Tacrine-O-protected phenolics heterodimers as multitarget-directed ligands against Alzheimer's disease: Selective subnanomolar BuChE inhibitors

Article abstract of DOI:10.1016/j.ejmech.2019.07.053

Concerned by the devastating effects of Alzheimer's disease, and the lack of effective drugs, we have carried out the design of a series of tacrine-phenolic heterodimers in order to tackle the multifactorial nature of the disease. Hybridization of both pharmacophores involved the modification of the nature (imino, amino, ether) and the length of the tether, together with the type (hydroxy, methoxy, benzyloxy), number and position of the substituents on the aromatic residue. Title compounds were found to be strong and selective inhibitors of human BuChE (from low nanomolar to subnanomolar range), an enzyme that becomes crucial in the more advanced stages of the disease. The lead compound, bearing an ether-type tether, had an IC50 value of 0.52 nM against human BuChE, and a selectivity index of 323, with an 85-fold increase of activity compared to parent tacrine; key interactions were analysed using molecular modelling. Moreover, it also inhibited the self-aggregation of Aβ42, lacking neurotoxicity up to 5 μM concentration, and showed neuroprotective activity in primary rat neurons in a serum and K+ deprivation model, widely employed for reproducing neuronal injury and senescence. Moreover, low hepatoxicity effects and complete stability under physiological conditions were found for that compound. So, overall, our lead compound can be considered as a promising multitarget-directed ligand against Alzheimer's disease, and a good candidate for developing new drugs.

Full text of DOI:10.1016/j.ejmech.2019.07.053

Accepted Manuscript
Tacrine-O-protected phenolics heterodimers as multitarget-directed ligands against
Alzheimer's disease: Selective subnanomolar BuChE inhibitors
Jesús M. Roldán-Peña, Valle Romero-Real, Javier Hicke, Inés Maya, Antonio
Franconetti, Irene Lagunes, José M. Padrón, Sabrina Petralla, Eleonora Poeta,
Marina Naldi, Manuela Bartolini, Barbara Monti, Maria L. Bolognesi, Óscar López,
José G. Fernández-Bolaños
PII:
S0223-5234(19)30674-9
DOI:
https://doi.org/10.1016/j.ejmech.2019.07.053
EJMECH 11550
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 1 January 2019
Revised Date: 15 July 2019
Accepted Date: 19 July 2019
Please cite this article as: Jesú.M. Roldán-Peña, V. Romero-Real, J. Hicke, Iné. Maya, A. Franconetti,
I. Lagunes, José.M. Padrón, S. Petralla, E. Poeta, M. Naldi, M. Bartolini, B. Monti, M.L. Bolognesi, Ó.
López, José.G. Fernández-Bolaños, Tacrine-O-protected phenolics heterodimers as multitarget-directed
ligands against Alzheimer's disease: Selective subnanomolar BuChE inhibitors, European Journal of
Medicinal Chemistry (2019), doi: https://doi.org/10.1016/j.ejmech.2019.07.053.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Tacrine-O-protected phenolics heterodimers as multitarget-directed
ligands against Alzheimer’s disease: selective subnanomolar BuChE
inhibitors
Jesús M. Roldán-Peña,a Valle Romero-Real,a Javier Hicke,a Inés Maya,a Antonio
Franconetti,b Irene Lagunes,c José M. Padrón,c Sabrina Petralla,d Eleonora Poeta,d
Marina Naldi,d Manuela Bartolini,d Barbara Monti,d Maria L. Bolognesi,d Óscar
Lópeza,* José G. Fernández-Bolañosa,*
aDepartamento de Química Orgánica, Facultad de Química, Universidad de Sevilla, Apartado 1203, E-
41071 Sevilla, Spain. Email: bolanos@us.es, osc-lopez@us.es
b Department de Química, Universitat Autònoma de Barcelona, 08193 Cerdanyola del Vallè, Spain
cBioLab, Instituto Universitario de Bio-Orgánica “Antonio González” (IUBO-AG), Centro de
Investigaciones Biomédicas de Canarias (CIBICAN), Universidad de La Laguna, Apartado 456, E-38071
La Laguna, Spain
dDepartment of Pharmacy and Biotechnology, Alma Mater Studiorum University of Bologna, via
Belmeloro 6, 40126 Bologna, Italy
Graphical abstract
R= H, Me, Bn
Tether
Imine, amine, ether
(OR)n
HN
N
New tacrine-phenolics as
multi target anti-Alzheimer derivatives
Up to subnanomolar
inhibitors of BuChE
Inhibition of
42 self-aggregation
Neuroprotective
effects
Lack of
neurotoxicity hepatotoxicity
Low
A
Highlights
A series of tacrine-phenolic hybrids have been accessed as MTDL’s for Alzheimer’s disease
•
•
Nature/length of the linker (imine, amine, ether) and type of aryl substituents have been
modified
•
•
Strong and selective BuChE inhibitors have been obtained (low nanomolar-subnanomolar)
Good Aβ inhibition and neuroprotection profiles were observed
1

ACCEPTED MANUSCRIPT
•
Low neurotoxicity and hepatotoxicity for the lead compound were observed
Abstract
Concerned by the devastating effects of Alzheimer’s disease, and the lack of effective
drugs, we have carried out the design of a series of tacrine-phenolic heterodimers in
order to tackle the multifactorial nature of the disease.
Hybridization of both pharmacophores involved the modification of the nature (imino,
amino, ether) and the length of the tether, together with the type (hydroxy, methoxy,
benzyloxy), number and position of the substituents on the aromatic residue. Title
compounds were found to be strong and selective inhibitors of human BuChE (from low
nanomolar to subnanomolar range), an enzyme that becomes crucial in the more
advanced stages of the disease.
The lead compound, bearing an ether-type tether, had an IC50 value of 0.52 nM against
human BuChE, and a selectivity index of 323, with an 85-fold increase of activity
compared to parent tacrine; key interactions were analysed using molecular modelling.
Moreover, it also inhibited the self-aggregation of Aβ42, lacking neurotoxicity up to 5
µM concentration, and showed neuroprotective activity in primary rat neurons in a
serum and K+ deprivation model, widely employed for reproducing neuronal injury and
senescence. Moreover, low hepatoxicity effects and complete stability under
physiological conditions were found for that compound.
So, overall, our lead compound can be considered as a promising multitarget-directed
ligand against Alzheimer’s disease, and a good candidate for developing new drugs.
Keywords: Alzheimer’s disease, tacrine, heterodimers, multitarget, BuChE
2

ACCEPTED MANUSCRIPT
1.
Introduction
Alzheimer’s disease was firstly reported as a rare disorder by the German
psychiatrist Alois Alzheimer in 1906 [1]. Such uncommon neuropsychiatric disorder
has become nowadays the main form of senile dementia; roughly 70% of cases can be
attributed to Alzheimer’s disease [2]. According to the World Alzheimer Report 2018
[3], 50-million cases have already been reached, and an exponential rise is expected to
afford 152-million cases in 2050; this means a new reported case each 3.2 s [4].
Alzheimer’s remains as one of the most current devastating diseases; although
the most recognizable symptom is memory loss, many others, like depression, cognitive
impairment, behavioural changes, or psychosis, among others are also associated with
the progression of this pathology [5]. Alzheimer’s disease also involves a severe
economic burden [3] to the World Health Systems, and to the families. More frustrating,
no drug is available for the cure of Alzheimer’s; the only four marketed drugs only
allow palliative and temporal improvement of the cognitive functionality for the early to
moderate stages.
From a pathophysiological point of view, Alzheimer’s disease is an
extraordinary complex pathology, with a multifactorial etiology, lacking a single
triggering cause. Accordingly, this situation hinders the development of a successful
drug [6,7]. From a microscopic point of view, two hallmarks have been reported:
cortical amyloidogenesis, which leads to the extracellular deposits of toxic β-amyloid,
mainly Aβ42 [8], and the intraneuronal accumulation of neurofibrillary tangles, as a
result from the hyperphosphorylation of tau protein [9]. Some other events that also
contribute to neurodegeneration are mitochondrial dysfunction, oxidative stress, pro-
inflammatory responses, or the alteration of the homeostasis of biometals [9−11].
Moreover, abnormally low levels of the neurotransmitter acetylcholine, which is
responsible for the cognitive functionality, are also found in the brain of Alzheimer’s
patients [12].
Due to the extremely complex nature of the disease, it is currently claimed
[13,14] that a shift is required for the successful treatment of Alzheimer’s disease, from
the classical one-target-one-drug philosophy to multitarget-directed ligands (MTDLs)
approach [15], which tackle several therapeutic targets simultaneously.
3

ACCEPTED MANUSCRIPT
In this context, we envisioned the possibility of preparing tacrine heterodimers as
multitarget ligands against Alzheimer’s disease; tacrine, a cholinesterase inhibitor, was
the first Alzheimer’s drug approved by FDA (1993), but it was withdrawn shortly after
because of its hepatotoxicity, probably due to its pro-oxidant character [16]. In spite of
that, tacrine has been largely investigated as starting scaffold for the synthesis of
MTDLs for Alzheimer therapy [17].
Amalgamating tacrine structure in new hybrid molecules should improve its
biological profile and overcome undesirable side effects. In particular, we propose the
hybridization of tacrine (ensuring cholinesterase inhibition), with a phenolic residue,
which might provide access to other relevant therapeutic targets, like the modification
of the neuronal redox status, or the accumulation of toxic Aβ peptides in brain.
In order to get some insight into the structural requirements for obtaining the
lead-compounds, we planned the modification of the nature of the tether linking both
pharmacophores, as well as the type and position of substituents on the aromatic
scaffold (Figure 1). The inhibition of cholinesterases and deposition of amyloid plaque,
neuroprotection, hepatotoxicity, and neurotoxicity will be evaluated in order to get a
more defined pharmacological profile of the new compounds.
R= H, Me, Bn
Imine, amine, ether
(OR)n
HN
N
Figure 1. General structure for the tacrine-polyphenolic heterodimers prepared herein
2.
2.1. Chemistry
In a first approach, the condensation of parent tacrine 1 with O-protected phenolic
Results and discussion
aldehydes 2−5 was attempted (Scheme 1) to furnish the corresponding imines 6−9,
featured with restricted rotation. Initially, a classical acid-catalysed procedure was
assayed, but the yields were sensibly lower than using basic catalysis (Et2NH);
4

ACCEPTED MANUSCRIPT
resonances at 8.10−8.25 (1H-NMR, azomethine proton) and roughly 160 ppm (13C-
NMR, N=CH) confirmed the proposed structures. Moreover, due to the extended
conjugation, such imines were endowed with high stability, even in solution.
Next, the hydrogenolysis of the Schiff base was accomplished, in order to provide
the compounds with conformational flexibility around the nitrogen atom (benzyl amines
10, 12, 14); Pd/C, Pd(OH)2/C and Raney Ni were evaluated as catalysts, the two former
being the optimal ones.
The low reactivity of the imine functionality was also demonstrated in this
reaction; thus, hydrogenolysis of dimethoxy derivative 6 allowed the isolation of 10
[18] in a 52% yield, and the recovery of 48% of the starting material after a 6-h reaction
(Scheme 1). Moreover, when reduction of the p-benzyl substituted 7 was accomplished
in a short reaction time (30 min), only p-hydroxy-imine 11 was isolated (48% yield);
more prolonged treatments (3 h) afforded the isolation of 12 (29% yield). In the case of
the dibenzylated imine 8, a mixture of catechol-containing imine 13 (20%), together
with benzylamine 14 (55%) was obtained. Nevertheless, for tribenzylated imine 9, only
the removal of benzyl protecting groups was achieved (15, 50%); attempts to reduce the
imine functionality led to unexpected extensive decomposition.
5

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis of tacrine-phenolic hybrids with an imine or amine tether
In order to increase the length of the tether connecting tacrine and the
polyphenolic scaffolds, we also designed the preparation of phenethylamines 20a and
20b (Scheme 2); such compounds can be considered as hybrid structures of tacrine and
tryramine/dopamine. Moreover, the influence of the number of phenolic hydroxyl
groups will also be analysed for the bioactivities of such compounds. For that purpose,
9-chloro-1,2,3,4-tetrahydroacridine 16, obtained by POCl3-promoted condensation
between anthranilic acid and cyclohexanone, was proposed as a valuable intermediate.
Next, nucleophilic attack exerted by O-protected derivatives 18a and 18b in refluxing
phenol, furnished derivatives 19a,b from modest to quantitative yields, which in turn
were deprotected to give the hitherto unknown heterodimers 20a,b.
Scheme 2. Preparation of tacrine-tyramine/dopamine hybrids
6

ACCEPTED MANUSCRIPT
In order to increase further the length of the tether, and to introduce a second
nitrogen atom that might be involved in non-covalent interactions with cholinesterases,
the synthetic pathway depicted in Scheme 3 was followed. The first step involves the
monoalkylation of parent tacrine with α,ω-dibromoalkanes 21−23 under basic
conditions. Nevertheless, for 1,4-dibromobutane, no monoalkylated product was
detected, and 9-pyrrolidine derivative 25 was obtained as the only product through a
second intramolecular nucleophilic substitution (42%). Use of 1,5-dibromopentane
afforded the isolation of monoalkylated derivative 26 as the major compound (45%),
together with the corresponding 9-piperidinyl derivative 27 (30%); only when 1,6-
dibromohexane was employed, the corresponding monoalkylated product (28, 47%)
was obtained as the only product.
Next, nucleophilic substitution of the terminal bromine atom with NaN3,
followed by reduction of the azido group furnished ω-amino derivatives 30, 31 (Scheme
3). Finally, base-catalysed condensation of such compounds with O-protected
benzaldehydes, followed by catalytic hydrogenation of the benzyl groups afforded
dimethoxy and dihydroxylated derivatives 32 and 33, respectively (Scheme 3).
NH2
m
Br
HN
N
N
N
n
KOH
+
Br
Br
+
n
Acetonitrile
rt
N
21 n = 4
22 n = 5
23 n = 6
1
24 (n = 4, 0%)
26 (n = 5, 45%)
28 (n = 6, 47%)
25 (m = 1, 42%)
27 (m = 2, 30%)
29 (m = 3, 0%)
1) NaN3, DMF
2) H2, Pd/C
RO
RO
CHO
1)
N
NH2
HN
N
HN
N
n
n
Et2NH, Toluene
Reflux
H
OR
OR
2) H2, Pd/C
32 (n = 5, R = Me, 30%)
33 (n = 6, R = H, 30%)
30 (n = 5, 98%)
31 (n = 6, 66%)
Scheme 3. Synthesis of tacrine-phenolic hybrids with an aminoalkyl linker
7

ACCEPTED MANUSCRIPT
Finally, ether isosters of 32 and 33 were prepared following the synthetic
pathway showed in Scheme 4; isosteric replacement of the nitrogen atom with oxygen
might confer different binding properties within the enzyme active site, thus modulating
the inhibitory properties. Alkylation of substituted 2-phenyl ethanol derivatives 34−37
with 1,6-dibromopentane to give the monobromo derivatives 38−41, followed by
nucleophilic displacement of the terminal bromine atom with tacrine furnished
derivatives 42−45 from moderate to good yields (25−66%); deprotection of the benzyl
functionalities of 44-45 gave access to mono- and dihydroxylated derivatives 46, 47.
Scheme 4. Preparation of tacrine-polyphenols hybrids with an alkyl-ether linker
2.2. Biological assessment
2.2.1. Cholinesterases and Aβ-self aggregation inhibition
A common feature affecting patients suffering from Alzheimer’s disease is the
low level of neurotransmitter acetylcholine, which in turn provokes a decline of the
cognitive function, including memory and thinking skills [19]. Some years ago it was
established that inhibitors of cholinesterases (acetylcholinesterase-AChE and
butyrylcholinesterase-BuChE) can be useful in restoring the levels of acetylcholine, and
8

ACCEPTED MANUSCRIPT
are therefore of high therapeutic interest; this is known as the cholinergic hypothesis
[20].
It has been reported that AChE, which accounts for 90% of the termination of
the signal in the cholinergic neurons in healthy brains, undergoes a significant reduction
within the progression of the disease; whereas BuChE, whose biological roles have not
been completely elucidated, either remains unchanged, or increases [21], particularly in
the brain regions associated with cognition and behaviour [22]. BuChE has also been
found to associate with amyloid plaques and neurofibrillary tangles both in human
tissues, and in mouse models, and is probably involved in their maturation to initiate
neurodegeneration [23]. Moreover, administration of selective BuChE inhibitors is not
only proposed to be more beneficial for advanced stages of Alzheimer’s disease, but
could also avoid the onset of the cholinergic side-effects commonly associated with the
administration of AChE inhibitors [24].
Compounds prepared herein were initially screened in vitro against AChE
(Electrophorus electricus) and BuChE (equine serum), which are currently considered
as good model enzymes. Those exhibiting strong inhibitory properties against the model
enzymes were selected for further evaluation against the human ones. For that purpose,
the Ellman’s colorimetric assay [25] was used, in which acetyl- and butyryl-thiocholine
iodides are used as analogues of the natural substrate; reaction takes place in the
presence of 5,5’-dithiobis(2-nitrobenzoic acid) (DTNB) as a chromogen reagent, and
activity is measured in an indirect way by monitoring the absorbance increase at 405 nm
associated to the in situ generation of 5-thio-2-nitrobenzoate anion (pH 8.0).
Valuable structure-activity relationships (SARs) can be extracted from data
listed in Table 1; tacrine and galantamine are included as reference drugs.
Regarding AChE and the first set of compounds (imines and amines 6−15), the
presence of one or two benzyl groups (compounds 7, 8) provoke an impairment of
activity, leading to only a moderate inhibitory activity (Ki= 20−39 µM). Interestingly,
replacement of two benzyl groups with methoxy moieties (compound 6) furnished a
roughly 10-fold increase of activity, with inhibition constants within the low
micromolar range. A similar effect was observed upon removal of the benzyl group to
achieve the free phenolic imines 11 and 13. However, and unexpectedly, incorporation
9

ACCEPTED MANUSCRIPT
of three benzyl groups (compound 9) allowed virtually the same degree of inhibitory
activity as found for counterpart 6.
Moreover, in most cases, conformationally-flexible and more basic amines were
found to exhibit higher inhibitory properties than parent imines, shifting from activities
within the low micromolar range to the submicromolar range (11 vs. 12, 13 vs. 14). Two
effects might be responsible for such results; on the one hand, the increased
conformational flexibility of the benzylamino functionality compared to the more rigid
imino group might enable the establishment of favourable interactions within the
catalytic site of the enzyme. On the other hand, the higher basicity of the amino group
compared to the imine would favour the formation of the ammonium cation, which
better resembles the natural substrate of the enzyme. We have recently found that
conversion of tacrine into amides strongly diminished the inhibitory action on AChE
[26].
Furthermore, regarding imines, an increase in the number of free hydroxyl
groups was proved to significantly improve inhibitory properties against AChE; thus,
comparison between derivatives 11, 13 and 15 (mono-, di- and trihydroxylated imines)
shows that trihydroxylated derivative 15 is endowed with an up to 10-fold higher
activity (Table 1) compared to the other two partners of the series (Ki’s= 0.40 and 0.17
µM). The effect is less pronounced when considering more active amine-counterparts;
thus, dimethoxy derivative 10 [18] is already a strong AChE inhibitor, and the activity
is mildly improved (up to 3-fold) by inserting one (12) or two hydroxyl groups (14).
Elongation of the tether in the polydroxylated amines (e.g. 10 vs. 32 [27]) usually
improved slightly the activity. However, comparison of tacrine/dopamine hybrid
(compound 20b) with counterpart 14 showed a reversed behaviour for AChE, as
depicted in Table 1. Furthermore, regarding BuChE, tyramine- and dopamine
derivatives (20a and 20b, respectively) showed an impairment of activity compared to
most of the compounds evaluated herein.
Finally, most of the tested compounds turned out to be more potent inhibitors
than galantamine, which is currently in clinical use for the treatment of mild to
moderate Alzheimer’s cases.
Free polyhydroxylated compounds are not only endowed with potent inhibitory
properties against AChE, but also with antioxidant capacity, capable of altering the
10

ACCEPTED MANUSCRIPT
redox status of the cells in Alzheimer’s patients. Moreover, it has recently been reported
that the use of bulkyl aromatic fragments on C-9 can avoid the interaction of the
tetrahydroacridine core with the heme-iron center of cytochrome P450 enzyme, thus,
reducing the inherent hepatotoxicity of tacrine [28].
Interestingly, all the tacrine derivatives prepared herein showed remarkable
selectivity towards BuChE, which exerts a more prominent role at later stages of the
disease. The volume of the active site gorge is considerably higher (roughly 200 Å3)
than the one found in AChE [29], so BuChE can accommodate bulkier inhibitors, and
this may constitute the basis for the selectivity of these derivatives.
Inhibition of BuChE spanned over a quite wide range of potencies, ranging from
low micromolar to low nanomolar values (Table 1). The most potent compounds turned
out to be the ones bearing an ether-type linker, with more than 4000-fold higher activity
in comparison with galantamine.
Some of the most remarkable examples within each family (polyhydroxylated
imines, amines and ether-types) were selected for further studies as potential anti-
Alzheimer’s agents. Firstly, they were evaluated as inhibitors of the human enzymes
(hAChE and hBuChE). The calculated IC50 values are listed in Table 2.
Regarding the influence of the number of hydroxyl groups for the amino-type
tether, comparison of dihydroxylated derivative 33 and its mono-hydroxylated
counterpart, the latter already reported by Luo et al. [27] showed virtually no change in
the AChE inhibition (Table 2; 29.0 vs. 25.6 nM, respectively), and a moderate increase
in selectivity towards BuChE (1.3 vs. 3.4, respectively).
Cleary, imines and amines behaved as much weaker inhibitors than ether-
containing derivatives; compounds 42, 44, 46 were strong hBuChE inhibitors with
activities in the low nanomolar range. In particular, the dimethoxy derivative 43 and
dibenzyloxy derivative 45 were found to be extremely potent inhibitors of title enzyme,
within the subnanomolar range; this represents a roughly 85-fold increase compared to
the parent tacrine, used herein as a reference drug. Accordingly, the key structural
motifs for achieving the highest inhibitory potency are the presence of the tacrine core,
an ether-type tether, and a dialkoxylated phenyl residue. Dialkoxylated derivatives were
found to be much stronger inhibitors of hBuChE than their mono-substituted
counterparts (42 vs. 43, 44 vs. 45).
11

ACCEPTED MANUSCRIPT
All the ether-containing derivatives were tested as potential in vitro inhibitors for
the Aβ self-aggregation (1:1 inhibitor/Aβ ratio), another relevant hallmark in
Alzheimer’s disease (data shown in Table 2); for that purpose, the thioflavin-T
fluorescence assay was used [30]. All tested derivatives exerted a significant inhibition
of amyloid self-aggregation with inhibitory potencies in a relatively close range (from
54.4 to 74.3%). Despite the limited number of tested compounds, some considerations
could be drawn. In particular, the type of substituent at the aromatic ring seems to
influence activity and the following trend can be observed: OBn>OMe>OH. On the
other hand, the mono or disubstitution pattern does not seem to influence the inhibitory
activity.
Thus the mono- and di-substituted benzyloxy derivatives 44 and 45 were the
most potent derivatives, showing inhibitory potencies similar to that of the known
multipotent compound bis(7)tacrine, a tacrine homodimer, which is one of the most
prominent tacrine-based cholinesterase inhibitors reported so far [31]. Inhibitory activity
of compound 47 could not be determined because of a significant interference in the
assay conditions (quenching of the thioflavin T fluorescence signal).
12

ACCEPTED MANUSCRIPT
Table 1. Inhibitory properties of tacrine-polyphenols (Ki ± SD)
AChE (electric eel)
BuChE (equine serum)
Type of
Type of
Compound
Series
inhibiton
inhibiton
Kia (µM)
Kib (µM)
Kia (nM)
Kib (nM)
2.5 ± 0.6
1.9 ± 0.4
271 ± 36
35 ± 8
6
7
643 ±
327
39 ± 10
23 ± 6
2925 ± 1183
Imine
Mixed
24 ± 4
20 ± 1
637 ± 79
704 ± 82
173 ± 42
192 ± 74
8
9
2.3 ± 0.8
3.0 ± 0.8
Mixed
Amine
(1C)
31 ± 6
---
10
11
12
13
14
15
0.64 ± 0.08
3.0 ± 0.6
0.21 ± 0.06
1.8 ± 0.1
6.4 ± 1.3
17 ± 5
Imine
Amine
(1C)
Uncompetitive
0.37 ± 0.04
2.1 ± 0.3
0.081 ± 0.013
0.82 ± 0.27
0.19 ± 0.06
0.17 ± 0.04
12 ± 2
11 ± 4
25 ± 6
51 ± 17
3.2 ± 0.5
1.9 ± 0.8
3.2 ± 1.2
5.8 ± 1.5
Imine
Amine
(1C)
Mixed
Non-
0.19 ± 0.06
0.40 ± 0.19
competitive
Imine
Mixed
13

ACCEPTED MANUSCRIPT
Mixed
1.6 ± 0.5
22 ± 7
184 ± 46
(Uncompetitive)
19
0.61 ± 0.15
1.1 ± 0.6
Amine
(2C)
Mixed
0.11±0.01
0.032±0.005
20a
Mixed
0.87 ± 0.43
3.2 ± 0.8
0.59 ± 0.38
---
128 ± 20
28 ± 9
20b
25
Competitive
1202 ± 130
1898 ± 590
30 ± 14
304 ± 81
Cyclic
amine
275 ±
105
1.8 ± 1.2
0.16 ± 0.09
0.12 ± 0.05
2.5 ± 1.5
0.12 ± 0.03
0.058 ± 0.015
27
32
33
Mixed
Amine
9.9 ± 3.7
(5- and
6C)
7.9 ± 1.0
1.7 ± 0.4
Non-
29 ± 5
42
43
44
0.075 ± 0.028
0.17 ± 0.02
0.11 ± 0.02
0.075 ± 0.028
0.055 ± 0.016
0.11 ± 0.02
2.9 ± 1.1
1.9 ± 0.8
1.5 ± 0.9
competitive
Mixed
7.0 ± 0.4
1.9 ± 0.9
Non-
Ether
competitive
45
46
47
0.34 ± 0.05
0.14 ± 0.02
0.21 ± 0.02
0.13 ± 0.03
0.10 ± 0.06
4.6 ± 0.6
---
2.7 ± 0.9
4.5 ± 1.5
9.0 ± 2.2
Mixed
Uncompetitive
Mixed
43 ± 7
0.077 ± 0.016
14

ACCEPTED MANUSCRIPT
0.023 ± 0.010
3.0 ± 0.9
---
---
17 ± 3
---
---
Tacrine
Competitive
Competitive
7700 ± 1700
Galantamine
15

ACCEPTED MANUSCRIPT
Table 2. Anti-Alzheimer’s properties of selected compounds
%Inhibition
Aβ42
hBuChE
Selectivity
index (SI)
Compound
hAChE (IC50, nM)
self-aggregation
(IC50, nM)
(at 50 µM)
11
(R1= R3= H,
R2= OH)
1475 ± 311
609 ± 26
194±15
41.5 ± 4.06
120± 49
7.6
14.8
8.7
---
---
---
R2
R1
R3
13
(R1= H, R2=
R3= OH)
N
15
N
(R1= R2= R3=
OH)
1040 ± 40
100 ± 13
74.8 ± 3.4
1.3
---
14
16

ACCEPTED MANUSCRIPT
25.6 ± 0.9
Ref [27]
7.50 ± 0.02
Ref [27]
---
3.4
1.3
---
---
29.0 ± 1.3
22.9 ± 1.2
33
42
(R1= H, R2=
OMe)
216 ± 45
168 ± 37
454 ± 69
74.5
323.1
65.8
63.5 ± 1.0
63.8 ± 4.8
70.0 ± 4.8
2.92 ± 0.26
0.515 ± 0.050
6.95 ± 0.21
43
(R1= R2=
OMe)
44
(R1= H, R2=
OBn)
45
(R1= R2=
OBn)
506 ± 64
142 ± 16
1012
23.3
75.3 ± 0.4
54.4 ± 6.6
0.497 ± 0.059
6.06 ± 0.59
46
(R1= H, R2=
17

ACCEPTED MANUSCRIPT
OH)
47
542 ± 16
412 ± 15
35.6 ± 2.1
44.2 ± 1.7
15.2
9.3
(*)
<5
(R1= R2= OH)
Tacrine
2010 ± 150
[32]
20700 ± 1500
[32]
<5
0.1
Galantamine
[32]
5.66
[31]
0.81 ± 0.09
[31]
71.1 ± 0.7
Bis(7)tacrine
0.14
(*) Not tested, due to significant interference (quenching of the Thioflavin T fluorescence signal)
18

ACCEPTED MANUSCRIPT
2.2.2. Molecular modelling
Molecular modelling calculations were used to analyze the interaction between
tacrine and its derivatives 43, 45−47, and hBuChE.
Crystallographic coordinates (PDB code: 4bds) were used as jumping-off place
to study the molecular conformation of compounds 43, 45−47 in the binding site of the
enzyme. The analysis of contacts of native tacrine reveals that the main stabilization of
tacrine is due to π-π stacking interactions involving the Trp82 residue (Figure 2). Two
different states, protonated and non-protonated, were found for tacrine ligand. These
states modify the strength of π-π interactions form −18.5 (non-protonated) to −21.9
kcal/mol (protonated). Therefore, the protonated state would be predominant at
physiological pH.
The value of these interaction energies is too high for an individual π-π stacking
(ca. 6 kcal/mol). This is because the π-π stacking interaction works together with CH-π
interactions in which two protons clearly point towards the Trp aromatic ring.
Figure 2. (a) Crystallographic coordinates of tacrine (PDB: 4bds) and (b) molecular interactions
involving this compound with the binding site of BuChE
This stacking interaction of native tacrine is also found in all the derivatives
subjected to modelization. However, the values of the interaction energies are slightly
modified due to additional hydrophobic contacts and CH-π bonding with His438. We
focused our attention on the other interactions tacrine derivatives might establish. In this
context, the introduction of an alkyl chain provides enough conformational flexibility to
maximize interactions.
19

ACCEPTED MANUSCRIPT
In order to understand the role of both hydroxyl groups of hybrid 47 (IC50 = 36 ±
2 nM), the non-covalent contacts involving this moiety were analyzed. The catechol
moiety (R1 = R2 = OH) enables additional hydrogen bonding interactions (involving
Gln119 and Ser287, Figure 3), and the stabilization (15.3 kcal/mol) of these combined
H-bonds was estimated by means of DFT-based models.
Figure 3. (a) Molecular modelling of tacrine-phenolic heterodimer 47 in the binding site of BuChE and
(b) molecular interactions involving this compound. Only the main residues are shown for a better clarity.
The outcomes from molecular modelling highlighted the role of hydroxyl groups
of the aromatic ring. When these hydroxyl groups are protected (OBn, compound 45),
the hydrogen bonding with Gln119 and Ser287 is broken, and an increase in the
inhibition potency (IC50 = 0.50 ± 0.06 nM) was observed. In fact, conformational
flexibility allows the aromatic ring to be orientated towards the protein surface, showing
intramolecular CH-π interactions and two intermolecular CH-π contacts (∆Eint = −8.0
kcal/mol for one of them) with Ala277 residue (Figure 4a). In addition, van der Waals
interactions with residues Ile69 and Asp70 and the alkyl chain stabilize the complex.
The comparison between compound 45 and 43 (R1 = R2 = OMe, IC50 = 0.515 ±
0.050 nM) provides information about subtle differences on active pocket. The penalty
caused by hindered benzyl groups in compound 45 is balanced by aforementioned CH-π
interactions, however, this kind of contacts are missing in compound 43 bearing
methoxy substituents. Despite this, two new CH-π contacts are observed with Gln69-
Asn68 (∆Eint = −8.5 kcal/mol) and Leu88-Ser89 (∆Eint = −2.3 kcal/mol) dipeptides
moieties. Significant differences between both binding energies of CH-π contacts can be
20

ACCEPTED MANUSCRIPT
explained from differences of computed MEP surface (−19 vs. −17 kcal/mol,
respectively for each 3,4-dimethoxyphenyl face).
Remarkably, the inhibitory potency of compound 46 (IC50 = 6.1 ± 0.6 nM)
bearing a single hydroxyl group is significantly higher than that of derivative 47.
Although 46 is unable to establish the H-bond with Ser287 (∆Ecorr = −8.7 kcal/mol), two
stronger H-bond (∆Eint = −17.5 kcal/mol) involving the negatively charged –COO−···H
of Glu276, and Gln119, as well as a new NH-π contact (∆Eint = −14.1 kcal/mol)
between the phenol ring and Asn68 (Figure 4b) are likely responsible for the stronger
activity. These interactions provide a better stabilization.
Figure 4. Protein-ligand interactions found in the interplay between BuChE and compounds 45 (a) and 46
(b). Only the main residues are shown for clarity
.
In order to get further insights about the selectivity of compound 43 towards
BuChE instead of AChE, we have also modeled this compound in its corresponding
active site (PDB code: 4w63). Herein, compound 43 is only stabilized by non-covalent
contacts involving Trp84 and His440 residues with tacrine ring. As criteria, we have
considered that a better inhibition should be characterized by a more favorable binding
energy value. Therefore, we have roughly estimated the overall calculated binding
energy values in binding sites (Figure 5) of AChE (∆Eint = −10.2 kcal/mol) and BuChE
21

ACCEPTED MANUSCRIPT
(∆Eint = −33.4 kcal/mol). Based on these values, our results follow the similar trend of
experimental selectivity index for compound 43. The main reason seems to be the
absence of interactions involving the phenolic moiety due to a short spacer regarding
crystallized tacrine-based ligand of 4w63.
Figure 5. (a) Molecular modelling of tacrine-phenolic heterodimer 43 in the binding site of AChE and (b)
modelling involving this compound and hBuChE.
2.2.4. Neuroprotection
Neurodegeneration in rat primary neurons was induced by serum and K+
deprivation, which, upon partial ATP depletion, has been reported to cause oxidative
stress and neuronal death through an apoptotic pathway [33]. This represents a good
model of neuronal injury and senescence [46]; neuronal viability was determined using
the MTT assay. Tacrine was used as a reference compound, and the two more potent
BuChE inhibitors, 43 and 45, were evaluated. Interestingly, only the dimethoxy
derivative 43 showed relevant neuroprotective activity up to a concentration of 10 µM
(Figure 6).
22

ACCEPTED MANUSCRIPT
Figure 6. Neuroprotection assay (serum and K+ deprivation) for tacrine and tacrine hybrids 43, 45. White
bars represent control condition, whereas black bars represent serum/K+ deprivation in absence or
presence of increasing concentrations of the tested compounds. Results are expressed as percentage of
controls and are the mean ± SE of at least 3 different experiments, each run in triplicate. *p < 0.05; **p <
0.01 relative to control CGNs at 0µM; #p < 0.05, ##p < 0.01 relative to control CGNs treated with the
same concentration of the compound. Bonferroni's post-hoc test following one-way ANOVA.
2.2.3. Neurotoxicity
With these data in hand, heterodimers 43 and 45, the most potent hBuChE
inhibitors, were further tested to assess their safety profile on neurons. Thus, potential
neurotoxicity against immortalized rat primary cerebellar granule neurons (CGNs; MTT
viability test) was evaluated for these compounds (Figure 7). Whereas dibenzylated
derivative 45 showed significant neurotoxicity even at 1 µM concentration (roughly
50% of neuronal survival), dimethoxy counterpart 43 exhibited no significant
neurotoxicity up to 5 µM.
23

ACCEPTED MANUSCRIPT
Figure 7. Neurotoxicity of tacrine (A) and tacrine-phenolic heterodimers 43 (B), 45 (C). Results are
expressed as percentage of controls and are the mean ± SE of at least 3 different experiments, each run in
triplicate. **p < 0.01, ***p < 0.001 relative to control CGNs at 0µM. Bonferroni's post-hoc test
following one-way ANOVA
We also searched for potential antiproliferative properties in a small panel of the
compounds prepared herein; dual activity (inhibition of AChE-antiproliferative
properties) has been reported for some marketed chemotherapeutic agents [34].
A panel of six human tumor cell lines was used, namely A549 (non-small cell
lung), HBL-100 (breast), HeLa (cervix), SW1573 (non-small cell lung), as drug
sensitive lines, T-47D (breast) and WiDr (colon) as drug resistant lines.
Antiproliferative activities, in terms of GI50 values (µM) are depicted in Supporting
Information file (Table S1). Interestingly, 9-piperidine-tetrahydroacridine 27 exhibited
strong antiproliferative activity (GI50 = 5.2, 4.5 µM) against the two drug resistant lines
(T-47D and WiDr); this represents an increase of activity of up to 10-fold when
compared to classical chemotherapeutic agents, like 5-fluorouracil used as a reference
compound. Moreover, for such lines, a selectivity index of roughly 5 was found when
comparing with non-tumor cell lines (human fibroblast), in contrast with
chemotherapeutic agents, which turned out to be more potent against fibroblasts than
against most of the tested tumor cell lines.
2.2.4. Hepatotoxicity assessment
Hepatotoxicity of tacrine and the two tacrine-phenolic heterodimers considered
as potential lead compounds in terms of anti-cholinesterases (derivatives 43 and 45) was
evaluated after 24 hours by MTT viability test (Figure 8). Whereas dibenzylated
derivative 45 showed significant hepatotoxicity at all concentrations, dimethoxy
24

ACCEPTED MANUSCRIPT
counterpart 43 exhibited a constant hepatic survival (roughly 60-70% of hepatic
survival).
Figure 8. Hepatotoxicity of tacrine (A) and tacrine-phenolic heterodimers 43 (B), 45 (C). Results are
expressed as percentage of controls and are the mean ± SE of at least 2 different experiments, each run in
duplicate. *p < 0.05 **p < 0.01 relative to control HepG2 at 0µM. Bonferroni's post-hoc test following
one-way ANOVA
Taken all together, data collected herein clearly indicate that compound 43
exhibits the best pharmacological profile and thus, can be considered as a valuable lead
multitarget drug for the treatment of Alzheimer’s disease, in contraposition of
dibenzylated derivative 45. The latter, despite the good profile concerning inhibition of
BuChE and inhibition of Aβ deposition, exhibits significant neuro- and hepatoxicity, as
indicated by the bioassays conducted herein; as a result, compound 45 must be
discarded as a potential anti-Alhzeimer’s agent.
2.2.5. Stability in plasma
Stability of new inhibitors in plasma is an important parameter in order to assess
potential degradation which may affect in vivo activity. Stability of 43, the compound
with the best pharmacological profile, at 37°C in plasma was assayed using a LC-MS
approach and propanolol as internal standard (IS). Three time points were considered,
namely 60, 180 and 360 min. The ratio of 43 and IS peak areas at t = 0 and at the
selected time points was compared (Figure 9). Within the selected time-frame no
significant degradation was observed indicating good plasma stability for 43, suggesting
the stability of the ether-based linker under physiological conditions.
25

ACCEPTED MANUSCRIPT
3
2
1
0
0
100
200
300
400
Incubation time (min)
Figure 9. Plasma stability of derivative 43. The ratio between the areas of 43 and IS is plotted
as a function of the incubation time. The study was performed at 37°C using a concentration of
43 equals to 10 µM.
3. Conclusions
In conclusion, we have accomplished the preparation of a plethora of tacrine-
polyphenol heterodimers, modifying the type (imino, amino, ether) and length of the
tether, together with the nature and position of the substituents on the aryl core.
Extensive analysis of bioactivities has revealed that the dimethoxy derivative 43 is a
promising lead candidate for Alzheimer’s disease treatment. This compound hits two
key hallmarks of the Alzheimer’s disease. Indeed, it behaves as an extremely potent and
selective inhibitor of hBuChE, with activity in the subnanomolar range and inhibits to a
good extent the self-aggregation of β-amyloid peptide. The inhibitory profiles against
cholinesterases were explained via molecular modelling, by analysing the key
intramolecular interactions. Compound 43 showed to be safe for neurons at
concentration as high as 5 µM, and to provoke only limited hepatotoxicity at relatively
high concentrations; interestingly, it exhibited neuroprotective properties in primary rat
neurons, using a model in which neuronal damage was induced by serum and K+-
deprivation, a well-validated method for reproducing neuronal injury and senescence.
Moreover, its ether-type linker proved to be stable under physiological conditions.
Overall, the tacrine-dimethoxyphenyl hybrid 43 exhibits excellent properties as a
novel multi-target drug for tackling Alzheimer’s disease.
26

ACCEPTED MANUSCRIPT
4. Materials and methods
4.1. Materials and methods
4.1.1. General procedures
13
1H (300.1 and 500.1 MHz) and C (75.5 and 125.7 MHz) NMR spectra were recorded
at 25 °C on a Bruker Avance 300//Avance III 500 MHz spectrometers using the solvent
1
indicated in each case. H and 13C assignments were confirmed by 2D COSY and
HSQC experiments. Mass spectra (ESI) were recorded on a Q Exactive mass
spectrometer. TLCs were performed on aluminium pre-coated sheets (E. Merck Silica
gel 60 F254); spots were visualized by UV light, and by charring with 10% vanillin in
EtOH containing 1% of H2SO4. Column chromatography was performed using E.
Merck Silica Gel 60 (40–63 µm), using the eluant indicated in each case.
4.1.2. Inhibition of cholinesterases
Ellman’s assay [25] was followed with minor modifications. For non-human
enzymes (AChE from Electrophorus electricus and BuChE from equine serum), acetyl-
and butyrylthiocholines iodides were used as substrates (14.5−116 µM and 9.2−73.6
µM). Stock solutions of the inhibitor were prepared in DMSO; maximum DMSO
concentration was 1.25%. Activity was monitored in 1.2-mL Polystyrene cuvettes
containing 0.1 mM phosphate buffer (pH 8.0), 3 mM DTNB, different substrate
concentrations (5 different concentrations), inhibitor (or solvent) plus water up to a
constant volume of 1.14 mL. Reaction was started by adding 60 µL of properly diluted
enzyme solution at 25 ºC. The formation of the chromophore was monitored at 405 nm
for 125 s. Initial rates were calculated from the slopes of the plots obtaining when
representing Abs vs. t. For the calculation of the inhibition constants (Ki’s) and the
mode of inhibiton, Lineweaver-Burk plot (double reciprocal plot) was used (1/V vs.
1/[S]) for 2-3 different inhibitor concentrations.
Ellman’s assay was also used for estimating the activity of human enzymes [26].
Human recombinant AChE (Sigma, Italia) and BuChE from human serum (Sigma,
Italia) were used for the evaluation.
27

ACCEPTED MANUSCRIPT
4.1.3. Inhibition of Aβ42 self-aggregation
Previously described methodology was used [26]. Inhibitors were solubilized in
methanol at 2.5 mM concentration and diluted in the assay buffer. Each compound was
assayed in duplicate in at least two independent experiments.
4.1.4. Molecular modelling
Human BuChE complexed with tacrine was obtained from Protein Data Bank (PDB
code: 4bds) [35]. This protein was prepared with MAESTRO software suite [36]. Then,
each tacrine derivative was built by modifying the tacrine moiety of 4bds structure.
Analogous methodology was applied for AChE complexed with tacrine-benzofuran
hydrid inhibitor (PDB code: 4w63) [37].
Molecular dynamic simulations were carried out using Chem3D program. The
equilibration protocol consisted of an initial molecular mechanic minimization using
MM2 force field (RMS gradient 0.01), followed by a minimization just moving the
tracrine derivative (10,000 steps) whereas the system was heated at a constant volume
until 300 K using a time constant for the heat bath coupling of 1 ps.
Non-covalent interactions between tacrine models and binding sites were evaluated
using the crystallographic coordinates at M06-2X/def2-TZVP level of theory with
Gaussian 09 package [38]. In addition, the Grimme’s dispersion correction [39] was
also included as implemented in Gaussian 09. The basis set superposition error for the
calculation of interaction energies has been corrected using the counterpoise method
[40] as followed:
∆ꢀꢁꢂꢃ = ꢀꢄꢅꢄꢃꢆꢇ − ꢀꢈ − ꢀꢉ
∆ꢀꢊꢋꢌꢌ = ∆ꢀꢁꢂꢃ + ꢀꢍꢎꢎꢏ
Models of compounds 45, 46 and 47 were fully optimized at the same level of theory.
Frequency analysis was performed to characterize these structures as minima (Nimag
=
0). Solvent effects (H2O) were evaluated during the optimization by applying the
polarizable continuum model (PCM) with the integral equation formalism variant
(IEFPCM).
28

ACCEPTED MANUSCRIPT
4.1.5. Neuroprotection assays
Two different models for analyzing the potential neuroprotection exerted by tested
compounds were used; in both cases, neuronal viability was quantified using the MTT
(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay.
In one of such models, 6-hydroxydopamine was used as the neurotoxic agent (See
Supporting Information).
In the other one, neurodegeneration was achieved by serum and K+-deprivation for 48 h
[33,41].
4.1.6. Neurotoxicity and cytotoxicity assays
Previously described methodologies were used [26]. For the in vitro antiproliferative
activity, the protocol of the USA National Cancer Institute (NCI), against the six human
solid tumor cell lines, and the one non-tumor (BJ-hTert, human fibroblasts) with minor
modifications, was followed [42].
4.17. Cell cultures for hepatotoxicity assays
Human hepatoma cell line (HepG2) was cultured in Dulbecco Modified Eagle Medium
(DMEM) supplemented with 10% heat inactivated Fetal Bovine Serum (FBS), 1%
Penicillin/Streptomycin and 2mM Glutamine (all cell cultures’ reagents were from
Aurogene Srl, Rome, Italy). At confluence, cells were trypsinized for 5 minutes at 37 °
C and trypsin was inactivated with complete DMEM medium. Detached cells were than
collected, centrifuged for 5 minutes at 300xg and resuspended to be counted. For the
experiments, hepatocytes were plated on 96 well plates, previously coated with 10
µg/mL poly-L-lysine (Sigma-Aldrich) at the density of 2x104/0.2 mL medium/well in
presence or absence of increasing concentrations of the compounds (0, 1, 2.5, 5, 10 and
25 µM) in serum-free medium for 24h.
29