Synthesis, physicochemical properties, and evaluation of N-substituted- 2-alkyl-3-hydroxy-4(1H)-pyridinones

Article abstract of DOI:10.1021/jm9707784

The synthesis of a range of 3-hydroxy-4(1H)-pyridinones with potential for the chelation of iron(III) is described. The pK(a) values of respective ligands and the stability constants of their iron(III) complexes are presented. The distribution coefficient values of a range of 48 hydroxypyridinones and their corresponding iron(III) complexes between 1- octanol and MOPS buffer (pH 7.4) are reported. The range of log D(complex) values covers 7 orders of magnitude. The results suggest the existence of a biphasic relationship between the distribution coefficient values of the chelator and the corresponding iron(III) complexes. For ligands with a log D(ligand) = -1, a linear relationship exists with a value of the slope 2.53, whereas with ligands with a log D(ligand) < -1, a linear relationship exists with a slope of 0.49. The reduced slope for the more hydrophilic molecules of the series offers some advantage for this type of hydroxypyridinone as the distribution coefficients for such complexes do not change so rapidly with increasing ligand hydrophilicity. The ability of selected 3- hydroxypyridinones to facilitate the excretion of iron in bile was investigated in non, iron-overloaded, bile duct-cannulated rats and in a [⁵⁹Fe]ferritin-loaded rat model. Both systems compare the ability of chelators to remove iron from the liver, the prime target organ in thalassemia. The N-(hydroxyalkyl)-3-hydroxypyridin-4-ones are demonstrated to be orally active under the in vivo conditions adopted. Thus both 1- (hydroxyalkyl)- and 1-(carboxyalkyl)pyridinones are able to remove iron from the liver. Although 1-(carboxyalkyl)hydroxypyridinones are: active, they do not demonstrate any clear advantage over Deferiprone (1,2-dimethyl-3- hydroxypyridin-4-one). Indeed 1-(hydroxyalkyl)hydroxypyridinones which are known to be rapidly converted to 1-(carboxYalkyl)hydroxypyridinones are also marginally superior to Deferiprone. In contrast, 2-ethyl-1-(2'- hydroxyethyl)3-hydroxypyridin-4-one, which is not metabolized to the corresponding (carboxyalkyl)hydroxypyridinone, was found to be superior to Deferiprone and therefore deserves further consideration as an orally active iron chelator with potential for the treatment of iron overload associated with transfusion-dependent thalassemia.

Full text of DOI:10.1021/jm9707784

J . Med. Chem. 1998, 41, 3347-3359
3347
Syn th esis, P h ysicoch em ica l P r op er ties, a n d Eva lu a tion of
N-Su bstitu ted -2-a lk yl-3-h yd r oxy-4(1H)-p yr id in on es†
Bijaya L. Rai, Lotfollah S. Dekhordi, Hicham Khodr, Yi J in, Zudong Liu, and Robert C. Hider*
Department of Pharmacy, King’s College London, Manresa Road, London SW3 6LX, U.K., and Novartis,
Pharmaceuticals Research, CH-4002 Basel, Switzerland
Received November 14, 1997
The synthesis of a range of 3-hydroxy-4(1H)-pyridinones with potential for the chelation of
iron(III) is described. The pKa values of respective ligands and the stability constants of their
iron(III) complexes are presented. The distribution coefficient values of a range of 48
hydroxypyridinones and their corresponding iron(III) complexes between 1-octanol and MOPS
buffer (pH 7.4) are reported. The range of log Dcomplex values covers 7 orders of magnitude.
The results suggest the existence of a biphasic relationship between the distribution coefficient
values of the chelator and the corresponding iron(III) complexes. For ligands with a log Dligand
) -1, a linear relationship exists with a value of the slope 2.53, whereas with ligands with a
log Dligand < -1, a linear relationship exists with a slope of 0.49. The reduced slope for the
more hydrophilic molecules of the series offers some advantage for this type of hydroxypyri-
dinone as the distribution coefficients for such complexes do not change so rapidly with
increasing ligand hydrophilicity. The ability of selected 3-hydroxypyridinones to facilitate the
excretion of iron in bile was investigated in non-iron-overloaded, bile duct-cannulated rats and
in a [59Fe]ferritin-loaded rat model. Both systems compare the ability of chelators to remove
iron from the liver, the prime target organ in thalassemia. The N-(hydroxyalkyl)-3-hydroxy-
pyridin-4-ones are demonstrated to be orally active under the in vivo conditions adopted. Thus
both 1-(hydroxyalkyl)- and 1-(carboxyalkyl)pyridinones are able to remove iron from the liver.
Although 1-(carboxyalkyl)hydroxypyridinones are active, they do not demonstrate any clear
advantage over Deferiprone (1,2-dimethyl-3-hydroxypyridin-4-one). Indeed 1-(hydroxyalkyl)-
hydroxypyridinones which are known to be rapidly converted to 1-(carboxyalkyl)hydroxypyri-
dinones are also marginally superior to Deferiprone. In contrast, 2-ethyl-1-(2′-hydroxyethyl)-
3-hydroxypyridin-4-one, which is not metabolized to the corresponding (carboxyalkyl)-
hydroxypyridinone, was found to be superior to Deferiprone and therefore deserves further
consideration as an orally active iron chelator with potential for the treatment of iron overload
associated with transfusion-dependent thalassemia.
Iron possesses two oxidation states, iron(II) and iron-
(III), which are relatively stable under most biological
conditions. The redox potential between these two
states can be adjusted such that oxidation processes
centered on iron can be readily coupled to metabolic
reactions. Furthermore, iron has a high affinity for
oxygen atoms. These two properties are widely utilized
by iron-containing proteins. However, these same two
properties render iron potentially toxic, particularly
when the metal is nonspecifically bound to proteins and
membranes. Under aerobic conditions, such iron leads
to the formation of oxygen radicals, including the
hydroxyl radical.1 It is for this reason that iron levels
are carefully controlled, normal individuals absorbing
between 1 and 2 mg/24 h. This intake matches the daily
loss of iron. In the extracellular compartment, iron is
transferred between cells bound to the high-affinity
iron-binding protein transferrin, whereas ferritin acts
as a powerful iron buffer in intracellular compartments.
Thus under normal physiological conditions only a
vanishingly small amount of nonspecifically bound iron
exists. However there are conditions where the iron
status can change, for instance with idiopathic hemo-
chromatosis2 and transfusion-induced iron overload.3 In
such circumstances, either by virtue of hyperabsorption
of iron from the intestine or by regular transfusion, iron
overload results. Once transferrin and ferritin become
saturated with iron, free radical-induced tissue damage
results. Iron overload associated with transfusion-
dependent â-thalassemia major results in irreversible
damage to endocrine organs and lethal cardiac toxicity.
In humans such excess iron cannot be excreted via the
normal routes, namely, the urine and the bile, and
consequently chelation therapy is essential.3
Ideally, administration of an iron-selective chelator
should not interfere with iron-dependent enzymes and
respiratory proteins, and the resulting complex should
prevent the coordinated iron from redox cycling, thereby
preventing the production of toxic free radicals.4
Parenteral chelation therapy using desferrioxamine
(DFO) is now well-established and this natural sidero-
phore is capable of promoting negative iron balance and
† Abbreviations: Bn, benzyl; DCCI, dicyclohexylcarbodiimide; DFO,
desferrioxamine; log D, log distribution coefficient; EDTA, ethylene-
diaminetetraacetic acid; EI, electron impact; Et, ethyl; HBED, N,N-
bis(2-hydroxybenzyl)ethylenediamine N,N-diacetic acid; HPOs, hy-
droxypyridin-4-ones; IR, infrared; k, constant; mp, melting point; Me,
methyl; MOPS, 3-(N-morpholino)propanesulfonic acid; MS, mass
spectrum; m/z, mass/charge; NMR, nuclear magnetic resonance spec-
troscopy; Pr, propyl; r, correlation coefficient; TBTU, 2-(1H-benzotria-
zol-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate.
S0022-2623(97)00778-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/05/1998

3348 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Rai et al.
Sch em e 1. Synthesis of
1-Alkyl-3-hydroxy-4(1H)-pyridinones
reversing cardiac toxicity. Unfortunately, desferriox-
amine (DFO) lacks oral activity, and this has a dramatic
influence on patient compliance.5
There has been considerable interest in the design of
orally active iron chelators over the past 20 years, and
many high-affinity iron(III)-binding compounds have
been prepared.6-8 In principle it is possible to design
hexadentate, tridentate, and bidentate ligands which
are capable of providing a single iron(III) cation with
the desired octahedral field. Bidentate and tridentate
ligands have the advantage of low molecular weight,
which favors efficient absorption from the gastrointes-
tinal tract. In contrast, hexadentate ligands possess
high molecular weights, typically falling in the range
350-1000, and consequently are less efficiently ab-
sorbed.9 Hexadentate ligands which fall toward the
lower end of this range belong to the amino carboxylate
family, for instance HBED,10 and these all possess an
appreciable affinity for zinc(II).4,11 In similar fashion
the majority of iron(III)-binding tridentate ligands
include nitrogen as one of the coordinating ligands;
again this endows the molecule with appreciable affinity
for zinc(II).4,12 In contrast, it is possible to optimize the
Fe(III)/Zn(II) selectivity using bidentate ligands, for
instance, with substituted catechols13,14 and 3-hydroxy-
pyridin-4-ones (HPOs).11
cient on the efficacy of selected HPOs in a cannulated
bile duct rat model and a [59Fe]ferritin-loaded rat model.
Ch em istr y
The 1-substituted-2-alkyl-3-hydroxy-4(1H)-pyridino-
nes in this study were synthesized utilizing the meth-
odology described by Dobbin and co-workers17 (Scheme
1, Table 1). The commercially available 2-alkyl-3-
hydroxy-4(1H)-pyranones maltol and ethylmaltol18 were
benzylated in 90% aqueous methanol to give 3 and 4,
respectively. Reactions of these adducts with primary
amines were invariably performed under reflux in 50%
aqueous ethanol in the presence of a catalytic amount
of sodium hydroxide. 1-Amino-2-methyl-2-propanol is
not commercially available and consequently was pre-
pared via reduction of acetone cyanohydrin with Li-
AlH4.19 The reaction of benzylmaltol with serine in
aqueous ethanol gave the expected benzylated product
in the form of free base, but this compound being very
hydrophilic was not extractable in organic solvents at
pH 7. Thus the more hydrophobic amine 1,3-bis-
(benzyloxy)-2-aminopropane was prepared from 1,3-bis-
(benzyloxy)-2-propanol via a two-step reaction20,21 and
was subsequently reacted with benzylmaltol. The re-
sulting protected HPO was found to partition from
aqueous solution (pH 7) into dichloromethane and was
finally purified by column chromatography. Where
possible the benzylated pyridinones 5 were isolated in
a crystalline form as either the free bases or the
hydrochloride salts. Removal of the protecting benzyl
group was achieved by catalytic hydrogenolysis invari-
ably to yield the corresponding bidentate chelators 6 as
either the free bases or the hydrochloride salts (Table
2).
The introduction of 3-hydroxypyridin-4-ones as iron-
selective chelators has demonstrated that the design of
orally active iron chelators is possible, but at the present
time a compound with acceptable therapeutic potential
has not been identified.15 The long-term efficacy and
toxicity of Deferiprone (L1 or CP20), 1,2-dimethyl-3-
hydroxy-4(1H)-pyridinone, the first HPO to enter clini-
cal trials, have yet to be established. One of the
problems with such N-alkyl-3-hydroxypyridin-4-ones is
the ability of the free ligand and the resulting iron
complex to rapidly penetrate cell membranes and other
biological barriers.11 A second major difficulty is that
N-alkyl-3-hydroxypyridin-4-ones are rapidly metabo-
lized by glucuronidation of the 3-hydroxyl function (eq
1).16 Such conjugation abolishes the iron-chelating
(1)
properties of the molecule. In principle both the dif-
ficulties could be minimized by using more hydrophilic
hydroxypyridinones, the lower log D value leading to
reduced membrane permeability and reduced suscep-
tibility toward phase 2 conjugation reactions. However
absorption by the gastrointestinal tract and the liver is
also likely to be reduced with such compounds. Fur-
thermore extremely hydrophilic chelators may become
trapped intracellularly thereby inducing toxic side ef-
fects.9 To investigate these possibilities we have ex-
tended the range of hydroxypyridinones reported pre-
viously, thus enabling us to investigate HPO-iron
complexes with log D values spanning a range of 7
orders of magnitude. We report the relationship be-
tween the log D values of the free ligand and their iron-
(III) complexes and the influence of distribution coeffi-
The esters 65 and 66 were synthesized by treating
the benzylated carboxylic acid derivatives with the
corresponding alcohol in the presence of triphenylphos-
phine and diethyl azodicarboxylate.22 The amide ana-
logues 69 and 76-79 were obtained via reaction of the
succinimidyl-activated ester 7a with primary amines in
two steps or the in situ formation of the benzotriazolyl-
activated ester 7c (Scheme 2) in one step in the presence
of 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium
tetrafluoroborate (TBTU) as an activating agent23,24
(Table 2).
Deter m in a tion of Solu tion P r op er ties
3-Hydroxy-4(1H)-pyridinones possess two pKa values,
one corresponding to the 4-oxo function (pKa ≈ 3.7) and
the second corresponding to the 3-hydroxyl function (pKa

N-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3349
≈ 9.8).17 These bidentate ligands also form a number
of complexes with iron(III) so that aqueous solutions
equilibrate to give mixtures in which the speciation
depends on the metal ion, ligand, and hydrogen ion
concentrations. A simple model of this system is shown
in equation (eq 2). This model has previously been
Ta ble 1. 1-Substituted-2-alkyl-3-(benzyloxy)- and
-3-hydroxy-4(1H)-pyridinones (Bn ) benzyl)
Fe3+ + L- h FeL2+ (K1)
compd
R′
R
X
ref
8
9
CH2C(CH3)2OH
CH(CH2OBn)2
CH2CH2CH2OH
CH2CH2CH2CH2OH
CH2CH(OMe)2
CH2CH(OEt)2
CH2CH2COOH
CH2CH2CH2COOH
CH2CH2COOH
CH2CH2CH2COOH
CH2CH2SO3H
CH2CH2COOCH3
CH2CH2COOCH2CH3
CH2CH2CH2CONHMe
CH2CH2CONHMe
CH2CH2CONHPr
CH2CH2CON(Me)2
CH2CH2CON(Et)2
H
Me
Me
Et
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
Bn
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
FeL2+ + L- h FeL2 (K2)
(2)
+
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
Et
FeL2+ + L- h FeL3 (K3)
Me
Me
Et
shown to apply to 3-hydroxy-4(1H)-pyridinones.25 Both
the pKa values and stability constants were obtained
by global optimization of parameters corresponding to
titrations at several different ligand and iron(III) con-
centrations. The experimental conditions ensured that
the most associated species predominated.
Distribution coefficients (D values) of the 3-hydroxy-
4(1H)-pyridinones and the iron(III) complexes were
determined in an aqueous/octanol system using a filter
probe device.17,26 However, reproducible values for
extremely hydrophobic compounds (log D values > 2)
or hydrophilic compounds (log D < -2) were difficult to
obtain using the filter probe. The distribution coef-
ficients for such compounds were determined by the
“shake-flask” method changing the ratio of the two
phases (aqueous/1-octanol) to, for instance, 100:1 for
hydrophobic compounds (log D values > 2) and 1:100
for hydrophilic compounds (log D < -2), respectively.27
Et
Me
Me
Me
Me
Me
Me
Et
Et
Et
Et
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Et
Et
Et
Et
Et
Et
Et
Me
Me
Me
Me
Me
Me
Et
17
17
17, 46
17, 46
17, 45
45
17, 45
17
17
17
17
17
17, 47
17
17, 45
17, 45
45
17
17
17
17
Me
Et
CH2CHdCH2
CH2CH2CH3
CH2CH2CH2CH3
CH(CH3)2
CH2CH(CH3)2
CH2CH2CH2CH2CH3
CH2CH2CH2CH2CH2CH3
C6H5
H
Me
Et
Biologica l Exp er im en ts
CH2CHdCH2
CH2CH2CH3
CH(CH3)2
CH2CH2CH2CH3
CH2CH2OH
CH2CH2CH2OH
CH2CH2CH2CH2OH
CH2C(CH3)2OH
CH2CH(OH)CH2OH
CH(CH2OH)2
A selection of hydroxyalkyl-substituted hydroxypyri-
dinones were tested for their ability to remove iron from
non-iron-overloaded rats. The bile duct was cannulated,
and the quantity of iron excreted by the bile in the
presence and absence of chelators was determined and
expressed as µg of Fe/kg of animal weight.28 In a
parallel study chelators were also investigated in a [59-
Fe]ferritin-loaded rat model.29 Ferritin, which has a
molecular weight of approximately 450 000, is one of the
major iron storage proteins in the body. It consists of
24 equivalent subunits which provide a hollow center
capable of accommodating up to 4000 iron atoms in the
form of a regular ferric-hydroxide-phosphate lattice.
Since 59Fe-labeled ferritin is predominantly taken up
by hepatocytes in rats, it can be used to label the liver
iron pool selectively.30 Thus this [59Fe]ferritin-loaded
rat model can be used to compare the ability of chelators
to remove iron from the liver, the major storage organ
in iron-overloaded conditions.
17
17
CH2CH2OH
17
CH2CH2CH2OH
CH2CH2CH2CH2OH
CH2CH2OMe
CH(CH3)CH2OMe
(CH2)3OEt
CH2CH(OMe)2
CH2CH(OEt)2
CH2CH2OMe
Et
Et
Me
Me
Me
Me
Me
Et
17
17
17
17
17
17
17
(CH2)3OEt
Et
CH2CH2COOH
CH2CH2CH2COOH
CH2CH2COOH
CH2CH2CH2COOH
CH2CH2SO3H
CH2CH2COOMe
CH2CH2COOEt
CH2CH2NH2
(CH2)2CONHMe
(CH2)3CONHMe
(CH2)2CONHEt
(CH2)2CONHPr
(CH2)2CONHiPr
(CH2)2CONMeEt
(CH2)2CON(Me)2
CH2CH2CON(Et)2
(CH2)2CONHMe
(CH2)2CONHPr
(CH2)2CON(Me)2
(CH2)2CON(Et)2
Me
Me
Et
Et
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Et
Resu lts
17
17
Distr ibu tion Coefficien ts. The distribution coef-
ficients (log D) of the free ligands and their iron(III)
complexes between an aqueous phase buffered at pH
7.4 and octanol are presented in Table 3. In general,
as expected the increase in the length of the alkyl chain
on the heterocyclic nitrogen results in an increase in
the log D values of both the ligand and the iron(III)
complex. In most cases iron(III) complexes are more
hydrophilic than their corresponding free ligands. How-
ever, this trend did not hold for those compounds which
17
17
17
17
17
17
Et
Et
Et
H

3350 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Rai et al.
Ta ble 2. Synthesis of 1-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones Isolated as Free Bases or Hydrochloride Salts
compd
R
R′
mp, °C
% yield
formula
anal.
47
49
51
52
56
57
62
63
64
65
66
69
76
77
78
79
Me
Me
Et
CH2C(CH3)2OH
CH(CH2OH)2
CH2CH2CH2OH
CH2CH2CH2CH2OH
CH2CH(OMe)2
CH2CH(OEt)2
CH2CH2COOH
CH2CH2CH2COOH
CH2CH2SO3H
CH2CH2COOMe
CH2CH2COOEt
CH2CH2CH2CONHMe
CH2CH2CONHMe
CH2CH2CONHPr
CH2CH2CONMe2
CH2CH2CONEt2
204-205
225-226
147-148
166-167
213-214
116-117
167-168
190-191
80
71
96
98
78
80
87
64
85
85
90
95
96
95
86
85
C10H15NO3
CHN
C19H14NO4Cl
C10H16NO3Cl
C11H18NO3Cl
C10H15NO4
C12H19NO4
C10H13NO4
CHNCl
CHNCl
CHNCl
CHN
CHN
CHN
Et
Me
Me
Et
Et
C11H15NO4
C8H11NO5S
CHN
Me
Me
Me
Me
Et
Et
Et
Et
>300
CHNS
CHNCl
CHNCl
CHNCl
CHNCl
CHNCl
CHNCl
CHNCl
142-144
149-151
207-209
165-166
157-158
136-137
153-154
C10H14NO4Cl
C11H16NO4Cl
C11H17N2O3Cl
C11H17N2O3Cl
C13H21N2O3Cl
C12H19N2O3Cl
C14H23N2O3Cl
Sch em e 2. Synthesis of 1-[(N-Alkylcarbamoyl)alkyl]-2-alkyl-3-hydroxy-4(1H)-pyridinone
have log D values greater than 3 (i.e., 31, 33-36, 41-
43, and 59). In these examples the iron(III) complexes
were found to be more hydrophobic than their corre-
sponding free ligands. Surprisingly, the unsubstituted
pyridinones 26 and 37 possess higher values than the
N-methylpyridinones 27 and 38. This trend also holds
for the corresponding iron(III) complexes. The charged
molecules 60-63 and 67 possess the lowest values.
Among the neutral compounds, 35 and 49 possess the
highest and the lowest log D values, respectively, and
not surprisingly they form the most hydrophobic and
hydrophilic iron(III) complexes, respectively.
A plot of the log D values of the neutral ligands
against the log D values of the corresponding iron(III)
complexes demonstrates the existence of a biphasic
linear relationship between the log D values of ligand
and complex (Figure 1). A simple equation (eq 3) was
derived relating these two parameters for ligands with
log Dligand ) -1. The more hydrophilic ligands (log
Dligand < -1) possess a different relationship yielding
the corresponding equation (eq 4). The entire range of
log Dligand values of the ligands covers 3 orders of
magnitude, whereas for the corresponding iron(III)
complexes, this range spans 7 orders of magnitude.
Liga n d p Ka Va lu es. Some representative ligands,
the hydroxyalkyl derivatives (44 and 50) and the
carboxyalkyl derivatives (60, 62, and 63), have been
studied by simultaneous spectrophotometric/potentio-
metric titrations. The (hydroxyalkyl)pyridinones pos-
sess two pKa values, pKa1 and pKa2, whereas the
(carboxyalkyl)pyridinones possess three, pKa1, pKa2, and
pKa3. For the (hydroxyalkyl)pyridinones, pKa1 is as-
sociated with the protonation of the 4-oxo group and
pKa2, the protonation of the 3-hydroxy function (Scheme
3). The pKa2 value of the (carboxyalkyl)pyridinone
corresponds to the carboxylic function and can be
distinguished from the similar value of pKa1 by the
absence of a spectrophotometric change. The optimized
pKa values obtained with NONLIN1525 are shown in
Table 4. The values were found to be relatively constant
irrespective of the N-substituent.
Speciation plots relating to these values are presented
in Figure 2. The neutral form of the (hydroxyalkyl)-
pyridinones, as typified by 50, dominates the plot over
the pH range 4-9, whereas the neutral form of the
(carboxyalkyl)pyridinones, as typified by 62, only makes
a minor contribution, the monoanion dominating over
the pH range 5-9.
Sta bility Con sta n ts of Ir on (III) Com p lexes. For
pyridinones 50 and 62, the K1, K2, and K3 values were
determined and compared with the corresponding â3
values, the latter resulting from competitive titration
log Dligand ) 2.53 log Dcomplex - 0.80 (r ) 0.99) (3)
log Dligand ) 0.49 log Dcomplex - 2.45 (r ) 0.99) (4)

N-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3351
Ta ble 3. log Distribution Coefficient (log D) Values of Ligands
and Fe Complexes (n ) no. of replications)
log D
ligand
log D
Fe complex
compd
R′
R
n
n
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
65
66
67
68
69
70
71
72
73
74
75
H
Me
Et
Me -0.50 ( 0.02
Me -0.77 ( 0.02
Me -0.31 ( 0.01
Me 0.03 ( 0.01
5
-2.38 ( 0.02a
-2.60 ( 0.05a
-1.59 ( 0.08
-0.70 ( 0.02
4
6
5
5
4
6
6
6
6
5
6
4
4
4
6
6
CH2CHdCH2
Pr
Me 0.18 ( 0.01 10 -0.08 ( 0.01
(CH2)3CH3
CH(CH3)2
CH2CH(CH3)2
(CH2)4CH3
(CH2)5CH3
C6H5
H
Me
Et
Me 0.70 ( 0.01 10 1.36 ( 0.03
Me 0.05 ( 0.01
Me 0.65 ( 0.02
5
5
-0.54 ( 0.03
1.26 ( 0.02
Me 1.24 ( 0.01 10 2.74 ( 0.02a
Me 1.89 ( 0.02
Me 1.01 ( 0.01
6
6
8
5
3.40 ( 0.02a
2.16 ( 0.02
-1.16 ( 0.01
-1.50 ( 0.02
Et
0.05 ( 0.01
Et -0.21 ( 0.01
Et
Et
Et
Et
Et
0.23 ( 0.01 10 -0.62 ( 0.02 10
CH2CHdCH2
Pr
0.47 ( 0.01
0.70 ( 0.01
0.73 ( 0.01
5
6
5
0.36 ( 0.01
0.96 ( 0.01
1.01 ( 0.01
8
8
5
4
4
4
4
4
4
4
4
5
5
5
5
4
CH(CH3)2
(CH2)3CH3
CH2CH2OH
CH2CH2CH2OH
(CH2)4OH
CH2C(CH3)2OH
CH(CH2OH)2
CH2CH2OH
CH2CH2CH2OH
(CH2)4OH
CH2CH2OMe
CH(CH3)CH2OMe Me -0.26 ( 0.01
(CH2)3OEt
CH2CH(OMe)2
CH2CH(OEt)2
CH2CH2OMe
(CH2)3OEt
1.22 ( 0.01 10 2.36 ( 0.02a
Me -1.10 ( 0.01
Me -0.89 ( 0.01
Me -0.75 ( 0.01
Me -0.77 ( 0.02
Me -1.70 ( 0.04
Et -0.66 ( 0.02
Et -0.32 ( 0.01
Et -0.28 ( 0.01
Me -0.41 ( 0.01
5
5
7
5
4
5
5
5
5
5
5
5
6
5
5
4
4
5
5
8
5
5
5
8
5
5
5
5
5
5
-3.00 ( 0.08a
-2.92 ( 0.07a
-2.75 ( 0.05a
-2.33 ( 0.04a
-3.30 ( 0.04a
-2.75 ( 0.07a
-2.10 ( 0.01a
-1.68 ( 0.02
-1.85 ( 0.05
-1.40 ( 0.03
-0.52 ( 0.04
-2.52 ( 0.06a
F igu r e 1. Relationship between the log D value of Fe complex
(log Dcomplex) and log D value of hydroxypyridinone (log Dligand).
Distribution coefficients were determined using a MOPS buffer
(50 mM, pH 7.4)/octanol system.
Sch em e 3
Me 0.12 ( 0.01
Me -0.46 ( 0.03
Me 0.21 ( 0.01
-0.10 ( 0.02 10
-1.16 ( 0.03
0.99 ( 0.01
Et
Et
0.04 ( 0.01
0.72 ( 0.01
5
6
4
5
5
4
4
6
4
4
4
4
5
5
4
4
5
CH2CH2CO2H
Me -2.92 ( 0.07a
-4.00 ( 0.15a
-3.70 ( 0.06a
-3.22 ( 0.03a
-3.10 ( 0.03a
-2.70 ( 0.04a
-1.16 ( 0.01
-3.22 ( 0.03
-3.05 ( 0.03a
-3.00 ( 0.11a
-2.70 ( 0.06a
-1.52 ( 0.01
-1.89 ( 0.07
-2.52 ( 0.04a
-3.00 ( 0.11a
-1.70 ( 0.02
CH2CH2CH2CO2H Me -2.75 ( 0.07a
CH2CH2CO2H
Et -1.60 ( 0.01
CH2CH2CH2CO2H Et -1.38 ( 0.01
CH2CH2CO2Me
CH2CH2CO2Et
CH2CH2NH2
(CH2)2NHCOMe
(CH2)3CONHMe
(CH2)2CONHEt
(CH2)2CONHPr
(CH2)2CONHiPr
Me -0.77 ( 0.02
Me -0.24 ( 0.01
Me -1.55 ( 0.01
Me -1.15 ( 0.01
Me -1.08 ( 0.01
Me -0.82 ( 0.02
Me -0.40 ( 0.01
Me -0.47 ( 0.01
(CH2)2CONMeEt Me -0.66 ( 0.02
(CH2)2CON(Me)2 Me -1.00 ( 0.01
CH2CH2CONEt2
Me -0.36 ( 0.01
a
Determination by hand-shake method. Determinations made
for 1-octanol/water (pH 7.4) system at 25 °C.
by a proportionate decrease in the value of log K3 (Table
4). Thus the log â3 value remains remarkably constant.
Indeed log â3 values for 26, 28, 30,17 39,31 and 50 (Khodr
and Hider, unpublished) all fall within the range 36.8-
37.7. For this reason we did not determine the â3 values
of the entire set of the ligands. The speciation plot for
pyridinone 62 in the presence of iron(III) (Figure 4)
indicates that the FeL3 complex dominates over the pH
range 6.5-10. An energy-minimized structure of the
iron(III) complex of 50 using Hyperchem Software
(Molecular Modeling, release II for SGI) is presented
in Figure 5.
Ability of 3-Hyd r oxyp yr id in on es To F a cilita te
th e Excr etion of Ir on in Bile. The 3-hydroxypyridin-
4-one Deferiprone (27) is orally active in this rat test
system, although as demonstrated in humans less
effective than desferrioxamine when administered pa-
with EDTA. The optimized values are presented in
Table 4, where it can be seen that there is excellent
agreement between the summation of log K1, K2, and
K3 values with the directly determined log â3 value for
compounds 50 and 62. The pH titration curve for
pyridinone 50 in the presence of iron(III) is displayed
in Figure 3: at low pH values the λmax of the iron(III)
complex (545 nm) corresponds to the FeIIIL12+ complex;
over the pH range 3.5-5.5, the FeIIIL2+ complex (λmax
,
510 nm) dominates; and above pH 7.0 the neutral FeIIIL3
complex (λmax, 460 nm) dominates.
The â3 values are not markedly influenced by the net
charge of the FeL3 complex, which is neutral for 50, and
3- for 62. Although the value of log K1 is predictably
larger for the charged pyridinone 62, this is balanced

3352 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Rai et al.
Ta ble 4. Spectrophotometric and Potentiometric Determined pKa Values for Ligands and Affinity Constants for Fe(III) Complexes
affinity constants for Fe(III)
a
pKa1
pKa2/pKa3
log â3
potentio- spectrophoto-
potentio- spectrophoto-
summation
K1, K2, K3
EDTA competition
value
ligand
metric
metric
pKa2
metric
metric
log K1 log K2 log K3
27
44
50
60
62
63
3.56
3.43
3.34
3.19
3.22
3.42
9.64
9.72
9.75
9.51
9.91
9.99
14.9
15.7
16.6
12.2
11.6
11.4
9.8
9.4
7.9
37.2
36.7
36.9
3.44
3.31
3.10
3.12
3.21
9.65
9.77
9.49
9.89
9.89
36.8
37.0
3.79
3.91
4.27
a
pKa2 for 27, 44, and 50; pKa3 for 60, 62, and 63.
F igu r e 3. pH dependence of the spectrum of ligand 50 in the
presence of Fe(III). [Fe(III)] ) 1.889 × 10-4 M, [50] ) 1.901 ×
10-3 M. The isosbestic point at 510 nm corresponds to the
conversion of the 2:1 complex to the 3:1 complex.
F igu r e 2. Speciation plots of hydroxypyridinones 50 and 62.
The following parameters were incorporated into the model
for the system: for 50, pKa1 ) 3.34, pKa2 ) 9.75; 62, pKa1
3.22, pKa2 ) 3.91, pKa3 ) 9.91.
)
F igu r e 4. Speciation plot of 62 (5 × 10-5 M) in the presence
of iron(III) (1 × 10-5 M). The following parameters were
incorporated into the model for the system: log â1ML ) 16.61,
rentally (Table 5). Similar profiles were observed for
the pyridinones 45 and 51, both of which had similar
efficacies to that of 27 (Figure 6). The extremely
hydrophilic pyridinone 48 possessed an extremely low
activity. In contrast, the 1-(2′-hydroxyethyl)pyridinones
(44 and 50) induced more prolonged periods of iron
excretion, and compound 50 was found to be markedly
more active than both Deferiprone (27) and desferriox-
amine in this test system (Table 5, Figure 6).
Ability of 3-Hyd r oxyp yr id in on es To F a cilita te
th e Excr etion of 59F e fr om F er r itin -Loa d ed Ra ts.
Rats were infused with 59Fe-rat ferritin (10 µg), and
after a 1-h period the pyridinones were administered
log â2ML2 ) 28.05, log â3ML3 ) 36.93, pKa1 ) 3.22, pKa2
)
3.91, pKa3 ) 9.91, pK(FeOH) ) -3.03, pK((FeOH)2) ) -6.30,
and pK((FeOH)3) ) -2.9.
by oral gavage. During this period the [59Fe]ferritin is
efficiently cleared from the systemic circulation by the
hepatocytes where it is catabolized and incorporated
into endogenous ferritin via the cytoplasm labile iron
pool. It is likely that the orally administered chelators
scavenge iron from this pool. Again the hydroxyl-
substituted-N-alkylpyridinones were demonstrated to be

N-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3353
F igu r e 5. Energy-minimized structure of the iron(III) com-
plex of 50: white, hydrogen; gray, carbon; black, oxygen and
nitrogen.
Ta ble 5. Influence of Chelators on Iron Excretion in the Bile
Duct-Cannulated Ratsa
dose
iron excretion, µg kg-1
mg iron-binding equivalent
compd kg-1
27 100
DFO 100
44
45
48
50
51
in mequiv kg-1
oral
subcutaneous
0.24
0.15
0.24
0.24
0.24
0.24
0.24
351 ( 38
292 ( 35
527 ( 131
353 ( 61
383 ( 33
142 ( 50
698 ( 58
634 ( 266
122
132
143
132
142
272 ( 100
414 ( 44
104 ( 75
819 ( 22
265 ( 153
F igu r e 6. Hydroxypyridinone-induced iron excretion by the
rat bile duct. The chelators 27, 44, 45, 48, 50, and 51 were
given orally. Chelator dose was 450 µmol/kg (n ) 3).
a
The results are the mean ( SD iron excreted in bile and urine
from 3-4 animals.
orally active in this test system, although only 45 was
found to be significantly better than Deferiprone (27)
(Figure 7). The carboxylate-containing hydroxypyridi-
none 60 was found to be equieffective as 27, but none
of these relatively hydrophilic chelators were as effective
as the more lipophilic pyridinone 39.
Discu ssion
Fifty-four 3-hydroxypyridin-4-ones have been synthe-
sized covering a wide range of distribution coefficients.
All bind iron(III) tightly to form neutral 3:1 complexes
over the pH range 5-9.5. The relationship between the
noncharged ligand D values (Dligand) and the corre-
sponding D values of the iron complexes (Dcomplex) for
ligands with a log Dligand ) -1 is linear with a r value
of 0.99 (Figure 1, eq 3). The more hydrophilic ligands
(log Dligand < -1) gave a different relationship (Figure
1, eq 4). In principle log Dligand and log Dcomplex will be
an additive function of the varying contribution of the
R and R′ groups. The electrically neutral iron(III)
complex has three bidentate chelating molecules octa-
hedrally coordinated to iron(III)17,32 (Figure 5) so that
it will have approximately 3 times the surface area of
the ligand alone, the contribution of the iron atom being
negligible33,34 as it is completely coordinated by the six
coordinating ligands. Consequently it might be ex-
F igu r e 7. Biliary 59Fe excretion induced by hydroxypyridi-
nones from ferritin-loaded rats. The chelators 27, 39, 44, 45,
50, and 60 were given orally. Chelator dose was 450 µmol/kg
(n ) 5-9).
pected that a relationship similar to eq 3 would result34
where k is a constant representing the change in
hydrophobicity in converting the six hydrophilic ligating
groups of the three hydroxypyridinones into the coor-
dinating sphere of the iron(III) complex. The effective
masking (or partial masking) of the oxygen atoms on
the 3- and 4-positions of the hydroxypyridinone ring on
complexation leads to an increase in the lipophilicity

3354 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Rai et al.
contribution of the complex, and this difference should
be constant irrespective of the nature of the substitution
R and R′. This constant difference is represented by
the negative constant -0.80 in eq 3. The slope of 2.53
is appreciably smaller than the theoretical slope of 3.00,
and this probably represents the relatively efficient
packing of the three ligands in the iron(III) complex;
i.e., the surface area is <3 times that of the free ligand.
Significantly Edward et al.34 observed a similar effect
in a parallel study centered on the chelation of iron-
(III) by pyridoxal hydrazones, which are trident ligands.
The observed slope was 1.94 as compared to the
theoretical value of 2.0 and k ) -0.83 (r ) 0.92). In
the same study these authors also included a limited
range of hydroxypyridinones and obtained slightly dif-
ferent values of slope and k to those reported in this
study. However some of the D values adopted in the
earlier study have since been modified. It is not always
convenient to measure the log Dcomplex directly as the
filter probe system can easily become contaminated with
iron (Dobbin et al.17 and this work). The relationships
depicted in Figure 1 render it possible to make a reliable
estimation of log Dcomplex for a directly determined log
that the dominate species over the pH range 5.0-9.0
(Figure 2) is the monoanionic form (Scheme 3), and
although the corresponding 3:1 iron(III) complex will
possess a net charge close to 3- at pH 7.4, this complex
dominates the speciation plot above pH 7.0 (Figure 4).
As indicated in Figure 1, the ratio of log Dcomplex/log
Dligand for HPOs possessing a log Dligand value > -0.1 is
much lower than for neutral ligands, and thus the rate
of efflux of 3:1 iron complexes is predicted not to
markedly decrease with decreasing log Dligand value.
Thus this preliminary analysis and search for nontoxic
hydroxypyridinones favors compounds with a log Dligand
value falling in the range -1 to -2. Hydroxypyridino-
nes possessing a carboxylate function on the ring
nitrogen substituent, for instance 60-63, fall into this
range (Table 3).
Two animal models were selected to monitor the
ability of iron chelators to mobilize iron under in vivo
conditions: the bile duct cannulated rat and the [59Fe]-
ferritin-labeled rat. Both have a proven record for
successfully identifying potentially useful iron chelators.
Both models are designed to maintain the removal of
iron from the hepatocyte, the liver being the major
target organ in transfused â-thalassemia patients.3 With
the [59Fe]ferritin model, the labeled ferritin is rapidly
removed from the circulation by hepatocytes, which then
degrade the exogenous ferritin to iron which subse-
quently enters the so-called “labile iron pool”. It is this
pool of iron which is the specific target of orally active
iron chelators. Unfortunately when administered orally,
the carboxylate-containing hydroxypyridinone was found
not to be highly efficient at extracting iron (Figure 7).
Although Molenda and co-workers37 reported oral activ-
ity for 6-carboxyl-3-hydroxy-1-methylpyridin-4-one (80),
Dligand
.
With the more hydrophilic hydroxypyridinones (log
Dligand < -1) (44, 49, 60-63, 67-69), a deviation from
eq 3 results with the log Dcomplex values leveling out
(Figure 1). The reason for this deviation is not clear
but would indicate that the solvation in the aqueous
phase is similar for both the free ligand and the complex
for the more hydrophilic molecules and that therefore
the difference between the log Dcomplex and log Dligand is
less marked with this group of molecules.
When extremely hydrophilic pyridinones enter a cell,
albeit slowly, and scavenge iron(III), the complex is
predicted to efflux even more slowly due to its higher
molecular weight and more hydrophilic nature. In
principle this could lead to toxic side effects resulting
from trapped intracellular iron. Furthermore as the
systemic concentration of the hydroxypyridinone de-
creases, leading to a net pyridinone efflux, the 3:1
complex would be predicted to partially dissociate
forming the charged 2:1 complex which would possess
an even lower log D, by virtue of its net charge (eq 5).
the compound was found to be more effective when given
intravenously (80% of activity of Deferiprone, 27) than
when administered orally (38% activity of Deferiprone,
27). In the present study a similar finding was made
with 60 which is equiefficient with 27 (Figure 7). The
difference between 60 and 80 will largely result from
the higher Dligand value of the former. Thus although
monoanionic pyridinones are likely to be less toxic than
Deferiprone (27), they are unlikely to be more efficient
at scavenging iron, and the general opinion is that a
ligand possessing a higher efficacy than Deferiprone (L1
or CP20) is required.15 Similar arguments could be
made for monobasic pyridinones, for instance 67, al-
though there is a greater possibility of such molecules
inhibiting critical enzymes such as tyrosine hydroxy-
lase.38 Some (hydroxyalkyl)hydroxypyridinones are
rapidly and efficiently metabolized to the corresponding
monoanionic derivatives (eq 6)35 and thus can be
considered as prodrugs of the latter class of pyridinones.
Such (hydroxyalkyl)pyridinones are also capable of
facilitating the excretion of iron when administered
orally (Figures 6 and 7). However the finding that 45
and 51 have similar activities to that of 27 (Deferiprone)
H
+ + [Fe(HPO)3]0 h [Fe(HPO)2]+ + HPO0 (5)
The presence of such partially coordinated iron com-
plexes may generate hydroxyl radicals.1 Thus although
(hydroxyalkyl)pyridinones possess the advantage of slow
phase II metabolism, for instance glucuronidation (eq
1), and thereby retain iron-scavenging properties,35 they
may lead to the retention of iron by cells. However the
flattening of the relationship between log Dcomplex and
log Dligand with increasing hydrophilicity of the ligand
will tend to minimize the effect.
One method of avoiding the above difficulty is to
design hydroxypyridinones which are so hydrophilic
that they are effectively unable to enter cells via passive
diffusion. This is probably one of the reasons that
desferrioxamine is relatively nontoxic.36 In principle
anionic hydroxypyridinones, by virtue of their hydro-
philicity, could function in a similar manner. Speciation
plots of carboxylate-containing hydroxypyridinones show

N-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3355
sulfate, filtered, and concentrated in vacuo to yield a colorless
oil40 (7.24 g, 51.7%).
2-Am in o-1,3-bis(ben zyloxy)p r op a n e. 1,3-Bis(benzyloxy)-
2-propanol (2.3 g, 8.4 mmol), triphenylphosphine (2.43 g, 9
mmol), and phthalimide (1.37 g, 9 mmol) were dissolved in
dry tetrahydrofuran (40 mL) at room temperature under
nitrogen atmosphere. Diethyl azodicarboxylate (1.62 g, 9
mmol) was then added. The resulting mixture was stirred at
room temperature for 24 h. Solvent was removed at reduced
pressure, and the product was purified by column chromatog-
raphy (silica, eluant 20% EtOAc/petroleum ether, bp 40-60
°C) to give white crystals of 1,3-bis(benzyloxy)-2-phthalimi-
dopropane (2.60 g, 70%): mp 82-83 °C. Anal. (C25H23NO4)
C, H, N.
To a solution of 1,3-bis(benzyloxy)-2-phthalimidopropane
(4.60 g, 12 mmol) in absolute ethanol (60 mL) was added
hydrazine hydrate (55%, 14.7 mL, 13 mmol), and the mixture
refluxed for 3 h. The reaction mixture was cooled at 0 °C and
acidified with concentrated hydrochloric acid (pH < 2). The
resulting suspension was filtered and the collected phthalhy-
drazide washed with water (2 × 60 mL). The filtrate was
evaporated to dryness under reduced pressure. The resulting
residue was dissolved in water (100 mL), diethyl ether (100
mL) was added, and the aqueous layer was made alkaline (pH
) 12) at 0 °C by addition of saturated potassium hydroxide.
The resulting mixture was stirred at 0 °C for 30 min. The
organic layer was separated, and the aqueous layer was
extracted with diethyl ether (3 × 100 mL). The combined ether
extracts were dried (Na2SO4), filtered, and concentrated in
vacuo to dryness to yield 2-amino-1,3-bis(benzyloxy)propane
as a yellow oil (2.80 g, 86%).
1-(2′-Met h yl-2′-h yd r oxyp r op yl)-2-m et h yl-3-h yd r oxy-
4(1H)-p yr id in on e, 47. Sodium hydroxide solution (10 M) was
added to a mixture of benzylmaltol (3) (2.0 g, 9 mmol) in
ethanol (25 mL), water (25 mL), and 1-amino-2-methyl-2-
propanol (2.4 g, 27 mmol) until the mixture reached pH 13.
The resulting suspension was heated under reflux for 12 h.
After removal of the solvent by rotary evaporation, water (50
mL) was added, and the mixture was adjusted to pH 1 with
concentrated hydrochloric acid and washed with diethyl ether
(2 × 25 mL). The resultant aqueous solution was then
adjusted to pH 7 with 10 M sodium hydroxide solution and
extracted with dichloromethane (3 × 50 mL). The organic
extracts were combined, dried over anhydrous sodium sulfate,
and filtered. The solvent was removed by rotary evaporation
to give a yellow solid. Recrystallization from ethanol/diethyl
ether afforded 1-(2′-methyl-2′-hydroxypropyl)-2-methyl-3-(ben-
zyloxy)-4(1H)-pyridinone (8) as a white powder (1.15 g, 45%):
mp 153-154 °C. Anal. (C17H21NO3) C, H, N.
indicates that such a metabolic change does not offer
an advantage over Deferiprone (27) in their ability to
scavenge and remove iron under in vivo conditions.
In contrast to 45 and 51, the (hydroxyalkyl)pyridinone
50 is not converted to a carboxylate derivative nor
rapidly glucuronidated.35 These properties result in the
molecule inducing a more prolonged period of iron
excretion (Figure 6). Thus whereas iron excretion
resulting from the oral administration of 27 is virtually
complete 12 h after oral administration, with 50 excre-
tion continues over the entire 24-h period. Indeed this
compound was found to be the most efficient HPO tested
(Table 5), being more effective than both Deferiprone
(27) and desferrioxamine. However in the [59Fe]ferritin
model the activity of 50 was found not to be significantly
different from that of Deferiprone (27) (Figure 7).
Thus, although the (hydroxyalkyl)pyridinone 50 may
offer some advantages over the simple N-alkylhydroxy-
pyridinones, the influence will probably be marginal.
For treatment of thalassemia, delivery of such molecules
to the liver is required, and ester prodrugs of (hydroxy-
alkyl)pyridinones are currently under investigation.39
Exp er im en ta l Section
Gen er a l P r oced u r e. Melting points are uncorrected. IR
spectra were recorded on a Perkin-Elmer 298 IR spectrometer.
Proton NMR spectra were determined with a Perkin-Elmer
R32 (90 MHz) NMR spectrometer. Mass spectra (EI) or
positive fast atom bombardment (FAB) were recorded on J EOL
AX505W and KRATOS MS890 mass spectrometers. Elemen-
tal analyses were performed by Butterworth Laboratories
Limited, Teddington, Middlesex, or Micro Analytical Labora-
tories, Department of Chemistry, The University of Manches-
ter, Manchester. Full spectroscopic and analytical data are
available as Supporting Information.
2-Meth yl-3-(ben zyloxy)-4-(1H)-p yr a n on e, 3. To a stirred
solution of maltol (1) (178.0 g, 1.41 mol) in methanol (1800
mL) was added sodium hydroxide (60 g, 1.5 mol) in water (200
mL) followed by benzyl chloride (209 g, 1.65 mol), and the
mixture was heated to reflux for 12 h. After removal of solvent
by rotary evaporation the resultant orange oil was taken up
in dichloromethane (800 mL) and washed with 5% (w/v)
aqueous sodium hydroxide (5 × 300 mL) and water (2 × 500
mL). The organic fraction was dried over anhydrous sodium
sulfate and filtered. The solvent was removed by rotary
evaporation to give a colorless oil which was triturated with
diethyl ether and cooled at 0 °C for 1 h to give solid product.
Recrystallization from diethyl ether afforded 3 as colorless
needles (243 g, 83%): mp 56-57 °C (lit. value17 51-52 °C).
An analogous procedure using ethylmaltol (2) gave 2-ethyl-
3-(benzyloxy)-4-(1H)-pyranone (4) (80%): mp 33-34 °C (lit.
value17 33-34 °C).
To a solution of 1-(2′-methyl-2′-hydroxypropyl)-2-methyl-3-
(benzyloxy)-4(1H)-pyridinone (8) (2.7 g, 10 mmol) in ethanol
(90 mL) and water (10 mL) was added 5% palladium on
charcoal catalyst (0.25 g). The mixture was stirred at room
temperature under a constant stream of hydrogen for 24 h.
After removal of used catalyst by filtration, the solution was
boiled with activated charcoal and filtered. The solvent was
removed by rotary evaporation to give a cream-colored solid.
Recrystallization from ethanol/diethyl ether yielded 1-(2′-
methyl-2′-hydroxypropyl)-2-methyl-3-hydroxy-4(1H)-pyridino-
1
ne (47) as a white powder (1.50 g, 80%): mp 204-205 °C; H
NMR (DMSO-d6) δ 1.1 (6H, s, 2 CH3), 3.32 (3H, d, 2-CH3), 3.86
(2H, s, NCH2), 5.8-6.1 (2H, br, 2 OH), 6.14 (1H, d, 5-H), 7.55
(1H, d, 6-H). Anal. (C10H15NO3) C, H, N.
Analogous syntheses reacting either 3 or 4 with 2-amino-
1,3-bis(benzyloxy)propane, 3-aminopropan-1-ol, 4-aminobutan-
1-ol, aminoacetaldehyde dimethyl acetal, and aminoacet-
aldehyde diethyl acetal gave 1-(1′-(benzyloxymethyl)-2-(ben-
zyloxy)ethyl)-2-methyl-3-(benzyloxy)-4(1H)-pyridinone (9), 1-(3′-
hydroxypropyl)-2-ethyl-3-(benzyloxy)-4(1H)-pyridinone (10),
1-(4′-hydroxybutyl)-2-ethyl-3-(benzyloxy)-4(1H)-pyridinone (11),
1-(2′,2′-dimethoxyethyl)-2-methyl-3-(benzyloxy)-4(1H)-pyridi-
none (12), and 1-(2′,2′-diethoxyethyl)-2-methyl-3-(benzyloxy)-
4(1H)-pyridinone (13), respectively, as free base or hydrochlo-
ride salts on acidic workup.
1-Am in o-2-m eth yl-2-p r op a n ol. Lithium aluminum hy-
dride (12.5 g, 0.33 mol) was suspended in dry ether (200 mL)
and cooled to 0 °C under a nitrogen atmosphere. Acetone
cyanohydrin (13.35 g, 0.157 mol) in dry ether (50 mL) was
added dropwise during 1 h, and the reaction mixture stirred
overnight at room temperature. With continued cooling and
vigorous stirring, water (12 mL), sodium hydroxide (20%, 9
mL), and water (40 mL) were added. The ether solution was
decanted and the residue washed with ether (2 × 50 mL). The
ether portions were all combined, dried over anhydrous sodium

3356 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Rai et al.
Analogous hydrogenolysis procedure using 9, 10, and 11-
13 varying the reaction period from 6 to 24 h afforded 49, 51-
52, and 56-57, respectively, as shown in Table 2.
white precipitate was removed by filtration, dimethylforma-
mide was removed by rotary evaporation under high vacuum
to give an oil which was extracted into dichloromethane (200
mL) and washed with water (3 × 100 mL). The organic layer
was dried over anhydrous sodium sulfate and filtered. The
solvent was removed by rotary evaporation to give a light-
yellow oil which was dissolved in ethyl acetate and cooled at
0 °C for 1 h to give a white solid. Recrystallization from ethyl
acetate yielded white crystals (7.15 g, 95%): mp 99-101 °C;
1H NMR (DMSO-d6) δ 1.75-2.08 (2H, m, CH2CH2CH2COO),
2.16 (3H, s, 2-CH3), 2.73 (2H, t, CH2CO), 2.83 (4H, s, succin-
imideCH2), 3.95 (2H, t, NCH2), 5.06 (2H, s, CH2Ph), 6.23 (1H,
d, 5-H), 7.2-7.5 (5H, m, ArH), 7.63 (1H, d, 6-H).
1-(2′-Ca r b oxyet h yl)-2-et h yl-3-(b en zyloxy)-4(1H )-p yr i-
d in on e, 14. To a solution of 4 (23 g, 0.1 mol) in ethanol (300
mL) and water (300 mL) was added â-alanine (11.0 g, 0.12
mol) followed by 10 M sodium hydroxide solution until pH 13
was attained. After heating under reflux for 18 h, the
resulting solution was reduced in volume (∼100 mL) by rotary
evaporation and water was added. The aqueous solution was
adjusted to pH 9 by adding hydrochloric acid and washed with
diethyl ether (2 × 100 mL) prior to adjustment to pH 4 (by
adding hydrochloric acid) and extraction with dichloromethane
(5 × 100 mL). The organic fractions were combined, dried over
anhydrous sodium sulfate, and filtered. The solvent was
removed by rotary evaporation to a yellow oil which was
triturated with acetone and then solidified by cooling at -15
°C. Recrystallization from hot water/acetone yielded light-
Analogous procedure using 14 gave 1-[2′-((succinimidyloxy)-
carbonyl)ethyl]-2-ethyl-3-(benzyloxy)-4(1H)-pyridinone (7b) as
white crystals (66%): mp 165-166 °C.
1-[3′-(Meth ylcar bam oyl)pr opyl]-2-m eth yl-3-(ben zyloxy)-
4(1H)-p yr id in on e Hyd r och lor id e, 21. Methylamine (2 g,
26 mmol, 40% in water) was added dropwise to a solution of
7a (8.76 g, 22 mmol) in dichloromethane (200 mL). After the
reaction mixture stirred overnight at room temperature, the
solution was filtered, washed with water (3 × 100 mL), dried
over anhydrous sodium sulfate, and filtered. The solvent was
removed by rotary evaporation to give a colorless oil which
was dissolved in ethanol, and hydrochloric acid was added.
The product was reconcentrated to yield a white solid. Re-
crystallization from ethanol/diethyl ether yielded white crys-
tals (7.2 g, 93%): mp 161-162 °C. Anal. (C18H23N2O3Cl) C,
H, N, Cl.
yellow crystals (21 g, 70%): mp 156-157 °C. Anal. (C17H19
NO4) C, H, N.
-
Analogous procedure reacting 4 with 4-aminobutyric acid
and 3 with â-alanine or 4-aminobutyric acid gave 15, 16, and
17, respectively.
Analogous procedure using 3 and taurine at pH > 12 gave
1-(2′-sulfoxyethyl)-3-methyl-3-(benzyloxy)-4(1H)-pyridinone (18)
in zwitterionic form (80%): mp > 300 °C. Anal. (C15H17NO5S)
C, H, N, S.
1-(2′-Ca r boxyeth yl)-2-eth yl-3-h yd r oxy-4(1H)-p yr id in o-
n e, 62. To a solution of 14 (1.5 g, 5 mmol) in tetrahydrofuran
(30 mL) and water (20 mL) was added 2 drops of 1 M
hydrochloric acid followed by 5% palladium on charcoal
catalyst (0.2 g). The mixture was stirred at room temperature
for 6 h under a constant stream of hydrogen. After removal
of used catalyst by filtration, the solution was concentrated
to dryness by rotary evaporation to give a cream-colored solid.
Recrystallization from water yielded cream-colored crystals
(0.91 g, 87%): mp 167-168 °C; 1H NMR (DMSO-d6) δ 1.12
(3H, t, CH2CH3), 2.68 (2H, t, CH2CO2-), 2.7 (2H, q, CH2CH3),
4.16 (2H, q, NCH2), 6.15 (1H, d, 5-H), 6.8 (2H, br, s, 2OH), 7.6
(1H, d, 6-H). Anal. (C10H13NO4) C, H, N.
Analogous procedure reacting 1-[2′-((succinimidyloxy)car-
bonyl)ethyl]-2-ethyl-3-(benzyloxy)-4(1H)-pyridinone (7b) with
methylamine and propylamine gave 22 and 23 in yields of 40%
and 55%, respectively.
Rou te 2. Syn th esis of Am id es via Ben zotr ia zolyl
Active Ester . 1-[2′-(Dim eth ylca r ba m oyl)eth yl]-3-(ben -
zyloxy)-2-eth yl-4(1H)-p yr id in on e Hyd r och lor id e, 24. N-
Methylmorpholine (2.80 g, 26 mmol) was added to a solution
of 1-(2′-carboxyethyl)-2-ethyl-3-(benzyloxy)-4(1H)-pyridinone
(14) (7.53 g, 25 mmol) in dimethylformamide (100 mL) followed
by 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium tet-
rafluoroborate (TBTU; 8 g, 25 mmol) under an atmosphere of
nitrogen. After the reaction mixture stirred for 20 min at room
temperature, dimethylamine (1.3 g, 25 mmol) was added
dropwise and the mixture was stirred for 4 h. Dimethylfor-
mamide was removed, and the product was taken into dichlo-
romethane (100 mL). The organic fraction was washed with
5% hydrochloric acid (2 × 50 mL), 5% aqueous sodium
hydroxide (3 × 50 mL), and water (2 × 50 mL), dried over
anhydrous sodium sulfate, filtered, and concentrated to give
an oil which was dissolved in ethanol and hydrochloric acid.
A white solid was obtained after removing the solvent by rotary
evaporation. Recrystallization from ethanol/diethyl ether
yielded a white powder (6.5 g, 72%): mp 147-148 °C. Anal.
(C19H25N2O3Cl) C, H, N, Cl.
Analogous procedure using 15 and 18 in aqueous ethanol
gave 63 and 64, respectively, as shown in Table 2.
1-[2′-(Me t h oxyca r b on yl)e t h yl]-2-m e t h yl-3-h yd r oxy-
4(1H)-p yr id in on e Hyd r och lor id e, 65. A solution of metha-
nol (0.5 g, 15 mmol) and triphenylphosphine (2.88 g, 11 mmol)
in dry tetrahydrofuran (30 mL) was added to a solution of
diethyl azodicarboxylate (1.92 g, 11 mmol) and 16 (2.88 g, 10
mmol) in dry tetrahydrofuran (100 mL) at room temperature.
After the reaction mixture stirred for 16 h, the solvent was
removed under reduced pressure. The residue was purified
by column chromatography on silica gel (eluant 10% methanol
in chloroform) to afford a light-yellow oil of 1-[2′-(methoxycar-
bonyl)ethyl]-2-methyl-3-(benzyloxy)-4(1H)-pyridinone (19) (2.65
g, 88%). The oily product was dissolved in dimethylformamide
(50 mL) and subjected to hydrogenolysis reaction in the
presence of 5% Pd/C (10% (w/w) of the compound) catalyst for
12 h. The mixture was warmed and filtered. The filtrate was
acidified using hydrogen chloride gas followed by rotary
evaporation to give a white solid. Recrystallization from
methanol/diethyl ether gave a white powder (85%): mp 142-
144 °C. Anal. (C10H14NO4Cl) C, H, N, Cl.
Analogous synthesis using ethanol gave 1-[2′-(ethoxycarbo-
nyl)ethyl]-2-methyl-3-(benzyloxy)-4(1H)-pyridinone (20) (92%).
Hydrogenolysis procedure as above gave 66 (Table 2).
Rou te 1. Syn th esis of Am id es via Su ccin im id e Active
Ester . 1-[3′-((Su ccin im id yloxy)ca r bon yl)p r op yl]-2-m eth -
yl-3-(ben zyloxy)-4(1H)-p yr id in on e, 7a . To a solution of
1-(3′-carboxypropyl)-2-methyl-3-(benzyloxy)-4(1H)-pyridino-
ne (17) (5.7 g, 19 mmol) in dimethylformamide (100 mL) were
added a solution of N-hydroxysuccinimide (2.3 g, 20 mmol) and
dicyclohexylcarbodiimide (DCCI; 4.13 g, 20 mmol) each in dry
dimethylformamide (25 mL). The resulting mixture was
stirred in darkness at 0 °C for 1 h under an atmosphere of
nitrogen and at room temperature for 16 h. After the dense
Analogous procedure using diethylamine, methylamine, and
propylamine yielded 25, 22, and 23 in yields of 70%, 74%, and
73%, respectively.
1-[3′-(Met h ylca r b a m oyl)p r op yl]-2-m et h yl-3-h yd r oxy-
4(1H)-p yr id in on e Hyd r och lor id e, 69. To a solution of 1-[3′-
(methylcarbamoyl)propyl]-2-methyl-3-(benzyloxy)-4(1H)-pyri-
dinone hydrochloride (21) (7 g, 20 mmol) in ethanol (100 mL)
and water (25 mL) was added 5% Pd/C catalyst (0.8 g). The
mixture was stirred at room temperature under a constant
stream of hydrogen for 5 h. After used catalyst was removed
by filtration, the solution was boiled with activated charcoal
for 5 min and filtered. The solvent was removed by rotary
evaporation to give a white solid. Recrystallization from
ethanol/diethyl ether gave white crystals (4.9 g, 95%): mp
207-209 °C; 1H NMR (D2O, Ext TMS) δ 1.96-2.26 (2H, m,
+NCH2CH2CH2CO), 2.36 (2H, t, CH2CO), 2.59 (3H, s, 2-CH3),
2.66 (3H, s, CONHCH3), 4.38 (2H, t, +NCH2), 7.12 (1H, d, 5-H),
8.1 (1H, d, 6-H). Anal. (C11H17O3N2Cl) C, H, N, Cl.
Analogous hydrogenolysis procedure using 22-25 gave 76-
79, respectively, as shown in Table 2.

N-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3357
Deter m in a tion of Distr ibu tion Coefficien ts Usin g th e
F ilter P r obe Meth od . 1. F r ee Liga n d s. Distribution
coefficients were determined using an automated continuous
flow technique similar to that described by Tomlinson.26 The
system was comprised of an IBM-compatible PC running the
TOPCAT program,41,42 which controlled both a Metrohm 665
Dosimat autoburet and a Pye-Unicam Lambda 5 UV/vis
spectrophotometer, as well as performing all calculations of
distribution coefficients. All distribution coefficient determi-
nations were performed using AnalaR grade reagents under
a nitrogen atmosphere using a flat-based glass vessel equipped
with a sealable lid at 25 °C. The aqueous and octanol phases
were presaturated with respect to each other before use. The
filter probe consisted of a poly(tetrafluoroethylene) (PTFE)
plunger associated with a gel-filtration column. The aqueous
phase (50 mM MOPS buffer, pH 7.4, prepared using Milli-Q
water) was separated from the two-phase system (1-octanol/
MOPS buffer, pH 7.4) by means of a hydrophilic cellulose filter
(5-µm diameter, 589/3 Blauband filter paper, Schleicher and
Schuell) mounted in the gel-filtration column adjuster (SR 25/
50, Pharmacia). A known volume (normally 25-50 mL) of
MOPS buffer (saturated with octanol) was taken in the flat
base mixing chamber. After a baseline was obtained the
of the aqueous layer was then carefully removed using a glass
Pasteur pipet ensuring that the sample was not contaminated
with 1-octanol. The distribution coefficient was calculated
from the ratio of the equilibrium absorbance to the decrease
in absorbance of the aqueous phase. For each sample, the
experiment was repeated at least four times which led to
calculation of a mean distribution coefficient value and stan-
dard deviation.
Deter m in a tion of P h ysica l Con sta n ts of Liga n d s. 1.
p Ka Det er m in a t ion . Equilibrium constants of protonated
ligands were determined using an automated computerized
system, consisting of a Metrohm 665 dosimat, a Perkin-Elmer
λ5 UV/vis spectrophotometer, a Corning ion analyzer 225, and
a 286 Opus PC which controlled the integrated system.25
A
combined Sirius electrode was used to calibrate the electrode
zero and to measure pH values. This system is capable of
undertaking simultaneous spectrophotometric and potentio-
metric measurements. A blank titration of 0.1 M KCl (25 mL)
was carried out to determine the electrode zero using Gran’s
plot method.43 The solution (0.1 M KCl, 25 mL), contained in
a jacketed titration cell, was acidified by 0.15 mL of 0.2 M HCl.
Titrations were carried out against 0.3 mL of 0.2 M KOH using
0.01-mL increments dispensed from a Metrohm 665 dosimat.
Solutions were maintained at 25 ( 0.5 °C under an argon
atmosphere. The above titration was repeated in the presence
of ligand. The data obtained from titrations were analyzed
by the TITRFIT program, a modified version of NONLIN15.25
The pKa values were obtained to an accuracy of (0.02 pH unit.
2. Ir on (III) Affin ity Con sta n t Deter m in a tion . The Ki
value for the iron(III)-ligand interaction was determined by
a spectrophotometric competition study of ligand-Fe(III)-
EDTA using the automated system.42 The iron(III) complexes
of the ligand were prepared in 5:1 molar ratio in 0.1 M 3-(N-
morpholino)propanesulfonic acid (MOPS) buffer, pH 7.4. This
solution was then titrated against EDTA. The resulting
spectrophotometric data were inserted into the COMPT1
program to evaluate the affinity constants of the complex.
Affinity constants were obtained to an accuracy of (0.02 log
unit. Speciation plots were obtained by running the program
MLCOM50.25 These programs require concentrations of metal
and ligand, Ki(FeIII) values of complexes, Ki values for the
FeIII-OH interaction, and pKa value of ligand.
solution was used for reference absorbance. A 1 × 10-4
M
solution of the ligand was prepared in the aqueous phase
(typically 50 mL) to give an absorbance of between 0.5 and
1.5 absorbance units at the preselected wavelength (∼280 nm).
The flow rate of the aqueous circuit was limited to 1 mL m-1
.
The computer program was initiated. When
a constant
absorbance was obtained (defined as an absorbance change of
less than 0.002 absorbance unit over a minimum of 100
individual readings) a suitable aliquot of 1-octanol was added
to the aqueous phase from the automatic dispenser. The two
immiscible phases were continuously stirred to ensure effective
partitioning of the sample compound. Subsequent absorbance
readings were recorded until the system reached equilibrium
when a further aliquot of 1-octanol was added. The cycle was
repeated for at least five additions of 1-octanol. An estimate
of the distribution coefficient was obtained from each 1-octanol
addition, which enabled calculation of a mean distribution
coefficient value and standard deviation.
2. Ir on (III) Com p lexes. Distribution coefficients were
measured using a 10:1 molar ratio of ligand to iron to ensure
that the 3:1 neutral complexes formed. The ligand was
dissolved in 10 mL of MOPS buffer (pH 7.4) to give a stock
solution with a concentration of 0.1 M. Iron(III) nitrate
nonahydrate was dissolved in 10 mL of HCl (0.1 M) at a
concentration of 0.1 M. The iron(III) complexes were prepared
in the glass vessel by the addition of 400 µL of ligand solution
to 39.56 mL of MOPS buffer, and then the iron solution (40
µL) was added to give a total volume of 40 mL resulting in
concentrations of ligand and iron of 10-3 and 10-4 M, respec-
tively. Absorbance of the iron(III) complexes was monitored
at 460 nm (λmax). The partitioning was carried out as described
above.
Biologica l Meth od s. The guidelines provided by the Home
Office (Animals (Scientific Procedures) Act 1986) were adopted
throughout this study.
1. Excr etion of Ir on in Bile. In vivo testing of HPOs
was carried out in non-iron-overloaded, bile duct-cannulated
rats.28 Male rats (Fischer) weighing 200-240 g were switched
from standard rodent food (Nafag 890 rat maintenance diet;
Gossau, Switzerland) to an iron-poor diet (Nafag 9046) 5 days
prior to the start of the test. One day before surgery, the
animals were fasted and adapted to metabolic cages where
drinking water was freely accessible.
Rats were anesthetized by inhalation of 4% isofluane
(Forene, Abbott) mixed with oxygen. After complete anesthe-
tization, the animal body was transferred to an operation table
and the concentration of isofluane reduced to about 2%. The
abdomen and shoulder areas were shaved and sterilized with
70% ethanol. The abdominal cavity was opened by an incision
of 2 cm allowing the stomach and duodenum to be visualized,
after retraction of the abdominal wall. After locating the bile
duct leading from the liver to the duodenum, forceps were used
to free the duct from surrounding connective tissue. A small
opening was made in the duct about 1 cm from the duodenum
using iris scissors. A cannula was prepared by cutting
polyethylene tubing (Portex, U.K.; i.d. 0.58 mm, o.d. 0.96 mm)
at an angle to leave a bevel point. The cannula was carefully
inserted 1.5 cm into the bile duct. After bile flow was
established the cannula was tied snugly in place with surgical
thread. The cannula was further secured by placing a second
suture below the insertion. The abdominal layer was then
closed with thread, allowing the collection tubing to emerge
from the lower end of the incision. A skin tunneling needle
was inserted at the shaved shoulder area and fed around to
Molecu la r Mod elin g. The structure of the iron(III)-50
complex was obtained by energy minimization procedures
using the Polak-Ribiere algorithm of Hyper Chem program;
Hyper Chem (Release 2) for Silicon Graphics (Hypercube),
1992.
Deter m in a tion of Distr ibu tion Coefficien t Va lu es Us-
in g th e Sh a k e-F la sk Meth od . The distribution coefficients
of the iron(III) complexes with values greater than 100 were
monitored using MOPS buffer/1-octanol volume ratio of 100:1
at 25 °C. In contrast, for both ligands and iron(III) complexes
which possess distribution coefficient values less than 0.01,
MOPS buffer/1-octanol volume ratio of 1:100 was used.27
A
solution of ligands with concentration of 10-4 M was prepared
in MOPS buffer (pH 7.4), and the absorbance of the solution
was measured in the ultraviolet region at a wavelength of
approximately 289 nm using the buffer as a blank. A 10-mL
sample of the solution was stirred vigorously with 1000 mL of
1-octanol in a glass vessel for 1 h. The two layers were
separated by centrifugation at 3000 rpm for 5 min. An aliquot

3358 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Rai et al.
the abdominal incision. The collection tubing was threaded
through the needle until it emerged from the shoulder opening.
The needle was then removed and the abdominal dermal flaps
closed with surgical thread. The collection tubing was passed
through a rodent jacket (Harvard Instruments) worn around
the chest to a miniature single channel fluid swivel inside a
metal torque-transmitting tether. The animal was observed
at 30 °C constant temperature under an infrared lamp for 30
min and its bile collected in a 10-mL plastic tube.
thank Mrs. S. Lu for skillful assistance with the
biological investigations.
Su p p or tin g In for m a tion Ava ila ble: Full elemental and
spectral analyses of compounds 8-25 and 47, 49, 51, 52, 56,
57, 62-66, 69, and 76-79 (14 pages). Ordering information
is given on any current masthead page.
Refer en ces
Rats were transferred back to their individual metabolic
cages. The collection tubing was directed from the animal to
a fraction collector by a fluid swivel mounted above the
metabolic cage. This system allowed the animal to move freely
in the cages while continuous bile samples were collected.
Control samples of bile and urine were collected for 3 h. After
administration of the test compound (subcutaneously or by
gavage), bile samples were collected in plastic disposable tubes
for 24 h at 3-h intervals and urine was collected for 24 h. To
reduce postoperative pain, the rats received three subcutane-
ous doses of Temgesic (Reckitt & Colman, Hull, England) (the
first immediately after operation, the second 5 h after opera-
tion, and the third 16 h later, each 0.6 mL/kg of body weight).
Temgesic contains the long-acting and potent analgesic bu-
prenorphine. Food was given ad libitum 2 h after the drug
application.
(1) Halliwell, B.; Gutteridge, J . M. C. Free Radicals in Biology and
Medicine, 2nd ed.; Clarendon Press: Oxford, 1998; pp 22-85.
(2) Barton, J . C.; Bertoli, L. F. Hemochromatosis: The genetic
disorder of the twenty-first century. Nature Med. 1996, 2, 394-
395.
(3) Pippard, M. J . Secondary Iron Overload. In Iron Metabolism in
Health and Disease; Brock, J . H., Halliday, J . W., Pippard, M.
J ., Powell, L. W., Eds.; W. B. Saunders Co. Ltd.: London, 1994;
pp 271-309.
(4) Hider, R. C.; Choudhury, R.; Rai, B. L.; Dekhordi, L. S.; Singh,
S. Design of Orally active iron chelators. Acta Haematol. 1996,
95, 6-12.
(5) Hershko, C. Iron Chelators in Medicine. Mol. Aspects Med. 1992,
13, 114-165.
(6) Raymond, K. N.; Mu¨ller, G.; Maczanke, B. Complexation of iron
by siderophores: A review of their solution and structural
chemistry and biological function. Top. Curr. Chem. 1984, 123,
49-102.
At the end of the experiment, animals were sacrificed with
CO2. The volume of each bile and urine fraction was mea-
sured, and the samples were stored at -20 °C until determi-
nation of their iron concentration.
2. Bile a n d Ur in e Ir on Assa y. Iron concentration in bile
and urine was determined colorimetrically using the Batho-
phenanthrolin method.44 Experimental operations were per-
formed with the aid of a Biomek 2000 pipet robot (Beckman).
(7) Martell, A. E.; Motekaitis, R. J .; Murase, I.; Sala, L. F.; Stoldt,
R.; Ng, C. Y. Development of iron chelators for Cooley’s Anaemia.
Inorg. Chim. Acta 1987, 138, 215-230.
(8) Porter, J . Oral iron chelators: Prospects for future development.
Eur. J . Haematol. 1989, 43, 271-285.
(9) Hider, R. C.; Porter, J . B.; Singh, S. The design of therapeutically
useful iron chelators. In The Development of Iron Chelators for
Clinical Use; Bergeron, R. J ., Brittenham, G. M., Eds.; CRC
Press: Ann Arbor, 1994; pp 353-371.
(10) (a) Martell, A. E.; Motekaitis, R. J .; Clarke, E. T. Synthesis of
N,N′-di(2-hydroxybenzyl)ethylenediamine-N,N′-diacetic acid
(HBED) derivatives. Can. J . Chem. 1986, 64, 449-456. (b)
Hershko, C.; Grady, R. W.; Link, G. Phenolic ethylenediamine
derivatives: A study of orally effective iron chelators. J . Lab.
Clin. Med. 1984, 103, 337-346.
(11) Hider, R. C.; Hall, A. D. Clinically useful chelators of tripositive
elements. Prog. Med. Chem. 1991, 28, 31-173.
(12) Anderegg, G.; Raber, M. Metal complex formation of a new
siderophore desferrithiocin and of three related ligands. J . Chem.
Soc., Commun. 1990, 1194-1196.
(13) Garrett, T. M.; Miller, P. W.; Raymond, K. N. 2,3-Dihydroxy-
terephthalamides: Highly efficient iron(III)-chelating agents.
Inorg. Chem. 1989, 28, 128-133.
(14) Orama, M.; Tilus, O.; Taskinen, J .; Lotta, T. Iron(III)-chelating
properties of the novel catechol O-methyl inhibitor entacapone
in aqueous solution. J . Pharm. Sci. 1997, 86, 827-831.
(15) Brittenham, G. M. Development of iron-chelating agents for
clinical use. Blood 1992, 80, 569-575.
(16) Singh, S.; Epemolu, R. O.; Dobbin, P. S.; Tilbrook, G. S.; Ellis,
B. L.; Damani, L. A.; Hider, R. C. Urinary metabolic profiles in
human and rat of 1,2-dimethyl and 1,2-diethyl-substituted-3-
hydroxypyridin-4-ones. Drug Met. Disp. 1992, 20, 256-261.
(17) Dobbin, P. S.; Hider, R. C.; Hall, A. D.; Taylor, P. D.; Sarpong,
P.; Porter, J . B.; Xiao, G.; van der Helm, D. Synthesis, physico-
chemical properties, and biological evaluation of N-substituted-
2-alkyl-3-hydroxy-4(1H)-pyridinones: Orally active iron chela-
tors with clinical potential. J . Med. Chem. 1993, 36, 2448-2458.
(18) Le Blanc, D. T.; Akers, H. A. Maltol and ethyl maltol. Food
Technol. (Chicago) 1989, 43, 78-84.
(19) Amundsen, L. H.; Nelson, L. S. Reduction of nitriles to primary
amines with lithium aluminium hydride. J . Am. Chem. Soc.
1951, 73, 242-244.
(20) Mitsunobu, O.; Wada, M.; Sano, T. Stereospecific and stereose-
lective reactions 1. Preparation of amines and alcohols. J . Am.
Chem. Soc. 1972, 94, 679-680.
(21) Mitsunobu, O. The use of diethyl azodicarboxylate and tri-
phenylphosphine in synthesis and transformation of natural
products. Synthesis 1981, 1, 1-28.
(22) Mitsunobu, O.; Yamada, Y. Preparation of esters of carboxylic
acid phosphoric acid via phosphonium salts. Bull. Chem. Soc.
J pn. 1967, 40, 2380-2382.
(23) Dourtoglou, V.; Ziegler, J . C.; Gross, B. L1 Hexafluorophosphate
De O-benzo-triazolyl-N,N-tetramethyl uronium. Tetrahedron
Lett. 1978, 15, 1269-1272.
(24) Dourtoglou, V.; Gross, B. O-Benzo-triazolyl-N,N,N-tetramethyl
uronium hexafluorophosphate as coupling reagent for the syn-
thesis of peptides of biological interest. Synthesis 1984, 7, 572-
574.
Bile and urine were briefly vortexed, and 0.1 mL of each
sample was pipetted twice into two wells of a conical-bottom
microplate (96 wells; Nunc, Denmark). Water and standard
iron solution (10 µg/mL and 20 µg/mL; J anssen Chimica,
Belgium) were included on each plate. To each well, 0.1 mL
of acid solution (10% TCA, 3 M HCl) was added with an
Eppendorf multipipet. After incubation in a 65 °C bath for 4
h, the 96-well plates were centrifuged for 10 min at 2000g
(2500 rpm, Omnifuge 2.ORS; Heraeus). A portion of each
supernatant (0.1 mL) was automatically transferred (Biomek
2000, Beckman) to the corresponding well of the flat-bottom
microplate (Dynatech). Then 0.1 mL of iron indicator solution
(9.48 mL of H2O, 120 µL of mercaptoacetic acid, 18.6 mg of
bathophenanthrolin sulfonate, 4.74 mL of saturated sodium
acetate, and 6.33 mL of NaOH, 5.9 M) was added. Blank
solution (no color reagent bathophenanthrolin) was also pi-
petted to another well. The plates were agitated for 30 min,
and the absorbance was measured at 535 nm with a Biomek
plate reader. The iron concentration of each sample was
estimated by comparison with the standard solutions after
subtraction of individual blank absorbance. Calculation was
performed on a Macintosh computer using the Excel software.
3. Met h od of Ir on Mob iliza t ion E ffica cy St u d y in
Ra ts. Hepatocytes of normal fasted rats (190-230 g) were
labeled with 59Fe by administration of [59Fe]ferritin via the
tail vein. One hour later, each rat was administered orally
with chelator (450 µmol/kg). Control rats were administered
with an equivalent volume of water. The rats were placed in
individual metabolic cages and urine and feces collected. Rats
were allowed access to food 1 h after oral administration of
chelator. There was no restriction of water throughout the
study period. The investigation was terminated 24 h after the
[59Fe]ferritin administration, rats were sacrificed and the liver
and gastrointestinal tract (including its content and feces) were
removed for gamma counting.
Ack n ow led gm en t. L.S.D. acknowledges the Iranian
Government for financial support, and Z.L. acknowl-
edges the KC Wong Foundation for a studentship and
the CVPC for an ORS award. The authors also wish to

N-Substituted-2-alkyl-3-hydroxy-4(1H)-pyridinones
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3359
(25) Taylor, P. D.; Morrison, I. E. G.; Hider, R. C. Microcomputer
application of nonlinear regression analysis to metal-ligand
equilibria. Talanta 1988, 35, 507-512.
(36) Hider, R. C. Potential protection from toxicity by oral iron
chelators. Toxicol. Lett. 1995, 82/83, 961-967.
(37) Molenda, J . J .; J ones, M. M.; Basinger, M. A. Enhancement of
iron excretion via monoanionic 3-hydroxypyridin-4-ones. J . Med.
Chem. 1994, 37, 93-98.
(38) Hider, R. C.; Lerch, K. The inhibition of tyrosine by pyridinones.
Biochem. J . 1989, 257, 289-290.
(39) Choudhury, R.; Epemolu, R. O.; Rai, B. L.; Hider, R. C.; Singh,
S. Metabolism and pharmacokinetics of 1-(2′-trimethylacetoxy-
ethyl)-2-ethyl-3-hydroxypyridin-4-one (CP117) in the rat. Drug
Metab. Dispos. 1997, 25, 332-339.
(40) Pollock, J . R. A.; Stevens, R. Dictionary of organic compounds,
4th ed.; Eyre and Spottiswoode Ltd.: London, 1965; Vol. 1, p
162.
(41) Hall, A. TOPCAT program for Determination of distribution
coefficients. Department of Pharmacy, King’s College London,
U.K., 1990.
(26) Tomlinson, E. Filter-probe extractor.
A tool for the rapid
determination of oil-water partition coefficients. J . Pharm. Sci.
1982, 71, 602-604.
(27) Ellis, B. L.; Duhme, A. K.; Hider, R. C.; Hossain, M. B.; Rizvi,
S.; van der Helm, D. Synthesis, physicochemical properties, and
biological evaluation of hydroxypyranones and hydroxypyridi-
nones: Novel bidentate ligands for cell-labeling. J . Med. Chem.
1996, 39, 3659-3670.
(28) Bergeron, R. J .; Wiegand, J . B.; Dionis, J . B.; Egli-Karmakka,
M.; Frei, J .; Huxley-Tencer, A.; Peter, H. H. Evaluation of
desferrithiocin and its synthetic analogues as orally effective iron
chelators. J . Med. Chem. 1991, 34, 2072-2078.
(29) Pippard, M. J .; J ohnson, D. K.; Finch, C. A. A rapid assay for
evaluation of iron-chelating agents in rats. Blood 1981, 58, 685-
692.
(30) Hershko, C.; Cook, J . D.; Finch, C. A. Storage iron kinetics III.
Study of desferrioxamine action by selective radioiron labels of
RE and parenchymal cell. J . Lab. Clin. Med. 1973, 81, 876-
886.
(31) Ma, R.; Reibenspies, J . J .; Martell, A. E. Stabilities of 1,2-diethyl-
3-hydroxy-4-pyridinone chelates of divalent and trivalent metal
ions. Inorg. Chim. Acta 1994, 223, 21-29.
(32) Hider, R. C.; Taylor, P. D.; Walkinshaw, M.; Wang, J . L.; van
der Helm, D. Crystal structure of 3-hydroxy-1,2-dimethylpyridin-
4(1H)-one. An iron(III) chelation study. J . Chem. Res. 1990, 316-
317.
(33) Edward, J . T. Iron transport across biological membranes; Some
fundamental considerations. In Intestinal absorption of metal
ions, trace metals and radionucleotides; Skoryna, S. C., Edward,
D. M., Eds.; Pergamon Press: Oxford, 1970; pp 8-26.
(34) Edward, J . T.; Ponka, P.; Richardson, D. R. Partition coefficients
of the iron(III) complexes of pyridoxal isonicotinoyl hydrazone
and its analogues and the correlation to iron chelation efficacy.
Biometals 1995, 8, 209-217.
(42) Khodr, H. KDHK94 Program (a modified version of TOPCAT
Program) for determination of distribution coefficients. Depart-
ment of Pharmacy, King’s College London, U.K., 1994.
(43) Gran, G. Determination of the equivalence point in potentio-
metric titrations. Analyst 1952, 77, 661-671.
(44) Smith, G. F.; McCurdy, W. H.; Diehl, H. The colorimetric
determination of iron in raw and treated municipal water
supplies by use of 4:7-diphenyl-1:10-phenanthroline. Analyst
1952, 77, 418-422.
(45) Kontoghiorghes. G. J .; Barr, J .; Nortey, P.; Sheppard, L. Selection
of a new generation of orally active R-ketohydroxypyridine iron
chelators intended for use in the treatment of iron overload. Am.
J . Haematol. 1993, 42, 340-349.
(46) Nelson, W. O.; Karpishin, T. B.; Rettig, S. J .; Orvig, C. Physical
and structural studies of N-substituted-3-hydroxy-2-methyl-
4(1H)-pyridinones. Can. J . Chem. 1988, 66, 123-131.
(47) Zhang, Z.; Rettig, S. J .; Orvig, C. Lipophilic coordination
compounds: aluminium, gallium, indium complexes of 1-aryl-
3-hydroxy-2-methyl-4-pyridinones. Inorg. Chem. 1991, 30,
509-515.
(35) Singh, S.; Choudhury, R.; Epemolu, R. O.; Hider, R. C. Metabo-
lism and pharmacokinetics of 1-(2′-hydroxyethyl)- and 1-(3′-
hydroxyethyl)-2-ethyl-3-hydroxypyridin-4-ones in the rat. Eur.
J . Drug Metab. Pharmacokinet. 1996, 21, 33-41.
J M9707784