Oligo-α-hydroxy ester cross-linkers: Impact of cross-linker structure on biodegradable hydrogel networks

Article abstract of DOI:10.1021/ma0518306

We describe the synthesis of a series of biodegradable oligo-α-hydroxy ester cross-linkers and evaluate their impact on the degradation kinetics and macromolecule diffusion from a hydrogel network. By changing the steric and electronic environment at the site of degradation in the cross-linker, we were able to modulate the degradation, swelling kinetics, and corresponding release profiles of macromolecules from poly(HPMA) hydrogel networks under physiologically relevant conditions. As the steric hindrance and electron demand at the site of hydrolysis for three different cross-linkers was increased, the total time for the hydrogel network to completely dissolve increased from 2 to over 30 days while incubated in pH 7 buffer. As the number of hydrolyzable sites and the electron demand at the side of hydrolysis decreased, the time to completely dissolve decreased from weeks to several days. Increasing the cross-linking density for one of the degradable cross-linkers (1.5% to 3.0% feed ratio) increased the degradation time by several weeks. Burst release was absent for high molecular weight solutes because the release rate depended on controlled degradation of the polymer network and an increase in average network mesh size. The synthetically adaptable cross-linkers described herein offer a new approach for controlling the rate and extent of release from biodegradable hydrogel networks.

Full text of DOI:10.1021/ma0518306

Macromolecules 2005, 38, 10757-10762
10757
Oligo-R-hydroxy Ester Cross-Linkers: Impact of Cross-Linker Structure
on Biodegradable Hydrogel Networks
Kenneth D. Eichenbaum,† Allen A. Thomas,‡ Gary M. Eichenbaum,§
Brian R. Gibney,| David Needham,§ and Patrick F. Kiser*,§,
Department of Materials Science and Engineering, University of Pennsylvania,
Philadelphia, Pennsylvania 19104; Access Pharmaceuticals, 2600 Stemmons Frwy,
Dallas, Texas 75207; Department of Mechanical Engineering and Materials Science,
Duke University, Durham, North Carolina 27708; Department of Biophysics, University of
Pennsylvania, Philadelphia, Pennsylvania 19104; and Department of Bioengineering,
University of Utah, Salt Lake City, Utah 84112
Received August 18, 2005; Revised Manuscript Received October 6, 2005
ABSTRACT: We describe the synthesis of a series of biodegradable oligo-R-hydroxy ester cross-linkers
and evaluate their impact on the degradation kinetics and macromolecule diffusion from a hydrogel
network. By changing the steric and electronic environment at the site of degradation in the cross-linker,
we were able to modulate the degradation, swelling kinetics, and corresponding release profiles of
macromolecules from poly(HPMA) hydrogel networks under physiologically relevant conditions. As the
steric hindrance and electron demand at the site of hydrolysis for three different cross-linkers was
increased, the total time for the hydrogel network to completely dissolve increased from 2 to over 30
days while incubated in pH 7 buffer. As the number of hydrolyzable sites and the electron demand at the
side of hydrolysis decreased, the time to completely dissolve decreased from weeks to several days.
Increasing the cross-linking density for one of the degradable cross-linkers (1.5% to 3.0% feed ratio)
increased the degradation time by several weeks. Burst release was absent for high molecular weight
solutes because the release rate depended on controlled degradation of the polymer network and an
increase in average network mesh size. The synthetically adaptable cross-linkers described herein offer
a new approach for controlling the rate and extent of release from biodegradable hydrogel networks.
Introduction
macromolecular cross-linkers have been synthesized
which contain short oligomeric sections of R-hydroxy
esters terminated in vinyl groups.13 Similarly, small
molecule cross-linkers composed of acid labile groups,18
anhydrides,19,20 and peptides21 have been studied for
anticancer, dental, and orthopaedic applications.
Biodegradable hydrogels offer a promising approach
for the controlled release delivery of small molecules and
macromolecules.1,2 They offer a safe and effective means
for controlling the release kinetics of entrained agents
and the biodistribution of a drug.3-5 The cross-linker
plays an important role in determining the swelling and
pore size of the hydrogel matrix, making it an ideal
target for synthetic modifications to control the corre-
sponding release kinetics of entrained molecules.6-8
A significant amount of work has been directed
toward developing cross-linker technology for biomedical
applications.2,9 Previous research has demonstrated that
the mechanism of solute release from degradable cross-
linked polymers is impacted by a variety of factors,
including pH, solute molecular weight, cross-link den-
sity, enzymes, and hydrolytically labile spacers.1,7,8,10,11
Water-soluble biodegradable macromolecular cross-link-
ers have been the subject of study by several groups.
These polymers are designed to contain R-hydroxy ester
monomers, such as lactate and glycolate esters for
degradability,5 and are attached to poly(ethylene glycol)
(PEG)12 or dextran.6,7,13-17 They are particularly well
suited for biological applications because they are
readily synthesized and degrade under physiological
conditions into biocompatible products. More recently,
In this work, we describe a new class of synthetic
cross-linkers composed of symmetrical oligo-glycolate
and oligo-lactate esters terminated with vinylic 2-hy-
droxypropyl methacrylamide (HPMA) polymerizable
moieties22 (Figure 1). The incorporation of these cross-
linkers, which have homogeneous degradation kinetics,
into a hydrogel network permits control over the deg-
radation and release profile of the bulk polymer under
biologically relevant conditions (Figure 1a-c). We se-
lected the carboxylic ester functionality to be the cleav-
able moiety because this ester has a well-understood
mechanism of hydrolysis,23 and its kinetics can be
controlled under physiological conditions by varying the
electronic and steric factors at the site(s) of hydrolysis.
Here we report a series of homologous cross-linkers that
is readily synthesized and purified, is comprised of
biocompatible components, and can be easily adjusted
to degrade with a range of time constants under
physiological conditions. We then designed these cross-
linkers (Figure 1d) based on the hypothesis that by
varying the steric environment of the cleavable ester
moiety and the number of such moieties (Table 1) within
a cross-linker, as well as the cross-link density within
the network, we could create hydrogel systems with a
wide range of degradation and release rates.2,11,13,24
Advantages of these small molecule cross-linkers in
comparison to polydispersed macromolecular cross-
linkers are their ease of synthesis and characterization
† Department of Materials Science and Engineering, University
of Pennsylvania.
‡ Access Pharmaceuticals.
§ Duke University.
| Department of Biophysics, University of Pennsylvania.
University of Utah.
* To whom correspondence should be addressed. E-mail:
patrick.kiser@utah.edu.
10.1021/ma0518306 CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/03/2005

10758 Eichenbaum et al.
Macromolecules, Vol. 38, No. 26, 2005
allowed to warm to 25 °C, and then HPMA (2.57 g, 17.0 mmol)
was added. The reaction was stirred at 25 °C for 2 h and then
washed with 1 M NaH2PO4 (10 mL), saturated Na2CO3, (10
mL), and brine (10 mL). The dichloromethane phase was then
dried (Na2SO4) and concentrated in vacuo to give a pale yellow,
viscous oil. Yield of 2a: 4.08 g (95%). Although the purity was
>90% by TLC and NMR, the purity could be improved by flash
chromatography. Elution on silica gel (300 mL) using 3%
methanol/dichloromethane resulted in 3.22 g (75%) of 2a: [R]D
1
) -21.3 (c ) 1.0, CHCl3). H NMR (CDCl3): δ 1.24-1.29 (m,
6H), 1.47-1.51 (m, 6H), 1.96 (s, 6H), 2.70-2.74 (m, 4H), 3.20-
3.38 (m, 2H), 3.57-3.72 (m, 2H), 4.87-5.00 (m, 2H), 5.03-
5.16 (m, 2H), 5.33-5.36 (m, 2H), 5.71-5.75 (m, 2H), 6.25-
6.55 (m, 2H). 13C NMR (62.9 MHz, DMSO-d6, several peaks
exhibited duality which may be due to diastereomers): δ 16.53,
17.21, 17.37, 18.55, 28.20, 42.99, 54.88, 68.74, 70.17, 70.22,
119.11, 139.83, 139.87, 167.68, 167.83, 169.72, 169.89, 171.27,
171.35. HRMS (FAB+) Calcd for C24H27N2O10 (M + H)
513.2448. Found: 513.2418. The complete synthetic methodol-
ogy for each cross-linker will be published elsewhere.
Figure 1. (a) Schematic diagram showing a cross-linked
element of a degradable hydrogel: (I) linear polymer backbone
p(HPMA); (II) the degradable region of the cross-linkers; (III)
water which swells the gel and aids in degradation; (IV)
entrained macromolecule. (b) Degradation of the polymer and
an increase in the mesh size of the polymer. (c) Final product
of linear polymer (V) and release of macromolecule. (d) General
structure of the cross-linkers in which n and R refer to the
number of R-hydroxy esters on each side of the diacid and
substitution at the R-hydroxy ester, respectively.
Gel Synthesis and Testing. The vinyl groups on the
terminus of the cross-linking monomer can be used to form a
gel network structure. Gels were synthesized using the am-
monium persulfate (APS) N,N,N′,N′-tetramethylethylenedi-
amine (TMED) couple as the initiation system.6,11 The gels
described below were made at a mole feed ratio of 1.5 mol %
cross-linker as a copolymer with 98.5 mol % HPMA. Before
the gels were polymerized, three 1.0 mL plastic syringes, to
be used as a slab gel template, were silylanized by briefly
incubating them in a heptane solution containing Sigmacote
and oven-drying at 90 °C. A representative procedure to form
gels was as follows: a 7 mL test tube was charged with HPMA
(2.115 g, 14.8 mmol, [HPMA]final ) 5 M), and the cross-linker
was placed on the end of a tarred spatula (109.0 mg, 0.225
mmol, [XL]final ) 0.075 M). The end of the spatula was
submerged in the HPMA, and DI water (1.5 mL) was added
to the mixture and then was sonicated until all of the material
was dissolved. To this solution was added a solution of APS
in water (99 mg, 0.438 mmol, 166 µL of a 2.63 M solution,
[APS]final ) 0.143 M). To this mixture was then added TMED
to initiate the polymerization (49 mg, 0.429 mmol, 204 µL of
a 2.10 M solution of TMED adjusted to pH 7 with HCl).
Immediately after the TMED was added, the mixture of
monomers and APS was vigorously mixed on a vortexer for
15 s and then drawn into the 1.0 mL plastic syringes (which
act as a mold for gel formation) by plunger aspiration. The
gels were allowed to polymerize for 4 h. The gels were cut into
10 mm tall cylinders and were placed in 15 mL vials containing
the desired buffer (10 mL). The initial dry mass of the gel was
determined by drying seven of the gels from each composition
in their relaxed state. The incubation solutions were changed
each time the gel was weighed. The gels were incubated in a
gyratory water-bath shaker. The temperature was regulated
to be 37 ( 2 °C, and the shaker was set to 30 rpm. The change
in volume of the polymer network was measured by weighing
the gel at different time points. Knowing the initial mass of
the dry polymer and the final swollen mass, the swelling ratio
Qv can be determined by eq 110
Table 1. Description of the Cross-Linkers
compd
ID
degradation
rate
n
R
name
1
0
1
1
2
NR
CH3
H
HPMA2succinate
ND, control
slowest
medium
fastest
2a
2b
3
(HPMALac)2succinate
(HPMAGly)2succinate
(HPMAGlyGly)2succinate
H
which could be of benefit in fundamental studies of
degradable polymer network behavior and in regulatory
considerations related to devices built with these sub-
stances.
Experimental Section
Materials and Instrumentation. All chemicals were
reagent grade and were used without purification unless
otherwise noted. Dichloromethane was distilled from P2O5 and
stored over 4 Å molecular sieves. All other solvents were
obtained in their anhydrous state or stored over 4 Å molecular
sieves before use. 1H NMR and 13C NMR spectra were recorded
at 400 and 100.4 MHz, respectively, on a Varian INOVA-400
spectrometer. Solvent mixtures are given in volume-to-volume
ratios unless otherwise stated. Flash chromatography was
performed on SiO2 Kieselgel 60 (70-230 mesh, E. Merck). The
FITC dextrans used in release studies were obtained from
Sigma-Aldrich.
Cross-Linker Synthesis. 2 equiv of benzyl lactate (or
benzyl glycolate) is reacted with succinyl chloride in the
presence of pyridine, providing a diester (see Figures 1d and
2). Reductive deprotection of the carboxylate groups of the
diester renders a dicarboxylic acid. Addition of HPMA to the
dicarboxylic acid activated via carbonyl diimidazole (CDI) pro-
vides a diacrylate (HPMALac)2succinate, 2a, and (HPMAGly)2-
succinate, 2b. The length of the cross-linker can be extended
by reacting the diacid intermediate with an additional 2 equiv
of benzyl glycolate (or benzyl lactate) and repeating the
deprotection step and lastly coupling with HPMA, forming for
example (HPMAGlyGly)2succinate, 3. Cross-linker 1, which is
lacking lactate or glycolate ester moieties, was prepared as a
control compound since its hydrolysis rate at pH 7 was
expected to be negligible.25 A sample preparation for the
degradable cross-linker 2a in Figure 1 is given below.
Preparation of (HPMALac)2Succinate, 2a. Compound
5a (2.2 g, 8.3 mmol) was dissolved in dichloromethane (30 mL)
and cooled to 0 °C under an argon atmosphere in a three-
necked flask equipped with a stir bar and a powder addition
funnel. The reaction vessel was then charged with carbonyl
diimidazole (CDI, 2.75 g, 17.0 mmol) via the powder addition
funnel (warning: CO2 gas is released). The reaction vessel was
-1
Vs V2 + V
(mt - m2)F1
Qv ) ν2,s
)
)
1 ) 1 +
(1)
-1
-1
V2
V2
m2F2
where ν2,s-1 is the polymer volume fraction in the swollen state,
Vs is the total volume of the gel in the swollen state, V2 is the
initial volume of the dry gel network, V1 is the volume of water
entrained in the gel, mt is the total mass of the gel, F1 is the
density of water, m2 is the initial dry mass of polymer, and F2
is the density of the polymer.26,27 Therefore, by measuring the
mass of the swelling gel during its degradation, the swelling
kinetics were obtained.
Dextran Solution Preparation, Calibration, and Re-
lease Measurements. Three separate stock solutions con-
taining FITC-labeled dextrans of 12, 42, and 148 kDa were
prepared at concentrations of 28.5, 20.9, and 24.1 mg mL-1
,

Macromolecules, Vol. 38, No. 26, 2005
Oligo-R-hydroxy Ester Cross-Linkers 10759
respectively. These solutions were then used as above to
construct the monomer solutions for polymerization. The
release of FITC dextran was measured by UV/vis spectropho-
tometry (Perkin-Elmer Lambda 2S UV/vis spectrophotometer).
Samples of ∼1 mL of solution were drawn into cuvettes, and
the absorption spectrum for each sample was measured over
a 350-600 nm range. The intensity value was recorded at the
absorption maximum, and each measurement was rescaled to
a constant baseline, since there was variation in the initial
baseline of many of the spectra. The measured value of the
absorption at any one time was then normalized according to
the maximum amount released. In the case of degradable
hydrogels that exhibited complete degradation behavior, the
entire quantity of fluorescent dextran was released into
solution, or equivalently 100% of the contents was measured
at the final measurements when degradation occurred. For the
control hydrogels that did not undergo degradation, 100%
release was mathematically determined on the basis of the
expected calculation of 100% release from the initial unpoly-
merized solution.
Estimation of Network Mesh Size. Mesh size has been
estimated on the basis of equilibrium swelling for hydrogels
composed of PVA, PEG, and PHEMA backbones.14-16,24 The
procedure for calculating mesh size, adopted from Canal and
Peppas24 (Figure 5b), is as follows. We used eq 2 to relate the
polymer swelling to the molecular weight between cross-links,
Mh c, in the network,28,29 thereby linking the volume fractions
of polymer before and after swelling, υ2,r and υ2,s, to the cross-
link density. However, this equation was only used with gels
made of cross-linkers 2a and 2b which degraded slowly enough
to be in equilibrium.
and the characteristic ratio Cn was chosen to be 10. We were
unable to find the characteristic ratio of HPMA in the
literature, and therefore we choose an average value of 10
because HPMA is expected to have a characteristic ratio
between HEMA (Cn ) 6.9), a relatively hydrophobic polymer,
and PEG (Cn ) 14),32 a more hydrophilic polymer. Mesh size,
ê, was then computed as shown in eq 6. The Stokes radii for
the fluorescent FITC dextrans used in our studies were
obtained from the Sigma-Aldrich catalog.
2 1/2
ê ) ν2,s1/3(rj0 )
(6)
Results and Discussion
Three new degradable symmetrical cross-linkers con-
taining oligo-glycolate and oligo-lactate esters (Figures
1d and 2) were synthesized according to the scheme
depicted in Figure 2 and copolymerized with 2-hydrox-
ypropyl methacrylamide (HPMA), using free radical
polymerization to give biodegradable hydrogels (Figure
3). The swelling kinetics (measured swelling ratio Q vs
time) of the four gel systems were measured at intervals
in pH 7.4 media (Figure 3). Figure 3a shows a photo-
graph of biodegradable gels composed of 1.5 mol % cross-
linker 2b (HPMAGly)2succinate and 98.5 mol % HPMA,
after incubation at pH 7.4 for 2, 5, and 15 days. These
experiments demonstrate the well-known effect that
cross-link density has on the degree of equilibrium gel
swelling.10 As the cross-linker was hydrolyzed, the pore
size increased and the elastic modulus decreased, al-
lowing increasing amounts of water to be imbibed by
the hydrogel. Ultimately, the gels dissolved into water-
soluble components, completing a gel to sol transition.
As can be seen in Figures 3a and 4, the structure of the
cross-linker has a large impact on the rate of this
process.
3
1
N
2/3
υj/Vh1[ln(1 - υ2,s) + υ2,s + øυ2,s2] 1 - (υ2,s
)
[
]
1
2
)
-
2
1
2
1
N
1/3
1/3
Mh c Mh n
υ2,r (υ2,s/υ2,r
)
-
(υ2,s/υ2,s) 1 + (υ2,s/υ2,s
)
[
][
]
(2)
Figure 4 graphically illustrates the range of rates of
degradation that can be obtained from the cross-linkers.
In Figure 4, it can be seen that cross-linker 1 behaves
in a static fashion similar to known nondegradable
cross-linkers. As described above, the degradation rate,
which is set by the chemical structure of the cross-
linker, determines the swelling ratio. Finally, by simply
manipulating the initial concentration or mole fraction
of cross-linker in the network, one can also adjust the
time course to allow for faster or slower overall gel
degradation rates because there is more cross-linker in
the network to degrade.
In this equation, Mh n is the number-average molecular weight
of the polymer in the polymer network which was determined
experimentally to be Mh n ) 93 000 for these pHPMA networks
by degrading the hydrogel samples at pH 12.0, dialyzing the
solution in a 500 MW cutoff membrane, and measuring the
molecular weight distribution by GPC with a light scattering
detector (Wyatt Technologies). The value of υj is the specific
volume of HPMA (0.841 cm3/g), Vh1 is the molar volume of water
(18.14 cm3/mol), and ø is the Flory-Huggins parameter. υ2,r
was experimentally determined to be 0.57 as an average for
gel compositions studied. The value of ø for poly(HPMA), given
by eq 3, appears in the literature27 and is an empirical function
of the polymer volume fraction υ2,s and the temperature. It is
given by the relation at 20 °C27
The slowest gel of the series dissolved after ∼800 h
(∼30 days) and was composed of cross-linker 2a, which
had a single lactate ester on each side of the succinate
core. In comparison, the unmethylated homologue of
this material, cross-linker 2b, dissolved after ∼350 h
(∼14.6 days), which indicated that this cross-linker was
cleaved at nearly twice the rate as cross-linker 2a. We
attribute this to differences in the steric crowding at
the site of reaction during attack by the hydroxide
nucleophile.33 The methyl group in 2a increases crowd-
ing, resulting in slower ester hydrolysis than 2b. Not
surprisingly, the fastest cross-linker to degrade was 3,
which swelled and dissolved after 50 h (∼2.1 days). In
comparison to cross-linker 2b, this gel was the more
labile due to increased electronic demands on the
carboxylate groups resulting from the additional glyco-
late functionality and the statistical effect of a greater
number of labile sites. These factors in concert signifi-
cantly increased the overall rate of degradation. Finally,
by simply manipulating the initial concentration or mole
fraction of cross-linker in the network, one can also
ø ) 0.473 + 0.489ν2,s
(3)
N represents the number of bond vectors in the chain and is
given by eq 4, where Mr is the molecular weight of the
monomer unit (144 g/mol).
2Mh c
Mr
N )
(4)
This parameter is added to account for the random non-
Gaussian distribution of the effective chains.30,31 To calculate
the mesh size, ê, the number-average molecular weight
between cross-links, Mh c,was used to estimate the end-to-end
distance of the unperturbed (or solvent-free) state (rj0 )1/2 along
2
with the characteristic ratio, Cn, of the HPMA chain, as shown
in eq 5.
2
1/2
(rj0 ) ) l(2Mh c/Mr)1/2Cn
(5)
In this equation the C-C bond length is given by l (1.54 Å),

10760 Eichenbaum et al.
Macromolecules, Vol. 38, No. 26, 2005
Figure 2. Conditions for the synthesis of the cross-linkers. Conditions: (a) succinyl dichloride, CH2Cl2, pyridine, 0 °C; (b) Pd/C,
50 psi of H2, i-PrOH; (c) DMF, CDI, 0 °C, HPMA.
Figure 3. (a) Photograph of the swelling profile for biodegradable hydrogels composed of cross-linkers 2b,8,22 (HPMAGly)2succinate
(X ) 1.5 mol % cross-linker feed ratio and 98.5% HPMA), after incubation in pH 7.4 buffer (100 mM phosphate buffer, ionic
strength of 200 mM; T ) 37 °C). From left to right: control gel made with nondegradable cross-linker 1 (HPMA2succinate) after
15 days and gel made with cross-linker 2b after 2, 5, and 15 days, respectively (see Figure 4 for swelling profiles of hydrogels
composed of cross-linkers 1-3). (b) Photograph displaying the dynamic release of entrained macromolecules: four hydrogel samples
(3.0 mol % feed cross-linker and 97 mol % HPMA) incubated at pH 7.4 solution for ∼300 h (13 days) containing three different
molecular weights of dextran. From left to right: nondegradable control (cross-linker 1) containing 148 kDa fluorescenated dextran
and cross-linker 2b with 148, 42, and 12 kDa fluorescenated dextran samples, respectively. Scale bars: 1 cm (see Figure 5a for
release profiles).
adjust the time course to allow for faster or slower
overall gel degradation rates. As shown in Figure 4, by
doubling the amount of cross-linker 2b, we were able
to dramatically decrease the rate of change of the
swelling ratio and thus obtain a total degradation time
nearly equal to the slowest cross-linker 2a.
To evaluate the influence of swelling kinetics on the
release rate of macromolecules, we synthesized degrad-
able hydrogels with entrained dextrans of varying sizes.
This allowed us to study the relationships of a cross-
linker’s chemical structure, the hydrodynamic size of
an entrapped macromolecule, the hydrogel cross-link
density, and the pH dependence of degradation on the
swelling and release characteristics. Figure 3b shows a
photograph of biodegradable gels that were loaded with
three different molecular weight fluorescenated dex-
trans. As the gel dynamically swelled and the cross-
linkers degraded, the mesh size expanded, allowing for
the release of the entrained macromolecules.
Figure 5a demonstrates that one can control the rate
of release from the gel by changing either the cross-
linker or the MW of the entrained macromolecule. Of
particular interest in Figure 5a is the demonstrated
ability to closely match the release profile of a 148 kDa
entrained molecule to that of a 42 kDa entrained
molecule by simply changing the cross-link density (i.e.,
1.5% for the 42 kDa dextran and 3% for the 148 kDa
dextran). In this way, it was possible to construct two
different gels that released ∼50% of their fluorescenated
dextran (42 and 148 kDa) at 200 h. Furthermore, by
comparing the two cross-link densities (1.5% and 3%)
used with the 148 kDa dextran, one can see that after
200 h the 3% cross-linked hydrogel has released only
20% of the macromolecule vs 50% for the 1.5% cross-
linked polymer. In the higher cross-link density case,
the release profile approximates zero-order release,
which is advantageous in many therapeutic dose regi-
mens.34

Macromolecules, Vol. 38, No. 26, 2005
Oligo-R-hydroxy Ester Cross-Linkers 10761
cross-link density. As a result, fluorescenated dextran
molecules of 12, 42, and 148 kDa entrained in six
distinct gels are observed to converge into three mesh
size/release profiles that depend on the size of the
entrained dextran.
Figure 5b also demonstrates how a larger mesh size
is required to achieve the release of larger entrained
dextran molecules, since 50% release of the 12kD
entrained molecules was achieved after the mesh size
expanded to 7 nm compared with a 15 nm mesh size
required for release of the 148 kDa entrained molecules.
Given that the Stokes radius of the 148 kDa entrained
dextran molecule is ∼9 nm, Figure 5b illustrates how
the 148 kDa dextran is largely contained in the gel until
the mesh size reaches 9 nm, at which point release
increases significantly. This relationship between mesh
size, entrained macromolecule size, and release has
clear implications for the design of biodegradable gels
engineered for different sized macromolecules in sys-
tems that largely avoid burst release.
The results above show that these ester-containing
cross-linkers conformed to the known order of impor-
tance of steric, electronic, and pH effects in ester
hydrolysis.23 The data confirm that the primary factors
that impact gel degradation rates are the steric hin-
drance and electronic properties at the site(s) of degra-
dation, the number of sites available for hydrolysis, and
the degree of initial cross-linking within the network.
In comparison to the normal bulk degradation kinetics
of PLGA polymers, where the hydrolysis mechanism is
dominated by autocatalytic hydrolysis, the degradation
mechanism for these polymers is dominated by normal
base hydrolysis at pH 7. It is well-known that for the
polymers of lactic and glycolic acids base hydrolysis is
slower than the corresponding acid hydrolysis. However,
this is not true for polymer systems made of these cross-
linkers.
Figure 4. Time course of swelling ratio (Q) of degradable gels
made from four different cross-linkers (1-3) at pH 7.0, 100
mM PBS; I ) 200 mM; T ) 37 °C, error bars ( SD, n ) 6 (X
) mol % feed cross-linker). See images in Figure 3a for images
of gels made with cross-linker 2b.
Conclusion
We have successfully developed22 and tested a new
class of small molecule biodegradable cross-linkers that
offer several advantages over existing similar cross-
linking systems.12,13 First, the methodology allows for
the cross-linker structure to be readily modified and
easily purified into a single compound after synthesis.
Accordingly, the method allows for control over the
hydrolysis rate by increasing or decreasing the number
of monomers in the oligomer. Third, the end groups of
the oligomers can be readily modified to accommodate
either condensation or free radical polymerizations.
Finally, by varying the cross-linker structure(s), their
proportions, and the cross-link density, one can readily
“dial in” the degradation and release properties of a
hydrogel across a range of several orders of magnitude
in time. This opens up a range of material properties
that one could imagine using to obtain different con-
trolled release profiles.7,9 However, loading of thera-
peutic proteins into gel networks by polymerization in
the presence of the protein presents a number of
unsolved challenges related to purification and protein
stability. We circumvented some of these issues by
loading the protein after forming the polymer network
using electrostatic interactions between the polymer
network and the protein.35 Biodegradable hydrogels
prepared using these synthetically adaptable cross-
linkers offer biomaterials researchers a new tool for
applications spanning drug delivery, tissue engineering,
Figure 5. (a) Release rate of FITC dextran from gels
composed of 3% (open symbols) and 1.5% (closed symbols) mole
feed ratio cross-linker 2b and p(HPMA) at pH 7.0 (see Figure
3). The control cross-linker was nondegradable. Error bars SD,
n ) 3. (b) Plot of fractional release as a function of mesh size
for a single cross-linker 2b at pH 7.0 and for various MW FITC
dextrans (X ) mol % feed cross-linker).
To better understand the molecular basis for the
release rates of the different MW dextrans, we esti-
mated the mesh size of the hydrogels at different
swelling ratios.14-16,24 Figure 5b shows the relationship
between the release rates and the estimated average
mesh size for different MW dextrans and cross-link
densities. Consistent with our explanation of the deg-
radation mechanism of the gel, the release behavior for
a given mesh size is independent of the gel’s initial

10762 Eichenbaum et al.
Macromolecules, Vol. 38, No. 26, 2005
(15) Metters, A. T.; Anseth, K. S.; Bowman, C. N. Polym. Prepr.
(Am. Chem. Soc., Div. Polym. Chem.) 2000, 41 (2), 1663-
1664.
and medical devices. Studies examining therapeutic
applications and physical behavior of these materials
are underway.
(16) Metters, A. T.; Anseth, K. S.; Bowman, C. N. Polymer 2000,
41, 3993-4004.
Acknowledgment. The authors thank Glynn Wil-
son, Ned Porter, and Joseph Eichenbaum for helpful
discussions. We also acknowledge both the Duke Uni-
versity and the University of Texas Southwestern
Medical Center NMR Spectroscopy Centers for use of
their facilities. We are grateful to Access Pharmaceu-
ticals, Dallas, TX, for financial support.
(17) Martens, P. J.; Bryant, S. J.; Anseth, K. S. Biomacromolecules
2003, 4, 283-292.
(18) Ulbrich, K.; Subr, V.; Seymour, L. W.; Duncan, R. J.
Controlled Release 1993, 24, 181-190.
(19) Anseth, K. S.; Shastri, V. R.; Langer, R. Nat. Biotechnol. 1999,
17, 156-159.
(20) Muggli, D. S.; Burkoth, A. K.; Keyser, S. A.; Lee, H. R.;
Anseth, K. S. Macromolecules 1998, 31, 4120-4125.
(21) Kim, S.; Healy, K. E. Biomacromolecules 2003, 4, 1214-1223.
(22) Kiser, P. F.; Thomas, A. Degradable cross-linking agents and
cross-linked network polymers formed therewith. US Patent
6,521,431.
(23) Euranto, E.; Esterification and ester hydrolysis. In The
Chemistry of Carboxylic Acids and Esters; Patai, S., Ed.;
Interscience: London, UK, 1969.
References and Notes
(1) Park, K. Biodegradable Hydrogels for Drug Delivery; Tech-
nomic Publishing Co., Inc.: Lancaster, PA, 1993.
(2) Hennink, W. E.; van Nostrum, C. F. Adv. Drug Delivery Rev.
2002, 54, 13-36.
(3) Jeong, B.; Bae, Y. H.; Lee, D. S.; Kim, S. W. Nature (London)
1997, 388, 860-862.
(4) Langer, R. S. Science 2001, 293, 58-59.
(5) Chasin, M.; Langer, R. Biodegradable Polymers as Drug
Delivery Systems; Marcel Decker: New York, 1990; Vol. 45.
(6) van Dijk-Wolthius, W. N. K.; Tsang, S. K. Y.; Kettees-van
den Bosch, J. J.; Hennink, W. E. Polymer 1997, 38, 6235-
6242.
(24) Canal, T.; Peppas, N. A. J. Biomed. Mater. Res. 1989, 23,
1183-1193.
(25) Thomas, A. A.; Kiser, P. F.; Kim, I. Book of Abstracts, 219th
ACS National Meeting, San Francisco, CA, March 26-30,
2000; ORGN-394.
(26) Kopecek, J.; Bazilova, H. Eur. Polym. J. 1974, 9, 7-14.
(27) Kopecek, J.; Bazilova, H. Eur. Polym. J. 1974, 10, 465-470.
(28) Peppas, N. A.; Mikos, A. G. Preparation methods and
structure of hydrogels. In Hydrogels in Medicine and Phar-
macy; Peppas, N. A., Ed.; CRC Press: Boca Raton, FL, 1986;
Vol. 1, pp 1-83.
(7) Lu, S.; Anseth, K. S. Macromolecules 2000, 7, 2509-2515.
(8) Kiser, P. F.; Thomas, A. A.; Eichenbaum, G. M.; Needham,
D.; Kim, I. Book of Abstracts, 219th ACS National Meeting,
San Francisco, CA, March 26-30, 2000; POLY-362.
(9) Qiu, Y.; Park, K. Adv. Drug Delivery Rev. 2001, 53, 321-
339.
(29) Peppas, N. A.; Moynihan, H. J.; Lucht, L. M. J. Biomed.
Mater. Res. 1985, 19, 397-411.
(10) Peppas, N. Hydrogels in Medicine and Pharmacy; CRC
Press: Boca Raton, FL, 1986; Vol. 1.
(30) Fixman, M.; Kovac, J. J. Chem. Phys. 1973, 58, 1564-1568.
(31) Kovac, J. Macromolecules 1977, 11, 362-365.
(32) Flory, P. J. Statistical Mechanics of Chain Molecules; Inter-
science Publishers: New York, 1969; p 425.
(11) van Dijk-Wolthius, W. N. E.; Hoogeboom, J. A. M.; van
Steenbergen, M. J.; Tsang, S. K. Y.; Hennink, W. E. Macro-
molecules 1997, 30, 4639-4645.
(33) Taft, R. W. J. Am. Chem. Soc. 1952, 74, 2729-2732.
(34) Tomlinson, E. Adv. Drug Delivery Rev. 1987, 1, 87-198.
(35) Eichenbaum, G. M.; Kiser, P. F.; Dobrynin, A. V.; Simon, S.
A.; Needham, D. Macromolecules 1999, 32, 4867-4878.
(12) Sawhney, A. S.; Pathak, C. P.; Hubbell, J. A. Macromolecules
1993, 26, 581-587.
(13) Cadee, J. A.; De Kerf, M.; De Groot, C. J.; Den Otter, W.;
Hennink, W. E. Polymer 1999, 40, 6877-6881.
(14) Metters, A. T.; Bowman, C. N.; Anseth, K. S. AIChE J. 2001,
47, 1432-1437.
MA0518306