Structure-activity relationships and mechanism of action of antitumor bis 8-hydroxyquinoline substituted benzylamines

Article abstract of DOI:10.1016/j.ejmech.2009.11.006

A series of twenty six 8-hydroxyquinoline substituted amines, structurally related to compounds 2 and 3, were synthesized to evaluate the effects of structural changes on antitumor activity and understand their mechanism of action. The studies were performed on a wide variety of cancer cell lines within glioma and carcinoma models. The results obtained from chemical models and biological techniques such as microarrays suggest the following hypothesis that a quinone methide intermediate which does not react with DNA but which gives covalent protein thiol adducts. Micro-array analysis showed that the drugs induce the expression of a variety of stress related genes responsible for the cytotoxic and cytostatic effects in carcinoma and glioblastoma cells respectively. The described analogues could represent new promising anti-cancer candidates with specific action mechanisms, targeting accessible thiols from specific proteins and inducing potent anti-cancer effects.

Full text of DOI:10.1016/j.ejmech.2009.11.006

European Journal of Medicinal Chemistry 45 (2010) 623–638
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Structure–activity relationships and mechanism of action of antitumor bis
8-hydroxyquinoline substituted benzylamines
a
,
1
d,
Sebastien Madonna , Christophe Beclin 1, Younes Laras a, Vincent Moret a, Aline Marcowycz b,
´
´
Delphine Lamoral-Theys c, Jacques Dubois c, Magali Barthelemy-Requin d, Gae¨lle Lenglet e,
Sabine Depauw , Thierry Cresteil , Genevieve Aubert , Valerie Monnier , Robert Kiss ,
´ `
Marie-Helene David-Cordonnier , Jean-Louis Kraus
e
f
f
g
b
`
´
e
a,
*
a Laboratoire de Chimie Biomole´culaire, CNRS, IBDML-UMR-6216, Campus de Luminy, Case 907, 13288 Marseille cedex 09, France
b Laboratoire de Toxicologie, Institut de Pharmacie, Universite´ Libre de Bruxelles, Campus de la Plaine, Blvd du Triomphe, 1050 Brussels, Belgium
c Institut de Pharmacie , Universite´ libre de Bruxelles (ULB)-CP205/12-Boulevard du Triomphe 1050, Brussels, Belgium
d CNRS, IBDML-UMR-6216, IMVT, Campus de Luminy, Case 907, 13288 Marseille cedex 09, France
e INSERM U-837, Team Molecular and cellular Targeting for Cancer Treatment, Jean-Pierre Aubert Research Center, Institut de Recherches sur le Cancer de Lille,
Place de Verdun, F-59045 Lille, IMPRT-IFR114, Lille, France
f Institut de Chimie des Substances Naturelles, CNRS UPR, 2301 Avenue de la terrasse, 91198 Gif sur Yvette, France
g Spectropole-Service de Spectrome´trie de Masse, Faculte´ de St Je´roˆme, Avenue Escadrille Normandie Nie´men, 13397 Marseille Cedex 20, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of twenty six 8-hydroxyquinoline substituted amines, structurally related to compounds 2 and 3,
were synthesized to evaluate the effects of structural changes on antitumor activity and understand their
mechanism of action. The studies were performed on a wide variety of cancer cell lines within glioma
and carcinoma models. The results obtained from chemical models and biological techniques such as
microarrays suggest the following hypothesis that a quinone methide intermediate which does not react
with DNA but which gives covalent protein thiol adducts. Micro-array analysis showed that the drugs
induce the expression of a variety of stress related genes responsible for the cytotoxic and cytostatic
effects in carcinoma and glioblastoma cells respectively. The described analogues could represent new
promising anti-cancer candidates with specific action mechanisms, targeting accessible thiols from
specific proteins and inducing potent anti-cancer effects.
Received 16 February 2009
Received in revised form
19 October 2009
Accepted 2 November 2009
Available online 10 November 2009
Keywords:
Bis-8-hydroxyquinoline
Quinone methide
Cancer
Protein thiol alkylation
Ó 2009 Elsevier Masson SAS. All rights reserved.
1. Introduction
for chemical transformations toward novel key intermediate
skeletons.
Privileged substructures from bioactive natural products have
played and continue to play an invaluable role in the drug discovery
process [1–3] The diversification of these heterocyclic compounds
using diversity-oriented synthesis (DOS) has been proven to be an
essential tool to rapidly discover biologically active small molecules
[4–6]. Therefore, the development of a concise and efficient
strategy leading to skeletal and stereo chemical diversity has
gained much attention in scientific communities involved in drug
discovery and biomedical research [7]. To achieve this goal, it is
crucial to identify a new privileged substructure with the potential
Hydroxyquinoline is a privileged structural moiety observed in
many biologically active natural products, and is used as the source
for many drugs diversely prescribed among a wide range of
pathologies including neurodegenerative diseases [8], parasitic
amoebic dysentery disease [9], and herpes viral diseases [10]. More
specifically, 8-hydroxyquinoline moiety has been mostly used for
its capacity to strongly chelate metal ions, particularly Cuþþ and
Znþþ [11].
We have recently reported preliminary results on the antitumor
activity of a family of bis 8-hydroxyquinoline substituted benzyl-
amines [12], but the mechanism(s) of action by which they exert
their antitumor activity remains to be elucidated. The aim of the
present work is thus to enlarge the initial panel of bis 8-hydrox-
yquinoline benzylamine analogues, in order to first proceed with
SAR analyses, and then to provide data which could account, at least
partly, for the mechanism of action of this new family of compounds.
* Corresponding author. Tel.: þ33 4 91 82 91 41.
E-mail address: kraus@luminy.univ-mrs.fr (J.-L. Kraus).
Both authors equally contributed to this work.
1
0223-5234/$ – see front matter Ó 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2009.11.006

624
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
analogues are not reactive enough to form covalent DNA adducts,
R
but are sufficiently reactive to form adducts with cell proteins at
their thiol sites. This finding is corroborated by direct character-
ization of quinone methide electrophile intermediates generated
from these new derivatives, which react preferentially with
nucleophiles like thiols. The compounds 2 and 3 under study are
also able to induce a strong stress response in cells leading to cell
division arrest and/or cell death depending on cell types.
OH
N
N
N
2. Chemistry
OH
Fig. 1. Bis 8-hydroxyquinoline benzylamine derivatives.
The synthetic scheme which has been developed in order to
elaborate this library of 26 analogues is outlined in Schemes 1 and
2. Syntheses of analogues 1–4, 15–17, 20 and 26 have already been
reported [12]. (Compounds 2 and 3 are also named respectively
JLK1472 and JLK1486 since they have been described under these
code numbers in Ref. [12])
The new analogues 5–14, 18 and 19 have been synthesized by
direct condensation of 5-chloromethyl 8-hydroxyquinoline with
various substituted amines (Scheme 1) [12]. This reaction was
performed in the presence of potassium carbonate in CH3CN and
gives a mixture of mono- and bis-hydroxyquinoline derivatives. By
reacting bis-8-hydroxyquinoline 3 with methyl iodide and sodium
halide the corresponding bis-8-methoxyquinoline analogue 25 was
obtained. Boc protected analogues 21 and 22 have been synthe-
sized by condensation of the mono-quinoline compounds with
Boc2O.
We first investigated the possible role for these bis 8-hydrox-
yquinoline substituted benzylamine drugs (Fig. 1) to behave as
alkylating agents because their 8-hydroxyquinoline moiety is
coupled to a benzyl amino group, which can induce through
chemical reactivity the formation of an alkylating quinone methide
intermediate.
To our chemical knowledge this alkylating property of the bis
8-hydroxyquinolinyl benzylamine scaffold has never been repor-
ted. Given this, we found much interest in searching for new bis
8-hydroxyquinoline substituted benzylamines, modified on the
quinoline nucleus as well on the benzylamine moiety, in order to
challenge the proposed mechanism(s) of action. Combining
mimicking chemistry, in vitro cell biology and molecular biology
experiments, we investigated the reactivity of these analogues in
the presence of chemical and biological nucleophiles in order to
further decipher their mechanism of action. This study led to the
main conclusion that the bis 8-hydroxyquinoline benzylamine
The synthesis of 23 and 24 is illustrated in Scheme 2. The
required starting material 5-aminomethyl 8-hydroxyquinoline was
prepared by reacting 5-chloromethyl 8-hydroxyquinoline with
1 : R1 = -benzyl, X= OH
2 (JLK1472):
R1 = -CH2-(4-CH3-phenyl), X= OH
3 (JLK1486):
R
1 = -CH2-(4-CF3-phenyl), X= OH
(ii)
25 : R1 = -CH2-(4-CF3-phenyl), X= OMe
4 :
X
R1 = -CH2-(2-CF3-phenyl), X= OH
R1
N
5 :
R1 = -CH2-(3-CF3-phenyl), X= OH
N
(i)
6 : R1 = -CH2-(3,5-CF3-phenyl), X= OH
7 :
R1 = -CH2-(4-OCF3-phenyl), X= OH
8 :
R1 = -CH2-(3-I-phenyl), X= OH
N
9 : R1 = -CH2-(2-thiophene), X= OH
10 :
X
R
1 = -CH2-(2-tetrahydrofuran), X= OH
11 :
R1 = -CH2-(1-methyl-1-H-pyrrole), X= OH
Cl
12 : R1 = -CH2-CH2-(2-pyridine), X= OH
13 :
+
R1 = -CH2-CH2-(1-pyrrolidine), X= OH
14 :
R1 = -CH2-(4-cyclohehanecarboxylic acid), X= OH
N
OH
R1
N
R1
N
1
R2
R3
(iii)
(i)
N
N
OH
R1 = -benzyl, R2 = H
R1 = -CH2-(4-CH3-phenyl), R2 = H
OH
15 :
16 :
20 :
21 :
R
1 = -CH2-(4-CF3-phenyl), R3 = Boc
1 = -CH2-(3-CF3-phenyl), R3 = Boc
R
17 : R1 = -CH2-(4-CF3-phenyl), R2 = H
18 :
22 : R1 = -CH2-(3,5-CF3-phenyl), R3 = Boc
R1 = -CH2-(3-CF3-phenyl), R2 = H
Scheme 1. Preparation of the mono (15–18 and 20–22) and bis 8-hydroxyquinoline 1–14 and 25. Reagents and conditions: (i) R1-NH2, K2CO3, CH3CN, 50 ꢀC, 24 h; (ii) NaH, CH3CN,
reflux, 3 h then MeI, reflux, 2 h; (iii) (tBoc)2O, CH2Cl2, 6 h.

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
625
R1
N
Cl
NH2
R2
N
ii or iii
i
N
N
OH
OH
OH
19 : R1 = R2 = -CH2-phenyl
23 : R1 = -(CO)-(4-CH3-phenyl), R2= H
24 :
R1 = R2 = -5 -CH2-(8-hydroxyquinoline)
Scheme 2. Preparation of the 8-hydroxyquinoline 19, 23 and 24. Reagents and conditions: (i) potassium phthalimide, DMF, 150 ꢀC, 5 h then 32% HCl, 150 ꢀC, 9 h (ii) 4-methylbenzoyl
chloride, K2CO3, CH3CN, 18 h; (iii) Chloromethylbenzene or 5-chloromethyl 8-hydroxyquinoline, K2CO3, CH3CN, 50 ꢀC, 18–24 h.
potassium phthalimide in DMF at 150 ꢀC to give the intermediate 2-
((8-hydroxyquinolin-5-yl)methyl)isoindoline-1,3-dione [13].
Deprotection of phthalimide group with concentrated hydro-
chloric acid led to the corresponding primary amine hydrochloride
in good yield. Acylation of 5-aminomethyl 8-hydroxyquinoline
with the 4-methylbenzoyl chloride gives compound 23. The reac-
tion of 5-chloromethyl 8-hydroxyquinoline with 5-aminomethyl
8-hydroxyquinoline gives the 5,50,500-nitrilotris(methylene)-
triquinolin-8-ol 24.
compound, the cell density never went below the initial cell
density, even at 20 M.
m
As shown in Fig. 2, compound 2 was found to be two orders of
magnitude more potent than Temozolomide (TMZ), the most active
drug against glioblastomas [14].
As reported in Table 3, we next examined the antitumor activity
of the most active selected compounds on a panel of human cell
lines organized into two subpanels representing carcinomas and
gliomas. This screening has been performed according to a meth-
odology already described [15,16].
3. Results and discussion
As reported in Table 3, it can be observed that compounds 2, 3, 5
and 9 were the most active at nanomolar range on glioma cell lines
while on carcinoma cell lines the most active compounds were 2, 3,
6 and 9. With exception for BxPC3 carcinoma cell lines, which
exhibits a lower sensitivity to analogues 2 and 3. Only compounds 6
and 9 were found to be as active on BxPC3 as on the other carci-
noma cell lines shown in Table 3.
3.1. Antitumor activity
As a primary screening, all analogues were first submitted for
evaluation of in vitro antitumor activity at a single concentration
(10
mM) in the human mouth epidermal carcinoma cell line (KB3)
and at a single concentration (30
m
M) in the human glioblastoma
Taken together, the data make evident that compounds 2, 3, and
9 are the most promising hits.
U87-GFP cell line. The active compounds which met our criteria for
CC50 (KB3) or ED50 (U87) determination were then used for further
evaluation against a panel of 5 carcinoma and 3 glioma cell lines.
Table 1 reports the antitumor activity of the bis 8-hydrox-
yquinoline analogues. Table 2 reports the in vitro antitumor activ-
ities of the mono and other quinoline analogues specifically
designed for deciphering the mechanism(s) of action of this family
of compounds.
3.2. Mechanism of action
In considering the electronic properties and the chemical reac-
tivity of the N-benzyl 8-hydroxyquinoline, the possibility that this
scaffold may behave as an alkylating agent promoter was envis-
aged. In specific conditions (temperature, solvent, and nucleophile
reactants) 8-hydroxyquinoline derivatives can undergo the
formation of quinone methide intermediates. Several examples of
drugs whose antitumor activities have been attributed to their
capacity to generate alkylating intermediates have been reported
[17–24]. In this perspective, we hypothesized the possibility for the
N-benzyl 8-hydroxyquinoline derivative 3 to generate quinone
methide intermediate 27 through the pathway underlined in
Scheme 3. This proposed mechanism of action requires two steps,
a protonation step followed by a nucleophilic attack leading to the
generation of the quinone methide intermediate.
From this primary screening, it can be observed that among the
whole 26 screened analogues
- Compounds 2, 3 and 9, the most active analogues on KB3 cell
line with respectively CC50 values of 0.003, 0.001 and 0.002 mM,
are only one order less active than the well known drug
docetaxel CC50 value 0.1 nM.
- The most active compounds on U87 glioma cell line were
analogues 2, 3 and 9 with EC50 values ranging around 0.2 mM.
These values which are two orders magnitude less potent
indicating that as expected U87 Glioma cells line are less
sensitive to the drugs than adenocarcinoma cell lines such as
KB3.
- Bis 8-hydroxyquinoline analogues 2 and 3 were more active
than their mono 8-hydroxyquinoline corresponding analogues
16 and 17.
- Dose–response curves of U87 cells treated with increasing
concentrations of the most active analogues (2, 3 and 9)
showed that, whereas the compounds at concentrations up to
In order to support this hypothesis of a possible cytotoxic
activity through a quinone methide alkylating intermediate, we
first established a structure–cytotoxic activity relationship based
on the results obtained against KB3 and U87 cell lines reported in
Table 2.
- Compound 3 was the most active analogue of this series of bis
8-hydroxyquinoline substituted benzylamines with two 8-
hydroxyquinoline moieties.
- Compound 25 in which both hydroxyl groups have been pro-
tected by methoxy substituents was, as expected, found to be
or equal to 0.2 mM were able to completely abolish cell prolif-
eration, apparent cell death was never observed even at very
high doses. Only the dose response – curve for 2 is shown in
Fig. 2, and a similar curve was obtained with compound 3 (data
not shown). During the three days after addition of the
totally inactive at 10
mM, a feature that confirms the

626
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
Table 1
Structure and cytotoxic activity of bis-8-hydroxyquinoline analogues.
General structure
Compd
R1
Cell lines (mM)
KB3a (CC50
)
U87b (ED50
)
1
0.02
<0.5
H3C
2 (JLK1472)
0.003
0.001
0.1
<0.2
<0.2
<0.2
2
F3C
3 (JLK1486)
CF3
4
5
2.2
F3C
F3C
CF3
6
1
<0.2
OH
N
R1
N
O
F3C
7
8
0.1–1
n.d.c
2–5
N
OH
1
I
9
0.002
10
<0.2
5–10
2
S
10
11
O
N
n.d.c
N
12
13
>10
n.d.
1
0.2–2
N

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
627
Table 1 (continued )
General structure
Compd
R1
Cell lines (mM)
KB3a (CC50
)
U87b (ED50
)
O
C
HO
14
1.1
n.d.c
–
Docetaxel
–
0.25
a
Cytotoxic Concentration (CC50). Cell growth inhibition was determined by the MTS assay; the number of viable cells were proportional to the extent of formazan
production in the presence or absence of drugs. The cytotoxicity index was plotted against the drug concentration which ranged between 0.5 and 10 nM and the value
resulting in 50% cytotoxicity was determined (the experiments were performed in triplicate).
b
Efficacy Dose (ED50). Antiproliferative activity is given as ED50; antiproliferative effect (U87-GFP cells) is measured by fluorescence using a flash-cytophotometer. In this
assay, the antiproliferative effects were observed at five concentrations 0.1, 1, 5, and 10
observed is given. These experiments are performed in quadruplicate.
mM, versus control. The lowest concentration at which an antiproliferative effect is
c
N.D not determined.
Table 2
Structure and antitumor activity of mono 8-hydroxyquinoline and other related analogues.
General structure
Compd
R1
R2
R3
Cell lines (mM)
KB3a CC50
U87b EC50
n.d.c
a
b
15
H
OH
>1
H3C
F3C
16
17
H
H
OH
OH
n.d.c
>10
>10
10
18
H
OH
0.2
0.18
F3C
R1
N
R2
19
20
OH
OH
1
5–10
2–5
F3C
F3C
N
Boc
Boc
10
R3
21
22
OH
OH
1
1
>10
CF3
Boc
n.d.c
F3C
H3C
23
H
OH
1
10
O
(continued on next page)

628
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
Table 2 (continued )
General structure
Compd
R1
R2
R3
Cell lines (mM)
KB3a CC50
U87b EC50
a
b
N
N
N
N
R3
R3
24
OH
>10
>10
F3C
F3C
R3
25
26
OMe
>10
n.d.c
R3
H
10
>10
a
Cytotoxic Concentration (CC50). Cell growth inhibition was determined by the MTS assay; the number of viable cells were proportional to the extent of formazan
production in the presence or absence of drugs. The cytotoxicity index was plotted against the drug concentration which ranged between 0.5 and 10 nM and the value
resulting in 50% cytotoxicity was determined (the experiments were performed in triplicate).
b
Efficacy Dose (ED50). Antiproliferative activity is given as ED50; antiproliferative effect (U87-GFP cells) is measured by fluorescence using a flash-cytophotometer. In this
assay, the antiproliferative effects were observed at five concentrations 0.1, 1, 5, and 10
observed is given. These experiments are performed in quadruplicate.
mM, versus control. The lowest concentration at which an antiproliferative effect is
c
N.D not determined.
requirement of free hydroxyl groups at the 8 position of the
quinolinyl nucleus.
step is mandatory for initiating the formation of the quinone
methide alkylating intermediate.
- Compound 26, in which the two 8-hydroxyl groups at the 8
position of the heterocyclic nucleus have been removed [12],
cannot undergo the formation of the suggested quinone
methide alkylating species, and was found to be 4 orders of
magnitude less potent than its corresponding hydroxylated
analogue 3 on KB3 cell line.
This structure–cytotoxicity correlation is consistent with the
proposed quinone methide formation hypothesis; since structural
changes favouring or impeding the generation of this intermediate
directly influence their cytotoxicity against the KB3 cell line. The
formation of the quinone methide intermediate required two major
structural conditions: (1) Presence of a protonable function repre-
sented by the tertiary amine function. Analogues bearing non-
protonable functions, such as 20 and 25 were found inactive. (2)
Presence of at least one free hydroxyl group at the 8 position of the
quinoline nucleus appears to be a favouring feature for promoting
activity, but not a sufficient structural feature since in the mono-8-
hydroxy series, compounds 16 and 17 are found inactive on both
cell lines KB3 and U87.
- Compounds 20 (CC50 ¼ 10
mM) and 23 (CC50 ¼ 1
mM) in which
the tertiary amine was respectively replaced by a carbamate or
amide function, showed cytotoxic effects noticeably reduced.
This drop in cytotoxic effect can be attributed to the presence of
carbamate or amide functions which are non-protonable
functions; a feature that once more confirms that a protonation
Compounds 20, 23, 25, and 26 whose chemical structure does
not fulfill these two structural requirements are inactive or only
weakly active. Such structures do not favour or impede the
formation of the quinone methide intermediate.
Table 3
Cytotoxic activity (IC50) against cancer cell panel of selected analogues.
Carcinoma (nM)
Glioma (nM)
Compd.
A549 BxPC3 LoVo MCF7 PC3
HS683 T98G U373
1
544
44
8
46
38
9
350
27
661
7442
325
4299
274
167
79
2593
89
4337
91
43
10
93
94
10
393
40
97
47
34
88
30
38
374
35
258
41
66
179
37
71
394
44
9939 2857
76
51
10
64
36
181
479
35
249
71
99
210
57
72
446 5595
88 85
9755 3726
357
53
44
78
65
96
2 (JLK1472)
3 (JLK1486)
4
5
6
8
9
14
Fig. 2. Effect of compound 2 concentrations on proliferation and motility of U87 cells.
Values were given at 3 days after compound addition and were obtained using the
scratch assay described in the experimental procedures section. Proliferation rate is
calculated as the ratio between cell content in the wells at day 3 and cell content at day
0. Motility represents the re-colonization rate of the scratch area by the cells divided
by the proliferation rate. Bars: SD. Similar results were obtained using compound 3
instead of compound 2 (data not shown).
422
9461
Inhibition of cell proliferation measured by the MTT colorimetric assay as detailed
previously [15,16]. The values reported in this Table are means obtained on six
distinct values. The standard errors are not reported for the sake of clarity because
they are <5% as compared to the mean values.

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
629
-
F3C
F3C
F3C
Nu
O
N
O
N
H
H
N
H
H
N
NH
N
O
+
N
N
alkylating specie
N
OH
OH
17
OH
27
JLK 1486 (3)
F3C
H
N
F3C
N
H
O
+
NH2
N
O
H
Nu
alkylating specie
28
Scheme 3. Hypothesis on the mechanism of action of compound 3 (JLK1486).
A second quinone methide intermediate could also be generated
from 17, after protonation of the newly formed secondary amine
leading subsequently to the formation of tri-fluorobenzylamine 28
as shown in Scheme 3. This observation is consistent in part with
the reduced cytotoxic activity observed for most of the mono-8-
hydroxyquinoline N-benzyl analogues compared to the bis-quino-
line analogues, but some exceptions are observed for analogue 5
(bis-8-hydroxyquinoline series analogue) compared to analogue 18
belonging to the mono-8-hydroxyquinoline series.
methide intermediate, we used a chemical model similar to that
reported by Pande et al. [26,27] in order to provide evidence that O-
(tert-butyldimethylsilyl)-2-bromo-methylphenol can generate an
ortho-quinone methide in the presence of fluoride anion. According
to this model, it appears that when compound 3 was submitted to
a nucleophilic fluoride anionic attack in a DMF/H2O solvent system
(for 48 h at 50 ꢀC), product analysis by mass spectrometry allowed
a clear identification of the presence of the quinone methide
intermediate 27 (m/z [M þ H]þ ¼ 158) as well as the corresponding
expected secondary amine product 17 (m/z [M þ H]þ ¼ 333). The
structure of 17 was confirmed by comparison with a genuine
sample synthesized separately [12]. We also identified the presence
of 4-trifluorobenzyl amine 28 in the reaction mixture products,
suggesting that the secondary amine 17 could also generate
a second quinone methide intermediate, as shown in Scheme 3.
3.3. Identification of a quinone methide intermediate through
a chemical mimicking model
Since relatively stable quinone methide intermediates have
been described [25] and in order to identify the postulated quinone
F3C
F3C
OH
(i)
N
NH
N
+
N
O
N
N
OH
JLK1486
OH
27
17
SH
NO2
S
H
S
NO2
NO2
N
N
OH
O
29
Scheme 4. Alkylation of quinone methide with 4-nitrothiophenol. Reagents and conditions: (i) KF, 4-nitrothiophenol, DMF/H20 (1:1), 50 ꢀC, 48 h.

630
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
From the obtained mixture, all attempts to isolate and purify 27
could lead to a subsequent cleavage of the DNA [32], which could be
seen using denaturing polyacrylamide gel electrophoresis. Once
more, no cleavage of the DNA was evidenced (data not shown).
Altogether, these experiments strongly suggest that DNA is not
a target for compound 3, at least from in vitro experiments where
no bioactivation could occur. EI-MS experiments were performed
on compound 3 incubated with desoxyguanosine (37 ꢀC, 24 h in
1 mM ammonium acetate, pH 7.15) in conditions described in
Ref. [33] Direct ion spray mass spectrometry analysis presented in
Fig. 3A evidenced the formation of a product at m/z ¼ 346 which is
absent in the sample containing compound 3 alone or desoxy-
guanosine alone. MS/MS of the product at m/z ¼ 346 is presented in
Fig. 3B and reveals that it is derived from desoxyguanosine (m/
z ¼ 268). An additional peak at 152 corresponds to a guanine
residue. The difference in mass between 346 and 268 brings to light
a mass of M ¼ 78. EI-MS experiments were also performed on
compound 3 incubated with a short hairpin oligonucleotide as
described in Ref. [31]. However, no DNA alkylation by compound 3
was observed even after 24 h incubation time with the DNA.
Altogether these results support the finding that the cytotoxicity
induced by this family of compounds against a large set of human
cancer cell types suggest that the identified quinone methide
electrophile intermediate preferentially reacts with other highly
active biological nucleophiles, such as thiol, phenol, or nucleophilic
amino groups.
failed, but NMR analysis of the reaction product mixture which,
besides 27, contained also products 3 (un-reacted starting material)
and 4-trifluorobenzyl amine 28, confirmed the obtained mass
spectrometry results. While a peak at
d
¼ 4.40 (ppm) corresponding
to two methylenyl proton signals of the quinone methide 27 was
clearly identified, this 1H NMR proton signal was not present in the
NMR spectra of compounds 3 and 4-trifluorobenzyl amine 28.
3.4. Potential reactivity of the quinone methide intermediate on
DNA and nucleotides
We then used this chemical model to investigate the possibility
for this quinone methide intermediate to alkylate DNA. For this
purpose we carried on the same reaction in the same conditions
(50 ꢀC, KF, DMF/H2O) but in the presence of dideoxyadenosine as
a DNA mimicking substrate. In these conditions, HPLC/Mass spec-
trometry reaction products analysis revealed that no adducts
resulting from the addition of the quinone methide intermediate
on the dideoxyadenosine substrate were detected or identified.
Results were similar even though the reaction was performed for
a longer time or at a higher temperature (data not shown).
Dideoxyadenosine was probably not a strong enough nucleophile
to react with the weak electrophile quinone methide intermediate
27. We then performed a similar reaction in the same conditions
(solvent and temperature) but in the presence of 4-nitrothiophenol
instead of dideoxyadenosine, since the thiol function is a stronger
nucleophile which should be more reactive than the 4-amino group
of dideoxyadenosine. Under these conditions and among the other
reaction products, which contained un-reacted starting compound
3 and secondary amine derivative 17, a new adduct, un-reacted
nitrothiophenol 29 (M þ H ¼ 311), was identified by HPLC/MS. Its
structure corresponds to the addition of 4-nitrothiophenol on
quinoline 27, and subsequent oxidation as represented in Scheme 4.
These results suggest that the quinolinyl methide intermediate
27 is a poorly or moderately reactive electrophile which reacts
preferentially with strong nucleophiles such as thiols. These results
strongly agree with results reported previously [28], which showed
that peptide alkylation by quinone methide electrophiles should
occur in the following current order: cysteine thiol > N-terminal
6 107
268
A
5 107
4 107
JLK1486
279
3 107
2 107
1 107
0
490
292
535
319 346
333
303
508
411
250
300
350
400
450
500
550
600
M/z
268
amino > N-
3
Lysine ¼ N-Im-Histidine. The moderate reactivity and
8 106
6 106
4 106
2 106
0
the relative stability of this intermediate 27 could be attributed to
its extended conjugation with the neighbouring nitrogen contain-
ing heterocycle.
dG
B
3.5. Evaluation of DNA binding/bonding ability of compound 3
G
From thermodynamic considerations, one can argue that
chemical models are far from a biological model. Nevertheless,
examples are described in the literature showing DNA alkylation by
various anti-cancer drugs occurred through a quinone methide-like
intermediate, in the minor groove site [29]. With the hope of
confirming the results obtained with the chemical mimicking
models, in vitro experiments were performed to bring to light the
possible formation of adducts between compound 3 and DNA. For
this purpose, UV/visible spectrometry, circular dichroism analysis
and DNA melting temperature were performed in the presence of
compound 3 and CT-DNA to evaluate DNA binding ability of
compound 3 according to standard spectroscopic methods [30].
From those experiments, no binding of compound 3 to DNA was
identified (data not shown). A DNA alkylation reaction was also
analyzed by incubation of compound 3 with a radiolabeled DNA
fragment [31]. EMSA did not reveal any band shift, suggesting that
the compound does not bind/bond the DNA (data not shown). In
a similar manner, alkylation of DNA at the N7 group of guanines
152
346
50
100
150
200
250
300
350
400
450
500
M/z
Fig. 3. (A) Electrospray ionization mass spectrometry for compound 3 (JLK1486);
(M þ H ¼ 490, in red) incubated with desoxyguanosine (dG, M þ H ¼ 268, in blue) for
24 h at 37 ꢀC in 1 mM ammonium acetate, pH 7.15. Peaks present in samples con-
taining compound 3 alone for which MS/MS relates them to compound 3 (presented in
Fig. 4B) are also presented in red, whereas peaks present in samples containing dG
alone and for which MS/MS relates them to dG are also presented in blue (dG at
2M þ H ¼ 535). An additional peak that was not present in samples containing
compound 3 or dG alone is evidenced in purple. (B) MS/MS for peak at M þ H ¼ 346
evidenced peaks at M þ H ¼ 268 (dG) or 152 (guanine, resulting from dG) in blue. (For
interpretation of the references to colour in this figure legend, the reader is referred to
the web version of this article).

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
631
3.6. Detection of GSH/compound 3 adducts
(GSSG), or S-methyl-glutathione (GS-Me), since both molecules are
lacking a thiol group (data not shown).
Glutathione (GSH) is a natural bis-nucleophile, known to be
responsible for the observed increased resistance to alkylating anti-
cancer drugs [33,34]. GSH is known to react with numerous alky-
lating agents. When compound 3 was incubated with glutathione,
a new adduct was characterized by ion spray mass spectrometry
(Fig. 4A) corresponding to M þ H ¼ 465.
Specific alkylation of GSH could take the part of the cytotoxic
activity of an alkylating agent or a detoxification process. In order
to address the positive or negative effects of GSH adduct formation
on the cytotoxic activity of compound 3, GSH depletion of cells was
undertaken prior to treatment with compound 3. For this purpose
cells were treated with buthionine sulfoximine (BSO), an inhibitor
When this adduct was submitted to MS/MS, two new sub-
products, e.g. M þ H ¼ 158 (quinone methide) and M þ H ¼ 308
(GSH) were identified in addition to a product at M þ H ¼ 179
of GSH biosynthesis specifically blocking g-glutamylcysteine
synthetase which is known to reduce endogenous GSH cellular
levels [35], prior to treatment using increasing concentrations of
compound 3. These experiments were performed on the KB3-1
epidermoid carcinoma cells [36]. As presented in Fig. 5A,
increasing the concentration of BSO leads to a large decrease in the
amount of cellular reactive thiols in cells (observed decrease in
absorbance at 500 nm). The concentration of 1 mM BSO was
selected for pre-treatment of KB3-1 cells in the MTT assays pre-
sented. Comparison of untreated cells with BSO treated cells
evidences a shift of the CC50 from 70 nM (plain line, Fig. 5B) to
33 nM (dashed line, Fig. 5B). Such a 2-fold shift demonstrates an
increase in the cytotoxicity of compound 3 when cells are depleted
in GSH. Therefore, alkylation of GSH might be considered as
corresponding to a-glutamic acid obtained from cleavage of GSH.
Such a product at m/z ¼ 465 could therefore be the adduct whose
structure is embedded in Fig. 4C, and that could result from the
reaction presented in Scheme 5.
These results clearly confirm those obtained with the mimicking
chemical model, e.g. that the identified quinone methide inter-
mediate is a weak reactive specie, which can react only with highly
nucleophilic substrates such as GSH or any protein thiols. Such
reactivity on thiol occurs using GSH but not other thiols such as L-
cysteine or N-acetyl-cysteine. In the same manner, the reaction
does not occur in the presence of methionine, oxidized glutathione
6 107
GSH
490 JLK1486
A
5 107
4 107
3 107
2 107
1 107
0
308
319
GSSG
613
411
465
484
333
350
400
450
500
M/z
550
600
650
700
300
MSMS 490
MSMS 465
6 106
4.8 106
3.6 106
2.4 106
1.2 106
0
2.5 106
331
158
B
C
2 106
1.5 106
1 106
5 105
0
158
179
308
300
490
465
100
200
300
M/z
400
500
100
200
400
500
M/z
Fig. 4. (A) Electrospray ionization mass spectrometry for compound 3 (JLK1486) (M þ H ¼ 490, in red) incubated with GSH (M þ H ¼ 308, in blue) for 24 h at 37 ꢀC in 1 mM
ammonium acetate pH 7.15. Peaks present in samples containing compound 3 alone and for which MS/MS relates them to compound 3 (presented in panel B) are also presented in
red, whereas peaks present in samples containing GSH alone and for which MS/MS relates them to GSH are also presented in blue (GSSG, oxidized glutathione). An additional peak
that was not present in samples containing compound 3 or GSH alone is evidenced in green. (C) MS/MS for peak at M þ H ¼ 465 evidenced sub-products at M þ H ¼ 308 (GSH) or
179 (acide a-glutamic, resulting from GSH) in blue and at M þ H ¼ 158 (quinone methide, deriving from compound 3, in red). The potential structure of the adduct is presented in
this panel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article).

632
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
F3C
F3C
OH
N
NH
N
+
N
O
N
N
OH
OH
JLK1486
27
17
SH
O
O
O
H
N
HO
N
H
OH
NH2
O
GSH
N
N
OH
O
S
S
O
O
O
O
O
O
H
N
H
N
HO
N
H
OH
HO
N
H
OH
NH2
O
NH2
O
GS-QM
Scheme 5.
a detoxification process but not as being the molecular target of
compound 3.
500
400
300
200
100
0
A
3.7. Gene expression patterns in U87 glioblastoma cells treated with
compound 2 show a stress signature and a down-regulation of cell-
cycle
The search for molecular mechanisms of action has demon-
strated that compounds 2 and 3 are able to react with some thiol
radicals. These reactions are likely to affect the properties of thiol
containing molecules. They may lead to mis-folding of proteins
harboring cysteine residues. They also may inactivate enzymes
with a cysteine in their active site or reduce the anti-oxidant
capacity of the cells due to depletion of most of the reducing agents
in the cell such as GSH. To investigate these different possibilities
and to attempt to understand the observed cell phenotypes we
decided to explore the response of cells submitted to a treatment
with compound 2. We thus analyzed the change in gene expression
when compound 2 was added to the culture medium. We chose to
perform this analysis in U87 cells because these cells do not die (as
KB3 cells do) when treated with compound 2. Micro-array analyses
were performed 3 h and 16 h after compound addition. Table 4
contains the genes that are significantly disregulated by compound
2 treatment. These genes were classified into families based on
their cellular function.
400
450
500
wavelength (nm)
550
600
650
120
100
80
60
40
20
0
B
0,1
1
10
100
1000
In accordance with the intended detrimental consequences of
thiol radical alkylation many stress responsive genes are over-
expressed in U87 glioblastoma cells upon treatment with
compound 2:
JLK1486 (nM)
Fig. 5. Role of GSH addition on compound 3 (JLK1486) cytotoxicity in KB3-1 cells. (A)
Depletion of endogenous GSH upon treatment of the cells using increasing concentra-
tions of BSO. KB3-1 cells (grey line) were treated with 0.1, 0.25, 0.5, 1 or 10 mM of BSO
(black lines, top to bottom) for 24 h before measuring the level of endogenous global
thiol in cell extracts using ThioGlo-1Ô (lex ¼ 360 nm; lem ¼ 400–650 nm). (B) Cyto-
toxicity measurement of compound 3 in KB3-1 cells, pre-treated (black boxes and
dashed lane) or not (open circle and plain lane) with 1 mM BSO for 24 h. Increasing
amounts of compound 3 (JLK1486) from 0.1 nM to 1 mM were incubated during 72 h
before measuring the cell survival. Results are expressed as the percent of cell survival
over drug concentration (log scale). The 50% survival level (CC50) is represented as a line.
- Genes responsive to Endoplasmic Reticulum stress (ER stress
(Chop, ATF3, ATF4, RGS16, ERN1, PUMA)), probably due to the
accumulation of mis-folded proteins in the ER [37].
- Genes responsive to DNA-damage or to an oxidative stress
(DDIT4, GADD45a, SESN2) which may be a consequence of
lower anti-oxidant capacities from the cell stress [38].

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
633
Table 4
Genes significantly disregulated by treatment with compound 2 (JLK1472).
ER stress
DNA-damage and
Oxidation stress
Inflammation
Apoptosis
Cell survival
Cell cycle
Development
a
b
Chop, ATF3,
ATF4, RGS16,
PUMA, ARN1
SESN2, GADD45A,
DDIT4
IL6, IL8, IL1A, IL1F5,
EREG, IFRD1, CCL17,
CXCL3
NOXA, IGFBP3,
BIM, FAS, GDF15
BCL2L10 FGF2
FGF18
cyclinD2, CCPG1
PROX1 HQXA6
HQXB9 ISL2
E2F1
CDK2, cyclinF, CDC6,
CDC25A, cyclinE2,
EDN1, SKP2, CDCA7
Genes showing more than 3 fold increase (top of the table) or 3 fold reduction (lower part) in mRNA expression 16 h after JLK1472 addition in the culture medium were
classified based on their cellular function. In bold are indicated the genes that are disregulated (>1.4 fold change in mRNA content) as soon as 3 h after the beginning of the
experiment.
- Genes related to inflammation (IL6, IL8, IL1A, IL1F5, EREG,
IFRD1, CCL17, CXCL3) as it is often found when cells harbor ER
stress [39,40].
NAD(P)H-dependent quinone oxidoreductase (NQO1) induction
probably via thiol-rich sensor proteins. Some quinoid drugs have
been reported to inhibit cell growth and induce apoptosis via
reaction with GSH and protein thiols [46,47]. Results obtained from
BSO experiments indicate that alkylation of GSH by compound 3
should be considered as a detoxification process and not a molec-
ular target.
Another consequence of stress induction is the expression of
genes related to apoptosis (NOXA, BIM, IGFBP3, FAS, GDF15) that
can explain the cell death processes observed with the compounds
under study in carcinoma cells. In glioblastoma cells (U87, U373,
Hs683 or T989G) no cell death is observed despite the activation of
apoptosis-related genes. This resistance to apoptosis may be
explained by the intrinsic resistance of glioblastoma cells to
apoptosis [41] or by the induction of survival genes in treated cells
such as two genes coding for FGF factors (FGF2 and FGF18) and one
anti-apoptotic gene (BCL2L10), for example. Another set of genes
that are significantly disregulated in the presence of compound 2
are genes involved in the control of cell-cycle; there was an
induction of cell cycle inhibitors (cyclinD2, CCPG1) and an inhibi-
tion of genes involved in cell-cycle progression (CDK2, cyclinF,
CDC6, CDC25A, cyclinE2, EDN1, SKP2, CDCA7). This inhibition of cell
division is generally observed in cells submitted to a stress. In
particular, many reports establish the inhibition of cell division as
a direct consequence of ER stress [42]. Finally, several develop-
mental genes are also induced by treatment of cancer cells with
compound 2 (for example isl2 and the homeotic genes HOXA6,
HOXB9 and PROX1). This may be interpreted as a consequence of
cell-cycle exit that may induce some developmental processes at
the cellular level.
It therefore seems that the data reported here could account, at
least partly, for a relatively well defined mechanism of action
involving the fact that N-benzyl bis 8-hydroxyquinoline substituted
analogues actually generate stable quinone methide intermediates,
which can in turn react not only with thiols (at least with GSH) but
possibly also with other protein thiols to form adducts. It is prob-
ably due to their electrophilicity and consequently low reactivity
that these quinone methide intermediates do not react with DNA.
Chemical models and cellular in vitro assays have confirmed this
finding. Furthermore, this latter observation could account for the
cytostatic effect of the compounds observed in the case of specific
glioma cell lines (U373, Hs683, U87, or T98G), and for the cytoxicity
observed in the case of the epidermoid carcinoma cell line KB3.
Additional experiments are still required to confirm the results
obtained with the KB3 cell line and the glioma cell lines (U87,
U373), in particular to investigate if these latter cell lines yield DNA
and/or GSH adducts when treated with compound 3 or derivatives.
Reaction of quinone methides with thiol is prone to affect the
properties of thiol containing molecules which may lead to the
abnormal folding of proteins harboring cysteine residues. These
mis-folded proteins will as a consequence be retained and accu-
mulate in the endoplasmic reticulum leading to an ER stress [37].
Micro-array analyses have further shown that, in addition to an ER
stress signature, U87 glioblastoma cells treated with compound 2
exhibited a strong induction of several oxidative stress responsive
genes. Accordingly, depletion of thiol containing reducing mole-
cules (such as GSH) may affect the anti-oxidant capacity of the cells
rendering them more susceptible to oxidative stress [38].
3.8. Synthetic analysis of the obtained biological data
Cellular defense mechanisms that respond to damage from
oxidative stress and electrophiles (key cause of malignant trans-
formation) have been shown to be a target for chemo-preventive
agents [43]. Thiol group oxidation or alkylation by quinoids can
trigger chemo-preventive mechanisms. Several observations have
pointed out that quinoid metabolites, including O-quinones,
quinone methides and diquinone methides, can generate Reactive
Oxygen Species (ROS) leading to oxidation and alkylation of various
types of biomolecules, such as proteins and nucleic acids, a feature
which in turn can explain, at least partly, why these compounds are
cytostatic and/or cytotoxic [44,45]. According toThatcher et al. [45],
the balance between the harmful and the beneficial properties
associated with oxidative bioactivation are key to drug discovery
and chemical toxicology. Induction of cytoprotective enzymes such
as NAD(P)H-dependent quinone oxidoreductase (NQO1), gluta-
thione-S-transferase (GST) and UDP-glucuronyl-transferase, known
as phase 2 enzymes, may be cytoprotective not via chemical
detoxification but via biochemical pathways instead. It has been
proposed that quinone methide metabolites, after rapid reaction
with GSH, can activate so-called anti-oxidant responsive element
(ARE) via electrophilic mechanisms, which in turn, leads to
4. Conclusions
The antitumor activities of the new 8-hydroxy N-benzylamines
appear to be strongly correlated with their ability to generate
relatively stable and moderately reactive quinone methide inter-
mediates, which we show in turn to be able to react with thiols (at
least with GSH), but not with DNA probably due to their electro-
philicity and consequently low reactivity. The decrease in antitu-
moral activity observed with analogues whose structures impede
the formation of this alkylating intermediate, as well as the results
obtained simultaneously with mimicking chemical models and in
vitro cellular assays, confirm the suggested action mechanism.
Numerous alkylating agents are known to react with GSH. The first
known group consists of antitumor agents that are activated by
reaction with GSH, such as the natural product leinamycin, after

634
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
which a reaction with thiols generates an electrophilic epis-
ulfonium ion intermediate which then reacts with the N7 position
of guanines in DNA [47]. This is also the case for irofulven (MGI 114),
which reacts with GSH to give a diene intermediate implicated in
further nucleophilic attack of DNA [48,49] or brostallicin (PNU-
166196), whose cytotoxic activity derives from the formation of
GSH–brostallicin covalent adducts [50]. The second group corre-
sponds to compounds for which bonding to GSH inhibits alkylation
of the functional target such as DNA; this group includes
compounds S23906-1 [34], mitomycin C [51], cisplatin [52,53], and
nitrogen mustards [54,55]. The increase in cytotoxicity of
compound 3 upon GSH depletion by BSO suggests that this
compound belongs to the second group. Micro-array analysis
showed that despite their low reactivity, quinone methide inter-
mediates provoke a generalized stress induction in cells treated
with compounds 2 or 3, activating several apoptotic pathways in
carcinoma cells and the inhibition of cell proliferation in glioblas-
toma cells.
5.2.3. 5,50-(3,5-Bis(trifluoromethyl)benzylazanediyl)bis(methylene)-
diquinolin-8-ol (6)
It was synthesized from 3,5-di(trifluoromethyl)benzylamine
and 5-(chloromethyl)quinolin-8-ol following the general proce-
dure A. Yield 62%. mp ¼ 193–195 ꢀC. 1H NMR (250 MHz, CDCl3):
8.76–8.74 (m, 2H), 7.96–7.92 (m, 2H), 7.64 (s, 1H), 7.43–7.38 (m, 4H),
7.19–7.13 (m, 2H), 7.08–7.05 (m, 2H), 3.89 (br s, 4H), 3.61 (s, 2H). MS,
m/z (C29H21F6N3O2): calcd 558.5 [M þ H]þ; found, 557.5. Anal.
(C29H21F6N3O2) C, H, N.
5.2.4. 5,50-(4-(Trifluoromethoxy)benzylazanediyl)bis(methylene)di-
quinolin-8-ol. (7)
It was synthesized from 4-(trifluoromethyl)benzylamine and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
Yield 90%. mp ¼ 108–112 ꢀC. 1H NMR (250 MHz, CDCl3): 8.73 (m,
2H), 7.86 (m, 2H), 7.51 (m, 2H), 7.41 (m, 2H), 7.17 (m, 2H), 7.12–7.07
(m, 4H), 3.84 (br s, 4H), 3.55 (s, 2H). MS, m/z (C28H22F3N3O3): calcd
506.1 [M þ H]þ; found, 506.1. Anal. (C28H22F3N3O3) C, H, N.
5. Experimental section
5.2.5. 5,50-(3-Iodobenzylazanediyl)bis(methylene)diquinolin-8-ol (8)
It was synthesized from 3-Iodobenzylamine and 5-(chlor-
omethyl)quinolin-8-ol following the general procedure A. Product
purified by chromatography (SiO2 from CH2Cl2/MeOH 95:5 to
80:20). Yield 50%. mp ¼ 82–83 ꢀC. 1H NMR (250 MHz, MeOD): 8.79
(m, 2H), 8.13 (m, 2H), 7.61–6.92 (m, 10H), 4.02 (s, 2H), 3.72 (br s,
4H). Anal. (C27H22IN3O2) C, H, N.
5.1. Chemistry
Unless otherwise noted, starting materials and reagents were
obtained from commercial suppliers and were used without puri-
fication. Tetrahydrofuran (THF) was distilled over sodium benzo-
phenone ketyl immediately prior to use. Methylene dichloride
(CH2Cl2) was distilled over P2O5 just prior to use. Nuclear magnetic
resonance spectra were recorded at 250 MHz for 1H on a Bru¨ker
AC-250 spectrometer. Elemental analyses were within ꢁ0.4% of
theoretical values for all compounds. Chemical shifts are expressed
5.2.6. 5,50-(Thiophen-2-ylmethylazanediyl)bis(methylene)-
diquinolin-8-ol (9)
It was synthesized from 2-Thiophenemethylamine and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
The product was filtrated and washed by CH3CN (2 ꢂ 20 mL) and
ether (40 mL). The solvent was concentrated under vacuum. Yield
95%. 1H NMR (250 MHz, MeOD): 8.79 (m, 2H), 8.13 (m, 2H), 7.43 (m,
3H), 7.23 (m, 2H), 7.01 (m, 4H), 3.86 (br s, 4H), 3.76 (s, 2H). MS, m/z
(C25H21N3O2S): calcd 428.1 [M þ H]þ; found, 428.1. Anal.
(C25H21N3O2S) C, H, N.
as
d units (part per million) downfield from TMS (Tetramethyl-
silane). Electrospray mass spectra were obtained on a Waters
Micromass ZMD spectrometer by direct injection of the sample
solubilized in CH3CN. Analytical thin layer chromatographies (TLC)
and preparative thin layer chromatographies (PLC) were performed
using silica gel plates 0.2 mm thick and 1 mm thick respectively
(60F254 Merck). Preparative flash column chromatographies were
carried out on silica gel (230–240 mesh, G60 Merck). Melting of the
final compounds were obtained on a Thomas Hoover apparatus and
are uncorrected.
5.2.7. 5,50-((Tetrahydrofuran-2-yl)methylazanediyl)bis(methylene)-
diquinolin-8-ol (10)
It was synthesized from (tetrahydrofuran-2-yl)methanamine
and 5-(chloromethyl)quinolin-8-ol following the general proce-
dure A. Yield 90%. mp ¼ 98–99 ꢀC. 1H NMR (250 MHz, MeOD): 8.79
(m, 2H), 8.64 (d,1H, J ¼ 1.5 Hz), 8.12 (d, 1H, J ¼ 1.5 Hz), 7.35–6.90 (m,
6H), 4.01–3.99 (m, 3H), 3.76–3.65 (m, 4H), 2.76–2.47 (m, 2H), 1.99–
1.18 (m, 4H). MS, m/z (C25H25N3O3): calcd 416.2 [M þ H]þ; found,
416.2. Anal. (C25H25N3O3) C, H, N.
5.2. General procedure A
5.2.1. Preparation of the mono and bis 8-hydroxyquinoline 1–17
To a stirred solution of 5-chloromethylquinolin-8-ol hydro-
chloride (5.74 mmol) and K2CO3 (8.5 mmol) in CH3CN (20 mL), the
suitable amine (2.87 mmol) was added and the resulting mixture
was heated at 50 ꢀC for 24 h. The reaction was monitored by TLC.
The mixture was cooled to 0 ꢀC and filtrated; the filter cake was
washed with cold CH3CN (10 mL). The residue was purified by silica
gel column chromatography (CH2Cl2/MeOH 95:5) as eluent to give
the required compounds 1–17.
5.2.8. 5,50-((1-Methyl-1H-pyrrol-2-yl)methylazanediyl)bis-
(methylene)diquinolin-8-ol (11)
It was synthesized from (1H-pyrrol-2-yl)methanamine and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
Product purified by chromatography (SiO2 from CH2Cl2/MeOH 95:5
to 80:20). Yield 86%. mp ¼ 108–112 ꢀC. 1H NMR (250 MHz, MeOD):
8.83 (dd, 2H, J ¼ 1.5 HzF, 1.5 Hz), 7.80 (m, 2H), 7.62 (m, 2H), 7.34 (m,
2H), 7.20 (m, 2H), 6.81 (d, 1H, J ¼ 0.75 Hz), 6.40 (d, 1H, J ¼ 0.75 Hz),
6.01 (d, 1H, J ¼ 0.75 Hz), 3.86 (br s, 4H), 3.76 (s, 3H), 3.41 (s, 2H). MS,
m/z (C26H24N4O2): calcd 426.7 [M þ H]þ; found, 426.8. Anal.
(C26H24N4O2) C, H, N.
Compounds (1), (2), (3) and (4) were prepared following the
previously reported method described by Moret et al. [12].
5.2.2. 5,50-(3-(Trifluoromethyl)benzylazanediyl)bis(methylene)-
diquinolin-8-ol (5)
It was synthesized from 3-(trifluoromethyl)benzylamine and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
Yield 58%. mp ¼ 131–133 ꢀC. 1H NMR (250 MHz, CDCl3): 8.76 (m,
2H), 7.93 (m, 2H), 7.52–7.48 (m, 2H), 7.43–7.40 (m, 2H), 7.31–7.29
(m, 2H), 7.15–7.08 (m, 4H), 3.85 (s, 4H), 3.55 (s, 2H). MS, m/z
(C28H22F3N3O2): calcd 490.5 [M þ H]þ; found, 490.5. Anal.
(C28H22F3N3O2) C, H, N.
5.2.9. 5,50-(2-(Pyridin-2-yl)ethylazanediyl)bis(methylene)-
diquinolin-8-ol (12)
It was synthesized from 2-(pyridin-2-yl)ethanamine and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
Product purified by chromatography (SiO2 from CH2Cl2/MeOH 95:5

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
635
to 80:20). Yield 13%. mp ¼ 67–68 ꢀC. 1H NMR (250 MHz, MeOD):
8.79 (m, 2H), 8.22 (m, 2H), 7.90 (m, 4H), 7.51–6.98 (m, 6H), 3.88 (br
s, 4H), 2.95 (m, 4H). Anal. (C27H24N4O2) C, H, N.
chromatography eluting with cyclohexane–CH2Cl2 (1:1) to give the
required compounds 20–22.
5.3.2. Preparation of the (8-hydroxyquinolin-5-yl)methyl(4-(trifluoro-
methyl)benzyl) carbamate (20)
Compound (20) was prepared following the previously reported
method described by Moret et al.[12].
5.2.10. 5,50-(2-(Pyrrolidin-1-yl)ethylazanediyl)bis(methylene)di-
quinolin-8-ol (13)
It was synthesized from 2-(pyrrolidin-1-yl)ethanamine and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
Product purified by chromatography (SiO2 from CH2Cl2/MeOH 95:5
to 80:20). Yield 26%. mp ¼ 106–107 ꢀC. 1H NMR (250 MHz, MeOD):
8.61 (m, 2H), 8.20–8.06 (m, 2H), 7.40–6.90 (m, 6H), 3.78 (br s, 4H),
2.50 (s, 4H), 2.38 (m, 4H), 2.21 (s, 2H), 1.56 (s, 2H). Anal.
(C26H28N4O2) C, H, N.
5.3.3. Tert-butyl-(8-hydroxyquinolin-5-yl)methyl(3-(trifluoromethyl)-
benzyl) carbamate (21)
It was synthesized from compound (18) following the general
procedure B. Yield 27%. mp ¼ 114–116 ꢀC. 1H NMR (250 MHz,
CDCl3): 8.78 (m, 1H), 8.56–8.52 (m, 1H), 7.49–7.34 (m, 2H), 7.37–
7.30 (m, 2H), 7.26–7.21 (m, 2H), 7.09–7.06 (m, 1H), 4.82 (s, 2H), 4.28
(s, 2H), 1.95–1.55 (m, 9H). MS, m/z (C23H23F3N2O3): calcd 433.4
[M þ H]þ; found, 433.4. Anal. (C18H15F3N2O) C, H, N.
5.2.11. 4-((Bis((8-hydroxyquinolin-5-yl)methyl)amino)methyl)-
cyclohexanecarboxylic acid (14)
It was synthesized from 4-(aminomethyl)cyclohexanecarboxylic
acid and 5-(chloromethyl)quinolin-8-ol following the general
procedure A. Product purified by chromatography (SiO2 from
CH2Cl2/MeOH 95:5 to 80:20). Yield 95%. 1H NMR (250 MHz, CDCl3):
8.60–8.65 (m, 2H), 8.23–8.01 (m, 2H), 7.26–7.41 (m, 2H), 7.15–6.98
(m, 4H), 3.71 (br s, 4H), 2.18–2.12 (m, 2H), 2.00–1.92 (m, 1H), 1.75–
1.68 (m, 2H), 1.57–1.50 (m, 3H), 1.24–1.18 (m, 2H), 0.36–0.21 (m,
2H). MS, m/z (C28H29N3O4): calcd 472.5 [M þ H]þ; found, 472.5.
Anal. (C28H29N3O4) C, H, N.
5.3.4. Tert-butyl-3,5-bis(trifluoromethyl)benzyl((8-hydroxyquinolin-
5-yl)methyl) carbamate (22)
It was synthesized from 5-((3,5-difluorobenzylamino)methyl)
quinolin-8-ol following the general procedure B. Yield 23%.
mp ¼ 149–152 ꢀC. 1H NMR (250 MHz, CDCl3): 8.69 (m, 1H), 8.45–
8.38 (m, 1H), 7.53–7.47 (m, 1H), 7.40–7.34 (m, 1H), 7.30–7.21 (m,
2H), 7.19–7.16 (m, 1H), 6.98–6.95 (m, 1H), 4.79 (s, 2H), 4.28 (s, 2H),
1.39–1.52 (m, 9H). MS, m/z (C24H22F6N2O3): calcd 501.4 [M þ H]þ;
found, 501.4. Anal. (C24H22F6N2O3) C, H, N.
5.2.12. Preparation of the 5-((benzylamino)methyl) quinolin-8-ol
dihydrochloride (15)
5.3.5. Preparation of N-((8-hydroxyquinolin-5-yl)methyl)-4-methyl-
benzamide (23)
5-((4-methylbenzylamino)methyl)quinolin-8-ol
dihydrochlo-
ride (16), 5-((4-(trifluoromethyl)benzylamino)methyl)quinolin-
8-ol (17). Compounds (15), (16) and (17) were prepared following the
previously reported method described by Moret et al. [12].
0.998 g of aminomethyl-8-hydroxyquinoline (5.74 mmol) and
0.883 g of 4-methylbenzoyl chloride (5.74 mmol) were dissolved
by 20 mL of CH3CN. The reaction was heated at 50 ꢀC for two
days. Product purified by chromatography (SiO2 from CH2Cl2/
MeOH 95:5 to 80:20). Yield 75%. mp ¼ 137–139 ꢀC. 1H NMR
(250 MHz, CDCl3, ppm): 8.72 (d, 1H, J ¼ 3.75 Hz); 8.31 (d, 1H,
J ¼ 8.25 Hz); 8.11 (d, 2H, J ¼ 8.25 Hz), 7.60 (t, 1H, J ¼ 5.25 Hz), 7. 47
(d, 1H, J ¼ 7.5 Hz), 7.36 (d, 2H, J ¼ 7.5 Hz), 7.24 (d, 2H, J ¼ 8.25 Hz),
7.12 (d, 1H, J ¼ 8.25 Hz), 4.90 (d, 2H), 2.4 (s, 3H). Anal.
(C18H16N2O2)C, H, N.
5.2.13. 5-((3-(Trifluoromethyl)benzylamino)methyl)quinolin-8-ol (18)
0.76 g of 3-(trifluoromethyl)benzylamine, 1 g of 5-(chlor-
omethyl)quinolin-8-ol and 1 g of K2CO3 were added in 20 mL of
CH3CN. The reaction was heated at 50 ꢀC for one day. The product
was filtrated and washed by acetonitrile (2 ꢂ 20 mL) and ether
(40 mL). The solvent is concentrated under vacuum. Product puri-
fied by chromatography (SiO2 from DCM/MeOH 95:5 to DCM/
MeOH 80:20). Yield 28%. mp ¼ 127–129 ꢀC. 1H NMR (250 MHz,
MeOD): 8.86 (m, 1H), 8.52–8.49 (m, 1H), 7.42–7.38 (m, 2H), 7.35–
7.23 (m, 2H), 7.19–7.14 (m, 2H), 7.02–6.99 (m, 1H), 4.75 (s, 2H), 4.21
(s, 2H). MS, m/z (C18H15F3N2O): calcd 333.2 [M þ H]þ; found, 333.2.
Anal. (C18H15F3N2O) C, H, N.
5.3.6. Preparation of 5,50,500-nitrilotris(methylene)triquinolin-8-ol(24)
It was synthesized from 5-(aminomethyl)quinolin-8-ol and 5-
(chloromethyl)quinolin-8-ol following the general procedure A.
Yield 73%. mp ¼ 175–178 ꢀC. 1H NMR (250 MHz, MeOD): 8.82–8.79
(m, 3H), 8.59–8.39 (m, 3H), 7.56–7.51 (m, 3H), 7.45–7.42 (m, 3H),
7.05–7.01 (m, 3H), 3.34 (br s, 6H). MS, m/z (C30H24N4O3): calcd
489.18 [M þ H]þ; found, 489.1. Anal. (C30H24N4O3) C, H, N.
5.2.14. 5-((Dibenzylamino)methyl)quinolin-8-ol (19)
0.5 g of 5-(aminomethyl)quinolin-8-ol, 0.76 g of K2CO3 and
0.72 g of (chloromethyl)benzene were added in 20 mL of CH3CN.
The reaction was heated at 50 ꢀC for one day. The product was fil-
trated and washed by acetonitrile (2 ꢂ 20 mL) and ethyl acetate
(40 mL). The solvent is concentrated under vacuum. Product puri-
fied by chromatography (SiO2 from DCM/MeOH 95:5). Yield 75%.
mp ¼ 142–143 ꢀC. 1H NMR (250 MHz, MeOD): 8.79 (d, 2H, J ¼ 1 Hz),
8.39 (d, 1H, J ¼ 1 Hz), 7.47–7.01 (m, 13H), 3.82 (br s, 4H), 3.67 (s, 2H).
Anal. (C24H22N2O) C, H, N.
5.3.7. Preparation of 5,50-(4-(trifluoromethyl)benzylazanediyl)bis-
(methylene) diquinolin-8-ol (25)
0.1 g of 5,50-(4-(Trifluoromethyl)benzylazanediyl)bis(methylene)
diquinolin-8-ol and 0.018 g of NaH were dissolved in 15 mL of CH3CN.
The reaction was heated at 70 ꢀC for 3 h. Then the temperature was
decreased, and 0.578 g of iodomethane in 7 mL of CH3CN was added.
The reaction was heated again at 70 ꢀC for 2 h. The product was fil-
trated and washed by CH2Cl2 and CH3CN. Yield 81%. 1H NMR
(250 MHz, CDCl3): 8.73 (m, 2H), 7.51 (m, 4H), 7.41 (m, 2H), 7.17 (m,
2H), 7.07–7.12 (m, 2H), 4.05 (s, 6H), 3.84 (s, 4H), 3.55 (s, 2H). MS, m/z
(C30H26F3N3O2): calcd 518.2 [M þ H]þ; found, 518.2. Anal.
(C30H26F3N3O2) C, H, N.
5.3. General procedure B
5.3.1. Preparation of the Boc protected analogues 20–22
To a stirred solution of the suitable mono 8-hydroxyquinoline
derivatives (0.58 mmol) in CH2Cl2 (10 mL), Boc2O (0.58 mmol) was
added. The solution was stirred over night at room temperature
then washed with water (4 mL) and brine (2 mL). It was dried over
MgSO4 then concentrated in vaccuo and purified by
5.3.8. Preparation of 5,50-(4-(trifluoromethyl)benzylazanediyl)bis-
(methylene) diquinolin-8-ol (26)
Compound 26 was prepared following the previously reported
method described by Moret et al. [12].

636
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
5.4. Biological methods
5.5.2. Detection of adducts using electrospray ionization mass
spectroscopy (EI-MS) and MS/MS
5.4.1. Antitumor activity
The alkylation by compound 3 of desoxyguanosine, GSH or an
18-mer hairpin oligonucleotide (sequence described in [31]) was
monitored by EI-MS as described in [31,32,34]. Each potential target
5.4.1.1. In vitro KB3-1 cell line assay. Human mouth epidermal
carcinoma (KB) cell line was incubated for 72 h with different
concentrations of drugs. The final volume in each experiment was
prepared with the media containing 1% DMSO final volume.
Docetaxel was used as positive control. The experiments were
performed in triplicate. Cell growth inhibition was determined by
the MTS assay according to the recommendations of the manu-
facturer [Promega]. The optical density was measured at 490 nm.
The number of viable cells was proportional to the extent of for-
mazan production. The percent cytotoxicity index [(OD490 treated/
OD490 control) ꢂ 100] was calculated from three experiments. For
CC50 determinations, the cytotoxicity indices were plotted against
drug concentrations ranging from 0.05 to 10000 nM.
was incubated with the compound of interest (50
mM each) alone
(controls) or together in 200 L of 1 mM ammonium acetate, pH
m
7.15, prior to being injected in a simple-quadrupole mass spec-
trometer API 3000 (Perkin–Elmer Sciex) equipped with an ion
spray (nebulizer-assisted electrospray) source (Sciex, Toronto,
Canada) using a needle pre-washed with methanol. The solutions
were continuously infused with a medical infusion pump (Model
11, Harvard Apparatus, South Natick, USA) at a flow rate of 5 mL/min.
Polypropylene glycol was used to calibrate the quadrupole. Ion
spray mass spectra were acquired at unit resolution by scanning
from m/z 200 to 1000 with a step size of 0.1 Da and a dwell time of
2 ms. Twenty spectra were summed and recorded at an orifice
voltage of þ50 V whereas the potential of spray needle was held at
þ5 kV for all analyses, except for the oligonucleotide analysis which
was analyzed by scanning from m/z 600 to 1300 at an orifice voltage
of ꢃ60 V with the potential of spray needle held at ꢃ4.5 kV.
5.4.1.2. U87-GFP cell line, culture and assay. U87-gfp cells were
cultivated in MEM (Invitrogen) medium 10% FBS. Stable GFP
expressing U87 cells were generated by nucleofecting (Amaxa) cells
with the pEGFP-C1 vector (Clontech). U87-gfp cells were cultivated
in MEM (Invitrogen) medium 10% FBS.
For the scratch assay, U87-GFP cells were seeded in 96 well
black plates (GreinerBio-One) at a density of 3000 cells per well.
The following day (start of the experiment), compounds were
added to the cells and the cell layer was wounded in each well
using the VP 408FH device (VP-Scientific). An image of each well
was taken daily by the Flash-cytometer (Trophos, France) during
the course of the experiment (3 days). Images were then analyzed
using an application derived from the Metamorph software
(Universal Imaging) to determine cell density in the whole well and
in the scratch area at the different time-points. These values were
then used to characterize the effect of each compound on the
survival/proliferation and migration of cells.
5.5.3. Cytotoxic activity of compound 3 (JLK1486) upon GSH
depletion
Depletion of KB3-1 cells was performed upon treatment of the
cells (1 ꢂ106 cells/dish) by buthionine sulfoximine (BSO, Sigma, at
0.1; 0.25; 0.5; 1 or 10 mM) for 24 h as described in [34]. Cells were
then collected, washed with 10 mL PBS and resuspended in 200
of PBS prior to lysis by freezing and thawing twice. The ThioGlo-1Ô
(Calbiochem) reagent (20 M) was then added from a fresh 10 mM
mL
m
stock solution in DMSO to the lysate and the fluorescence was
immediately measured with an excitation wavelength of 360 nm
and an emission wavelength from 400 to 650 nm. The cytotoxicity
of compound 3 was assessed using the MTS cell proliferation assay
developed by Promega (CellTiter 96Ò AQueous one solution cell
proliferation assay). Briefly, 3 ꢂ 103 exponentially growing KB3-1
cells were seeded in 96-well microculture plates for 24 h prior to
being subjected to treatment with buthionine sulfoximine (BSO, at
the optimal concentration of 1 mM) for 24 h before subsequent
treatment using graded concentrations of compound 3 from 1 nM
5.5. Mechanism of action
5.5.1. Binding to DNA
Spectroscopic analyses (UV/visible absorbancy, melting
temperature studies and circular dichroism) were performed as
described in [30]. EMSA was performed as described in [31]. DNA
to 1
20
m
M in 6 independent points. After 72 h of incubation at 37 ꢀC,
m
L of the tetrazolium dye was added to each well and the
cleavage assays were developed as in [32]. Briefly, 8
labeled DNA, 2 L of buffer (10 mM ammonium acetate, pH 7.15),
and 10 L of the drug (from 2 to 100 M) were incubated in the dark
at room temperature for 2 or 24 h. For EMSA, 5 L of a 50% glycerol
mL of radio-
samples were incubated for a further 2 h at 37 ꢀC. Plates were
analyzed on a Labsystems Multiskan MS (type 352) reader at
492 nm.
m
m
m
m
solution containing tracking dyes was added to the DNA samples
prior to electrophoresis under non-denaturing conditions in 10%
polyacrylamide gels for about 5 h at 300 V at room temperature in
1X TBE buffer (89 mM boric acid, 2.5 mM Na2EDTA, pH 8.3). For
5.5.4. Microarrays
U87 cells were incubated in 0.5 mM compound 2 for 3 or 20 h.
After incubation treated and control cells were harvested and RNA
was extracted by Trizol. Micro-array experiments were performed
by Genoscience (France) using AGILENT arrays representing around
10,000 genes. Data analysis was performed using the Genoscience
proprietary software ‘‘PredictSearch’’.
cleavage assays, 4 mL of denaturing loading buffer (80% formamide
solution containing tracking dyes) was added. The DNA samples
were heated at 90 ꢀC for 4 min and chilled on ice for 4 min prior to
electrophoresis. DNA cleavage products were resolved under
polyacrylamide electrophoresis in denaturing conditions (0.35 mm
thick, 8% polyacrylamide containing 8 M urea). After 1.5 h of elec-
trophoresis at 65 W in 1X TBE buffer, gels were soaked in 10% acetic
acid for 10 min. Both gels were transferred to Whatman 3 MM
paper, dried under vacuum at 80 ꢀC, and then analyzed on a phos-
phorimager (Molecular Dynamics 445SI). For DNA cleavage assays,
each resolved band was assigned to a particular bond within the
DNA fragments by comparison of its position relative to sequencing
standards generated by treatment of the DNA with dimethylsulfate
(DMS) followed by piperidine-induced cleavage at the modified
guanine bases.
Acknowledgments
The authors thank G. Garcia and F. Gaubert for excellent tech-
nical assistance and Prof. K. Dudley for critical reading of the
manuscript. This work was supported by INCA (Institut National du
Cancer), CNRS (Centre National de la Recherche Scientifique),
IBDML (Institut de Biologie du Developpement de Marseille
Luminy) and Fonds Yvonne Boe¨l (Brussels, Belgium). Marie-Helene
´
´ `
´
David-Cordonnier thanks the Service de Spectrometrie de Masse de
´
l’Universite de Lille 2 (Dr Mostafa Kouach and Dr Gilbert Briand) for
access to the simple-quadrupole mass spectrometer API 3000

S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
637
[20] J.L. Bolton, L. Shen, p-Quinone methides are the major decomposition products
of catechol estrogen o-quinones. Carcinogenesis 17 (1996) 925–929.
[21] Q. Zhou, S.E. Rokita, A general strategy for target-promoted alkylation in
biological systems. Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 15452–15457.
[22] Y. Chiang, A.J. Kresge, Alkylation of guanosine and 20-deoxyguanosine by
o-quinone alpha-(p-anisyl)methide in aqueous solution. Org. Biomol. Chem. 2
(2004) 1090–1092.
´
equipment, the Ligue Nationale Contre le Cancer, Comite du Nord,
´
´
for a grant and the Universite de Lille 2/Conseil Regional Nord/Pas-
de-Calais for a PhD fellowship to Gae¨lle Lenglet.
Appendix. Supplementary data
[23] Y. Song, T. Tian, P. Wang, H. He, W. Liu, X. Zhou, X. Cao, X.L. Zhang, X. Zhou,
Phenol quaternary ammonium derivatives: charge and linker effect on their
DNA photo-inducible cross-linking abilities. Org. Biomol. Chem. 4 (2006)
3358–3366.
Supplementary data associated with this article can be found in
the online version, at doi:10.1016/j.ejmech.2009.11.006
[24] D.C. Thompson, K. Perera, R. London, Spontaneous hydrolysis of 4-tri-
fluoromethylphenol to a quinone methide and subsequent protein alkylation.
Chem. Biol. Interact. 126 (2000) 1–14.
References
[25] M.A. Lewis, D.G. Yoerg, J.L. Bolton, J.A. Thompson, Alkylation of 20-deoxy-
nucleosides and DNA by quinone methides derived from 2,6-di-tert-butyl-4-
methylphenol. Chem. Res. Toxicol. 9 (1996) 1368–1374.
[1] R. Balamurugan, F.J. Dekker, H. Waldmann, Design of compound libraries
based on natural product scaffolds and protein structure similarity clustering
(PSSC). Mol. Biosyst. 1 (2005) 36–45.
[2] D.A. Horton, G.T. Bourne, M.L. Smythe, The combinatorial synthesis of
bicyclic privileged structures or privileged substructures. Chem. Rev. 103
(2003) 893–930.
[26] P. Pande, J. Shearer, J.H. Yang, W.A. Greenberg, S.E. Rokita, Alkylation of nucleic
acids by a model quinone methide. J. Am. Chem. Soc. 121 (1999) 6773–6779.
[27] M.A. Brimble, M.R. Nairn, Reductive thioalkylation of
thoquinone. J. Chem. Society-Perkin Trans. 1 (2000) 317–322.
a pyranonaph-
[28] J.L. Bolton, S.B. Turnipseed, J.A. Thompson, Influence of quinone methide reac-
tivity on the alkylation of thiol and amino groups in proteins: studies utilizing
amino acid and peptide models. Chem. Biol. Interact. 107 (1997) 185–200.
[29] R.M. Williams, B. Herberich, DNA interstrand cross-link formation induced by
bioxalomycin A2. J. Am. Chem. Soc. 120 (1998) 10272–10273.
[3] R.E. Ziegert, J. Torang, K. Knepper, S. Brase, The recent impact of solid-phase
synthesis on medicinally relevant benzoannelated oxygen heterocycles. J.
Comb. Chem. 7 (2005) 147–169.
[4] S.L. Schreiber, Target-oriented and diversity-oriented organic synthesis in
drug discovery. Science 287 (2000) 1964–1969.
[30] M.H. David-Cordonnier, M.P. Hildebrand, B. Baldeyrou, A. Lansiaux, C. Keuser,
K. Benzschawel, T. Lemster, U. Pindur, Design, synthesis and biological eval-
uation of new oligopyrrole carboxamides linked with tricyclic DNA-inter-
calators as potential DNA ligands or topoisomerase inhibitors. Eur. J. Med.
Chem. 42 (2007) 752–771.
[5] M.D. Burke, S.L. Schreiber, A planning strategy for diversity-oriented synthesis.
Angew. Chem. Int. Ed. Engl. 43 (2004) 46–58.
[6] T.U. Mayer, T.M. Kapoor, S.J. Haggarty, R.W. King, S.L. Schreiber, T.J. Mitchison,
Small molecule inhibitor of mitotic spindle bipolarity identified in a pheno-
type-based screen. Science 286 (1999) 971–974.
[31] M.H. David-Cordonnier, W. Laine, A. Lansiaux, M. Kouach, G. Briand, A. Pierre,
J.A. Hickman, C. Bailly, Alkylation of guanine in DNA by S23906-1, a novel
potent antitumor compound derived from the plant alkaloid acronycine.
Biochemistry 41 (2002) 9911–9920.
[7] H.E. Pelish, N.J. Westwood, Y. Feng, T. Kirchhausen, M.D. Shair, Use of biomi-
metic diversity-oriented synthesis to discover galanthamine-like molecules
with biological properties beyond those of the natural product. J. Am. Chem.
Soc. 123 (2001) 6740–6741.
[32] M.H. David-Cordonnier, M. Casely-Hayford, M. Kouach, G. Briand,
L.H. Patterson, C. Bailly, M. Searcey, Stereoselectivity, sequence specificity and
mechanism of action of the azinomycin epoxide. Chembiochem 7 (2006)
1658–1661.
[33] N.H. Cnubben, A.J. Rommens, M.J. Oudshoorn, P.J. Van Bladeren, Glutathione-
dependent biotransformation of the alkylating drug thiotepa and transport of
its metabolite monoglutathionylthiotepa in human MCF-7 breast cancer cells.
Cancer Res. 58 (1998) 4616–4623.
[34] M.H. David-Cordonnier, W. Laine, A. Joubert, C. Tardy, J.F. Goossens, M. Kouach,
G. Briand, H.D. Thi Mai, S. Michel, F. Tillequin, M. Koch, S. Leonce, A. Pierre,
C. Bailly, Covalent binding to glutathione of the DNA-alkylating antitumor
agent, S23906-1. Eur. J. Biochem. 270 (2003) 2848–2859.
[8] P.A. Adlard, R.A. Cherny, D.I. Finkelstein, E. Gautier, E. Robb, M. Cortes,
I. Volitakis, X. Liu, J.P. Smith, K. Perez, K. Laughton, Q.X. Li, S.A. Charman,
J.A. Nicolazzo, S. Wilkins, K. Deleva, T. Lynch, G. Kok, C.W. Ritchie, R.E. Tanzi,
R. Cappai, C.L. Masters, K.J. Barnham, A.I. Bush, Rapid restoration of cognition
in Alzheimer’s transgenic mice with 8-hydroxy quinoline analogs is associated
with decreased interstitial. Abeta. Neuron 59 (2008) 43–55.
[9] P.E. Thompson, J.W. Reinertson, Chemotherapy of amebic hepatitis in hamsters
with emetine, chloroquine, amodiaquin (camoquin), quinacrine and other
drugs. Am. J. Trop. Med. Hyg. 31 (1951) 707–717.
[10] N.L. Oien, R.J. Brideau, T.A. Hopkins, J.L. Wieber, M.L. Knechtel, J.A. Shelly,
R.A. Anstadt, P.A. Wells, R.A. Poorman, A. Huang, V.A. Vaillancourt, T.L. Clayton,
J.A. Tucker, M.W. Wathen, Broad-spectrum antiherpes activities of 4-hydrox-
yquinoline carboxamides, a novel class of herpesvirus polymerase inhibitors.
Antimicrob. Agents Chemother. 46 (2002) 724–730.
[35] H.H. Bailey,
L-S,R-buthionine sulfoximine: historical development and clinical
issues. Chem. Biol. Interact. 111–112 (1998) 239–254.
[36] S. Leonce, A. Pierre, M. Anstett, V. Perez, A. Genton, J.P. Bizzari, G. Atassi, Effects
of a new triazinoaminopiperidine derivative on adriamycin accumulation and
retention in cells displaying P-glycoprotein-mediated multidrug resistance.
Biochem. Pharmacol. 44 (1992) 1707–1715.
[37] Y. Wu, H. Zhang, Y. Dong, Y.M. Park, C. Ip, Endoplasmic reticulum stress signal
mediators are targets of selenium action. Cancer Res. 65 (2005) 9073–9079.
[38] E. Isaki, T.J. Spaulding, E. Plante, Contributions of language and memory
demands to verbal memory performance in language-learning disabilities. J.
Commun. Disord. 41 (2008) 512–530.
[39] K. Zhang, R.J. Kaufman, From endoplasmic-reticulum stress to the inflamma-
tory response. Nature 454 (2008) 455–462.
[40] M. Schroder, Endoplasmic reticulum stress responses. Cell. Mol. Life Sci. 65
(2008) 862–894.
[41] A. Chakravarti, G. Zhai, Y. Suzuki, S. Sarkesh, P.M. Black, A. Muzikansky,
J.S. Loeffler, The prognostic significance of phosphatidylinositol 3-kinase
pathway activation in human gliomas. J. Clin. Oncol. 22 (2004) 1926–1933.
[42] C.H. Jakobsen, G.L. Storvold, H. Bremseth, T. Follestad, K. Sand, M. Mack,
K.S. Olsen, A.G. Lundemo, J.G. Iversen, H.E. Krokan, S.A. Schonberg, DHA
induces ER stress and growth arrest in human colon cancer cells: associations
with cholesterol and calcium homeostasis. J. Lipid Res. 49 (2008) 2089–2100.
[43] W.D. Holtzclaw, A.T. Dinkova-Kostova, P. Talalay, Protection against elec-
trophile and oxidative stress by induction of phase 2 genes: the quest for
the elusive sensor that responds to inducers. Adv. Enzym. Regul. 44 (2004)
335–367.
[44] T.S. Dowers, Z.H. Qin, G.R. Thatcher, J.L. Bolton, Bioactivation of selective
estrogen receptor modulators (SERMs). Chem. Res. Toxicol. 19 (2006)
1125–1137.
[45] T. Dunlap, R.E. Chandrasena, Z. Wang, V. Sinha, Z. Wang, G.R. Thatcher,
Quinone formation as a chemoprevention strategy for hybrid drugs: balancing
cytotoxicity and cytoprotection. Chem. Res. Toxicol. 20 (2007) 1903–1912.
[46] S. Kar, I.M. Lefterov, M. Wang, J.S. Lazo, C.N. Scott, C.S. Wilcox, B.I. Carr, Binding
and inhibition of Cdc25 phosphatases by vitamin K analogues. Biochemistry
42 (2003) 10490–10497.
[11] H.F. Ji, H.Y. Zhang, A new strategy to combat Alzheimer’s disease. Combining
radical-scavenging potential with metal-protein-attenuating ability in one
molecule. Bioorg. Med. Chem. Lett. 15 (2005) 21–24.
[12] V. Moret, Y. Laras, T. Cresteil, G. Aubert, Q.P. D, C. Di, M. Barthelemy-Requin,
C. Beclin, V. Peyrot, D. Allegro, A. Rolland, F. De Angelis, E. Gatti, P. Pierre,
L. Pasquini, E. Petrucci, U. Testa, J.L. Kraus, Discovery of a new family of bis-8-
hydroxyquinoline substituted benzylamines with pro-apoptotic activity in
cancer cells: synthesis, structure-activity relationship, and action mechanism
studies. Eur. J. Med. Chem. 44 (2009) 558–567.
[13] L. Fan, W. Zhu, J. Li, H. Tian, Novel red-light emitting metal complex based on
asymmetric perylene bisimide and 8-hydroxyquinoline dyads. Synth. Met. 145
(2004) 203–210.
[14] E.S. Newlands, M.F.G. Stevens, S.R. Wedge, R.T. Wheelhouse, C. Brock, Temo-
zolomide: a review of its discovery, chemical properties, pre-clinical devel-
opment and clinical trials. Cancer Treat. Rev. 23 (1997) 35–61.
[15] L. Ingrassia, P. Nshimyumukiza, J. Dewelle, F. Lefranc, L. Wlodarczak,
S. Thomas, G. Dielie, C. Chiron, C. Zedde, P. Tisnes, R. van Soest, J.C. Braekman,
F. Darro, R. Kiss, A lactosylated steroid contributes in vivo therapeutic benefits
in experimental models of mouse lymphoma and human glioblastoma. J. Med.
Chem. 49 (2006) 1800–1807.
[16] E. Van Quaquebeke, T. Mahieu, P. Dumont, J. Dewelle, F. Ribaucour, G. Simon,
S. Sauvage, J.F. Gaussin, J. Tuti, M. El Yazidi, F. Van Vynckt, T. Mijatovic,
F. Lefranc, F. Darro, R. Kiss, 2,2,2-Trichloro-N-({2-[2-(dimethylamino)ethyl]-
1,3-dioxo-2,3-dihydro-1H-be nzo[de]isoquinolin-5-yl}carbamoyl)acetamide
(UNBS3157), a novel nonhematotoxic naphthalimide derivative with potent
antitumor activity. J. Med. Chem. 50 (2007) 4122–4134.
[17] D. Chen, Q.C. Cui, H. Yang, R.A. Barrea, F.H. Sarkar, S. Sheng, B. Yan, G.P. Reddy,
Q.P. Dou, Clioquinol, a therapeutic agent for Alzheimer’s disease, has protea-
some-inhibitory, androgen receptor-suppressing, apoptosis-inducing, and
antitumor activities in human prostate cancer cells and xenografts. Cancer Res.
67 (2007) 1636–1644.
[18] L.H. Hurley, V.L. Reynolds, D.H. Swenson, G.L. Petzold, T.A. Scahill, Reaction of
the antitumor antibiotic CC-1065 with DNA: structure of a DNA adduct with
DNA sequence specificity. Science 226 (1984) 843–844.
[47] Y. Nishikawa, B.I. Carr, M. Wang, S. Kar, F. Finn, P. Dowd, Z.B. Zheng, J. Kerns,
S. Naganathan, Growth inhibition of hepatoma cells induced by vitamin K and
its analogs. J. Biol. Chem. 270 (1995) 28304–28310.
[19] T. Li, Q. Zeng, S.E. Rokita, Target-promoted alkylation of DNA. Bioconjugate
Chem. 5 (1994) 497–500.

638
S. Madonna et al. / European Journal of Medicinal Chemistry 45 (2010) 623–638
[48] T.C. McMorris, Discovery and development of sesquiterpenoid derived
hydroxymethylacylfulvene: a new anticancer drug. Bioorg. Med. Chem. 7
(1999) 881–886.
[49] S.E. Wolkenberg, D.L. Boger, Mechanisms of in situ activation for DNAtargeting
antitumor agents. Chem. Rev. 102 (2002) 2477–2495.
[50] C. Geroni, S. Marchini, P. Cozzi, E. Galliera, E. Ragg, T. Colombo, R. Battaglia,
M. Howard, M. D’Incalci, M. Broggini, Brostallicin, a novel anticancer agent
whose activity is enhanced upon binding to glutathione. Cancer Res. 62 (2002)
2332–2336.
[51] M. Sharma, M. Tomasz, Conjugation of glutathione and other thiols with
bioreductively activated mitomycin C. Effect of thiols on the reductive acti-
vation rate. Chem. Res.Toxicol 7 (1994) 390–400.
dominant platinum-thiol binding to its anticancer action. Pharm. Res. 19
(2002) 124–131.
[53] A. Bernareggi, L. Torti, R.M. Facino, M. Carini, G. Depta, B. Casetta, N. Farrell,
S. Spadacini, R. Ceserani, S. Tognella, Characterization of cisplatinglutathione
adducts by liquid chromatography-mass spectrometry. Evidence for their
formation in vitro but not in vivo after concomitant administration of cisplatin
and glutathione to rats anc cancer patients. J. Chromatogr. B. Biomed. Appl.
669 (1995) 247–263.
[54] H.A. Dirven, B. van Ommen, P.J. van Bladeren, Glutathione conjugation of
alkylating cytostatic drugs with a nitrogen mustard group and the role of
glutathione S-transferases. Chem. Res.Toxicol 9 (1996) 351–360.
[55] J. Rothbarth, A.L. Vahrmeijer, G.J. Mulder, Modulation of cytostatic efficacy of
melphalan by glutathione: mechanisms and efficacy. Chem. Biol. Interact 140
(2002) 93–107.
[52] E. Volckova, L.P. Dudones, R.N. Bose, HPLC determination of binding of
cisplatin to DNA in the presence of biological thiols: implications of