Design, synthesis, and evaluation of novel bifunctional iron-chelators as potential agents for neuroprotection in Alzheimer's, Parkinson's, and other neurodegenerative diseases

Article abstract of DOI:10.1016/j.bmc.2004.10.037

Several novel antioxidant-iron chelators bearing 8-hydroxyoxyquinoline moiety were synthesized, and various properties related to their iron chelation, and neuroprotective action were investigated. All the chelators exhibited strong iron(III) chelating and high antioxidant properties. Chelator 9 (HLA20), having good permeability into K562 cells and moderate selective MAO-B inhibitory activity (IC₅₀ 110 μM), displayed the hightest protective effects against differentiated P19 cell death induced by 6-hydroxydopamine. EPR studies suggested that Chelator 9 also act as radical scavenger to directly scavenge hydroxyl radical.

Full text of DOI:10.1016/j.bmc.2004.10.037

Bioorganic & Medicinal Chemistry 13 (2005) 773–783
Design, synthesis, and evaluation of novel bifunctional
iron-chelators as potential agents for neuroprotection in
AlzheimerÕs, ParkinsonÕs, and other neurodegenerative diseases
Hailin Zheng,a Lev M. Weiner,a Orit Bar-Am,b Silvina Epsztejn,c Z. Ioav Cabantchik,c
Abraham Warshawsky,a Moussa B. H. Youdimb and Mati Fridkina,*
aDepartment of Organic Chemistry and Neurobiology, The Weizmann Institute of Science, Rehovot 76100, Israel
bEve Topf and USA National Parkinson Foundation Centers of Excellence for Neurodegenerative Diseases and
Department of Pharmacology, Technion—Faculty of Medicine, Haifa, Israel
cDepartment of Biological Chemistry, Institute of Life Sciences, Hebrew University of Jerusalem, Jerusalem 91904, Israel
Received 25 August 2004; revised 18 October 2004; accepted 18 October 2004
Available online 5 November 2004
Abstract—Several novel antioxidant-iron chelators bearing 8-hydroxyoxyquinoline moiety were synthesized, and various properties
related to their iron chelation, and neuroprotective action were investigated. All the chelators exhibited strong iron(III) chelating
and high antioxidant properties. Chelator 9 (HLA20), having good permeability into K562 cells and moderate selective MAO-B
inhibitory activity (IC50 110lM), displayed the hightest protective effects against differentiated P19 cell death induced by 6-hydroxy-
dopamine. EPR studies suggested that Chelator 9 also act as radical scavenger to directly scavenge hydroxyl radical.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
consist mainly of L-dopa and/or dopamine (DA) agon-
ists, monoamine oxidase B inhibitors such as rasagiline
and selgiline, catechol-methyl transferase inhibitor, enta-
capone.2 However, these drugs can only improve clini-
cal symptoms (symptomatic) but cannot mitigate
progression of the disease process underlying PD. For
AD the only therapeutically active drugs are cholinest-
erase inhibitors3 and glutamate antagonist, memantine.4
Moreover, severe side effects will appear in the long-
term treatment.2 Therefore, other approaches are criti-
cally needed for these progressive neurodegenerative
diseases.
ParkinsonÕs disease (PD) and AlzheimerÕs disease (AD)
are the most common neurodegenerative disorders.
They affect at least 5% of the population above the
age of 65years.1 At present, drugs used for PD therapy
Abbreviations: L1, 1,2-dimethyl-3-hydroxypyridin-4-one; DFO, des-
feral; Boc, tert-butoxycarbonyl; DMSO, dimethyl sulfoxide; Fmoc, 9-
fluorenylmethoxycarbonyl; Boc2O, di-tert-butyl dicarbonate; TFA,
trifluoroacetic acid; Trt, trityl; DMF, dimethylformamide; TV3326,
N-propargyl-3R-aminoindan-5yl-ethyl methylcarbamate; TV3279,
N-propargyl-3S-aminoindan-5yl-ethyl
methylcarbamate;
MAO,
Although the etiology of neurodegenerative diseases
(ParkinsonÕs disease, AlzheimerÕs disease, Huntington
disease, amyotrophic lateral sclerosis, Friedreich ataxia)
is not yet well established, accumulating evidences have
shown that iron-dependent oxidative stress, increased
levels of iron and monoamine oxidase (MAO)-B activ-
ity, and depletion of antioxidants in the brain may
be major pathogenic factors in PD and other neurode-
generative diseases.1,2 In fact, a number of iron chela-
tors, antioxidant or MAO-B inhibitors have been
shown to possess neuroprotective activity in animal
models. For example, deferoxamine (DFO), a prototype
monoamine oxidase; MPTP, 1-methyl-4-phenyl-1,2,5,6-tetrahydro-
pyridinium; 6-OHDA, 6-hydroxydopamine; ICV, intraventricu-
larly; SIH, salycylaldehyde isonicitinoyl hydrazone; ROS, reactive
oxygen species; EGCG, (ꢀ)-epigallocatechin-3-gallate; EPR, elec-
tron paramagnetic resonance; NMR, nuclear magnetic resonance;
MTT, 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyl tetrazolium bro-
mide; ELISA, enzyme-linked immunosorbent assay; HPLC, high
performance liquid chromatography; IC50, the concentration req-
uired for 50% inhibition; THF, tetrahydrofuran; tBu, tert-butyl.
Keywords: Antioxidant-iron chelator; Lipid peroxidation; Neuropro-
tection; 8-Hydroxyquinoline.
*
Corresponding author. Tel.: +972 08 934 2505; fax: +972 08 934
4142; e-mail: mati.fridkin@weizmann.ac.il
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.10.037

774
H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
iron chelator drug that does not cross the blood brain
barrier, when injected intraventricularly (ICV), protects
against the dopaminergic neurodegeneration induced by
6-hydroxydopamine (6-OHDA).5 Desferal also has neu-
roprotective activity in preventing iron and 1-methyl-4-
phenyl-1,2,5,6-tetrahydropyridinium (MPTP)-induced
neurotoxicity in mice.6 More recently, the antibiotic iron
chelator, 5-chloro-7-iodo-8-hydroxyquinoline (clioqui-
nol) has been shown to be able to prevent MPTP-in-
duced neurotoxicity in vivo.7 Oral administration of
clioquinol was reported to inhibit b-amyloid accumula-
tion in an AlzheimerÕs disease (AD) transgenic mouse
model via its actions as a metal chelator.8 Antioxidants
such as vitamin E and ebselen when injected into rats
and mice prevent the neurotoxic action of 6-OHDA
and MPTP.9,10 The green tea catechin, (ꢀ)-epigallacto-
catechin-3-gallate (EGCG), which is known for its iron
chelating and antioxidant properties, is one of the most
potent neuroprotective agents against MPTP neurotoxi-
city in mice.11 In animal models, L-deprenyl, as well as
other MAO-B inhibitors, such as rasagiline, has been
shown to protect against the toxic damage induced by
(MPTP) and 6-OHDA.2 However, none of these anti-
oxidants and iron chelators have been successfully intro-
duced for clinical use.2,7,9
neuroprotective activities of rasagiline and TVP-1022
and other propaergylamines.18 A piperazine moiety
was selected as a spacer since it contains two amino moie-
ties available for convenient chemical modifications.
Moreover, due to its hydrophobicity, it may enhance
penetration through the BBB.
In this paper, we reported the synthesis and characteri-
zation of these novel antioxidant chelators.
2. Results and discussion
2.1. Chemical synthesis
The preparation of chelators 2, 3, and 4 is outlined
in Scheme 1, starting from commercially available
8-hydroxyquinoline. According to standard methods,
8-hydroxyquinoline was converted to 5-chloromethyl-
8-quinoline 1 by treating with hydrochloric acid and
formaldehyde. Reaction of 1 with N-substituted
piperazines or propargylamine yielded the target com-
pounds 2, 3, or 4.
The synthesis of 9 (HLA20) is shown in Scheme 2. Pip-
erazine 5 was protected to afford its N-Boc derivative 6.
Compound 6 reacted with propargyl bromide to obtain
N-t-Boc propargylpiperazine 7. Deprotection of 7 using
trifluroacetic acid yielded propargylpiperazine 8. Final-
ly, 8 was alkylated with 5-chloromethyl-8-quinoline 1
to give propargyl hydroxyquinoline 9.
Recently a new neuroprotective strategy has been pro-
posed, that is, neuroprotection in neurodegenerative dis-
eases may require a drug combining iron chelating with
antioxidant capacity, iron chelator, and MAO-B inhibi-
tory properties.2
The goal of this study was to develop novel antioxidant-
iron chelators with potent selective MAO-B inhibitory
activity, which can be used as drug candidates for the
treatment and/or prevention of the neurodegenerative
diseases. For this purpose, several novel antioxidant
chelators with 8-hydroxyoxyquinoline moiety were de-
signed, synthesized, and characterized. Here we selected
8-hydroxyoxyquinoline as a chelating moiety based on
the following reasons: first, 8-hydroxyquinoline is a
strong iron chelator with antioxidant property.12,13 Sec-
ondly, it can protect plants against oxidative damage
caused by herbicide paraquat via blocking the Fenton
reaction.14 Moreover, 8-hydroxyquinoline has been
shown to be able to cross the blood brain barrier
(BBB).15 And more recently, as already mentioned, the
activity of the antibiotic iron chelator, 5-chloro-7-iodo-
8-hydroxyquinoline (clioquinol) in preventing MPTP
neurotoxicity in mice has been reported.7 We also intro-
duced a propargyl moiety and/or piperazine moiety (as a
linker) to these new chelators. The propargyl moiety is
believed to be responsible for the potent MAO B inhib-
itory activity since N-propargyl-containing compounds
such as rasagiline, pargyline, clorgyline, and deprenyl
show high MAO inhibitory activity; removal of the
propargyl moiety abolishes the inhibitory activity.16,28,29
This moiety is also responsible for neuroprotective activ-
ity unrelated to MAO inhibition, since the S-isomer of
rasagiline, TVP-1022, is an extremely weak inhibitor of
MAO, but has similar neuroprotective activity in vitro
and in vivo to rasagiline.17 Furthermore recent studies
have shown that propargylamine itself possesses the
Compounds 11 and 14 were synthesized by the pathways
depicted in Scheme 3. Starting from quinolinyl alanine
ethyl ester 10.19 Compound 11 was obtained by reacting
ethyl ester 10 with propargylamine in DMSO with
sodium bicarbonate as base. To obtain compound 14,
we started from commercially available L-cysteine ethyl
ester, which was first protected as its S-Trt derivative 12.
Compound 12 was then treated with 5-chloro-methyl-
8-quinoline 1 to give intermediate 13. Removal of the
trityl protecting group in 13 by trifluroacetic acid
afforded target compound 14.
Cl
N
N R
b
a
N
N
N
OH
OH
OH
8-hydroxyquinoline
1
2: R = CH2CH2OH
3: R = COOEt
c
N
4
N
OH
Scheme 1. Reagents and conditions: (a) HCl (32% water), HCHO
(37% in water), 0ꢁC ! rt, 8h; (b) N-substituted piperazine, (Me2CH)2-
NEt, CHCl3, rt, 1day; (c) propargylamine, (Me2CH)2NEt, CHCl3, rt,
1day.

H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
775
Boc
N
Boc
N
1.5
1.3
1.1
0.9
0.7
0.5
0.3
0.1
-0.1
H
N
a
b
N
H
N
H
N
6
7
5
OH
H
N
N
c
d
b
N
N
N
8
9
a
Scheme 2. Reagents and conditions: (a) Boc2O, MeOH, 0ꢁC ! rt,
1day; (b) propargyl bromide, (Me2CH)2NEt, 0ꢁC ! rt, 1day, (c)
F3CCOOH, 0ꢁC! rt, overnight; (d) 5-chloromethyl-8-quinoline 2,
CHCl3, (Me2CH)2NEt, rt, 1day.
300
400
500
600
700
Wavelength (nm)
Figure 1. Spectrophotometric detection of Fe(III) complexation with
HLA20: (a) spectrum of 0.8mM compound 9 (HLA20) in Tris buffer
pH7.4; (b) spectrum of 0.8mM compound 9 (HLA20) in Tris buffer
pH7.4 after the addition of 0.15mM Fe2(SO4)3.
COOEt
COOEt
NH2
NH
a
A
N
1.35
N
OH
OH
b
10
11
1
O
O
OEt
STrt
B
O
NH
H3N
HS
H3N
OEt
0.65
b
c
OEt
STrt
N
c
L-Cysteine ethyl ester
0.3
OH
12
a
13
O
OEt
SH
-0.05
NH
250
350
450
550
650
d
Wavelengh (nm)
N
Figure 2. Spectrophotometric detection of the formation of Fe(III)–9
(HLA20) complex from Cu(II)–9 (HLA20) complex: (a) spectrum of
0.6mM compound 9 (HLA20) in Tris buffer pH7.4; (b) spectrum of (a)
after the addition of 0.2mM Cu(SO4)2. (c) spectrum of (b) after the
addition of 0.1mM Fe2(SO4)3.
OH
14
Scheme 3. Reagents and conditions: (a) propargyl bromide, NaHCO3,
DMSO, rt; (b) TrCl, DMF, rt; (c) 5-chloromethyl-8-quinoline 2,
CHCl3, (Me2CH)2NEt, rt; (d) CF3COOH, rt.
(HLA20) had a higher binding selectivity for iron over
copper. The fact that the new chelators have binding
capability both for iron and copper but with higher
selectivity for iron may be important factors for the anti-
oxidative-type drugs, since it is the excessive iron stores
and iron-mediated generation of free radicals in the
brain that are thought to be associated with neurodegen-
erative diseases.2,20 Therefore, the novel chelators with
these properties would be expected to chelate iron
instead of copper and hence would have potential use
as drug candidates in neurodegenerative diseases.
2.2. Iron-chelating properties
2.2.1. Complex formation with Fe(III), and Cu(II). Spec-
trophotometric study reveals the complex formation of
the new chelators with both Fe(III), and Cu(II), but with
higher selectivity for Fe(III) over Cu(II). One represent-
ative is shown in Figures 1 and 2. The appearance of a
pronounced absorption maxima around 460 and
590nm upon the addition of Fe2(SO4)3 (Fig. 1) showed
the formation of Fe(III)–9 complex. As shown in Figure
2b a new band around 375nm appeared upon the addi-
tion of CuSO4 demonstrating the formation of Cu-com-
plex. Addition of Fe2(SO4)3 resulted in disappearance of
the 375nm band and appearance of new bands 460 and
590nm (Fig. 2c), which are the characteristic maximal
absorption of Fe(III)–9 complex. This indicates that 9
2.2.2. Composition of the iron-complex. The composition
of the Fe(III)-complex of the novel chelators was estab-
lished by spectroscopic measurement and mass spectro-
metry. One example is reported here. Figure 3 indicates
that under the experimental conditions a complex with
1:3 [iron:chelator 9 (HLA20)] molar ratio was formed.

776
H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
0.5
0.4
0.3
0.2
0.1
0
60000
50000
40000
30000
20000
10000
9
2
3
L1
0
0.2
0.4
0.6
0.8
1
1.2
0
Equivalents of Iron / HLA20
0
20
40
µ
[Chelator]
M
Figure 3. Change in absorption at kmax = 590nm with [Fe(III)]/[HLA20]
molar ratio in Tris buffer (pH7.4).
Figure 5. Dequenching of CA–Fe complexes by chelators. Iron precom-
plexed calcein (CA–Fe) solutions of 1lM in Hepes 20mM buffered
saline (pH7.4 = HBS) were incubated with various concentrations of
chelators at rt for 1h. The graph shows the original fluorescence data
given in arbitrary units (a.u.) against chelator concentration.
(5.4) > 11 (6.0) > 4 (6.5) > L1 (8.4). DFO is a prototype
iron chelator used clinically for treating iron overload
with log stability constant (logb) = 31;22 L1 or 1,2-di-
methyl-3-hydroxy-4-pyridinone is an orally active iron
chelator with overall log stability constant
(logb3) = 36.22 The above data suggested that all the
tested compounds were good iron chelators with rather
similar Fe(II) binding ability. Their Fe(II) binding
capacity were less potent than that of DFO, but slightly
higher than that of L1, capable, as well, in iron binding
in 3:1M ratio,22 respectively. This affinity for iron
should be high enough to inhibit the iron redox cycle
in the biological environment. The high affinity of
DFO for iron preclude its use for prolonged periods of
time in situations unrelated to iron overload due to seri-
ous cytotoxicity.20 It has been reported that this serious
cytotoxicity of DFO is very likely attributed to its inter-
action with iron containing enzymes and protein, result-
ing in mobilizing iron from these enzymes and
proteins.23 The new chelators with moderate affinities
for iron may very well prevent interference with iron
metabolism and, therefore, could be used in the therapy
of neurodegenerative diseases in which treatment for
prolonged periods of time is required. In fact, it has been
reported7,8 that treatment with clioquinol, a 8-hydroxyl-
quinoline derivative with moderate iron binding affinity,
does not elicit any apparent adverse general health or
behavioral effects, unlike chelators currently used as
therapy for iron overload conditions, which have severe
side effects.23
Figure 4. Positive ion electrospray ionization mass spectrum of a solu-
tion of iron(III) sulfate and compound 9 (HLA20).
The mass spectra of the complex (Fig. 4) shows that the
HLA20–Fe(III) complex gave a value of m/z 935.95
corresponding to the ion [Fe(HLA20)3 + K]+, that is,
stoichiometry of 1:3. The same stoichiometry (1:3
iron:chelator) was also found for the other new chela-
tors (data not shown).
2.3. Relative binding affinities
The relative Fe(II) binding affinities of the chelators
were determined by a fluorescent probe calcein (CA).21
In this method, the fluorescene of CA can be quenched
by metal ions such as Fe(II), and then dequenched by
adding iron chelators. The Fe(II) binding affinity is in-
versely proportional to the half maximal dequenching
concentration [M]1/2. Figure 5 shows the fluorescence
recovery as a function of the concentration of three
new chelators. Similar profiles were also obtained for
the other chelators (data not shown). From these pro-
files, [M]1/2 for each chelator was estimated, and the
order of chelator capacity for complexing iron in solu-
tion was obtained: ([M]1/2 in lM): DFO (1.5) (not
shown) > compound 14 (4.0) > 3 (4.4) > 2 (4.8) > 9
2.4. Antioxidant properties
As shown in Table 1, all the novel chelators inhibited
lipid peroxidation with similar IC50 values (12–16lM)
comparable to that of desferal, a prototype iron chela-
tor, which does not cross the blood brain barrier. This
inhibition may be caused by iron-chelating activity of

H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
777
Table 1. Inhibitory properties on lipid peroxidation (LPO) and in vitro
monoamine oxidase (MAO) of the new chelators
(FAD) cofactor at the active site of MAO enzyme to in-
hibit it.28,29 The poor activity of compounds 9 and 11 as
compared with 4 may be due, at least partially, to the
presence of a bulky somewhat group (piperazinyl in 9,
ethyloxyacyl in 11) near the propargyl moiety. The steric
effect from a bulky group may affect the chelators bind-
ing to the enzyme. The fact that the new chelators, ex-
cept 4, do not inhibit MAO-A may be advantageous
for their use in the treatment of neurodegenerative dis-
eases since the inhibition of MAO-A activity is related
to the serious side effect, known as Ôcheese reactionÕ.30
Compd
Inhibition of LPOa
In vitro inhibitionb
IC50 (lM)
IC50 (lM)
MAO-A
MAO-B
2
13
14 1.5
2
>200
>200
<0.1
>200
>200
<0.1
3
4
16
12
15
3
2
3
9
>200
>200
>200
>1000
110 10
>200
>200
11
14
DFO
12 2.5
1.5
8
>1000
a Inhibition of lipid peroxidation by the new chelators in the presence
of 1.5lM FeSO4/50lM ascorbic acid. The results are the mean
SEM, n = 3, p < 0.05.
2.6. Transport properties
2.6.1. Cell permeability. The method used for assessing
the relative cell permeability of the test chelators is based
on chelator-evoked increase in the fluorescence of iron-
quenched calcein.31 Generally the rate of fluorescence
dequenching induced by addition of a chelator depends
on its cell permeability and binding affinity for iron.
However, in a biological system such as in K562 cells
(erythroleukemia cell lines, which have been extensively
studied in term of their iron-regulatory properties),31,32
chelator permeation into cells is the rate-limiting step
in the chelation of iron. Fast permeating chelators such
as salycylaldehyde isonicitinoyl hydrazone (SIH) evokes
swift increases in cell calcein fluorescence but poorly
permeable chelators such as DFO acts only after pro-
longed incubations with cells.31 Thus, the rate constant
of fluorescence recovery is considered to be proportional
to the permeation of chelators across the cell membrane.
In the present experiments, the permeation properties of
the test chelators were assessed in K562 cells. Figure 9
shows the time-dependent fluorescence recovery of
CA-loaded cells evoked by adding (at 310s) various
chelators at equivalent concentrations (5lM). As indi-
cated in Figure 9, the fluorescence dequenching rates
varied greatly with different chelators. While chelators
DFO had no effect on dequenching in the employed con-
ditions (data not shown), SIH, 4, and 9 acted quickly
with t1/2 145, 170, and 181s, respectively. The rate con-
stant of chelator permeation was determined from the
reciprocal of the half time (t1/2) of the maximum
fluorescence recovery after chelator addition. Taking
SIH as 100, the order was SIH (100 7) > 4
(85 3) > 9 (80 5) > 3 (75 8) > 11 (36 4) > 2
(27 6) > 14 (20 8) ꢁ DFO (not shown). As seen in
Figure 10, 4 and 9 had a good permeability in K562
cells, about 80% permeation into K562 as compared
with control (SIH 100%). In contrast, the other test che-
lators showed a much lower permeability [<40% perme-
ation as compared to control (100%)]. As expected,
DFO, a poorly permeant chelator, did not pass K562
cell membrane in our experimental conditions.
b In vitro inhibitory action against rat brain MAO-A and MAO-B. The
chelators were added to buffer containing 0.05lM deprenyl (MAO-
A) or 0.05lM clorgylin (MAO-B) and were incubated with the tissue
homogenate for 60min at 37ꢁC before addition of 14C-5-hydroxy-
tryptamine (MAOA) or 14C-phenylethylamine (MAO-B). MAO-A
and -B activity in presence of drug was expressed as a percentage of
that in control samples. Mean values shown SEM, n = 3, p < 0.01.
the test compounds. It has been reported that the anti-
oxidative properties of iron chelators, such as desferal,
rutin, and quercetin, are most likely due to their iron
chelation since iron chelators may form inert complexes
with iron and interfere with the Fention reaction leading
to a decrease of hydroxyl free radical production.5,6,24
The new chelators might act by similar mechanism to in-
hibit the free radical formation. However, it is also pos-
sible that the novel chelators may, as other antioxidants,
directly scavenge free radicals and show inhibitory activ-
ity. In a word, whatever the mechanism of new com-
pounds works by, their high potency in inhibiting on
lipid peroxidation make them relatively good candidates
for the treatment of neurodegenerative disease since
lipid peroxidation has been implicated in the pathogen-
esis of these disease.25,26
2.5. Monoamine oxidase (MAO) inhibitory activity
in vitro
Selective MAO-B inhibitors have great potential for
treating PD.26,27 Thus, the effects of the novel chelators
on MAO-B activity were investigated. The results (Table
1) show that there were distinct differences in MAO
inhibitory activities among the tested compounds. While
compound 4 caused a significant inhibition of both
MAO-A and -B activities with the IC50 values less than
0.1lM. Compounds 2, 3, 11, and 14 have almost no ef-
fect on MAO-A and MAO-B activity (IC50 > 200lM).
Chelator 9 exhibited more potent inhibition on MAO-
B (IC50 110 10lM) than on MAO-A (IC50 > 200lM).
The high MAO inhibitory activity of 4 may be attribu-
ted to the N-propargyl group in the molecule since
N-propargyl-containing compounds such as rasagiline,
pargyline, clorgyline, and deprenyl are potent MAO
inhibitors. Removal of the propargyl moiety from these
molecules abolishes the inhibitory activity.28,29 It is well
established that the propargyl moiety binds covalently
mole/mole with the N-5 of flavin-adenosine-dinucleotide
2.6.2. Lipophilicity. Lipophilicity is thought to be a
major determinant in the permeability of a drug. In
order to correlate the permeability of the new chelators
to their lipophilicity, we used the shake-flask method to
assess their lipophilicity in water (pH7).33 The logD
(logarithm of the n-octanol/water distribution coeffi-
cients) is the most commonly used measure of the lipo-
philicity of a drug candidate.33 The logD values of the

778
H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
be induced to differentiate into post-mitotic neuron-like
800
700
600
500
400
300
200
100
0
cells in the presence of high dose of retinoic acid.37,38 It
has been reported that 6-OHDA triggered mostly necro-
sis and less than 5% apoptosis in PC12 cells, whereas 6-
OHDA-induced death in P19 cells was apoptotic.35,36
2
3
9
11
SIH
4
Figure 8 shows protective effects of a selection of chela-
tors on differentiated P19 cells against neurotoxin 6-
hydroxydopamine. These results indicate that chelator
9 (HLA20) exhibits very good neuroprotection (61%),
and 3 also had moderate protective activity (31%). How-
ever, no neuroprotection were obtained with both 2 and
apomorphine. Even for DFO, very poor neuroprotective
action (12%) was observed. We chose these compounds
for the above experiments based on the following rea-
sons: (a) both 2 and 3 do not show appreciable inhibi-
tion on MAO-A or -B in vitro in the range of 0.1–
100lM, 2 having poor but 3 good membrane permeabil-
ity. (b) Chelator 9 has good membrane permeability and
selectively inhibits MAO B in vitro with IC50 of 110lM.
(c) DFO is a prototype iron chelator and potent radical
scavenger, and apomorphine is a highly potent radical
scavenger with iron-chelating potency.39 The good neu-
roprotection of chelator 9 in differentiated P19 cells may
be due to the combined effects of iron chelation, antioxi-
dant, MAO B inhibition, and good permeability. This
finding is consistent with the hypothesis that neuropro-
tection in neurodegenerative diseases requires a permea-
ble drug combining iron chelating with antioxidant
capacity, iron chelator, and MAO-B inhibitory proper-
ties.2 Another possible explanation might be related to
glutamate excitotoxicity. Differentiated P19 cells exhibit
N-methyl-D-aspartate (NMDA) receptor-mediated
intracellular calcium responses to glutamate.40 Studies
have shown that pretreatment of differentiated P19 cells
with NMDA antagonists such as dizocilpine, protected
against glutamate-induced cell death.41 Compound 9
-100
0
500 1000 1500 2000
Time (sec)
Figure 6. Dequenching of CA/iron complexes in K562 cells by extra-
cellularly added the test chelators. CA-loaded K562 cells were treated
with various chelators (5lM concentration) at the time indicated
(arrow). Scale, arbitrary fluorescence units (a.u.) against time (s).
120
100
80
60
40
20
0
SIH
4
9
3
11
2
14 DFO
Chelators [5µM]
Figure 7. Relative permeability of iron chelators in K562 cells. Values
were mean SEM (n = 6), given relative to those obtained with SIH,
for which 100 represents an apparent rate constant of 0.0069sꢀ1
(equivalent to a t1/2 of ingress of 145s for a 5lM chelator solution).
100
80
60
40
20
0
new chelators were as follows (logD value): compound 4
(1.92) > 9 (1.79) > 3 (1.57) > 11 (0.71) > 2 (0.54) > 14
(0.21). These data indicated that there is a good correla-
tion between the lipophilicity values and the permeabil-
ity of the new chelators, that is, the permeability is
proportional to their lipophilicity values. Chelators 4,
9, and 3 with high logD (good lipophilicity) were found
to have a good permeability in K562 cells; 11, 2, and 14
with low logD exhibited poor permeability in K562
cells. The above results suggest that chelators 4, 9, and
3 may cross the blood brain barrier due to their lipid solu-
bility, as opposed to 11, 2, and 14 with higher hydrophi-
licity (Figs. 6 and 7).
CON DFO APO
9
2
3
Compounds
Figure 8. Effects of apomorphine and iron chelators 2, 3, and 9 and
desferal on 6-OHDA-toxicity in differentiated P19 cells. Differentiated
P19 cells were treated for 6h with 50lM 6-hydroxydopamine
(6OHDA) in full medium containing either 5lM of indicated chelators
or 1lM apomorphine. The cell viability was measured with Alamar
blue assay 4h later. % Protection = the % activity relative to the system
treated with no 6OHDA. Data are means SEM, n = 3. p < 0.05,
compared with control.
2.7. In vitro neuroprotective activity
The cell death induced by 6-hydroxydopamine (6-
OHDA) either in PC12 cells or P19 cells is a good model
for studying oxidative stress associated with PD and
AD.34–36 As a nerve-cell-model substitute, we used P19
cells (mouse embryonal carcinoma cell line) that can

H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
779
(HLA20), but neither 2 nor DFO may act as NMDA
100
80
60
40
20
0
antagonist to block NMDA receptors and thus protect
against cell death. However, this hypothesis need to be
further confirmed.
2.8. Radical-scavenging property
Chelator 9, which shows very good neuroprotection in
differentiated P19 cells, was also evaluated for its radi-
cal-scavenging property by EPR. The EPR spectrum [a
quartet (1:2:2:1) with hyperfine splitting of 14.9G] in
Figure 9A is characteristic for the DMPO-ÅOH spin
adduct.42 As can seen in Figure 9B–D, the addition of
the increasing concentration of 9 to the control (Fig.
9A) reduces the intensity of the DMPO-ÅOH signal. To
0
0.2 0.4 0.6 0.8
HLA20 [mM]
1
1.2
Å
quantify the relative amount of OH present in a given
Figure 10. Concentration–response curves for inhibition at room temp-
erature and pH = 7.4 (PBS buffer containing 0.1M DMPO) of EPR
DMPO-ÅOH peaks by HLA20.
sample, the third peak of each spectrum was used for
calculation. The result (Fig. 10) demonstrated that com-
pound 9 (HLA20) reduced the DMPO-ÅOH signal in a
concentration-dependent manner. At room temperature
and the reagent concentration used in this study, the
IC50 for compound 9 (HLA20) is about 0.15mM (Fig.
10). It is known that the formation of ÅOH by photolysis
of H2O2 is iron-independent but catalyzed by light.43
Compound 9 (HLA20) remarkably decrease the inten-
sity of the DMPO-ÅOH signal produced in photolysis
system suggesting that 9 (HLA20) acts as a radical scav-
pounds radical-scavenging activity. Therefore, it is per-
haps not surprising that 9 (HLA20), due to the
presence of the phenolic group, possesses radical-scav-
enging property.
The clinical application of neuroprotective drugs pos-
sessing hydroxyquinoline moiety depends, of course,
on minimization of potential toxicity. In this regard, it
is worth noting that an old antibiotic iron chelator 5-
chloro-7-iodo-8-hydroxyquinoline (clioquinol or CQ)
was withdrawn from the market due to its possible
neurotoxicity, including nerve cell degeneration in the
hippocampus, ataxia in rats, and subacute myelo-neur-
opathy (SMON) in patients.45,46 As with most drugs,
toxicity may occur at very high doses, but for useful
agents, not within their effective therapeutic window.
More recent studies have found no evidence for CQ toxi-
city with bioeffective dosing in mice.47 It was proposed
that indiscriminate synaphigh-dose CQ use aggravated
B12 deficiency in postwar Japan and led to SMON in
a subset of vulnerable patients. More significantly, in a
phase II clinical trial, therapeutic CQ has been coadmin-
istered with B12 to Alzheimer patients with no evidence
of drug-dependent adverse events.7,47
Å
enger to directly scavenge OH. This result is consistent
with published data since, as reported before,24,44 flava-
noids, and other phenolic compounds act as radical
scavenger due to the presence of phenolic moiety. It is
the ability of the phenolic moiety to neutralize lipid radi-
cals by donating hydrogen atoms to them and form a
poorly reactive phenoxy radical that offers these com-
3. Conclusion
Several novel antioxidant-iron chelators with 8-
hydroxyquinoline moiety have been synthesized, and
their various properties related to iron chelation, anti-
oxidant, and neuroprotective activities have been inves-
tigated. Studies have shown that all the new compounds
are strong chelators for iron and copper with higher
selectivity for iron, and chelate iron(III) in a 3:1M ratio,
respectively. They have slightly higher affinity for
iron(III) than that of L1, an orally active iron chelator
with the overall log stability constant (logb3) = 36.
The new compounds inhibited Fe/ascorbate induced
mitochondrial membrane lipid peroxidation with IC50
in the range of 12–16lM, which are comparable to that
of desferal, a prototype iron chelator which does not
Figure 9. Typical EPR spectra of DMPO-ÅOH. Derived from hydroxyl
radicals generated by photolysis of a 0.6% H2O2 pH7.4 PBS buffer
containing 0.1M DMPO, with different irradiation time: A, B, and C:
40 and 46min; D 40min; E 46min. (A) Control. (B) 0.1mM 9
(HLA20). (C) 0.5mM 9 (HLA20). (D) 1.0mM 9 (HLA20). (E) 1.0mM
9 (HLA20).

780
H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
cross the blood brain barrier. Further studies selected
Chelator 9 (HLA20) as a possible lead for further devel-
opment, based on the following criteria: (a) strong iron
chelator with high free radical-scavenging capability, (b)
moderate MAO-B inhibitory activities and good perme-
ability through K562 cell membranes, and (c) effective
neuroprotection in differentiated P19 cell model. Fur-
ther experiments both in vitro and in vivo are underway
to evaluate 9 as a potential drug candidate for treating
ParkinsonÕs disease.
121.35, 124.48, 127.84, 128.88, 134.06, 138.63, 147.49,
151.76. Mass spectrometry: calcd for C16H21N3O2 m/z
[M+H]+ = 288.36, found [M+H]+ = 288.21.
4.1.3. Ethyl 4-(8-hydroxyquinolin-5-ylmethyl)-1-piper-
azinecarboxylate (3). Compound 3 was synthesized
following the procedure above for 2. (42% yield,
mp = 95–96ꢁC). 1H NMR (250MHz, CDCl3), 1.25
(dd, J = 7.1, 7.1Hz, 1H), 2.42 (s, 4H), 3.43 (s, 4H),
3.81 (s, 2H), 4.14 (dd, J = 14.21, 7.12Hz, 2H), 7.08 (d,
J = 7.72Hz, 1H), 7.31 (m, 1H), 7.47 (dd, J = 8.52,
4.20Hz, 1H), 8.66 (dd, J = 8.56, 1.58Hz, 1H), 8.79
(dd, J = 4.18, 1.54Hz, 1H). 13C NMR (100MHz,
CDCl3) 14.49, 43.47, 52.49, 60.40, 61.09, 108.73,
121.22, 123.77, 127.67, 128.92, 133.83, 138.48, 147.42,
151.91, 151.27. Mass spectrometry: calcd for C17H21-
N3O3 m/z [M+Na]+ = 238.37, found [M+Na]+ = 238.22.
4. Experimental
4.1. Chemical synthesis
All chemicals were obtained from commercial suppliers
(Aldrich, Merck, or Fluka). Proton NMR spectra were
measured on a Bruker DPX-250 spectrometer. Flash
column chromatography separations were performed
on silica gel Merck 60 (230–400 mesh ASTM). UV–vis
spectra were measured on a Hewlett-Packard 8450A
diode array spectrophotometer. TLC was performed
on E. Merck Kieselgel 60 F254 plates. Mass spectra
(DI, EI-MS) were measured on a VG-platform-II elec-
trospray single quadropole mass spectrometry (Micro
Mass, UK). Reverse-phase HPLC was performed with
a Spectra-Physics SP8800 liquid chromatography system
(Spectra-Physics, San Jose, CA) equipped with an
applied Biosystem 757 variable-wavelength absorbance
detector.
4.1.4. 5-(N-Methyl-N-propargylaminomethyl)-8-hydroxy-
quinoline (4). Chelator 4 was prepared as described
above for 2. (80%). Mp = 232–233ꢁC (hydrochloric
1
salt); H NMR (250MHz, CDCl3), 2.30 (dd, J = 2.15,
2.14Hz, 1H), 2.33 (s, 3H), 3.27 (d, J = 2.20Hz, 2H),
3.86 (s, 2H), 7.06 (d, J = 7.72Hz, 1H), 7.31 (m, 1H),
7.37 (d, J = 7.73Hz, 1H), 7.46 (dd, J = 8.52, 4.2Hz,
1H), 8.60 (dd, J = 8.52, 1.47Hz, 1H), 8.76 (dd,
J = 4.01, 1.50Hz, 1H). 13C NMR (100MHz, hydrochlo-
ric salt in D2O) 42.46, 47.54, 56.35, 74.18, 83.77, 118.22,
1198.17, 15.83, 131.63, 132.06, 138.96, 145.46, 145.97,
152.20. Mass spectrometry: calcd for C17H21N3O3 m/z
[M+Na]+ = 249.27, found [M+Na]+ = 249.23.
4.1.1. 5-Chloromethyl-8-quinolinol hydrochloride (1). A
mixture of 14.6g (0.1mol) of 8-quinolinol, 16mL of
32% HCl in water, and 16mL (0.1mL) of 37% formalde-
hyde in water at 0ꢁC was treated with hydrogen chloride
gas for 6h. The solution was allowed to stand at room
temperature for 2h without stirring. The yellow solid
obtained was collected on a filter, washed with 90%
alcohol, and dried under vacuum to give 5-chloro-
4.1.5. tert-Butyl 1-piperazinecarboxylate (6). A solution
of di-tert-butyl dicarbonate (2.93g, 12.77mmol) in
MeOH (25mL) was slowly added to a stirring solution
of piperazine (2.00g, 23.22mmol) in MeOH (50mL) at
0ꢁC. The mixture was then stirred for 2days at room
temperature, and the solvent removed in vacuum. The
crude solid was redissolved in diethyl ether (100mL)
with warming, and the white precipitate left was filtered
off. The product was extracted from the mother liquor
with 1M citric acid solution (3 · 50mL), and the aque-
ous layer was washed with Et2OAc (3 · 50mL), basified
with Na2CO3 (pH11), and extracted with Et2OAc
(3 · 50mL). The organic layer was dried over Na2SO4
and evaporated in vacuum to give tert-butyl 1-piperazine-
carboxylate 6 as a waxy white solid (crude, 1.57g, 66%),
mp = 53–54ꢁC. 1H NMR (250MHz, CDCl3) 1.42 (s,
9H), 1.89 (s, 1NH), 2.78 (m, 4H), 3.36 (m, 4H).
1
methyl-8-quinolinol hydrochloride 1 (19.0g, 98%); H
NMR (250MHz, CDCl3): 5.32 (s, 2H), 7.53 (m, 1H),
7.85 (m, H), 8.12 (m, 1H), 9.12 (m, 1H), 9.28 (m, 1H).
4.1.2. 5-(4-(2-Hydroxyethyl)piperazin-1-ylmethyl)-8-hydr-
oxyquinoline (2). To a mixture of 5-chloromethyl-8-qui-
nolinol hydrochloride
1 (2.707g, 11.8mmol) and
diisopropylethylamine (2.1mL, 20.4mmol, 2equiv) in
50mL CHCl3 at 0ꢁC was added 4-(2-hydroxyethyl)-pip-
erazine (1.7mL, 10.2mmol, 1equiv). The mixture was
stirred for 24h at room temperature. CHCl3 (100mL)
was added and the solution obtained was washed with
5% NaHCO3 (3 · 50mL), brine (2 · 50mL), and then
dried over Na2SO4. The solution was filtered and evap-
orated to dryness. The residue was crystallized from a
mixture of benzene–hexane (1:1) to yield 2 as white
solid. (2.2g, 75% yield). Mp = 127–128ꢁC; 1H NMR
(250MHz, CDCl3), 2.51 (m, 8H), 2.54 (dd, J = 5.58,
3.56Hz, 2H), 3.59 (dd, J = 5.28, 5.49Hz, 2H), 3.80 (s,
2H), 7.06 (d, J = 7.71Hz, 1H), 7.32 (m, 1H), 7.49 (dd,
J = 8.42, 4.10Hz, 1H), 8.66 (dd, J = 8.46, 1.55Hz, 1H),
8.79 (dd, J = 4.28. 1.34Hz, 1H). 13C NMR (100MHz,
CDCl3) 52.86, 52.94, 57.65, 59.20, 60.50, 108.72,
4.1.6. tert-Butyl 4-propargylpiperazine-1-carboxylate (7).
Propargyl bromide (356.9mg, 3mmol) was added slowly
to a mixture of tert-butyl 1-piperazinecarboxylate 6.
(558.8mg,
3mmol)
and
diisopropylethylamine
(407.1mg, 3.15mmol) in CHCl3 (25mL) at 0ꢁC. The
mixture was stirred for 24h at room temperature.
CHCl3 (50mL) was then added and the solution ob-
tained was washed with 5% NaHCO3 (3 · 50mL), brine
(2 · 50mL), and then dried over Na2SO4. The solution
was filtered and evaporated to dryness. The residue
was crystallized from a mixture of benzene–hexane
(1:1) and gave tert-butyl 4-propargylpiperazine-1-carb-
1
oxylate 7 (337mg, 86%). H NMR (250MHz, CDCl3):

H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
781
1.42 (s, 9H), 2.22 (s, 1H), 2.46 (s, 4H), 3.26 (s, 2H), 3.41
(s, 4H).
70.34, 84.72, 109.31, 121.58, 123.76, 127.17, 128.89,
132.57, 138.61, 147.44, 151.40, 174.96. Mass spectrome-
try: calcd for C17H18N2O3 m/z [M+H]+ = 299.34, found
[M+H]+ = 299.30.
4.1.7. N-Propargylpiperazine (8). tert-Butyl 4-propargyl-
piperazine-1-carboxylate 7 (570mg, 2.545mmol) was
dissolved in a mixture of trifluoroacetic acid (10mL)
and water (2.5mL). The solution was stirred at room
temperature overnight, and then was evaporated to
dryness in vacuum. The residue was dissolved in
water (10mL) and then based with Na2CO3 (pH11),
and extracted with Et2OAc (3 · 50mL). The organic
layer was washed with brine (2 · 50mL) and dried
over Na2SO4 overnight. Evaporation in vacuum
gave N-propargylpiperazine 8 as white solid. (193mg,
62% yield). 1H NMR (250MHz, CDCl3) 1.64 (s,
1NH), 2.26 (m, 1H), 2.55 (dd, J = 4.73, 4.50Hz, 4H),
2.93 (dd, J = 4.96, 4.84Hz, 4H), 3.29 (d, J = 2.44Hz,
2H).
4.1.10.
N-(8-Hydroxyquinolin-5-ylmethyl)-L-cysteine
ethyl ester (14). L-Cysteine hydrochloride (250mg,
1.1mmol) and trityl chloride (675mg, 2.4mmol) were
stirred in 1mL DMF for 2days at room temperature.
A 10% sodium acetate solution (9mL) was then added,
and the precipitate obtained was filtered and washed
with distilled water. The solid residue was stirred in ace-
tone at 50ꢁC for 30min and filtered after cooling to 0ꢁC,
and washed with acetone and diethyl ether. After drying
in vacuo, 500mg (86%) of 12 was obtained as a white
powder. Compound 12 (39mg, 0.1mmol) was dissolved
in DMSO (1mL), to the solution solid NaHCO3 (17mg,
0.2mmol) was added, and the mixture was stirred for
30min at room temperature. Then powdered 5-chloro-
methyl-8-quinolinol hydrochloride (19mg, 0.1mmol)
was added, and the solution was stirred at room temper-
ature for 24h. The solvent was removed by vacuum, and
the crude product was dissolved in 3mL solution of
TFA–H2O2–triethylsilane–thioanisole (85:5:5:5). After
2h, the solvent was removed in vacuo, and the crude
product was purified by preparative HPLC [VYDAC
RP-18 column (250 · 22mm, Hesperia, CA); solvent
A = water, 0.1% v/v TFA; solvent B = MeCN–
water = 3:1, 0.1% v/v TFA; tR = 15.4min (linear gradi-
ent 0–80% B over 55min)] to give the title compound
14: 21mg 60% based on 12. Mp = 117–118ꢁC. Mass
spectrometry: calcd for C15H18N2O3S m/z [M+Na]+ =
4.1.8.
quinoline (9). To a mixture of 5-chloromethyl-8-quinoli-
nol hydrochloride (323mg, 1.407mmol) and
5-(4-Propargylpiperazin-1-ylmethyl)-8-hydroxy-
2
diisopropylethylamine (0.26mL, 1.477mmol, 1.05equiv)
in 6mL CHCl3 at 0ꢁC was added N-propargylpiperazine
8 (173mg, 1.407mmol, 1equiv). The mixture was stirred
for 24h at room temperature. CHCl3 (10mL) was then
added and the solution was washed with 5% NaHCO3
(3 · 50mL), brine (2 · 50mL), and then dried over
Na2SO4. The crude product was purified by FC to give
5-(4-propargylpiperazin-1-ylmethyl)-8-hydroxyquinoline
12 as white solid. (337mg, 86% yield). Mp = 183–184ꢁC
(hydrochloric salt). Mass spectrometry: calcd for
C17H21N3O3 m/z [M+H]+ = 282.35, found [M+H]+ =
282.29. 1H NMR (250MHz, CDCl3), 2.23 (m, 1H),
2.54 (s, 8H), 3.28 (d, J = 2.44Hz, 2H), 3.80 (s, 2H),
7.08 (m, 1H), 7.36 (d, J = 10.59, 1H), 7.46 (dd,
J = 8.55, 4.18Hz, 1H), 8.67 (dd, J = 8.56, 1.55Hz, 1H),
8.78 (dd, J = 4.18. 1.51Hz, 1H); 13C NMR (100MHz,
CDCl3) 46.77, 51.98, 52.87, 60.55, 73.05, 78.92, 108.60,
121.40, 124.58, 127.87, 128.87, 134.10, 138.68, 147.50,
151.78.
1
229.38, found [M+Na]+ = 229.27. H NMR (250MHz,
D2O) 0.95 (dd, J = 7.1, 7.1Hz, 3H), 2.94 (m, 1H), 3.42
(dd, J = 14.2, 7.1Hz, 2H), 4.15 (m, 2H), 4.75 (s, 2H),
7.01 (d, J = 8.0Hz, 1H), 7.38 (d, J = 7.8Hz, 1H), 7.83
(m, 1H), 8.73 (m, 1H), 8.99 (m, 1H).
4.2. EPR spectra
The EPR spectra were recorded on E-12 (Varian)
spectrometer in a flat sealed (volume 200lL) at room
temperature (22ꢁC). The experimental conditions were
as follows: microwave power 20mW; sweep width
100G; modulation amplitude 1.0G; receive gain
4.0 · 105.
4.1.9. DL-N-Propargyl-3-(8-hydroxyquinolin-5-yl)alanine
ethyl ester (11). A mixture of NaHCO3 (17mg,
0.2mmol) and compound 10 (26mg, 0.1mmol, synthe-
sized as described before4) was dissolved in 5mL
DMSO, and the solution was stirred at room tempera-
ture for 2h. To the solution propargyl bromide
(11.9mg, 0.1mmol) was slowly added, and the solution
was stirred at room temperature for 24h. The solvent
was removed by vacuum, and the crude product was
purified by preparative HPLC [VYDAC RP-18 column
(250 · 22mm, Hesperia, CA). Solvent A = water, 0.1%
v/v TFA; solvent B = MeCN–water = 3:1, 0.1% v/v
TFA; tR = 25.4min (linear gradient 0–80% B over
55min)] to give the title compound 11: 21mg, 70% yield.
Mp = 34–35ꢁC. 1H NMR (250MHz, D2O) 0.92 (dd,
J = 28.5, 8.2Hz, 3H), 2.91 (s, 1H), 3.68 (m, 2H), 4.01
(m, 2H), 4.36 (dd, J = 7.7, 7.5Hz, 1H), 5.04 (d, J =
2.3Hz, 1H), 7.59 (d, J = 8.0Hz, 1H), 7.70 (d,
J = 8.2Hz 1H), 8.06 (dd, J = 8.7, 5.4Hz, 1H), 8.97 (d,
J = 5.4Hz, 1H), 9.13 (d, J = 8.7Hz, 1H). 13C NMR
(100MHz, CDCl3) 14.04, 31.08, 37.12, 55.53, 60.98,
4.3. Determination of distribution coefficients
Distribution coefficients of the novel chelators were
determined using the shake-flask method. 1-Octanol
(spectrophotometric grade from Aldrich Chemie,
Steinheim, Germany) and distilled deionized water were
used as the partitioning solutions. The water was shaken
with an excess of 1-octanol for presaturating and was
then allowed to stand overnight before use. 1-octanol
was also presaturated with an excess of water and
allowed to settle overnight. A 1mL of the test chelators
(0.1mM) in water (pH7) was stirred vigorously with
9mL of 1-octanol in10-mL stoppered centrifuge tubes
for 1h. The tubes were centrifuged for 20min at 1000–
2000g, and the two layers were separated and analyzed
by UV spectrophotometry.

782
H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
4.4. Lipid peroxidation assay
traces of extracellular-associated fluorescence. The sus-
pensions were placed in 96 well plates (100lL final vol-
ume) and analyzed while maintaining the system at
37ꢁC. At the indicated time, the chelator was added so
as to reach a final 5lM concentration.
Lipid peroxidation was measured in rat brain
mitochondrial homogenates as previously described.39
Lipid peroxidation was induced by 50lM ascorbic
acid and 1.5lM FeSO4 The absorption of thiobarbit-
uric acid derivatives is measured photometrically at
532nm.
4.7. Determination of neuroprotective effects on differen-
tiated P19 cells
P19 cells were obtained from the American Type Culture
Collection. They were maintained in alpha minimal
essential medium (MEM) supplemented with 10% fetal
bovine serum (FBS). Before stimulation, P19 cells were
allowed to differentiate into neurons during 4days in
complete medium containing 100nM of retinoic acid as
previously described.34 Differentiated P19 neuronal cells
grown in 96 well culture plate were treated for 6h with
50lM 6-OHDA in full medium containing either 5lM
of the indicated chelator or 1lM apomorphine. The cells
were subsequently washed and resuspended in fresh
medium containing 5% Alamar blue, incubated for 4h
and the fluorescence monitored at 450nm exc 590nm
emission (Tecan Safire fluorescence plate reader).
Data are given as % protection which is equivalent to
the % activity relative to the system treated with no
6OHDA.
4.5. Monoamine oxidase (MAO) assay
4.5.1. Preparation of brain MAO. Rats were decapitated
and the brains were quickly taken into a weighted ice-
cold sucrose buffer (10mM Tris–HCl buffer, pH7.4,
containing 0.32M sucrose), and their weights were
determined. All subsequent procedures were performed
at 0–4ꢁC. The brains were homogenized in 0.32M
sucrose (one part tissue to 20 parts sucrose) in a Teflon
glass homogenizer followed by the addition of sucrose
buffer to a final concentration of 10% homogenate.
The homogenates were centrifuged at 600g for 15min.
The supernatant fraction was taken out and centrifuged
at 4500g for 30min. The pellet was diluted in 0.32M
sucrose buffer and kept frozen for later assaying of
MAO. Protein concentration was determined with Brad-
ford reagent at 595nm, using bovine serum albumin as a
standard.
Acknowledgements
4.5.2. Determination of MAO activity. The activity of
MAO-A and MAO-B were determined by the adapted
method of Tipton and Youdim.48 Briefly, the drug under
test was added to a suitable dilution of the enzyme prep-
aration (70lg protein for MAO-B and150lg MAO-A
assay) in 0.05M phosphate buffer (pH7.4). The mixture
was incubated together with 0.05lM deprenyl (for
determination of MAO-A) or 0.05lM clorgylin (for
determination of MAO-B). Incubation was carried on
1h at 37ꢁC before addition of 14C-5-hydroxytryptamine
binoxalate (100lM) for determination of MAO-A, or
14C-phenylethylamine 100lM for determination of
MAO-B and incubation continued for 30min or
20min, respectively. The reaction was stopped with
2M ice-cold citric acid, and the metabolites were
extracted and determined by liquid-scintillation count-
ing in cpm units.
M.B.H.Y. wishes to thank National Parkinson Founda-
tion (Miami, USA), Michael F. Fox Foundation (New
York, USA), Israel Psychobiology Institute (Jerusalem,
Israel) and Technion Friedman Research Fund (Haifa,
Israel) for their generous support of this work. We are
grateful to Yotam Sagi and Shunit Gal for technical
assistant.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmc.2004.
10.037.
References and notes
4.6. Fluorescence assay
1. McDowell, I. Aging 2001, 13, 143.
2. Mandel, S.; Grunblatt, E.; Riederer, P.; Gerlach, M.;
Levites, Y.; Youdim, M. B. H. CNS Drugs 2003, 17,
729.
3. Bentue-Ferrer, D.; Tribut, O.; Polard, E.; Allain, H. CNS
Drugs 2003, 17, 947.
4. Koch, H. J.; Szecsey, A.; Haen, E. Curr. Pharm. Design
2004, 10, 253.
5. Ben-Shachar, D.; Eshel, G.; Finberg, J. P.; Youdim, M. B.
H. J. Neurochem. 1991, 56, 1441.
6. Lan, J.; Jiang, D. H. J. Neural Transm. 1997, 104, 469.
7. Kaur, D.; Yantiri, F.; Rajagopalan, S.; Kumar, J.; Mo, J.
Q., et al. Neuron 2003, 37, 899.
8. Cherny, R. A.; Atwood, C. S.; Xilinas, M. E.; Gray, D.
N.; Jones, W. D., et al. Neuron 2001, 30, 665.
9. Cadet, J. L.; Katz, M.; Jackson-Lewis, V.; Fahn, S. Brain
Res. 1989, 476, 10.
The relative binding affinity and permeability of the new
compounds were measured by fluorescence assay as pre-
viously described.31,32 For determination of relative
binding affinity, iron precomplexed calcein (Fe-CAL)
solutions of 1lM in Hepes buffered saline (20mM,
pH7.4) were incubated with various concentrations of
different chelators at room temperature. The fluores-
cence intensity (475nm > 520nm) was followed with
time in a Tecan fluorescence plate reader, using 96 well
culture plates (Nunc) at 100lL final volume. For meas-
ure of permeability, K562 cells were preloaded with calc-
ein by incubation with 0.25mM calcein-acetoxymethyl
ester for 5min at 37ꢁC, washed, and incubated in HBS
medium containing anti-calcein antibodies to quench

H. Zheng et al. / Bioorg. Med. Chem. 13 (2005) 773–783
783
10. Moussaoui, S.; Obinu, M. C.; Daniel, N.; Reibaud, M.;
Blanchard, V.; Imperato, A. Exp. Neurol. 2000, 166, 235.
11. Levites, Y.; Weinreb, O.; Maor, G.; Youdim, M. B. H.;
Mandel, S. J. Neurochem. 2001, 78, 1073.
12. Turnquist, T. D.; Sandell, E. B. Anal. Chim. Acta 1968, 42,
239.
13. Kayyali, R.; Pannala, A. S.; Khodr, H.; Hider, R. C.
Biochem. Pharmacol. 1998, 55, 1327.
14. Rogachev, I.; Kampel, V.; Gusis, V.; Cohen, N.; Gressel,
J.; Warshawasky, A. Pestici. Biochem. Phys. 1998, 60,
133.
29. Youdim, M. B. H. J. Neural. Transm. 1978, 43, 199.
30. Anderson, M. C.; Hasan, F.; McCrodden, J. M.; Tipton,
K. F. Neurochem. Res. 1993, 18, 1145.
31. Epsztejn, S.; Kakhlon, O.; Glickstein, H.; Breuer, W.;
Cabantchik, Z. I. Anal. Biochem. 1997, 248, 31–40.
32. Zanninelli, G.; Glickstein, H.; Breuer, W.; Milgram, P.;
Brissot, P.; Hider, R. C.; Konijn, A. M.; Libman, J.;
Shanzer, A.; Cabantchik, Z. I. Mol. Pharmacol. 1997, 51,
842.
33. Leo, A.; Hansch, C.; Elkins, D. Chem. Rev. 1971, 71,
525.
15. Frederickson, C. J.; Kasarskis, E. J.; Ringo, D.; Freder-
ickson, R. E. J. Neurosci. Methods 1987, 20, 91.
16. Huang, W.; Chen, Y.; Shohami, E.; Weinstock, M. Eur. J.
Pharmacol. 1999, 366, 127.
17. Youdim, M. B.; Wadia, A.; Tatton, W.; Weinstock, M.
Ann. N.Y. Acad. Sci. 2001, 939, 450.
34. Jones-Villeneuve, E. M.; McBurney, M. W.; Rogers, K.
A.; Kalnins, V. I. J. Cell Biol. 1982, 94, 253.
35. Woodgate, A.; MacGibbon, G.; Walton, M.; Dragunow,
M. Brain Res. Mol. Brain Res. 1999, 69, 84.
36. Nakao, N.; Yokote, H.; Nakai, K.; Itakura, T. J.
Neurosurg. 2000, 92, 659.
18. Yogev-Falach, M.; Amit, T.; Bar-Am, O.; Youdim, M. B.
FASEB J. 2003, 17, 2325.
37. Finley, M. F. A.; Kulkarni, N.; Huettner, J. E. Neuro-
science 1996, 16, 1056.
19. Warshawsky, A.; Youdim, M.; Fridkin, M.; Zheng, H. L.;
Warshawsky, R. WO Patent 2,004,041,151, A2, 2004.
20. Gassen, M.; Youdim, M. B. H. Pharmacol. Toxicol. 1997,
80, 159.
38. Staines, W. A.; Morassutti, D. J.; Reuhl, K. R.; Ally, A. I.;
McBurney, M. W. Neuroscience 1994, 58, 735.
39. Gassen, M.; Glinka, Y.; Pinchasi, B.; Youdim, M. B. H.
Eur. J. Pharmacol. 1996, 308, 219.
21. Cabantchik, Z. I.; Glickstein, H.; Milgram, P.; Breuer, W.
Anal. Biochem. 1996, 233, 221.
40. Meldrum, B.; Garthwaite, J. Trends Pharmacol. Sci. 1990,
11, 379.
22. Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1991,
183, 71.
23. Singh, S.; Khodr, H.; Taylor, M. I.; Hider, R. C. Biochem.
Soc. Symp. 1996, 61, 127.
24. Van Acker, S. A.; Van den Berg, D. J.; Tromp, M. N.;
Griffioen, D. H.; Van Bennekom, W. P., et al. Free
Radical Biol. Med. 1996, 20, 331.
41. Grant, E. R.; Errico, M. A.; Emanuel, S. L.; Benjamin, D.;
McMillian, M. K.; Wadsworth, S. A.; Zivin, R. A.;
Zhong, Z. Biochem. Pharmacol. 2001, 62, 283.
42. Harbour, J. R.; Chow, V.; Bolton, J. R. Can. J. Chem.
1974, 52, 3549.
43. Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. J. Am.
Chem. Soc. 1980, 102, 4994.
25. Sayre, L. M.; Smith, M. A.; Perry, G. Curr. Med. Chem.
2001, 8, 721.
44. Cheng, Z.; Ren, J.; Yan, G.; Li, Y.; Chang, W.; Chen, Z.
Bioorg. Chem. 2003, 31, 149.
26. Olanow, C. W. Adv. Neurol. 1993, 60, 666 (ParkinsonÕs
Disease).
45. Nakao, N.; Yokote, H.; Nakai, K.; Itakura, T. Neurosci.
Lett. 1989, 107, 85.
27. Fowler, J. S.; Logan, J.; Volkow, N. D.; Wang, G. J.;
MacGregor, R. R.; Ding, Y. S. Methods 2002, 27,
263.
46. Tabira, T.; Bush, A. I.; Masters, C. L. Science 2001, 292,
2251.
47. Cole, G. M. Neuron 2003, 37, 889.
28. Youdim, M. B. H.; Weinstock, M. Mech. Aging Dev. 2002,
123, 1081.
48. Tipton, K. F.; Youdim, M. B. H. Methods Biog. Amine
Res. 1983, 441.