Carbohydrate-interactive pDNA and siRNA gene vectors based on boronic acid functionalized poly(amido amine)s

Article abstract of DOI:10.1016/j.jconrel.2013.02.008

In order to evaluate the influence of incorporation of boronic acid groups on the properties of poly(amido amine)s as gene vectors, a novel poly(amido amine) copolymer p(CBA-ABOL/2AMPBA) containing ortho-aminomethylphenylboronic acid (2AMPBA) moieties was prepared by Michael-type polyaddition of a mixture of 1,4-aminobutanol (ABOL) and 2-((4-aminobutylamino)methyl)phenyl boronic acid to N,N′-cystamine bisacrylamide (CBA). It appeared that the presence of the boronic acid moieties as side groups along the polymer chain strongly enhances the stability of the self-assembled nanoparticles and nanosized polyplexes formed from this polymer; no aggregation was observed after storage for 6 days at 37°C. This strong stabilization can be attributed to intermolecular Lewis acid-base interactions between the 2AMPBA groups and the alcohol and amine groups present in the polymer, leading to dynamical (reversible) crosslinking in the nanoparticles. Moreover, since the boronic acids can reversibly form boronic esters with vicinal diol groups, the presence of the 2AMPBA groups add carbohydrate-interactive properties to these polymers that strongly influence their behavior as gene delivery vectors. DNA transfection with p(CBA-ABOL/2AMPBA) polyplexes gave transfection efficiencies that were approximately similar to commercial PEI in different cell lines (COS-7, HUH-6 and H1299-Fluc), but lower than those obtained with reference polyplexes from p(CBA-ABOL). It is hypothesized that the uptake of the boronated polyplexes is suppressed by binding to the glycocalyx of the cells. This is supported by the observation that addition of sorbitol or dextran to the transfection medium significantly enhances the transfection efficiency, which can be attributed to increased cellular uptake of the polyplexes due to boronic ester formation with these agents. AFM, SEM and confocal microscopy showed that polyplexes of p(CBA-ABOL/2AMPBA) become decorated with a dextran layer in the presence of 0.9% (w/v) dextran in the polyplex medium. In siRNA-mediated gene silencing experiments the use of p(CBA-ABOL/2AMPBA) as polymeric vector gave a knockdown of luciferase expression in H1229-Fluc cells of 35%. Also in this case addition of 0.9% (w/v) sorbitol or dextran to the transfection medium strongly increased the knockdown efficiency to 59% and 76%, respectively.

Full text of DOI:10.1016/j.jconrel.2013.02.008

Journal of Controlled Release 169 (2013) 266–275
Contents lists available at SciVerse ScienceDirect
Journal of Controlled Release
journal homepage: www.elsevier.com/locate/jconrel
Carbohydrate-interactive pDNA and siRNA gene vectors based on boronic acid
functionalized poly(amido amine)s
⁎
Martin Piest, Marc Ankoné, Johan F.J. Engbersen
Department of Biomedical Chemistry, MIRA Institute for Biomedical Technology and Technical Medicine, Faculty of Science and Technology, University of Twente, P.O. Box 217,
7500 AE Enschede, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
In order to evaluate the influence of incorporation of boronic acid groups on the properties of poly(amido
amine)s as gene vectors, a novel poly(amido amine) copolymer p(CBA-ABOL/2AMPBA) containing ortho-
aminomethylphenylboronic acid (2AMPBA) moieties was prepared by Michael-type polyaddition of a
mixture of 1,4-aminobutanol (ABOL) and 2-((4-aminobutylamino)methyl)phenyl boronic acid to N,N′-
cystamine bisacrylamide (CBA). It appeared that the presence of the boronic acid moieties as side groups
along the polymer chain strongly enhances the stability of the self-assembled nanoparticles and nanosized
polyplexes formed from this polymer; no aggregation was observed after storage for 6 days at 37 °C. This strong
stabilization can be attributed to intermolecular Lewis acid–base interactions between the 2AMPBA groups and
the alcohol and amine groups present in the polymer, leading to dynamical (reversible) crosslinking in the
nanoparticles. Moreover, since the boronic acids can reversibly form boronic esters with vicinal diol groups, the
presence of the 2AMPBA groups add carbohydrate-interactive properties to these polymers that strongly influence
their behavior as gene delivery vectors. DNA transfection with p(CBA-ABOL/2AMPBA) polyplexes gave transfec-
tion efficiencies that were approximately similar to commercial PEI in different cell lines (COS-7, HUH-6 and
H1299-Fluc), but lower than those obtained with reference polyplexes from p(CBA-ABOL). It is hypothesized
that the uptake of the boronated polyplexes is suppressed by binding to the glycocalyx of the cells. This is support-
ed by the observation that addition of sorbitol or dextran to the transfection medium significantly enhances the
transfection efficiency, which can be attributed to increased cellular uptake of the polyplexes due to boronic
ester formation with these agents. AFM, SEM and confocal microscopy showed that polyplexes of p(CBA-ABOL/
2AMPBA) become decorated with a dextran layer in the presence of 0.9% (w/v) dextran in the polyplex medium.
In siRNA-mediated gene silencing experiments the use of p(CBA-ABOL/2AMPBA) as polymeric vector gave a
knockdown of luciferase expression in H1229-Fluc cells of 35%. Also in this case addition of 0.9% (w/v) sorbitol
or dextran to the transfection medium strongly increased the knockdown efficiency to 59% and 76%, respectively.
© 2013 Elsevier B.V. All rights reserved.
Received 29 October 2012
Accepted 9 February 2013
Available online 18 February 2013
Keywords:
Gene delivery
Poly(amido amine)
Boronic acid
Carbohydrate interaction
Dextran
1. Introduction
from the Michael-type polyaddition of 4-aminobutanol to cystamine
bisacrylamide, having pending hydroxybutyl side chains (Scheme 1A),
Administration of therapeutic ‘healthy’ genes or silencing of
malfunctioning genes in the body is a promising technique to cure var-
ious diseases that have a genetic component [1]. The greatest challenge
in this approach is the delivery of the therapeutic genes into the target
cells. Although the use of viral vectors have shown some success, inher-
ent disadvantages like immune response and mutagenicity direct
further expectations in progress in this field towards application of
multifunctional cationic polymeric vectors [2–6]. However, using
these non-viral systems, the efficiency of transfection has to be signifi-
cantly improved. We previously developed a range of poly(amido
amine)s with repetitive disulfide linkages in the main chain (SS-PAAs)
that combine high transfection efficiencies with low cytotoxicity profiles
[7–14]. Among these polymers, the polymer (p(CBA-ABOL), derived
was the most successful in vitro gene delivery vector to date [7].
In our studies to further evaluate the influence of different function-
alities on the physicochemical and biological properties of poly(amido
amine)s for gene delivery [9], we have recently started to explore the
effects of the presence of boronic acid moieties in SS-PAAs [15]. Boronic
acids are Lewis acids that reversibly form boronic esters with vicinal
diols, including carbohydrates [16–20]. Incorporation of boronic acids
in various materials is known to improve cell adhesion by reaction
with the glycosylated membrane surface [21,22] and allows for new
strategies in drug and gene delivery [15,23–27]. For example, boronic
acids bind to sialic acid which is overexpressed at the surface of various
types of cancer cells, enabling targeting of tumor cells [28,29] and the
use of carbohydrate binding boronic agents has been proposed as a po-
tential therapeutic route to the treatment of infections attributed to
viruses possessing a highly glycosylated envelope such as HIV [30].
Since carbohydrates play a pivotal role in many cellular processes, like
⁎
Corresponding author. Tel.: +31 53 489 2926; fax: +31 53 489 2479.
E-mail address: j.f.j.engbersen@utwente.nl (J.F.J. Engbersen).
0168-3659/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jconrel.2013.02.008

M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
267
H
H
NH2
NH2
N
N
+
+
S S
O
O
ABOL
2AMPBA
CBA
H
H
*
N
N
N
∗
S S
NH
HO
OH
B
O
O
OH
HO
p(CBA-ABOL)
H
H
H
H
*
N
N
N
N
N
N
*
S S
S S
O
O
O
O
23%
77%
NH
HO
p(CBA-ABOL/2AMPBA)
OH
B
OH
A
B
Scheme 1. A: Structure of p(CBA-ABOL) obtained by Michael-type polyaddition of N,N′ cystamine bisacrylamide (CBA) and 4-aminobutanol (ABOL). B: Michael addition polymerization
of CBA with 75% ABOL and 25% of 2-((4-aminobutylamino)methyl)phenyl boronic acid (2AMPBA), resulting in the boronated analog p(CBA-ABOL/2AMPBA) containing 23%
2-aminomethylphenylboronic acid groups.
cell adhesion and molecular recognition processes by membrane
glycocalyxes [31–35], it is of particular interest to gain information
about the effect of introduction of carbohydrate binding boronic acid moi-
eties in polymeric gene vectors. In a previous study we have modified an
SS-PAA polymer with aminobutyl side chains with phenylboronic acid by
grafting the boronic acid moieties to 30% of the aminobutyl side chains.
This resulted in improved polyplex stability and transfection efficiencies
similar to linear PEI [15]. However, the high content of positively charged
primary amine groups (70% of the side chains) in this polymer dominated
the overall properties of this polymer and cell toxicity at higher doses was
observed. In order to obtain a more direct reflection of the potential
effects of boronic acids on relevant physicochemical and biological prop-
erties of gene transfection polymers, we have designed and synthesized a
novel SS-PAA copolymer, p(CBA-ABOL/2AMPBA), which possesses, be-
sides the aimed 30% boronic acid moieties, 70% neutral hydroxybutyl
groups in the side chains. The use of hydroxybutyl groups as appending
groups has the additional advantage that the properties of the copoly-
mer p(CBA-ABOL/2AMPBA) can be well compared with those of the
parent polymer p(CBA-ABOL), which has already extensively studied
in various in vitro transfection studies and appeared to be essentially
nontoxic and among the most effective vectors for pDNA transfec-
tion in vitro [7,36,37]. The copolymer p(CBA-ABOL/2AMPBA) can be
easily obtained by Michael-type polyaddition of a 3:1 mixture of
aminobutanol (ABOL) and 2-((4-aminobutylamino)methyl)phenyl
boronic acid (2AMPBA) to an equimolar amount of N,N′-cystamine
bisacrylamide (CBA) (Scheme 1B). 2AMPBA was selected as the boron-
ic acid derivative since the aminomethyl group in the ortho position rela-
tive to the boronic center in 2AMPBA can give a dative B–N (or N→B)
interaction causing enhanced binding affinity of the boronic acid for vici-
nal diols (and thus carbohydrates) by reversible boronic ester formation
at neutral pH [38].
readily degradable in the reductive intracellular environment by cleavage
of the disulfide linkages. This enables rapid unpacking of the therapeutic
payload once the polyplex has been internalized by the cell. (3) The pres-
ence of hydroxybutyl groups in the side chains of SS-PAAs has shown to
be favorable for transfection efficiency. Although the specific role of
this side group has not yet been elucidated, it is assumed that this
group provides additional endosomolytic properties to the polymer
The structure of p(CBA-ABOL/2AMPBA) has several features which
makes this polymer of interest to evaluate as potential vector for gene
delivery (Fig. 1).
Fig. 1. Overview of the special features of p(CBA-ABOL/2AMPBA). (1) Protonable
tertiary amines for DNA condensation and endosomal pH buffering. (2) Repetitive
disulfide linkages for fast intracellular polymer degradation through disulfide reduc-
tion. (3) Hydroxybutyl side chains to contribute to transfection efficiency of DNA
polyplexes. (4) Intra- and intermolecular Lewis acid–base interactions of 2AMPBA
groups with amine and alcohol groups lead to stabilization of nanoparticles and
polyplexes. (5) Negatively charged boronate anions and boronate esters formed in
nanoparticles and polyplexes give stabilizing charge interaction with proximal cationic
amines.(6) 2AMPBA groups at the surface of polyplexes allow carbohydrate decoration and
reversible interaction with glycocalixes at the cell surface. (7) 2AMPBA groups can reversibly
bind siRNA at the 3′-terminal end of the nucleotide chain, stabilizing polyplexes with siRNA
and allowing release of siRNA in the cell. For further explanations see text.
(1) The polymer has protonable tertiary amino groups which makes
this polymer a polycation suitable for self-assembled DNA condensation
to nanosized polyplexes. Moreover, since only part of the amino groups
is protonated at neutral pH (≤30%) the amino groups in the PAA main
chain give the polymer buffer capacity in the pH range 7.4–5.1, the pH
region where endosomal acidification occurs [7]. This buffering effect
favorably contributes to the endosomal escape of PAA polyplexes.
(2) The repetitive disulfide moieties in the main chain make the polymer

268
M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
through a beneficial influence on the hydrophilic/hydrophobic balance.
(4) In nanoparticles and polyplexes the 2AMPBA groups interact with
neighboring hydroxybutyl side chains and unprotonated tertiary amine
groups, giving rise to non-covalent dynamic crosslinking interactions
that provide additional stability to these particles. (5) The 2AMPBA
groups can form negatively charged boronate anions and boronate esters
which can interact with cationic protonated amines, giving stabilizing
charge interactions in nanoparticles. (6) 2AMPBA groups at the polyplex
surface can bind to carbohydrates to form reversibly boronate esters, en-
abling carbohydrate decoration at the surface of polyplexes for introduc-
tion of stealth properties and/or cell specificity. The reversible character of
non-covalent Lewis acid–base interactions and dynamically covalent bo-
ronic ester formation implies that 2AMPBA groups also possess binding
affinity to the glycocalyx of the cell membrane. This could either lead to
increased uptake by the cells due to increased resident time at the cellular
surface and/or receptor-mediated uptake, or reduced uptake due to limit-
ed migration of the polyplexes through the glycocalix layer. (7) Another
important property of p(CBA-ABOL/2AMPBA) is that the 2AMPBA moie-
ties in the polymer can reversibly bind siRNA at the 3′-terminal end of
the nucleotide chain. It has been established that boronic acids have a
strong interaction with RNA due to reversible boronic ester formation
with the vicinal hydroxyl groups at the 3′ ribose end [39–41]. Therefore,
it can be expected that boronic ester formation with 2AMPBA gives addi-
tional stabilization to siRNA polyplexes, which is advantageous for siRNA
delivery where polyplex stability usually poses a problem. Moreover, the
dynamic reversible nature of the covalent boronic ester bonds could allow
for release of siRNA in the cells.
5 J(H–H)=8.4 Hz, CH2–CH2–CH2–CH2–NH2). MS (ES+-TOF): m/z=
239.3 (100) ([M+OH+H+], calculated for [C11H20BN2O3]: 239.16).
2.2. Synthesis of the polymers p(CBA-ABOL) and p(CBA-ABOL/2AMPBA)
The reference polymer p(CBA-ABOL) was synthesized according to
the procedure described in reference [7]. After ultrafiltration the poly-
mer was obtained in 32% overall yield.
The polymer p(CBA-ABOL/2AMPBA) was synthesized by the Michael
addition of N,N′-cystamine bisacrylamide (1.18 g, 4.37 mmol) with
a mixture of 78 mol% of 4-aminobutanol (0.31 g, 3.41 mmol) and
22 mol% of 2AMPBA (0.23 g, 0.96 mmol). The reactants were dissolved
in a mixture of methanol (1.6 ml), water (0.4 ml) and triethylamine
(0.5 ml). After 8 days reacting at 45 °C under nitrogen in the dark,
the reaction was terminated by addition of excess of 2AMPBA
(0.33 g, 1.47 mmol) dissolved in methanol (0.9 ml), water (0.6 ml)
and triethylamine (0.23 ml). The termination reaction was left to
proceed for another 5 days and the polymer was isolated using an ultra-
filtration membrane (MwCO 3000 g/mol), yielding 1.14 g of off-white
solid (66.5% yield). The composition of the polymer was established by
1H NMR (D2O, 300 MHz) and the content of boronic acid side groups
was 23% with respect to the total amount of side groups.
1H NMR (p(CBA-ABOL/2AMPBA), HCl salt, D2O, 300 MHz): δ 7.7–7.8
(0.23H, m, ArH), δ=7.4–7.5 (0.69H, m, ArH), δ=4.2–4.4 (0.46H, m,
ArH–CH2–NH), δ=3.65 (4H, t, (2H, t, NHCO–CH2–CH2–N), δ=3.55 (4H,
t, SS–CH2–CH2–N), δ=3.30 (1.64H, t, CH2–CH2–OH), δ=3.30 (0.46H, t,
CH2–CH2–NH–CH2–Ar), δ=3.15 (2H, t, R2–N–CH2–CH2), δ=2.90 (4H, t,
NHCO–CH2–CH2–N), δ=2.75 (4H, t, CH2–SS–CH2), δ=1.75 (2H, m,
N–CH2–CH2–CH2–CH2), δ=1.55 (2H, m, N–CH2–CH2–CH2–CH2).
2. Materials and methods
N,N′-cystamine bisacrylamide (CBA, Aldrich), N-BOC-1,4-diamino-
butane (Aldrich), benzoylchloride (Aldrich), 2-formylphenylboronic acid
(Aldrich) and 4-aminobutanol (Aldrich) were of commercial grade and
used without further purification. All reagents and solvents were of re-
agent grade and were used without further purification.
2.3. Determination of the molecular weight of the polymers
Molecular weight information was obtained using a TDA302 system
(operated without a column) equipped with a TDA302 triple detector
array, consisting of a refractive index (RI) detector, a light scattering de-
tector (7° and 90°) and a viscosity detector, together with the OmniSec
4.1 software provided by Viscotec (Oss, The Netherlands). First, the sam-
ples were dried overnight over Siccapent, and subsequently stirred to
dissolve for 24 h at a concentration of 5 mg/ml in a mixture of water/
methanol 1/4 (v/v). Next, 20 μl of the sample was injected and dn/dc
was calculated based on the RI response using the flow injection polymer
analysis (FIPA) method. The weight average molecular weight of the
polymers was determined to be Mw=4.1 kDa for p(CBA-ABOL) and
NMR spectra were recorded on a Varian Unity 300 (1H NMR
300 MHz) using the solvent residual peak as the internal standard.
ES-TOF-MS spectra were recorded on a Waters/Micromas LCT mass
spectrometer.
2.1. Synthesis of 2-((4-aminobutylamino)methyl)phenyl boronic acid
(2AMPBA)
2-Formylphenylboronic acid (5.10 g, 34.0 mmol) was dissolved in
45 ml methanol, together with N-BOC-1,4-diaminobutane (6.69 g,
34.7 mmol) and 3 equivalents triethylamine (10.08 g, 99.6 mmol) and
the mixture was stirred overnight under nitrogen at ambient tempera-
ture. The mixture turned yellow and 1H NMR confirmed the formation
of the Schiff's base. Next, 3 equivalents NaBH4 (3.91 g, 103.3 mmol)
were added to the reaction mixture to reduce the Schiff's base and the so-
lution turned colorless. The product was extracted thrice from water/
chloroform. The organic layer was dried over MgSO4 and evaporated
under reduced pressure. The product was isolated as an off-white solid
(8.78 g, 75% yield). The BOC-protected product was dissolved in 30 ml
methanol and deprotected by perfusion of HCl-gas through the reaction
mixture for 30 min at ambient temperature. The solvent was evaporated
to dryness and the product was isolated as the HCl salt. 1H NMR of the
product showed complete deprotection and contained 50% of 2AMPBA
according to 1H NMR using p-toluene sulfonic acid as internal standard.
Residual mass was due to bound water to the boronic acid and counter
ions (HCl-salt of the amines). The product was used without further puri-
fication in the synthesis of p(CBA-ABOL/2AMPBA).
Mw=4.2 kDa for p(CBA-ABOL/2AMPBA).
2.4. Polyplex preparation
Polyplexes were prepared in HEPES buffer solutions (50 mM, pH 7.4).
The use of PBS buffer in the polyplex samples is avoided since interaction
of phosphate anion to boronic acid would conceal the binding properties
of 2AMPBA. The DNA plasmid solution (1 mg/ml) as supplied by manu-
facturer (PlasmidFactory, Bielefeld) was diluted to a final concentration
of 0.075 mg/ml in HEPES buffer. A series of polymer solutions of different
concentrations was prepared by dilution of a solution of 0.9 mg/ml poly-
mer in HEPES buffer solution repeatedly 1:1 with HEPES buffer solution.
To prepare polyplexes at 48/1, 24/1, and 12/1 polymer/DNA weight
ratio 0.80 ml of each polymer solution was added to 0.20 ml of DNA solu-
tion in a 1.5 ml Eppendorf tube. A 48/1 polymer/DNA weight ratio corre-
sponds with an N/P ratio of ca. 45/1 and 40/1 for p(CBA-ABOL) and
p(CBA-ABOL/2AMPBA), respectively. The polyplex solutions were
vortexed for 5 s and incubated for 30 min at ambient temperature
prior to use. Polyplexes with siRNA (Anti-Firefly Luciferase, Eurogentec;
AllStars Negative Control siRNA, Qiagen) were prepared from
0.375 mg/ml polymer in HEPES buffer. Next, 16 μl of polymer solu-
tion was added to 4 μl of siRNA (31.25 μg/ml), resulting in polyplex
samples of 48/1 polymer/siRNA weight ratio.
1H NMR (D2O, 300 MHz): δ=7.1–7.4 (4H, m, ArH), δ=3.950
(2H, s, ArH–CH2–NH); δ=2.905 (2H, t, 3 J(H–H)=7.2 Hz, NH–CH2–
CH2), δ=2.825 (2H, t, 3 J(H–H)=7.5 Hz, CH2–CH2–NH2), δ=1.673
(2H, q, 5 J(H–H)=7.7 Hz, CH2–CH2–CH2–CH2–NH2), δ=1.601 (2H, q,

M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
269
For the study of the polyplex properties and transfection experiments
in the presence of sugars, 2.0 μl of 10% w/v of the appropriate carbohy-
drate solution was added to 20 μl of the polyplex dispersion. The resulting
samples, containing 0.9% w/v of the carbohydrate, were vortexed for 5 s
and were stored for 30 min at ambient temperature prior to use.
4 readings per well at different positions. For GFP excitation was at
480 nm and optimal emission was determined at 503 nm. The per-
centage of GFP positive cells was determined after fixation of the
cells with 4% formaldehyde in PBS, followed by a DAPI staining
according to manufacturer (Sigma). The 96-well plates were ana-
lyzed using a BD Pathway 435 bioimaging system operated with
AttoVision™ Software and the percentage of GFP positive cells was
calculated using Cellprofiler 2.0 Software.
2.5. Determination of polyplex properties by dynamic light scattering
(DLS)
The cell viability was measured using an XTT assay, in which the XTT
value for untreated cells (cells not exposed to transfection agents) was
taken as 100% cell viability. XTT measurements were performed in
quattro using the Perkin-Elmer LS50B luminescence spectrometer and
analyzed in Kineticalc for Windows V2.0.
Size and zeta-potential of the nanoparticles formed by spontaneous
self-assembly of the poly(amido amine)s, of poly(amido amine)/plasmid
DNA polyplexes and of poly(amido amine)/plasmid siRNA polyplexes
were measured at 25 °C with a Zetasizer Nano (Malvern Instruments
Ltd, Malvern, UK) and the dynamic light scattering results were processed
using Dispersion Technology Software V5.0. For the stability measure-
ments in time the polyplexes were stored at 37 °C and measured at 25 °C.
2.8. Luciferase knockdown using siRNA
Anti-luciferase siRNA was used to knockdown luciferase expression
in H1299-Fluc cells. H1299-Fluc is a non-small cell lung carcinoma cell
line derived from the lymph node, stably expressing firefly luciferase.
Cells were cultured in completed RPMI 1640 medium completed with
L-glutamine (1% v/v of GlutaMAX 200 mM) and penicillin–streptomycin
solution (2% v/v of stock solution) and 10% v/v FBS). The cells were plated
at 8000 cells/well in a 96-well plate 24 h prior to the experiment and in-
cubation at 37 °C in a humidified 5% CO2-containing atmosphere. Cells
were 40–50% confluent.
Two parallel series of siRNA-polyplex incubations were performed,
one series using anti-luciferase siRNA (Eurogentec) to determine the
knockdown and one series using Allstars Negative Control siRNA
(Qiagen) to determine the cell viability, respectively. First, the medium
was replaced with 100 μl plain RPMI 1640 medium and after 30 min of
incubation in the refreshed medium, 20 μl of the polyplexes were added
to the corresponding wells. After 2 h incubation at 37 °C and 5% CO2,
the medium/polyplex mixture was removed from the cells, cells were
washed once with 100 μl PBS and finally 100 μl completed medium
(with FBS) was added. The cells were incubated for 48 h at 37 °C and
5% CO2 before lysis of the cells. The cell lysates were treated with Lucif-
erase Assay Reagent (Promega) at ambient temperature and the lucifer-
ase activity was measured within the next 10 min, using a VICTOR3™
1420 multilabel counter (Perkin-Elmer) analyzed with Wallac 1420
Software. Lipofectamine2000 (Invitrogen) was used as reference and
transfection formulations were prepared according to manufacturer's
protocol with the same amount of siRNA as in the SS-PAA polyplexes.
2.6. Agarose gel retardation
Polyplex solutions at different polymer/DNA weight ratios were pre-
pared as described in section 2.5, followed by vortexing for 5 s and incu-
bation at ambient temperature for 30 min. DTT was added to a final
concentration of 2.5 mM to the appropriate samples to mimic intracellu-
lar reduction by glutathione. After another 30 min (with or without
DTT incubation), polyplexes were loaded on a 0.8% w/v agarose
gel containing 1.25 μM ethidium bromide. Electrophoresis was
performed for 60 min at 90 V in a TAE running buffer (40 mM
tris(hydroxymethyl)aminomethane, 20 mM acetic acid, 10 mM EDTA,
pH 8.0) supplemented with 1.25 μM ethidium bromide. After electro-
phoresis, pictures were taken on a Biorad Gel Doc 2000 under UV illumi-
nation and analyzed using Biorad Multi Analyst software version 1.1.
2.7. In vitro transfection and cell viability
The cell culture medium consisted of Dulbecco's Modified Eagle Medi-
um (Invitrogen) with 1 g/l D-glucose, completed with L-glutamine (1% v/v
of GlutaMAX 200 mM) and penicillin–streptomycin solution (2% v/v of
stock solution).
Transfection and cell viability studies were performed with COS-7
cells (SV-40 transformed African Green monkey kidney cells), HUH-6
cells (human hepatoblastoma cells) and H1299-Fluc (non-small cell
human lung carcinoma cell line stably transfected with firefly lucifer-
ase) in separate 96-well plates, n=4. Serum free conditions were
applied in order to observe directly the effects of interactions of the
boronic acid moieties with the cells, avoiding interference with plas-
ma proteins. Cells were pleated 24 h prior to transfection (COS-7
at 10,000 cells/well, HUH-6 at 8000 cells/well and H1299-Fluc at
12,000 cells/well) and were 60–80% confluent after 24 h. Plasmid
pCMV/GFP was used as the reporter gene. In all experiments two paral-
lel transfection series, one for the determination of reporter gene ex-
pression (GFP) and the other for the evaluation of the cell viability by
XTT assay, were carried out in separate 96-well plates. Different poly-
mer/DNA weight ratios were used to prepare the polyplexes. In a typical
transfection experiment, the cells were incubated with the desired
amount of polyplexes (100 μl dispersion or 110 μl after addition of 10 μl
a 10% (w/v) sorbitol or dextran solution, resulting in 1 μg plasmid DNA
per well) for 2 h at 37 °C in a humidified 5% CO2-containing atmosphere.
Next, the polyplex solution was removed, 100 μl of fresh culture medium
was added and the cells were cultured for another 48 h. Transfection
efficiencies (TE) were compared to the transfection of linear PEI
(Exgen 500, Fermentas) at its optimal 6/1 N/P ratio (TE was set at 1.0).
In all transfection studies p(CBA-ABOL/2AMPBA) was compared with
the non-boronated analog p(CBA-ABOL), that has shown excellent trans-
fection efficiencies in our previous studies [7]. The TE was determined by
measuring the GFP (RFP) expression by fluorescence spectroscopy, using
a TECAN Safire2 plate reader using SW Magellan Software V6.4 with
3. Results and discussion
3.1. Synthesis of p(CBA-ABOL/2AMPBA)
The aminobutyl monomer containing the boronic acid 2AMPBA was
obtained according to the synthesis depicted in Scheme 2. Schiff's con-
densation of o-formylphenylboronic acid with N-BOC-diaminobutane,
followed by reduction with sodium borohydride [15] gave the N-BOC-
protected precursor compound in 75% yield, which could be quantita-
tively converted into the monomer 2AMPBA.
Copolymer p(CBA-ABOL/2AMPBA) was prepared by reacting N,N′-
cystamine bisacrylamide with an 78/22 mixture of 4-aminobutanol
and 2AMPBA according Scheme 1B. After ultrafiltration, the copolymer
was obtained in 67% yield with a weight average molecular weight of
4.2 kDa, which is close to the Mw of 4.1 kDa of the p(CBA-ABOL) poly-
mer used in this study. The percentage of boronic acid containing side
groups was 23% according to 1H NMR, which is close to the feed ratio.
From the Mw it can be calculated that the polymer chains contain
ca. 12 repeating units; therefore on average each p(CBA-ABOL/2AMPBA)
polymer chain contains approximately three 2AMPBA moieties.
By comparing the gene delivery properties of p(CBA-ABOL/2AMPBA)
and p(CBA-ABOL), possible effects of the presence of the 2AMPBA groups
in the polymer can be evaluated. Binding of phenylboronic acids with

270
M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
NH2
NH2
O
NH
i.
ii.
O
O
OH
B
+
OH
NH
HN
O
OH
B
O
OH
NH
OH
B
OH
2AMPBA
Scheme 2. Synthesis of monomer 2AMPBA. i) Schiff's base formation in methanol followed by reduction with NaBH4 ii) deprotection with HCl (g) in methanol.
diols to form phenylboronic esters is generally optimal at the pH that is
equal to the pKa of the phenylboronic acids. The parent compound,
phenylboronic acid (PBA), has a pKa of 8.8, and therefore boronic ester
formation at physiological relevant pH is poor [19,20]. However in
2AMPBA the 2-aminomethyl group can form a dative B\N (or N→B)
bond that reduces the energy for the transition of the sp2 hybridized
boronic center to the sp3 hybridized boronate and this effect is beneficial
for diol binding and ester formation. As a result, polymer p(CBA-ABOL/
2AMPBA) can form reversible covalent esters with diols at physiological
pH, including carbohydrates at the cell surface. It was hypothesized that
this property would also affect the transfection properties of pDNA- and
siRNA-polyplexes formed with this polymer.
the side chains and unprotonated tertiary amines in the main chain
(Fig. 1, @4). The hydroxybutyl interaction can lead to the reversible
formation of boronic esters and boronate esters (Fig. 1, @5), resulting
in the dynamic covalent crosslinking in the spontaneously formed
nanoparticles.
Formation of polyplexes of pDNA and siRNA with cationic polymers
occurs primarily by electrostatic interactions. It is likely that in the
boronated polymer p(CBA-ABOL/2AMPBA) also intra- and intermolecular
forces as described above for the self-assembled nanoparticles contribute
to the stability of polyplexes formed by this polymer. In addition, in siRNA
polyplexes additional stabilization can occur by boronic ester formation
between the 2AMPBA groups and the 3′ ribose end groups of siRNA
(Fig. 1, @7). Fig. 3 shows that indeed for both DNA and siRNA
p(CBA-ABOL/2AMPBA) forms considerably smaller polyplexes than
p(CBA-ABOL) at the different weight ratios.
3.2. The buffer capacity of p(CBA-ABOL/2AMPBA) compared to
p(CBA-ABOL)
At 12/1 polymer/DNA weight ratio p(CBA-ABOL) forms relatively
large polyplexes (400 nm) that are not stable, but increase of the polymer
ratio to 24/1 and 48/1 results in improved DNA condensation with
polyplex sizes of ca. 200 and ca. 100 nm, respectively (Fig. 2A). However,
p(CBA-ABOL/2AMPBA) forms already at 12/1 ratio stable polyplexes of
ca. 150 nm size and the size decreases to ca. 100 nm at higher polymer/
DNA weight ratios. Similar good polyplex formation was observed for
p(CBA-ABOL/2AMPBA) polyplexes with siRNA (Fig. 2B). Increasing the
polymer/siRNA weight ratio from 12/1 to 48/1 leads to a decrease of the
particle size from ca. 240 nm to 100 nm with a narrow size distribution
(PDI=ca. 0.2) and a concomitant increase in zeta potential from 21 to
43 mV. In contrast, p(CBA-ABOL) polyplexes are much larger with a
size of ca. 185 nm and zeta potential of 18 mV at 48/1 weight ratio.
Fig. 2 furthermore shows that the zeta potentials of polyplexes of
p(CBA-ABOL/2AMPBA) are generally higher than those of p(CBA-ABOL).
The higher positive surface charge may be attributed to an extra contribu-
tion of protonated aminomethyl groups of 2-AMPBA in these polyplexes,
since it is conceivable that the vicinity of the negative nucleotide charges
the pKa of these groups shifts from ca. 5.3 [38] towards a value closer to
neutral pH.
Typically, SS-PAAs with tertiary amines in the polymer backbone
(Fig. 1, @1) have high buffer capacity in the pH range 7.4 to 5.1. For exam-
ple, p(CBA-ABOL) has a buffer capacity of 72% as determined by acid–base
titration [7], indicating that this polymer undergoes a large increase in the
degree of protonation upon endosomal acidification from pH 7.4 (extra-
cellular pH) to pH 5.1 (pH of late endosomes). Such strong increase
in polymer charge density is believed to cause endosomal membrane dis-
ruption [42]. Compared to p(CBA-ABOL), in the copolymer p(CBA-ABOL/
2AMPBA) 23% of the hydroxyl groups in the side chains are replaced by
2AMPBA groups. Anslyn et al. reported that the pKa of the secondary
amine group in 2-aminomethylphenylboronic acid is ca. 5.3, and the B–
N interaction raises the pKa of boronic acid (OH− addition) from ca. 8.8
in PBA to ca. 12 [38]. Consequently, protonation of the secondary amino
group of AMPBA contributes to the buffering action of the polymer,
whereas the boronic acid group does not add to the buffer capacity. The
buffer capacities as calculated from the pH titration curves (see Fig. S1,
supplementary content) are 69% for p(CBA-ABOL/2AMPBA) and 72% for
p(CBA-ABOL) which is in accordance to the expected contributions
based on the pKa values of the proton active groups in the polymer.
The DLS experiments further revealed that polyplex solutions from
p(CBA-ABOL/2AMPBA) with pDNA remained remarkably stable when
stored at 37 °C. Whereas in solutions at 24/1 polymer/DNA ratio the
polyplex size increases slowly from ca 105 to 185 nm after 4 days and
290 nm after 6 days, polyplexes at 48/1 ratio showed only a minor in-
crease from ca. 90 to 100 nm (Fig. 4). This stable polyplex formation is
an outstanding property of this polymer since polyplexes generally
tend to aggregate upon standing.
Since the 2AMPBA moieties in the polyplexes are expected to have
also affinity for carbohydrate binding, it was of interest to examine
whether addition of the weakly binding sugars D-glucose, D-mannose
and D-galactose, and in particular the more strongly binding D-sorbitol
[19,20] would influence the polyplex properties. Moreover, addition of
the polysaccharide dextran was included in this study to evaluate wheth-
er it was possible to decorate polyplexes of p(CBA-ABOL/2AMPBA) with a
carbohydrate layer by post-modification (Fig. 1, @6). Decoration of these
3.3. Particle formation and characterization
Solutions of p(CBA-ABOL) in the concentration range 0.18–
0.72 mg/ml in 20 mM HEPES buffer tend to aggregate upon standing
at ambient temperature, and eventually form a precipitate on a time-
scale of 4–10 h. Therefore always fresh solutions were prepared prior
to polyplex formation. However, p(CBA-ABOL/2AMPBA) under identi-
cal conditions forms stable solutions that are slightly opaque. Dynamic
light scattering (DLS) measurements revealed that self-assembled
nanoparticles were formed in this concentration range. These
nanoparticles were stable in time and have particle sizes of ca. 100 nm
with narrow size distribution (PDI=0.2–0.3) and zeta potentials of ca.
40 mV. As the boron atom in boronic acids is a Lewis acid that can bind
alcohols and amines, the stability of these nanoparticles is most likely
due to interactions of the AMPBA groups with hydroxybutyl groups in

M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
271
Fig. 3. Polyplex size of p(CBA-ABOL/2AMPBA) with pDNA at 24/1 and 48/1 polymer/DNA
weight ratio as a function of time in HEPES (pH 7.4, 20 mM, kept at 37 °C). The solid lines
represent the polyplexes in the absence of carbohydrate (dark squares 24/1 ratio, light
squares 48/1 ratio); the dotted lines represent polyplexes in the presence of 0.9% w/v
D-sorbitol (dark circles 24/1 ratio, light circles 48/1 ratio), and the dashed lines represent
polyplexes in the presence of 0.9% w/v dextran (dark triangles 24/1 ratio, light triangles
48/1 ratio).
of the polyplexes, it is expected that the coating of the polyplexes with
a dextran layer cause additional stabilization and prevent particle
growth. In Fig. 3 it can be seen that in the 24/1 polymer/DNA solution
dextran has a significant retarding effect on particle growth, which is
in favor of the particle coating possibility. Further support for the pos-
sibility of dextran coating was found in the change in zeta potential
from 32 mV to 25 mV for polyplexes upon addition of dextran to a
48/1 polymer/DNA solution.
The influence of addition of dextran to the polyplexes was further in-
vestigated with atomic force microscopy (AFM) and scanning electron
microscopy (SEM). In earlier AFM experiments it was observed that
polyplexes of disulfide-containing poly(amido amine)s have the tenden-
cy to spread out on a gold surface, whereas this does not occur on a mica
substrate. This behavior can most probably be attributed to the strong in-
teraction of the disulfide groups in the polymeric backbone with the gold
surface, causing the polyplexes to spread out. Here it was investigated
whether the presence of dextran could shield and stabilize the polyplexes
sufficiently from the disrupting interaction with the gold surface. In the
absence of dextran, a DNA polyplex solution of p(CBA-ABOL/2AMPBA)
at 48/1 ratio showed a sticky smear on the gold surface (Fig. 4A), whereas
in the presence of dextran more spherical polyplexes of 100–200 nm
were observed together with larger particles up to 2 μm (Fig. 4B). The
same samples were also probed by SEM. In the absence of dextran no
particles could be detected, but in the presence of dextran polyplexes of
ca 100–200 nm are visible in large quantity, besides larger particles of
1–2 μm that may be formed during the drying process (Fig. 4C).
Fig. 2. Particle size and zeta potential of polyplexes of p(CBA-ABOL) (black bars) and
p(CBA-ABOL/2AMPBA) (gray bars) with pDNA (A) and siRNA (B) in HEPES buffer (20 mM,
pH 7.4). The zeta potentials for p(CBA-ABOL) (triangles) and for p(CBA-ABOL/2AMPBA)
(squares) are shown in connecting lines for clarity.
polyplexes with carbohydrates would result in particles with altered cel-
lular interactions and could open new possibilities for glycotargeting. It
has been shown that both endosomal uptake and intracellular trafficking
of nucleic acids are influenced by glycosylation of polymeric vectors, and
that glycolysation of gene vectors has a beneficial effect on the transfec-
tion efficiency [33,35].
It was found that addition of 0.9% (w/v) of the weakly binding glu-
cose, mannose and galactose to a polyplex solution of p(CBA-ABOL/
2AMPBA) with pDNA at 24/1 and 48/1 ratio did not have an effect on
particle size and zeta potential (data not shown). However, at 24/1
ratio the addition of sorbitol to the polyplex solution showed a clearly
destabilizing effect, whereas addition of dextran has a stabilizing effect
on the polyplexes, as became manifest in a significantly accelerated and
decelerated particle growth, respectively (Fig. 3). The destabilizing ef-
fect of D-sorbitol supports the idea that binding of sorbitol to 2AMPBA
competes with the dynamic covalent crosslinking by hydroxybutyl
and tertiary amino groups with 2AMPBA in the polyplex.
3.4. Release of DNA from the polyplexes by reductive cleavage of disulfide
bonds in the polymer
At 48/1 ratio the polyplexes are stable in time both in the absence
and presence of carbohydrates, and addition of sorbitol has a negligible
effect on the particle stability. Addition of dextran slightly increases the
polyplex size (to ca 111 nm), but the size remains constant in the sub-
sequent 6 days. It can be rationalized that the competing effect of sorbi-
tol is more dominant at 24/1 polymer/DNA weight ratio than at the 48/1
ratio where a higher polymer concentration and thus a higher 2AMPBA
content is present. In principle, one could expect from dextran addition
to the polyplexes either a destabilizing or a stabilizing effect. If dextran
strongly interferes with the internal crosslinks in the polyplexes, parti-
cle destabilization and particle growth could be expected. On the other
hand, if dextran primarily interacts with 2AMPBA groups at the surface
Whereas stable condensation of DNA in the polyplex is needed in
the extracellular environment, rapid release of the nucleotide is re-
quired once the polyplexes have been arrived in the cytosol. The pres-
ence of the disulfide moieties in the polymer backbone (Fig. 1, @2)
allows for rapid cleavage of these linkages in a reductive environment
such as the cytosol, with vector unpacking and release of therapeutic
content as a result.
This process is mimicked for p(CBA-ABOL/2AMPBA)/DNA polyplexes
in the electrophoresis experiments shown in Fig. 5. Complete DNA retar-
dation is observed at polymer/DNA weight ratios 12/1, 24/1, and 48/1,
indicating that DNA is well packed in these polyplexes. However, after
treatment for 30 min with 2.5 mM DTT as the reducing agent, a clear

272
M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
release of DNA was observed. Also in the presence of the monosaccha-
rides and dextran identical results were obtained (data not shown).
Thus, although polyplexes of p(CBA-ABOL/2AMPBA possess very high sta-
bility in solution (Fig. 3), complete nucleotide release still occurs after
cleavage of the polymer chains by reduction of the disulfide linkages.
3.5. Transfection properties of p(CBA-ABOL/2AMPBA)polyplexes
The first studies to evaluate the effects of the presence of the
2AMPBA moieties in the polymer on transfection properties were
performed by comparing the transfection efficiencies and cytotoxic-
ities of pDNA polyplexes of p(CBA-ABOL/2AMPBA) and p(CBA-ABOL)
in the absence of serum. Polyplexes of p(CBA-ABOL) at 48/1 polymer/
DNA weight ratio yield approximately two times higher expression of
green fluorescent protein (GFP) in COS-7 cells than polyplexes of lin-
ear poly(ethylene imine) (Exgen500 at optimal N/P ratio of 6/1). For
p(CBA-ABOL/2AMPBA) the transfection efficiency under these condi-
tions was similar to that of poly(ethylene imine) (PEI) (Fig. 6). Only
minor cytotoxicity was observed (85–95% cell viability), exceeding
PEI under these conditions.
Besides COS-7 cells, also transfections in HUH-6 and H1299-Fluc cell
lines were investigated. All three cell lines are cancer cells, but are de-
rived from different organs, thereby expressing different glycoproteins
at the cell surface. COS-7 is a kidney derived cell line from simians
(green African monkey), HUH-6 is a human hepatoblastoma cell line,
and H1299-Fluc is a human lung carcinoma cell line which is stably ex-
pressing firefly luciferase. In Table 1 the transfection efficiencies are
shown for the three cell lines (COS-7, HUH-6, and H1299-Fluc) under
serum free conditions, relative to p(CBA-ABOL) (set to 100%).
Compared to p(CBA-ABOL), transfection efficiencies with p(CBA-
ABOL/2AMPBA) are ca. 50 ( 13) % for COS-7 cells, 21 ( 13) % for
HUH-6 cells, and 61 ( 11) % for H1299-Fluc cells. No significant cytotox-
icity was observed in any of the different cell lines (data not shown). The
lower transfection efficiency of p(CBA-ABOL/2AMPBA) compared to
p(CBA-ABOL) indicates that the 2AMPBA moieties have an inhibitory ef-
fect in one or more of the steps in the transfection process, and that the
effect is to a certain extent cell specific. Due to the carbohydrate binding
properties of the 2AMPBA groups, it may be hypothesized that these dif-
ferences are at least partly caused by binding of the polyplexes to carbo-
hydrates at the cell surface (Fig. 1, @6), which hampers cellular uptake
due to a limited access of the polyplexes to the cell membrane and/or
Fig. 5. Agarose gel retardation of free DNA (first lane) followed by polyplexes of
p(CBA-ABOL/2AMPBA) at 12/1 24/1 and 48/1 polymer/DNA weight ratio in the absence
of DTT (left) and after treatment with 2.5 mM DTT for 30 min (right).
endocytic uptake. It is noted that improved intracellular uptake and trans-
fection have been reported upon boronic acid functionalization of a neu-
tral Pluronic/glycerol block copolymer, which possesses by itself only low
intracellular uptake and transfection efficiency [27]. However, it cannot
be ruled out that at these low levels of transfection the relative increase
observed with the boronic acid functionalized polyplexes is due to less
complete removal of these polyplexes from the cells upon medium
change after incubation, compared to the neutral polyplexes.
3.6. Effects of the presence of carbohydrates on the transfection efficiency
In order to gain additional insight in the influence of the carbohy-
drate binding properties of p(CBA-ABOL/2AMPBA) polyplexes, further
transfection studies were performed in the presence of D-sorbitol and
dextran. In particular these carbohydrates have shown the highest af-
finity for the 2AMPBA moieties in the polyplex studies, and it can be
expected that these carbohydrates can compete with binding of the
polyplexes to the glycocalyx. Such competitive binding could result in
reduced binding to the glycocalyx and consequently increased mobility
of the boronic-acid modified polyplexes for intracellular uptake.
The influence of the presence of 0.9% (w/v) D-sorbitol and dextran
was examined in transfections with p(CBA-ABOL) and p(CBA-ABOL/
2AMPBA) polyplexes with a plasmid encoding for GFP at 48/1 polymer/
DNA weight ratio in COS-7 cells. Fig. 7 shows that in the absence of carbo-
hydrate, polyplexes of p(CBA-ABOL) generally give about three times
higher GFP expression than polyplexes of p(CBA-ABOL/2AMPBA). From
the fluorescence microscopy pictures (Fig. S2, supplementary content) it
was determined that the percentage of transfected cells by p(CBA-ABOL)
polyplexes was ca 38 ( 3) %. In contrast, polyplexes from p(CBA-ABOL/
2AMPBA) yielded only ca. 5 ( 1) % transfected cells under these condi-
tions. However, addition of D-sorbitol or dextran to the p(CBA-ABOL/
2AMPBA) polyplexes gave an increase of the percentage of positive cells
to ca 26 ( 5) % and a concomitant increase in transfection efficiency.
This indicates that binding of D-sorbitol and dextran with 2AMPBA moie-
ties mitigates the inhibitory binding of the polyplexes to glycoproteins of
the cells, resulting in higher cell internalization and higher transfection ef-
ficiency. The addition of the carbohydrates to p(CBA-ABOL) polyplexes
shows a slightly promoting effect on the transfection efficiency, despite
it even seems to decrease the number of transfected cells (Fig. S2). How-
ever, overinterpretation of this observation has to be avoided since the ef-
fects are close within the error margins. In all cases the addition of the
carbohydrates to the transfection medium did not significantly affect
the cell viabilities, which remained 90–100% (data not shown).
A
B
C
40 nm
20 nm
0 nm
In order to obtain more information whether dextran is functioning
as a coating of the boronated polyplexes, or is only co-transported to-
gether with the polyplexes into the cells, additional experiments
were performed with FITC-labeled dextran and polyplexes with
pDNA encoding for red fluorescent protein (RFP).
Fluorescence microscopy pictures of COS-7 cells incubated for 20 min
with p(CBA-ABOL) polyplexes in the presence of FITC-dextran in the me-
dium did not show fluorescent dextran inside the cells (Fig. S3, supple-
mentary content). The cells appear as dark non-fluorescent areas in a
homogeneous fluorescent background (intensity 77 1 a.u.) originating
1 μm
1 μm
5 μm
Fig. 4. Images of p(CBA-ABOL/2AMPBA)/DNA polyplexes at 48/1 weight ratio in the ab-
sence (A) and presence (B,C) of dextran. Panel A displays a 3×3-μm height image by
AFM of polyplexes in the absence of dextran, where no particles are visible but “sticky
layers” are formed on the surface. Panel B shows a 10×10-μm height image by AFM in
the presence of dextran where polyplexes of 100–200 nm are observed together with
larger particles up to 2 μm. Panel C gives a SEM picture of the polyplexes in the pres-
ence of dextran, where particles of 100–200 nm can be observed together with larger
particles of 1–2 μm. Scale bar is 1 μm in A and B and 5 μm in C.

M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
273
Fig. 7. GFP expression in COS-7 cells after transfection with p(CBA-ABOL) or p(CBA-
ABOL/2AMPBA) polyplexes at a 48/1 polymer/DNA weight ratio. Fluorescent intensi-
ties (bars) and the corresponding percentage of GFP positive cells (squares) resulting
from GFP expression in the absence and presence of 0.9% of D-sorbitol and dextran
are given, respectively.
Fig. 6. Transfection efficiency of polyplexes p(CBA-ABOL) (black bars) and p(CBA-
ABOL/2AMPBA) (gray bars) at 24/1 and 48/1 polymer/DNA weight ratio in COS-7
cells, relative to linear PEI (Exgen at N/P=6). Corresponding cell viabilities are
shown in symbols with connecting lines for clarity for p(CBA-ABOL) (triangles) and
for p(CBA-ABOL/2AMPBA) (squares). Cell viability for PEI is included (circles).
diol group at the 3′ ribose that the 2AMPBA groups can siRNA [39–41]
can play a role. Fig. 8 shows that anti-luciferase siRNA polyplexes of
p(CBA-ABOL) at 48/1 weight ratio gave a knockdown of ca. 28% of lucif-
erase expression in H1299-Fluct cells with good cell viability. The pres-
ence of D-sorbitol seems to induce a slight increase of gene silencing,
but more remarkable is that the presence of dextran in the transfection
medium strongly promotes the knockdown efficiency to 76%, while
maintaining full cell viability. Similarly, the gene silencing efficiency of
p(CBA-ABOL/2AMPBA) in the absence of the carbohydrates is 35%,
which increases to 59% in the presence of sorbitol and to 76% in the
presence of dextran, respectively, again without loss of cell viability.
From these results it can be concluded that the inhibitory effect of
the 2AMPBA groups, as was noticed in the uptake of DNA polyplexes,
is more than compensated in the case of siRNA delivery and gene si-
lencing. Most likely this can be attributed to the higher stability of
the p(CBA-ABOL/2AMPBA) polyplexes, due to the dynamic covalent
crosslinking of the Lewis bases and the 3′ ribose end groups of the
incorporated siRNA's. DLS experiments showed that polyplexes of
p(CBA-ABOL/2AMPBA) at 24/1 and 48/1 polymer/siRNA ratio were
still stable after storage at 37 °C for 4 days, whereas siRNA polyplexes
of p(CBA-ABOL), even at 48/1 ration appeared not to be stable under
from the FITC–dextran in the medium. In contrast, p(CBA-ABOL/
2AMPBA) polyplexes in the presence of FITC–dextran under identical
conditions appeared as highly fluorescent particles outside the cells,
while the intensity of the fluorescent background was decreased to
48 1 a.u.
This clearly indicates that the p(CBA-ABOL/2AMPBA) polyplexes
interact with dextran under the formation of condensed fluorescent
particles. After 2 h of incubation with the polyplex medium, and subse-
quent washing with PBS and replacement with fresh culture medium,
the cells were allowed for further transfection for 48 h. No fluorescent
FITC signal was then apparent in the cell samples treated with
p(CBA-ABOL)-dextran, indicating that all FITC–dextran was removed
by the washing step with PBS and was not internalized by the cells.
Moreover, a parallel experiment under the same conditions showed
that also free FITC–dextran (without the presence of polyplexes) was
not internalized in the cells. In contrast, for p(CBA-ABOL/2AMPBA)
polyplexes with FITC–dextran clearly fluorescent particles could be ob-
served (Fig. S4). Pictures taken of cross-sections based on a z-stack
(Fig. S5) indicate that the particles were predominantly present inside
the cells. These results indicate that dextran becomes only internalized
when bound to the p(CBA-ABOL/2AMPBA) polyplexes.
3.7. Luciferase knockdown with siRNA polyplexes in the presence of D-
sorbitol and dextran
The polymers p(CBA-ABOL) and p(CBA-ABOL/2AMPBA) were also
evaluated for their ability to deliver siRNA for gene silencing, both in
the absence and in the presence of D-sorbitol and dextran. Besides the
differences in interaction of the two types of polyplexes with the cellu-
lar environment as observed in DNA transfection, in the siRNA gene si-
lencing also the reversible covalent binding of siRNA in p(CBA-ABOL/
2AMPBA) polyplexes by formation of a boronate ester with the vicinal
Table 1
Transfection efficiency (%) of p(CBA-ABOL/2AMPBA) polyplexesa.
Polymer\cell line
COS-7 cells
HUH-6 cells
H1299-Fluc cells
61 11%b
a
a
p(CBA-ABOL/2AMPBA)
50 13%
21 13%
Fig. 8. Luciferase expression in H1299-Fluc cells after anti-luciferase siRNA administration
with polyplexes of p(CBA-ABOL) (black bars) and p(CBA-ABOL/2AMPBA) (gray bars) at
48/1 polymer/siRNA weight ratio, relative to similar treatment with non-coding siRNA
samples (black and gray triangles, respectively). Lipofectamine (LF) is included as positive
control.
a
Relative to transfection efficiency of p(CBA-ABOL) polyplexes; GFP expression of
p(CBA-ABOL) polyplexes under identical conditions was set to 100%.
b
Gene expression (GFP) of different cells, measured 48 h after incubation for 2 h
with polyplexes of p(CBA-ABOL/2AMPBA) at a 48/1 polymer/DNA weight ratio.

274
M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
these conditions. In previous work we have incorporated additional cat-
ionic segments in the main chain of p(CBA-ABOL) to modify this poly-
mer suitable for siRNA delivery [43].
References
[1] I.M. Verma, N. Somia, Gene therapy — promises, problems and prospects, Nature
389 (1997) 239–242.
[2] S.Y. Wong, J.M. Pelet, D. Putnam, Polymer systems for gene delivery-past, present,
and future, Prog. Polym. Sci. 32 (2007) 799–837.
[3] E. Wagner, J. Kloeckner, Gene delivery using polymer therapeutics, Adv. Polym.
Sci. 192 (2006) 135–173.
4. Conclusion
[4] S.B. Zhang, Y.M. Xu, B. Wang, W.H. Qiao, D.L. Liu, Z.S. Li, Cationic compounds used
in lipoplexes and polyplexes for gene delivery, J. Control. Release 100 (2004)
165–180.
[5] J. Luten, C.F. van Nostruin, S.C. De Smedt, W.E. Hennink, Biodegradable polymers as
non-viral carriers for plasmid DNA delivery, J. Control. Release 126 (2008) 97–110.
[6] T.G. Park, J.H. Jeong, S.W. Kim, Current status of polymeric gene delivery systems,
Adv. Drug Deliv. Rev. 58 (2006) 467–486.
[7] C. Lin, Z.Y. Zhong, M.C. Lok, X.L. Jiang, W.E. Hennink, J. Feijen, J.F.J. Engbersen, Novel
bioreducible poly(amido amine)s for highly efficient gene delivery, Bioconjug.
Chem. 18 (2007) 138–145.
[8] C. Lin, C.J. Blaauboer, M.M. Timoneda, M.C. Lok, M. van Steenbergen, W.E. Hennink, Z.Y.
Zhong, J. Feijen, J.F.J. Engbersen, Bioreducible poly(amido amine)s with oligoamine side
chains: synthesis, characterization, and structural effects on gene delivery, J. Control.
Release 126 (2008) 166–174.
[9] C. Lin, J.F.J. Engbersen, Effect of chemical functionalities in poly(amido amine)s
for non-viral gene transfection, J. Control. Release 132 (2008) 267–272.
[10] C. Lin, Z.Y. Zhong, M.C. Lok, X.J. Jiang, W.E. Hennink, J. Feijen, J.F.J. Engbersen, Ran-
dom and block copolymers of bioreducible poly(amido amine)s with high- and
low-basicity amino groups: study of DNA condensation and buffer capacity on
gene transfection, J. Control. Release 123 (2007) 67–75.
[11] C. Lin, Z.Y. Zhong, M.C. Lok, X.L. Jiang, W.E. Hennink, J. Feijen, J.F.J. Engbersen,
Linear poly(amido amine)s with secondary and tertiary amino groups and vari-
able amounts of disulfide linkages: synthesis and in vitro gene transfer proper-
ties, J. Control. Release 116 (2006) 130–137.
[12] L.V. Christensen, C.W. Chang, W.J. Kim, S.W. Kim, Z.Y. Zhong, C. Lin, J.F.J. Engbersen, J.
Feijen, Reducible poly(amido ethylenimine)s designed for triggered intracellular
gene delivery, Bioconjug. Chem. 17 (2006) 1233–1240.
[13] M.A. Mateos-Timoneda, M.C. Lok, W.E. Hennink, J. Feijen, J.F.J. Engbersen,
Poly(amido amine)s as gene delivery vectors: effects of quaternary nicotinamide
moieties in the side chains, ChemMedChem 3 (2008) 478–486.
[14] M. Piest, C. Lin, M.A. Mateos-Timoneda, M.C. Lok, W.E. Hennink, J. Feijen, J.F.J.
Engbersen, Novel poly(amido amine)s with bioreducible disulfide linkages in their
diamino-units: structure effects and in vitro gene transfer properties, J. Control.
Release 130 (2008) 38–45.
A multifunctional poly(amido amine), p(CBA-ABOL/2AMPBA),
was developed containing tertiary amines that are protonable in
the endosomal acidification range, disulfide groups that are cleavable
in the reducing intracellular environment, hydroxybutyl groups that in-
teract with the endosomal membrane, and 2-aminomethylboronic acid
groups that reversibly interact with amino and hydroxyl groups in
nanoparticles and polyplexes and form boronic esters with vicinal diol
groups as are present in e.g. carbohydrates. The boronic acid groups in
p(CBA-ABOL/2AMPBA) stabilize self-assembled nanoparticles and DNA-
and siRNA-polyplexes formed at physiological pH, due to reversible
crosslink formation between the Lewis acidic 2AMPBA group and neigh-
boring Lewis bases. This architectural picture is supported by the fact
that D-sorbitol, that can compete with the crosslinking groups for bind-
ing to 2AMPBA, destabilizes DNA polyplexes of p(CBA-ABOL/2AMPBA)
at 24/1 polymer/DNA ratio. The high stability of p(CBA-ABOL/2AMPBA)
polyplexes with pDNA (at least 6 days at 37 °C) and with siRNA (at
least 4 days at 37 °C) is a remarkable and favorable property of these
polyplexes. The stability can be further enhanced by self-assembled dec-
oration of dextran to the surface of these polyplexes. Since p(CBA-ABOL)
polyplexes gave higher transfection than p(CBA-ABOL/2AMPBA)
polyplexes in three different cell lines tested, COS-7, HUH-6, and
H1299-Fluc, it can be concluded that the presence of the 2AMPBA
moieties has an overall unfavorable effect on the DNA transfection
efficiency. Most probably this can be primarily attributed to binding
of the p(CBA-ABOL/2AMPBA) polyplexes to glycoproteins on the sur-
face of the cells, thereby hampering the cellular uptake of the
polyplexes. Support for this explanation was found in the observation
that the presence of D-sorbitol and dextran in the polyplex medium in-
creases the transfection efficiency. Then, available 2AMPBA groups bind
with sorbitol and dextran to form boronic esters, which reduces the
binding affinity of the polyplexes with the glycocalix of the cells. The re-
sults furthermore show that polyplexes of siRNA with p(CBA-ABOL/
2AMPBA) are more effective delivery vectors in gene silencing than
polyplexes from p(CBA-ABOL). This can be attributed to the higher sta-
bility of the p(CBA-ABOL/2AMPBA) polyplexes with their dynamical
crosslinking boronated moieties. The knockdown of gene expression
by the siRNA polyplexes could be enhanced by the addition of sorbitol,
and in particular dextran, to the polyplex solution. For polyplexes of
p(CBA-ABOL/2AMPBA) this effect can be explained by reduced binding
interactions with the glycocalix, however the origin of this effect is less
clear for p(CBA-ABOL) polyplexes which are lacking specific binding
sites for carbohydrates. The enhanced gene silencing together with the
excellent stability of siRNA polyplexes with p(CBA-ABOL/2AMPBA), as
well as their good possibilities for carbohydrate decoration, makes this
multifunctional polymer a very promising candidate for siRNA delivery.
[15] M. Piest, J.F.J. Engbersen, Role of boronic acid moieties in disulfide containing
poly(amido amine)s for non-viral gene delivery, J. Control. Release 155 (2011)
331–340.
[16] D.G. Hall, Boronic Acids — Preparation and Applications in Organic Synthesis and
Medicine, John Wiley & Sons, 2005.
[17] T.D. James, M.D. Phillips, S. Shinkai, Boronic Acids in Saccharide Recognition,
Royal Society of Chemistry, Cambridge, 2006.
[18] J. Yan, H. Fang, B.H. Wang, Boronolectins and fluorescent boronolectins: an examina-
tion of the detailed chemistry issues important for the design, Med. Res. Rev. 25
(2005) 490–520.
[19] G. Springsteen, B. Wang, A detailed examination of boronic acid–diol complexa-
tion, Tetrahedron 58 (2002) 5291–5300.
[20] J. Yan, G. Springsteen, S. Deeter, B.H. Wang, The relationship among pK(a), pH,
and binding constants in the interactions between boronic acids and diols — it
is not as simple as it appears, Tetrahedron 60 (2004) 11205–11209.
[21] A.E. Ivanov, J. Eccles, H.A. Panahi, A. Kumar, M.V. Kuzimenkova, L. Nilsson, B.
Bergenståhl, N. Long, G.J. Phillips, S.V. Mikhalovsky, I.Y. Galaev, B. Mattiasson,
Boronate-containing polymer brushes: characterization, interaction with saccha-
rides and mammalian cancer cells, J. Biomed. Mater. Res. A 88A (2009) 213–225.
[22] N.D. Winblade, H. Schmokel, M. Baumann, A.S. Hoffman, J.A. Hubbell, Sterically
blocking adhesion of cells to biological surfaces with a surface-active copolymer
containing poly(ethylene glycol) and phenylboronic acid, J. Biomed. Mater. Res.
59 (2002) 618–631.
[23] W.Q. Yang, X.M. Gao, B.H. Wang, Boronic acid compounds as potential pharmaceutical
agents, Med. Res. Rev. 23 (2003) 346–368.
[24] S. Moffatt, S. Wiehle, R.J. Cristiano, Tumor-specific gene delivery mediated by a novel
peptide–polyethylenimine–DNA polyplex targeting aminopeptidase N/CD13, Hum.
Gene Ther. 16 (2005) 57–67.
[25] G.A. Ellis, M.J. Palte, R.T. Raines, Boronate-mediated biologic delivery, J. Am. Chem.
Soc. 134 (2012) 3631–3634.
[26] Q. Peng, F. Chen, Z. Zhong, R. Zhuo, Enhanced gene transfection capability of
polyethylenimine by incorporating boronic acid groups, Chem. Commun. 46
(2010) 5888–5890.
[27] F. Chen, Z. Zhang, M. Cai, X. Zhang, Z. Zhong, R. Zhuo, Phenylboronic-acid-modified
amphiphilic polyether as a neutral gene vector, Macromol. Biosci. 12 (2012) 962–969.
[28] M.R. Dreher, W. Liu, C.R. Michelich, M.W. Dewhirst, F. Yuan, A. Chilkoti, Tumor
vascular permeability, accumulation, and penetration of macromolecular drug
carriers, J. Natl. Cancer Inst. 98 (2006) 335–344.
Acknowledgments
The authors thank the Netherlands Organization for Scientific Re-
search (NOW) for the financial support in the framework of their ECHO
program, Anika Embrechts for her help with the AFM measurements,
Tom Groothuis for his help with the confocal microscopy measurements,
and Anouk Mentink-Leusink and Hemant V. Unadkat for their help with
the automated fluorescence microscopy.
[29] A. Matsumoto, H. Cabral, N. Sato, K. Kataoka, Y. Miyahara, Assessment of tumor
metastasis by the direct determination of cell-membrane sialic acid expression,
Angew. Chem. Int. Ed. 49 (2010) 5494–5497.
Appendix A. Supplementary data
[30] A. Mahalingam, A.R. Geonnotti, J. Balzarini, P.F. Kiser, Activity and safety of syn-
thetic lectins based on benzoboroxole-functionalized polymers for inhibition of
HIV entry, Mol. Pharm. 8 (2011) 2465–2475.
Supplementary data to this article can be found online at http://
dx.doi.org/10.1016/j.jconrel.2013.02.008.

M. Piest et al. / Journal of Controlled Release 169 (2013) 266–275
275
[31] A. Varki, R.D. Cummings, J.D. Esko, H.H. Freeze, P. Stanly, C.R. Bertozzi, G.W. Hart,
Essentials of Glycobiology, 2nd ed. Cold Spring Harbor Laboratory Press, Cold
Spring harbor, NY, 2009.
[32] H. Zhang, Y. Ma, X.L. Sun, Recent developments in carbohydrate-decorated
targeted drug/gene delivery, Med. Res. Rev. 30 (2010) 270–289.
[33] H. Yan, K. Tram, Glycotargeting to improve cellular delivery efficiency of nucleic
acids, Glycoconj. J. (2007) 107–123.
[38] S.L. Wiskur, J.J. Lavigne, H. Ait-Haddou, V. Lynch, Y.H. Chiu, J.W. Canary, E.V.
Anslyn, pK(a) values and geometries of secondary and tertiary amines complexed
to boronic acids — Implications for sensor design, Org. Lett. 3 (2001) 1311–1314.
[39] B. Elmas, M.A. Onur, S. Senel, A. Tuncel, Temperature controlled RNA isolation by
N-isopropylacrylamide-vinylphenyl boronic acid copolymer latex, Colloid Polym.
Sci. 280 (2002) 1137–1146.
[40] E. Uguzdogan, H. Kayi, E.B. Denkbas, S. Patir, A. Tuncel, Stimuli-responsive properties
of aminophenylboronic acid-carrying thermosensitive copolymers, Polym. Int. 52
(2003) 649–657.
[34] B.G. Davis, M.A. Robinson, Drug delivery systems based on sugar-macromolecule
conjugates, Curr. Opin. Drug Discov. Devel. 5 (2002) 279–288.
[35] M. Monsigny, P. Midoux, R. Mayer, A.C. Roche, Glycotargeting: influence of the
sugar moiety on both the uptake and the intracellular trafficking of nucleic acid
carried by glycosylated polymers, Biosci. Rep. 19 (1999) 125–132.
[36] R. Namgung, J.H. Brumbach, J.H. Jeong, J.W. Yockman, S.W. Kim, C. Lin, Z. Zhong, J.
Feijen, J.F.J. Engbersen, W.J. Kim, Dual bio-responsive gene delivery via reducible
poly(amido amine) and survivin-inducible plasmid DNA, Biotechnol. Lett. 32
(2010) 755–764.
[37] D. Vercauteren, M. Piest, L.J. van der Aa, M. Al Soraj, A.T. Jones, J.F.J. Engbersen, S.C. De
Smedt, K. Braeckmans, Flotillin-dependent endocytosis and a phagocytosis-like
mechanism for cellular internalization of disulfide-based poly(amido amine)/DNA
polyplexes, Biomaterials 32 (2011) 3072–3084.
[41] A.R. Martin, I. Barvik, D. Luvino, M. Smietana, J.J. Vasseur, Dynamic and program-
mable DNA-templated boronic ester formation, Angew. Chem. Int. Ed. 50 (2011)
4193–4196.
[42] M. Jager, S. Schubert, S. Ochrimenko, D. Fischer, U.S. Schubert, Branched and linear
poly(ethylene imine)-based conjugates: synthetic modification, characterization,
and application, Chem. Soc. Rev. 41 (2012) 4755–4767.
[43] L.J. van der Aa, P. Vader, G. Storm, R.M. Schiffelers, J.F.J. Engbersen, Optimization of
poly(amido amine)s as vectors for siRNA delivery, J. Control. Release 150 (2011)
177–186.