Novel synthetic LPDs consisting of different cholesterol derivatives for gene transfer into hepatocytes

Article abstract of DOI:10.3109/10611860903548370

In the present study, LPDs composing of a series of novel synthetic cholesterylated derivatives bearing a cluster of galactose residues and different spacer lengths were prepared for performing target gene delivery to hepatocytes and their physiochemical properties as well as gene transfer efficiency were investigated. In agreement with the "clustering effect" known to occur with more complex oligomeric structures, the addition of galactose residues under optimized spatial arrangement condition invariably increased the transfect efficiency into hepatoma cells, which can be owed to the sufficient binding of galactose ligands to the ASGPR on hepatocytes. However, the gene transfer ability to hepatocytes was not always improved with extended spacer arms, suggesting a spatial binding sites arrangement of the receptor. Moreover, our research has established galactosylated LPDs, specifically, LPDIIb, LPDIIIc, and LPDIVe as potential vectors to deliver special genes into hepatocytes with low toxicity, combining the condensing effect of protamine and the targeting capability of cholesterylated thiogalactosides.

Full text of DOI:10.3109/10611860903548370

Journal of Drug Targeting, 2010; 18(7): 520–535
RESEARCH ARTICLE
Novel synthetic LPDs consisting of different cholesterol
derivatives for gene transfer into hepatocytes
Jiao Lu*, Di Zhu*, Zhi-Rong Zhang, Li Hai, Yong Wu, and Xun Sun
Key Laboratory of Drug Targeting and Drug Delivery Systems, Ministry of Education, West China School of Pharmacy,
Sichuan University, Chengdu, Sichuan, P.R. China
Abstract
In the present study, LPDs composing of a series of novel synthetic cholesterylated derivatives bearing a
cluster of galactose residues and different spacer lengths were prepared for performing target gene delivery
to hepatocytes and their physiochemical properties as well as gene transfer efficiency were investigated.
In agreement with the “clustering effect” known to occur with more complex oligomeric structures, the
addition of galactose residues under optimized spatial arrangement condition invariably increased the
transfect efficiency into hepatoma cells, which can be owed to the sufficient binding of galactose ligands
to the ASGPR on hepatocytes. However, the gene transfer ability to hepatocytes was not always improved
with extended spacer arms, suggesting a spatial binding sites arrangement of the receptor. Moreover, our
research has established galactosylated LPDs, specifically, LPDIIb, LPDIIIc, and LPDIVe as potential vectors
to deliver special genes into hepatocytes with low toxicity, combining the condensing effect of protamine
and the targeting capability of cholesterylated thiogalactosides.
Keywords: ASGPR; galactosylated LPD; cholesterylated thiogalactosides; hepatotrophic gene targeting
Introduction
subunits, designated H1 and H2 in the human system,
which form a noncovalent heterooligomeric complex
with an estimated ratio of 2-5:1, respectively. Both
subunits are single-spanning membrane proteins with a
calcium-dependent galactose/N-acetyl-gal-actosamine
recognition domain (Bider et al., 1996). Recently, the
X-ray crystal structure of the carbohydrate recognition
domain (CRD) of the major subunit H1 was elucidated
(Meier et al., 2000).
Many studies have been performed with both natural
and synthetic carbohydrates to establish the structure–
affinity relationship for the ASGPR. Baenzinger et al.
(Baenziger & Fiete, 1980; Baenziger & Maynard, 1980)
have shown that the human receptor exhibits specificity
for terminal Gal and GalNAc (with an approximately
50-fold higher affinity for the latter) on desialylated
glycoproteins. Triantennary ligands displayed a higher
affinity than their mono- and diantennary counterparts.
Furthermore, the studies led to the conclusion that only
One of the most successful cationic lipids based gene
delivery system reported to date is liposomes/protamine/
DNA (LPD), known for their superior gene transfer abil-
ity over conventional liposomes (Li et al., 1998; Sorgi,
Bhattacharya, & Huang, 1997; Li & Huang, 1997; Gao &
Huang, 1996). For targeting purposes, LPD was incorpo-
rated with cell- or tissue-specific ligands as an effective
approach to improve gene transfer efficiency in target tis-
sue or cells. Among those ligands, galactose is the most
extensively studied to target genes to liver parenchymal
cells since galactose moiety can be specifically recognized
by asialoglycoprotein receptors (ASGPR) on hepatocytes
(Schwartz et al., 1981).
e affinity and specificity of the ASGPR is a conse-
quence of oligovalent interactions with its physiological
ligands, a process termed “cluster glycoside effect” by Lee
et al. (1983). e receptor consists of two homologous
*ese authors contributed equally to this work.
Address for Correspondence: Xun Sun, Key Laboratory of Drug Targeting and Drug Delivery Systems, Ministry of Education, West China School of Pharmacy,
Sichuan University, Chengdu, Sichuan, P.R. China; E-mail: xunsun22@gmail.com
(Received 09 October 2009; revised 26 November 2009; accepted 01 December 2009)
ISSN 1061-186X print/ISSN 1029-2330 online © 2010 Informa UK Ltd
DOI: 10.3109/10611860903548370
http://www.informahealthcare.com/drt

Hepatoma cell targeting with galactosylated LPD
521
the terminal residues are necessary for specific recogni-
tion, and that the binding process proceeds through a
simultaneous interaction of 2–3 sugar residues with 2–3
binding sites of the heterooligomeric receptor (Khorev
et al., 2008). Studies on rabbit hepatocytes by Lee et al.
(Lee et al., 1983; Conolly et al., 1982) using synthetic oli-
gosaccharides, further reinforced the binding hierarchy
of polyvalent ligands: tetraantennary > triantennary >>
diantennary >> monoantennary (Khorev et al., 2008).
e optimal distance of the Gal moieties in these oli-
gosaccharides was determined by binding assays with
synthetic carbohydrates representing partial structures
of N-linked glycans (Khorev et al., 2008) high-resolution
NMR, and molecular modeling studies (Bock, Arnarp, &
Lönngren 1982). Based on these results, Lee et al. (Lee
et al., 1983; Khorev et al., 2008) presented a model for
the optimal spatial arrangement of the terminal sugar
residues (Khorev et al., 2008).
which result in the variability in their gene transfer effi-
ciency. To further clarify the unaddressed problems, we
prepared a series of Gal-LPDs composing cholesterylated
thiogalactosides possessing different galactose densi-
ties and spacer lengths and studies on those Gal-LPDs
were investigated such as charactertization, in vitro gene
transfer study and cell viability, etc., which might prove
valuable in optimizing the strategy for hepatotrophic
gene targeting.
Materials and methods
Materials
e plasmid pORF lacZ (3.54 kb) was purchased from
Invitrogen (USA). Protamine sulfate (derived from
salmon), dimethyldioc tadecyl ammonium bromide
(DDAB)
and
dioleoylphosphatidylethanol-amine
Despite attempts made to elucidate the correlation
between the structural differences of the ligands and
their binding affinity to the receptor; the frequently used
methods to construct multivalent structures include
conjugation of carbohydrate ligands with proteins (i.e.,
neoglycoproteins) or other polymers (Stowell & Lee,
1980; Pazur, 1981), polymerization of glycosylated mono-
mers (i.e., glycopolymers) (Horejsí, Smolek, & Kocourek,
1978; Roy, Tropper, & Romanowska, 1992) and formation
of liposomes using glycolipids or neoglycolipids. As for
liposomes, how the variability in ligands’ structures on its
surface influences its specific gene transfer ability when
conjugated with gene delivery vectors has been rarely
discussed. Besides, although these approaches have been
successful, the products are ambiguous in composition
and structure (Eggens et al., 1989; Stoll et al., 1988); the
basic knowledge of the structure-activity relationship is
still unsatisfactory and the optimization of these systems
remains largely a result of trial and error (Masotti et al.,
2009).
erefore, it is in our study that the galactosylated cho-
lesterol bearing a cluster of galactose residues and dif-
ferent spacer lengths were synthesized and the possible
relation between the structural diversity of LPD and its
gene transfer ability was investigated for the first time.
In our previous study, cholesterylated thiogalactosides
bearing different spacer lengths were synthesized to for-
mulate LPD complexes and the lipid-protamine-DNA
complex containing cholesterylated thiogalactoside 1c
(Sun et al., 2005) was established as the most promis-
ing vector for hepatocyte-specific gene delivery in vitro
among LPDs, combining the condensing effect of pro-
tamine and the targeting capability of cholesterylated
thiogalactosides. However, several questions remain
unclear concerning other structural factors especially
the amount of galactosidase moieties that might attribute
to the affinity between ASGPR and its ligands on LPDs,
(DOPE) were purchased from Sigma. Qiagen Giga
Endo-free plasmid purification kit was purchased from
Qiagen (CA, USA). Cell culture medium RPMI-1640
was obtained from Gibco Co. (USA). Lipofectamine™
2000 Transfection Reagent and β-gal assay kit were from
Invitrogen (USA). Mouse fibroblasts L929, Hepatoma
cells HepG2 and A549 were obtained from Shanghai
Cell Institute, China Academy of Sciences. All the other
chemicals and reagents used were of the analytical grade
obtained commercially.
Synthesis of galactosylated cholesterols
Synthesis of target compound IIa–f (Figure 1,
Scheme 1C)
Cholesterol was methylsulfonylated, followed by coupling
with the corresponding olig-polyethylene glycol, yield-
ing the intermediate 3a. According to previous methods
(Loiseau, Hii, & Hill, 2004; Burns et al., 1999), using 2,3-
dihydro-2H-pyran to protect one hydroxyl functional-
ity, the other hydroxyl group of triethylene glycol was
sulfonylated with methylsulfonyl chloride, leading to the
mono-mesylate-ester 6, which was subsequently coupled
with the intermediate 3a, affording the compound 7a.
After deprotection of 7a, the target compound 3d (Scheme
1B) bearing the prolonged oligo-diethylene glycol chain
was achieved. To prepare the potential scaffold 9a, 3a was
sulfonylated with methylsulfonyl chloride, resulting in the
compound 8a, which can convert into iodide 9a via substi-
tution of the mesyl group by NaI in refluxing butanone. By
conjugation of 9a with the known 2,3,4,6-tetra-O-acetyl-1-
thio-β-D-galacto-pyranose, using diisopropylethylamine
(DIPEA) as catalyst, the monosaccharide derivative 11a
was achieved. After deacetylation under mild condi-
tion, the desired product IIa was yielded. Compound IIa
(C37H64O7S): 1HNMR (δ ppm, 400 MHz, CD3OD): 5.36 (m,

522
Jiao Lu et al.
1H, chol H-6), 4.35–3.43 (13H, gal. protons and 4 ×CH2O),
3.19 (m, 1H, chol H-3), 2.94 (m, 1H, SCH-a), 2.81 (m, 1H,
SCH-b), 2.39–0.71 (remaining chol protons) with 1.01 (s,
3H, CH3-19), 0.93 (d, 3H, CH3-21, J20, 21 =6.4 Hz), 0.87 (d,
6H, CH3-26 and CH3-27, J25, 26 = J25, 27 =6.8 Hz), 0.71 (s, 3H,
CH3-18). MS (m/z): (M+Na)+ 675.
CH2OH
e synthetic route of compound IIb–f is similar to that
of compound IIa. Compound IIb (C39H68O8S): 1HNMR (δ
ppm, 400 MHz, CD3OD): 5.34 (m, 1H, chol H-6), 4.30–3.40
(17H, gal. protons and 6 × CH2O), 3.20 (m, 1H, chol H-3),
2.96 (m, 1H, SCH-a), 2.88 (m, 1H, SCH-b), 2.41–0.67
(remaining chol protons) with 1.0 (s, 3H, CH3-19), 0.91
(d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.87 (d, 6H, CH3-26 and
HO
S
O
O
O
n
OH
OH
IIa–f n=1–6
Figure 1. Structure of mono-antennary galactosides (target com-
pound IIa–f).
TsCl
Py
TsO
HO
1
2
O
OH
HO
n
Dioxane
HO
O
O
n
3a–c n=1–3
Scheme 1A. e synthesis route of intermediate 3a–c.
O
Dihydropyran
O
O
OH
5
O
O
O
OH
4b
HO
O
MsCl
O
O
OMs
6
O
Et3N,THF
3a + 6
3b + 6
3c + 6
NaH, DMSO, THF
O
O
O
O
m
7a–c m=4–6
TsOH, EtOH
HO
O
O
m
3d–f m=4–6
Scheme 1B. e synthesis route of intermediate 3d–f.

Hepatoma cell targeting with galactosylated LPD
523
(remaining chol protons) with 1.01 (s, 3H, CH3-19), 0.93
(d, 3H, CH3-21, J20, 21 = 6.8 Hz), 0.87 (d, 6H, CH3-26 and
CH3-27, J25, 26 = J25, 27 = 7.2 Hz), 0.71 (s, 3H, CH3-18). MS (m/
z): (M+Na)+ 807.
MsCl
Compound IIe (C45H80O11S): 1HNMR (δ ppm, 400
MHz, CD3OD): 5.35 (m, 1H, chol H-6), 4.33–3.43 (29H,
gal. protons and 12×CH2O), 3.20 (m, 1H, chol H-3), 2.94
(m, 1H, SCH-a), 2.82 (m, 1H, SCH-b), 2.39–0.71 (remain-
ing chol protons) with 1.01 (s, 3H, CH3-19), 0.93 (d, 3H,
CH3-21, J20, 21 =6.8 Hz), 0.87 (d, 6H, CH3-26 and CH3-27,
O
HO
n
O
3a–f n=1–6
NaI
J25, 26 = J25, =6.4 Hz), 0.71 (s, 3H, CH3-18). MS (m/z):
27
O
O
MsO
(M+Na)+ 851.
n
8a–f n=1–6
Compound IIf (C47H84O12S): 1HNMR (δ ppm, 400 MHz,
CD3OD): 5.35 (m, 1H, chol H-6), 4.33–3.58 (33H, gal. pro-
tons and 14 × CH2O), 3.20 (m, 1H, chol H-3), 2.94 (m, 1H,
SCH-a), 2.82 (m, 1H, SCH-b), 2.39–0.71 (remaining chol
protons) with 1.01 (s, 3H, CH3-19), 0.93 (d, 3H, CH3-21,
CH2OAc
AcO
O
SH
10
OAc
J
20, 21 = 6.8 Hz), 0.86 (d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27
=
OAc
6.4 Hz), 0.71 (s, 3H, CH3-18). MS (m/z): (M+Na)+ 895.
I
O
O
n
Synthesis route of target compound IIIa–f (Figure 2,
Scheme 2)
9a–f n=1–6
e compound 13 (Hukkamaki & Pakkanen, 2001) was
prepared by the Michael addition reaction of pentaerythri-
tol with acrylonitrile in the presence of sodium hydrocho-
ride, which was converted into the compound 14 through
ethanolysis. With the four ester functionalities reduced by
LiAlH4, the product tetra-(ω-hydroxypropyloxymethyl)
methane 15 (Newkome, Mishra, & Moorefield, 2002) was
obtained, which was then methylsulfonylated, affording
mesylate 16. Followed by coupling with 3a, the conju-
gate 17a was prepared and subsequently converted into
iodide 18a by treating with NaI in refluxing butanone.
Conjugation of 18a with 2,3,4,6-tetra-O-acetyl-1-thio-
β-D-galactopyranose resulted in the clustered trisac-
charide derivative 19a. en the target compound IIIa
was obtained by deacetylation of 19a. Compound IIIa
(C66H118O22S3): 1HNMR (δ ppm, 400 MHz, CD3OD): 5.36 (m,
1H, chol H-6), 4.31 (d, 3H, 3×gal. H1, J=9.6 Hz), 3.88 (d,
3H, 3×gal. H4, J=3.2Hz), 3.76–3.45 (m, 33H, 3×gal. H2, H3,
H5, H6, H6’ protons and 9×CH2O), 3.37 (s, 8H, 4×CCH2O),
3.19 (m, 1H, chol H-3), 2.78 (m, 6H, SCH-a, SCH-b),
2.35–0.72 (remaining chol protons, 4×CCH2C) with 1.02
(s, 3H, CH3-19), 0.91 (d, 3H, CH3-21, J20, 21 =6.4Hz), 0.86 (d,
6H, CH3-26 and CH3-27, J25, 26 = J25, 27 =6.4 Hz), 0.72 (s, 3H,
CH3-18). MS (m/z): (M+Na)+ 1381.6.
MeONa
CH2OAc
AcO
O
O
O
S
n
OAc
AcO
11a–f n=1–6
CH2OH
HO
O
O
O
S
n
OH
OH
IIa–f n=1–6
Scheme 1C. e synthesis route of target compound IIa–f.
CH3-27, J25, 26 = J25, 27 = 6.8 Hz), 0.67 (s, 3H, CH3-18). MS
(m/z): (M+Na)+ 719.
Compound IIc (C41H72O9S): 1HNMR (δ ppm, 400 MHz,
CD3OD): 5.35 (m, 1H, chol H-6), 4.33–3.58 (21H, gal. protons
and 8×CH2O), 3.20 (m, 1H, chol H-3), 2.96 (m, 1H, SCH-a),
2.88 (m, 1H, SCH-b), 2.39–0.67 (remaining chol protons)
with 1.0 (s, 3H, CH3-19), 0.91 (d, 3H, CH3-21, J20, 21 = 6.0Hz),
0.86 (d, 6H, CH3-26 and CH3-27, J25, 26 = J25,27 =6.4Hz), 0.67
(s, 3H, CH3-18). MS (m/z): (M+Na)+ 763.
e synthetic route of compound IIIb-f is similar to that of
1
compound IIIa. Compound IIIb (C68H122O23S3): HNMR
(δ ppm, 400 MHz, CD3OD): 5.36 (m, 1H, chol H-6), 4.31(d,
3H, 3 × gal. H1, J= 9.6 Hz), 3.88 (d, 3H, 3 × gal. H4, J= 3.6
Hz), 3.77-3.45 (m, 37H, 3 × gal. H2, H3, H5, H6, H6’ protons
and 11 × CH2O), 3.37 (s, 8H, 4 × CCH2O), 3.19 (m, 1H, chol
H-3), 2.79 (m, 6H, SCH-a, SCH-b), 2.37–0.71 (remaining
chol protons, 4 × CCH2C) with 1.02 (s, 3H, CH3-19), 0.91
(d, 3H, CH3-21, J20, 21 = 6.8 Hz), 0.86 (d, 6H, CH3-26 and
Compound IId (C43H76O10S): 1HNMR (δ ppm, 400
MHz, CD3OD): 5.35 (m, 1H, chol H-6), 4.33–3.43 (25H,
gal. protons and 10 × CH2O), 3.20 (m, 1H, chol H-3),
2.94 (m, 1H, SCH-a), 2.82 (m, 1H, SCH-b), 2.39–0.71

524
Jiao Lu et al.
HO
HO
OH
OH
CN
CN
NC
NC
H2C CHCN
KOH
C2H5OH
O
O
O
O
Conc H2SO4
13
HO
12
O
COOEt
EtOOC
EtOOC
OH
OH
O
O
O
O
O
LiAlH4
HO
O
O
COOEt
14
15
OMs
MsO
O
O
O
O
MsCl
MsO
OMs
16
HO
O
O
n
3a–f n=1–6
NaH
OMs
MsO
MsO
O
O
O
O
O
O
O
n
17a–f n=1–6
NaI
I
I
O
O
O
O
I
O
O
O
n
18a–f n=1–6
OAc
SH
AcO
O
OAc
OAc
10
DIPEA
AcO
AcO
OAc
S
O
AcO
AcO
OAc
O
S
O
O
O
OAc
OAc
OAc
OS
AcO
O
O
O
n
O
OAc
OAc
19a–f n=1–6
OH
OH
S
O
HO
HO
HO
OH
O
O
O
O
S
O
OH
OH
OH
OS
MeONa
HO
O
O
O
n
OH
OH
IIIa–f n=1–6
Scheme 2. e synthesis route of target compound IIIa-f.

Hepatoma cell targeting with galactosylated LPD
525
OH
OH
O
OH
OH
S
S
O
HO
HO
S
O
O
O
O
OH
HO
OH
OH
O
O
O
O
n
OH
OH
IIIa–f n=1–6
Figure 2. e structure of tri-antennary galactosides (target compound IIIa–f).
Synthesis route of target compound IVa–f (Figure 3,
Scheme 3B)
CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.71 (s, 3H, CH3-18). MS
(m/z): (M+Na)+ 1426.2.
1
e sexi-(β-cyanoethyl) cyclohexanehexo-lether 21 could
be afforded by Michael reaction of cyclohexanehexol with
acrylonitrile, followed by ethanolysis, resulting in the
sexi-(β-carbethoxyethyl) cyclohexanehexolether 22. e
reduction of four ester functionalities of 22 was achieved
under anhydrous condition, yielding the hexa-antennary
alcohol 23, which was subsequently methysulfonylated
to the intermediate 24 (Scheme 3A). By coupling with 3a,
25a was obtained and could be converted into iodide 26a
by treating with NaI in refluxing butanone. Using 2,3,4,6-
tetra-O-acetyl-1-thio-β-D-galactopyranose as glycosyl
donor, the clustered penta-saccharide derivative 27a
was synthesized, the deacetylation of which yielded the
desired product IVa (C85H150O34S5). 1HNMR (δ ppm, 400
MHz, CD3OD): 5.37 (m, 1H, chol H-6), 4.34 (t, 5H, 5 × gal.
H1, J = 9.6 Hz), 3.93-3.43 (m, 56H, 5 × gal. H2, H3, H4, H5,
H6, H6’ protons and 6 × CHOCH2, 2 × OCH2CH2O), 3.19
(m, 1H, chol H-3), 2.89-2.81 (m, 10H, SCH-a, SCH-b),
2.40–0.71 (remaining chol protons, 6 × CCH2C) with 1.02
(s, 3H, CH3-19), 0.93 (d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.87
(d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.71 (s,
3H, CH3-18). MS (m/z): (M+Na)+ 1898.1.
Compound IIIc (C70H126O24S3): HNMR (δ ppm, 400
MHz, CD3OD): 5.36 (m, 1H, chol H-6), 4.30 (d, 3H,
3 × gal. H1, J= 9.2 Hz), 3.88 (d, 3H, 3 × gal. H4, J= 3.6 Hz),
3.75-3.43 (m, 41H, 3 × gal. H2, H3, H5, H6, H6’ protons and
13 × CH2O), 3.36 (s, 8H, 4 × CCH2O), 3.18 (m, 1H, chol
H-3), 2.79 (m, 6H, SCH-a, SCH-b), 2.37–0.71 (remaining
chol protons, 4 × CCH2C) with 1.02 (s, 3H, CH3-19), 0.91
(d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.86 (d, 6H, CH3-26 and
CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.71 (s, 3H, CH3-18). MS
(m/z): (M+Na)+ 1469.9.
Compound IIId (C72H130O25S3): 1HNMR (δ ppm,
400 MHz, CD3OD): 5.35 (m, 1H, chol H-6), 4.31 (d, 3H,
3 × gal. H1, J= 9.2 Hz), 3.88 (d, 3H, 3 × gal. H4, J= 3.6 Hz),
3.78-3.43 (m, 45H, 3 × gal. H2, H3, H5, H6, H6’ protons and
15 × CH2O), 3.34 (s, 8H, 4 × CCH2O), 3.18 (m, 1H, chol
H-3), 2.78 (m, 6H, SCH-a, SCH-b), 2.37-0.70 (remaining
chol protons, 4 × CCH2C) with 1.02 (s, 3H, CH3-19), 0.91
(d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.86 (d, 6H, CH3-26 and
CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.70 (s, 3H, CH3-18). MS
(m/z): (M+Na)+ 1514.1.
Compound IIIe (C74H134O26S3): 1HNMR (δppm, 400 MHz,
CD3OD): 5.35 (m, 1H, chol H-6), 4.31 (d, 3H, 3×gal. H1, J=9.8
Hz), 3.88 (d, 3H, 3×gal. H4, J=3.2 Hz), 3.77-3.43 (m, 49H,
3×gal. H2, H3, H5, H6, H6’ protons and 17×CH2O), 3.34 (s,
8H, 4×CCH2O), 3.18 (m, 1H, chol H-3), 2.78 (m, 6H, SCH-a,
SCH-b), 2.37-0.70 (remaining chol protons, 4×CCH2C) with
1.02 (s, 3H, CH3-19), 0.91 (d, 3H, CH3-21, J20, 21 =6.4 Hz), 0.86
(d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27 =6.4 Hz), 0.70 (s, 3H,
CH3-18). MS (m/z): (M+Na)+ 1557.9.
e synthetic route of compound IVb-f is similar to that of
compound IVa. Compound IVb (C87H154O35S5): 1HNMR
(δ ppm, 400 MHz, CD3OD): 5.36 (m, 1H, chol H-6), 4.34
(m, 5H, 5 × gal. H1), 3.95–3.43 (m, 60H, 5 × gal. H2, H3,
H4, H5, H6, H6’ protons and 6 × CHOCH2, 3 × OCH2CH2O),
3.18 (m, 1H, chol H-3), 2.90–2.81 (m, 10H, SCH-a, SCH-
b), 2.39-0.71 (remaining chol protons, 6 × CCH2C) with
1.02 (s, 3H, CH3-19), 0.93 (d, 3H, CH3-21, J20, 21 = 6.8 Hz),
0.87 (d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27 = 6.4 Hz),
0.71 (s, 3H, CH3-18). MS (m/z): (M+Na)+ 1941.9.
Compound IIIf (C76H138O27S3): 1HNMR (δ ppm, 400 MHz,
CD3OD): 5.35 (m, 1H, chol H-6), 4.31 (d, 3H, 3×gal. H1,
J=9.8Hz), 3.88 (d, 3H, 3×gal. H4, J=3.2 Hz), 3.77-3.43 (m,
53H, 3×gal. H2, H3, H5, H6, H6’ protons and 19×CH2O), 3.34 (s,
8H, 4×CCH2O), 3.18 (m, 1H, chol H-3), 2.78 (m, 6H, SCH-a,
SCH-b), 2.37–0.71 (remaining chol protons, 4×CCH2C) with
1.02 (s, 3H, CH3-19), 0.91 (d, 3H, CH3-21, J20, 21 =6.4 Hz), 0.87
(d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27 =6.8 Hz), 0.71 (s, 3H,
CH3-18). MS (m/z): (M+Na)+ 1601.9.
1
Compound IVc (C89H158O36S5): HNMR (δ ppm, 400
MHz, CD3OD): 5.35 (m, 1H, chol H-6), 4.34 (m, 5H,
5 × gal. H1), 3.93–3.43 (m, 64H, 5 × gal. H2, H3, H4, H5, H6,
H6’ protons and 6 × CHOCH2, 4×OCH2CH2O), 3.19 (m, 1H,
chol H-3), 2.89–2.81 (m, 10H, SCH-a, SCH-b), 2.40- 0.71
(remaining chol protons, 6 × CCH2C) with 1.01 (s, 3H,

526
Jiao Lu et al.
CH3-19), 0.93(d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.87 (d, 6H,
CH3-26 and CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.71 (s, 3H,
CH3-18). MS (m/z): (M+Na)+ 1985.9.
chol H-3), 2.89–2.81 (m, 10H, SCH-a, SCH-b), 2.40–0.70
(remaining chol protons, 6 × CCH2C) with 1.01 (s, 3H,
CH3-19), 0.93(d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.87 (d, 6H,
CH3-26 and CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.70 (s, 3H,
CH3-18). MS (m/z): (M+Na)+ 2073.8.
Compound IVd (C91H162O37S5): 5.34 (m, 1H, chol H-6),
4.33(m, 5H, 5 × gal. H1), 3.91–3.43 (m, 68H, 5 × gal. H2, H3,
H4, H5, H6, H6’ protons and 6 × CHOCH2, 5 × OCH2CH2O),
3.18 (m, 1H, chol H-3), 2.89–2.81 (m, 10H, SCH-a, SCH-b),
2.40–0.71 (remaining chol protons, 6 × CCH2C) with 1.01
(s, 3H, CH3-19), 0.93 (d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.87
(d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.71 (s,
3H, CH3-18). MS (m/z): (M+Na)+ 2029.9.
1
Compound IVf (C95H170O42S5): HNMR (δ ppm, 400
MHz, CD3OD): 5.36 (m, 1H, chol H-6), 4.34 (m, 5H,
5 × gal. H1), 3.93–3.43 (m, 76H, 5 × gal. H2, H3, H4, H5,
H6, H6’ protons and 6 × CHOCH2, 7 × OCH2CH2O), 3.19
(m, 1H, chol H-3), 2.89–2.81 (m, 10H, SCH-a, SCH-b),
2.40–0.71 (remaining chol protons, 6 × CCH2C) with 1.02
(s, 3H, CH3-19), 0.93(d, 3H, CH3-21, J20, 21 = 6.4 Hz), 0.87
(d, 6H, CH3-26 and CH3-27, J25, 26 = J25, 27 = 6.4 Hz), 0.71 (s,
3H, CH3-18). MS (m/z): (M+Na)+ 2117.6.
1
Compound IVe (C93H166O38S5): HNMR (δ ppm, 400
MHz, CD3OD): 5.35 (m, 1H, chol H-6), 4.34 (m, 5H,
5×gal. H1), 3.93–3.43 (m, 72H, 5×gal. H2, H3, H4, H5, H6, H6’
protons and 6 × CHOCH2, 6 × OCH2CH2O), 3.18 (m, 1H,
NC
OH
O
CN
NC
NC
HO
HO
OH
OH
O
O
O
O
C2H5OH
H2C CHCN
KOH
Conc H2SO4
CN
O
OH
CN
20
21
OH
O
EtOOC
O
COOEt
EtOOC
EtOOC
O
O
O
HO
OH
O
O
O
LiAlH
4
HO
O
OH
O
COOEt
O
O
COOEt
HO
23
22
OMs
O
O
O
O
O
MsO
MsO
OMs
MsCl
OMs
O
MsO
24
Scheme 3A. e synthesis route of intermediate compound 24.

Hepatoma cell targeting with galactosylated LPD
527
OMs
O
O
O
O
n
NaI
3a-f
OMs
OMs
O
O
O
O
MsO
MsO
NaH
O
OMs
OMs
O
O
O
O
MsO
MsO
O
O
MsO
24
25a–f n=1–6
MsO
OAc
SH
AcO
O
O
O
O
n
OAc
OAc
10
O
I
O
O
O
O
I
I
DIPEA
I
O
I
26a–f n=1–6
O
O
O
O
O
O
O
n
n
OAc
OH
OH
OAc
AcO
O
HO
O
O
O
OAc
S
S
O
OH
OH
OH
HO
HO
O
AcO
S
S
S
O
O
O
O
O
O
O
OAc
MeONa
HO
OH
AcO
AcO
OH
OH
OAc
OAc
AcO
HO
O
O
S
S
O
OH
OAc
O
O
O
AcO
S
OAc
HO
AcO
OH
OAc
OH
OAc
S
O
OH
HO
S
O
OAc
AcO
OH
OAc
27a–f n=1–6
IVa–f n=1–6
Scheme 3B. Synthesis route of target compound IVa–f.
quantified by UV absorbance at 260 and 280nm on a GBC
UV cintra 10e Spectrophotometer. e structural integrity
and topology of purified DNA was analyzed by agarose gel
electrophoresis.
Plasmid DNA preparation
e plasmid pORF lacZ (3.54kb) is a eukaryotic expres-
sion vector which contains the EF-1a/HTLV hybrid
promoter within an intron. e lacZ gene codes for the
enzyme β-galactosidase, whose activity allows for the
quick determination of cells expressing the lacZ gene.
pORF-lacZ plasmid DNA was isolated and purified from
DH5α E. coli using the Qiagen Giga Endo-free plasmid
purification kit. DNA concentration and purity were
LPD preparation
A control cationic liposome was composed of DDAB,
DOPE and cholesterol at a molar ratio of 3:1:1. By contrast,
Gal-liposomes, specifically, Liposome IIa–f, Liposome

528
Jiao Lu et al.
Electron microscopy
LPD0 and Gal-LPD were examined by transmission elec-
tron microscopy (JEM-100SX, Japan). Samples were pre-
pared by placing a drop of LPD suspension onto a copper
grid and air-drying, followed by negative staining with a
drop of 2% aqueous solution of uranyl acetate for contrast
enhancement. e air dried samples were then directly
examined under the transmission microscope.
O
O
O
O
n
OH
OH
OH
S
O
OH
HO
HO
O
S
O
O
O
O
OH
HO
OH
OH
OH
Size and zeta potential
OH
S
O
Diameter and surface charge of DNA polyplexes, the con-
trol cationic liposome LPD0 and LPDs were measured by
photon correlation spectroscopy (PCS) (Zetasizer Nano
ZS90, Malvern instruments Ltd., UK) with a 50 mV laser. A
measure of 20 mL of LPD was diluted by 2mL of 5% dex-
trose and added into the sample cell. e measurement
time was set to 2min (rapid measurement) and each run
consisted of 10 subruns. e measurements were done
at 25°C at an angel of 90°C. e size distribution follows
a lognormal distribution. e potential of the lipid car-
riers at the surface of spheres, called the zeta potential,
which was derived from mobility of particles in electric
field by applying the smoluchowsky relationship, was
measured at least three times at appropriate concentra-
tions of samples.
OH
O
S
O
OH
HO
OH
OH
S
HO
O
OH
IVa–f n=1–6
OH
Figure 3. Structure of penta-antennary galactosides (target com-
pound IVa–f).
IIIa–f, and Liposome IVa–f, consisted of DDAB, DOPE,
and cholesterylated galactosides with different amount
of galactose residues and various spacer lengths, at a
molar ratio of 3:1:1. e lipid mixture was dissolved in
appropriate chloroform and a thin lipid film was formed
in a round-bottomed flask by drying the solvent under a
stream of nitrogen gas. Residual chloroform was further
removed by placing the flask in a desiccator vacuum for
30 min. e film was hydrated with the addition of 5%
dextrose. en the lipid suspension was ultrasounded by
ultrasonic probe.
LPDs were prepared as described previously (Sun
et al., 2005). To form protamine/DNA polycation, the
protamine/DNA ratio was used 1.5:1. In brief, DNA
and protamine were both diluted with 5% dextrose to
get stock concentrations of 500 µg/mL and 150 µg/mL,
respectively. Appropriate amount of DNA stock solution
was added dropwise to a certain amount of protamine
stock solution with mild stirring. After mixing, the
resulted polyplexes (DNA-polycation complexes) were
incubated for 10 min at room temperature. Preformed
cationic liposomes were subsequently added to the
DNA/protamine polyplexes under mild vortexing to
achieve the desired final component concentrations
and ratios. us, the nongalactosylated formulation
LPD0 was obtained by mixing the polyplexes and the
control cationic liposome mentioned above while the
galactosylated formulations LPDIIa–f, LPDIIIa–f, and
LPDIVa–f were prepared by mixing the polyplexes and
Gal-liposomes consisting of DDAB, DOPE and corre-
sponding cholesterylated galactosides at a molar ratio
of 3:1:1.
Protection assay of DNA
Briefly, 4 µL of DNase I (2 units) or PBS in DNase digestion
buffer (50 mmol, Tris-Cl, pH 7.6 and 10 mmol MgCl2) was
added to 4 µg of naked plasmid DNA or LPDs, and incu-
bated at 37°C for different time periods. For DNase inac-
tivation, all samples were treated with EDTA (0.5mol).
en Triton X-100 was mixed with each sample to destroy
the lipolayers. After 2 min’ brief shaking, heparin solu-
tion was added to the mixture at a final concentration of
0.9% (w/v). e final samples were incubated for 2 h and
electrophoresis was performed with 1.0% agarose gel in
TAE running buffer for 2 h at 80 V. Images were analyzed
using Kodak Digital Science 1D software to obtain in the
form of volume per area.
Cell transfection
HepG2 and A549 cells were cultured in RPMI-1640 with
10% fetal bovine serum and streptomycin (100mg/mL).
e cells were seeded at 2 × 105 cells/well onto 12-well
plates 24h before transfection so that they would be 70%
confluence at the time of transfection. After the cells were
washed twice by PBS, 0.5mL of serum-free, antibiotics-
free medium were added into each well. For each well,
LPD containing 4 µg (if not mentioned specifically)
of pORF-1acZ was diluted to 0.5 mL by the serum-free

Hepatoma cell targeting with galactosylated LPD
Results
529
medium, and then gently overlaid onto the cells. e
cells were incubated with LPD for 5 h at 37°C in a CO2
incubator. Following incubation, LPD was removed and
the cell surfaces were rinsed thoroughly and treated
with 2 mL fresh complete medium. en the cells were
returned to the incubator for a further 45 h to allow intra-
cellular gene expression to proceed. As a control, the
commercial transfection reagent Lipofectamine™ 2000
was used to transfect cells according to manufacturers’s
instructions.
LPD characterization
To prepare stable liposomes with high transfect effi-
ciency, liposomes consisting of DDAB, DOPE, and
cholesterol as well as protamine–DNA complexes
protamine at different ratios were formulated. e
protamine/DNA and the DDAB/ DOPE/cholesterol
ratio were optimized at 1.5 and 3:1:1, respectively,
considering their size, zeta potential and transfect
efficiency (Data not shown). Besides, we investigated
the influence of the DDAB/DNA ratio on the particle
size and zeta potential, as well as examined the in vitro
transfect activity of LPD0, LPDIIb, LPDIIIc, and LPDIVe
composing of different DDAB/DNA ratios. As shown in
Table 1, for LPDs with the same DDAB/DNA ratio, the
zeta potential decreased generally when cholesteryl-
ated thiogalactosides containing more galactosidase
residues were incorporated into the structure. Also,
it is notable that for LPD0 and LPDIIb groups, when
the DDAB/DNA ratio was 2, the gene transfer ability
was superior in HepG2 cells. However, for LPDIIIc and
LPDIVe groups, though the transfect efficiency was
not the highest with a DDAB/DNA ratio of 2, there was
not significant difference (P > 0.05) compared with the
highest within each group. erefore, considering the
possible negative effects of DDAB on cell viability, the
DDAB/DNA ratio was fixed at 2 in the following study.
LPD0 (without galactosylated cholesterols) and Gal-
LPDs ( LPDIIa–f, LPDIIIa–f, and LPDIVa–f) were prepared
by adding cationic liposomes to protamine-DNA com-
plexes. In general, zeta potential and particle size of LPDs
were measured to between 19–57 mV and 107–221 nm,
respectively (data not shown). Based on electron micros-
copy, the structure and morphology of the resulting LPD0
was described. As shown in Figure 4, the LPD0 particles
and Gal-LPD particles (LPDIVe) were all nearly round-
shaped, indicating the addition of galactose residues did
not significantly change the shape of LPD while bringing
an increase in particle size.
β-Galactosidase assay
Expression of β-galactosidase genes was measured with
β-gal assay kit according to the manufacturer’s instruc-
tion. e transfected HepG2 cells were washed once with
PBS and lysed with lysis buffer (30 mL/well). Cell debris
was removed by centrifugation at 12,000 g and 10 mL of
the supernatant added to 50 mL of cleavage buffer con-
taining β-mercaptoethanol and 17 mL of ONPG solution.
After incubation for 30 min at 37°C, the absorbance at
415 nm was measured. e total protein concentrations
in cell lysates were determined using BCA assay (Pierce
Chemical).
Cytotoxicity assay
Cytotoxicity of LPDs was studied using a colorimetric
MTT assay. Briefly, cells were collected, counted and
seeded in growth medium (200 µL) into 96-well plates
at a density of 15,000 cells/well. One day later, the cells
were transfected as described above. After incubation
for 45 h, 20 µL of MTT (5 mg/mL) solution was added
into each well and was allowed to react for 4 h at 37°C.
en the medium of each well was replaced with 150
µL of DMSO and the plate was incubated for 10 min at
room temperature. Absorbance at 570 nm was meas-
ured with an ELISA plate reader (Bio-Rad, Microplate
Reader 550).
Competitive inhibition assay of galactose
e DNA in the gene vehicle should be protected from
degradation by DNase in the extracellular environment.
erefore, DNA protection assay was performed. As
shown in Figure 5, after incubation with DNase I, naked
plasmid DNA were completely digested within 5min (lane
2). In contrast, the plasmid DNA incorporated in LPD
still remained their integrity in supercoiled forms after
incubation with DNase for 5min and 2h (lanes 3 and 4),
suggesting its capability to be a potential gene vector.
HepG2 cells were cultured in RPMI-1640 with 10% fetal
bovine serum and streptomycin (100 mg/mL). e cells
were seeded at 2 × 105 cells/well onto 12-well plates 24 h
before transfection, so that they would be 70% confluence
at the time of transfection. After the cells were washed
twice by PBS, 0.5mL serum-free, antibiotics-free medium
were added into each well. For each well, LPD containing
4 µg pORF-1acZ, LPD/80 mM D-galactose or LPD/80 mM
D-mannose were diluted to 0.5 mL by the serum-free
medium, and then gently overlaid onto the cells. Other
steps were performed as previously mentioned in the
cell transfection study and the transfection efficiency
was measured.
Cell transfection studies
e modified LPDs containing the synthesized ligands
were examined for their transfection efficiency. Two

530
Jiao Lu et al.
Table 1. e influence of DDAB/DNA ratios on zeta potential, particle size, and β-galactosidase activity in HepG2 cells for LPD0, LPDIIb,
LPDIIIc and LPDIVe. e DDAB, DOPE, and cholesterol was fixed at 1.5 and 3: 1: 1, respectively. e data points represent the mean SD of three
experiments.
LPD
DDAB/DNA (w/w)
Zeta potential (mV)
33.6 1.8
49.7 3.8
58.9 4.0
56.8 3.2
29.0 2.7
48.5 3.5
55.9 3.4
52.2 5.2
19.4 2.5
22.0 2.2
26.2 4.8
34.5 2.9
20.3 2.6
21.8 3.7
27.8 3.5
32.6 4.6
Particle size (nm)
109 8.5
β-Galactosidase activity (mU/mg protein)
22.01 1.98
LPD0
2
4
6
8
2
4
6
8
2
4
6
8
2
4
6
8
107 10.2
110 5.3
18.60 2.35
15.97 2.61
115 7.8
14.28 3.67
LPD IIb
LPD IIIc
LPD IVe
112 13.2
109 18.0
107 15.8
112 16.5
134 12.8
125 21.6
109 10.5
109 15.9
221 14.7
208 12.8
194 23.1
193 21.5
35.43 3.25
28.60 3.41
18.50 3.52
11.50 2.17
39.79 4.49
43.21 5.20
38.33 2.86
15.77 2.88
47.20 6.28
50.21 3.47
52.26 4.38
30.75 3.69
A
B
4
3
2
1
Figure 4. Transmission electronic micrographs of (A) LPD0, bar 200 nm,
(B) LPDIVe, bar 200nm. e nongalactosylated formulation LPD0 was
obtained by mixing the DNA–protamine polyplex and the control
cationic liposome and LPDIVe (Gal-LPD) was prepared by mixing
the polyplex and Gal-liposome (DDAB/DOPE/cholesterylated galac-
tosides 3:1:1). en they were examined by transmission electron
microscopy (JEM-100SX, Japan).
different cell lines were used: HepG2 cells derived
from a human hepatocellular carcinoma expressing the
ASGP-R (Schwartz et al., 1981), and the human alveo-
lar epithelial cell line A549, which lacks the receptor
(Nikken et al., 2009). e transfection efficacy of LPDs
in gene delivery to HepG2 and A549 cells was stud-
ied to identify the receptor-mediated internalization.
As indicated in Figure 7, compared with LPD0 which
did not contain galactose residues, LPDIIa, LPDIIb,
LPDIIIb, LPDIIIc, and LPDIVd–f has improved trans-
fect efficiency in HepG2 cells. Also, it is notable that
LPDIIb, LPDIIIc and LPDIVe, which possessed the most
superior transfect efficiency among each galactosylated
LPD group with the same spacer length, showed sig-
nificant higher transfect efficiency than the commercial
Figure 5. DNA protection assay by agarose gel electrophoresis of
naked DNA and LPD subjected to DNase degradation. Lane 1: Naked
DNA. Lane 2: Naked DNA incubated with DNase 5min. Lane 3: LPD
incubates with DNase 5min. Lane 4: LPD incubates with DNase 2h.
transfect reagent Lipofectamine™ 2000 (P < 0.05) In
contrast, all the LPDs containing cholesterylated thi-
ogalactosides did not increase the transfect efficiency
in A549 cells compared with LPD0. Instead, their trans-
fect efficiency was significantly decreased in contrast to
LPD0 (P < 0.05) except LPDIIc in A549 cells (Figure 6).
Besides, it is worth mentioning that those Gal-LPDs,
including LPDIIc–f, LPDIIIa, LPDIIId–f, LPDIVa–c,
did not significantly improve the transfect efficiency

Hepatoma cell targeting with galactosylated LPD
531
A
40
30
20
10
0
*
*
*
*
**
**
**
B
40
30
20
10
0
**
*
*
*
**
**
**
**
C
40
30
20
10
0
*
**
**
**
**
**
**
**
Figure 6. Comparison of transfection efficiencies of LPDs and Lipofectamine™ 2000 Transfection Reagent in A549 cells. (A) LPD0, Lipofectamine™
2000 and LPDIIa–f, (B) LPD0, Lipofectamine™ 2000 and LPDIIIa–f, (C) LPD0, Lipofectamine™ 2000 and LPDIVa–f. Transfect efficiency was com-
pared between LPD with the highest transfect ability (LPD0) and other LPDs as well as Lipofectamine™ 2000 within each group. e data points
represent the mean SD of three experiments (P <0.05*, P<0.01**). e transfect efficiency in A549 cells did not increase using LPDs bearing
galactose residues compared with LPD0.
into HepG2 cells in contrast to the control group LPD0
(P > 0.05) (Figure 7).
LPDIVe, each of which performed the highest transfect
efficiency to HepG2 cells within the same galactose
density group. As demonstrated in Figure 8, the trans-
fect activities of LPDIIb, LPDIIIc and LPDIVe can be sig-
nificantly inhibited at the presence of 80 mmol galactose
(P< 0.05), while the existence of mannose at the same
concentration failed to suppress their gene transfer. In
addition, the inhibition effect by galactose was slightly
Galactose competitive inhibition assay
To further validate the recognition of galactosidase
residues of LPDs by the ASGP-R, galactose competitive
inhibition assay was carried out on LPDIIb, LPDIIIc, and

532
Jiao Lu et al.
A
60
50
40
30
20
*
*
*
**
**
**
10
0
**
B
60
50
40
30
20
10
0
*
*
*
*
*
**
**
C
60
50
40
30
20
10
0
*
**
**
**
**
**
Figure 7. Comparison of transfection efficiencies of LPDs and Lipofectamine™ 2000 Transfection Reagent in HepG2 cells. (A) LPD0, Lipofectamine™
2000 and LPDIIa–f; (B) LPD0, Lipofectamine™ 2000 and LPDIIIa–f; (C) LPD0, Lipofectamine™ 2000 and LPDIVa–f. Transfect efficiency was com-
pared between LPD with the highest transfect ability (A: LPDIIb, B: LPDIIIc, C: LPDIVe) and other LPDs as well as Lipofectamine™ 2000. e data
points represent the mean SD of three experiments (P <0.05*, P<0.01**). LPDIIa–b, LPDIIIb–c, and LPDIVd–f performed significant higher
transfect efficiencies in HepG2 cells.
enhanced in LPDs with more galactose residues, but
without significant difference (P> 0.05). Taken together,
these results were in accordance with in vitro gene
transfer studies into HepG2 and A549 cells, validating
the specific recognition of galactose residues of LPDs by
ASGPR on the surface of HepG2 cells and the strength-
ened affinity between the ASGPR and the ligand with
more galactose residues.

Hepatoma cell targeting with galactosylated LPD
533
60
50
40
30
20
10
0
A
110
105
LPD only
LPD+mannose
LPD+galactose
100
95
LPD II
LPD III
LPD IV
90
85
80
LPD0 LPDIIb LPDIIIc LPDIVe
a
b
c
d
e
f
Figure 8. Galactose competitive inhibition assay. Effects of man-
nose or galactose (80mM) on the transfection activities of LPDIIb,
LPDIIIc, and LPDIVe in HepG2 cells. e data points represent the
mean S.D. of three experiments. Cells were seeded at 2×105 cells/
well onto 12-well plates 24h before transfection. After the cells were
washed twice by PBS, 0.5mL serum and antibiotics-free medium were
added into each well. For each well, LPD containing 4 µg pORF-1acZ,
LPD/80mM D-galactose or LPD/80mM D-mannose were diluted
to 0.5mL by the serum-free medium, and then gently overlaid onto
the cells. After incubation of 6 h, medium were moved and cells were
washed twice with PBS, then the transfection efficiency was measured
by β-galactosidase assay.
B
110
105
LPD II
LPD III
LPD IV
100
95
90
85
80
a
b
c
d
e
Cytotoxicity assay
f
Since the toxicity of nonviral cationic vectors may inter-
fere with their transfection activities, the relative viabili-
ties of two cell types (HepG2 and L929) after exposure
to LPDs with the DNA concentration of 4 mg/mL was
measured by MTT method, in contrast to the control
group LPD0. Figure 9 shows all the Gal-LPDs presented
no obvious toxicities to L929 cell line and HepG2 cell
line. Moreover, they did not have significant difference
(P< 0.05) in toxicity between each other.
Figure 9. Cytotoxicity assay. Cytotoxicity of LPDs in (A) L929, (B)
HepG2 cell lines were measured by MTT test. Cells were plated at
15,000 cells/well into 96-well plates one day before transfection. en
cells were transfected using LPDs. After incubation of 45 h, cell viabil-
ity was measured by MTT test. Data are means of three experiments
(SD < 10%).
of multi-antennary galactosyl ligands, differing in their
inter-galactose distances.
It is well recognized that the mode of formation of the
lipoplexes strongly determines the final physicochemical
characteristics of the lipoplexes, including size, charge,
density and colloidal stability, which consequently mod-
ulates their biological activity both in vivo and in vitro
(Pedroso et al., 2001). erefore, it is of crucial importance
to optimize the formulation of lipid-polycation-DNA
complexes as well as characterize these parameters that
govern transfection profiles. In the LPD characterization
assay, we determined the zeta potential and particle size
of LPDs. As shown in Table 1, the particle size of these
LPDs briefly increased with more galactose residues
(LPDIVe > LPDIIIc > LPDIIb). is might be attributed
to a shielding effect on part charges by galactose residues,
which leads to the increase of the hydrophilicity of Gal-
LPD, resulting in enhanced hydration and thus cause the
size to increase.
In cell transfection studies, we measured
β-galactosidase activity in both A549 and HepG2 cells.
As indicated in Figure 7, LPDIIb, LPDIIIc, and LPDIVd–f
exhibited higher transfect efficiency compared with
LPD0 and lipofectamine™ 2000 in HepG2 cells whereas
the transfect activity in A549 cells decreased using LPDs
containing cholesterylated thiogalactosides, in contrast
Discussion
Our present study aims to further elucidate how struc-
tural differences of the galactose ligands on the gene
transfer vector LPD influence the transfect ability by
affecting the binding affinity between the ligands and
the receptor ASGPR. Previous structure–activity studies
of synthetic galactosides and isolated oligoantennary
glycopeptides with the asialoglycoprotein receptor have
provided some insight into the basic structural require-
ments for ligand recognition. e nature of the branching
pattern of the oligosaccharide component as well as the
distance between the galactosyl residues were suggested
to be major factors determining the affinity of a ligand
(Lee et al., 1983; Baenziger & Maynard, 1980; Kawaguchi
et al., 1980; Connolly et al., 1982; Lee, Lin, & Lee, 1984).
erefore, to illuminate the effects of these factors on
transfect activity, we synthesized a series of cluster lig-
ands varying in galactose substitution degree and spacer
lengths of the cholesterol derivatives by virtue of a flex-
ible chemical strategy which allows for the preparation

534
Jiao Lu et al.
to the control group LPD0. is can be rationalized by
the receptor mediated internalization. ASGPR, highly
expressed on the surface of HepG2 cells, can recognize
galactosidase residues of cholesterylated thiogalactos-
ides and therefore facilitate the internalization of LPD
into HepG2 cells. However, the human alveolar epi-
thelial cell line A549 cell lacks ASGPR; moreover, the
presence of galactose residues may cause an enhanced
surface hydrophilicity and facilitate the formation of
water shell to the surfaces of galactosylated LPDs, which
was hypothesized to hinder the membrane fusion and
consequently suppressed the gene expression of galac-
tosylated LPDs in A549 cells. Besides, it is quite obvi-
ous that regarding the gene transfer ability, LPDIVe >
LPDIIIc > LPDIIb. is is in consistent with previous
research by Reiko T. Lee et al., which has established
each addition of a galactosyl residue to an existing lig-
and structure invariably increased the binding affinity
of such a ligand (Lee, Lin, & Lee, 1984). And according
to Kawasaki et al., multivalent ligands showed a strong
“cluster effect,” which is defined as affinity enhance-
ment over and beyond what would be expected from
the concentration increase of the determinant sugar in
a multivalent ligand and geometries (Balaji et al., 1993).
erefore, for those LPDs bearing optimized spatial
arrangement and restricited geometry, the increase in
galactose densities undoubtedly improved the ligand’s
binding affinity, leading to enhanced gene delivery into
hepatocytes.
Also, it is notable that among those Gal-LPDs, includ-
ing LPDIIa, LPDIIc–f, LPDIIIa, LPDIIId–f, LPDIVa–c,
though bearing higher galactose dentisities than LPD0,
did not significantly increase the transfect efficiency
into HepG2 cells. is can be explained by unoptimized
spacer lengths. At a given level of valency, the binding
strength of a cluster ligand depended mainly on two fac-
tors: (1) the maximum spatial inter-galactose distances
and (2) the flexibility of the arm connecting galactosyl
residues and the branch points (Lee, Lin, & Lee, 1984).
Moreover, optimal interactions with the clustered
galactosides (“cluster effect”) also require well defined
inter-galactose distances (Lee, Lin, & Lee, 1984) and
geometries (Ozaki et al., 1995). Furthermore, previous
work has established a model in which only glycopep-
tides bearing terminal Gal or GalNAc residues that fall
within a restricted spatial relationship can induce a con-
formational alteration in the receptor which is required
for uptake to occur and the spacing of terminal Gal
and GalNAc residues is critical for interaction with the
lectin binding site and that this would be reflected in
the susceptibility of these glycopeptides to endocytosis
(Baenziger & Fiete, 1980), As a consequence, LPDIIa,
LPDIIc–f, LPDIIIa, LPDIIId–f, LPDIVa–c, which lack the
proper dimensions for the binding to the receptor, pre-
sented relatively weak gene transfer ability. By contrast,
LPDIIb, LPDIIIc, and LPDIVe, which attain appropriate
spatial and conformational arrangement for the tight
binding to the receptor, exhibited superior gene transfer
ability to HepG2 cells.
In conclusion, we have established LPDIVe, LPDIIIc,
and LPDIIb, bearing the most appropriate inter-galactose
distances were the most sufficient gene transfer vector at
each given level of valency, with low toxicity. And regard-
ing the substitution degree of galactose residues, under
optimized spatial arrangement, the gene transfer ability
was: tetra- > tri- > mono-antennary galactose contained
LPDs. erefore, our work would prove valuable for future
research in optimizing the strategy for hepatotrophic
gene targeting.
Declaration of interest
We are thankful for the financial support of the National
Natural Science Foundation of China (Grant No:
30600786).
References
Baenziger JU, Fiete D. (1980). Galactose and N-acetylgalactosamine-
specific endocytosis of glycopeptides by isolated rat hepatocytes.
Cell 22: 611–620.
Baenziger JU, Maynard Y. (1980). e asialoglycoprotein receptor. J
Biol Chem 255: 4607–4613.
Balaji PV, Qasba PK, Rao VS. (1993). Molecular dynamics simula-
tions of asialoglycoprotein receptor ligands. Biochemistry
32:12599–12611.
Bider MD, Wahlberg JM, Kammerer RA, Spiess M. (1996). e
oligomerization domain of the asialoglycoprotein receptor pref-
erentially forms 2:2 heterotetramers in vitro. J Biol Chem 271:
31996–32001.
Bock K, Arnarp J, Lönngren J. (1982). e preferred conformation of
oligosaccharides derived from the complex-type carbohydrate
portions of glycoproteins. Eur J Biochem 129: 171–178.
Burns CJ, Field LD, Hashimoto K, Petteys BJ, Ridley DD, Samankumara
Sandanayake KRA. (1999). A convenient synthetic route to dif-
ferentially functionalized long chain polyethylene glycols. Synth
Commun 29: 2337–2347.
Connolly DT, Townsend RR, Kawaguchi K, Bell WR, Lee YC. (1982).
Binding and endocytosis of cluster glycosides by rabbit hepato-
cytes. Evidence for a short-circuit pathway that does not lead to
degradation. J Biol Chem 257: 939–945.
Eggens I, Fenderson B, Toyokuni T, Dean B, Stroud M, Hakomori
S. (1989). Specific interaction between Lex and Lex determi-
nants. A possible basis for cell recognition in preimplantation
embryos and in embryonal carcinoma cells. J Biol Chem 264:
9476–9484.
Gao X, Huang L. (1996). Potentiation of cationic liposome-mediated
gene delivery by polycations. Biochemistry 35: 1027–1036.
Horejsí V, Smolek P, Kocourek J. (1978). Studies on lectins. XXXV.
Water-soluble O-glycosyl polyacrylamide derivatives for specific
precipitation of lectins. Biochim Biophys Acta 538: 293–298.
Hukkamaki J, Pakkanen TT. (2001). Study on catalytic hydrogenation
in synthesis of four-directional amine-terminated dendritic
molecules. Journal of molecular catalysis. A, Chemical. 174:
205–211.
Kawaguchi K, Kuhlenschmidt M, Roseman S, Lee YC. (1980). Synthesis
of some cluster galactosides and their effect on the hepatic galac-
tose-binding system. Arch Biochem Biophys 205: 388–395.

Hepatoma cell targeting with galactosylated LPD
535
Khorev O, Stokmaier D, Schwardt O, Cutting B, Ernst B. (2008).
Trivalent, Gal/GalNAc-containing ligands designed for the
asialoglycoprotein receptor. Bioorg Med Chem 16: 5216–5231.
Lee RT, Lin P, Lee YC. (1984). New synthetic cluster ligands for
galactose/N-acetylgalactosamine-specific lectin of mammalian
liver. Biochemistry 23: 4255–4261.
Lee YC, Townsend RR, Hardy MR, Lönngren J, Arnarp J, Haraldsson
M, Lönn H. (1983). Binding of synthetic oligosaccharides to the
hepatic Gal/GalNAc lectin. Dependence on fine structural fea-
tures. J Biol Chem 258: 199–202.
Li S, Rizzo MA, Bhattacharya S, Huang L. (1998). Characterization of
cationic lipid-protamine-DNA (LPD) complexes for intravenous
gene delivery. Gene er 5: 930–937.
Li S, Huang L. (1997). In vivo gene transfer via intravenous adminis-
tration of cationic lipid-protamine-DNA (LPD) complexes. Gene
er 4: 891–900.
Nikken W, Yen WT, Yi YY. (2009). Targeted co-delivery of drug and
gene with self-assembled galactosylated amphiphilic oligopep-
tide nanostructures for cancer treatment.
Ozaki K, Lee RT, Lee YC, Kawasaki T. (1995). e differences in structural
specificity for recognition and binding between asialoglycoprotein
receptors of liver and macrophages. Glycoconj J 12: 268–274.
Pazur JH. (1981). Affinity chromatography of macromolecular sub-
stances on adsorbents bearing carbohydrate ligands. Adv
Carbohydr Chem Biochem 39: 405–447.
Pedroso de Lima MC, Simões S, Pires P, Faneca H, Dűzqűnes N. (2001).
Cationic lipid-DNA complexes in gene delivery: from biophysics
to biological applications. Adv Drug Deliv Rev 47: 277–294.
Roy R, Tropper FD, Romanowska A. (1992). New strategy in
glycopolymer syntheses. Preparation of antigenic water-
solublepoly(acrylamide-co-p-acrylamido-Phenyl.beta.-lactos-
ide). Bioconjugate Chem 3: 256–261.
Loiseau FA, Hii KK, Hill AM. (2004). Multigram synthesis of well-de-
fined extended bifunctional polyethylene glycol (PEG) chains. J
Org Chem 69: 639–647.
Masotti A, Mossa G, Cametti C, Ortaggi G, Bianco A, Grosso ND,
Malizia D, Esposito C. (2009). Comparison of different commer-
cially available cationic liposome-DNA lipoplexes: Parameters
influencing toxicity and transfection efficiency. Colloids Surf B
Biointerfaces 68: 136–144.
Schwartz AL, Fridovich SE, Knowles BB, Lodish HF. (1981).
Characterization of the asialoglycoprotein receptor in a continu-
ous hepatoma line. J Biol Chem 256: 8878–8881.
Sorgi FL, Bhattacharya S, Huang L. (1997). Protamine sulfate enhances
lipid-mediated gene transfer. Gene er 4: 961–968.
Stowell CP, Lee VC. (1980). Neoglycoproteins: the preparation and
application of synthetic glycoproteins. Adv Carbohydr Chem
Biochem 37: 225–281.
Meier M, Bider MD, Malashkevich VN, Spiess M, Burkhard P. (2000).
Crystal structure of the carbohydrate recognition domain of the
H1 subunit of the asialoglycoprotein receptor. J Mol Biol 300:
857–865.
Newkome GR, Mishra A, Moorefield CN. (2002). Improved synthesis
of an ethereal tetraamine core for dendrimer construction. J Org
Chem 67: 3957–3960.
Stoll MS, Mizuochi T, Childs RA, Feizi T. (1988). Improved procedure
for the construction of neoglycolipids having antigenic and lec-
tin-binding activities, from reducing oligosaccharides. Biochem
J 256: 661–664.
Sun X, Hai L, Wu Y, Hu HY, Zhang ZR. (2005). Targeted gene delivery to
hepatoma cells using galactosylated liposome-polycation-DNA
complexes (LPD). J Drug Target 13: 121–128.