Novel degradable poly(ethylene glycol) hydrogels for controlled release of protein

Article abstract of DOI:10.1021/js980065o

Hydrogels have become increasingly important in the biomedical field. This paper describes synthesis and characterization of two types of novel degradable poly(ethylene glycol) (PEG) hydrogels with potential utility as delivery carriers for bioactive drug

Full text of DOI:10.1021/js980065o

Novel Degradable Poly(ethylene glycol) Hydrogels for Controlled
Release of Protein
†
XUAN ZHAO AND J. MILTON HARRIS*
Contribution from Department of Chemistry, The University of Alabama in Huntsville, Huntsville, Alabama 35899.
Received February 26, 1998. Accepted for publication May 27, 1998.
studied as depot devices for drug delivery, and they have
Abstract 0 Hydrogels have become increasingly important in the
biomedical field. This paper describes synthesis and characterization
of two types of novel degradable poly(ethylene glycol) (PEG) hydrogels
with potential utility as delivery carriers for bioactive drugs. The
simplest gel is prepared by one-step polycondensation of difunctional
PEG acid and branched PEG “polyol”. The second type of the novel
degradable PEG hydrogel was prepared in a two-step process, in
which an ester-containing, amine-reactive PEG derivative was
synthesized and then reacted with a branched PEG amine to form
the gel. The two-step gels are formed in very mild conditions, and
therefore fragile drugs such as proteins can be loaded during gel
formation. Because most proteins have free amino groups in the
sequence, these proteins can be covalently linked to the hydrogel
network. This covalent attachment provides a new way to achieve
long-term controlled release of proteins. These hydrogels have a wide
range of degradation rates. Upon hydrolysis, these PEG hydrogels
will degrade into low molecular weight PEG derivatives, which can
be easily cleared by the body.
been prepared in several fundamentally different ways. The
simplest route is to cross-link PEG by γ-irradiation27 or
electron-beam irradiation.28 The hydrogels formed in this
way are nondegradable.
Graham and co-workers23,24,29 pioneered use of coupling
reactions between a difunctional PEG, a triol, and a
multifunctional isocyanate. This type of hydrogel was
recently studied for release of proteins.30,31 These hydro-
gels can also be made degradable by inclusion of copolymers
containing ester linkages.24 Suitably activated PEGs can
be cross-linked by a variety of cross-linkers. Fortier and
co-workers32 have developed a hydrogel by cross-linking
PEG p-nitrophenyl carbonate with bovine serum albumin
(BSA). Similarly, Barrows and co-workers33 have prepared
a hydrogel by reaction between difunctional PEG succin-
imidyl succinate and human serum albumin (HSA).33
Because of presence of an ester linkage in PEG succinate,
the hydrogels are hydrolytically degradable. This hydrogel
was developed as a tissue sealant and has not yet been
utilized for drug delivery.
The degradable PEG hydrogels studied most were de-
veloped in Hubbell’s laboratory.34-36 These workers pre-
pared PEG acrylates with a short segment of poly(lactide/
glycolide) between the PEG and the acrylate group. Acrylate
polymerization then provides a cross-linked network in
which the poly(lactide/glycolide) provides degradability.
Other degradable PEG hydrogels that take advantage of
ester linkages include gels made from PEG-lysine acry-
late37 and PEG-poly(vinyl alcohol).38,39 Hoffman40 dis-
closed a novel hydrolytically degradable hydrogel prepared
from PEG dialdehyde and poly(vinylamine) in which the
Schiff base between the aldehyde and the amine provides
degradability. Recently, West and Hubbell41 reported
enzymatically degradable PEG-based hydrogels made from
photopolymerization of block copolymers of PEG and short
oligopeptides that are further capped at each end with an
acrylate group. Russell and co-workers42 have formed a
photodegradable PEG hydrogel by photochemical [2 + 2]
coupling of vinyl PEGs. This reaction is reversible upon
irradiation at a different wavelength.42
In our laboratory, we have developed two types of novel
degradable PEG-based hydrogels. The first type of hydro-
gel was formed in a one-step polycondensation reaction
between difunctional PEG acids and branched PEG polyols.
The resulting ester linkages provide degradability. Upon
hydrolysis of the ester linkages, these gels degrade into only
PEG and PEG derivatives. The degradation rate can be
controlled by slightly varying the molecular structure in
the esters and the degree of branching in the cross-linker.
The second type of hydrogel was formed under mild
condition (aqueous solution and room temperature) via a
two-step process in which an ester-containing active PEG
derivative is first synthesized and then coupled to an amine
cross-linker. In these hydrogels, protein drugs can be
covalently attached to the hydrogel network via ester
Introduction
Hydrogels are generally considered as biocompatible
materials because of their high water content.1 They have
been used in a variety of biomaterial and biotechnology
applications, such as tissue engineering, artificial organs,
and drug delivery.2 Recently, hydrogels have been exten-
sively studied as drug carriers, especially for proteins.
Hydrogels have been made from different polymers such
as dextran,3 gelatin,4-6 amylose,7-9 cellulose,10 chitosan,11
fibrin,12 collagen,13,14 alginate,15,16 poly(vinyl alcohol),17
poly(hydroxyethyl methacrylate),18,19 poly(hydroxyethyl acry-
late),20 poly(acrylic acid) (Carbopol),21 polyasparthydrazide,22
and poly(ethylene glycol) (PEG).23,24
Most hydrogel-based drug delivery systems are designed
as implants that release drug locally at a predetermined
rate. Therefore, degradability of the gels is very important
because surgical removal of the drug-depleted device is not
required, and, drug release kinetics can be controlled.
However, it is necessary that degradation products be
biocompatible and nontoxic.
Poly(ethylene glycol) is a biocompatible, nontoxic, non-
immunogenic and water-soluble polymer of much use in
biomaterials, biotechnology and medicine.25 Two PEG-
protein drugs, Adagen (Enzon) and Oncaspar (Enzon), have
been approved by the FDA.26 Therefore, PEG could be an
ideal material for internal use in hydrogel-based drug-
delivery devices. In fact, several PEG hydrogels have been
* Corresponding author: Shearwater Polymers, Inc., 2305 Spring
Branch Rd., Huntsville, AL 35801. Telephone: 205-533-4201, ext 227.
Fax: 205-533-4805. E-mail: jmharris@swpolymers.com.
† Current address: Shearwater Polymers, Inc., 2305 Spring Branch
Rd., Huntsville, AL 35801.
1450 / Journal of Pharmaceutical Sciences
10.1021/js980065o CCC: $15.00
Published on Web 09/04/1998
© 1998, American Chemical Society and
Vol. 87, No. 11, November 1998
American Pharmaceutical Association

linkages. Thus, release of the protein drug would be
controlled by hydrolysis of the ester linkage between the
gel and the protein, by diffusion from the gel, and by
degradation of the gel. There are several types of esters
used in these hydrogels, which provide a range of degrada-
tion rates.
In this paper, we report the synthesis and characteriza-
tion of these hydrogels as well as the potential of these
hydrogels for long-term controlled release of protein drugs.
Experimental Section
Ma ter ia lssGlycolic acid (GA), 3-hydroxybutyric acid (HBA),
benzyl glycolate, thionyl chloride, and disperse yellow 9 were
purchased from Aldrich (Milwaukee, WI). Carboxymethylated
PEG 2000 PEG-CM, PEG-propionic acid 2000 (PA), branched
PEGs (4-arm PEG 2000, 4-arm PEG 10,000, and 8-arm PEG
10,000), and m-PEG-silane (5 kD) were gifts from Shearwater
Polymers, Inc. (Huntsville, AL). Fluorescein isothiocyanate-
labeled BSA (FITC-BSA) and stannous octoate [Sn(Oct)2] were
purchased from Sigma (St. Louis, MO). All other reagents were
purchased from either Aldrich or Sigma and used without further
purification.
On e-Step Hyd r ogel P r ep a r a tion sIn an aluminum pan (1-
in. diameter), 600 mg of dry difunctional PEG 2000 acid (0.6 mmol
end groups) and one equivalent of dry branched PEG (e.g., 750
mg 8-arm PEG 10,000) were mixed with 15-20% stannous octoate
and melted. After a thin film of the melt covered the pan surface
uniformly, the pan was put into a vacuum oven and heated at 130
°C and 100 milliTorr for 6-24 h. A firm and transparent gel
formed. After cooling in a stream of N2, the gel became translu-
cent. The gel was cut into thin disks. The shaped crude gels were
then swollen in glacial acetic acid and washed three times with
this solvent over a 2-3-day period. After washing, the gels were
dried under reduced pressure. The tin content of the gel was
determined by the inductively coupled plasma (ICP) method
conducted at SouthEastern Analytical Services, Inc. (Huntsville,
AL).
Tw o-St ep H yd r ogel P r ep a r a t ion sA Two-Step Gel Made
from Difunctional Double-Ester PEGsFifty milligrams of difunc-
tional PEG-CM-HBA-N-hydroxysuccinimide (NHS) 2000 (vide
infra) was dissolved in 0.25 mL of deionized water. To the solution
was added 0.5 mL of a buffered solution with or without FITC-
BSA (10 mg/mL) and 0.25 mL of 8-arm PEG-amine solution (250
mg/mL). After vigorous shaking, the solution was allowed to sit.
The gel formed in a few minutes. A suitable pH range was 5.5 to
8.
A Two-Step Gel Made from Difunctional PEG-Succinimidyl
Carbonate Containing an Ester in the MiddlesFifty milligrams
of difunctional ester-containing PEG-succinimidyl carbonate 6800
(vide infra) was dissolved in 0.3 mL of deionized water. To the
solution was added 0.3 mL of a buffered solution with or without
FITC-BSA (10 mg/m) and 0.13 mL of 8-arm PEG-amine (250 mg/
mL). After vigorous shaking, the solution was allowed to sit. The
gel formed in a few minutes. A suitable pH range was 5.5 to 8.
Kin etics of th e Ester Lin k a gessA solution of m-PEG-ester-
PEG-m prepared as reported elsewhere43 (1.0 wt %) and PEG
20,000 (0.2 wt %, as internal standard) was prepared in a desired
buffer. The concentration of m-PEG-ester-PEG-m (C) and its
hydrolysis products were monitored by high-performance liquid
chromatography (HPLC)-gel permeation chromatography (GPC)
(Ultrahydrogel 250 column, Waters). Assuming first-order kinet-
ics, the hydrolytic half-life was calculated from the slope of the
natural logarithm plot of C versus time.
Figure 1sSchematic of the circulation system used in release study: (A) UV
spectrophotometer; (B) peristaltic pump; (C) UV flow cuvette; (D) temperature
control device; (E) a gel; (F) circulating water bath; (G) magnetic stir bars;
(H) magnetic stir plate; (I) teflon tubing.
flow cell were coated (vide infra) with PEG to minimize nonspe-
cific protein absorption. The buffer solution was maintained at
37 °C with gentle stirring. At the same time, the solution was
continuously pumped into the flow cell of a Hewlett-Packard
(8452A) UV spectrophotometer to monitor the release of the drug
at its maximum absorption wavelength. The UV absorption was
collected automatically at predetermined time intervals of 5 min.
This method was mainly used for release experiments that lasted
for <3 days.
Manual MethodsRelease of proteins was studied mainly by the
manual method. A hydrogel disk was immersed (t ) 0) in
phosphate buffer (0.1 M, pH 7.0) with the amount of the buffer
>20 times the wet gel weight, and the buffer solution was
maintained at 37 °C with a water bath and gently shaken. At a
predetermined time, a small amount of buffer solution was
removed for drug concentration determination and then returned
after measurement.
Coa tin g of th e Gla ss Su r fa ce of th e Cir cu la tion System
w ith m -P EG-Sila n e44sThe surface of the flask and the UV flow
cell were treated with aqueous sodium hydroxide (1 wt %) at 90
°C for 10 min, aqueous hydrogen chloride (3 wt %) at 90 °C for 10
min, and hot H2O2 solution (30 wt %) for 1 h. The surface was
then washed with a large amount of deionized water and dried in
a
vacuum oven at 150 °C overnight. After cooling to room
temperature, the flask and the flow cell were filled with m-PEG-
silane 5000 in dry toluene (2 wt %) and kept under N2 overnight.
The toluene solution was decanted and the surface was washed
twice with dry toluene. The surface was then cured in a vacuum
oven at 130 °C for 4 h.
P r ep a r a t ion of Difu n ct ion a l Dou ble-E st er P E GsPEG
Ester Acids (CM-R-COOH and PA-R-COOH)sMethod 1sPEG
2000 acid (10 g, 10 mmol end-group, PEG-CM or PEG-PA) was
azeotropically dried with 60 mL of toluene under N2. After 2 h,
the solution was cooled to room temperature, and thionyl chloride
solution (20 mL, 40 mmol) in CH2Cl2 was injected. The solution
was stirred at room-temperature overnight. The solvent was
condensed using a rotary evaporator, and the residual syrup was
dried under reduced pressure for ∼4 h over P2O5 powder. Hydroxy
acid (GA or HBA, 27 mmol) was azeotropically dried with 140 mL
of 1,4-dioxane, and the distillation was stopped when ∼70 mL of
solution remained. The solution was slowly cooled to room
temperature under N2. The glycolic acid-dioxane solution was
then added to the PEG acyl chloride. After the PEG was dissolved,
6 mL of dry triethylamine was injected to the system (precipitate
formed immediately), and the solution was stirred overnight. The
salt was filtered off and the filtrate was condensed on a rotary
evaporator at 55 °C and dried under reduced pressure. The crude
product was then dissolved in 100 mL of distilled water and the
pH of the solution was adjusted to 3.0. The aqueous phase was
extracted three times with a total of 200 mL of methylene chloride.
The organic phase was dried over sodium sulfate, filtered to
remove salt, condensed on a rotary evaporator, and precipitated
into 1000 mL of ethyl ether. The precipitate was collected by
filtration and dried under reduced pressure to yield 8.5-9.2 g (85-
Relea se Stu d ies of th e Mod el Dr u gs fr om th e GelssIn the
one-step gel, model drugs were first loaded into the hydrogels by
placing a purified gel in a solution of the model drug. Depending
on the drug molecular weight and its concentration, the soaking
time ranged from a few days for m-PEG-dyes to a few weeks for
FITC-BSA. All drug-loaded hydrogel disks were weighed and their
diameters were measured before release studies were undertaken.
Automatic MethodsThe release study was conducted in
a
circulation system with an ultraviolet (UV) flow cell. The system
is shown schematically in Figure 1. A drug-loaded hydrogel was
immersed, at time t ) 0, in phosphate buffer (0.1 M, pH 7.0). In
protein-release measurements, the surface of the flask and the
Journal of Pharmaceutical Sciences / 1451
Vol. 87, No. 11, November 1998

92%). 1H NMR (DMSO-d6): PEG-GA: δ 3.5 (br m, PEG), 4.3-
4.6 (s, PEGCOOCH2COOH), 2.59 [t, PEGOCH2CH2COO (PA)],
4.19 [s, PEGOCH2COO (CM)]; PEG-HBA: δ 3.5 (br m, PEG), 2.54
(d, PEGCOOCH(CH3)CH2COOH), 5.1 (h, PEGCOOCH(CH3)CH2-
COOH), 1.2 (d, PEGCOOCH(CH3)CH2COOH), 2.54 [t, PEG-
OCH2CH2COO (PA)], 4.055 [s, PEGOCH2COO (CM)].
was dissolved in 3.0 mL of buffer solution and shaken promptly
to obtain solution as soon as possible. Then, the solution was
transferred into a UV cuvette and the absorbance at 260 nm for
NHS ester or at 402 nm for p-nitrophenyl ester was followed as a
function of time. The hydrolysis half-life was determined from
the first-order kinetic plot (natural logarithm of final absorbance
minus absorbance at time t versus time).
Meth od 2sm-PEG 5000 acid (8 g, 1.6 mmol, PEG-CM or PEG-
PA), benzyl glycolate (GA-BZ), or 3-hydroxybenzylbutyrate (HBA-
BZ, vide infra; 2.24 mmol), 1-hydroxybenzotriazole (HOBT, 1.6
mmol), 4-(dimethylamino)pyridine (DMAP, 2 mmol), and dicyclo-
hexylcarbodiimide (DCC, 2.4 mmol) were dissolved in 80 mL of
dry methylene chloride. The solution was stirred under nitrogen
overnight, and the solvent was removed by rotary evaporation.
The resulting syrup was redissolved in 30 mL of dry toluene, and
the insoluble solid was filtered off. Then the solution was
precipitated into 500 mL of dry ethyl ether. The product was
reprecipitated with ether, then collected by filtration and dried
under reduced pressure to yield 7.8 g (98%). 1H NMR(DMSO-d6):
PA-GA-BZ: δ 3.5 (br m, PEG), 4.74 (s, PEGCOOCH2COOBZ), 2.62
(t, PEGOCH2CH2COOGA), 5.17 (GACOOCH2C6H5), 7.37 (GA-
COOCH2C6H5); CM-HBA-BZ: δ 3.5 (br m, PEG), 2.72 (m,
PEGCOOCH(CH3)CH2COOBZ), 5.22 (h, PEGCOOCH(CH3)CH2-
COOBZ), 1.23 (d, PEGCOOCH(CH3)CH2COOH), 4.0 (q, PEGOCH2-
COOHBA), 5.09 (s, HBACOOCH2C6H5), 7.36 (br s, HBACOO-
CH2C6H5).
A solution of PEG acid-hydroxyacid benzyl ester obtained in
the previous step (5 g, 1 mmol) in 1,4-dioxane (50 mL) was
hydrogenated with H2 (2 atm pressure) and 1.5 g of Pd/C (10%)
overnight. The catalyst was removed by filtration, and the solvent
was condensed by rotary evaporation. The residual syrup was
added to 200 mL of ethyl ether. The product was collected by
filtration and dried under reduced pressure to yield 4 g (80%).
PEG Ester-NHS Esters (Double Ester, CM-R-COO-NHS and
PA-R-COO-NHS)sThe PEG ester acid (3.3 g, ∼3.3 mmol) and
420 mg NHS (3.5 mmol) were dissolved in 50 mL of methylene
chloride. Then, DCC (800 mg, 3.8 mmol) in 5 mL of dry methylene
chloride was added. The solution was stirred under nitrogen
overnight and the solvent was removed by rotary evaporation. The
resulting syrup was redissolved in 10 mL of dry toluene and the
insoluble solid was filtered off. The solution was then precipitated
into 100 mL of dry ethyl ether. The precipitate was collected by
filtration and dried under reduced pressure to yield 3.1 g (94%).
1H NMR (DMSO-d6): PEG-GA-NHS: δ 3.5 (br m, PEG), 5.15-
5.21 (s, PEGCOOCH2COONHS), 2.67 (t, PEGOCH2CH2COO (PA)),
4.27 (s, PEGOCH2COO ppm (CM)), 2.82 (s, NHS, 4H); PEG-HBA-
NHS: δ 3.5 (br m, PEG), 3.0-3.2 (m, COOCH(CH3)CH2COONHS),
5.26 (h, COOCH(CH3)CH2COONHS), 1.3 (d, COOCH(CH3)CH2-
COONHS), 2.54 (t, OCH2CH2COO(PA)), 4.1 (s, OCH2COO (CM)),
2.81 (s, NHS).
PEG Ester-Glycolic Acid-p-Nitrophenyl ester (CM-GA-NP or
PA-GA-NP)sPEG ester-glycolic acid 2000 (1.5 g, ∼1.5 mmol),
p-nitrophenol (420 mg, 3 mmol), and DMAP (195 mg, 1.575 mmol)
were dissolved in 90 mL of dry methylene chloride. To it was
added DCC (360 mg, 1.725 mmol) in 5 mL of dry methylene
chloride. The solution was stirred under N2 overnight, and the
solvent was removed by rotary evaporation. The resulting syrup
was redissolved in 30 mL of dry toluene, and the insoluble solid
was filtered off. Then the solution was precipitated into 300 mL
of dry ethyl ether. The product was reprecipitated with ether, then
collected by filtration and dried under reduced pressure to yield
1.425 g (95%). 1H NMR (DMSO-d6): δ 3.5 (br m, PEG), 5.01 (s,
PEGCOOCH2COONP), 2.69 (t, PEGOCH2CH2COO(PA)), 8.35 &
7.48 (d & d, Ha & Hb in NP, 4H).
Synthesis of 3-Hydroxybenzylbutyrates3-Hydroxybutyric acid
sodium salt (2 g, 15.8 mmol) and benzylbromide (16 mmol) were
dissolved in 40 mL of acetonitrile. The solution was refluxing over
2 days under N2. The solid in the solution was removed by
filtration, and the solvent was condensed under reduced pressure.
The product was obtained by chromatography of a silica gel with
Rf of ∼0.45 when 50% hexane and 50% ethyl acetate were used
as eluent; yield, 2.72 g (88%). 1H NMR (DMSO-d6): δ 1.09 (d,
HOCH(CH3)CH2COOCH2C6H5), 2.40 (d, HOCH(CH3)CH2COO-
CH2C6H5), 4.03 (h, HOCH(CH3)CH2COOCH2C6H5), 4.79 (d,
HOCH(CH3)CH2COOCH2C6H5), 5.09 (s, HOCH(CH3)CH2COO-
CH2C6H5) and 7.36 (br s, HOCH(CH3)CH2COOCH2C6H5).
Determination of Hydrolysis Half-Lives of PEG Active Esters
Measurements were conducted using a HP8452a UV-vis spec-
trophotometer. In a typical experiment, 1 mg of PEG active ester
P r ep a r a tion of Difu n ction a l P EG Ester -Con ta in in g Ac-
tive Ca r bon a tesBenzyloxy-PEG-tert-Butyl Carboxymethyl Es-
ter (C6H5CH2O-PEG-OCH2COOC(CH3)3sBenzyloxy-PEG 3400
(40 g, 11.7 mmol end groups) was azeotropically dried with 250
mL of toluene under N2. After 2 h, the solution was cooled to room
temperature. Potassium tert-butoxide (2.8 g, 23.5 mmol) in 90 mL
of tert-butyl alcohol and 90 mL toluene was added to the
aforementioned PEG solution. The mixture was stirred for 2 h at
room temperature. tert-Butyl bromoacetate (4 mL, 26.3 mmol) was
added, and the solution was stirred under N2 at room-temperature
overnight. The solution was filtered, condensed under reduced
pressure, and precipitated into 300 mL of ether. The product was
collected by filtration and dried under reduced pressure at room
temperature.
R-Benzloxy-ω-carboxymethyl PEG [(C6H5CH2O-PEG-OCH2-
COOH]sBenzyloxy-PEG-tert-butyl carboxymethyl ester (20 g)
was dissolved in distilled water and the pH was adjusted to 12.0
with 1 N NaOH solution. The solution was kept at pH 12.0 for
the first 2 h by continuously adding NaOH solution (1 N) and then
stirred overnight. The pH of the solution was adjusted to 2.5 by
addition of 1 N HCl aqueous solution. The solution was extracted
three times with methylene chloride. The combined organic phase
was dried over anhydrous sodium sulfate, filtered to remove salt,
condensed by rotary evaporation, and then precipitated into 250
mL of ether. The product was collected by filtration and dried
under reduced pressure at room temperature to yield 18 g (90%).
1H NMR (DMSO-d6): δ 3.5 (br m, PEG), 4.01 (s, PEGOCH2COOH),
4.49 (s, C6H5CH2OPEG), 7.33 (br s, C6H5CH2OPEG).
Benzyloxy-PEG-CM-PEG-benzyloxy (C6H5CH2O-PEG-OCH2-
COO-PEG-OCH2C6H5)sIn a 100-mL round-bottomed flask, ben-
zyloxy-PEG carboxymethyl acid 3400 (3.4 g, 1 mmol) in toluene
was azeotropically distilled for 2 h and then cooled to room
temperature. A solution of thionyl chloride (2 M, 4 mL, 8 mmol,
Aldrich) in methylene chloride was injected and the mixture was
stirred under N2 overnight. The solvent was condensed by rotary
evaporation and the syrup was dried under reduced pressure for
∼4 h over P2O5 powder. To the residue was added 5 mL of
anhydrous methylene chloride and azeotropically dried benzyloxy-
PEG 3400 (2.55 g, 0.75 mmol) in toluene (20 mL). After benzyl-
oxy-PEG acyl chloride was dissolved, freshly distilled triethyl-
amine (0.6 mL) was added. The mixture was stirred overnight.
The triethylamine salt was removed by filtration and the product
collected by precipitation with ethyl ether. It was further purified
by dissolving in water and extracting with methylene chloride.
HPLC-GPC) of the product showed that 100% of benzyloxy-PEG
had been converted into the PEG ester and ∼15 wt % benzyloxy-
PEG acid remained.
The mixture was then purified by ion-exchange chromatography
(DEAE sepharose fast flow, Pharmacia) to remove the benzyloxy-
PEG acid. This purification yielded 100% pure R-benzyloxy-ω-
benzyloxy-PEG ester 6800 was obtained (4.1 g, 80%).
Ester-Containing PEG diol (HO-PEG-OCH2COO-PEG-OH)s
A solution of benzyloxy-PEG-CM-PEG-benzyloxy 6800 prepared
in the previous step (2 g, 0.59 mmol) in 1,4-dioxane (20 mL) was
hydrogenated with H2 (2 atm pressure) and 1 g of Pd/C (10%)
overnight. The catalyst was removed by filtration, and the solvent
was condensed by rotary evaporation. The residual syrup was
added into 100 mL of ethyl ether. The product was collected by
filtration and dried under reduced pressure to yield 1.5 g (75%).
1H NMR(DMSO-d6): δ 3.5 (br m, PEG), 4.14 (s, PEGOCH2-
COOPEG), 4.18 (t, PEGOCH2COOCH2CH2OPEG).
Difunctional Ester-Containing PEG-Succimidyl Carbonate
(NHS-OOCO-PEG-OCH2COO-PEG-OCOONHS)sEster-con-
taining PEG diol 6800 (2 g, 0.29 mmol) was azeotropically distilled
with 100 mL acetonitrile and then slowly cooled to room temper-
ature. To the resulting solution were added DSC (310 milligram,
0.58 mmol) and pyridine (0.15 mL), and the solution was stirred
at room-temperature overnight. The solvent was removed under
reduced pressure. The resulting solid was dissolved in 35 mL of
dry methylene chloride, and the insoluble solid was removed by
filtration. The filtrate was washed with pH 4.5 sodium chloride
1452 / Journal of Pharmaceutical Sciences
Vol. 87, No. 11, November 1998

Table 1sHydrolysis Half-Lives (Days, ±10%) of the Esters in 0.1 M
Phosphate Buffer (pH 7.0)
PA ester linkagea
CM ester linkageb
pKa of the acid
4.45 ± 0.1
3.67 ± 0.05
pH
5.5
7.0
8.1
5.5
7.0
8.1
room temp. (22−23 °C)
37 °C
50 °C
>500
s
s
250
43
15
37
s
s
>150
s
s
30
4
1.5
5
s
s
a PA is propionic acid. b CM is carboxymethylated PEG.
Table 2sDegradation Half-Lives of Esters and Dissolution Times of
One-Step Hydrogels (pH 7.0, 37 °C)
complete dissolution
of the gel (days)
c
gel
type
t1/2
4-arm
PEG
8-arm
PEG
(days)
a
PA
43
4
25−30
2−3
60−70
5−7
Scheme 1
b
CM
saturated acetate buffer. The organic phase was dried over
anhydrous sodium sulfate, condensed by rotary evaporation, and
precipitated into ethyl ether. The product was collected by
filtration and dried under reduced pressure to yield 1.8 g (90%).
1H NMR (DMSO-d6): δ 3.5 (br m, PEG), 4.14 (s, PEGOCH2-
COOPEG), 4.18 (t, PEGOCH2-COOCH2CH2OPEG-), 4.45 (t,
-PEGOCH2CH2OCONHS), 2.81 (s, NHS).
a Ester linkage made from PEG propionic acid (PA). b Ester linkage made
from carboxymethylated PEG (CM). c Hydrolysis half-lives of corresponding
ester linkage measured in water-soluble compounds.
Results and Discussion
1. Syn th esis a n d Ch a r a cter iza tion sOne-Step PEG
HydrogelssThe one-step PEG hydrogels were prepared in
a condensation reaction as shown in Scheme 1. The R
group in this Scheme is a branched core moiety, generally
a polyol that is ethoxylated to form a branched PEG (e.g.,
for 4-arm PEG, this core moiety is simply C, and the polyol
that is ethoxylated is pentaerythritol). The PEG carboxylic
acids of Scheme 1 were prepared using linear PEG (MW
2000 to 20 000) and the branched PEG alcohols were 4-arm
or 8-arm (MW 2000 to 15 000). For the one-step gel, the
weak linkage W in the Scheme is the ester linkage -O-
(CH2)n-CO2-CH2CH2O-, where n ) 1 (carboxymethy-
lated PEG) and n ) 2 (PEG propionic acid). Stannous
octoate (Sn(Oct)2) was chosen as a Lewis acid catalyst for
hydrogel formation. This compound has been documented
as having low toxicity45,46 and is widely used in synthesis
of biomaterials of polylactide and polyglycolide.47 Also,
stannous octoate is not covalently attached to the polymer
chains47 and thus can be removed when necessary. Indeed,
all the one-step hydrogels were washed in glacial acetic
acid to remove the catalyst. The tin content in the dry gel
dropped from 0.78 wt % before purification to 0.006 wt %
after purification according to ICP measurement.
The gels were characterized in terms of gel conversion,
swelling degree, and water content. All gels made by this
method have a high gel conversion with >92% of PEG
reactants incorporated into the gel. Also these gels have
high water content with the equilibrium state of being
>80% water. A slight temperature dependence of equilib-
rium swelling was also observed.
Hydrolysis half-lives of the ester linkages were deter-
mined by the GPC method, and are listed in Table 1. The
data are collected from hydrolysis of water-soluble ester-
containing PEG derivatrives.48
Scheme 2
of hydrolysis is also substantial; gels are most stable at
pH 5.5, and hydrolysis rate increases with pH >5.5.
Degradation of the gels was investigated by following
weight loss with time. Two distinct time periods were
observed during degradation. One can be referred to as
the “stable period” during which the gel maintains its shape
and mass loss is slow. The other is the “unstable period”
during which the gel starts to lose shape, becomes a viscous
“slime”, and mass loss is much faster. Table 2 gives the
degradation times of the gels. Note that the degree of
branching in the PEG alcohol used in preparing the gel
(4-arm or 8-arm) also influences the degradation process.
Two-Step PEG HydrogelssIn preparing a two-step hy-
drogel, an ester-containing amine-reactive PEG is first
synthesized and then coupled to an amine cross-linker. Two
types of ester-containing amine-reactive PEGs were syn-
thesized. The first type was prepared from difunctional
“double-ester” PEGs, as shown in Scheme 2.
In this synthetic route (Scheme 2), PEG has only 90% of
end groups converted into double ester as shown by NMR
analysis. To improve the yield of the desired compound,
another method was developed, as shown in Scheme 3. In
this way, the double ester PEG is >99% pure according to
NMR analysis.
As can be seen from Table 1, there is a large difference
between hydrolysis rates of the PA ester and the CM ester.
At the same temperature and pH, the CM ester hydrolyzes
7-10 times faster than the PA ester. The pH dependence
Journal of Pharmaceutical Sciences / 1453
Vol. 87, No. 11, November 1998

Figure 2sRelease of m-PEG−dyes from PEG gels made from difunctional
PEG 2k and 4-arm PEG 2k (±5%).
In all of the gel precursors, organic solvent residues are
<100 ppm in concentration according to GC-MS analysis.
The gels can be lyophylized and stored at low temperature,
and then be used after reconstitution.
Scheme 3
2. Relea se Stu d iessRelease of model drugs from the
hydrogels was studied. The first type of model drug was
an m-PEG-dye, which was synthesized in our laboratory.43
The m-PEG-dyes were loaded into the one-step hydrogels
by a diffusion process, whereas the two-step hydrogels
formed in the presence of m-PEG-dye. The release
experiment was conducted in a circulation system at 37
°C in pH 7.0 buffer. The components of the circulation
system are shown in Figure 1.
Release profiles of m-PEG-dyes from these hydrogels
are shown in Figure 2. Compared with those from the one-
step gels, the release of m-PEG-dyes is slower from the
two-step gels. Presumably this result is due to the fact
that for the two-step gels, dye is loaded throughout the gel,
not just in a layer near the outer surface, as is the case
with the diffusion-loaded, one-step gel.
The m-PEG-dyes have no active functional groups and
are not covalently attached to the hydrogel networks.
Because gel degradation is much slower than the release
process (hours days), the release of m-PEG-dyes should
be controlled only by diffusion. The diffusional release
mechanism was supported by a linear relationship between
the amount of drug released and the square root of time
for the first 60% of drug to release.43 From diffusion theory,
larger molecules have smaller diffusion coefficients and
diffuse more slowly. This phenomenon can be seen in
Figure 2, where m-PEG-dyes of higher molecular weight
are released more slowly. This phenomenon is more
obvious in the two-step gel than in the one-step gel,
probably because of the more uniform distribution of the
drugs throughout the two-step gels.
Scheme 4
As shown in the Schemes, two PEG carboxylic acids (PA
and CM) and two natural metabolites (GA and HBA) make
four combinations with four hydrolysis rates. The hydroly-
sis kinetics were determined by GPC and were reported
previously.43,48 If a protein is attached to this type of PEG
compound, hydrolysis of the ester linkage leaves a small
hydroxy “tag” attached to the protein.
PEG-COO-R-COO-NHS + NH2-protein f
PEG-COO-R-CONH-protein
PEG-COO-R-CONH-protein f
The FITC-BSA was used as a protein model drug in a
release study. It was loaded into the two-step hydrogels
in the gel formation step. As shown in Figure 3, the release
profile of FITC-BSA is approximately linear until the gel
dissolved, at which point the remaining protein is released
in a “burst”. This type of release is totally in contrast to
the diffusional release of m-PEG-dyes, which was shown
in Figure 2. Because of the presence of amino groups, the
protein is incorporated covalently into the degradable gel
matrix. Therefore, its release is controlled by hydrolysis
of the ester linkage in the tether that attaches the protein
to the hydrogel, by degradation of the matrix, and by
diffusion of unattached proteins from inside to outside of
the hydrogel.
PEG-COOH + HO-R-CONH-protein
The second type of ester-containing amine-reactive PEG
was synthesized as illustrated in Scheme 4. This com-
pound leaves a PEG chain attached to the protein after
hydrolysis, compared with the small molecule “tag” in the
previous case.
PEG-COO-PEG-OCOO-NHS + NH2-protein f
PEG-COO-PEG-OCONH-protein
PEG-COO-PEG-OCONH-protein f
PEG-COOH + HO-PEG-OCONH-protein
The two-step hydrogels were then prepared from these
PEG derivatives and multi-arm PEG amines in aqueous
solution (pH 5.5-8) at room temperature.
Although the initial “burst”, which is commonly seen in
diffusional release, disappears in this hydrolysis-controlled
1454 / Journal of Pharmaceutical Sciences
Vol. 87, No. 11, November 1998

Figure 3sRelease of FITC-BSA from two-step gels with different number of
cross-linking (±10%). The vertical dashed lines indicate the point of gel
dissolution. HAS is human serum albumin as a cross-linker.
Figure 5sRelease of FITC-BSA from different thickness gels (±10%).
gel by hydrolysis of the weak linkage in the tether. As
shown in Figure 4, covalent attachment to the gel signifi-
cantly alters release kinetics from a diffusion-controlled
case, and thus this approach could be used to control
release of proteins. In general, the release profile is
affected by the number of attachments of protein drug, gel
degradation time, and diffusion rate. The maximum
number of attachment sites depends on the number of
accessible reactive groups on the drug. For each drug, the
number of attachments is achieved in a random fashion.
However, the distribution of drug as a function of number
of attachments per drug follows statistical law.
Once a protein is attached to the hydrogel, it must be
released first by degradation of the linkage inside the gel
and then by a diffusion process to the outside of the gel. A
few papers have studied the effect of linkage cleavage on
drug release.49-52 Pitt and co-workers53,54 mathematically
analyzed drug release from a polymeric system that
combined drug cleavage and diffusion. In their work, the
drug was attached to the polymer through only one labile
linkage, so that drug liberation followed first-order kinetics.
Mathematically, Pitt and co-workers53 analyzed the diffu-
sion of a drug formed by first-order cleavage of drug-
polymer bonds in a polymer film with the following
equation:
Figure 4sRelease of FITC-BSA from CM-HBA gels with different gel formation
times with 8-arm PEG amine 10k as cross-linker (±10%).
process, there was also an undesirable late “burst” caused
by gel dissolution. If the drug can be released before gel
dissolution, the late “burst” will be avoided.
Several factors can be adjusted to remove the burst
release. As shown in Figure 3, increased gel degradation
life helps to minimize the late burst. Reducing the degree
of attachment of protein to the network is also effective.
As shown in Figure 4, speeding up the time for gel
formation results in fewer covalent attachments of the
FITC-BSA and thus a faster release. A small portion of
initial diffusional release is observed in this case. Thus it
appears that if the gelation condition is appropriate, release
of protein can approach zero-order kinetics.
In a two-step hydrogel, if all the protein is attached to
the hydrogel network and if diffusion is much faster than
the hydrolysis process, the total release profile should
follow the kinetics of degradation of the tether as well as
of the gels. Consequently, the release profile will be
independent of gel thickness, but dependent on tempera-
ture and pH in vitro. Because drug delivery devices are of
interest at physiological conditions (pH ∼7 and 37 °C),
release studies were conducted only at 37 °C and pH 7. In
Figure 5 are shown the release profiles of FITC-BSA from
two different thickness of the same two-step PEG hydrogels
made from difunctional CM-HBA double-ester and 8-arm
PEG. Because gel size is not a factor, it is possible with
two-step hydrogel to achieve controlled release of proteins
in an injectable microgel formulation. This subject will be
the focus of future work.
∂C(x,t)
∂t
∂C(x,t)
∂x
∂
)
D
+ kCbe-kt
(1)
(
)
0
∂x
where C(x,t) is local concentration of free drug inside the
film, D is the diffusion coefficient of the free drug, Cb0 is
the uniform initial concentration of bound drug, k is the
rate constant of the first-order hydrolysis, t is time, and x
is the distance from one surface. It is important to note
that bound drugs and free drug inside the film are
considered as two different species.
Equation 1 describes the case in which there is only one
hydrolytic weak linkage along each tether, and the bonded
protein is attached through only one tether, as shown in
Scheme 5a. This type of gel is formed in two-step hydrogels
prepared with ester-containing PEG carbonates (the second
type of ester-containing amine-reactive compound).
If two degradable linkages are present in one tether, such
as in a difunctional double-ester PEG that has two esters
in each PEG chain, breaking any one of them will lead to
release of the protein (Scheme 5b). Note that breaking the
linkage near the protein leads to release of protein with a
small tag on it, whereas breaking the linkage at the far
end results in protein attached with a degradable PEG
chain. These two kinds of released proteins have different
molecular weights and thus should have different diffusion
coefficients. However, for large proteins and small PEGs,
the difference in diffusion may be negligible. In this case,
the diffusion of a unattached drug formed by cleavage of
drug-polymer bonds in a gel film follows eq 2, which differs
3. Ma th em a tica l Discu ssion of P r otein Relea se
fr om th e Tw o-Step Hyd r ogelssBecause of the amino
groups in a protein, in situ loading of protein drugs in the
two-step hydrogels results in covalent attachment of
protein drugs to the hydrogel network. These proteins are
attached to the hydrogel through hydrolytically degradable
PEG tethers, and therefore they could be released from the
Journal of Pharmaceutical Sciences / 1455
Vol. 87, No. 11, November 1998

Figure 6sComputer simulation of release profile of proteins from a two-step
hydrogel. Curve A: assuming 20% free protein, 70% single-PEG attached
protein, and 10% double-PEG attached protein; Curve B: assuming no free
protein, 10% single-PEG attached protein, 60% double-PEG attached protein,
and 40% triple-PEG attached protein.
Scheme 5
By means of computer simulation, release profiles of
FITC-BSA that was loaded in situ into the two-step gels
can be theoretically evaluated. However, it is still difficult
to precisely predict the release of the proteins without doing
release experiments. The big hurdle in prediction is to
obtain the precise distribution of proteins as a function of
the number of attachments.
from eq 1 simply by multiplying k by two.
∂C(x,t)
∂t
∂C(x,t)
∂x
∂
)
D
+ 2kC e-2kt
(2)
(
)
0
∂x
In a two-step gel-protein system, some proteins are
attached to the hydrogel network by more than two tethers.
In this case, the diffusion of an unattached protein formed
by cleavage of drug-polymer bonds in such a gel film can
be described as in eq 3.
Conclusion and Future Work
A series of degradable PEG hydrogels have been devel-
oped with excellent potential to serve as drug delivery
carriers. These gels are hydrolytically degradable, and the
degradation products are nontoxic PEG derivatives. The
first type of hydrogel is prepared by simple one-step
polycondensation of difunctional PEG acid and branched
PEG polyol. The second type is prepared in a two-step
process. In the first step, amine-reactive, ester-containing
PEG derivatives were synthesized. These derivatives react
with branched PEG amines to form a gel in the second step.
These hydrogels can release physically entrapped drugs in
a sustained manner by a diffusion mechanism. In addition,
the two-step gels can covalently link protein drugs to the
gel network, and therefore release of the drug is controlled
by degradation of the linkages as well as by diffusion.
Mathematical analysis shows that covalent attachment
provides control of protein release. Because release of
attached proteins is independent of dimension, two-step
gels can be formulated into injectable microgels, which will
be studied in the future. Also, the stability of proteins in
terms of conformational integrity and bioactivity during
both incorporation and release will be investigated.
∂C(x,t)
∂t
∂C(x,t)
∂x
∂
n-1
)
D
+
nmkCb0ne-mkt(1 - e-mkt
)
∑
(
)
∂x
n)1
(3)
where Cb0n is the initial concentration of protein bonded
with n PEG tethers, and m is the number of weak linkages
along each tether. In the two-step gels, m is either 1 or 2.
Solution of eq 3 can be complicated. However, if diffusion
of protein is much faster than hydrolysis, the release profile
can be simplified into eq 4. In obtaining eq 4, it was
assumed that the amount of drug generated from cleavage
of the polymer-drug bonds is negligible during the release
of initial free drugs, and it was assumed that diffusivity of
the drug and gel thickness do not change during the initial
diffusion period.
(2i + 1)2π2Dt
∞
8
1
Mt ) M00(1 -
exp -
+
∑
(
)
π2
(2i + 1)2
L2
i)0
Mb0n (1 - e-mkt
)
(4)
n
∑
n)1
Here, M00 is the initial amount of free protein and Mb0n is
the initial amount of protein attached by n PEG tethers.
If the initial concentration of different types of proteins
is known, the release profile can be simulated according
to eq 4. Figure 6 shows the simulation results of release
profiles of proteins with different numbers of attachments
from a double-ester gel, in which diffusion is a much faster
process than hydrolysis of weak linkages and degradation
of gel.
If gel degradation is taken into consideration, the release
profile needs to be reevaluated. For example, if the gel
starts to dissolve from 200 to 280 h, curve A will become
more like the experimental result shown in Figure 4 (short
gelation time). Similarly, curve B will become very close
to the release profile of the long gelation time of Figure 4
if the gel dissolves sharply from 150 to 200 h.
References and Notes
1. Bell, C. L.; Peppas, N. A. Poly(methacrylic acid-g-ethylene
glycol) Hydrogels as pH Responsive Biomedical Materials.
N. A. Mater. Res. Soc. Symp. Proc. 1994, 331, 199-204.
2. Hydrogels in Medicine and Pharmacy; Volumes I, II, and III,
Peppas, N. A., Ed.; CRC: Boca Raton, FL, 1986.
3. Hennink, W. E.; Talsma, H.; Borchert; J . C. H.; De Smedt,
S. C.; Demeester, J . Controlled Release of Proteins from
Dextran Hydrogels. J . Controlled Release 1996, 39, 47-55.
4. Emara, L. H. The Development of Slow-Release Molluscicide
in a Biodegradable Gelatin Matrix. J . Controlled Release
1994, 31, 255-261.
5. Li, J . K.; Wang, N.; Wu, X. S. Gelatin Hydrogel Nanoparticles
as a Protein Drug Carrier. Proceeding 29th Int. Biomater.
Symp., New Orleans, LA, 1997; p 199.
6. Chun, S.-W.; Kim, J .-D. A Novel Hydrogel-dispersed Com-
posite Membrane of Poly(N-isopropylacrylamide) in a Gelatin
Matrix and Its Thermal Actuated Permeation of 4-Acetami-
dophen. J . Controlled Release 1996, 38, 39-47.
1456 / Journal of Pharmaceutical Sciences
Vol. 87, No. 11, November 1998

7. Milojevic, S.; Newton, J . M.; Cummings, J . H.; Gibson, G.
R.; Botham, R. L.; Ring, S. G.; Stockham, M.; Allwood, M. C.
Amylose as a Coating for Drug Delivery to the Colon:
Preparation and in vitro Evaluation Using 5-Aminosalicylic
Acid Pellets. J . Controlled Release 1996, 38, 75-84.
8. Milojevic, S.; Newton, J . M.; Cummings, J . H.; Gibson, G.
R.; Botham, R. L.; Ring, S. G.; Stockham, M.; Allwood, M. C.
Amylose as a Coating for Drug Delivery to the Colon:
Preparation and in vitro Evaluation Using Glucose Pellets.
J . Controlled Release 1996, 38, 85-94.
29. McNeil, M. E.; Graham, N. B. Properties Controlling the
Diffusion and Release of Water Soluble Solutes from Poly-
(ethylene oxide) Hydrogels. In Polymer Biomaterials in
Solution, as Interfaces and as Solids; Cooper, S. L.; Bamford,
C. H.; Tsuruta, T., Eds.; VSP: Utrecht, The Netherland,
1995, pp 971-1008.
30. Bromberg, L. Cross-linked Poly(ethylene glycol) Networks as
Reservoirs for Protein Delivery. J . Appl. Polym. Sci. 1996,
59, 459-466.
31. Carelli, V.; Di Colo, G.; Nannipieri, E.; Serafini, M. F. Study
of Controlled-Release Systems for Progesterone Based on
Cross-linked Poly(ethylene oxides). Int. J . Pharm. 1993, 94,
103-113.
32. Gayet, J .-C.; Fortier, G. High Water Content BSA-PEG
Hydrogel for Controlled Release Device: Evaluation of the
Drug Release Properties. J . Controlled Rel. 1996, 38, 177-
184.
33. Barrow, T. H.; Truong, M. T.; Lewis, T. W.; Grussing, D. M.;
Kato, K. H.; Gysbers, J . E.; Lamprecht, E. G. Evaluation of
a New Tissue Sealant Material: Serum Albumin Cross-
linked In Vivo with Polyethylene Glycol, Proc. 5th World
Biomater. Congr. Toronto, Canada, 1996, #8.
9. Moussa, I. S.; Cartilier, L. H. Characterization of Moving
Fronts in Cross-linked Amylose Matrixes by Image Analysis.
J . Controlled Release 1996, 42, 47-55.
10. Tahara, K.; Yamamoto, K.; Nishihata, T. Overall Mechanism
Behind Matrix Sustained Release (SR) Tablets Prepared with
Hydroxypropyl Methylcellulose 2910. J . Controlled Release
1995, 35, 59-66.
11. Yao, K. D.; Peng, T.; Feng, H. B.; He, Y. Y. Swelling Kinetics
and Release Characteristics of Cross-linked Chitosan: Poly-
ether Polymer Network (Semi-IPN) Hydrogels. J . Polym. Sci.,
Part A: Polym. Chem. 1994, 32, 1213-1223.
12. Ho, H.-O.; Hsiao, C.-C.; Sokoloski, T. D.; Chen, C.-Y.; Sheu,
M.-T. Fibrin-based Drug Delivery Systems III: The Evalu-
ation of the Release of Macromolecules from Microbeads. J .
Controlled Release 1995, 34, 65-70.
13. Gilbert, D. L.; Kim, S. W. Macromolecular Release from
Collagen Monolithic Devices. J . Biomed. Mater. Res. 1990,
24, 1221-1239.
14. Singh. M. P.; Lumpkin, J . A.; Rosenblatt, J . Effect of
Electrostatic Interactions on Polylysine Release Rates from
Collagen Matrixes and Comparison with Model Predictions.
J . Controlled Release 1995, 35, 165-179.
15. Aslani, P.; Kennedy, R. A. Studies on Diffusion in Alginate
Gels. I. Effect of Cross-linking with Calcium or Zinc Ions on
Diffusion of Acetaminophen. J . Controlled Release 1996, 42,
75-82.
16. Murata, Y.; Miyamoto, E.; Kawashima, S. Additive Effect of
Chondroitin Sulfate and Chitosan on Drug Release from
Calcium-induced Alginate Gel Beads. J . Controlled Release
1996, 38, 101-108.
17. Peppas, N. A. Hydrogels of Poly(Vinyl Alcohol) and its
Copolymers. In Hydrogels in Medicine and Pharmacy, Vol-
ume II; Peppas, N. A., Ed.; CRC: Boca Raton, FL, 1987; pp
1-48.
18. Peppas, N. A.; Moynihan, H. J . Structure and Physical
Properties of Poly(2-Hydroxyethyl Methacrylate) Hydrogels.
In Hydrogels in Medicine and Pharmacy; Volume II; Peppas,
N. A., Ed.; CRC: Boca Raton, FL, 1987; pp 49-64.
19. Mack, E. J .; Okano, T.; Kim, S. W. Biomedical Applications
of Poly(2-Hydroxyethyl Methacrylate) and its Copolymers.
In Hydrogels in Medicine and Pharmacy; Volume II; Peppas,
N. A., Ed.; CRC: Boca Raton, FL, 1987; pp 65-93.
20. Dong, L. C.; Hoffman, A. S.; Yan, Q. Dextran Permeation
through Poly(N-isopropylacrylamide) Hydrogels. In Polymer
Biomaterials in Solution, as Interfaces and as Solids; Cooper,
S. L.; Bamford, C. H.; Tsuruta, T., Eds.; VSP: Utrecht, The
Netherland, 1995; pp 1009-1020.
21. French, D.; Himmelstein, K.; Mauger, J . W. Physicochemical
Aspects of Controlled Release of Substituted Benzoic and
Naphthoic Acids from Carbopol Gels. J . Controlled Release
1995, 37, 281-289.
22. Giammona, G.; Pitarresi, G.; Tomarchio, V.; Cavallaro, G.;
Mineo, M. Cross-linked R,â-Polyasparthydrazide Hydro-
gels: Effect of Cross-linking Degree and Loading Method on
Cytarabine Release Rate. J . Controlled Release 1996, 41,
195-203.
23. Graham, N. B. Poly(Ethylene Oxide) and Related Hydrogels.
In Hydrogels in Medicine and Pharmacy; Volume II; Peppas,
N. A., Ed.; CRC: Boca Raton, FL, 1987; pp 95-113.
24. Graham, N. B. Poly(Ethylene Glycol) Gels and Drug Delivery.
In Poly(ethylene glycol) Chemistry. Biotechnical and Bio-
medical Applications; Harris, J . M., Ed.; Plenum: New York,
1992; pp 263-281.
25. Poly(ethylene glycol) Chemistry. Biotechnical and Biomedical
Applications; Harris, J . M., Ed., Plenum: New York, 1992.
26. Zalipsky, S.; Harris, J . M. Introduction to Chemistry and
Biological Applications of Poly(ethylene glycol). In Poly(eth-
ylene glycol) Chemistry and Biological Applications; Harris,
J . M.; Zalipsky, S., Eds.; ACS Symposium Series 680,
Washington, DC, 1997; pp 1-13.
27. Villain, F. L.; Parel, J .-M. A.; Lee, W. G.; Simon, G. Injectable
Poly(ethylene oxide) Gel Implant and Method for Production.
U.S. Patent 5, 634, 943, 1997.
34. Sawhney, A. S.; Pathak, C. P.; Hubbell, J . A. Bioerodible
Hydrogels Based on Photopolymerized Poly(ethylene glyol)-
co-poly(a-hydroxy acid) Diacrylate Macromers. Macromol-
ecules 1993, 26, 581-587.
35. Han, D. K.; Hubbell, J . A. Lactide-Based Poly(ethylene glycol)
Polymer Networks for Scaffolds in Tissue Engineering.
Macromolecules 1996, 29, 5233-5235.
36. Sawhney, A. S.; Pathak, C. P.; van Rensburg, J . J .; Dunn,
R. C.; Hubbell, J . A. Optimization of Photopolymerized
Bioerodible Hydrogel Properties for Adhesion Prevention. J .
Biomed. Mater. Res. 1994, 28, 831-838.
37. Nathan, A.; Bolikal, D.; Vyavahare, N.; Zalipsky, S.; Kohn,
J . Hydrogels Based on Water-soluble Poly(ether urethanes)
Derived from L-Lysine and Poly(ethylene glycol). Macromol-
ecules 1992, 25, 4476-4484.
38. Llanos, G. R.; Sefton, M. V. Immobilization of Poly(ethylene
glycol) onto a Poly(vinyl alcohol) Hydrogel: 1. Synthesis and
Characterization. Macromolecules 1991, 24, 6065-6072.
39. Llanos, G. R.; Sefton, M. V. Immobilization of Poly(ethylene
glycol) onto a Poly(vinyl alcohol) Hydrogel: 2. Evaluation of
Thrombogenicity. J . Biomed. Mater. Res. 1993, 27, 1383-
1391.
40. Hoffman, A.; Saito, H.; Harris, M. J . Drug Delivery from
Biodegradable PEG Hydrogels with Schiff Base Linkages.
Proceedings of 24th Int. Symp. Contol. Release Bioact. Mater.,
Stokholm, Sweden, 1997.
41. West, J . L.; Hubbell, J . A. Proteolytically Degradable Hy-
drogels. Proc. 23th Int. Symp. Control. Release Bioact. Mater.,
Kyoto, J apan, 1996, pp 224-225.
42. Andreopoulos, F. M.; Deible, C. R.; Stauffer, M. T.; Weber,
S. G.; Wagner, W. R.; Beckman, E. J .; Russell, A. J .
Photoscissable Hydrogel Synthesis via Rapid Photopolym-
erization of Novel PEG-based Polymers in the Absence of
Photoinitiators. J . Am. Chem. Soc. 1996, 118, 6235-6240.
43. Zhao, X. Ph.D. Dissertation; University of Alabama in
Huntsville, 1997.
44. Emoto, K.; Harris, J . M.; Van Alstine, J . M. Grafting Poly-
(ethylene glycol) Epoxide to Amino-Derivatized Quartz:
Effect of Temperature and pH on Grafting Density. Anal.
Chem. 1996, 68, 3751-3757.
45. Tanzi, M. C.; Verderio, P.; Lampugnani, M. G.; Resnati, M.;
Dejana, E.; Sturani, E. Cytotoxicity of Some Catalysts
Commonly Used in the Synthesis of Copolymers for Biomedi-
cal Use. J . Mater. Sci.: Mater. Med. 1994, 5, 393-396.
46. Sodergard, A. Chemical and Morphological Changes in
Peroxide-Stabilized Poly(L-Lactide). In Hydrogels and Bio-
degradable Polymers for Bioapplications; Ottenbrite, R. M.;
Huang, S. J .; Park, K., Eds.; ACS: Washington D. C., 1996;
pp 103-117.
47. Nijenhuis, A. J .; Grijpma, D. W.; Pennings, A. J . Lewis Acid-
Catalyzed Polymerization of L-Lactide. Kinetics and Mech-
anism of the Bulk Polymerization. Macromolecules 1992, 25,
6419-6424.
48. Zhao, X.; Harris, J . M. Novel Degradable Poly(ethylene
glycol) Esters for Drug Delivery. In Poly(ethylene glycol)
Chemistry and Biological Applications; Harris, J. M.; Zalipsky,
S., Eds.; ACS Symposium Series 680, Washington, DC, 1997;
pp 458-472.
49. Shah, S. S.; Kulkarni, M. G.; Mashelkar, R. A. Release
Kinetics of Pendent Substituted Bioactive Molecules from
Swellable Hydrogels: Role of Chemical Reaction and Dif-
fusive Transport. J . Membr. Sci. 1990, 51, 83-95.
50. Shah, S. S.; Kulkarni, M. G.; Mashelkar, R. A. A Mechanistic
Interpretation of the Zero Order Release from Pendent
Chain-linked Glassy and Swollen Hydrogels. J . Appl. Polym.
Sci. 1991, 41, 2437-2451.
28. Royappa, A. T.; Lopina, S. T.; Cima, L. G. Synthesis and
Characterization of Tetraethylene Glycol-Poly(ethylene ox-
ide) Hydrogels for Tissue Engineering Applications. Mater.
Res. Soc. Symp. Proc. 1994, 331, 245-250.
Journal of Pharmaceutical Sciences / 1457
Vol. 87, No. 11, November 1998

51. Shah, S. S.; Kulkarni, M. G.; Mashelkar, R. A. Zero-order
Release of Pendent Substituted Active Ingredients from
Swollen Hydrogel Matrixes. J . Controlled Release 1990, 12,
155-158.
54. Pitt, C. G.; Shah, S. S. Manipulation of the Rate of Hydrolysis
of Polymer-drug Conjugates: the Degree of Hydration. J .
Controlled Release 1995, 33, 397-403.
52. Shah, S. S.; Kulkarni, M. G.; Mashelkar, R. A. Swellable
Hydrogel Matrixes for Release of Pendent Chain-linked
Active Ingredients over Extended Time Periods. J . Appl.
Polym. Sci. 1991, 43, 1879-1884.
53. Pitt, C. G.; Shah, S. S. The Kinetics of Drug Cleavage and
Release from Matrixes Containing Covalent Polymer-drug
Conjugates. J . Controlled Release 1995, 33, 391-395.
Acknowledgments
X. Zhao gratefully acknowledges the financial support of this
work by Shearwater Polymers, Inc.
J S980065O
1458 / Journal of Pharmaceutical Sciences
Vol. 87, No. 11, November 1998