Robust polymeric nanoparticles for the delivery of aminoglycoside antibiotics using carboxymethyldextran-b-poly(ethyleneglycols) lightly grafted with n-dodecyl groups

Article abstract of DOI:10.1039/c0sm00316f

Aminoglycoside antibiotics are effective in the treatment of infections caused by aerobic Gram negative bacilli, but their widespread use is hampered by serious side effects that may be alleviated through the use of tailored delivery systems. Robust polyi

Full text of DOI:10.1039/c0sm00316f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/softmatter | Soft Matter
Robust polymeric nanoparticles for the delivery of aminoglycoside antibiotics
using carboxymethyldextran-b-poly(ethyleneglycols) lightly grafted with
n-dodecyl groups†
a
Ghareb M. Soliman, Janek Szychowski, Stephen Hanessian and Franc¸oise M. Winnik
b
b
ab
*
*
Received 2nd May 2010, Accepted 23rd June 2010
DOI: 10.1039/c0sm00316f
Aminoglycoside antibiotics are effective in the treatment of infections caused by aerobic Gram negative
bacilli, but their widespread use is hampered by serious side effects that may be alleviated through the
use of tailored delivery systems. Robust polyion complex (PIC) micelles, incorporating up to 50 weight
% drug, were prepared using two aminoglycosides: paromomycin and neomycin, and
a dihydrophilic block copolymer consisting of a poly(ethyleneglycol) (PEG) chain linked to
a carboxymethyldextran fragment (CMD) lightly grafted with n-dodecyl groups. The micelles were
stable under physiological conditions (pH 7.4, 150 mM NaCl), in contrast to micelles formed by the
unmodified CMD-PEG and the aminoglycosides or their guanidinylated derivatives. The
aminoglycosides were released from the n-dodecyl-CMD-PEG micelles in a pharmacologically active
form as indicated by their ability to kill test micro-organisms in culture. This study opens up new
opportunities in the biomedical applications of PIC micelles with inherently enhanced stability.
small size, and narrow size distribution. They are thermody-
namically stable and resist disintegration upon dilution as long as
their concentration exceeds their critical association concentra-
tion (CAC), which usually is very low. PIC micelles often have
a poly(ethyleneglycol) (PEG) corona, which prolongs their
circulation time in blood allowing them to accumulate passively
in tissues of leaky vasculature, such as tumors and inflamed
tissues.13,14
1. Introduction
The intelligent self-assembly of functional polymers is used
extensively in the fabrication of soft materials and devices with
applications in medicine, pharmaceutics, cosmetics, biosensing,
and food technology.1–3 Much thought needs to be given to the
design of the macromolecules themselves, so that, in a given
environment, they associate in a predictable and repetitive
manner. Although recent advances in polymer synthesis, in
particular the advent of controlled free radical polymerization,
allow one to synthesize at will complex macromolecular archi-
tectures, simple diblock copolymers still remain the key building
blocks of soft nanoparticles.4–7 Dihydrophilic copolymers which
consist of two suitable water-soluble blocks readily self-assemble
in water in response to external stimuli, such as a change of
temperature, pH, salinity or a light impulse. If the copolymer has
a neutral hydrophilic block and an ionic block, micellization can
also be triggered by interaction with an oppositely charged
compound. The resulting nanoparticles, known as polyion
complex (PIC) micelles, have a core-shell morphology: the core
entraps charged guests neutralized by the oppositely charged
blocks, while the shell is formed by the neutral blocks which
confer colloidal stability to the nanoparticles via steric interac-
tions.8–10 PIC micelles are effective delivery vehicles for charged
hydrophilic compounds, such as drugs, proteins, plasmid DNA,
and siRNA.5,11,12 They present a number of advantages for
biomedical applications, including ease of fabrication, excellent
colloidal stability in aqueous media, high drug loading capacity,
We reported recently the use of the dihydrophilic anionic
copolymer carboxymethyldextran-b-PEG (CMD-PEG) (Fig. 1)
as host in the formation of PIC micelles with a number of
cationic drugs.15,16 CMD-PEG copolymers are non-toxic and
yield PIC micelles of small size, high drug loading capacity and
are stable upon freeze drying and in presence of serum proteins.16
These attractive features prompted us to evaluate CMD-PEG
copolymers as delivery agents for aminoglycosides, a group of
structurally diverse polyamines frequently used in the treatment
of serious infections caused by aerobic Gram negative bacilli,
such as pneumonia, urinary tract infections and peritonitis.17,18
The antibacterial activity of aminoglycosides results from their
interaction with prokaryotic ribosomal RNA (rRNA).19 Unfor-
tunately, aminoglycosides therapy is associated with a host of
side-effects due to drug accumulation in healthy non-target
tissues resulting in nephrotoxicity and ototoxicity. Moreover,
aminoglycosides are mostly administered parenterally or locally,
since their poor absorption in the gastro-intestinal tract as
a consequence of their highly polar cationic nature prevents their
oral administration.20,21
In view of the clinical importance of aminoglycosides, much
effort has been directed towards their encapsulation into vectors
to modify their biodistribution, limit their toxicity, and increase
their oral bioavailability. Drug delivery systems tested include
liposomes,22–24 polymeric nanoparticles,25 solid lipid nano-
particles26 and polyelectrolyte multilayers.27 Although liposome-
encapsulated aminoglycosides showed enhanced activity against
a
ꢀ
ꢀ
Faculty of Pharmacy, Universite de Montreal, CP 6128 Succ. Centre
Ville, Montreꢀal, QC, Canada H3C 3J7. E-mail: francoise.winnik@
umontreal.ca; Tel: +(1) 514 340 5179
bDepartment of Chemistry, Universiteꢀ de Montreꢀal, CP 6128 Succ. Centre
Ville, Montreꢀal, QC, Canada H3C 3J7
† Electronic supplementary information (ESI) available: Further
information. See DOI: 10.1039/c0sm00316f
4504 | Soft Matter, 2010, 6, 4504–4514
This journal is ª The Royal Society of Chemistry 2010

View Article Online
In our system, however, the critical ionic strength was too low
for eventual medical applications and we were brought to modify
the design of the aminoglycoside PIC micelles in order to
enhance their resistance against increased salinity. We present
here the implementation of two strategies towards this goal. One
approach was to enhance the ionic interactions between the
aminoglycosides and the CMD-PEG carboxylates by intro-
ducing a basic functional group in selected positions in the
antibiotic structure. The guanidine (Gua) moiety (pKa of gua-
nidinium ¼ 12.5) was selected in order to increase the basicity of
the aminoglycoside.32 The second strategy was to strengthen the
core of the micelles via non-covalent interactions between
hydrophobic groups linked to the charged block of the CMD-
PEG. The formation and stability of the two types of PIC
1
micelles were assessed by H NMR spectroscopy, dynamic light
scattering, and turbidimetry. The kinetics of drug release from
the PIC micelles were determined by the dialysis bag method.
In vitro experiments confirmed that the drugs retained their
ability to kill test micro-organisms upon entrapment in the core
of PIC micelles. PIC micelles stabilization through hydrophobic
modification of CMD-PEG could be applied to other PIC
micelles-forming ionic copolymers. This is expected to widen the
biomedical utility of PIC micelles since many of their drug-
loaded formulations are unstable under physiological
conditions.33,34
Fig. 1 Chemical structures of neomycin, paromomycin, guanidinylated
paromomycins (top) and the CMD-PEG block copolymer (bottom).
2. Results and discussion
resistant strains of Pseudomonas aeruginosa due to enhanced
entry into the bacterial cell, their use suffers from several draw-
backs, in particular their limited stability in the presence of blood
lipoproteins, low encapsulation efficiency, osmotic fragility and
poor storage stability.28,29 Poly(lactic-co-glycolic acid) (PLGA)
nanoparticles were used for the encapsulation of gentamicin, but
their drug loading efficiency was very low (ꢀ 1 wt%).25
Gentamicin microspheres, although effective in reducing splenic
infections in mice, triggered pulmonary embolism due to
particles aggregation.30
Synthesis and chemical characterization of the copolymers and
the guanidinylated aminoglycosides
As previously reported, the dihydrophilic copolymer CMD-PEG
was synthesized by end-coupling of u-dextran-lactone with an
The objectives of this study were to assemble and characterize
PIC micelles of CMD-PEG and two aminoglycoside antibiotics:
neomycin and paromomycin (Fig. 1), which are examples of
4,5-disubstituted 2-deoxystreptamine aminoglycosides. Their
structures differ only by the substituent at the 60 position, which
is a hydroxyl group in paromomycin and an amino group in
neomycin. Neomycin and paromomycin are positively charged at
pH 7.4, making them suitable for PIC micelles formation with
polyanions, such as CMD-PEG. Evidence from isothermal
titration calorimetry (ITC), 1H NMR spectroscopy, and dynamic
light scattering indicated that PIC micelles of CMD-PEG and
aminoglycosides form in water or buffers of low salinity.
However, to our disappointment, CMD-PEG/aminoglycoside
PIC micelles exhibited poor stability against increased ionic
strength, severely hampering their in vivo applications. Of course,
all PIC micelles dissociate above a critical ionic strength, since
both the chemical potential of the counterions and the coulombic
forces between the ionic block and the charged guest are directly
influenced by the external ionic strength. This ionic strength
dependent association/dissociation of PIC micelle is useful in
some cases. For instance, it has been applied as an on-off control
of enzymatic activity.31
Scheme 1 Synthesis of Dod-CMD-PEG. Reagents and conditions:
(a) EEDQ, 1 : 1 v/v ethanol/water, pH 4.0, 30 min, (b) n-dodecylamine,
pH 9.0, 60 min, (c) soxhlet extraction with hexane and (d) cation
exchange resin, water.
This journal is ª The Royal Society of Chemistry 2010
Soft Matter, 2010, 6, 4504–4514 | 4505

View Article Online
Table 1 Molecular characteristics of the copolymers prepared
excess of a-amino-u-methoxy-PEG, followed by carboxy-
methylation of the dextran block and extensive purification to
remove unreacted a-amino-u-methoxy-PEG.35 This polymer
was used as such for the preparation of PIC micelles or converted
to Dod-CMD-PEG by reaction with n-dodecylamine in aqueous
ethanol in the presence of N-ethoxycarbonyl-2-ethoxy-1,2-dihy-
droquinoline (EEDQ) acting as coupling agent (Scheme 1).36 The
successful outcome of the synthesis was ascertained from
the FTIR spectrum of the Dod-CMD-PEG copolymers, in
a
a
Mw
(g molꢁ1
Mn
(g molꢁ1
%
Polymer
)
)
xb yc md ne
dodecylf
CMD-PEG
Dod18-CMD-PEG
Dod38-CMD-PEG
14 800
16 100
17 600
10 800
12 100
13 600
0
7
34 40 140
27 40 140
—
18
38
15 19 40 140
a
From GPC measurements in aqueous NaNO3 (0.2 M)/NaH2PO4
(0.01 M)/NaN3 (0.8 mM); pH 7.02. Number of glucopyranose units
b
c
bearing
an
carboxymethylated glucopyranose units.
n-dodecyl
chain.
Number
of
Average number of
unmodified
particular the appearance of bands at 1546 cmꢁ1 and 1644 cmꢁ1
,
d
e
glucopyranosyl repeat units of the CMD chain. Number of –CH2-
attributed to the amide II and amide I vibration modes (Fig. 2A).
The 1H NMR spectrum of the modified copolymers in DMSO-d6
presented signals at 1.18 and 0.78 ppm due, respectively, to the
methylene and methyl protons of the n-dodecyl groups, in
f
CH2-O- repeat units of PEG. Mole percent of n-dodecyl-modified
glucopyranose units (1H NMR measurements in DMSO-d6).
addition to signals characteristic of the CMD block (d 3.61–3.78,
4.21–4.32, 4.68 and 4.94–5.01 ppm) and of the PEG block
(d 3.51 ppm) (Fig. 2B). The n-dodecyl grafting density was
1
calculated from the H NMR spectra of the copolymers. The
chemical characteristics of the copolymers are listed in Table 1.
This chemical modification effectively converted CMD-PEG
from a double hydrophilic ionic copolymer to an amphiphilic
polyelectrolyte. Thus, while CMD-PEG (0.2 g/L) is soluble in
water with no sign of aggregation within a wide pH range
(pH > 2.0) the modified polymers Dod-CMD-PEG self–assemble
in acidic aqueous solution (pH 3.0) into nanoparticles
(RH ꢀ 100 nm). The association of Dod-CMD-PEG in the
absence of aminoglycosides can be tuned by the solution pH and
salinity, as described in a forthcoming publication.
Paromomycin possesses five primary amine groups that can be
converted into a guanidinyl function in order to increase the
basicity of the antibiotic and consequently enhance the strength of
its electrostatic interaction with CMD-PEG.32 We prepared 6000-
G-Par, and 500-G-Par. The minimum inhibitory concentrations
for 6000-G-Par and 500-G-Par were comparable to the parent anti-
biotic paromomycin (MIC ¼ 2–8 mg/mL in both cases). 6000-G-Par
was obtained by treatment of compound 1 with aqueous NaOH to
deprotect selectively the 6000-NH-Cbz group (Scheme 2).37 The
resulting free amino group was guanidinylated with reagent 2,
followed by hydrogenolysis of the N-Cbz group to afford the
desired 6000-G-Par.38 In order to obtain 500-G-Par, the known
compound 4 was treated with PPh3 under Staudinger conditions
and the resulting amine was guanidinylated with reagent 2
(Scheme 2).38,39 Benzylidene deprotection with aqueous AcOH
followed by N-Cbz hydrogenolysis afforded the desired 500-G-Par.
Preparation and characterization of PIC micelles formed by the
electrostatically-driven assembly of aminoglycosides and their
derivatives with CMD-PEG and Dod-CMD-PEG
Microcalorimetric measurements by ITC were conducted first to
ascertain that the thermodynamic parameters of the interactions
between CMD-PEG and neomycin sulfate or paromomycin
sulfate were favorable to the formation of PIC micelles. Aliquots
of a concentrated solution of a drug in a phosphate buffer
Fig. 2 A: FTIR spectra of CMD-PEG sodium salt (1), Dod38-CMD-
PEG free acid (2), and Dod38-CMD-PEG sodium salt (3) (powder
sample) in the region of 1200–1900 cmꢁ1. B: 1H NMR spectra of CMD-
PEG (top spectrum) and Dod38-CMD-PEG (bottom spectrum)
(DMSO-d6, room temperature).
ꢂ
(10 mM, pH 7.0, 50 mM NaCl) at 25 C were added in a time-
controlled manner to a solution of CMD-PEG in the same buffer
also maintained at 25 ꢂC. Interaction between the two entities
was accompanied by release of heat, which diminished as the
4506 | Soft Matter, 2010, 6, 4504–4514
This journal is ª The Royal Society of Chemistry 2010

View Article Online
Scheme 2 Synthesis of 6000-guanidino-paromomycin (3) and 500-deoxy-500-guanidino-paromomycin (5).
population of free polyelectrolyte in solution decreased as
a result of PIC micelle formation. Eventually, the signal leveled
off, indicating completion of the electrostatic interactions that
led to micellization. Further addition of drug triggered constant
heat absorption due to the dilution of the drug in buffer.
Enthalpograms recorded upon titration of each drug into
a CMD-PEG solution are presented in Fig. 3, in which the top
panels present the heat released/absorbed upon injection of the
drug. The bottom panels show the integrated heat data that were
used to extract the binding parameters after subtraction of the
heat of dilution of the drugs in the same buffer. The data are
plotted as a function of the molar ratio of amine groups provided
by the drugs to carboxylates provided by the copolymer.
fitting are presented in Table 2. The number of binding sites (N)
was ꢀ 1, confirming that each CMD-PEG carboxylate group is
bound to one amino group of either neomycin or paromomycin.
The enthalpy change upon binding of one mol amine
was ꢀ ꢁ1.0 kcal in each case, giving a total enthalpy change
of ꢀ ꢁ6.0 kcal/mol for neomycin and ꢀ ꢁ5.0 kcal/mol for
paromomycin, since neomycin and paromomycin have 6 and 5
amines, respectively. Indeed, the raw ITC data (Fig. 3, top
panels) confirm that the binding of neomycin to CMD-PEG is
more exothermic than that of paromomycin. This may indicate
that the binding of neomycin to CMD-PEG is stronger than that
of paromomycin.40 Stronger neomycin binding is also confirmed
by the higher binding free energy (DG) of ꢁ28.0 kcal/mol for
neomycin, compared to ꢁ22.7 kcal/mol for paromomycin. In
both cases, the binding is mainly entropically driven since the
entropic contribution (TDS) accounted for ꢀ 75% of the binding
free energy. This observation is in good agreement with a report
on the binding of the same drugs to the A site of 16S rRNA for
which it was reported that 72% of the driving force for the
binding of paromomycin to RNA was derived from entropic
contributions.41 Electrostatic interactions between aminoglyco-
sides and either RNA or CMD-PEG copolymers are associated
with counter ions release, which results in entropy gain of the
system.42,43
The ITC profiles were fit with a model for one set of binding
sites. The thermodynamic parameters resulting from the data
The formation of core-shell nanoparticles upon addition
of either neomycin sulfate or paromomycin sulfate into
Table 2 Thermodynamic parameters for the binding of neomycin
sulfate and paromomycin sulfate to CMD-PEG (25 ꢂC, 10 mM phos-
phate buffer, pH 7.0, 50 mM NaCl)
Parameter
Neomycin sulfate
Paromomycin sulfate
N
0.85 ꢃ 0.06
4.56 ꢃ 2.20
ꢁ1.04 ꢃ 0.10
3.63
0.81 ꢃ 0.01
3.71 ꢃ 0.41
ꢁ1.08 ꢃ 0.03
3.46
K (x103)(Mꢁ1
)
Fig. 3 Corrected integrated injection heats plotted as a function of the
[amine]/[carboxylate] ratio for the titration of either neomycin sulfate or
paromomycin sulfate into CMD-PEG copolymer in phosphate buffer
(10 mM, pH 7.0) at 25 ꢂC.
DH (kcal molꢁ1
TDS (kcal molꢁ1
DG (kcal molꢁ1
)
)
)
ꢁ4.67
ꢁ4.54
This journal is ª The Royal Society of Chemistry 2010
Soft Matter, 2010, 6, 4504–4514 | 4507

View Article Online
achieved for a mixture having [amine]/[carboxylate] ¼ 2.5 as seen
also by ITC measurements (Fig. 3).
1H NMR spectroscopy was used also to confirm that hydro-
phobic modification did not affect the ability of the drugs to
form core/shell micelles via interactions with the carboxylate
groups of the Dod-CMD-PEG copolymers. The 1H NMR
spectra of Dod38-CMD-PEG in water in the absence and pres-
ence of neomycin ([amine]/[carboxylate] ¼ 2.5) are presented in
Fig. 4E and 4F. The spectrum of the mixed solution presents only
the singlet at 3.61 ppm associated with the PEG protons. Signals
due to the drug protons and to the protons of the Dod38-CMD
block cannot be detected, as in the case of a mixed CMD-PEG/
neomycin solution of identical composition (Fig. 4C).
Dynamic light scattering (DLS) experiments were conducted
in order to determine the hydrodynamic size of the PIC micelles
formed upon interaction of each drug with the copolymers listed
in Table 1, as well as the micelles formed by the guanidinylated
aminoglycosides and CMD-PEG. For each drug/copolymer pair,
mixed solutions of different [amine]/[carboxylate] ratios were
prepared in phosphate buffer (10 mM, pH 7.0) and the hydro-
dynamic radius (RH) of the resulting micelles was determined
(Fig. 5). In all cases, the distributions of nanoparticles were
monomodal. Paromomycin/CMD-PEG micelles of [amine]/
[carboxylate] ¼ 1.0 had a RH of ꢀ 100 nm and a polydispersity
index (PDI) ꢀ 0.3. Increasing the [amine]/[carboxylate] ratio to
1.5 resulted in a drop of RH to ꢀ 70 nm and of PDI to 0.2
(Fig. 5A). The RH of the micelles gradually increased upon
further increase of the [amine]/[carboxylate] ratio, probably as
a consequence of the incorporation of additional drug in the
micellar core. The RH value levelled off for [amine]/[carboxylate] ¼
2.5 and remained constant upon further addition of drug, con-
firming observations by NMR spectroscopy and ITC that the
maximum drug loading is achieved at this point. Also, drug/
polymer mixtures prepared for [amine]/[carboxylate] < 1.0 did
not scatter light indicating the absence of nanoparticles in mixed
solutions within this composition range. Similar trends were
observed in the DLS analysis of mixed neomycin/CMD-PEG
solutions, although neomycin micelles were smaller in size,
compared to paromomycin/CMD-PEG micelles (Table 3)
presumably as a result of tighter electrostatic interactions in the
core of neomycin micelles due to the presence of an additional
amino group in neomycin (Fig. 1). This amino group has a pKa of
9.24, hence it is fully ionized at pH 7.0.41 ITC experiments
Fig. 4 1H NMR spectra of neomycin sulfate (A), CMD-PEG (B),
neomycin/CMD-PEG micelles (pH 7.4, 0 mM NaCl) (C), neomycin/
CMD-PEG micelles (pH 7.4, 150 mM NaCl) (D), Dod38-CMD-PEG (E),
neomycin/Dod38-CMD-PEG micelles (pH 7.4, 0 mM NaCl) (F) and
neomycin/Dod38-CMD-PEG micelles (pH 7.4, 150 mM NaCl) (G).
(D2O: polymer concentration: 2.0 g/L, neomycin concentration: 2.1 g/L
and [amine]/[carboxylate] ¼ 2.5).
a CMD-PEG solution was confirmed by 1H NMR spectroscopy
measurements which take advantage of the fact that signals of
protons entrapped within the micellar core are either very broad
or do not appear at all, as a consequence of the restricted
mobility of such protons in this environment. In contrast,
protons of the polymer segments forming the micelle corona
maintain their mobility and their signals appear well resolved.
1
The H NMR spectrum of neomycin sulfate in D2O (Fig. 4A)
presents three singlets at 5.2, 5.34 and 5.87 ppm attributed to the
anomeric protons on the three amino sugars. The axial and
equatorial methylene protons of the neomycin cyclohexyl ring
resonate at 1.66 and 2.2 ppm, respectively. The other neomycin
protons resonances appear as a series of signals between 3.16 to
1
4.45 ppm.44 The H NMR spectrum of CMD-PEG (Fig. 4B) in
D2O presents characteristic signals at d 4.08–4.15, 4.89, and
5.07 ppm, ascribed to protons of the CMD block, and a broad
strong singlet centered at d 3.61 ppm due to the PEG methylene
protons (-CH2-CH2-O-).15 The spectrum of neomycin/CMD-
PEG micelles ([amine]/[carboxylate]
¼
2.5, pH 7.4,
[NaCl] ¼ 0 mM) (Fig. 4C) features only a strong signal at d 3.61
ppm, ascribed to the PEG methylene protons. The signals due to
the protons of neomycin and of the CMD segment of the poly-
mer cannot be seen any more, confirming the formation of PIC
micelles with a neomycin/CMD core and a PEG corona. Micelles
of this composition contain 50 wt% drug, which can be taken as
the actual drug loading, since the NMR spectrum for this
composition shows no signal due to free drug (Fig. 4C).
Spectra of mixed solutions in the 1.5 # [amine]/[carboxylate] #
2.5 composition range were similar to that presented in Fig. 4C.
However, mixed solutions having [amine]/[carboxylate] < 1.5 or
> 2.5 showed signals characteristic of free drug and of the CMD
block of the copolymer, in addition to the signal at 3.61 ppm due
to the PEG protons resonance. This indicates, on the one hand,
that micelles do not form until the [amine]/[carboxylate] ratio
reaches 1.5 or, if they form, their core is sufficiently loose and
hydrated to permit free motion of the drug and CMD protons
and, on the other hand, it confirms that maximum drug loading is
Fig. 5 Effect of the [amine]/[carboxylate] molar ratio on the hydrody-
namic radius of paromomycin sulfate (panel A) and neomycin sulfate
(panel B) micelles with: CMD-PEG (:), Dod18-CMD-PEG (C), Dod38
-
CMD-PEG (-); phosphate buffer (10 mM, pH 7.0), polymer
concentration ¼ 0.2 g/L.
4508 | Soft Matter, 2010, 6, 4504–4514
This journal is ª The Royal Society of Chemistry 2010

View Article Online
micelles.15,45 This feature is very important for in vivo applications
where micelles are subjected to drastic dilution upon intravenous
injection.
Table
3 Characteristics of aminoglycoside/CMD-PEG micelles
([amine]/[carboxylate] ¼ 2.5) in a phosphate buffer (10 mM, pH 7.0),
[polymer] ¼ 0.2 g/L
PDIa
% Drugb
a
We carried out next a light scattering study of micelles formed
by hydrophobically modified copolymers and aminoglycosides.
In Fig. 5, we present the [amine]/[carboxylate] dependence of the
RH of PIC micelles formed between each aminoglycoside and the
hydrophobically modified copolymers, Dod18-CMD-PEG and
Dod38-CMD-PEG at a polymer concentration of 0.2 g/L. For
identical [amine]/[carboxylate] ratios paromomycin/Dod-CMD-
PEG micelles are significantly smaller than micelles formed by
CMD-PEG, but no significant difference in size was detected
between paromomycin micelles formed with Dod18-CMD-PEG
and Dod38-CMD-PEG copolymers. Neomycin micelle formation
followed similar trends, except that neomycin/Dod38-CMD-PEG
micelles were significantly smaller than neomycin/Dod18-CMD-
PEG (Table 3). The core of aminoglycosides/Dod-CMD-PEG
micelles may be less hydrated than the core of micelles formed by
CMD-PEG due to the presence of hydrophobic dodecyl chains
which can act as internal non-covalent crosslinking point.46 Gao
et al. reported similar trends in a comparative study of lysozyme/
poly(1-tetradecene-alt-maleic acid) and lysozyme/poly-
(isobutylene-alt-maleic acid) micelles. The smaller size of the
former micelles was attributed to hydrophobic interactions
between the tetradecyl chains.47
Both 6000-G-Par and 500-G-Par form PIC micelles with
CMD-PEG, as expected since the guanidinyl group was selected
on the basis of its stronger basicity, compared to the amine
groups. Micelles formed in mixed solutions of CMD-PEG and
both 6000-G-Par and 500-G-Par for all [amine]/[carboxylate] > 1.
They were significantly smaller than the micelles formed by this
copolymer and unmodified paromomycin, independently of the
[amine]/[carboxylate] ratio (Figure SI1† and Table 3 for data on
micelles with [amine]/[carboxylate] ¼ 2.5). Particularly striking is
the decrease in size triggered by a guanidinyl substituent when it
is linked to the 500 position (Table 3). It should be recalled here
that a primary –OH group in 500-G-Par was converted into
a guanidine group whereas in 6000-G-Par a primary –NH2 was
converted into a guanidine group. This makes 500-G-Par more
basic than 6000-G-Par allowing stronger electrostatic interactions
with CMD-PEG carboxylate groups.
Micelle
RH
Neomycin/CMD-PEG
Paromomycin/CMD-PEG
Neomycin/Dod18-CMD-PEG
74.9 ꢃ 1.8 0.03 ꢃ 0.03 50
130.1 ꢃ 0.5 0.08 ꢃ 0.03 49.8
63.3 ꢃ 0.6 0.08 ꢃ 0.05 50
Paromomycin/Dod18-CMD-PEG 48.5 ꢃ 0.4 0.02 ꢃ 0.03 49.8
Neomycin/Dod38-CMD-PEG
Paromomycin/Dod38-CMD-PEG 54.5 ꢃ 1.2 0.03 ꢃ 0.02 49.8
40.5 ꢃ 0.4 0.06 ꢃ 0.03 50
110 ꢃ 2.2 0.08 ꢃ 0.02 50
6
000-G-Par/CMD-PEG
500-G-Par/CMD-PEG
83.8 ꢃ 2.6 0.04 ꢃ 0.05 50
a
Mean of six measurements ꢃ S.D. b % drug loading ¼ weight of drug/
(weight of micelles)ꢄ100.RH: Hydrodynamic radius, PDI: Polydispersity
index.
also showed stronger interactions between neomycin and
CMD-PEG, compared to paromomycin/CMD-PEG.
Solutions of micelles of CMD-PEG and either paromomycin or
neomycin in a phosphate buffer (10 mM, pH 7.0, [amine]/
[carboxylate] ¼ 2.5, CMD-PEG 0.5 g/L) were diluted serially in
the same buffer in order to determine by light scattering the
polymer concentration below which micelles do not form, or at
least cannot be detected by light scattering. Plots of the changes of
the relative scattering intensity, defined as the ratio of the intensity
of a solution of given concentration to the scattering intensity of
the solution of CMD-PEG ¼ 0.5 g/L, are presented in Fig. 6
together with the hydrodynamic radius of the micelles (Fig. 6
inset). The size of the micelles was not affected by dilution and
micelles were detected in solutions of [CMD-PEG] $ 0.062 g/L
and 0.125 g/L for neomycin and paromomycin, respectively. The
minimal CMD-PEG concentrations necessary for micelle
formation were determined graphically (see arrows in Fig. 6) as
the polymer concentration corresponding to the onset of the
increase in scattered light intensity. Both types of micelles exhibit
the remarkable resistance against dilution characteristic of PIC
Effects of electrolyte concentration on the size and integrity of
micelles of aminoglycosides and their derivatives with CMD-
PEG and Dod-CMD-PEG
Evidence of the remarkably enhanced stability against increased
salinity of aminoglycosides/Dod-CMD-PEG micelles, compared
to micelles formed by CMD-PEG, was obtained from 1 H NMR
experiments. Micelles of neomycin were prepared in D2O with
either CMD-PEG or Dod38-CMD-PEG (polymer concentration:
2.0 g/L, [amine]/[carboxylate] ¼ 2.5) in the absence of salt. The
spectra of the micellar solutions were recorded. For both solu-
tions, only the signal at d 3.61 ppm attributed to the PEG protons
was detected (Fig. 4). Then, we performed the same experiments
with micellar solutions containing NaCl of physiological
concentration (150 mM). The spectra recorded are presented in
Fig. 4. The spectrum of the neomycin/Dod38-CMD-PEG
micellar solution in the presence of 150 mM NaCl (Fig. 4G) was
Fig. 6 Dependence of the relative intensity of scattered light on CMD-
PEG concentation for neomycin/CMD-PEG micelles (-) and paromo-
mycin/CMD-PEG micelles (C). Inset: CMD-PEG concentration
dependence of the hydrodynamic radius of neomycin/CMD-PEG
micelles (-) and paromomycin/CMD-PEG micelles (C). Relative
scattering intensity ¼ intensity of scattered light for a given CMD-PEG
concentration/intensity of scattered light of a micellar solution of
[CMD-PEG] ¼ 0.5 g/L.
This journal is ª The Royal Society of Chemistry 2010
Soft Matter, 2010, 6, 4504–4514 | 4509

View Article Online
identical to the spectrum of the micellar solution in D2O in the
absence of salt (Fig. 4F), indicating that the micelles form and
persist even in the presence of salt. In contrast, the spectrum of
a neomycin/CMD-PEG solution (150 mM NaCl) presents
signals characteristic of neomycin, indicated by an arrow in
Fig. 4D as well as small signals due to protons of the CMD block.
This observation implies that neomycin and CMD fragments
exist as soluble species rather than entrapped within the micellar
core, and, consequently that the formation of micelles in this case
was severely impeded by the presence of salt.
participation of electrostatic and hydrophobic interactions in the
formation of tighter micelle cores.
We assessed also the resistance against salt of micelles formed
by the modified drugs, 500-G-Par and 6000-G-Par (Figure SI2).
Solutions of 6000-G-Par/CMD-PEG retained ꢀ 40% of their
initial scattering intensity at 50 mM NaCl, compared to ꢀ 20%
for Par/CMD-PEG micelles at the same salt concentration, while
500-G-Par/CMD-PEG solutions maintained the same scattered
light intensity for [NaCl] # 50 mM and ꢀ 30% of their initial
scattered light intensity at [NaCl] ¼ 100 mM (Figure SI2). This
enhanced stability of guanidinylated paromomycin micelles
might result from stronger interactions between guanidine
groups of the drug and carboxylate groups of CMD-PEG.
Nonetheless, the light scattering data imply that guanidinylation
of one position of the paromomycin nucleus is not effective in
stabilizing PIC micelles under the salinity conditions imposed by
the physiological milieu.
We determined also by light scattering experiments the effect
of added salt, from 0 to 200 mM, on the intensity of the light
scattered by micellar solutions and on the size of micelles (Fig. 7).
The intensity of the light scattered by the micellar solutions
(200 mM NaCl) was ꢀ 40% and ꢀ 80%, respectively, for
neomycin/Dod18-CMD-PEG and neomycin/Dod38-CMD-PEG
micelles (Fig. 7A). Neomycin/Dod18-CMD-PEG and neomycin/
Dod38-CMD-PEG micelles maintained their initial size at
[NaCl] ¼ 200 mM (Fig. 7B). Also, neomycin/Dod38-CMD-PEG
Drug release study and antibacterial activity of aminoglycoside/
CMD-PEG micelles
micelles
prepared
under
physiological
conditions
([NaCl] ¼ 150 mM and pH 7.4) showed no signs of disintegration
after a three-months storage at room temperature.
The release of neomycin from its PIC micelles with CMD-PEG
and Dod38-CMD-PEG was evaluated by the dialysis bag method
(Fig. 8) with micellar solutions in phosphate buffer of various pH
values and different salt concentrations since both factors can
affect drug release rate from PIC micelles.15,48 Neomycin rapidly
diffused out of the dialysis membrane in the absence of polymers
and almost complete release was achieved within 4 h (Fig. 8).
In contrast, micelle-encapsulated neomycin exhibited a lower
rate of release under all conditions studied. For instance,
the lowest release rate was detected in phosphate buffer at pH
7.0–7.4 and 0 mM NaCl. Under these conditions, ꢀ 30% of
neomycin was released after 24 h. The release rate was
significantly increased by increasing [NaCl] from 0 to 150 mM.
Thus, after 24 h, ꢀ 70% neomycin was released at pH 7.4,
[NaCl] ¼ 150 mM. The higher drug release rate in the presence of
150 mM NaCl confirms that the release involves an ion exchange
mechanism,49 as reported previously.48 No burst drug release
occurred even in conditions of high salt concentration, con-
firming that the drug is located in the micelles core, and despite
higher neomycin release rate under physiological conditions (pH
7.4, [NaCl] ¼ 150 mM), the micelles were still able to sustain drug
release for more than 24 h (Fig. 8). DLS studies of the solution
recovered inside the dialysis bags after a 24-hr dialysis against
phosphate buffer (pH 7.0, [NaCl] ¼ 150 mM) confirmed the
presence of nanoparticles (RH ꢀ 230 nm), with an estimated
neomycin content ꢀ 30% of the initial value. Changes in pH from
7.0 to 7.4 did not affect the release rate, in the absence or presence
of NaCl (150 mM). Moreover, hydrophobic modification of the
copolymer did not have any marked effect on release rates. For
instance, the release rates of neomycin from neomycin/Dod38-
CMD-PEG and neomycin/CMD-PEG micelles were identical,
within experimental uncertainty.
Hydrophobic modification of CMD-PEG had
a less
pronounced effect on paromomycin micelles stability against
increase in salinity. Yet, micelles of paromomycin/Dod38-CMD-
PEG maintained their initial size and their solution kept ꢀ 80%
of their initial scattered light intensity up to [NaCl] ¼ 100 mM
(Fig. 7C and D). Paromomycin/CMD-PEG micelles did not
form at all at this salt concentration. From these results it can be
concluded that the level of CMD-PEG hydrophobic modifica-
tion, as well as the basicity of the aminoglycoside amino groups
are major factors determining micelles stability against increase
in salinity. The enhanced stability of neomycin/Dod38-CMD-
PEG against salt-induced disintegration is probably due to the
Fig. 7 Changes in the intensity of scattered light and hydrodynamic
radius of neomycin (panels A and B) and paromomycin (panels C and D)
micelles as a function of NaCl concentration; Micelles were formed with:
Dod38-CMD-PEG (-), Dod18-CMD-PEG (C), CMD-PEG (:);
phosphate buffer (10 mM, pH 7.0), polymer concentration ¼ 0.5 g/
L;[amine]/[carboxylate] ¼ 2.5. Relative scattering intensity ¼ intensity at
a given [NaCl]/intensity for [NaCl] ¼ 0.
The antibacterial activity of free aminoglycosides and ami-
noglycosides encapsulated in CMD-PEG micelles was evaluated
by exposing the test organism E. coli XL-1 blue strain to different
drug concentrations in order to determine the lowest drug
concentration that prevents detectable bacterial growth (minimal
inhibitory concentration, MIC). In addition to neomycin,
4510 | Soft Matter, 2010, 6, 4504–4514
This journal is ª The Royal Society of Chemistry 2010

View Article Online
(162 mg, 0.87 mmol) was added to the solution. The pH was
increased to 9.0 by addition of NaOH (1.0 N). The reaction
mixture was stirred for 1.0 h at this pH. Ethanol was removed
from the mixture under reduced pressure at 50 ꢂC and the
product was recovered by freeze drying. The recovered product
was subjected to soxhlet extraction with hexane for 24 h to
remove the unreacted dodecylamine and EEDQ. The purified
Dod-CMD-PEG was converted to its free acid form by treat-
ment with a cation exchange resin (Amberliteꢀ IR 120). The
grafting density (defined as the number of dodecyl chains per
100 glycopyranose units) was determined from the 1H NMR
spectrum of Dod-CMD-PEG in DMSO-d6 using the ratio of the
areas of the signals due to the terminal methyl protons of the
dodecyl chain (0.85 ppm) and to dextran anomeric protons
(4.66 ppm) (Fig. 2). Two Dod-CMD-PEG samples of grafting
densities 18 and 38 were prepared.
Fig. 8 Release profiles of neomycin from: a neomycin solution (-);
neomycin/CMD-PEG micelles, pH 7.0, [NaCl] ¼ 0 mM (C); neomycin/
CMD-PEG micelles, pH 7.4, [NaCl] ¼ 0 mM (;); neomycin/CMD-PEG
micelles, pH 7.0, [NaCl] ¼ 150 mM (A); neomycin/CMD-PEG micelles,
pH 7.4, [NaCl] ¼ 150 mM (:); neomycin/Dod38-CMD-PEG micelles,
pH 7.4, [NaCl] ¼ 150 mM (B). ([neomycin] ¼ 2.0 g/L, [amine]/
[carboxylate] ¼ 2.5), temperature 37 ꢂC, 10 mM phosphate buffer.
Synthesis of 6000-guanidino-paromomycin peracetate (6000-G-Par)
Aqueous NaOH (0.80 g, 20 mmol, in 5 mL H2O) was added to
a solution of penta-N-Cbz-paromomycin (compound 1, 0.50 g,
0.26 mmol) (Scheme 2) in 1,4-dioxane (15 mL).37,52 The selective
deprotection of the 6000 amine was complete after 16 h,
as indicated by TLC on silica plates (mobile phase:
CHCl3:AcOEt:MeOH, 20 : 5 : 3) and confirmed by mass spec-
trometry (MS) analysis (m/z calcd for C65H77N5O22 [M + H]+:
1280.5, MS found: 1280.6). 1,4-Dioxane was evaporated under
reduced pressure. The separated white gum formed was dissolved
in MeOH (3 mL). The methanolic solution was transferred in
water (50 mL) to obtain a white precipitate that was recovered by
filtration and dried by lyophilization. The resulting product was
paromomycin and guanidinylated-paromomycin analogues
(500-G-Par and 6000-G-Par), several other aminoglycosides were
tested, including the clinically relevant tobramycin and amikacin.
Both 6000-G-Par and 500-G-Par showed MIC of 2–8 mg/mL. In no
case was the MIC of the drug altered upon encapsulation in
CMD-PEG micelles. Thus, whether drugs were free or encap-
sulated in PIC micelles, MICs were 2–8, 4–8, 2–8 and 2–8 mg/mL
for amikacin, neomycin, paromomycin and tobramycin,
respectively. Similar results were reported for ciprofloxacin
encapsulated in polyethylbutylcyanoacrylate nanoparticles and
amphotericin B encapsulated in poly(lactic-co-glycolic acid)
nanoparticles.50,51
added to
a
solution of N,N0-diCbz-N00-triflylguanidine
(compound 2, 0.21 g, 0.47 mmol) and Et3N (0.11 mL, 0.78 mmol)
in chloroform (20 mL).38 The reaction mixture was kept refluxing
for 18 h. After evaporation of the solvent under reduced pres-
sure, the residue was dissolved in the minimum amount of
CH2Cl2 and loaded onto a silica gel column eluted with 0 to 5%
MeOH in CH2Cl2 to yield N-Cbz protected guanidinylated
paromomycin (0.41 g, 72%). m/z calcd for C72H84N7O26 [M +
H]+: 1462.5, MS found: 1462.7. Water was added to a solution of
N-Cbz protected guanidinylated paromomycin (0.41 g,
0.28 mmol) in MeOH (5 mL) until the solution became cloudy.
20% Pd(OH)2/C (80 mg) and few drops of AcOH were added to
the cloudy solution and the suspension was stirred under
hydrogen atmosphere (hydrogen balloon) until the conversion of
the starting material into the product was completed as indicated
by MS analysis (6 h). The mixture was filtered through a layer of
Celite on cotton, concentrated under vacuum, washed
with CH2Cl2 twice, dissolved in water and lyophilized to afford
6000-G-Par (240 mg, 90%) as a peracetate salt. m/z calcd for
3. Experimental
Materials
Neomycin sulfate, paromomycin sulfate and all other chemicals
were purchased from Sigma-Aldrich Chemicals (St. Louis, MO)
unless otherwise specified. The block copolymer CMD-PEG
(Fig. 1) was synthesized starting with dextran (Mn 6 000 g/mol,
Fluka Chemicals) and a-amino-u-methoxy-poly(ethyleneglycol)
(Mn 5 000 g/mol) as described previously.35 The degree of car-
boxymethylation of the dextran block, defined as the number of
glucopyranose units having carboxymethyl groups per 100 glu-
copyranose units, was 85%. The average number of glucopyr-
anosyl and of -CH2-CH2-O- repeat units of the CMD and PEG
segments, were 40 and 140, respectively. Water was deionized
using a Millipore MilliQ system.
Synthesis of n-dodecyl-carboxymethyldextran-b-
poly(ethyleneglycol) (Dod-CMD-PEG)
C24H48N7O14 [M + H]+: 658.3, MS found: 658.4.); H NMR
1
(400 MHz, D2O), d (ppm): 5.69 (m, 1H), 5.33 (m, 1H), 5.11
(m, 1H), 4.28 (m, 1H), 414.18 (m, 1H), 4.02 ꢁ 3.98 (m, 3H),
3.88 ꢁ 3.72 (m, 4H), 3.68 ꢁ 3.01 (m, 11H), 2.29 (m, 1H), 1.87
(s, 15H), 1.61 (m, 1H), 1.12 (m, 1H); 13C NMR (101 MHz, D2O),
d (ppm): 180.6, 156.9, 109.4, 95.6, 95.3, 83.9, 81.1, 77.4, 75.5,
73.3, 73.0, 72.3, 68.8, 68.5, 67.3, 65.9, 62.1, 60.1, 53.5, 50.6, 49.3,
48.4, 46.2, 41.4, 28.1, 22.7.
A solution of N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline
(EEDQ) (216 mg, 0.87 mmol) in absolute ethanol (120 mL) was
added gradually to
a solution of CMD-PEG (288 mg,
0.97 mmol carboxylate) in water (120 mL, pH 4.0). The
apparent pH of the water/ethanol mixture was adjusted to 4.0
and kept at this value for 30 min. Subsequently, n-dodecylamine
This journal is ª The Royal Society of Chemistry 2010
Soft Matter, 2010, 6, 4504–4514 | 4511

View Article Online
Synthesis of 500-deoxy-500-guanidino-paromomycin peracetate
(500-G-Par)
concentration: 0.5 g/L; [amine]/[carboxylate] ¼ 2.5) in 10 mM
phosphate buffer (10 mM, pH 7.0). Aliquots of a NaCl stock
solution (2.5 M) in the same buffer were added to the micellar
solutions in volumes such that [NaCl] in the sample ranged from
10 to 400 mM. The mixtures were stirred for 5 min and the
volume of each sample was adjusted to 2.5 mL with the same
buffer. The pH of solutions of [NaCl] ¼ 150 mM was increased to
7.4 by the addition of aqueous NaOH (1.0 N). The samples were
stirred overnight prior to analysis.
Water (0.1 mL) and PPh3 (0.27 g, 1.0 mmol) were added to
a solution of (penta-N-Cbz-40,60-O-benzylidene-500-deoxy-500-
azido-paromomycin) (compound 4, 1.2 g, 0.86 mmol) in THF
(30 mL) (Scheme 2).39 The mixture was kept at room temperature
for 18 h. The solvent was evaporated under reduced pressure.
The residue, dissolved in the minimum amount of CH2Cl2, was
loaded on a silica gel column eluted with 4 to 8% MeOH in
CH2Cl2 yielding the corresponding 500-amino compound (0.20 g,
17%, m/z calcd for C70H81N6O23 [M + H]+: 1373.5, MS found:
1373.8. Subsequently, Et3N (0.041 mL, 0.30 mmol) and reagent 2
(0.80 mg, 0.18 mmol) were added to a solution of the 500-amino
compound (0.20 g, 0.15 mmol) in CHCl3 (20 mL). The reaction
mixture was refluxed for 18 h. The residue recovered after
evaporation of the solvent under reduced pressure was dissolved
in the minimum amount of CH2Cl2and loaded onto a silica gel
column eluted with 2 to 7% MeOH in CH2Cl2 yielding the
desired N-Cbz protected guanidinylated paromomycin (0.20 g,
83%, m/z calcd for C87H95N8O27 [M + H]+: 1683.6, MS found:
1684.0.) The N-Cbz protected guanidinylated paromomycin
(0.20 g, 0.12 mmol) was dissolved in 80% aqueous acetic acid
Aminoglycoside release studies
A micellar solution (3.0 mL, neomycin: 2.0 g/L, [amine]/
[carboxylate]
¼
2.5) was introduced into a dialysis bag
(MWCO ¼ 6.0–8.0 kDa), which was immersed into a phosphate
ꢂ
buffer (10 mM, 25 mL) kept at 37 C and containing either no
NaCl (pH 7.0 or 7.4) or 150 mM NaCl (pH 7.0 or 7.4). At pre-
determined time intervals, 5 mL aliquots were taken from the
release medium and replaced by an equal volume of fresh buffer.
A control experiment to determine the drug diffusion through
the dialysis membrane was carried out starting with a neomycin
solution in the absence of the polymer. The neomycin concen-
tration in each aliquot was determined spectrophotochemically
after treatment with o-phthaldialdehyde.53 Each sample (1 mL)
was treated with a o-phthaldialdehyde solution in isopropanol
(1 mL, 1 mg/mL), diluted with isopropanol (1.5 mL), and the
sample volume was adjusted to 5.0 mL by addition of aqueous
borate buffer (50 mM, pH 9.0). The samples were allowed to
stand for 1 h at room temperature. Neomycin concentration was
determined from the solution absorbance at 335 nm and using
a calibration curve. The cumulative percent neomycin released
was plotted as a function of the dialysis time.
ꢂ
(5 mL) and the solution was heated at 60 C until total conver-
sion of the starting material into the benzylidene deprotected
product (5 h) as ascertained by MS analysis. The solvent was
evaporated under reduced pressure. The residue was dissolved in
MeOH (3 mL) and H2O was added dropwise until the solution
became cloudy. The resulting mixture was treated with 20%
Pd(OH)2/C (40 mg) and a few drops of acetic acid. The suspen-
sion was stirred under hydrogen atmosphere (hydrogen balloon)
to achieve total conversion of the starting material into 500-G-Par
(6 h), as indicated by MS analysis (m/z calcd for C80H91N8O27 [M
+ H]+: 1595.6, MS found: 1595.9). The mixture was filtered
through celite, concentrated under vacuum, washed with CH2Cl2
twice, dissolved in water and lyophilized to afford 500-G-Par (105
mg, 87%) as a peracetate salt. m/z calcd for C24H49N8O13 [M +
H]+: 657.3, MS found: 657.4. 1H NMR (400 MHz, D2O),
d (ppm): 5.71 (m, 1H), 5.22 (m, 1H), 5.13 (m, 1H), 4.35 (m, 1H),
414.27 (m, 1H), 4.14 (m, 2H), 4.06 (m, 2H), 3.87 ꢁ 3.11 (m, 15H),
2.28 (m, 1H), 1.76 (s, 18H), 1.63 (m, 1H), 1.11 (m, 1H); 13C NMR
(101 MHz, D2O), d (ppm): 180.7, 156.8, 109.8, 94.9, 94.0, 83.8,
78.4, 76.7, 76.2, 73.6, 72.5, 71.8, 69.8, 69.3, 68.7, 67.2, 66.9, 59.8,
53.3, 50.4, 49.2, 48.5, 43.5, 40.0, 28.0, 22.8.
Minimal inhibitory concentration (MIC) determination
Free aminoglycosides or aminoglycosides micelles (aminoglyco-
side concentration: 0.3 g/L, [amine]/[carboxylate] ¼ 2.5) in
deionized water were diluted using sterile Luria-Bertani media
(LB) in a 96 wells plate to reach drug concentrations of 0.25, 0.50,
0.75, 1.00, 2.00, 4.00, 8.00, 16.0 and 32.0 mg/mL. E. coli XL-1 blue
strain was grown at 37 ꢂC in 2 mL sterile LB to mid log phase (until
absorbance at 595 nm reaches 0.6). The resulting suspension was
shaken for homogeneity. Aliquots (1 mL) were placed in the wells
containing the micellar solutions. Blank samples were prepared
without E. coli XL-1 blue strain. After shaking the plate at 37 ꢂC
for 5 h, the absorbance at 595 nm of each well was monitored. The
lowest concentration at which the absorbance at 595 nm was the
same as the blank samples was taken as the MIC. MIC determi-
nation was done in triplicates in all cases.
Preparation of aminoglycoside/copolymer micelles
Stock solutions of CMD-PEG or Dod-CMD-PEG (1.0 g/L) and
aminoglycosides (4.96 g/L, 6.94 mmol/L and 5.27 g/L,
5.79 mmol/L for paromomycin sulfate and neomycin sulfate,
respectively) were prepared in a phosphate buffer (10 mM,
pH 7.0). Specified volumes of the aminoglycoside solution were
added dropwise to a magnetically stirred polymer solution over
a 10-min period to obtain solutions ranging in [amine]/
[carboxylate] ratios from 1.0 to 5.0. The volume of each sample
was adjusted to 2.5 mL using the same buffer. Samples had
a copolymer concentration of 0.5 g/L, unless otherwise specified.
The effect ionic strength on the stability of aminoglycoside/
copolymer micelles was studied using solutions (copolymer
Measurements
1H NMR spectra were recorded using a Bruker AV-400 MHz
spectrometer operating at 400 MHz. Chemical shifts are given
relative to external tetramethylsilane (TMS ¼ 0 ppm). FTIR
spectra were recorded on a Perkin Elmer Spectrum One spec-
trometer with a resolution of 8 cmꢁ1. Lyophilization was per-
formed with a Virtis (Gardiner, NY) Sentry Benchtop (3L)
freeze-dryer. UV–vis absorption spectra were recorded with an
Agilent 8452A photodiode array spectrometer. Dynamic light
4512 | Soft Matter, 2010, 6, 4504–4514
This journal is ª The Royal Society of Chemistry 2010

View Article Online
scattering experiments (DLS) were performed using a CGS-3
goniometer (ALV GmbH) equipped with an ALV/LSE-5003
multiple-s digital correlator (ALV GmbH), a He-Ne laser
(l ¼ 632.8 nm), and a C25P circulating water bath (Thermo
Haake). A cumulant analysis was applied to obtain the diffusion
coefficient (D) of the micelles in solution. The hydrodynamic
radius (RH) of the micelles was obtained using the Stokes–Ein-
stein eqn (1),
Joelle Pelletier are acknowledged for their assistance in
measuring MIC of aminoglycosides.
References
1 M. A. C. Stuart, W. T. S. Huck, J. Genzer, M. Muller, C. Ober,
M. Stamm, G. B. Sukhorukov, I. Szleifer, V. V. Tsukruk,
M. Urban, F. M. Winnik, S. Zauscher, I. Luzinov and S. Minko,
Nat Mater, 2010, 9, 101–113.
2 A. Mata, L. Hsu, R. Capito, C. Aparicio, K. Henrikson and
S. I. Stupp, Soft Matter, 2009, 5, 1228–1236.
kBT
D ¼
(1)
6phsRH
3 K. Channon and C. E. MacPhee, Soft Matter, 2008, 4, 647–652.
4 A. J. Dirks, S. S. van Berkel, H. I. V. Amatdjais-Groenen, F. Rutjes,
J. Cornelissen and R. J. M. Nolte, Soft Matter, 2009, 5, 1692–1704.
5 Y. Lee and K. Kataoka, Soft Matter, 2009, 5, 3810–3817.
6 K. K. Upadhyay, H. G. Agrawal, C. Upadhyay, C. Schatz, J. F. Le
Meins, A. Misra and S. Lecommandoux, Crit. Rev. Ther. Drug
Carr. Syst., 2009, 26, 157–205.
7 S. Kim, J. H. Kim, O. Jeon, I. C. Kwon and K. Park, Eur. J. Pharm.
Biopharm., 2009, 71, 420–430.
8 A. Harada and K. Kataoka, Macromolecules, 1995, 28, 5294–5299.
9 A. V. Kabanov, T. K. Bronich, V. A. Kabanov, K. Yu and
A. Eisenberg, Macromolecules, 1996, 29, 6797–6802.
10 M. A. C. Stuart, N. A. M. Besseling and R. G. Fokkink, Langmuir,
1998, 14, 6846–6849.
11 K. Osada, R. J. Christie and K. Kataoka, J. R. Soc. Interface, 2009, 6,
S325–S339.
12 C. H. Wang, W. T. Wang and G. H. Hsiue, Biomaterials, 2009, 30,
3352–3358.
13 H. Pinto-Alphandary, A. Andremont and P. Couvreur, Int.
J. Antimicrob. Agents, 2000, 13, 155–168.
where hs is the viscosity of the solvent, kB is the Boltzmann
constant, and T is the absolute temperature. The constrained
regularized CONTIN method was used to obtain the particle size
distribution. The data presented are the mean of six measurements
ꢃ S.D. Solutions for analysis were filtered through a 0.45 mm
Millex Millipore PVDF membrane prior to measurement.
Isothermal titration calorimetry (ITC) measurements were carried
out with a Microcal VP-ITC instrument. Solutions of neomycin
sulfate, paromomycin sulfate and CMD-PEG were prepared in
a phosphate buffer (10 mM, pH 7.0). Prior to measurements all
solutions were degassed under vacuum for about 10 min to elim-
inate any air bubbles. Aliquots of the neomycin sulfate solution
(10 mL, 6.0 g/L, 6.6 mM, 39.6 mM amine) or paromomycin sulfate
solution (10 mL, 5.65 g/L, 7.92 mM, 39.6 mM amine) were injected
from a 300 mL continuously stirred (300-rpm) syringe into a solu-
tion of CMD-PEG (1.43 mL, 0.75 g/L, 2.61 mM carboxylate) at
14 D. E. Owens and N. A. Peppas, Int. J. Pharm., 2006, 307, 93–102.
15 G. M. Soliman and F. M. Winnik, Int. J. Pharm., 2008, 356, 248–258.
16 G. M. Soliman, A. O. Choi, D. Maysinger and F. M. Winnik,
Macromol. Biosci., 2010, 10, 278–288.
ꢂ
25 C. Heats of dilution and mixing were determined in control
experiments by injecting aliquots (10 mL) of each drug solution
into the same buffer (1.43 mL). A total of 28 aliquots were injected
into the sample cell in intervals of 300 s. The calorimetric data were
analyzed and converted to enthalpy change using Microcal
ORIGIN 7.0.
17 E. Durante-Mangoni, A. Grammatikos, R. Utili and M. E. Falagas,
Int. J. Antimicrob. Agents, 2009, 33, 201–205.
18 M. P. Mingeot-Leclercq, Y. Glupczynski and P. M. Tulkens,
Antimicrob. Agents Chemother., 1999, 43, 727–737.
19 D. Moazed and H. F. Noller, Nature, 1987, 327, 389–394.
20 C. Roberta, B. Alessandro, P. Valerio, M. Elisabetta, Z. Gian Paolo
and G. Maria Rosa, J. Pharm. Sci., 2003, 92, 1085–1094.
€
21 J. Hombach, H. Hoyer and A. Bernkop-Schnurch, Eur. J. Pharm.
Sci., 2008, 33, 1–8.
4. Conclusion
22 M. Halwani, C. Mugabe, A. O. Azghani, R. M. Lafrenie, A. Kumar
and A. Omri, J. Antimicrob. Chemother., 2007, 60, 760–769.
23 C. Mugabe, M. Halwani, A. O. Azghani, R. M. Lafrenie and A. Omri,
Antimicrob. Agents Chemother., 2006, 50, 2016–2022.
24 A. M. Abraham and A. Walubo, Int. J. Antimicrob. Agents, 2005, 25,
392–397.
25 M. C. Lecaroz, M. J. Blanco-Prieto, M. A. Campanero, H. Salman
and C. Gamazo, Antimicrob. Agents Chemother., 2007, 51, 1185–1190.
26 R. Cavalli, M. R. Gasco, P. Chetoni, S. Burgalassi and
M. F. Saettone, Int. J. Pharm., 2002, 238, 241–245.
27 H. F. Chuang, R. Smith, C. xe and P. T. Hammond,
Biomacromolecules, 2008, 9, 1660–1668.
28 C. Mugabe, A. O. Azghani and A. Omri, Int. J. Pharm., 2006, 307,
244–250.
29 P. A. Bridges and K. M. G. Taylor, J. Pharm. Pharmacol., 2001, 53,
393–398.
30 S. Prior, B. Gander, J. M. Irache and C. Gamazo, J. Antimicrob.
Chemother., 2005, 55, 1032–1036.
31 A. Harada and K. Kataoka, J. Am. Chem. Soc., 1999, 121, 9241–9242.
32 N. W. Luedtke, T. J. Baker, M. Goodman and Y. Tor, J. Am. Chem.
Soc., 2000, 122, 12035–12036.
Electrostatic interactions between two aminoglycosides:
neomycin and paromomycin and different CMD-PEG copoly-
mers led to the formation of PIC micelles with a drug/CMD core
and a PEG corona. The interactions were entropically driven and
stronger in the case of neomycin/CMD-PEG compared to paro-
momycin/CMD-PEG. Aminoglycosides/CMD-PEG micelles
were unstable under physiological conditions (pH 7.4, [NaCl] ¼
150 mM). Micelles stability was significantly improved by
hydrophobic modification of CMD-PEG or guanidinylation of
paromomycin. Neomycin/Dod38-CMD-PEG micelles resisted
salt-induced disintegration for up to 200 mM. Smaller micelle size
and better stability against salt were observed for drugs having
more cationic groups and polymers having both carboxylate and
dodecyl groups. The proposed approaches of micelle stabilization
could be applied to other unstable PIC micelles.
33 N. Nishiyama, M. Yokoyama, T. Aoyagi, T. Okano, Y. Sakurai and
K. Kataoka, Langmuir, 1999, 15, 377–383.
34 X. Yuan, Y. Yamasaki, A. Harada and K. Kataoka, Polymer, 2005,
46, 7749–7758.
Acknowledgements
35 O. S. Hernandez, G. M. Soliman and F. M. Winnik, Polymer, 2007,
48, 921–930.
36 M. Mauzac and J. Jozefonvicz, Biomaterials, 1984, 5, 301–304.
ꢀ
37 S. Hanessian, T. Takamoto, R. Masse and G. Patil, Can. J. Chem.,
1978, 56, 1482–1491.
The work was supported in part by a grant of the Natural
Sciences and Engineering Research Council of Canada to FMW.
GMS thanks the Ministry of Higher Education, Egypt for
granting him a scholarship. Prof. Jeffrey W. Keillor and Prof.
This journal is ª The Royal Society of Chemistry 2010
Soft Matter, 2010, 6, 4504–4514 | 4513

View Article Online
38 K. Feichtinger, C. Zapf, H. L. Sings and M. Goodman, J. Org. Chem.,
1998, 63, 3804–3805.
47 G. Gao and P. Yao, J. Polym. Sci., Part A: Polym. Chem., 2008, 46,
4681–4690.
ꢀ
39 S. Hanessian, R. Masse and M. L. Capmeau, J. Antibiot., 1977, 30,
893–896.
40 Y. Tian, P. Ravi, L. Bromberg, T. A. Hatton and K. C. Tam,
Langmuir, 2007, 23, 2638–2646.
48 K. W. Yang, X. R. Li, Z. L. Yang, P. Z. Li, F. Wang and Y. Liu,
J. Biomed. Mater. Res., Part A, 2009, 88a, 140–148.
49 H. M. Aliabadi and A. Lavasanifar, Expert Opin. Drug Delivery,
2006, 3, 139–162.
41 M. Kaul, C. M. Barbieri, J. E. Kerrigan and D. S. Pilch, J. Mol. Biol.,
2003, 326, 1373–1387.
42 Z. OuandM. Muthukumar, J. Chem. Phys., 2006, 124, 154902–154911.
43 B. Hofs, I. K. Voets, A. d. Keizer and M. A. C. Stuart, Phys. Chem.
Chem. Phys., 2006, 8, 4242–4251.
44 D. G. Reid and K. Gajjar, J. Biol. Chem., 1987, 262, 7967–7972.
45 A. Harada, H. Togawa and K. Kataoka, Eur. J. Pharm. Sci., 2001, 13,
35–42.
50 M. E. Page-Clisson, H. Pinto-Alphandary, M. Ourevitch,
A. Andremont and P. Couvreur, J. Controlled Release, 1998, 56,
23–32.
51 J.-Y. Bang, C.-E. Song, C. Kim, W.-D. Park, K.-R. Cho, P.-I. Kim,
S.-R. Lee, W.-T. Chung and K.-C. Choi, Arch. Pharmacal Res.,
2008, 31, 1463–1469.
52 T. Takamoto and S. Hanessian, Tetrahedron Lett., 1974, 15, 4009–
4012.
46 E. Akiyama, N. Morimoto, P. Kujawa, Y. Ozawa, F. o. M. Winnik
and K. Akiyoshi, Biomacromolecules, 2007, 8, 2366–2373.
53 S. SuchitraSampath and D. H. Robinson, J. Pharm. Sci., 1990, 79,
428–431.
4514 | Soft Matter, 2010, 6, 4504–4514
This journal is ª The Royal Society of Chemistry 2010