Lactic acid- and carbonate-based crosslinked polymeric micelles for drug delivery

Article abstract of DOI:10.1002/pola.26392

Our objective was to synthesize and evaluate lactic acid- and carbonate-based biodegradable core- and core-corona crosslinkable copolymers for anticancer drug delivery. Methoxy poly(ethylene glycol)-b-poly(carbonate-co- lactide-co-5-methyl-5-allyloxycarbo

Full text of DOI:10.1002/pola.26392

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
Lactic Acid- and Carbonate-Based Crosslinked Polymeric Micelles
for Drug Delivery
Michael Danquah,1 Tomoko Fujiwara,2 Ram I. Mahato1
1Department of Pharmaceutical Sciences, University of Tennessee Health Science Center, Memphis, Tennessee, 38103
2Department of Chemistry, The University of Memphis, Memphis, Tennessee 38152
Correspondence to: R. I. Mahato 19 South Manassas, CRB RM 224, Memphis, Tennessee 38103-3308 (E-mail: rmahato@uthsc.edu)
Received 4 June 2012; accepted 12 September 2012; published online 19 October 2012
DOI: 10.1002/pola.26392
ABSTRACT: Our objective was to synthesize and evaluate lactic
acid- and carbonate-based biodegradable core- and core-co-
rona crosslinkable copolymers for anticancer drug delivery.
Methoxy poly(ethylene glycol)-b-poly(carbonate-co-lactide-co-
5-methyl-5-allyloxycarbonyl-1,3-dioxane-2-one) [mPEG-b-P(CB-
co-LA-co-MAC)] and methoxy poly(ethylene glycol)-b-poly(acry-
loyl carbonate)-b-poly(carbonate-co-lactide) [mPEG-b-PMAC-b-
P(CB-co-LA)] copolymers were synthesized by ring-opening
polymerization of LA, CB, and MAC using mPEG as an macroi-
nitiator and 1,8-diazabicycloundec-7-ene as a catalyst. These
amphiphilic copolymers which exhibited low polydispersity
and critical micelle concentration values (0.8–1 mg/L) were
used to prepare micelles with or without drug and stabilized
by crosslinking via radical polymerization of double bonds
introduced in the core and interface to improve stability.
mPEG114-b-P(CB8-co-LA35-co-MAC2.5) had a higher drug encap-
sulation efficiency (78.72% 6 0.15%) compared to mPEG114-b-
PMAC2.5-b-P(CB9-co-LA39) (20.29% 6 0.11%).1H NMR and IR
spectroscopy confirmed successful crosslinking (ꢀ70%) while
light scattering and transmission electron microscopy were
used to determine micelle size and morphology. Crosslinked
micelles demonstrated enhanced stability against extensive
dilution with aqueous solvents and in the presence of physio-
logical simulating serum concentration. Furthermore, bicaluta-
mide-loaded crosslinked micelles were more potent compared
to non-crosslinked micelles in inhibiting LNCaP cell prolifera-
tion irrespective of polymer type. Finally, these results suggest
crosslinked micelles to be promising drug delivery vehicles for
C
V
chemotherapy. 2012 Wiley Periodicals, Inc. J Polym Sci Part
A: Polym Chem 51: 347–362, 2013
KEYWORDS: biomaterials; core-crosslinked micelles; drug deliv-
ery systems; interface-crosslinked micelles; micelles; polycar-
bonate; poly lactic acid
INTRODUCTION Polymeric micelles are presently under
intense investigation as nanovehicles for delivery of low-molec-
ular weight chemotherapeutic drugs. Key advantages of poly-
meric micelles include efficient solubilization of hydrophobic
agents in their core1–3 and the ability to be made site-specific.
Furthermore, they are nanosized supramolecules which self-
assemble into core-corona structures from amphiphilic block
copolymers above the critical micelle concentration (CMC).3
These intrinsic physicochemical properties of micelles often
leads to prolonged blood circulation kinetics4 and preferential
tumor accumulation via the enhanced permeation and reten-
tion effect.5–7 Nonetheless, poor in vivo stability resulting in
premature dissociation and subsequent untimely drug release
has limited their clinical application.8–11 Polymeric micelles
become unstable either due to destabilization by plasma
proteins (kinetic instability) or dilutions below their CMC
(thermodynamic instability) upon systemic administration.
micelle thermodynamic and kinetic stability using physical
or chemical strategies. As micelles with low CMC values have
better thermodynamic stability, typical physical approaches
involve strategically altering properties of amphiphilic
copolymers, which can affect the CMC (e.g., chain length and
chemical composition).12,13 For example, tailoring the hydro-
phobic core to contain aromatic moieties has been shown to
improve micelle stability through p–p interactions as demon-
strated by decreased CMC values.12 In practice, physically
stabilized micelles are limited by their inability to prevent
micelle destabilization caused by blood components.
In contrast, introduction of chemical covalent crosslinks into
the shell, core, or core-corona interface of micelles is a
potential approach to stabilize micelles against both plasma
proteins and sub-CMCs.13–16 Over the past decade, shell-
crosslinked micelles have been extensively studied.17 It has
been shown that shell-crosslinked micelles remain stable
upon infinite dilution compared to precursor micelles. High
core mobility, greater versatility in core composition and
To facilitate clinical translation of polymeric micelle-based
therapeutics, recent research effort has focused on increasing
Additional Supporting Information may be found in the online version of this article.
C
V
2012 Wiley Periodicals, Inc.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
347

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
properties, and the membrane-like characteristics of the
crosslinked shell layer, which can improve encapsulation are
advantages associated with shell-crosslinked micelles. How-
ever, one longstanding issue with shell-crosslinked micelles
is their low crosslinking efficiency due to their need to be
prepared at high dilution to prevent intermicellar crosslink-
ing. Furthermore, shell crosslinking tends to suppress the
mobility of the hydrophilic corona chains (e.g., poly(ethylene
glycol) [PEG]).
allyloxycarbonyl-1,3-dioxan-2-one)-b-poly(carbonate-co-lactide)
[mPEG-b-PMAC-b-P(CB-co-LA)] biodegradable copolymers for
core- and core-corona interface-crosslinked micelles. Effect
of polymer composition and crosslinking moiety location on
key micelle properties was also investigated. Self-aggregation
behavior, morphology, and stability of non-crosslinked (NCM)
and crosslinked micelles (CM) were determined by dynamic
light scattering (DLS), transmission electron microscopy
(TEM), proton nuclear magnetic resonance (1H NMR)
spectroscopy, and fluorescence spectroscopy. Furthermore,
influence of crosslink location on micelle stability against
physiological simulating dilution and serum conditions as
well as drug release was investigated.
Core-crosslinked micelles exhibit better crosslinking efficien-
cies as they do not need to be prepared at high dilution to
prevent intermicellar crosslinking. Several groups have intro-
duced crosslinkable groups into the hydrophobic portion of
the block copolymer which may be polymerized after micel-
lation to lock the micelle structure. For instance, Kataoka’s
group reported core-crosslinked micelles prepared using
poly(ethylene glycol)-b-polylactide (PEG-b-PLA) possessing
methacryloyl groups at the end of the PLA chain.18,19 These
methacryloyl groups could be photopolymerized using ultra-
violet (UV) light or polymerized thermally via the addition of
a radical initiator. In another paper, Hu et al.20 covalently sta-
bilized micelle architecture using a hydrophobic polymer
block containing crosslinkable double bonds. Recently, Garg
et al.21 used click chemistry to introduce hydrolyzable cross-
links in the core of poly(ethylene oxide)-block-poly(e-capro-
lactone) (PEO-b-PCL) micelles. These crosslinked micelles
also demonstrated enhanced stability under diluted condi-
tions without any detrimental effect on the drug encapsula-
tion and in vitro drug release profiles.
EXPERIMENTAL
Materials
2,2-bis(Hydroxymethyl) propionic acid, tin(II) 2-ethylhexa-
noate (Sn(Oct)2), dimethylaminopyridine, allyl bromide, and
benzyl bromide were purchased from Sigma Aldrich (St.
Louis, MO) and used as received. mPEG (Mn ¼ 5000, polydis-
persity index [PDI] ¼ 1.03) was obtained from Sigma Aldrich
(St. Louis, MO) dried by lyophilization from benzene solution
unless it was freshly opened reagent. L-Lactide (LA) was
purchased from PURAC Biochem bv (Gorinchem, The
Netherlands) and recrystallized from toluene several times.
1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochlor-
ide (EDC) and hydroxybenzotriazole (HOBt) were obtained
from AK Scientific (Palo Alto, CA). 1,8-Diazabicyclo[5.4.0]
undec-7-ene (DBU) was purchased from Sigma Aldrich, dried
over calcium hydride, and distilled. All other reagents were
obtained from Sigma Aldrich and used without further
purification.
Armes and coworkers22 have elegantly stabilized micelles by
crosslinking the core-corona interface using triblock copoly-
mers in which the reactive groups capable of undergoing
crosslinking are located in the central block. Interfacial
crosslinking is potentially more beneficial than shell- or
core-crosslinked micelles as it combines the advantages of
both methods and may be prepared at high concentrations
without intermicellar crosslinking and aggregation. In
another study, Xu and coworkers showed reduction-sensitive
reversibly interfacially crosslinked micelles to be stable
against extensive dilution and physiological salt conditions
and retained most drugs even at concentrations below the
CMC.23 Yang et al.24 recently reported preparing interface-
crosslinked micelles using PEG-b-poly(acryloyl carbonate)-b-
polycaprolactone (PEG-b-PAC-b-PCL) triblock copolymer.
Once micelles were prepared, the double bonds in the
acryloyl carbonate were photo-crosslinked using UV light
and a photoinitiator.
Synthesis of 5-Methyl-5-allyl-1,3-dioxane-2-one
Method 1
5-Methyl-5-benzyloxycarbonyl-1,3-dioxane-2-one (CB) mono-
mer was first synthesized as described by Danquah et al.12
Briefly, 2,2-bis(hydroxymethyl)propionic acid (0.168 mol)
and potassium hydroxide (0.169 mol) were dissolved in 125
ꢁ
mL of dimethylformamide (DMF) by heating at 100 C for 1
h. Subsequently, benzyl bromide (0.202 mol) was added
dropwise to the reaction mixture and allowed to stir at 100
ꢁC for 15 h. The solvent was then removed under reduced
pressure, the residue was dissolved in ethyl acetate (150
mL), hexanes (150 mL), and water (100 mL), and dried over
Na2SO4. The organic layer solvent was removed under vac-
uum and recrystallized from toluene to yield pure benzyl
2,2-bis(methylol)propionate. Next, benzyl 2,2-bis(methylol)-
propionate (0.05 mol) was dissolved in a mixture of CH2Cl2
(150 ꢁmL) and pyridine (25 mL) and the solution chilled to
ꢂ78 C. A solution of triphosgene (25 mmol) in CH2Cl2 was
added dropwise to the reaction mixture over 1 h and stirred
for an additional 2 h at the room temperature. The reaction
was then quenched with saturated aqueous NH4Cl (75 mL),
and the organic layer was sequentially washed with 1 M
aqueous HCl (3 ꢃ 100 mL) and saturated aqueous NaHCO3
(1 ꢃ 100 mL), dried with Na2SO4, and evaporated under
Our group has developed a series of thermodynamically sta-
ble (low CMCs) lactic acid- and carbonate-based copolymers
for micellar drug delivery.12,25 To further enhance micelle
stability and controlled drug release to make our delivery
systems clinically relevant, we herein report on the synthe-
sis, characterization, and in vitro evaluation of methoxy
poly(ethylene glycol)-b-poly(carbonate-co-lactide-co-5-methyl-
5-allyloxycarbonyl-1,3-dioxan-2-one) [mPEG-b-P(CB-co-LA-co-
MAC)] and methoxy poly(ethylene glycol)-b-poly (5-methyl-5-
348
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
vacuum to give CB. The crude product was purified by
recrystallization from ethyl acetate to obtain white crystals.
Synthesis of mPEG-b-PMAC-b-P(Lactide-co-carbonate)
The following reaction mixture was prepared in a fume hood
under ambient atmosphere. mPEG (150 mg, 30 mmol) and
MAC (40 mg, 0.2 mmol) were dissolved in CH2Cl2 (3 mL) in
a 25-mL reaction vessel. DBU (15 mL, 0.1 mmol) was added,
and the reaction was carried out for 2 h to yield mPEG-b-
PMAC. To this solution, a mixture (3 mL) of CB (100 mg, 0.4
mmol) and LA (100 mg, 0.693 mmol) was infused into the
reaction vessel at a rate of 1 mL/min. DBU (15 mL) was then
added, and the reaction was allowed to proceed for 3 h, after
which benzoic acid (40 mg, 0.327 mmol) was added. PEG-b-
PMAC-b-P(CB-co-LA) copolymer was purified by concentrat-
ing the reaction mixture to approximately 50% of the initial
volume, added dropwise into excess of cold isopropanol, and
dried under vacuum.
5-Methyl-5-allyl-1,3-dioxane-2-one (MAC) monomer was then
synthesized by first subjecting CB to catalytic hydrogenation
and reacting the subsequent product with allyl alcohol. In
brief, CB (3 g) was dissolved in ethyl acetate (30 mL) con-
taining 600 mg Pd/C. The Parr bottle was purged thrice
with H2, charged to 50 psi, and the reaction was allowed to
proceed for 3 h. Subsequently, Pd/C catalyst was separated
by centrifugation, the solvent was filtered and evaporated
under vacuum to give 5-methyl-2-oxo-1,3-dioxane-5-carbox-
ylic acid (MTC-OH) as a white crystal. Next, MTC-OH (3
mmol), allyl alcohol (2.5 mmol), HOBt (3.6 mmol), and EDC
(4.5 mmol) were dissolved in DMF (20 mL) to which trie-
thylamine (TEA) (3.5 mmol) was later added, and the reac-
tion was allowed to proceed for 18 h. Afterward, 20 mL of
ethyl acetate was added to the mixture and then washed
with water. The organic layer was dried over Na2SO4, evapo-
rated under vacuum, and the crude product purified by col-
umn chromatography to give MAC as white crystals (30%).
Polymer Characterization
NMR
1H NMR, 13C NMR, 2D-COSY, and 1H-13C HSQC spectra were
recorded on a Bruker (400 MHz, T ¼ 25 ꢁC) using deuter-
ated chloroform (CDCl3) and deuterated dimethyl sulfoxide
(DMSO-d6) as solvents. For crosslinked micelles, they were
first prepared, subjected to crosslinking, lyophilized and
redissolved in CDCl3 prior to 1H NMR studies. The chemical
shifts were calibrated using tetramethylsilane as an internal
reference and given in parts per million.
Method 2
MAC was synthesized as described by Hu et al.20 Briefly, a
mixture of 2,2-bis(hydroxymethyl)propionic acid (9.0 g,
67.11 mmol), potassium hydroxide (88% assay; 4.30 g,
76.79 mmol), and DMF (50 mL) was heated to 100ꢁC for 1 h
with stirring at which point a homogenous potassium salt
solution was formed. Allyl bromide (5.8 mL, 67.11 mmol)
was added dropwise to the warm solution, and stirring was
continued at 45ꢁC for 48 h. Upon completion of the reaction,
the mixture was cooled, solvent was removed under vacuum,
and the residue was dissolved in methylene chloride (200
mL) and water (100 mL). The organic layer was retained,
washed with water (100 mL), dried (Na2SO4), and evapo-
rated to yield a viscous yellowish liquid (70%).
Gel Permeation Chromatography
A Shimadzu 20A gel permeation chromatography (GPC) sys-
tem equipped with two Jordi Gel DVB500 columns and a dif-
ferential refractive index detector was used to determine the
molecular weight and PDI of the copolymers. THF was used
as eluent at a flow rate of 1 mL/min at 35 ꢁC. A series of
narrow polystyrene standards (900–100,000 g/mol) were
used for calibration, and the data were processed using a
LCSolution ver.1.21 GPC option software.
Infrared Spectra
Allyl 2,2-bis(methylol)propionate (10 g, 58 mmol) and ethyl
chloroformate (16.5 mL, 0.173 mol) were dissolved in tetra-
hydrofuran (THF) (200 mL), and the solution was stirred at
0ꢁC for 30 min under N2. Subsequently, TEA (24.2 mL, 0.173
mol) was added dropwise over 30 min, after which the reac-
tion mixture was removed from the ice bath and stirred at
room temperature overnight. TEA-HCl precipitate was fil-
tered and the filtrate concentrated under reduced pressure.
The ensuing solid was recrystallized from THF/ether to
obtain white crystals (70%).
Copolymer composition and degree of crosslinking were con-
firmed with Fourier transform infrared (FTIR) spectra using
a Perkin-Elmer FT-IR spectrometer.
CMC
Fluorescence spectroscopy was used to estimate the CMC of
mPEG-b-P(LA-co-CB-co-MAC) and mPEG-b-PMAC-b-P(CB-co-
LA) copolymers using pyrene as a hydrophobic fluorescent
probe as described previously.26 Fluorescence spectra of py-
rene were recorded with a Molecular Devices SpectraMax
M2/M2e spectrofluorometer (Sunnyvale, CA) with excitation
wavelength 338 nm (I3) and 333 nm (I1) and emission wave-
length of 390 nm. The intensity ratio (I3/I1) was plotted
against the logarithm of polymer concentration and CMC
obtained as the point of intersection of two tangents drawn
to the curve at high and low concentrations, respectively.
Synthesis of mPEG-b-P(Lactide-co-carbonate-co-MAC)
The following reaction mixture was prepared in a fume hood
under ambient atmosphere. mPEG (150 mg, 30 mmol), LA
(100 mg, 0.693 mmol), CB (100 mg, 0.4 mmol), and MAC
(20 mg, 0.1 mmol) were dissolved in CH2Cl2 (6 mL) in a 25-
mL reaction vessel. DBU (30 mL, 0.2 mmol) was added, and
the reaction was carried out for 3 h after which benzoic acid
(40 mg, 0.327 mmol) was added. The resulting solution was
concentrated to approximately 50% of the initial volume and
added dropwise into excess of cold isopropanol with stirring.
The polymer was then dried under vacuum.
Preparation of Micelles and Crosslinking
mPEG-b-P(LA-co-CB-co-MAC) and mPEG-b-PMAC-b-P(CB-co-
LA) micelles were prepared using the film sonication method
as previously described with slight modifications.26 Briefly,
20 mg of the copolymer was dissolved in CH2Cl2 (2 mL). The
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
349

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
mixture was sonicated for 5 min to ensure homogeneity, and
the solvent was evaporated under a flow of N2. The resulting
film was hydrated (10 mL) and sonicated for 10 min using a
Misonix ultrasonic liquid processor (Farmingdale, NY) with
an amplitude of 70. The ensuing formulation was then cen-
trifuged at 5000 rpm for 10 min and the supernatant filtered
using a 0.22 -lm nylon filter.
validated UV spectrophotometer by measuring the absorb-
ance of the solution at 270 nm. Samples were injected back
into the original media once after being analyzed. The cumu-
lative amount of drug released into the media at each time
point was evaluated as the percentage of total drug release
to the initial amount of the drug. All experiments were per-
formed in triplicate, and the data were reported as the mean
of the three individual experiments.
Micelles were stabilized by polymerization of the allyl moi-
eties. Briefly, the micelle solution was first deoxygenated by
bubbling with argon for 1 h. A solution of 2,2-azo-bis(isobu-
tyronitrile) (AIBN) (1.0 wt % of polymer) in THF was then
added and stirred for 2 h followed by heating to 70ꢁC for 24 h.
Assessment of In Vitro Micelle Stability
The stability of NCM and CM in PBS (pH 7.4) and under
physiologically simulating conditions (45 mg/mL bovine
serum albumin [BSA]) was assessed by DLS. For stability in
serum, micelle solution at a final polymer concentration of
20 mg/mL was incubated in BSA (45 mg/mL) at 37ꢁC for 24
and 48 h with gentle shaking at 100 rpm. Particle size distri-
bution of 1-mL aliquots was determined by DLS (n ¼ 3).
Drug Loading and Encapsulation Efficiency
Bicalutamide-loaded micelles were prepared as described
above. 1 mg drug and 19 mg copolymer were dissolved in
CH2Cl2 (2 mL) and evaporated under a stream of N2. The
film was then hydrated with 10 mL of phosphate buffered
saline (PBS), sonicated, and filtered through a 0.22 mm filter.
Acetonitrile was used to extract the loaded drug, and the
amount of bicalutamide was determined using UV spectros-
copy at 270 nm. Drug loading content and encapsulation effi-
ciency were then determined using eqs 1 and 2 as follows:
In Vitro Cytotoxicity of Bicalutamide-Loaded Micelles
The ability of bicalutamide-loaded PEG-b-PMAC-b-P(CB-co-LA)
and PEG-b-P(CB-co-LA-co-MAC) NCM and CM to inhibit cell
proliferation was evaluated using LNCaP human prostate can-
cer cell line. About 3 ꢃ 103 cells were incubated with bicaluta-
mide-loaded micelle formulations (0, 25, and 50 mM of drug)
for 24 and 48 h. At the end of treatment, (3-(4,5-Dimethylthia-
zol-2-yl)-2,5-diphenyltetrazolium bromide, a yellow tetrazole)
(MTT) solution (10% v/v) was added to each well, incubated
for 4 h, residual formazan crystals were solubilized with
DMSO, and the plate was analyzed using a microplate reader
(560 nm). Cell viability was expressed as a percentage of con-
trol and data reported as the mean of triplicate experiments.
drug loading density ð%Þ
weight of drug in micelles
¼
ꢃ 100
ð1Þ
ð2Þ
weight of micelles
drug encapsulation efficiency ð%Þ
weight of drug in micelles
¼
ꢃ 100
weight of drug originally fed
RESULTS
Particle Size Distribution and Morphology
Synthesis and Characterization of MAC
Mean particle size (intensity mean) and size distribution of
NCM and CM were determined by DLS using a Zetasizer
(Malvern Instruments, Worcestershire, UK) at a 1 mg/mL
polymer concentration. Samples were analyzed at room tem-
perature with a 90ꢁ detection angle, and the mean micelle
size was obtained as a Z-average. Five repeat measurements
were performed, and data were reported as the mean diame-
ter 6 SD. Zeta potential was measured in PBS (pH ¼ 7.4).
NCM and CM micelles prepared using mPEG-b-P(LA-co-CB-
co-MAC) and mPEG-b-PMAC-b-P(CB-co-LA) copolymers were
visualized using a JEM-100S (Japan) TEM. Micelles were
loaded on a copper grid, followed by blotting of excess liquid
prior to negative staining with 1% uranyl acetate. The grid
was visualized under the electron microscope at 60 kV and
magnifications ranging from 50,000ꢃ to 100,000ꢃ.
To facilitate micelle stabilization via crosslinking, we synthe-
sized MAC which is a cyclic monomer containing a double
bond functional moiety. Figure 1(A) illustrates the reaction
pathway for synthesizing the MAC monomer. Drawing from
our experience in synthesizing six-membered cyclic carbonate
monomers, we first synthesized MAC using Method 1 accord-
ing to the procedure previously reported by Pratt et al.27 with
slight modification. This method allowed the synthesis of CB
intermediate which is an integral component of our lactic
acid- and carbonate-based copolymers. To this end, 2,2-bis(h-
ydroxymethyl)propionic acid was reacted with benzyl bro-
mide to give benzyl 2,2-bis(methylol)propionate, which was
then reacted with triphosgene to obtain CB (MW: 250 g/mol).
The chemical structure of CB was confirmed using mass and
1H NMR spectroscopy and our results matched the litera-
ture.12,25 CB was then dehydrogenated using Pd/C catalyst to
give MTC-OH (30% of CB) (MW: 160 g/mol). Next, MTC-OH
was coupled to allyl alcohol using EDC HOBt chemistry and
purified using column chromatography to obtain MAC.
In Vitro Drug Release from Micelles
The dialysis technique was used to study the release profile
of bicalutamide from NCM and CM in PBS (pH 7.4) with
0.1% Tween-80. Bicalutamide-loaded micelles with a final
bialutamide concentration of 0.2 mg/ml were placed into a
dialysis membrane with a molecular weight cut-off of 2000
Da and dialyzed against 50 mL PBS (pH 7.4) in a thermo-
controlled shaker with a stirring speed of 150 rpm. 1 ml
samples were taken at specified times and assayed with a
However, as we had large quantities of CB from previous stud-
ies, coupled with significant yield drops in step 1C [Fig. 1(A)]
and the time consuming column purification step associated
with this method, we explored an alternative synthesis route
(Method 2). In this method, 2,2-bis(hydroxymethyl)propionic
350
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 1 (A) Synthesis scheme for MAC. Conditions: (1A) KOH, DMF, 100ꢁC, 15 h. (1B) Triphosgene, pyridine, CH2Cl2, ꢂ78 to 0ꢁC.
(1C) Pd/C (10%), H2, ethyl acetate, RT, 40 psi, 3 h. (1D) EDC, HOBt, TEA, DMF, RT, 18 h. (2 A) KOH, DMF, preheat to 100ꢁC, 1 h,
45ꢁC, 48 h. (2B) ClCO2Et, THF, TEA, 0ꢁC, 30 min, RT, 12 h. (B) 1H NMR spectrum of MAC in CDCl3. (C) FTIR spectrum of MAC.
acid was reacted with allyl bromide to yield allyl 2,2-bis(me-
thylol)propionate (MW: 174 g/mol), which was subsequently
reacted with ethyl chloroformate and TEA to give MAC (70%)
and purified by recrystallization. Figure 1(B) shows the 1H
NMR spectrum (CDCl3) of MAC with peak assignments. The
signal d at 5.8 ppm and e at 5.23–5.42 ppm are characteristic
of the acryloyl protons in the monomer; while the signal b at
4.2 and 4.7 ppm denote methylene protons present in the car-
bonate ring. Other peak assignments include signals a and c at
1.38 and 4.72 ppm, assigned to methyl protons in the carbon-
ate ring and methylene protons next to the ester, respectively.
The structure of MAC was further confirmed using FTIR spec-
trometry [Fig. 1(C)]. Prominent absorbance peaks were
observed at 1651 and 1731 cmꢂ1 for C¼¼C and C¼¼O stretches,
respectively.
Synthesis and Characterization of
mPEG-b-PMAC-b-P(CB-co-LA) Copolymer
mPEG-b-PMAC-b-P(CB-co-LA) copolymer was obtained by
first reacting mPEG and MAC to give mPEG-b-PMAC copoly-
mer followed by the addition of CB and LA [Scheme 1(A)]. A
kinetic study was performed to determine the optimal time
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
351

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
SCHEME 1 Synthesis method for preparation of lactic acid- and carbonate-based crosslinked micelles. Conditions: (i) DBU, CH2Cl2,
RT, 3 h. (ii) AIBN, 60ꢁC. (A) Interface-crosslinked micelles. (B) Core-crosslinked micelles.
required before the addition of CB and LA monomer. From
Figure 2(A,B), a maximum MAC conversion of 80% was
achieved at a time of 105 min and maintained at 120 min
when DBU amount was 5 mL/mL. The observed conversion
saturation could possibly be due to lower MAC concentra-
tions. A plot of ln (1/(1 ꢂ x)) versus time resulted in a
straight line with a 70 minꢂ1 rate constant. Since MAC con-
version remained constant after 105 min, PEG-b-PMAC poly-
merization was carried for 2 h and characterized using 1H
NMR spectroscopy before the addition of CB and LA. A
typical 1H NMR spectrum for mPEG-b-PMAC copolymer is
displayed in Figure 2(C).
molecular weight was determined to be 10,558 g/mol from
1H NMR spectrum. GPC measurements also confirmed suc-
cessful copolymer synthesis and showed the resulting copol-
ymer to have a PDI of 1.19 and Mn of 9462 g/mol.
Synthesis and Characterization of
mPEG-b-P(CB-co-LA-co-MAC) Copolymer
The synthetic procedure for mPEG-b-P(CB-co-LA-co-MAC)
copolymer is shown in Scheme 1(B). mPEG-b-P(CB-co-LA-
co-MAC) copolymer was obtained by ring-opening polymer-
ization (ROP) of CB, MAC, and LA using PEO as a macroini-
tiator and DBU as a catalyst. The monomer to initiator
mole ratio was set at 20:1 with MAC monomer constituting
8 mol %. Addition of 5 mol % DBU resulted in greater
than 80% monomer conversion in 3 h affording copolymer
with controlled molecular weights and low polydispersity.
1H NMR spectrum [Fig. 3(B)] of mPEG-b-P(CB-co-LA-co-
MAC) copolymer showed the following resonance peaks at
d3.65, 5.12, 5.28–5.8, and 7.3 attributable to the methylene
protons of mPEG, the methylene protons of lactide, the
acryloyl protons of MAC, and the phenyl protons of carbon-
ate, respectively. Furthermore, the signals at 4.25–4.35
reflect the methylene protons in the carbonate main chain
and confirm successful ROP of the carbonate monomers.
The molecular weight of mPEG-b-P(CB-co-LA-co-MAC) co-
polymer was estimated to be 10,020 g/mol using the
The successful synthesis of mPEG-b-PMAC-b-P(CB-co-LA)
copolymer was confirmed using 1H NMR spectroscopy and
GPC. The presence of characteristic resonance peaks for
mPEG (3.65 ppm), LA (1.5–1.57, 5.12 ppm), CB (1.24, 4.25–
4.35, 5.2, 7.3 ppm), and MAC (4.2-4.4, 4.6, 5.3–5.8 ppm)
clearly confirmed the co-occurrence of all the monomers
[Fig. 3(A)]. Some characteristics of the synthesized mPEG-b-
PMAC-b-P(CB-co-LA) copolymer are summarized in Table 1.
Copolymer molecular weight was determined by comparative
analysis of the mPEG (i.e., CH2-CH2- at 3.65 ppm), lactide
(i.e., CH at 5.12), CB (i.e., C6H5 at 7.3 ppm), and MAC
(ACH2ACH¼¼CH2 at 5.8 ppm) characteristic proton peaks
from 1H NMR spectrum. mPEG-b-PMAC-b-P(CB-co-LA)
352
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 2 mPEG and MAC copolymerization in CH2Cl2 at room temperature. Reaction progress was monitored by 1H NMR spec-
troscopy. (A) Observed MAC conversion during synthesis of mPEG-b-PMAC copolymer. (B) Plot of ln (1/(1 ꢂ x)) versus time (x ¼
monomer conversion). (C) 1H NMR spectrum of mPEG-b-PMAC copolymer in CDCl3.
mPEG (d ¼ 3.65), LA (d ¼ 5.12), CB (d ¼ 7.3), and MAC
(d ¼ 5.8) characteristic peaks. GPC revealed mPEG-b-P(CB-
co-LA-co-MAC) copolymer to have a polydispersity of 1.22
and Mn of 9875 g/mol which was close to the theoretical
molecular weight and that determined by 1H NMR
spectroscopy.
FIGURE 3 1H NMR spectra for lactic acid- and carbonate-based copolymer for crosslinked micelles. (A) mPEG-b-PMAC-b-P(CB-co-
LA) copolymer in CDCl3. (B) mPEG-b-P(CB-co-LA-co-MAC) copolymer in CDCl3.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
353

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
TABLE 1 Characteristics of Lactic Acid- and Carbonate-Based Copolymers
Block Copolymer
Mn (1H NMRa)
Mn (GPCb)
Mw (GPCb)
Mw/Mn (GPCb)
CMC (g/l)
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
)
10,020
10,558
9,875
9,462
b
12,106
11,260
1.22
1.19
0.001
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39
)
0.0008
a
Molecular weight calculated from 1H NMR spectroscopy.
Determined using GPC (polystyrene standards).
Microstructure Analysis and Polymer Toxicity
arrangement, the 13C NMR spectra of mPEG-b-PMAC, mPEG-
b-PMAC-b-P(CB-co-LA), mPEG-b-P(CB-co-LA-co-MAC), and
mPEG-b-P(CB-co-LA) copolymers were analyzed by observing
peaks at 65.82 ppm (secondary carbon) and 46.48 ppm
(quaternary carbon), respectively.
MAC arrangement is expected to be blocky in mPEG-b-
PMAC-b-P(CB-co-LA) and random in mPEG-b-P(CB-co-LA-co-
MAC) copolymers. Since 13C NMR spectroscopy is sensitive
to chemical shifts of the quaternary carbon in the polymer
backbone and secondary carbon in the side group, it was
used to confirm MAC arrangement in both copolymers. The
13C NMR spectra of mPEG-b-PMAC-b-P(CB-co-LA) and mPEG-
b-P(CB-co-LA-co-MAC) copolymers are shown in Figure
4(A,B), respectively, and further confirm successful copoly-
mer synthesis. However, to determine MAC monomer
From Figure 5, a distinctive peak can be observed at 65.82
ppm for mPEG-b-PMAC copolymer synonymous with the sec-
ondary carbon on the allyl group. This characteristic peak at
65.82 ppm is also evident for mPEG-b-PMAC-b-P(CB-co-LA)
copolymer suggesting the presence of mPEG-b-PMAC block
in mPEG-b-PMAC-b-P(CB-co-LA) copolymer. The additional
FIGURE 4 13C NMR spectra for lactic acid- and carbonate-based copolymer for crosslinked micelles. (A) mPEG-b-PMAC-b-P(CB-co-
LA) copolymer in CDCl3. (B) mPEG-b-P(CB-co-LA-co-MAC) copolymer in CDCl3.
354
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 5 13C NMR spectra comparative plot for mPEG-b-PMAC, mPEG-b-PMAC-b-P(CB-co-LA), mPEG-b-P(CB-co-LA-co-MAC), and
mPEG-b-P(CB-co-LA) demonstrating differences in monomer arrangement.
peak shifted downfield is assigned as the same carbon of
MAC unit adjacent to P(CB-co-LA), which is reasonable since
PMAC is short (n ¼ 2.5). In contrast, the characteristic sec-
ondary carbon peak of MAC shifts downfield for mPEG-b-
P(CB-co-LA-co-MAC) copolymer, indicating lack of MAC block
sequence. This MAC peak is clearly absent in mPEG-b-P(CB-
co-LA) copolymer. Furthermore, the 46.48 ppm peak repre-
senting the quaternary carbon in mPEG-b-PMAC main chain
is seen in the spectra of mPEG-b-PMAC-b-P(CB-co-LA) copol-
ymer. The same peak is seen for mPEG-b-PMAC-b-P(CB-co-
LA) overlapped with the broad peak which is corresponding
to the quaternary carbon from the CB unit. Unlike mPEG-b-
PMAC and mPEG-b-PMAC-b-P(CB-co-LA), the 46.48 ppm
peak is shifted downfield for mPEG-b-P(CB-co-LA-co-MAC)
demonstrating random distribution of MAC monomer. A simi-
lar broad peak can be observed for mPEG-b-P(CB-co-LA)
reflecting contribution from the CB quaternary carbon. As
allyl compounds can be toxic, we evaluated the toxicity of
mPEG-b-PMAC-b-P(CB-co-LA) and mPEG-b-P(CB-co-LA-co-
MAC) copolymers in LNCaP prostate cancer cell lines. Our
data reveal both copolymers are not toxic up to 2000 mg/mL
(Fig. 6). In this study, the effect of micelles (NCM and CM)
on cancer cells was investigated to determine the influence
of these delivery vehicles on cancer cell proliferation. This
study confirmed that the therapeutic effect observed follow-
ing treatment with drug-loaded micelles was as a result of
the drug and not an artifact of the carrier.
0.001 g/L. In contrast, mPEG-b-PMAC-b-P(CB-co-LA) had a
CMC of 0.0008 g/L. These results when juxtaposed with our
previous findings12 suggests the presence of the intermedi-
ate MAC block results in copolymers that form more thermo-
dynamically stable micelles compared to random distribution
of MAC in the hydrophobic core.
Subsequently, micelle size and drug-loading content were
determined following preparation by the film sonication
method. We observed that the introduction of MAC monomer
did not affect the micelle forming ability or morphology of
the block copolymers regardless of its location. Micelle size
and morphology were examined by DLS and TEM, respec-
tively. Average hydrodynamic size was 123 6 3.02 nm for
Effect of Polymer Composition and MAC Location on
CMC, Particle Size, and Drug Loading
FIGURE 6 Cytotoxicity of mPEG-b-PMAC-b-P(CB-co-LA) and
mPEG-b-P(CB-co-LA-co-MAC) copolymers in prostate cancer
cells. LNCaP cells were incubated with various concentrations
of copolymers (0–2000 mg/mL) for 48 h and cell viability was
determined by MTT assay. Data are presented as mean 6 SD
(n ¼ 3).
Slight changes in polymer composition (e.g., monomer con-
tent and arrangement) can affect micelle properties. There-
fore, we examined the effect of polymer composition, MAC
location and crosslinking on CMC, size, and drug loading.
The CMC of mPEG-b-P(CB-co-LA-co-MAC) was found to be
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
355

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
TABLE 2 Effect of Polymer Composition and MAC Location on Particle Size and Drug Loading
Drug loading
(%) 6 SDb
Encapsulation
a
Block Copolymer
Non-crosslinked
Size (nm)c
PDI
Efficiency (%) 6 SD
Before drug loading
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
)
123.0 6 3.02
128.0 6 1.10
0.15 6 0.01
0.12 6 0.01
–
–
–
–
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39
)
After drug loading
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
)
139.0 6 1.05
133.0 6 1.49
0.10 6 0.02
0.04 6 0.01
3.79 6 0.56
1.01 6 0.73
75.72 6 0.15
20.29 6 0.11
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39
)
Crosslinked
Before drug loading
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39
After drug loading
)
151.0 6 1.69
135.0 6 1.89
0.12 6 0.04
0.20 6 0.03
–
–
–
–
)
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
)
118.0 6 1.70
122.0 6 0.81
0.10 6 0.01
0.12 6 0.01
3.21 6 2.11
0.89 6 0.10
64.08 6 0.48
17.77 6 0.02
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39
)
a
b
c
Subscripts reflect degree of polymerization of each monomer
Percentage of drug loaded into micelles based on 5% theoretical loading.
Mean particle size was determined by dynamic light scattering.
obtained from 1H NMR spectroscopy.
mPEG-b-P(CB-co-LA-co-MAC) copolymer while that of mPEG-
b-PMAC-b-P(CB-co-LA) was 128 6 1.10 nm (Table 2) with
both copolymers resulting in micelles possessing low PDI
values (0.12–0.15; Table 2) implying a narrow micelle size
distribution. Furthermore, our findings suggest inclusion of
drug resulted in a modest increase in micelle size. Specifi-
cally, micelle size was 139 6 1.05 nm for mPEG-b-P(CB-co-
LA-co-MAC) copolymer while that of mPEG-b-PMAC-b-P(CB-
co-LA) was 133 6 1.49 nm. However, drug-loaded micelles
still exhibited low PDI values.
a prominent peak at d5.8 ppm (acryloyl proton) can clearly
be seen in the NCM. In contrast, the signal significantly
weakened for the crosslinked sample (Fig. 7(B)]. Quantitative
1
analysis from the H NMR spectra revealed 69% crosslinking
efficiency. IR spectroscopy also confirmed successful cross-
linking as indicated by the weakening of the 1650 cmꢂ1
peak corresponding to the alkene group (Fig. 8).
From our studies, crosslinking of non-drug-loaded micelles
generally led to an increase in micelle size compared to their
non-crosslinked counterparts (Table 2). This size change was
greater for mPEG-b-P(CB-co-LA-co-MAC) which increased
from 123 6 3.02 nm to 151 6 1.69 nm compared to mPEG-b-
PMAC-b-P(CB-co-LA) where size was 128 6 1.10 nm and 135
6 1.89 nm before and after crosslinking, respectively. Inter-
estingly, we observed that crosslinking of drug-loaded
micelles resulted in a decrease in size compared to their non-
crosslinked counterparts (Table 2). Here again, the size
change was greater for mPEG-b-P(CB-co-LA-co-MAC) com-
pared to mPEG-b-PMAC-b-P(CB-co-LA) (21 nm vs. 11 nm).
Our findings also suggest that crosslinking resulted in a mod-
est decrease in drug loading regardless of copolymer type.
Furthermore, TEM analysis demonstrates that crosslinking
does not significantly alter micelle morphology (Fig. 9).
We next quantified the amount of drug loaded into the
micelles in terms of drug loading density (eq 1) and encap-
sulation efficiency (eq 2). From Table 2, mPEG-b-P(CB-co-LA-
co-MAC) copolymer had superior drug loading characteristics
compared to mPEG-b-PMAC-b-P(CB-co-LA). Based on a 5%
theoretical loading, mPEG-b-P(CB-co-LA-co-MAC) had a drug
loading density of 3.79% 6 0.56% indicating an encapsula-
tion efficiency of 78.72% 6 0.15%. In contrast, drug loading
density for mPEG-b-PMAC-b-P(CB-co-LA) was computed to be
1.01% 6 0.73% reflecting a 20.29% 6 0.11% encapsulation
efficiency. Furthermore, surface charge of the micelles pre-
pared from both copolymers were observed to be slightly neg-
ative: ꢂ0.96 6 0.05 and ꢂ5.52 6 0.13 mV for mPEG-b-P(CB-
co-LA-co-MAC) and mPEG-b-PMAC-b-P(CB-co-LA), respectively.
Effect of Crosslinking on Stability
Effect of Crosslinking on Particle Size, Morphology, and
Drug Loading
After confirming successful crosslinking of micelles, we next
examined the effect of crosslinking on micelle stability in
physiologically relevant and extreme micelle destabilization
conditions. Since micelles encounter sink-like conditions upon
systemic administration, we first examined the stability of
NCM and CM in PBS at micelle concentrations of 200, 20, 2,
and 0.20 mg/mL. Both NCM and CM remained intact up to 2
mg/mL micelle concentration further demonstrating the ther-
modynamic stability of both micelle systems. Under these
Next, mPEG-b-P(CB-co-LA-co-MAC) and mPEG-b-PMAC-b-
P(CB-co-LA) copolymers were suitably crosslinked in a mix-
ture of THF and water at 70ꢁC using AIBN as the initiator.
Pendant allyl moieties in the core and interface of copoly-
mers allowed convenient crosslinking in the presence of
AIBN initiator. 1H NMR and IR spectroscopy was used to
confirm the success of crosslinking. As shown in Figure 7(A),
356
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 7 1H NMR spectra for mPEG-b-P(MAC-co-CB-co-LA) copolymer in CDCl3. (A) Before crosslinking. (B) After crosslinking.
conditions, NCM had an average size of 123 6 3.02 nm with
0.15 6 0.01 PDI while CM exhibited an average size of 127 6
3.69 nm with 0.16 6 0.02 PDI. However, the NCM dissociated
when micelle concentration was 0.20 mg/mL, whereas the CM
maintained their nanostructure [Fig. 10(A)].
In Vitro Cytotoxicity of mPEG-b-PMAC-b-P(CB-co-LA) and
mPEG-b-P(CB-co-LA-co-MAC) NCM and CM
The inhibitory effect of bicalutamide-loaded mPEG-b-PMAC-
b-P(CB-co-LA) and mPEG-b-P(CB-co-LA-co-MAC) NCM and
CM was determined in LNCaP human prostate cancer cell
line for 24 h. From Figure 12(A), bicalutamide-loaded
We next investigated the stability of micelles with or without
drug in the presence of serum for 24 and 48 h. From Figure
10(B), NCM appeared to aggregate with time as indicated by
increase in particle size. For example, the di/d0 (diameter at
specified time to diameter at t ¼ 0) ratio increased from
approximately 1 to 4 and 8 at 24 and 48 h, respectively. We
also determined the CMC of mPEG-b-P(CB-co-LA-co-MAC)
NCM and CM as a secondary measure of stability (Support-
ing Information). Our results reveal NCM had a CMC value of
1 mg/L based on where the two tangent lines intersect. In
contrast, CMC of crosslinked micelles was determined to be
0.06 mg/L. These results demonstrate that CM enhanced
stability compared to NCM.
In Vitro Release Studies of Bicalutamide from
Polymeric Micelles
We next performed in vitro bicalutamide release study in PBS
(pH 7.4) at 37ꢁC and 150 rpm with 0.1% Tween-80 to deter-
mine the effect of crosslinking on drug release. Figure 11
shows the cumulative percentage of bicalutamide released
from mPEG-b-P(CB-co-LA-co-MAC) and mPEG-b-PMAC-b-P(CB-
co-LA) NCM and CM. Our results clearly reveal that the release
of bicalutamide from mPEG-b-P(CB-co-LA-co-MAC) CM to be
slower than their non-crosslinked counterpart. The cumulative
amount of bicalutamide released after 72 h was ꢀ76% for
NCM and 60% for CM. However, ꢀ78% of bicalutamide was
released from NCM compared to 68% from CM in the case of
mPEG-b-PMAC-b-P(CB-co-LA) copolymer.
FIGURE 8 FTIR spectra for lactic acid- and carbonate-based co-
polymer (A) before crosslinking and (B) after crosslinking.
Arrow indicates weakening of 1650 cmꢂ1 peak reflecting cross-
linking of double bonds. (C) Peaks at 1650 cmꢂ1 zoomed in and
offset for clarity; Top (before crosslinking) and Bottom (after
crosslinking).
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
357

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 9 TEM images of (A) NCM and (B) CM.
crosslinked micelles more potently inhibited LNCaP cell
growth compared to bicalutamide-loaded NCM regardless of
polymer type. For bicalutamide-loaded mPEG-b-PMAC-b-
P(CB-co-LA) micelles, IC50 for NCM was ꢀ47.2 mM while that
of CM was ꢀ15.0 mM. Similarly, bicalutamide-loaded mPEG-
b-P(CB-co-LA-co-MAC) NCM had an IC50 of ꢀ36.9 mM while
CM had IC50 of ꢀ15.4 mM. Bicalutamide is known to affect
expression of prostate-specific antigen (PSA), which is down-
stream of and regulated by the androgen–androgen receptor
signaling axis. Therefore, we examined the effect of
bicalutamide-loaded CM and NCM on secreted PSA following
incubation in LNCaP cells for 24 h. Our results suggests bica-
lutamide micelles reduce secreted PSA expression by approx-
imately 50% compared to control, irrespective of polymer
type and presence of crosslinking [Fig. 12(B)].
areas, one reason the pharmaceutical industry has not fully
adopted micelle formulations is their perceived instability
especially in biological environments. The purpose of our
study is to address this pertinent issue through the synthe-
sis, characterization, and in vitro evaluation of innovative
lactic acid- and carbonate-based CM for enhanced drug
delivery.
We have successfully synthesized mPEG114-b-PMAC2.5-b-
P(CB9-co-LA39
)
and mPEG114-b-P(CB8-co-LA35-co-MAC2.5)
copolymers for core-corona interface-crosslinked and core-
crosslinked micelles, respectively (Scheme 1). The design of
these copolymers was guided by our desire to improve upon
the existing micelle properties obtained using our previously
synthesized mPEG-b-P(CB-co-LA) copolymers, especially with
regards to sub-CMC and serum stability. Therefore, we
decided to strategically introduce chemical crosslinks at the
core-corona interface with the goal of creating a molecular
fence to improve micelle integrity and sustained release. We
also introduced crosslinks in the core as a complimentary
approach and to demonstrate the flexibility of our crosslink-
ing approach. Although a variety of chemistries (e.g., click
chemistry, disulfide bond) can be used for crosslinking, we
judiciously selected the MAC monomer as our crosslinking
DISCUSSION
Polymeric micelles are promising drug delivery plat-
forms.9,12,25,26,28–30 Insight accruing from micelle research
particularly in oncology has led to the strategic design and
synthesis of novel polymers to address the key issues of
adequate drug loading, targetability, sustained drug release
and degradation. Despite significant advances made in these
FIGURE 10 Stability of mPEG-b-PMAC-b-P(CB-co-LA) NCM and CM. (A) Micelle stability against 1000-fold dilution. (B) Micelle sta-
bility in BSA after 24 and 48 h.
358
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
reactivity of the cyclic carbonate monomers. The DBU-cata-
lyzed conversion of our cyclic carbonates is higher than that
of trimethyl carbonate (80% vs. 70%) reported by Watanabe
et al.;33 however, they used 4 -(chloromethyl)benzyl alcohol
as an initiator. Our kinetic study suggests MAC polymeriza-
tion with mPEG to be first order as demonstrated by the lin-
ear nature of the ln(1/(1 ꢂ x)) versus time plot [Fig. 2(B)].
This result is not surprising as DBU and mPEG concentra-
tions do not change during polymerization. Qian et al.34 have
reported similar first-order dependency with respect to
monomer, DBU, and macroinitiator concentrations for ROP of
LA and glycolide.34 Furthermore, ROP of valerolactone using
similar amidine catalysts have been shown to follow first-
order kinetics with respect to catalyst, alcohol, initiator and
monomer concentration.35 Although we did not systemati-
cally study the polymerization kinetics upon addition of CB
and LA, a time of 3 h was sufficient to obtain polymers with
FIGURE 11 Effect of crosslinking on bicalutamide release from
mPEG-b-PMAC-b-P(CB-co-LA)
and
mPEG-b-P(CB-co-LA-co-
MAC) micelles. Release experiments were performed in tripli-
cate in PBS with 0.1% Tween 80 at 37ꢁC and 150 rpm.
targeted mass ratios of MAC, CB, and LA for both mPEG114
-
-
b-PMAC2.5-b-P(CB9-co-LA39
) and mPEG114-b-P(CB8-co-LA35
moiety as we did not want to dramatically alter the superior
attributes of mPEG-b-P(CB-co-LA) copolymer (e.g., drug load-
ing, thermodynamic stability, micelle architecture, biodegrad-
ability). MAC has close structural similarity to our well-stud-
ied cyclic carbonate (CB) monomer except it has an allyl
pendant group in contrast to the phenyl pendant of CB.
Therefore, we were confident that the introduction of MAC
will not significantly change the backbone or hydrophilic lip-
ophilic balance of the polymer. We synthesized MAC mono-
mer using two approaches (Fig. 1) and used 1H NMR and
FTIR spectrometry to confirm its structure. No FTIR data
was found in the literature, however, our 1H NMR spectrum
matched the literature.20 While the first approach had a
shorter overall reaction time, the second approach was pre-
ferred as it resulted in higher MAC yields (70% vs. 30%)
and had fewer steps (2 vs. 4). Our MAC yield of 70% was
comparable to that reported by Hu et al.20 Furthermore,
unlike the first approach which required column chromatog-
raphy, pure compounds could be obtained with recrystalliza-
tion in the second approach making it less labor intensive.
Recently, Chen et al.31 also synthesized a couple of cyclic car-
bonate monomers similar to MAC. However, their synthetic
scheme required four steps with overall yields of 40% which
is lower than our preferred second approach.
co-MAC2.5) copolymers (inferred from 1H NMR spectra).
mPEG-b-P(CB-co-LA) copolymers previously synthesized
using Sn(Oct)2 catalyzed ROP exhibited relatively high PDIs
(around 1.4) possibly due to undesired transesterification.12
Furthermore, increasing CB content affected LA reactivity.
Therefore, we used the organic base catalyst DBU (pKa
¼
24.3)32 for synthesizing mPEG114-b-PMAC2.5-b-P(CB9-co-
LA39) and mPEG114-b-P(CB8-co-LA35-co-MAC2.5) copolymers.
DBU is known to mitigate unwanted transesterification
resulting in polymers with well-controlled molecular weights
and low PDIs. In synthesizing mPEG114-b-PMAC2.5-b-P(CB9-
co-LA39), we first sought to understand how the reactivity of
MAC compares with CB using mPEG as a macroinitiator. Our
findings suggest the reactivity of MAC (80% conversion in 2
h (Fig. 2)) to be similar to that of CB (data not shown) indi-
cating the pendant group does not significantly impact the
FIGURE 12 Anticancer effect of mPEG-b-PMAC-b-P(CB-co-LA)
and mPEG-b-P(CB-co-LA-co-MAC) NCM and CM. (A) 3 ꢃ 103
LNCaP cells were incubated with bicalutamide-loaded micelle
formulations (0, 25, and 50 mM of drug) for 24 h. Cell viability
was determined using MTT assay and results represented as
the mean 6 SD of triplicates. *p < 0.05; **p < 0.01 using Stu-
dent’s unpaired t test. (B) Effect of bicalutamide-loaded mPEG-
b-PMAC-b-P(CB-co-LA) and mPEG-b-P(CB-co-LA-co-MAC) NCM
and CM (50 mM of drug) on secreted PSA following 24 h
incubation.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
359

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
Furthermore, DBU-catalyzed polymerization allowed us to
obtain polymers in a relatively short time and avoided MAC
crosslinking observed during initial polymerization studies
attempted using Sn(Oct)2 at 120ꢁC.
PMAC2.5-b-P(CB9-co-LA39) micelles had a CMC of 0.0008 g/L
while previously reported CMC of mPEG114-b-P(CB9-co-LA5)
micelles was 0.004 g/L. As CB content is the same in both
copolymers, this fivefold improvement in CMC value is due
to the presence of MAC and LA. However, CMC depends on
both composition and molecular mass of the hydrophobic
core.36,37 Therefore, it is difficult to clearly ascertain the indi-
vidual contribution of MAC due to considerable difference in
overall core length between mPEG114-b-PMAC2.5-b-P(CB9-co-
LA39) (Mn ¼ 9992 g/mol) and mPEG114-b-P(CB9-co-LA5)
Monomer arrangement in a copolymer significantly impacts
its physicochemical properties and consequently key micelle
properties. Consequently, we performed microstructure anal-
ysis on mPEG114-b-PMAC2.5-b-P(CB9-co-LA39) and mPEG114
-
b-P(CB8-co-LA35-co-MAC2.5 copolymers to confirm MAC
)
arrangement in the copolymers. By systematically observing
the shift in 65.82 ppm (secondary carbon) and 46.48 ppm
(quaternary carbon) peaks in the 13C NMR spectra of mPEG-
b-PMAC, mPEG-b-PMAC-b-P(CB-co-LA), mPEG-b-P(CB-co-LA-
co-MAC), and mPEG-b-P(CB-co-LA) copolymers, we were able
to show MAC arrangement to be block in mPEG-b-PMAC-b-
P(CB-co-LA) and random in mPEG-b-P(CB-co-LA-co-MAC).
Typically, reported microstructure analyses of ester-based
copolymers use the carbonyl group in the polymer main
chain. However, the close similarity between the contribution
of MAC and CB to copolymer backbone complicates such
analysis, and it is not feasible to use carbonyl groups in the
polymer backbone for microstructure analysis in this study.
(Mn ¼ 7510 g/mol). Additionally, mPEG114-b-P(CB8-co-LA35
-
co-MAC2.5) micelles had a CMC of 0.001 g/L compared to
0.002 g/L for mPEG114-b-P(CB8-co-LA24) micelles reported
previously. Here, the modest improvement is most likely due
to the presence of MAC since CB content is the same and the
LA amounts are relatively close. Our results demonstrate
that the CMC of mPEG114-b-PMAC2.5-b-P(CB9-co-LA39) and
mPEG114-b-P(CB8-co-LA35-co-MAC2.5) micelles is sensitive to
the presence of MAC but more importantly to its location
and arrangement: block versus random distribution among
CB and LA.
Crosslinking based on double bonds requires a radical initia-
tor and exposure to heat or UV light. Our copolymers were
successfully crosslinked using AIBN as radical initiator at a
temperature of 70ꢁC for 24 h. We observed crosslinking effi-
ciency of approximately 70% which was slightly lower than
the 80% value reported in the literature.23 Complete cross-
linking may not be possible as all the allyl moieties present
will have to be in close proximity for this to occur. However,
it is unlikely that micelle architecture would permit all the
double bonds to be within the effective distance required for
crosslinking. Theoretically, a few rightly positioned crosslinks
would be sufficient to improve micelle mechanical integrity
and 10 % crosslinking efficiency is not required. We found
that crosslinking of blank micelles caused a modest increase
in size. This we believe may be an artifact of the crosslinking
process as THF used in the process is known to cause
micelles to swell. Unexpectedly, crosslinking of drug-loaded
micelles resulted in a decrease in micelle size regardless of
MAC location. This was accompanied by a small loss in drug
loading and encapsulation efficiency post-crosslinking. The
reduction in drug loading and encapsulation efficiency may
be due to leakage during the crosslinking process and may
account for the modest size decrease observed after cross-
linking. Finally, it is worth mentioning that while the CM
possess better mechanical integrity than the non-crosslinked
counterparts they are designed ultimately to be biodegrad-
able. The polymer main chain consists of ester linkage and
carbonate linkage, which degrades enzymatically regardless
of side-chain crosslinking. The number of crosslinkable
monomers in these polymers is limited to 2–3 units, thus we
do not expect any long polymerization to occur by crosslink-
ing reaction in the micelle form. However, further studies
investigating the actual structure of the crosslinked micelles,
biodegradability, and their safety are required.
Although inclusion of MAC monomer added a layer of com-
plexity in terms of microstructure analysis, our results sug-
gest that it does not detrimentally affect micelle forming
ability or key properties. Both mPEG114-b-PMAC2.5-b-P(CB9-
co-LA39) and mPEG114-b-P(CB8-co-LA35-co-MAC2.5) micelles
exhibited similar size (with or without drug), which was
slightly larger than mPEG-b-P(CB-co-LA) micelles prepared
in our previous study.12 As micelle size is dependent on core
composition and length (molecular weight), it appears this
increase may be due to the contribution of the MAC mono-
mer and relatively larger quantities of LA present in
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39) and mPEG114-b-P(CB8-
co-LA35-co-MAC2.5) copolymers compared to mPEG114-b-
P(CB8-co-LA24 and mPEG114-b-P(CB9-co-LA5) copolymers.
)
Our results are consistent with the fact that hydrophobic
blocks with higher molecular weight have generally been
shown to have larger sizes. Notwithstanding, drug loading
and encapsulation efficiency in crosslinked micelles were
comparable to mPEG114-b-P(CB8-co-LA24) and mPEG114-b-
P(CB9-co-LA5) micelles possessing similar CB content. The
difference in bicalutamide loading between mPEG114-b-
PMAC2.5-b-P(CB9-co-LA39) and mPEG114-b-P(CB8-co-LA35-co-
MAC2.5) micelles may be attributed to CB content. Our previ-
ous studies clearly showed that CB degree of polymerization
(DP) of nine resulted in a precipitous decrease in bicaluta-
mide loading compared to DP of 8.12 Furthermore, our results
provide indirect evidence supporting our hypothesis that the
structure and quantities of MAC present in the copolymer do
not significantly alter drug loading characteristics.
Interestingly,
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39
)
and
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
)
micelles exhibited
CMCs better than or comparable to mPEG-b-P(CB-co-LA)
micelles with similar CB content suggesting equivalent if not
better thermodynamic stability. Specifically, mPEG114-b-
Polymeric micelles are dynamic in nature and gradually dis-
integrate under sub-CMC conditions. Therefore, monitoring
360
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
changes in micelle size (using DLS) at different dilutions is
an elegant way to determine micelle stability. Our results
reveal that CM remain intact even below the CMC [Fig. 8(A)]
while NCM disintegrated. CMC for CM was found to be
20-fold lower compared to NCM confirming the fact that CM
were several times more stable than their non-crosslinked
counterparts. It is worth mentioning that, although CMC of
CM has little practical meaning it still provides a reasonable
measure of the extent in improvement in micelle mechanical
integrity. Micelles have also been shown to become unstable
once they encounter blood components.11,36,38 Instability
may result from protein adsorption, protein penetration, or
drug extraction.36 As the most abundant protein in blood
plasma is serum albumin, we investigated the effect of physi-
ological simulating concentrations of BSA (45 mg/mL) on
Cm and NCM stability using DLS to observe time-dependent
changes in micelle size. Our results showed significant
increase in size of NCM with time and reflect aggregation of
micelles in BSA. This phenomenon may be due to protein
adsorption which is undesirable as it can result in fast clear-
ance by the mononuclear phagocyte system.39 It has been
shown that BSA–micelle interaction is typically driven by
hydrophobic aggregation. For instance, micelles with high
density hydrophilic corona experience greater steric stabili-
zation and encounter less BSA interaction.40 Other instances
of polymeric micelle aggregation and interaction with BSA
have been reported in the literature.40,41 In contrast, there
was no appreciable increase in the size of CM following incu-
bation with BSA. One reason may be the less dynamic nature
of CM, which limits the continuous interchange of unimers
thereby reducing the possibility of BSA interacting with the
hydrophobic core. It is also likely that the presence of cova-
lent crosslinks make it difficult for BSA to disrupt micelle
architecture thus preventing BSA–micelle aggregates.
dilution and physiological simulating serum (BSA ꢀ45 mg/
mL). CM size remained unchanged while NCM disintegrated
at a thousand fold dilution. Besides, CM size remained
unchanged in BSA whereas there was a time-dependent
increase in the size of NCM. These results showed CM to be
more stable compared to their non-crosslinked counterparts.
Additionally, bicalutamide-loaded CM were found to be more
potent in inhibiting proliferation of LNCaP prostate cancer
cells compared to NCM regardless of polymer type. In all, we
have demonstrated that these new biodegradable copolymer
systems are potentially useful for cancer therapy. Future
studies on influence of MAC block length, extent of crosslink-
ing efficiency, and ratio of MAC to CB and LA on key micelle
properties are required to generate material design rules
which can be used for customized fabrication of improved
micelle delivery platforms.
ACKNOWLEDGMENT
This work is supported by an Idea Award (W81XWH-10-1-
0969) from the Department of Defense Prostate Cancer
Research Program.
REFERENCES AND NOTES
1 Kwon, G. S. Crit. Rev. Ther. Drug Carrier Syst. 2003, 20,
357–403.
2 Soga, O.; van Nostrum, C. F.; Fens, M.; Rijcken, C. J.; Schiff-
elers, R. M.; Storm, G.; Hennink, W. E. J Control Release 2005,
103, 341–353.
3 Torchilin, V. P. Pharm. Res. 2007, 24, 1–16.
4 Solink, S.; Illum, L.; Davus, S. S. Adv. Drug Deliv. Rev. 1995,
16, 195–214.
5 Iyer, A. K.; Khaled, G.; Fang, J.; Maeda, H. Drug Discov.
Today 2006, 11, 812–818.
6 Maeda, H. Adv. Enzyme Regul. 2001, 41, 189–207.
7 Matsumura, Y.; Maeda, H. Cancer Res. 1986, 46 (12 Pt 1),
CONCLUSIONS
6387–6392.
To address the current need for biodegradable, biocompati-
ble, and stable polymeric micelle delivery systems, we report
on the synthesis, characterization, and evaluation of lactic-
acid- and carbonate-based copolymers containing an allyl
moiety for preparing interface and core crosslinked micelles.
A series of complimentary spectroscopic techniques was
used to confirm successful synthesis by ROP, elucidate poly-
mer microstructure, and confirm crosslinking. mPEG114-b-
PMAC2.5-b-P(CB9-co-LA39) and mPEG114-b-P(CB8-co-LA35-co-
MAC2.5) copolymers for interface and core crosslinking,
respectively, exhibited relatively low PDI (1.19–1.22) and
CMC values (0.8–1 mg/L). Inclusion of MAC monomer did
not significantly alter micelle size or morphology irrespective
of polymer composition or MAC location. Importantly,
8 Bronich, T. K.; Keifer, P. A.; Shlyakhtenko, L. S.; Kabanov, A.
V. J. Am. Chem. Soc. 2005, 127, 8236–8237.
9 Lavasanifar, A.; Samuel, J.; Kwon, G. S. Adv. Drug Deliv.
Rev. 2002, 54, 169–190.
10 Liu, J.; Zeng, F.; Allen, C. J. Control Release 2005, 103,
481–497.
11 Savic, R.; Azzam, T.; Eisenberg, A.; Maysinger, D. Langmuir
2006, 22, 3570–3578.
12 Danquah, M.; Fujiwara, T.; Mahato, R. I. Biomaterials 2010,
31, 2358–2370.
13 Rosler, A.; Vandermeulen, G. W.; Klok, H. A. Adv. Drug
Deliv. Rev. 2001, 53, 95–108.
14 Joralemon, M. J.; Murthy, K. S.; Remsen, E. E.; Becker, M.
L.; Wooley, K. L. Biomacromolecules 2004, 5, 903–913.
mPEG114-b-P(CB8-co-LA35-co-MAC2.5
encapsulation efficiency (78.72% 6 0.15%) compared to
mPEG114-b-PMAC2.5-b-P(CB9-co-LA39 (20.29% 0.11%).
) had a higher drug
15 O’Reilly, R. K.; Hawker, C. J.; Wooley, K. L. Chem. Soc. Rev.
2006, 35, 1068–1083.
)
6
16 Read, E. S.; Armes, S. P. Chem. Commun. (Camb) 2007,
3021–3035.
Subsequent micelle crosslinking did not dramatically change
drug loading or micelle morphology. However, there was a
strong link between micelle crosslinking and improved sta-
bility when incubated in aqueous media under extensive
17 Hawker, C. J.; Wooley, K. L. Science 2005, 309, 1200–1205.
18 Iijima, M.; Nagasaki, Y.; Okada, T.; Kato, M.; Kataoka, K.
Macromolecules 1999, 32, 1140–1146.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362
361

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
19 Kim, J.-H.; Emoto, K.; Iijima, M.; Nagasaki, Y.; Aoyagi, T.;
Okano, T.; Sakurai, Y.; Kataoka, K. Polym. Adv. Technol. 1999,
10, 647–654.
30 Li, F.; Danquah, M.; Singh, S.; Wu, H.; Mahato, R. I. Drug
Deliv. Transl. Res. 2011, 1, 420–428.
31 Chen, W.; Yang, H.; Wang, R.; Cheng, R.; Meng, F.; Wei, W.;
Zhong, Z. Macromolecules 2010, 43, 201–207.
20 Hu, X.; Chen, X.; Xie, Z.; Liu, S.; Jing, X. J. Polym. Sci.: Part
A: Polym. Chem. 2007, 45, 5518–5528.
32 Kaljurand, I.; Kutt, A.; Soovali, L.; Rodima, T.; Maemets, V.;
Leito, I.; Koppel, I. A. J. Org. Chem. 2005, 70, 1019–1028.
21 Garg, S. M.; Xiong, X.-B.; Lu, C.; Lavasanifar, A. Macromole-
cules 2011, 44, 2058–2066.
33 Watanabe, J.; Amemori, S.; Akashi, M. Polymer 2008, 49,
3709–3715.
22 Butun, V.; Wang, X.-S.; de Paz Banez, M. V.; Robinson, K. L.;
Billingham, N. C.; Armes, S. P.; Tuzar, Z. Macromolecules 1999,
33, 1–3.
34 Qian, H.; Wohl, A. R.; Crow, J. T.; Macosko, C. W.; Hoye, T.
R. Macromolecules 2011, 44, 7132–7140.
23 Hu, X.; Chen, X.; Wei, J.; Liu, S.; Jing, X. Macromol. Biosci.
2009, 9, 456–463.
35 Lohmeijer, B. G. G.; Pratt, R. C.; Leibfarth, F.; Logan, J. W.;
Long, D. A.; Dove, A. P.; Nederberg, F.; Choi, J.; Wade, C.;
Waymouth, R. M.; Hedrick, J. L. Macromolecules 2006, 39,
8574–8583.
24 Yang, R.; Meng, F.; Ma, S.; Huang, F.; Liu, H.; Zhong, Z. Bio-
macromolecules 2011, 12, 3047–55.
36 Kim, S.; Shi, Y.; Kim, J. Y.; Park, K.; Cheng, J. X. Expert
Opin. Drug Deliv. 2010, 7, 49–62.
25 Li, F.; Danquah, M.; Mahato, R. I. Biomacromolecules 2010,
11, 2610–2620.
37 Attwood, D.; Booth, C.; Yeates, S. G.; Chaibundit, C.;
Ricardo, N. M. Int. J. Pharm. 2007, 345, 35–41.
26 Danquah, M.; Li, F.; Duke, C. B., 3rd; Miller, D. D.; Mahato,
R. I. Pharm. Res. 2009, 26, 2081–2092.
38 Chen, H.; Kim, S.; He, W.; Wang, H.; Low, P. S.; Park, K.;
Cheng, J. X. Langmuir 2008, 24, 5213–7.
27 Pratt, R. C.; Nederberg, F.; Waymouth, R. M.; Hedrick, J. L.
Chem. Commun. (Camb) 2008, 114–116.
39 Alexis, F.; Pridgen, E.; Molnar, L. K.; Farokhzad, O. C. Mol.
Pharm. 2008, 5, 505–515.
28 Danquah, M.; Duke, C. B., 3rd; Patil, R.; Miller, D. D.;
Mahato, R. I., Combination Therapy of Antiandrogen and XIAP
Inhibitor for Treating Advanced Prostate Cancer. Pharm. Res.
2012, 29, 2079–2091.
40 Li, W.; Nakayama, M.; Akimoto, J.; Okano, T. Polymer 2011,
52, 3783–3790.
29 Danquah, M. K.; Zhang, X. A.; Mahato, R. I. Adv. Drug Deliv.
Rev. 2011, 63, 623–639.
41 Yang, C.; Attia, A. B.; Tan, J. P.; Ke, X.; Gao, S.; Hedrick,
J. L.; Yang, Y. Y. Biomaterials 2012, 33, 2971–2979.
362
JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY 2013, 51, 347–362