Hydrogels for combination delivery of antineoplastic agents
Kamal H. Bouhadir , Eben Alsberg , David J. Mooney *
Department of Biologic and Materials Sciences, University of Michigan, 3074 H.H. Dow Building, 2300 Hayward Street, Ann Arbor, MI 48109-2136, USA
Department of Chemical Engineering, University of Michigan, 3074 H.H. Dow Building, 2300 Hayward Street, Ann Arbor, MI 48109-2136, USA
Department of Biomedical Engineering, University of Michigan, 3074 H.H. Dow Building, 2300 Hayward Street, Ann Arbor, MI 48109-2136, USA
Received 1 September 2000; accepted 27 December 2000
The systemic delivery of anticancer agents has been widely investigated during the past decade but localized delivery may o!er
a safer and more e!ective delivery approach. We have designed and synthesized a novel hydrogel to locally deliver antineoplasticagents, and demonstrate the di!erent types of release that can be achieved from these hydrogels using three model drugs:methotrexate, doxorubicin, and mitoxantrone. Alginate was chemically modi"ed into lowmolecular weight oligomers and cross-linked with a biodegradable spacer (adipic dihydrazide) to form biodegradable hydrogels. The model antineoplastic agents wereloaded into the hydrogel via three di!erent mechanisms. Methotrexate was incorporated within the pores of the hydrogel and wasreleased by di!usion into the surrounding medium. Doxorubicin was covalently attached to the polymer backbone via a hydrolyti-cally labile linker and was released following the chemical hydrolysis of the linker. Mitoxantrone was ionically complexed to thepolymer and was released after the dissociation of this complex. These three release mechanisms could potentially be used to delivera wide selection of antineoplastic agents, based on their chemical structure. This novel delivery system allows for the release of singleor combinations of antineoplastic agents, and may "nd utility in localized antineoplastic agent delivery.
Alginate; Controlled release; Biodegradable; Antineoplastic agents; Doxorubicin; Methotrexate; Mitoxantrone
reduce the side e!ects associated with the systemic deliv-ery of anticancer agents [3}6]. Several drugs have been
A broad spectrum of antineoplastic agents has been
found to amplify the anticancer activity of other drugs
found to be e!ective in combating di!erent types of
[7}10]. This synergistic e!ect can potentially lead to
cancer. However, to achieve complete eradication of tu-
reduced doses for each drug administered [11}13].
mors, antineoplastic agents are administered systemically
Hence, the administration of several drugs simulta-
in high doses, and almost all drugs e!ective in killing
neously could reduce the side e!ects caused by high doses
cancer cells cause damage to other healthy tissues and
of single drugs and could prevent the development of
organs. This is due to the non-speci"c uptake of these
multi-drug resistance (MDR) [14,15].
agents by healthy organs such as the kidney, liver, bone
An alternative approach to systemic delivery of anti-
marrow, and heart. The adverse side e!ects include
neoplastic agents is the localized release from a polymer
severe immune suppression, myelosuppression, neph-
. We have designed and synthesized a novel hydrogel
rotoxicity, and cardiotoxicity [1,2]. During the past dec-
to deliver anticancer agents locally. We have oxidized
ade, many researchers have investigated the sequential
sodium alginate to form lowmolecular weight oligomers
and simultaneous delivery of drug combinations to
that are cross-linked with adipic dihydrazide to formhydrogels. We hypothesize that a variety of antineoplas-tic agents could be locally delivered, alone or in combina-tion from these hydrogels. The kinetics of drug release
Correspondence address: Department of Chemical Engineering,
could potentially be controlled by exploiting three types
University of Michigan, 3074 H.H. DowBuilding, 2300 Hayward St.,
interaction, and covalent coupling via degradable
E-mail address: [email protected]
linkages. We have incorporated three model drugs into
2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 0 0 3 - 5
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
cross-linked oxidized alginate hydrogels: methotrexate,
mixture to reduce any unreacted periodate. The reaction
doxorubicin, and mitoxantrone to test this possibility. All
was stirred for 0.5 h at ambient temperature, and the
three drugs are potent antineoplastic agents that have
been extensively utilized in cancer chemotherapy.
(Spectra/Pro membrane, MWCO 3500) against double-
Methotrexate is a folic acid antimetabolite inhibitor of
distilled water (dd. HO) for three days. The water was
dihydrofolate reductase that has been widely used in the
changed at least 3 times a day. The solutions were then
treatment of neoplastic diseases . Doxorubicin is an
concentrated to around 100 ml and freeze dried under
antineoplastic agent from the anthracycline antibiotic
reduced pressure to yield a white product (6.9 g, 86%). IR
family that has been most commonly used to treat solid
(KBr pellet, cm\) 3336, 2942, 1730, 1622, 1406, 1321,
tumors [18,19]. Mitoxantrone, an anthracenedione, is an
intercalating agent that is e!ective in treating varioustumors . Methotrexate, an anionic drug, is used in
2.3. Determination of the degree of oxidation
this study as it is expected to rapidly di!use out into thesurrounding medium. We have used doxorubicin as
The degree of oxidation of alginate was determined
a model drug for chemical-controlled release. Doxo-
by measuring the percentage of periodate that was
rubicin is expected to be released following the hydrolysis
consumed before quenching with ethylene glycol. The
of the degradable bond linking it to the gel . In
consumption of sodium periodate was determined by
addition, we have used mitoxantrone as a model drug for
spectrophotometrically measuring the formation of
ionic-controlled release. Mitoxantrone is expected to
a complex between unreacted periodate anion and
form ionic complexes with the carboxylate groups on the
thyodene. Brie#y, equal volumes of freshly prepared
sugar residues of the polymer backbone.
aqueous solutions of potassium iodide (20% w/v in pH7.0 sodium phosphate bu!er) and thyodene solution(10% w/v in pH 7.0 sodium phosphate bu!er) were mixed
as an indicator solution. An Erlenmeyer #ask (100 ml)was covered with aluminum foil and charged with an
aqueous solution of alginate (50 ml, 1.0% w/v) and anaqueous solution of sodium periodate (10.1 ml, 0.25 M).
Sodium alginate was purchased from Pronova Bio-
The mixture was stirred at room temperature. At di!er-
materials (Drammen, Norway). Sodium periodate, adipic
ent time intervals, aliquots (0.3 ml) were rapidly removed
dihydrazide, ethylene glycol, and anhydrous KBr were
and diluted to a volume of 100 ml using dd. HO.
purchased from the Aldrich Chemical Company (Mil-
A 0.5 ml aliquot of this solution was immediately mixed
waukee, WI) and used as received. Ethanol (95%),
with 1.0 ml of the indicator solution in a cuvette. The
methanol, and concentrated hydrochloric acid were pur-
concentration of the unreacted periodate was measured
chased from Fisher Scienti"c Company (Fair Lawn, NJ)
spectrophotometrically at 486 nm. This number was then
and were used as received. Doxorubicin and mitoxan-
subtracted from the original concentration of periodate
trone hydrochlorides were purchased from Sigma
to yield the amount of periodate that was consumed.
Chemical Company (St. Louis, MO). Methotrexate waspurchased from Fluka Chemical Corporation (Ronkon-
2.4. Size exclusion chromatography (SEC)
koma, NY). Phosphate bu!er saline (PBS) and Dul-becco's
SEC analysis was performed on a liquid chromato-
purchased from Life Technologies (Grand Island, NY).
graph consisting of a SpectraSystem P1000 pump (Ther-
Eagle's Minimum Essential Medium (EMEM), fetal bo-
mal Separation Products), a Rheodyne 7010 manual
vine serum, and MCF-7 breast epithelial cell lines were
injector, a dual di!erential viscometer and right-angle
purchased from the American Tissue Culture Collection
laser light scattering (RALLS) detector (Viscotek T 60,
(Manassas, VA). Cell culture plates (96-well) were pur-
" 670 nm) and a laser refractometer detector (Viscotek
chased from Falcon (Lincoln Park, NJ).
LR40, "670 nm), the detectors being connected in par-allel. The mobile phase consisted of aqueous sodium
nitrate (0.1 M) and was periodically degassed with anon-line degasser. The mobile phase was delivered at am-
A 1 l Erlenmeyer #ask was wrapped with aluminum
bient temperature with a nominal #owrate of 0.7 ml/min.
foil and charged with sodium alginate (8.0 g). Double-
The separations were carried out on two TSK GMPW6*
distilled water (800 ml) was added, and the mixture was
(TosoHaas, 7.8;300 mm) mix bed columns. Polymers
stirred until the solid dissolved. An aqueous solution of
were dissolved in mobile phase solvent at a concentration
sodium periodate (0.25 M, 162 ml) was added and the
of 1}3 mg/ml by mechanical stirring for a minimum of 6 h
reaction was stirred for 24 h at room temperature.
until completely hydrated. A 100 l injection volume was
Ethylene glycol (2.3 ml) was then added to the reaction
used for all analyses. The chromatograms were analyzed
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
using the TriSEC 3.0 GPC software (Viscotek). A di!er-
ance of the drug at 327.5 nm (methotrexate), 480 nm
ential index of refraction (dn
) of 0.154 ml/g was used
spectra were collected on a Perkin Elmer Lambda 12UV/VIS spectrophotometer.
2.8. In vitro drug release from cross-linked oxidized
Hydrogels were formed at various concentrations of
oxidized alginate, adipic dihydrazide and calcium chlor-ide in 24-well plates. The contents of each well were
Sterile tubes were charged with aqueous solutions of
mixed and allowed to gel for 1 h at ambient temperature
adipic dihydrazide (150 l, 0.5 M), calcium chloride (20 l,
on a mechanical shaker. The hydrogels were immersed in
1.0 M), and DMEM (20 l). Solutions of methotrexate,
de-ionized water and incubated at 373C for 24 h to reach
doxorubicin hydrochloride, or mitoxantrone (10 l,
the equilibrium swelling condition. The hydrogels were
25 mg/ml in DMSO) were then added to the above aque-
transferred to 2 ml vials and weighed (wet weight). The
ous solutions, and the mixtures were mixed for 15 min.
gels were then frozen, lyophilized, and the dried samples
An aqueous solution of oxidized alginate (300 l,
were weighed (dry weight). The swelling ratio was de"ned
10% w/w) was then added, and the contents of the tubes
as the ratio of (wet weight ! dry weight)/(dry weight).
were mixed thoroughly and allowed to gel for 1 h. Aque-
Infrared spectra were recorded as percent transmittance
ous DMEM solutions (5 ml) containing penicillin and
using a Nicolet 5DX FTIR spectrophotometer and
streptomycin were added to each tube, and the tubes
a Hewlett Packard 7470A plotter. Samples were pressed
were incubated at 373C. The medium was removed
as KBr pellets using a hydraulic press (Carver, Inc.). IR
periodically and replaced with fresh DMEM, and the
(KBr pellet, cm\) 3554, 3472, 3414, 3236, 1660, 1622,
released drug was quanti"ed as described previously.
A total of 0.25 mg of each drug was always loaded per gelsample. Release data are reported as a percentage of this
20% w/w) and adipic dihydrazide (125 l, 0.5 M) contain-
ing calcium chloride (80 mM) were mixed in 15 ml conicaltubes (in quadruplicates) and allowed to gel for 5 h.
Aqueous solutions of sodium alginate were oxidized in
Solutions of Dulbecco's Modi"ed Eagles Medium
the dark using sodium periodate at room temperature
(DMEM, 10 ml) were added, and the tubes were incu-
following a modi"ed procedure reported previously
bated at 373C. The medium was replaced with fresh
[23}25]. The amount of sodium periodate used in these
medium on a weekly basis. Four tubes were removed
reactions was varied to form alginates with di!erent
every week, and the medium was decanted. The gels were
degrees of oxidation. FTIR analysis of the oxidized prod-
frozen, lyophilized, and the dry solid was weighed.
uct revealed a newpeak at 1730 cm\ corresponding forthe vibrational symmetric stretching vibration of the
2.7. In vitro drug release from alginate hydrogels
aldehyde groups. The degree of oxidation was deter-mined indirectly by measuring the percentage of sodium
An aqueous alginate solution was prepared by dissolv-
periodate that was consumed in each reaction. Sodium
ing sodium alginate (2 g), sodium chloride (0.8 g), and
periodate was almost quantitatively consumed in all con-
sodium hexametaphosphate (0.4 g) in 96.8 ml of dd. HO.
ditions except when 100 percentage equivalents was used
A 50 ml conical tube was charged with 5 g of the above
(Table 1). In this case, only 69% of the sodium periodate
solution. A solution of the drug (165 l, 25 mg/ml in
was consumed after 24 h. This is consistent with earlier
DMSO) was then added and mixed thoroughly. Aqueous
papers reporting that the aldehyde groups of the oxidized
calcium sulfate slurry (200 l, 158 mg/ml) was added,
uronate react with neighboring alcohol groups to form
mixed thoroughly, and the mixture was cast between
hemiacetals . The formation of hemiacetals protects
parallel glass plates separated with glass spacers (2 mm in
the alcohol groups of neighboring uronates from further
thickness). The gels were allowed to set for 8 h. The glass
plates were separated, and the disks were punched outwith a 12.7 mm hole puncher (McMaster-Carr). The hy-
3.1. Molecular weight distribution of oxidized alginate
drogel disks were placed in scintillation vials (2 disks ineach vial). Aqueous PBS (pH 7.4) was added, and the
The molecular weight distribution of oxidized alginate
vials were incubated at 373C. The medium was replaced
was analyzed by aqueous gel permeation chromatogra-
periodically, and the amount of drug that was released
phy. The weight-average molecular weight of the
in the medium was quanti"ed by measuring the absorb-
starting alginate was 394 kDa. Alginate oxidized with "ve
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
Table 1The experimental degree of oxidation of cross-linked oxidized alginatehydrogels. Reactions were run at a concentration of 0.8% w/w alginatein the dark at room temperature for 24 h
Table 2Molecular weight distributions of alginate and oxidized alginates
Represents the percentage equivalents of sodium periodate that was
initially added to the reaction mixture (based on the uronate groups).
) is the weight fraction of the polymer that have a molecular
equivalents of sodium periodates formed a polymer witha weight-average molecular weight of 198 kDa (Table 2).
The weight-average molecular weight then decreased asthe percentage equivalents of periodate was increased toreach 29 kDa with 100% equivalents of sodium peri-odate. As a result, the intrinsic viscosity of the polymers
Fig. 1. Synthesis and cross-linking of oxidized alginate: (a) sodium
decreased as periodate concentration was increased.
Only 12.5% weight fraction of the original unmodi"edalginate has a molecular weight below 80 kDa. This num-ber increased to reach a value of 96% for alginate thatwas oxidized with 100 equivalents of sodium periodate
The degree of swelling of cross-linked oxidized alginate
hydrogels was analyzed after the hydrogels reached the
equilibration swelling in dd. HO. The swelling ratio of
these hydrogels varied signi"cantly depending on the
Hydrogels were subsequently formed by the reaction
concentrations of both the ionic and the covalent
of adipic dihydrazide and the oxidized alginates. The
cross-linkers. The swelling ratio of hydrogels made with
hydrazide group reacts with the aldehyde groups in oxi-
oxidized alginate (100 equivalents periodate) and cross-
dized alginate to form hydrazone bonds (Fig. 1). The
linked at 150 mM adipic dihydrazide was 29.9$1.2 in dd.
hydrogels were washed with water and soaked in dd.
HO (Table 3). The swelling ratio then decreased with
HO for 24 h to release the unreacted adipic dihydrazide.
increasing concentrations of calcium to reach a minimum
The gels were frozen and lyophilized. FTIR spectroscopic
of 11.7$0.3 at 40 mM calcium ions. A similar trend was
analysis of the discs indicated the disappearance of the
observed when the concentration of covalent cross-links
peak at 1730 cm\ and the appearance of another peak
was increased. The swelling ratio was 20.1$1.2 at 40 mM
at 1660 cm\ that corresponds to the stretching vibra-
calcium ions at 50 mM adipic dihydrazide, and decreased
tion of the carbonyl in the hydrazide group.
to 11.7$0.3 at 40 mM as the concentration of adipic
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
Table 3Swelling ratio of cross-linked oxidized alginate hydrogels as a functionof the concentrations of the ionic and covalent cross-linkers. Hydrogelswere formed at 6% w/w oxidized alginate in dd. HO
Fig. 2. Percentage weight loss of cross-linked oxidized alginate hydro-
gels as a function of time. Hydrogels were formed at (ⅷ) 100 mM and
(*) 150 mM adipic dihydrazide and 40 mM CaCl
prepared with 10% w/w oxidized alginates (69% oxidized) in dd. H
dihydrazide increased to 150 mM (Table 3). The swelling
Methotrexate was quantitatively released within 2 days
ratio slightly increased as the adipic dihydrazide con-
of incubation from hydrogels formed with 50 mM adipic
centration was further increased. This latter result is
dihydrazide. The overall release time of methotrexate
consistent with past studies which indicated a decreased
increased to 3 days at 75 mM adipic dihydrazide and
cross-link density, and increased content of dangling
7 days at 150 mM adipic dihydrazide. However, 75% of
cross-linkers, above 150 mM adipic dihydrazide .
the loaded drug was released initially at a constant rate of37.5% per day in all conditions. Drug release from hy-
drogels formed with concentrations of adipic dihydrazideabove 150 mM exhibited similar release kinetics (not
Alginate hydrogels degrade in an uncontrolled manner
following the release of calcium ions into the surrounding
Doxorubicin was covalently incorporated into the hy-
medium. To evaluate whether the degradation of cross-
drogel by reacting it with excess adipic dihydrazide as
linked oxidized alginates can be controlled, gels were
reported previously for daunomycin . The ketone
formed with 10% w/w oxidized alginates (oxidized with
group on the C13 position of doxorubicin reacts with the
100 equivalents of periodate) and cross-linked with
hydrazide group to form the doxorubicin}adipoyl hy-
adipic dihydrazide and/or calcium. The percentage
drazide conjugate (Fig. 3). Upon mixing with oxidized
weight loss of these gels was determined following incu-
alginate, the free hydrazide group on this conjugate re-
bation in medium (Fig. 2). Hydrogels formed at 100 mM
acts with the pendant aldehyde group on the backbone
adipic dihydrazide degraded after 3 weeks of incubation
of oxidized alginate to form a labile hydrazone bond
at a rate of 5% per day. Hydrogels cross-linked at
(Fig. 3). The release of doxorubicin from cross-linked
150 mM adipic dihydrazide and 40 mM calcium chloride
oxidized alginate was dependent on the concentration of
degraded at a lower rate of 2.5% per day (Fig. 2). Only
the covalent cross-linker (Fig. 4b). Doxorubicin was
40% of the gel weight dissolved after 15 weeks. Therefore,
quantitatively released within 2 days from all hydrogels
hydrogels may be formed with this approach that de-
formed at 50 mM adipic dihydrazide. At 75 mM adipic
grade in time frames from 3 weeks to more than 4 months
dihydrazide, doxorubicin was released after 3 days. In-
by simply varying the number of covalent and ionic
creasing the concentration of adipic dihydrazide to
100 mM prolonged the total release of doxorubicin to6 days of incubation, and at a higher concentration of
150 mM adipic dihydrazide, only 20% of doxorubicin wasreleased after 22 days of incubation at 373C.
The three model drugs were "rst separately incorpor-
Mitoxantrone was expected to form ionic complexes
ated in cross-linked oxidized alginate hydrogels. Metho-
with the hydrogels (Fig. 3), and would thus be released
trexate (Fig. 3) was not expected to interact ionically or
only as the gels degraded. The release of mitoxantrone
covalently with the hydrogel. The release pro"le of
from cross-linked oxidized alginate hydrogels was
methotrexate was not signi"cantly dependent on the con-
in#uenced by the concentration of adipic dihydrazide
centration of the covalent cross-linker (Fig. 4a), as ex-
(Fig. 4c). The entire loaded drug was released after only
pected at the high degrees of swelling in these hydrogels.
2 days of incubation from hydrogels formed with 50 mM
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
Fig. 3. Proposed mechanism for drug incorporation. Doxorubicin is chemically linked to oxidized alginate with a hydrazone bond. Mitoxantroneforms an ionic complex with the carboxylate groups.
adipic dihydrazide. Hydrogels formed at 75 and 100 mM
of incubation as expected for an ionically interacting
adipic dihydrazide released the entire drug after 3 and
6 days, respectively. However, hydrogels formed at
To test the utility of oxidized alginate hydrogels in
150 mM adipic dihydrazide released only 7% of the
delivering combinations of drugs, cross-linked oxidized
loaded drug after 21 days of incubation.
alginate hydrogels have been loaded with all three drugs
Methotrexate, doxorubicin, and mitoxantrone have
simultaneously. Methotrexate was quantitatively re-
also been separately incorporated into calcium cross-
leased after 9 days of incubation (Fig. 6). Doxorubicin
linked alginate hydrogels, and their release monitored
was released slowly at a rate of 1.7% per day for 13 days
spectrophotometrically as a control. Methotrexate was
followed by a rapid release of 19.5% per day during the
quantitatively released after 8.5 h of incubation at 373C
last 4 days (Fig. 6). Mitoxantrone, on the other hand, was
as expected. Doxorubicin was released over a 3.5 day
not signi"cantly released during the initial 10 days. How-
time period (Fig. 5). The release of mitoxantrone
ever, a rapid release of 24% per day was observed during
from alginate hydrogels was negligible over the "rst week
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
Fig. 5. The cumulative release of (ⅷ) methotrexate and () doxorubicinfrom alginate hydrogels. Hydrogels were formed with 2% w/w alginatein dd. HO and cross-linked with calcium sulfate.
Fig. 6. The release of (ⅷ) methotrexate, () doxorubicin, and (᭡)mitoxantrone from cross-linked oxidized alginate hydrogels loadedwith all three drugs. Hydrogels were formed at 100 mM adipic dihyd-razide and 6% w/w oxidized alginates (69% oxidized) in dd. HO.
utilized the reactive aldehyde groups to cross-link thesepolymers and form hydrogels. These polymers could besynthesized in a relatively short period of time and could
Fig. 4. The cumulative release of (a) methotrexate, (b) doxorubicin, and
be formed with a wide range of molecular weights. The
(c) mitoxantrone from cross-linked oxidized alginate hydrogels. Hydro-
weight fraction of the product that had a molecular
gels were formed at (ⅷ) 50 mM adipic dihydrazide, (*) 75 mM adipicdihydrazide and () 100 m
weight below 80 kDa could also be readily controlled.
M adipic dihydrazide and (᭡) 150 mM adipic
dihydrazide. All hydrogels were prepared with 6% w/w oxidized al-
This is very attractive for biomedical applications of
alginate derivatives since polymers with molecularweights lower than 80 kDa are expected to be clearedfrom the body in a similar manner to lowmolecular
We have separately incorporated three model drugs
We have synthesized a novel hydrogel derived from
into cross-linked oxidized alginate hydrogels: metho-
alginate for the single or simultaneous delivery of anti-
trexate, doxorubicin, and mitoxantrone to demonstrate
neoplastic agents. We have previously demonstrated the
that various drug}hydrogel interactions can be exploited
synthesis and cross-linking of poly(aldehyde guluronate),
to control the kinetics of drug release. Methotrexate was
PAG to form hydrogels . However, we were limited
quantitatively released over a 2 days period from hydro-
with that polymer by the low molecular weight of PAG
gels formed at 50 mM adipic dihydrazide, and the release
(6 kDa), and the synthesis required several labor intensive
was extended up to 7 days from hydrogels formed at
puri"cation steps. In the present study, we oxidize so-
higher concentration of adipic dihydrazide. This rapid
dium alginate directly, bypassing the hydrolysis step, and
release is expected for a drug with minimal hydrogel
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
interaction. The higher concentrations of adipic dihyd-
The simultaneous delivery of a combination of thera-
razide led to a lower swelling ratio, and this likely caused
peutical active agents has recently been shown to be
the slower release. These "ndings are consistent with
bene"cial in combating cancer and HIV infection. How-
many past reports of the rapid release of non-coupled or
ever, it might even be more attractive to deliver each drug
non-interacting drugs from alginate hydrogels .
with a di!erent release pro"le. We have incorporated all
To demonstrate the utility of these hydrogels to deliver
the three drugs simultaneously into cross-linked oxidized
drugs via a chemical-controlled release mechanism, we
alginate hydrogels and have observed three di!erent re-
have incorporated doxorubicin by covalent coupling to
lease pro"les (Fig. 6). Methotrexate was completely re-
the alginate backbone. Doxorubicin was released from
leased after 9 days of incubation. Both doxorubicin and
days to weeks depending on the concentration of adipic
mitoxantrone were completely released following the
dihydrazide used (Fig. 5b). This clearly indicates that
degradation of the hydrogel after 17 days of incubation.
doxorubicin is not merely interacting ionically with the
However, over 20% of doxorubicin was slowly released
hydrogel but is covalently linked to the polymer back-
during the "rst 13 days whereas mitoxantrone was not
bone . Doxorubicin is released following the hy-
signi"cantly released during that time. This demonstrates
drolysis of the hydrazone bond connecting it to the
clearly that we can indeed load a variety of drugs and
hydrogel. We can potentially release a variety of drugs
release them simultaneously over a wide range of time
that contain an aldehyde, a ketone, or a hydrazide group
frame. Another approach to achieve delivery of multiple
in a similar manner. A detailed analysis of this mecha-
drugs simultaneously or in sequence is to deliver each
nism of drug release from these types of polymers has
from a di!erent polymer, and either mix the formulations
been provided in a previous publication .
prior to delivery or introduce each separately. However,
To con"rm that doxorubicin was covalently linked to
the data in this paper demonstrate a single hydrogel may
oxidized alginate hydrogels, we used calcium cross-lin-
be used to deliver multiple drugs either simultaneously or
ked alginate hydrogels as a control. The release of
sequentially. This may simplify multi-drug delivery, and
doxorubicin is expected to followa di!usion-controlled
biomaterial development and regulatory approval.
release in a similar manner to methotrexate in these gels,due to the lack of potential for covalent coupling.
Doxorubicin was released from these gels over 3.5 days
(Fig. 5), which was somewhat longer than expected. Thissuggests that doxorubicin, a positively charged molecule,
We have designed and synthesized novel hydrogels
may be interacting ionically with the hydrogel, and this
derived from alginate to simultaneously deliver a variety
interaction is slowing down the release of doxorubicin.
of drugs. We can control the degradation pro"le of the
However, doxorubicin was released over a longer time
hydrogel from days to months and the release of model
from oxidized alginate hydrogels (formed at 150 mM
antineoplastic agents over a similarly wide range of time
adipic dihydrazide) indicating a chemical attachment be-
frames. Three di!erent release mechanisms: di!usion-
tween the drug and the polymer backbone .
controlled, covalent bond degradation, and ionic dis-
We have also incorporated and released mitoxantrone
sociation-controlled mechanisms, can be utilized in this
from oxidized alginate hydrogels to determine if ionic
system to control the kinetics of drug release. This novel
interactions can be used to control drug release. The
delivery system could be potentially used for the control-
release of mitoxantrone was coupled with the degrada-
led delivery of a variety of anticancer compounds sequen-
tion of the hydrogels, as expected for this mechanism.
tially or simultaneously, and in a localized manner.
Mitoxantrone was completely released from 2 to 6 daysfrom hydrogels cross-linked with 50}100 mM adipicdihydrazide (Fig. 4). However, less than 10% of the drug
was released after 22 days from hydrogels formed at highconcentrations of adipic dihydrazide (150 mM) suggesting
The authors would like to thank Gavin Sy for his help
that mitoxantrone is forming ionic junctions with the
in this project. The authors would also like to thank
carboxylate groups on the polymer backbone similar to
Reprogenesis for "nancial support of this research. The
divalent cations. Mitoxantrone was not signi"cantly re-
research was also supported in part by a grant from
leased from calcium cross-linked alginate hydrogels dur-
the John and Suzanne Munn Endowed Research Fund of
ing the initial 4 weeks of incubation (data not shown).
the University of Michigan Comprehensive Cancer
These gels are not expected to degrade in vitro. This
supports our prediction that mitoxantrone is
forming ionic junctions between the carboxylate groups
on the polymer backbone (Fig. 3). Ionic interactionsbetween alginate and many drugs are possible, and have
 Lowenthal RM, Eaton K. Toxicity of chemotherapy. Oncol Clin
K.H. Bouhadir et al.
/ Biomaterials 22 (2001) 2625
 Klein-Szanto AJ. Carcinogenic e!ects of chemotherapeutic com-
damaging agents in a human lung-cancer cell line. Int J Cancer
pounds. Prog Clin Biol Res 1992;374:167}74.
 Ikeda M, Okada S, Ueno H, Okusaka T, Tanaka N, Kuriyama H,
 von Minckwitz G, Costa SD, Eiermann W, Blohmer JU, Tulusan
Yoshimori M. A phase II study of sequential methotrexate and
AH, Jackisch C, Kaufmann M. Maximized reduction of primary
5-#uorouracil in metastatic pancreatic cancer. Hepato-Gastroen-
breast tumor size using preoperative chemotherapy with
doxorubicin and docetaxel. J Clin Oncol 1999;17(7):1999}2005.
 Andersson M, Madsen EN, Overgaard M, Rose C, Domber-
 Le Cesne A, Vassal G, Farace F, Spielmann M, Le chevalier T,
nowsky P, Mouridsen HT. Doxorubicin versus methotrexate
Angevin E, Valteau-Couanet D, Fizazi K, Cojean I, Llombard A,
both combined with cyclophosphamide, 5-#uorouracil and
Tursz T, Escudier B. Combination interleukin-2 and doxorubicin
tamoxifen in postmenopausal patients with advanced breast can-
in advanced adult solid tumors: circumvention of doxorubicin
cer * a randomised study with more than 10 years follow up from
resistance in soft-tissue sarcoma? J Immunother 1999;22(3):
Danish breast cancer cooperative group. Eur J Cancer 1999;
 Langer R, Brem H, Langer LF. Newdirections in CNS drug
 Ghielmini M, Zappa F, Menafoglio A, Cauduro L, Pampallona S,
delivery. Neurobiol Aging 1998;10:642}4.
Gallino A. The high dose sequential chemotherapy/PBSC trans-
 Calabresi P, Parks RE. Alkylating agents, antimetabolites, hor-
plantation regimen for patients with lymphoma is not cardiotoxic.
mones and other antiproliferative agents. In: Goodman LS, Gil-
man A, editors. The pharmalogical basis of therapeutics. New
 Miller KD, Stevens WM, Sisk J, Loesch DM, Manaco F, Seshadri
York: Macmillan Publishing, 1975. p. 1254}307.
R, Sledge GW. Combination versus sequential doxorubicin and
 Aubel-Sadron G, Londos-Gagliardi D. Daunorubicin and
docetaxel as primary chemotherapy for breast cancer: a random-
doxorubicin, anthracycline antibiotics, a physiological and biolo-
ized pilot trial of the Hoosier Oncologr Group. J Clin Oncol
gical review. Biochemie 1984;66:333}52.
 Arcamone F, Animati F, Capranico G, Lombardi P, Pratesi G,
 Presant CA, Wolf W, Walush V, Wiseman CL, Weitz I, Shani J.
Manzini S, Supino R, Zunino F. Newdevelopments in antitumor
Enhancement of #uorouracil uptake in human colorectal and
anthracyclines. Pharmacol Therapeutics 1997;76:117}24.
gastric cancers by interferon or by high dose methotrexate: an
 Bouhadir KH, Kruger GM, Lee KY, Mooney DJ. Sustained and
in vivo human study using noninvasive 19-F-magnetic resonance
controlled release of daunomycin from cross-linked poly(al-
spectroscopy. J Clin Oncol 2000;18(2):255}61.
dehyde guluronate) hydrogels. J Pharm Sci 2000;89(7):910}9.
 Britten CD, Hilsenbeck SG, Eckhardt SG, Marty J, Mangold G,
 Mackie W, Nor R, Sellen DB. Solution properties of sodium
MacDonald JR, Rowinsky EK, Von Ho! DD, Weitman S. En-
alginate. Biopolymers 1980;19:1839}60.
hanced antitumor activity of 6-hydroxymethylacylfulvene in com-
 Horton JC, Harding SE, Mitchel JR. Gel permeation chromatog-
bination with irinotecan and 5-#uorouracil in the HT29
raphy-multi-angle laser light scattering characterization of the
human colon tumor xenograft model. Cancer Res 1999;59(5):
molecular mass distribution of &Pronova' sodium alginate. Bio-
 Kogawa K, Muramatsu H, Tanaka M, Nishihori Y, Hagiwara S,
 Painter T, Larsen B. Formation of hemiacetals between neighbor-
Kuribayashi K, Nakamura K, Koike K, Sakamaki S, Niitsu Y.
ing hexuronic acid residues during the periodate oxidation of
Enhanced inhibition of experimental metastasis by the combina-
alginate. Acta Chem Scand 1970;24:813}33.
tion chemotherapy of Cu}Zn SOD and adriamycin. Clin Exp
 Bouhadir KH, Hausman DS, Mooney DJ. Synthesis of cross-
 Ravid A, Rocker D, Machlenkin A, Rotem C, Hochman A,
Kessler-Icekson G, Liberman UA, Koren R. 1,25-Dihydroxy-
 Lee KY, Bouhadir KH, Mooney DJ. Degradation behavior of
vitamin D3 enhances the susceptibility of breast cancer cells to
covalently cross-linked poly(aldehyde guluronate) hydrogels.
doxorubicin-induced oxidative damage. Cancer Res 1999;
 Al-Shamkhani A, Duncan R. Radioiodination of alginate via
 Buzdar AU, Hortobagyi GN. Recent advances in adjuvant ther-
covalently-bound tyrosinamide allows for monitoring of its fate
apy of breast cancer. Sem Oncol 1999;26(4):21}7.
in vivo. J Bioact Compat Polym 1995;10:4}13.
 Reich S, Overberg-Schmidt US, Buhrer C, Henze G. Low-dose
 Gombotz WR, Pettit DK. Biodegradable polymers for protein
chemotherapy with vinblastin and methotrexate in childhood
and peptide drug delivery. Bioconjugate Chem 1995;6(4):332}51.
desmoid tumors. J Clin Oncol 1999;17(3):1086.
 Imai T, Fujii K, Shiraishi S, Otagiri M. Alteration of phar-
 Raynal S, Nocentini S, Croisy A, Lawrence DA, Jullien P. Trans-
macokinetics and nephrotoxicity of cisplatin by alginates.
forming growth factor-1 enhances the lethal e!ects of DNA-
Hydroxylation reactions with P450 enzymes Hydroxylation reactions catalysed using P450 enzymes have a range ofapplications in drug metabolite production and lead diversificationFacilitating the identification of toxicmetabolites early in the drug developmentprocess and enabling the diversification ofidentification and preliminary toxicity andlead compounds through the generation of aacti
Diurnal Variation of Cerebrospinal Fluid Hypocretin-1(Orexin-A) Levels in Control and Depressed SubjectsRonald M. Salomon, Beth Ripley, John S. Kennedy, Benjamin Johnson,Dennis Schmidt, Jamie M. Zeitzer, Seiji Nishino, and Emmanuel Mignot Background: Hypocretins, excitatory neuropeptides at Introduction monoaminergic synapses, appear to regulate humansleep-wake cycles. Undetectable cereb