Toward a rational design of multitarget-directed antioxidants: Merging memoquin and lipoic acid molecular frameworks

Article abstract of DOI:10.1021/jm901123n

Novel multitargeted antioxidants 3 - 6 were designed by combining the antioxidant features, namely, a benzoquinone fragment and a lipoyl function, of two multifunctional lead candidates. They were then evaluated to determine their profile against Alzheime

Full text of DOI:10.1021/jm901123n

J. Med. Chem. 2009, 52, 7883–7886 7883
DOI: 10.1021/jm901123n
Toward a Rational Design of Multitarget-Directed Antioxidants: Merging Memoquin and Lipoic Acid
Molecular Frameworks
Maria Laura Bolognesi,*,† Andrea Cavalli,†,§ Christian Bergamini,‡ Romana Fato,‡ Giorgio Lenaz,‡ Michela Rosini,†
Manuela Bartolini,† Vincenza Andrisano,† and Carlo Melchiorre†
†Department of Pharmaceutical Sciences, University of Bologna, Via Belmeloro 6, 40126 Bologna, Italy, §Department of Drug Discovery and
Development, Italian Institute of Technology, Via Morego 30, 16163 Genova, Italy, and ‡Department of Biochemistry “G. Moruzzi”, University of
Bologna, Via Irnerio 48, 40126 Bologna, Italy
Received May 31, 2009
Novel multitargeted antioxidants 3-6 were designed by combining the antioxidant features, namely, a
benzoquinone fragment and a lipoyl function, of two multifunctional lead candidates. They were then
evaluated to determine their profile against Alzheimer’s disease. They showed antioxidant activity,
improved following enzymatic reduction, in mitochondria and T67 cell line. They also displayed a
balanced inhibitory profile against amyloid-β aggregation and acetylcholinesterase, emerging as
promising molecules for neuroprotectant lead discovery.
Experimental evidence has greatly expanded our under-
standing of the role of oxidative stress (OSa) in the pathoge-
nesis of Alzheimer’s disease (AD). This evidence comes from
the established interconnections among OS and other key
events of AD, such as apoptosis, amyloid-β (Aβ) processing
and secretion, τ phosphorylation, and the disruption of Ca2þ
homeostasis.1 The brain exhibits particularly high sensitivity
to OS, and the brains of AD patients reveal loss of synapses
and OS damage by reactive oxygen species (ROS).2 In trans-
genic mice, Aβ colocalizes with several OS markers confir-
ming the in vivo link between Aβ deposition and oxidative
damage.2 OS promotes Aβ toxicity through the production of
free radicals,2 and the application of Aβ to neuronal cultures
elevates intracellular H2O2.2 Intriguingly, ROS generation is
not just an outcome of the disease process but actually
precedes extracellular amyloid deposition as one of the initial
events in the AD cascade.3 OS elevates β-secretase (BACE-1)
expression and activity and Aβ levels in vivo4 and triggers the
amyloidogenicpathway in human cell lines.5 These data imply
that not only can Aβ induce OS but its generation is also
increased as a consequence of OS, thus forming a vicious
circle.
mitochondria-targeted antioxidant and a precursor of an
essential cofactor for mitochondrial enzymes, may be useful
for preventing or delaying mitochondrial decay, thus prevent-
ing or treating AD and related dementias.8 However, the
disappointing results of clinical trials of these and other
antioxidants suggest that single antioxidants may not be
adequate for the treatment of neurodegenerative diseases
while optimal combinations of antioxidants may provide a
more effective strategy.9
We have proposed that additive neuroprotective effects can
be produced by a single molecule endowed with antioxidant
properties and are able to act at different steps in the neuro-
degenerative process. This concept, termed multitarget-
directed ligands (MTDLs) design strategy,10,11 has been em-
bodied in the development of two drug candidates memoquin
(1) and lipocrine (2), which display a multitarget profile
against AD. 1, which bears the benzoquinone moiety of
CoQ inserted into a polyamine chain, is able to modulate in
vitro and in vivo different AD crucial molecular targets, such
as acetylcholinesterase (AChE) and BACE-1 enzymes, Aβ,
and oxidative processes.12 A 9-amino-6-chloro-1,2,3,4-tetra-
hydroacridine moiety coupled with LA led to 2. In addition to
its AChE inhibitory activity, 2 also behaved as an antioxidant
(Chart 1).13
Following the same rationale, we reasoned that merging the
antioxidant features of 1 and 2 in a single chemical entity
could, hopefully, lead to ligands (3-6) with multiple antioxi-
dant mechanisms, hence a better potential for AD prevention
and therapy. These hybrid molecules can be defined as multi-
functional antioxidants, i.e., MTDLs with other pharmaco-
logical effects in addition to their antioxidant activity. Since
LA14 and CoQ15 have shown antiaggregating properties, the
new compounds 3-6 may also have intrinsic antiamyloido-
genic capacities. Finally, since LA reduces CoQ to ubiquinol
(the actual antioxidant form) by the transfer of a pair of
electrons,16 they should in principle possess a more effective
antioxidant profile in vivo.
In this context, on the basis of what is defined as the OS
theory of AD,6 antioxidants may bekey factors in fighting this
disease. Coenzyme Q (CoQ) is an endogenous antioxidant
and a powerful free radical scavenger in mitochondrial mem-
branes. Synthesis of CoQ decreases with aging, and CoQ has
potentially neuroprotective effects in aging and neurodegen-
erative diseases in which excessive OS is present.7 Simila-
rly, supplementation with R-lipoic acid (LA), which is a
*To whom correspondence should be addressed. Phone: þ39
0512099718. Fax: þ39 0512099734. E-mail: marialaura.bolognesi@unibo.it.
a Abbreviations: AChE, acetylcholinesterase; LA, R-lipoic acid; AD,
Alzheimer’s disease; Aβ, amyloid-β; BChE, butyrylcholinesterase;
BACE-1, β-secretase; CoQ, coenzyme Q; MTDLs, multitarget-directed
ligands; NQO1, NAD(P)H/quinone oxidoreductase 1; OS, oxidative
stress; ROS, reactive oxygen species; SMP, submitochondrial particles.
r
2009 American Chemical Society
Published on Web 10/08/2009
pubs.acs.org/jmc

7884 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Chart 1. Design Strategy for Hybrids 3-6
Bolognesi et al.
Scheme 1a
a Reaction conditions: (a) CH2Cl2, air, room temp, 6 h; (b) NH2(CH2)nNHBOC (12, n = 3; 13, n = 6), CH2Cl2, room temp, overnight; (c) CF3COOH,
CH2Cl2, room temp, 4 h; (d) LA, EDCI, HOBt, NEt3, CH2Cl2, room temp, overnight.
Results and Discussion
Because of the heterodimeric structure, 3-6 could not be
synthesized following the efficient convergent synthetic
protocol developed for 1 and analogues, which relies on
the disubstitution reaction of diamines with 2,5-dimethoxy-
1,4-benzoquinone (7).17 Conversely, a stepwise linear synth-
esis approach was developed (Scheme 1). Monosubstitution
of 7 was the key step of the new synthetic strategy. Thus,
reaction of 8 equiv of 7 with diamines 817 and 917 gave
monosubstituted derivatives 10 and 11 in acceptable yield.
The second methoxy group could then participate in a sub-
sequent substitution reaction with the monoprotected dia-
mines 12 and 13. The ready removal in mildly acidic
conditions, compatible with the stability of the ami-
no-quinone bond, made BOC a suitable protecting group
to obtain 18-21, by treating 14-17 with CF3COOH. Since in
linear synthesis the overall yield quickly drops with each step
First, we replaced one terminal 2-methoxybenzyl func-
tion of 1 with an LA fragment to afford lipoamide-related
compounds. LA bound in amide linkage is the essential
cofactor in the mitochondrial oxoacid dehydrogenase
complexes, and lipoamide was even more efficient than
LA in protecting cells against oxidative insults.8 Clearly,
the same amido functionality is present in 2. The possibility
of having a cationic head, fundamental for recognizing
some of the selected molecular targets (in cholinesterases
molecular recognition is critically mediated by a Trp),17 is
guaranteed by the presence of a remaining benzylamino
group in 3-6 (Chart 1). A three- or six-methylene spacer
between the nitrogen atoms was used because this chain
length was shown to confer a better MTDL profile in SAR
studies on 1.17

Brief Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7885
Figure 1. Effect of 3-6 on ROS production in SMP induced by
2 μM antimycin A (AA) using NADH as substrate. Experiments
were performed with compounds in oxidized and reduced forms.
Reduction was achieved by 30 min of preincubation with purified
human NQO1 using NADH as substrate. After this time the
reaction was stopped by adding 10 μM dicumarol: CTRL=control;
(/) p e 0.05 with respect to AA treated samples.
and since the method reported for 213 was not very efficient,
efforts were made to optimize the coupling conditions of
18-21 to LA. After variation of solvents and reaction times,
it was found that addition of HOBT and NEt3 gave the best
yield to 3-6.
Figure 2. Effect of 1-6 on ROS formation in T67 cells. The
antioxidant activity was evaluated against ROS formation induced
by t-BuOOH. Experiments were performed with T67 cells treated or
not with 2.5 μM sulforaphane (SFP): CTRL=control; (/) p e 0.05
with respect to t-BuOOH treated samples; (#) p e 0.05 with respect
to t-BuOOH þ SFP treated samples.
Table 1. Inhibition of AChE and BChE Activities and Self-Induced Aβ
Aggregation by 3-6 and Reference Compounds 1, 2, and Propidium
The presence in 3-6 of the benzoquinone and lipoyl
moieties of CoQ and LA, respectively, suggests that their
antioxidant action may be targeting mitochondria, which
are structurally and functionally damaged in AD.18 Thus,
their antioxidant activity was first tested on bovine heart
submitochondrial particles (SMP). Addition of antimycin
A to SMP treated with NADH induced a strong increase in
ROS production. A 10 μM concentration of 3-6 in their
quinone form significantly decreased ROS production.
This was even more pronounced in their reduced form,
after reduction by isolated NAD(P)H/quinone oxidoreduc-
tase 1 (NQO1) (Figure 1). This deserves comment. We have
previously demonstrated that the antioxidant property of 1,
in analogy with CoQ, pertains mainly to its hydroquinone
form.17 NQO1, an inducible enzyme that catalyzes the two-
electron reduction of quinones to hydroquinones, was
shown to be responsible for the generation of the CoQ-
reduced state, as well as that of 1. More interestingly,
NQO1 is increased in AD19 in response to the OS typical
of the pathology. In light of this, we advanced that, being
specifically reduced by NQO1 into the corresponding hy-
droquinone, 1 may exert its antioxidant activity specifically
in those brain regions affected by AD, where a colocaliza-
tion of NQO1 activity with AD hallmarks coexists.20 There-
fore, since 1 and 3-6 share the same benzoquinone nucleus,
their antioxidant action might be similarly mediated by
NQO1, being able to convert them into the more-antiox-
idant hydroquinone form. This would explain the higher
antioxidant potential of 3-6, which was observed following
NQO1 reduction (see Supporting Information (SI)). To
further explore the antioxidant potential of 3-6, cell via-
bility and neuroprotective activity against OS were assayed
in the human glioma cell line T67, following treatment with
t-BuOOH and in the absence or presence of pretreatment
with sulforaphane (Figure 2). As previously verified,17
sulforaphane was able to increase NQO1 activity by 50%
with respect to control cells (Figure 1 in SI). Figure 2 clearly
IC50 a ( SEM (μM)
inhibition of Aβ
aggregationb (
SEM (%)
compd
m
n
AChE
BChE
1
2
(1.5 ( 0.11) ꢀ 10-3 1.44 ( 0.1
66.8 ( 4.4c
0.011 ( 0.003 <10
(0.253 ( 0.016) ꢀ
10-3
3
4
5
6
4
4
1
1
4
1
4
1
0.10 ( 0.01
0.15 ( 0.01
0.16 ( 0.01
0.19 ( 0.01
4.99 ( 0.13
21.0 ( 1.20
11.1 ( 0.80
212 ( 76
45.4 ( 0.1
34.4 ( 1.8
31.8 ( 2.1
19.2 ( 2.3
61.1 ( 4.6c
propidium
a Human recombinant AChE and BChE from human serum were
used. IC50 values represent the concentration of inhibitor required to
decrease enzyme activity by 50% and are the mean of two independent
measurements, each performed in triplicate. b Inhibition of self-induced
Aβ(1-42) aggregation (50 μM) produced by the tested compound at
10 μM. SEM=standard error of the mean. c Data from ref 17.
shows that 3-6 (at 10 μM) in their oxidized form show a
basal antioxidant activity, which is probably due to the LA
fragment that does not require activation by NQO1. This
activity was increased in cells pretreated with sulforaphane,
confirming that NQO1 is involved in the activation of these
compounds. As expected, the antioxidant activity of 2 is not
influenced by the overexpression of NQO1.
In light of the antiaggregating properties of CoQ15 and LA,14
the ability of 2 and 3-6 to reduce Aβ(1-42) self-aggregation was
also studied, using 1 and propidium as reference compounds.
Data in Table 1 show that 3-6 at 10 μM inhibited Aβ
aggregation in a range from 19% to 45%, whereas 2 at that
concentration displayed a negligible activity, perhaps because,
compared to LA, 2 lacks the free carboxylic group, an essential
feature in some potent aggregation inhibitors.21
Since 1 and 2 are effective cholinesterases inhibitors, to
achieve a wider picture of the multifunctional profile, 3-6
were evaluated for their activity against AChE, a validated
molecular target in AD, and butyrylcholinesterase (BChE)

7886 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Bolognesi et al.
(Table 1). All compounds were fair inhibitors of AChE and
BChE and markedly less potent than both prototypes, sug-
gesting that the insertionof the lipoyl fragment atposition2 of
the benzoquinone resulted in a highly negative effect on the
interaction with the enzyme. This was not unexpected because
the second protonated nitrogen of 1 is missed in 3-6. How-
ever, we wanted to achieve a balanced multitarget profile
rather than a highly potent AChE inhibitor. Where connec-
tions exist between two or more targets, multifunctional
ligands with only modest activity at one or more targets
may still produce superior in vivo effects compared to high-
er-affinity target-selective compounds.22 Because most links
in cellular networks are weak, a low-affinity MTDL may be
enough to accomplish a significant modification.23
biological methods. This material is available free of charge via
the Internet at http://pubs.acs.org.
References
(1) Smith, M. A.; Perry, G.; Pryor, W. A. Causes and consequences of
oxidative stress in Alzheimer’s disease. Free Radical Biol. Med.
2002, 32, 1049.
(2) Chong, Z. Z.; Li, F.; Maiese, K. Oxidative stress in the brain: novel
cellular targets that govern survival during neurodegenerative
disease. Prog. Neurobiol. 2005, 75, 207–246.
(3) Zhu, X.; Su, B.; Wang, X.; Smith, M. A.; Perry, G. Causes of
oxidative stress in Alzheimer disease. Cell. Mol. Life Sci. 2007, 64,
2202–2210.
(4) Xiong, K.; Cai, H.; Luo, X. G.; Struble, R. G.; Clough, R. W.; Yan,
X. X. Mitochondrial respiratory inhibition and oxidative stress
elevate beta-secretase (BACE1) proteins and activity in vivo in the
rat retina. Exp. Brain Res. 2007, 181, 435–446.
In conclusion, we discovered new multitargeted antioxi-
dants that might represent a step forward in the search for new
MTDLs against AD. 3-6 were superior to 1, for the basal
antioxidant activity (not dependentbyNQO1activation), and
to 2 for their additional ability to inhibit Aβ aggregation.
Moreover, they displayed balanced activityprofiles, which are
promising if properly corroborated by the definitive in vivo
proof of principle.
All these considerations establish the new hybrids as pro-
mising molecular structures for the development of new AD
agents that, beyond the modulation of OS, might have extra
beneficial activities. In particular, the simultaneous modula-
tion of brain levels of Aβ and ROS might break the vicious
cycle that further promotes OS, elevates Aβ levels, and
accelerates AD development. Clearly, proof of the concept
willinvolveaninvestigationof theirneuroprotectant profile in
vivo.
(5) Coma, M.; Guix, F. X.; Ill-Raga, G.; Uribesalgo, I.; Alameda, F.;
Valverde, M. A.; Munoz, F. J. Oxidative stress triggers the
amyloidogenic pathway in human vascular smooth muscle cells.
Neurobiol. Aging 2008, 29, 969–980.
(6) Markesbery, W. R. Oxidative stress hypothesis in Alzheimer’s
disease. Free Radical Biol. Med. 1997, 23, 134–147.
(7) Wadsworth, T. L.; Bishop, J. A.; Pappu, A. S.; Woltjer, R. L.;
Quinn, J. F. Evaluation of coenzyme Q as an antioxidant strategy
for Alzheimer’s disease. J. Alzheimer’s Dis. 2008, 14, 225–234.
(8) Holmquist, L.; Stuchbury, G.; Berbaum, K.; Muscat, S.; Young,
S.; Hager, K.; Engel, J.; Munch, G. Lipoic acid as a novel treatment
for Alzheimer’s disease and related dementias. Pharmacol. Ther.
2007, 113, 154–164.
(9) Liu, J. The effects and mechanisms of mitochondrial nutrient alpha-
lipoic acid on improving age-associated mitochondrial and cognitive
dysfunction: an overview. Neurochem. Res. 2008, 33, 194–203.
(10) Cavalli, A.; Bolognesi, M. L.; Minarini, A.; Rosini, M.; Tumiatti, V.;
Recanatini, M.; Melchiorre, C. Multi-target-directed ligands to com-
bat neurodegenerative diseases. J. Med. Chem. 2008, 51, 347–372.
(11) Bolognesi, M. L.; Rosini, M.; Andrisano, V.; Bartolini, M.;
Minarini, A.; Tumiatti, V.; Melchiorre, C. MTDL design strategy
in the context of Alzheimer’s disease: from lipocrine to memoquin
and beyond. Curr. Pharm. Des. 2009, 15, 601–613.
(12) Cavalli, A.; Bolognesi, M. L.; Capsoni, S.; Andrisano, V.;
Bartolini, M.; Margotti, E.; Cattaneo, A.; Recanatini, M.;
Melchiorre, C. A small molecule targeting the multifactorial nature
of Alzheimer’s disease. Angew. Chem., Int. Ed. 2007, 46, 3689–3692.
(13) Rosini, M.; Andrisano, V.; Bartolini, M.; Bolognesi, M. L.; Hrelia,
P.; Minarini, A.; Tarozzi, A.; Melchiorre, C. Rational approach to
discover multipotent anti-Alzheimer drugs. J. Med. Chem. 2005,
48, 360–363.
(14) Ono, K.; Hirohata, M.; Yamada, M. Alpha-lipoic acid exhibits
anti-amyloidogenicity for beta-amyloid fibrils in vitro. Biochem.
Biophys. Res. Commun. 2006, 341, 1046–1052.
(15) Ono, K.; Hasegawa, K.; Naiki, H.; Yamada, M. Preformed beta-
amyloid fibrils are destabilized by coenzyme Q10 in vitro. Biochem.
Biophys. Res. Commun. 2005, 330, 111–116.
(16) Kozlov, A. V.; Gille, L.; Staniek, K.; Nohl, H. Dihydrolipoic acid
maintains ubiquinone in the antioxidant active form by two-electron
reduction of ubiquinone and one-electron reduction of ubisemiqui-
none. Arch. Biochem. Biophys. 1999, 363, 148–154.
(17) Bolognesi, M. L.; Banzi, R.; Bartolini, M.; Cavalli, A.; Tarozzi, A.;
Andrisano, V.; Minarini, A.; Rosini, M.; Tumiatti, V.; Bergamini,
C.; Fato, R.; Lenaz, G.; Hrelia, P.; Cattaneo, A.; Recanatini, M.;
Melchiorre, C. Novel class of quinone-bearing polyamines as
multi-target-directed ligands to combat Alzheimer’s disease.
J. Med. Chem. 2007, 50, 4882–4897.
The present study adds to the potential benefits of using
rationally designed multitargeted antioxidants to study and
ultimately address the complex mechanisms underlying neu-
rodegeneration.
Experimental Section
Satisfactory elemental analysis results were obtained for all
new compounds, confirming >95% purity.
Synthesis of 3. A solution of 18 (35 mg, 0.072 mmol) in
CH2Cl2 (5 mL), NEt3 (30 μL), hydroxybenzotriazole (HOBt)
(15 mg, 0.12 mmol), 1-ethyl-3-[3-(dimethylamino)propyl]car-
bodiimide hydrochloride (EDCI) (40 mg, 0.21 mmol), and
additional NEt3 (30 μL) were added successively to a solution
of LA (16 mg, 0.077 mmol) in CH2Cl2 (5 mL). The mixture was
stirred overnight at 25 °C and diluted with water (30 mL). The
product was extracted with CH2Cl2 (3 ꢀ 30 mL). Evaporation of
the dried extracts afforded crude 3 that was purified by flash
chromatography. Eluting with CH2Cl2/CH3OH/NH3 (9.5:0.5:0.05),
3 was obtained (red waxy solid, 83% yield).
Synthesis of 4. It was synthesized in 75% yield from LA
(35 mg, 0.17 mmol) and 19 (60 mg, 0.14 mmol), following the
procedure described for 3.
Synthesis of 5. It was synthesized in 37% yield from LA
(66 mg, 0.32 mmol) and 20 (130 mg, 0.3 mmol), following the
procedure described for 3.
(18) Reddy, P. H.; Beal, M. F. Amyloid beta, mitochondrial dysfunction
and synaptic damage: implications for cognitive decline in aging and
Alzheimer’s disease. Trends Mol. Med. 2008, 14, 45–53.
(19) Raina, A. K.; Templeton, D. J.; Deak, J. C.; Perry, G.; Smith, M.
A. Quinone reductase (NQO1), a sensitive redox indicator, is
increased in Alzheimer’s disease. Redox Rep. 1999, 4, 23–27.
(20) Bolognesi, M. L.; Cavalli, A.; Melchiorre, C. Memoquin: a multi-
target-directed ligand as an innovative therapeutic opportunity for
Alzheimer’s disease. Neurotherapeutics 2009, 6, 152–162.
(21) Frid, P.; Anisimov, S. V.; Popovic, N. Congo red and protein
aggregation in neurodegenerative diseases. Brain Res. Rev. 2007,
53, 135–160.
Synthesis of 6. It was synthesized in 70% yield from LA
(38 mg, 0.186 mmol) and 21 (70 mg, 0.17 mmol), following the
procedure described for 3.
Acknowledgment. This research was supported by grants
from MUR.
(22) Morphy, R.; Rankovic, Z. Fragments, network biology and de-
signing multiple ligands. Drug Discovery Today 2007, 12, 156–160.
€
(23) Korcsmaros, T.; Szalay, M. S.; Bode, C.; Kovacs, I. A.; Csermely,
P. How to design multi-target drugs: target search options in
Supporting Information Available: Analytical characteriza-
tion of 3-6, experimental details of the synthesis of 10-21, and
cellular networks. Expert Opin. Drug Discovery 2007, 2, 1–10.