Pyridylthiourea-grafted polyethylenimine offers an effective assistance to siRNA-mediated gene silencing in vitro and in vivo

Article abstract of DOI:10.1016/j.jconrel.2011.10.007

Success of synthetic interfering nucleic acids (siRNAs)-based therapy relies almost exclusively on effective, safe and preferably nanometric delivery systems which can be easily prepared, even at high concentrations. We prepared by chemical synthesis vari

Full text of DOI:10.1016/j.jconrel.2011.10.007

Journal of Controlled Release 157 (2012) 418–426
Contents lists available at SciVerse ScienceDirect
Journal of Controlled Release
journal homepage: www.elsevier.com/locate/jconrel
Pyridylthiourea-grafted polyethylenimine offers an effective assistance to
siRNA-mediated gene silencing in vitro and in vivo
Gaelle Creusat a, Jean-Sébastien Thomann a, Anne Maglott b, Bénédicte Pons c, Monique Dontenwill b,
Eric Guérin d, Benoit Frisch a, Guy Zuber a,
⁎
a
Laboratoire de Conception et Application de Molécules Bioactives, CNRS — Université de Strasbourg UMR 7199, Faculté de Pharmacie, 74, route du rhin, 67400 Illkirch, France
Laboratoire de Biophotonique et Pharmacologie, CNRS — Université de Strasbourg UMR7213, Faculté de Pharmacie, 74, route du rhin, 67400 Illkirch, France
Laboratoire d'Innovation Thérapeutique CNRS — Université de Strasbourg UMR 7200, Faculté de Pharmacie, 74, route du rhin, 67400 Illkirch, France
Physiopathologie et Recherche Translationnelle, EA 4438, Université de Strasbourg, 3 avenue Molière, 67200 Strasbourg, France
b
c
d
a r t i c l e i n f o
a b s t r a c t
Article history:
Success of synthetic interfering nucleic acids (siRNAs)-based therapy relies almost exclusively on effective,
safe and preferably nanometric delivery systems which can be easily prepared, even at high concentrations.
We prepared by chemical synthesis various self-assembling polymers to entrap siRNAs into stable poly-
plexes outside cells but with a disassembly potential upon sensing endosomal acidity. Our results revealed
that pyridylthiourea-grafted polyethylenimine (πPΕΙ) followed the above-mentioned principles. It led to
above 90% siRNA-mediated gene silencing in vitro on U87 cells at 10 nM siRNA concentration and did not
have a hemolytic activity. Assembly of siRNA/πPΕΙ at high concentration was then studied and 4.5% glucose
solution, pH 6.0, yielded stable colloidal solutions with sizes slightly below 100 nm for several hours. A sin-
gle injection of these concentrated siRNA polyplexes into luciferase-expressing human glioblastoma tu-
mors, which were subcutaneously xenografted into nude mice, led to a significant 30% siRNA-mediated
luciferase gene silencing 4 days post-injection. Our results altogether substantiate the potential of self-
assembling cationic polymers with a pH-sensitive disassembly switch for siRNA delivery in vitro and also
in vivo experiments.
Received 16 July 2011
Accepted 6 October 2011
Available online 12 October 2011
Keywords:
siRNA delivery
Nanovector
Transfection
Polymer
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
delivery [14]. Encouraging siRNA-mediated gene silencing results were
obtained in vivo with low molecular weight branched PEI [8,15] despite
Small interfering RNA duplex (siRNA) generates sequence-selective
mRNA degradation, resulting in an effective silencing of the targeted
gene [1]. This gene-silencing mechanism is of great value for analyzing
the function of a single gene in cultured cells and holds great potential
as gene-specific therapeutics [2]. High pressure tail vein injection of siR-
NAs (hydrodynamic delivery method) enables translocation of nucleic
acids into the liver cells and provided the first demonstrations of the
therapeutic potential of siRNAs [3,4]. Unfortunately, this method is too
risky to be used on humans, especially on unhealthy individuals. Fur-
ther progress has been achieved by formulating siRNAs within lipidic
vehicles either for delivery to the liver [5] or to lung cancer cells [6]. Cat-
ionic polymers have also been used for siRNA complexation and in vivo
delivery. Promising results were obtained for tumor therapy either
using simple cationic systems [7–9], PEG-protected [10] or targeted
multifunctional ones [11,12].
the weak cohesion of siRNA/PEI polyplexes in presence of cell surfaces
[16]. At an in vitro cellular level, successful siRNA delivery with PEI
showed to rely on high siRNA dose, on a branched polymer and on spe-
cific polyplex formation conditions [17]. Strategies to increase the stabil-
ity of the siRNA polyplexes were developed [18]. One is to modify the
polycation backbone with hydrophobic elements [9,19]. In this vein,
we recently reported that modification of the “proton-sponge” polyethy-
lenimine (PEI) 25 kDa with tyrosine led to a self-assembling polymer
(PEIY) with effective siRNA delivery abilities at a cellular level [20]. Be-
sides providing increased extracellular stability, non-covalent interac-
tions have the advantage of potential reversibility. A further detailed
investigation indicated that mildly acidic conditions (pH 6.0), such as
ones encountered in PEI-buffered endosomes [21], weaken the non-
covalent tyrosine–tyrosine interactions and may facilitate release of
siRNA at the right time for optimal delivery [22]. However, effective
siRNA-mediated gene silencing was observed with large polymolecular
aggregates (>500–800 nm) [20]. This feature asks question about suit-
ability of such polyplexes for in vivo administration that relies on prepa-
ration of concentrated and stable colloids of sizes around 100 nm. We
also found tyrosine-modified PEI to have a toxicological profile unsuited
for in vivo administration.
The polyethylenimine (PEI) is one of the most studied cationic DNA
transfection reagents [13] and has been evaluated as well for siRNA
⁎
Corresponding author. Fax: +33 3 68 85 43 06.
E-mail address: zuber@unistra.fr (G. Zuber).
0168-3659/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jconrel.2011.10.007

G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
419
In this study, we tuned the chemical structure of self-assembling
polymers for in vivo administration and focused our efforts to improve
the biocompatibility of the siRNA carrier. We synthesized various
novel hydrophobic polyethylenimines starting from the commercial
branched polymer of 25 kDa. We also evaluated their siRNA delivery ef-
ficiency and their toxicological profile and selected N-3-pyridyl, N′-PEI
thiourea (πPEI) as the most suitable siRNA vehicle for in vivo adminis-
tration. Single injection of siRNA polyplexes in subcutaneous implanted
U87 tumors resulted in a significant siRNA-mediated luciferase silenc-
ing at about 30% relative to controls.
NaHCO3 (twice 200 mL), citric acid 5% (200 mL), saturated NaCl
(200 mL), dried over MgSO4 and the solvent was removed under re-
duced pressure to give the product as a yellow solid (10.1 g, 95%
yield). 1H NMR (CDCl3) δppm: 2.9 (s, 4H), 6.95 (t, J=7.2 Hz, 1H), 7.05
(d, J=8.0 Hz, 1H), 7.58 (td, J=7.2 Hz, J=1.6 Hz, 1H), 8.0 (dd,
J=8.0 Hz, J=1.6 Hz, 1H). 13C NMR (CDCl3) δppm: 25.6, 108.1, 118.0,
120.0, 130.1, 137.9, 161.9, 165.0, 169.1. ES-MS: (M calculated for
C
11H9NO5: 235.048) found: 258.038 ([MNa]+).
2.4. Succinimidyl ester of nicotinic acid
2. Materials and methods
A solution of DCC (7.0 g, 33.98 mmol) in CH2Cl2 (10 mL) was added
dropwise at 0–4 °C to a solution of N-hydroxysuccinimide (4.1 g,
35.6 mmol) and nicotinic acid (3.89 g, 31.6 mmol) in DMF (30 mL).
The reaction mixture was then stirred overnight at room temperature
and the DCU was removed by filtration and washed with ethyl acetate
(200 mL). The combined organic phases were washed with saturated
NaHCO3 (twice 200 mL), citric acid 5% (200 mL), saturated NaCl
(200 mL), dried over MgSO4 and the solvent was removed under re-
duced pressure to give the product as a white solid (5.7 g, 80% yield).
1H NMR (CDCl3) δppm: 2.8 (s, 4H), 7.2–7.43 (m, 1H), 8.31–8.35 (m,
1H), 8.82–8.84 (m, 1H), 9.26 (d, J=1.6 Hz). 13C NMR (CDCl3) δppm:
25.7, 121.7, 123.6, 137.8, 151.4, 155.2, 160.8, 168.9. ES-MS: (M calculated
for C10H8N2O4: 220.049) found: 221.056 ([MH]+).
2.1. Materials
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide
(MTT) was purchased from Invitrogen (Cergy Pontoise, France).
Branched PEI 25 kDa (40,872-7, batch 09529KD-466) and the other che-
micals were purchased from Sigma-Aldrich (St Quentin, France) and
were used as supplied. Water was deionized on a Millipore Milli-Q appa-
ratus. Before use, regenerated cellulose dialysis membranes (SpectraPor
4, 12–14 kD, SpectrumLabs) were soaked in MilliQ water (200 mL, 3
times, 8 h each) to remove preservatives. Chemical synthesis and
work-ups were performed under a chemical fume hood and plastic
tubes were guaranteed RNAse-free by the manufacturers. Experiments
involving cell lines were performed according to the biosafety level 2
guidance. NMRI nude mice (strain Crl:NMRI-Foxn1nu) were obtained
from Elevage Janvier, Le Genest-St-Isle, France. Animal experimentation
was conducted according to French regulations at the Inserm U682 ani-
mal facility. HEPES buffered saline ×2 (HBS) contained 20 mM HEPES,
240 mM NaCl and 1.5 mM Na2HPO4. Buffered solutions and water were
sterilized by filtration through 0.22 μm pore membrane. The polymers
and siRNA solutions were prepared using sterile media. All solutions
were kept sterile by working under a class II microbiological safety cab-
inet. UV/Vis analyses were performed on a Shimadzu UV2401PC spec-
trometer. NMR spectra were performed on a Bruker DPX 400 MHz
spectrometer. The modification degree of the polymer was determined
relative to ethylenimine (EI) residues by integration of 1H NMR signals
and was of 30+/−5% for all polymer otherwise indicated.
2.5. N-3-pyridyl-, N′-PEI–thiourea (πPEI) 1
A solution of 3-pyridyl isothiocyanate (460.4 mg; 3.46 mmol) in
CH2Cl2 (40 mL) was added dropwise at room temperature to a solution
of PEI (500 mg; 11.62 mmol) in CH2Cl2 (40 mL). After 30 min, the sol-
vent was evaporated under reduced pressure. The residue was dissolved
in water (40 mL) and the solution was adjusted to pH 4.0 by addition of
hydrochloric acid 3 M. Dialysis against water (1 L volume; 2 changes
over 48 h) and freeze-drying gave 650 mg of the pyridyl PEI–thiourea.
The modification degree was estimated at 28% relative to ethylenimine.
1H NMR (D2O) ∂ppm: 3.95–2.5 (m, 4H, NHCH2CH2NH), 4.2 (t, 0.5H,
CH2NCS), 7.5 (m, 0.25H, CHaro), 7.86 (m, 0.25H, CHaro), 8.45 (m, 0.5H,
CHaro). λmax (ε calculated for ethylenimine unit): 245 nm
(2660 M−1.cm−1). Average molecular weight (MW): 122.0 g/mol.
2.2. Succinimidyl ester of vanillic acid
2.6. N-3-pyridyl-, N′-PEI–thiourea 15% 2
A
solution of N, N′-DiCyclohexylCarbodiimide (DCC) (20 g,
A solution of 3-pyridyl isothiocyanate (230 mg; 1.73 mmol) in
CH2Cl2 (50 mL) was added dropwise at room temperature to a solu-
tion of branched polyethylenimine (500 mg; 11.62 mmol) in CH2Cl2
(40 mL). After 30 min, TLC indicated full consumption of the isothio-
cyanate. The solvent was then evaporated under reduced pressure.
The residue was dissolved in water (40 mL) and the solution was ad-
justed to pH 4.0 by addition of hydrochloric acid 3 M. Dialysis against
water (1 L volume; 2 changes over 48 h) and freeze-drying gave
515 mg of pyridyl PEI-thiourea 15%. 1H NMR (D2O) ∂ppm: 4.02–
2.65 (m, 4H, NHCH2CH2NH) 4.24 (t, 0.3H, CH2NCS), 7.82 (m, 0.15 H,
CHaro), 9.1–8.1 (m, 0.3H, CHaro). MW: 100 g/mol.
97 mmol) in CH2Cl2 (25 mL) was added dropwise at 0–4 °C to a solution
of N-hydroxysuccinimide (8.9 g, 77 mmol) and vanillic acid (11.8 g,
70 mmol) in ethylacetate (80 mL)/DMF (20 mL). The reaction mixture
was then stirred overnight at room temperature. The DiCyclohexylUrea
(DCU) was removed by filtration, washed with ethyl acetate (100 mL)
and the combined organic phases were washed with saturated NaCl
(100 mL), saturated NaHCO3 (twice 100 mL), saturated NaCl (100 mL)
and then dried over MgSO4. The solvent was then removed under re-
duced pressure to give the product as a yellow solid (16.5 g; 70%
yield). 1H NMR (CDCl3) δppm: 2.9 (s, 4H), 3.9 (s, 3H), 6.9 (d,
J=8.4 Hz, 1H), 7.55 (d, J=2.0 Hz, 1H), 7.75 (dd, J=2.0 Hz, J=8.4 Hz,
1H), 8.0 (s, 1H). 13C NMR (CDCl3) δppm: 25.7, 56.2, 112.4, 114.8,
116.5, 125.9, 146.7, 152.2, 161.5, 169.5. ES-MS: (M calculated for
2.7. N-4-aminophenyl, N′-PEI–thiourea 3
C
12H11NO6: 265.057) found: 288.049 ([MNa]+).
A solution of tert-butyl 4-isothiocyanatophenylcarbamate (500 mg;
2.02 mmol) in CH2Cl2 (50 mL) was added dropwise to a solution of PEI
(265 mg; 6.16 mmol) in CH2Cl2 (40 mL). The reaction was then stirred
for 30 min at room temperature and the solvent was removed by evapo-
ration under reduced pressure. The residue was dissolved in aqueous HCl
3 M (30 mL) and the solution was carefully adjusted to pH 4.0 with aque-
ous NaOH 6 M. Dialysis against water (2 changes over 24 h) and freeze
drying provided N-(4-aminophenyl), N′-PEI–thiourea (650 mg) as a yel-
low powder. 1H NMR (D2O) ∂ppm: 4.0–2.67 (m, 3.3H, NHCH2CH2NH),
4.21 (m, 0.7H, CH2NCS), 7.35 (bm, 1.4H, CHaro). Average MW: 140 g/mol.
2.3. Succinimidyl ester of salicylic acid
A solution of DCC (10.9 g, 53 mmol) in CH2Cl2 (5 mL) was added
dropwise at 0–4 °C to a solution of N-hydroxysuccinimide (6.19 g,
53.8 mmol) and salicylic acid (6.16 g, 44.7 mmol) in DMF (30 mL).
The reaction mixture was then stirred overnight at room temperature
and the DCU was removed by filtration and washed with ethyl acetate
(200 mL). The combined organic phases were washed with saturated

420
G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
2.8. 4-hydroxybenzamido-polyethylenimine 4
A. Untargeted RNA duplex (sic) was of sequence 5′-CGU ACG CGG AAU
ACU UCG A. Polymers were dissolved in water and the solutions were ad-
justed to pH 6.0 with NaOH 1 M. Solutions were stored at 4 °C. Cells were
maintained at 37 °C in a 5% CO2 humidified atmosphere and grown in
DMEM medium with 10% Fetal bovine serum (FBS) (Perbio, Brebières,
France), 100 units/mL penicillin, 100 μg/mL streptomycin and 2 mM L-
glutamine (Eurobio, Courtaboeuf, France). U87 cells (human glioblastoma
ATCC HTB-14) were transformed to stably express the Photinus pyralis
luciferase-enhanced green fluorescence protein fusion gene originating
from the pEGFPluc plasmid (Clontech, Mountain View, CA) [23]. The plas-
mid coded as well for a resistance gene to G418, selection was thus per-
formed by adding to a final concentration of 0.8 μg/mL to the cell
culture medium. The day before experiments U87egfpluc cells were seed-
ed into 96-well plates at a density of 5000 cells per well in 100 μL cell cul-
ture medium.
A solution of freshly prepared 4-hydroxybenzoic acid (635 mg;
4.6 mmol) and BOP (1.7 g; 5.06 mmol) in DMF (20 mL) was added drop-
wise at room temperature to a solution of PEI (1 g, 23.2 mmol) in DMF
(50 mL). After 2 h under stirring, DMF was removed by evaporation
under reduced pressure. The crude product was taken in water, dissolved
by addition of aqueous sodium hydroxide solution 1 M (pH 11) and sub-
jected to dialysis against water (1 L, 2 changes over 24 h). Lyophilization
provided 4-hydroxybenzamide-PEI (0.7 g) at a modification degree of
28%. 1H NMR (D2O) ∂ppm: 2.6 (bm, 2.9H, \NHCH2CH2NH\), 3.22 (m,
0.55H, Phe-CONHCH2CH2NH\), 3.35 (m, 0.55H, Phe-CONHCH2CH2NH)
6.57 (d, J=7.3 Hz, 0.55H, CHaro), 6.97 (m, 0.55H, CHaro).
2.9. Vanillamido-polyethylenimine 5
A solution of succinimidyl ester of vanillic acid (2.8 g, 10.5 mmol) in
DMF (20 mL) was added at room temperature to a solution of polyethy-
lenimine (0.9 g, 20.9 mmol in ethylenimine) in methanol (5 mL). After
2 days under stirring, the residue was treated with NaOH 1 M (5 mL)
for 1 h at room temperature. The pH of the solution was then adjusted
to pH 7.0 by addition of aqueous HCl 0.5 M and the solution was sub-
jected to dialysis against water (1 L, 5 changes over a 48 h period). Lyoph-
ilization afforded the product as yellow powder (1.4 g, 67% yield). 1H
NMR (D2O) δppm: 2.0–3.9 (16H), 6.6 (s br, 1H), 6.9–7.4 (m, 2H). 13C
NMR (D2O) δppm: 36.6, 49.9, 56.0, 115.0, 121.0, 146.9, 150.3. λmax (ε cal-
culated for ethylenimine unit): 290 nm (2000 M−1.cm−1), 260 nm
(4000 M−1.cm−1). Average MW: 127 g/mol.
2.13. Determination of polyplexes stability
Polyplexes (EI/P of 60) were prepared by mixing the control siRNA
(735 ng, 50 pmol) and each polymer (120 nmol in ethylenimine) ei-
ther in RPMI (15 μL, final pH of 7.8) or in water (15 μL, final pH of
6.0). After 20 min of incubation, the polyplexes were treated with in-
creasing charge excess of heparin, for 30 min, loaded onto a 1.5% aga-
rose gel containing 1 mM EDTA and 40 mM Tris acetate buffer pH 8.0,
and subjected to electrophoresis for 30 min at 90 V. After staining
with ethidium bromide solution (0.5 μg/mL, 15 min), released siRNA
were visualized with a UV transilluminator and quantified using
NIH ImageJ analysis software.
2.10. 2-hydroxybenzamido-polyethylenimine 6
2.14. Acido-basic titration of the polymer
A solution of succinimidyl ester of salicylic acid (1.47 g, 6.2 mmol) in
DMF (15 mL) was added at room temperature to a solution of polyethy-
lenimine (0.9 g, 20.9 mmol in ethylenimine) in CH2Cl2 (3 mL). Two
hours later, the reaction mixture was completed with methanol
(20 mL) and stirred for 24 h. The solvents were removed under reduced
pressure and the residue was treated with NaOH 1 M (20 mL). The solu-
tion was subjected to dialysis against water (1 L, 2 changes over a 24 h
period), aqueous HCl (1 L, 6 changes over a 48 h period) and water (1 L,
once for 6 h). Lyophilization afforded the product as a powder (1.8 g,
86% yield). 1H NMR (D2O) δppm: 2.0–3.9 (13.2 H), 6.7–7.6 (m, 4H).
λmax (ε calculated for ethylenimine unit): 300 nm (790 M−1.cm−1).
Average molecular weight: 117 g/mol.
The buffering capacities of the polymers were determined using
an automatic titrator (Abu 901, radiometer, Copenhagen) and a Met-
tler Toledo Inlab423 glass electrode. Two mL-aqueous solutions of the
polymer (50 mM amino nitrogen atoms) containing 100 mM KCl
were titrated with 0.1 M NaOH. The pH-sensible aggregation ability
of the polymers was determined by monitoring the turbidity and
measuring the pH of a 20 mM aqueous solution of the polymer during
an acido basic titration.
2.15. Determination of the polymer toxicity (MTT assay)
MTT assays were performed in triplicate in 96-well plates. Cells
(U87efgpluc) were seeded the night before experiment at 8000 cells/
well. Different amounts of polymers in a final volume of 11 μL water,
were added to the cells to obtain the final concentrations as indicated
in the graph. After 48 h incubation, the cell culture medium was removed
from each well and replaced with a 0.5 mg/mL MTT solution in DMEM
without serum (220 μL). After 2 h incubation at 37 °C, excess reagent
was removed by aspiration. The formazan crystals were then dissolved
in DMSO (100 μL) and quantified spectrophotometrically in a microplate
reader at a wavelength of 570 nm. The cell viability was plotted relative
to untreated cells that were grown the same day in the same plate.
2.11. Nicotinamido-polyethylenimine 7
A solution of succinimidyl ester of nicotinic acid (3.52 g, 16 mmol) in
DMF (15 mL) was added at room temperature to a solution of polyethy-
lenimine (1.4 g, 32 mmol in ethylenimine) in CH2Cl2 (10 mL). Two hours
later, the reaction mixture was completed with methanol (20 mL) and
stirred for 24 h. Solvents were removed under reduced pressure and
the residue was treated with NaOH 1 M (20 mL). The solution was sub-
jected to dialysis against water (1 L, 2 changes over a 24 h period), aque-
ous HCl (1 L, 6 changes over a 48 h period) and water (1 L, once for 6 h).
Lyophilization afforded the product as a powder (2.2 g, 75% yield). 1H
NMR (D2O) δppm: 2.6–3.9 (10.1 H),7.1–9.0 (m, 4H). λmax (ε calculated
for ethylenimine unit): 260 nm (1800 M−1.cm−1). Average MW:
112 g/mol.
2.16. Hemolysis experiments
Before experiments, sheep red blood cells (RBC) (Eurobio, Courta-
boeuf, France) were recovered by centrifugation at 400 RCF for 10 min
and washed three times with NaCl aqueous solution (150 mM). RBC
were then resuspended in phosphate buffer saline (PBS) and plated in
96-well plates at a rate of 15×106 cells in 50 μL. Solutions of polymer
(50 μL) in PBS at different concentrations were then added to the eryth-
rocytes and incubated for 1 h at 37 °C. The release of hemoglobin was
determined after centrifugation at 700 RCF for 10 min by spectrophoto-
metric analysis of the supernatant at 550 nm. Complete hemolysis (100%
control value) was achieved using TritonX-100 to a final concentration
2.12. Materials for the siRNA delivery experiments
PAGE-purified oligonucleotides terminated with two 2′-deoxythymi-
dines at their 3′-ends were purchased from Eurogentec (Seraing, Bel-
gique), supplied at 100 μM concentration and stored at −20 °C. The
egfpluc fusion gene silencing experiments were performed with a RNA
duplex of the sense sequence (siluc): 5′-CUU ACG CUG AGU ACU UCG

G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
421
of 0.1% w/v. The negative control was defined by suspension of RBC in
phosphate buffer alone. The experiments were performed in triplicate.
grinded with a T25 ultra-turrax at 0–4 °C. The solutions were then clari-
fied by centrifugation (10,000 RCF, 10 min) and luciferase activity of the
supernatant was measured and expressed as relative light units (RLU) in-
tegrated over 10 s and normalized per μg of tumor protein by using the
BCA assay (Pierce, Brebières, France). Each condition was performed on
group of 5 tumors. Statistical analysis (Student t-test) and calculation of
the t probability (p) were performed using KaleidaGraph Software 4.1
on paired data.
2.17. Formation of siRNA polyplexes
First procedure (assembly of complex at EI/P of 60 at low concentra-
tions — typical for in vitro experiment): a 100 nM solution of siRNA in
RPMI medium without serum (40 μL) was added to a 0.5 mL polypropyl-
ene tube containing an aqueous solution of the polymer (10 mM in ethy-
lenimine, 0.96 μL). The mixture was then homogenized and used within
2 h. Second procedure (assembly of complex at high concentrations —
for in vivo experiment): typically for EI/P of 9.4, polyplexes were formed
at room temperature by mixing a 100 μM siRNA solution (80 μL) with a
9% (w/v) glucose solution (100 μL). This mixture was then rapidly
added to 20 μL of a 150 mM (18.3 mg/mL) πPEI solution, pH 6.0. The
same procedure was applied using other medium by replacing the glu-
cose solution with 300 mM NaCl, HEPES buffered saline (HBS) ×2, pH
7.0 and pH 7.8 or lowering the initial siRNA solution to 15.6 μM.
3. Results and discussion
3.1. Polymer synthesis and initial in vitro siRNA delivery activities
Most synthetic delivery systems use initial electrostatic anchorage to
sulfated (polyanionic) proteoglycans of the external cell surfaces to accu-
mulate in large amounts into endosomes. The use of this route relies on
cationic complexes stable enough to sustain electrostatic competition
with the external polyanionic cell surface receptors but capable of disso-
ciating inside the cell for payload delivery [25,26]. After cell-anchorage to
anionic proteoglycans, polyplexes may be directed into endosomal com-
partments and sense a pH-decrease from pH 7.4 to 4.5. It has been dem-
onstrated that PEI polyplexes interfere with the normal acidification
process [27]. The buffering ability of PEI blocks the cell-induced endoso-
mal acidification at measured pH of 5.9 [21] or 6.1 [27], leading eventu-
ally to a rupture of the endosome membrane and subsequent access of
the polyplexes to the cytosol. Chemical modification of the water-
soluble 25 kDa polyethylenimine (PEI) with tyrosine improves consider-
ably its oligonucleotide delivery abilities at a cellular level [20,28]. Our
mechanistic investigation [22] indicated that the efficiency of PEIY may
be due to an optimal hydrophobic/hydrophilic balance. We therefore
attempted to provide more ground to this assumption as well as to im-
prove the overall siRNA delivery properties of the delivery agent by im-
proving the toxicological profile. At a starting point, we explored
variations in the chemical structure of the PEI-grafted hydrophobic do-
mains and in the type of linkage for possible improvement of delivery ac-
tivity and/or toxicological profile (Fig. 1A).
2.18. SiRNA delivery experiments in vitro
For polyplexes prepared according to the first procedure, the com-
plexes (11 μL containing 16.17 ng siRNA, EI/P of 60) were directly
added to cells by dilution into each well of the 96-well plate. Polyplexes
prepared according to the second procedure were diluted at a 1/50 ratio
in 10% FBS cell culture medium. Diluted complexes (10 μL) were then
added to the cells. Cells were then let to grow in the incubator without
further handling. Luciferase gene expression was usually assessed 48 h
after delivery using a commercial kit using manufacturer's protocol
(Promega, Charbonnières, France) using 20 μL of lysis buffer. The lumi-
nescence was measured from 0.8 μL of lysate during 1 s with a lumin-
ometer (Centro LB960 XS; Berthold, Thoiry, France). The error bars
represent standard deviation derived from triplicate experiments and
luciferase activity was calculated relative to untreated cells.
2.19. Estimation of complex size
The commercial branched 25 kDa PEI was chosen as the core cat-
ionic element because it readily accepts chemical modification [29].
The various hydrophobic elements were grafted to this cationic
core at a ratio of 30% relative to ethylenimine unit (EI) unless indicat-
ed. In principle, this value corresponds to a full modification of the
PEI primary amines [30]. The first polymer, named πPEI 1 was pre-
pared by reaction with 3-pyridyl-isothiocyanate (0.3 equivalent rel-
ative to PEI ethylenimine unit) in a DMF/dichloromethane. The effect
of the modification degree was investigated by synthesizing a second
polymer, πPEI 2, with a grafting of 15%. This polymer was prepared
using the same protocol but with 0.15 equivalent of isothiocyanate
relative to PEI. Polymer 3 was made from N-tert-butyloxycarbonyl-
4-aminophenylisothiocyanate and obtained after removal of the
protecting group with trifluoroacetic acid. Polymer 4 was prepared
from 4-hydroxybenzoic acid using benzotriazole-1-yl-oxy-tris-(dimethy-
lamino)-phosphonium hexafluorophosphate (BOP) as the condensing
reagent. Finally, polymers 5, 6 and 7 were prepared by reacting PEI
with succinimidyl esters of vanillic acid, salicylic acid and nicotinic acid,
respectively, also in DMF/dichloromethane. After completion of the reac-
tions, each polymer was purified by dialysis and isolated as hydrochloride
salts. Yields were typically in the 50–60% range and 1H NMR integration
of characteristic peaks confirmed the modification degree to be close to
the expected values of 15% for 2 and 30% for all the others. The ability
of each polymer to make up electrostatic complexes with oligonucleo-
tides was then assessed by agarose gel electrophoresis. Regardless of
the modification, the onset of full complexation happens at EI to oligonu-
cleotide phosphate (P) ratio of 3.0 (Fig. S2, Supporting information).
siRNA duplexes can effectively and selectively silence the expression
of a gene only if the duplexes can be transferred into the cell. The deliv-
ery abilities of the polymer were hence directly deducted by measuring
After assembly, the complexes (40 μL) were diluted in water (0.5 mL).
Dynamic light scattering measurements were performed using a NanoZS
apparatus (Malvern instruments, Paris, France) at 25 °C using a refractive
index of particles of 1.49. Data were analyzed using the multimodal num-
ber distribution software included with the instrument. Diameters of
complexes were calculated from the intensity of the signal. Transmission
electron microscopy was performed on a Philips CM120 apparatus [24].
EM samples were prepared by transferring polyplexes onto carbon
grids (Ted Pella 1822-F) and were stained with uranyl acetate.
2.20. SiRNA delivery experiments in vivo
Particles were freshly prepared according to the second procedure
and were injected within 60 min. Male NMRI nude mice (6–8 weeks of
age, weights in the 25–30 g) were subcutaneously inoculated in both
flanks with 5×106 U87egfpluc cells in 100 μL PBS. Mouse behavior and
weight were monitored every 2 to 4 days and tumor growth was evaluat-
ed with a digital caliper. Tumor volume was calculated using the formula
[(length×width2)×0.52]. Twenty-five days after inoculation, mice were
anesthetized using isoflurane and complexes (40 μL) were injected into
the tumor mass using a 0.3 mL syringe equipped with a 30 G needle
(BD MicroFine, Becton Dickinson, Franklin Lakes, NJ, USA). Average vol-
ume of the tumors was 260 mm3. Mice were anesthetized again 4 days
later. The volume of tumors was measured and the animals were sacri-
ficed by cervical dislocation. The tumors were dissected, washed in PBS
and cut in two pieces. One piece was fixed in PFA for histological analysis.
The other piece was wiped up with an adsorbing paper and freeze-dried
in liquid nitrogen. For luciferase activity measurement, the freeze-dried
tumor was weighted, diluted in cold PBS at 80 mg of tumor/mL and

422
G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
A
(Supporting information) showed that the polymers gave a similar
siRNA delivery efficiency trend on the A549 cell line.
R
N
NH2
HN
S
3
4
R=
R=
N
H
3.2. Physicochemical properties of the polymers
O
O
O
N
H
Cationic delivery systems usually do not promote nucleic acid trans-
location by direct action on the plasma membrane. They rather divert
the cell machinery and enter into endosomal compartments in where
they undergo a pH change from 7.8 to as low as 4.5 within minutes. It
is generally assumed that the high buffering abilities (the so-called
“proton-sponge” properties) of PEI in that range contribute strongly to
its nucleic acid translocation efficiency by promoting an osmotic pres-
sure imbalance in between PEI-loaded endosomes and the cytosol, lead-
ing eventually to the rupture of endosomal lipid membrane and hence
to the desired translocation [21].
2 HCl
180
PEI
OH
O
O
5
6
R=
R=
R=
OH
NH2
OH
N
PEIY
HO
O
S
1 or 2
R=
πPEI
N
H
7
R=
N
We examined whether the observed difference in siRNA delivery
properties between the effective polymers (1, 3, 4, 5, 6) and the poorly
effective ones (PEI, 2 and 7) was directly due to modified hydrogen buff-
ering properties. In this aim, the pH of 50 mM aqueous solution of the
various polymers, was measured upon incremental addition of NaOH
(Fig. 2A). The PEI 4 was hardly soluble at concentration above 40 mM
and was dropped out from this assay. Polyornithine, a cationic polymer
without any nucleic acid delivery ability on its own was used as a con-
trol. As expected, the high basicity of polyornithine amines does not en-
able to buffer the solution pH in the 4.5 to 7.8 range as seen by the sharp
increase in the pH solution with as low as 0.04 equivalent of base per
monomer (black triangle). Acid base titration of PEI (red square) con-
firmed the weaker basicity of the ethylenimine residues because as
much as 0.2 equivalent of base is required to raise the pH solution
from 4.5 to 7.8. Chemical modification of PEI altered the acid base titra-
tion profiles but only to a limited extend. In particular, the titration lines
of the modified PEIs superposed almost exactly in the pH 6.0–7.8 range
and had slopes similar to the PEI one. Considering that PEI blocks endo-
some near pH 6.0, this experiment strongly indicates that the high siRNA
delivery efficiencies of the polymers 1, 3, 5 and 6 relative to PEI, 2 or 7 do
not directly derived from a peculiar buffering abilities. Nonetheless, each
polymer undergoes a variation of its protonation state. From the slope of
the acid base titration curve, we can determine that for all polymers, 6%
of ethylenimine residues go from neutral to cationic when the pH of the
solution decreased from pH 7.8 to pH 6.0. This phenomenon can impact
the aqueous solubility of the polymer and was analyzed in the next ex-
periment (Fig. 2B). In here, we determined the pH value at which mac-
roscopic state of the polymer goes from soluble to particulate (Fig. 2B).
Aqueous solutions of each polymer were carefully treated with incre-
mental addition of NaOH. The pHs were then measured at the onset of
turbidity. The bared PEI never formed a turbid solution, even above pH
9.0 (not shown) and the less effective polymers 2 and 7 precipitated
only at pHs above 7.8. Interestingly, the effective siRNA delivery poly-
mers precipitated upon ethylenimine deprotonation within the pH
6.0–7.8 indicating that this pH sensitive aggregation ability is important.
For effective delivery, siRNA polyplexes should remain stable enough
to sustain electrostatic competition in extracellular media and conditions.
Assembly of the siRNA polyplexes by other means than electrostatic inter-
actions such as hydrophobic ones may actually limit electrostatic-induced
dissociation of the complex upon binding with sulfated proteoglycans
present on the external face of the plasma membrane. After internaliza-
tion, siRNA polyplexes should however be able to dissociate for a con-
trolled release of siRNAs. We examined the electrostatic stability of the
siRNA polyplexes at pH 7.8 or pH 6.0 to mimic interaction with the sulfat-
ed proteoglycans receptors either on the external plasma membrane
faces [31] or in PEI-loaded endosomes [21], respectively. The siRNA poly-
plexes from πPEI 1 and 2, 6 or 7 were prepared at the same EI/P ratio of 60
which previously used for the in vitro delivery assay (Fig. 1). The cationic
particles were then incubated with increasing amounts of the sulfated
polysaccharide heparin and release of siRNA was quantified from agarose
gel electrophoresis assays (Fig. 3).
B
120
100
80
60
40
20
0
sic
siluc
UC PEI PEIY
1
2
3
4
5
6
7
Fig. 1. A. Chemical structure of the siRNA delivery polyethylenimines (PEIs). Polymers
were modified at about 30% extend relative to ethylenimine (EI) residues, except for
πPEI 2, which has the same chemical structure of 1 but modified at 15%. B. Efficiency
of each polymer to deliver siRNA into U87 cells stably transformed to express an egf-
pluc fusion protein (U87egfpluc). Each PEI was mixed with either luc-targeting siRNA
(siluc) or untargeting siRNA (sic) and added to cells. Luciferase activity was measured
48 h later and plotted relative to untreated cells. Data show the means of triplicate and
standard deviations. Final concentrations were 10 nM in siRNA and 24 μM of the indi-
cated polymer to give siRNA polyplexes at a EI/P of 60. UC: untreated cells.
the siRNA-mediated silencing of a luciferase-enhanced green fluores-
cent chimera gene using either targeting (siluc) or control (sic) siRNAs.
To avoid flaws from transient transfection experiment, the human glio-
bastoma U87 cell line was previously genetically transformed to ex-
press a luciferase-green fluorescent protein chimera. The resulting
U87egfpluc has the advantage to express a fusion protein that can be si-
lenced with a unique targeted siRNA and be easily quantified either by
monitoring the green fluorescent protein or by measuring the enzymat-
ic activity of the firefly luciferase domains. The siRNAs polyplexes were
prepared by mixing the polymer with siRNA in RPMI cell culture medi-
um at final concentrations of 240 μM in ethylenimine unit (EI) and
100 nM in siRNA. The ensuing cationic polyplexes with an ethylenimine
to siRNA phosphate (EI/P) ratio of 60 were then simply added to cells by
dilution to 1/10 with the serum-containing cell culture medium to
reach final concentrations of 24 μM EI and 10 nM siRNA. Forty-eight
hours later, the luciferase activity of the cell lysate was measured and
expressed relative to untreated cells (UC) (Fig. 1B). As previously seen
on A549 cells [17], unmodified 25 kDa PEI showed to be a poor in vitro
siRNA carrier. In contrast, PEIY, πPEI 1 and hydroxybenzamido-PEI de-
rivatives 4, 5 and 6 enabled a high (above 90%) and selective siluc-
mediated luciferase silencing. Decreasing the pyridylthiourea content
on PEI (πPEI 2) considerably diminished the siRNA-mediated luciferase
gene silencing while a less dramatic diminution was obtained with the
polymer 3. Replacement of the thiourea linkage of πPEI 1 with an amide
bond (polymer 7) almost abolished the polymer delivery ability. Fig. S1

G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
423
A
A. Complex stability at extracellular pH 7.8
10
8
100
80
60
40
20
0
PEI
2, 7
polyOrn
PEI
6
1
6
6
2, 5
πPEI,
3, 7
0
1
2
3
4
Ratio of anionic charge of heparin per
ethylenimine residue
0.06
B. Complex stability at pH 6.0
2
0
0.2
0.4
0.5
0.8
100
80
60
40
20
0
2, 6
1
Molar equivalent of NaOH per monomer
1, PEI
7
B
particulate form
PEIY
6.3
6.4
1
2
3
4
5
6
7
8.4
7.7
6.1
6.2
6.4
0
1
2
3
Ratio of anionic charge of heparin per
ethylenimine residue
Fig. 3. Electrostatic stability of the polyplexes at pH 7.8 (extracellular value; graph A)
and pH 6.0 (PEI-buffered endosomal value, graph B). siRNA polyplexes made from
PEI (black dot), πPEI 1 (blue square) and 2 (yellow square), 7 (green diamond) or 6
(red triangle) at EI/P of 60 (full line) or 9.4 (dotted line) were incubated for 20 min
with increasing amounts of heparin. Release of siRNA was determined from agarose
gel electrophoresis analyses.
7.95
9
8
7
6
5
pH of the polymer solution
Cytotoxicity can be a limiting factor for the development of any drug
carrier. Polycations and especially the ones used for nucleic acid delivery
are known to damage cellular membranes [32,33] and to induce cell
death [32]. The potential damaging effect of PEI, 6, πPEI and PEIY on cel-
lular membranes was estimated from hemolysis assays using sheep red
blood cells (Fig. 4A).
Fig. 2. A. Acido-basic titration of polyornithine and of the various PEI polymers as indi-
cated. B. Sensitivity of the polymers to form particulate matter as a function of the pH.
The value indicated the transition pH at which the 20 mM aqueous polymer solution
becomes turbid. The colored box indicates the putative cellular pH range encountered
by the particles.
The PEI, but also the hydroxyphenyl-grafted 6 and PEIY polymers
possess a hemolytic activity at concentrations above 1000 μg/mL. In con-
trast, the πPEI did not induce any release of hemoglobin from the red
blood cells up to a concentration level of 5000 μg/mL. This result suggests
that πPEI, which has the same delivery activity as PEIY and 6, is the most
suitable carrier for in vivo administration. Next, πPEI cytotoxicity was
compared to that of PEIY, 6 and PEI by exposing the U87-egfpluc to
increasing concentrations of each polymer. Cell viability was then
extrapolated 48 h later by measuring the cellular mitochondrial activity
using the MTT reagent (Fig. 4B). Except for 6, the polymer affected the
cell viability with a sigmoidal type dose–response profile, as previously
observed on other cell lines [34]. The πPEI polymer appears more
innocuous to U87 cells (IC50 =90 μg/mL) than PEIY and PEI
(IC50 =40 μg/mL) and onset of toxicity is much above the πPEI final
concentration of 3 μg/mL used for siRNA delivery.
The siRNA polyplexes made from the soluble polymer PEI, 2 and 7
dissociated almost quantitatively upon challenge with heparin, at pH
7.8 (Fig. 3A). This suggests that their weak siRNA delivery efficiencies
may be due to a too early extracellular siRNA release. In contrast, the
polymers πPEI and 6, which exist in a particulate state above pH 6.4,
improved the electrostatic cohesion of the siRNA/πPEI complexes. De-
creasing the incubation pH to 6.0 (Fig. 3B), value of measured pH of
PEI-buffered endosomes [21] and at which any polymer is or becomes
fully soluble (Fig. 2B), led to almost quantitative siRNA release in
presence of heparin. Increasing the content of siRNA within πPEI
polyplexes by decreasing the EI/P ratio from 60 to 9.4 (dotted line)
did not modify the siRNA release profile at both pHs, suggesting
that the stability is given by the auto-interaction between the PEI-
appended elements.
Altogether, these in vitro data were similar to the ones obtained on
PEIY [22]. They provided further evidences that the effectiveness of
these siRNA delivery polymers is not due to a specific dangling chem-
ical group but is rather related to an overall pH-dependent solubility
of the polymer that is tuned to sense the cellular pH variation during
the intracellular delivery process.
Polymer toxicity on cells followed nonetheless the same trend than
the hemolytic activity, suggesting that direct PEI-induced damage to
membrane is undesired but can be worked out not only by chemical
modification of PEI with succinate [14] but also with aromatic groups.
The reason behind the singular low hemolytic and cellular toxicity of
πPEI is unclear. The most obvious speculation is that the specific

424
G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
NaCl or in HEPES buffered saline (HBS), pH 7.0 or pH 7.8. After a 2-h
A
100
80
60
40
20
0
incubation period at 25 °C, the mixtures were analyzed by dynamic
light scattering (DLS) measurements (Fig. 5A). In the electrolyte-free
glucose solution, the polymer alone was actually not fully soluble and
slightly self-assembled into a population of 21 nm-diameter particles.
The presence of siRNA at an EI/P ratio of 9.4 drove the two partners to
assemble into larger 96 nm-diameter particles. Increased πPEI to
siRNA ratio (EI/P ratio of 60) led to similar particles (138 nm-diameter).
Yet, the smaller particles (now 27 nm in diameter), representing pre-
sumably the excess of πPEI, were detected again. This type of polyplex
formation appears similar to the one described for the soluble PEI [13]
but starting with a cationic πPEI polymer being already a nanoaggregate
instead of a single molecule and suggests an assembly process essential-
ly driven by electrostatic interaction and a fast kinetic. Presence of salts
in the medium and increased pHs led to slightly larger complexes pre-
sumably because these elements favor hydrophobic interactions and
lower the electrostatic repulsion between cationic particles [35–37]. Be-
fore undertaking siRNA delivery experiments, we examined the effect
of dilution of the particles in a cell culture medium of higher pH and
containing serum. siRNA/πPEI polyplexes of about 100 nm in diameter
(prepared in 4.5% glucose, EI/P=9.4) showed to grow in size over
times but at a slow pace. After 4 h, sizes of particles remained below
200 nm (data not shown). Next, the siRNA delivery efficiency of these
so formed polyplexes was evaluated in vitro on the U87egfpluc cells
(Fig. 5B) by dilution into the 10% FBS containing cell culture medium
to a final πPEI concentration of 27 μM. Results showed the assembly
media, the stoichiometry, and the particle sizes, to have little impact
on the final siRNA-mediated luciferase silencing except using HBS pH
7.8, for which luciferase activity diminished to about 20% relative to
untreated cells versus less than 10% for the other media.
PEI
6
PEIY
πPEI 1
0
1
2
3
4
5
polymer concentration (mg/mL)
B
140
120
100
80
PEI
6
PEIY
πPEI 1
60
40
20
0
1
10
100
1000
polymer concentration (µg/mL)
The concentrated particles prepared in 4.5% glucose appeared the
most interesting because of their smallest sizes and highest efficiency.
Preparation of particles at EI/P ratio of 9.4 versus 60 appeared also more
interesting on a first base because of the occurrence of a single population
with the narrower size distribution and slightly highest siRNA-mediated
gene silencing. They were hence characterized further. The gel agarose
electrophoresis analysis (Fig. 5C) showed the 4.5% glucose solution to
permit full electrostatic complexation of the siRNAs by the cationic πPEI
above EI/P of 3. Transmission electron microscopy (Fig. 5D) showed
4.5% Glc solution to conduce πPEI and siRNA at an EI/P ratio of 9.4 to as-
semble into a population of spherical particles with an average diameter
of 60.0 nm (SD: 12 nm, n=40).
Fig. 4. A. Hemolytic activity of the polymers. 100% hemolysis was obtained using triton
X-100 (final concentration of 0.1% w/v) B. Effect of the polymer concentration on the
viability of U87egfpluc cells. Cell viability was monitored by measuring the cell meta-
bolic activity using the MTT assay.
polar/hydrophobic nature of thiourea forms a shield around the PEI,
hence preventing insertion of ethylenimine residues into lipid bilayers
and hence damage.
3.3. Optimization of complexes for in vivo administration
The duration of siRNA-mediated luciferase gene silencing after car-
riage using concentrated πPEI/siRNA polyplexes prepared in 4.5% glc
at EI/P of 9.4 was as well evaluated at the cellular level (Fig. 5E). The
πPEI showed to assist a selective siRNA-mediated luciferase gene silenc-
ing since delivery of an untargeted siRNA (sic) did not diminish the cel-
lular luciferase activity over the experiment time course. Maximum
protein activity inhibition (superior to 90%) was reached 48 h after ad-
ministration and lasted at least 4 days. It is worth mentioning that the
luciferase activity in untreated and πPEI/sic-treated wells increased
over times at the same rate. This indicates that the polyplexes did not
modify the cellular proliferation rate because the amount of luciferase,
produced in the U87egfpluc, is proportional to the number of cells.
The previous experiments showed πPEI to be an excellent in vitro
siRNA delivery vehicle with a favorable toxicological profile when formu-
lated in commercial RPMI medium at a concentration of 0.03 mg/mL
(0.24 mM in EI). Cumulative experiences with PEI for in vivo delivery of
nucleic acid suggest to administer nucleic acid polyplexes at a much
higher PEI concentration of about 1–2 mg/mL [8,35]. However, the kinetic
and polyplex assembly processes depend not only on the incubation me-
dium [36] but also on the concentration of the partners. Concentrated so-
lutions, and especially ones made with self-assembling polymers, can
yield to sizes of polyplexes too large for in vivo injection. For instance,
self-assembly of πPEI in RPMI at 29.3 μg/mL (diluted conditions used
for in vitro experiments), yielded particles with an average diameter of
240 nm after 30 min incubation. The same polymer, in the same condi-
tions but at higher concentration (1.83 mg/mL) grew faster and ended
after 30 min incubation to particles with diameter of 1200 nm. We there-
fore aimed to find a suitable assembly medium enabling formation of
stable πPEI/siRNA polyplexes, even at high concentration. To limit the
auto-assembly of the polymer during the siRNA polyplex formation
while remaining as close as possible to a physiological pH, the πPEI
stock solution (18.3 mg/mL) was adjusted to pH 6.0. The πPEI polymer
was then diluted alone or in the presence of siRNA at a final 1.83 mg/mL
concentration either in aqueous 4.5% (w/v) glucose solution, in 150 mM
3.4. In vivo experiment
The final experiment was to demonstrate the in vivo potential of
πPEI to carry any siRNA. Subcutaneous injection of U87egfpluc human
glioblastoma cells into flanks of athymic mice led to well-developed
solid tumors with exponential growth mode and expressing luciferase.
At this development stage, the idea was only to provide evidence of the
delivery ability of the polymer in an in vivo context. We thought it is rea-
sonable to attempt silencing the innocuous and easy to monitor lucifer-
ase reporter gene of U87egfpluc cells instead of silencing a putative
gene involved in tumor growth [38].

G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
425
p = 0.9
18
16
14
12
10
8
Cationic polyplexes are known to promote erythrocyte aggrega-
tion [39], to dissociate in the blood [40], and to accumulate preferen-
tially in lung or liver [41] over tumor, leading to possible systemic
toxicity [42]. Before undertaking sophistication by equipping the
p = 0.007
A
πPEI/siRNA
Average
PDI
Medium pH
ratio (EI/P)
diameters (nm)
6
Glucose
(Glc)
6
no siRNA
9.4
21
96
0.35
0.26
4
2
60
27 (30%); 138 (70%) 0.8
0
NaCl
6
no siRNA
9.4
615
274
0.35
0.28
Control
πPEI/sic
πPEI/siluc
Fig. 6. Luciferase activity of the U87egfpluc tumors 4 days after polyplex injection (in
MegaRLU/10 s/μg tumoral protein). Polyplexes (72 μg πPEI, 23.5 μg siRNA in 40 μL
4.5% glucose) were injected into the solid tumor 25 days after subcutaneous inocula-
tion of the U87egfpluc cells. Each condition was performed on group of 5 tumors. Sta-
tistical analysis (Student t-test) and calculation of the t probability (p) were performed
using KaleidaGraph Software 4.1 on paired data.
HBS 7.0
HBS 7.8
no siRNA
9.4
no siRNA
9.4
147
630
0.24
0.8
193
700
0.225
0.7
B
120
100
80
sic
siluc
nucleic acid delivery systems with targeting or/and surface-shielding
elements [12, 24], we aimed to prove that the bared cationic self-
assembling πPEI polymer conserved its intrinsic siRNA delivery ability
into cells also in an in vivo setting. As recommended [41], problems re-
lated to systemic injection were bypassed by administration of the
siRNA polyplexes directly into the tumors. A single dose of siRNA poly-
plexes was slowly injected into tumors of average volume of
260 mm3. Four days later, the luciferase activity of the U87egfpluc tu-
mors was measured and reported in Fig. 6. A luciferase gene silencing
of 30% was observed with the siluc/πPEI polyplexes by comparison to
controls 4 days after injection. A statistical analysis using the Student
t-test on paired data indicates a high probability of similarity (p of
0.9) between luciferase expression in the tumor controls and (sic/
πPEI)-treated tumors and a significant difference (p of 0.007) between
(sic/πPEI)- and (siluc/πPEI)-treated tumors. A 30% gene silencing ac-
tivity may appear modest but increased dose and design of protocol
with repeated injection should help to get better gene silencing activ-
ity. Nonetheless this level may be sufficient for cancer immunotherapy
where modulation of cytokine level can amplify immune response
[38]. If literature led us to believe that chemical sophistication is man-
datory for systemic injection, these bared cationic systems may con-
ceivably be already of practical application for intracellular siRNA
delivery of accessible tissues such as lung or bladder as previously
demonstrated with analogous delivery systems [10, 43].
60
40
20
0
medium Glc
Glc
6.0
60
NaCl HBS HBS
pH
6.0
9.4
6.0
9.4
7.0
9.4
7.8
9.4
EI/P
C
D
EI/P ratio
0.5
0
1
2
3
4. Conclusion
In summary, we synthesized a novel self-assembling polymeric
siRNA carrier by conjugating the branched PEI 25 kDa to pyridines
via thiourea linkages. The particular kind of hydrophobicity given by
200 nm
E
cell
400
Fig. 5A. Effect of the medium, pH and stoichiometry of the partners on πPEI particles
formation. Diameters and polydispersity index were determined from light scattering
data. B. Effect of the medium, pH and stoichiometry of the partners on the efficiency
of the πPEI polyplexes to deliver siRNA into U87egfpluc cells. Polyplexes were diluted
in 10% FBS medium to reach final concentrations of 27 μM πPEI and either 72.7 nM
siRNA or 11.4 nM for polyplexes with EI/P of 9.4 and 60, respectively. C. Electrophoresis
analysis of the formation of πPEI/siRNA polyplex in 4.5% Glc as a function of the poly-
mer etylenimine monomer (EI) to siRNA phosphate (P) ratio. D. Transmission electron
microscopy analysis of πPEI particles at an EI/P ratio of 9.4, prepared in 4.5% glucose.
Scale bar represents 200 nm. E. Duration of siRNA-mediated luciferase silencing using
πPEI/siRNA polyplexes (EI/P 9.4) prepared in 4.5% glucose (Glc). Luciferase activity
was plotted relative to untreated cells of day 1. All particles in this figure were pre-
pared at a πPEI concentration of 1830 μg/mL or 15 mM in EI. NaCl: 150 mM NaCl solu-
tion, HBS: HEPES buffered saline.
sic
siluc
300
200
100
0
24
48
72
96
Incubation time (hours)

426
G. Creusat et al. / Journal of Controlled Release 157 (2012) 418–426
[18] D. Edinger, E. Wagner, Bioresponsive polymers for the delivery of therapeutic
nucleic acids, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 3 (1) (2011) 33–46.
[19] S.M. Noh, S.E. Han, G. Shim, K.E. Lee, C.W. Kim, S.S. Han, Y. Choi, Y.K. Kim, W.K. Kim,
Y.K. Oh, Tocopheryl oligochitosan-based self assembling oligomersomes for siRNA
delivery, Biomaterials 32 (3) (2011) 849–857.
the pyridylthiourea element appeared to confer to PEI 25 kDa a high
siRNA delivery ability in vitro and an appreciable low toxicological
profile. The πPEI carrier can be formulated with siRNAs to form nano-
metric delivery systems with sizes and concentrations suitable for
local in vivo administration. A proof of concept experiment showed
already a 30% siRNA-mediated gene silencing 4 days after local intra-
tumoral in vivo administration. Our current effort is now directed to-
ward the performance of a comprehensive in vivo study.
[20] G. Creusat, G. Zuber, Self-assembling polyethylenimine derivatives mediate effi-
cient siRNA delivery in mammalian cells, Chembiochem
9 (17) (2008)
2787–2789.
[21] N.D. Sonawane, F.C. Szoka Jr., A.S. Verkman, Chloride accumulation and swelling
in endosomes enhances DNA transfer by polyamine–DNA polyplexes, J. Biol.
Chem. 278 (45) (2003) 44826–44831.
[22] G. Creusat, A.S. Rinaldi, E. Weiss, R. Elbaghdadi, J.S. Remy, R. Mulherkar, G. Zuber,
Proton sponge trick for pH-sensitive disassembly of polyethylenimine-based
siRNA delivery systems, Bioconjug. Chem. 21 (5) (2010) 994–1002.
[23] S. Foillard, G. Zuber, E. Doris, Polyethylenimine–carbon nanotube nanohybrids for
siRNA-mediated gene silencing at cellular level, Nanoscale 3 (4) (2011) 1461–1464.
[24] G. Zuber, M. Dontenwill, J.P. Behr, Synthetic viruslike particles for targeted gene
delivery to alphavbeta3 integrin-presenting endothelial cells, Mol. Pharm. 6 (5)
(2009) 1544–1552.
[25] A.W. Fraley, B. Pons, D. Dalkara, G. Nullans, J.P. Behr, G. Zuber, Cationic
oligonucleotide-peptide conjugates with aggregating properties enter efficiently
into cells while maintaining hybridization properties and enzymatic recognition,
J. Am. Chem. Soc. 128 (33) (2006) 10763–10771.
[26] D. Dalkara, C. Chandrashekhar, G. Zuber, Intracellular protein delivery with a
dimerizable amphiphile for improved complex stability and prolonged protein
release in the cytoplasm of adherent cell lines, J. Control. Release 116 (3)
(2006) 353–359.
[27] A. Akinc, M. Thomas, A.M. Klibanov, R. Langer, Exploring polyethylenimine-
mediated DNA transfection and the proton sponge hypothesis, J. Gene Med. 7
(5) (2005) 657–663.
[28] E.M. Zaghloul, J.R. Viola, G. Zuber, C.I. Smith, K.E. Lundin, Formulation and deliv-
ery of splice-correction antisense oligonucleotides by amino acid modified poly-
ethylenimine, Mol. Pharm. 7 (3) (2010) 652–663.
[29] A.M. Doody, J.N. Korley, K.P. Dang, P.N. Zawaneh, D. Putnam, Characterizing the
structure/function parameter space of hydrocarbon-conjugated branched poly-
ethylenimine for DNA delivery in vitro, J. Control. Release 116 (2) (2006)
227–237.
[30] A. von Harpe, H. Petersen, Y. Li, T. Kissel, Characterization of commercially avail-
able and synthesized polyethylenimines for gene delivery, J. Control. Release 69
(2) (2000) 309–322.
[31] M.H. Dufresne, M. Elsabahy, J.C. Leroux, Characterization of polyion complex mi-
celles designed to address the challenges of oligonucleotide delivery, Pharm. Res.
25 (9) (2008) 2083–2093.
[32] D. Fischer, Y. Li, B. Ahlemeyer, J. Krieglstein, T. Kissel, In vitro cytotoxicity testing
of polycations: influence of polymer structure on cell viability and hemolysis, Bio-
materials 24 (7) (2003) 1121–1131.
[33] S. Hong, P.R. Leroueil, E.K. Janus, J.L. Peters, M.M. Kober, M.T. Islam, B.G. Orr, J.R. Baker Jr.,
M.M. Banaszak Holl, Interaction of polycationic polymers with supported lipid bilayers
and cells: nanoscale hole formation and enhanced membrane permeability, Bioconjug.
Chem. 17 (3) (2006) 728–734.
[34] O.M. Merkel, A. Beyerle, B.M. Beckmann, M. Zheng, R.K. Hartmann, T. Stoger, T.H.
Kissel, Polymer-related off-target effects in non-viral siRNA delivery, Biomaterials
32 (9) (2011) 2388–2398.
[35] D. Goula, J.S. Remy, P. Erbacher, M. Wasowicz, G. Levi, B. Abdallah, B.A. Demeneix,
Size, diffusibility and transfection performance of linear PEI/DNA complexes in
the mouse central nervous system, Gene Ther. (1998) 712–717.
[36] S. Werth, B. Urban-Klein, L. Dai, S. Hobel, M. Grzelinski, U. Bakowsky, F. Czubayko,
A. Aigner, A low molecular weight fraction of polyethylenimine (PEI) displays in-
creased transfection efficiency of DNA and siRNA in fresh or lyophilized com-
plexes, J. Control. Release 112 (2) (2006) 257–270.
[37] M. Ogris, P. Steinlein, M. Kursa, K. Mechtler, R. Kircheis, E. Wagner, The size of
DNA/transferrin-PEI complexes is an important factor for gene expression in cul-
tured cells, Gene Ther. 5 (10) (1998) 1425–1433.
[38] X. Yang, Y. Peng, B. Yu, J. Yu, C. Zhou, Y. Mao, L.J. Lee, R.J. Lee, A covalently stabi-
lized lipid-polycation-DNA (sLPD) vector for antisense oligonucleotide delivery,
Mol. Pharm. 8 (3) (2011) 709–715.
[39] R. Kircheis, S. Schuller, S. Brunner, M. Ogris, K.H. Heider, W. Zauner, E. Wagner,
Polycation-based DNA complexes for tumor-targeted gene delivery in vivo, J.
Gene Med. 1 (2) (1999) 111–120.
Acknowledgments
This work was funded by the non-profit organizations CEFIPRA
and the AFM. We would like to thank Yannick Schwab and Coralie
Spiegelhalter for TEM imaging, Dominique Bagnard (INSERM) for lu-
ciferase measurement, Philippe and Dany Zuber for hosting support
and Béatrice Heurtault and Carole Schanté for proof-reading.
Appendix A. Supplementary material
Supplementary data to this article can be found online at doi:10.
1016/j.jconrel.2011.10.007.
References
[1] S.M. Elbashir, J. Harborth, W. Lendeckel, A. Yalcin, K. Weber, T. Tuschl, Duplexes of
21-nucleotide RNAs mediate RNA interference in cultured mammalian cells, Na-
ture 411 (6836) (2001) 494–498.
[2] D.M. Dykxhoorn, D. Palliser, J. Lieberman, The silent treatment: siRNAs as small
molecule drugs, Gene Ther. 13 (6) (2006) 541–552.
[3] A.P. McCaffrey, L. Meuse, T.T. Pham, D.S. Conklin, G.J. Hannon, M.A. Kay, RNA in-
terference in adult mice, Nature 418 (6893) (2002) 38–39.
[4] E. Song, S.K. Lee, J. Wang, N. Ince, N. Ouyang, J. Min, J. Chen, P. Shankar, J. Lieberman,
RNA interference targeting Fas protects mice from fulminant hepatitis, Nat. Med. 9
(3) (2003) 347–351.
[5] T.S. Zimmermann, A.C. Lee, A. Akinc, B. Bramlage, D. Bumcrot, M.N. Fedoruk, J. Harborth,
J.A. Heyes, L.B. Jeffs, M. John, A.D. Judge, K. Lam, K. McClintock, L.V. Nechev, L.R. Palmer,
T. Racie, I. Rohl, S. Seiffert, S. Shanmugam, V. Sood, J. Soutschek, I. Toudjarska, A.J. Wheat,
E. Yaworski, W. Zedalis, V. Koteliansky, M. Manoharan, H.P. Vornlocher, I. MacLachlan,
RNAi-mediated gene silencing in non-human primates, Nature 441 (7089) (2006)
111–114.
[6] Y. Chen, J. Sen, S.R. Bathula, Q. Yang, R. Fittipaldi, L. Huang, Novel cationic lipid
that delivers siRNA and enhances therapeutic effect in lung cancer cells, Mol.
Pharm. 6 (3) (2009) 696–705.
[7] P. Ofek, W. Fischer, M. Calderon, R. Haag, R. Satchi-Fainaro, In vivo delivery of
small interfering RNA to tumors and their vasculature by novel dendritic nano-
carriers, FASEB J. 24 (9) (2010) 3122–3134.
[8] S. Hobel, I. Koburger, M. John, F. Czubayko, P. Hadwiger, H.P. Vornlocher, A. Aigner,
Polyethylenimine/small interfering RNA-mediated knockdown of vascular endothe-
lial growth factor in vivo exerts anti-tumor effects synergistically with Bevacizumab,
J. Gene Med. 12 (3) (2010) 287–300.
[9] M.S. Huh, S.Y. Lee, S. Park, S. Lee, H. Chung, S. Lee, Y. Choi, Y.K. Oh, J.H. Park, S.Y. Jeong,
K. Choi, K. Kim, I.C. Kwon, Tumor-homing glycol chitosan/polyethylenimine nano-
particles for the systemic delivery of siRNA in tumor-bearing mice, J. Control. Release
144 (2) (2010) 134–143.
[10] O.M. Merkel, A. Beyerle, D. Librizzi, A. Pfestroff, T.M. Behr, B. Sproat, P.J. Barth, T. Kissel,
Nonviral siRNA delivery to the lung: investigation of PEG–PEI polyplexes and their in
vivo performance, Mol. Pharm. 6 (4) (2009) 1246–1260.
[11] X.L. Wang, R. Xu, X. Wu, D. Gillespie, R. Jensen, Z.R. Lu, Targeted systemic delivery
of a therapeutic siRNA with a multifunctional carrier controls tumor proliferation
in mice, Mol. Pharm. 6 (3) (2009) 738–746.
[12] M.E. Davis, J.E. Zuckerman, C.H. Choi, D. Seligson, A. Tolcher, C.A. Alabi, Y. Yen,
J.D. Heidel, A. Ribas, Evidence of RNAi in humans from systemically adminis-
tered siRNA via targeted nanoparticles, Nature 464 (7291) (2010) 1067–1070.
[13] O. Boussif, F. Lezoualc'h, M.A. Zanta, M.D. Mergny, D. Scherman, B. Demeneix, J.P. Behr,
A versatilevector for geneand oligonucleotide transfer into cells in culture and in vivo:
polyethylenimine, Proc. Natl. Acad. Sci. U. S. A. 92 (16) (1995) 7297–7301.
[14] A. Zintchenko, A. Philipp, A. Dehshahri, E. Wagner, Simple modifications of
branched PEI lead to highly efficient siRNA carriers with low toxicity, Bioconjug.
Chem. 19 (7) (2008) 1448–1455.
[15] P. Kaestner, A. Aigner, H. Bastians, Therapeutic targeting of the mitotic spindle
checkpoint through nanoparticle-mediated siRNA delivery inhibits tumor growth
in vivo, Cancer Lett. 304 (2) (2011) 128–136.
[16] A.L. Bolcato-Bellemin, M.E. Bonnet, G. Creusat, P. Erbacher, J.P. Behr, Sticky over-
hangs enhance siRNA-mediated gene silencing, Proc. Natl. Acad. Sci. U. S. A. 104
(41) (2007) 16050–16055.
[40] R.S. Burke, S.H. Pun, Extracellular barriers to in vivo PEI and PEGylated PEI polyplex-
mediated gene delivery to the liver, Bioconjug. Chem. 19 (3) (2008) 693–704.
[41] O.M. Merkel, D. Librizzi, A. Pfestroff, T. Schurrat, K. Buyens, N.N. Sanders, S.C. De Smedt,
M. Behe, T. Kissel, Stability of siRNA polyplexes from poly(ethylenimine) and poly(-
ethylenimine)-g-poly(ethylene glycol) under in vivo conditions: effects on pharmaco-
kinetics and biodistribution measured by Fluorescence Fluctuation Spectroscopy and
Single Photon Emission Computed Tomography (SPECT) imaging, J. Control. Release
138 (2) (2009) 148–159.
[42] M. Meyer, C. Dohmen, A. Philipp, D. Kiener, G. Maiwald, C. Scheu, M. Ogris, E. Wagner,
Synthesis and biological evaluation of a bioresponsive and endosomolytic siRNA–
polymer conjugate, Mol. Pharm. 6 (3) (2009) 752–762.
[43] C.-J. Arum, Y. Kodama, N. Rolim, M. Wideroe, E. Anderssen, T. Viset, M. Otterlei,
S. Lundgren, D. Chen, C.-M. Zhao, A rat model of intravesical delivery of small in-
terfering RNA for studying urinary carcinoma, World J. Urol. 28 (4) (2010)
479–485.
[17] A.C. Grayson, A.M. Doody, D. Putnam, Biophysical and structural characterization
of polyethylenimine-mediated siRNA delivery in vitro, Pharm. Res. 23 (8) (2006)
1868–1876.