Self-assembled nanocages for hydrophilic guest molecules

Article abstract of DOI:10.1021/ja065462c

Reverse polymeric micelles are obtained following the association of polymeric amphiphiles in apolar media. To this date, reports of pharmaceutical applications for such micelles have been scarce, mainly because these systems have been studied in solvents that are not suitable for medical use. Here, alkylated star-shaped poly(glycerol methacrylate) polymers have been proposed in the design of oil-soluble reverse polymeric micelles. Micellar behavior was studied in various apolar solvents, including ethyl oleate, a pharmaceutically acceptable vehicle. The polymers were shown to assemble into spherical nanostructures (<40 nm) as determined by cryogenic transmission electron microscopy and atomic force microscopy studies. Interestingly, the reverse micelles were able to encapsulate various peptides/proteins (vasopressin, myoglobin, and albumin) in substantial amounts and facilitate their solubilization in oil. The nature of both the polymer used in micelle formation and the guest molecules was found to influence the ability of the micelle to interact with hydrophilic compounds.

Full text of DOI:10.1021/ja065462c

Published on Web 10/21/2006
Self-Assembled Nanocages for Hydrophilic Guest Molecules
Marie-Christine Jones,† Pankaj Tewari,† Claudia Blei,† Kelly Hales,‡
Darrin J. Pochan,‡ and Jean-Christophe Leroux*,†,§
Contribution from the Canada Research Chair in Drug DeliVery, Faculty of Pharmacy,
UniVersite´ de Montre´al, P.O. Box 6128, Downtown Station, H3C 3J7 Montreal, Quebec,
Canada, and Materials Science and Engineering and Delaware Biotechnology Institute,
UniVersity of Delaware, Newark, Delaware 19176
Received July 28, 2006; E-mail: jean-christophe.leroux@umontreal.ca
Abstract: Reverse polymeric micelles are obtained following the association of polymeric amphiphiles in
apolar media. To this date, reports of pharmaceutical applications for such micelles have been scarce,
mainly because these systems have been studied in solvents that are not suitable for medical use. Here,
alkylated star-shaped poly(glycerol methacrylate) polymers have been proposed in the design of oil-soluble
reverse polymeric micelles. Micellar behavior was studied in various apolar solvents, including ethyl oleate,
a pharmaceutically acceptable vehicle. The polymers were shown to assemble into spherical nanostructures
(<40 nm) as determined by cryogenic transmission electron microscopy and atomic force microscopy
studies. Interestingly, the reverse micelles were able to encapsulate various peptides/proteins (vasopressin,
myoglobin, and albumin) in substantial amounts and facilitate their solubilization in oil. The nature of both
the polymer used in micelle formation and the guest molecules was found to influence the ability of the
micelle to interact with hydrophilic compounds.
1. Introduction
micelles has mainly been studied in organic solvents such as
saturated hydrocarbons, and few examples of their use in
pharmaceutically relevant vehicles have been reported.10,11
Hydrophobically modified dendrimers and hyperbranched
polymers were recently proposed for the design of unimolecular
reverse polymeric micelles. Both types of polymers are char-
acterized by relatively compact structures with high densities
of surface functionalities.12 Compared to dendrimers, hyper-
branched polymers present the added benefit of being easier to
prepare.13 Reverse unimolecular micelles are typically generated
by the partial modification of surface hydroxyl or amine
functions with fatty acid derivatives. Such polymeric structures
have been shown to promote the extraction and solubilization
of various hydrophilic model compounds in nonpolar organic
solvents. For instance, Cooper et al.14 highlighted the use of
reverse polymeric micelles (RPMs) for improving the solubility
of a water-soluble dye in supercritical carbon dioxide. Later,
reverse micelles obtained from hyperbranched poly(glycerol)s15
or poly(ethylenimine)s16 were shown to extract the hydro-
philic dye Congo red from water into chloroform. The potential
The ability of micellar aggregates to act as solubility
enhancers stems from the dichotomy of their architecture.
Indeed, the segregation of the lyophobic segment of an
amphiphilic compound in the micelle core allows for the creation
of a favorable micro-environment for molecules with limited
affinity for the dispersing phase. Theoretically, reverse micelles
should be obtained from any amphiphilic molecules placed in
an organic, nonpolar solvent. However, for most low-molecular-
weight surfactants, the presence of water or another polar solvent
appears to be necessary for micelle formation.1-5 Consequently,
these systems consist of a polar phase stabilized by a molecular
surfactant layer in an arrangement reminiscent of a water-in-
oil microemulsion. Nevertheless, several reports have emerged
on potential applications for these reverse micelles, including
their use as nanoreactors,6 for protein encapsulation,7 for
extraction of water-soluble biomolecules,8 or in protein con-
formational studies.9 The assembly of surfactant-based reverse
† Universite´ de Montre´al.
‡ University of Delaware.
§ Canada Research Chair in Drug Delivery.
(1) Price, K. E.; McQuade, D. T. Chem. Commun. 2005, 1714-1716.
(2) Pileni, M. P. Nat. Mater. 2003, 2, 145-150.
(3) Carvalho, C. M. L.; Cabral, J. M. S. Biochimie 2000, 82, 1063-1085.
(4) Riter, R. E.; Kimmel, J. R.; Undikis, E. P.; Levinger, N. E. J. Phys. Chem.
B 1997, 101, 8292-8297.
(10) Ichikawa, S.; Sugiura, S.; Nakajima, M.; Sano, Y.; Seki, M.; Furusaki, S.
Biochem. Eng. J 2000, 6, 193-199.
(11) New, R. R. C.; Kirby, C. J. AdV. Drug DeliVery ReV. 1997, 25, 59-69.
(12) Stevelmans, S.; Hest, J. C. M.; Jansen, J. F. G. A.; Van Boxtel, D. A. F.
G.; Da Brabander-van den Berg, E. M. M.; Meijer, E. W. J. Am. Chem.
Soc. 1996, 118, 7398-7399.
(5) Khougaz, K.; Gao, Z. S.; Eisenberg, A. Langmuir 1997, 13, 623-631.
(6) Mecking, S.; Thomann, R.; Frey, H.; Sunder, A. Macromolecules 2000,
33, 3958-3960.
(13) Sunder, A.; Bauer, T.; Mu¨lhaupt, R.; Frey, H. Macromolecules 2000, 33,
1330-1337.
(14) Cooper, A. I.; Londono, J. D.; Wignall, G.; McClain, J. B.; Samulski, E.
T.; Lin, J. S.; Dobrynin, A.; Rubinstein, M.; Burke, A. L. C.; Frechet, J.
M. J.; Desimone, J. M. Nature (London) 1997, 389, 368-371.
(15) Sunder, A.; Kra¨mer, M.; Hanselmann, R.; Mu¨lhaupt, R.; Frey, H. Angew.
Chem., Int. Ed. 1999, 38, 3552-3555.
(7) Luisi, P. L.; Giomini, M.; Pileni, M. P.; Robinson, B. H. Biochim. Biophys.
Acta 1988, 947, 209-246.
(8) Fadnavis, N. W.; Satyavathi, B.; Deshpande, A. A. Biotechnol. Prog. 1997,
13, 503-505.
(9) Van Horn, W. D.; Simorellis, A. K.; Flynn, P. F. J. Am. Chem. Soc. 2005,
127, 13553-13560.
(16) Chen, Y.; Shen, Z.; Frey, H.; Perez-Prieto, J.; Stiriba, S.-E. Chem. Commun.
2005, 755-757.
9
10.1021/ja065462c CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 14599-14605
14599

A R T I C L E S
Jones et al.
ARX400 spectrometer in deuterated chloroform (CDCl3). 1H NMR (δ,
ppm, CDCl3): 4.31 (m, 112H); 4.1 (m, 7.99H); 3.8 (m, 113H); 3.22
(m, 114H); 2.84 (m, 115H); 2.63 (m, 116H); 1.8-2.2 (m, 208H); 0.9-
1.1 (m, 349H).
of RPMs as drug delivery systems was demonstrated using
hydrophobically modified poly(amidoamine) dendrimers for the
oral administration of 5-fluorouracil, an anticancer drug.17 While
they are good at solubilizing monomolecular entities, uni-
molecular reverse micelles present limited core space, which
may not be suitable for the incorporation of larger guests such
as proteins. In the present study, it was hypothesized that the
encapsulation of more complex structures could be better
achieved using multimolecular reverse micelles. Accordingly,
RPMs were prepared from amphiphilic star-shaped poly-
(alcohol)s, and their behavior was studied in various organic
phases. More specifically, RPMs were evaluated for their ability
to enhance the solubility of peptides/proteins in ethyl oleate, a
solvent accepted for pharmaceutical use. It is expected that the
formulation of these compounds in an oleaginous environment
will improve their stability against enzymatic degradation and
promote their absorption through biological barriers such as the
intestinal membrane.11 If administered subcutaneously, the RPM
formulation could also serve as a reservoir for the sustained
release of peptidic drugs. Here, four-arm poly(glycerol meth-
acrylate)s were employed as scaffolds in the elaboration of
RPMs. The hydrophobicity of the micellar shell was tailored
using fatty acid derivatives of varying length. The micelles were
characterized with regard to their assembly in tetrahydrofuran
(THF) and ethyl oleate, and their ability to interact with
hydrophilic macromolecules was studied using three peptidic
molecules, namely vasopressin, myoglobin, and albumin.
2.2.2. Synthesis of Star-Shaped Poly(glycerol methacrylate)
(PGOHMA). PGOHMA was obtained following the hydrolysis of the
epoxy ring of PGMA. In a typical procedure, PGMA (6 g) was dissolved
in 1-methyl-2-pyrrolidinone (NMP, 75 mL) under gentle stirring. After
complete dissolution of the polymer, 15 mL of water was added
dropwise, and the mixture was left to react at 140 °C for 24 h. The
hydrolyzed polymer was dialyzed (Spectra/Por no. 1, MW cutoff 2000,
Spectrum Laboratories, Rancho Dominguez, CA) against water for 48
1
h and then freeze-dried. H NMR spectra were recorded on a Bruker
ARX400 spectrometer in deuterated dimethyl sulfoxide (DMSO-d6).
1H NMR (δ, ppm, DMSO-d6): 4.7-4.9 (m, 2H); 3.7-3.9 (m, 3H);
3.3 (m, 48H); 1.82 (m, 1.4H); 0.4-1.2 (m, 3H).
2.2.3. Synthesis of Hydrophobically Modified Star-Shaped
(PGOHMA-CX). PGOHMA (228 equiv of hydroxyl group) was dried
by azeotropic distillation and solubilized in pyridine (0.8 M) in the
presence of catalytic amounts of 1-methylimidazole. Acyl chloride
derivatives (137 equiv) were dissolved in toluene and slowly added to
the reaction mixture under anhydrous and inert conditions. The reaction
was left to proceed overnight under reflux. K2CO3 (5 g) was added
following completion of the reaction. Excess pyridine was removed
under reduced pressure. The mixture was then transferred to a separation
funnel with 200 mL of chloroform, and pyridine residues were extracted
successively with HCl (0.1 M) and brine. The organic extracts were
dried over MgSO4, filtered, and dialyzed against chloroform for 24 h
(Spectra/Por Biotech regenerated cellulose, MW cutoff 6000-8000,
Spectrum Laboratories). The solvent was removed under reduced
pressure to yield the alkylated polymer as off-white waxy flakes. All
star-shaped amphiphilic polymers are referred to as PGOHMA-CX,
where X stands for the total number of carbon atoms present on the
fatty acid chain.
2. Experimental Section
2.1. Materials. Glycidyl methacrylate was obtained from Poly-
sciences (Warrington, PA). Other reagents employed in polymer
synthesis were purchased from Aldrich (Oakville, ON, Canada) and
used without further purification. THF was dried over sodium using
benzophenone as a dryness indicator. Myoglobin and fluorescein-
isothiocyanate conjugate albumin (FITC-albumin) were obtained from
Sigma (Oakville, ON, Canada).
2.2.4. Determination of the Degree of Alkylation. The degree of
alkylation (DA) was determined from the 1H NMR spectra of PGOHMA-
CXs using eq 1,
I(5DP + 8)
20DP(n - 4) + 10DP
DA )
× 100
(1)
2.2. Methods. 2.2.1. Synthesis of Star-Shaped Poly(glycidyl
methacrylate) (PGMA). Atom-transfer radical polymerization (ATRP)
was carried out in THF using copper bromide (CuIBr) as a catalyst
and 1,2-bipyridyl as a ligand. ATRP initiator, tetrakis(2-bromoiso-
butyryl) pentaerythritolate18 (1 equiv) was added to CuIBr (1 equiv),
1,2-bipyridyl (1 equiv), and glycidyl methacrylate (106 equiv) in THF
([monomer] ) 0.3 M). The mixture was degassed under argon for 15
min at room temperature and then heated to 90 °C. The reaction was
left to proceed for 24 h. The mixture was then poured into THF
containing ethanol (95:5) and passed through a silica gel column to
remove the copper. THF was used as eluent. The solvent was evaporated
under reduced pressure, and the crude polymer extract was precipitated
twice in diethyl ether. The polymer was further purified by Soxhlet
extraction in diethyl ether. A mean degree of polymerization of 26
units per initiator arm was targeted for a total molecular weight (MW)
of 15 000. Number- (Mn) and weight-average (Mw) MWs were
determined using a Waters 1525 size exclusion chromatography (SEC)
system equipped with a 2410 refractometer (Waters, Milford, MA) and
a low-angle laser light scattering detector (PD2000DLS, Precision
Detectors, Bellingham, MA). Analyses were performed in N,N-
dimethylformamide with 10 mM LiBr at a flow rate of 1 mL/min using
a set of three Styragel HT2, HT3, and HT4 columns (Waters),
maintained at 45 °C. 1H NMR spectra were recorded on a Bruker
where I is the intensity of the peak at δ ) 1.2-1.3 ppm, DP is the
degree of polymerization calculated from the 1H NMR of PGMA, and
n is the number of carbon atoms in the alkyl chain.
2.2.5. Particle Size Analysis. Dynamic light scattering experiments
were performed in THF and ethyl oleate (Crodamol EO, Croda Canada
Ltd., Vaughan, ON, Canada) at 25 °C and at a angle of 173° using a
Malvern Zetasizer ZS instrument (Malvern, Worcesteshire, UK). The
polymer concentration was kept constant at 1 g‚L-1, and peptide
concentration was kept at 1 wt % where applicable. All polymer
solutions were filtered twice prior to analysis using 0.22-µm hydrophilic
propylene (GHP) filters. Viscosity values of 0.455 and 9.29 cP
(determined experimentally) were employed for THF and ethyl oleate,
respectively. The theoretical diameter of a fully extended PGOHMA
backbone was calculated on the basis of the length of a saturated
carbon-carbon bond length (1.54 Å) and the degree of polymerization
1
obtained by H NMR.
2.2.6. Atomic Force Microscopy (AFM) Imaging. Forty microliters
of a solution of the polymer in THF (0.1 mg‚mL-1) was deposited on
a mica sheet (1 × 1 cm2) and centrifuged at 2000 rpm during 2 min
using a spin coater (model WS400-A6NPP-lite, Laurell Technologies,
North Wales, PA). The samples were vacuum-dried for 24 h before
AFM imaging. AFM images were recorded in tapping mode on a
Nanoscope IIIa Dimension 3100 instrument (Digital Instruments, Santa
Barbara, CA).
(17) Tripathi, P.; Khopade, A. J.; Nagaich, S.; Shrivastava, S.; Jain, S.; Jain, N.
Pharmazie 2002, 57, 261-264.
(18) Jones, M.-C.; Ranger, M.; Leroux, J.-C. Bioconjugate Chem. 2003, 14,
774-781.
2.2.7. Cryogenic Transmission Electron Microscopy (cryo-TEM)
Analysis. For cryo-TEM imaging, a thin film of polymer solution in
9
14600 J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006

Self-Assembled Nanocages for Hydrophilic Guests
A R T I C L E S
Scheme 1. Synthesis of Alkylated Star-Shaped PGOHMAa
ethyl oleate was transferred to carbon-coated copper grids. The grids
were plunged into liquid ethane at -170 °C using a FEI Vitrobot (FEI
Co., Hillsboro, OR), an automated cryo preparation system. The vitrified
samples were transferred to a Gatan 626 cryoholder and cryotransfer
stage. During observation of the vitrified samples, the cryoholder
temperature was maintained at -170 °C. Images of the vitrified samples
were taken in bright-field mode with a Tecnai G2 12 transmission
electron microscope (FEI Co.) at 120 kV accelerating voltage.
2.2.8. Determination of the Partition Coefficient of Vasopressin
between Ethyl Oleate and Water. Hydrophobically modified
PGOHMAs were dissolved in dichloromethane (DCM) (2.5 or 5
mg‚mL-1; 1 mL of DCM) and mixed with increasing volumes of an
ethanolic solution of vasopressin (Peptide International, Louisville, KY)
spiked with 3H-vasopressin (Perkin-Elmer, Boston, MA). Ethyl oleate
(500 µL) was added to the mixture. The volatile organic solvents were
removed under vacuum. At that point, 500 µL of water was added.
The two phases were gently mixed together for 24 h. Following
complete phase separation, 100-µL aliquots of both the aqueous and
organic layers were assayed for radioactivity on a LKB-Wallac 1217
Rackbeta liquid scintillation counter (Perkin-Elmer) using Ultima Gold
scintillation cocktail (Perkin-Elmer). Controls without polymer were
also prepared.
a Reaction conditions: (A) Pentaerythritol (1) was esterified with
2-bromoisobutyryl bromide in the presence of triethylamine in THF at room
temperature. (B) The ATRP initiator (2) was reacted with glycidyl
methacrylate (GMA) in THF in the presence of CuIBr/1,2-bipyridyl. (C)
PGMA (3) was hydrolyzed in a mixture of NMP/H2O (5:1) to yield star-
shaped PGOHMA. (D) PGOHMA (4) was partially esterified using alkanoyl
chlorides in pyridine in the presence of 1-methylimidazole at 140 °C.
2.2.9. Extraction of Myoglobin and Albumin. Small volumes of
a concentrated solution (100 g‚L-1) of the polymer in ethyl oleate were
mixed with predetermined volumes of a solution of myoglobin (1 g‚L-1
)
by ATRP from a four-arm multifunctional initiator, namely
tetrakis(2-bromoisobutyryl) pentaerythritolate18 (Scheme 1).
The reaction proceeded smoothly, with over 85% of the
monomer converted in 24 h. Polymerization was initiated
equally at all sites, and the star conformation was retained as
evidenced by AFM (see Supporting Information). The molecular
or FTIC-albumin (1 g‚L-1) in phosphate-buffered saline (PBS, 75 mM
NaCl, 53 mM Na2HPO4, 17 mM NaH2PO4, pH 7.4) and left to stand
at room temperature. In the event of gel formation, the mixture was
heated at 37 °C in order to destroy the organogel. Both phases were
completed to their final volume (polymer concentration 5 g‚L-1) and
gently stirred for 24 h. Sample were centrifuged at 5000g (Hermle
Z252M microcentrifuge, Mandel Scientific, Guelph, ON, Canada) for
3 min to induce phase separation. The protein content was assayed in
1
weight of the polymer was evaluated by SEC and H NMR
(Table 1). The results correlated well with the expected value.
The pendant epoxides were hydrolyzed in a mixture of NMP
and water to yield the star-shaped PGOHMA core. The hydro-
philic PGOHMA core was rendered more hydrophobic through
chemical modification of the pendant -OH functionalities
(Scheme 1). The latter were partially esterified (60 mol %) using
alkanoyl chlorides. The number of modified hydroxyl groups
was controlled by the feed ratio. In the present case, fatty acid
derivatives of varying chain length, namely lauric (C12),
myristic (C14), and stearic (C18) acids, were employed. These
derivatives were purposely chosen in order to study the influence
of alkyl chain length on the properties of the reverse polymeric
micelles.
The resulting PGOHMA-CX derivatives (Figure 1) were water-
insoluble but easily dissolved in organic solvents such as DCM
and THF. Furthermore, all polymers were soluble in ethyl oleate,
an oil approved for pharmaceutical use. Oleaginous polymeric
solutions were prepared by dissolving the polymer in ethyl oleate
at high temperature (60 °C) and then allowing the mixture to
cool to room temperature. All star-shaped alkylated PGOHMAs
formed thermoreversible ethyl oleate organogels. Gel formation
was observed at high concentration (>100 g‚L-1) for both the
lauric and myristic acid derivatives, while PGOHMA-C18
induced gelation at a much lower concentration (g10 g‚L-1).
In all cases, the sol-gel transition occurred upon cooling and
was attributed to a decrease of the affinity of the polymer for
the dispersing phase.21 Organogelation remains an intricate and
often unpredictable process in which a combination of inter-
molecular attractive interactions, such as hydrogen bond forma-
the aqueous layer by spectrophotometry for both myoglobin (λmax
)
410 nm) and FITC-albumin (λmax ) 494 nm) on a BioTek PowerWave
X (Winooski, VT) plate reader. Controls without polymer were also
prepared.
2.2.10 Analysis of Extraction Data. The percentage of peptide/
protein retained in ethyl oleate was calculated from eq 2,
CEO
% retained in ethyl oleate )
× 100
(2)
CEO + CW
where CEO and Cw are the final peptide/protein concentrations in the
oil and aqueous phase in the presence of RPMs, respectively. Results
are presented as the mean ( SEM of at least three independent
experiments.
The apparent partition coefficient (Kapp) of peptide/protein between
the organic and aqueous phases was calculated using eq 3,
CEO
Kapp
)
(3)
CW
3. Results and Discussion
3.1. Polymer Synthesis. Amphiphilic PGOHMA-CXs were
selected as building blocks to form multimolecular RPMs. The
rationale was that a star-shaped conformation would allow
sufficient mobility within the polymer structure to permit
hydrogen bond formation between neighboring polymers19,20 and
subsequent aggregation. Star-shaped PGMA was synthesized
(19) Kim, C.; Kim, K. T.; Chang, Y. J. Am. Chem. Soc. 2001, 123, 5586-
5587.
(20) Garamus, V. M.; Maksimova, T.; Kautz, H.; Barriau, E.; Frey, H.;
Schlotterbeck, U.; Mecking, S.; Richtering, W. Macromolecules 2004, 37,
8394-8399.
(21) Murdan, S.; Gregoriadis, G.; Florence, A. T. J. Pharm. Sci. 1999, 88, 608-
614.
9
J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006 14601

A R T I C L E S
Jones et al.
Table 1. Characterization of Alkylated PGOHMAs
PGMA properties
fatty acid derivatives
mean diameter (nm)
[polydispersity index]
exptl Mn
DA (%)
exptl
polymer
theor Mn
by 1H NMRa
by SEC
Mw/Mn
chain length
theor
ethyl oleate
THF
PGOHMA-C12
PGOHMA-C14
PGOHMA-C18
lauric (-OC(O)-(CH2)10-CH3
myristic (-OC(O)-(CH2)12-CH3
stearic (-OC(O)-(CH2)16-CH3
60
60
60
65.3
60.1
61.0
19 [0.3]
19 [0.3]
19 [0.3]
33 [0.24]
32 [0.25]
32 [0.24]
15 000
16 900
20 600
1.17
{
a Mn was calculated from the spectrum of PGMA using the intensities of the peaks at δ ) 3.22 and 4.2 ppm (8H).
Figure 1. 1H NMR spectrum of PGOHMA-C14 in CDCl3.
tion and van der Waals forces, are thought to participate.19,22
In the present case, free stearoyl chloride failed to produce an
organogel in ethyl oleate at a similar concentration (10 g‚L-1),
thus confirming the determining role of hydrogen bonding. It
should be mentioned that in all further experiments involving
PGOHMA-C18, polymer concentrations were adjusted to prevent
gelation.
3.2. Reverse Micelle Characterization. Micelle size was
evaluated in both ethyl oleate and THF (Table 1). In ethyl oleate,
micellar aggregates presented mean diameters of approximately
19 nm with a relatively broad size distribution (Table 1). The
observed size did not vary between the different alkylated
polymers. In THF, the average diameters experienced a 1.7-
fold increase for all formulations (Table 1). The heightened
polarity (ꢀ ) 7.6 vs 3.2 for ethyl oleate) of this solvent is thought
to contribute to the observed occurrence. Indeed, it is possible
that the more polar surrounding provided by THF induced the
swelling of the hydrophilic core and exposed some of the
hydroxyl functionalities, thus promoting the aggregation of the
polymer. When compared to the diameters reported for den-
drimers or hyperbranched structures with similar MWs (2-10
nm),12,20 the sizes observed for star-shaped micelles tend to
indicate that multimolecular rather than unimolecular micelles
are formed.23 Furthermore, the diameters observed by DLS in
both solvents are higher than the theoretical diameter of a fully
extended star-shaped PGOHMA-CX (ca. 9-10 nm), which
further supports the hypothesis that alkylated PGOHMAs as-
semble into supramolecular structures in apolar media.
The morphology of the self-assembled nanocages in oil and
THF was elucidated by AFM and cryo-TEM, respectively
(Figure 2). In both solvents, spherical aggregates were observed,
and the average diameters and polydispersity correlated well
with the light scattering results (see Supporting Information).
To date, cryo-TEM has been applied mainly to imaging in
aqueous environments,24,25 low-viscosity organic solvents,26,27
or ionic liquids;28 its utilization for oleaginous media is restricted
due to inherent difficulties in sample preparation.29 However,
the results presented here (Figure 2b) confirm the feasibility of
using cryo-TEM to study the self-assembly of polymeric
amphiphiles in oils.
3.3. Protein Partition between Ethyl Oleate and Water.
3.3.1. Extraction of Vasopressin. The ability of PGOHMA-
(23) Kra¨mer, M.; Stumbe´, J.-F.; Tu¨rk, H.; Krause, S.; Komp, A.; Delineau, L.;
Prokhorova, S.; Kautz, H.; Haag, R. Angew. Chem., Int. Ed. 2002, 41,
4252-4256.
(24) Pochan, D. J.; Chen, Z. Y.; Cui, H. G.; Hales, K.; Qi, K.; Wooley, K. L.
Science 2004, 306, 94-97.
(25) Wittemann, A.; Drechsler, M.; Talmon, Y.; Ballauff, M. J. Am. Chem. Soc.
2005, 127, 9688-9689.
(26) Balmes, O.; Malm, J.-O.; Karlsson, G.; Bovin, J.-O. J. Nanopart. Res. 2004,
6, 569-576.
(27) Boettcher, C.; Schade, B.; Fuhrhop, J.-H. Langmuir 2001, 17, 873-877.
(28) He, Y. Y.; Li, Z. B.; Simone, P.; Lodge, T. P. J. Am. Chem. Soc. 2006,
128, 2745-2750.
(22) Motulsky, A.; Lafleur, M.; Couffin-Hoarau, A. C.; Hoarau, D.; Boury, F.;
(29) Danino, D.; Gupta, R.; Satyavolu, J.; Talmon, Y. J. Colloid Interface Sci.
2002, 249, 180-186.
Benoit, J. P.; Leroux, J. C. Biomaterials 2005, 26, 6242-6253.
9
14602 J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006

Self-Assembled Nanocages for Hydrophilic Guests
A R T I C L E S
Figure 2. Study of the morphology of reverse PGOHMA-C18 micelles.
The morphology of PGOHMA-C18 was studied by (a) AFM in THF and
(b) cryo-TEM in ethyl oleate. The inset shows a magnified RPM.
CX RPMs to take up a water-soluble macromolecule in ethyl
oleate was first demonstrated through partition studies using
vasopressin, a natural peptide (1084 Da) which plays an
important role in the regulation of extracellular fluid levels in
the human organism. The encapsulation of vasopressin was
studied as a function of peptide concentration (0.25-10 wt %
of polymer). RPMs were loaded with vasopressin by mixing
solutions of the polymer in DCM/ethyl oleate with ethanolic
solutions of the peptide. The volatile organic solvents were then
removed to allow the incorporation of vasopressin. Particle size
analysis performed at that point revealed a 3- to 4-fold increase
in micelle size, with PGOHMA-C18 experiencing the most
change with diameters around 65-70 nm. The aqueous phase
was then added in order to extract the un-incorporated peptide.
In the absence of micelles, vasopressin was only marginally
retained in the oleaginous phase (<5%; Kapp < 0.05) and par-
titioned preferentially into the aqueous layer. Linear PGOHMA-
CX also failed to take up considerable amounts of vasopressin
in ethyl oleate (data not shown), emphasizing the importance
of the polymer structure.30,31 The addition of RPMs, however,
led to a significant increase in the quantity of peptide solubilized
in the organic phase. As vasopressin exhibited limited affinity
for the oleaginous phase, its partition in the organic phase is
attributed to its loading in the micellar core.
Figure 3. Partition of vasopressin between ethyl oleate and water in the
presence of RPMs of (a) PGOHMA-C12, (b) PGOHMA-C14, and (c)
PGOHMA-C18. The results for the amount (%) of vasopressin retained
(closed symbols) are presented as the mean ( SEM of at least three
independent experiments. The partition coefficient (Kapp; open symbols)
was estimated using eq 3. The initial loadings are expressed in wt % of
polymer (2.5 or 5 mg).
The retention efficiency of vasopressin in the RPMs was
influenced by both the nature of the star-shaped polymer and
the initial peptide loading (Figure 3). Under the tested experi-
mental conditions, PGOHMA-C12 reverse micelles were able
to retain 0.2-2.4% of their weight of vasopressin in the oil
phase, corresponding to entrapment efficiencies between 46.4
( 6.6% and 69.7 ( 6.2% and Kapp values between 0.9 and 2.3
(Figure 3a). PGOHMA-C14 showed higher retention efficacy,
with partition coefficients between 1.8 and 3.4 (Figure 3b). The
stearic acid derivative was the most effective, with up to 10%
(30) Stiriba, S.-E.; Kautz, H.; Frey, H. J. Am. Chem. Soc. 2002, 124, 9698-
9699.
(31) Chen, Y.; Shen, Z.; Pastor-Pe´rez, L.; Frey, H.; Stiriba, S.-E. Macromolecules
2005, 38, 227-229.
9
J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006 14603

A R T I C L E S
Jones et al.
Table 2. Maximum Protein Loading in RPMs
vasopressin
(MW 1048)
myoglobin
(MW 17 600)
FITC-albumin
(MW 66 000)
a
a
a
DLmax
DLmax
DLmax
RPM
xmax
(% w/w)
xmax
(% w/w)
xmax
(% w/w)
PGOHMA-C12
PGOHMA-C14
PGOHMA-C18
0.64
0.85
0.99
6.5
17.8
77.0
0.08
0.07
0.07
5.0
4.2
3.8
n.d.
n.d.
0.009
n.d.
n.d.
1.7
a Maximum drug loading. “n.d.” indicates not determined.
3.3.2. Extraction of Myoglobin and Albumin. The results
obtained with vasopressin demonstrate that RPMs are able to
interact with a hydrophilic molecule and to significantly increase
its solubility in an oleaginous phase. The RPM’s extraction
capability was further challenged using a more complex
polypeptide, namely myoglobin. Myoglobin is a 17 600 Da
protein responsible for the transport and storage of oxygen in
the muscles. Unlike vasopressin, myoglobin presents an elabo-
rate tri-dimensional structure, which could be destroyed in the
presence of ethanol. As a result, the loading method was adapted
to avoid the use of potentially deleterious solvents.
Figure 4. Fitting of the partition data according to the Langmuir isotherm
for PGOHMA-C12 (closed circles), PGOHMA-C14 (open triangles), and
PGOHMA-C18 (closed diamonds).
of its weight in vasopressin solubilized in ethyl oleate and
efficiencies ranging from 94.5 ( 2.5% to 98.7 ( 0.7%.
Accordingly, in the presence of PGOHMA-C18 micelles, the
peptide concentrations were 17-77 times higher in oil than in
water (Figure 3c). The remarkable ability of this formulation
to retain vasopressin in ethyl oleate may be attributed to the
higher lipophilicity of the C18 shell. The C18 chains may act
as a barrier that prevents the entrapped peptide from migrating
into the apolar external phase, where it could be extracted by
water molecules.15,23
In all cases, the Kapp values were inversely related to guest
concentration. Peptide molecules that are taken up by the reverse
micelles may slowly fill the core-shell interface until the space
is completely occupied; from this point forward, excess peptide
is transferred to the aqueous phase, shifting the Kapp toward
lower values.13 The Langmuir adsorption isotherm was thus
applied to evaluate the maximum amount of peptide that can
be accommodated32 (Figure 4).
As expected, increasing the MW of the guest molecule
resulted in decreases of the loading efficiency and the Kapp value
(see Supporting Information). Still, the reverse micelles were
able to load approximately 1.3-2.9% of their weight in
myoglobin, with PGOHMA-C18 showing the best results. At low
concentrations (1-2 wt%), the latter extracted over 80% of the
myoglobin present in the aqueous phase, while PGOHMA-C12
and PGOHMA-C14 took up approximately 50% of the protein.
The slight difference in efficacy previously observed between
the two shorter-chained RPMs disappeared with the increase
in guest size. The C12 derivative even proved to be somewhat
more efficient in extracting myoglobin than PGOHMA-C14.
The Kapp values remained above unity at the lowest initial
loading tested, confirming that RPMs were able to improve the
partition of myoglobin between oil and water. For all formula-
tions, the Kapp values decreased steadily over the range of
concentrations studied (see Supporting Information).
The Langmuir isotherm is based on the premise that solutes
will adsorb onto a surface to form a monomolecular layer and
is described by eq 4,
The myoglobin partition data were fitted using the Langmuir
isotherm (eq 4). Contrary to what was observed with vaso-
pressin, there was no difference between the slopes of the
different RPM formulations. Replacing vasopressin by myo-
globin led to an 8- to 13-fold decrease in the xmax value (Table
2), which correlates well with the fact that this protein is 16
times larger than vasopressin. The maximum loading values for
myoglobin are reported in Table 2. The loading was highest
for PGOHMA-C12 (5% w/w) and decreased slightly with
lengthening of the hydrophobic chain, in opposition to what
was seen for vasopressin. This difference may be ascribed to
both the MW of the encapsulated compounds and the extraction
methods. Vasopressin was first loaded inside the micelles with
the help of ethanol and then extracted back with the aqueous
phase. To avoid using an organic solvent in the presence of a
protein, myoglobin was first dissolved in the aqueous buffer
and then incubated with the oleaginous RPM solution. As a
consequence, augmenting the length of the hydrocarbon chain
may increase steric hindrance and thus lower the drug loading.
Moreover, since myoglobin forms aggregates of 3-4 nm in
aqueous solutions (one-sixth of the micellar size), its diffusion
through a highly shielded micelle core may be impaired.
C
x
1
C
)
+
(4)
xmaxKads xmax
where C is the molar concentration of peptide, x is the mole
fraction of peptide in the micellar phase, xmax is the maximum
of that mole fraction value, and Kads is the Langmuir adsorption
constant.
The partition data fitted well with the Langmuir model (r 2
> 0.99), supporting the hypothesis that loading can reach
saturation and occurs at the core-shell interface.32,33 Interest-
ingly, the value of the slope was influenced by the nature of
the hydrophobic segment (Figure 4). The slope was more
pronounced for the C12 and C14 derivatives compared to
PGOHMA-C18, yielding xmax values of 0.64, 0.85, and 0.99,
respectively (Table 2). These experimental xmax values were used
to compute the maximum amount of drug that can be accom-
modated within the RPMs (Table 2). As expected, PGOHMA-
C18 showed the highest solubilization capacity of all formula-
tions, which corroborates the partition data.
(32) Choucair, A.; Eisenberg, A. J. Am. Chem. Soc. 2003, 125, 11993-12000.
(33) Croy, S. R.; Kwon, G. S. J. Pharm. Sci. 2005, 94, 2345-2354.
9
14604 J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006

Self-Assembled Nanocages for Hydrophilic Guests
A R T I C L E S
To further investigate the effect of MW on uptake, a 66 000
Da serum protein (i.e., albumin) was selected for encapsulation
into PGOHMA-C18 RPMs. Albumin is known to interact with
hydrophobic compounds, such as lauric and myristic acids.
However, it has been shown that conformational rigidity limits
its interactions with neutral fatty acids containing over 14 carbon
atoms;34,35 the nonspecific adsorption of PGOHMA-C18 RPMs
and oil adsorption should therefore remain negligible. The
extraction ability of the PGOHMA-C18 RPMs was influenced
by the initial protein concentration, similar to what was observed
for vasopressin and myoglobin. Although significantly lower
than those observed for vasopressin or myoglobin, the maximal
partition coefficient remained above unity at initial concentra-
tions below 2% (see Supporting Information). Albumin parti-
tioned preferentially in water at concentrations above this
threshold. Fitting the partition data to the Langmuir isotherm
yielded an xmax value of 9.0 × 10-3, which corresponds to a
116- and 8.5-fold decrease compared to the values obtained for
vasopressin and myoglobin, respectively (Table 2). Further proof
of the more favorable interaction of PGOHMA-C18 RPM with
smaller proteins was provided through a kinetic study of the
loading. In the case of vasopressin, where the peptide is loaded
in the RPM prior to extraction with water, the time dependency
of guest loading was not evaluated. For both myoglobin and
albumin, maximum loading was achieved in approximately 1
h. After 30 min, RPMs had extracted ∼20% of albumin from
the aqueous phase, compared to ∼40% for myoglobin. Although
PGOHMA-C18 micelles remain somewhat effective in promoting
the solubilization of albumin, albeit at lower concentration than
vasopressin or myoglobin, the analysis of the partition data
indicates that these micelles are better suited for the encapsula-
tion of poly(amino acid)s with a MW between those of
vasopressin and myoglobin.
4. Conclusion
Hydrophobically modified star-shaped PGOHMAs were suc-
cessfully employed as building blocks for supramolecular
assemblies in a pharmaceutically acceptable oil. These RPMs
were able to drastically increase the solubility of various proteins
including vasopressin, myoglobin, and, to a lesser extent,
albumin in ethyl oleate. The nature of the hydrophobic segment
and the guest molecule had a marked influence on the solubi-
lizing ability of the reverse micelles. After additional processing,
the system described here could be applied to the oral delivery
of fragile peptidic molecules11 or for their sustained release
following parenteral extravascular injection.36
Acknowledgment. This work was financially supported by
the Natural Sciences and Engineering Research Council of
Canada and Mistral Pharma Inc. (Montreal, QC, Canada). M.C.J.
acknowledges a scholarship from the Fonds de la Recherche
en Sante´ du Que´bec. Financial contribution from the Canada
Research Chair program and NSF NIRT (Grant No. DMR-
0210247) is also recognized.
Supporting Information Available: AFM picture of four-
arm star-shaped PGOHMA, extraction of myoglobin by RPMs,
and extraction of albumin by RPMs. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA065462C
(34) Bird, D. A.; Laposata, M.; Hamilton, J. A. J. Lipid Res. 1996, 37, 1449-
1458.
(35) Ashbrook, J. D.; Spectore, A. A.; Santos, E. C.; Fletcher, J. E. J. Biol.
Chem. 1975, 250, 2333-2338.
(36) Plourde, F.; Motulsky, A.; Couffin-Hoarau, A. C.; Hoarau, D.; Ong, H.;
Leroux, J. C. J. Controlled Release 2005, 108, 433-441.
9
J. AM. CHEM. SOC. VOL. 128, NO. 45, 2006 14605