Synthesis and characterization of PEG-based ether-anhydride terpolymers: Novel polymers for controlled drug delivery

Article abstract of DOI:10.1021/ma049853s

A series of biodegradable poly(ether-anhydrides) composed of polyethylene glycol) (PEG), sebacic acid (SA), and 1,3-bis(carboxyphenoxy)propane (CPP) were synthesized for use in advanced drug delivery applications. PEG (M_n = 8000 Da) was incorporated to reduce polymeric particle clearance rates by the immune system and improve particle resuspension and aerosolization efficiencies. CPP and SA were selected to render the polymer insoluble in water and allow control over polymer degradation and drug release rates. In particular, CPP incorporation caused a significant decrease in polymer degradation rates and release kinetics of model drugs incorporated into poly(ether-anhydride) microparticles. Terpolymers were synthesized with weight-average molecular weights over 65 kDa without catalyst. The first thermal transition in polymers containing ≤ 10 wt % PEG was ～80°C (well above typical storage conditions and body temperature), and there was no evidence of a glass transition (-100 to 200°C). Several of the polymers were used to produce particles suitable for injection or inhalation; these particles released model drugs, with molecular weights ranging from 443 to 5 143 000 Da, in a continuous fashion for up to 7 days.

Full text of DOI:10.1021/ma049853s

7
174
Macromolecules 2004, 37, 7174-7180
Synthesis and Characterization of PEG-Based Ether-Anhydride
Terpolymers: Novel Polymers for Controlled Drug Delivery
J ie F u , J en n ifer F iegel, a n d J u stin Ha n es*
Department of Chemical and Biomolecular Engineering, The J ohns Hopkins University,
Baltimore, Maryland 21218
Received J anuary 21, 2004; Revised Manuscript Received J uly 1, 2004
ABSTRACT: A series of biodegradable poly(ether-anhydrides) composed of poly(ethylene glycol) (PEG),
sebacic acid (SA), and 1,3-bis(carboxyphenoxy)propane (CPP) were synthesized for use in advanced drug
delivery applications. PEG (M ) 8000 Da) was incorporated to reduce polymeric particle clearance rates
n
by the immune system and improve particle resuspension and aerosolization efficiencies. CPP and SA
were selected to render the polymer insoluble in water and allow control over polymer degradation and
drug release rates. In particular, CPP incorporation caused a significant decrease in polymer degradation
rates and release kinetics of model drugs incorporated into poly(ether-anhydride) microparticles.
Terpolymers were synthesized with weight-average molecular weights over 65 kDa without catalyst. The
first thermal transition in polymers containing e10 wt % PEG was ∼80 °C (well above typical storage
conditions and body temperature), and there was no evidence of a glass transition (-100 to 200 °C).
Several of the polymers were used to produce particles suitable for injection or inhalation; these particles
released model drugs, with molecular weights ranging from 443 to 5 143 000 Da, in a continuous fashion
for up to 7 days.
In tr od u ction
many applications, especially those that require injec-
tion of polymer drug carriers into the blood or, alter-
natively, their inhalation into the lungs. A second
limitation is that PLGA devices undergo bulk erosion,
which leads to a variety of undesirable outcomes,
including exposure of unreleased drug to a highly acidic
Biodegradable polymers have been used for many
applications in medicine, including controlled release
1
-3
drug delivery systems,
resorbable bone pins and
screws,4 and scaffolds for cells in tissue engineering.
-6
7,8
Systems based on biodegradable polymers obviate the
need for surgical removal since their degradation prod-
ucts are absorbed or metabolized by the body. Micron-
sized systems made using polymers can be used to
deliver precise amounts of drugs, including proteins and
2
0
environment. Third, it is difficult to release drugs in
a continuous manner from PLGA particles owing to the
polymers’ bulk-erosion mechanism. Instead, special
1
9,21,22
preparation methods
are required with PLGA to
avoid the typical intermittent drug release pattern (i.e.,
burst of drug followed by a period of little or no drug
release and then by the onset of a second phase of
genes, over prolonged periods to local tissues or the
systemic circulation.9,10 Of particular interest is the
development of drug delivery vehicles that exhibit
reduced detection rates by the immune system (e.g.,
long-circulating carriers for intravenous administra-
2
3,24
significant drug release).
Fourth, the particularly
fine PLGA particles needed for intravenous injection or
inhalation can agglomerate significantly, making re-
suspension for injection or aerosolization for inhalation
1
1-13
tion
) or that can be administered via noninvasive
1
4-17
delivery routes (such as inhalation
polymers that safely erode in the body, preferably at a
). Biodegradable
1
6,25
difficult.
Finally, small, insoluble particles with
hydrophobic surfaces, like those made with PLGA, are
rapidly removed and destroyed by the immune system
rate that closely coincides with the rate of drug deliv-
ery,1
0,18
are required for these advanced applications.
2
6-28
(due to fast opsonization).
Despite their wide and growing need in medicine, only
two synthetic biodegradable polymers are currently used
routinely in humans as carriers for drug delivery: ester
copolymers of lactide and glycolide (PLGA family) and
anhydride copolymers of sebacic acid (SA) and 1,3-bis-
Implants composed of poly(CPP:SA) were approved
for use in humans in the late 1990s to deliver chemo-
therapeutic molecules directly at the site of a resected
2
9,30
brain tumor.
CPP:SA copolymers erode from the
3
1
surface in (called surface erosion), leading to desirable
steady drug delivery rates over time. Proven biocom-
patibility, current clinical use, and steady drug release
profiles make polymers composed of CPP and SA good
candidates for new drug delivery applications. However,
like PLGA particles, small particles made with poly-
(
carboxyphenoxy)propane (CPP). PLGA is the most
widely used due to its history of safe use as surgical
sutures and in current drug delivery products like the
1
9
Lupron Depot. While the development of PLGA re-
mains among the most important advances in medical
biomaterials, there are some limitations that signifi-
cantly curtail its use. First, PLGA particles typically
take a few weeks to several months to completely
degrade in the body, but the device is typically depleted
of drug more rapidly.16 Repeated dosing of such a system
leads to an unwanted buildup of drug-depleted polymer
in the body. This may preclude the use of PLGA for
(
CPP:SA) possess hydrophobic surfaces that lead to
rapid removal by the immune system and poor resus-
pension and aerosolization properties.
3
2
In this paper, we describe the synthesis of a new
family of terpolymers for advanced drug delivery ap-
plications that may overcome some important limita-
tions of existing synthetic biodegradable polymers. The
polymers are composed of three monomers, poly-
(ethylene glycol) (PEG), SA, and CPP, polymerized in
*
To whom correspondence should be addressed: Ph (410) 516-
3
484; Fax (410) 516-5510; e-mail hanes@jhu.edu.
1
0.1021/ma049853s CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/25/2004

Macromolecules, Vol. 37, No. 19, 2004
Poly(ether-anhydrides) for Drug Delivery 7175
various ratios to develop a series of PEG:SA:CPP poly-
using double-sided graphite carbon tape and sputter-coated
with gold-palladium using a Hummer VI sputtering system
(
ether-anhydrides). These polymers are similar to the
(Bethesda, MD). Populations representative of each micro-
FDA-approved poly(CPP:SA) polymers currently in use,
except they contain PEG built into their backbone
structure. Each monomer is currently used in humans,
but they have not yet been polymerized together to
produce a useful material.
sphere sample were photographed.
Stability studies were performed in solid state and in
anhydrous chloroform at 25, 5, -20, and -80 °C under P
2 5
O .
Polymer molecular weight was followed by GPC with time.
Syn th esis of P olyoxyeth ylen e Dicar boxylic Acid Mon o-
m er . Hydroxyl-terminated PEG (40.0 g) was dissolved in
chloroform (300 mL). Succinic anhydride (5.0 g) and pyridine
(5 mL) were added to the chloroform solution, and the mixture
was reacted at 60 °C for 72 h. The solution was cooled, filtered,
and concentrated to dryness by rotary evaporation. The crude
product was dissolved in 30 mL of 1 N HCl, washed with
diethyl ether, extracted with chloroform, and dried with
We previously reported on the synthesis of PEG:SA
copolymers that utilized PEG with a lower molecular
1
5
weight (Mn ) 600 Da) and that did not include CPP.
In the current paper, we show that larger molecular
weight PEG (8000 Da) chains can be incorporated into
high molecular weight poly(ether-anhydrides) under
optimized synthetic conditions. This is critical since PEG
with molecular weights in the range of 2-20 kDa are
anhydrous sodium sulfate. Excess solvents were removed
1
under vacuum. H NMR (CDCl
3
): δ 3.65 (s, OCH
OCH
P r ep a r a tion of Acyla ted P r ep olym er s. Polyoxyethylene
dicarboxylic acid (10.0 g) was refluxed in 200 mL of acetic
anhydride for 30 min under N and evaporated to dryness by
rotary evaporation. The residue was extracted with anhydrous
2
CH
2
), 2.48
most effective at protecting particles from removal by
-1
(t, CH
2
). IR (KBr, cm ): 1735 (CdO), 1110 (CH
2
2
).
the immune system.2
6-28
Therefore, the successful
incorporation of PEG8000 into the backbone in high
amounts is expected to significantly improve the per-
formance of fine nano- and microparticles designed for
more advanced drug delivery applications. We also
incorporate a third monomer component into the poly-
mer backbone in this study, the hydrophobic monomer
2
1
ether and dried under vacuum. H NMR (CDCl
OCH CH ), 2.32 (s, CH ), 2.47 (t, CH ). IR (KBr, cm ): 1807,
743 (CdO anhydride), 1110 (CH OCH ).
Sebacic acid (10.0 g) was refluxed in 100 mL of acetic
anhydride under N for 15 min and evaporated to dryness by
3
): δ 3.64 (s,
-
1
2
2
3
2
1
2
2
1
,3-bis(carboxyphenoxy)propane (CPP), which signifi-
cantly improves the range of properties of the poly-
ether-anhydride) family for drug and gene delivery
2
rotary evaporation. The crude prepolymer was recrystallized
from dried toluene, washed with anhydrous ethyl ether/
(
applications. We use the materials to produce drug-
loaded microparticulates capable of injection via small
needles or aerosolization as a dry powder, and we
demonstrate the utility of this new terpolymer family
as controlled release vehicles for drug molecules ranging
1
petroleum ether (1:1), and finally dried under vacuum.
H
NMR (CDCl ): δ 2.45 (t, 4H, CH ), 2.33 (s, 6H, CH ), 1.66 (m,
3
2
3
-
1
2
4H, CH ), 1.33 (m, 8H, CH2). IR (KBr, cm ): 1813, 1742 (Cd
O anhydride).
CPP (10.0 g) was refluxed in 200 mL of acetic anhydride
for 30 min under N , followed by removal of the unreacted
2
6
in molecular weight from 443 to over 5 × 10 Da.
diacid by filtration and solvent by evaporation. The residue
was recrystallized from dimethylformamide and ethyl ether,
then washed with dry ethyl ether, and dried under vacuum.
Exp er im en ta l Section
1
Ma ter ia ls. Sebacic acid (Sigma) was recrystallized three
times from ethanol. Acetic anhydride (Aldrich) was purified
by distillation. Toluene (J .T. Baker) and chloroform (Aldrich)
were refluxed over and distilled from calcium hydride (Sigma).
H NMR (CDCl ): δ 7.14, 7.99 (d, 4H, ArH), 4.29 (t, 4H, CH
3
2
-1
), 2.38 (s, 6H, CH ), 2.25 (m, 2H, CH ). IR (KBr, cm ): ∼1798,
3
2
∼1773, and 1718 (CdO anhydride).
Melt P olym er iza tion . Acyl-PEG, acyl-SA, and acyl-CPP
were mixed in a defined ratio (with or without 0.5-2.0%
catalyst) in a round-bottom flask with a stopcock adapter. Poly-
(ether-anhydrides) of eight different compositions were syn-
thesized by melt polycondensation of prepolymers in the bulk
under high vacuum.34 Briefly, the flask was immersed in an
oil bath at the selected temperature (150-220 °C), and the
monomers were allowed to melt. High vacuum was applied,
and the condensation byproduct, acetic anhydride, was col-
lected in a liquid nitrogen trap. Throughout the polymeriza-
tion, a strong nitrogen sweep was performed for 30 s every 15
min to agitate the reacting melt. At the end of the reaction,
the polymers were allowed to cool completely and then
dissolved in chloroform. The solution was precipitated drop-
wise into excess petroleum ether. The precipitate was collected
Hydroxyl-terminated poly(ethylene glycol) (PEG, M
n
) 8000)
Sigma) was dried by lyophilization before use. 1,3-Bis-
carboxyphenoxy)propane (CPP) was synthesized according to
(
(
3
3
the method described by Conix. Succinic anhydride (Sigma),
cadmium acetate (Aldrich), poly(vinyl alcohol) (88 mol %
hydrolyzed, 20 kDa MW, Polysciences), pyridine (Aldrich),
diethyl ether (J .T. Baker), petroleum ether (Fisher), dimeth-
ylformamide (Aldrich), methylene chloride (Fisher), and
rhodamine B base (Sigma) were used as received without
further purification.
Meth od s. 1H NMR spectra were recorded in CDCl
on a
3
Varian UNITY 400 MHz spectrometer. The composition of the
poly(ether-anhydrides) was determined by using the ratio of
average intensities per proton of each of the monomers.
Infrared (IR) spectra were obtained using a Perkin-Elmer 1600
series spectrometer. The samples were ground and pressed into
KBr pellets for analysis.
The molecular weight and polydispersity of the polymers
were determined by gel permeation chromatography (GPC)
analysis (J ASCO PU-980 intelligent HPLC pump, 1560 intel-
ligent column thermoset, RI-1530 intelligent refractive index
detector). Samples were filtered and eluted in chloroform
through a series of Styragel columns (guard, HR4, and HR3
Waters Styragel columns) at a flow rate of 0.3 mL/min. The
molecular weights were determined relative to polystyrene
standards (Fluka, Milwaukee, WI).
1
by filtration and dried under vacuum to constant weight. H
NMR (CDCl ): δ 6.95, 7.98 (d, ArH), 4.25 (s, CH ), 3.65 (s,
3
2
OCH CH O), 2.44 (t, CH ), 2.33 (m, CH ), 1.65 (m, CH ), 1.32
2
2
2
2
2
-
1
(s, CH ). IR (KBr, cm ): ∼1813-1773, ∼1742 (CdO anhy-
2
dride), 1112 (CH OCH ).
2
2
P r ep a r a tion of Dr u g-Loa d ed P oly(P EG:SA:CP P ) P a r -
ticles. Drug-loaded particles were prepared using a double
emulsion solvent evaporation method.1
5,16
The primary water-
in-oil emulsion was created by probe sonication (Sonics and
Materials Inc., Newtown, CT) of 100 µL of an aqueous solution
((2 mg/mL DNA) in a 50 mg/mL polymer solution in 4 mL of
methylene chloride ((5 mg/mL rhodamine B base). The
primary emulsion was then poured into 100 mL of 1% PVA
solution and homogenized at 6000 rpm for 1 min to form the
double emulsion (Silverson Machines Inc., East Longmeadow,
MA). The particles were stirred for 3 h to allow hardening,
collected by centrifugation at 3400 rpm (18.5 cm rotor,
International Equipment Co., Needham heights, MA), washed
twice with deionized water, resuspended in 10 mL of water,
Thermal analysis was performed using a SEKIO DSC220
differential scanning calorimeter. An average sample weight
of 5-10 mg was heated from -100 to 200 °C at a rate of 10
°
C/min.
Polymer particle morphology was evaluated by scanning
electron microscopy (SEM) with an AMRAY 1860 FE micro-
scope. Microparticle samples were attached to SEM mounts

7
176 Fu et al.
Macromolecules, Vol. 37, No. 19, 2004
Sch em e 1
and freeze-dried. Model drug molecules were added either to
the primary aqueous phase (LacZ plasmid DNA; 5148 kDa)
or to the organic phase (rhodamine B base; 443 Da) dependent
on which phase they were soluble.
Resu lts a n d Discu ssion
We have synthesized a family of poly(ether-anhy-
drides) in which the ratio of monomers, poly(ethylene
glycol) with Mn ) 8000 Da (PEG), sebacic acid (SA), and
Deter m in a tion of P a r ticle Size a n d Den sity. Particle
size distribution was determined using a Coulter Multisizer
IIe (Beckman-Coulter Inc., Fullerton, CA). Approximately 2
mL of Isoton II electrolyte solution (Beckman-Coulter Inc.) was
added to 5-10 mg of microparticles. The solution was briefly
vortexed to suspend the microparticles and added dropwise
to 100 mL of isoton II solution until the coincidence of particles
was between 8% and 10%. Greater than 100 000 particles were
sized for each batch of microparticles to determine the mean
particle size and size distribution. The bulk density of the
particles was determined by tap density. Briefly, particles were
loaded into 0.3 mL sections of a 1 mL plastic pipet, capped
with NMR tube caps, and tapped approximately 300-500
times until the volume of the powder did not change. The tap
density was determined from the difference between the
weight of the pipet before and after loading, divided by the
volume of powder after tapping.
1
,3-bis(carboxyphenoxy)propane (CPP), were varied. An
outline of the monomer and polymer synthesis is shown
in Scheme 1. As discussed in subsequent sections, poly-
(PEG:SA:CPP) with up to 50 wt % PEG can be used to
generate drug-loaded microparticles for inhalation or
injection. These particles can efficiently release drugs
with molecular weights ranging from 443 to 5 148 000
Da in a continuous fashion for up to 7 days. Importantly,
the versatile SA:CPP polymers have been modified to
incorporate various percentages of PEG into their
backbone, which should significantly enhance their
applicability to various advanced drug delivery applica-
tions. PEG with Mn between ∼2 and 20 kDa is known
to render insoluble particles less susceptible to phago-
cytosis and destruction by macrophages in the blood-
2
7,35,36
stream or the lung,
thereby allowing the particles
Degr a d a tion of P oly(P EG:SA:CP P ) P a r ticles. 10 mg of
particles was suspended in 1 mL of phosphate-buffered saline
to release drug for longer periods of time. PEG also
reduces the interparticle adhesion forces, which facili-
(pH 7.4) and incubated at 37 °C on a rotator. At predetermined
3
6
time intervals, the remaining particles were collected by
centrifugation at 13 000 rpm (7.3 cm rotor, National Labnet
Co., Woodbridge, NJ ), and the supernatant was removed.
Samples were dissolved in chloroform and filtered, and the
molecular weight was determined by GPC.
tates improved dispersion for injection and enhances
particle aerosolization from dry powder inhalers.32
Ch a r a cter iza tion of P EG:SA:CP P Ter p olym er s.
The structure of PEG:SA:CPP terpolymers was con-
firmed by FT-IR, 1H NMR, and GPC. The 1H NMR
Dr u g Relea se fr om P oly(P EG:SA:CP P ) P a r ticles. 10 mg
of drug-loaded particles was suspended in 1 mL of phosphate-
buffered saline (pH 7.4) and incubated at 37 °C on a rotator.
At selected time points, supernatants were removed following
centrifugation at 13 000 rpm (7.3 cm rotor, National Labnet
Co., Woodbridge, NJ ) and were replaced with 1 mL of fresh
phosphate buffer. Supernatants were either analyzed for
rhodamine B base fluorescence using a TD-700 fluorometer
(
Figure 1) resonance line of the methylene protons of
PEG appeared at 3.65 ppm. The three peaks at 2.44,
.65, and 1.32 ppm were attributed to the methyl
1
protons of SA and peaks at 2.33, 4.25, 6.95, and 7.98
ppm attributed to the protons of CPP. The peak at 7.25
ppm corresponds to the solvent, deuterated chloroform.
Importantly, the actual weight percentages of PEG, SA,
and CPP in the polymer (estimated by integration and
comparison of the corresponding NMR peaks) showed
good agreement with the monomer feed ratio. The
amount of SA in the copolymer was usually only slightly
(Turner Designs, Sunnyvale, CA) at excitation and emission
wavelengths of 550 and 570-700 nm or assayed for pDNA
content using the PicoGreen dye exclusion assay (Molecular
Probes, Eugene, OR).

Macromolecules, Vol. 37, No. 19, 2004
Poly(ether-anhydrides) for Drug Delivery 7177
F igu r e 1. 1H NMR spectra of poly(PEG:SA:CPP) (30:50:20) poly(ether-anhydride): a, peaks attributed to CPP; b, peaks attributed
to SA; c, peak attributed to PEG.
F igu r e 3. Molecular weight of poly(PEG:SA) (30:70) as a
F igu r e 2. Molecular weight of poly(PEG:SA) (30:70) as a
function of reaction time (with catalyst, vacuum ∼0.08 Torr).
function of reaction time at different temperatures (without
Polymers were melt-polymerized at 180 °C in the presence of
catalyst, vacuum ∼0.08 Torr).
2
cadmium acetate (CdAc ).
higher than the amount fed, while the CPP amount was
slightly less than fed. This may be due to the decreased
flexibility of CPP owing to steric hindrance. The rela-
tively limited mobility of CPP may cause a decrease in
polymer chain interactions with CPP, subsequently
decreasing the extent of CPP polymerization. Chro-
matographs from GPC analysis contained one peak
corresponding to the molecular weight of the polymer
determined to be the optimal polymerization conditions
without catalyst since a high molecular weight polymer
could be produced in a relatively short time.
The effect of catalyst on polymer molecular weight
was also studied (Figure 3). Cadmium acetate (CdAc2)
has previously been shown to be effective in producing
34,37
high molecular weight polyanhydrides.
The addition
(
not shown). NMR studies combined with GPC data
of catalyst led to only slightly higher polymer molecular
weights, with the maximum being achieved at 20-30
min. However, since the maximum molecular weight
with catalyst (t ) 20 min, 1% CdAc2, Mw ) 43.3 kDa)
was not significantly different than without (t ) 30 min,
Mw ) 42.0 kDa), a catalyst was not deemed necessary.
Since it is preferable to avoid the use of potentially toxic
catalysts during polymerization, reaction conditions of
180 °C for 30 min (without CdAc2) were used for all
further polymerizations.
indicated that PEG was successfully copolymerized with
SA and CPP.
P olym er Syn th esis Op tim iza tion . The effects of
polymerization time and temperature on poly(ether-
anhydride) molecular weight were studied using PEG:
SA with a weight feed ratio of 30:70 (Figures 2 and 3).
Subsequently, we discovered that a lowered vacuum
pressure during polymerization led to significantly
enhanced polymer molecular weights at the same reac-
tion temperatures and times (e.g., higher values re-
ported in Table 1).
We used the optimized reaction temperature and
time, as determined with PEG:SA:CPP 30:70:0, to
produce high molecular weight terpolymers with up to
35 wt % PEG and 45 wt % CPP incorporated into the
polymer backbone (Table 1). The molecular weight of
polymers generally decreased with increasing amounts
of PEG or CPP in the polymer backbone (Table 1) under
the same polymerization conditions (180 °C for 30 min).
The higher molecular weight polymers (up to 67 kDa
for terpolymers with significant percentages of all three
monomers) summarized in Table 1 were produced using
a slightly lower vacuum pressure than that used to
Figure 2 shows the dependence of polymerization time
on polymer weight- and number-average molecular
weight. With increased polymerization temperature, a
maximum in molecular weight was obtained in a shorter
amount of time (90 min at 150 °C to 10 min at 220 °C);
however, the maximum molecular weight that could be
achieved decreased at the highest temperature. When
polymerization was conducted at 180 °C, polymer mo-
lecular weight achieved a maximum after polymeriza-
tion for 30 min. Reaction at 180 °C for 30 min was

7
178 Fu et al.
Macromolecules, Vol. 37, No. 19, 2004
Ta ble 1. Ch a r a cter iza tion of P oly(eth er -a n h yd r id es) Syn th esized w ith P EG8000a
PEG:SA:CPP in
the feed (wt %)
PEG:SA:CPP
H NMR (wt %)
b
1
c
yield (%)
Mw (Da)
Mn (Da)
PDI
Tm1 (°C)
Tm2 (°C)
5
:95:0
86.4
82.5
85.6
81.4
83.1
81.2
78.3
77.1
4.4:95.6:0
9.2:91.8:0
80 500
80 800
56 500
49 100
41 600
67 000
66 500
58 100
26 000
31 800
25 600
20 600
18 700
28 900
27 000
24 200
3.10
2.54
2.21
2.39
2.22
2.32
2.46
2.40
d
d
49.8
e
81.3
80.4
79.9
e
1
3
4
5
3
3
3
0:90:0
0:70:0
0:60:0
0:50:0
0:50:20
0:35:35
0:20:50
28.6:71.4:0
37.6:62.4:0
46.8:53.2:0
33.1:49.4:17.5
34.7:35.4:29.9
30.7:24.6:44.7
e
e
50.9
e
50.5
63.0
e
d
a
No glass transition detected between -100 and 200 °C. b Poly(ether-anhydrides) were polymerized at 180 °C, ∼0.04 Torr for 30 min.
c
Estimated from the integral height of hydrogen atoms in the 1H NMR spectra. d Not detectable. e Not tested.
produce the polymers described in Figures 2 and 3
(
∼0.04 Torr compared to ∼0.08 Torr). High molecular
weight polymers are necessary to impart mechanical
strength to the polymers and to render them water-
insoluble (so that devices made from them do not readily
dissolve in the body) but should not be so high that
processing becomes difficult or impossible. Polyanhy-
dride molecular weights above approximately 15 kDa
were sufficient for preparation of microparticles capable
of controlled drug delivery. Polymers containing up to
5
0 wt % PEG or up to 35 wt % CPP all had molecular
F igu r e 4. Stability profiles of poly(PEG:SA:CPP) (30:50:20)
in the solid state and in anhydrous chloroform at four storage
temperatures. Polymerization conditions were 30 min at 180
°C without catalyst.
weights significantly above 15 kDa. Therefore, polymers
produced in this study could be processed into micro-
spheres using emulsion techniques, even after molecular
weight stabilization during storage, as shown later.
Polymer polydispersities remained fairly constant
regardless of the polymer composition, with most poly-
dispersity index (PDI) values between 2.2 and 2.5 (Table
a stabilization of the molecular weight for at least 18
days (i.e., the duration of the study). A lower storage
temperature provided more protection against degrada-
tion (∼28 kDa at -80 °C vs ∼20 kDa at 25 °C in the
solid state). Additionally, polymers stored in the solid
form (∼28 kDa at -80 °C) maintained higher molecular
weights than those stored in chloroform (∼15 kDa at
-80 °C). The final molecular weights obtained after
stabilization was sufficient to allow the formation of
microspheres by emulsion techniques and the controlled
release of therapeutic drugs from the microspheres for
all polymers tested.
1
). PDI values in this range are within the range of both
the PLGA and poly(CPP:SA) families of polymers cur-
rently used in humans.
Th er m a l An a lysis. Thermal analysis was performed
on the poly(ether-anhydrides), as shown in Table 1.
Only the SA melting point was observed (∼80 °C) when
there was less than 10% PEG in PEG:SA polymers. The
melting point of PEG (49.8 °C) and SA (79.9 °C) were
both observed when PEG content was increased to 30%.
The appearance of two distinct melting points implies
that PEG- and SA-rich regions phase separate at high
PEG percentages. The SA melt point disappeared, and
a new melt point appeared at 60.3 °C, when 20% CPP
was introduced into a polymer chain containing 30%
PEG. This likely indicates that CPP disrupted SA
crystallinity and that CPP and SA phases are colocal-
ized. The addition of 50% CPP caused the SA melting
point to be undetectable, leaving only the PEG melting
point. Therefore, the CPP:SA-rich phase of the polymer
was amorphous, and the CPP and SA monomers are
likely distributed uniformly throughout the polymer
backbone. No glass transition temperature was observed
for any of the polymers (-100 to 200 °C). Additionally,
since the melting temperature of the entire family of
poly(ether-anhydrides) was 50 °C or higher (∼80 °C for
polymers composed of e10% PEG), these polymers are
not expected to undergo a thermal phase transition
under typical storage or use conditions.
P oly(P EG:SA:CP P ) P a r ticles for In h a la tion a l
Dr u g Deliver y. Large (5-20 µm), low-density dry
powder aerosols can be efficiently aerosolized into the
1
4,16
deep lungs.
In particular, a decrease in particle
density allows the efficient aerosolization of geometri-
cally large particles (e.g., diameter > 5 µm) since they
possess low aerodynamic diameters (e.g., < 3 µm) (see
eq 1). Large particle size also reduces particle clearance
rates by phagocytic cells, allowing them to remain in
the deep lungs and deliver drugs for extended periods
of time. We used the poly(ether-anhydrides) to produce
geometrically large, but aerodynamically small, par-
ticles (owing to high porosity) and found that density
decreased as the amount of PEG in the polymer back-
3
bone increased (Table 2), from 0.34 g/cm (0% PEG) to
3
0.06 g/cm (50% PEG). One possible explanation for this
significant change in density is that the addition of the
hydrophilic PEG monomer increases water uptake by
the particles during their preparation, thus causing
them to swell. Pores would then be formed as the water
is removed during freeze-drying (Figure 5), leading to
a decrease in density.
Sta bility of P EG:SA:CP P Ter p olym er s. The sta-
bility of poly(PEG:SA:CPP) (30:50:20), a model terpoly-
mer with significant percentage of all three monomers,
was studied in the dry state (as a high surface area
powder) and in solution (anhydrous chloroform) at
temperatures ranging from -80 to 25 °C (Figure 4).
Polymers stored under all conditions showed an initial
decrease in molecular weight within a few days and then
Particle aerodynamic diameter (da) was calculated by
the following relation:
d ) d F/F /γ
(1)
a
x
a

Macromolecules, Vol. 37, No. 19, 2004
Poly(ether-anhydrides) for Drug Delivery 7179
Ta ble 2. Ch a r a cter iza tion of P oly(eth er -a n h yd r id e)
Micr osp h er es
bulk
a
geometric
diam (µm)
density
aerodynamic
diam (µm)
a
3
(g/cm )
PSA
PEG:SA (5:95)
6.26
6.05
5.97
5.66
5.86
6.45
6.54
6.28
0.34
0.32
0.28
0.14
0.08
0.06
0.15
0.11
3.66
3.43
3.15
2.13
1.67
1.51
2.57
2.07
PEG:SA (10:90)
PEG:SA (30:70)
PEG:SA (40:60)
PEG:SA (50:50)
PEG:SA:CPP (10:70:20)
PEG:SA:CPP (30:50:20)
a
Geometric size and density are the average of three measure-
ments.
F igu r e 5. Morphology, determined by scanning electron
microscopy, of (A) PSA, (B) poly(PEG:SA) (10:90), and (C) poly-
F igu r e 7. Release of model drugs, (A) rhodamine B base (MW
) 443 Da) and (B) LacZ plasmid DNA (MW ) 5143 kDa), from
poly(ether-anhydride) microspheres.
(
PEG:SA) (30:70) microspheres, showing an increase in the
number of visible pores in the microspheres with an increase
in the amount of PEG in the polymer backbone.
h, whereas PEG:SA (30:70) was degraded to 10% its
original molecular weight after only 5 h. In a related
1
5
study, we used PEG with a lower molecular weight
600 Da vs 8000 Da used here) and showed that
(
PEG600:SA (30:70) degraded to 20% of its original
molecular weight in 2.5 h, compared to 50% for PEG8000:
SA (30:70) in the same time period (Figure 6). Thus, an
increase in the size of PEG monomer added into the
polymer backbone (8000 Da vs 600 Da) caused the
polymer to degrade more slowly. The addition of very
hydrophobic CPP into the polymer backbone, however,
had the most significant effect on decreasing the poly-
mer degradation rate (Figure 6).
F igu r e 6. Degradation profiles of poly(ether-anhydride)
It is important to note that although the polymer
molecular weight decreased significantly in a few hours
in PBS at 37 °C, dissolution of the hydrophobic mono-
mers is slow, thus producing steady drug release over
longer time periods. Therefore, drug release is controlled
mainly by monomer dissolution rates (i.e., particle
erosion) rather than polymer degradation rates (see next
section). Also, because degradation in the lungs occurs
on a thin layer of fluid (∼0.2 µm in the alveoli where
the majority of oxygen is exchanged with carbon di-
oxide), the in vivo degradation of the relatively large
particles may be significantly slower compared to
particles completely immersed in PBS. We are currently
developing methods to model the degradation of par-
ticles on pulmonary epithelial cells covered by a thin
microspheres in PBS (pH 7.4, 37 °C).
where d ) geometric diameter, F ) particle bulk density
3
3
(
g/cm ), Fa ) water mass density (1 g/cm ), and γ )
shape factor ) 1 for a sphere. Therefore, an aerosolized
particle with a geometric diameter of 1 µm and standard
3
density of 1 g/cm will deposit in the respiratory tract
following inhalation as a dry powder in roughly the
same manner as a 10 µm particle with a density of 0.01
3
g/cm . The decreased density of particles as PEG content
increased resulted in a significant decrease in particle
aerodynamic diameter (da) from 3.7 (0% PEG) to 1.5 µm
(50% PEG) (Table 2). Thus, these particles are within
the necessary aerodynamic size range (1-5 µm) for
efficient aerosolization into the deep lungs, even though
the geometric diameters are significantly larger (6.3-
38
aqueous film.
Dr u g Relea se. Varying the monomer composition in
the polymer backbone allowed control over the release
of rhodamine B base, a small molecular weight hydro-
phobic fluorescent molecule, from microparticles (Figure
6
.5 µm). PEG also aids in the control of particle
aggregation that can prevent efficient aerosolization. In
particular, we recently showed that the inclusion of PEG
in the polymer backbone of ether-anhydrides signifi-
7
A). Increasing the hydrophobicity of the particles (by
3
2
cantly improves aerosolization of the microspheres.
increasing the percentage of SA or CPP relative to PEG)
decreased the drug release rate. PEG:SA (10:90) par-
ticles released 60% of their drug load in approximately
4 h, whereas PSA particles (more hydrophobic) released
60% in approximately 18 h and PEG:SA:CPP (10:70:
20) particles (most hydrophobic) released 60% in about
2.5 days. No initial burst of drug was seen for any of
the particles studied.
Degr a d a tion of P oly(P EG:SA:CP P ) P a r ticles.
The degradation rate of poly(PEG:SA:CPP) particles
immersed in phosphate buffered saline (PBS) at 37 °C
increased with increasing amounts of PEG (Figure 6),
likely because the polymer becomes more hydrophilic
upon addition of PEG in the backbone. PSA was
degraded to 10% its original molecular weight after 13

7
180 Fu et al.
Macromolecules, Vol. 37, No. 19, 2004
The release of LacZ plasmid DNA (5148 kDa) from
(11) Stolnik, S.; Illum, L.; Davis, S. S. Adv. Drug Deliv. Rev. 1995,
1
6, 195-214.
the poly(ether-anhydride) particles (Figure 7B) was
similar to that of the smaller molecule, rhodamine B
base, and lasted approximately 7 days (compared to 5
days for rhodamine B base). The slightly protracted
release time is likely due to the large size of the DNA
molecule inhibiting its diffusion through the polymer
matrix. Similar to the small molecule release, the
composition of the polymer backbone controlled the
DNA release rate, with the more hydrophilic polymer
(
12) Gref, R.; Minamitake, Y.; Peracchia, M. T.; Trubetskoy, V.;
Langer, R. Science 1994, 263, 1600-1603.
(
13) Li, Y. P.; Pei, Y. Y.; Zhang, X. Y.; Gu, Z. H.; Zhou, Z. H.; Yuan,
W. F.; Zhou, J . J .; Zhu, J . H.; Gao, X. J . J . Controlled Release
2001, 71, 203-211.
(
(
(
14) Hanes, J .; Edwards, D. A.; Evora, C.; Langer, R. U.S. Patent
No. 5,855,913, 1999.
15) Fu, J .; Fiegel, J .; Krauland, E.; Hanes, J . Biomaterials 2002,
2
3, 4425-4433.
16) Edwards, D. A.; Hanes, J .; Caponetti, G.; Hrkach, J .; Ben-
J ebria, A.; Eskew, M. L.; Mintzes, J .; Deaver, D.; Lotan, N.;
Langer, R. Science 1997, 276, 1868-1871.
(
PEG:SA 30:70) releasing DNA faster than the more
hydrophobic polymer (PEG:SA 5:95). Again, no initial
burst of drug (DNA) was seen for the particles exam-
ined.
(
(
17) Suarez, S.; O’Hara, P.; Kazantseva, M.; Newcomer, C. E.;
Hopfer, R.; McMurray, D. N.; Hickey, A. J . Pharm. Res. 2001,
1
8, 1315-1319.
18) Tamada, J . A.; Langer, R. Proc. Natl. Acad. Sci. U.S.A. 1993,
0, 552-556.
Con clu sion s
9
(
(
19) Okada, H. Adv. Drug Deliv. Rev. 1997, 28, 43-70.
20) Zhu, G. Z.; Mallery, S. R.; Schwendeman, S. P. Nat. Biotech-
nol. 2000, 18, 52-57.
A new family of poly(ether-anhydrides) containing
various ratios of SA, CPP, and PEG was synthesized
for use as drug carriers for injection or inhalation. The
polymers were designed to enable the creation of
particles that erode over time periods roughly equiva-
lent to the rate of drug release, to provide steady
controlled release of various drug molecules, and to
potentially avoid rapid destruction by the immune
system due to the incorporation of PEG8000 into the
polymer backbone. These new polymers may have
significant advantages compared to those currently used
in humans for a variety of applications and are com-
posed of monomers already safely used in humans.
(
21) Lee, H. J .; Riley, G.; J ohnson, O.; Cleland, J . L.; Kim, N.;
Charnis, M.; Bailey, L.; Duenas, E.; Shahzamani, A.; Marian,
M.; J ones, A. J . S.; Putney, S. D. J . Pharmacol. Exp. Ther.
1
997, 281, 1431-1439.
(
22) J ohnson, O. L.; J aworowicz, W.; Cleland, J . L.; Bailey, L.;
Charnis, M.; Duenas, E.; Wu, C. C.; Shepard, D.; Magil, S.;
Last, T.; J ones, A. J . S.; Putney, S. D. Pharm. Res. 1997, 14,
730-735.
(23) Batycky, R. P.; Hanes, J .; Langer, R.; Edwards, D. A. J .
Pharm. Sci. 1997, 86, 1464-1477.
(
24) Cleland, J . L.; Lim, A.; Barron, L.; Duenas, E. T.; Powell, M.
F. J . Controlled Release 1997, 47, 135-150.
(
(
25) Visser, J . Powder Technol. 1989, 58, 1-10.
26) Evora, C.; Soriano, I.; Rogers, R. A.; Shakesheff, K. M.; Hanes,
J .; Langer, R. J . Controlled Release 1998, 51, 143-152.
Ack n ow led gm en t. This work was supported by The
Whitaker Foundation (RG 990046) and by a fellowship
to J . Fiegel from the National Science Foundation (DGE
(27) Redhead, H. M.; Davis, S. S.; Illum, L. J . Controlled Release
001, 70, 353-363.
28) Tabata, Y.; Ikada, Y. J . Biomed. Mater. Res. 1988, 22, 837-
58.
29) Wang, P. P.; Frazier, J .; Brem, H. Adv. Drug Deliv. Rev. 2002,
4, 987-1013.
30) Dang, W.; Daviau, T.; Brem, H. Pharm. Res. 1996, 13, 683-
91.
31) Leong, K. W.; Brott, B. C.; Langer, R. J . Biomed. Mater. Res.
985, 19, 941-955.
32) Fiegel, J .; Fu, J .; Hanes, J . J . Controlled Release 2004, 96,
11-423.
33) Conix, A. Macromol. Synth. 1966, 2, 95-99.
2
(
(
(
(
(
(
9
616062).
8
Refer en ces a n d Notes
5
(
(
1) Langer, R. Nature (London) 1998, 392, 5-10 (Suppl).
2) Uhrich, K. E.; Cannizzaro, S. M.; Langer, R. S.; Shakesheff,
K. M. Chem. Rev. 1999, 99, 3181-3198.
3) Kumar, M. N. V. R.; Kumar, N.; Domb, A. J .; Avora, M. Adv.
Polym. Sci. 2002, 160, 45-117.
6
1
(
4
(
4) Rokkanen, P. U.; Bostman, O.; Hirvensalo, E.; Makela, E.
A.; Partio, E. K.; Patiala, H.; Vainionpaa, S.; Vihtonen, K.;
Tormala, P. Biomaterials 2000, 21, 2607-2613.
5) Veiranto, M.; Tormala, P.; Suokas, E. J . Mater. Sci.: Mater.
Med. 2002, 13, 1259-1263.
(34) Domb, A. J .; Langer, R. J . J . Polym. Sci. 1987, 25, 3373-
(
(
(
(
(
3386.
(35) M u¨ ller, B. G.; Kissel, T. Pharm. Pharmacol. Lett. 1993, 3,
6) Turvey, T. A.; Bell, R. B.; Tejera, T. J .; Proffit, W. R. J . Oral
Maxil. Surg. 2002, 60, 59-65.
67-70.
(36) Gref, R.; Domb, A.; Quellec, P.; Blunk, T.; Muller, R. H.;
Verbavatz, J . M.; Langer, R. Adv. Drug Deliv. Rev. 1995, 16,
215-233.
7) Sipe, J . D.; Kelley, C. A.; McNichol, L. A. Ann. N.Y. Acad.
Sci. 2002, 961, 1-389.
8) Marler, J . J .; Upton, J .; Langer, R.; Vacanti, J . P. Adv. Drug
Deliv. Rev. 1998, 33, 165-182.
(37) Hanes, J .; Chiba, M.; Langer, R. Macromolecules 1996, 29,
5279-5287.
9) Edlund, U.; Albertsson, A. C. Adv. Polym. Sci. 2002, 157, 67-
(
38) Fiegel, J .; Ehrhardt, C.; Schaefer, U. F.; Lehr, C.-M.; Hanes,
1
12.
J . Pharm. Res. 2003, 20, 788-796.
(
10) Hanes, J .; Chiba, M.; Langer, R. Biomaterials 1998, 19, 163-
1
72.
MA049853S