Tris(2-aminoethyl)amine-based α-branched fatty acid amides - Synthesis of lipids and comparative study of transfection efficiency of their lipid formulations This article is dedicated to Professor Gerald Brezesinski on occasion of his 65th birthday.

Article abstract of DOI:10.1016/j.ejpb.2015.08.011

The synthesis of a new class of cationic lipids, tris(2-aminoethyl)amine-based α-branched fatty acid amides, is described resulting in a series of lipids with specific variations in the lipophilic as well as the hydrophilic part of the lipids. In-vitro structure/transfection relationships were established by application of complexes of these lipids with plasmid DNA (pDNA) to different cell lines. The α-branched fatty acid amide bearing two tetradecyl chains and two lysine molecules (T14diLys) in mixture with the co-lipid 1,2-di-[(9Z)-octadec-9-enoyl]-sn-glycero-3-phosphoethanolamine (DOPE) (1/2, n/n) exhibits effective pDNA transfer in three different cell lines, namely Hep-G2, A549, and COS-7. The presence of 10% serum during lipoplex incubation of the cells did not affect the transfection efficiency. Based on that, detailed investigations of the complexation of pDNA with the lipid formulation T14diLys/DOPE 1/2 (n/n) were carried out with respect to particle size and charge using dynamic light scattering (DLS), ζ-potential measurements, and transmission electron microscopy (TEM). Additionally, the lipoplex uptake was investigated by confocal laser scanning microscopy (CLSM). Overall, lipoplexes prepared from T14diLys/DOPE 1/2 (n/n) offer large potential as lipid-based polynucleotide carriers and further justify advanced examinations.

Full text of DOI:10.1016/j.ejpb.2015.08.011

European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
Contents lists available at ScienceDirect
European Journal of Pharmaceutics and Biopharmaceutics
journal homepage: www.elsevier.com/locate/ejpb
Research Paper
Tris(2-aminoethyl)amine-based
a-branched fatty acid
amides – Synthesis of lipids and comparative study
of transfection efficiency of their lipid formulations
Nicole Erdmann a, Christian Wölk a,b, , Ingo Schulze a, Christopher Janich a, Manuela Folz a,
⇑
Simon Drescher c, Matthias Dittrich d, Annette Meister e, Jürgen Vogel f, Thomas Groth f, Bodo Dobner a,
,
⇑
Andreas Langner a
a Institute of Pharmacy, Biochemical Pharmacy, Martin Luther University (MLU) Halle-Wittenberg, Wolfgang-Langenbeck-Str. 4, 06120 Halle (Saale), Germany
b Department of Pharmaceutical Technology and Biopharmaceutics, Philipps University Marburg, Ketzerbach 63, 35037 Marburg, Germany
c Centre for Innovation Competence – HALOmem, MLU Halle-Wittenberg, Kurt-Mothes-Str. 3, 06120 Halle (Saale), Germany
d Max Planck Institute of Colloids and Interfaces, Am Mühlenberg 1, 14476 Potsdam, Germany
e Martin Luther University Halle-Wittenberg, Center for Structure and Dynamics of Proteins (MZP), Weinbergweg 22, 06120 Halle (Saale), Germany
f Institute of Pharmacy, Biomedical Materials Group, MLU Halle-Wittenberg, Heinrich-Damerow-Str. 4, 06120 Halle (Saale), Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 June 2015
Revised 21 August 2015
Accepted in revised form 24 August 2015
Available online 29 August 2015
The synthesis of a new class of cationic lipids, tris(2-aminoethyl)amine-based
amides, is described resulting in a series of lipids with specific variations in the lipophilic as well as
the hydrophilic part of the lipids. In-vitro structure/transfection relationships were established by appli-
cation of complexes of these lipids with plasmid DNA (pDNA) to different cell lines. The a-branched fatty
a-branched fatty acid
acid amide bearing two tetradecyl chains and two lysine molecules (T14diLys) in mixture with the co-
lipid 1,2-di-[(9Z)-octadec-9-enoyl]-sn-glycero-3-phosphoethanolamine (DOPE) (1/2, n/n) exhibits effec-
tive pDNA transfer in three different cell lines, namely Hep-G2, A549, and COS-7. The presence of 10%
serum during lipoplex incubation of the cells did not affect the transfection efficiency. Based on that,
detailed investigations of the complexation of pDNA with the lipid formulation T14diLys/DOPE 1/2 (n/
n) were carried out with respect to particle size and charge using dynamic light scattering (DLS), f-
potential measurements, and transmission electron microscopy (TEM). Additionally, the lipoplex uptake
was investigated by confocal laser scanning microscopy (CLSM). Overall, lipoplexes prepared from
T14diLys/DOPE 1/2 (n/n) offer large potential as lipid-based polynucleotide carriers and further justify
advanced examinations.
This article is dedicated to Professor Gerald
Brezesinski on occasion of his 65th birthday.
Keywords:
Cationic lipids
Gene therapy
Lipofection
Lipoplex
Non-viral gene transfer
Synthetic methods
Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction
but virus-based vehicles are connected with higher safety risk (e.g.
virus protein based immune reactions, oncogenesis) [5,6]. Further-
Gene therapy, the therapeutic insertion of genes into cells [1,2],
is a growing field in medical research. Furthermore, the RNA-based
strategies of gene silencing extended the application of polynu-
cleotide delivery systems in this field of medicine [2–4]. However,
effective polynucleotide transfer needs special vehicle systems,
called vectors. Currently, viral and nonviral vectors are used. The
viral ones are more effective in comparison with nonviral systems,
more, the loading capacity is limited by the virus capsid and the
upscaling process with respect to the industrial production is
difficult [7]. Because of these disadvantages of viral vectors, the
nonviral vehicle systems have become a promising alternative.
Cationic lipids (cytofectins) are the most common representatives
of nonviral vectors [2,8]. Unfortunately, current cationic lipid-
based polynucleotide vehicles are less effective compared with
viral vectors. Therefore, the synthesis of new cytofectins is
required. Because the mechanism of lipofection is still not fully
understood, the design of novel lipids has an empirical character.
Consequently, substance libraries of different lipids in varying
combinations with co-lipids have to be screened regarding their
transfection efficiency [9,10].
⇑
Corresponding authors at: Institute of Pharmacy, Biochemical Pharmacy, Martin
Luther University (MLU) Halle-Wittenberg, Wolfgang-Langenbeck-Str. 4, 06120
Halle (Saale), Germany. Tel.: +49 345 55 25078; fax: +49 345 55 27018 (C. Wölk).
Tel.: +49 345 55 25120; fax: +49 345 55 27018 (B. Dobner).
E-mail addresses: christian.woelk@pharmazie.uni-halle.de (C. Wölk), bodo.
dobner@pharmazie.uni-halle.de (B. Dobner).
http://dx.doi.org/10.1016/j.ejpb.2015.08.011
0939-6411/Ó 2015 Elsevier B.V. All rights reserved.

350
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
In this report we will present the synthesis of a lipid substance
library based on -branched fatty acid amides of tris(2-
ues. For the elemental analyses a CHNS-932 apparatus of the
a
LECO-corporation (St Joseph, Michigan, USA) was used. Electro-
spray ionization (ESI)-mass spectra were recorded with a Finnigan
MAT 719 (Thermo Separation Products, San Jose, USA) with an
electron spray ionization energy of 4.5 kV in negative and positive
modus. Electron impact ionization (EI)-mass spectra were pre-
pared with an AMD 402 apparatus (70 eV, AMD Intecta GmbH,
Harpstedt, Germany). High resolution mass spectrometry (HRMS)
was prepared with the hybrid mass spectrometer Q-TOF-2 (Waters
Micromass, Manchester, GB). For substances with a molecular
mass higher than 600 g/mol a LTQ-Orbitrap-mass spectrometer
(Thermo Fisher Scientific, Bremen, Germany) was used. The 1H
NMR spectra were recorded on Varian Gemini 2000 or a Varian
Inova 500 with CDCl3 as internal standard.
aminoethyl)amine that are suitable as polynucleotide carriers
(Fig. 1). The synthesis strategy is based on well-established reac-
tions, which require little preparative effort. The modular character
of the lipid synthesis allows a fast preparation of lipid libraries. The
synthesized lipid structures exhibit the amide bond as connective
unit. This allows degradation by enzymatic hydrolysis in combina-
tion with a higher stability against chemical hydrolysis compared
to ester bonds [11]. The directed structure variations allow the
investigation of the transfection efficiency in Hep-G2 cells (human
hepatocellular carcinoma) in dependence on structural variations
in the lipophilic as well as the hydrophilic part of the lipid. Potent
lipid formulations were further screened in A549 (human lung car-
cinoma) and COS 7 (African green monkey kidney) cells in presence
and absence of 10% fetal bovine serum (FBS) during lipoplex incu-
bation. It is remarkable that lipoplex formulations with a higher
serum stability compared to the standard LipofectamineTM were
found. Additionally, the most efficient lipid formulation, T14diLys
in mixture with 1,2-di-[(9Z)-octadec-9-enoyl]-sn-glycero-3-
phosphoethanolamine (DOPE) (1/2, n/n), in complex with pDNA,
was characterized by dynamic light scattering (DLS) experiments,
f-potential measurements, and transmission electron microscopy
(TEM). Finally, fluorescence probe labelled lipoplexes prepared
from T14diLys/DOPE (1/2, n/n) were investigated in time depen-
dent confocal laser scanning microscopic (CLSM) experiments to
gain further information about the cellular uptake.
Monoalkyl malonic acid diethylester (1a/b): The compounds 1a/b
were prepared according to literature [10,12] and purified by
column chromatography. The analytical data of synthesized com-
pounds are in agreement with the literature [13,14].
Dialkyl malonic acid diethylester (2a–c): The second alkylation
was carried out according to Bier [15] and Staudinger et al. [16]
starting from 0.1 mol of the corresponding monoalkyl malonic acid
diethylester in xylene to 0.1 mol of sodium hydride. After the
sodium salt has been generated (clear solution), 0.12 mol of the
alkyl bromide of the corresponding chain length was added, and
the mixture was heated for 10 h under reflux. Afterwards, 1 mL
water were added and stirred for 10 min. Then, 20 mL brine was
added and the mixture was extracted three times with CHCl3.
The organic layer was separated, dried over Na2SO4, and
evaporated. The raw dialkyl esters were purified using column
chromatography. The analytical data of ditetradecyl malonic acid
diethylester (2a) and dihexadecyl malonic acid diethylester (2c)
are in accordance with literature [16].
2. Materials and methods
2.1. Materials
a
-Branched fatty acids (3a–c): The fatty acids 3a–c were
If not mentioned otherwise, all materials were purchased from
Sigma–Aldrich (Steinheim, Germany). MilliQ-water with a specific
obtained from purified esters 2a–e according to the procedures
described by Staudinger et al. [17].
resistance of 18.2 MX cm was filtered through a 0.22 lm cellulose
a
-Branched chain fatty acid amides (T14, T14016, and T16):
acetate membrane before use. Cell lines were acquired from
German Collection of Microorganisms and Cell Cultures (DSMZ,
Braunschweig, Germany). Cell culture media, Dulbecco’s Modified
Eagle’s Medium (DMEM), Minimal Essential Medium (MEM), and
fetal bovine serum (FBS) were supplied from PAA Laboratories
GmbH (Cölbe, Germany).
100 mmol of tris(2-aminoethyl)amine was poured in a flask. Then,
5 mmol of the branched fatty acid (3a–c), 5 mmol (benzotriazol-1-
yloxy)tripyrrolidinophosphonium hexafluorophosphate (PyBOP),
and 5 lL triethylamine (TEA) dissolved in 100 mL CH2Cl2 were
dropped into the solution under stirring. The mixture was stirred
for further 12 h at room temperature. The mixture was filtered
and the organic solution was washed three times with water.
The organic layer was separated, dried over Na2SO4, and evapo-
rated. The crude product was purified by MPLC with CHCl3/
MeOH/NH3 using gradient technique.
2.2. Synthetic work
The solvents were analytically pure and dried before use. The
thin layer chromatography (TLC), which was used to check the pur-
ity of all compounds, was carried out on aluminium sheets coated
with silica gel 60 F254 (Merck, Darmstadt, Germany) and devel-
oped with bromothymol blue spray. Column chromatography
was performed using silica gel 60 (0.063–0.200 mm) for the nor-
mal pressure procedure and silica gel 60 (0.040–0.063) for middle
pressure liquid chromatography (MPLC). The melting points were
assigned with a Boetius apparatus representing uncorrected val-
a
-Branched chain fatty acid amides (T14diLys, T14016diLys, and
T16diLys): 5 mmol of the amino functionalized lipids T14,
0
´
T14 16, and T16, respectively, and 10 mmol N,N-di-Boc-L-lysine
hydroxysuccinimide (Boc-Lys(Boc)-OSu) were dissolved in 10 mL
CH2Cl2 and stirred for 12 h at room temperature. Afterwards the
solvent was removed by rotary evaporation. For the BOC cleavage
the crude product was dissolved in 10 mL ethyl acetate and 0.5 mL
HCl was added. The mixture was stirred for 2 h. The solvent was
removed and the crude product was dried over P2O5 in vacuum.
The product was purified and converted in the amine by MPLC
using CHCl3/MeOH/NH3 and the gradient technique.
Synthesis of fluorescence labelled transfection lipid (BFA-Rho6G):
see Supporting information.
The analytical data of Hexadecyltetradecyl malonic acid diethylester
(2b), 2-Tetradecylhexadecanoic acid (3a), 2-Tetradecyloctadecanoic
acid (3b), 2-Hexadecyloctadecanoic acid (3c), 2-Tetradecylo-
ctadecanoic acid-[2-bis(2-aminoethyl)aminoethyl]amide (T14016), 2-
Hexadecyloctadecanoic acid-[2-bis(2-aminoethyl)aminoethyl]amide
(T16), 2-Tetradecylhexadecanoic acid-(2-bis{[2-(2,6-diamino-1-oxohexyl)
amino]ethyl}aminoethyl)-amide (T14diLys), 2-Tetradecyloctadecanoic
Fig. 1. General structure of the tris(2-aminoethyl)amine-based
acid amides and the directed structural variations of the lipids reported herein.
a-branched fatty

N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
351
acid-(2-bis{[2-(2,6-diamino-1-oxohexyl)amino]ethyl}aminoethyl)amide
2.7. Transmission electron microscopy (TEM)
(T14016diLys), and 2-Hexadecyloctadecanoic acid-(2-bis{[2-(2,6-
diamino-1-oxohexyl)amino]ethyl}aminoethyl)-amide (T16diLys) are
presented in the Supporting information. The lipid 2-Tetradecylhexa-
decanoic acid-[2-bis(2-aminoethyl)aminoethyl]amide (T14) C36H76N4O
was already characterized by Antipina et al. [18].
Liposome or lipoplex dispersions were diluted with water to a
concentration of 0.05 mg/mL. The negatively stained samples were
prepared by spreading the dispersion (5 lL) onto a Copper grid
coated with a Formvar film (Plano, Wetzlar, Germany). After
1 min, excess liquid was removed by blotting with filter paper
and 1% aqueous uranyl acetate (5 lL) was placed onto the grid
2.3. Liposome and lipoplex formulation
and drained off after 1 min. The dried specimens were examined
using an EM 900 transmission electron microscope (Carl Zeiss
Microscopy GmbH, Oberkochen, Germany). Micrographs were
acquired using a SSCCD SM-1k-120 camera (TRS, Moorenweis,
Germany).
The cationic lipids and the co-lipids (DOPE or Cholesterol) were
dissolved in a CHCl3/MeOH (1/1, v/v) solution and mixed at differ-
ent molar ratios. The solvent was removed by rotary evaporation
yielding a thin lipid film, which was dried for further 2 h at 37 °C
and 15 mbar. The lipid film was rehydrated in MilliQ-water to a
concentration of 1 mg/mL and allowed to swell at 37 °C. Subse-
quently, the lipid dispersion was treated by three cycles of vortex-
ing and sonication (15 min, 37 °C).
2.8. Plasmid DNA (pDNA) isolation
The plasmid DNA pCMV sport b-gal, which is used for the trans-
fection experiments and encodes b-galactosidase, was isolated
For lipoplex preparation the liposome dispersion was combined
from Escherichia coli DH5a (Invitrogen GmbH, Darmstadt,
with the pDNA (0.15
l
g/well) at different charge ratios (+/ꢀ, with
Germany) using the Qiagen plasmid mega kit (Qiagen, Hilden,
Germany) according to the manufacturers protocol. Plasmid DNA
with an A260/A280 value between 1.8 and 2.0 was used for
experiments.
number of primary amino groups as + and number of DNA phos-
phates as ꢀ) and incubated for 15 min at 25 °C before adding to
the cells.
2.9. Cell lines and cell culture conditions
2.4. Dynamic light scattering (DLS)
All cells used in the experiments grow adherently as monolayer.
A549 cells (human lung carcinoma) were cultured in DMEM with
Size distributions of aggregate dispersions were characterized
with
a
Zetasizer Nano ZS ZEN3600 (Malvern Instruments,
2 mM glutamine, 50
(African green monkey kidney) were cultured in DMEM with
2 mM glutamine, 50 g/mL gentamicin, and 10% FBS. Hep-G2 cells
lg/mL gentamicin, and 10% FBS. COS-7 cells
Worcestershire, UK). The scattering angle was 173°. Three measure-
ments consisting of 10 runs with a duration time of 20 s for each
run were performed at 25 °C. For the calculations, a viscosity
l
(human hepatocellular carcinoma) were cultured in MEM with
g
= 0.8872 mPa s and a refractive index of 1.33 were assumed.
Earle’s salts containing 1 mM sodium pyruvate, 10 mM non-
The z-average diameters d were calculated from the correlation
functions using Zetasizer Software 6.34 (cumulant analysis), based
essential amino acid solution, 2 mM glutamine, 50 lg/mL gentami-
cin, and 10% FBS. All cell lines were cultured in 75 cm2 tissue flasks
at 37 °C and 5% CO2. Cells were dissociated after confluent growing
by trypsinization and transferred to new flasks. Hep-G2 cells were
passaged once a week, COS-7 and A549 twice a week.
on Stokes–Einstein equation [d = (k ꢁ T)/(3 ꢁ
p
ꢁ
g
ꢁ D)]. To gain size
distribution curves of polydisperse systems the autocorrelation
function was evaluated by ALV-correlation software version 3.0
using an exponential regularized fit.
2.10. Transfection efficiency and cytotoxicity assay
2.5. f-Potential measurements
Cells were seeded into 96-well sterile culture plates at a density
of 1–1.2 ꢁ 105 cells/mL for Hep-G2, 2.0–2.5 ꢁ 104 cells/mL for A549,
and 5.0–5.5 ꢁ 104 cells/mL for COS-7 and grown 16–24 h to approx-
imately 75% confluence. Afterwards, the medium was removed, the
cells were washed with phosphate buffered saline (PBS), and lipo-
The electrophoretic measurements were carried out with the
laser Doppler electrophoresis technique using a Zetasizer Nano
ZS ZEN3600 (Malvern Instruments, Worcestershire, UK). Three
measurements consisting of 15 runs with a voltage of 50 V
plex dispersion was added (0.15 lgpDNA/well). Transfection was
were performed at 25 °C. For the calculations,
a
viscosity
allowed to proceed for 4 h at 37 °C, in a humidified atmosphere
containing 5% CO2 in serum-free growing medium. Afterwards,
FBS was added to the culture to reach a final concentration of
10%. The medium was replaced 24 h after transfection with fresh
medium.
The viability and the b-galactosidase activity were determined
using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT)-assay, o-nitrophenyl-b-galactoside (ONPG), and
bicinchoninic acid (BCA)-assay as previously described [10].
g
= 0.8872 mPa s, a dielectric constant e = 78.5 F/m, and a refractive
index of 1.33 were assumed. The mobility
l of the diffusing aggre-
gates was converted into the f-potential using the Smoluchowski
relation f = ( )/e (Zetasizer Software 6.34).
l g
ꢁ
2.6. Differential scanning calorimetry (DSC)
DSC measurements were performed on a microcalorimetry
system (MCS DSC, MicroCal Inc., Northampton, MA). Samples were
degassed under vacuum before use. The DSC cells, with sample and
reference, were kept under 2 bar of nitrogen pressure to prevent
evaporation at higher temperatures. The heating rate was 60 K/h.
Each heating and cooling scan was repeated to confirm repro-
ducibility. The first scan was discarded. The reference cell was
filled with pure water and the water baseline was subtracted from
the thermograms of the samples. The DSC scans were evaluated
using MicroCal Origin 8.0 software.
2.11. Confocal laser scanning microscopy (CLSM) experiments
The lipoplexes were double tagged: YOYO-1 (kmexax = 491 nm,
kmemax = 509 nm) labelled pDNA was prepared according to manufac-
turer’s protocol and a fluorescence labelled lipid (a-branched fatty
acid BFA-Rho6G) (kemxax = 535 nm, kemmax = 560 nm) was incorporated
into the liposomes (T14diLys/DOPE/BFA-Rho6G, 1/2/0.01, n/n/n).
The lipoplexes and liposomes were prepared as mentioned
above.

352
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
COS-7 cells were plated on coverslips (18 ꢁ 18 mm) on the bot-
The a-branched fatty acids (3a–c) were transformed into the
tom of 6-well plates 24 h before transfection at an initial density of
fatty acid amides (T14, T14016, T16) with the use of tris(2-
aminoethyl)amine (Scheme 2). This amine was used as head group
structure itself as well as a spacer structure in enlarged head
groups. For the linkage of the fatty acid with the amine, the amide
formation with PyBOP was the method of choice as described pre-
0.5–1.0 ꢁ 105 cells in 250
lL culture medium. After 4 h the medium
was removed with caution, replaced by 2.5 mL fresh medium, and
incubated overnight. After medium removal and a subsequent
washing step, the labelled lipoplexes were added to the cells and
incubated for 4 h without serum. Afterwards, the transfection
medium was replaced by medium with 10% FBS. After 2, 6, and
24 h, respectively, the cells were washed and the plasma mem-
brane was labelled using the wheat germ agglutinin (WGA) Alexa
FluorÒ 633 conjugate (kemxax = 632 nm, kemmax = 647 nm). WGA selec-
tively binds to N-acetylglucosamine and N-acetylneuraminic acid
residues [19]. A dilution of the conjugate was added at a concentra-
viously [12,20]. To gain an enlarged head group, the
a-branched
fatty acid amides T14, T14016, and T16 were converted into the
compounds T14diLys, T14016diLys, and T16diLys by reaction with
the activated ester Boc-Lys(Boc)-OSu in an efficient reaction with-
out any condensation reagent followed by the HCl induced BOC
cleavage (Scheme 2) [12].
The described synthesis has a modular character. Consequently,
an easy variation of the alkyl chain pattern as well as the head
groups is possible. Individual modules can be exchanged and
additional modules coupled to the primary amino functions by
the formation of amide bonds. This allows the synthesis of lipid
libraries, which can exceed the number of lipids presented in this
study by far.
tion of 5 lg/mL to the cells and incubated for 10 min at 37 °C
followed by a washing step. Afterwards, the cells were fixed for
15 min with a mixture of 2% formaldehyde and 0.2% glutaralde-
hyde in PBS buffer and subsequently washed two times with PBS.
Finally, the cells were prepared on a glass slide with the embed-
ding agent AquatexÒ.
The fluorescence microscopic images were acquired with a
confocal laser scanning microscope (Zeiss LSM 710, Carl Zeiss, Jena)
For CLSM studies a fluorescence labelled lipid was needed. To
ensure an incorporation of the labelled lipid in the cationic lipid
assemblies the hydrophobic part should exhibit a structure similar
to the cationic lipid. Due to the fact that CLSM studies were
performed with T14diLys lipoplexes, the fluorescence labelled
lipid BFA-Rho6G was a 2-tetradecylhexadecanoic acid amide, too.
Contrary to the transfection lipid, ethylene diamine was used as
linker structure (Scheme S1, Supporting information). This allows
homogeneous labelling because the ATTO Rho6GÒ label binds to
the one free amino group.
using an inverse microscope (AxioObserver.Z1) with
a fast
z-scanning stage and 40ꢂ and 63ꢂ Plan APOCHROMAT objectives.
Zeiss Software ZEN2008 was used for data processing, generation
of images, and control of the microscope.
2.12. Statistical analysis
One-way ANOVA in conjunction with Bonferroni’s multiple
comparison tests was used to assess the statistical significance.
The level of significance was set as 95% (p = 0.05).
3.2. Transfection efficiency screening
A summary of the screening with regard to the transfection effi-
ciency and cell viability in Hep-G2 cells of the six cationic lipids
described above is shown in Fig. 2. The transfection efficiency
was determined by the established ONPG assay using a plasmid
DNA encoding for b-galactosidase [21,22]. The viability was ascer-
tained using the MTT assay [23]. For this screening the cationic
lipids were used in mixture with DOPE as co-lipid in a molar ratio
of 1/2 (cationic lipid/DOPE). This ratio turned out to be the most
effective mixture in screening experiments with different mixing
ratios (data not shown). Moreover, the results depicted in Fig. 2
represent the optimized and most efficient charge ratio (+/ꢀ) of
the lipoplexes used.
3. Results and discussion
3.1. Synthesis of the
a-branched fatty acid amides
The synthesis of the new
a
-branched fatty acid amides starts
from monoalkyl malonic acid diethylesters (1a/b) with 14 or 16
carbon atoms in the alkyl chains (Scheme 1). Such compounds
are easy to prepare by alkylation of malonic acid diethylester and
the appropriate alkyl bromides using a strong base. The monoalkyl
malonic acid diethylesters were transformed into dialkyl malonic
acid diethylesters by a common second alkylation step whereas
the chain length of the second hydrophobic chain is controllable
by the alkylation reagent. After the commonly used saponification
The lipids T14, T14016, and T16 consist only of the
a-branched
fatty acid which is linked to a tris(2-aminoethyl)amine unit.
These lipids exhibit transfection efficiency values in Hep-G2 cells
similar to the reference LipofectamineTM. However, the lipid compo-
sitions containing one of these three lipids bear a higher cytotoxic
and decarboxylation steps, the
isolated in good yields.
a-branched fatty acids (3a–c) were
Scheme 1. Synthesis of the a-branched fatty acids: (i) NaH, hexadecyl bromide or tetradecyl bromide, xylene; (ii) NaOH, ethanol; (iii) H2O, HCl; (iv) 200 °C.

N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
353
Scheme 2. Synthesis of the amino functionalized a-branched fatty acid amides: (i) PyBOP, triethylamine, tris(2-aminoethyl)amine, CH2Cl2; (ii) BocLys(Boc)-OSu, TEA, CH2Cl2;
(iii) HCl, ethyl acetate.
Table 1
f-Potential of liposome formulations (1 mg/mL) consisting of one cationic lipid in
mixture with DOPE (1/2, n/n). The results are given as mean standard deviation
(n P 12).
Cationic lipid/DOPE 1/2
f-Potential (mV)
s
T14
19 11
T14016
T16
32
21
38
56
42
2
6
6
5
5
T14diLys
T14016diLys
T16diLys
Fig. 2. Results of the transfection efficiency 48 h after lipoplex incubation (ONPG-
assay, bottom) and the cell viability (MTT-assay, top) of the cationic lipids in
mixture with DOPE (1/2, n/n) compared to the reference LipofectamineTM in Hep-G2
cells. The presented values are the most effective determined charge ratio for each
transfection system. The results are given as mean standard deviation (n P 12).
The symbols at the bars highlight significant differences.
potential. The viability for these formulations is in the range of 50–
55%, a disadvantage compared to the reference LipofectamineTM
with a viability of 84%. The insertion of two lysine molecules in
the polar head group (T14 ? T14diLys, T14016 ? T14016diLys,
and T16 ?T16diLys) increases the transfection efficiency signifi-
cantly: The efficiency was 2- to 6-fold higher compared to the lipid
analogues without lysine in the head group (Fig. 2). Thereby, a sig-
nificant higher transfection compared to LipofectamineTM was
achieved with lysine-containing
The transfection efficiency values determined with the lysine-
containing compounds illustrate distinct chain-effect. The
a-branched fatty acid amides.
a
exchange of at least one hexadecyl chain by a tetradecyl chain
results in an increased transfection efficiency by a factor of 2
(T16diLys ? T14016diLys or T16diLys ? T14diLys). Possible rea-
sons will be discussed below. T14diLys and T14016diLys exhibit
the highest efficiency, which exceed the reference 5-fold. The
enhancing effect of lysine in the head group of cationic lipids on
transfection efficiency was observed in previous investigations in
our research group [10,20].
Fig. 3. (a) Intensity-weighted size distribution curves of liposome formulations
(1 mg/mL) consisting of one cationic lipid in mixture with DOPE (1/2, n/n). The
corresponding cationic lipid is indicated at the curves. The curves are shifted
vertically for clarity. (b) The corresponding correlation functions of the distribution
curves presented in (a).
Table 1 demonstrates that the lipids without lysine units in the
head group exhibit a lower f-potential compared to their lysine-
containing counter partners. The particle size is comparable for

354
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
all 6 lipid formulations (major size fraction with radii between
10 nm and 200 nm, see Fig. 3). T14016, T14diLys, T14016diLys,
and T16diLys formulations exhibit a minor size fraction with radii
above 200 nm and T14, T16, and T14016diLys formulations exhibit
an additional size population with radii below 5 nm.
manner [Tm(T14diLys) < Tm(T14016diLys) < Tm(T16diLys)]. The alkyl
chain fluidity (number of gauche-conformers) has effects on the
lipid shape (critical packing parameter P established by Israelach-
vili) and the connected curvature effects of the lipids [24] and, as
a consequence, determines the preferred mesophases and the lipid
mixing behaviour. The latter two characteristics determine the
transfection efficiency (read Rehman et al. [25]).
The most efficient formulations in Hep-G2 cells, T14diLys/DOPE
(1/2, n/n) and T14016diLys/DOPE (1/2, n/n) were tested in two
other cell lines: A549 and COS-7 cells. The transfection results in
all three cell lines and the viability are summarized in Fig. 4a.
The good efficiency of T14016diLys in Hep-G2 cells was not repro-
ducible in the other two cell lines. The formulation only reached
about 45% of the efficiency of LipofectamineTM in A549 and COS-7
cells in combination with an increased cytotoxicity. In contrast,
the lipid formulation with T14diLys achieved transfection efficien-
cies comparable with LipofectamineTM in A549 cells. In COS-7 cells
T14diLys tended to be more efficient than the commercially avail-
able standard. Nevertheless, the lipid system T14diLys/DOPE (1/2,
n/n)) is a high effective polynucleotide transfer system in three cell
lines comparable with and surpassing, respectively, Lipofec-
tamineTM. For this reason, this lipid formulation will be character-
ized in more detail in the next chapter.
The results presented in Fig. 4 also indicate the distinct influ-
ence of the alkyl chain pattern on the transfection efficiencies.
While in Hep-G2 cells the lipids T14diLys and T14016diLys show
comparable results (in Hep-G2 cells the chain effect is more pro-
nounced for the exchange with T16diLys as mentioned above),
T14diLys exhibits 2-fold higher transfection efficiency compared
with T14016diLys in COS-7 and A549 cells. DSC experiments show
a pronounced shift of the phase transition (Tm) from 40.5 °C
(T16diLys) to 9 °C (T14016diLys) if only one hexadecyl chain is
exchanged by a tetradecyl chain. T14diLys shows no phase transi-
tion between 2 °C and 95 °C. Presumably, the transition is below
0 °C. Assuming, that the transitions are caused by alkyl chain melt-
ing, the stepwise exchange of the tetradecyl chains by hexadecyl
chains increases the chain melting temperature in a distinct
3.3. Detailed characterization of T14diLys-containing lipid
formulations as polynucleotide carrier
3.3.1. Transfection efficiency and cytotoxicity in dependence of the
co-lipid
To investigate the influence of co-lipids on the transfection
efficiency and cytotoxic behaviour the cationic lipid T14diLys
was screened in mixture with DOPE or cholesterol in different
molar ratios.
Without a co-lipid the lipid T14diLys is an inefficient transfec-
tion system (Fig. 5a). The transfection efficiency increases signifi-
cantly after a co-lipid (DOPE or cholesterol) was added to the
vehicle system. DOPE exhibits a larger enhancing effect on the
transfection efficiency than cholesterol. This becomes particularly
obvious in the formulation T14diLys/DOPE (1/2, n/n): The effi-
ciency increases threefold compared to the same molar mixing
ratio with cholesterol. The observed transfection-enhancing effect
of the co-lipids may be associated with the fact that the addition
of neutral co-lipids to the cationic lipid reduces the charge density
of the lipid formulation, whereupon the addition of either choles-
terol or DOPE to T14diLys yields different charge densities because
of their different head group areas. The charge density determines
the binding strength between the lipid formulation and the DNA,
which influences the vector unpacking (DNA-release),
a key
parameter of the transfection process [26,27]. Ahmad et al. dis-
cussed a bell curve for transfection efficiency versus membrane
charge density with an optimum in charge density for efficient
Fig. 4. (a) Results of the transfection efficiency 48 h after lipoplex incubation (ONPG-assay, bottom) and the cell viability (MTT-assay, top) values of the cationic lipids
T14diLys and T14016diLys in mixture with DOPE (1/2, n/n) compared to LipofectamineTM in A549, COS-7, and Hep-G2 cells. The results are given as mean standard deviation
(n P 12). The values presented are at the optimized charge ratio. (b) DSC heating curves of T14diLys, T14016diLys, and T16diLys dispersions (1 mg/mL) in H2O. The curves are
shifted vertically for clarity.

N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
355
Fig. 5. (a) Transfection efficiency 48 h after lipoplex incubation (ONPG-assay, bottom) and cell viability (MTT-assay, top) values of T14diLys in mixture with DOPE and
cholesterol, respectively, in different molar ratios (n/n) compared to LipofectamineTM in Hep-G2 cells. The most effective determined charge ratio for each transfection system
is shown. (b) Cell viability of different T14diLys formulations as a function of the molar amount (ncat.) of T14diLys in each well. All results are given as mean standard
deviation (n P 12). (c) Intensity-weighted and number-weighted size distribution curves of T14diLys in aqueous dispersion (1 mg/mL). The curves are shifted vertically for
clarity. The inset shows the corresponding correlation function.
gene transfer [28]. Further, the ability of DOPE to form inverse
hexagonal mesophases could explain the enhancing effect of DOPE
in the lipid formulation. The formation of inverse hexagonal phases
is discussed as one key step of interactions between membranes of
endosomes and lipoplexes resulting in an efficient endosomal
escape of the polynucleotide cargo [29,30].
The first population is in a size range typical for lipid-based
micelles and dominates in the number weighted size distribution
curve (Fig. 5c), which is more representative for polydisperse for-
mulations. T14diLys in mixture with DOPE builds vesicles in
aqueous dispersions (see Section 3.3.3). Consequently, T14diLys
alone forms micelles which are smaller and exhibit a higher charge
density compared to vesicles built of T14diLys/DOPE mixtures,
what can consequently affect the abovementioned receptor medi-
ated cell toxicity. Additionally, the curves in Fig. 5b demonstrate
the increasing effect of the increasing amount of cationic lipids
on the cytotoxicity, which is also a known phenomenon [11,33].
The viability values of Hep-G2 cells treated with T14diLys lipo-
plexes alone, as well as in mixture with the co-lipids DOPE and
cholesterol, respectively, are shown in Fig. 5b as a function of the
amount of cationic lipid per well. Lipid mixtures with DOPE are
less toxic than cholesterol containing mixtures at the same values.
Nevertheless, lipoplexes without co-lipid exhibit better viability
values. The toxic effects of co-lipids are described ambivalently
in the literature: Promoting [11] as well as suppressive [10] effects
on cell survival are described. While the mechanism of cationic
lipid assembly mediated cytotoxicity is not fully understood yet,
the activation of cell signal pathways resulting in cell death can
be also involved in the process of cationic lipid induced cell death
[31,32]. For this effect the charge density and size of the lipid
aggregates seem to play an important role in receptor activation
[31,32]. The charge density is influenced by the co-lipid. Whether
the addition of the co-lipid shifts the charge density to the opti-
mum of receptor activation or not may explain the ambivalent
behaviour of co-lipids on the cytotoxicity. Also, the type and, con-
sequently, the size of the preferred aggregates of T14diLys in pres-
ence or absence of a co-lipid are different. The large head group of
T14diLys results in a conical molecular shape (high positive curva-
ture) and, accordingly, micelles are the preferred lipid aggregates
in aqueous dispersions. DLS measurements of T14diLys dispersions
result in a bimodal intensity-weighted size distribution curve
(Fig. 5c) with size populations at r = 2–5 nm and r = 20–100 nm.
3.3.2. Influence of serum
Keeping further in vivo applications in mind, first serum stabil-
ity experiments were performed. For this purpose, 10% FBS was
added during the lipoplex incubation of the cells. The results are
summarized in Fig. 6. Usually, the addition of serum results in a
decrease in lipofection efficiency due to interactions between pro-
teins of the serum and lipoplexes, which further results in lipoplex
disintegration and DNA release before the lipoplex can enter the
cell [34]. For instance the transfection efficiency of LipofectamineTM
decreases by around 50% in Hep-G2 cells (determined for this
study using the ONPG assay) and by around 40% in A549 cells
[10] in the presence of 10% serum.
T14diLys/DOPE (1/2, n/n) is stable against the presence of 10%
FBS in all three cell lines since no significant decrease in transfec-
tion efficiency was observed. In the experiments with T14diLys/
DOPE (1/2, n/n) in A459 cells with a serum free incubation interval
an optimized charge ratio of 3.5 was found out. The same ratio
applies for the experiments in the presence of serum during lipo-
plex incubation. During transfection of COS-7 cells the optimized

356
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
that allows the estimation of the particle charge, gives an explana-
tion for this bell-shaped size curve. Nevertheless, it has to be taken
into account that the f-potential measurement (potential at the
shearing plane of the particle surrounding ion cloud) cannot yield
the real surface potential (Nernst-potential). The f-potential of
lipoplexes at different charge ratios results in a sigmoidal curve
with two plateau regions: the first at charge ratios below 1
(f = ꢀ10 mV, negative overcharging, DNA excess) and the second
at charge ratios above 3 (f = 30 mV, positive overcharging, cationic
lipid excess). The positive overcharging effect is typical for a one-
step lipoplex formation procedure used in this work [35]. In the
region between the two plateaus the electrostatic stabilization of
the lipoplex dispersions is lower than in the plateau regions result-
ing in aggregation effects and, hence, in higher particle sizes. This
effect should be most pronounced at the point where the curve
passes f = 0 mV, the isoelectric point (IEP). The determined IEP is
at charge ratio 1.7, while the size maximum is at 2.6. This mis-
match can result from the mismatch between f- and Nernst-
potential mentioned above.
Following, comparisons between the optimized (most effective)
charge ratios of the lipid formulation in the three different cell
lines (Hep-G2: (+/ꢀ) = 2.5, COS-7: (+/ꢀ) = 3, and A549: (+/ꢀ)
= 3.5) and the size and f-potential curves presented in Fig. 7 were
made. All three charge ratios are in the region where the largest
Fig. 6. Results of the transfection efficiency 48 h after lipoplex incubation (ONPG-
assay, bottom) and the cell viability (MTT-assay, top) values of the cationic lipids
T14diLys in mixture with DOPE (1/2, n/n) in A549, COS-7, and Hep-G2 cells in
presence as well as absence of 10% FBS during lipoplex incubation. The results are
given as mean standard deviation (n P 12). The values presented are at the
optimized charge ratio.
lipoplexes occur. Nevertheless, the diameters are below 1 lm
and therefore the formulations are theoretically suitable for
in vivo experiments. The lipoplexes at all three charge ratios are
characterized by a positive f-potential above 20 mV, a fact often
discussed as necessity for effective lipofection but which is also
connected with problems for in vivo applications due to interac-
tions with biologic material [36].
charge ratio was 3 in the absence of serum and 4.5 in the presence
of serum. For Hep-G2 the optimized charge ratio of 2.5 was the
same in absence and presence of serum.
Finally, the IEP allows to determine the effective charge (qeff) of
the cationic lipid within the lipid formulation: [37] qeff = n(NH2)/
IEP with n(NH2) as the number of primary amines in one cationic
lipid. In the transfection system T14diLys/DOPE (1/2, n/n) qeff = 2.4.
This means that T14diLys acts like a lipid with 2.4 positive charges
assuming that one positive charge can complex one DNA phos-
phate (a situation assumed at the IEP). Nevertheless, it has to be
taken into account that qeff can deviate from the real protonation
degree due to steric effects, which can shield charged groups from
the complex formation process [38].
3.3.3. Particle characterization
T14diLys/DOPE (1/2, n/n) was intensively characterized with
respect to particle size and particle charge (f-potential), respec-
tively, as a function of the ongoing complex formation with pDNA.
The results are given in Fig. 7. The particle size of the lipoplexes
increases with increasing charge ratio (addition of lipid to a con-
stant amount of pDNA, increasing pDNA complexation) until a
maximum of about d (z-average) = 800 nm is reached at a charge
ratio of 2.57. A further increase of the charge ratio results in a
decrease of the particle size, finally resulting in a bell-shaped
curve. The investigation of the corresponding f-potential, a value
To get further information of the size distribution of selected
z-average values, selected correlation functions from the DLS
Fig. 7. (a) Particle size given as diameter (z-average) of T14diLys/DOPE (1/2, n/n) lipoplexes [c(pDNA) was 10.1 mg/mL] as a function of the charge ratio resulting from the
complex stoichiometry of the complex formation with pCMVꢁSPORT b-gal. The curve results from a Gauss fit (r2 = 0.922, maximum = 2.57). (b) f-potential of T14diLys/DOPE
(1/2, n/n) lipoplexes [c(pDNA) was 10.1 mg/mL] as a function of the charge ratio resulting from the complex stoichiometry of the complex formation with pCMVꢁSPORT b-gal.
The curve results from a Boltzmann fit (r2 = 0.987, inflection point (+/ꢀ) = 2.09). All values are given as mean standard deviation (n = 3).

N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
357
measurements of lipoplexes were evaluated by a regularized fit.
The results are presented in Fig. 8a–f as intensity- and number-
weighted size distribution functions. The number-weighted curve
considers the lower light scattering of smaller particles during
the DLS measurement and is more representative for polydisperse
systems. The size distribution curves of lipoplexes at a charge ratio
of 0.5 are monomodal for the intensity-weighted curve (Fig. 8a)
while the number-weighted shows a bimodal shape (Fig. 8d). A
comparison to the z-average diameter shown in Fig. 7a demon-
strates the disadvantage of the cumulant analysis in the case of
broad size distributions [d(z-average) = 200–300 nm]. The
z-average value is dominated by the larger particles. The distribu-
tion curves determined by the regularized fit show the dominant
particle species at d = 80–140 nm. At a charge ratio of 2 (near the
IEP) a bimodal intensity-weighted size distribution curve shows
maxima at 130 nm and 880 nm (Fig. 8b). The determined
z-average was between 400 nm and 500 nm (Fig. 7a). In the light
of the number-weighted function the smaller particles with radii
below 330 nm are dominating the lipoplex dispersion (Fig. 8e).
Nevertheless, the increase of the particle size near the IEP
mentioned above is still present. At a charge ratio of 4.5 the size
of lipoplexes decreases and size distribution curves are comparable
to the charge ratio of 0.5 and occur with a slight shift of the first
maximum to smaller diameter. In conclusion, this investigation
shows that z-average values calculated by cumulant analysis
allow to determine the particle size of the bigger lipoplexes in a
Fig. 8. (a–f) Intensity- (a–c) and number-weighted (d–f) size distribution curves of T14diLys/DOPE (1/2, n/n) lipoplexes [c(pDNA) was 10.1 mg/mL] at three different charge
ratios 0.5 (a and d), 2 (b and e), and 4.5 (c and f). (g–l) TEM images of negatively stained samples prepared from an aqueous T14diLys/DOPE (1/2, n/n) lipoplex dispersion at
different charge ratios (c = 0.05 mg/mL): (g and j) charge ratio 0.5; (h and k) charge ratio 2; (i and l) charge ratio 4.5.

358
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
polydisperse system neglecting smaller particles. However, the z-
average curve presented in Fig. 7a allows estimating the size of lar-
ger particles with respect to further in vivo applications.
particles without DNA occur and compete with the lipoplexes for
cell attachment and entering.
The structure of selected T14diLys/DOPE (1/2, n/n) lipoplexes
was investigated by means of negative stain TEM (Fig. 8g–l). The
lipoplex formulation T14diLys/DOPE (1/2, n/n) with a charge ratio
of 0.5 exhibits DNA excess (negative f-potential, below the deter-
mined IEP). The image shows lipoplexes (Fig. 8g, white arrow)
being composed of stacked lamellae (Fig. 8j, white arrow) beside
more diffuse structures (Fig. 8g and j, black arrows) which can be
interpreted as uncomplexed pDNA. The pDNA (negatively charged)
containing structures exhibit higher contrast due to the interaction
with positively charged uranyl-ions. The estimated size of the lipo-
plexes with diameters below 500 nm fits the determined size using
DLS (compare Fig. 8a, d, and g). The charge ratio of the second
investigated lipoplex formulation T14diLys/DOPE (1/2, n/n) is with
2 slightly above the determined IEP (charge ratio 1.7). Neverthe-
less, uncomplexed cationic liposomes (Fig. 8k black arrow) are pre-
sent beside lamellar lipoplex assemblies (Fig. 8k white arrow). The
T14diLys/DOPE (1/2, n/n) lipoplexes at charge ratio 2 give more
clustered aggregates compared to the formulation with a charge
ratio of 0.5 (compare Fig. 8h and k with j). This is in line with
the charge dependent lipoplex aggregation discussed above
(Fig. 7). An insert of Fig. 8k is shown in the Supporting information
(Fig. S1). At cationic lipid excess (T14diLys/DOPE (1/2, n/n) with a
charge ratio of 4.5) several lipid aggregates without complexed
DNA (structures of low contrast, Fig. 8i, black arrows) are present
beside lamellar lipoplexes (structures of high contrast, Fig. 8i,
white arrows, Fig. 8l). The estimated size of both types of struc-
tures with diameters below 500 nm fits the determined aggregate
size determined by DLS (compare Fig. 8c, f, and i). The electron
micrographs also allow to explain why lipoplex charge ratios far
below (T14diLys/DOPE (1/2, n/n) charge ratio 0.5) and far above
(T14diLys/DOPE (1/2, n/n) charge ratio 4.5) the IEP are less effec-
tive: Below a high fraction of uncomplexed DNA is accessible for
nucleases, while above a high amount of positively charged lipid
3.3.4. Cellular uptake of lipoplexes investigated by CLSM
To investigate interactions of the most effective lipoplex formu-
lation T14diLys/DOPE (1/2, n/n) with a charge ratio of 2.5 in com-
plex with pCMV sport b-gal in COS-7 cells, time dependent CLSM
investigations were performed (2, 6, and 24 h after lipoplex incu-
bation). This allows the visualization of the lipoplex uptake and
the DNA release and is therefore a complementary method to the
ONPG-assay. COS-7 cells were chosen because of the low
background fluorescence required for CLSM investigations. The
lipoplexes were double tagged with YOYO-1Ò iodide as DNA tag
(shown in red) and BFA-Rho6G as lipid probe (shown in green)
while the cell membranes were labelled with WGA-Alexa FluorÒ
633 (white). Hence, a co-localization of lipid and DNA (intact
lipoplex) results in a yellow colour.
Confocal images along the z-direction allow to determine the
position of objects on and in the cell. Two hours after transfection
lipoplexes are still localized extracellularly and they seem to inter-
act with the cell membrane (see Supporting information, white
triangles in Fig. S2). In this timescale the processes of adhesion of
lipoplexes were observed. Lipoplexes were detected on the cell
surface at different focal planes along the z-direction (perpendicu-
lar to the glass slide) (white triangles pointing to yellow structures
in Fig. S2). This observation is further demonstrated in Fig. 9 as an
enlarged section of Fig. S2 (white box, image C). The view of the
planes in x-, y-, and z-direction allows to determine the position
of objects relative to the cell. The fluorescence intensities of all
three dyes are shown in separate images (Fig. 9A–C) to observe
the labelled structures without mixed colours and also as multi-
colour image (Fig. 9D). The green (Fig. 9A) and red (Fig. 9B) staining
results in yellow (mixed) staining (Fig. 9D) indicating the presence
of lipoplexes (co-localization of lipid and DNA label). In Fig. 9C,
which shows the staining of the cell membrane with WGA-Alexa
Fluor 633, the triangles point to the membrane areas where the
Fig. 9. Enlarged section of Fig. S2 (white box). COS-7 cells 2 h after transfection with labelled lipoplexes. Following colours are chosen: lipid label (red), pDNA label YOYOÒ-1
iodide (green), membrane label Alexa FluorÒ 633 (white). The single presentations of the YOYOÒ-1-iodide fluorescence intensity (Image A), the lipid label fluorescence
intensity (Image B), and the Alexa FluorÒ 633 fluorescence intensity (Image C) are shown. Image D: multistained presentation of all three fluorescence intensities. Image E:
multistained presentation at three room directions: square image = x–y top view at a given z, rectangular image on top = y–z side view at a given x along the red line,
rectangular image on right site = x–z side view at a given y along the yellow line. The white triangles point to lipoplexes.

N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
359
lipoplexes were located (compare images A and B). A diffuse struc-
ture of the membrane was visualized here, which results from lipo-
plex membrane interactions and can be interpreted as a starting
invagination process of endocytosis. In contrast to this area the
remaining cell membrane is well defined (straight clear line).
Fig. 9E is a multi-stained image and shows 3 planes in x-, y-, and
z-direction. The image demonstrates that the lipoplex (white trian-
gle) is still outside the cell but in direct contact with the cell mem-
brane. Such cell membrane interaction with the lipoplex could be
observed on several locations in the sample (see Supporting infor-
mation, Fig. S2).
The image series in Fig. 10A1–A3 shows COS-7 cells 2 h after
transfection in three focus planes along the z-direction, in which
the triangles point to lipoplexes which interact with the cell mem-
brane (white triangles). The detailed image series B (Fig. 10B1–B4)
shows an enlarged section of image A1 (white box). The images in
Fig. 10B1–B4 show an early interaction of cytoplasmic protrusions
with the lipoplexes (interacting protrusion is indicated by blue
arrows Fig. 10A1, B3, and B4). The co-localization of DNA label
(green colour, Fig. 10B1) and lipid label (red colour, Fig. 10B2),
resulting in a yellow mix colour (Fig. 10B4), indicates that the
observed objects are lipoplexes. Fig. 10B3 and B4 clearly demon-
strates that these lipoplexes are co-localized with cell membrane
domains (grey arrows) attached to cytoplasmic protrusions.
Filopodia and lamellipodia in fibroblasts such as the COS-7 cells
carry out functions in cell migration, cell–cell and cell–substrate
interaction. They can form focal adhesion plaques in order to fix
particles to the cell surface before the cellular uptake starts. Such
interactions are also described in literature earlier by Zuhorn,
Hoekstra, and co-workers for nanoparticles [25,39]. Rehman et al.
could demonstrate that lipoplexes and polyplexes are captured
by actin-rich filopodial extensions, followed by a migration to
the cell surface. Afterwards, the nanoparticles were internalized
by endocytosis [39]. The mentioned interactions of T14diLys-
based lipoplexes with cytoplasmic protrusions indicate a compara-
ble mechanism of lipoplex uptake. Furthermore, Rehman et al.
discussed the role of syndecans (polyanionic heparan sulphate pro-
teoglycans) in the attachment of positively charged lipoplexes and
polyplexes to the filopodia [39]. The electrostatic interaction of
positively charged lipoplexes with negatively charged proteogly-
cans in the glycocalyx of cells is an established theory for the first
interaction step [36]. Also the T14diLys/DOPE (1/2, n/n) lipoplexes
at charge ratio 2.5 have a positive surface charge (Fig. 7b,
f-potential of 20 mV) and offer the possibility of electrostatic
interactions with the cell surface.
After 6 h the lipoplex uptake seems to be completed (Fig. 11A1–
A5). All lipoplexes are localized in the cell interior and no
lipoplexes could be visualized outside the cells (lipoplexes can be
Fig. 10. COS-7 cells 2 h after transfection with lipoplexes. Following colours are chosen: lipid label (red), pDNA label YOYOÒ-1 iodide (green), and membrane label Alexa
FluorÒ 633 (white). The images (A1–A3) show different focus planes along the z-direction (space between the planes 0.8
lm). The images (B1–B4) show an enlarged section of
(A1) (white box): B1 = single presentation of YOYOÒ-1-iodide fluorescence intensity, B2 = single presentation of the lipid label fluorescence intensity, B3 = single presentation
of Alexa FluorÒ 633 fluorescence intensity, and B4 = multistained presentation of all three fluorescence intensities.

360
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
Fig. 11. Fixed COS-7 cells 6 h (A series) and 24 h (B series) after transfection with lipoplexes. Following colours are chosen: lipid label (red), pDNA label YOYOÒ-1 iodide
(green), and membrane label Alexa FluorÒ 633 (white). Single presentations of the YOYOÒ-1-iodide fluorescence intensity (A1 and B1), the lipid label fluorescence intensity
(A2 and B2), and the Alexa FluorÒ 633 fluorescence intensity (A3 and B3) are shown. The images (A4 and B4) show multistained presentations of all three fluorescence
intensities. The images (A5 and B5) show multistained presentation at three room directions: square image = x–y top view at a given z, rectangular image on top = y–z side
view at a given x along the red (A5) or green (B5) line, rectangular image on right site = x–z side view at a given y along the yellow (A5) or red (B5) line.
detected by yellow colour in Fig. 11A4 and A5). The triangle in
image A1 points to a lipoplex localized in the cell interior. The
triangle in image A2 marks the distribution of lipid labelled struc-
tures (BFA-Rho6G intensity) across the cell interior (accept of the
nucleus). This indicates lipoplex disintegration processes with lipid
release. The lipid is localized in the whole cytoplasm. This is under-
lined in image 11A4 by the presence of yellow stained structures
(lipoplexes) and red areas (lipid label alone). The enlarged multi-
stained presentation at three room directions (Fig. 11A5) clearly
shows a lipoplex (white triangle), which is located inside the
COS-7 cell. In summary, the cellular uptake seems to be completed
6 h after lipoplex incubation.
would decrease the amount of translated protein per cell, but only
the degradation leads to a decrease of total expressed protein
(b-galactosidase) of all cells after transfection. Nevertheless, the
triangles in Fig. 11B1, B4, and B5 point to incorporated lipoplexes.
Also the background lipid fluorescence decreases in comparison
with the image series A taken 6 h after incubation (compare
Fig. 11A2 with B2). The images demonstrate that the lipoplex
disintegration is nearly finished. Nevertheless, it is too early for
effective expression of the internalized gene. After 48 h, the time
where the expression product was determined by the ONPG-
assay, no lipoplex structures could be determined in the cells (data
not shown).
After 24 h of lipoplex incubation the micrographs (Fig. 11, series
B) show rarely lipoplexes. Labelled DNA is rarely detected
(Fig. 11B1). This could be caused by the diluting effect during the
cell growth and also by degradation processes. Both processes
The effectivity of a cationic lipid formulation depends on
several factors. Varga et al. summarize the complex interplay of
vector adhesion, internalization, transport, and unpacking as well
as intracellular polynucleotide transport and processing in a

N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
361
kinetic model [27,40]. Also the above presented CLSM studies
investigate the lipoplex processing as the function of time. 2 h after
lipoplex incubation, lipoplexes are visualized in the stage of adhe-
sion and beginning internalization. The mechanism of lipoplex
uptake used by the T14diLys-lipoplexes cannot be determined
without the use of specific inhibitors of endocytotic mechanisms
and/or fluorescent labels for characteristic endocytotic vesicles
[41], but was out of scope in this study. Nevertheless, endocytosis
is discussed as major mechanism of lipoplex uptake [25,42]. The
lipoplex associated membrane invaginations in Fig. 9A–E and the
fact that the lipoplexes were not disintegrated by uncharged
membrane-mimicking lipids (DNA release experiments in the Sup-
porting information, Fig. S3), as well as the filopodia/lipoplex inter-
actions discussed above (Fig. 10), indicate that also T14diLys/DOPE
(1/2, n/n) lipoplexes at charge ratio 2.5 were internalized by endo-
cytotic pathways. In other cell lines and with other vector systems
the uptake was already completed after 2 h (as shown by Varga
et al. with LipofectamineTM in C3A cells [27]; and Rehman et al. with
PEI in HeLa cells [39]). We assume that the longer time of about 6 h
required for lipoplex uptake in the current study is due to the dif-
ferent types of cells used. It can be also assumed that lipoplexes
were released from endosomal/lysosomal compartments indicated
by the diffuse localization of lipid label in the whole cytoplasm
after 6 h. This assumption is also supported by the release experi-
ments with anionic endosomal membrane-mimicking vesicles at
pH 5.2. The incubation of T14diLys/DOPE (1/2, n/n) lipoplexes at
charge ratio 2.5 with anionic vesicles in acidic milieu resulted in
DNA release, while the lipoplexes were stable after incubation with
uncharged vesicles. This experiment gives evidence of a anionic
lipid induced lipoplex disintegration, resulting in DNA release from
endosomes into the cytoplasm as postulated by Xu and Szoka [43].
Also the above discussed fusion-inducing behaviour of DOPE,
which can adopt the inverse hexagonal phase, supports the theory
of endosome destabilization [44]. Therefore, it is possible to
assume that endosomal escape arises or is nearly finished 6 h after
lipoplex incubation (see Section 3.3.1). Although no tagged DNA
was observable in the nucleus after 24 h, the detected enhanced
activity of b-galactosidase after 48 h indicates a distinct gene
expression activity of the transfected marker gene and, hence, a
successful entering of the nucleus. However, more detailed studies
are required to get clear evidence for endosomal release and
nucleus entering.
values surpassing the standard LipofectamineTM at comparable via-
bility values. The in vitro transfection efficiency tests with respect
to serum stability demonstrate that the presence of 10% FBS during
the lipoplex incubation did not affect the transfection efficiencies
significantly. The size measurements of lipoplexes at different
charge ratios demonstrate that at the most effective lipoplex for-
mulations in the three screened cell lines the lipoplex dispersions
exhibit values suitable for in vivo experiments in combination with
positively charged lipoplexes. CLSM studies could clearly demon-
strate the efficient uptake of pDNA loaded lipoplexes. This process
is finished after 6 h while intact lipoplexes could be proved after
24 h. Effective expression of the reporter gene b-galactosidase
was achieved 48 h after lipoplex incubation.
First studies of the effect of serum on the transfection efficiency
as well as the lipoplex particle sizes with diameter below 1 lm
allow further in vivo application. Such studies will be performed
in the future.
Acknowledgements
This work was supported by the Deutsche Forschungsgemein-
schaft DFG (LSM: INST 271/250-1). We thank Mrs. Angela Munk
(Institute of Pharmacy, MLU Halle-Wittenberg, Germany) for
her excellent technical support in the cell culture. The support
of Dr. Gerd Hause (Biocenter, Martin-Luther-University Halle-
Wittenberg) by providing us access to the electron microscope
facility is greatly appreciated.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ejpb.2015.08.011.
References
[1] R.M. Blaese, K.W. Culver, A.D. Miller, C.S. Carter, T. Fleisher, M. Clerici, G.
Shearer, L. Chang, Y. Chiang, P. Tolstoshev, J.J. Greenblatt, S.A. Rosenberg, H.
Klein, M. Berger, C.A. Mullen, W.J. Ramsey, L. Muul, R.A. Morgan, W.F.
Anderson, T lymphocyte-directed gene therapy for ADAꢀSCID: initial trial
results after 4 years, Science 270 (1995) 475–480.
[2] S.L. Ginn, I.E. Alexander, M.L. Edelstein, M.R. Abedi, J. Wixon, Gene therapy
clinical trials worldwide to 2012 – an update, J. Gene Med. 15 (2013) (2012)
65–77.
[3] A. Aigner, Cellular delivery in vivo of siRNA-based therapeutics, Curr. Pharm.
Des. 14 (2008) 3603–3619.
[4] S. Zhang, D. Zhi, L. Huang, Lipid-based vectors for siRNA delivery, J. Drug
Target. 20 (2012) 724–735.
4. Conclusion
[5] E. Marshall, What to do when clear success comes with an unclear risk?,
Science 298 (2002) 510–511
[6] D. Ferber, Safer and virus-free?, Science 294 (2001) 1638–1642
[7] M.A. Mintzer, E.E. Simanek, Nonviral vectors for gene delivery, Chem. Rev. 109
(2008) 259–302.
[8] A.D. Miller, Cationic liposomes for gene therapy, Angew. Chem. Int. Ed. 37
(1998) 1768–1785.
[9] S. Bhattacharya, A. Bajaj, Advances in gene delivery through molecular design
of cationic lipids, Chem. Commun. (2009) 4632–4656.
[10] M. Heinze, G. Brezesinski, B. Dobner, A. Langner, Novel cationic lipids based on
malonic acid amides backbone: transfection efficacy and cell toxicity
properties, Bioconjug. Chem. 21 (2010) 696–708.
[11] H. Lv, S. Zhang, B. Wang, S. Cui, J. Yan, Toxicity of cationic lipids and cationic
polymers in gene delivery, J. Control. Release 114 (2006) 100–109.
[12] C. Wölk, S. Drescher, A. Meister, A. Blume, A. Langner, B. Dobner, General
synthesis and physicochemical characterisation of a series of peptide-mimic
lysine-based amino-functionalised lipids, Chem. – Eur. J. 19 (2013) 12824–
12838.
[13] D.E. Floyd, S.E. Miller, Synthesis of alkylmalonic esters by oxalate
condensations1, 2, J. Am. Chem. Soc. 69 (1947) 2354–2355.
[14] R.R. Schmidt, K. Jankowski, New types of nonionic surfactants with sugar head
groups, Liebigs Ann. 1996 (1996) 867–879.
[15] v.G. Bier, Über die Viskosität von Verzweigten verbindungen, Die Makromol.
Chem. 4 (1949) 124–133.
Structure/transfection relationships of the new class of the tris
(2-aminoethyl)amine-based
a-branched fatty acid amides were
generated. The screening in cell culture models demonstrates that
an introduction of two lysine molecules in the head group is
accompanied by a large increase in the transfection efficiency. Fur-
thermore, the shortening of the alkyl chains by two methylene
units at one or both chains (alkyl chain pattern T16 ? T14016 ?
T14) also increases the transfection efficiency. Nevertheless, the
most effective lipid composition in combination with a co-lipid
as well as the most effective charge ratio has to be determined
by screening methods. The most effective lipid composition in this
study, which shows good results in three cell lines (Hep-G2, A549,
COS-7), was the composition T14diLys/DOPE (1/2, n/n). In this
composition DOPE was the co-lipid of choice because of the more
cytotoxic potential of compositions with cholesterol. Although
the mixtures with DOPE also show an increase of cytotoxicity in
comparison with the pure cationic lipid, the DOPE mixture shows
high increasing effects on the transfection efficiency. Concluding,
[16] v.H. Staudinger, G. Bier, G. Lorentz, Viskositätsuntersuchungen an
niedermolekularen estern mit verzweigten molekülen, Die Makromol. Chem.
3 (1949) 251–280.
the use of the mixture with DOPE is
a compromise. The
T14diLys/DOPE 1/2 (n/n) formulation shows high transfection

362
N. Erdmann et al. / European Journal of Pharmaceutics and Biopharmaceutics 96 (2015) 349–362
[17] H. Staudinger, W. Kern, Über hochpolymere Verbindungen, 76. Mitteil. 1):
viscositäts-Messungen an Lösungen von Fadenmolekülen mit verzweigten
Ketten, Berichte deutschen chem. Gesellschaft (A and B Ser.) 66 (1933) 373–
378.
[31] C. Lonez, M. Vandenbranden, J.-M. Ruysschaert, Cationic lipids activate
intracellular signaling pathways, Adv. Drug Deliv. Rev. 64 (2012) 1749–1758.
[32] C. Lonez, M. Vandenbranden, J.-M. Ruysschaert, Cationic liposomal lipids: from
gene carriers to cell signaling, Prog. Lipid Res. 47 (2008) 340–347.
[18] M.N. Antipina, B. Dobner, O.V. Konovalov, V.L. Shapovalov, G. Brezesinski,
Investigation of the protonation state of novel cationic lipids designed for gene
transfection, J. Phys. Chem. B 111 (2007) 13845–13850.
[33] K.B. Knudsen, H. Northeved, P. Kumar Ek, A. Permin, T. Gjetting, T.L. Andresen,
S. Larsen, K.M. Wegener, J. Lykkesfeldt, K. Jantzen, S. Loft, P. Møller, M.
Roursgaard, In vivo toxicity of cationic micelles and liposomes, Nanomed.
Nanotechnol. Biol. Med. 11 (2015) 467–477.
[34] I.S. Zuhorn, W.H. Visser, U. Bakowsky, J.B.F.N. Engberts, D. Hoekstra,
Interference of serum with lipoplex–cell interaction: modulation of
[19] C.S. Wright, Structural comparison of the two distinct sugar binding sites in
wheat germ agglutinin isolectin II, J. Mol. Biol. 178 (1984) 91–104.
[20] C. Wölk, M. Heinze, P. Kreideweiß, M. Dittrich, G. Brezesinski, A. Langner, B.
Dobner, Synthesis and DNA transfection properties of new head group
modified malonic acid diamides, Int. J. Pharm. 409 (2011) 46–56.
[21] A. Lazik, Y. Liu, P. Bringas, F. Sangiorgi, R. Maxson, A sensitive method for
analyzing b-galactosidase reporter gene expression in tissue sections of mouse
embryos, Trends Genet. 12 (1996) 445–447.
[22] J.H. Miller, in: Experiments in Molecular Genetics, Cold Spring Harbor
Laboratory Press, New York, 1977.
[23] T. Mosmann, Rapid colorimetric assay for cellular growth and survival:
application to proliferation and cytotoxicity assays, J. Immunol. Methods 65
(1983) 55–63.
intracellular processing, Biochim. Biophys. Acta (BBA)
(2002) 25–36.
– Biomembr. 1560
[35] G. Caracciolo, D. Pozzi, R. Caminiti, H. Amenitsch, Formation of overcharged
cationic lipid/DNA complexes, Chem. Phys. Lett. 429 (2006) 250–254.
[36] I. Zuhorn, J.F.N. Engberts, D. Hoekstra, Gene delivery by cationic lipid vectors:
overcoming cellular barriers, Eur. Biophys. J. 36 (2007) 349–362.
[37] C. Wölk, C. Janich, D. Pawlowska, S. Drescher, A. Meister, G. Hause, B. Dobner,
A. Langner, G. Brezesinski, Lamellar versus micellar structures—aggregation
behavior of a three-chain cationic lipid designed for nonviral polynucleotide
transfer, ChemPhysChem 16 (2015) 2115–2126.
[24] J.N. Israelachvili, in: Intermolecular and Surface Forces, Academic Press, San
Diego, 2011, pp. 535–561.
[25] Z.u. Rehman, I.S. Zuhorn, D. Hoekstra, How cationic lipids transfer nucleic acids
into cells and across cellular membranes: recent advances, J. Control. Release
166 (2013) 46–56.
[38] C. Wölk, D. Pawlowska, S. Drescher, A. Auerswald, A. Meister, G. Hause, A.
Blume, A. Langner, G. Brezesinski, B. Dobner, New micellar transfection agents,
Langmuir 30 (2014) 4905–4915.
[39] Z.u. Rehman, K.A. Sjollema, J. Kuipers, D. Hoekstra, I.S. Zuhorn, Nonviral gene
delivery vectors use syndecan-dependent transport mechanisms in filopodia
to reach the cell surface, ACS Nano 6 (2012) 7521–7532.
[40] C.M. Varga, N.C. Tedford, M. Thomas, A.M. Klibanov, L.G. Griffith, D.A.
Lauffenburger, Quantitative comparison of polyethylenimine formulations
and adenoviral vectors in terms of intracellular gene delivery processes, Gene
Ther. 12 (2005) 1023–1032.
[41] G. Sahay, D.Y. Alakhova, A.V. Kabanov, Endocytosis of nanomedicines, J.
Control. Release 145 (2010) 182–195.
[42] Z.u. Rehman, D. Hoekstra, I.S. Zuhorn, Protein kinase A inhibition modulates
the intracellular routing of gene delivery vehicles in HeLa cells, leading to
productive transfection, J. Control. Release 156 (2011) 76–84.
[43] Y. Xu, F.C. Szoka, Mechanism of DNA release from cationic liposome/DNA
complexes used in cell transfection, Biochemistry 35 (1996) 5616–5623.
[44] I. Koltover, T. Salditt, J.O. Rädler, C.R. Safinya, An inverted hexagonal phase of
cationic liposome-DNA complexes related to DNA release and delivery, Science
281 (1998) 78–81.
[26] M. Thomas, A.M. Klibanov, Non-viral gene therapy: polycation-mediated DNA
delivery, Appl. Microbiol. Biotechnol. 62 (2003) 27–34.
[27] C.M. Varga, K. Hong, D.A. Lauffenburger, Quantitative analysis of synthetic
gene delivery vector design properties, Mol. Ther. 4 (2001) 438–446.
[28] A. Ahmad, H.M. Evans, K. Ewert, C.X. George, C.E. Samuel, C.R. Safinya, New
multivalent cationic lipids reveal bell curve for transfection efficiency versus
membrane charge density: lipid–DNA complexes for gene delivery, J. Gene
Med. 7 (2005) 739–748.
[29] I.S. Zuhorn, U. Bakowsky, E. Polushkin, W.H. Visser, M.C.A. Stuart, J.B.F.N.
Engberts, D. Hoekstra, Nonbilayer phase of lipoplex-membrane mixture
determines endosomal escape of genetic cargo and transfection efficiency,
Mol. Ther. 11 (2005) 801–810.
[30] B. Ma, S. Zhang, H. Jiang, B. Zhao, H. Lv, Lipoplex morphologies and their
influences on transfection efficiency in gene delivery, J. Control. Release 123
(2007) 184–194.