The bivalent ligand approach as a tool for improving the in vitro anti-alzheimer multitarget profile of dimebon

Article abstract of DOI:10.1002/cmdc.201300263

Inspired by the concept of bivalent ligands, we prepared a small set of analogues of the drug candidate dimebon. They were shown to inhibit AChE, Aβ₄₂ aggregation, and NMDA receptor activation to a greater extent than dimebon. Some of these compounds also enhanced the survival of chicken neurons under apoptosis-inducing conditions. Copyright

Full text of DOI:10.1002/cmdc.201300263

CHEMMEDCHEM
COMMUNICATIONS
DOI: 10.1002/cmdc.201300263
The Bivalent Ligand Approach as a Tool for Improving the
in vitro Anti-Alzheimer Multitarget Profile of Dimebon
Michela Rosini,*[a] Elena Simoni,[a] Manuela Bartolini,[a] Elena Soriano,[b] Josꢀ Marco-
Contelles,[b] Vincenza Andrisano,[c] Barbara Monti,[a] Manfred Windisch,[d] Birgit Hutter-Paier,[d]
David W. McClymont,[e] Ian R. Mellor,[e] and Maria Laura Bolognesi*[a]
Alzheimer’s disease (AD) is an extremely challenging and often
frustrating area of drug discovery. Since the withdrawal of ta-
crine (Cognex) in 2006, more than 200 AD drug candidates
have failed in late-stage clinical trials.[1] The antihistamine drug
dimebon (1, Figure 1) belongs to this long list. In 2008, 1 at-
tracted considerable interest within the AD community when
it successfully completed a small six-month clinical trial in
Russia, in which it showed impressive cognition-enhancing ef-
fects in patients suffering from mild to moderate AD.[2,3] How-
ever, these positive outcomes were unconfirmed in a replica-
tion trial in the United States.[4]
We were particularly interested in 1 because, from initial re-
ports on its in vitro activity profile, it seemed to fulfill the
promise of an effective multitarget drug for treating AD.[5] In
recent years, multitarget drug development has emerged as
an effective approach in the search for disease-modifying
drugs against AD.[6] This approach involves the development
of single chemical entities that can simultaneously modulate
multiple targets critically involved in the neurotoxic pathway.[7]
It therefore runs parallel to drug combination in the search for
appropriate therapeutic interventions against the complex
pathogenesis of AD.[8,9]
Indeed, early research suggested that the clinical benefits of
dimebon were related to its ability to simultaneously inhibit
Figure 1. Design strategy for generating compounds 2–6.
two crucial AD molecular targets: acetylcholinesterase (AChE)
and N-methyl-d-aspartate receptor (NMDAR).[10] The effective-
ness of simultaneous inhibition of both targets was further
supported by higher short- and long-term efficacies observed
in clinical trials involving co-administration of the NMDAR an-
tagonist memantine and an AChE inhibitor (AChEI). Combining
memantine and AChEIs is the current standard of care for AD
patients.[11] Furthermore, pre-clinically, the combination was
demonstrated to act synergistically, which may explain the ob-
served clinical effects.[12] In this respect, we have already suc-
cessfully combined the symptomatic relief of AChE inhibition
and the neuroprotective action of NMDAR antagonism in
a single multitarget molecule.[13,14]
[a] Dr. M. Rosini, Dr. E. Simoni, Dr. M. Bartolini, Dr. B. Monti,
Prof. M. L. Bolognesi
Department of Pharmacy and Biotechnology
University of Bologna, Via Belmeloro 6, 40126 Bologna (Italy)
E-mail: michela.rosini@unibo.it
marialaura.bolognesi@unibo.it
In the case of 1, however, its low in vitro activity against
these two key targets (IC50 =42 mm and 10–70 mm against
AChE and NMDAR, respectively) might be one of the causes of
its clinical failure. Indeed, Giorgetti et al. demonstrated that the
brain concentration reached after acute oral administration of
1 to rats is much lower than that required to significantly
affect AChE or NMDA pathways (nanomolar vs. micromolar).[15]
Although exposure levels of dimebon in plasma and cerebro-
spinal fluids in humans have not been published, it has been
implied that the same situation might be replicated in the AD
patients that were enrolled in the trial.[16] This raised much un-
certainty on the real mechanism by which dimebon may bene-
[b] Dr. E. Soriano, Prof. J. Marco-Contelles
Instituto de Quimica Organica General (CSIC)
Juan de la Cierva, 3, 28006 Madrid (Spain)
[c] Prof. V. Andrisano
Department for Life Quality Studies, Alma Mater Studiorum-University of
Bologna, Rimini Campus, Italy
[d] Dr. M. Windisch, Dr. B. Hutter-Paier
JSW Lifesciences GmbH, Parkring 12, 8074 Grambach (Austria)
[e] Dr. D. W. McClymont, Dr. I. R. Mellor
School of Biology, University of Nottingham
University Park, Nottingham NG7 2RD (UK)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300263.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1276 – 1281 1276

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
fit AD. In this puzzling scenario, we believe that dimebon’s
contradictory story does not refute the multitarget drug
design strategy. Rather, we consider that it may represent an
interesting starting point in pursuing this approach.[5]
All these considerations prompted us to manipulate the
structure of 1 in an attempt to enhance the cholinesterase and
NMDAR-blocking activities and to gain an actual multitarget
profile. In this respect, we note that the anticholinesterase ac-
tivity of the marketed drugs varies from the low- to sub-micro-
molar range,[17] while a relatively low-affinity (apparent affinity
of ~1 mm) NMDAR open-channel blocker profile is fundamental
for the clinical efficacy and tolerability of memantine.[18]
To achieve a suitable activity profile, we turned our attention
to the bivalent ligand approach. Linking two pharmacophoric
units through a spacer is a widely used strategy for enhancing
the binding affinity of ligands to therapeutically relevant pro-
teins. For both enzymes and receptors, it has been verified
that such ligands exploit a bivalent binding site. They thus
foster a more potent response by increasing the affinity
toward their targets by some hundredfold.[19] In the AD field,
this approach has inspired the design of numerous drug candi-
dates over the past decade, motivated by the peculiar topolo-
gy of a classical AD target, such as the enzyme AChE.[20–22]
Thanks to the presence of the catalytic (CAS) and the peripher-
al anionic (PAS) sites, which share common molecular features,
bivalent AChEIs, i.e., compounds able to simultaneously bind
the two recognition sites, resulted in agents with markedly im-
proved potency.[23] More interestingly, the possibility has re-
cently emerged that bivalent ligands may be capable of linking
independent recognition sites on other AD-validated targets,
such as amyloid-b (Ab) peptides.[24,25] In fact, several amyloid
binding compounds share a common bivalent structure, and
so-called “bivalent tweezers” have been envisaged as the next
generation of ligands.[25] As a further support to our rationale,
bivalent b-carbolines were recently discovered as effective
NMDAR blockers.[26]
Figure 2. Representation of the putative binding modes of dimebon (1) at
the CAS (pink) and PAS (blue) of AChE. Selected relevant residues of the pro-
tein are indicated. The X-ray structure of bis(5)-tacrine (green) is shown for
comparison.
Phe331. The binding mode found at the CAS showed a lower
(more favorable) binding energy than the former and placed
the tricyclic core of 1 stacked against Trp84 and Phe330. The
charged nitrogen atom of the tetrahydropyridinium ring is ori-
ented in a suitable position for a hydrogen bonding interaction
with the carboxylate moiety of the Glu199 side chain. Addi-
tional interactions were formed with the hydrophobic pocket
formed by Trp432, Met436, and Ile439. The methylpyridine
moiety rested on a cavity formed by Phe330, Phe331, and
Tyr334 (see the Supporting Information for details). Altogether,
the superposition of bis(5)-tacrine and 1 at the CAS and PAS
(Figure 2) revealed a good correlation of the g-carboline moiet-
ies with the tetrahydroacridine fragments and supported the
design of bivalent ligands 2–6.
Compounds 2–6 were then synthesized using standard reac-
tion conditions, as outlined in Scheme 1. g-Carboline 8 was ef-
ficiently prepared by reaction of ethyl 4-oxopiperidine-1-car-
boxylate and para-tolylhydrazine according to the Fischer
indole synthesis, following alkaline deprotection of the piperi-
dine nitrogen atom of 7. Compounds 2–6 were then accessed
by condensing intermediate 8 with the appropriate bromo de-
rivatives, either commercially available or prepared under mild
bromination conditions using CBr4/PPh3.
In the present study, we sought to design dimebon-based
bivalent derivatives. We initially investigated the binding mode
of 1 with AChE by means of docking simulations. Then, on the
basis of the obtained results (see below), we designed and
synthesized new chemical entities 2–6 by connecting the
g-carboline moieties of 1 with variable-length polymethylene
spacers and heteroatom or aromatic linkers (Figure 1). The new
compounds were then biologically investigated to assess their
ability to inhibit cholinesterases (human AChE and BChE),
NMDAR, and Ab aggregation.
To determine the multitarget profile of 2–6, we initially in-
vestigated their inhibitory activity toward cholinesterase by the
method of Ellman and co-workers[29] on human recombinant
AChE and BChE from human serum. To allow comparison of
the results, dimebon was used as the reference compound
(Table 1). In accordance with previously published data,[10]
under our experimental conditions 1 confirmed very weak
AChE inhibitory potency (sub-millimolar range) and a modest
BChE blocking efficacy (micromolar range). In contrast, 2 and
3, carrying six- or seven-methylene-unit spacers, respectively,
showed a marked increase of inhibitory potency. This was in
agreement with our starting hypothesis. Both 2 and 3 were
sub-micromolar AChE inhibitors (IC50 =0.19 and 0.69 mm, re-
spectively), which were 454- and 128-fold more potent than
the prototype. Notably, they were also more potent inhibitors
The feasibility of developing dimebon-related bivalent li-
gands as AChE inhibitors was first established by means of
computational studies taking advantage of the crystallographic
complex of AChE (PDB code: 2CMF) with the bivalent AChEI
bis(5)-tacrine.[27] Docking of 1 showed a plausible accommoda-
tion at both the PAS and CAS (Figure 2), which was the most
populated cluster in the first region. At the PAS, the tricyclic
core of 1 stacked against the aromatic side chain of Trp279
and Tyr334 (thus mimicking a similar interaction found with
the aromatic scaffold of propidium),[28] whereas the protonated
amine group was hydrogen bonded to the carbonyl group of
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1276 – 1281 1277

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
the p–p interactions of the aro-
matic group mainly with Trp279
(see the Supporting Informa-
tion).
The nature of the spacer was
also relevant for the AChE inhibi-
tion profile (see the activities of
compounds 5 and 6). In particu-
lar, 5, carrying
a heteroatom
linker, was tenfold more potent
than its polymethylene homo-
logue, 4. This suggests for 5 the
possibility of establishing addi-
tional interactions with the
enzyme by means of the
spacer’s oxygen atoms. A dra-
Scheme 1. Reagents and conditions: a) EtOH, 1008C, 4 h; b) KOH, EtOH/H2O, overnight; c) Cs2CO3/K2CO3, DMF,
reflux, 48 h.
Table 1. Inhibition of AChE and BChE activities and self-induced Ab ag-
gregation by compounds 1–6.
Compd
IC50 [mm][a]
Inhibition of Ab42
self-aggregation [%][b]
hAChE
hBChe
[I]=10 mm
[I]=50 mm
1
2
3
4
5
6
88.5Æ4.5
0.195Æ0.011
0.692Æ0.015
21.5Æ0.6
1.03Æ0.02
NA
10.8Æ1.7
67.0Æ0.9
71.1Æ2.7
72.7Æ2.1
43.4Æ2.0
39.3Æ4.6
0.629Æ0.010
0.737Æ0.027
0.508Æ0.020
1.73Æ0.08
35.6Æ2.4
35.2Æ1.2
40.2Æ3.6
24.8Æ0.4
17.7Æ0.1
2.52Æ0.05
225Æ36
1.24Æ0.06
[a] Human recombinant AChE and BChE from human serum were used.
The method of Ellman et al.[29] was used to assess enzyme activity in the
presence and absence of tested inhibitor. IC50 values represent the con-
centration of inhibitor required to decrease enzyme activity by 50% and
are the meanÆSEM of two independent measurements, each performed
in triplicate. Detailed experimental conditions are given in the Supporting
Information. [b] Inhibition of Ab42 self-aggregation (50 mm) produced by
the tested compound. Detailed experimental conditions are given in the
Supporting Information; NA: not active.
Figure 3. Schematic binding mode of 2 predicted by docking studies. Rele-
vant residues are shown and are colored by atom type (with C atoms in
light blue). Ligand 2 is also colored by atom type, with C atoms in orange.
One of the pyridoindole moieties binds to the enzyme, displaying a parallel
p–p stacking interaction between Trp84 and Phe330 in a “sandwich” form.
The other tricyclic moiety is linked at the top of the gorge (PAS) and stacks
against the indole ring of Trp279 through a classical p–p interaction. In addi-
tion, the NH of this group is hydrogen bonded to His440, while the methyl
substituent fits into a hydrophobic pocket formed by Trp432, Met436, and
Ile439. The X-ray structure of bis(5)-tacrine (green) is shown for comparison.
than the marketed drugs galantamine (IC50 =2.01 mm)[17] and ri-
vastigmine (3.03 mm).[30] However, the AChE inhibition data of
the polymethylene-linked dimers 2–4 clearly indicate that
spacer length is a critical issue for AChE binding. In fact,
lengthening the flexible tether up to eight methylene units (in
4) resulted in an evident drop in activity, leading to an AChE-
inhibition profile similar to that of 1. On these bases, we as-
sumed that the spacer plays a key role in allowing a proper
docking of the new compounds into the AChE gorge, and in
simultaneously contacting both sites of the enzyme.
matic loss of activity was observed for 6, revealing that the ad-
ditional aromatic ring of the tether cannot be properly accom-
modated inside the gorge. This was also supported by the pos-
itive (disfavored) and high calculated binding energy values of
the most populated clusters obtained by docking simulations
(Supporting Information).
Concerning the inhibition of BChE, the structural modifica-
tions we performed on 1 did not generally lead to significant
changes in the inhibitory profile. This resulted in a variation of
the AChE/BChE selectivity profile, with 2 and 3 presenting
a balanced dual AChE/BChE efficacy. On the basis of the work
by Greig et al.,[31] this balanced action might be beneficial, as
both enzymes are recognized therapeutic targets at different
stages of the pathology. Indeed, as AD progresses, the activity
of AChE decreases while that of BChE increases,[32] in an at-
tempt to modulate acetylcholine levels in cholinergic neurons.
Therefore, the simultaneous inhibition of both cerebral en-
To analyze this hypothesis, computational studies were per-
formed on 2 and 4, the shortest and the longest analogues, re-
spectively. Figure 3 shows the position of 2 with respect to the
key residues in the binding site. Interestingly, this binding
mode shows the same interaction pattern of the bivalent
ligand bis(5)-tacrine, suggesting an effective dual binding
mode for 2. The same studies performed on 4 afforded differ-
ent results. Indeed, the longer spacer seems not to allow
a proper accommodation of 4 in the gorge, thus disrupting
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1276 – 1281 1278

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
zymes might be more beneficial for moderate to severe forms
of AD.
pendence between À50 and À80 mV (~1.5- to 4-fold) relative
to 1 (>12-fold). This behavior was similar to that of meman-
tine, albeit with slightly lower potency. Compounds 1–6 were
further explored by considering their onset times of inhibition
in comparison with memantine. Current decay time constants
following addition of 1–6 at 100 mm were estimated by fitting
an exponential decay (Supporting Information). Compound 5
showed the fastest onset, with its tonset being significantly
lower (p<0.05) than that of 1 at all holding potentials and 2
and 3 at À100 mV. Compared with memantine, none of the
compounds showed a significantly different onset, with the ex-
ception of 1 at À100 mV; the onset in this case was ~3-fold
slower (p=0.032). There were no significant effects of voltage
on onset time.
Because of the anti-aggregating properties of tacrine-based
bivalent ligands[33,34] and the claimed protective effects of di-
mebon against Ab-induced cytotoxicity,[10] we used a thiofla-
vin-T-based fluorescence assay to investigate the ability of
compounds 2–6 to decrease Ab42 self-aggregation.[35,36] This
was compared with the ability of 1 (Table 1). An almost negligi-
ble anti-aggregating effect (~10%) was verified for 1 when
tested at a 1:1 ratio with Ab42 (50 mm). However, compounds
2–6 significantly inhibited Ab aggregation, with inhibitory per-
centages ranging from 39 to 73%. Remarkably, 2–6 showed
a concentration-dependent inhibitory profile and were still sig-
nificantly active at 10 mm. This finding reinforces the concept
that bivalent ligands may be effective tools in amyloid recogni-
tion. It also indicates a key role for the spacer. Indeed, for the
present series of compounds, an aliphatic and flexible linker, as
in 2–4, is associated with the improved anti-aggregating activi-
ty. This suggests it can promote positive g-carboline interac-
tions with amyloid peptides.
To corroborate the neuroprotective profile of 2–6 in a cellular
context, their effects in a low-serum cell stress in vitro model
were investigated.[37] Primary cortical chicken neurons were
maintained under apoptosis-inducing low-serum conditions,
which mimic senescence, and 2–6 were added to the culture
from the first day onward.[37] Potential protective effects of 2–6
were investigated by MTT assays at day 8 (Figure 4). Results
show a clear effect on neuron viability for 3, 4, and 5, but not
6 and 2, which failed to show any significant effect. The effects
were more pronounced with 3, which markedly enhanced the
survival of cortical neurons at 100 nm, but also showed some
protective activity at a tenfold lower concentration (10 nm,
Figure 4). Similarly, 4 and 5 were effective (twofold increase in
viability) at 100 nm, although all tested compounds seemed to
develop some cytotoxicity starting from 1 mm. Therefore, in
this lesion model 3–5 effectively protect cells from degenera-
tion, displaying neuroprotective activity.
In parallel, we sought to verify the capability of 2–6 as
NMDAR antagonists to link the potential neuroprotective
effect with the cholinergic and anti-aggregating actions; they
were thus studied at recombinant NMDARs. In particular, their
profile was assessed at Ca2+-permeable GluN1/GluN2A
NMDARs expressed in Xenopus laevis oocytes, using 1 as the
reference compound. Memantine was also added to this study,
because it is the only NMDAR antagonist that has been mar-
keted for AD, thanks to its capacity to preferentially block ex-
cessive NMDAR activity without disrupting normal neuronal
function.[18]
With oocytes voltage-clamped at À100 mV, compounds 2–6
were co-applied with NMDA (100 mm plus 10 mm Gly) revealing
an NMDAR antagonism similar to or slightly greater than that
of compound 1 (Table 2). To clarify the mechanism of inhibition
As a preliminary evaluation of the CNS penetration of the
synthesized compounds, 2 was subjected to an ex vivo deter-
mination of its AChE inhibitory activity, as reported by MuÇoz-
Torrero and colleagues.[38] Unfortunately, under these experi-
mental conditions (see Supporting Information for experimen-
tal details), we could not detect any inhibitory activity. Howev-
er, it is well known that, among other possibilities, this might
be due to dissociation of the reversible enzyme–inhibitor com-
plex owing to the necessary dilution of the tissue sample. No-
tably, a similar result was reported for the drug tacrine.[39]
In summary, in the present investigation we successfully ex-
ploited the bivalent ligand approach to generate dimebon
congeners with a markedly improved in vitro biological profile,
despite being unable to experimentally verify their blood–
brain barrier permeation. The molecular duplication applied
while increasing the inhibitory potency against AChE had no
negative effect on molecular recognition at NMDARs. The in-
hibitory profiles of 2–6 were similar to or better than that of
1 and close to that of the drug memantine. More importantly,
the strategy allowed us to change 1 from an ineffective com-
pound against in vitro amyloid aggregation into a series of ef-
fective inhibitors.
Table 2. NMDAR inhibition by 2–6 and reference compounds 1 and
memantine.
Compd
IC50 [mm][a]
À80 mV
À50 mV
149Æ42
À100 mV
1
2
3
4
5
12.0Æ4.2
17.9Æ2.1
29.1Æ4.6
14.8Æ2.0
8.22Æ0.95
15.2Æ3.9
1.04Æ0.19
13.9Æ4.0
11.0Æ1.5
12.6Æ2.1
9.66Æ1.03
4.90Æ0.58
8.72Æ2.62
0.71Æ0.06
27.5Æ3.8
51.1Æ7.4
35.2Æ3.3
13.9Æ2.4
37.4Æ8.3
3.84Æ0.24
6
Mem
[a] Inhibitor concentration required to elicit 50% maximum inhibition of
NR1/NR2A-mediated current evoked by 100 mm NMDA plus 10 mm Gly at
holding potentials of À50, À80, and À100 mV. IC50 values were estimated
from concentration–inhibition curves using 5–9 oocytes. Mem=meman-
tine.
by 2–6, their activity was also tested at more positive holding
potentials (À50 and À80 mV). Under these latter conditions,
the inhibitory activities were less prominent than those record-
ed at À100 mV. Interestingly, 2–6 exhibited less voltage de-
We suggest that these positive results might warrant further
evaluation of the bivalent ligand approach. They might serve
as the basis for developing new dimebon-based bivalent li-
gands that could yield more consistent clinical benefits follow-
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1276 – 1281 1279

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
[1] R. E. Becker, N. H. Greig, Alzheimer’s Dementia 2013, 9, 50.
[2] A. Burns, R. Jacoby, Lancet 2008, 372, 179.
[3] R. S. Doody, S. I. Gavrilova, M. Sano, R. G. Thomas, P. S. Aisen, S. O. Bach-
urin, L. Seely, D. Hung, Lancet 2008, 372, 207.
[4] Dimebon Disappoints: Is There Hope for Novel Alzheimer’s Agent?, Med-
scape News Today, Mar 12, 2010 (updated March 25): http://www.
medscape.com/viewarticle/718401.
[5] A. Cavalli, M. L. Bolognesi in Polypharmacology in Drug Discovery (Ed.:
J.-U. Peters), John Wiley & Sons, New York, 2012, p. 441.
[6] A. Cavalli, M. L. Bolognesi, A. Minarini, M. Rosini, V. Tumiatti, M. Recana-
tini, C. Melchiorre, J. Med. Chem. 2008, 51, 347.
[7] M. L. Bolognesi, Curr. Med. Chem. 2013, 20, 1639.
[8] M. L. Bolognesi, E. Simoni, M. Rosini, A. Minarini, V. Tumiatti, C. Mel-
chiorre, Curr. Top. Med. Chem. 2011, 11, 2797.
[9] R. Leꢂn, A. G. Garcia, J. Marco-Contelles, Med. Res. Rev. 2013, 33, 139.
[10] S. Bachurin, E. Bukatina, N. Lermontova, S. Tkachenko, A. Afanasiev, V.
Grigoriev, I. Grigorieva, Y. Ivanov, S. Sablin, N. Zefirov, Ann. N. Y. Acad.
Sci. 2001, 939, 425.
[11] S. Gauthier, J. L. Molinuevo, Alzheimer’s Dementia 2013, 9, 326.
[12] G. Rammes, W. Danysz, C. G. Parsons, Curr. Neuropharmacol. 2008, 6, 55.
[13] M. Rosini, E. Simoni, M. Bartolini, A. Cavalli, L. Ceccarini, N. Pascu, D. W.
McClymont, A. Tarozzi, M. L. Bolognesi, A. Minarini, V. Tumiatti, V. Andri-
sano, I. R. Mellor, C. Melchiorre, J. Med. Chem. 2008, 51, 4381.
[14] E. Simoni, S. Daniele, G. Bottegoni, D. Pizzirani, M. L. Trincavelli, L. Gold-
oni, G. Tarozzo, A. Reggiani, C. Martini, D. Piomelli, C. Melchiorre, M.
Rosini, A. Cavalli, J. Med. Chem. 2012, 55, 9708.
[15] M. Giorgetti, J. A. Gibbons, S. Bernales, I. E. Alfaro, C. D. La Rochelle, T.
Cremers, C. A. Altar, R. Wronski, B. Hutter-Paier, A. A. Protter, J. Pharma-
col. Exp. Ther. 2010, 333, 748.
[16] I. Okun, S. E. Tkachenko, A. Khvat, O. Mitkin, V. Kazey, A. V. Ivachtchenko,
Curr. Alzheimer Res. 2010, 7, 97.
[17] M. L. Bolognesi, R. Banzi, M. Bartolini, A. Cavalli, A. Tarozzi, V. Andrisano,
A. Minarini, M. Rosini, V. Tumiatti, C. Bergamini, R. Fato, G. Lenaz, P.
Hrelia, A. Cattaneo, M. Recanatini, C. Melchiorre, J. Med. Chem. 2007, 50,
4882.
[18] S. A. Lipton, Nat. Rev. Drug Discovery 2006, 5, 160.
[19] P. S. Portoghese, J. Med. Chem. 2001, 44, 2259.
[20] D. MuÇoz-Torrero, P. Camps, Curr. Med. Chem. 2006, 13, 399.
[21] H. Haviv, D. M. Wong, I. Silman, J. L. Sussman, Curr. Top. Med. Chem.
2007, 7, 375.
[22] M. L. Bolognesi, A. Minarini, M. Rosini, V. Tumiatti, C. Melchiorre, Mini-
Rev. Med. Chem. 2008, 8, 960.
[23] Y. P. Pang, P. Quiram, T. Jelacic, F. Hong, S. Brimijoin, J. Biol. Chem. 1996,
271, 23646.
[24] M. Bartolini, V. Andrisano, ChemBioChem 2010, 11, 1018.
[25] A. A. Reinke, J. E. Gestwicki, Chem. Biol. Drug Des. 2011, 77, 399.
[26] Y. Rook, K. U. Schmidtke, F. Gaube, D. Schepmann, B. Wunsch, J. Heil-
mann, J. Lehmann, T. Winckler, J. Med. Chem. 2010, 53, 3611.
[27] E. H. Rydberg, B. Brumshtein, H. M. Greenblatt, D. M. Wong, D. Shaya,
L. D. Williams, P. R. Carlier, Y. P. Pang, I. Silman, J. L. Sussman, J. Med.
Chem. 2006, 49, 5491.
Figure 4. Low-serum assay for a) 3, b) 4, and c) 5. Values represent the mean
neuronal viability plus SEM from one experiment in 96-well plates with n=6
for each compound concentration and n=6 for vehicle controls (VC). Data
are shown as percentage of vehicle controls (100%). Statistical significance
is indicated by *p<0.05, **p<0.01, ***p<0.001 as determined by a paramet-
ric one-way ANOVA with Bonferroni’s multiple comparison test.
ing the fine-tuning of the multitarget activities and pharmaco-
kinetic properties.
[28] Y. Bourne, P. Taylor, P. E. Bougis, P. Marchot, J. Biol. Chem. 1999, 274,
2963.
[29] G. L. Ellman, K. D. Courtney, V. Andres, Jr., R. M. Feather-Stone, Biochem.
Pharmacol. 1961, 7, 88.
[30] M. Rosini, E. Simoni, M. Bartolini, A. Tarozzi, R. Matera, A. Milelli, P.
Hrelia, V. Andrisano, M. L. Bolognesi, C. Melchiorre, Eur. J. Med. Chem.
2011, 46, 5435.
[31] N. H. Greig, T. Utsuki, D. K. Ingram, Y. Wang, G. Pepeu, C. Scali, Q. S. Yu,
J. Mamczarz, H. W. Holloway, T. Giordano, D. Chen, K. Furukawa, K. Sam-
bamurti, A. Brossi, D. K. Lahiri, Proc. Natl. Acad. Sci. USA 2005, 102,
17213.
[32] S. Darvesh, D. A. Hopkins, C. Geula, Nat. Rev. Neurosci. 2003, 4, 131.
[33] M. L. Bolognesi, M. Bartolini, F. Mancini, G. Chiriano, L. Ceccarini, M.
Rosini, A. Milelli, V. Tumiatti, V. Andrisano, C. Melchiorre, ChemMedChem
2010, 5, 1215.
Acknowledgements
This work was supported by grants from the University of Bolo-
gna, Italian Ministry of Education, Universities and Research
(MIUR) (PRIN 2009ESXPT2_001 and PRIN 2009Z8YTYC_003), Uni-
Rimini S.p.A., the UK Medical Research Council (DTP) and
MINECO (Spain) (SAF2009-07271). M.L.B. and M.R. thank Prof. C.
Melchiorre for fruitful discussions. The authors thank Grace Fox
for proofreading and editing the manuscript.
[34] A. Minarini, A. Milelli, V. Tumiatti, M. Rosini, E. Simoni, M. L. Bolognesi, V.
Andrisano, M. Bartolini, E. Motori, C. Angeloni, S. Hrelia, Neuropharma-
cology 2012, 62, 997.
[35] M. Bartolini, C. Bertucci, M. L. Bolognesi, A. Cavalli, C. Melchiorre, V. An-
drisano, ChemBioChem 2007, 8, 2152.
Keywords: acetylcholinesterase · Alzheimer’s disease · amyloid
aggregation · butyrylcholinesterase · dual binding inhibitors ·
NMDA receptors
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1276 – 1281 1280

CHEMMEDCHEM
COMMUNICATIONS
www.chemmedchem.org
[36] M. Bartolini, M. Naldi, J. Fiori, F. Valle, F. Biscarini, D. V. Nicolau, V. Andri-
sano, Anal. Biochem. 2011, 414, 215.
[37] M. Hartbauer, B. Hutter-Paier, G. Skofitsch, M. Windisch, J. Neural
Transm. 2001, 108, 459.
[39] P. Camps, R. El Achab, J. Morral, D. MuÇoz-Torrero, A. Badia, J. E. BaÇos,
N. M. Vivas, X. Barril, M. Orozco, F. J. Luque, J. Med. Chem. 2000, 43,
4657.
[38] C. Galdeano, E. Viayna, I. Sola, X. Formosa, P. Camps, A. Badia, M. V. Clos,
J. Relat, M. Ratia, M. Bartolini, F. Mancini, V. Andrisano, M. Salmona, C.
Minguillꢂn, G. C. Gonzꢃlez-MuÇoz, M. I. Rodrꢄguez-Franco, A. Bidon-
Chanal, F. J. Luque, D. MuÇoz-Torrero, J. Med. Chem. 2012, 55, 661.
Received: June 11, 2013
Published online on July 3, 2013
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2013, 8, 1276 – 1281 1281