Preparation and Characterization of Chitosan and Trimethyl chitosan modified Poly (e-caprolactone) Nanoparticles as DNA Carriers

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
Preparation and Characterization of Chitosan and Trimethyl-chitosan-
modified Poly-(e-caprolactone) Nanoparticles as DNA Carriers
Submitted: July 16, 2004; Accepted: October 28, 2004; Published: August 10, 2005
Jochen Haas,1 M. N. V Ravi Kumar,2 Gerrit Borchard,3 Udo Bakowsky,4 and Claus-Michael Lehr1
1Department of Biopharmaceutics and Pharmaceutical Technology, Saarland University, 66123 Saarbr€ucken, Germany
2Department of Pharmaceutics, National Institute of Pharmaceutical Education & Research, SAS Nagar-160 062,
Punjab, India
3Division of Pharmaceutical Technology, LACDR, Leiden University, 2300 RA Leiden, The Netherlands
4Department of Pharmaceutical Technology and Biopharmacy, Philipps University of Marburg, 35032 Marburg, Germany
ABSTRACT
immune response, can be ÔÔtailored to measureÕÕ by chem-
ical and immunological modifications and can be easily
The purpose of this research was to prepare poly-(e-capro-
produced in large quantities. Because of these advantages
lactone) (PCL) particles by an emulsion-diffusion-evapora-
they are enjoying an increasing interest in gene transfer
tion method using a blend of poly-(vinyl alcohol) and chi-
research. Among the strategies that have been employed
tosan derivatives as stabilizers. The chitosan derivatives
to condense DNA can be found dendrimers,2,3 cationic
used were chitosan hydrochloride and trimethyl chitosans
peptides,4,5 cationic polymers,6-9 cationic lipids,10-12 as
(TMC) with varying degrees of quaternization. Particle
well as liposomes.13-18 Even if nonviral gene transfer
characteristics—size, zeta potential, surface morphology,
systems have been much improved during the past dec-
cytotoxicity, and transfection efficiency—were investigated.
ade, they still cannot compete with viral vectors in terms
The developed method yields PCL nanoparticles in the size
of transfection efficiency, so research activities continue
range of 250 to 300 nm with a positive surface charge (2.5
in order to unite the advantages of both approaches.
to 6.8 mV). The cytotoxicity was found to be moderate and
DNA could be transferred to the cells either by encapsu-
virtually independent of the stabilizersÕ concentration with
lating in or by surrounding the particles11,19—21 or by
the exception of the highly quaternized TMC (degree of
having the DNA complexed on the outer surface of par-
substitution 66%) being significantly more toxic. In immo-
ticles.22,23 However, one cannot ignore the possibility of
bilization experiments with gel electrophoresis, it could be
damaging the active substance during the encapsulation
shown that these cationic nanoparticles (NP) form stable
process; above all there is no immediate availability of
complexes with DNA at a NP:DNA ratio of 3:1. These
the DNA for the cells.24 Investigations were performed
nanoplexes showed a significantly higher transfection effi-
to encapsulate DNA-cation-complexes into microspheres,
ciency on COS-1 cells than naked DNA.
but the preparation is complicated and yields microscaled
particles.25 Successful transfection in vitro and in vivo
K
has been achieved with positively charged silica nanopar-
EYWORDS: biocompatible, cytotoxicity, gene transfer,
nanoparticles, trimethyl chitosan, AFM
ticles;21,26 however, the fate of the inorganic carrier sub-
stance silica may remain a problem. In order to address
this issue it was decided to develop nanoparticulate DNA
INTRODUCTION
carriers based on biodegradable polymers.22,23 Poly-
(e-caprolactone) (PCL) was chosen because of its bio-
Nonviral gene transfer systems have become increasingly
compatibility, lipophilicity (to support passive uptake
popular as an alternative to viral vectors.1 While the lat-
processes), and cost-effectiveness compared with other
ter ones offer a high transfection efficacy owing to
polyesters such as poly-(lactic-co-glycolic acid) (PLGA).
highly specialized intrusion mechanisms, they also suffer
from a certain risk of backwards mutation toward the
Besides the common parameters for particle characteriza-
infectious wild type. In contrast to this, artificial trans-
tion (ie, size, zeta potential, and microscopy), the biocom-
fection agents are considered a much safer alternative;
patibility of the carriers was investigated by cytotoxicity
they can be administered repeatedly with a low risk for
studies, since some nonviral transfection agents (notably
short linear poly-[ethylene imine], PEI) are characterized
by a cytotoxicity, which limits their use.27,28 Since nanopar-
ticles were prepared using amphiphilic molecules or surfac-
Corresponding Author: Claus-Michael Lehr, Department
tants as excipients and stabilizers, possible effects on the
of Biopharmaceutics and Pharmaceutical Technology,
cell membrane are an important parameter to investigate.
Saarland University, Building 8.1, 66123 Saarbr€
ucken,
Germany; Tel: 149-681-302-3039; Fax: 149-681-302-
Finally, the ability of the nanoparticles to complex DNA
4677; E-mail: lehr@mx.uni-saarland.de
and transfect cells was tested in COS-1 cells, since this
E22

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
parameter can be regarded as the pivotal point for their
rotary evaporator at 50 mbar at 40°C (water bath) for
suitability as transfection agents.
20 minutes. The finished nanoparticle suspension was
stored at 4°C for further use. To determine the concentra-
tion of nanoparticles (weight per volume), 2 mL of this
MATERIALS AND METHODS
suspension was spun down in Eppendorf tubes at 20 000g
Materials
for 30 minutes; the supernatant was removed and the pellet
was allowed to dry under a nitrogen stream before weighing.
PCL with an average molecular weight of 10 000 Da was
purchased from Fluka (Deisenhofen, Germany). The poly-
(vinyl alcohol) (PVA) used was Mowiol 4-88 (88% hydrol-
Determination of Particle Size and Zeta Potential
ysis) from Hoechst (Frankfurt/Main, Germany). Seacure
Nanoparticles and NP-DNA complexes were analyzed for
CL 210 chitosan hydrochloride (CS, 83% deacetylated)
their size and zeta potential with a Malvern Zetasizer 3000
was obtained from Pronova Biopolymer (Drammen, Nor-
photon correlation spectroscopy (PCS) system (Malvern
way). Trimethyl chitosan (TMC) with varied degrees of
Instruments, Southborough, UK). The nanoparticle and
substitution (4%, 10%, 18%, 66%) were synthesized at
NP-DNA suspensions were diluted 100-fold with a 20-mM
Leiden University, The Netherlands. The COS-1 cells and
N-(2-hydroxyethyl)piperazine-N0-(2-ethane sulfonic acid)
the beta-galactosidase expression plasmid pCMVb were
(HEPES) buffer to minimize particle interactions and multi-
obtained from ATCC (Manassas, VA). We further prepar-
ple scattering and were adjusted to pH 7.4 to maintain a con-
ed the plasmid; it was transformed into E. coli DH5a and a
stant dispersion medium. For particle size determination,
Gigaprep from 2500 mL of an overnight culture was per-
3 mL diluted suspension was placed in a disposable poly-
formed according to instructions provided by the supplier
styrene cuvette and measured at room temperature using a
(Qiagen, Hilden, Germany). The lactate dehydrogenase
scattering angle of 90°. For zeta potential determination,
(LDH), detection Kit, was obtained from Roche Applied
samples were measured in a fixed-glass cell; and the instru-
Sciences (Penzberg, Germany, catalog no. 1644793) as was
ment was calibrated with 250 mV latex standard. Samples
the WST-1 cell proliferation reagent (catalog no. 1644807).
were analyzed using Malvern PCS software. The signal
Cell culture media were obtained from Promocell
intensity mean was used to calculate the mean particle
(Heidelberg, Germany) and Sigma-Aldrich (Taufkirchen,
diameter, and all measurements were performed in dupli-
Germany). All other chemicals were purchased from
cate. DNA-modified particles were prepared by mixing
Sigma-Aldrich. All solvents were of high-performance
equal parts of a particle suspension of 60 mg/mL with a
liquid chromatography (HPLC) grade, and reverse osmosis
20 mg/mL DNA solution and incubating for 15 minutes
treated water (Millipore, Schwalbach, Germany) was used.
minimum, thus keeping the 3:1 NP:DNA ratio described
below. The product was diluted 1:100 with 20 mM HEPES
buffer as described above prior to PCS measurements.
Particle Preparation
Arithmetic mean and SD were calculated from 3 consecu-
Particles were prepared by an emulsion-diffusion-evapora-
tive runs, and samples were analyzed using Malvern PCS
tion method.22 A 100-mg amount of PCL was dissolved in
software.
10 mL organic phase consisting of 9 mL ethyl acetate (EA)
and 1 mL acetone (AC) for 1 hour under mild heating at
30°C (water bath) in a closed container. As for the aqueous
Gel Electrophoresis and Determination of Unbound DNA
phase, 100 mg PVA and 15 mg CS or TMC were stirred in
NP-DNA complexes were prepared by mixing 25 mL of a
5 mL water for 2 hours at room temperature until a clear
serial dilution of NP in water with 25 mL DNA in 50 mM
solution was obtained. The organic phase was passed
HEPES buffer. Plasmid DNA was kept at a constant con-
through a 0.22-mm syringe filter to remove any undis-
centration of 10 mg/mL throughout the experiment; the
solved solids and subsequently added drop wise to the
particle suspension varied from 100 mg/mL to 0 mg/mL.
aqueous phase under constant stirring. The resulting micro-
The particles were allowed to incubate and complex DNA
emulsion was kept under constant agitation on magnetic
for 15 minutes at room temperature. Ten microliters of this
stirrer at 1000 rpm for 1 hour and was subsequently homo-
suspension was added to 2 mL of a loading buffer contain-
genized with an Ultra-Turrax homogenizer (Janke and
ing coomassie blue dye for monitoring. Ten microliters of
Kunkel, Staufen, Germany) at 13 500 rpm for 10 minutes.
this blend were electrophoresed in a 1% agarose gel,
This colloidal preparation was diluted to a volume of
stained with 1% ethidium bromide for better visualization;
50 mL by adding water drop wise under stirring conditions
the electrophoresis chamber (Bio-Rad, Munich, Germany)
(1000 rpm, magnetic stirrer), which resulted in nanopreci-
was set to 90 minutes and 90 V (ie, 10 V/cm). A control
pitation. In order to remove the organic solvent and to
was done by electrophoresing a DNA ÔÔladderÕÕ mix with
harden the nanospheres, the suspension was treated with a
DNA fragments of different molecular masses. Data
E23

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
analysis was done by acquiring the images on a Geldoc
turerÕs instructions) was added, and absorbance was read at
2000 gel documentation system with a UV transilluminator
492 nm immediately after addition of reagent and after
(Bio-Rad, Munich, Germany) and analyzing the images
60 minutes of incubation at room temperature under light
with Molecular Analyst 1.1 software (Bio-Rad).
protection. Cells were washed again with PBS and 100 mL
fresh medium was added; then cells were kept in an incuba-
tor for another 48 hours to simulate the conditions in a
Nanoparticle Characterization by Atomic Force
transfection study. After this period, 10 mL of WST-1 pro-
Microscopy
liferation agent was added and samples were analyzed using
The nanoparticle formulations were prepared as described
a UV/Visible photometric plate reader at 450 nm. Data
above and diluted in demineralized water. After 1 hour, the
were corrected for blank samples (no cells) and expressed
particles or particle-DNA complexes were directly trans-
in percentage of survival compared with a positive control.
ferred onto a silicon chip by dipping into the nanoplex
To dissect the influence of the used excipients from a whole
solution. Atomic force microscopy was performed on a
particle suspension, cytotoxicity studies were also performed
Digital Nanoscope IV Bioscope (Veeco Instruments, Santa
with PVA and the various chitosan derivatives. A stock solu-
Barbara, CA) as described elsewhere.21 The microscope
tion of PVA was prepared starting with the concentration
was vibration damped. Commercial pyramidal Si3N4 tips
present in the native particle suspension by dissolving PVA
(NCH-W, Veeco Instruments) on a cantilever with a length
in water (1.43 mg/mL) and mixing with double-concentrated
of 125 mm, a resonance frequency of ~220 kHz, and a
nominal force constant of 36N/m were used. All measure-
DMEM. A serial dilution of this was prepared in DMEM
ments were performed in tapping mode to avoid damage of
and added to the cell monolayers. Further treatment was
the sample surface. The scan speed was proportional to the
performed according to the procedures mentioned above.
scan size, and the scan frequency was between 0.5 and
Chitosan and trimethyl chitosan derivatives were equally
1.5 Hz. Images were obtained by displaying the amplitude
investigated, starting with stock solutions of 428 mg/mL.
signal of the cantilever in the trace direction, and the height
Both excipients were applied in the same concentrations as
signal in the retrace direction, both signals being simulta-
in corresponding particle preparations.
neously recorded. The results were visualized either in
height or in amplitude mode.
Transfection Studies
To examine the actual suitability of the prepared nanopar-
Cytotoxicity Studies
ticles for gene transfer, complexes of particles and the plas-
COS-1 cells were cultured in DulbeccoÕs modified eagleÕs
mid DNA were prepared and added to cell cultures. Cos-1
medium (DMEM, ready-made liquid media from Promo-
cells were cultured as described above and seeded at a den-
cell, Heidelberg, Germany) with 4.5 g/L glucose, supple-
sity of 10 000 cells per well on 96-well plates, and cells were
mented with 1% nonessential amino acids (MEM-NEAA);
used for transfection after 24 hours of settling and adhesion.
pH was adjusted to 7.4. Subculture was performed by
The DNA concentration was kept constant at 10 mg/mL;
detaching the cells with trypsine-EDTA solution and split-
particles were used in 2 concentrations deemed optimal
ting them at a 1:10 ratio. Double concentrated DMEM was
from binding and electrophoresis experiments (10 and
prepared by using powdered media (Sigma-Aldrich) dis-
30 mg/mL, respectively). Complexes of nanoparticles and
solved in half the required amount of water.
plasmid DNA were prepared by adding freshly prepared
For cytotoxicity studies, COS-1 cells were seeded onto 96-
particle suspension to 23 concentrated DMEM to obtain
well plates with a density of 10 000 cells per well and
the desired concentration. This suspension was mixed with
allowed to adhere and grow for 24 hours. Serial dilutions of
an equal volume of a solution of DNA in 50 mM HEPES
the nanoparticle suspensions were prepared in DMEM in a
buffer and complexes were allowed to incubate and form
concentration range from 2000 mg/mL to 31.25 mg/mL; the
for 15 minutes. Cells were washed with warm PBS prior to
first dilution was prepared by mixing the original suspen-
addition of the transfection agent. A solution of DNA in
sion (4 mg/mL) with double-concentrated DMEM. Cells
medium was used as a control. The transfection agent was
were washed with phosphate-buffered saline (PBS) prior to
allowed to incubate for 4 hours at 37°C. After incubation,
addition of 100 mL of the nanoparticle suspensions. Nega-
the cells were washed once with PBS and fresh medium was
tive control was performed by incubating cells with
added. Subsequently cells were kept for another 48 hours in
medium only, positive control by 0.1% Triton X-100 in
an incubator to allow gene expression. Controls were per-
medium. The samples were allowed to incubate for 4 hours
formed by incubating cells with either naked DNA or DNA
at 37°C at 5% CO
with a commercial transfection agent, PolyFect.
2. After incubation, 50 mL of the superna-
tant was removed for determination of LDH release, 50 mL
To assay the cells content of b-galactosidase, cells were
of LDH reagent mixture (prepared according to manufac-
lysed with a solution of 0.1% Triton X-100 in PBS
E24

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
Table 1. Photon Correlation Spectroscopy Analysis of Trimethyl Chitosan-modified Nanoparticles*
Size [nm]
Polydisp
Size [nm]
Polydisp
f pot [mV]
f pot [mV]
Batch
Without DNA
Index
With DNA
Index
Without DNA
With DNA
Chitosan
280.9
0.10
298.8
0.07
1 2.6 6 0.5
26.8 6 0.4
TMC, 4%
302.5
0.13
514.6
0.52
1 2.5 6 1.2
n/d
TMC, 10%
279.0
0.09
293.2
0.08
1 4.0 6 0.2
26.8 6 0.5
TMC, 18%
261.5
0.03
352.1
0.15
1 6.0 6 0.9
24.5 6 0.3
TMC, 66%
245.0
0.04
268.6
0.10
1 6.8 6 1.2
25.5 6 0.6
*Polydisp index indicates polydispersity index; z pot, zeta potential; TMC, trimethyl chitosan; n/d, not determined. Size is given as diameter in nm;
zeta potential is noted as mean 6 SD.
(100 mL/well) for 30 minutes at 4°C. Fifty microliters of
TMC formulations with higher degrees of substitution
this cell lysate was mixed with an equal amount of reaction
yielded slightly smaller particles with a lower polydisper-
mixture, prepared from 5 mM 4-methyl-umbelliferyl-b-D-
sity. The addition of DNA increases the measured diameter
galactoside, 100 mM D-galactose and 2 mM MgCl2 in
of the particle/DNA nanoplexes by 20 to 200 nm.
PBS. After an incubation period of 4 hours at 37°C, the flu-
The zeta potential of freshly prepared particles was always
orescence was determined with a Cytofluor II microplate
found to be in a slightly positive range, between +2.5 and
fluorescence reader at lex 5 360 nm and lem 5 460 nm
+6.8 mV. After incubation with DNA, the zeta potential
(PE Biosystems, Weiterstadt, Germany).
dropped to negative values in the range of À4.5 to À6.8 mV,
also indicating that anionic DNA molecules had attached
to the cationic surface. Results from PCS measurements
Statistical Analysis
are compiled in Table 1 .
Statistical analysis and determination of significance was per-
formed with SigmaStat software (SPSS Inc., Chicago, IL).
Data sets were tested using Student t test on raw data, and dif-
Atomic Force Microscopy Analysis of Nanoparticles
ferences were considered statistically significant if P < .05.
and DNA-nanoplexes
Atomic Force Microscopy (AFM) imaging showed all par-
R
ticles to be of an isodiametric shape. The preparation with
ESULTS
plain chitosan yielded particles with highly spherical shape
Particle Preparation
(Figure 1A). Comparison with the results from PCS meas-
The preparation resulted in a colloidal dispersion of nano-
urements showed the particles to be larger in diameter,
particles in water measuring 35 mL in total after the final
between 300 and 500 nm. The preparations with the lower
solvent evaporation step. In order to accurately prepare
substituted TMC derivatives (ie, 4% and 10%) were char-
particle-DNA-complexes the concentration of solid matter
acterized by particles embedded in a matrix of excipients
more than the homogeneously distributed chitosan prepara-
in the dispersions was determined. After drying the pellets
tion. In the case of TMC 4%, a spherical shape was
under nitrogen to a constant weight, concentration was
observed as well, with some particles surrounded by a
found to be 4 mg/mL for all preparations (or 140 mg solids
halo-like structure (Figure 1B). Size is similar to those pre-
in 35 mL). Since the polymer, PCL, is considered insoluble
pared with chitosan and hence $1.5 to 2 times as much as
in water, it can be assumed that only 40 mg or $30% of
PCS measurements. With an increasing degree of substitu-
the particlesÕ dry matter is made up from the water-soluble
tion, the discrete particles become less clearly visible and
excipients PVA and TMC, while the remaining 75 mg of
are to an increasing extent embedded in a matrix, as can be
the excipient blend (ie, 115 mg in total) remain dissolved
seen in Figure 1 (comparison of images B to D). Prepara-
in the aqueous phase.
tions with TMC 10% showed a trend toward less agglom-
eration of particles, while a preparation with TMC 18%
showed hardly any particles at all. Particles prepared using
Determination of Particle Size and Zeta Potential
TMC 66% showed a different result: all particles appear
The PCS analysis of particle sizes showed a unimodal size
clustered together in rows, and their size was found to be
distribution in most cases. All batches ranged in size
in a similar range as PCS measurements (Figure 1E). How-
between 250 and 300 nm with low polydispersity indices
ever, there is no spherical shape to be seen; furthermore
(0.03 to 0.13) indicating a narrow size distribution. The
many particles seem to have collapsed.
E25

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
Figure 1. AFM micrograph, particles prepared with chitosan derivatives: (A) chitosan hydrochloride; (B) trimethyl chitosan,
4% degree of substitution; (C) trimethyl chitosan, 10% degree of substitution; (D) trimethyl chitosan, 18% degree of substitution;
and (E) trimethyl chitosan, 66% degree of substitution. Length of scale bar 5 1 mm.
Gel Electrophoresis and Determination
The effect on LDH release is only little dependent on the
of Unbound DNA
particle concentration (Figure 2). For all nanoparticle for-
The binding capacity of cationic nanoparticles for plasmid
mulations, survival rates range between 40% and 80%
DNA was investigated by assessing the complexesÕ electro-
whether at higher concentrations of particles or at the
phoretic mobility in an agarose gel. If a complex is formed
lower end. Only in case of particles modified with TMC
efficiently, that is, all DNA is bound to the nanoparticles,
(66% modification) a different effect was observed: it
no bands for free DNA can be observed, and the immobile
showed a direct correlation between the particles concen-
complex remains in the starting zone in the same way as
tration and the survival rate.
the much larger particles do. Judging from the binding
The incubation of cells with nanoparticles over a period of
curves, an efficient binding started at a particle concentra-
4 hours has an influence on the cells as shown by the meta-
tion of 20 mg/mL (NP:DNA ratio of 2:1), and at a con-
bolic activity (Figure 3). However, there is no indication as
centration of 30 mg/mL (3:1) the DNA was nearly fully
to a correlation between particle concentration and survival
complexed. Therefore, concentrations of 10 mg/mL and
rate. As mentioned above in the LDH assay, the amount of
30 mg/mL were chosen as appropriate for transfection stud-
viable cells ranged between 40% and 80% survival rate of
ies. Results from DNA binding measured by gel electro-
a negative control. While the particles modified with chito-
phoresis are compiled in Table 2.
san or TMC of a low-substitution degree (4%) showed a
minimal survival at medium concentrations (see dip at
250 mg/mL), the TMC derivatives of an intermediate sub-
Cytotoxicity Studies
stitution degree (10% and 18%) seemed to influence the
The effects of chitosan-modified particles on COS-1 cells
cell growth even positively at higher concentrations. As in
were investigated by testing membrane integrity via the
case of the LDH release, only the highly substituted TMC
LDH release and metabolic activity via mitochondrial
66% shows a direct relation between dose and effect, with
enzymes.
a metabolic activity higher than an untreated control at the
E26

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
Table 2. Immobilization of Plasmid DNA by Cationic Nanoparticles*
Concentration
TMC
TMC
TMC
TMC
NP [lg/mL]
Chitosan
4%
10%
18%
66%
Ratio NP:DNA
0
100.00
100.00
100.00
100.00
100.00
0:1
1
68.27
95.45
133.95
118.20
64.64
0.1:1
10
0
57.88
103.80
105.20
23.94
1:1
20
0
23.04
29.77
45.40
0
2:1
30
0
0
2.58
13.67
0
3:1
50
0
0
18.95
7.35
0
5:1
100
0
5.16
15.33
13.19
0
10:1
*NP indicates nanoparticles; TMC, trimethyl chitosan. Numbers indicate percentage of free, uncomplexed DNA.
ing no effect on mitochondrial activity at all (Figure 5). The
LDH release followed no discernible pattern and could
therefore only be explained by an influence of handling
stress or the general surface-active properties of the sub-
stance, which might disturb the bilayer structures (Figure 5).
Transfection Studies
After 4 hours of transfection and 48 hours of gene expres-
sion, cells were still viable as could be expected from the
cytotoxicity studies with particle suspensions. First results
indicated the particle-DNA-complexes to be suitable for
transfection but much less effective than the commercial
transfection agent, PolyFect. Batches prepared with CS,
TMC 4%, and TMC 10% showed expression levels of
Figure 2. Cytotoxicity of trimethyl chitosan-modified
b-galactosidase (in fluorescence units) significantly higher
nanoparticles—membrane integrity measured in release of LDH.
than naked DNA (see Figure 6); while nanoplexes prepared
Error bars represent standard deviation of the mean.
with TMC 18% and TMC 66% remained below the chosen
level of significance. Further experiments are in progress
lower end of the concentration range and a total inhibition
to demonstrate the potentiality of the new PCL nanopar-
at the upper end.
ticles coated with chitosan and modified chitosans.
Statistical analysis revealed no significant differences
between different preparations, either for LDH test or for
DISCUSSION
WST-1 assay.
The preparation process resulted in a homogenous colloi-
The membrane degrading dose-effect relation of pure chito-
dal suspension of particles, which was stable for up to 7
san derivatives was more pronounced than of whole particle
days. Visualization via AFM showed that the particlesÕ
preparations. With the notable exception of TMC 66%, all
morphologies in the prepared batches are different. A
chitosans affected the cellsÕ membranes in a similar fashion
direct look at the particles revealed differences between
and intensity, with a survival rate of around 90% at the low-
the preparations. While some appeared as discretely dis-
est concentrations and 20% to 40% at the highest. The
tributed and spherically shaped particles, as in the case of
curves ran closely parallel with marginally higher variations
the preparations with chitosan hydrochloride, others
in the upper concentration range (see Figure 4).
(eg, TMC 10% and 18% degree of substitution) showed a
In contrast to this, PVA showed no linear relations between
thick matrix covering sparsely visible and irregularly
dose and cytotoxicity. The effect of PVA on the cellsÕ
shaped particles. This finding could perhaps be explained
metabolism was not discernible, since at all concentrations
by the increasing solubility in water or the polarity of the
the survival rate ranged around a 100% value, thus indicat-
chitosan derivatives. While pure chitosan is the least polar
E27

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
Figure 3. Cytotoxicity of trimethyl chitosan-modified
Figure 5. Cytotoxicity of PVA. Black circles: membrane
nanoparticles—metabolic activity of mitochondria measured by
integrity measured in release of LDH. Open triangles:
WST-1 assay. Error bars represent standard deviation of the
metabolic activity of mitochondria measured by WST-1
mean.
assay. Error bars represent standard deviation of the mean.
a new outer layer. This phenomenon should lead to an
increase in particle size, which was observed (see Figure 7).
Supercoiled plasmid DNA chains show a diameter of $8
nm, so the particle diameter should increase by $16 nm as
was observed in case of batches prepared with chitosan,
TMC 10%, and TMC 66%. The other batches (TMC 4%
and 18%) were found to have increased much more in
terms of particle size after incubation; this result was
attributed to particle agglomeration.
The positive zeta potentials were considered sufficient for
binding DNA long enough to safely transport it to the
desired site of action and subsequently release it. Particu-
late DNA carriers are reported20 with a zeta potential of up
to +50 mV, binding DNA to a higher extent than the ones
Figure 4. Cytotoxicity of chitosan and trimethyl chitosan
described here, but the question arises whether these highly
derivatives—membrane integrity measured in release of LDH.
charged carriers are actually capable of completely releas-
Error bars represent standard deviation of the mean.
ing their drug load at the target cells at all. It is interesting
to note that while the unmodified particles showed an
of the excipients, the polarity increases with the percentage
interdependence of substitution degree and zeta potential,
of cationic moieties in the molecule. Therefore, chitosan
no correlation of this kind could be detected with the
should adsorb better than other derivatives to the lipophilic
DNA-modified particles. It was concluded that the surface
polymer, a property that is reduced with the degree of
was saturated with DNA chains, while additional free
hydrophilicity (see Figure 1). These effects are reduced
phosphate groups determined the measured surface charge.
when the suspensions are highly diluted for PCS measure-
Insights into the cytotoxic effects of transfection agents are
ments, hence the similar sizes as determined with this
essential when dealing with new developments in this area.
method. The lower particle sizes and polydispersity indices
Since one of the best-known and widely used transfection
of particles prepared with high-substitution TMC (see
agents—PEI—is known for both its efficiency and toxicity,
Table 1) can be explained by a higher surface activity; the
the development of less harmful alternatives is still inter-
TMC derivatives can be regarded as surfactants similar to
esting. The impact on membrane integrity and metabolic
other cationic amphiphiles (eg, cetrimide).
activity were investigated, but no correlation or inter-
The plasmid chains associate themselves with the oppo-
dependence could be detected for particle concentration
sitely charged particles by wrapping around them, forming
and cytotoxicity. The different behavior of particles pre-
E28

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
between dose and effect (Figure 3). The substances are
however less toxic than the particles at corresponding con-
centrations. The different behavior of highly substituted
TMC 66% both pure and in a particle preparation can be
explained by the high degree of quaternization: cationic
amphiphilic molecules like benzalkonium chloride or cetyl
trimethyl ammonium bromide (cetrimide) have been
widely used as preservatives owing to their membrane-dis-
rupting properties. Concluding from this, TMC 66% seems
to show much more amphiphilic properties than the other
trimethyl chitosans and hence shows a higher cytotoxicity;
therefore its use as an excipient in a biological matrix
should be considered carefully, and it should be replaced
by lower substituted derivatives when possible.
First experiments aimed at transfection showed the
particle-DNA-complexes as being superior to naked DNA;
Figure 6. Transfection efficiency of nanoparticle-DNA-
however, experiments are underway aiming at the im-
complexes compared with pure DNA and a commercial
proved transfection rates, whereas similar studies using
transfection agent, PolyFect. Data are presented as fluorescence
units (mean 6 standard deviation). Asterisks (*) denote
chitosan-modified PLGA nanoparticles showed transfec-
a statistically
tion in vitro and in vivo. 23
pared with TMC 66% with its direct correlation can be
CONCLUSION
explained by the high amount of cationic moieties (see
The described method seems suitable to prepare cationic
below). The pure excipients seem to be less cytotoxic than
nanoscale carriers for nucleic acids. The main advantage of
the particle preparations when applied in the same concen-
the method can be seen in the fact that all degrading influ-
trations but in absence of any nanoparticles. PVA is little
ences such as contact of the sensitive DNA with extreme
or not at all toxic to the cell cultures (Figures 4), while chi-
pH values or organic solvents during the preparation proc-
tosan and its trimethyl derivatives with a low and inter-
ess is avoided; the DNA is added after all preparation steps
mediate degree of substitution show a linear correlation
are finished. The particlesÕ diameter depends only to a
Figure 7. Nanoparticle-DNA complex; particles prepared with trimethyl chitosan 4% degree of substitution. (A) amplitude signal,
lateral resolution 10 mm; (B) amplitude signal, lateral resolution 1.25 mm. Black square in panel A is zoomed in as panel B. Length of
scale bar 5 1 mm.
E29

---

AAPS PharmSciTech 2005; 6 (1) Article 6 (http://www.aapspharmscitech.org).
minor extent from the used chitosan derivative, which
12. Oberle V, Bakowsky U, Zuhorn IS, Hoekstra D. Lipoplex formation
allows this parameter to be varied in a relatively broad
under equilibrium conditions reveals a three step mechanism. Biophys
range. The cytotoxicity of the nanoparticle suspensions
J. 2000;79:1447-1454.
appears to be moderate and independent from the concen-
13. Mahato RI, Kawabata K, Nomura T, Takakura Y, Hashida M.
Physicochemical and pharmacokinetic characteristics of plasmid
tration (with the notable exception of 66% substituted
DNA/cationic liposome complexes. J Pharm Sci.
TMC, which appears to be significantly more toxic).
1995;84:1267-1271.
Transfection ability seems to be present but needs further
14. Farhood H, Serbina N, Huang L. The role of dioleyl
investigation.
phosphatidylethanolamine in cationic liposome mediated gene transfer.
Biochim Biophys Acta. 1995;1235:289-295.
15. Sternberg B, Hong K, Zheng W, Papahadjopoulos D.
ACKNOWLEDGMENTS
Ultrastructural characterization of cationic liposome-DNA complexes
Support from Leopoldina Foundation BMBF (9901/8—6),
showing enhanced stability in serum and high transfection activity in
vivo. Biochim Biophys Acta. 1998;1375:23-35.
Germany, to U. Bakowsky is gratefully acknowledged.
16. Meyer O, Kirpotin D, Hong K, Sternberg B, Park JW, Woodle
M.N.V. Ravi Kumar is grateful to Alexander von
MC, Papahadjopoulos D. Cationic liposomes coated with
Humboldt foundation, Germany, for the personal fellow-
polyethylene glycol as carriers for oligonucleotides. J Biol Chem.
ship.
1998;273:15621-15627.
17. Behr JP, Demeneix B, Loeffler JP, Perez-Mutul J. Efficient gene
transfer into mammalian primary endocrine cells with lipopolyamine-
REFERENCES
coated DNA. Proc Natl Acad Sci U S A. 1989;86:6982-6986.
1. Crystal RG. The gene as drug. Nat Med. 1995;1:15-17.
18. Torchilin VP, Levchenko TS. Rammohan R, Volodina N,
Papahadjopoulos-Sternberg B, DÕSouza GG. Cell transfection in vitro
2. Florence AT. Sakthivel T, Toth I. Oral uptake and translocation of a
and in vivo with nontoxic TAT peptide-liposome-DNA complexes.
polylysine dendrimer with a lipid surface. J Control Release.
Proc Natl Acad Sci U S A. 2003;100:1972-1977.
2000;65:253-259.
19. Cui Z, Mumper RJ. Plasmid DNA-entrapped nanoparticles
3. Ramaswamy C, Sakthivel T, Wilderspin AF, Florence AT.
engineered from microemulsion precursors: in vitro and in vivo
Dendriplexes and their characterization. Int J Pharm. 2003;254:17-21.
evaluation. Bioconjug Chem. 2002;13:1319-1327.
4. Pouton CW. Lucas P, Thomas BJ, Uduehi AN, Milroy DA, Moss SH.
20. Kneuer C, Sameti M, Haltner EG, Schiestel T, Schirra H, Schmidt
Polycation-DNA complexes for gene delivery: a comparison of the
H, Lehr CM. Silica nanoparticles modified with aminosilanes as carriers
biopharmaceutical properties of cationic polypeptides and cationic
for plasmid DNA. Int J Pharm. 2000;196:257-261.
lipids. J Control Release. 1998;53:289-299.
21. Kneuer C, Sameti M, Bakowsky U, et al. A nonviral DNA delivery
5. Ramsay E, Gumbleton M. Polylysine and polyornithine gene transfer
system based on surface modified silica-nanoparticles can efficiently
complexes: a study of complex stability and cellular uptake as a basis
for their differential in-vitro transfection efficiency. J Drug Target.
transfect cells in vitro. Bioconjug Chem. 2000;11:926-932.
2002;10:1-9.
22. Ravi Kumar MNV, Bakowsky U, Lehr C-M. Preparation and
6. Oupicky D, Konak C, Ulbrich K, Wolfert MA, Seymour LW.
characterization of cationic PLGA nanospheres as DNA carriers.
DNA delivery systems based on complexes of DNA with synthetic
Biomaterials. 2004;25:1771-1777.
polycations and their copolymers. J Control Release.
23. Ravi Kumar MNV, Mohapatra SS, Kong X, Jena PK, Bakowsky
2000;65:149-171.
ULehr C-M. Cationic poly(lactide-co-glycolide) nanoparticles as
7. Wagner E, Plank C, Zatloukal K, Cotten M, Birnstiel ML. Influenza
efficient in vivo gene transfection agents. J Nanosci Nanotechnol.
virus hemagglutinin HA-2 N-terminal fusogenic peptides augment gene
2004;4:990-994.
transfer by transferrin-polylysine-DNA complexes: toward a synthetic
24. Leong KW, Mao HQ, Truong Le VL, Roy K, Walsh SM, August JT.
virus-like gene-transfer vehicle. Proc Natl Acad Sci U S A.
DNA-polycation nanospheres as non-viral gene delivery vehicles.
1992;89:7934-7938.
J Control Release. 1998;53:183-193.
8. Fischer D, Bieber T, LiY, Elsasser HP, Kissel T. A novel non-viral
25. Capan Y, Woo BH, Gebrekidan S, Ahmed S, DeLuca PP.
vector for DNA delivery based on low molecular weight, branched
Preparation and characterization of poly (D,L-lactide-co-glycolide)
polyethylenimine: effect of molecular weight on transfection efficiency
microspheres for controlled release of poly(L-lysine) complexed
and cytotoxicity. Pharm Res. 1999;16:1273-1279.
plasmid DNA. Pharm Res. 1999;16:509-513.
9. Kunath K, von Harpe A, Fischer D, et al. Low-molecular-weight
26. Ravi Kumar MNV, Sameti M, Mohapatra SS, Kong X, Lockey RF,
polyethylenimine as a non-viral vector for DNA delivery: comparison
Bakowsky U, Lindenblatt G, Schmidt H, Lehr CM. Cationic silica
of physicochemical properties, transfection efficiency and in vivo
nanoparticles as gene carriers: synthesis, characterization and
distribution with high-molecular-weight polyethylenimine. J Control
transfection efficiency in vitro in vivo. J Nanosci Nanotechnol.
Release. 2003;89:113-125.
2004;