.4 is 181 nm (Fig. 7). This show that a hydrophobic drug in nanospheres partial-
value is smaller than that of drug-free DS 5.6 ly crystallized at higher drug loading content. How-
nanoparticles. This result is observed with all sam- ever, at lower drug loading, the hydrophobic drug
ples. The addition of the hydrophobic drug may was a molecular dispersion in the nanoparticles. The
reduce the swelling capacity of hydrogel nanoparti- crystallized drug is expected to dissolve and diffuse
cles by increasing the hydrophobic interactions be-
tween sulfonylurea (SU) groups and hydrophobic Table 1
drugs. This hydrophobic interaction affected the ATRA loading content and loading efficiency in CM-curdlan
drug-loading efficiencies and loading contents of the nanoparticles a
self-assembled hydrogel nanoparticles. While the Sample Mean diameter Drug loading Loading efficiency
drug-loading efficiencies and the loading contents of 6S.D. (nm) 6S.D. (mg / mg) 6S.D. (wt.%)
CM-curdlan / SU nanoparticles increased as DS in- DS 2.4 2276106 42616 22610
creased, the particle size of CM-curdlan / SU DS 5.6 181666 104625 58613
nanoparticles decreased under the same conditions DS 7.2 164659 117627 66614
(Table 1). a
The amount of drug loading was calculated by loaded drug
To study the ATRA release kinetics, ATRA- weight / nanoparticle weight. The loading efficiencies were mea-
loaded CM-curdlan / SU nanoparticles were dispersed sured as total loaded drug weight / initial drug weight3100.
 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
Fig. 8. ATRA release behaviors of CM-curdlan / SU nanoparticles
(DS 2.4, 5.6, and 7.2).
more slowly into the outer aqueous phase relative to
the molecular dispersion. Therefore, the release rate
of hydrophobic drug from the higher-drug-loaded
nanoparticles was slower than that from the lower-
drug-loaded nanoparticles [25]. Jeong et al. have
reported that the release rate of clonazepam at higher
drug loading was slower than that at lower drug
loading [24]. Therefore, control of the drug-release
kinetics can be achieved by control of the drug-
loading contents or by the DS.
3.3. HepG2 targeting of self-assembled hydrogel
nanoparticles
Hepatocytes are known to recognize galactose-
and N-acetylgalactosamine-terminated glycoproteins
via asialoglycoprotein receptors located on the cell
surface. Thus, CM-curdlan / SU was conjugated with
lactobionic acid (LBA; galactose moiety), for target-
ing the hepatoma cell (HepG2) line. The obtained
LBA / CM-curdlan / SU derivative contained 5.5 LBA
molecules per hundred glucose units. To observe
interactions between LBA / CM-curdlan / SU and
HepG2, LBA / CM-curdlan / SU was labeled with
RITC as a fluorescence probe. The interactions
between the labeled nanoparticles and HepG2 cells
233
were quantified by a microplate-fluorescence-reader.
Chang cells, which do not possess asialoglycoprotein
receptors, were used as a control cell line. RITC-
labeled LBA / CM-curdlan / SU hydrogel nanoparti-
cles showed very strong adsorption to HepG2 cells,
while RITC-labeled CM-curdlan / SU hydrogel
nanoparticles without LBA did not show any inter-
action. The nanoparticles exhibited little interaction
with Chang cells (Fig. 9). Therefore, we are conv-
inced that the LBA / CM-curdlan / SU hydrogel
nanoparticle has a specific interaction with HepG2
cells by ligand–receptor recognition. To clarify the
specific interaction between LBA / CM-curdlan / SU
hydrogel nanoparticles and HepG2 cells, we attempt-
ed to use confocal laser microscopy. HepG2 cells
were strongly luminated by specific interactions with
LBA / CM-curdlan / SU hydrogel nanoparticles. Fur-
thermore, the luminescence was not observed with
the other controls (Chang cells with LBA / CM-cur-
dlan / SU hydrogel nanoparticles, HepG2 and Chang
cells with CM-curdlan / SU hydrogel nanoparticles)
(Fig. 10). Fig. 11 shows the changes of fluorescence
intensity as a function of (a) LBA / CM-curdlan / SU
hydrogel nanoparticle concentration and (b) HepG2
Fig. 9. Changes of fluorescence intensity by specific interaction
between cells and polymer. HepG2 (white bar) and Chang (black
bar) cells were incubated at 378C for 1 h. RITC-labeled polymer
concentration was 10 24 mg / ml.
234 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236

Fig. 10. Different degree of luminescence of RITC by specific interaction between cell and polymer. HepG2 and Chang cells were
incubated at 378C for 1 h. RITC-labeled polymer concentration was 10 24 mg / ml. Luminescence images of RITC were observed by confocal
microscope. (a) LBA / CM-curdlan / SU and HepG2 cell interaction, (b) LBA / CM-curdlan / SU and Chang cell interaction, (c) CM-curdlan
and HepG2 cell interaction and (d) CM-curdlan and Chang cell interaction.

cultivation time. The fluorescence intensity of
HepG2 cells linearly increased with the concen-
tration of LBA / CM-curdlan / SU hydrogel nanoparti-
cles and the cultivation time of HepG2 cells for the
time span tested.
the hepatoma cell (HepG2) by ligand–receptor me-
diated interaction, and release the anti-cancer drug in
a controlled release fashion.

Acknowledgements

4. Conclusion This work was financially supported by the Korea

LBA / CM-curdlan / SU hydrogel nanoparticles are
a potential drug delivery system for the treatment of
liver cancer. The carrier may induce the immuno-
logical enhancement activity in the body, attach to
Organization Science and Engineering Foundation
(KOSEF) through the Hyperstructured Organic Ma-
terials Research Center and the first author was
supported from KOSEF in the program of post-
doctoral fellowship.
 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
Fig. 11. Changes of fluorescence intensity by specific interaction between HepG2 and LBA / CM-curdlan / SU against polymer concentration
(a) and incubation time (b).
References
[1] T. Nishikawa, K. Akiyoshi, J. Sunamoto, Microscopic
structure and thermoresponsiveness of hydrogel nanoparticle
by self-assembly of a hydrophobized polysaccharide, Macro-
molecules 30 (1997) 857–861.
[2] T. Nishikawa, K. Akiyoshi, J. Sunamoto, Macromolecular
complexation between bovine serum albumin and self-as-
sembled hydrogel nanoparticle of hydrophobized polysac-
charide, J. Am. Chem. Soc. 118 (1996) 6110–6115.
[3] K.Y. Lee, W.H. Jo, I.C. Kwon, Y.H. Kim, S.Y. Jeong,
Physicochemical characteristics of self-aggregates of hydro-
phobically modified chitosans, Langmuir 14 (1998) 2329–
2332.
[4] K.Y. Lee, W.H. Jo, I.C. Kwon, Y.H. Kim, S.Y. Jeong,
Structural determination and interior polarity of self-aggre-
gates prepared from deoxycholic acid-modified chitosan in
water, Macromolecules 31 (1998) 378–383.
[5] T. Yoshida, Y. Tasuda, T. Uryu, H. Nakashima, N. Yama-
moto, T. Mimura, Y. Kaneko, Synthesis and in vitro inhib-
itory effect of L-glucosyl-branched curdlan sulfates on AIDS
virus infection, Macromolecules 27 (1994) 6272–6276.
[6] K. Katsuraya, T. Shoji, K. Inazawa, H. Nakashima, N.
Yamamoto, T. Uryu, Synthesis of sulfated alkyl Laminara-
oligosaccharides having potent anti-HIV activity and rela-
tionship between structure and biological activity, Macro-
molecules 27 (1994) 6695–6699.
[7] A. Greinacher, S. Alban, V. Dummel, G. Franz, C. Mueller-
Eckhardt, Characterization of the structural requirements for
a carbohydrate based anticoagulant with a reduced risk of
235
inducing the immunological type of heparin-associated
thrombocytopenia, Thromb. Haem. 74 (1995) 886–892.
[8] T. Sasaki, N. Abiko, Y. Sugino, K. Nitta, Dependent on chair
length of antitumor activity of (1-3)-b-D-glucan from Al-
caligenes var faecalis myxogenes IFO 13140, and its acid-
degraded productions, Cancer Res. 38 (1978) 379–384.
[9] T. Sasaki, N. Abiko, K. Nitta, N. Takasuka, Y. Sugino,
Antitumor activity of carboxymethylglucans obtained by
carboxymethylation of (1-3)-b-D-glucan from Alcaligenes
var faecalis myxogenes IFO 13140, Eur. J. Cancer 15 (1979)
211–215.
[10] H. Saito, Y. Yoshioka, N. Uehara, Relationship between
conformation and biological response for (1-3)-b-D-glucans
in the activation of coagulation factor G from limulus
amebocyte lysate and host-mediated antitumor activity.
Demonstration of single-helix conformation as a stimulant,
Carbohydr. Res. 217 (1991) 181–190.
[11] Y. Ohya, T. Nishimoto, J. Murata, T. Ouchi, Immunological
enhancement activity of muramyl dipeptide analogue / CM-
curdlan conjugate, Carbohydr. Polym. 23 (1994) 47–54.
[12] L.C. Groop, Sulfonylurea in NIDDM, Diabetes Care 15
(1992) 737–754.
[13] S.S. Davis, L. Illum, Drug targeting using colloidal carriers,
Proc. Int. Symp. Control. Release Bioact. Mater. 12 (1985)
326–332.
[14] J. kreuter, Nanoparticle-based drug delivery systems, J.
Control. Release 16 (1991) 169–176.
[15] E. Allemann, R. Gurny, E. Doelker, Drug-loaded nanoparti-
cles preparation methods and drug targeting issues, Eur. J.
Pharm. Biopharm. 39 (1993) 173–191.
[16] S.E. Dunn, A.G.A. Coombes, M.C. Garnett, S.S. Davis,
236 K. Na et al. / Journal of Controlled Release 69 (2000) 225 – 236
M.C. Davies, L. Illum, In vitro cell interaction and in vivo
biodistribution of poly(lactide-co-glycolide) nanospheres sur-
face modified by poloxamer and poloxamine copolymers, J.
Control. Release 44 (1997) 65–76.
[17] G. Ashwell, J. Harford, Carbohydrate-specific receptors of
the liver, Annu. Rev. Biochem. 51 (1982) 531–554.
[18] R. Duncan, J. Kopecek, P. Rejmanova, J.B. Lloyd, Targeting
of N-(2-hydroxypropyl)methacrylate copolymer liver by
incorporation of galactose residues, Biochem. Biophys. Acta
926 (1987) 270–279.
[19] M. Nishikawa, A. Kamijo, T. Fujita, Y. Takakura, H. Sezaki,
M. Hashida, Synthesis and pharmacokinetics of a new liver-
specific carrier, glycolated carboxymethyl-dextran and its
application to drug targeting, Pharm. Res. 10 (1993) 1253–
1261.
[20] M. Hashida, H. Hirabayashi, M. Nishikawa, Y. Takakura,
Targeted delivery of drugs and proteins to the liver via
receptor-mediated endocytosis, J. Control. Release 46 (1997)
129–137.
[21] S. Han, J.H. Choi, Highly specific cytochrome P450-like
enzyme for all-trans retinoic acid in T47D human breast
cancer cells, J. Clin. Endocrinol. Metab. 81 (1996) 1–7.
[22] P.G. Sccks, D. Harris, T.C. Chou, Modulation of growth and
proliferation in squamous cell carcinoma by terinoic acid: a
rationale for combination therapy with chemotherapeutic
agents, Int. J. Cancer 61 (1995) 409–415.
[23] J.S. Hwang, S.Y. Chae, M.K. Lee, Y.H. Bae, Synthesis of
sulfonylurea conjugated copolymer via PEO spacer and its in
vitro short-term bioactivity in insulin secretion from islets of
Langerhans, Biomaterials 19 (1998) 1189–1195.
[24] Y.I. Jeong, J.B. Cheon, S.H. Kim, J.W. Nah, Y.M. Lee, Y.K.
Sung, T. Akaike, C.S. Cho, Clonazepam release from core-
shell type nanoparticles in vitro, J. Control. Release 51
(1998) 169–178.
[25] R. Gref, Y. Minamitake, M.T. Peracchia, V. Trubetskoy, V.
Torchilin, R. Langer, Biodegradable long-circulating poly-
meric nanospheres, Science 263 (1994) 1600–1603.