Synthesis, physicochemical properties and biological evaluation of aromatic ester prodrugs of 1-(2'-hydroxyethyl)-2-ethyl-3-hydroxypyridin-4-one (CP102): Orally active iron chelators with clinical potential

Article abstract of DOI:10.1211/0022357991772655

The synthesis of seven aromatic ester derivatives of 1-(2'-hydroxyethyl)-2-ethyl-3-hydroxypyridin-4-one is described. These ester prodrugs have been designed to target iron chelators to the liver, the major iron storage organ. In principle this should improve chelation efficacy and minimize toxicity. The distribution coefficients of these ester prodrugs between 1-octanol and MOPS buffer pH 7.4 were measured together with their rates of hydrolysis at pH 2 and pH 7.4, in rat blood and liver homogenate. Esters with heteroaromatic acid moieties were found to be less stable than benzoyl analogues. The in-vivo iron mobilisation efficacy of these ester prodrugs has been compared with that of the parent drug using a ⁵⁹Fe-ferritin loaded rat model. Many prodrugs were found to enhance the ability of the parent hydroxypyridinone to facilitate ⁵⁹Fe excretion. However, not all prodrugs provided increased efficacy, demonstrating that lipophilicity is not the only factor which influences drug efficacy. Furthermore, no clear correlation between efficacy and susceptibility to hydrolysis was detected. The picolinic and nicotinic ester derivatives appear to offer the best potential as prodrugs as they have a relatively low LogP value and yet lead to enhanced efficacy over the parent hydroxypyridinone.

Full text of DOI:10.1211/0022357991772655

J. Pharm. Pharmacol. 1999, 51: 555±564
Received November 23, 1998
Accepted December 12, 1998
# 1999 J. Pharm. Pharmacol.
Synthesis, Physicochemical Properties and Biological Evaluation
of Aromatic Ester Prodrugs of 1-(20-Hydroxyethyl)-2-ethyl-3-
hydroxypyridin-4-one (CP102): Orally Active Iron Chelators
with Clinical Potential
ZU DONG LIU, DING YONG LIU, SHU LI LU AND ROBERT C. HIDER
Department of Pharmacy, King's College London, Manresa Road, London SW3 6LX, UK
Abstract
The synthesis of seven aromatic ester derivatives of 1-(20-hydroxyethyl)-2-ethyl-3-hydro-
xypyridin-4-one is described. These ester prodrugs have been designed to target iron
chelators to the liver, the major iron storage organ. In principle this should improve
chelation ef®cacy and minimize toxicity.
The distribution coef®cients of these ester prodrugs between 1-octanol and MOPS buffer
pH 7Á4 were measured together with their rates of hydrolysis at pH 2 and pH 7Á4, in rat
blood and liver homogenate. Esters with heteroaromatic acid moieties were found to be
less stable than benzoyl analogues. The in-vivo iron mobilisation ef®cacy of these ester
prodrugs has been compared with that of the parent drug using a 59Fe-ferritin loaded rat
model. Many prodrugs were found to enhance the ability of the parent hydroxypyridinone
to facilitate 59Fe excretion. However, not all prodrugs provided increased ef®cacy,
demonstrating that lipophilicity is not the only factor which in¯uences drug ef®cacy.
Furthermore, no clear correlation between ef®cacy and susceptibility to hydrolysis was
detected.
The picolinic and nicotinic ester derivatives appear to offer the best potential as prodrugs
as they have a relatively low LogP value and yet lead to enhanced ef®cacy over the parent
hydroxypyridinone.
Transfusion-dependent patients such as those suf-
fering from b-thalassaemia develop a fatal sec-
iron-overloaded animal models and in thalassaemic
patients. Most results have shown that excretion of
iron can be enhanced via both the urinary and
biliary routes, and some compounds have potential
as clinically useful chelators (Porter et al 1989).
Unfortunately, the majority of effort with human
studies has centred on the 1,2-dimethyl analogue 1
(L1, Deferiprone; Table 1), which is a relatively
inef®cient chelator under in-vivo conditions and
consequently high doses must be utilised in order to
achieve clinically useful levels of iron excretion.
One of the major reasons for the limited ef®cacy of
1 in clinical use is that it undergoes extensive phase
II metabolism in the liver (Singh et al 1992a). The
3-hydroxyl functionality, which is crucial for
scavenging iron, is also a prime target for phase II
conjugation. Urinary recovery studies conducted on
1 in both rats and man have shown that, respec-
tively, >44% and >85% of the administered dose
is recovered in the urine as the non-chelating 3-O-
glucuronide conjugate (Figure 1) (Singh et al
ondary haemosiderosis and consequently
a
selective iron chelator must be used to relieve such
iron overload (Pippard et al 1978a). Currently,
desferrioxamine is the only clinically useful drug
which is widely available for this purpose. Unfor-
tunately patient compliance is poor, because of the
lack of oral activity and rapid plasma clearance
(Pippard et al 1978a, b ). Consequently there is an
urgent need for an orally active iron chelating
agent.
3-Hydroxypyridin-4-ones are currently one of the
main candidates for the development of orally
active iron chelators (Tilbrook & Hider 1998). To
date, several 3-hydroxypyridin-4-one ligands have
been widely investigated for iron chelation, both in
Correspondence: R. C. Hider, Department of Pharmacy,
King's College London, Manresa Road, London SW3 6LX,
UK.
E-mail: robert.hider@kcl.ac.uk

556
ZU DONG LIU ET AL
limited clinical evaluation of 2 in thalassaemic
patients (Porter et al 1994). Unfortunately, the
metabolism of 2 in man did not parallel that in the
rat. The main urinary metabolite of 2 in man is the
3-O-glucuronide conjugate (>85%; Porter et al
1993). Extensive conversion to this metabolite was
found to severely limit the clinical ef®cacy of 2.
This critical dependence of chelator ef®cacy on
metabolic behaviour has led to a concept of ligand
design which minimises conjugation reactions with
glucuronic acid. Despite the limited ef®cacy of 2 in
man, the superior extracellular and intracellular
Figure 1. Metabolism of 1 in rat and man to form the
corresponding non-chelating glucuronide conjugates.
1992a). This extensive biotransformation of 1
limits its ability to mobilise excess body iron in
thalassaemic patients (Oliveri et al 1992).
The 1,2-diethyl analogue 2 (Table 1) has also
been introduced into man. This chelator has been
found to be more ef®cient at iron removal than 1 in
several mammalian species, for example the rat
(Porter et al 1990a, b). A likely explanation for this
greater ef®cacy in the rat model is the difference in
metabolism of 2 and 1 (Figure 2) (Singh et al
1992a). The 3-O-glucuronide of 2, unlike that of 1,
is only a minor metabolite in the rat. The major
metabolic pathway for 2 in the rat is phase I
metabolism which leads to the formation of the 2-
(10-hydroxyethyl) metabolite (metabolite I; 3). This
metabolite does not undergo further phase II
metabolism to form a glucuronide conjugate and
hence retains the ability to chelate iron (Figure 2).
Promising results obtained in rat models led to the
Figure 2. Metabolism of 2 via phase I and phase II metabolic
pathways. Metabolite I (3) is the major metabolite in rat;
metabolite II is the major metabolite in man.
Table 1. Chemical structure of selected 3-hydroxypyridin-4-ones and aliphatic ester derivatives of CP102.
Compound
R1
R2
MW
D7Á4
1
2
CH3
CH2CH3
CH2CH3
CH2CH2OH
CH2CH2OCOCH3
CH2CH2OCOCH(CH3)2
CH2CH2OCOC(CH3)3
CH3
139
167
183
183
225
253
267
0Á17
1Á78
0Á25
0Á22
0Á58
5Á06
14Á5
CH2CH3
CH(OH)CH3
CH2CH3
CH2CH3
CH2CH3
CH2CH3
3
CP102 (4)
5
6
7

HYDROXYPYRIDINONE PRODRUGS
557
iron mobilisation ability in the rat provided
important information for chelator design. The
lack of glucuronidation of the 2-(10-hydroxy-
ethyl) metabolite of 2 led to the investi-
gation of the possibility of developing structurally
related compounds. Indeed, 1-hydroxyalkyl deri-
vatives of 3-hydroxypyridin-4-ones such as 1-(20-
hydroxyethyl)-2-ethyl analogue CP102 (4; Table
1), which are not extensively metabolised via phase
II reactions, have been identi®ed (Singh et al
1992b, 1996). Ef®cacy studies of CP102 in rat
indicated that the iron excretion was at least twice
that seen with 1 (Jin et al 1993). Further investi-
gation in primates con®rmed that CP102 is even
more effective; urinary and biliary iron excretion
being almost three times greater than that of 1 (Jin
et al 1993).
Although the use of CP102 may offer a sig-
ni®cant improvement over previously evaluated 3-
hydroxypyridin-4-ones, the disadvantage of this
hydrophilic compound (D7Á4 ˆ 0Á22) is poor oral
absorption and insuf®cient extraction by the liver,
which is the major iron storage organ. The devel-
opment of hydrophobic ester prodrugs of CP102 is
one route which has been considered for improving
both drug absorption and hepatic extraction. A
range of 1-positioned aliphatic ester derivatives of
CP102 has been prepared by Rai et al (1999) (Table
1). In-vitro esterase studies indicate that 7, which is
the pivaloyl ester of CP102, may partially ful®l the
requirements for relatively ef®cient liver extrac-
tion. This is due to its resistance towards hydrolysis
in pH 2 and 7Á4 buffer and a much lower hydrolysis
rate in plasma than the liver, whereas 5 and 6 show
much faster hydrolysis rates in plasma (Choudhury
1995). Preliminary pharmacokinetic and absorption
studies of 7 in the rat have demonstrated that
selective targeting of the liver is possible
(Choudhury et al 1997). Results obtained indicate
that 7 is rapidly absorbed from the gastrointestinal
tract in the intact form and subsequently undergoes
extensive ®rst pass metabolism. Unfortunately,
pivalic acid possesses undesirable side effects in
man when used in high doses. For instance, the
carnitine cycle is inhibited by branched acids which
can lead to serious muscle side effects (Melegh et
al 1987; Holme et al 1989, 1992 ). In order to
overcome this problem, alternate ester prodrug
derivatives of CP102 need to be identi®ed.
1982; Poon & Pang 1995). It is well established
that benzoic acid is conjugated with glycine to form
hippuric acid in humans and that this conjugate is
not toxic (Ansel & Levy 1969; Borshop et al 1989).
By selecting different aromatic acid moieties, it is
possible to synthesise a range of ester derivatives
with differing lipophilicity and possibly differing
biological activity.
Materials and Methods
Chemistry
Melting points were determined using an Electro-
thermal IA 9100 Digital Melting Point Apparatus
and are uncorrected. H-NMR spectra were recor-
1
ded using a Perkin-Elmer (60 MHz) NMR spec-
trometer. Chemical shifts (d) are reported in ppm
down®eld from the internal standard tetra-
methylsilane (TMS). Elemental analyses were
performed by Micro Analytical Laboratories,
Department of Chemistry, The University of
Manchester, Manchester, M13 9PL, UK. The
common synthetic pathway of the ester prodrugs of
CP102 is summarised in Figure 3.
2-Ethyl-3-benzyloxypyran-4(1H)-one (8). To
a
solution of ethyl maltol (56 g, 0Á4 mol, 1 equiva-
lent) in methanol (500 mL) was added sodium
hydroxide (17Á6 g, 0Á44 mol, 1Á1 equivalent) dis-
solved in water (40 mL) followed by benzyl
chloride (55Á7 g, 0Á44 mol, 1Á1 equivalents) and the
mixture was re¯uxed for 6 h. After removal of
solvent by rotary evaporation, the residue was
mixed with water (200 mL) and extracted with
dichloromethane (3 6 100 mL). The combined
extracts were washed with 5% aqueous sodium
hydroxide (2 6 200 mL), followed by water
(2 6 200 mL). The organic fraction was then dried
over anhydrous sodium sulphate, ®ltered and rotary
evaporated to yield an orange oil which solidi®ed
on cooling. Recrystallization from diethyl ether
gave the pure product (8) as colourless needles
1
(78 g, 84Á8%), mp 33±34ꢀC. H-NMR (CDCl3) d:
1Á0 (t, 3H, 2-CH2 CH3), 2Á55 (q, 2H, 2-CH2CH3),
5Á13 (s, 2H, CH2Ph), 6Á3 (d, 1H, 5- H(pyranone)),
7Á35 (s, 5H, Ar), 7Á6 (d, 1H, 6-H (pyranone)).
2-[2-Ethyl-3-benzyloxypyridin-4(1H)-one-1-yl]-
ethanol (9). To a solution of (8) (25 g, 0Á11 mol, 1
equivalent) in ethanol (200 mL)=water (200 mL)
was added ethanolamine (10 g, 0Á165 mol, 1Á5
equivalent) followed by 2 N sodium hydroxide
solution until pH 13Á5 was obtained. The mixture
was re¯uxed overnight. After adjustment to pH 1
with concentrated hydrochloric acid, the solvent
In this study we report the synthesis and prop-
erties of aromatic esters of CP102. The advantage
of using aromatic moieties is that they do not
interfere with the carnitine transport system.
Indeed, benzoic acid is used clinically to treat
hyperammonaemia caused by a genetic defect in
urea synthesis (Brusilow et al 1979; Batshaw et al

558
ZU DONG LIU ET AL
Figure 3. The synthesis of aromatic ester derivatives of CP102.
was removed by rotary evaporation before addition
of water (200 mL) and washing with diethyl ether
(2 6 200 mL). Subsequent adjustment of the aqu-
eous fraction to pH 7 with 10 N sodium hydroxide
solution was followed by extraction into dichlor-
omethane (4 6 200 mL). The combined organic
layers were dried over anhydrous sodium sulphate,
®ltered and rotary evaporated to give a yellow
solid. Recrystallization from ethyl acetate gave the
pure product (9) as yellow needles (17Á85 g, 60%),
mp 120±122ꢀC. 1H-NMR (CDCl3) d: 1Á0 (t, 3H, 2-
CH2CH3), 2Á65 (q, 2H, 2-CH2CH3), 3Á8 (br., s, 4H,
NCH2CH2O), 5Á0 (s, 2H, CH2Ph), 5Á3±6Á2 (br., 1H,
OH), 6Á1 (d, 1H, 5-H (pyridinone)), 7Á0±7Á5 (m,
6H, Ar & 6-H (pyridinone)). Anal. calcd for
C16H19NO3: C 70Á31, H 7Á01, N 5Á12%. Found: C
70Á58, H 6Á89, N 4Á96%
Pd=C (5±10% w=w of the compound) catalyst for
12 h. The mixtures were warmed and ®ltered. The
®ltrates were acidi®ed using hydrogen chloride gas
followed by rotary evaporation to give white solids.
Recrystallization from methanol=diethyl ether gave
the pure products (11a±g) as white powders.
Distribution coef®cient determination
Distribution coef®cients were determined using an
automated continuous ¯ow technique as described
by Rai et al (1998). The system comprised an IBM
compatible PC running the Omniferous Personal
Computer Auto-Titrator (TOPCAT) program,
which controlled both a Metrohm 665 Dosimat
autoburette and a Pye-Unicam Lambda 5 UV=vis
spectrophotometer, as well as performing all cal-
culations of distribution coef®cients. All distribu-
tion coef®cient determinations were performed
using analytical-grade reagents under a nitrogen
atmosphere using a ¯at-based glass vessel equipped
with a sealable lid at 25ꢀC. The aqueous and
octanol phases were presaturated with respect to
each other before use. The ®lter probe consisted of
a polytetra¯uoroethylene plunger associated with a
gel-®ltration column. The aqueous phase (50 mM
MOPS buffer, pH 7Á4, prepared using Milli-Q
water) was separated from the two-phase system
(1-octanol±MOPS buffer, pH 7Á4) by means of a
hydrophilic cellulose ®lter (5 mm diameter), 589=3
Blauband ®lter paper, Schleicher and Schuell)
mounted in the gel-®ltration column adjuster (SR
25=50, Pharmacia). A known volume (normally
25±50 mL) of MOPS buffer (saturated with octa-
nol) was taken in the ¯at base mixing chamber.
After a base line was obtained, the solution was
used for reference absorbance. The compound to be
examined was dissolved in buffer (saturated with
octanol) so as to give an absorbance of between
General method for the preparation of benzylated
ester derivatives of CP102 (10a±g). A solution of
triphenyl phosphine (2Á88 g, 11 mmol, 1Á1 equiva-
lents) and benzylated CP102 (9) (2Á73 g, 10 mmol,
1 equivalents) in dry tetrahydrofuran (75 mL) was
added dropwise to a solution of diethyl azodi-
carboxylate (1Á914 g, 11 mmol, 1Á1 equivalents) and
the corresponding aromatic acid (10 mmol, 1
equivalent) in dry tetrahydrofuran (15 mL) at room
temperature. After stirring the mixture overnight at
room temperature, the solvent was removed under
reduced pressure. The residue thus obtained was
puri®ed by column chromatography on silica gel
(eluant, methanol±chloroform, 8 : 92) to afford the
product (10a±g) as a yellow oil.
General method for the preparation of ester
derivatives of CP102 (11a±g). Solutions of the
benzylated ester derivatives of CP102 (10a±g) (2±
3 g) in N,N-dimethylformamide (40±50 mL) were
subjected to hydrogenolysis in presence of 5%

HYDROXYPYRIDINONE PRODRUGS
559
0Á5±1Á5 absorbance units at the preselected wave-
length ( ꢁ 280 nm). The ¯ow rate of the aqueous
Extraction and HPLC procedures for analysis of
hydrolytic product. All of the incubated samples,
in phosphate buffer, whole blood and liver homo-
genate, were added to 5 mL organic solvent (aceto-
nitrile±isopropanol, 80 : 20 v=v). Solid NaCl (0Á5 g)
was added and the mixture was acidi®ed using 2 N
HCl (100 mL). The samples were then mixed and
centrifuged. The entire upper organic layer (4±
5 mL) was separated, evaporated and analysed by
an established HPLC method. A Hewlett Packard
Model 1090M Series-II HPLC system, complete
with auto injector, auto sampler and diode array
detector, linked to a HP 900-300 data station was
used for analysis. A polymer PLRP-S column
(15 cm 6 0Á46 cm) and a gradient mobile phase
system, utilising PBS (10 mM, pH 2Á90, containing
2 mM EDTA) and acetonitrile, were used for the
separation of the analytes. Eluents were monitored
at 285 nm.
1
circuit was limited to 1 mL min , as greater ¯ow
rates would result in the overloading of the
hydrophilic ®lter and subsequent octanol con-
tamination of the aqueous phase. The on-line
spectrophotometer allowed continuous assessment
of the equilibrium of the aqueous phase. Once a
stable UV absorbance was obtained, an aliquot of
octanol was added and re-equilibration assessed.
This cycle was repeated until a pre-de®ned total
volume of added octanol was reached. The dis-
tribution coef®cient was calculated for each octanol
addition.
Ester hydrolysis
In order to identify a lead prodrug by which speci®c
drug delivery could be achieved, it was essential to
determine the stability of the ester under different
conditions. The hydrolysis rates of all aromatic
ester derivatives were compared with the pivaloyl
ester analogue 7 at different pH values (pH 2Á0 and
pH 7Á4), in rat blood and in the presence of liver
homogenate.
Ef®cacy study
In-vivo iron mobilisation ef®cacy of ester prodrugs
has been compared with 1 and the parent compound
CP102 in a 59Fe-ferritin loaded rat model (Liu et al
1999). Hepatocytes of normal fasted rats (190±
230 g) were labelled with 59Fe by administration of
59Fe-ferritin by the tail vein. One hour later, each
rat was orally administered with chelator
Each ester prodrug, 10 mmol in 0Á1 mL of acet-
onitrile solution (50% v=v in water), was added to
0Á9 mL of phosphate buffer (100 mM, pH 7Á4 or pH
2Á0). The samples were incubated at 37ꢀC for var-
ious time-intervals up to 24 h. At the end of the
incubation, the samples were removed from the
incubator and immediately subjected to extraction
as described below. Fresh heparinized rat whole
blood (0Á5 mL) was added to the phosphate buffer
(0Á5 mL, 0Á139 M, pH 7Á4) containing the ester
substrate (2Á0 mmol). Subsequently the samples
were incubated at 37ꢀC for 0, 10, 20, 40 or 60 min.
The hydrolytic reaction was terminated by remov-
ing the samples onto an ice bath and extracting the
hydrolytic product immediately. For studies utilis-
ing liver homogenate, rats were killed by cervical
dislocation and the livers removed and immersed in
ice-cold phosphate buffer (100 mM, pH 7Á4) in
order to remove excess blood. The livers were
blotted dry, weighed and cut into small pieces.
Thereafter the tissue was homogenised in phos-
phate buffer (100 mM, pH 7Á4) in the ratio of 1 g
liver tissue to 4 mL buffer. Each ester (2 mmol) was
added to 1Á0 mL of liver homogenate, which was
diluted before use with phosphate buffer (100 mM,
pH 7Á4) to a concentration of 25 mg liver tissue in
10 mL of incubation mixture. The incubation was
carried out at 37ꢀC for 0, 10, 20, 40 or 60 min and
termination of the hydrolysis was achieved by
immediate extraction.
1
(450 mmol kg ). Due to the poor water solubility
associated with several of the hydrophobic prodrug
molecules, 70% 1,2-propanediol in water (v=v) was
adopted as the solvent. This compound did not
in¯uence the ef®cacy of 1 (data not shown). Con-
trol rats were administered with an equivalent
volume of 70% 1,2-propanediol. Rats were placed
in individual metabolic cages and urine and faeces
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 59Fe-
ferritin administration. Rats were killed and the
liver, gastrointestinal tract (including contents) and
faeces were removed for gamma counting.
Results
Chemistry
The properties of seven aromatic ester derivatives
1
of CP102 (11a±g) and the H-NMR spectra of the
corresponding benzyl ethers (10a±g) are presented
in Tables 2, 3 and 4. The diethyl azodicarboxy-
late=triphenylphosphine system (Mitsunobu 1981)

560
ZU DONG LIU ET AL
Table 2. Synthesis of benzylated ester derivatives of CP102 (10a±g).
Compound
Ar
Yield (%) 1H-NMR (CDCl3) d ppm
10a
68
73
1Á0 (t, 3H,2-CH2CH3), 2Á55 (q, 2H, 2-CH2CH3), 4Á0 (t, 2H, NCH2CH2O), 4Á4 (t,
2H, NCH2CH2O), 5Á1 (s, 2H,CH2Ph), 5Á2 (s, 2H, CH2Ph), 6Á25 (d, 1H, 5-H
(pyridinone)), 6Á8±7Á8 (m, 15H, Ar & 6-H(pyridinone))
10b
1Á2 (t, 3H,2-CH2CH3), 2Á65 (q, 2H, 2-CH2CH3), 4Á1 (t, 2H, NCH2CH2O),4Á4 (t, 2H,
NCH2CH2O), 5Á1 (s, 2H,CH2Ph), 5Á23 (s, 2H,CH2Ph), 6Á4 (d, 1H, 5-H(pyridi-
none)), 6Á95 (d, 1H, 6-H(pyridinone)), 7Á1 (d, 2H, Ar, H-meta to the carboxyl
group), 7Á15±7Á65 (m, 10H, Ar),7Á9 (d, 2H, Ar, H-ortho to the carboxyl group)
10c
10d
89
82
1Á0 (t, 3H, 2-CH2CH3), 2Á6 (q, 2H, 2-CH2CH3), 4Á1 (t, 2H, NCH2CH2O), 4Á4 (t, 2H,
NCH2CH2O), 5Á18 (s, 2H, CH2Ph), 6Á45 (d, 1H, 5-H(pyridinone)), 7Á0±8Á1 (m,
11H,Ar & 6-H(pyridinone))
1Á0 (t, 3H,2-CH2CH3), 2Á65 (q, 2H, 2-CH2CH3), 4Á15 (t, 2H, NCH2CH2O), 4Á55 (t,
2H, NCH2CH2O), 5Á21 (s, 2H, CH2Ph), 6Á45 (d, 1H, 5-H(pyridinone)), 7Á1±8Á2 (m,
9H, Ar & 6-H(pyridinone)), 8Á75 (d, 1H, 6-H(pyridine))
10e
10f
10g
74
72
81
1Á0 (t, 3H,2-CH2CH3), 2Á5 (q, 2H, 2-CH2CH3), 4Á0 (t, 2H, NCH2CH2O), 4Á35 (t,
2H, NCH2CH2O), 5Á0 (s, 2H, CH2Ph), 6Á12 (d, 1H, 5-H (pyridinone)), 6Á8±7Á4 (m,
7H, Ar;5-H(pyridine) & 6-H(pyridinone)), 7Á9 (m, 1H, 4-H(pyridine)), 8Á5 (d, 1H,
6-H(pyridine)), 8Á8 (s, 1H, 2-H(pyridine))
1Á0 (t, 3H,2-CH2CH3), 2Á55 (q, 2H, 2-CH2CH3), 4Á0 (t, 2H, NCH2CH2O), 4Á35 (t,
2H, NCH2CH2O), 5Á05 (s, 2H, CH2Ph), 6Á1 (d, 1H, 5-H(pyridinone)), 6Á8±7Á3 (m,
6H, Ar &6-H(pyridinone)), 7Á5 (d, 2H, 3,5-H(pyridine)), 8Á5 (d, 2H, 2,6-
H(pyridine))
1Á0 (t, 3H, 2-CH2CH3), 2Á6 (s, 3H, 6-CH3), 2Á65 (q, 2H, 2-CH2CH3), 4Á1 (t, 2H,
NCH2CH2O), 4Á45 (t, 2H, NCH2CH2O), 5Á2 (s, 2H,CH2Ph), 6Á35 (d, 1H, 5-
H(pyridinone)), 7Á1±7Á6 (m, 7H, Ar; 5-H(pyridine) & 6-H(pyridinone)), 8Á05 (m,
1H, 4-H(pyridine)), 9Á05 (s, 1H, 2-H(pyridine))
Table 3. Aromatic ester derivatives CP102 (11a±g hydrochloride) and their distribution coef®cients between an aqueous phase
buffered at pH 7Á4 and octanol (n ˆ 5).
Compound Ar
mp (ꢀC)
Yield (%)
Formula
D7Á4
Log P
11a
11b
11c
11d
11e
11f
166±168
159±161
155±158
160±163
188±190
196±197
198±201
92
88
93
86
90
92
85
C16H17NO5.HCl
C16H17NO5.HCl
C16H17NO4.HCl
C15H16N2O4.2HCl
C15H16N2O4.2HCl
C15H16N2O4.2HCl
C16H18N2O4.2HCl
50Á3Æ 0Á8
17Á7Æ 0Á7
32Á8Æ 1Á7
1Á3Æ 0Á1
1Á8Æ 0Á1
1Á8Æ 0Á2
4Á9Æ 0Á1
1Á70
1Á25
1Á52
0Á10
0Á25
0Á25
0Á69
11g
was selected due to the mild and neutral reaction
conditions required and the excellent yields.
be orally active. The para-hydroxybenzoyl ester
derivative of CP102 (11b), had a lower partition
coef®cient value than the unsubstituted analogue,
11c. In contrast, 11a, the ortho-hydroxybenzoyl
ester derivative of CP102, was more lipophilic than
11c, probably due to intramolecular hydrogen
bonding between the phenolate group and the car-
bonyl function (Figure 5), resulting in both the
hydroxyl and carbonyl groups being less exposed to
the aqueous solvent. The picolinic, nicotinic and
Distribution coef®cients
The UV spectra associated with compound 11f
(Figure 4) are presented as an example. The dis-
tribution coef®cient values of the ester prodrugs
covered the range 1Á3±50Á3 (Table 3), which sug-
gested that all the compounds had the potential to

HYDROXYPYRIDINONE PRODRUGS
561
Table 4. 1H-NMR spectra and elemental analysis of 11a±g hydrochloride.
Compound
1H-NMR (DMSO-d6) d ppm
Elemental analysis (%)
C
H
N
11a
11b
1Á2 (t, 3H, 2-CH2CH3), 2Á9 (q, 2H, 2-CH2CH3), 4Á6 (br., s, 4H,
NCH2CH2O), 6Á5±7Á6 (m, 5H, Ar & 5-H(pyridinone)), 8Á1 (d,
1H, 6-H(pyridinone)), 7Á5±10Á0 (br., 3H, OH)
Calcd (%)
Found (%)
56Á56
56Á39
5Á34
5Á10
4Á12
4Á02
1Á12 (t, 3H, 2-CH2CH3), 2Á96 (q, 2H, 2-CH2CH3), 4Á6 (br., s, 4H,
NCH2CH2O), 6Á85 (d, 2H, Ar 1H, H-meta to the carboxyl group),
7Á35 (d, 1H, 5-H(pyridinone)), 7Á7 (d, 2H, Ar, H-ortho to the
carboxyl group), 8Á3 (d, 1H, 6-H(pyridinone)), 8Á5±10Á2 (br.,
3H, OH)
Calcd (%)
Found (%)
56Á56
56Á43
5Á34
5Á21
4Á12
3Á96
11c
11d
0Á95 (t, 3H, 2-CH2CH3), 2Á8 (q, 2H, 2-CH2CH3), 4Á5 (br., s, 4H,
NCH2CH2O), 7Á10±7Á95 (m, 6H, Ar & 5-H(pyridinone)), 8Á2 (d,
1H, 6-H(pyridinone)), 7Á9±9Á6 (br., 2H,OH)
Calcd (%)
Found (%)
59Á35
59Á58
5Á60
5Á47
4Á33
4Á25
1Á15 (t, 3H, 2-CH2CH3), 3Á03 (q, 2H, 2-CH2CH3), 4Á77 (br., s,
NCH2CH2O), 7Á5 (d, 1H, 5-H(pyridinone)), 7Á6±8Á3 (m, 3H, 3,4,5-
H(pyridine)), 8Á45 (d, 1H, 6-H(pyridinone)), 8Á75 (d, 1H,
6-H(pyridine))
Calcd (%)
Found (%)
49Á88
49Á63
5Á02
4Á95
7Á76
7Á85
11e
11f
11g
1Á2 (t, 3H, 2-CH2CH3), 3Á0 (q, 2H, 2-CH2CH3), 4Á7 (br., s, 4H,
NCH2CH2O), 7Á3 (d, 1H, 5-H(pyridinone)), 7Á6±8Á0 (m, 1H,
6-H(pyridinone)), 5-H(pyridine)), 8Á1±8Á65 (m, 2H,4-H(pyridine)
8Á8(d, 1H, 6-H(pyridine)), 9Á0 (s, 1H, 5-H,2-H(pyridine))
Calcd (%)
Found (%)
49Á88
50Á15
5Á02
5Á09
7Á76
7Á69
1Á15 (t, 3H, 2-CH2CH3), 2Á95 (q, 2H, 2-CH2CH3), 4Á7 (br., s, 4H,
NCH2CH2O), 7Á3 (d, 1H, 5-H(pyridinone)), 7Á95 (d, 2H, 3,5-H(pyri-
dine)), 8Á25 (d, 1H, 6-H(pyridinone)), 8Á8 (d, 2H, 2,6-H(pyridine)),
9Á9 (s, br., 3H, OH & NH)
Calcd (%)
Found (%)
49Á88
49Á79
5Á02
5Á24
7Á76
7Á57
1Á2 (t, 3H, 2-CH2CH3), 2Á7 (s, 3H, 6-CH3), 3Á05 (q, 2H, 2-none)),
CH2CH3), 4Á8 (br., s, 4H, NCH2CH2O), 7Á5 (d, 1H, 5-H(pyridinone-
7Á8 (d, 1H, 5-H(pyridine)), 8Á3±8Á85 (m, 2H, 4-H(pyridine) & 6-H
(pyridinone)), 9Á05 (s, 1H, 2-H(pyridine))
Calcd (%)
Found (%)
51Á21
51Á21
5Á37
5Á24
7Á47
7Á35
Figure 5. Proposed intramolecular hydrogen bonding of 11a.
isonicotinic ester derivatives, 11d, 11e and 11f, had
much lower partition coef®cients than the benzoyl
ester derivatives.
Ester hydrolysis studies
The stability of the pivaloyl ester 7 was compared
with the aromatic esters (11a±11g) (Table 5). The
stability of these esters at both pH 7Á4 and pH 2Á0
Figure 4. UV spectra for the determination of the distribution
coef®cient (pH 7Á4) of 11f. Spectra 1±6 are of the aqueous
phase and result from the sequential addition of 5-mL aliquots
of 1-octanol.
varied quite appreciably, the range of values being
0Á2±105Á2 nmol mL 1 h
(pH 2Á0) and 0Á1±
1

562
ZU DONG LIU ET AL
Table 5. Hydrolysis of CP102 ester prodrugs at different pH values, in rat blood and in rat liver homogenate.
Compound
Rate of hydrolysis
1
)
1
)
1
1
pH 2Á0 (nmol mL 1 h
pH 7Á4 (nmol mL 1 h
Rat blood (nmol mL 1 h
)
Rat liver (nmol g 1 h
)
7
5Á0Æ 1Á1
<0Á1
185Á6Æ 27Á2
16Á8Æ 0Á8
27 600Æ 700
7400Æ 3600
11a
11b
11c
11d
11e
11f
11g
0Á2Æ 0Á1
17Á3Æ 1Á4
<0Á1
0Á5Æ 0Á1
38Á4Æ 0Á8
15 900Æ 2100
180 400Æ 5700
432 800Æ 2900
43 400Æ 900
1Á9Æ 0Á5
0Á8Æ 0Á3
57Á5Æ 0Á5
28Á5Æ 1Á0
97Á9Æ 1Á2
16Á7Æ 0Á2
591Á2Æ 16Á0
3541Á6Æ 93Á6
794Á4Æ 12Á0
1313Á6Æ 182Á4
563Á2Æ 62Á4
97Á7Æ 5Á6
30Á4Æ 4Á6
105Á2Æ 15Á0
17Á1Æ 1Á0
95 200Æ 8200
153 500Æ 2200
Values are expressed as meansÆ s.d. (n ˆ 3).
1
97Á9 nmol mL 1 h (pH7Á4). In particular, esters
containing a nitrogen atom in the aromatic moiety
(i.e. 11d±f) were found to be less stable than the
corresponding benzoyl analogue (11c). All esters
were hydrolysed more rapidly in rat blood, activ-
Table 6. Iron mobilization studies of CP102 ester prodrugs
and the parent compound in the 59Fe-ferritin loaded rat model.
Chelator
% Faecal iron excretion
Control
1
CP102
7
11a
11b
11c
11d
11e
11f
3Á87Æ 1Á0
13Á2Æ 5Á2
16Á8Æ 2Á7
23Á0Æ 7Á3
26Á3Æ 6Á2
15Á1Æ 4Á9
13Á5Æ 6Á8
28Á4Æ 3Á5
27Á6Æ 8Á8
13Á4Æ 3Á7
22Á2Æ 2Á6
ities
falling
in
the
range
16Á8±
3541Á6 nmol mL 1 h . However, some esters,
such as 11a and 11b, possessed appreciable stabi-
lity in rat whole blood, with plasma esterase
activities of 16Á8 and 38Á4 nmol mL h , respec-
tively. The nitrogen-containing esters (e.g. 11d±f)
demonstrated higher hydrolysis rates when com-
pared with the other esters. Without exception, the
rate of hydrolysis was found to be much greater in
1
1
1
11g
All chelators were given orally (450 mmol kg 1). Values are
expressed as meansÆ s.d. (n ˆ 5).
the presence of liver homogenate, activities ranging
from 7400 nmol (g tissue) 1 h
(11a) to
1
432 800 nmol (g tissue) 1 h (11d). Generally the
hydrolytic rates in the presence of liver homo-
genate were several-hundred-fold faster than those
observed in the rat blood.
1
compounds at removing iron from the liver were 7,
11a, 11d and 11e despite the hydrolysis rate being
spread over the range 7400±432 800 nmol
1
(g tissue) 1 h
(Table 5). Clearly as long as
hydrolysis occurred at an appreciable rate, then iron
chelation occurred with acceptable ef®cacy. In a
similar fashion D7Á4 values were not critical, 11a
being hydrophobic (D7Á4 ˆ 50Á3) and 11c being
relatively hydrophilic (D7Á4 ˆ 1Á8).
Ef®cacy study
The iron mobilisation ef®cacy of ester prodrugs was
compared with the parent compound (CP102) and
the 1,2-dimethyl derivative 1 (Table 6). All the
ester prodrugs were found to be superior to 1.
Several, such as 11a, 11d and 11e, with D7Á4 values
of 50Á3, 1Á3 and 1Á8, respectively, provided a sig-
ni®cant improvement over CP102 (D7Á4 ˆ 0Á22) in
the ef®cacy model. The optimal effect was
observed with the picolinic ester derivative 11d,
which provided 70% improvement in the ability of
facilitating iron excretion over that of the parent
drug (CP102). However, not all the prodrugs were
associated with an increased ef®cacy; 11b, 11c and
11f were less effective than the parent compound
(CP102), suggesting that lipophilicity was not the
only factor which in¯uenced drug ef®cacy. There
was no correlation between ef®cacy and ester
hydrolysis in the liver. Thus, the four most effective
Discussion
In order to obtain selective delivery of the chelator
CP102 to the major chelatable iron pool in the body ±
the liver, it is essential that prodrug esters are
absorbedintact from the gastrointestinaltract, and are
reasonably stable in the plasma, thereby providing
suf®cient time for the prodrug to perfuse the liver.
Subsequently, rapid hydrolysis by hepatic carboxy-
esterases will generate hydrophilic metabolites
within hepatocytes. Hydrolysis studies conducted at
selected pH values provide useful information relat-
ing to the likely in-vivo stability of the corresponding
compounds at different sites in the gastrointestinal

HYDROXYPYRIDINONE PRODRUGS
563
tract. Although the stability of these aromatic esters
was found to vary quite appreciably, they were found
to be relatively stable in aqueous solution at the pH
values of 2Á0 and 7Á4 (Table 5). That the esters with
heteroaromatic acid moieties (e.g. 11d±f) are less
stable than the benzoyl analogue is probably due to
the aromatic nitrogen exerting a negative inductive
effect and thereby rendering the ester bond more
susceptible to hydrolysis.
Ideally, ester prodrugs which are designed to
selectively deliver CP102 to the liver, should
undergo minimal pre-systemic hydrolysis in the
plasma relative to the rate of hydrolysis in the liver.
This would ensure that the ester reaches the liver
intact where extensive intracellular hydrolysis will
occur to produce the more hydrophilic CP102.
Extensive extracellular hydrolysis will reduce the
effectiveness of the hydrophobic ester by virtue of
the poor extraction of the hydrophilic compound by
the liver. However, it is also essential that there is
some esterase activity within the systemic circula-
tion in order to ensure that the small quantities of
ester which are not extracted by the liver during the
®rst pass are hydrolysed to the hydrophilic meta-
bolite. The ratio between the hydrolytic rates of the
esters in the presence of liver homogenate and that
in the rat blood varies from 50 to 450.
In-vivo iron mobilisation ef®cacies of the ester
prodrugs were investigated with 59Fe-ferritin loa-
ded rats. In this model the chelator was adminis-
tered at a time when 59Fe released by lysosomal
degradation of ferritin was maximally available
(Pippard et al 1981; Liu et al 1999). Since a major
reason for developing ester prodrugs of hydroxy-
pyridinones is to target the liver iron pool, the 59Fe-
ferritin=rat model was considered to be an ideal in-
vivo test system. Preliminary analysis of the results
shows that in many cases the ester prodrug leads to
superior iron excretion via the bile than does the
corresponding alcohol (CP102) (Table 6). Several
aromatic esters, for instance 11a, 11d and 11e were
found to be even more effective than the pivaloyl
ester analogue (7). Such enhancement is almost
certainly due to the increased lipophilicity of these
ester prodrugs as compared with the parent com-
pound, which
benzoyl analogue (11b), which suggests that the
position of the substitution on the aromatic ring has
a dominating in¯uence on ef®cacy. The compar-
ison between 11e and 11f is particularly signi®cant
as, although their log P values are identical (Table
3), their ef®cacies are markedly different (Table 6).
This con®rms that lipophilicity is not the only
factor which in¯uences the drug ef®cacy. Further-
more, no clear correlation between the hydrolytic
rates and ef®cacy of iron removal could be identi-
®ed for this series of compounds, indicating that
other factors are important for controlling their in-
vivo disposition. A likely additional factor is pro-
drug metabolism. Although CP102 does not
undergo extensive glucuronic conjugation (Singh et
al 1992b, 1996 ), the more lipophilic ester deriva-
tives may do so. Although the esters are predicted
to be rapidly hydrolysed in the liver, a competition
may exist between ester hydrolysis and phase II
metabolism. Detailed metabolism studies need to
be undertaken in order to verify this hypothesis.
In summary, the selective delivery of CP102 to
target organs such as the liver can be achieved
using aromatic ester prodrugs. Although several
aromatic ester prodrugs provide a signi®cant
improvement over the parent hydroxyalkyl com-
pound, the selection of ester moieties is critical.
The nature and position of the substituent on the
aromatic ring has an appreciable in¯uence on the
ef®cacy of the compound.
Acknowledgements
The research described in this manuscript was
supported by Apotex Research Inc., Canada and
Biomed EC grant BMH4-CT97-2149. Z. D. Liu
would like to thank the KC Wong Education
Foundation (Hong Kong) and the Committee of
Vice-Chancellors and Principals of the Universities
of the United Kingdom for the joint scholarships
which enabled him to carry out the studies.
References
Ansel, L. P., Levy, G. (1969) Drug biotransformation interac-
tions in man II. A pharmacokinetics study of the simulta-
neous conjugation of benzoic and salicylic acids with
glycine. J. Pharm. Sci. 58: 321±326
Batshaw, M. L., Brusilow, S. W., Blom, W., Brubakk, A. M.,
Burton, B. K., Cann, H. M., Kerr, D., Mamuners, P.,
Matalon, R., Myerberg, D., Shafer, I. A. (1982) Treatment
of inborn errors of urea synthesis. N. Engl. J. Med. 306:
1387±1392
Borshop, B. A., Breuer, J., Holm, J., Leslie, J., Nyhan, W. L.
(1989) Excretion of hippuric acid during sodium benzoate
therapy in patient with hyperglycinaemia. J. Inher. Dis. 12:
72±79
indicates that by analogy with 7, selective delivery
of the drug to the liver has been achieved. Phar-
macokinetic studies have demonstrated that 7 is
rapidly cleared by the liver and undergoes exten-
sive hydrolysis to generate the hydrophilic meta-
bolite CP102 (Choudhury et al 1997).
Some interesting features emerge from the
structure±activity studies, for instance the ef®cacy
of the para-hydroxybenzoyl ester derivative of
CP102 (11a), is twice that of the ortho-hydroxy-
Brusilow, S. W., Valle, D. L., Batshaw, M. L. (1979) New
pathways of nitrogen excretion in inborn errors of urea
synthesis. Lancet ii: 452±454

564
ZU DONG LIU ET AL
Choudhury, R. (1995) Metabolism and Pharmacokinetics of 1-
Hydroxyalkyl-3-hydroxypyridin-4-one Chelating Agents.
Ph.D. Thesis, King's College, London
Porter, J. B., Morgan, J., Hoyes, K. P., Burke, L. C., Huehns, E.
R., Hider, R. C. (1990a) Relative oral ef®cacy and acute
toxicity of hydroxypyridin-4-one iron chelators in mice.
Blood 76: 2389±2396
Porter, J. B., Huehns, E. R., Hider, R. C. (1990b) Update on
the hydroxypyridinone oral-iron chelating agents. Seminars
in Haematology 27: 95±100
Porter, J. B., Abeysinghe, R. D., Hoyes, K. P., Barra, C.,
Huehns, E. R., Brooks, P.N., Blackwell, M. P., Araneta, M.,
Britenham, G., Singh, S., Dobbin, P., Hider, R. C. (1993)
Contrasting interspecies ef®cacy and toxicology of 1,2-
diethyl-3-hydroxypyridin-4-one CP94, relates to differing
metabolism of the iron chelating site. Br. J. Haematol. 85:
159±168
Porter, J. B., Singh, S., Katherine, P. H., Epemolu, O.,
Abeysinghe, R. D., Hider, R. C. (1994) Lessons from
preclinical and clinical studies with 1,2-diethyl-3-hydroxy-
pyridin-4-one, CP94 and related compounds. Adv. Exp.
Med. Biol. 356: 361±370
Rai, B. L., Dekhordi, L. S., Khodr, H., Jin, Y., Liu, Z., Hider,
R. C. (1998) Synthesis, physicochemical properties and
evaluation of N-substituted-2-alkyl-3-hydroxy-4(1H)-pyri-
dinones. J. Med. Chem. 41: 3347±3359
Rai, B. L., Liu, Z. D., Liu, D. Y., Lu, S. L., Hider, R. C. (1999)
Synthesis, physicochemical properties and biological eva-
luation of ester prodrugs of 3-hydroxypyridin-4-ones: design
of orally active chelators with clinical potential. Eur. J. Med.
Chem. In press
Choudhury, R., Epemolu, R. O., Rai, B. L., Hider, R. C.,
Singh, S. (1997) Metabolism and pharmacokinetics of 1-(20-
trimethylacetoxyethyl)-2-ethyl-3-hydroxypyridin-4-one
(CP117) in the rat. Drug Metab. Disposit. 25: 332±339
Holme, E., Greter, J., Jacobson, C. E., Nordin, I., Lindstedt, S.,
Kristiansson, B. (1989) Carnitine de®ciency induced by
pivampicillin and pivampicillinam therapy. Lancet ii:
469±472
Holme, E., Jodal, U., Linstedt, S., Nordin, I. (1992) Effects of
pivalic acid-containing prodrugs on carnitine homeostasis
and on response to fasting in children. Scand. J. Clin. Invest.
52: 361±372
Jin, Y., Singh, S., Epemolu, R. O., Porter, J. B., Hider, R. C.,
Nick, H. P., Bergeron, R. J., Wiegand, J., Schnebli, H. P.
(1993) CP102; a potent orally active iron chelator. Proceed-
ings of the European Iron Club Meeting, London
Liu, Z. D., Lu, S. L., Hider, R. C. (1999) In-vivo iron
mobilisation evaluation of hydroxypyridinones in 59Fe-fer-
ritin loaded rat model. Biochem. Pharmacol. 57: 559±566
Melegh, B., Kerner, J., Bieber, L. L. (1987) Pivampicillin-
promoted excretion of pivaloylcarnitine in humans. Bio-
chem. Pharmacol. 36: 3405±3409
Mitsunobu, O. (1981) The use of diethylazodicarboxylate and
triphenylphosphine in synthesis and transformation of nature
products. Synthesis 1: 1±28
Oliveri, N. F., Koren, G., Matsui, D., Liu, P. P., Blendis, L.,
Cameron, R., McClelland, R. A., Templeton, D. M. (1992)
Reduction in tissue iron stores and normalisation of serum
ferritin during treatment with the oral iron chelator L1 in
thalassaemia intermedia. Blood 79: 2741±2748
Singh, S., Epemolu, O., Dobbin, P. S. Tilbrook, G. S., Ellis, B.
L., Damani, L. A. Hider, R. C. (1992a) Urinary metabolic
pro®les in man and rat of 1,2-dimethyl- and 1,2-diethyl
substituted 3-hydroxypyridin-4-ones. Drug Metab. Disposit.
20: 256±261
Pippard, M. J., Callender, S. T., Letsky, E. A., Weatherall, D.
J. (1978a) Prevention of iron loading in transfusion-depen-
dent thalassaemia. Lancet i: 1178±1180
Pippard, M. J., Callender, S. T., Weatherall, D. J. (1978b)
Intensive iron-chelation therapy with desferrioxamine in
iron-loading anaemias. Clin. Sci. Mol. Med. 54: 99±106
Pippard, M. J., Johnson, D. K., Finch, C. A. (1981) A rapid
assay for evaluation of iron chelating agents in rats. Blood
58: 685±692
Poon, K., Pang, K. S. (1995) Benzoic acid glycine conjugation
in the isolated perfused rat kidney. Drug Metab. Disposit.
23: 255±260
Porter, J. B., Huehns, E. R., Hider, R. C. (1989) The develop-
ment of iron chelating drugs. Bailliere's Clin. Haematol. 2:
257±292
Singh, S., Epemolu, R. O., Ackerman, R., Porter, J. B.,
Hider, R. C. (1992b) Development of 3-hydroxypyridin-4-
ones which do not undergo extensive phase II metabolism.
3rd NIH-Sponsored Symposium on The Development of
Iron Chelators for Clinical Use, Gainsville, FL, Abstract
No. 52
Singh, S., Choudhury, R., Epemolu, R. O., Hider, R. C. (1996)
Metabolism and pharmacokinetics of 1-(20-hydroxyethyl)-
and 1-(30-hydroxypropyl)-2-ethyl-3-hydroxypyridin-4-ones
in the rat. Eur. J. Drug Metab. Pharmacokinet. 21: 33±41
Tilbrook, G. S., Hider, R. C. (1998) Iron chelators for clinical
use. In: Sigel, A., Sigel, H. (eds) Metal Irons in Biological
Systems. Vol. 35. Iron Transport and Storage in Microor-
ganisms, Plants and Animals. Marcel Dekker, New York, pp
691±730