Inter-micellar dynamics in block copolymer micelles: FRET experiments of macroamphiphile and payload exchange

Article abstract of DOI:10.1016/j.reactfunctpolym.2010.10.010

The co-formulation of micelles bearing different targeting groups and different payloads could allow the selective and contemporaneous treatment of various cell types with different drugs. The selectivity of such a system, however, would be compromized if

Full text of DOI:10.1016/j.reactfunctpolym.2010.10.010

Reactive & Functional Polymers 71 (2011) 303–314
Contents lists available at ScienceDirect
Reactive & Functional Polymers
journal homepage: www.elsevier.com/locate/react
Inter-micellar dynamics in block copolymer micelles: FRET experiments
of macroamphiphile and payload exchange
Ping Hu a, Nicola Tirelli b,c,
⇑
a School of Pharmacy and Pharmaceutical Sciences, University of Manchester, Oxford Road, Manchester M13 9PT, United Kingdom
b School of Materials, University of Manchester, Oxford Road, Manchester M13 9PT, United Kingdom
c School of Biomedicine, University of Manchester, Oxford Road, Manchester M13 9PT, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 2 November 2010
The co-formulation of micelles bearing different targeting groups and different payloads could allow the
selective and contemporaneous treatment of various cell types with different drugs. The selectivity of
such a system, however, would be compromized if macroamphiphiles and/or payloads would undergo
inter-micellar exchange, homogenizing the bio-functionalization and the content of the co-formulated
micelles.
Here we have investigated the occurrence of exchange phenomena in micelles of poly(propylene sul-
fide)–poly(ethylene glycol) (PPS–PEG) block copolymers, employing fluorophores (dansyl groups) and
quenchers (dabsyl groups) either as terminal groups in macroamphiphiles or as encapsulated hydropho-
bic payloads. Upon exchange, the increased proximity between dansyl and dabsyl groups would signifi-
cantly increase the quenching efficiency. Our results showed that even employing a rather hydrophilic
block copolymer (PPS10–PEG44) no significant macroamphiphile exchange could be detected within
24 h from preparation. The payload exchange was temperature-dependent and could be substantially
avoided for days if appropriately low storage temperatures are used.
Keywords:
Block colymer micelles
PEG
Polysulfides
Vinyl sulfone
FRET
We also present an improved experimental procedure for the synthesis of vinyl sulfone-terminated
PEG and PPS–PEG and for the conjugation of these structures with labels or possibly bioactive groups.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
polyelectrolyte complexation between oppositely charged blocks,
such as poly(lysine) and poly(aspartic acid) [8] or nucleic acids
Micelles based on block copolymers have been extensively used
as colloidal carriers for poorly water soluble molecules [1,2].
Among their distinctive advantages, they have considerably lower
critical micellar concentrations (CMCs) than those of low molecu-
lar weight compounds, and therefore higher stability against dilu-
tion [3].
The self-assembly of micellar aggregates is generally based on
hydrophobic association. The pioneering work of the group of
Kataoka has provided a number of such systems where the associ-
ating blocks have been based e.g. on poly (b-benzyl aspartate) [4],
[9]. Among the most popular micellar systems for the solubiliza-
tion of drugs, the poly(ethylene glycol)/poly(propylene glycol)
di- and tri-block (Pluronics or Poloxamers) copolymers occupy a
position of preminence [10,11]. The presence of poly(ethylene
glycol) (PEG) ensures a prolonged circulation of Pluronic in vivo,
but, although these polymers are non-biodegradable, the possibility
of renal excretion reduces concerns about their long-term perma-
nence [12,13]. Their relatively high CMC values, ranging between
0.01 and 10 wt.% for both di- and triblocks [1,14], appear to increase
the permeation biological barriers through a double effect of
unimers or poorly aggregated micelles that deplete intracellular
ATP and inhibit drug efflux transporters [15,16], such as the P-glyco-
protein, e.g. allowing to overcome multi-drug resistance of cancer
cells [17] and allow to cross the blood–brain-barrier [18].
or drug-functionalized poly(aspartic acid) [5] or poly(L,D-lactide)
[6,7]. Other associative mechanisms have been used too, e.g.
Polymeric micelles can be easily rendered environmentally
responsive: for example, drugs can be released upon (endosomal)
acidification because of the presence of pH-sensitive linkages, such
as acetals [19] or Schiff bases [20,21], or because of a hydrophobic-
to-hydrophilic transition of the micelle core following the proton-
ation of aliphatic amines [22] or histidines [23]. Active biological
Abbreviations: PS, Propylene sulfide; TEA, Triethylamine; DBU, 1,8-Diazabicyclo
undec-7-ene; TBP, Tributyl phosphine; DVS, Divinyl sulfone; PPS, Poly(propylene
sulfide); PEG, Poly(ethylene glycol); DA, Dansyl; DB, Dabsyl.
⇑
Corresponding author at: School of Materials and School of Biomedicine,
University of Manchester, Oxford Road, Manchester M13 9PT, United Kingdom. Tel.:
+44 1612752480.
E-mail address: nicola.tirelli@manchester.ac.uk (N. Tirelli).
1381-5148/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.reactfunctpolym.2010.10.010

304
P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
targeting can been achieved through the decoration with antibod-
ies [24], adhesion or cell-penetrating peptides [25], biotin [23], etc.
Our group has focused its attention on Pluronic-like macromol-
ecules with a specific response to oxidation, which is obtained by
replacing the Pluronic lipophilic component (poly(propylene gly-
col)) with the more hydrophobic poly(propylene sulfide) (PPS).
PPS is hydrophobic but can be rendered hydrosoluble via oxidation
of sulphur (II) to higher oxidation states [35] and this can allow for
a response to the oxidative environment typical of inflammatory
conditions [26–28], where the production of oxidants (e.g. Reactive
Oxygen Species, ROS) is generally induced by the presence of
inflammatory cytokines [29] and regulate the activity of inflamma-
tory and phagocytic cells [30]. The cornerstone of the synthesis of
PPS–PEG block copolymers is the living, anionic ring-opening poly-
merization of propylene sulfide. As a result of their mild character
of the propagating species (thiolates) the polymerization is toler-
ant to a wide variety of functional groups and can even be con-
ducted in the presence of water [31]; further, end-capping can be
conveniently performed using terminal thiolates, e.g. in Michael-
type additions with acrylates [32] or in substitution reactions on
organic halides [33,34]. Using low MW mono- or multifunctional
protected thiols (thioacetates) as initiators, and thiol-reactive
PEG derivatives as end-cappers it is possible to prepare linear or
star diblock structures [33,34], while PEG macroinitiators would
lead to asymmetric or symmetric PEG–PPS–PEG triblock polymer
structures [32]. This class of block copolymers has been used for
the decoration of gold [36] or nanotube [37] surfaces, but it has
also been extensively employed to generate vesicular [35,38] or
micellar [34] aggregates for the encapsulation of, respectively,
hydrophilic [39] or hydrophobic [40] payloads.
micelles is most often seen as the release of unimers in the water
phase from a micelle and their uptake by a different one, a process
generally referred to as the Aniansson–Wall mechanism [42]. The
exchange kinetics would be essentially determined by the length
of the hydrophobic block, although a retardation effect for thicker
hydrophilic coronas is possible. This picture gives a central role to
the equilibrium concentration of unimers, and this also implies a
faster exchange for amphiphiles with higher CMC; indeed, this
model provides a good description of the behavior of polymers
with very hydrophobic cores, such as blocks of poly(ethylene-alt-
propylene) [43–45], or poly(a-methylstyrene) [46]. However, it
has also been suggested that the exchange may be in some cases
dominated by events of micellar fusion and fission [47], which
have also been predicted in computer simulations [48]; in this
case, as reviewed by Denkova et al. [49], the concentration of block
copolymer would strongly influence the exchange kinetics.
Here we have tried to develop a qualitative picture of the inter-
micellar exchange of both macromolecular amphiphiles and
hydrophobic low molecular weight payloads by using Förster Res-
onance Energy Transfer (FRET, also called Fluorescence Resonance
Energy Transfer). The non-radiative energy transfer between a
fluorophore (a ‘‘photon donor’’) and a quencher (a ‘‘photon accep-
tor’’) typically shows a 1/R6 dependence (R being the distance be-
tween fluorophore and quencher) and generally becomes
negligible at distances larger than ꢀ10 nm [50]. However, the in-
ter-chromophoric distance for an effective quenching is propor-
tionally larger for fluorophore/quencher couples with
a
substantial spectral overlap. We are interested in the co-localiza-
tion of acceptor and donor groups present in different (macro)mol-
ecules, which therefore are unlikely to be in intimate contact; we
have therefore chosen dansyl and dabsyl chromophores, which
are characterized by an ample overlap of emission and absorption
spectra. Additionally, they have good stability under irradiation
and are commercially available as amine-reactive labels. The emis-
sion of dansylated amines (e.g. terminal groups on dendrimers [52]
or side chains of organic polymers [53]) upon excitation in their
near UV absorption band (330–340 nm) covers the spectral range
between 450 and 600 nm with a maximum at about 500–520 nm
in solvents from moderate-to-strong polarity (dichloromethane
to water) [51], Dansyl groups have also been used as acceptors in
FRET experiments, emitting upon non-radiative energy transfer
from fluorophores such as ethenoadenosine [54] . More commonly,
however, they have been used as FRET donors in combination with
Here we have investigated the possibility to prepare formula-
tions of micelles composed of identical amphiphilic polymers but
bearing different payloads and surface (targeting) groups, which
could address multiple cellular targets at the same time
(Scheme 1); for example, it could be possible to use a single formu-
lation to deliver a cytotoxic drug to cancer cells, an anti-inflamma-
tory drug to inflammatory cells and a protective principle (e.g. an
anti-oxidant) to normal tissue cells. In order to achieve selectivity
in both targeting and delivery, however, any exchange between
functional (targeting) macroamphiphiles and between payloads
should be minimized.
Following the initial interpretation by Halperin and Alexander
[41], the exchange of macromolecular amphiphiles between
Scheme 1. The possibility of using a library or at least a co-formulation of identical micellar carriers differing in targeting groups and payloads could open the way to
contemporaneously operate different cellular-targeted therapies. This would be possible only if exchange phenomena are minimized.

P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
305
azobenzene groups, for example allowing to monitor the formation
of linkages between molecular fragments [55]; for example, dansyl
FRET experiments have been performed using p-phenylazoben-
zene sulfonyl (PABS) [56], 4-[(N,N-diethylamino)-phenylazo]
benzene-40-sulfonyl (DPBS) [56], 4-[4-(N,N-dimethylamino)pheny-
lazo]-40-benzoyl (dabcyl) [57,58] and 4-[4-(N,N-dimethylamino)
phenylazo]-40-sulfonyl (dabsyl) [57] derivatives as acceptors.
In particular, dabsyl chloride is a very stable, amine-reactive
azobenzene, which was originally introduced to allow detection
of peptides [59] and proteins [60] at visible light wavelengths;
its push–pull structure determines a shift of the azobenzene
absorption band in the visible range, thus dabsylated amines pres-
ent a broad absorption in the region 370–560 nm (kmax ꢀ 430–
460 nm for dabsylated amines on polymer termini or colloidal par-
ticles [61,62]), which almost completely overlaps dansyl emission.
In order to follow the payload exchange, we have prepared dansyl
and dabsyl hexylamides, as model hydrophobicpayloads (Scheme 2).
The macrophiphile exchange was studied reacting amine-
terminated PPS–PEG derivatives with dansyl or dabsyl chloride. It
is worth to notice that the macroamphiphile structure was free of
cleavable groups: esters [32–35,63] or disulfides [40] have also been
used as linkers, but their lability could determine the release of free
or PEGylated chromophores, compromising the FRET experiments.
We have therefore employed sulfone linkers, which were obtained
through the Michael-type addition of thiols (PPS thiolate or cyste-
amine) on vinyl sulfone-terminated PEG (Scheme 2).
methoxide, 1,8-diazabicyclo[5.4.0]undec-7-ene(1, 5-5)(DBU). THF
was degassed by bubbling argon under inert atmosphere for 1 h
before use.
2.2. Characterization
2.2.1. Molecular characterization
1H NMR spectra were recorded on 1 wt.% polymer solutions in
deuterated chloroform using a 300 MHz Bruker spectrometer. FT-
IR spectra were recorded in ATR mode (Golden gate) on a Tensor
27 Bruker spectrometer. GPC was performed in THF on a Polymer
Laboratories GPC 50 equipped with refractive index and viscosity
detectors, using universal calibration with poly(styrene) standards.
2.2.2. Dynamic light scattering
Size distributions and scattering intensity of micelles were
measured with the help of Zetasizer Nano ZS Instrument (Model
ZEN2500, Malvern Instruments Ltd., UK). All the samples were ana-
lyzed at an angle of 114° and a temperature of 25 °C.
2.2.3. Fluorescence
Fluorescent intensities of the samples were measured by a Bio-
Tek Synergy 2 multi-mode microplate reader at a temperature of
25 °C (filters at excitation of 360 40 nm and emission of
528 20 nm).
2.2.4. UV–Vis
The analyses of the UV spectra of the samples were carried out
using a PerkinElmer lamda25 spectrophotometer.
2. Experimental section
2.1. Materials
2.3. Polymer synthesis and derivatization
All materials were used as received from the supplier (Aldrich,
Gillingham, United Kingdom, for acetic acid, molecular sieves,
poly(ethylene glycol) diol, poly(ethylene glycol) monomethyl ether
(both Mn = 2000 g/mol), propylene sulfide, sodium hydride, acetic
acid, triethylamine, tributyl phosphine, divinyl sulfone, dansyl
chloride, dabsyl chloride; Fluka, Gillingham, United Kingdom, for
sodium methoxide solution (0.5 M in methanol), solid sodium
2.3.1. Mono-functional and di-functional end-cappers (PEG–VS and
VS–PEG–VS)
The reaction was performed using variable molar ratios be-
tween monofunctional or difunctional PEG, the base (NaH) and
Michael-type acceptor (DVS). Optimized conditions for the synthe-
sis of VS–PEG–VS were as follows: 2 g (2 mmoles of OH groups) of
Scheme 2. Structures of fluorophores (dansyl derivatives) and quenchers (dabsyl derivatives) and sketch of the micellar assemblies: prior to the FRET experiments
fluorophores and quenchers are present in different micellar assemblies, maximizing the fluorescence of the system; at time = 1 both macroamphiphiles and payloads will
equilibrate, reducing the average distance between fluorphores and quenchers and thus producing a substantial reduction to the system fluorescence.

306
P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
PEG 2000 diol were dissolved in 30 mL toluene under dry nitrogen
atmosphere and dried by azeotropic distillation, in a Dean–Stark
apparatus, corresponding to the removal of about 10 mL of toluene.
9.6 mg of NaH (0.4 mm, corresponding to a 1:0.2 OH/NaH molar ra-
tio) were then introduced into the reactor. When hydrogen evolu-
tion ceased, the partially deprotonated PEG solution was syphoned
into another degassed reactor containing 2.96 g of DVS (25 mmo-
les, corresponding to 50 mmoles of double bonds, i.e. to 1:0.2:50
OH/NaH/double bond molar ratio) previously dissolved in 30 mL
toluene, leaving the reaction under stirring for 24 h. The resulting
mixture was filtered to remove any formed salt, neutralized adding
a drop of acetic acid, reduced to small volume by rotary evapora-
tion and finally precipitated in diethyl ester. Yield: 73% (weight
of recovered polymer/theoretical amount of polymer). Conversion:
100% (molar percentage of reacted OH groups).
1144, 1100 (mas CAOAC), 961, 839 cmꢁ1 (in italics and underlined
the absorptions characteristic of PEG, in bold those characteristic
of PPS).
1H NMR (CDCl3): PPS–PEG: d = 1.35–1.45 (d, CH3 in PPS chain),
2.55–2.75 (m, 1 diastereotopic H of CH2 in PPS chain), 2.85–3.05
(m, CH and 1 diastereotopic H of CH2 in PPS chain), 3.40 (s, 3H,
AOCH3), 3.93 (t, 2H, AOACH2ACH2ASO2A), 3.6–3.8 (broad, PEG
chain protons), 7.28–7.34 (m, 5H, ACH2APh) ppm. PPS–PEG–VS:
d = 1.35–1.45 (d, CH3 in PPS chain), 2.55–2.75 (m, 1 diastereotopic
H of CH2 in PPS chain), 2.85–3.05 (m, CH and 1 diastereotopic H of
CH2 in PPS chain), 3.93 (t, 4H, AOACH2ACH2ASO2A), 3.6–3.8
(broad, PEG chain protons),), 6.10 (d, 1 H, CH2 = CHASO2A), 6.24
(d, 1H, CH2@CHASO2A), 6.81 (d, 1 H, CH2@CHASO2A), 7.06–7.14
(m, 5H, ACH2APh) ppm.
FT-IR (film on ATR crystal): PEG–VS: 3059 (mas @CH2, very
weak), shoulder at 2947 (mas CH2), 2855 (ms CH2), 1639 (ms C@C),
1451 (ds CH2), 1348 (mas SO2), 1292, 1247, 1094 (mas CAOAC),
946, 850 cmꢁ1. VS–PEG–VS: 3059 (mas @CH2, very weak), shoulder
2.4. Synthesis of PPS–PEG–dansyl and PPS–PEG–dabsyl
150 mg of PPS10PEG44–VS (0.02 mmol of vinyl sulfone groups)
and 39 mg of cysteamine. (0.2 mmol) were dissolved in 30 mL of
previously degassed THF under dry nitrogen atmosphere (1:10
VS-to-thiol molar ratio). 1 mL of THF containing 5 eq.s of DBU
(15.2 mg, 0.1 mmol) was then introduced dropwise into the reac-
tor, the mixture was allowed to react for 2 h at room temperature.
The THF was then evaporated and the resulting viscous syrup was
precipitated in ice cold diethyl ether twice. The product was re-dis-
solved with 20 mL of THF and introduced into a reactor under dry
nitrogen atmosphere; 2 mL of THF containing five equivalents of
dansyl chloride (27 mg/0.1 mmol) or dabsyl chloride (32 mg/
0.1 mmol) and triethylamine (10.1 mg/0.1 mmol) was added to
the reactor. The reaction was allowed to react for 24 h with protec-
tion from light. Then the solvent was removed at the rotary evap-
orator and precipitated in ice cold diethyl ether twice. Yield: 57%
for PPS–PEG–DA, 42% for PPS–PEG–DB. Conversion: 90% for PPS–
PEG–DA, 86% for PPS–PEG–DB.
at 2952 (
m
as CH2), 2856 (
as SO2, roughly twice as intense as for PEG–VS), 1277, 1242, 1104
as CAOAC), 961, 839 cmꢁ1 (in bold the absorptions characteristic
ms CH2), 1634 (ms C@C), 1450 (ds CH2), 1343
(m
m
(
of PEG).
1H NMR (CDCl3): PEG–VS: d = 3.4 (s, 3H, AOCH3), 3.6–3.8
(broad, PEG chain protons), 6.06 (d, 1 H, CH2@CHASO2A, cis to sul-
fone group), 6.22 (d, 1H, CH2@CHASO2A, trans to sulfone group),
6.84 (d, 1 H, CH2@CHASO2A) ppm. VS–PEG–VS: d = 3.6–3.8 (broad,
PEG chain protons), 6.11 (d, 2 H, CH2@CHASO2A), 6.24 (d, 2H,
CH2@CHASO2A), 6.75 (d, 2 H, CH2@CHASO2A) ppm.
2.3.2. PPS–PEG and PPS–PEG–VS block copolymers
A literature procedure based on the use of a reducing agent
(TBP) during the polymerization and of a buffer containing non-
nucleophilic base in the end-capping step was adopted [33]. In a
typical experiment, the polymerization environment (parallel reac-
tor FirstMate from Argonaut Technologies) was purged with nitro-
gen for 5 min before polymerization and 5 mL of previously
degassed THF were introduced in each reactor. one milli litre of a
previously degassed THF solution of S-benzyl thioacetate (contain-
ing 33.2 mg/0.2 mmol of compounds) and 1 mL of a TBP stock solu-
tion (corresponding to a 5-fold TBP: thioacetate molar ratio) were
introduced in the reactor. Separately, a stock solution of sodium
methoxide was prepared by mixing 380 mg of 0.5 M sodium met-
hanoate solution in methanol (0.42 mL, 0.21 mmol, corresponding
to 1.05 eq.s) with 1 mL of previously degassed THF; the solution
was then added via a syringe, and the mixture was stirred at room
temperature and allowed to react for 5 min. A variable quantity of
PS (corresponding to 10, 20, 30 or 40 equivalents compared to thio-
acetate groups) was then introduced into the reactor and allowed
to react for 45 min. 2 eq.s of acetic acid and 1 eq. of DBU were
added to neutralize the pH. 5 mL of a THF solution containing an
excess of end-capping agent (1.5 eq.s of PEG–VS for the synthesis
of PPS–PEG or 10 equivalents of VS–PEG–VS polymer for the syn-
thesis of PPS–PEG–VS) were finally added, and the mixture was
stirred to react for another 2 h at room temperature.
The same procedure was adopted to synthesize PEG–DB start-
ing from PEG–VS. Yield: 66%, Conversion: 92%.
FT-IR (film on ATR crystal): PPS–PEG–DA: 3456 (
ms CH3), 2881 (mas CH2), 2855 (mas CH3 and ms CH2), 1613, 1572 (ms
C = C), 1470 (ds CH2), 1358 ( as SO2), 1343, 1277, 1242, 1144, 1104
mas CAOAC), 961, 839 (ms CAOAC), 793, 737 (
@CH) cmꢁ1. PPS–
mas NH), 2952
(
m
(
x
PEG–DB: 3431 (mas NH), 2957 (ms CH3), 2886 (mas CH2), 2855 (mas
CH3 and ms CH2), 1603, 1522 (ms C@C), 1450 (ds CH2), 1419 (ms
N@N), 1358 (mas SO2), 1343, 1277, 1236, 1134, 1104 (mas CAOAC),
961, 839 (ms C-O-C), 824, 686 (
x .
@CH) cmꢁ1
1H NMR (CDCl3): PPS–PEG–DA: d = 1.35–1.45 (d, CH3 in PPS
chain), 2.0 (s, 1H, ACH2ANHASO2A), 2.55–2.75 (m, 1 diastereotop-
ic H of CH2 in PPS chain), 2.85–3.05 (m, CH and 1 diastereotopic H
of CH2 in PPS chain), 3.20 (s, 6H, NA(CH3)2), 3.43 (t, 2H,
ACH2ACH2ANHA), 3.78 (t, 4H, ASO2ACH2ACH2ANHA), 3.93 (t,
4H, AOACH2ACH2ASO2A), 3.6–3.8 (broad, PEG chain protons),
7.18 (d, 1H, aromatic CH ortho to aniline), 7.28–7.34 (m, 5H,
ACH2APh), 7.45–7.55 (dt, 2H, aromatic CH meta to aniline and
meta to sulfone group), 8.22–8.38 (dd, 2H, aromatic CH para to ani-
line and para to sulfone group), 8.56 (d, 1H, aromatic CH ortho to
sulfone group) ppm.
The solvent was removed at the rotary evaporator, and the
resulting viscous liquid was precipitated with ice cold diethyl
ether. The oil was re-dissolved in 1 mL THF and precipitated,
repeating the procedure again before transferring to water and
purified through ultrafiltration using membranes with
MWCO = 30,000 Da. Average yields after freeze drying: 62–88 wt.%
FT-IR (film on ATR crystal): PPS–PEG: 2956 (ms CH3), 2890 (mas
PPS–PEG–DB: d = 1.35–1.45 (d, CH3 in PPS chain), 2.0 (s, 1H,
ACH2ANHASO2A), 2.55–2.75 (m, 1 diastereotopic H of CH2 in
PPS chain), 2.85–3.05 (m, CH and 1 diastereotopic H of CH2 in
PPS chain), 3.15 (s, 6H, NA(CH3)2), 3.43 (t, 2H, ACH2ACH2ANHA),
3.78 (t, 4H, ASO2ACH2ACH2ANHA), 3.93 (t, 4H, AOACH2A
CH2ASO2A), 3.6–3.8 (broad, PEG chain protons), 6.78 (d, 2H,
aromatic CH ortho to aniline), 7.28–7.34 (m, 5H, ACH2APh),
7.85–7.95 (m, 4H, aromatic CH ortho to azo group), 8.05 (d, 2H,
aromatic CH ortho to sulfone group) ppm.
CH2), 2862 (
1280, 1240, 1145, 1100 (
3069 ( as @CH2), 2957 (
CH2), 1649 (ms C@C), 1456 (ds CH2), 1343 (mas SO2), 1282, 1242,
mas CH3 and
m
m
m
s CH2), 1456 (ds CH2), 1348 (
as CAOAC), 959, 843 cmꢁ1. PPS–PEG–VS:
s CH3), 2881 ( as CH2), 2861 ( as CH3 and ms
mas SO2), 1344,
m
m
m
mPEG–DB: d = 2.0 (s, 1H, ACH2ANHASO2A), 3.15 (s, 6H,
Nꢁ(CH3)2), 3.4 (s, 3H, AOCH3), 3.53 (t, 2H, ACH2ACH2ANHA),

P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
307
3.78
(t,
2H,
ASO2ACH2ACH2ANHA),
3.93
(t,
2H,
allowed to evaporate to form a thin film at the bottom of the vial.
1 mL of PPS10–PEG44 micelle dispersions with different concentra-
tions ranging from 0.0001 to 10 mg/mL were added to the vial to
have a final 6.7 ꢂ 10ꢁ7 M pyrene concentration. The solutions were
kept on a shaker at room temperature for 24 h to reach equilibrium
prior to fluorescence measurement. Fluorescence spectra were re-
corded using a PerkinElmer LS55 luminescence spectrophotometer
spectrometer at room temperature. Excitation wavelength:
335 nm; emission recorded at 373 nm (I1) and 393 nm (I3).
AOACH2ACH2ASO2A), 3.6–3.8 (broad, PEG chain protons), 6.78
(d, 2H, aromatic CH ortho to aniline), 7.28–7.34 (m, 5H, ACH2APh),
7.85–7.95 (m, 4H, aromatic CH ortho to azo group), 8.05 (d, 2H,
aromatic CH ortho to sulfone group) ppm.
2.5. Synthesis of dansyl hexylamide (DA) and dabsyl hexylamide (DB)
Under an inert atmosphere, 20 mL of dichloromethane contain-
ing 54 mg/0.2 mmol dansyl chloride or 64 mg/0.2 mmol dabsyl
chloride was introduced into a parallel reactor. Then 1 mL of
dichloromethane containing five equivalents of hexylamine
(40 mg/1 mmol) and five equivalents triethylamine (40 mg/
1 mmol) were added dropwise to the reactor. After leaving the
reaction stirring for 2 h at room temperature, the resulting mixture
was extracted with 20 ml of water three times to remove the salts
generated by the reaction. After extraction, the organic solution
was collected and concentrated by rotary evaporation and fol-
lowed by further purification using silicon gel chromatography.
The purified products were collected and dried, followed by verifi-
cation by TLC (dichloromethane) and characterization by 1H NMR.
The yields of the products were 51% for dansyl hexylamide and 64%
for dabsyl hexylamide.
2.9. Drug loading
2 mg of DA-hexylamide or of DB-hexylamide were solubilized
in 1 mL of dichloromethane. PPS–PEG micellar dispersions were
prepared through direct dispersion in water as described above.
0.1 mL of the dichloromethane solutions were taken and mixed
with 1 mL of the 1.0 mg/mL PPS–PEG micelle dispersion, evaporat-
ing the organic at the rotary evaporator. Any non-encapsulated
payload was removed by centrifugation (solids) and dialysis (pos-
sibly over-saturated payload; MWCO = 3000 g/mol). 0.1 mL of DMF
was then added to 0.1 mL of drug loaded micellar dispersion in or-
der to disrupt the micellar aggregates and solubilize the hydropho-
bic compounds. The concentration of hydrophobe was then
measured via its UV–Vis absorbance (DA: 340 nm, DB: 450 nm;
PPS–PEG micelles have an absorbance maximum at 260 nm, with
negligible absorbance at 340 or 450 nm). The concentrations of
samples were calculated by comparing the readings to the stan-
dard curves of DA and DB. Hydrophobe (drug) loading (DL) and
encapsulation efficiency (EE) were calculated as follows: DL (w/
w) = (amount of loaded drug)/(amount of polymer), EE(wt.%) = (a-
mount of loaded drugs)/(amount of added drugs) ꢂ 100.
FT-IR (film on ATR crystal): DA Hexylamide: 3298 (mas NH),
2952 (ms CH3), 2932 (mas CH2), 2861 (mas CH3 and ms CH2), 1613,
1572 (ms C@C), 1450, 1313 (ds CH2), 1353 (mas SO2), 1160 (ms SO2),
1139, 1068 (ms CAC), 961, 839 (ms CAOAC), 788, 732 (x .
@CH) cmꢁ1
DB Hexylamide: 3293 (mas NH), 2942 (ms CH3), 2927 (mas CH2),
2855 (mas CH3 and ms CH2), 1608, 1522 (ms C = C), 1450, 1323 (ds
CH2), 1419 (ms N@N), 1374 (mas SO2), 1160 (ms SO2), 1139, 1089
(
ms CAC), 819, 686 (
x .
@CH) cmꢁ1
1H NMR (CDCl3): DA hexylamide: d = 0.95 (t, 3H, ACH2ACH3),
1.15 (m, 6H, ACH2A(CH2)3ACH3), 1.25 (m, 2H, ACH2A(CH2)3A
CH3), 2.0 (s, 1H, ACH2ANHASO2A), 3.00 (m, 2H, ANHACH2A
CH2A), 3.15 (s, 6H, NA(CH3)2), 7.18 (d, 1H, aromatic CH ortho to
aniline), 7.45–7.55 (dt, 2H, aromatic CH meta to aniline and meta
to sulfone group), 8.22–8.38 (dd, 2H, aromatic CH para to aniline
and para to sulfone group), 8.56 (d, 1H, aromatic CH ortho to
sulfone group) ppm.
DB hexylamide: d = 0.95 (t, 3H, ACH2ACH3), 1.15 (m, 6H,
ACH2A(CH2)3ACH3), 1.25 (m, 2H, ACH2A(CH2)3ACH3), 2.0 (s, 1H,
ACH2ANHASO2A), 3.00 (m, 2H, ANHACH2ACH2A), 3.15 (s, 6H,
NA(CH3)2), 6.78 (d, 2H, aromatic CH ortho to aniline), 7.85–7.95
(m, 4H, aromatic CH ortho to azo group), 8.05 (d, 2H, aromatic
CH ortho to sulfone group) ppm.
2.10. FRET experiments
2.10.1. Exchange of macroamphiphiles
In a typical experiment (case C: DA micelles + DB micelles.
Preparation from water), 18 mg of PPS10–PEG44 and 2 mg of
PPS10–PEG44–DA were dissolved in 2 mL of THF; after evaporation
of the solvent at the rotary evaporator 20 mL of deionised water
were added and the polymer was stirred at room temperature
for 2 h. The total polymer concentration of the resulting dispersion
was 1 mg/mL (3.1 ꢂ 10ꢁ4 M), where the concentration of PPS10
–
PEG44–DA was 3.1 ꢂ 10ꢁ5 M (10% of overall polymer molar concen-
tration). An identical procedure was adopted to prepare 9:1 PPS10
–
PEG44/PPS10–PEG44–DB dispersions in water at DB concentrations
ranging between 3.1 ꢂ 10ꢁ6 and 2.5 ꢂ 10ꢁ4 M. 1 mL of the PPS–
PEG–DB was then added to 1 mL of the PPS–PEG–DA dispersion
and aliquots were transferred into a 96 well plate measuring the
fluorescence emission of dansyl groups (excitation wavelength:
360 nm, emission wavelength: 528 nm) at 0, 0.5, 1, 2, 4, 16, 24 h.
All measurements were conducted at 25 °C and repeated three
times. In ‘‘case B’’ experiments, 1 ꢂ 10ꢁ6 to 2.5 ꢂ 10ꢁ4 M solutions
of PEG–DB in water were used instead of micellar suspensions of
PPS10–PEG44/PPS10–PEG44–DB. In ‘‘case A’’ experiments ((DA + DB)
micelles), PPS10–PEG44 was dissolved in THF with PPS10–PEG44–DA
and PPS10–PEG44–DB always in 9:1 non-labelled/labeled macroam-
phiphile ratio. Identical procedures were followed for the experi-
ments on micellar suspensions prepared by diluting THF
solutions in water.
2.7. Preparation and use of PPS10–PEG44 micelles
2.7.1. Preparation from water
10 mL of deionized water were added to 10 mg (3.5 lmol) of
PPS10–PEG44 at room temperature and the dispersion was stirred
for 2 h.
2.7.2. From organic solvents
10 mg of PPS10–PEG44 were dissolved in 1 mL of organic solvent
(THF or DCM). This solution was added dropwise to 10 mL of deion-
ised water under stirring and stirred at room temperature for 2 h.
The organic solvent was removed by rotary evaporation (80 mbar,
25 °C, 30 min), adding water to bring the total volume to 10 mL.
2.8. CMC measurement
2.10.2. Exchange of payloads
20 mL of DA- and DB-loaded PPS10–PEG44 dispersions (1.0 mg/
mL polymer, 0.042 mg/ml DA hexylamide, 0.034 mg/mL DB hex-
ylamide) were prepared as described above. 2 mL of both micellar
dispersions were conditioned at the target temperature (4, 25 or
6.8 mg of pyrene were dissolved in 1 mL of THF and were fur-
ther diluted to 6.7 ꢂ 10ꢁ3 mg/mL. Then a certain amount of this di-
luted stock solution was added to a glass vial and the solvent was

308
P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
37 °C), then rapidly mixed and stored in an incubator, being sam-
pled at 0, 0.1, 0.5, 1, 2, 4, 6 h.
3. Results and discussion
3.1. Synthetic procedures
In this study, PPS–PEG amphiphilic block copolymers were pre-
pared using a procedure based on vinyl sulfone-terminated PEG re-
agents. PEG vinyl sulfones have been originally proposed by
Veronese as selective reagents for protein PEGylation and were ob-
tained via reaction of PEG terminal OH groups using chloroethyl-
sulfones as intermediates [64]. In more recent studies these
groups were introduced by reacting a large excess of divinyl sul-
fone with an appropriate nucleophile, e.g. an in situ produced
PEG alcoholate [65]. Such vinyl sulfone-terminated PEGs have been
used to produce gels by reacting with multifunctional thiols
[66,67] or to decorate nanoaggregates [68].
Here we have used this procedure to prepare mono- and later
bis-vinyl sulfone-terminated PEG derivatives, optimizing the syn-
thetic conditions reported in literature: PEG terminal OH groups
were deprotonated with a stoichiometric defect (0.2 eq.s) of so-
dium hydride and the resulting alcoholates were reacted with a
large excess of DVS (Scheme 3).
Fig. 1. GPC traces of the products of Michael-type addition of PEG diol (MW 2000 g/
mol) on DVS obtained with a constant OH/NaH ratio and variable amounts of DVS.
In most experiments DVS was added to a solution of deprotonated PEG; also at low
OH/DVS molar ratios, this procedure provided significant amounts of chain
extended products, e.g. about 20% of the total polymeric material for both 1:40
and 1:50 OH/DVS ratios (calculation based on the integrals of the GPC peaks). In the
experiment identified as 1:0.2:50ꢃ the PEG solution was slowly dropped into a DVS
solution, allowing to reduce the amount of chain extended material: the amount of
dimer corresponded to less than 6% of the total polymeric material. The trace of the
original PEG diol (dashed line) is reported for comparison.
It is worth pointing out that it is necessary to use an almost cat-
alytic amount of base to avoid parasite reactions: the Michael-type
addition product is a carboanion, which can be quenched via pro-
ton transfer from protic groups (e.g. non-deprotonated alcohols),
regenerating alcoholates that can carry on additional Michael-type
additions. In the absence of protic groups, i.e. when stoichiometric
quantities of NaH were used, viscosity sharply increased and in
some cases even gelation was recorded; this is possibly due to mul-
tiple Michael-type additions on the same PEG residue and branch-
ing or cross-linking of the resulting multifunctional reagents
(Scheme 3, bottom right).
to avoid chain extension reactions due to double addition of alco-
holates on the same sulphur centre (Fig. 1).
The thiol-reactive PEG derivatives were then used as end-cap-
pers to functionalize poly(propylene sulfide) chains in a one pot
procedure comprising initiation, propagation of episulfide poly-
merization and end-capping of the thiol-terminated polymers; a
recent review provides ample detail of the mechanism of episulfide
polymerization. The optimized literature procedure is based on the
use of an in situ deprotected initiator (in this case benzyl thioace-
tate), the presence of a reducing agent (tributyl phosphine) during
polymerization and buffered conditions (DBU/acetic acetic acid)
For the synthesis of bis-vinyl sulfone-terminated PEG it is also
necessary to employ a large stoichiometric excess of DVS, in order
Scheme 3. Functionalization of monomethoxy PEG and PEG diol with terminal vinyl sulfone groups. NaH was employed in stoichiometric defect to the PEG OH groups
(typically 20% in moles), since the Michael-type addition product can on its turn deprotonate OH groups and re-initiate the reaction. When it was used in excess, gelation
occurred, possibly because of formation of multifunctional derivatives.

P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
309
for the end-capping, to avoid disulfide formation and side-reac-
tions on the vinyl sulfone groups due to the use of excess base [63].
All polymers were characterized by quantitative end-capping
and narrow molecular weight distribution. (Fig. 2 and Table 1).
Finally, the labeled polymers were prepared by reacting vinyl
sulfone-terminated macromolecules first with cysteamine and
then with the amine-reactive dansyl or dabsyl chloride, as depicted
in Scheme 2.
been demonstrated to form also larger aggregates, such as worm-
like micelles and vesicles [40]. It is worth mentioning that the
above diameters refer to micelles prepared by stirring PPS–PEG
polymers in deionized water at room or higher temperature; dif-
ferently from the results reported by Velluto et al., the dispersion
of dichloromethane or THF solutions in water followed by the
evaporation of the solvent provided considerably larger aggregates
(Fig. 3B), which are possibly formed by clustered/entangled
micelles.
The Critical Aggregation Concentrations showed a clear depen-
dence on the length of PPS block (Fig. 3C and D), which in our case
are 2–5 smaller than in the work of Velluto et al. (Table 1, last col-
umn). Since the same analytical method was used, these discrep-
ancies are likely to be ascribed to the small structural differences
of the polymers (PPS terminus pyridine vs. benzyl group; PEG–
PPS junction ether vs. sulfone).
We have then used the two model hydrophobic compounds
prepared in this study (dansyl hexyl hexylamide and dabsyl hex-
ylamide) to evaluate the loading capacity of the PPS–PEG micelles
as a function of the PPS length (Table 2). For both compounds the
loading and the encapsulation efficiency increased with increasing
PPS length for degrees of polymerization 10 to 30. The performance
of PPS40–PEG44 is apparently poorer, but we are inclined to con-
sider this to be an artefact caused by the partial sedimentation of
PPS40–PEG44 large aggregates, when the dispersions were centri-
fuged to remove any non-encapsulated drug in crystalline form.
3.2. Preparation and drug loading of non-functional PPS–PEG micelles
In a recent paper by Velluto et al. [40], PPS–PEG micelles have
been produced and loaded with cyclosporine A using polymers
(PPS10–PEG44, PPS20–PEG44 and PPS30–PEG44) with only very minor
differences from those prepared in this study: those structures had
no sulfone linking PPS and PEG blocks and a thiopyridine PPS ter-
minal group where we have used a benzylic group; in addition, the
polymerization conditions adopted in that reference are more
prone to the formation of PPS homopolymer due to disulfide-med-
iated chain transfer [63,69].
Since these differences do not seriously affect the hydrophilic/
lipophilic balance of the polymers, the aggregates of the polymers
produced in this study show a substantially identical size to those
previously reported, with a Z-average size of 20–23 nm for PPS10
–
PEG44 and PPS20–PEG44, 24–27 nm for PPS30–PEG44, and a bimodal
and broad distribution for PPS40–PEG44, (Fig. 3A) which has already
Fig. 2. Left: a benzyl mercaptane anion was generated in situ from the corresponding thioacetate and used to initiate the ring-opening polymerization of propylene sulfide.
The resulting thiolate-terminated macromolecules were then end-capped with PEG–VS (used stoichiometrically) or VS–PEG–VS (used in large excess). Right: The sequence of
reactions always provided polymers with monomodal MW distribution and low polydispersity.
Table 1
Summary of the characterization data for PEG and PPS–PEG derivatives.
End-capping conversion (% mol)a
Yield (wt.%)b
Theor. DP of each PPS armc
CACf (mg/mL)
d
NMR Mn
GPCe Mn
Mw=Mn
Sample
PEG44–VS
VS–PEG44–VS
PPS10PEG44
PPS20PEG44
PPS30PEG44
100
100
95
93
94
93
94
90
86
78
73
88
85
78
81
62
57
42
n.a.
n.a.
10
20
30
40
10
10
10
2100
2200
2900
3600
4600
5400
3000
3200
3300
2400
2600
3400
4200
5300
6100
3300
3400
3600
1.05
1.08
1.15
1.16
1.08
1.14
1.11
1.15
1.17
n.a.
n.a.
0.185 (0.392[40])
0.130 (0.350 [40])
0.031
0.011 (0.052 [40])
=
=
=
PPS40PEG44
PPS10PEG44–VS
PPS10PEG44–DA
PPS10PEG44–DB
a
Calculated from the ratio of the 1H NMR resonance of PEG terminal groups (CH3 at 3.4 ppm or vinyl sulfone at 6.1–6.2 ppm) and the resonance of PEG main chain for the
PEG–VS and VS–PEG–VS, or that of the aromatic group of the initiator (7.4 ppm) for the other polymers.
b
Weight of recovered polymer/weight of monomer + deprotected initiator + stoichiometric amount of PEG derivative.
c
Theoretical degree of polymerization (DP), expressed as the ratio [PS]/[thiol].
d
From the ratio of the 1H NMR resonance of the PPS CH3 group (1.4 ppm) of the aromatic group of the initiator (7.4 ppm).
e
Calculated from GPC data in THF by the means of the universal calibration with poly(styrene) standards and viscosimetric and RI detection.
In brackets data from Ref. [40]. The CACs for the functional derivatives of PPS10PEG44 were assumed identical to those of PPS10PEG44
f
.

310
P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
Fig. 3. (A) Size distributions of micelles of PPS10–PEG44 to PPS40–PEG44. The dispersions were prepared at room temperature directly in water. (B) Influence of the preparation
method on the size of the micelles; the presence of organic solvents shifted the size distribution to larger values. (C) I1/I3 Emission ratio for pyrene as a function of the
polymer concentration; the lines represent sigmoidal (Hill) fits of the experimental points; the CACs were determined as the end point of the sigmoids as indicated by dashed
lines in the figure. (D) Dependence of the CAC on the degree of PS degree of polymerization; although the CAC values have been reported with a logarithmic scale, there are
too few data points to highlight a logarithmic relation between the two variables.
Table 2
Loading and encapsulation efficiency for the two hydrophobic reporters (drugs) in PPS–PEG micellar dispersionsa.
Polymer
DA
DB
Drug loadingb (mg/mg)
Encapsulation efficiencyc (wt.%)
Drug loadingb (mg/mg)
Encapsulation efficiencyc (wt.%)
PPS10PEG44
PPS20PEG44
PPS30PEG44
PPS40PEG44
0.0422 0.008
0.0637 0.007
0.0931 0.009
0.0799 0.006
21.1 4.0
31.9 3.6
46.6 4.6
39.9 3.2
0.0344 0.002
0.0449 0.007
0.0694 0.008
0.058 0.006
17.2 1.1
22.5 3.6
34.7 3.8
29.4 3.1
a
b
c
Polymer concentration: 1 mg/mL; 1:5 drug/polymer weight ratio.
Amount of drug loaded/amount of polymer.
Amount of drug loaded/total amount of drug used * 100.
For further studies we have focused our attention on PPS10
–
sible situations (Scheme 4), employing PPS10–PEG44 at a concentra-
tion at least one order of magnitude higher than its CMC:
PEG44: its high hydrophilic content makes it the polymer with pos-
sibly the quickest inter-micellar exchange kinetics and therefore
the ‘‘worst case’’ scenario candidate. However, although character-
ized by the smallest PPS content, PPS10–PEG44 still showed a well-
defined molecular structure (low polydispersity) and micellar
organization (no large aggregates, rather low CAC and therefore
good stability against dilution) and a reasonable drug loading.
Finally, it is worth noting that the size of the micelles formed by
vinyl sulfone-terminated PPS10–PEG44 or its labeled derivatives,
alone or in mixture with non-functional PPS10–PEG44, was indistin-
guishable from that of PPS10–PEG44-only micelles (Fig. 4). This en-
sures that the presence of the chromophores did not significantly
alter the micellar self-assembly of PPS–PEG.
Case A: (DA + DB) micelles, i.e. micelles prepared from pre-
mixed DA- and DB-labelled PPS–PEG polymers (intra-micellar
quenching). This situation forces the closest proximity between
the chromophores while hydrophobic aggregation boosts the local
concentration of both, therefore it is expected quenching (ex-
pressed as FRET efficiency, e) to be maximal. We also expect it to
increase with increasing DB/DA ratio and not to depend on time,
since any exchange between micelles will not alter the average
DB/DA ratio.
Case B: DA micelles + PEG–DB, i.e. micelles containing PPS–PEG–
DA exposed to a hydrophilic polymeric quencher, PEG–DB (interfa-
cial quenching). Since PEG–DB lacks of a hydrophobic anchor, (a)
FRET to DB could only take place on the micellar surfaces; (b) the
DB concentration in proximity to DA would be considerably smal-
ler than in case A. As a result, the quenching efficiency should be
strongly limited; however, as above, quenching should signifi-
cantly depend on the PEG–DB concentration and, since exchange
phenomena are impossible, also in this case we expect to observe
no time dependence.
3.4. FRET experiments
3.4.1. Exchange between macroamphiphiles
Spatial proximity is the major determinant of the quenching
efficiency of dansyl emission. Here we have considered three pos-

P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
311
time, as in cases A and B. Quenching is likely to be higher than in
case B, because DB is present in micellar aggregates and therefore
its local concentration is higher than its macroscopic one. How-
ever, FRET should be less intense than in case A, due to the less inti-
mate contact between fluorophore and quencher. (2) If exchange
occurs, quenching efficiency should increase with time. The depen-
dence of quenching on DB concentration may show a similarity to
case B at shorter times (interfacial quenching) while becoming
similar to case A at longer ones (intra-micellar quencher).
We have investigated PPS–PEG micelles produced both by di-
rectly dispersing the polymers in water and by diluting in water
a THF solution, i.e. in the two cases of smaller and larger micellar
aggregates. In all cases the micelles were composed of unlabelled
macroamphiphiles in a 9-fold molar excess of labelled PPS–PEG,
in order to avoid self-quenching effects of the fluorophores.
Fig. 5 reports The FRET efficiency as a function of the quencher
concentration and time for the three cases reported above for both
aggregates produced from THF solutions (size 50–110 nm; Fig. 5
top row) and those obtained through direct dispersion of the poly-
mers in water (size 20–30 nm; Fig. 5 bottom row). It is immedi-
ately apparent that all systems exhibited a rather flat time
dependence, and an asymptotic dependence on DB concentration.
A closer look at the time dependence (1:1 DA/DB molar ratio;
Fig. 6, left graphs) confirms that in all cases FRET efficiency was
substantially constant for at least 24 h. A small increase in quench-
ing is recorded within the first hour after mixing; however, since
this is also recorded for pre-mixed macroamphiphiles, it cannot
be related to exchange phenomena. The above is therefore a first
indication of the absence of rapid inter-micellar exchange between
macroamphiphiles. At all time points the FRET efficiency was in the
order case A (pre-mixed polymers: (DA + DB) micelles) > case C
(DA micelles + DB micelles) > case B (DA micelles + PEG–DB), indi-
cating the absence of intra-micellar quenching in case C, which
otherwise would increase to the level of case A.
Fig. 4. The size the micellar aggregates is not influenced by the replacement of the
terminal methoxy group of PPS10–PEG44, with the vinyl sulfone of PPS10–PEG44–VS;
identical distributions are obtained for the polymers independently or in 1:1 M
mixture.
It is also noteworthy that the quenching shown by large aggre-
gates was significantly, although not dramatically higher than that
of smaller ones; this was more clear for case A ((DA + DB) micelles)
and case B (DA micelles + PEG–DB). The large aggregates are possi-
bly characterized by a less compact packing and therefore by a
quicker internal dynamics, which may allow more intimate contact
between fluorophores and quenchers.
The differences between the three cases and in particular the
intermediate behavior of case C were quantified using the depen-
dence of FRET efficiency on DB concentration at 0.5 h after mixing
(Fig. 6, right graphs). The data points were fitted with a double
exponential model e([DB]) = e1 ꢁ elow exp (ꢁ[DB]/Slow) ꢁ ehigh exp
( ꢁ [DB]/Shigh); Slow and Shigh B are in the essence the slopes of
the curves at low and high DB concentration, respectively. The
curves showed strong similarities at large DB concentration, with
Shigh ꢀ 1 L/mmol, indicating that at millimolar or higher concentra-
tions FRET happens similarly disregarding the physical organiza-
tion of the chromophores. The parameter Slow, on the other hand,
showed a clear trend increasing from about 0.1 L/mmol for case
B (interfacial quenching) to 0.03 L/mmol for case A (intra-micellar
quenching.
Scheme 4. FRET experiments aiming at assessing the exchange of PPS–PEG
macroamphiphiles between micelles were conducted on three different systems:
(A) micelles containing both FRET donor and acceptor, where any quenching would
be predominantly intra-micellar. (B) Donor (DA)-containing micelles exposed to a
solution of PEGylated acceptor, where any quenching would be predominantly
interfacial, since PEG–DB cannot get anchored in the micellar core. (C) Donor-
containing micelles and acceptor-containing ones mixed together, where the
quenching could be time-dependent and intra-micellar when exchange occurs
appreciably, or predominantly interfacial when it does not.
The intermediate behavior of case C in terms of FRET efficiency
and of its dependence on concentration, and the absence of a sound
time dependence within the first 24 h have therefore led us to con-
clude that no rapid amphiphile exchange occurs between PPS10
PEG44 micelles.
–
3.4.2. Exchange between payloads
Case C: DA micelles + DB micelles, i.e. micelles separately pre-
pared from PPS–PEG–DA and PPS–PEG–DB (inter- or intra-micellar
quenching). Two cases are given: (1) if no rapid exchange between
macroamphiphiles occurs, the FRET efficiency does not depend on
DA hexylamide and DB hexylamide were employed as model
hydrophobic payloads and separately loaded in micellar suspen-
sions of unlabelled PPS10–PEG44 (1 mg/mL polymer in both cases,
0.042 mg/mL DA, 0.034 mg/mL DB); the two suspensions were

312
P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
Fig. 5. FRET efficiency (=1 – normalized fluorescence) as a function of the molar concentration of DB groups in the three cases of DB- and DA-terminated PPS–PEG pre-mixed
and then dispersed in micelles (case A), of PPS–PEG–DA micelles exposed to PEG–DB (case B),and of PPS–PEG–DA and PPS–PEG–DB first dispersed in micelles and then mixed
together (case C). The DA concentration was kept constant to 3.1 ꢂ 10ꢁ5 M and the ratio between labeled and un labelled polymer was kept constant to 1:10. All data are
averages over three measurements. Top row: aggregates produced via dilution in water of a THF solution. Bottom row: aggregates produced via direct dispersion of the bulk
polymers in water.
Fig. 6. Left: FRET efficiency as a function of time for PPS–PEG–DA micelles with a DA concentration = 3.1 ꢂ 10ꢁ5 M (1 mg/mL PPS10–PEG44 concentration, 10% of it labeled
with DA) and prepared by diluting a concentrated THF solution (Top) or directly dispersing the polymer in water (Bottom) exposed to equimolar amounts of DB initially
present in the same micelles (white circles), in different micelles prepared in identical fashion (black squares) or in solution (white triangles). Right: Dependence of the FRET
efficiency (time = 0.5 h after mixing) on DB concentration for micelles prepared from a THF solution (Top) or via direct dispersion in water (Bottom), same concentration as
above. The solid lines represent the result of a double exponential fit; the values of the characteristic parameter A (i.e. the slope of the initial part of the curves) are reported.

P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
313
Fig. 7. Left: Fluorescence intensity of DA-loaded PPS–PEG micelles as a function of time and temperature upon exposure to DB-loaded PPS–PEG micelles (1:1 DB/DA molar
ratio). Micelles loaded with pre-mixed hydrophobes (same molar ratio) were used as controls. The solid lines represent the results of fittings using a model of biexponential
decay. Right: The relative amount of ‘‘instantaneous’’ quenching (F1/Ft=0) increases with temperature, possibly due to the increased DB hexylamide solubility in water; the
characteristic time of the slower quenching component (s2) decreases with temperature, due to accelerated inter-micellar diffusion.
then mixed together in 1:1 weight ratio, in order to obtain a
roughly 1:1 DA/DB molar ratio, while micelles loaded with the
two drugs (1:1 M ratio) were used as a control. DA fluorescence
was then monitored as a function of time and temperature
(Fig. 7), providing information on the kinetics of exchange of DB.
The fluorescence intensity exhibited a clear biexponential
behavior, which we have modeled as, F(t) = F1 ꢁ F1 exp ( ꢁ t/
no rapid macroamphiphile exchange to take place with either
mechanism.
FRET experiments between payloads have shown a quicker
dynamics (at room or body temperature) in comparison to the
macroamphiphiles, which is possibly to be ascribed to the higher
solubility in the water medium. However, it is also possible to sub-
stantially slow down the payload exchange at low temperature.
In the perspective of co-formulating differently functional and
differently loaded micellar carriers, our results suggest that such
preparations could be stored for relatively long periods (days) at
low temperature and may have a life time of a few hours at body
s
1) ꢁ F2 exp ( ꢁ t/
the controls (Fig. 7, left). The first phase is an almost instantaneous
quenching, whose characteristic time ( 1 < 3 min) does not appear
s2), where F is the fluorescence intensity of
1
s
to have a sound dependence on temperature, while the corre-
sponding extent of quenching (F1/Ft=0) increases with increasing
temperature. Since DB solubility will increase with temperature,
we are inclined to ascribe this phase to the action of DB in solution,
occurring immediately after mixing and without penetration in the
micellar cores. This is then followed by a slower phase, where the
temperature, requiring therefore
action.
a relatively quick targeting
References
characteristic time (
dence (Fig. 7, right).
s2) shows the expected temperature depen-
[1] A.V. Kabanov, E.V. Batrakova, N.S. Meliknubarov, N.A. Fedoseev, T.Y. Dorodnich,
V.Y. Alakhov, V.P. Chekhonin, I.R. Nazarova, V.A. Kabanov, J. Control. Release 22
(1992) 141–157.
The results obtained for DB hexylamide indicate a substantial
payload exchange in a few hours at 37 °C; however, this phenom-
enon is considerably retarded at 4 °C, and at this temperature the
payload exchange appears negligible for the first day(s) after prep-
aration. These data are payload-specific and, furthermore, they also
depend on the concentration of both micellar carrier and payload;
however, as a general indication it appears that storage at cold
temperature could allow for a reasonably long storage of a micellar
co-formulation with different payloads. On the other hand, the per-
manence at body temperature may be a critical point and the effi-
cacy of a micellar library will strongly rely on the efficiency of the
targeting groups.
[2] V.P. Torchilin, Cell. Mol. Life Sci. 61 (2004) 2549–2559.
[3] V.P. Torchilin, J. Control. Release 73 (2001) 137–172.
[4] G. Kwon, M. Naito, M. Yokoyama, T. Okano, Y. Sakurai, K. Kataoka, Langmuir 9
(1993) 945–949.
[5] M. Yokoyama, M. Miyauchi, N. Yamada, T. Okano, Y. Sakurai, K. Kataoka, S.
Inoue, Cancer Res. 50 (1990) 1693–1700.
[6] Y. Nagasaki, T. Okada, C. Scholz, M. Iijima, M. Kato, K. Kataoka, Macromolecules
31 (1998) 1473–1479.
[7] Y. Yamamoto, Y. Nagasaki, Y. Kato, Y. Sugiyama, K. Kataoka, J. Control. Release
77 (2001) 27–38.
[8] A. Harada, K. Kataoka, Macromolecules 28 (1995) 5294–5299.
[9] S. Katayose, K. Kataoka, Bioconjucate Chem. 8 (1997) 702–707.
[10] A.V. Kabanov, E.V. Batrakova, V.Y. Alakhov, J. Control. Release 82 (2002) 189–
212.
[11] D.A. Chiappetta, A. Sosnik, Eur. J. Pharm. Biopharm. 66 (2007) 303–317.
[12] E.V. Batrakova, S. Li, Y.L. Li, V.Y. Alakhov, W.F. Elmquist, A.V. Kabanov, J.
Control. Release 100 (2004) 389–397.
[13] J.M. Grindel, T. Jaworski, O. Piraner, R.M. Emanuele, M. Balasubramanian, J.
Pharm. Sci. 91 (2002) 1936–1947.
[14] D. Attwood, C. Booth, S.G. Yeates, C. Chaibundit, N. Ricardo, Int. J. Pharm. 345
(2007) 35–41.
4. Conclusions
This study has provided a semi-quantitative assessment of in-
ter-micellar exchange phenomena occurring in PPS–PEG disper-
[15] E.V. Batrakova, S. Li, S.V. Vinogradov, V.Y. Alakhov, D.W. Miller, A.V. Kabanov, J.
Pharmacol. Exp. Ther. 299 (2001) 483–493.
[16] E.V. Batrakova, S. Li, Y.L. Li, V.Y. Alakhov, A.V. Kabanov, Pharm. Res. 21 (2004)
2226–2233.
[17] T. Yamagata, H. Kusuhara, M. Morishita, K. Takayama, H. Benameur, Y.
Sugiyama, J. Control. Release 124 (2007) 1–5.
[18] E.V. Batrakova, D.W. Miller, S. Li, V.Y. Alakhov, A.V. Kabanov, W.F. Elmquist, J.
Pharmacol. Exp. Ther. 296 (2001) 551–557.
[19] E.R. Gillies, J.M.J. Frechet, Bioconjucate Chem. 16 (2005) 361–368.
[20] Y. Bae, N. Nishiyama, S. Fukushima, H. Koyama, M. Yasuhiro, K. Kataoka,
Bioconjucate Chem. 16 (2005) 122–130.
[21] H.S. Yoo, E.A. Lee, T.G. Park, J. Control. Release 82 (2002) 17–27.
[22] Y.Q. Tang, S.Y. Liu, S.P. Armes, N.C. Billingham, Biomacromolecules 4 (2003)
1636–1645.
[23] E.S. Lee, K. Na, Y.H. Bae, Nano Lett. 5 (2005) 325–329.
[24] V.P. Torchilin, A.N. Lukyanov, Z.G. Gao, Proc. Natl. Acad. Sci. USA 100 (2003)
6039–6044.
sions. We have selected
a
polymer with
a rather small
hydrophobic tail, PPS10–PEG44
.
Its predominantly hydrophilic
structure should provide the quickest exchange rate within an
Aniansson–Wall mechanism; the relatively high concentration
(1 mg/mL) used should also boost exchanges through micellar fu-
sion and fission. Our FRET experiments have used two control sys-
tems: as expected, significant differences were recorded between
the quenching efficiency in the three cases of fluoro-
phore + quencher in the same micelle (maximum quenching), in
different micelles (intermediate quenching) and fluorophore in mi-
celle + quencher in solution (lowest quenching). However, in no
case we have recorded a clear time dependence, which suggests
[25] X.B. Xiong, H. Uludag, A. Lavasanifar, Biomaterials 31 5886-5893.

314
P. Hu, N. Tirelli / Reactive & Functional Polymers 71 (2011) 303–314
[26] J. Emerit, A. Edeas, F. Bricaire, Biomed. Pharmacother. 58 (2004) 39–46.
[27] T.M. Paravicini, R.M. Touyz, Diabetes Care 31 (2008) S170–S180.
[28] F. Bonomini, S. Tengattini, A. Fabiano, R. Bianchi, R. Rezzani, Histol.
Histopathol. 23 (2008) 381–390.
[29] Y.S. Kim, M.J. Morgan, S. Choksi, Z.G. Liu, Mol. Cell 26 (2007) 675–687.
[30] L. Fialkow, Y.C. Wang, G.P. Downey, Free Rad. Bio. Med. 42 (2007) 153–164.
[31] A. Rehor, N. Tirelli, J.A. Hubbell, Macromolecules 35 (2002) 8688–8693.
[32] A. Napoli, N. Tirelli, G. Kilcher, J.A. Hubbell, Macromolecules 34 (2001) 8913–
8917.
[33] L. Wang, P. Hu, N. Tirelli, Polymer 50 (2009) 2863–2873.
[34] L. Wang, G. Kilcher, N. Tirelli, Macromol. Biosci. 7 (2007) 987–998.
[35] A. Napoli, M. Valentini, N. Tirelli, M. Muller, J.A. Hubbell, Nat. Mater. 3 (2004)
183–189.
[46] C. Honda, Y. Abe, T. Nose, Macromolecules 29 (1996) 6778–6785.
[47] E.E. Dormidontova, Macromolecules 32 (1999) 7630–7644.
[48] R. Pool, P.G. Bolhuis, Phys. Rev. Lett. 97 (2006) 4.
[49] A.G. Denkova, E. Mendes, M.O. Coppens, Soft Matter 6 (2010) 2351–2357.
[50] M.K. Johansson, R.M. Cook, Chem.-Eur. J. 9 (2003) 3466–3471.
[51] B.Y. Ren, F. Gao, Z. Tong, Y. Yan, Chem. Phys. Lett. 307 (1999) 55–61.
[52] V. Balzani, P. Ceroni, S. Gestermann, M. Gorka, C. Kauffmann, F. Vogtle,
Tetrahedron 58 (2002) 629–637.
[53] M. Asano, F.M. Winnik, T. Yamashita, K. Horie, Macromolecules 28 (1995)
5861–5866.
[54] M. Miki, T. Kouyama, Biochemistry 33 (1994) 10171–10177.
[55] S.R. Stauffer, J.F. Hartwig, J. Am. Chem. Soc. 125 (2003) 6977–6985.
[56] W. Veatch, L. Stryer, J. Mol. Biol. 113 (1977) 89–102.
[36] J.P. Bearinger, S. Terrettaz, R. Michel, N. Tirelli, H. Vogel, M. Textor, J.A. Hubbell,
Nat. Mater. 2 (2003) 259–264.
[37] E.M. Di Meo, A. Di Crescenzo, D. Velluto, C.P. O’Neil, D. Demurtas, J.A. Hubbell,
A. Fontana, Macromolecules 43 (2010) 3429–3437.
[38] M. Valentini, A. Napoli, N. Tirelli, J.A. Hubbell, Langmuir 19 (2003) 4852–4855.
[39] A. Napoli, M.J. Boerakker, N. Tirelli, R.J.M. Nolte, N. Sommerdijk, J.A. Hubbell,
Langmuir 20 (2004) 3487–3491.
[40] D. Velluto, D. Demurtas, J.A. Hubbell, Mol. Pharmacol. 5 (2008) 632–642.
[41] A. Halperin, S. Alexander, Macromolecules 22 (1989) 2403–2412.
[42] E.A.G. Aniansson, S.N. Wall, M. Almgren, H. Hoffmann, I. Kielmann, W. Ulbricht,
R. Zana, J. Lang, C. Tondre, J. Phys. Chem. 80 (1976) 905–922.
[43] R. Lund, L. Willner, D. Richter, E.E. Dormidontova, Macromolecules 39 (2006)
4566–4575.
[57] L.G. Reddy, L.R. Jones, D.D. Thomas, Biochemistry 38 (1999) 3954–3962.
[58] R.A. Melnyk, A.W. Partridge, C.M. Deber, J. Mol. Biol. 315 (2002) 63–72.
[59] J.K. Lin, J.Y. Chang, Anal. Chem. 47 (1975) 1634–1638.
[60] D. Parkinson, J.D. Redshaw, Anal. Biochem. 141 (1984) 121–126.
[61] C. Cannizzo, S. Amigoni-Gerbier, M. Frigoli, C. Larpent, J. Polym. Sci. Polym.
Chem. 46 (2008) 3375–3386.
[62] C. Sasaki, T. Hamada, H. Okumura, S. Maeda, J. Muranaka, A. Kuwae, K. Hanai,
K.K. Kunimoto, Polym. Bull. 57 (2006) 747–756.
[63] L. Wang, G. Kilcher, N. Tirelli, Macromol. Chem. Phys. 210 (2009) 447–456.
[64] M. Morpurgo, F.M. Veronese, D. Kachensky, J.M. Harris, Bioconjucate Chem. 7
(1996) 363–368.
[65] M.P. Lutolf, J.A. Hubbell, Biomacromolecules 4 (2003) 713–722.
[66] S.C. Rizzi, J.A. Hubbell, Biomacromolecules 6 (2005) 1226–1238.
[67] S.P. Zustiak, J.B. Leach, Biomacromolecules 11 (2010) 1348–1357.
[68] J.W. Bae, E. Lee, K.M. Park, K.D. Park, Macromolecules 42 (2009) 3437–3442.
[69] G. Kilcher, L. Wang, N. Tirelli, J. Polym. Sci. Polym. Chem. 46 (2008) 2233–2249.
[44] L. Willner, A. Poppe, J. Allgaier, M. Monkenbusch, D. Richter, Europhys. Lett. 55
(2001) 667–673.
[45] R. Lund, L. Willner, M. Monkenbusch, P. Panine, T. Narayanan, J. Colmenero, D.
Richter, Phys. Rev. Lett. 102 (2009) 4.