Cell transfection with polycationic cyclodextrin vectors

Article abstract of DOI:10.1016/j.ejps.2004.01.001

Polycationic cyclodextrins (CDs) were complexed with plasmid DNA and their effectiveness as vectors was tested on COS-7 cells. These CDs were modified with pyridylamino, alkylimidazole, methoxyethylamino or primary amine groups at 6-positions of the glucose units. Uncharged CDs, β-CD, hydroxypropyl- β-CD, and dimethyl-β-CD were also tested, but these did not form stable complexes with the DNA and produced only a slight improvement in transfection level over DNA alone. The polycationic CDs neutralised DNA to form stable nanoparticulate complexes. The transfection efficiency of these CDs was dependent on the substituents present, with the most efficient having either an amino, pyridylamino or butylimidazole group at the 6-positions and unmodified 2- and 3-hydroxyls. One of the most effective vectors, heptakispyridylamino CD, produced a 4000-fold increase in transfection level over DNA alone. Levels were improved 10-fold by use of the endosomolytic agent, chloroquine. The transfection efficiency of the best of these systems in serum equals that of DOTAP in serum. Studies with ³²P-labelled plasmid DNA indicate that the polycationic CDs are exceptional promoters of DNA cellular-uptake, the most efficient surpassing DOTAP. Uptake is dependent on proteoglycan-mediated binding to cells. The data imply that intracellular trafficking but not cellular uptake, may be the rate-limiting step in the transfection process. These initial results indicate that CDs are useful templates for further modification to produce molecular constructs capable of enhanced gene delivery.

Full text of DOI:10.1016/j.ejps.2004.01.001

European Journal of Pharmaceutical Sciences 21 (2004) 625–633
Cell transfection with polycationic cyclodextrin vectors
Sally-Ann Cryana,1, Ann Holohanb, Ruth Donohueb, Raphael Darcyb,
Caitriona M. O’Driscolla,∗
a
Department of Pharmaceutics and Pharmaceutical Technology, Trinity College, University of Dublin, Dublin 2, Ireland
b
Department of Chemistry, Centre for Synthesis and Chemical Biology, Conway Institute, University College, Dublin, Ireland
Received 7 August 2003; received in revised form 17 November 2003; accepted 7 January 2004
Abstract
Polycationic cyclodextrins (CDs) were complexed with plasmid DNA and their effectiveness as vectors was tested on COS-7 cells. These
CDs were modified with pyridylamino, alkylimidazole, methoxyethylamino or primary amine groups at 6-positions of the glucose units.
Uncharged CDs, ␤-CD, hydroxypropyl-␤-CD, and dimethyl-␤-CD were also tested, but these did not form stable complexes with the DNA
and produced only a slight improvement in transfection level over DNA alone. The polycationic CDs neutralised DNA to form stable
nanoparticulate complexes. The transfection efficiency of these CDs was dependent on the substituents present, with the most efficient having
either an amino, pyridylamino or butylimidazole group at the 6-positions and unmodified 2- and 3-hydroxyls. One of the most effective
vectors, heptakispyridylamino CD, produced a 4000-fold increase in transfection level over DNA alone. Levels were improved 10-fold by
use of the endosomolytic agent, chloroquine. The transfection efficiency of the best of these systems in serum equals that of DOTAP in
serum. Studies with 32P-labelled plasmid DNA indicate that the polycationic CDs are exceptional promoters of DNA cellular-uptake, the most
efficient surpassing DOTAP. Uptake is dependent on proteoglycan-mediated binding to cells. The data imply that intracellular trafficking but
not cellular uptake, may be the rate-limiting step in the transfection process. These initial results indicate that CDs are useful templates for
further modification to produce molecular constructs capable of enhanced gene delivery.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Gene delivery; Non-viral vectors; Polycationic cyclodextrins; Transfection
1. Introduction
Cyclodextrins (CDs) have many advantages as potential
core molecules for the development of new vectors. The nat-
ural, and some modified, cyclodextrins have well-researched
pharmaceutical and toxicological profiles as drug delivery
agents (Szejtli and Osa, 1996). As oligosaccharides they
have low immunogenicity, with multiple sites for attachment
of functions such as cations or cell-targeting groups, and
potential for complexing with macromolecules as well as
developing supramolecular chemistry (Wenz, 1994; Ravoo
and Darcy, 2000; Ravoo et al., 2001). Properties of CDs as
excipients in conventional drug delivery include the ability
to protect drugs from physical, chemical and enzymatic
degradation, and to enhance membrane permeability (Irie
et al., 1982; Kilsdonk et al., 1995). Neutral CDs have been
shown to interact with nucleic acids (Hoffman and Bock,
1970) and nucleotides (Zhao et al., 1995) and to enhance
their transfection efficiency in vivo (Freeman and Niven,
1996). Cationic CD derivatives have shown an even greater
ability to bind nucleotides (Eliseev and Schneider, 1993,
1994; Schwinté et al., 1998) and enhance delivery by viral
The basic premise of somatic gene therapy is that genes
can be used to cause in vivo production of therapeutic pro-
teins. Safety concerns raised in clinical trials due to reaction
to viral vectors have favoured research into non-viral gene
vectors. These include lipid, peptide, polymeric or den-
drimeric carriers (Luo and Saltzman, 2000). Correlations
of in vitro with in vivo results however have so far been
poor, and there is an obvious need to expand the range
of molecular templates or core molecules both to develop
more effective vectors and to gain a greater understanding
of the factors that limit their activity.
∗
Corresponding author. Present address: School of Pharmacy, Bio-
sciences Institute, University College, Cork, Ireland.
Tel.: +353-21-490-1396; fax: +353-21-490-1377.
E-mail address: caitriona.odriscoll@ucc.ie (C.M. O’Driscoll).
1
Present address: School of Pharmacy, Royal College of Surgeons in
Ireland, St. Stephen’s Green, Dublin 2, Ireland.
0928-0987/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejps.2004.01.001

626
S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
vectors (Croyle et al., 1998). In addition, cyclodextrins have
been incorporated into polycationic polymer and dendrimer
vectors (Gonzalez et al., 1999; Hwang et al., 2001; Arima
et al., 2001).
The aims of this work were to characterise novel plas-
mid DNA complexes formed by polycationic CDs in terms
of size, surface charge, DNA binding ability, stability and
toxicity and to compare the cationic with neutral CDs for
transfection ability in COS-7 cells. Preliminary studies were
carried out on their mechanism of transfection.
2. Materials and methods
2.1. Materials
Polyethylenimine (PEI) (high molecular weight, 25 kDa)
and poly(l-lysine) (pLL) were purchased from Sigma–
Aldrich;
1,2-dioleoyl-3-trimethylammonium
propane
(DOTAP) from Roche Molecular Biochemicals; Super-
fect from Qiagen; beta-cyclodextrin (␤-CD) from Sterling
Organics; 2,6-dimethyl-␤-cyclodextrin (DM-CD) from
Janssen; 2-hydroxypropyl-␤-cyclodextrin (HP-CD) from
Pharmatec. Water was doubly distilled deionised water
(DDDW).
Fig. 1. Structures of the polyamino CDs.
2.2. Cell line and cell culture
(using the Hind III restriction enzyme) pCMVluc plasmid
(25 ng) was denatured and mixed with 4 ␮l of High Prime
Solution (Boehringer Mannheim) and 5 ␮l of ␣32P-dCTP
(1.85 × 109 Bq) on ice and reacted for 10 min at 37 ◦C.
Unincorporated ␣32P-dCTP was removed using a freshly
prepared Sepharose column (Uduehi et al., 1999; Cryan and
O’Driscoll, 2003).
COS-7 cells, a fibroblast-like cell line established by
SV40 transformation of CV-1 African green monkey kidney
cells, were obtained from European Collection of Animal
Cell Cultures (ECACC), UK. The cell line was cultured
using Dulbecco’s Modified Eagle’s Medium (DMEM) sup-
plemented with foetal calf serum (10%, v/v), l-glutamine
(1%, w/v), gentamicin (100 mg/ml) and sodium pyruvate
(1 mM). Cells were maintained at 37 ◦C in a 5% CO2 hu-
midified atmosphere. The COS-7 cells were subcultured 1
in 5, twice weekly.
2.5. Synthesis of the polyaminocyclodextrins
Heptakis(6-amino-6-deoxy)-␤-cyclodextrin was prepared
by a literature method (Breslow et al., 1980) but starting from
the heptabromo rather than the heptatosylated cyclodextrin.
The other polyaminocyclodextrins (Fig. 1) were prepared
from heptakis(6-bromo-6-deoxy)-␤-cyclodextrin (Chmurski
and Defaye, 1999) or its heptakis (2,3-di-O-acetate) by
treatment with the corresponding amine. The bromi-
nated cyclodextrin (0.1 mmol) was reacted with amine
(4–10 mmol, as solvent or dissolved in DMF) at 45–80 ◦C
for 2–3 days. Completion of reaction was confirmed
by TLC on Merck Kieselgel 60 plates by elution with
chloroform–methanol–water (50:10:1) or ethyl acetate–iso-
propanol–ammonium hydroxide–water (7:7:5:2), and de-
velopment with 5% sulfuric acid in ethanol followed by
heating. The products were obtained in yields of 60–80%
after purification by precipitation from methanol with ether
followed by treatment with basic ion-exchange resin. The
structures of the polyaminocyclodextrins were confirmed
using elemental microanalysis, high-field proton and carbon
2.3. Plasmid DNA
The pCMVluc plasmid contains the firefly luciferase gene
under the control of the cytomegalovirus enhancer/promoter.
This plasmid was maintained and propagated in the E. coli
strain DH5␣. It was isolated by alkaline lysis and puri-
fied using anion exchange chromatography using the Plas-
mid Maxi Kit (Qiagen Inc., USA) in accordance with the
manufacturer’s instructions and stored in TE buffer (Tris,
10 mM; EDTA, 1 mM), pH 8 at −20 ◦C. Purity of the plas-
mid was determined using 1.0% (w/v) agarose gel elec-
trophoresis and its concentration was determined from the
absorbance at 260 nm according to established protocols
(Sambrook et al., 1989).
2.4. Preparation of radiolabelled (32P) DNA
Plasmid DNA was end-labelled with ␣32P-dCTP (de-
oxy cytosine triphosphate) (Amersham). The linearised

S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
627
NMR, and MALDI-MS, which showed complete substitu-
tion by amine groups at the six-positions.
20–24 h. Prior to transfection the cells were rinsed twice with
serum-free medium and then incubated in medium (unless
otherwise stated) for up to 1 h. In experiments examining the
effect of the lysosomotropic agent, chloroquine, on the level
of gene expression, chloroquine was added 10 min before
transfection to give a final concentration of 100 ␮M.
The compound names have been abbreviated as fol-
lows—AP1: heptakis(6-deoxy-6-pyrid-4-ylamino)-␤-cyclo-
dextrin; AP2: heptakis(2,3-di-O-acetyl-6-deoxy-6-pyrid-
4- ylamino)-␤-cyclodextrin; IM1: heptakis[2,3-di-O-acet-
yl-6-deoxy-6-(1-methyl-1H-imidazol-2-yl)]-␤-cyclodextrin;
IM2: heptakis[6-(1-n-butyl-1H-imidazol-2-yl)-6-deoxy]-␤-
cyclodextrin; IM3: heptakis[2,3-di-O-acetyl-6-(1-n-butyl-
1H-imidazol-2-yl)-6-deoxy]-␤-cyclodextrin; ME1: hepta-
kis(6-deoxy-6-methoxyethylamino)-␤-cyclodextrin; AM1:
heptakis(6-amino-6-deoxy)-␤-cyclodextrin.
The CD.DNA complexes for transfection were formed by
adding an equal volume of CD at the required concentration
in water to 1 ␮g DNA made up in the same diluent. Nitro-
gen/phosphate (N/P) ratios equivalent to mass ratios (MR) of
between 10:1 and 1000:1 CD:DNA were tested to determine
the optimum N/P for transfection with each system. Trans-
fection complexes were left for 30 min before addition to
the cells to give a final concentration of 3 ␮g/ml DNA. After
4 h serum-containing medium was added, diluting the over-
all concentration to 1 ␮g/ml DNA. The complexes were left
in contact with the cells for a further 20 h after which time
they were removed and replaced with fresh serum-containing
medium. After a further 24 h at 37 ◦C the medium was re-
moved and the wells washed with PBS twice. A negative
and positive control of cells transfected with DNA alone and
DOTAP.DNA complex (charge ratio +/− 2.4) respectively
were run in triplicate in each experiment.
2.6. Ethidium bromide exclusion assay
For the ethidium bromide exclusion assay complexes were
formed in water and were then made up to 3 ml with 20 mM
NaCl solution to give a final plasmid DNA concentration
of 1.3 ␮g/ml (Gershon et al., 1993). Ethidium bromide was
added to a concentration of 0.5 ␮g/ml and the resulting flu-
orescence was monitored on a Perkin Elmer LS 50B instru-
ment at λexcitation (526 nm) and λemission (592 nm). The flu-
orescence of DNA alone was taken as 100%.
2.7. Electron microscopy
2.10. Luciferase assay
For EM studies 400-mesh carbon-coated copper grids
were glow-discharged for several minutes and used within
30 min. Complexes were made up at a concentration of
0.5 ␮g/ml in water and a drop placed on the grid for 2 min.
Excess was blotted off, the grid was negatively stained with
a drop of uranyl acetate (2%, w/v) for 1 min and again blot-
ted. The grids were stored at 4 ◦C and viewed the same day
using a Hitachi H7000 transmission electron microscope.
The cells were lysed using lysis buffer (Promega, USA)
and a freeze–thaw cycle. The lysate was assayed us-
ing a luciferase assay kit (Promega, USA) according to
manufacturer’s instructions in a Promega TD 20/20 lumi-
nometer. The relative light units (RLU) were compared
against a calibration curve constructed using standard lu-
ciferase (Promega) and the results were expressed as RLU
of luciferase/mg protein (Bio-Rad protein assay kit).
2.8. Particle size and zeta potential measurements
2.11. Cell association and internalisation of
radiolabelled DNA
Complexes were formed between pCMVluc and CDs over
a range of CD Nitrogen: DNA Phosphate (N/P) ratios. Equal
volumes of DNA and CD in water were mixed to give a
solution of 3 ml with a final DNA concentration of 4 ␮g/ml.
The particle size was measured at 25 ◦C using a Malvern
PCS Series 4700 instrument. The sample was allowed to
equilibrate prior to readings (n = 10) taken with a 100 mW
laser set at an angle of 90◦ to the sample.
For zeta potential measurements, complexes were formed
between pCMVluc and CDs over a range of N/P ratios. Equal
volumes of DNA and CD in 10 mM NaCl were mixed to give
a solution of 3 ml with a final DNA concentration of 2 ␮g/ml.
The zeta potential of the complex was then measured using
a Malvern Zetasizer 3000. Readings (n = 5) were taken at
25 ◦C, at an angle of 90◦.
The radiolabelled DNA was mixed with pCMVluc plas-
mid, so that there were approximately 40,000 CPM/␮g
DNA. The DNA was complexed as described for transfec-
tion with the relevant vector, for 30 min. The cells were
rinsed twice and then incubated with Opti-MEM prior to
studies. The complexes were added to the cells to give a final
concentration of 3 ␮g/ml DNA, and left in contact with the
cells at 37 ◦C (except in the case of experiments specifically
carried out at 4 ◦C) for 24 h. After 4 h serum-containing
medium was added, diluting the overall concentration to
1 ␮g/ml DNA. After this time, the medium was removed and
the cells were rinsed with PBS twice. Lysis buffer was added
to each well and the cells were put through a freeze–thaw
cycle to aid lysis. The cells were collected and centrifuged
at 10,000 rpm for 10 min. The radioactivity of each sample
was measured by adding 200 ␮l of the supernatant to 1 ml
of Emulsifier-SafeTM scintillation cocktail and measuring
the radioactivity (disintegrations/min) on a Tri-Carb 2100
2.9. Transfection protocol
COS-7 cells were seeded at a density of 6×104 cells/cm2
in 24-well plates and grown in full medium (DMEM) for

628
S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
series liquid scintillation counter. The protein content of
each sample was determined using the Bio-Rad Dc protein
assay kit. Data represent the mean of four samples S.E.
At 37 ◦C the radioactivity represents the sum of bound
and internalised plasmid. At 4 ◦C, where internalisation is
halted, the radioactivity represents bound plasmid only.
2.12. Sodium chlorate treatment of cells
Cells were seeded and left to grow for 12 h, sodium chlo-
rate (35 mM) was then added. The cells were incubated with
chlorate for 48 h. Prior to cell association studies it was re-
moved and replaced with fresh serum-free media.
Fig. 2. Effect of nitrogen:phosphate (N/P) ratio for CD.DNA on the
accessibility of DNA to ethidium bromide (expressed as % of maximum
fluorescence) (n = 3, S.D.).
2.13. MTT assay
Ghosh et al., 2000). For example, the inability of ME1-CD
to bind DNA except at much higher N/P than AM-1 is
probably related to the steric effect of the N-methoxyethyl
group. These differences were also seen in the zeta potential
data and transfection vector activities, as discussed below.
COS-7 cells were seeded at a density of 50,000 cells/100 ␮l
per well onto 96-well plates and cultured for 24–48 h in
DMEM medium until they reached confluency. Then 100 ␮l
of the CD.DNA complex in water was added to the wells
and incubated for 4 h at 37 ◦C. The complex was removed,
100 ␮l of fresh medium was added and the cells were incu-
bated overnight. Medium was removed and fresh medium
(100 ␮l) was added to each well along with (20 ␮l) MTT
solution (5 mg/ml in PBS) before incubating the cells for a
further 4 h at 37 ◦C. The reaction product was solubilised
with sodium dodecyl sulfate (100 ␮l, 10%, w/v) in 0.01 N
HCl overnight before quantifying the product using a Dy-
natec plate reader at 590 nm and comparing with control
cells.
3.2. CD.DNA complex characterisation
3.2.1. Electron microscopy
When plasmid DNA was condensed with AP1-CD at the
lower N/P ratios such as 30:1 CD:DNA, toroid-shaped par-
ticles of 200–300 nm were observed (Fig. 3a). The toroidal
morphologies resemble those observed with polylysine
(Wagner et al., 1990, 1991) and have been shown to be
generic at the initial stages of cationic condensation of DNA
(Hud et al., 1995). Also, similarly to other polyamine vec-
tors used in high concentration, larger aggregates of 1 ␮m
approximately (Fig. 3b) were seen at the N/P ratio of 280
used for transfection.
2.14. Statistics
Unpaired Student’s t-test (two-tailed) was used to test the
significance of the difference between two means.
3.2.2. Dynamic light scattering
For dynamic light scattering complexes were prepared
over a range of increasing N/P ratios. Neutral CDs gave no
significant light scattering, in agreement with EM results,
which did not show particle formation. For the charged CDs,
CD.DNA particle sizes increased with increasing N/P ratio.
The mean particle sizes of the complexes used in transfec-
tion and cell viability studies are shown in Table 1. The poly-
dispersity was always greater than 0.1, indicating variation
in particle size.
3. Results and discussion
3.1. CD.DNA complex formation
The ethidium bromide exclusion assay was carried out
to assess complexation of plasmid DNA with the neu-
tral and cationic CDs over a range of nitrogen:phosphate
(N/P) ratios, which are equivalent to nominal charge ratios.
␤-Cyclodextrin produced no significant fluorescence reduc-
tion. Both AM1-CD and AP1-CD caused pronounced re-
duction even at lower N/P (Fig. 2). In contrast, the IM2-CD
produced a gradual decrease in fluorescence, to approxi-
mately 50%, and the ME1-CD had an effect only at the
higher N/P ratios.
The affinity of the CD molecules for DNA is clearly in-
fluenced by the type of substituent on the CD derivative. As
with CD complexes of nucleoside phosphates, differences
are probably due to a combination of polar and steric effects
(Eliseev and Schneider, 1993, 1994; Schwinté et al., 1998;
Table 1
Mean particle size of CD.DNA complexes at the N/P ratios (and equivalent
mass ratios) used for transfection and cell viability experiments
CD
N/P ratio
Mass ratio
Size (nm)
S.D.
AP1-CD
AP2-CD
IM1-CD
IM2-CD
IM3-CD
AM1-CD
ME1-CD
280
204
215
248
189
207
1440
100:1
100:1
100:1
100:1
100:1
100:1
1000:1
114.3
112.7
110.7
117.2
131.2
140.0
942.8
17.0
11.3
8.0
5.5
10.7
12.0
126.5

S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
629
neutralising the charge on DNA, but some produce higher
positive charges at the highest N/P ratios.
The relatively high N/P ratios CD:DNA (nominal charge
ratios) required to produce charge inversion imply that only
a fraction of the amino groups on each CD molecule is ef-
fective in binding. Potentiometric titrations carried out on
polyamino CDs (Hamelin et al., 1995; Schwinté et al., 1998)
have shown that their pKa values extend over a wide range,
about 6–9. Since potentiometry has also shown that the as-
sociation constants between protonated polyamino CDs and
polyanions are very high, these high N/P ratios must be
due to incomplete protonation and steric effects. Studies on
polyamine-mediated DNA condensation have shown the in-
fluence of interamine distances and substituents on pKas and
biological activities (Geall et al., 1998). The molecular pre-
sentation of the positive charges to the DNA appears to be
inefficient in these polyamino CDs. Their clustering on the
CD rim gives an optimum N/P ratio for transfection of about
200 (although this transfection effect is only 10–30 times
that at N/P 100) in contrast to PEI (N/P 10) (see below).
We have also used polycationic amphiphilic CDs for trans-
fection, in the next logical development of CDs as vectors
(Cryan, 2001; Cryan et al., 2004) and these are effective
with typical N/P ratios of 7 since they probably form more
ordered liquid crystalline complexes with DNA, as do small
cationic lipids (Chesnoy and Huang, 2000).
3.3. Transfection in COS-7 cells
The neutral and cationic CDs were tested for transfec-
tion over a range of N/P ratios. The optimal conditions
for transfection were: complexation of the CD vector and
plasmid DNA in water at the N/P ratios shown in Table 1
and addition to the cells at a concentration of 3 ␮g/ml for
24 h, followed by 24 h expression time. Complexes formed
in Hanks balanced salt solution (HBSS), instead of water,
did not enhance transfection above that of the neutral CDs.
Ghosh et al. (2000) have reported that the composition and
concentration of a buffer system influence the association
constant between positively charged CDs and phosphate
esters. The influence of buffer on transfection efficiency is
therefore probably due to effects on association between
the positively charged CDs and the DNA.
Fig. 3. Electron micrographs of AP1-CD.DNA complexes at N/P ratio
(a) 30:1 and (b) 280:1.
3.2.3. Zeta potential
The charge on vector-DNA complexes has been shown to
have an important influence on transfection, as a surplus of
positive charge is needed for electrostatic interaction with
the cell surface. Zeta potentials of complexes were measured
in 10 mM NaCl at different N/P ratios. In contrast to the
neutral CDs, as the quantity of cationic CD was increased
relative to that of DNA, zeta potentials changed from nega-
tive to positive (Fig. 4). All the cationic CDs are capable of
Under optimum conditions, the relative order of transfec-
tion enhancement was AM1-CD > AP1-CD > IM2-CD >
AP2-CD
> ME1-CD > DM-CD (dimethyl-CD) >
IM3-CD > ␤-CD. The CDs produced up to a 4000-fold
increase in luciferase expression relative to uncomplexed
DNA (Fig. 5). This is 100-fold less than that produced
by DOTAP. However the transfection levels attained with
AP1-CD.DNA and with DOTAP.DNA (1.0E+03 RLU/mg
protein) were equal in the presence of serum (Fig. 6).
Chloroquine, a known endosomolytic agent, increased
transfection levels for the cationic CD.DNA complexes by
10–400-fold, with the greatest increases for the alkylimida-
zole (IM-CD) derivatives (Fig. 5).
Fig. 4. Effect of N/P ratio on the zeta potential of CD.DNA complexes
(n = 3, S.D.).

630
S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
Fig. 5. Transfection (RLU/mg protein) (black columns) of COS-7 cells with CD.DNA complexes formed at optimum N/P ratios (as shown in Table 1)
in water; and effect of chloroquine (100 ␮M) (white columns). Control represents transfection with DNA alone (n = 4, S.E.). ∗P < 0.05, ∗∗P < 0.01.
As with their complexation of DNA, transfection efficien-
cies of various polycationic CDs can be expected to vary in
view of the differences in cationic group structures, which
will also affect interaction with the cell. Borrajo et al. (1997)
has shown that per(6-amino-6-deoxy)-CDs bound less to
cell surfaces on 2,3-methylation. When AP- and IM-CDs
are acetylated at these positions, the resulting AP2-, IM1-
and IM3-CDs are considerably less efficient vectors than the
unacetylated compounds.
Neutral CDs, despite their inability to complex and neu-
tralise plasmid DNA, did produce a significant (180-fold
for DM-CD) increase in transfection levels above control.
␤-CD, HP-CD and DM-CD have been found to have a ca-
pacity for removal of cholesterol from various cell types,
including fibroblasts (such as the COS-7 cells), resulting in
membrane instability (Irie et al., 1982). Such CDs have been
shown to increase both transcellular and paracellular traf-
ficking of macromolecules and to increase cell association
and uptake of oligonucleotides (Zhao et al., 1995). ␤-CD has
been used for in vivo transfection studies in rat lung, where
it increased the level of reporter gene expression (Freeman
and Niven, 1996).
3.4. Cell binding and internalisation
To further elucidate the mechanism of transfection of the
CD.DNA complexes, the degree of complex binding and in-
ternalisation by COS-7 cells were compared at 4 and 37 ◦C.
The CDs that produced the highest levels of cell-associated
DNA, namely AM1-CD and AP1-CD, were also those most
influenced by the change in incubation temperature (Fig. 7).
There was a significant difference between the total levels
of cell-associated (bound and internalised) plasmid DNA (at
37 ◦C) and the amount of plasmid DNA that was simply
bound to the COS-7 cells (at 4 ◦C) for AP1-CD.DNA and
AM1-CD.DNA complexes. The results imply that these two
CDs actively internalised a high proportion of the DNA with
20% of the total added per well being internalised by the
COS-7 cells when complexed with AM1-CD. This is sig-
nificantly greater than the percentage DNA, approximately
5%, actively internalised using DOTAP.
3.5. Chlorate cell-treatment and role of sulfated
proteoglycans
A net positive charge on vector-DNA complexes gener-
ally leads to non-specific endocytosis caused by attraction
between cationic complexes and negative charges on the
extracellular surface. A number of groups have established
the role of sulfated proteoglycans in non-viral gene delivery
(Ruponen et al., 1999; Belting and Petersson, 1999) includ-
ing a role in the attachment of complexes to the cell surface
and internalisation (Hyvönen et al., 2002). There is evidence
for endocytosis based on the ability of cationic complexes
to bind to sulfated cell surface proteoglycans (Mislick and
Baldeschwieler, 1996). Pretreatment of the cells with sodium
chlorate, to inhibit the sulfation of glycosaminoglycans, had
a pronounced effect on the levels of cell-associated DNA
Fig. 6. Transfection (RLU/mg protein) of COS-7 cells in the presence of
serum with DOTAP.DNA (charge ratio +/− 2.4) and selected CD.DNA
complexes at their previously determined optimum N/P ratio for transfec-
tion: AP1-CD.DNA, 280:1; AP2-CD:DNA 204:1; IM2-CD.DNA, 248:1.
Control represents transfection with DNA alone (n = 4, S.E.).

S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
631
Fig. 7. Effect of incubation at 37 ◦C (black columns) and 4 ◦C (white columns) on the levels of DNA (dpm/mg protein) associated with COS-7 cells
after incubation for 24 h with CD.DNA complexes (at N/P ratios shown in Table 1) and DOTAP charge ratio +/− 2.4. Controls of DNA alone are also
shown (n = 4, S.E.). ∗P < 0.05, ∗∗P < 0.01.
mediated by AP1-CD, IM3-CD and DM-CD. The great-
est effect was with AP1-CD, resulting in a 35-fold reduc-
tion, while IM3-CD and DM-CD were less affected, with a
three-fold reduction. The effect of chlorate pretreatment on
AP1-CD performance is compared to the effect on commer-
cial vectors in Fig. 8. The levels of cell-associated DNA for
DOTAP, PEI and Superfect were reduced only by factors of
1.5, 4 and 5, respectively. The more successful CD vectors
therefore appear to exploit the negatively charged proteogly-
cans as a means of binding to cells and internalising DNA.
Borrajo et al. (1997) showed for COS-7 cells that polyamino
CD derivatives bound to glycosaminoglycan (GAG) sulfates.
Modification of the secondary face of the aminated CD sig-
nificantly decreased the binding efficiency to GAGs and this
may be a factor in the decreased transfection efficiency of
the acetylated polyamino CDs, AP2-, IM1- and IM3-CD,
compared with the corresponding unacetylated compounds.
Although AP1-CD and AM1-CD promote cell internal-
isation of DNA better than the liposomal vector DOTAP
in serum-free medium, they fail to improve on its transfec-
tion ability indicating that they are not as efficient at other
points in the gene delivery pathway. Increase in transfec-
tion in the presence of chloroquine implies that the com-
plexes are endocytosed and that release from the endosome
may be rate-limiting. In contrast, the lack of any effect by
chloroquine on the transfection efficiency of the neutral CD,
DM-CD, supports the theory that the neutral CDs do not
lead to internalisation of plasmid DNA via endocytosis but
rather via membrane fluidisation. The alkylimidazole and
pyridylamino groups when incorporated into cationic lipids
have been shown to increase transfection via enhanced endo-
somolysis, by acting as pH-sensitive groups (Budker et al.,
1996), however such an influence is not evident here.
3.6. Toxicity
The MTT assay indicated that the CD.DNA com-
plexes display some toxicity at the concentrations used for
Fig. 8. Effect of sodium chlorate pretreatment (white columns) on association of DNA (black columns) (dpm/mg protein) with COS-7 cells mediated by
AP1-CD (N/P 280) and commercial vectors (DOTAP charge ratio +/− 2.4, PEI N/P 10, Superfect N/P 3.3) (n = 4, S.E.).

632
S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
Belting, M., Petersson, P., 1999. Protective role for proteoglycans against
cationic lipid cytotoxicity allowing optimal transfection efficiency in
vitro. Biochem. J. 342, 281–286.
Borrajo, A.M.P., Gorin, B.I., Dostaler, S.M., Riopelle, R.J., Thatcher,
G.R.J., 1997. Derivatised cyclodextrins as peptidomimetics: influence
on neurite growth. Bioorg. Med. Chem. Lett. 7, 1185–1190.
Breslow, R., Czarniecki, M.F., Emert, J., Hamaguchi, H., 1980. Improved
acylation rates within cyclodextrin complexes from flexible capping
of the cyclodextrin and from adjustment of the substrate geometry. J.
Am. Chem. Soc. 102, 762–770.
Budker, V., Gurevich, V., Hagstrom, J.E., Bortzov, F., Wolff, J.A., 1996.
pH-Sensitive, cationic liposomes: a new synthetic virus-like vector.
Nature Biotech. 14, 760–764.
Chesnoy, S., Huang, L., 2000. Structure and function of lipid-DNA com-
plexes for gene delivery. Ann. Rev. Biophys. Biomol. Struct. 29, 27–
47.
Fig. 9. MTT assay (dehydrogenase activity, %) for AP1-CD.DNA and
AM1-CD.DNA complexes over a range of N/P ratios in COS-7 cells
(n = 3, S.D.).
Chmurski, K., Defaye, J., 1999. An improved synthesis of per(6-
deoxyhalo) cyclodextrins using N-halosuccinimide-triphenylphosphine
in dimethylformamide. In: Torres Labandeira, J.J., Vila-Jato, J.L.
(Eds.), Proceedings of the Ninth International Symposium on Cy-
clodextrins. Kluwer Academic Publishers, Dordrecht, pp. 183–186.
Croyle, M.A., Roessler, B.J., Hsu, C.P., Sun, R., Amidon, G.L., 1998.
␤-Cyclodextrins enhance adenoviral-mediated gene delivery to the
intestine. Pharm. Res. 15, 1348–1355.
transfection. The toxicity of the complexes increased with
increasing concentrations of either the neutral (DM-CD) or
cationic CDs tested. The effects of increasing N/P ratio on
the percentage dehydrogenase activity for the best of the
cationic CD vectors, AP1-CD and AM1-CD, are shown in
Fig. 9.
Cryan, S.-A., 2001. Development of novel cyclodextrins as non-viral gene
delivery vectors and assessment of differentiated cell culture models to
predict in vivo gene expression. Ph.D. Thesis. Trinity College, Dublin.
Cryan, S.-A., O’Driscoll, C.M., 2003. Mechanistic studies on nonviral
gene delivery to the intestine using in vitro differentiated cell culture
models and an in vivo rat intestinal loop. Pharm. Res. 20 (4), 569–
575.
Cryan, S.-A., Donohue, R., Ravoo, B.J., Darcy, R., O’Driscoll, C.M.,
2004. Cationic cyclodextrin amphiphiles as gene delivery vectors. J.
Drug. Del. Sci. Tech., in press.
Eliseev, A.V., Schneider, H.-J., 1993. Amino-cyclodextrins as selective
hosts with several binding sites for nucleotides. Angew. Chem. Int.
Edn. Engl. 32, 1331–1333.
Eliseev, A.V., Schneider, H.J., 1994. Molecular recognition of nucleotides,
nucleosides and sugars by aminocyclodextrins. J. Am. Chem. Soc.
116, 6081–6088.
4. Conclusions
Polycationic CDs successfully bound, neutralised and
formed nanoparticulate complexes with plasmid DNA. Of
the cationic CDs, AP1-CD produced a 4000-fold increase
in transfection over DNA alone and in serum equalled that
of DOTAP.DNA complexes. The best DNA-uptake me-
diators, AP1-CD and AM1-CD, are highly dependent on
proteoglycan-mediated binding and active uptake processes.
Combinations of cyclodextrins with polyamino struc-
tures have been made before where polyamino polymers
and dendrimers have incorporated CDs (Gonzalez et al.,
1999; Hwang et al., 2001; Arima et al., 2001). The first
of these studies showed that the incorporation of the CDs
reduced toxicity by a factor of 2–3 relative to the CD-free
polyamides, however there was no significant effect on
transfection ability. It appears therefore that in those cases
any advantages to be obtained from individual CD molecules
were not accessed. In contrast, the overall results here show
that cyclodextrins can function as promising molecular
templates with potential for development as a new generic
series of gene delivery vectors.
Freeman, D.J., Niven, R.W., 1996. The influence of sodium glycocholate
and other additives on the in vivo transfection of plasmid DNA in the
lungs. Pharm. Res. 13, 202–209.
Geall, A.J., Taylor, R.J., Earll, M.E., Eaton, M.A.W., Blagbrough, I.S.,
1998. Synthesis of cholesterol–polyamine carbamates: pKa studies and
condensation of calf thymus DNA. Chem. Commun., 1403–1404.
Gershon, H., Ghirlando, R., Guttman, S.B., Minsky, A., 1993. Mode of
formation and structural features of DNA-cationic liposome complexes
used for transfection. Biochemistry 32 (28), 7143–7151.
Ghosh, M., Zhang, R., Lawler, R., Seto, C., 2000. The effects of buffers
on the thermodynamics and kinetics of binding between positively-
charged cyclodextrins and phosphate ester guests. J. Org. Chem. 65,
735–741.
Gonzalez, H., Hwang, S.J., Davis, M.E., 1999. New class of polymers
for the delivery of macromolecular therapeutics. Bioconjugate Chem.
10, 1068–1074.
Acknowledgements
The authors wish to thank Professor Colin Pouton for
assistance in establishing transfection techniques.
Hamelin, B., Jullien, L., Guillo, F., Lehn, J.-M., Jardy, A., De Robertis,
L., Driguez, H., 1995. Formation of highly stable heterodimers in
aqueous solution between ␤-cyclodextrin derivatives bearing multiple
opposite charges. J. Phys. Chem. 99, 17877–17885.
Hoffman, J.L., Bock, R.M., 1970. The interaction of cyclodextrins with
nucleic acids. A study of secondary structure in three transfer ribonu-
cleic acids. Biochemistry 9, 3542–3550.
References
Arima, H., Kihara, F., Hirayama, F., Uekema, K., 2001. Enhancement of
gene expression by polyamidoamine dendrimer conjugates with ␣-,
␤-, and ␥-cyclodextrins. Bioconjugate Chem. 12, 476–484.
Hud, N.V., Dowing, K.H., Balhorn, R., 1995. A constant radius of cur-
vature model for the organization of DNA in toroidal condensates.
Proc. Natl. Acad. Sci. U.S.A. 92, 3581–3585.

S.-A. Cryan et al. / European Journal of Pharmaceutical Sciences 21 (2004) 625–633
633
Hwang, S.J., Bellocq, N.C., Davis, M.E., 2001. Effects of structure of
␤-cyclodextrin-containing polymers on gene delivery. Bioconjugate
Chem. 12, 280–290.
Hyvönen, Z., Ruponen, M., Rönkkö, S., Suhonen, P., Urtti, A., 2002.
Extracellular and intracellular factors influencing gene transfection
mediated by 1,4-dihydropyridine amphiphiles. Eur. J. Pharm. Sci. 15
(5), 449–460.
glycans: physicochemical and transfection studies. Biochim. Biophys.
Acta 1415, 331–341.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning—A
Laboratory Manual. Cold Spring Harbor Press, New York.
Schwinté, P., Darcy, R., O’Keeffe, F.J., 1998. Ditopic binding of nu-
cleotides by heptakis(6-hydroxyethylamino-6-deoxy)-␤-cyclodextrin.
J. Chem. Soc., Perkin Trans. 2, 805–808.
Irie, T., Otagiri, M., Sunada, M., Uekema, K., Ohtani, Y., Yamada,
Y., Sugiyama, Y., 1982. Cyclodextrin-induced hemolysis and shape
changes of human erythrocytes in vitro. J. Pharmaco-Biodynam. 5,
741–744.
Kilsdonk, E.P.C., Yancey, P.G., Stoudt, G.W., Bangeerter, F.W., Johnson,
W.J., Phillips, M.C., Rothblat, G.H., 1995. Cellular cholesterol efflux
mediated by cyclodextrins. J. Biol. Chem. 270, 17250–17256.
Luo, D., Saltzman, W.M., 2000. Synthetic DNA delivery systems. Nature
Biotech. 18, 33–37.
Szejtli, J., Osa, T. (Eds.), 1996. Cyclodextrins. Comprehensive Supramole-
cular Chemistry, vol. 3. Pergamon Press, UK.
Uduehi, A.N., Moss, S.H., Nuttall, J., Pouton, C.W., 1999. Cationic
lipid-mediated transfection of differentiated CaCo-2 cells: a filter cul-
ture model of gene delivery to a polarized epithelium. Pharm. Res.
16 (12), 1805–1811.
Wagner, E., Zenke, M., Cotten, M., Beug, H., Birnstiel, M.L., 1990.
Transferrin-polycation conjugates as carriers for DNA uptake. Proc.
Natl. Acad. Sci. U.S.A. 87, 3410–3414.
Mislick, K.A., Baldeschwieler, J.D., 1996. Evidence for the role of pro-
teoglycans in cation-mediated gene transfer. Proc. Natl. Acad. Sci.
U.S.A. 93, 12349–12354.
Ravoo, B.J., Darcy, R., 2000. Cyclodextrin bilayer vesicles. Angew. Chem.
Int. Edn. Engl. 39, 4324–4326.
Ravoo, B.J., Darcy, R., Mazzaglia, A., Nolan, D., Gaffney, K., 2001.
Supramolecular tapes formed by a catanionic cyclodextrin in water.
Chem. Commun., 827–828.
Ruponen, M., Yla-Herttuala, S., Urtti, A., 1999. Interactions of polymeric
and liposomal gene delivery systems with extracellular glycosamino-
Wagner, E., Cotten, M., Foisner, R., Birnstiel, M.L., 1991. Transferrin-
polycation-DNA complexes: the effect of polycations on the structure
of the complex and DNA delivery to cells. Proc. Natl. Acad. Sci.
U.S.A. 88, 4255–4259.
Wenz, W., 1994. Cyclodextrins as building blocks for supramolecular
structures and functional units. Angew. Chem. Int. Edn. Engl. 33,
803–822.
Zhao, Q., Temsamani, J., Agrawal, S., 1995. Use of cyclodextrin and
its derivatives as carriers for oligonucleotide delivery. Antisense Res.
Dev. 5, 185–192.