Polyoxazoline thermoresponsive micelles as radionuclide delivery systemsa

Article abstract of DOI:10.1002/mabi.201000034

Thermoresponsive polymer micelles are promising drug and radionuclide carriers with a strong passive targeting effect into solid tumors. We have synthesized ABA triblock copolymers poly[2-methyl-2-oxazoline-block-(2- isopropyl-2-oxazoline-co-2-butyl-2-oxa

Full text of DOI:10.1002/mabi.201000034

Full Paper
Polyoxazoline Thermoresponsive Micelles as
Radionuclide Delivery Systemsa
Martin Hruby,* Sergey K. Filippov, Jiri Panek, Michaela Novakova,
Hana Mackova, Jan Kucka, David Vetvicka, Karel Ulbrich
Thermoresponsive polymer micelles are promising drug and radionuclide carriers with a
strong passive targeting effect into solid tumors. We have synthesized ABA triblock copoly-
mers poly[2-methyl-2-oxazoline-block-(2-isopropyl-2-oxazoline-co-2-butyl-2-oxazoline)-block-
2-methyl-2-oxazoline]. These polymers are
molecularly dissolved in aqueous millieu below
the cloud point temperature (CPT) of the ther-
moresponsive central block and above CPT form
polymer micelles at CMC 5–10 ꢀ 10ꢁ5 g ꢂ mLꢁ1
with diameter ꢃ200 nm. The phenolic moiety
introduced into the copolymer allowed radio-
nuclide labeling with iodine-125 ongoing in
good yield with sufficient in vitro stability
under model conditions.
Introduction
(alkyl ¼ methyl or ethyl)[4] through thermal sensitivity of
thermoresponsive polymers (alkyl ¼ isopropyl)[2,5] to
Poly(2-alkyl-2-oxazolines) are attracting increasing atten-
tion in biomedicinal research due to their peptide-related
structure and wide range and adjustable physicochemical
and biological properties depending on the alkyl substi-
tuent.[1–3] Their properties range from high hydrophilicity
enabling synthesis of hydrophilic water soluble biocompa-
tible polymers with excellent antibiofouling properties
hydrophobicity typical for hydrophobic aromatic or ali-
phatic polymers (substituent ¼ phenyl, butyl, nonyl
etc.).[1,6] Moreover, polyoxazolines may be synthesized by
controlled living cationic ring-opening polymerization
with polydispersities typically below 1.3 and this method
enables synthesis of block copolymers by subsequent
addition of different monomers.[7,8] The choice of initiator
and terminating agent also enables synthesis of polymers
with defined chain ends (see scheme in Scheme 1).[9]
Polymer micelles rank highly among prospective anti-
cancer drug and radionuclide delivery systems, enabling
efficient tumor targeting due to their high apparent
molecular weight (EPR effect).[10,11] Unimers may be
eliminated rapidly from organisms by glomerular filtration
after disassembly of the micellar system that fulfilled its
task as a drug carrier.[10,11] Many micelles with a thermo-
responsive core (made of a polymer block with a lower
critical solubility temperature (LCST)) and a hydrophilic
corona have the additional advantage of simple preparation
by heating of an aqueous solution of the thermoresponsive
M. Hruby, S. K. Filippov, J. Panek, M. Novakova, H. Mackova,
J. Kucka, K. Ulbrich
Institute of Macromolecular Chemistry, v.v.i., Academy of
Sciences of the Czech Republic, Heyrovsky Sq. 2, 162 06 Prague 6,
Czech Republic
Fax: þ 420 296 809 410.; E-mail: mhruby@centrum.cz
D. Vetvicka
Institute of Microbiology, v.v.i., Academy of Sciences of the Czech
Republic, Videnska 1083, 142 20 Prague 4, Czech Republic
a : Supporting information for this article is available at the bottom
of the article’s abstract page, which can be accessed from the
journal’s homepage at http://www.mbs-journal.de, or from the
author.
Macromol. Biosci. 2010, 10, 916–924
916
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/mabi.201000034

Polyoxazoline Thermoresponsive Micelles as Radionuclide . . .
R
R
R
by condensation of ethanolamine with the
particular carboxylic acid (isobutyric or valeric
for IprOX and BuOX, respectively), or by a
multistep method according to ref. [18] (EnOX).
N-(2-sulfanylethyl)-2-(4-hydroxyphenyl) acet-
amide was synthesized by condensation of 4-
hydroxyphenylacetic acid and cystamine with
(EEDQ) as a condensation reagent according to
thefollowingprocedure(seeScheme2).Cystea-
mine (2.00 g, 25.9 mmol) was dissolved by
gentle heating in tetrahydrofuran (THF,
200 mL), 2-ethoxy-1-ethoxycarbonyl-1,2-
dihydroquinoline (6.41 g, 25.9 mmol) was
dissolved in THF (50 mL) and 4-hydroxyphe-
nylacetic acid (3.44 g, 22.6 mmol) was also
dissolved in THF (20 mL). The solution of EEDQ
was mixed with the solution of 4-hydroxyphe-
nylacetic acid and then the solution of
cysteamine was added. The mixture was
stirred overnight at ambient temperature,
heated at 60 8C for 1 h and then evaporated
in vacuo. The solid residue was vigorously
shaken with water and diethyl ether (aa
δ+
R 1
δ−
X
initiation
O+
N+
O
R 1
R 1
N
O
N
X
X
R
R
R
O
O+
propagation
propagation
R 1
R 1
N
O
N
R
N
.........
X
N+
O
Nu-
termination
O
N
Nu
n
R 1
Scheme 1. Mechanism of living cationic polymerization of 2-alkyl-2-oxazolines.
polymer.[12] Since there are thermoresponsive poly(2-alkyl-
2-oxazolines) with suitable properties described in the
literature,[13–17] we decided to use this class of polymers for
the construction of our thermoresponsive micellar radio-
nuclide delivery systems.
100 mL). The suspension was filtered and the collected precipitated
product was recrystallized from methanol. The yield was 3.34 g
(70%).
1HNMR(CD3OD):d ¼ 2.75(t, ꢁCH2ꢁSꢁ), 3.33(s, ArylꢁCH2ꢁCOꢁ),
3.43 (t, NꢁCH2ꢁ), 6.72 (d, ArylꢁH on positions 2 and 6), 7.08 (d,
ArylꢁHonpositions3and5). 13CNMR(CD3OD):d ¼ 38.4(ꢁCH2ꢁSꢁ),
39.7 (NꢁCH2ꢁ), 43.1 (ArylꢁCH2ꢁ), 116.4 (aromatic carbons in
positions 3 and 5), 127.5 (aromatic carbon in position 1), 131.2
(aromatic carbons in positions 2 and 6), 157.5 (aromatic carbon in
position 4), 175.0 (ꢁCOꢁ). C10H13NO2S (211.3): Calcd. C 56.85, H 6.20,
N 6.63, S 15.18; Found C 56.90, H 6.03, N 6.50, S 15.10.
All other chemicals were purchased from Sigma-Aldrich Ltd.
(Prague, Czech Republic). 2-Methyl-2-oxazoline (MeOX) was dis-
tilledwithcalciumhydridebeforeuse, andallotherchemicalswere
used as obtained.
In this paper we describe the synthesis and study of the
properties of ABA triblock copolymer poly[2-methyl-2-
oxazoline-block-(2-isopropyl-2-oxazoline-co-2-butyl-2-oxa-
zoline)-block-2-methyl-2-oxazoline] with two hydrophilic A
blocks (poly(2-methyl-2-oxazoline)) and one central ther-
moresponsiveBblockcopolymer(poly(2-isopropyl-2-oxazo-
line-co-2-butyl-2-oxazoline)) with different monomer unit
ratios. These polymers are molecularly soluble in aqueous
milieu below the cloud point temperature (CPT) of the
thermoresponsive block and self-assemble into micelles at
higher temperature. Micelles are formed within a narrow
temperature range. The CPT of the thermoresponsive block
may be adjusted with the 2-butyl-2-oxazoline (hydrophobic
monomer lowering the CPT) to 2-isopropyl-2-oxazoline
(main monomer giving thermoresponsive properties to its
copolymers) ratio and the size of the micelles can be
controlled by the A to B block weight ratio. A phenolic
moiety was introduced into the above stated polymer to
allow radionuclide labeling with iodine radioisotopes for
both diagnostics (123I, 124I) and therapy(131I) ofsolid tumors.
The polymers were isotopically labeled and the in vitro
stability of the radiolabel was checked.
Synthesis of Polymers
Living cationic ring opening polymerization was performed in
acetonitril using tosylates as initiators. Thermoresponsive poly-
mers poly(IprOX-co-BuOX) as models of the thermoresponsive
central block were synthesized as follows. The mixture of
monomers (IprOX with 0, 10 and 20 mol-% BuOX, respectively,
total weight 1.00 g in all cases) was polymerized in anhydrous
acetonitrile (1.00 mL) using methyl p-toluenesulfonate (37 mg,
0.2 mmol) as an initiator. The polymerization was carried out at
42 8C for 8 d in a 15 mL Ace pressure tube (Sigma-Aldrich Ltd.,
Prague, Czech Republic) under a dry nitrogen atmosphere. The
polymerization mixture was mixed with ethanol (30 mL), left
overnight at room temperature and evaporated in vacuo. The raw
polymer was purified by gel permeation chromatography on
Sephadex LH-20 using methanol as the eluent and evaporation of
the polymer-containing fractions in vacuo. The polymer was
redissolved in ethanol and evaporated in vacuo to obtain solid
foam, which waseasierto manipulate. A typical yieldwasca. 0.90 g
(90%).
Experimental Part
Low Molecular Weight Precursors
2-Isopropyl-2-oxazoline (IprOX), 2-butyl-2-oxazoline (BuOX) and 2-
butenyl-2-oxazoline (EnOX) were synthesized according to ref.[5,18]
Macromol. Biosci. 2010, 10, 916–924
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.mbs-journal.de
917

M. Hruby et al.
initiator in anhydrous acetonitrile (1.00 mL).
Thepolymerizationwascarriedoutat42 8Cfor
7 d in a 15 mL Ace pressure tube (Sigma-
Aldrich Ltd., Prague, Czech republic) under
HO
N
H
S
O
a
dry nitrogen atmosphere. After this,
O
AIBN
O
O
2-methyl-2-oxazoline (to the total weight of
the sum of oxazoline monomers 2.00 g, i.e.,
1.33, 1.00 and 0.66 g, respectively) and anhy-
drous acetonitrile (1.00 mL) were added and
the polymerization continued for 7 d. The
polymerization mixture was mixed with
ethanol (30 mL) and the mixture was worked
up as described above for the polymerization
of model poly(IprOX-co-BuOX) copolymers. A
typical yield was ca. 1.85 g (93%).
N
n
N
*I
n
HO
N
H
SH
EEDQ
Inthecaseofthepolymerstobeisotopically
labeled, EnOX (25 mg, 0.20 mmol) was used
instead of the equimolar part of BuOX thus
*I
forming
poly[MeOX-b-(IprOX-co-BuOX-co-
EnOX)-b-MeOX].
HO
HO
OH
Molecular weights of the polymers were
determined by size exclusion chromatogra-
phy(SEC)inamixtureofacetatebuffer(pH6.5;
0.3 mol ꢂ Lꢁ1) and methanol (20:80 v/v) as a
mobile phase on a TSK 3000 column (Polymer
Laboratories Ltd., UK) using HPLC System
N
H
S
O
O
+
O
H2N
SH
N
¨
AKTA Explorer (Amersham Biosciences; Swe-
n
den) equipped with RI, UV and multi-angle
light-scattering DAWN DSP-F (Wyatt, USA)
detectors. The refractive index increments of
Scheme 2. Scheme of synthesis of sulfanylethyl-2-(4-hydroxyphenyl)acetamide and
polyMeOX
(dn/dc ¼ 0.181 ꢄ 0.001 mL ꢂ gꢁ1
)
)
introduction of isotopically labelable phenolic moiety into the block copolymers.
andpolyIprOX(dn/dc ¼ 0.177 ꢄ 0.002 mLꢂ gꢁ1
in the mobile phase used for molecular weight
The triblock ABA copolymers poly[MeOX-b-(IprOX-co-BuOX)-b-
determinationweremeasuredonaBrice-Phoenix visuallaboratory
type differential refractometer BP-2000-V (Phoenix Precision
Instrument Co., USA), and the dn/dc of the copolymers was
2-MeOX] were synthesized as follows in nine alternatives with
three different central block compositions (10, 15 and 20 mol-%
BuOX, respectively) and thermoresponsive to hydrophilic block
ratios (1:2, 1:1 and 2:1 w/w, respectively) in all possible
combinations (see Scheme 3). The mixture of 2-isopropyl-2-
oxazoline with 10, 15 and 20 mol-% BuOX, respectively, and total
weights 0.66, 1.00 and 1.33 g, respectively, was polymerized using
diethylene glycol di(p-toluenesulfonate) (83 mg, 0.20) as the
calculated as
composition.
a
weighted average reflecting monomeric
The content of BuOX monomeric moiety in the thermorespon-
sive block poly[IprOX-co-BuOX] was calculated according to
Equation (1) from the 1H NMR spectra measured in CDCl3.
w1 ¼ S0:91=ðS0:91 þ ðS1:09=2ÞÞ
(1)
where S0.91 is the integral signal of the terminal ꢁCH3 hydrogen
nuclei in the BuOX monomeric unit at d ¼ 0.91, S1.09 is the integral
signal of the ꢁCH3 hydrogen nuclei in the IprOX monomeric units
at d ¼ 1.09 ppm and w1 is the mole fraction of the BuOX
monomeric unit in the thermoresponsive block.
The content of the MeOX monomeric moiety in the copolymer
was calculated according to Equation (2) from the 1H NMR spectra
measured in CDCl3.
w2 ¼ ðS2:09=3Þ=ððS3:46=4Þ ꢁ ðS1:09=6ÞÞ
(2)
where S2.09 is the integral signal of the ꢁCOꢁCH3 hydrogen nuclei
in the MeOX monomeric unit at d ¼ 2.09, S3.46 is the integral signal
of the oxazoline backbone NꢁCH2ꢁCH2ꢁN hydrogen nuclei at
Scheme 3. Schematic imaging of formation of triblock copoly-
mers.
Macromol. Biosci. 2010, 10, 916–924
918
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/mabi.201000034

Polyoxazoline Thermoresponsive Micelles as Radionuclide . . .
d ¼ 3.46 ppm, S1.09 is the integral signal of the ꢁCH3 hydrogen
nuclei in the IprOX monomeric units at d ¼ 1.09 ppm and w2 is the
mole fraction of the MeOX monomeric unit in the copolymer.
heating or cooling, five measurements were performed after
reaching steady state conditions.
The accurate determination of the shape factor of the particle
Rg/Rh was carried out by static and dynamic light scattering
measurements at a temperature just above a threshold, to avoid
multiple scattering influence, in the angular range 30–1508 using
an ALV instrument equipped with a 30 mW He-Ne laser (vertically
polarized light at l ¼ 632.8 nm). The light scattering data were
taken after a fixed waiting time of 20 min to achieve equilibration
of the sample. The Zimm plot procedure was used for the Rg
evaluation. Dynamic light scattering measurements were carried
out multiple times at a 908 angle. The obtained correlation
functions were analyzed by REPES analytical software providing a
hydrodynamic radius distribution function, G(Rh). To account for
the logarithmic scale on the Rh axis, all DLS distribution diagrams
are shown in the equal area representation, RhG(Rh). The values of
Addition of N-(2-Sulfanylethyl)-2-(4-
hydroxyphenyl)acetamide to Double Bonds of
Poly[MeOX-b-(IprOX-co-BuOX-co-EnOX)-b-MeOX]
Poly[MeOX-b-(IprOX-co-BuOX-co-EnOX)-b-MeOX] (100 mg) and N-
(2-sulfanylethyl)-2-(4-hydroxyphenyl)acetamide (100 mg) were
dissolved in N,N-dimethylacetamide (DMAc, 1.2 mL). Azobis(iso-
butyronitril) (AIBN, 50 mg) was dissolved in DMAc (800 mL). The
solution of polymer and thiol was heated at 100 8C for 24 h while
thesolutionofAIBNwasaddedinfour200 mLaliquotsevery2 h(i.e.,
at t ¼ 0, 2, 4 and 6 h). The solution was cooled to ambient
temperature, methanol (3.0 mL) was added and the polymer was
isolated by gel permeation chromatography on Sephadex LH-20
column using methanol as the eluent and evaporation of the
polymer containing fraction. The yield was 82 mg (82%).
Rh for each system were averaged over three runs.
If not otherwise stated, all measurements were done with a
polymer concentration of 0.5 g ꢂ Lꢁ1 in water and all solutions were
filtered through a 0.22 mm PVDF syringe filter before a measure-
ment.
The content of the phenolic moiety was determined by 1H NMR
in CDCl3 according to Equation (3):
cphenol ¼ 2 ꢀ 106 ꢀ S7:12=ðM ꢀ S3:46
Þ
(3)
Determination of Critical Micelle Concentration
(CMC) by Static Light Scattering Measurements
For polymers that have a cloud point temperature below body
temperature, CMC values were determined by static light
scattering experiments as the intersection points of straight lines
drawn through the data at small and large concentrations (see
where S7.12 is the integral signal of the Aryl-H hydrogen nuclei in
the ortho- position relative to the hydroxy group on the aromatic
ring at d ¼ 7.12 ppm, S3.46 is the integral signal of the oxazoline
backbone NꢁCH2ꢁCH2ꢁN hydrogen nuclei at d ¼ 3.46 ppm, M is
the average molecular weight of monomeric units weighted in
respect to their content in the copolymer and cphenol is the phenolic
unit content in micromol per gram of polymer.
Supporting Information for
a typical example). Static light
scattering measurements were carried out multiple times at a
908 angle using an ALV instrument equipped with a 30 mW He-Ne
laser (vertically polarized light at l ¼ 632.8 nm). Heating was
achieved by immersion of the polymer solution (ca 1 mL) in a
measuring cell placed into a pre-heated thermostated bath (37 8C).
Cloud Point Temperature Determination
Cloudpointtemperatures(CPT)ofthepolymersweredeterminedin
aqueous solutions of the particular polymer (25 mg ꢂ mLꢁ1) at a
heating rate of 1 8C ꢂ minꢁ1 with laser scattering detection.
Radioisotope Labeling Studies
Radioisotope labeling with 125I and stability studies were
performed as described in the literature.[19]
Static (SLS) and Dynamic (DLS) Light Scattering
The temperature-induced micelle formation in aqueous solutions,
temperature dependences of the apparent hydrodynamic radius of
the particle, Rh, and scattering intensity, Is, were automatically
measured at a scattering angle u ¼ 1738 on a Zetasizer Nano-ZS,
Model ZEN3600 (Malvern Instruments, UK). For evaluation of data,
the DTS (Nano) program was used. It provides a Rh intensity,
volume and number-weighted distribution function G(Rh). A
volume-weighted value of apparent Rh was chosen for the
monitoring of temperature changes in the system because it gives
a more realistic view of the characteristic particle size in a solution.
Such information is more valuable for assessing possible biome-
dical impact. The temperature dependences of micelle formation
and disintegration (a hysteresis measurement) were measured by
heating from 10 to 70 8C and back in variable steps: 0.5 or 1.2 8C
(slow temperature variations). After every temperature change,
Haemolytic Assay
Human blood from a healthy volunteer was collected in 5 mL
Heparin coated vacutainers (Greiner Bio-One, Austria) and
centrifuged for 5 min at 3 000 rpm. The blood sediment was twice
washed in fresh PBS (pH 7.4) and 4% human red blood cell (RBC)
suspension was prepared. The micellar delivery system was added
to RBC suspensions to achieve final concentrations in the range
1–1 000 mg ꢂ mLꢁ1. All samples were incubated at 37 8C in a
humidified 5% CO2 – 95% air atmosphere for either 30 min or 24 h
andthencentrifugedfor5 minat3000 rpm. Thesupernatantswere
spectrophotometricaly analyzed at 550 nm (Spectra Rainbow,
Tecan, Austria). Distilled water was used as a positive control of
haemolysis and PBS as a negative control. The value of haemolytic
activity was calculated by considering absorbance of positive
Macromol. Biosci. 2010, 10, 916–924
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.mbs-journal.de
919

M. Hruby et al.
control as 100% haemolysis. The results are the means of two
independent experiments, and the standard deviation was lower
than 5%.
solvation shell into the bulk water resulting in an entropy
gain is the driving force).
We have chosen BuOX as a more hydrophobic monomer
for this study. Synthesis of the model thermoresponsive
copolymers was carried out under the same conditions as
synthesis of the block copolymers (anhydrous acetonitrile,
42 8C). The molar weight was set to be the average value as
in the triblock copolymers (5.0 kDa). Methyl tosylate was
used as an initiator in this case. The copolymers were
obtainedinhighyields(>90%afterpurification), had molar
weights in good correspondence with theory (weight-
average molecular weights Mw ¼ 4.72 kDa in the ran-
ge ꢄ 0.19 kDa) and narrow molecular weight distributions
(polydispersities I ¼ Mw=Mn ¼ 1.15 in the range ꢄ 0.02,
where Mn is the number-average molecular weight); see
Supporting Information for a typical size-exclusion chro-
matogram. The content of the BuOX monomeric unit in the
copolymers is within the experimental error identical
(correlation coefficient 0.997, i.e., within 15 relative % of the
theoretical value in all cases) with the composition of the
polymerization mixture, consistent with the nearly quan-
titative polymer yield and thus nearly complete monomer
consumption. The monomer unit composition in the
copolymer was followed by 1H NMR (see Experimental
Part for details and Supporting Information for typical
NMR). An increase inthe contentofBuOX inthe copolymers
within a range 0–20 mol-% of BuOX causes a nearly linear
(R2 ¼ 0.996) decrease in their CPT (0.81 8C per mol-% BuOX).
The cloud point temperature was slightly higher at lower
copolymer concentrations in a solution and this concentra-
tion dependence was somehow more pronounced in
statistic IprOX-co-BuOX copolymers compared to pure
IprOX homopolymer (see Supporting Information), prob-
ablyduetostatisticalintermolecularheterogeneity(similar
to that reported for other polyoxazoline copolymers).[20,21]
The triblock copolymers poly[MeOX-b-(IprOX-co-BuOX)-
b-MeOX] were synthesized in high yields (> 90% after
purification). Their molar weights were only slightly lower
than theoretically expected (found Mw ¼ 8.58 kDa in the
range ꢄ0.34 kDa) with slightly higher polydispersities
(I ¼ 1.39 in the range ꢄ0.04) compared to the model
copolymers poly(IprOX-co-BuOX); see Supporting Informa-
tion for a typical size-exclusion chromatogram. The slightly
higher polydispersities may be due to the use of ethylene
glycol ditosylate as an initiator, which results in slower
initiation compared to methyl tosylate which yields
broader molar mass distributions. This is comparable to
the recently reported difference using butyne tosylate and
propargyl tosylate initiators.[22] Monomeric unit composi-
tion, as followed by 1H NMR (see Experimental Part for
details and Supporting Information for typical NMR) was
identical within the experimental error with the monomer
feed (correlation coefficient 0.996 within 15 relative-% of
the theoretical value in all cases), consistent with the near
Results and Discussion
Polymer Synthesis
The thermoresponsive polymeric micellar drug delivery
systemwasmadeofaABAtriblockcopolymerspoly[MeOX-
b-(IprOX-co-BuOX)-b-MeOX] with two hydrophilic A blocks
(polyMeOX) and one central thermoresponsive B block
[poly(IprOX-co-BuOX)] with different monomer unit ratios.
Thetriblockcopolymersweresynthesizedbylivingcationic
ring-opening polymerization using diethylene glycol dito-
sylate as the initiator. This highly reactive commercially
available bifunctional initiator allows synthesis of triblock
copolymer by a one pot method in two steps starting with
the middle copolymer block. In the first step, the mixture of
monomers forming the thermoresponsive block (IprOX þ
BuOX) is polymerized with diethylene glycol ditosylate
forming a thermoresponsive polymer block with two living
cationic ends. In the second step, the monomer forming the
hydrophilic terminal blocks (MeOX) was added and the
polymerization continued on both chain ends forming two
terminal hydrophilic polymer chains. The polymerization
was done under conditions (anhydrous acetonitrile, 42 8C)
described in the literature[2] to keep the polymerization
living without significant chain transfer or termination.
Copolymerization was set by the initiator-monomer ratio
to synthesize copolymers of a theoretical molar weight of
10 kDa. This molar weight is sufficiently low to allow
eliminationofunimersbykidneys(therenalthresholdisca.
45 kDa for hydrophilic polymers) after disassembly of the
system, but still sufficiently high to suppress intermole-
cular heterogeneity (given by the statistical nature of the
IprOX-BuOX copolymerization) among the polymer chains.
Before using central thermoresponsive IprOX-BuOX
copolymer for the synthesis of the triblock copolymers,
we optimized its composition with respect to the cloud
point temperature. IprOX homopolymer itself has rela-
tively high cloud point temperature only slightly below
body temperature (33 8C at cpolymer ¼ 25 mg ꢂ mLꢁ1). This is
why it should be decreased to assure formation of micelles
at body temperature if also more hydrophilic corona-
forming blocks are present in the macromolecule. The
hydrophilic blocks significantly increase the micelle-form-
ing temperature compared to the cloud point temperature
of the thermoresponsive block alone (see below). The
presence of hydrophobic groups (monomer units) in the
polymer makes solvation interactions weaker, which
results in a decrease in CPT of such copolymers (LCST is
an entropy-driven transition, i.e., the release of water in the
Macromol. Biosci. 2010, 10, 916–924
920
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/mabi.201000034

Polyoxazoline Thermoresponsive Micelles as Radionuclide . . .
quantitative conversion, as was also the case for the model
copolymers.
monodisperse aggregates or mesoglobules are formed.[23]
It was noticed that the mesoglobules occurring above the
CPT adopt a core-shell structure resembling a micelle.
Inthemajorityofcases, thehysteresisinthetemperature
dependence of apparent Rh is manifested. One can see that
the mesoglobules disintegrate at a lower temperature in
comparison with the one of creation. The Rh value of the
mesoglobules also shows non-monotonous behavior near
the CPT. Occurrence of hysteresis was explained by the
presence of intra- or interchain hydrogen bonds that are
broken during cooling.[23,25]
Temperature-responsive Aggregation Behavior
Static and dynamic light scattering methods are common
tools to characterize temperature-dependent structural
changes of thermoresponsive polymers with LCST. The
intensityofthescatteredlightissensitivetothemolarmass
andthesizeofthescattersandthuscanbeusedtofollowthe
phase separation. A detailed review of the properties of
temperature-sensitive polymers, including light scattering
studies, has been published.[23] Recently, light and neutron
scattering methods were exploited extensively for the
investigation of these polymers, e.g., of poly(N-isopropyla-
crylamide)[24,25] (PNIPAM), poly(N-vinyl caprolactam)[24,26]
(PVCL), poly(methyl vinyl ether)[24] (PVME), hydrophobi-
cally modified poly(N-isopropylacrylamide) (HM-PNI-
PAM)[27–31] or of hydrophobically modified polyoxazo-
lines.[32–34]
Our light scattering experiments prove that the synthe-
sized thermosensitive polyoxazolines have similar features
to the thermosensitive polymers described above. We have
completed a basic characterization in water to make the
system as principally transparent as possible. In physiolo-
gical solution (0.9 wt-% NaCl), the temperature-dependent
behavior is exactly the same in shape, but everything is
shifted 3 8C to lower temperatures. All samples visually
become turbid when heated above CPT where dynamic
light scattering reveals the formation of nanosize objects
(seeSupportingInformationforchart). Oncoolingdownthe
samples to below CPT, turbidity disappears, so a transpar-
ent solution is formed. Formation of nanoparticles is
completely reversible in all cases. Because the perspective
structure of nanoparticles built from the studied polymers
will be a core-shell one, we will exploit hereafter the term
micelle rather than mesoglobule.
Several scenarios for the structural transition in a
solution during phase separation could come to life,
depending on the chemical composition and concentration
of the thermoresponsive polymer. It was proven that
PNIPAM with high molecular weight at low concentrations
exists in a coil conformation below the LCST. On approach-
ing the LCST, a homopolymer undergoes a coil-globule
transition, wherein the loose macromolecules at first
collapse into a compact globule with further aggregation
of globules into a so-called mesoglobule. The PNIPAM
mesoglobules are uniform and colloidally stable particles
with some amount of water. Once created, they have a
tendency to shrink with increasing temperature. At high
concentrations, PNIPAM macromolecules form intermole-
cularaggregatesbelowtheCPT.Thepropensitytoaggregate
in cold water below the CPT was found for other
thermosensitive homopolymers - PVCL and PVME, dis-
regarding their molecular weight and concentration. Loose
and polydisperse aggregates transform above the CPT into
nearly monodisperse compact (although non-uniform)
mesoglobules with a sponge-like structure. The density
of such mesoglobules declines from the center to the
periphery. The third scenario is realized when a macro-
molecule contains chemically different moieties: thermo-
responsive/hydrophobicorthermoresponsive/hydrophilic.
Inclusion of hydrophilic groups in a macromolecule shifts
CPT to higher values; hydrophobic moieties have an
opposite effect. Below CPT at low concentrations, hydro-
phobically modified macromolecules form multi-chain
assemblies, presumably flower-like micelles, with low
aggregation numbers.[25,27,30–33] At higher concentrations,
flower-like micelles form inter-micelle aggregates with
lower density. When approaching the CPT, flower-
like micelles undergo a coil-to-globule transition and
The CPT value is a function of the polymer composition
(see Supporting Information for chart and Table 1). One can
seethatthroughthewholepolymerseriestheCPT increases
as the content of hydrophilic MeOX groups increases.
Conversely, increasing the content of hydrophobic BuOX
moieties results in CPT reduction.
In all cases, the micelle-formation temperatures of the
triblock copolymers are significantly higher than CPTs of
the thermoresponsive copolymers of the same composition
as the thermoresponsive block of the triblock copolymer. If
we also compare triblock copolymers with different
thermoresponsiveness to hydrophilic block ratios (but
the same ratio of monomeric units in the thermoresponsive
block), the increase in the hydrophilic block content
considerably increases CPT. One can thus conclude that
the overall hydrophilicity/hydrophobicity of the whole
triblock copolymer is at least of the same importance as the
ratio of monomeric units and subsequently the CPT of the
thermoresponsive block itself. These triblock copolymers
thusbehavemorelikePluronics[poly(ethyleneoxide-block-
propyleneoxide-block-ethyleneoxide)blockcopolymers]or
poly(ethyleneoxide-block-lactide)thanlike, forexample, N-
isopropyl acrylamide copolymers, where the CPT of the
thermoresponsive block itself is dominant in determina-
tion of the thermal behavior of the block and graft
copolymers. The effect also cannot be attributed to the
Macromol. Biosci. 2010, 10, 916–924
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.mbs-journal.de
921

M. Hruby et al.
Table 1. Chemical composition and properties of ABA triblock copolymers poly[2-methyl-2-oxazoline-block-(2-isopropyl-2-oxazoline-co-2-
butyl-2-oxazoline)-block-2-methyl-2-oxazoline]; nMeOX - average number of 2-methyl-2-oxazoline monomeric units per polymer chain,
nIprOX - average number of 2-isopropyl-2-oxazoline monomeric units per polymer chain, nBuOX - average number of 2-butyl-2-oxazoline
monomeric units per polymer chain.
Triblock
colymer
Thermoresponsive to
hydrophilic block
BuOX in
Tdem
CMC at 37 -C
nMeOX
nIprOX
nBuOX
thermoresponsive block
w/w
mol-%
-C
g ꢂ mLꢁ1
P2574-1:2
P2574-1:1
P2574-2:1
P2639-1:2
P2639-1:1
P2639-2:1
P2620-1:2
P2620-1:1
P2620-2:1
1:2
1:1
2:1
1:2
1:1
2:1
1:2
1:1
2:1
10
10
10
15
15
15
20
20
20
62
63
64
44
28
27
42
38
28
–
78
59
39
78
59
39
78
59
39
27
40
53
25
38
50
24
35
47
2.6
3.9
5.2
3.9
5.9
7.9
5.2
7.9
10.5
–
–
–
1.0 ꢀ 10ꢁ4
2.5 ꢀ 10ꢁ5
–
–
5 ꢀ 10ꢁ5
effect of the molecular weight of the thermoresponsive
blocks, because the molecular weights of the model
thermoresponsive copolymers and of the thermorespon-
sive block in the triblock copolymers are comparable.
Below the CPT, the values of the hydrodynamic radius of
the P2639-2:1 polymer correspond to the size of single
macromolecules (1–2 nm average), implying that that they
are fullydissolved (Figure1(a)). The scatteredlightintensity
increases approaching the CPT (Figure 1(b)). At the CPT it
sharply increases and a simultaneously increasing appar-
ent Rh indicates the formation of micelles (Figure 1).
Although all the triblock copolymers form micelles at
certain temperatures, their properties above the CPT differ,
depending on the ratio between hydrophobic, hydrophilic
and thermosensitive moieties. Thus for P2620-1:1 polymer
with the highest content of hydrophobic BuOX groups
(20 mol-% in the thermosensitive block) and medium
hydrophilicity, it was observed that the nanoparticle size
continuously decreases with increasing temperature
(Figure 2(a)). Meanwhile, the intensity of the scattered
light grows (Figure 2(b)). Heating and cooling measure-
mentsshowsminorhysteresisofabout2 8C(seeFigure2(a)).
In contrast to that, for solutions of P2639-1:2, the intensity
and size of the micelles above the CPT permanently grows
(Figure 3). Heating and cooling measurements show almost
complete absence of hysteresis. Hydrophobic/hydrophilic
interactions among whole macromolecules (as a result of
the overall hydrophilicity/hydrophobicity of the polymer
chains) thus become dominant above CPT for the P2639-1:2
polymer compared to polymer P2620-1:1 due to the shorter
length and lower hydrophobicity of the thermoresponsive
block (a switch from block polymer-type behavior shown in
Figure 2(a) and 2(b)).
The trend of changes during heating and cooling in
apparent Rh for P2639-2:1 is quite differ-
450
ent to those of P2639-1:2 and P2620-1:1
a)
350
(see Supporting Information for chart).
The apparent Rh dramatically increases
250
100
10
1
above the cloud point and the sharp peak
150
50
in the Rh vs. temperature dependence
0
20
40
50
T, oC
reveals the formation of large intermo-
2.5x104
2.0x104
1.5x104
1.0x104
10
20
30
40
60
70
lecular aggregates. At higher tempera-
tures, aggregates gradually shrink and
micelles are formed above 40 8C. The
shape factor (Rg/Rh) values at tempera-
tures 15–20 8C above the CPT are approxi-
mately 0.8–1. Minor hysteresis is detect-
b)
5.0x103
0.0
10
20
30
40
50
60
70
able for the polymer (see Supporting
T, oC
Informationforchart).Suchbehaviorisin
agreement with previous observations
Figure 1. Temperature dependence of volume-weighted Rh (a) and the intensity of
scattered light Is (b) of P2639-2:1 polymer at u ¼ 1738, cooling (cP ¼ 0.5 g ꢂ Lꢁ1).
for thermoresponsive polymers reported
Macromol. Biosci. 2010, 10, 916–924
922
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/mabi.201000034

Polyoxazoline Thermoresponsive Micelles as Radionuclide . . .
forming micelles at this temperature were measured) are
somehow higher but comparable with the CMC obtained
with purely amphiphilic block and graft copolymers, in
accordance with the literature. As mentioned above, this
may lead to shorter blood circulation times, but, on the other
hand, should facilitate polymer elimination from the
organism after the system fulfills its task.
200
a)
150
100
50
0
0
10
20
30
40
50
60
70
8000
4000
0
b)
Radionuclide Labeling
Since copolymer micelles are intended as carriers for
radiodiagnostics/radiotherapeutics and polyoxazolines
without suitable functional moieties cannot be directly
radiolabeled, we introduced radiolabelable phenolic moi-
ety into the copolymers. Phenol is a highly activated
aromatic moiety towards electrophilic radioiodination
with iodine radioisotopes suitable for radiodiagnostics
0
10
20
30
40
50
60
70
T, oC
Figure 2. Temperature dependence of volume-weighted Rh
(a) and the intensity of scattered light Is (b) of P2620-1:1 polymer
at u ¼ 1738, heating (*) and cooling (&) (cP ¼ 0.5 g ꢂ Lꢁ1).
(
123I, 124I) or radiotherapy (131I). We have chosen the
polymer P2639-1:1 containing 15 mol-% BuOX in the
thermoresponsive block and a 1:1 thermoresponsive to
hydrophilic block ratio as a starting copolymer for further
modification due to its suitable properties (micelle-forming
temperature and RH, see Table 1 and Supporting Informa-
tion). The radiolabelable moiety was introduced into
the triblock copolymer by the introduction of pendant
double bonds into the copolymer (part of BuOX in the
polymerization mixture was substituted with EnOX to
achieve on average one double bond per polymer molecule)
and thiol-clock addition in analogy to ref.[18] of sulfany-
lethyl)-2-(4-hydroxyphenyl)acetamide onto the double
bond (see Scheme 2). In analogy to ref.[35], AIBN was used
as an initiator instead of UV-light. UV-light photocatalysis
was originally described for another thiol-ene click reaction
to polyoxazolines,[36] but the photoinitiated reaction
offered unsatisfactory conversions only in our case. No
polymer-boundinitiatorfragmentscanbedetectedbyNMR
since there is an excess of the thiol in the reaction mixture.
The achieved phenolic moiety content (36.2 mmol ꢂ gꢁ1
polymer) and thus the theoretical labeling capacity
(61 GBq 125I/mg polymer) calculated from the phenol
600
a)
300
0
10
20
30
40
50
60
70
2000
1000
0
b)
10
20
30
40
T, oC
50
60
70
Figure 3. Temperature dependence of volume-weighted Rh
(a) and the intensity of scattered light Is (b) of P2639-1:2 polymer
at u ¼ 1738, heating (*) and cooling (&) (cP ¼ 0.5 g ꢂ Lꢁ1).
in the literature.[23] It is reasonable to assume that due to
the smallest content of hydrophilic MeOX groups in the
macromolecularstructure, thethermosensitiveIprOXblock
have control over behavior of the aggregates above the CPT.
A very important property of micelles designed for drug
delivery purposes is their critical micellar concentration
(CMC). The micelles must be sufficiently stable in dilution in
the bloodstream (total blood volume in humans is ca. 5 L).
They should not contain too much unimer in equilibrium
withthemicelles,sincehydrophobicdomainsofunimersare
notprotectedfromunwantedinteractionsinthebodybythe
hydrophilic corona and unimers may also be quickly
eliminated through the kidneys due to their relatively low
molecularweight.Ontheotherhand,toostablemicellesmay
have a problem with a too slow rate of elimination from the
organism. The CMC values for the polymers (Table 1, since
the measurements were carried out at 378C, only polymers
1
content assayed by H NMR was more than sufficient for
the possible application (the typical dose per human
patient is 0.02–3 GBq depending on the particular radio-
nuclide and application).
Radionuclide labeling was carried out using the chlor-
amine method in good yield (66%). Most of the radioiodine
was bound to polymer by a stable bond (see Supporting
Information for a chart) and the remaining part was
gradually released into low molecular weight fraction
during incubation in PBS buffer at 37 8C. This means that
part of the iodine is bound by a metastable bond; however,
if GPC separation on a PD-10 desalting column is repeated
after 2.5 h, a more stable product is obtained and should be
sufficient for the intended radiodiagnostic purposes.
Macromol. Biosci. 2010, 10, 916–924
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.mbs-journal.de
923

M. Hruby et al.
Haemolytic activity is a straightforward measure of
eventual membrane toxicity due to the amphiphilic
character of the copolymers. All the copolymers have
shown no toxicity (haemolysis less than 3.5% even at the
highest concentration used (1 mg ꢂ mLꢁ1) which corre-
sponds to the hypothetical total dose of 5 g per human
with a total blood volume of 5 L, (see Supporting Informa-
tion for a chart).
[7] R. Hoogenboom, F. Wiesbrock, H. Huang, M. A. M. Leenen,
H. M. L. Thijs, S. F. G. M. van Nispen, M. van der Loop, C.-A.
Fustin, A. M. Jonas, J. F. Gohy, U. S. Schubert, Macromolecules
2006, 39, 4719.
[8] R. Hoogenboom, F. Wiesbrock, M. A. M. Leenen, H. M. L. Thijs,
H. Huang, C.-A. Fustin, P. Guillet, J.-F. Gohy, U. S. Schubert,
Macromolecules 2007, 40, 2837.
[9] M. Einzmann, W. H. Binder, J. Polym. Sci. A, Polym. Chem. 2001,
39, 2821.
[10] A. Mahmud, X. B. Xiong, H. M. Aliabadi, A. Lavasanifar, J. Drug
Targeting 2007, 15, 553.
[11] V. P. Torchilin, Pharm. Res. 2007, 24, 1.
Conclusion
[12] M. Hruby, C. Konak, J. Kucka, M. Vetrik, S. K. Filippov,
D. Vetvicka, H. Mackova, G. Karlsson, K. Edwards,
B. Rihova, K. Ulbrich, Macromol. Biosci. 2009, 9, 1016.
[13] H. Uyama, S. Kobayashi, Chem. Lett. 1992, 9, 1643.
[14] D. Christova, R. Velichkova, W. Loos, E. J. Goethals, F. Du. Prez,
Polymer 2003, 44, 2255.
We have synthesized ABA triblock copolymers poly[2-
methyl-2-oxazoline-block-(2-isopropyl-2-oxazoline-co-2-
butyl-2-oxazoline)-block-2-methyl-2-oxazoline] with two
hydrophilic A blocks and one central thermoresponsive B
block with different monomer unit ratios. These polymers
are soluble in aqueous millieu, molecularly dissolved below
the cloud point temperature of the thermoresponsive block
and form micelles at higher temperature. Micelles are
formed within a narrow temperature range. The CPT of the
thermoresponsive block was adjusted with 2-butyl-2-
oxazoline (hydrophobic monomer lowering the CPT) to 2-
isopropyl-2-oxazoline (main monomer giving thermore-
sponsive properties to its copolymers) ratio and size of the
micelles was also influenced by the A to B block weight
ratio. A phenolic moiety was introduced into the above
stated polymer to allow radionuclide labeling with iodine
radioisotopes for both diagnostics and therapy of solid
tumors. Such polymer was then radiolabeled with 125I in
good yield with sufficient in vitro stability under model
conditions.
[15] J. S. Park, K. Kataoka, Macromolecules 2006, 39, 6622.
[16] J. S. Park, K. Kataoka, Macromolecules 2007, 40, 3599.
[17] R. Hoogenboom, H. M. L. Thijs, M. J. H. C. Joachems, B. M. van
Lankvelt, M. W. M. Fijten, U. S. Schubert, Chem. Commun.
2008, 44, 5758.
[18] A. Gress, A. Volkel, H. Schlaad, Macromolecules 2007, 40, 7928.
[19] M. Hruby, V. Subr, J. Kucka, J. Kozempel, O. Lebeda, A. Sikora,
Appl. Radiat. Isot. 2005, 63, 423.
[20] S. Huber, R. Jordan, Colloid Polym. Sci. 2008, 286, 395.
[21] L. Del Villano, R. Kommedal, M. W. M. Fijten, U. S. Schubert,
R. Hoogenboom, M. A. Kelland, Energy Fuels 2009, 23, 3665.
[22] M. W. M. Fijten, C. Haensch, B. M. van Lankvelt,
R. Hoogenboom, U. S. Schubert, Macromol. Chem. Phys.
2008, 209, 1887.
[23] V. O. Aseyev, H. Tenhu, F. M. Winnik, ‘‘Temperature Depen-
dence of the Colloidal Stability of Neutral Amphiphilic Poly-
mers in Water’’, in: Conformation-Dependent Design of
Sequences in Copolymers II, Advances in Polymer Science
Series Volume 196, Springer, Berlin/Heidelberg 2006, p. 1.
[24] V. Aseyev, S. Hietala, A. Laukkanen, M. Nuopponen,
O. Confortini, F. E. Du Prez, H. Tenhu, Polymer 2005, 46,
7118.
Acknowledgements: Financial support from the Academy of
Sciences of the Czech Republic (grant # KAN 200200651), the Grant
Agency of the Czech Republic (grant # P207/10/P054, # 202/09/2078)
and the Ministry of Education, Youth and Sports of the Czech
Republic (grant # IM 4635608802) is gratefully acknowledged.
[25] P. Kujawa, V. Aseyev, H. Tenhu, F. M. Winnik, Macromolecules
2006, 39, 7686.
[26] A. Laukkanen, L. Valtola, F. M. Winnik, H. Tenhu, Macromol-
ecules 2004, 37, 2268.
[27] P. Kujawa, H. Watanabe, F. Tanaka, F. M. Winnik, Eur. Phys. J. E
2005, 17, 129.
Received: January 25, 2010; Revised: March 8, 2010; Published
online: May 20, 2010; DOI: 10.1002/mabi.201000034
[28] K. Akiyoshi, E. C. Kang, S. Kurumada, J. Sunamoto, T. Principi,
F. M. Winnik, Macromolecules 2000, 33, 3244.
[29] T. Koga, F. Tanaka, R. Motokawa, S. Koizumi, F. M. Winnik,
Macromolecules 2008, 41, 9413.
Keywords: drug delivery systems; micelles; radiation; synthesis;
thermal properties
[30] P. Kujawa, F. Tanaka, F. M. Winnik, Macromolecules 2006, 39,
3048.
[31] P. Kujawa, F. Segui, S. Shaban, C. Diab, Y. Okada, F. Tanaka,
F. M. Winnik, Macromolecules 2006, 39, 341.
[32] R. Obeid, F. Tanaka, F. M. Winnik, Macromolecules 2009, 42,
5818.
[1] N. Adams, U. S. Schubert, Adv. Drug Delivery Rev. 2007, 59,
1504.
[2] J. S. Park, Y. Akiyama, F. M. Winnik, K. Kataoka, Macromol-
ecules 2004, 37, 6786.
[33] R. Obeid, E. Maltseva, A. F. Thunemann, F. Tanaka, F. M.
Winnik, Macromolecules 2009, 42, 2204.
[3] R. Hoogenboom, Angew. Chem., Int. Ed. 2009, 48, 7978.
[4] B. Pidhatika, J. Moller, V. Vogel, R. Konradi, Chimia 2008, 62,
264.
[34] N. Morimoto, R. Obeid, S. Yamane, F. M. Winnik, K. Akiyoshi,
Soft Matter 2009, 5, 1597.
[5] C. Diab, Y. Akiyama, K. Kataoka, F. M. Winnik, Macromolecules
2004, 37, 2556.
[35] M. Hruby, C. Konak, K. Ulbrich, J. Controlled Release 2005, 103,
137.
[6] P. Persigehl, R. Jordan, O. Nuyken, Macromolecules 2000, 33,
6977.
[36] C. Diehl, H. Schlaad, Macromol. Biosci. 2009, 9, 157.
Macromol. Biosci. 2010, 10, 916–924
924
ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/mabi.201000034