A new naphthalene derivative with anti-amyloidogenic activity as potential therapeutic agent for Alzheimer's disease

Article abstract of DOI:10.1016/j.bmc.2020.115700

The aggregation of β-amyloid peptides is associated to neurodegeneration in Alzheimer's disease (AD) patients. Consequently, the inhibition of both oligomerization and fibrillation of β-amyloid peptides is considered a plausible therapeutic approach for AD. Herein, the synthesis of new naphthalene derivatives and their evaluation as anti-β-amyloidogenic agents are presented. Molecular dynamic simulations predicted the formation of thermodynamically stable complexes between the compounds, the Aβ_1-42 peptide and fibrils. In human microglia cells, these compounds inhibited the aggregation of Aβ_1-42 peptide. The lead compound 8 showed a high affinity to amyloid plaques in mice brain ex vivo assays and an adequate log P_oct/PBS value. Compound 8 also improved the cognitive function and decreased hippocampal β-amyloid burden in the brain of 3xTg-AD female mice. Altogether, our results suggest that 8 could be a novel therapeutic agent for AD.

Full text of DOI:10.1016/j.bmc.2020.115700

Journal Pre-proofs
A new napthalene derivative with anti-amyloidogenic activity as potential
therapeutic agent for alzheimer's disease
Suchitil Marrero-Rivera, Alberto Bencomo Martinez, Erika Orta Salazar,
Marquiza Sablón-Carrazana, Laura García-Pupo, Florencia Zoppolo,
Florencia Arredondo, Rosina Dapueto, María Daniela Santi, Ingrid
Kreimerman, Tania Pardo, Laura Reyes, Lidice Galán, Samila León-
Chaviano, Luis A. Espinosa-Rodríguez, Roberto Menéndez-Soto del Valle,
Eduardo Savio, Sofia Diaz Cintra, Chryslaine Rodriguez-Tanty
PII:
S0968-0896(20)30530-7
https://doi.org/10.1016/j.bmc.2020.115700
BMC 115700
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Accepted Date:
24 July 2020
4 August 2020
Please cite this article as: S. Marrero-Rivera, A. Bencomo Martinez, E. Orta Salazar, M. Sablón-Carrazana, L.
García-Pupo, F. Zoppolo, F. Arredondo, R. Dapueto, M. Daniela Santi, I. Kreimerman, T. Pardo, L. Reyes, L.
Galán, S. León-Chaviano, L.A. Espinosa-Rodríguez, R. Menéndez-Soto del Valle, E. Savio, S. Diaz Cintra, C.
Rodriguez-Tanty, A new napthalene derivative with anti-amyloidogenic activity as potential therapeutic agent for
alzheimer's disease, Bioorganic & Medicinal Chemistry (2020), doi: https://doi.org/10.1016/j.bmc.2020.115700
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. 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.
©
2020 Published by Elsevier Ltd.

1
2
A NEW NAPTHALENE DERIVATIVE WITH ANTI-AMYLOIDOGENIC ACTIVITY AS
POTENTIAL THERAPEUTIC AGENT FOR ALZHEIMER'S DISEASE
3
§
[a]
§[a]
§[b]
4
5
6
7
8
Suchitil Marrero-Rivera , Alberto Bencomo Martinez
, Erika Orta Salazar , Marquiza
[
a]
[a]
[c]
[c]
Sablón-Carrazana , Laura García-Pupo , Florencia Zoppolo , Florencia Arredondo ,
[
c]
[c]
[c]
[c]
[c]
Rosina Dapueto , María Daniela Santi , Ingrid Kreimerman , Tania Pardo , Laura Reyes ,
[
a]
[a]
[d]
Lidice Galán , Samila León-Chaviano , Luis A. Espinosa-Rodríguez , Roberto Menéndez-
Soto del Valle , Eduardo Savio , Sofia Diaz Cintra* [b], Chryslaine Rodriguez-Tanty*[a].
[
a]
[c]
9
0
1
1
[a] Department of Neurochemistry, Cuban Center for Neurosciences, Street. 190 e/ 25 and 27.
Cubanacan. Playa. Havana, CP 11600. Cuba.
1
1
1
2
3
[b] Institute of Neurobiology (INB), Developmental Neurobiology and Neurophysiology, UNAM
Juriquilla Querétaro, México.
1
4
1
1
5
6
[c] Biomedical and Pharmaceutical Chemistry, Uruguayan Centre of Molecular Imaging
(CUDIM), Montevideo, Uruguay.
1
7
1
1
8
9
[d] Center of Genetic Engineering and Biotechnology (CIGB), Ave 31 e/ 158 and 190, Havana,
CP10600, Cuba.
2
0
2
2
1
2
§The first three authors contributed equally to this work.
2
2
3
4
Address all correspondence to *:
2
2
5
6
*Chryslaine Rodriguez-Tanty, Fax number (53 7) 208 67 07; E-mail: chris@cneuro.edu.cu
*Sofia Diaz Cintra, Fax number 52 (55) (442)238 1038; E-mail: yodi@unam.mx
1

2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
A NEW NAPTHALENE DERIVATIVE WITH ANTI-AMYLOIDOGENIC ACTIVITY AS
POTENTIAL THERAPEUTIC AGENT FOR ALZHEIMER'S DISEASE
Abstract
Alzheimer's disease (AD) is a highly prevalent neurodegenerative disorder that primarily
affects the elderly population worldwide. One of several theories explaining the
pathophysiology of AD relates to β-amyloid aggregates, which have emerged as the main
cause of synaptic loss and neurodegeneration in AD patients. Previous research has shown
that β-amyloids comprise a wide spectrum of species, including monomers, oligomers and
fibrils. While the physiological role of β-amyloid monomers has been highlighted, oligomers
and fibrils appear to be the main contributors to neurotoxicity in AD. Consequently, a
therapeutic approach to the disease should include the inhibition of both oligomerisation and
the fibrillation process. This paper outlines the synthesis of new naphthalene derivatives, as
well as their in silico, in vitro and in vivo evaluations of their properties as anti-β-amyloidogenic
agents. Molecular dynamic simulations predicted the formation of thermodynamically stable
(ΔG < 0) complexes between the compounds, the Aβ1-42 peptide and fibrils, as a result of
hydrophobic interactions, van der Waals forces and hydrogen bond formation, with the amino
acids fundamentally associated with the closure of the β-conformational structure. The in vitro
results obtained in the inhibition of Aβ1-42 peptide aggregation in human immortalised microglia
cells (at nanomolar concentrations) agreed with in silico evaluation. According to these results,
we selected one of the synthesised compounds as the most promising anti-amyloidogenic
agent (compound 8). In the in silico evaluations, this compound exhibited the highest Keq
between all the complexes studied. In addition, in the ex-vivo assays, the compound 8 had a
high affinity to the amyloid plaques (K = 16 nM) evaluated in the serial brain sections of
d
transgenic 3xTg-AD mice (female, aged 9 months), as well as an adequate log Poct/PBS value
2

5
5
5
5
5
5
5
3
4
5
6
7
8
9
(1.77 ± 0.07). This predicts its ability to passively cross the blood-brain barrier (BBB). In vivo
assays showed that daily oral administration of 8 (1 mg/kg for eight weeks) in 3xTg-AD female
mice (t : 10-12, t : 12-14 months) improved cognitive function and led to a significant decrease
o
f
(of around 50%) in amyloidogenic burden in the hippocampus, according to our
immunohistochemical studies. These preliminary results suggest that 8 could be a novel
therapeutic agent for AD.
6
6
6
0
1
2
Keywords: Alzheimer’s disease, anti-amyloidogenic agent, molecular dynamic simulation,
3xTg-AD, cognitive memory.
3

6
4
Introduction
Alzheimer’s disease (AD) is the most common neurodegenerative disorder affecting the
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
8
8
8
8
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
elderly population worldwide [1]. Current clinically approved therapeutic agents for AD include
an N-methyl-D-aspartate (NMDA) receptor antagonist (memantine) and acetylcholinesterase
inhibitors (donepezil, galantamine and rivastigmine). Although these compounds deliver slight
improvements in memory and cognition, they have no significant effect on the
neurodegenerative progression of AD [2]. The precise etiology of AD remains elusive,
however, several mechanisms and hypotheses related to the progression of AD have been
proposed, including a dysfunction in cholinergic transmission, homeostatic imbalance of
redox-active metals and oxidative stress [3,4]. Among the several theories explaining the
pathophysiology of AD, β-amyloid aggregates have emerged as the primary cause of synaptic
loss and neurodegeneration related to the disease. Previous evidence have shown that β-
amyloid comprises a wide spectrum of species, including monomers, oligomers and fibrils. It is
known that the progressive aggregation and accumulation of Aβ can lead to the formation of
extracellular senile plaques, one of the morphological hallmarks of this disease [5,6]. Aβ
peptides, which are the main component of these plaques, remain in equilibrium with
surrounding diffusible Aβ oligomers and protofibrils [7-11]. At the present, the amyloid cascade
hypothesis suggests that soluble β-amyloid oligomers or prefibrillar-aggregated intermediates
are the major toxic species in cases of AD [12-14]. Accordingly, soluble oligomers are known
to cause glial activation, neuro-inflammation [15], mitochondrial dysfunction [16], induction of
hyperphosphorylated tau protein [17], induction of lysosomal proteases [18] and deficits in
synaptic long-term potentiation [19]. According to clinical studies, free levels of Aβ1-42 peptides
in the cerebrospinal fluid (CSF) decrease long before the first symptoms of familial AD and
these seem to be the earliest species to accumulate in the brain [20]. However, over time,
these Aβ forms evolve into a complex array of isomerised, phosphorylated and/or truncated
4

8
9
9
9
9
9
9
9
9
9
9
9
0
1
2
3
4
5
6
7
8
9
peptides [21]. Although Aβ oligomers with high molecular weight are the most frequently seen
forms of Aβ in the soluble fraction of AD brain cortex, it is the population of smaller diffusible
Aβ oligomers which demonstrates neurotoxicity [22].
The mechanism by which the assembly of Aβ leads to neuronal death is still unclear. There
have been multiple molecular studies seeking to understand the process of Aβ fibre formation,
their interactions with neuronal structures and the subsequent cell damage [23,24]. To date, it
has been shown that 60% of Aβ peptide hydrophilic interactions with the cellular environment
originate in the N-terminal amino acidic region (D1-K16). In this region, the amino acid K16 is
believed to play a key function due to its proximity to the hydrophobic nucleus region
(“L17VFFA21”), which is essential for the formation of Aβ oligomers and fibrils. Several studies
[25-28] have suggested that this conformational region has a substantial influence on the
properties of Aβ aggregation. On the other hand, the Aβ peptide C-terminal region, located
between the Q15 and V36 residues, mainly adopts an α-helix conformation, with a hinge from
G25 to N27, allowing the van der Waals interactions and electrostatic forces to stabilise the β-
conformation. Around 25% of this surface region is uninterruptedly hydrophobic, and the α-
helix structure is metastable, which can lead to a global conformational rearrangement and
permit the formation of a secondary intermolecular β-sheet structure [28-30].
1
1
1
1
1
1
1
1
1
1
1
1
1
1
00
01
02
03
04
05
06
07
08
09
10
11
12
13
Although there are several remaining questions regarding the amyloid cascade hypothesis,
targeting Aβ species with anti-aggregation compounds may be a plausible therapeutic
approach [31-33]. Various small molecules [34] have been assessed as disruptors of the
dynamic balance of the association/dissociation process that takes place between the different
Aβ species during amyloidogenic fibrillation. Evidence suggests that these dissociated Aβ
structures then become a less toxic species that can be eliminated in a variety of ways,
including the absorption of neurovascular cells, proteolytic degradation by proteases or
transport through the BBB or perivascular drainage [35-40].
5

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
The design of these small molecules can be performed by virtual screening using
databases of chemical compounds, by chemical modification of compounds that have
demonstrated activity as Aβ1-42 aggregation inhibitors or by designing a compound according
to pharmacophore models against Aβ1-42. Examples of these compounds are derivatives of
naphthalene, polyphenols, curcumin, thioflavin, pyridine, and others [41]. These compounds
have aromatic rings in their structures, allowing them to bind to amyloid fibrils by means of π-π
stacking and hydrophobic interactions within the steric zipper’s dry surface [48]. In addition,
some of these compounds have donor/acceptor groups for forming hydrogen bonds with the
amino acids of Aβ target. These kinds of interactions could disrupt the Aβ structure and avoid
the aggregation process of the peptides.
In our work, taking into account to pharmacophore models against Aβ1-42, we designed a
new family of naphthalene derivatives that carry in their structures an amidoalkylic chain with
diverse terminal functional groups (see Figure 1). The naphthyl ring constitutes a nucleus that
should be able to allow the hydrophobic interactions, through π-π stacking with the F-19 and
F-20 aminoacids, as anchor point with the target. But, on the other hand, this core structure is
associated with solubility problems that entailing to a poor biodistribution of molecule.
Therefore, the introduction of amidoalkylic chain has the goal to give a hydrophilic character to
the molecule and at the same time, it should be able to provide donor/acceptor groups to yield
it the possibility to interact with the lateral chain of the key amino acids involved in the
formation of the Aβ configuration, through hydrophobic and electrostatic interactions, van der
Waals forces and hydrogen bonding formation.
6

1
35
1
1
1
1
1
1
36
37
38
39
40
41
Figure 1. Design strategy for the synthesis of naphthalene derivatives that can interact with the key
sequence of Aβ1-42 peptide (SGYEVHQKLVFF) to avoiding its structural closure towards a β-
conformation. The red, green and blue dashed lines represent the π-π stacking, donor/acceptor groups
and hydrophobic interactions, respectively.
Previous in silico and in vitro studies have demonstrated that some members of this novel
family have the capacity to directly interact with native and/or aggregation-prone species and
fibril structures in bovine serum albumin, as well as with two amyloidogenic peptides, such as
amylin. In addition, some of these compounds prevented the cytotoxicity and apoptosis
induced by monomeric and aggregated human islet amyloid polypeptides in cultures of
granule cells [42,43]. Given these findings, we are seeking to further study the Aβ anti-
aggregation activity of the most promising naphthalene compounds, as well as their
neuroprotective potential. To this end, this paper describes the process of synthesising these
compounds and outlines several potentially beneficial aspects of their reaction pathways. In
silico molecular dynamics simulations were performed to identify the type of interaction and
stability of the complex formed between these derivatives with the monomeric Aβ1-42 peptide,
as well as with a fibrillary structure thereof, in order to predict whether these compounds can
1
1
1
1
1
1
1
1
1
1
1
1
42
43
44
45
46
47
48
49
50
51
52
53
inhibit the oligomerisation and fibrillization processes. In vitro assays were carried out in
7

1
1
1
1
1
1
1
1
1
1
54
55
human microglia cells to evaluate the anti-Aβ1-42 aggregation activity of the compounds. In
addition, the affinity constant of a selected compound (8) for amyloid plaque was analysed ex
56 vivo in brain serial sections of triple-transgenic mice (3xTg-AD, female), using radiochemical
57 techniques. Finally, the in vivo effect of this naphthalene derivative on Aβ plaque burden and
58 related memory dysfunction in aged 3xTg-AD mice (female) was also evaluated.
59
60
61
62
63
Results and Discussion
Chemistry
General synthesis of 1-naphthyl derivatives
The synthesis of various 1-naphthyl derivatives was carried out in several steps, outlined in
Scheme 1 [44].
O
O
S
H
H
O
O
N
From 4,
N
N
HN
R
n
CS , Cs CO , TBAI,
HN
N
S
2
2
3
H
O
(CH ) SO
4
O
(CH2CO)2O, NMM
Diamines or carboxylamines
3
2
1,4-Dioxane
1,4-Dioxane
R
n
Compounds
DMF
1
NH2
1
4
8
Diamine
HN
NH2
3
1
5
COOH
COOH
6
7
4
O
O
N
H
N
NH
s
NH2
n
H
e
i
O
O
m
l
a
y
n
Compounds
x
e
o
n
b
a
1
3
9
r
x
a
i
o
c
10
r
-
D
o
s
1
,
e
i
n
m
a
D
i
ia
D
m
in
e
O
O
O
O
N
O
OH
HN
HN
(
CH2CO)2O
O
NHS, DCC
,4-Dioxane
O
1
,4-Dioxane
1
2
3
1
1
64
65
Scheme 1. General procedure for the synthesis of 1-naphthyl derivative compounds.
1
1
66
67
The 1-(1-naphthyl)-2,5-pyrrolidinedione (1) compound is an interesting intermediate for
obtaining various naphthyl derivatives. This compound [45,46] has previously been
synthesised from 1-naphthylamine (as a starting material), however, these procedures were
inefficient and cumbersome. We developed a one-pot procedure for the conversion of 1-
naphthylamine into 1 (see Scheme 1) using succinic anhydride in the presence of N-
methylmorpholine (NMM), as a tertiary amine, in 1, 4-dioxane. This novel method has the
1
1
1
1
1
1
68
69
70
71
72
73
advantage of avoiding the isolation of (1-naphthylamino)-4-oxobutanoic acid (2) as an
8

1
1
1
1
1
1
1
1
1
74
75
76
77
78
79
80
81
82
intermediate product, thus resulting in adequate yield and purity of the desired product 1.
Several experiments were performed to establish optimal reaction conditions. Total conversion
of 1 was achieved using two equivalents of NMM with regards to 1-naphthylamine (verified by
thin layer chromatography [TCL]), although the yield of compound 1 dropped by 72% after
recrystallization. The TLC experimental results suggested that 1 was obtained through the
formation of an ammonium quaternary salt of the 4-(1-naphthylamino)-4-oxobutanoic acid (2).
An excess of NMM could result in proton elimination of the amide group and a consequent
facilitation of the internal cyclisation of the alkyl chain. In the absence of NMM (see Scheme 1),
2 was produced as a white solid with a yield of 67% [45].
1
1
1
1
1
1
1
83
84
85
86
87
88
89
The ester N-hydroxysuccinimide of the 4-(1-naphthylamino)-4-oxobutanoic acid (3; see
Supplementary Material, Fig.1.SM) was obtained with a good yield (see Scheme 1), as a
convenient precursor to introducing spacer arms with different lengths and diverse terminal
functional groups in derivative 2. Thus, four new compounds were obtained using
1
4
1
4
intermediates 1 or 3: N -(2-aminoethyl)-N -(1-naphthyl)succinamide (4), N -(4-aminobutyl)-N -
(1-naphthyl) succinamide (5), N-[4-(1-naphthylamino)-4-oxobutanoyl]-β-alanine (6) and 6-{[4-
(1-naphthylamino)-4-oxobutanoyl]amino}hexanoic acid (7) (see Table 1).
Table 1. Acylation reaction conditions (RC) for obtaining naphthyl derivatives from respective
substrates (S) with selected nucleophilic agents (NA) at different molar ratios (S:NA).
Yield
Entry
S
NA
S:NA
RC[a]
Product
%
1
2
3
4
1b
1
1:3
1:0.5
1:3
A
Ad
A
4c
9[e]
74
71
59
38
H N(CH ) NH
2
2 2
2
2
1
5f
H N(CH ) NH
2
2 4
1
1: 0.5
Ad
10[e]
9

5
6
7
8
3g
3g
3
H N(CH ) NH
1:3
1:3
B
B
4
5
69h
66h
40
2
2 2
2
2
H N(CH ) NH
2
2 4
H N(CH ) COOH
1:1.72
1: 1.72
C
6i
7k
2
2 2
3
H N(CH ) COOH
Cj
34
2
2 5

a
1
1
1
1
1
1
1
1
1
90
91
92
93
94
95
96
97
98
RC A: 1, 4-dioxane anhydrous, 25ºC, 24 h. RC B: 2 (1 equiv.), NHS (1.2 equiv.), DCC (1.2 equiv.), 1, 4-dioxane
anhydrous, 101ºC, 3 h, for obtaining 3 in situ, then diamine, 1 h, 25ºC. RC C: TEA (1.72 equiv.), anhydrous DMF,
b
c
1
4
[d]
100ºC, 20 h. 1-(1-naphthyl)-2,5-pyrrolidenedione,
but 101ºC.
and N -(1-naphthyl)-N -(4-{[4-(1-naphthylamino)-4-oxobutanoyl]amino}butyl)succinamide, N -(4-aminobutyl)-N -
(1-naphthyl)succinamide,
naphthylamino)-4-oxobutanoyl]-β-alanine,
oxobutanoyl]amino}hexanoic acid.
N -(2-aminoethyl)-N -(1-naphthyl) succinamide, idem A,
[
e]
1
4
By-product (N -(1-naphthyl)-N -(2-{[4-(1-naphthylamino)-4-oxobutanoyl]amino}ethyl)succinamide,
1
4
[f]
1
4
[
g]
h
i
3 is obtained in situ.
Global yield from 2 (see Scheme 1).
N-[4-(1-
[
j]
[k]
idem RC C, but
6
h.
6-{[4-(1-naphthylamino)-4-
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
99
00
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
According to our results related to compounds 2 or 3 (see Table 1, entries 1 to 6), reaction
conditions such as temperature and equivalent amount of diamines should be controlled in
order to obtain the desired compounds and to minimise the formation of by-products: N, N´-
disubstituted diamines 9 and 10 or 1-naphthylamime. When the reaction was performed with
an excess of ethylenediamine (up to six equivalents) and at room temperature, the main
product obtained was 4 (see Table 1, entry 1). However, when a 1:0.5 molar ratio of
1:ethylenediamine was used or when the reaction was carried out at high temperature, the
1
4
N,N’-disubstituted
diamine
(N -(1-naphthyl)-N -(2-{[4-(1-naphthylamino)-4-oxobutanoyl]
amino}ethyl)succinamide, 9) was obtained. Similar results were achieved with the 1,4-
1
4
butylenediamine, yielding N -(4-aminobutyl)-N -(1-naphthyl)succinamide (5) and the by-
1
4
product
of
N -(1-naphthyl)-N -(4-{[4-(1-naphthylamino)-4-oxobutanoyl]amino}butyl)
succinamide (10) (see Table 1, entries 3 and 4, respectively). When compounds 1 or 3 were
used as starting materials, we did not observe any differences in the yield of the synthesised
compounds. All these new compounds were characterised using spectroscopic techniques
(see Supplementary Material, Fig.2.SM to Fig.9.SM and Fig.18.SM to Fig.19.SM).
Likewise, derivative 3 could be used to introduce a terminal carboxylic group, through its
reaction with carboxylamines (β-alanine and aminocaproic acid) as nucleophilic agents to
10

2
2
2
2
2
2
2
2
2
2
2
2
2
2
16
17
18
19
20
21
22
23
24
25
26
27
28
29
obtain N-[4-(1-naphthylamino)-4-oxobutanoyl]-β-alanine (6) and 6-{[4-(1-naphthylamino)-4-
oxobutanoyl]amino}hexanoic acid (7), respectively (see Table 1, entries 7 and 8). The
syntheses of 6 and 7 were carried out in the presence of triethylamine (TEA) in anhydrous
dimethylformamide (DMF) at 100ºC. They resulted in a high conversion grade (according to
TLC), however, their experimental yields dropped by 39% and 34%, respectively, after the
work-up. These compounds were also fully characterised (see Supplementary Material,
Fig.10.SM to Fig.17.SM).
Various authors have demonstrated the potential bioactive properties of dithiocarbamates.
With this in mind, a stable dithiocarbamate derivative was synthesised using an efficient
procedure outlined by Salvatore et al. [47]. The resultant dithiocarbamate was methyl (2-{[4-(1-
naphthylamino)-4-oxobutanoyl]amino}ethyl)dithiocarbamate (8), which was obtained from 4.
The desired product was purified using silica gel column chromatography and a white solid
was attained. The structure of compound 8 was confirmed by spectral data (see
Supplementary Material, Fig.20.SM to Fig.23.SM).
2
2
2
2
2
2
2
2
2
2
2
30
31
32
33
34
35
36
37
38
39
40
Molecular dynamics simulations of the complex between naphthyl derivatives, Aβ1-42
peptide and Aβ fibrils
Several in silico studies have sought to identify both, the binding affinity and type of
interaction between small molecules and Aβ peptide and fibrils, as well as to predict whether
these compounds may inhibit the formation of amyloid oligomers, fibrils and plaques [48,49].
As stated above, extensive evidence has accumulate to stablish that Aβ peptides trigger the
process of amyloidosis in AD, through a range of possible consecutive structural transitions
(dimers, oligomers, protofibrils, fibrils and, finally, plaques). Moreover, Aβ oligomers and fibrils
are considered the main cytotoxic species [21,22].
Molecular dynamics (MD) simulations are one of the most convenient and reliable methods
for obtaining a snapshot of macromolecular dynamic properties. This approach provides
11

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
detailed information on the fluctuations and conformational changes of biological systems,
such as protein-protein, ligand-peptide and ligand-fibril interactions [48-50]. This in silico
technique has been used to explore the amyloid assembly pathways, the conformational
ensembles of amyloidogenic peptides and fibril structures, and their thermodynamic stabilities
[51].
In this work, MD studies were carried out to evaluate the affinity and temporal stability of the
complexes: ligand-Aβ1-42 peptide (in its monomeric form) and ligand-Aβ fibril (as final products
in the Aβ amyloidogenic pathway) in a simulated biological environment. The docking studies
of the ligand-Aβ1-42-peptide complexes were performed to identify the most probable
interaction zone and the binding mode between them. The selected ligands of the 1-naphthyl
derivatives were the compounds: 4, 6, 7 and 8 and the Aβ1-42 structures chosen in the PDB
were 1IYT [29] and 5OQV [52].
Rigid docking simulations (see Supplementary Material, Fig.24.SM) revealed that around
80% of all ligands interact with Aβ1-42-peptide (Root-Mean Square Deviation, RMSD= 4.6 Å),
mainly with the amino acids E-11, V-12, H-13, Q-15, K-16, L-17, F-19 and F-20. This amino
acid region (the N-terminal) corresponds to those described in the formation of molecular
aggregates [53]. Essentially, the interactions took place through hydrogen bonding (H-bonds)
and hydrophobic interactions. H-bonds were observed between the O, N and S atoms of the
amidoalkylic chain of ligands with amino acids E-11, Q-15 and K-16 of Aβ1-42 peptide, which
probably increases the stability of the ligand-peptide complex formed. Hydrophobic
interactions, through the π-π cloud, were located between the naphthyl group of ligands and
the phenyl groups of the F-19 or F-20 amino acid residues. In general, these results agree
with prior predictions obtained by Sablon-Carrazana et al. [42], who used the fibril model of
Aβ17-42 peptide sequence. We subsequently developed a flexible docking study, wherein the
side chains of the amino acids region (which were found in the rigid docking for each complex)
12

2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
were left flexible during the evaluation (see Supplementary Material, Fig.25.SM). In 96% of the
simulations, all ligands linked in the same region via H-bonds, hydrophobic interactions and
van der Waals forces (cut off 4.6 Å). As in the rigid docking study, the same kind of π-π
stacking interactions were obtained. This could indicate that such interactions act as an
anchor point, allowing the rest of the molecule to acquire the most stable conformation. In this
simulation, H-bonds disappeared in the ligands-peptide complexes for 4 and 7, however, in
the case of ligands 6 and 8, there was an observable tendency to form H-bonds with Q-15 and
K-16 amino acids. This finding could give each of these complexes more stability than those
formed by 4 and 7 (see Supplementary Material, Fig.25.SM).
MD studies were performed considering the spatial resolution reached in flexible docking by
the ligand-Aβ1-42 peptide complexes. After 90 ns, the simulation concluded and 1.8 x103
structures were attained for each complex. According to the representative snapshot taken
between 0 and 45 ns, the spatial structure of the Aβ1-42 peptide in all complexes began to
“close in” on itself and originate interactions between the amino acids that could give rise to a
molecular closure structure (or zipper). This behaviour has been predicted with nuclear
magnetic resonance (NMR) techniques by Lührs et al. [25], using a fragment of Aβ17-42 peptide.
However, in all of these complexes, the ligands retain their tendency to interact with the Aβ1-42
peptide, thus avoiding its structural closure towards a β-conformation. Likewise, during the
simulation, all ligands interacted with the F-19 and F-20 amino acids of the Aβ1-42 peptide,
which could confirm that π-π hydrophobic interactions act as an “anchor” for the eventual
conformation. Thereby, the amidoalkylic chain of the ligands interacted with the rest of amino
acids (E-11 to F-20), in particular, with V-12, Q-15 and K-16, which remained steady during
the simulation (see Figure 2). In this evolution process, we estimated the temporal stabilities
by means of spatial positions (RMSD), as well as the attractive and repulsive forces (Coulomb
13

2
2
90
91
and Lennard-Jones energies) that were involved in the formation of the complexes in the
abovementioned region (see Supplementary Material, Fig.26.SM).
14

2
93
2
2
2
2
94
95
96
97
Figure 2. Representative snapshot of the conformational evolution of 4-Aβ1-42 peptide complex, 6-Aβ1-42 peptide
complex, 7-Aβ1-42 peptide complex and 8-Aβ1-42 peptide complex at 0, 45 and 90 ns, as observed during
molecular dynamics simulations, at pH 7.4 and 310K. The amino acids in black characters are those that
remained almost steady during simulation.
15

2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
98
99
00
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
In accordance with the conformational evolution of ligand-Aβ1-42 peptide complexes, the
influence of ligands binding to the secondary structure of the Aβ1-42 monomer appears to be
significant, and the peptide maintains an α-helix conformation due to the interaction between
the ligands and the different residues (fundamentally, 10 to 20). However, in the 4-Aβ1-42
peptide complex, initial H-bonds disappear in the N-terminal region after 40 ns, leaving only a
small amount of electrostatic and hydrophobic interactions in the C-terminal region. This could
explain why the ligand-peptide interaction breaks between 40 and 80 ns during the simulation
(see Supplementary Material, Fig.26.SM). On the contrary, the 6-Aβ1-42 peptide and 7-Aβ1-42
peptide systems seem to be more stable (Coulomb and LJ < 0), which can be explained by
the presence of almost one H-bond, as well as the van der Waals forces and hydrophobic
interactions. In the case of the 8-Aβ1-42 peptide complex, there was a notable enhancement in
stability. This was due to both an increase in the number of the H-bonds (almost three, with
the amino acids E11, Q15 and K16, starting at 40 ns; see Supplementary Material, Fig.26.SM)
and a near twofold increase in the number of contacts between the ligand and the amino acids
of the peptide (regions 11 to 24). This could be mean that the electrostatic and hydrophobic
interactions (π-π stacking with F19 and F20; see Supplementary Material, Fig.26.SM) were
more effective.
Taken together, these findings may suggest that ligands 6, 7 and 8 could reduce the
accumulation of Aβ fibrils by interfering with fibril elongation in the aggregation pathway.
To gain further insight into the mechanism of ligands-Aβ1-42 peptide complexes, we decided
to study the type of interactions taking place between the ligands and a fibrillar structure. To
do so, we used the S-shaped structure of Aβ1-42 fibril (PDB code: 5OQV). This structure was
obtained from cryo-electron microscopy and presents two filaments of five and four peptides
each, containing the full sequences of amino acids (42 a.a.) [52]. In MD simulation, we
considered only the pentamer of the two multimers, because this structure has an suitable
16

3
3
3
3
3
3
3
3
23
24
25
26
27
28
29
30
pharmacological site, which was studied by [54]. The MD strategy employed here was similar
to that used for the Aβ1-42 monomer, but the simulation time was 300 ns (see Figure 3) and the
stoichiometric ratio of the ligand:fibril was 1:1. The MD simulation of this pentamer´s fibril was
carried out for evaluating the fibril stability at timescale of 250 ns (see Supplementary Material,
Fig.27.SM). The 3D structure of pentamer remains steady stable during all simulation time
(RMSD ≤ 2.5 Å). Trajectory analysis of this structure revealed that the fibril is stable, without
changing its druggable site.
17

Complexes
Start simulation
Final simulaiton
4-Aβ1-42 fibril
6-Aβ1-42 fibril
7-Aβ1-42 fibril
8-Aβ1-42 fibril
3
3
3
31
32
33
Figure 3. Representative snapshot of the conformational evolution of 4-Aβ1-42 fibril complex, 6-Aβ1-42 fibril
complex, 7-Aβ1-42 fibril complex and 8-Aβ1-42 fibril complex at 0 and 300 ns, using molecular dynamics simulations,
at pH 7.4 and 310K.
3
34
3
35
Similarly, in the ligand-fibril complexes, the global shape of pentamer was stable after 300
ns of simulation, without any clear weakening in β secondary structure. The ligands 4, 6 and 7
interact with the convex surface of the β-fibril in the N-terminal region. Apparently, the internal
amino acid residues of U site seem to collapse closer to the pocket, possibly due to the
3
3
3
36
37
38
18

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
instability produced by hydrophobic interactions in the external surface between the amino
acid residues of the chain and the ligand in each case, with regard to the undocked
pentamer´s fibril. In sum, these findings and those obtained in the MD simulation of ligand-
peptide complexes could indicate that the ligands 4, 6 and 7 could prevent interactions
between Aβ-peptides and fibres and avoid the elongation process during the aggregation
pathway. In the case of ligand 8, the complex appear quite stable (RMSD ≤ 2 Å) during the
simulation time. Multiple interactions were established with the amino acids H14, K16, I17,
V18, F20 and I32; G33, L34, M35 in the N- and C-terminal regions, respectively, by means of
H-bonds, electrostatic and π-π stacking interactions (see Supplementary Material, Fig.28.SM).
These interactions produced an apparent opening of the groove of β-folder and seem to cause
a structural destabilisation of N- and C- terminal chains during the simulation whereby they
could modify this conformation. The hydrophobic interactions (including the π-π stacking) with
V, H, F residues and the H-bonds with the amino acids E, Q, K, involved in the binding pocket,
were also identified as dominant interactions in others fibril structures [41,54,55]. In line with
this notion, Hochdorffer and colleagues [56] have shown that these amino acids are involved
in the spatial U-shaped structure of the central cleft of the fibril, which suggests that ligand 8
could induce a change in this conformation.
The free binding energies of several compounds tested with MD or Monte Carlo simulations
were calculated using the linear interaction energy (LIE) method [51,57]. These theoretical
values are very close to experimental values, which has allowed the optimisation of resources
in the search for pharmacological targets [58]. The ΔG values for the ligand-Aβ1-42 peptide and
ligand-Aβ1-42 fibril complexes are presented in Table 2. It is notable that the 8-Aβ1-42 peptide
complex shows the highest binding free energy with the Aβ1-42 peptide and Aβ1-42 fibril.
3
3
3
62
63
64
Table 2. Binding free energies and predicted equilibrium constants of ligand-Aβ1-42 peptide
and ligand- Aβ1-42 fibril complexes, calculated using the LIE method based on MD simulations.
19

OH
O
NH2
NH
O
NH
O
O
O
NH
NH
Ligand 4
Ligand 6
HO
S
CH3
O
S
NH
NH
O
NH
O
O
O
NH
NH
Ligand 7
Ligand 8
3
65
Complexes
ΔG (kcal/mol)
-4.9*
Keq
Complexes
4-Aβ1-42 fibril
6-Aβ1-42 fibril
7-Aβ1-42 fibril
8-Aβ1-42 fibril
ΔG (kcal/mol)
-3.3**
Keq
4
6
7
8
-Aβ1-42 peptide
-Aβ1-42 peptide
-Aβ1-42 peptide
-Aβ1-42 peptide
2,85 x103
2.76 x104
7.54 x103
1.65 x105
2.12 x102
2.76 x104
1.75 x103
2.68 x105
-6.3*
-6.3**
-5.5*
-4.6**
-7.4*
-7.7**
3
3
66
67
*90 ns and **300ns of MD simulation times.
3
3
3
3
3
68
69
70
71
72
According to our MD results, the ligand 8 forms thermodynamically stable complexes
through stabilizing interactions with the Aβ1-42 peptide and with the structure of fibrils. This
allowed us to hypothesize that compound 8 could stabilize the conformation of the Aβ1-42
peptide, avoiding its aggregation and the evolution of the preformed fibrils to more complex
species.
373
374
375
376
377
378
379
380
In vitro inhibitory effect of naphthyl derivatives on Aβ1-42 peptide oligomerisation
Previous studies [42,43] have evaluated the effect of compounds 4, 6, 7 and 8 on BSA
protein and hIAPP1-37 peptide aggregation kinetics in vitro by means of kinetic assays, circular
dichroism and microscopy techniques (including transmission electron microscopy and atomic
force microscopy). These studies demonstrated that these compounds modify the
fibrillogenesis of these model proteins. However, compound 4 showed the lowest inhibitory
effect on hIAPP1-37 peptide aggregation. In addition, when the neuroprotective effects of these
compounds were assessed using cerebellar granular cells that had been exposed to the
20

381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
cytotoxic effects of hIAPP1-37 peptide, compound 4 showed the lowest neuroprotective effect.
The other compounds revealed favourable results in terms of inhibiting protein aggregation
and improving cell viability.
These previous results, together with our present in silico study, led us to select
compounds 6, 7 and 8 for the evaluation of inhibitory effects on Aβ1-42 peptide oligomerisation
in a cell-based model. Firstly, in order to identify a non-cytotoxic dose interval for these
compounds, we determined what effect they had on the cell viability of primary mixed
astroglial cultures from the cerebral cortex and hippocampus of mice, using MTT assay [59-
61]. The astroglial toxicity of compounds 6, 7 and 8 was very high (IC = 874, 433 and 195
50
µM, respectively), but did not reach physiologically relevant concentrations, making these
compounds safe. Compounds 6 and 7 in particular did not cause greater MTT reduction,
despite causing a slight attenuation at higher concentrations. Conversely, a 24-hour treatment
with 8 caused an increase in MTT reduction at low concentrations; this increase was then
attenuated at higher doses (see Supplementary Material, Fig.29.SM and Fig.30.SM). The
mechanistic significance of such biphasic effect is not yet fully understood. One possible
reason is that compound 8 may induce an increase in astrocyte proliferation rates at sub-toxic
concentrations, even though assays were made at confluent cultures and under serum-free
conditions, thus limiting the proliferative capacity of the cultures. Another possible explanation
is that compound 8 might alter cytosolic and mitochondrial dehydrogenases or other
NADH/NADPH-linked enzymes that are responsible for increases in MTT after sub-toxic
treatment. Further work is required to find correlations between these observations and glial-
specific functions.
Next, the Aβ1-42 anti-aggregation properties of compounds 6, 7 and 8 were evaluated in
human microglia, which is a relevant in vitro model for AD pathophysiology [62]. The assays
were carried out in human immortalised microglia cells with various concentrations of the
compounds (50, 1 and 0.1 μmol/L). Based on previous results showing the inhibitory activity of
21

407
408
409
410
galantamine hydrobromide (GH) on the aggregation of Aβ peptides, as well as on Aβ
clearance by microglial phagocytosis, this compound was included as a positive control of
inhibition [63,64] (see Figure 4).
4
11
4
4
4
4
4
12
13
14
15
16
Figure 4. Inhibitory effect of compounds 6, 7 and 8 on Aβ1-42 aggregation in human microglia. The compounds
were applied to a human microglia cell line at different concentrations (0,001-50 μmol/L) concomitantly with Aβ1-42
peptide. Galantamine hydrobromide (GH) was applied as a positive control for inhibition. After 48 h incubation,
Aβ1-42 aggregates were stained with thiazine red. Data points represent the mean ± SE percentage of staining
intensity in relation to the Aβ1-42 control, at each condition for a single experiment performed in triplicate.
4
4
4
4
4
4
4
4
4
4
4
4
4
17
18
19
20
21
22
23
24
25
26
27
28
As expected, mixed microglia cultures treated with Aβ1-42 peptide and GH showed
decreases in fluorescence intensity when compared to cells exposed only to Aβ1-42. This
indicates an inhibitory effect on Aβ1-42 peptide aggregation (see Figure 4). Compounds 6, 7
and 8 also inhibited Aβ1-42 aggregation in human microglia cultures, as revealed by the
thiazine red reagent (see Figure 4). Interestingly, compounds 6 and 7 showed no lineal dose
effect relationship, reaching a plateau in terms of inhibitory action; even at the lowest
concentration (1 nM), they were capable of inhibiting Aβ1-42 oligomerisation. On the other hand,
compound 8 inhibited Aβ1-42 aggregation in a dose-dependent fashion, with a maximum
inhibitory effect of 79% at a dose of 1 μM.
Ex vivo binding affinity of 8 to Aβ1-42 plaques in transgenic mice brain tissue
In order to evaluate the affinity of compound 8 for β-amyloid plaques, which were present in
11
29 the female 3xTg-AD transgenic mice, we carried out radiosynthesis of [ C]methyl (2-{[4-(1-
22

1
1
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
naphthylamino)-4-oxobutanoyl]amino}ethyl)dithiocarbamate ([ C]8) using an automated
synthesis platform. After a semi-preparative high-performance liquid chromatography and solid
11
phase extraction (SPE) (see Supplementary Material, Fig.31.SM) of [ C]8 was obtained with
a radiochemical purity of 94.2% ± 0.3 (n=3). In all cases, the difference between retention
11
times for [ C]8 and 8 was below 4%, and radionuclide identity and purity were confirmed by
measuring the two substances’ half-lives and gamma spectrum (see Supplementary Material,
11
Table 1.SM). The log Poct/PBS value of [ C]8 was 1.77 ± 0.07, determined by two independent
experiments (in triplicate). This value is optimal for a non-ionisable compound that can easily
cross the BBB, according to Dischino et al. (see Supplementary Material, Table 2.SM). Our
11
stability studies showed that the final formulation of [ C]8 was stable for 6 h at room
temperature and for 2 h after being added to fresh human plasma at 37ºC (see Supplementary
Material, Table 3.SM and Table 4.SM). Plasma protein binding (PPB) determination was
carried out using gel filtration. The product had a high PPB percentage, which remained
constant over time (see Supplementary Material, Table 5. SM).
11
The ex vivo affinity assessment of [ C]8 for Aβ plaques was performed on serial brain
sections taken from female 3xTg-AD mice aged 9 months by means of saturation and
competitive binding assays (see Supplementary Material, Fig.32.SM.), as described in [65].
The K value from Scatchard analyses was 15.8 nM (r = 0.85), which is very similar to the
d
18
affinity of [ F]Florbetaben (K = 16) that has been calculated in AD brain homogenates [66].
d
4
4
4
4
4
4
4
49
50
51
52
53
54
55
Beneficial effects of compound 8 on cognitive impairment and Aβ burden in 3xTg-AD
mice
There is sufficient evidence to support the hypothesis that soluble intermediate forms of Aβ
may drive early AD, rather than insoluble amyloid deposits [12-14,67]. Aβ aggregation and the
generation of prefibrillar intermediates hinder interactions between brain cells and cause
neurodegeneration, cognitive dysfunction and, ultimately, dementia [68-70]. This process
involves cellular and molecular mechanisms that have a negative impact on hippocampal-
23

4
4
4
4
4
4
4
4
4
4
4
56
57
58
59
60
61
62
63
64
65
66
dependent behavioral tasks. Aβ accumulates in AD-relevant brain regions, which are those
involved in the processes of cognition and memory (the temporoparietal and frontal cortices,
the cingulate gyrus and the hippocampus) [71,72].
Because the presence of Aβ deposits begins before the onset of AD symptoms, a
reasonable therapeutic goal for AD is to inhibit Aβ oligomerisation and the formation of larger
42
aggregates [31]. According to our in silico and in vitro results, discussed above, it is therefore
interesting to evaluate whether 8 is able to inhibit and/or disaggregate Aβ depositions in the
brains of aged transgenic animals. Moreover, the inhibitory activities of this naphthyl derivative
on Aβ aggregation must translate not only to in vivo decreases in Aβ burden, but also to
behavioural improvements in terms of learning and memory. To this end, the potential effect of
compound 8 on cognitive impairment in triple transgenic mice was studied.
467
468
469
470
471
472
473
474
475
476
477
478
479
480
Improvement in learning and spatial memory in 3xTg-AD mice
The experimental design used an oral administration of 8 to female 3xTg-AD transgenic
mice (t =10-12 months) for eight weeks at a daily dose of 1 mg/kg of body weight. The
0
experimental mice model over-expresses three AD-relevant human proteins (PS1M146V,
APPSWE and tauP301L), which produce the accumulation of Aβ with age. After 6 months, this
process generated fibrillar amyloid plaques [73] that have negative effects on memory [74];
this effect is most pronounced in female mice [75].
We assessed the cognitive and spatial memory functions of the transgenic mice that had
been treated with 8 (3xTg-AD-T) and compared the results to their age-matched transgenic
controls (3xTg-AD-nonT), using the Morris water maze (MWM), a classic and robust
behavioural test that evaluates memory and cognition [76,77]. When assessing the learning
performance of the animals over four days of training, the escape latency of the 3xTg-AD-T
mice tended to be lower on the second and third days, even though the differences did not
reach statistical significance (see Figure 5A). Probe analysis showed a non-significant
24

4
4
81
82
tendency for lower latencies to enter to the target quadrant, in comparison to the 3xTg-AD-
nonT control group (see Figure 5C).
4
83
4
4
4
4
4
4
84
85
86
87
88
89
Figure 5. Effect of compound 8 on learning and spatial memory of 3xTg-AD mice. The graphs show relevant
variables in the Morris water maze (MWM) test: (A) escape latency over four days of training and probe trials
regarding this on the fifth day; (B) number of entries to the target quadrant; (C) latency to enter to the target
quadrant; and (D) time spent in the target quadrant. Data are represented as mean ± standard error mean of
each group of age-matched transgenic mice treated with compound 8 for eight weeks (3xTg-AD-T) or with a
vehicle solution (3xTg-AD-nonT). n = 4, *p < 0.05, according to Mann Whitney U test.
490
491
492
493
494
495
496
497
498
499
500
We observed no clear trends or significant differences between groups in terms of time
spent in the target quadrant (see Figure 5D). The lack of statistical significance in results
related to the other cognitive parameters might be due to the low sample size per group. This
suggests that further studies are needed to confirm our findings.
Despite our experimental conditions, transgenic mice treated (3xTg-AD-T) with compound
8 entered the target quadrant (where the hidden platform was) significantly more frequently
than the control animals (3xTg-AD-nonT) (see Figure 5B). This implies that compound 8
improved the spatial strategy of mice, allowing them to search in a more targeted way for the
hidden platform, in turn suggesting better memory retrieval. Swim speed was checked during
the training and the probe trial; no differences were detected between groups.
25

501
502
503
504
505
506
507
508
Compound 8 decreases Aβ burden in the hippocampus of 3xTg-AD mice
According to immunohistochemical analysis (BAM-10 antibody), the subiculum area in the
brains of the older control transgenic mice that had been treated with vehicle (3xTg-AD-nonT;
tf = 12-14 months, female) showed a dense, positive region of Aβ plaques (see Figure 6A).
However, when transgenic mice were treated with compound 8 (3xTg-AD-T; from 10-12
months until 12-14 months of age), the subiculum showed a significant decrease in Aβ
plaques when compared to similar areas of the 3xTg-AD-nonT mice brains (see Figures 6B
and C).
5
09
5
5
5
5
5
5
10
11
12
13
14
15
Figure 6. Effect of compound 8 on Aβ burden in the brains of triple transgenic mice (3xTg-AD). Representative
immunohistochemistry photomicrographs of the subiculum and CA1 hippocampus, in sagittal sections, of both
non-treated transgenic mice (3xTg-AD-nonT) (A) and compound 8-treated transgenic animals (3xTg-AD-T) (B),
using a BAM-10 antibody at the symptomatic stage (12-14 months old). Graph C shows the average ± standard
error of the mean (SEM) of the area covered by Aβ in the subiculum as a relative percentage of the total area; n
= 4 animal per group, *p < 0.05 from Mann Whitney U test.
516
517
518
As previously stated, compound 8 is capable of interacting with Aβ1-42 peptides and fibrils,
primarily through hydrophobic interactions, van der Waals forces and H-bond formation. It can
26

5
5
5
5
5
5
5
5
19
20
21
22
23
24
25
26
therefore be stated that compound 8 not only inhibits the aggregation of the peptide but is
capable of inhibiting fibril elongation. The studies performed in silico, which are based on a
peptide and fibrillar structures of Aβ, suggest that 8 interacts with the amino acids involved in
the N-terminal region and with those that make up the groove of the folded β-structure (the C-
terminal), respectively. This interaction could also cause the structural destabilisation of
soluble fibrils. Oral treatment with compound 8 may inhibit the progressive amyloid depositions
that are evaluated with significant decreases in BAM-10 immunoreactivity (amyloid plaques) in
the subiculum, when compared to non-treated animals.
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
Conclusion
In silico evaluations of 1-naphthyl derivatives suggest that the selected compounds
interact with the Aβ1-42 peptide, in the region comprised between the residues E11 and F20,
which are considered as key in the formation of the molecular zipper. These interactions are
mainly π-π stacking, hydrophobic, van der Waals forces and by H-bond formation. The ligand-
Aβ1-42 complexes have high thermodynamic stability, similarly those attained with fibril and
ligand complexes. In the specific case of 8, this compound interacts with the amino acids
involved in the groove of the β-folder, causing a destabilization of this structure. All of these
interactions could inhibit the process of Aβ1-42 peptide oligomerization, disrupt fibril structure
and Aβ1-42 peptide-fibril interactions and thus slow down the aggregation process. This effect
was confirmed by in vitro assays, which showed that synthesized compounds 6, 7 and 8
interfere with Aβ1-42 aggregation, with a nanomolar-level potency in human microglia cell line
models. Compound 8 is also capable of interacting with Aβ fibrils and plaques with high affinity
(K = 16 nM), as shown by ex vivo assays using the brains of transgenic mice. Most
d
importantly, treatment with derivative 8 (log P = 1.77 ± 0.07) during AD progression in aged
3xTg-AD mice caused a decrease in Aβ burden in the animals’ cerebral subiculum region.
This could be down to compound 8’s ability to interact with Aβ peptide and fibrils in order to
disrupt their potential interaction and avoid the aggregation process, however, other
27

5
5
5
5
5
45
46
47
48
49
mechanisms should also be considered. In addition, treatment with compound 8 improves
search strategy for the MWM hidden platform, suggesting beneficial effects in terms of
cognition and memory. Our current and previous results highlight the potential of this
naphthalene derivative as promising lead compound and therapeutic candidate for AD and
support the need for future studies examining its mechanism of action.
5
5
50
51
Experimental Section
Chemical synthesis
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
General: All reagents were obtained from commercial suppliers and were pure for synthesis.
Reactions were monitored by thin layer chromatography (TLC) on pre-coated aluminum
sheets of silica gel 60 F254 (Merck), and the used mobile phases are: A: ethyl acetate; B:
ACN:NH :NH HCO (0.1 M) (7:1:1, v/v) and C: ACN:NH (7:1, v/v). Column chromatography
3
4
3
3
was performed on silica gel (35-70 mesh, Merck). The HPLC purity of 4, 6, 7 and 8 derivatives
was analyzed on a Shimadzu Prominence with a Chromolith® performance RP-18e (100 x 3
mm) or Phenomenex C18 (150 x 4.6 mm, 4 µm, 80Å) columns; mobile phases: H O/TFA
2
(0.1%) ACN:TFA (0.05%) or TEA/H O 0.1%: ACN (75:25). Flow rate: 1 mL/min, T = 37ºC,  =
2
254 nm (see Supplementary Material, Fig.6.SM, Fig.13.SM, Fig.17.SM and Fig.23.SM). IR
spectra were recorded on an ATI Mattson Genesis Series FTIR spectrophotometer (from 400
-1
1
13
to 4000 cm ). The 1D ( H, C, DEPT) NMR spectra were recorded on Bruker Model AC 250F,
1
1
JEOL ECLIPSE 400 and Varian Mercury-400 F spectrometers. In this last one, the 2D ( H, H-
1
COSY, HSQC, HMBC) NMR spectra were registered at 25ºC. H NMR multiplicities were
described using the following abbreviations: singlet (s), doublet (d), triplet (t), multiplet (m) and
broad signal (br. s). Hydrogen/deuterium exchange experiments were performed in order to
assign the NH signals. Melting points were determined on an Electrothermal (model 9100, UK)
apparatus. Mass spectra were recorded, in positive ion mode, on spectrometers Micromass
with orthogonal geometry QTOF-2 or Agilent Technologies model: LC/MSD TOF coupled to
HPLC 110, and processed with the program MassLinx v 3.5 (Micromass, UK).
28

5
71
Synthesis Procedures
5
5
5
5
5
5
72
73
74
75
76
77
1-(1-Naphthyl)-2, 5-pyrrolidinedione (1): Succinic anhydride (6.99 g, 69.92 mmol) and N-
methylmorpholine (NMM, 7.6 mL, 69.12 mmol) were added to a solution of 1-naphthylamine
(5 g, 34.96 mmol) in anhydrous 1, 4-dioxane (50 mL). The reaction mixture was refluxed for 5
h and then, the solvent was removed under vacuum. The solid residue was recrystallized from
º
º
ethanol to obtain 1, as white powder (4.1 g, 52%). R = 0.5 (A); m.p. 153.5-154.5 C (lit. 153 C)
f
[45].
5
5
5
5
5
78
79
80
81
82
4-(1-Naphthylamino)-4-oxobutanoic acid (2): 1-naphthylamine (10 g, 69.84 mmol) and
succinic anhydride (13.96 g, 140 mmol) were dissolved in anhydrous 1, 4-dioxane (100 mL).
The reaction mixture was refluxed for 2 h and later, cooled to precipitate a violet solid which
was filtered and washed with 1, 4-dioxane. The solid was recrystallized from ethanol to afford
º
º
a white solid (11.40 g, 67%). R = 0.32 (A); m.p. 167-169 C (lit. 169 C) [45].
f
5
5
5
5
5
5
5
5
5
5
5
5
5
83
84
85
86
87
88
89
90
91
92
93
94
95
4-(1-Naphthylamino)-4-oxobutanoic acid, N-hydroxysuccinimide ester (3): The
compound 2 (2 g, 8.23 mmol), N-hydroxysuccinimide (NHS, 1.42 g, 12.34 mmol) and N, N´-
dicyclohexylcarbodiimide (DCC, 2.54 g, 12.33 mmol) were dissolved in anhydrous 1,4-dioxane
(32 mL) and refluxed for 3 h. Then, the reaction mixture was filtered to separate the obtained
N, N´-dicyclohexylurea (DCU). The filtrate was concentrated to half volume under vacuum and
º
the mixture was cooled at 5 C for 5h. The white solid was filtered and washed with water and
º
diethyl ether (2.29 g, 82%). R = 0.5 (A); m.p. 168-171 C. IR (neat powder): ν = 3283, 2916,
f
-1 1
1729, 1666, 1562, 1208, 1104, 1083 cm . H-NMR (250 MHz, DMSO-d , 27ºC): δ = 9.73 (br.s,
6
1H, naphthyl-NHCO-), 7.86-7.83 (m, 1H, CH-8), 7.73-7.70 (m, 1H, CH-5), 7.68-7.66 (d, 3JH,H
=
8.04 Hz, CH-4), 7.51-7.48 (d, 3JH,H= 7.33 Hz, CH-2), 7.43-7.41 (m, 2H, CH-6 and CH-7), 7.30-
7.19 (t, 3JH,H= 7.81 Hz, CH-3), 2.91 (m, 4H, 2-CH CO), 2.30-2.23 (m, 4H, -NHCOCH and -
2
2
13
CH COO-). C-NMR (DMSO-d ): δ = 171.67 (3CO), 171.30 (NHCO), 133.73 (2C), 127.83 (C),
2
6
128.08 (CH-5), 126.00 (CH-6 or 7), 125.76 (CH-6 or 7), 125.57 (CH-3), 125.15 (CH-4), 122.90
29