A galactosamine-mediated drug delivery carrier for targeted liver cancer therapy
In order to minimize the side effect of cancer chemotherapy, a novel galactosamine-mediated drug delivery carrier, galactosamine-conjugated albumin nanoparticles (GAL-AN), was developed for targeted liver cancer therapy. The albumin nanoparticles (AN) and doxorubicin-loaded AN (DOX-AN) were prepared by the desolvation of albumin in the presence of glutaraldehyde crosslinker. Morphological study indicated the spherical structure of these synthesized particles with an average diameter of around 200 nm. The functional ligand of galactosamine (GAL) was introduced onto the surfaces of AN and DOX-AN via carbodiimide chemistry to obtain GAL-AN and GAL-DOX-AN. Cellular uptake and kinetic studies showed that GAL-AN is able to be selectively incorporated into the HepG2 cells rather than AoSMC cells due to the existence of asialoglycoprotein receptors on HepG2 cell surface. The cytotoxicity, measured by MTT test, indicated that AN and GAL-AN are non-toxic and GAL-DOX-AN is more effective in HepG2 cell killing than that of DOX-AN. As such, our results implied that GAL-AN and GAL-DOX-AN have specific interaction with HepG2 cells via the recognition of GAL and asialoglycoprotein receptor, which renders GAL-AN a promising anticancer drug delivery carrier for liver cancer therapy.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
Pharmacological Research 64 (2011) 410–419
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/yphrs
A galactosamine-mediated drug delivery carrier for targeted liver cancer therapy
, Wei Wei
, Hideyuki Tanaka
, Kazuhiro Kohama
, Guanghui Ma
, Toshiaki Dobashi
, Honghui Wang
, Jingxiu Bi
, Sheng Dai
School of Chemical Engineering, The University of Adelaide, Adelaide SA5005, Australia
Department of Molecular and Cellular Pharmacology, Gunma University Graduate School of Medicine, 3-39-22, Showa-machi, Maebashi, Gunma 371-8511, Japan
State Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China
Department of Biological and Chemical Engineering, Faculty of Engineering, Gunma University, 1-5-1, Tenjin-cho, Kiryu, Gunma 376-8515, Japan
Received 25 March 2011
Received in revised form 30 May 2011
Accepted 15 June 2011
Drug delivery carrier
In order to minimize the side effect of cancer chemotherapy, a novel galactosamine-mediated drug deliv-
ery carrier, galactosamine-conjugated albumin nanoparticles (GAL-AN), was developed for targeted liver
cancer therapy. The albumin nanoparticles (AN) and doxorubicin-loaded AN (DOX-AN) were prepared
by the desolvation of albumin in the presence of glutaraldehyde crosslinker. Morphological study indi-
cated the spherical structure of these synthesized particles with an average diameter of around 200 nm.
The functional ligand of galactosamine (GAL) was introduced onto the surfaces of AN and DOX-AN via
carbodiimide chemistry to obtain GAL-AN and GAL-DOX-AN. Cellular uptake and kinetic studies showed
that GAL-AN is able to be selectively incorporated into the HepG2 cells rather than AoSMC cells due to
the existence of asialoglycoprotein receptors on HepG2 cell surface. The cytotoxicity, measured by MTT
test, indicated that AN and GAL-AN are non-toxic and GAL-DOX-AN is more effective in HepG2 cell killing
than that of DOX-AN. As such, our results implied that GAL-AN and GAL-DOX-AN have speciﬁc interaction
with HepG2 cells via the recognition of GAL and asialoglycoprotein receptor, which renders GAL-AN a
promising anticancer drug delivery carrier for liver cancer therapy.
© 2011 Elsevier Ltd. All rights reserved.
To date, cancer remains one of the world’s most devastating
diseases, with more than 10 million new cases every year. As
one of the main cancer treatment approaches, chemotherapy is
often limited by the toxicity of anticancer drugs to normal tis-
sues and cells . To reduce the side-effect and increase the
therapeutic efﬁcacy of anticancer drugs, various drug delivery
vectors have been developed, such as polymer-drug conjugates
, nanoparticles [2–4], and microspheres . Among them,
nanoparticles are one of the most promising anticancer drug
vectors because they can be delivered to speciﬁc sites by pas-
sive targeting or active targeting . To obtain a high selectivity
to a speciﬁc organ and to enhance the uptake of drug-loaded
nanoparticles into the target cells, many distinctive active tar-
geting nanoparticles, such as monoclonal antibody conjugated
nanoparticles , magnetic nanoparticles [8–10], pH-responsive
Corresponding author. Tel.: +61 8 83034118; fax: +61 8 83034373.
Corresponding author. Tel.: +61 8 83131015; fax: +61 8 83034373.
E-mail addresses: firstname.lastname@example.org (J. Bi), email@example.com
nanoparticles , temperature-responsive nanoparticles [12,13],
and galactosamine-conjugated poly(␥-glutamic acid)-poly(lactide)
nanoparticles [14,15], have been systematically investigated in
drug delivery applications.
Galactosamine-conjugated poly(␥-glutamic acid)-poly(lactide)
nanoparticles can be employed as an active targeting anticancer
drug carrier for the treatment of liver cancer because galactosamine
(GAL) is able to recognize and bind to the asialoglycoprotein
(ASGP) receptors on the surfaces of hepatoma cells [16,17].
However, some problems are remaining unresolved: (1) poly(␥-
glutamic acid)-poly(lactide) nanoparticles were unstable because
they are constructed via the self-assembly. Such hydrophobic
interaction is relatively weak comparing with covalent and ionic
bonds . (2) Poly(␥-glutamic acid)-poly(lactide) nanoparticles
only can be used as the reservoir for water-insoluble anti-cancer
drugs since the nanoparticles are prepared by oil-in-water emul-
sion, and the hydrophobic domains are located in the core of
the nanomaterials. Therefore, they are not suitable for most of
the hydrophilic anti-cancer drugs [18,19]. (3) The toxic organic
phase and emulsiﬁers are always required for the preparation
of poly(␥-glutamic acid)-poly(lactide) nanoparticles [14,15]. (4)
Poly(␥-glutamic acid)-poly(lactide) nanoparticles cannot be used
to load polypeptide or protein drugs because they are prepared
by ultrasonication or homogeneous emulsiﬁcation, which could
destroy the microstructures of the polypeptide and protein drugs
1043-6618/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
Author's personal copy
Z. Shen et al. / Pharmacological Research 64 (2011) 410–419 411
[14,15]. It is evident that the development of novel anti-cancer
drug delivery systems with better performance is important for
In this paper, we report a novel actively targetable drug deliv-
ery carrier, GAL-conjugated albumin nanoparticles (GAL-AN), for
liver cancer treatment. The prepared GAL-AN possesses a few
advantages over traditional drug delivery systems. (1) GAL-AN is
constructed by glutaraldehyde cross-linking. The covalent bonding
makes such system more stable. (2) Due to the absence of emulsiﬁ-
cation procedure, the toxic organic phases and emulsiﬁers are not
involved during the preparation course, which results in good bio-
compatibility of GAL-AN. (3) GAL-AN can be used as the reservoir
for both water-soluble and water-insoluble drugs because of the
amphiphilic character. Additionally, such property also facilitates
the penetration of cell membranes. (4) Polypeptide and protein
drugs are suitable to be delivered using the GAL-AN vector because
the nanoparticle preparation approach is facile and mild without
ultrasonication or homogenization. (5) The GAL-AN crosslinked by
glutaraldehyde have great potential for biological tracing and quan-
titative analysis studies due to the speciﬁc autoﬂuorescent property
2. Materials and methods
Bovine serum albumin (BSA, fraction V, pH 7.0), L-Lysine (Lys),
EDAC (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide) and
(MTT) were purchased from Sigma (USA). 50% glutaralde-
hyde aqueous solution and potassium tetraborate tetrahydrate
O) were supplied from Kanto Chemical Co., Inc.
(Japan). Galactosamine (GAL), triethylamine, acetylacetone, pyri-
dine, boric acid and 4-(dimethylamino) benzaldehyde (DMBA)
were purchased from Wako Pure Chemical Industries, Ltd. (Japan).
Trypsin (≥250NF U/mg) was ordered from Nippon Becton Dick-
inson Co., Ltd. (Japan). Doxorubicin hydrochloride (DOX) was
purchased from TRC (Toronto Research Chemical, Canada) and
rhodamine-phalloidin (RP) was from Molecular Probes (USA). All
reagents were of analytical grade and used as received without
2.2. Preparation of albumin nanoparticles (AN) and DOX-loaded
albumin nanoparticles (DOX-AN)
AN was prepared using a modiﬁed desolvation approach com-
bined with chemical cross-linking by glutaraldehyde (Fig. 1(a))
. In a typical experiment, 2.0 mL ethanol (desolvation agent)
was added to 2.0 mL BSA aqueous solution (20 mg/mL in 10 mM
NaCl solution at pH 10.8) and stirred at room temperature for
10 min. 4.0 mL of ethanol was then charged at an addition speed of
2.3 mL/min under stirring (600 rpm) at room temperature. Imme-
diately after ethanol addition, 60 L of 8% glutaraldehyde aqueous
solution (diluted from the 50% glutaraldehyde aqueous solution
using Milli-Q water) was rapidly charged to induce cross-linking.
The cross-linking process was continued for over 24 h under stir-
1.0 mL of l-Lysine (Lys) aqueous solution (40 mg/mL) was then
introduced to end-cap the free aldehyde groups on the surface
of AN. After 2.0 h of reaction, the suspension was centrifuged
(30,000 × g, 20 min) at 15
C (Himac CP100WX Preparative Ultra-
centrifuge, HITACHI) to obtain the AN without free aldehyde groups
on surface. The harvested AN was washed three times by Milli-Q
water to eliminate the non-desolvated BSA and the excess Lys. The
puriﬁed AN samples were ﬁnally obtained by lyophilization for 48 h
(Eyela Freeze Dryer FD-1).
On the other hand, DOX-AN was prepared using the same
approach. The only difference is in the ﬁrst step, where 2.0 mL
ethanol was added to 2.0 mL BSA/DOX mixed aqueous solution
(500 g/mL DOX and 20 mg/mL BSA dissolved in 10 mM NaCl
solution at pH 10.8). The preparative details and the basic char-
acterization of these synthesized AN and DOX-AN particles are
summarized in Tables 1 and 2.
2.3. Conjugation of GAL to the surface of AN and DOX-AN
GAL-AN and GAL-DOX-AN were prepared by conjugating GAL
onto the surface of AN and DOX-AN (Fig. 1(b)). Typically, 10.0 mg of
EDAC was dissolved in 0.5 mL ice-cooled GAL PBS solutions, where
the GAL concentrations varied from 50 to 1000 g/mL. 4.5 mL of
AN (or DOX-AN) suspensions (2.0 mg/mL in PBS) was then charged
immediately. The mixtures were stirred at room temperature for
various conjugating times (4.0–24 h). After that, GAL-AN or GAL-
DOX-AN was ultracentrifuged and the supernatants were kept for
further analysis. The solid phases were collected and washed three
times using Milli-Q water. They were then harvested by lyophiliza-
tion for 48 h. The GAL-AN and GAL-DOX-AN sample details and their
basic characterization are also included in Tables 1 and 2.
2.4. Nanoparticles cellular uptake
The uptake of AN or GAL-AN3 by HepG2 or AoSMC cell lines was
investigated using a laser scanning confocal microscope (LSCM).
The ethanol treated glass slides were put in 6-well plates and coated
by 0.01% poly-L-Lysine. The HepG2 or AoSMC was seeded in these 6-
well plates at a density of 1.0 × 10
cells/mL and allowed to adhere
overnight. After that, the cell growth medium was changed to a
2.0 mL fresh growth medium containing 0.5 mg/mL AN or GAL-
AN3. Cells were then incubated at 37
C for 0–2.0 h. The cell layers
were washed three times with PBS to remove any absorbed free
For the LSCM imaging, the cells were ﬁxed in paraformaldehyde
for 0.5 h and the actin of the cells was stained with rhodamine
phalloidin. At dual excitation of 543 and 490 nm, the ﬂuorescent
images of rhodamine phalloidin and nanoparticles at emission
wavelengths of 570–600 and 510–540 nm were observed by LSCM.
2.5. Quantitative analysis of the internalized nanoparticles by
The nanoparticle uptake capability by HepG2 or AoSMC cell lines
was quantitatively analyzed using a steady-state ﬂuorescence spec-
trophotometer (F-4500, HITACHI). Typically, 7.0 mL of HepG2 (or
AoSMC) was seeded in a cell culture dish ( 90 mm × 20 mm) at a
density of 5.0 × 10
cells/mL and allowed to adhere overnight. The
growth medium was replaced with a fresh one (7.0 mL) containing
0.5 mg/mL of AN or GAL-AN3 (with or without 100 g/mL free GAL).
The cells were continued to incubate for another 0.5–24 h. After
that, the cells were washed twice using PBS, treated with trypsin
for 3.0 min and centrifuged at 2000 × g for 5.0 min to remove the
extracellular nanoparticles. The precipitated cells were dissolved
in 1.0 mL of DMSO and subjected to ﬂuorescent examination. The
cell samples were excited at 490 nm and the ﬂuorescence intensity
at the wavelength of 516 nm was recorded. The calibration curve
between AN or GAL-AN3 solutions and their ﬂuorescence emission
intensities were constructed as shown in Figs. S1 and S2 of Sup-
2.6. Viability of HepG2 treated with distinct DOX formulations
The cytotoxicity of various DOX formulations was examined
by the MTT tests. Typically, 150 L of HepG2 was seeded in
Author's personal copy
412 Z. Shen et al. / Pharmacological Research 64 (2011) 410–419
Fig. 1. Schematic illustration for the preparation of GAL-DOX-AN: (a) preparation of cross-linked DOX-AN nanoparticles without free aldehyde groups on surface; (b)
conjugation of GAL onto the surface of DOX-AN. The red small spheres represent DOX; the big spheres represent AN; the yellow lines inside the big spheres represent
glutaraldehyde crosslinkage. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of the article.)
96-well plates at a density of 1.0 × 10
cells/mL and allowed to
adhere overnight. The growth medium was replaced with a fresh
one (150 L) containing various concentrations (0.5–10 g/ml) of
distinct DOX formulations, namely, AN, DOX-AN, GAL-AN3, GAL-
DOX-AN3 and DOX. To determine the cell viability at different DOX
), the distinct DOX formulations with vary-
were then continuously incubated with cells for 48 h. To
determine the cell viability at different culture times, a two-step
cell culture approach was explored. HepG2 cells were ﬁrst cul-
tured in the presence of distinct DOX formulations at a ﬁxed C
of 5.0 g/mL for 2 h. Then, these cells were washed by PBS, and
150 L of fresh growth medium was introduced to culture the cells
Details on the synthesis and characterization of AN and GAL-AN.
Yield (%) Mean particle
Conjugated GAL amounts on
AN – – 74 177 0.088 −20.2 –
GAL-AN1 4 27.8 80 201 0.027 – 2.9
GAL-AN2 8 27.8 85 – – 9.8
GAL-AN3 16 27.8 83 199 0.038 −16.1 12.3
GAL-AN4 24 27.8 88 – – 13.2
GAL-AN5 16 2.8 85 – – 0.8
GAL-AN6 16 5.6 81 – – 1.9
GAL-AN7 16 13.9 85 – – 4.5
GAL-AN8 16 55.6 79 – – 14.0
Calculated from the weight ratio of feed GAL to AN.
Determined from dynamic light scattering.
Determined by the modiﬁed Seraﬁni-Cessi Assay.
Details on the synthesis and characterization of DOX-AN and GAL-DOX-AN.
Yield (%) Mean particle
DOX-AN – – 73 191 0.191 −19.8 2.03 61.2
GAL-DOX-AN1 4 27.8 77 194 0.174 – 1.92 57.6
GAL-DOX-AN3 16 27.8 80 187 0.182 −16.4 1.74 52.3
GAL-DOX-AN5 16 2.8 82 – – – 1.76 53.1
Calculated from the weight ratio of feed GAL to DOX-AN.
Determined from dynamic light scattering.
Determined by UV spectrophotometer and calculated from the standard calibration curve.
Author's personal copy
Z. Shen et al. / Pharmacological Research 64 (2011) 410–419 413
for different times (1–68 h). The total culture time was recorded in
our data presentation. Subsequently, 10 L of MTT (5.0 mg/mL in
PBS) was added and cells were incubated at 37
C. After 4.0 h, the
growth medium was removed, and 150 L of DMSO was added to
dissolve formazan crystals. The absorbance was determined using a
multi-well scanning spectrophotometer (VMax Kinetic Microplate
Reader with Softmax Pro from Molecular Devices) at a wavelength
of 570 nm. To eliminate the absorbance contribution from the inter-
nalized drugs in cells, the same experiments without MTT were
used as the references. The differential values were applied in our
2.7. Statistical analysis
Statistical signiﬁcance was determined by Student’s t-test or
one-way ANOVA followed by Student-Newman–Keuls test using
Sigma Stat version 3.5. The differences were considered signiﬁcant
for P < 0.05.
3. Results and discussion
3.1. Synthesis of GAL-conjugated albumin nanoparticles
Folic acid-conjugated albumin nanospheres (FA-AN) was devel-
oped to provide an actively targetable drug delivery system for
improved drug targeting of cancer cells in our previous study
. FA-AN could be used for the targeted therapy of ovarian
cancer, brain cancer, kidney cancer, breast cancer, lung cancer,
cervical cancer and nasopharyngeal cancer because folic acid recep-
tor alpha (FR␣), a glycosylphosphatidylinositol-linked protein, is
over-expressed on these cancer cell surfaces. However, FA-AN is
not useful for the targeted therapy of liver cancer because FR␣ is
not located on the surface of hepatoma cells. In order to enhance
the targeting of anticancer drugs for liver cancer therapy, in this
study, we developed another novel actively targetable drug deliv-
ery vector, GAL-conjugated albumin nanoparticles (GAL-AN). The
GAL-AN nanoparticles were prepared via the desolvation ,
chemical cross-linking and surface bioconjugation techniques. The
detailed approach of GAL-DOX-AN preparation is presented in
Fig. 1. DOX-loaded albumin nanoparticles (DOX-AN) is synthesized
by the desolvation approach, where spherical particles are formed
due to the protein desolvation in an alcoholic medium. During
the course of nanoparticle preparation, DOX can be encapsulated
due to the existence of both hydrophobic interaction and electro-
static interaction between DOX and BSA molecules. To enhance
the stability of formed nanoparticles, glutaraldehyde is introduced
to couple the amino groups of albumin molecules of DOX-AN in
an alcoholic-aqueous mixed solvent. On the other hand, in order
to eliminate the possible toxicity, l-lysine (Lys) is introduced to
inactive the dangling aldehyde functional groups on nanoparticle
surface. Finally, the galactosamine is conjugated onto the surface of
these preformed nanoparticles via carbodiimide chemistry. These
conjugated GAL molecules can be used as the ligand of asialoglyco-
protein (ASGP) receptors, which are located on the surface of liver
Tables 1 and 2 describe the details on the preparation of GAL-
AN and GAL-DOX-AN nanoparticles at various synthesis conditions.
For the GAL-AN (Table 1), at a ﬁxed feed ratio of GAL/AN, the yield
is about 85% beyond a conjugation time of 4 h. On the other hand,
the increase in the feed ratio of GAL/AN at a ﬁxed reaction time of
16 h could not signiﬁcantly increase the yield. For the GAL-DOX-
AN (Table 2), similar experimental results are evident. It seems the
prepared functional albumin nanoparticles are in their optimized
condition. As such, GAL-AN3 and GAL-DOX-AN3 were chosen as the
typical nanoparticles for following study.
The calculation of the aggregation number, or the number of BSA
single chains within one nanoparticle, will help us to determine
the GAL amount on the surface of one nanoparticle. The apparent
aggregation number can be approximately calculated from
where n is the aggregation number of BSA molecules,
d is the mean
diameter of one synthesized nanoparticle, and d
is the hydrody-
namic diameter of BSA (about 5.1 nm from the DLS (dynamic light
scattering) room temperature measurement at a concentration of
5.0 mg/mL). We found that the value of n is about 41,800 for AN,
d = 177nm. Further, the mole of AN nanoparticles per 1 mg
of dried AN samples can be simply computed. Based on the BSA
molecular weight of 66,000, the calculated value is found to be
3.6 × 10
The residual GAL amount in the supernatant after GAL con-
jugation was measured using the modiﬁed Seraﬁni-Cessi assay
(Supporting Information), and the amounts of GAL on nanopar-
ticle surface can be calculated by the difference of the feed and
residual GAL. The conjugated GAL on the surface of different AN
nanoparticles have been approximately calculated and listed in
Table 1. The conjugating amount of GAL on the surface of AN is inﬂu-
enced by both conjugating time and feeding concentration of GAL
(Fig. S3 in Supporting Information). Associated with the conjuga-
tion time increase, the amounts of GAL per AN raise from 2.9 × 10
to 13.2 × 10
. The conjugating amount of GAL initially increases
rapidly with increasing the conjugating time, but less dependence
is observed beyond a conjugation time of 16 h. On the other hand,
while keeping the reaction time of 16 h, the increase in the feed
GAL also gives rise to the increase of 0.8 × 10
to 14.0 × 10
AN. The conjugating amount of GAL initially increases quickly with
increasing the feeding concentration of GAL, but slow down as
the feeding concentration of GAL is higher than 27.8 g GAL/mg
AN. Similar experimental results are observed for DOX-AN related
3.2. Characteristics of nanoparticles by TEM and DLS
Particle size, particle size distribution and zeta potential play
important roles in intravenous administration . The properties
of our prepared nanoparticles with and without GAL conjugation
were characterized by TEM and DLS. Fig. 2(a)–(d) are the TEM mor-
phological images of AN, GAL-AN1, GAL-AN3 and GAL-DOX-AN3.
All these nanoparticles are spherical with their averaged diame-
ters of around 100 nm. The cumulant study from dynamic light
scattering (DLS) shows the mean hydrodynamic sizes of 177, 201
and 199 nm (Table 1) for AN, GAL-AN1 and GAL-AN3 nanoparti-
cles. Their corresponding polydispersity indices (PDIs) are found
to be 0.088, 0.027 and 0.038. Similarly, the mean particle sizes of
DOX-AN, GAL-DOX-AN1 and GAL-DOX-AN3 detected from DLS are
found to be 191, 194 and 187 nm (Table 2), with their correspond-
ing PDIs of 0.191, 0.174 and 0.182. The PDIs values of DOX-loaded
nanoparticles (DOX-AN or GAL-DOX-AN) are higher compared to
that of DOX-unloaded nanoparticles (AN or GAL-AN).
During the process of nanoparticles preparation, some amino
groups of DOX were inevitably coupled to the amino groups of
albumin molecules by glutaraldehyde, which was introduced to
couple the amino groups of albumin molecules to enhance the
stability of nanoparticles. Therefore, DOX-loaded nanoparticles are
more hydrophobic than DOX-unloaded nanoparticles due to some
hydrophobic DOX on the surface. The more hydrophobic surface
may cause aggregation to some extent and render higher PDIs val-
ues. The aggregation of DOX-loaded nanoparticles, however, is not
a problem for the future application because the PDIs values are
Author's personal copy
414 Z. Shen et al. / Pharmacological Research 64 (2011) 410–419
Fig. 2. TEM images of AN observed at 120,000 magniﬁcation (a), GAL-AN1 observed at 60,000 magniﬁcation (b), GAL-AN3 observed at 60,000 magniﬁcation (c), and GAL-
DOX-AN3 observed at 60,000 magniﬁcation (d). Particle size distributions of AN (e), GAL-AN1 (f), GAL-AN3 (g) and GAL-DOX-AN3 (h) measured by DLS at room temperature.
still comparatively low (lower than 0.2) and no precipitation was
The NNLS analysis of DLS autocorrelation functions is able to
provide more information related to the particle size distribution.
Fig. 2(e)–(h) shows the particle size distributions of AN, GAL-AN1,
GAL-AN3 and GAL-DOX-AN3. The nanoparticle size distributions
are narrow for these particles with their calculated apparent
particle sizes similar to those obtained from cumulant analysis.
However, the average diameters measured from TEM (around
100 nm) is much smaller than those from DLS (near 200 nm).
This result indicates that our prepared nanoparticles are partially
swelled in water since the TEM observes the morphology of dried
nanoparticles and DLS measure the hydrodynamic sizes. It has been
documented that nanoparticles with their size below 200 nm tends
to accumulate in tumor sites due to the enhanced permeability and
retention (EPR) effect [23,24]. Therefore, our synthesized nanopar-
ticles might be used as the potential vectors for anticancer drug
The surface charge characteristics of our prepared nanopar-
ticles were studied from the zeta-potential analysis. As shown
in Tables 1 and 2, the zeta potential values are −20.2 mV for
AN, −16.1 mV for GAL-AN3, −19.8 mV for DOX-AN, and −16.4 mV
for GAL-DOX-AN3. The negative surface charges arise from the
deprotonation of the hydrophilic carboxyl groups on the shell
of nanoparticles. Since the surfaces of many cells are negative-
charged, the negative surface charge of our prepared nanoparticles
will hinder the non-speciﬁc cellular uptake due to the electrostatic
repulsion. On the other hand, due to introduction of the functional
ligand on surface, the speciﬁc receptor-mediated interactions will
increase the rate of nanoparticles uptake into the target cells. The-
oretically, cationic nature of vectors lead to non-speciﬁc binding
to various cells after systemic administration , but the anionic
vectors might reduce the uptake capability. Our GAL-modiﬁed neg-
atively charged nanoparticles are suitable for the selective delivery
of anticancer drugs to liver cells because of the speciﬁc interaction
between GAL and asialoglycoprotein receptors.
Regarding the stability of albumin nanoparticles after lyophili-
sation procedure, we have performed and reported detailed study
in our previous paper . Neither precipitation of the albumin
nanoparticles nor change in particle size of the nanoparticles was
observed during storage periods of up to one month. This beneﬁcial
effect is likely a consequence of electrostatic repulsion between the
negatively charged nanoparticles. This result conﬁrmed that prepa-
rations of the albumin nanoparticles should exhibit a long shelf life
3.3. DOX loading analysis
The DOX contents were analyzed by digesting DOX-AN and
GAL-DOX-AN in trypsin solution, followed by UV–vis spectroscopy
analysis. The drug loading content (DLC, %) and drug loading efﬁ-
ciency (DLE, %) can be calculated from the following formula:
DLC (%) =
Weight of DOX in AN
Weight of AN
× 100 (2)
DLE (%) =
Weight of DOX in AN
Weight of the feeding DOX
× 100 (3)
As shown in Table 2, the drug loading content (DLC) and drug
loading efﬁciency (DLE) of GAL-DOX-AN1, GAL-DOX-AN3, GAL-
DOX-AN5 were slightly lower than those of DOX-AN, which might
be ascribed to the loss of drug during the conjugation process.
Author's personal copy
Z. Shen et al. / Pharmacological Research 64 (2011) 410–419 415
Fig. 3. XY-Z series of the LSCM images for HepG2 cells scanned upwardly from slide glass with 2.0 m interval. The HepG2 cells were incubated with GAL-AN3 for 2.0 h (a,
b, c, d); no nanoparticles were introduced (Control). The actin of the cells was stained with RP. The samples were excited at 490 and 543 nm and the ﬂuorescent images at
emission wavelengths of 510–540 and 570–600 nm were taken by LSCM.
3.4. In vitro release proﬁles of DOX from DOX-AN and
The DOX release proﬁles from our prepared nanoparticles with
or without trypsin were studied, where trypsin was used as the
enzyme to digest the BSA in the nanoparticles (Fig. S4 in Sup-
porting Information). In the absence of trypsin, both DOX-AN and
GAL-DOX-AN have a similar release proﬁle of DOX and shows a
burst release of DOX within the ﬁrst hours. About 5% of the loaded
drug in the DOX-AN or GAL-DOX-AN is released in 5 h. This may be
due to some of the drugs are adsorbed to the surface of the nanopar-
ticles. Finally, only about 12% of the loaded drug is released in 160 h.
That’s probably because most DOX molecules have been encapsu-
lated inside the synthesized nanoparticles rather than physically
absorbed on surface and the diffusion of DOX from the nanopar-
ticles is slow. However, in the presence of trypsin, the release of
DOX from DOX-AN or GAL-DOX-AN is dramatically accelerated.
This result indicates that the release of DOX from the nanoparticles
is mainly caused by enzymatic degradation of the nanoparticles.
Therefore, in cells, the loaded anticancer drugs can be released from
the nanoparticles due to the degradation of the nanoparticles by
3.5. Cellular uptake of the prepared nanoparticles
It has been well-documented that the ASGP (asialoglycopro-
tein) receptor is abundantly expressed on the surfaces of various
hepatoma cell lines [18,25]. The preliminary study on the ASGP
receptors’ expression for HepG2 and AoSMC cell lines were carried
out. The ﬂuorescent images of ASGP receptors are red after being
stained with goat anti-ASGPR1 IgG (N-18) and rhodamine red-X-
conjugated AfﬁniPure donkey anti-goat IgG. The cell nuclei are blue
Author's personal copy
416 Z. Shen et al. / Pharmacological Research 64 (2011) 410–419
Fig. 4. Comparison on the LSCM images of HepG2 or AoSMC incubating with AN or GAL-AN3 for distinct durations. The LSCM images in the left column were the AoSMC
incubated with GAL-AN3; those in the middle column were the HepG2 incubated with AN; those in the right column were the HepG2 incubated with GAL-AN3.
in color while being stained with To-Pro-3. From the ﬂuorescence
morphological results, large number of ASGP receptors are found
for HepG2 (Fig. S5, Supporting Information), but few are presented
for AoSMC. Such results indicate that ASGP receptors do express on
the surface of HepG2 and rarely express on the surface of AoSMC.
The speciﬁc characteristic of ASGP receptors is that they are
able to recognize galactose and galactosamine. Once a galac-
tose or galactosamine ligand binds to the ASGP receptor, the
ligand–receptor complex is rapidly internalized by hepatoma cells
and the receptor recycles back to the surface of cells . To assess
the extents of internalization of our prepared nanoparticles, HepG2
and AoSMC were incubated with AN or GAL-AN3 for different times,
where AoSMC does not possess ASGP receptors and is used as a
control cell line in our comparison experiments.
For the LSCM imaging, the samples were simultaneously excited
at 543 nm and 490 nm. At dual excitation of 543 nm and 490 nm,
red ﬂuorescent images of cytoskeleton with rhodamine phalloidin
(emission wavelengths of 570–600 nm) were observed by LSCM.
Red ﬂuorescent images of cytoskeleton with rhodamine phal-
loidin can also be observed at single excitation of 490 nm because
rhodamine phalloidin does absorb some of 490 nm light. Green ﬂu-
orescent images of the cytoskeleton with rhodamine phalloidin
Author's personal copy
Z. Shen et al. / Pharmacological Research 64 (2011) 410–419 417
(emission wavelength of 510–540 nm) were not found at excita-
tion of any wavelengths because the green emission is too weak.
On the other hand, the ﬂuorescent images of AN and GAL-AN3
nanoparticles are green (emission wavelength of 510–540 nm) at
an excitation of 490 nm, and are red (emission wavelength of
570–600 nm) at an excitation of 543 nm. As such, the dual excitation
at wavelengths of 490 and 543 nm gives rise to the red ﬂuorescence
images of cytoskeleton and yellow-colored nanoparticles, which
is associated with the combination of green and red ﬂuorescence
XY-Z series of LSCM images for HepG2 cells scanned upwardly
from slide glass with 2.0 m of interval are shown in Fig. 3. The
HepG2 cells were incubated with GAL-AN3 for Fig. 3(a)–(d) and
no nanoparticles were introduced for Fig. 3 (Control). In these
images, GAL-AN3 particles are in yellow color and the actin of
the cells stained with rhodamine phalloidin is red. In Fig. 3 (Con-
trol), no yellow points were found inside the HepG2 cells at 2.0 m
and 4.0 m of z-slice, which indicated that nothing inside HepG2
cells can disturb the visualizing of nanoparticles. There are no
nanoparticles, many nanoparticles, a few nanoparticles and almost
no nanoparticles inside the HepG2 cells at z-slice a (Z = 0.0 m), z-
slice b (Z = 2.0 m), z-slice c (Z = 4.0 m) and z-slice d (Z = 6.0 m),
respectively. The mean ﬂuorescent intensity inside one HepG2
at various z-slices (mean ± SD, n = 12) are found to be 8.3 ± 5.4,
42.4 ± 12.1, 16.2 ± 7.2 and 5.1 ± 1.7 at Z =0m slice, 2.0 m slice,
4.0 m slice and 6.0 m slice (Fig. S6 in Supporting Information).
The nanoparticles show different distributions with raising the z-
slice in XY-Z series of LSCM images for HepG2. This result indicates
that the nanoparticles have been internalized into the HepG2 cells
rather than absorbed on cell surfaces and our synthesized GAL-AN3
nanoparticles can be recognized and internalized by HepG2 cells.
The LSCM images of HepG2 and AoSMC after incubating with
AN or GAL-AN3 for 0–2.0 h are compared in Fig. 4. GAL-AN3 started
to be internalized into HepG2 (with ASGP receptors) after 0.5 h
incubation. However, after 1.0 h incubation, GAL-AN3 is almost
not internalized into AoSMC, and AN is almost not internalized
into HepG2. These results indicate that GAL-AN3 had a speciﬁc
interaction with HepG2 via the ligand (GAL) and receptor (ASGP)
recognition. Moreover, after 2.0 h incubation, only a small amount
of GAL-AN3 has been internalized into AoSMC (fourth picture in left
column), so does the AN-HepG2 (fourth picture in middle column)
system. In comparison, large amount of GAL-AN3 nanoparticles
have been internalized by HepG2 after 2.0 h incubation (fourth
picture in right column). Therefore, we can conclude that the non-
speciﬁc cellular uptake process of GAL-AN3 is insigniﬁcant while
comparing with the receptor mediated speciﬁc process.
ASGP receptors are localized exclusively in the parenchymal
cells of the mammalian liver, including hepatoma cells and normal
hepatocytes . Therefore, GAL-AN can recognize and bind to both
hepatoma cells and normal hepatocytes. Although the nanocarri-
ers with galactosamine conjugation cannot distinguish between
hepatoma cells and normal hepatocytes, they have more signif-
icant anti-tumor efﬁcacy in hepatoma-tumor-bearing nude mice
than the nanocarriers without galactosamine conjugation .
Therefore, the targeting ligand galactosamine is important for liver
targeting of nanocarriers and the hepatoma-tumor selectivity of the
nanocarriers with galactosamine conjugation is not only based on
a general EPR effect, but also active targeting of the galactosamine
3.6. Quantitative analysis of the internalized nanoparticles by
The quantiﬁcation of the internalized GAL-AN3 (with or without
presence of 100 g/mL free GAL) by HepG2, GAL-AN3 by AoSMC
and AN by HepG2 were studied with their internalization kinet-
Fig. 5. Quantiﬁcation of the internalized AN or GAL-AN3 (with or without excessive
GAL) by HepG2 or AoSMC analyzed by ﬂuorescence spectrophotometer (Mean ± SD,
n = 3),
P < 0.05 and
P < 0.01 compared with internalized GAL-AN3 (without exces-
sive GAL) by HepG2. The ﬂuorescence intensity was converted into the concentration
of AN or GAL-AN3 using a standard calibration curve.
ics as shown in Fig. 5. The internalization kinetics is dependent on
the formulation of nanoparticles and the cells. At any time inter-
val, the quantity of the internalized GAL-AN3 (without free GAL) by
HepG2 is larger than those of the internalized GAL-AN3 by AoSMC
and the internalized AN by HepG2 (P < 0.05 for 1.0 h incubation
and P < 0.01 beyond 2.0 h of incubation). Additionally, in the pres-
ence of excessive free GAL, the quantity of internalized GAL-AN3
by HepG2 reduces signiﬁcantly (P < 0.05 for 1.0 h incubation and
P < 0.01 beyond 2.0 h of incubation) because the uptake of GAL-AN3
could be disrupted by ligand GAL. The regular internalization pro-
cess of GAL was disturbed by the presence of the GAL-conjugated
nanoparticles because of the competitive interaction with ASGP
receptors. These results indicated that GAL-AN3 has speciﬁc inter-
actions with HepG2 via the recognition of ligand (GAL) and receptor
From Fig. 4, it was found that HepG2 cells incubated with AN
have a higher number of internalized nanoparticles than AoSMC
cells. However, this difference is only from several cells and its
statistical signiﬁcance need to be further veriﬁed. Fig. 5 shows
the quantiﬁcation of internalized nanoparticles by millions of cells
(HepG2 or AoSMC). It was found that HepG2 cells incubated with
AN have a slightly higher number of internalized nanoparticles than
AoSMC cells, but the difference is not enough to be statistically
signiﬁcant due to P > 0.05.
3.7. Viability of HepG2 treated with distinct DOX formulations
The killing capability of distinct DOX formulations on HepG2
was examined by the MTT test. Fig. 6(a) shows the concentration
dependence of the viability of HepG2 treated with AN, GAL-AN3,
DOX-AN, GAL-DOX-AN3 or DOX at an incubation time of 48 h. The
insert of Fig. 6(a) indicates that the cell viability is always around
100% when the cells are treated with AN or GAL-AN3 (without DOX
encapsulation). No cytotoxicity of the synthesized AN or GAL-AN3
nanoparticles to HepG2 cells indicates their good biocompatibility.
On the other hand, for the systems with DOX (DOX-AN, GAL-DOX-
AN3 and DOX), the cell viability decreases with increasing of DOX
concentrations. Additionally, free DOX seems more toxic than GAL-
DOX-AN3 nanoparticles, and DOX-AN has less cytotoxic effect than
Author's personal copy
418 Z. Shen et al. / Pharmacological Research 64 (2011) 410–419
Fig. 6. Cytotoxicity of HepG2 cell lines treated with distinct DOX formulations: (a) comparison on the DOX concentration (C
) at a ﬁxed incubation time of 48 h, the insert
of (a) examines cytotoxicity of the AN and GAL-AN3 without the presence of DOX, the drugs were continuously incubated with cells for 48 h; (b) comparison on the total
incubation time at a ﬁxed C
of 5.0 g/mL, the drugs were washed out by PBS after 2.0 h of incubation with cells (mean ± SD, n = 4).
DOX alone has a greater activity in inhibiting growth of HepG2
cells than GAL-DOX-AN because the release of DOX from GAL-
DOX-AN is sustained and the diffusion of small molecular DOX
into cells is very fast. However, the cancer chemotherapy using
DOX is often limited by the toxicity to normal tissues and cells.
Because GAL-DOX-AN has a speciﬁc interaction with HepG2 cells
via the recognition of GAL and asialoglycoprotein receptor, the side
effect of DOX to normal tissues and cells could be reduced using
GAL-DOX-AN shows more activity in inhibiting HepG2 growth
than that of DOX-AN due to the enhancement of internalization into
HepG2 by the speciﬁc ligand–receptor (GAL–ASGP) interaction.
Fig. 6(b) compares the effect of total culture time on the viabil-
ity of HepG2 treated with DOX-AN, GAL-DOX-AN3 or DOX. There
is no signiﬁcant difference among them within the ﬁrst 2 h. After-
wards, the GAL-DOX-AN3 displays a much higher cytotoxicity than
DOX-AN, but lower than DOX. The reason is ascribed to the differ-
ent amounts of internalized DOX, GAL-DOX-AN and DOX-AN in the
HepG2 cells during the ﬁrst 2.0 h incubation: excess DOX and GAL-
DOX-AN have been internalized into the cells, but the main portion
of DOX-AN has been washed out by PBS.
Additionally, we found that the release of DOX from the carri-
ers seems very slow (after 2 h there is any effect of DOX as shown
in Fig. 6(b)). That’s because the enzymatic degradation of the car-
riers is slow, which has been veriﬁed by Fig. S4.b (only around
60% of DOX was released from the carriers after treatment with
trypsin for 43 h). The reason of slow enzymatic degradation of the
carriers is probably ascribed to the cross-linking of the carriers by
With respect to the immunogenicity of albumin, the wide appli-
cation of albumin and drug–albumin conjugates in clinical practice
indicates that the immunogenicity of albumin is not a big problem.
For example, a methotrexate–albumin conjugate and an albumin-
binding prodrug of doxorubicin have been clinically evaluated ,
and the U.S. Food and Drug Administration (FDA) has approved
albumin-bound paclitaxel nanoparticles (Brand name: Abraxane
Abraxane) for the treatment of metastatic breast cancer.
In this study, GAL was conjugated onto the surface of nanopar-
ticles via formation of amide bonds. Because the amide bonds are
stable at neutral pH, the GAL-DOX-AN should be comparatively sta-
ble in the bloodstream (after intravenous administration) in the
future clinical application. However, the enzymatic hydrolysis of
amide bonds in the bloodstream may be a potential problem of
our nanoparticles. To protect amide bonds from enzymatic hydrol-
ysis, at the next stage of our study, we could graft pH-sensitive
polymers (with phase transition pH of 5.5) onto the surface of the
nanoparticles to hide the amide bonds. Because the pH value of the
interstitial space of solid tumors is usually more acidic (pH close
to 5) than blood plasma (pH 7.4), after arriving at tumor sites, the
spreading pH-sensitive polymers will shrink and the hidden GAL
will be exposed and recognized by asialoglycoprotein receptors of
A novel actively targetable drug delivery carrier of GAL-AN
has been developed by conjugating GAL onto the surface of AN,
which is spherical in shape with a diameter of around 200 nm.
The conjugated GAL amount on the surface can be elevated to
1.4 × 10
mol/mol AN. GAL-DOX-AN can be easily internalized into
HepG2 cells via the receptor-mediated interaction. The slow release
of DOX from the prepared nanoparticles in the absence of trypsin
and the obvious fast release proﬁle in the presence of trypsin
indicate the prepared nanoparticles can be used to reduce the
side-effect of anticancer drugs. Furthermore, the cell uptake and
internalization studies suggest that GAL-AN has speciﬁc interac-
tion with HepG2 cells via the ligand–receptor (asialoglycoprotein,
ASGP) recognition. As expected, the activity in inhibiting the
growth of HepG2 cells by GAL-DOX-AN is comparable to that of free
DOX, while DOX-AN displayed less activity. Consequently, GAL-
AN can be employed as a promising actively targetable anti-cancer
drug carrier for future liver cancer treatment.
Grants from Smoking Research Foundation and Grants-in-Aid
for Scientiﬁc Research of the Ministry of Education, Culture, Sports,
Science, and Technology of Japan are gratefully acknowledged. The
authors are grateful for the ﬁnancial support from the Australian
Author's personal copy
Z. Shen et al. / Pharmacological Research 64 (2011) 410–419 419
Research Council (ARC) (Discovery Grant No. DP110102877). The
authors thank Kohichi Hayakawa, who is at the Department of
Molecular and Cellular Pharmacology, Gunma University Graduate
School of Medicine, for his technical support and useful discussion.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.phrs.2011.06.015.
 Li C. Poly(l-glutamic acid)-anticancer drug conjugates. Adv Drug Deliver Rev
 Vlerken LE, Duan Z, Seiden MV, Amiji MM. Modulation of intracellular ceramide
using polymeric nanoparticles to overcome multidrug resistance in cancer.
Cancer Res 2007;67:4843–50.
 Lasic DD. Liposomes within liposomes. Nature 1997;387:26–7.
 Lasic DD. Mixed micelles in drug delivery. Nature 1992;355:279–80.
 Hoshino K, Kimura T, Grand AMD, Yoneyama R, Kawase Y, Houser S, et al. Three
catheter-based strategies for cardiac delivery of therapeutic gelatin micro-
spheres. Gene Ther 2006;13:1320–7.
 Xu X, Clarke P, Szalai G, Shively JE, Williams LE, Shyr Y, et al. Targeting and
therapy of carcinoembryonic antigen-expressing tumors in transgenic mice
with an antibody-interleukin 2 fusion protein. Cancer Res 2000;60:4475–84.
 Kirpotin DB, Drummond DC, Shao Y, Shalaby MR, Hong K, Nielsen UB, et al.
Antibody targeting of long-circulating lipidic nanoparticles does not increase
tumor localization but does increase internalization in animal models. Cancer
 Dobson J. Remote control of cellular behaviour with magnetic nanoparticles.
Nat Nanotechnol 2008;3:139–43.
 Reiss G, Hutten A. Magnetic nanoparticles: applications beyond data storage.
Nat Mater 2005;4:725–6.
 Amirfazli A. Magnetic nanoparticles hit the target. Nat Nanotechnol
 Lin YH, Sonaje K, Lin KM, Juang JH, Mi FL, Yang HW, et al. Multi-ion-crosslinked
nanoparticles with pH-responsive characteristics for oral delivery of protein
drugs. J Control Release 2008;132:141–9.
 Shen Z, Ma G, Dobashi T, Maki Y, Su Z. Preparation and characterization of
thermo-responsive albumin nanospheres. Int J Pharm 2008;346:133–42.
 Shen Z, Wei W, Zhao Y, Ma G, Dobashi T, Maki Y, et al. Thermosensitive
polymer-conjugated albumin nanoparticles as thermal targeting anti-cancer
drug carrier. Eur J Pharm Sci 2008;35:271–82.
 Liang HF, Yang TF, Huang CT, Chen MC, Sung HW. Preparation of nanopar-
ticles composed of poly(-glutamic acid)-poly(lactide) block copolymers and
evaluation of their uptake by HepG2 cells. J Control Release 2005;105:213–25.
 Liang HF, Chen CT, Chen SC, Kulkarni AR, Chiu YL, Chen MC, et al. Paclitaxel-
loaded poly(-glutamic acid)-poly(lactide) nanoparticles as a targeted drug
delivery system for the treatment of liver cancer. Biomaterials 2006;27:2051–9.
 Han JH, Oh YK, Kim DS, Kim CK. Enhanced hepatocyte uptake and liver tar-
geting of methotrexate using galactosylated albumin as a carrier. Int J Pharm
 Fallon RJ, Schwartz AL. Receptor-mediated delivery of drugs to hepatocytes.
Adv Drug Deliver Rev 1989;4:49–63.
 Neradovic D, Nostrum CF, Hennink WE. Thermoresponsive polymeric
micelles with controlled instability based on hydrolytically sensitive N-
isopropylacrylamide copolymers. Macromolecules 2001;34:7589–91.
 Soga O, Nostrum CF, Fens M, Rijcken CJF, Schiffelers RM, Storm G, et al. Ther-
mosensitive and biodegradable polymeric micelles for paclitaxel delivery. J
Control Release 2005;103:341–53.
 Shen Z, Li Y, Kohama K, Oneill B, Bi J. Improved drug targeting of cancer cells by
utilizing actively targetable folic acid-conjugated albumin nanospheres. Phar-
macol Res 2011;63:51–8.
 Weber C, Coester C, Kreuter J, Langer K. Desolvation process and surface char-
acterisation of protein nanoparticles. Int J Pharm 2000;194:91–102.
 Dong Y, Feng SS. Methoxy poly(ethylene glycol)-poly(lactide) (MPEG-PLA)
nanoparticles for controlled delivery of anticancer drugs. Biomaterials
 Hashida M, Takemura S, Nishikawa M, Takakura Y. Targeted delivery of plas-
mid DNA complexed with galactosylated poly(l-Lysine). J Control Release
 Peer D, Karp JM, Hong S, Farokhzad OC, Margalit R, Langer R. Nanocarriers as
an emerging platform for cancer therapy. Nat Nanotechnol 2007;2:751–60.
 Ong GL, Ettenson D, Sharkey RM, Marks A, Baumal R, Goldenberg DM, et al.
Galactose-conjugated antibodies in cancer therapy: properties and principles
of action. Cancer Res 1991;51:1619–26.
 Hata S, Ishii K. Effect of galactose on binding and endocytosis of asialoglyco-
protein in cultured rat hepatocytes. Ann Nucl Med 1998;12:255–9.
 Kratz F. Albumin as a drug carrier: design of prodrugs, drug conjugates and
nanoparticles. J Control Release 2008;132:171–83.