Synthesis and solution properties of deferoxamine amides

Article abstract of DOI:10.1002/1520-6017(200012)89:12<1525::AID-JPS3>3.0.CO;2-T

The poor membrane permeability and oral bioavailability of the iron chelating agent deferoxamine (DFO) mesylate result from the low octanol/water partition coefficient and high aqueous solubility. With the ultimate aim to improve biomembrane permeability while retaining the iron-binding ability of DFO, a series of more lipophilic amides were prepared by reacting the terminal primary amino group with fatty and aromatic acid chlorides or anhydrides. Octanol/water partition coefficients and equilibrium solubilities of these analogs in solvents, chosen to delineate physicochemical interactions, were determined as a function of temperature. Solid-state properties were evaluated by calorimetry. All DFO amide derivatives had higher melting points, indicating that derivatives formed strong intermolecular interactions in the solid phase. Formamidation of the primary amine of deferoxamine resulted in a 200-fold increase in the octanol/water partition coefficient and reduced aqueous solubility at least 2000-fold compared with the parent molecule. The partition coefficient increased and aqueous solubility decreased 2-fold with the addition of each methylene group in the homologous series of aliphatic amides. Solubilities of the derivatives in water-saturated octanol and hexane showed irregular profiles as a function of increasing aliphatic chain length that were attributed to intermolecular packing in the solid state. The temperature dependence of the partition coefficients was interpreted to indicate that interfacial transfer of the deferoxamine amides was, in part, affected by an apparent diminished ability to form energetically favorable interactions in the water-saturated organic phase. (C) 2000 Wiley-Liss, Inc.

Full text of DOI:10.1002/1520-6017(200012)89:12<1525::AID-JPS3>3.0.CO;2-T

Synthesis and Solution Properties of Deferoxamine Amides
PETER M. IHNAT,1,2 JONATHAN L. VENNERSTROM,1 DENNIS H. ROBINSON1
1 Department of Pharmaceutical Sciences, College of Pharmacy, University of Nebraska Medical Center, 986025
Nebraska Medical Center, Omaha, Nebraska 68198-6025
2 Schering-Plough Research Institute; K-11-2-J5, 2000 Galloping Hill Road, Kenilworth, New Jersey 07033
Received 26 August 1999; revised 3 July 2000; accepted 29 July 2000
ABSTRACT: The poor membrane permeability and oral bioavailability of the iron che-
lating agent deferoxamine (DFO) mesylate result from the low octanol/water partition
coefficient and high aqueous solubility. With the ultimate aim to improve biomembrane
permeability while retaining the iron-binding ability of DFO, a series of more lipophilic
amides were prepared by reacting the terminal primary amino group with fatty and
aromatic acid chlorides or anhydrides. Octanol/water partition coefficients and equilib-
rium solubilities of these analogs in solvents, chosen to delineate physicochemical in-
teractions, were determined as a function of temperature. Solid-state properties were
evaluated by calorimetry. All DFO amide derivatives had higher melting points, indi-
cating that derivatives formed strong intermolecular interactions in the solid phase.
Formamidation of the primary amine of deferoxamine resulted in a 200-fold increase in
the octanol/water partition coefficient and reduced aqueous solubility at least 2000-fold
compared with the parent molecule. The partition coefficient increased and aqueous
solubility decreased 2-fold with the addition of each methylene group in the homologous
series of aliphatic amides. Solubilities of the derivatives in water-saturated octanol and
hexane showed irregular profiles as a function of increasing aliphatic chain length that
were attributed to intermolecular packing in the solid state. The temperature depen-
dence of the partition coefficients was interpreted to indicate that interfacial transfer of
the deferoxamine amides was, in part, affected by an apparent diminished ability to
form energetically favorable interactions in the water-saturated organic phase. © 2000
Wiley-Liss, Inc. and the American Pharmaceutical Association J Pharm Sci 89:1525–1536, 2000
Keywords: deferoxamine; deferoxamine derivatives; iron chelators; partition coeffi-
cient; solubility; thermodynamics
iron result primarily from accidental or inten-
tional overdose, chronic blood transfusions, or
thalassemia, a genetically acquired anemia that
causes ineffective erythropoiesis and is endemic
to the Mediterranean regions. The World Health
Organization has identified thalassemia as the
most severe and widespread single locus genetic
disease in the world. Patients suffering from sys-
temic iron overload require chronic subcutaneous
injections or infusions of up to 80–120 mg/kg (4–
12g DFO/day) through an externalized central ve-
nous catheter to maintain a negative iron bal-
ance.2,3 The poor lipophilicity of DFO results in
absorption of only 15% of an orally administered
INTRODUCTION
Deferoxamine mesylate (DFO, Figure 1), the only
Federal Drug Administration (FDA) approved
drug used to remove toxic levels of iron from the
systemic circulation, strongly complexes the fer-
ric ion to form the hexadentate complex, ferriox-
amine (Kf ס 1030), which is readily excreted
through the kidneys.1 Toxic levels of systemic
Correspondence to: P.M. Ihnat (Telephone: 908-740-2101;
Fax: 908-740-2802; E-mail: peter.ihnat@spcorp.com)
Journal of Pharmaceutical Sciences, Vol. 89, 1525–1536 (2000)
© 2000 Wiley-Liss, Inc. and the American Pharmaceutical Association
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000
1525

1526
IHNAT, VENNERSTROM, AND ROBINSON
ability and antimalarial activity of the N-methyl-
anthranilamide derivative of DFO.6 In this work,
we wished to extend these efforts by constructing
a physicochemical database, starting with DFO
amide derivatives, that may lead to the identifi-
cation of an analog with optimal biomembrane
permeability. The specific objectives were to iden-
tify and evaluate the physicochemical solution
properties of DFO, synthesize a series of DFO am-
ide derivatives that demonstrate a range of solu-
tion properties, determine their aqueous and or-
ganic solvent solubilities as well as octanol/water
partition coefficients, and compare the properties
with reported characteristics necessary for bio-
membrane permeability.
Figure 1. The structure of deferoxamine (R ס H) and
the deferoxamine derivatives. The R groups are listed
in Table 2.
dose and inefficient mobilization of iron from
tissues following parenteral administration in
humans. Therefore, high dose infusions are re-
quired to mobilize iron from circulating transfer-
rin and tissue stores.2–4 Although high-dose DFO
therapy is usually free of major side effects, audi-
tory and visual abnormalities associated with
neurotoxicity have been reported.2,4 Because the
once promising orally bioavailable iron chelator
1,2-dimethyl-3-hydroxypyrid-4-one has not
advanced beyond clinical trials, DFO remains
the only available drug for the treatment for sys-
temic iron overload.3 In addition to mobilizing
systemic iron, DFO possesses moderate antima-
larial activity and has been shown to block the
iron-dependent generation of oxygen free radicals
in the body.5–7
The primary amino group of DFO does not par-
ticipate in the coordination of iron (III)(Figure 1)
and contributes to the low lipophilicity of this iron
chelator because it is protonated at physiological
pH.8 The overall hypothesis of this research is
that modification of the amino group of DFO with
lipophilic substituents should preserve the iron-
binding ability, eliminate the charge, and theoret-
ically improve biomembrane permeability allow-
ing for a more efficient access to tissue iron stores.
As a result, efficacy could be enhanced with de-
creased costs and improved patient compliance.
The more lipophilic amide derivatives are also ex-
pected to be stable to hydrolysis within the gas-
trointestinal pH range 2–7.9
Generally, the addition of one methylene group
into a molecular structure improves lipophilicity
by a factor of 3 and removal of a charged group
increases lipophilicity 100-fold.10,11 As a result,
acetylation of the primary amine of DFO could
theoretically improve lipophilicity by a factor of
300. Incremental increases in the octanol/water
partition coefficient of molecules as a result of
structural modifications have been positively cor-
related with increases in biomembrane perme-
ability, although the extent of improvement dif-
fers with each class of molecule.12–16 Indeed, this
approach has been partially successful for DFO as
exemplified by the improved membrane perme-
MATERIALS AND METHODS
Synthesis of Deferoxamine Derivatives
Aliphatic and Aromatic Amides of Deferoxamine
DFO (1 mmol) was dissolved in 10 mL of distilled
H2O and the pH adjusted to 8–9 with 5 M KOH.
An excess (6 mmol) of the appropriate acid chlo-
ride or anhydride was dissolved in 1–2 mL of
CHCl3 and added in a dropwise manner over a
period of 30 min to the DFO solution at room tem-
perature while maintaining the pH between 8
and 9. Reaction mixtures were extracted with
CHCl3 and dried with anhydrous Na2SO4. The
CHCl3 was removed in vacuo, and the oily resi-
dues of the tetracylated DFO intermediates were
redissolved in 20 mL of MeOH, placed in an ice
bath, and aerated with anhydrous ammonia gas
that was bubbled through the solutions for 1 h to
liberate the free hydroxamic acid functional
groups. Following refrigeration for at least 12 h,
the solid monoacylated deferoxamine derivatives
were collected after vacuum evaporation. The off-
white solids were recrystallized from absolute
ethanol or n-propanol17,18 (30–50% yields). The
propyl, butyl, valeryl, caproyl, octanoyl, benzoyl,
and phenylacetyl DFO derivatives were synthe-
sized from the acid chloride, and the acetamide
and succinamide derivatives were synthesized
from the corresponding anhydrides.
Acetamide Deferoxamine
Anal. Calcd for C27H50O9N6: C, 53.8% H, 8.3% N,
14.0%. Found: C, 54.0%; H, 8.4%; N, 13.9%; mp
1
175.7 °C; H NMR: ␦ 1.20–1.53 (m, 18H), 1.79 (s,
3H), 1.98 (s, 3H), 2.28 (m, 4H), 2.59 (m, 4H), 3.00
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

DEFROXAMINE AMIDES: SYNTHESIS AND SOLUBILITY
1527
(m, 6H), 3.48 (m, 6H), 7.81 (m, 3H), 9.68 (br s,
Phenylacetamide Deferoxamine
3H).
Anal. Calcd for C33H54O9N6: C, 58.4%; H, 8.0%;
N, 12.4%. Found: C, 58.2%; H, 7.9%; N, 12.2%; mp
174.5 °C; H NMR: ␦ 1.18–1.49 (m, 19H), 1.97 (s,
Propylamide Deferoxamine
1
3H), 2.26 (m, 4H), 2.57 (m, 4H), 3.00 (m, 7H),
3.38–3.45 (m, 7H), 7.30–7.81 (m, 7H), 8.04 (s, 1H),
9.6 (br s, 2H).
Anal. Calcd for C28H52O9N6: C, 54.5%; H, 8.4%;
N, 13.6%. Found: C, 54.6%; H, 8.4%; N, 13.4%; mp
176.4 °C; H NMR: ␦ 1.00 (m 4H), 1.22–1.52 (m,
1
18H), 1.97 (s, 2H), 2.04 (m, 2H), 2.29 (m, 4H), 2.60
(m, 5H), 3.00 (m, 6H), 3.40 (m, 7H), 7.84 (m, 3H),
9.79 (br s, 2H).
Succinamide Deferoxamine
Anal. Calcd for C29H51O11N6: C, 52.7%; H, 7.9%;
N, 12.7%; Found: C, 52.9%; H, 7.9%; N, 12.5%;;
mp 163 °C; 1H NMR: ␦ 1.18–1.54 (m, 19H), 1.99 (s,
3H), 2.31 (m, 8H), 2.40 (m, 4H), 3.03 (m, 6H), 3.47
(m, 7H), 7.86 (m, 3H), 9.70 (br s, 1H).
Butylamide Deferoxamine
Anal. Calcd for C29H54O9N6: C, 55.2%; H, 8.6%;
N, 13.3%. Found: C, 55.2%; H, 8.6%; N, 13.2%;
1
mp.177.9 °C; H NMR: ␦ 0.9 (m, 3H), 1.22–1.53
Formamide Deferoxamine
(m, 21H), 1.97 (s, 3H), 2.01 (m, 2H), 2.28 (m, 4H),
2.59 (m, 4H), 3.00 (m, 6H), 3.46 (m, 6H), 7.81 (m,
3H), 9.73 (br s, 2H).
DFO (1.5 mmol) and ethyl formate (20 mL) were
added to a solution of 1% triethylamine in 30 mL
of dimethylformamide (DMF), and the DFO sus-
pension was refluxed for 15 h under nitrogen. The
DFO dissolved during refluxing and the resulting
formylated DFO derivative precipitated after sub-
sequent cooling.19 The off-white precipitate of
DFO formamide was isolated by vacuum evapo-
ration and filtration followed by re-crystallization
from n-propanol (90% yield).
Valeramide Deferoxamine
Anal. Calcd for C30H56O9N6: C, 55.9%; H, 8.7%;
N, 13.0%. Found: C, 56.0%; H, 8.6%; N, 13.0%; mp
1
180.4 °C; H NMR: ␦ 0.85 (m, 3H), 1.20–1.52 (m,
22H), 1.97 (s, 3H), 2.04 (m, 3H), 2.29 (m, 4H), 2.58
(m, 4H), 3.01 (m, 6H), 3.48 (m, 6H), 7.79 (m, 3H),
9.69 (br s, 3H).
Formamide Deferoxamine
Caproylamide Deferoxamine
Anal. Calcd for C26H48O9N6: C, 53.1%; H, 8.2%;
N, 14.3%; Found: C, 53.3%; H, 8.2%; N, 14.3%; mp
157.6 °C; H NMR: ␦ 1.21–1.52 (m, 18H), 1.97 (s,
3H), 2.26 (m, 4H), 2.58 (m, 4H), 2.99-3.06 (m, 6H),
3.46 (m, 7H), 7.80 (m, 2H), 8.00 (s, 2H), 9.67 (br s,
2H).
Anal. Calcd for C31H58O9N6: C, 56.5%; H, 8.8%;
N, 12.8%; Found: C, 56.3%; H, 8.7%; N, 12.5%; mp
176.9 °C; H NMR: ␦ 0.85 (t, 3H), 1.18–1.52 (m,
25H), 1.97 (s, 3H), 2.03 (t, 2H), 2.29 (m, 5H), 2.57
(m, 4H), 3.01 (m, 6H), 3.45 (m, 6H), 7.80 (m, 3H),
9.72 (br s, 2H).
1
1
Methylsulfonamide Deferoxamine
Octanoylamide Deferoxamine
Equimolar (1.5 mmol) of DFO and anhydrous fer-
ric chloride were combined in distilled water (10
mL) and stirred for 15 min to form the deep-red,
iron (III) chelate ferrioxamine. After the pH was
adjusted to 8.5 with 5 M KOH, methylsulfonyl
chloride (2 mL) dissolved in CHCl3 (3 mL) was
added in a dropwise manner into the stirred fer-
rioxamine solution over a period of 30 min while
maintaining the pH at 8.5. The reaction solution
was extracted with CHCl3, and the volume of the
aqueous phase was reduced 50% in vacuo. A 10-
fold molar excess of EDTA was added to the con-
centrated methylsulfonamide ferroxamine solu-
tion and, after adjusting the pH to 4.0, the solu-
tion was equilibrated for 48 h at 4 °C.20 The more
Anal. Calcd for C33H62O9N6: C, 57.7%; H, 9.0%;
N, 12.2%, Found: C, 57.9%; H, 9.0%; N, 12.4%; mp
182.1 °C; H NMR: ␦ 0.86 (t, 3H), 1.23–1.49 (m,
28H), 1.97 (s, 3H), 2.02 (t, 2H), 2.26 (m, 4H), 2.57
(m, 4H), 2.98 (m, 6H), 3.45 (m, 7H), 7.80 (m, 3H),
9.72 (br s, 2H).
1
Benzoylamide Deferoxamine
Anal. Calcd for C32H52O9N6: C, 57.8%; H, 7.8%;
N, 12.7%; Found: C, 57.7%; H, 7.9%; N, 12.7%; mp
1
180.6 °C; H NMR: ␦ 1.25–1.54 (m, 18H), 1.98 (s,
3H), 2.28 (m, 4H), 2.60 (m, 4H), 3.02 (m, 4H), 3.26
(m, 3H), 3.47 (m, 5H), 7.60-7.80 (m, 7H), 8.20 (s,
1H), 9.60 (br s, 2H).
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

1528
IHNAT, VENNERSTROM, AND ROBINSON
water soluble iron(III)–EDTA complex remained
dissolved, whereas the methylsulfonamide–DFO
derivative precipitated and was collected by fil-
tration. Recrystallization from ethanol yielded
20%.
trifugation at 1000 rpm for 15 min and DFO con-
centrations in both phases were determined spec-
trophotometrically (octanol ␭ ס 210 nm, water
␭ ס 220 nm).
The partition coefficients of the DFO deriva-
tives were determined at 25 and 37 °C, and the
distribution coefficients in phosphate buffer (pH
7.4, I ס 0.1) were determined at 25 °C. Analog
concentrations in octanol were analyzed directly
by spectrophotometry (␭ ס 210 nm), whereas the
aqueous phase was diluted 4:1 with 0.5 N NaOH
prior to spectrophotometric analysis (␭ ס 230
nm).
Methylsulfonamide Deferoxamine
Anal. Calcd for C28H50O10N6S: C, 48.9%; H, 7.8%;
N, 13.2%; Found: C, 48.9%; H, 7.8%; N, 13.2%; mp
1
142.8 °C; H NMR: ␦ 1.23–1.50 (m, 18H), 1.97 (s,
3H), 2.27 (m, 5H), 2.58 (m, 4H), 2.86 (s, 4H), 3.00
(m, 4H), 3.46 (m, 8H), 7.81 (m, 2H), 9.67 (br s,
2H).
A
A
͓ ͔
͓ ͔
o
P =
(1)
Thermal Analysis
w
Differential scanning calorimetry (DSC; Shi-
madzu DSC-50) was used to determine the melt-
ing points (Tm) and heats of fusion (⌬Hf) of the
DFO amides. Approximately 0.8 to 1.2 mg of each
DFO analog was sealed in an aluminum pan and
heated at 10 °C/min under N2 atmosphere (20
mL/min flow rate). Thermal profiles of the ana-
logs revealed one corresponding endothermic
peak. Temperature cycling experiments con-
firmed the absence of polymorphism. Melting
points and heats of fusion were calculated using
associated software.
The intrinsic partition coefficient (P) was calcu-
lated from the ratio of the solute concentration
in water-saturated octanol ([A]O) and octanol-
saturated water ([A]W) at equilibrium. The ther-
modynamic partition coefficient (Px), which is
used to evaluate the thermodynamic parameters
associated with solute partitioning, was calcu-
lated from the ratio of the solute mole fraction
o
concentrations in water saturated octanol (X2 )
w
and octanol saturated water (X2 ):
X2o
X2w
Px =
(2)
Spectrophotometric Analysis
A Shimadzu UV 160U double beam spectropho-
tometer was used to generate calibration curves
of all DFO amides at 230 nm in 0.1 M NaOH and
at 208 or 210 nm in water-saturated octanol. For
both solvents, the concentration range was 5–55
␮g/mL, and the mean absorbance of triplicate
measurements were recorded.
The free energy of solute partitioning between
water and octanol (⌬Gtr, kJ/mol) was calculated
from the thermodynamic partition coefficient us-
ing eq. 3:
⌬Gtr = 2.303 R T log Px
(3)
where R is the universal gas constant (8.314
J/mol K).
Octanol/Aqueous Distribution Coefficients
All glassware used in these experiments was
rinsed with 0.1 M HCl and freshly distilled, deion-
ized H2O to ensure the absence of glass-bound
iron. Phosphate buffers of constant ionic strength
(0.1), were prepared at pH 5.0, 7.4, and 8.0. The
octanol and aqueous phases were mutually pre-
saturated by vigorous stirring for at least 12 h.
Equal volumes (10 mL) of the aqueous DFO
solutions and organic phases were combined in a
glass-stoppered flask, placed in a temperature-
controlled water-bath, and stirred for 24 h. After
equilibration, the phases were separated by cen-
Aqueous Solubilities
The solubilities of the DFO derivatives were
evaluated in triplicate in freshly distilled deion-
ized water, distilled water acidified with 0.01 M
HCl to pH 4–5, and phosphate buffer (pH 7.4, I ס
0.1) using glassware presoaked in 0.1 M HCI and
rinsed in deionized water. Vials, containing an
excess of each DFO analog, were placed in a tem-
perature-controlled environment and agitated for
at least 4 days while being periodically monitored
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

DEFROXAMINE AMIDES: SYNTHESIS AND SOLUBILITY
Table 1. Octanol/Water and Octanol/Buffer Distribution Coefficients of Deferoxaminea
1529
c
Temperature, °C
Solvent
water
Pb
Px
⌬G (kJ/mol)d
15
25
37
25
25
25
0.0013 0.0003
0.0022 0.0001
0.0025 0.0002
0.0037 0.0007
0.0141 0.0020
0.0084 0.0004
0.011 0.002
0.019 0.001
0.022 0.002
10.83 0.44
9.82 0.52
9.84 0.89
water
water
buffer pH 5.0e
buffer pH 7.4e
buffer pH 8.0e
—
—
—
—
—
—
a n ס 3.
b Calculated using eq 1.
c Calculated using eq 2.
d Calculated using eq 3.
e Phosphate buffer, ionic strength ס 0.1.
for saturation. At equilibrium, samples were fil-
tered and analyzed as already described.
ficient refers to the ratio of the concentrations of
the unionized, monomeric solute in both phases
whereas the distribution coefficient refers to the
ratio of the multiple ionic species partitioning be-
tween both phases.22 The profile of the pH-
dependent distribution of the ionic species of DFO
was determined using previously reported pKa
values and equations describing the relevant
ionic equilibria in solution (Figure 2).8,21
The lower dielectric constant of water-
saturated octanol (8.58) relative to octanol-
saturated water (76.6) indicates that octanol can-
not satisfy the electrical requirements of charged
moieties as well as water.12 In phosphate buffer
(pH 5.0, 7.4, and 8.0), ionic species of DFO are free
to form coulombic and hydrogen-bonding interac-
tions with water or cosolutes. The increase in the
distribution coefficient in buffer relative to water
may result from ion-pair formation at the organic
interface with the mesylate or phosphate counter-
ions that partially neutralize the charges and re-
Octanol and Hexane Solubilities
The solubilities of each DFO amide in water-
saturated octanol and in n-hexane were deter-
mined at 25 °C by the procedure previously de-
scribed. After saturation was confirmed, the sus-
pensions were centrifuged at 7000 rpm for 20
min, and the concentrations of the DFO analogs
in the clear octanol fraction were analyzed spec-
trophotometrically (␭ ס 210 nm). The suspen-
sions in hexane were centrifuged at 2000 rpm for
20 min, and the supernatant was evaporated in a
vacuum oven at ambient temperature. The resi-
due was dissolved in 0.1 N NaOH and analyzed
spectrophotometrically (␭ ס 230 nm).
RESULTS AND DISCUSSION
The hydrophilic nature of DFO (Figure 1) is char-
acterized by the low distribution coefficients
(Table 1) and high aqueous solubilities (>38.0
4.4 g/100 mL) in the pH range of 4.2 to 8.0 at
25 °C.21 Generally, the octanol/water distribution
coefficients increased with temperature and were
2–4-fold lower than the octanol/buffer distribu-
tion coefficients determined at 25 °C (Table 1).
Therefore, because DFO is a relatively large, flex-
ible molecule with polar hydroxamic acid and am-
ide groups, cosolutes will contribute to the array
of complex solvent–solute and solute–solute inter-
actions in solution and at the organic/aqueous in-
terface. For ionizable solutes, the interfacial dis-
tribution between an aqueous and organic phase
is pH dependent.12,22 The intrinsic partition coef-
Figure 2. The distribution profile of deferoxamine
species in solution as a function of pH. The symbol ␣
refers to the fraction of each species.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

1530
IHNAT, VENNERSTROM, AND ROBINSON
duce resistance to incorporation into the octanol
phase (Table 1).23 At pH 7.4 and 8.0, the opposite
charges of the zwitterionic form of DFO (H3DFO,
Figure 2) may be neutralized due to mutual at-
traction facilitated by conformational flexibility.
This neutralization may facilitate accommodation
of DFO into the lower dielectric environment of
the organic phase.23
The thermodynamics of the partitioning of
DFO was investigated in the octanol/water sys-
tem to reduce additional intermolecular interac-
tions that occur in presence of buffer ions (Table
1). The equation generated by linear regression of
the van’t Hoff plot was (r ס 0.931, n ס3):
gen bonds between water and DFO should be off-
set by the energy gained from hydrogen-bond
formation within water-saturated octanol and re-
flected in a lower enthalpy of partitioning. How-
ever, spectroscopic evidence in aqueous solution
indicates that an intramolecular hydrogen bond
is formed between the hydroxamic acid proton
and adjacent carbonyl group, in lower molecular
weight hydroxamic acids.25,26 Because of the
proximity of the carbonyl to the hydroxamic pro-
ton, the hydrogen bond may be too stable to re-
form in the water-saturated organic. Therefore, it
appears that attractive forces between water and
the charged amino group of DFO are more likely
than hydrogen bonding to be responsible for the
high aqueous solubility and low partition coeffi-
cients. Elimination of the charged amino group
through derivatization should improve the lipo-
philicity and may enhance biomembrane perme-
ability.11,14
The melting points and related thermal prop-
erties for the DFO derivatives are summarized in
Table 2. Formamidation of the primary amino
group of DFO increased the octanol/water parti-
tion coefficient >200-fold, and the octanol/buffer
distribution coefficient 44-fold at 25 °C (Table 3).
Addition of methyl or methylene groups further
increased the partition coefficients in 2-fold in-
crements up to the butylamide (COC3H7) deriva-
tive. There was an abrupt 2.9-fold increase in par-
titioning between the butylamide and valeramide
(COC4H9) DFO amides, after which no significant
1
T
log P = 1247.229 ע 490.614͒ + 2.401 ע 1.646
͑
x
(4)
The positive free energy (⌬G) values, calculated
from the thermodynamic partition coefficients
with eq. 3, indicate that interfacial transfer of
DFO is energetically unfavorable. The endother-
mic enthalpy (⌬H ס 23.88 9.39 kJ/mol) signifies
that the system must absorb energy to accommo-
date DFO into the hydrated organic phase.
Based on the number of hydrogen-bonding
groups, the empirical hydrogen-bonding potential
of DFO was determined to be 14.24 Consequently,
solute–solvent hydrogen bonding should contrib-
ute to the solubility of DFO in both water and
octanol, and the energy required to disrupt hydro-
Table 2. Molecular Weights and Thermal Properties of the Deferoxamine Derivatives
Melting
Point (°C)
⌬Hf
⌬Sf
Name
Derivative (R)a
MW
(kJ/mol)c
(kJ/mol K)d
deferoxaminea
formamide-
acetamide-
propylamide-
butylamide-
valerylamide-
caproylamide-
octanoylamide-
benzoylamide-
phenylacetamide-
succinamide-
methylsulfonamide-
DFOa
COH
COCH3
563
588
602
616
630
644
658
686
664
678
660
638
137.9b
157.6
175.7
176.4
177.9
180.4
176.9
182.1
180.6
174.5
163.0
142.8
105.3
92.93
118.4
116.9
111.4
123.1
119.2
130.7
107.0
119.0
101.0
117.8
0.26
0.22
0.26
0.26
0.25
0.27
0.26
0.29
0.24
0.27
0.24
0.28
COC2H5
COC3H7
COC4H9
COC5H11
COC7H15
COC6H5
COCH2C6H5
CO(CH2)2COOH
SO2CH3
a Figure 1.
b Deferoxamine mesylate.
c Heat of fusion.
d Entropy of fusion; ⌬Sf ס ⌬Hf /Tm; 0.26 0.02 (kJ/mol K).
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

DEFROXAMINE AMIDES: SYNTHESIS AND SOLUBILITY
1531
Table 3. Distribution Coefficients of the Deferoxamine Derivatives
Octanol/Water (P)b Octanol/Buffer (Pd)c
Derivative (R)a
T ס 25 °C
T ס 37 °C
T ס 25 °C
COH
COCH3
COC2H5
COC3H7
0.47 0.10
0.93 0.15
1.48 0.36
2.61 0.71
7.61 1.16
8.30 0.71
8.70 2.17
1.99 0.59
1.67 0.41
0.31 0.03
0.17 0.04
0.80 0.06
1.27 0.08
1.81 0.09
4.11 0.36
11.2 0.53
26.7 3.47
44.9 10.7
7.31 1.06
7.95 1.69
0.30 0.03
0.44 0.03
0.62 0.05
1.45 0.21
1.57 0.05
4.56 0.53
7.65 0.54
12.5 2.10
12.3 0.80
3.34 0.17
1.79 0.27
0.02 0.003
0.43 0.11
COC4H9
COC5H11
COC7H15
COC6H5
COCH2C6H5
CO(CH2)2COOH
SO2CH3
a n ס 3.
b Intrinsic partition coefficient.
c Phosphate buffer, pH 7.4; ionic strength ס 0.1.
increase in partition coefficient was observed
(Table 3). Similarly, the octanol/buffer distribu-
tion coefficient increased 2 -fold with each addi-
tion of methyl or methylene group through cap-
roylamide DFO (COC5H11). Commencing with
valeramide DFO (COC4H9), the partition coeffi-
cients of the methylsulfonic (SO2CH3), aromatic
(COC6H5, COCH2C6H5), and longer-chain ali-
phatic derivatives were more sensitive to tem-
perature than the short-chain aliphatic and suc-
cinamide (CO(CH2)2COOH) derivatives (Table 3).
The correlation between the distribution and par-
tition coefficients of all analogs except succin-
amide DFO (r ס 0.985), related by the fraction of
ionized species in solution at pH 7.4, shows that
the pK values for the analogs are similar.21,22
The aqueous solubilities of the DFO analogs in
phosphate buffer, pH 7.4, and distilled water are
summarized in Table 4. The intrinsic solubility
(So) or saturation solubility of the un-ionized spe-
cies was obtained in acidified, distilled water to
ensure that the solution pH was <8.5 or pK1 of the
first hydroxamic acid group.8,17 Formamidation of
the primary amino group of DFO reduced aque-
ous solubility at least 2000-fold. Additional meth-
ylene groups further reduced aqueous solubility
Table 4. Solubilities (␮M) of the Deferoxamine Derivatives at 25 °Ca
n-Octanole
n-Hexane
b
c
d
R
SPB
SO
SW
COH
COCH3
COC2H5
COC3H7
179 1.9
26.6 1.7
27.6 1.6
17.5 1.6
6.21 1.5
4.56 1.5
4.37 1.4
7.53 1.5
7.37 1.5
<9000
201 2.0
48.2 8.3
29.2 4.9
15.9 4.8
7.76 3.1
7.60 3.0
5.83 2.9
12.1 3.0
13.3 1.5
147 13
255 14
78.1 5.0
50.3 5.0
22.2 1.6
9.12 1.5
7.76 1.6
7.29 2.9
13.6 4.5
11.8 3.0
253 18
81.6 5.1
73.1 8.3
56.8 6.5
44.4 3.2
59.0 3.1
86.6 4.6
65.6 4.4
33.1 6.0
28.0 1.5
48.5 1.5
45.5 1.6
0.32 0.15
1.09 0.10
1.34 0.60
1.98 0.52
1.74 0.04
2.30 0.34
1.79 0.42
1.54 0.65
1.92 0.39
0.44 0.10
0.99 0.35
COC4H9
COC5H11
COC7H15
COC6H5
COCH2C6H5
CO(CH2)2COOH
SO2CH3
113 4.7
143 4.7
176 13
a n ס 3.
b Solubility in phosphate buffer, pH 7.4; ionic strength ס 0.1.
c Intrinsic solubility.
d Solubility in unbuffered, distilled water.
e Water-saturated octanol.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

1532
IHNAT, VENNERSTROM, AND ROBINSON
in 2-fold increments (Table 4). As expected,
there was an inverse relationship between parti-
tioning and intrinsic aqueous solubility for all 11
derivatives, but this correlation was most signifi-
cant for the aliphatic homologs.21 Although the
first hydroxamic group was 7% ionized at pH
7.4, the solubilities of the DFO analogs in phos-
phate buffer were comparable with their intrinsic
aqueous solubilities (Table 4). Conversely, the
analogs were most soluble in distilled water (5%
ionized, mean pH 7.3 0.7) due to partial ioniza-
tion in the absence of additional cosolute effects.
The solubilities of the DFO analogs were also
evaluated in water-saturated octanol and n-
hexane (Table 4). van der Waals interactions
acted to dissolve the DFO analogs in hexane,
whereas a combination of van der Waals forces
and hydrogen bonding facilitated solubility in wa-
ter-saturated octanol.12,13All derivatives were
more soluble in water-saturated octanol than wa-
ter or hexane because of intermolecular solute–
solvent interactions with both polar and nonpolar
regions (Table 4, Figure 3).13 However, (at 25 °C),
the solubilities of the aliphatic amides in hy-
drated octanol decreased by a factor of 1.23 0.10
with each additional carbon atom up to butyl-
amide DFO, after which there was an irregular
solubility relationship. The abrupt upward trend
in the octanol solubility profile of the aliphatic
DFO homologs may represent the point where the
aliphatic chain exerted more influence on solubil-
ity than the parent molecule (Figure 3). Con-
versely, the solubilities of the DFO analogs in
hexane increased in an irregular pattern with
each additional carbon atom (Table 4, Figure 3).
Solubility in hexane was lower than in water and
water-saturated octanol and relatively insensi-
tive to incremental addition of carbon atoms (Fig-
ure 3).
Saturated aliphatic homologs of alkoxy, hy-
droxy, and amino benzoates also exhibited irregu-
lar patterns of solubility. These irregularities
may be attributed to the structure-dependent
crystal lattice energies in the solid state and were
previously reported in homologs of 4 to 6 carbon
atom chain lengths.15,27,28 The melting points and
heats of fusion of the aliphatic DFO homologs also
increased irregularly with increasing carbon
number (Table 2). The irregular solubility pat-
terns and thermal properties may result from dis-
crete differences in packing efficiency in the solid
state, intramolecular degrees of freedom, and sol-
vent cavity accommodation of the solute as dic-
tated by carbon chain length.15,28
Except for formamide (COH), succinamide
(CO(CH2)2COOH), and methylsulfonamide DFO
(SO2CH3), the melting points of the remaining
analogs were within 10 °C. The compounds with
the highest solubilities in aqueous and octanol so-
lution also had the lowest melting points (Tables
2 and 4). The relatively more bulky groups prob-
ably formed weaker intermolecular packing inter-
actions that improved the escaping tendency from
the solid phase.23,28
For the aliphatic DFO amide analogs, the semi-
logarithmic correlation between the hexane solu-
bilities and the melting points (r ס 0.934, n ס 7)
indicated that the disruption of the crystal lattice,
and not solvent–solute van der Waals interac-
tions, dominated the solution processes in hex-
ane.15 Regardless of modification, the entropies of
fusion for all the DFO analogs were 0.26 0.02
kJ/mol K (Table 2). Therefore, in the solid state,
substituents affected the attractive and repulsive
forces, reflected in the enthalpy of fusion, more
than the conformational degrees of freedom. The
long aliphatic chains may stack to maximize con-
tacts and despite being bulky, the aromatic de-
rivatives may have formed ␲ bonds with adjacent
conjugated rings.27,28
The octanol/water partition coefficient of the
DFO analogs may be calculated by the ratio of the
analog solubility in octanol to water.29 The corre-
lation between the experimental octanol/water
partition coefficient and the solubility ratio at 25
°C (Table 5) was:
log P = 1.033 ע 0.058 log SR − 0.224
Figure 3. Comparison of water-saturated octanol,
water, and n-hexane solubilities of the aliphatic defer-
oxamine analogs (T ס 25 °C).
0.080
(5)
where n ס 11 and r ס 0.986. Omitting the suc-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

DEFROXAMINE AMIDES: SYNTHESIS AND SOLUBILITY
1533
Table 5. Solubility Ratios of the
Deferoxamine Derivatives
The enthalpy of partitioning represents the affin-
ity of the solute for the solvent phase as a function
of the solute–solvent and solvent–solvent interac-
tions. The magnitude of the entropy of partition-
ing is affected by the conformation of the solute in
the solvent and the conformational relationship
between the solute and the solvent.23 Assuming
linearity in the Van’t Hoff relationship between
25 and 37 °C, the enthalpy of partitioning may be
calculated when the thermodynamic partition co-
efficients at two temperatures, Px2 and Px1, are
known:
R
SRa
SRb
COH
COCH3
COC2H5
COC3H7
0.41
1.52
1.94
2.80
7.60
8.62
11.0
2.75
2.11
0.33
0.32
0.54
2.56
1.65
2.17
8.00
13.0
9.00
4.00
4.20
—
COC4H9
COC5H11
COC7H15
COC6H5
COCH2C6H5
CO(CH2)2COOH
SO2CH3
Px2
Px1
⌬Htr
2.303 R
T2 − T1
T2T1
0.53
log
=
(8)
ͩ ͪ ͩ ͪͩ ͪ
a
Octanol solubility/intrinsic solubility (S0) (data from
Table 4).
b Octanol solubility/phosphate buffer solubility (SPB) (data
Although the intercepts of the Van’t Hoff relation-
ship were not available, the entropy of partition-
ing at a temperature between 25 and 37 °C may
be estimated using eq. 7.22
from Table 4).
cinamide derivative (CO(CH2)2COOH) from the
regression due to a pH-dependent increase in
aqueous solubility at pH 7.4, the relationship be-
tween the octanol/buffer distribution coefficient
and solubility ratio at 25 °C was:
The octanol/water partition coefficients of the
DFO analogs were determined at 25 and 37 °C
(Tables 3 and 6). The free energies of partitioning
were negative and decreased with increasing tem-
perature. The enthalpies of partitioning were
positive for all compounds except succinamide
DFO (CO(CH2)2COOH). Similar to smaller mo-
lecular weight hydroxamic acids and DFO, the
analogs probably form an intramolecular hydro-
gen bond between the hydroxamic proton and ad-
jacent carbonyl. This bond may diminish their
ability to hydrogen bond within polar regions of
the water-saturated octanol. Moreover, the en-
ergy required to perturb the “structure” of water-
saturated octanol, by incorporation of the solute,
may also contribute to the positive enthalpy of
partitioning.13 Because the enthalpies of parti-
tioning were positive for the analogs, except suc-
cinamide DFO (CO(CH2)2COOH), a significant
entropic contribution to partitioning may have
helped to drive the reaction forward. The undis-
sociated carboxylic acid group of succinamide
DFO (CO(CH2)2COOH) may have formed a hy-
drogen bond within the water-saturated octanol
resulting in the negative enthalpy of partitioning.
In summary, these results demonstrate that
conversion of the protonated amino group of DFO
to a variety of amides markedly decreased aque-
ous solubility and substantially increased lipo-
philicity relative to the parent molecule. The ob-
served trends in the change in aqueous solubili-
ties and partition coefficients for these analogs
were consistent with those expected from incre-
log P = 0.989 ע 0.141 log SR − 0.064
0.210
(6)
where n ס 10 and r ס 0.927. Using the solubility
ratio to estimate the partition coefficient was
valid because the solubilities of the DFO analogs
in both solvents were < 0.01 M.29 Because the
solubility ratios were in excellent agreement with
the partition coefficients, the increase in parti-
tioning between formamide DFO (COH) and bu-
tylamide DFO (COC3H7) was related to a more
rapid decline in aqueous solubility relative to
solubility in water-saturated octanol with addi-
tional carbon atoms. The increase in partitioning
between butylamide DFO (COC3H7) and val-
eramide DFO (COC4H9) may correspond to the
point where the alkyl chain primarily influences
solubility in water-saturated octanol while aque-
ous solubility continues to decline (Figure 3).
The free energy of partitioning of the DFO ana-
logs between water and octanol (⌬Gtr, kJ/mol)
was calculated from the thermodynamic partition
coefficient (Px; eq. 3). The free energy of partition-
ing is comprised of the enthalpy (⌬Htr) and en-
tropy (⌬Str):
⌬Gtr = ⌬Htr − T⌬Str
(7)
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

1534
IHNAT, VENNERSTROM, AND ROBINSON
Table 6. Thermodynamic Parameters of the Partitioning of the Deferoxamine Analogs
⌬G (kJ/mol)a
⌬H
⌬S
R
25 °C
37 °C
(kJ/mol)b
(kJ/mol K)c
COH
COCH3
COC2H5
COC3H7
−3.49 0.71
−5.20 0.82
−6.37 0.88
−7.75 1.91
−10.5 1.30
−10.6 0.60
−10.5 2.00
−6.76 1.33
−6.47 1.29
−2.46 0.34
−1.21 0.28
−5.03 0.34
−6.20 0.39
−7.15 0.49
−9.21 0.88
−11.8 0.60
−14.1 1.90
−15.4 3.60
−10.7 1.60
−11.1 2.10
−2.45 0.24
−3.50 0.21
34.6 7.40
19.7 3.30
12.8 2.00
28.7 7.60
21.9 2.90
74.7 10.1
110 34.0
91.5 22.3
108 29.3
−2.66 0.45
55.6 13.3
0.13 0.03
0.08 0.01
0.06 0.01
0.12 0.03
0.11 0.01
0.29 0.03
0.41 0.11
0.33 0.08
0.38 0.10
−0.001 10−4
0.19 0.05
COC4H9
COC5H11
COC7H15
COC6H5
COCH2C6H5
CO(CH2)2COOH
SO2CH3
a Calculated using eq. 3.
b Calculated using eq. 8.
c Calculated using eq. 7.
mental addition of carbon atoms. However, aque-
ous solubilities decreased more than expected,
whereas lipophilicities did not increase to the de-
gree expected. Also, the aliphatic amide DFO de-
rivatives were relatively insoluble in the water-
saturated octanol and hexane, indicating that the
solid state was preferred to solution. Indeed, ali-
phatic and aromatic modifications, through inter-
molecular packing in the solid state, increased
the melting points compared with DFO. This re-
sult is contrary to previous work with homologous
series in which systematic declines in melting
points as a function of increasing chain lengths or
aromatic substituents were observed.15,28,30
Early studies of biomembrane permeability of-
ten concluded that the principal driving force was
the gain in entropy resulting from the randomiza-
tion of water molecules, surrounding a lipophilic
solute (hydrophobic effect), on incorporation into
the organic phase. These studies commonly used
a saturated alkane/water partition coefficient as a
reference system.14,24 A number of studies have
also shown that solutes that possess a strong abil-
ity to hydrogen bond in aqueous solution may not
readily diffuse due to a high desolvation energy at
the membrane interface.31 Because the octanol/
water system has been adopted as the reference
for screening solutes for permeability potential,
excellent correlations have been established be-
tween passive membrane diffusion and the parti-
tion coefficient for a variety of molecules. Hy-
drated octanol possesses polar and apolar local-
ized environments similar to the interior of
biomembranes.13 Hence, a substituent decreases
biomembrane permeability if it establishes at-
tractive forces between the solute and water and
increases membrane permeability if it enhances
attractive forces between the solute and mem-
brane. The solute must possess the ability to dif-
fuse across the aqueous boundary layer adjacent
to the membrane yet retain the capacity to form
energetically favorable interactions in the mem-
brane interior. Ideal modifications that facilitate
passive transmembrane diffusion must alter
physicochemical properties of a molecule by bal-
ancing energetically favorable solute–solvent in-
teractions in both aqueous and organic phases
with the entropic contribution.14
The permeabilities of selected deferoxamine
derivatives were evaluated in preliminary experi-
ments using the Caco-2 model. Approximately
1–1.6% of the acetamide (COCH3) and propyl-
amide (COC2H5) amides had permeated after 4 h.
Also, the benzoylamide (COC6H5) and methylsul-
fonamide (SO2CH3) accumulations in the receiver
compartments of the Transwell௠ plates were 1.7
and 2.2%, respectively, of the donor concentra-
tions after 4 h. By comparison, 0.05% DFO had
permeated after 4 h. It remains to be seen wheth-
er the water-soluble formamide (COH) and suc-
cinamide (CO(CH2)2COOH) DFO analogs exhibit
improved membrane permeabilities because their
partition coefficients were 50-fold higher than
DFO and the octanol solubilities were comparable
to the solubilities of the aliphatic and aromatic
analogs. Therefore, the conclusions derived from
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

DEFROXAMINE AMIDES: SYNTHESIS AND SOLUBILITY
1535
the physicochemical solution properties, regard-
REFERENCES
ing the biomembrane permeabilities of the DFO
amides, are supported by these preliminary
Caco-2 experiments.21
1. Keberle H. 1964. The biochemistry of desferriox-
amine and its relation to iron metabolism. Ann NY
Acad Sci 119:758–768.
The transcellular permeability constant is di-
rectly proportional to the diffusion coefficient
through the membrane and the diffusion coeffi-
cient is inversely proportional to the molecular
size (conformation) or approximately (molecular
weight)1/3. Therefore, the molecular weight of
the derivatives does affect passive transcellular
absorption. However, because the molecular
weights between the smallest and largest defer-
oxamine amides differed by only 17%, molecular
weight differences are not likely to affect passive
diffusion compared with other factors such as par-
tition coefficient, aqueous solubility and molecu-
lar size (conformation).24
Although the charge of the primary amino
group was eliminated and the analogs possessed
improved lipophilicity, they also demonstrated
weak solvent–solute interactions, which sug-
gested that passive, transmembrane diffusion
would be difficult. Nevertheless, as demonstrated
by the parabolic relationship between lipophilic-
ity and biological activity, an optimum partition
coefficient is usually identified within each group
of homologous analogs.16 Accordingly, it is pos-
sible that more water-soluble candidates, which
show moderate improvements in lipophilicities,
will demonstrate better transmembrane diffusion
than the relatively more hydrophobic aliphatic
amides. These permeability and biological activ-
ity studies remain to be completed.
2. Olivieri NF, Brittenham GM, Matsui D, Berko-
vitch M, Blendis LM, Cameron RG, McClelland
RA, Liu PP, Templeton DM, Koren G. 1995. N Engl
J Med 332:918–922.
3. Schnebli HP, Hassan I, Hamilton KO, Lynch S, Jin
Y, Nick HP, Peter HH, Junker Walker U, Ziel R,
Khanna SC, Dean R, Bergeron RJ. 1994. Toward
better chelation therapy: Current concepts and re-
search strategy. In: Bergeron RJ, Brittenham GM,
editors. The development of iron chelators for clini-
cal use. Boca Raton, FL: CRC Press, pp 131–147.
4. Lloyd JB, Cable H, Rice-Evans C. 1991. Evidence
that desferrioxamine cannot enter cells by passive
diffusion. Biochem Pharmacol 41:1361–1363.
5. Basco LK, LeBras J. 1993. In vitro activity of chlo-
roquine and quinine in combination with desferri-
oxamine against Plasmodium falciparum. Am
J Haematol 42:389–391.
6. Mabeza GF, Loyevsky M, Gordeuk VR, Weiss G.
1999. Iron chelation therapy for malaria. Pharma-
col Ther 81:53–75.
7. Halliwell B. 1989. Protection against tissue dam-
age in vivo by desferrioxamine: What is its mecha-
nism of action? Free Rad Biol Med 7:645–651.
8. Ihnat PM, Robinson DH. 1993. Potentiometric de-
termination of the thermodynamic ionization con-
stants of deferoxamine. J Pharm Sci 82:110–112.
9. Pitman IH, 1981. Prodrugs of amides, imides and
amines. Med Res Rev 1:189–214.
10. Leo A, Hansch C, Elkins D. 1971. Chem Rev 71:
525–616.
11. Ho NFH, Park YJ, Morozowich W, Higuchi WI.
1977. Physical model approach to the design of
drugs with improved intestinal absorption. In:
Roche EB, editor. Design of biopharmaceutical
properties through prodrugs and analogs. Wash-
ington, DC:American Pharmaceutical Association,
pp 136–227.
12. Smith RN, Hansch C, Ames MM. 1975. Selection of
a reference partitioning system for drug design
work. J Pharm Sci 64:559–606.
13. Franks NP, Abraham MH, Lieb WR. 1993. Molecu-
lar organization of liquid n-octanol: An x-ray dif-
fraction analysis. J Pharm Sci 82:466–470.
14. Diamond JM, Wright EM. 1969. Molecular forces
governing nonelectrolyte permeation through cell
membranes. Proc Roy Soc B 172:273–316.
15. Yalkowsky SH, Flynn GL, Slunick TG. 1972.
J Pharm Sci 61:852–857.
The DFO derivatives evaluated in this study
form a basis on which alternative modifications
may lead to analogs with optimized physicochem-
ical solution properties that lead to improved bio-
membrane permeabilities. The information gath-
ered from this study suggests that derivatization
of the primary amino group of DFO with more
bulky, polar, and unsaturated moieties would de-
crease the melting points and may enhance the
ability to establish solute–solvent interactions.
ACKNOWLEDGMENTS
This research was supported by the American
Heart Association (Nebraska Affiliate) and the
University of Nebraska Medical Center. P. Ihnat
was supported by the American Foundation for
Pharmaceutical Education.
16. Hansch C, Clayton JM. 1973. Lipophilic character
and biological activity of drugs II: The parabolic
case. J Pharm Sci 62:1–21.
17. Schwarzenbach G, Schwarzenbach K. 1963. Hy-
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000

1536
IHNAT, VENNERSTROM, AND ROBINSON
droxoamtkomplexe I: Die stabilitat der eisen (III)-
komplexe einfacher hydroxamsauren und des fer-
rioxamins B. Helv Chim Acta 46:1390–1400.
18. Rodgers SJ, Raymond KN. 1983. Ferric ion seques-
tering agents 11: Synthesis and kinetics of iron re-
moval from transferrin of catechoyl derivatives of
desferrioxamine B. J Med Chem 26:439–442.
19. Schmidhammer H, Brossi A. 1982. Synthesis of
(-)and (+)-2-hydroxy-6-keto-N-methylmorphinans,
their O-methyl ethers and 2-deoxy congeners. Can
J Chem 60:3055–3060.
20. Monzyk B, Crumbliss AL. 1983. Factors that influ-
ence siderophore mediated iron bioavailability: Ca-
talysis of interligand iron (III) transfer from ferri-
oxamine B to EDTA by hydroxamic acids. J Inorg
Biochem 19:19–39.
21. Ihnat PM. 1995. Design and characterization of de-
feroxamine derivatives with improved biomem-
brane permeability. Ph.D. Dissertation, University
of Nebraska Medical Center, Omaha, Nebraska.
22. Dearden JC, Bresnen GM. 1988. The measurement
of partition coefficients. Quant Struct-Act Relat 7:
133–144.
24. Stein WD. 1967 In: The movement of molecules
across cell membranes. New York: Academic Press.
25. Bauer L, Exner O. 1974. The chemistry of hydrox-
amic acids and N-hydroxyimides. Angew Chem, Int
Ed Engl 13:376–383.
26. Lipczynska-Kochany E, Iwamura H. 1982. Oxy-
gen-17 NMR studies of the structures of benzohy-
droxamic acids and benzohydroxamate ions in so-
lution. J Org Chem 47:5277–5282.
27. Buckton G, Beezer AE, Denyer SP, Russell SJ.
1991. Observations on the biopharmaceutical im-
portance of chain length in chemically related com-
pounds. Int J Pharm 73:1–7.
28. Forster S, Buckton G, Beezer AE. 1991. The impor-
tance of chain length on the wettability and solu-
bility of organic homologs. Int J Pharm 72:29–34.
29. Yalkowsky SH, Valvani SC, Roseman TJ. 1983.
Solubility and partitioning IV: Octanol solubility
and octanol-water partition coefficients. J Pharm
Sci 72:866–870.
30. Schwartz PA, Paruta AN. 1976. Solution thermo-
dynamics of alkyl p-aminobenzoates. J Pharm Sci
65:252–257.
23. Grand DJW, Higuchi T. 1990. In: Solubility behav-
ior of organic compounds. New York: John Wiley &
Sons Inc.
31. Conradi RA, Hilgers AR, Ho NFH, Burton PS.
1991. The influence of peptide structure on trans-
port across Caco-2 cells. Pharm Res 8:1453–1460.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 89, NO. 12, DECEMBER 2000