Prenylated Curcumin Analogues as Multipotent Tools to Tackle Alzheimer's Disease

Article abstract of DOI:10.1021/acschemneuro.8b00463

Alzheimer's disease is likely to be caused by copathogenic factors including aggregation of Aβ peptides into oligomers and fibrils, neuroinflammation, and oxidative stress. To date, no effective treatments are available, and because of the multifactorial nature of the disease, it emerges the need to act on different and simultaneous fronts. Despite the multiple biological activities ascribed to curcumin as neuroprotector, its poor bioavailability and toxicity limit the success in clinical outcomes. To tackle Alzheimer's disease on these aspects, the curcumin template was suitably modified and a small set of analogues was attained. In particular, derivative 1 turned out to be less toxic than curcumin. As evidenced by capillary electrophoresis and transmission electron microscopy studies, 1 proved to inhibit the formation of large toxic Aβ oligomers, by shifting the equilibrium toward smaller nontoxic assemblies and to limit the formation of insoluble fibrils. These findings were supported by molecular docking and steered molecular dynamics simulations which confirmed the superior capacity of 1 to bind Aβ structures of different complexity. Remarkably, 1 also showed in vitro anti-inflammatory and antioxidant properties. In summary, the curcumin-based analogue 1 emerged as multipotent compound worthy to be further investigated and exploited in the Alzheimer's disease multitarget context.

Full text of DOI:10.1021/acschemneuro.8b00463

Subscriber access provided by YORK UNIV
Article
PRENYLATED CURCUMIN ANALOGUES AS MULTIPOTENT
TOOLS TO TACKLE ALZHEIMER'S DISEASE
Federica Bisceglia, Francesca Seghetti, Massimo Serra, Morena Zusso, Silvia Gervasoni, Laura
Verga, Giulio Vistoli, Cristina Lanni, Michele Catanzaro, Ersilia De Lorenzi, and Federica Belluti
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.8b00463 • Publication Date (Web): 17 Dec 2018
Downloaded from http://pubs.acs.org on December 18, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
PRENYLATED CURCUMIN ANALOGUES AS MULTIPOTENT TOOLS TO TACKLE ALZHEIMER'S
DISEASE
7
8
Federica Bisceglia a, Francesca Seghetti b, Massimo Serraa, Morena Zusso c, Silvia Gervasoni d, Laura Verga e,
9
Giulio Vistoli d, Cristina Lanni a, Michele Catanzaro a, Ersilia De Lorenzi *,a Federica Belluti b
,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a Department of Drug Sciences, University of Pavia, Viale Taramelli 12, 27100, Pavia, Italy.
b Department of Pharmacy and Biotechnology, Alma Mater Studiorum-University of Bologna, Via Belmeloro
6, 40126 Bologna, Italy.
c Department of Pharmaceutical and Pharmacological Sciences, University of Padua, Largo Meneghetti 2,
35131 Padua, Italy.
d Department of Pharmaceutical Sciences, University of Milan, Via Mangiagalli 25, 20133 Milan, Italy.
e
Department of Molecular Medicine, Unit of Pathology, University of Pavia IRCCS Policlinico S. Matteo
Foundation, Via Forlanini 14, 27100, Pavia, Italy.
1
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 2 of 41
1
2
3
4
ABSTRACT
5
6
7
8
Alzheimer’s disease is likely to be caused by copathogenic factors including aggregation of Aβ peptides into
oligomers and fibrils, neuroinflammation and oxidative stress. To date, no effective treatments are
available and because of the multifactorial nature of the disease, it emerges the need to act on different
and simultaneous fronts. Despite the multiple biological activities ascribed to curcumin as neuroprotector,
its poor bioavailability and toxicity limit the success in clinical outcomes. To tackle Alzheimer’s disease on
these aspects, the curcumin template was suitably modified and a small set of analogues was attained. In
particular, derivative 1 turned out to be less toxic than curcumin. As evidenced by capillary electrophoresis
and transmission electron microscopy studies, 1 proved to inhibit the formation of large toxic Aβ oligomers,
by shifting the equilibrium towards smaller non-toxic assemblies and to limit the formation of insoluble
fibrils. These findings were supported by molecular docking and steered molecular dynamics simulations
which confirmed the superior capacity of 1 to bind Aβ structures of different complexity. Remarkably, 1
also showed in vitro anti-inflammatory and anti-oxidant properties. In summary, the curcumin-based
analogue 1 emerged as multipotent compound worth to be further investigated and exploited in the
Alzheimer’s disease multi-target context.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Keywords: Alzheimer’s disease; amyloid beta oligomers and fibrils; capillary electrophoresis; curcumin
analogues; neuroinflammation; oxidative stress.
2
ACS Paragon Plus Environment

Page 3 of 41
ACS Chemical Neuroscience
1
2
3
4
INTRODUCTION
5
6
7
8
9
Alzheimer’s disease (AD) is the most common form of dementia 1 and despite impressive efforts, so far the
set up of a successful anti-AD drug discovery strategy resulted extremely difficult, mainly because of the
multifactorial nature of the disease 1, 2. In the last years, numerous AD-modifying therapeutics failed clinical
trials and, to date, only five drugs mainly targeting cholinesterases have been approved by Food and Drug
Administration. Compelling evidences consider extracellular amyloid-beta (Aβ) deposits in the brain as one
of the main AD hallmarks 1, 3, 4: they trigger microglia and astrocytes activation, which in turn results in
chronic inflammation and cellular oxidative stress 3, 5, 6. All these mechanisms, together with mithocondria
and neurovascular dysfunctions, are mutually involved in a feed-forward loop, ultimately leading to
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
progressive neurodegeneration 1, 6
.
Aβ protein includes natively disordered peptides, ranging from 36 to 43 amino acids, that are produced
4
from the sequential metabolism of amyloid precursor protein by β- and γ-secretases ; in particular, the
predominant species found in plaques is the most toxic and amylodogenic Aβ1-42 (Aβ42). Monomeric Aβ42
peptide is prone to self-assembly into oligomeric species, which aggregate to form protofibrils and then
7
mature amyloid fibrils 2, 4. Some well-defined Aβ peptide regions are involved in the aggregation process .
The glycosaminoglycan (GAGs) binding site (HHQK) assists, via a α-helix intermediate, the conformational
transition of monomers from normal random coil to β-sheet structure which is characterized by an
increased tendency to aggregate into dimers and oligomers 8. The stabilization of the soluble Aβ monomers
conformation, or the prevention of the α-to-β conformational transition, represent suitable strategies to
avoid oligomers formation. The self recognition hydrophobic core (16KLVFFA21), located in the central region
of Aβ peptide, constitutes a nucleation site that initiates the Aβ-Aβ interaction 9. In particular, - stacking
involving the two Phe residues may play a significant role in the self-assembly process. The emerging --
hypothesis suggests that drugs able to block these interactions may effectively control the amyloid diseases
10. After this primary lag step, spherical oligomers transform into larger aggregates (nucleation phase) and
an equilibrium between hydrophobic and hydrophilic interactions, involving Glu22-Gly29 residues, is thought
to affect this stage.
Notably, in the last 25 years attention has been shifted from amyloid fibrils to soluble prefibrillar Aβ
oligomers which, unlike other amyloidoses, are found to be more toxic than mature fibrils 11. Aβ42
oligomers bind to hippocampal neurons and hamper synaptic plasticity, manifested as inhibition of the
long-term potentiation, which is involved in learning and memory 4, 12. To date, it is not clear which
particular oligomeric state is mainly involved in eliciting neurotoxic effects and by which specific
mechanism. Indeed, the transient and non-covalent nature makes it difficult their identification and
characterization 2, 13, 14. On the other hand, also insoluble fibril deposits are found to be neurotoxic since,
being in equilibrium with oligomers, they may serve as a reservoir and catalyze oligomer formation through
3
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 4 of 41
1
2
3
4
5
secondary nucleation pathways. Consequently, both oligomers and fibrils represent important targets. In
this scenario, a multitarget approach based on the simultaneous inhibition of some relevant AD targets
6
7
8
9
offers promises to achieve a successful therapeutic outcome 15
Curcumin (Figure 1), the primary bioactive compound found in the rhizome of Curcuma longa, has been
reported to interfere with Aβ aggregation and to attenuate oxidative stress and neuroinflammation 16-18
.
.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The 4-hydroxy,3-methoxy-phenyl (vanillin) rings of curcumin were shown to adopt a correct position to
establish π-π interactions with Phe19 and Phe20 residues of Aβ 19-21, as a result of the appropriate length and
flexibility of the heptatrienone central linker 22, 23. In addition, the curcumin coplanar rearrangement allows
the intercalation between the Aβ fibrils, thus hampering the β-sheets assembly and in turn the fibrillation
pathways 20. Nevertheless curcumin does not have success in clinical trials, due to its poor bioavailability
and possible toxic effects, mainly ascribed to the 4-hydroxy,3-methoxy substitution pattern of the two aryl
rings 22, 24. Together with a large number of multipotent compounds such as quinones, cathecols, and
isothiazolones, curcumin has been included among pan-assay interfering compounds (PAINS), since they
can display apparent bioactivity and/or interfere with assay readouts 24, 25. It is not definite whether
compounds containing a PAINS-related sub-structure behave as reactive chemicals rather than selective
drugs, indeed the reactivity of a compound or other effects might be also context-dependent 26. However, a
number of curcumin-based compounds, when properly decorated, have been shown to exert significant
biological activities 21, 27, thus confirming the great potential of the curcumin scaffold for drug discovery. In
this context, the key chemical features for an Aβ ligand have been identified to comprise two sterically
bulky aromatic moieties, connected by a flexible linker of suitable length, to ensure their optimal spatial
arrangement 23. The resulting elongated and flat molecule demonstrated to hamper the characteristic -
stacking interactions between the aromatic residues of Aβ monomers, oligomers, or filaments 19, 20
.
15, 28
In order to identify multipotent AD disease-modifying drug candidates
with an improved biological
profile as compared to the reference compound curcumin, a small series of newly synthesized analogues
was here tested for anti-oligomeric and anti-fibrillogenic activity, while toxicity was evaluated on both
neuroblastoma and microglia cells. Molecular docking and steered molecular dynamics (SMD) simulations
were performed in order to assess the molecular recognition of the derivatives with either monomeric or
aggregated Aβ42 and to assess the stability of the corresponding complexes which are responsible for the
inhibitory activity. The identification of probes able to modulate Aβ aggregation, inflammation and
oxidative stress could allow to shed light into this cross-talk and represents a crucial issue in the AD drug
discovery scenario. Therefore, the most promising compounds in terms of anti-oligomeric and anti-
fibrillogenic activity were also evaluated for their anti-inflammatory and anti-oxidant properties.
4
ACS Paragon Plus Environment

Page 5 of 41
ACS Chemical Neuroscience
1
2
3
4
RESULTS AND DISCUSSION
5
6
7
8
9
Design and synthesis of curcumin analogues. The aromatic 4-hydroxy,3-methoxy substitution patterns
of curcumin (cur, Figure 1) were sequentially replaced with the para-3,3-dimethylallyloxy (4-prenyloxy), to
obtain 1 as asymmetrical compound, bearing both vanillin and 4-prenyloxyphenyl functions, and 2 (Figure
1), as symmetrical counterpart in which both side rings were decorated with the 4-prenyloxy group. This
moiety, found in a large number of natural products, proved to be often associated to neuroprotective
effects, such as anti-inflammatory and anti-oxidant, among others 29, 30. This flexible mono-unsaturated
function could also impart to the molecule an increased capability to establish hydrophobic contacts with
Aβ monomers. Furthermore, the importance of the hepta-1,4,6-trien-3-one central linker was studied by its
proper simplification into a penta-1,4-dien-3-one, obtaining the corresponding compounds 3 and 4 (Figure
1). Since this structural modification has been successfully introduced in the design of curcumin derivatives
with increased anti-proliferative and anti-inflammatory activities 31, we wondered if the same increased
activity would have been observed in AD multi-target context. Indeed, the maintenance of the carbonyl
α,β-unsaturated electrophile reactive centre, together with the length of the linker (6–19 Å) and its rigidity
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
are some of the key chemical features required to effectively inhibit amyloid aggregation 23
.
Figure 1. Design strategy toward the curcumin-based analogues 1-4. Structures of curcumin (cur), the newly
synthesized curcuminoids (1, 2) and their monocarbonyl analogues (3, 4).
Chemistry. The synthetic route for the attainment of the target curcumin-based compounds (1-4) is
outlined in Schemes 1 and 2. Compounds 1 and 2 were prepared by two-consecutive steps, applying the
28,
Pabon reaction
32. Analogues 3 and 4 were synthesized through base-catalyzed Claisen-Schmidt
procedure. In the classical Pabon procedure pentane-2,4-dione was complexed with B2O3 in EtOAc to avoid
the methylene-centred reactivity toward the Knoevenagel reaction and to favor the nucleophilic attack at
the side methyl groups. The boric complex was then condensed with the suitable aldehyde and then a step-
5
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 6 of 41
1
2
3
4
5
6
7
8
9
wise addition of n-tributylborate and n-butylamine was carried out. Acidic treatment caused the
dissociation of the complex, providing the desired curcumin-based analogue. This reaction has been
reported to give the curcuminoid as β-keto-enol tautomer. In the classical Claisen-Schmidt aldol
condensation, acetone was reacted with the appropriate aldehyde in EtOH and in the presence of a 50%
KOH/H2O solution. In details, by condensing pentane-2,4-dione or acetone with two different selected
aldehydes, namely 4-hydroxy-3-methoxybenzaldehyde (vanillin) and 4-(3,3-dimethylallyloxy)benzaldehyde,
a mixture of the desired compounds could be obtained, among which the asymmetrical and the two
symmetrical curcuminoids, including the semi-reaction products. Consequently, obtaining the desired
compounds in a good purity grade would require several chromatographic purifications. To avoid this
drawback, the synthesis of asymmetric curcumin analogues 1 and 3 (Scheme 1) was performed through
two sequential steps, in which the semi-synthetic intermediates were first prepared reacting pentane-2,4-
dione and acetone with vanillin under the appropriate reaction condition to give 5 and 6, respectively.
Subsequently, a second reaction with 4-(3,3-dimethylallyloxy)benzaldehyde, allowed the attainment of the
desired asymmetrical compounds. In order to prepare the symmetrical analogues 2 and 4 (Scheme 2) a
one-pot procedure, by using a 1:2 stoichiometric ratio for the ketone and the benzaldehyde, was
performed. The 4-(3,3-dimethylallyloxy)benzaldehyde (7) was obtained by exploiting a Williamson ether
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
synthesis between 4-hydroxybenzaldehyde and 3,3-dimethylallybromide in the presence of K2CO3 as base.
O
H
O
OH
O
OH
O
7
H3CO
HO
H3CO
HO
5
i, ii, iii
1
O
O
O
O
i, ii, iii
B2O3 complex
H
O
O
O
H
O
7
H3CO
HO
H3CO
HO
H3CO
HO
O
iv
3
iv
O
6
Scheme 1. Reagents and conditions: i) B(n-BuO)3; ii) n-BuNH2, 80 °C; iii) HCl, 80 °C; iv) KOH 50%. EtOH, rt.
6
ACS Paragon Plus Environment

Page 7 of 41
ACS Chemical Neuroscience
1
2
3
OH
O
4
O
O
i, ii, iii
5
6
7
8
9
O
2
O
B2O3 complex
O
O
O
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
iv
O
4
7
Scheme 2. Reagents and conditions: i) B(n-BuO)3; ii) n-BuNH2, 80 °C; iii) HCl, 80 °C; iv) KOH 50%. EtOH, rt.
Cell toxicity of curcumin and analogues. In view of evaluating if the Aβ42 oligomers-induced toxicity is
ameliorated in presence of the newly synthesized analogues (1-4), the intrinsic toxicity was assessed on
neuroblastoma (SH-SY5Y) and microglia cells by using cur as reference compound. Cells were exposed to
compound concentrations ranging from 1 to 50 µM for 24 hours or from 1 to 40 µM for 16 hours,
respectively. As shown in Figure S1a), high concentrations (20-50 µM) of all tested compounds induced a
significant loss of neuroblastoma cell viability in comparison to cells exposed to culture medium only. The
di-prenoxylated derivatives 2 and 4 showed a significant and comparable profile, as toxic effects were
observed at 5 µM concentration. At high concentrations (20-50 µM), 3 considerably reduced cell viability,
conversely, at concentrations ranging from 1-10 µM it was found to be not toxic, alike cur and 1.
Compound 2 exerted a similar effect on microglia cell viability, showing a cytotoxic activity from 1 µM
concentration. Interestingly, treatment of microglia with 1, even at 20 and 40 µM concentrations, did not
significantly affect cell viability in comparison with vehicle-treated control cells, whereas the same
concentrations of cur were cytotoxic (Figure S1b)), confirming previously published results 33
.
Inhibition of Aβ42 aggregation. On the basis of the amyloid hypothesis, a variety of anti-amyloid agents
have been developed over the years, with the aim of interfering either with Aβ peptides oligomerization or
fibrillogenesis or with both events 2, 14. Despite the efforts made, results are not yet encouraging. The
assays employed in in vitro aggregation studies show limitations in providing simultaneous and comparable
information on both fibrils and oligomers, as well as among oligomers of different size. As a consequence, it
is clear that more than one technique has to be used, since a potential inhibitor of the oligomerization may
not be identified by using a specific assay for fibrils 14
.
The evidence of fibril formation or inhibition is commonly given by well established techniques, such as
transmission electron microscopy (TEM) and thioflavin T (ThT) fluorescence. Conversely, the identification
and characterization of soluble Aβ oligomers involved in the pathological cascade are extremely
challenging; indeed to date, a specific oligomeric state to target is missing 2, 14. In particular, the difficulties
7
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 8 of 41
1
2
3
4
5
in obtaining well characterized Aβ oligomers in solution, as well as the dynamic equilibrium among
oligomers which keep on aggregating during the in vitro experiments, represent important issues in the
6
7
search for modulators of fibrillogenesis 34
.
8
9
In this context, it has been demonstrated that capillary electrophoresis (CE) can play an alternative role in
the in vitro aggregation studies of Aβ peptides. Since it works in free solution and in the absence of
secondary equilibria, CE allows a real snapshot of Aβ while assembling into soluble oligomers inside the
capillary, until sample precipitation 35. The use of CE to identify agents able to interfere with Aβ42
oligomerization has been pioneered by us and also reported by others 36-38. Furthermore, the combination
of CE analysis of Aβ42 solutions in the presence of an anti-aggregating candidate with the TEM images of the
final precipitate, provides integrated information on the anti-oligomerization and anti-fibrillogenesis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
activity 34-37
.
Anti-oligomeric activity. In this work, we exploited one of the analytical platforms recently reported by us
that were optimized and standardized to correlate aggregation state, structure and toxicity of Aβ42
oligomers separated by CE, in view of coincubation experiments 34
.
Before carrying out any analysis of an Aβ42 sample in the presence of an inhibitor of aggregation, the setup
of a reproducible protocol for Aβ42 sample preparation and in turn of a reproducible CE separation of the
prepared oligomers are mandatory prerequisites. Aβ42 peptide aggregates and precipitates very rapidly,
therefore to appreciate differences in oligomer formation kinetics or to detect reduction/abrogation of
oligomer building-up by a coincubated compound, it is necessary to keep the assemblies in solution for a
suitable time window. Moreover, the toxicity of the oligomeric populations detected and separated by CE
has to be known, for identifying the target species of the small molecules evaluated.
In particular, for observing the effect of coincubated molecules, the least aggregating sample preparation
protocol was selected 39, as it allows a wider time window if compared with other methods 36-38. Since cur
and compounds 1-4 are soluble in pure ethanol, an adjustment of the original Aβ42 sample preparation
protocol was necessary (see Experimental Section), to account for the percentage of ethanol in buffer when
coincubating the peptide and the compounds. The comparison of the CE profile and mobility values
34
reported in with those of Aβ42 control (i.e. in the absence of compounds) shown in Figure 2a) and Table
S1 (Supplementary material), demonstrate that the kinetics of formation of oligomers and all
electrophoretic parameters are not affected by a concentration of ethanol in buffer as high as 3.26%.
Therefore, precise peak identification, molecular weight range and toxicity of the separated oligomers can
be considered the same. Briefly, Figure 2a) shows the analysis of Aβ42 peptide injected in the CE system
after different elapsed times from sample solubilization (t0): monomers and dimers (peaks 1 and 2)
progressively convert over time into neurotoxic aggregates larger than dodecamers (peak 3) 34. At late
times from solubilization, large aggregates are the most abundant species, until insoluble fibrils are formed
8
ACS Paragon Plus Environment

Page 9 of 41
ACS Chemical Neuroscience
1
2
3
4
5
and no more soluble species are detected. As small oligomers are not toxic 34, a compound is considered
active if it interferes with the formation of the assemblies migrating under peak 3.
6
7
8
9
All coincubation studies were carried out in triplicate and electrophoretic traces were recorded for cur and
all analogues, at different concentrations and at different elapsed times from sample redissolution.
Figure 2b) shows the effect on the normalized area percentage of peak 3 in the presence of a given
compound at the highest concentration tested (50 µM), at a time point where, for the control sample, the
formation of toxic species is maximized (first bar). All compounds show some activity, although it is clear
that the asymmetrical hepta-1,4,6-trien-3-one-based analogue 1 turns out to be the most effective of the
series in suppressing the building up of toxic oligomers, followed by cur and then 2, while the 1,4-dien-3-
one analogues 3 and 4 are remarkably less active. Complete information on many time points from
solubilization up to precipitation is reported in Figure S2, to appreciate how 1, even at later times, is
evidently more potent than cur.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Based on these results, experiments have been performed at 10 µM compound concentration, focusing the
attention on cur, 1 and 2, both endowed with the curcuminoid scaffold. Figure S3 (as compared to Figure 2
b)) shows how the activity is kept, in the same order, albeit with decreasing potency. Electropherograms in
Figures 2 c) d) and e) easily visualize how derivative 1 (red trace) is best in suppressing the formation of
large toxic oligomers (peak 3). The lower anti-aggregation activity of compound 2 (blue trace) could be
partially ascribed at the lack of aromatic substituents capable of taking part in hydrogen bonding, which,
conversely, are highly conserved in most amyloid ligands of natural origin 23
.
9
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 10 of 41
1
2
a)
60
50
40
30
20
10
3
4
5
6
7
8
9
38 days
b)
3
90
21 days
77.76%
80
70
60
50
40
30
20
2
1
1
2
21 days
47.31%
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
3
1
27.14%
24.48%
10 days
2
11.18%
10
0
3
t0
0%
0
Aß42
control
cur
1
2
3
4
4
6
8
10
12
Time (min)
1
2
2
1
d)
c)
e)
1
10 days
t 0
21 days
2
2
1
40
30
20
10
40
40
30
20
10
0
1
2
3
1
1
2
3
1
30
20
10
1
3
2
3
3
2
3
1
2
3
3
3
2
2
3
3
1
2
1
3
0
0
4
6
4
Time (min)
6
4
Time (min)
6
Time (min)
Figure 2. Oligomerization of 100 µM Aβ42 in the absence or in the presence of cur and of compounds 1-4. a) Electrophoretic profiles of Aβ42 control (solubilized in 3.26% EtOH in
20 mM phosphate buffer pH 7.4, see Experimental section) at selected elapsed times from t0 until precipitation: monitoring over time by CE. b) Anti-aggregation activity of 50
µM cur and compounds 1-4: normalized area % of peak 3 at 21 days from sample redissolution as observed in control peptide and in the presence of compounds. Data are
expressed as mean ± standard deviation, for n=3. c-e) Comparison between electrophoretic profiles of 100 µM Aβ42 control (black trace) and in the presence of 10 µM cur
(green traces), 1 (red traces) and 2 (blue traces) at different elapsed times from redissolution. Electropherograms are representative of n=3.
10
ACS Paragon Plus Environment

Page 11 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Complete evaluation at decreasing concentrations down to 1 µM is reported in Figure 3. Here the
observations on the reduction of the toxic large oligomers (peak 3) can be summarized as follows: i)
independently on the concentration tested, the symmetrical analogue 2 only slows down the formation of
toxic oligomers, that are eventually formed as abundant as they are in the control sample; ii) both cur and 1
show concentration-dependent activity down to 1 µM (see also Figure S3); iii) at the lowest concentration
and in comparison with cur, not only does 1 show a more consistent reduction (nearly double) of the toxic
oligomers at the end point of the analyses, but it also keeps their abundance constantly low over time.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a)
cur
100
90
80
70
60
Aß42 control
+ 25 µM cur
50
+ 10 µM cur
+ 1 µM cur
40
30
20
10
0
0
5
10
15
20
25
30
35
40
45
Time (days)
b)
1
100
90
80
70
60
50
40
30
20
10
0
Aß42 control
+ 25 µM 1
+ 10 µM 1
+ 1 µM 1
0
5
10
15
20
25
30
35
40
45
Time (days)
c)
2
100
90
80
70
60
50
40
30
20
10
0
Aß42 control
+ 25 µM 2
+ 10 µM 2
0
5
10
15
20
25
30
35
40
45
Time (days)
11
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 12 of 41
1
2
3
4
5
6
Figure 3. Anti-oligomeric effect of compounds on toxic large oligomers (peak 3) over time. Normalized area
percentage plot of peak 3 of 100 µM Aβ42 control and in the presence of decreasing concentrations of: a) cur, b) 1, c)
2. Data are expressed as mean± standard deviation, for n=3.
7
8
9
Figure 2e) also provides information on the effect that coincubation of compounds exerts on the
persistence in solution of non toxic monomers and dimers (earlier migrating peaks 1 and 2). In particular, it
is clear that dimers (peak 2) are always kept in solution for very long times, in comparison with Aβ42 control
sample. This effect is consistent with the stabilization of Aβ dimers induced by curcumin 20. Notably, from
Figure S4, where the trend of the normalized area of dimers over time is reported, it can be derived that
this effect is markedly induced by cur and by its mono 4-prenyloxy-aryl substituted 1, in a concentration-
dependent manner.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
It can be summarized that CE data are consistent with the activity expected by virtue of the progressive
structural modifications performed on 1. In particular, the further substitution of the 4-hydroxy,3-methoxy
group with the 4-prenyloxy (as for 2) and the simplification of the β-keto enol function (as for 3) result in a
gradual decrease of the anti-oligomeric activity, that is lowest for 4, in which both modifications are
combined. The order of potency observed is 1>cur>2>3>4.
Anti-fibrillogenic activity. Because of the intrinsic fluorescence of curcumin, ThT fluorescence assay is
precluded 40, therefore TEM has been here employed to appreciate the effect of cur and 1-4 on fibril
formation, fibril morphology and fibril density. Conclusions are drawn on data obtained for three
independent samples each one measured in duplicate.
As shown in Figure 4, Aβ42 peptide alone contains fibrils already at early stages from solubilization (24
hours), although, as explained in 34, these are not injected in the capillary of the CE system. Conversely to
the control peptide and already after 24 hours, Aβ42 samples coincubated with cur or 1 contain amorphous
aggregates only. The absence of amyloid fibrils is kept at later stages of oligomerization (after 21 days). This
evidence can be correlated with the anti-oligomeric activity induced by both compounds relative to large
oligomers (Figure S2, 21 days) and consistently with the persistence in solution of dimers (Figure 2 e)).
The sample containing cur eventually precipitates into dense amyloid fibrils. This is not overly surprising, as
also recent reports suggested that curcumin only inhibits oligomerization while promoting fibrillization 41, 42
;
despite a growing evidence on the anti-amyloid effect of curcumin, the exact mechanism of action is still
debated.
Contrary to samples coincubated with cur, only sporadic fibrils are detected in precipitated samples of Aβ42
in the presence of derivative 1. Surprisingly, compound 2 abrogates the formation of the fibrils observed
after 24 hours, as only amorphous granular assemblies are detected at sample precipitation. In accordance
with the poor anti-oligomeric activity, both monocarbonyl derivatives (3 and 4) lead to amyloid fibrils
already after 24 hours and as end products of the aggregation (Figure 4).
12
ACS Paragon Plus Environment

Page 13 of 41
ACS Chemical Neuroscience
1
2
3
Aβ42 control
cur
1
2
3
4
4
5
6
24 hours
7
8
9
10
11
12
13
14
15
16
21 days
sample
precipitation
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. TEM images of Aβ42 control peptide (100 µM) and 100 µM Aβ42 coincubated with 50 µM cur and 1-4 analyzed
at different times from redissolution. Scale bar=100nm. 60000x. n=3.
Modeling studies. Docking simulations were carried out to corroborate the effects of the curcumin-based
analogues on Aβ42 aggregation process and to elucidate by which specific interactions with Aβ residues the
tested compounds are able to modify it. Even though the Aβ42 peptide shows a marked flexibility and can
assume almost all secondary motifs, docking simulations were focused on the peptide in its alpha-helix
form because this is the only available resolved structure.
Figures 5 a) b) and c) show the best complexes obtained for cur, 1 and 2 with the monomeric Aβ42 peptide.
As expected, the ligands approach the amyloid region corresponding to the above mentioned 16KLVFFA21
sequence which is well-known to be involved in ligand recognition. In the computed complexes, the
curcumin analogues assume slightly bent conformations which are in line with the curcumin geometries
found in the resolved structures. For instance, the curcumin molecule included in the resolved complex
with human transthyretin (PDB Id: 4PMF) shows an angle, as defined by the central carbonyl and the two
aromatic rings, of 107.87 while, in the here generated complex, the same angle of cur is equal to 106.99
(see Figure 5a). Moreover and in all considered complexes, the β-keto-enol central core elicits H-bonds with
Glu22 and Ser26, while a ligand phenyl ring stabilizes a clear π-π stacking contact with Phe19. The other
interactions stabilized by the phenyl moieties depend on the polarity of the substituents. Thus, cur, through
a side vanillin group, elicits H-bonds with Gln15 and Asn27, while asymmetrical analogue 1 retains the same
polar contacts and, through its prenyloxy function, adds clear hydrophobic interactions with Ile31. The di-
prenoxylated 2 loses the H-bond with Gln15 and adds apolar contacts with Val18. Both monocarbonyl
analogues (3 and 4) lose the key polar interactions with Glu22 (complexes not shown). As depicted in Figure
S5, the complexes as computed by considering more extended amyloid structures show comparable ligand
poses even though the involvement of more amyloid segments increases the number of interacting
residues especially concerning the hydrophobic residues.
13
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 14 of 41
1
2
3
4
5
6
7
8
9
While contacting the same well-known amyloid region, Figure 5 d) e) and f) show that the simulated ligands
assume different poses when interacting with the fibril structures. In particular, the ligands assume rather
extended geometries by which they contact six different amyloid segments. As seen for 1, the ligands
contact the same residues of all involved segments: in detail, they elicit i) extended hydrophobic contacts
with six Leu18 residues which contact both phenyl ring and alkyl chains and ii) extended H-bonds with six
Lys16 residues, which approach both the β-keto-enol central core and the substituents on phenyl rings. Not
to mention that these protonated residues can also stabilize charge-transfer interactions with the phenyl
rings. Taken together, one may observe that, while the ligands assume comparable poses in all simulated
amyloid structures, the relevance of hydrophobic contacts increases when increasing the complexity of the
simulated amyloid structure and finds its maximum expression with the fibril structure where the ligands
are supported by a basis completely composed of apolar residues. These observations suggest that the
interaction with simple amyloid structures are enhanced by a fine balance between polar and apolar
contacts as seen for 1 in the monomeric protein, while the interaction with complex amyloid structures
appears to be strongly governed by hydrophobic contacts.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The described molecular docking results find encouraging validations by SMD simulations involving Aβ42
tetramers and fibrils and comparing 1 and 2 with cur. In these simulations the ligand is forced to unbound
from the Aβ42 structures and the required pulling force is monitored during the time. As compiled in Table
S2, each simulated ligand is thus characterized by the time required to unbound the ligand (i.e. until the
monitored force is 0 pN) as well as by the force mean required for unbinding. A third relevant parameter is
the AUC value of the plot limited to the residence time which accounts for both the total force and required
time. With regards to Aβ42 tetramers, Table S2 shows that the most polar cur compound requires an initial
marked force to break its key polar interactions (as reflected in the highest force mean) but then is unable
to prolong its binding and indeed it shows a residence time and AUC values clearly lower than 1 and 2.
These results confirm that only a proper balance between polar and hydrophobic interactions is able to
increase the stability of the complex with Aβ42 oligomers over the time. In contrast and when simulating the
fibrils, the ligands reveal rather similar values in all reported parameters. This observation may indicate that
the ligands are here similarly governed by the sole hydrophobic contacts and indeed the most apolar
compound (2) shows the best force mean and AUC value.
All proposed compounds were submitted to SwissADME online web-server (http://www.swissadme.ch) and
resulted completely devoid of PAINS alerts with a predicted pharmacokinetic profile roughly comparable
with that of the parent compound (results not shown).
14
ACS Paragon Plus Environment

Page 15 of 41
ACS Chemical Neuroscience
1
2
3
4
a)
b)
c)
5
6
7
8
9
F19
V18
F19
V18
K16
F19
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
K16
K16
E22
E22
F20
E22
N27
N27
N27
d)
e)
f)
ALA21
GLU22
VAL18
VAL18
LEU17
LYS16
LEU17
LYS16
VAL18
LYS16
LEU17
Figure 5. Main interactions stabilizing the putative complexes with monomeric Aβ42 for a) cur, b) 1 and c) 2, and with
the amyloid fibril structure for d) cur, e) 1 and f) 2.
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Evaluation of anti-inflammatory properties. The prenylation of aromatic natural products results in
derivatives with an improved pharmacological profile when compared with not-prenylated compounds and
prenylated natural compounds proved to be useful for the treatment of cancer and inflammation 29. As this
3,
latter represents an important pathological condition in AD 4, the functionalization of cur by the
introduction of a prenyl function could enhance the anti-inflammatory potential.
Thus, the most promising compound 1 in terms of anti-oligomeric and anti-fibrillogenic activity, was
selected to be tested for its anti-inflammatory properties, in comparison to cur. Lipopolysaccharide (LPS) is
the major component of the outer membrane of gram-negative bacteria and a potent immune activator of
a variety of mammalian cell types, including microglia. Inhibition of LPS-induced release of the pro-
inflammatory cytokines IL-1β and TNF-α by microglia was examined to assess the anti-inflammatory effects
of analogue 1 and cur. Microglia cells were exposed to non-cytotoxic concentrations of test compounds for
1 hour and then stimulated with LPS for 16 hours to induce an inflammatory response. In unstimulated cells
low or undetectable amounts of IL-1β and TNF-α were released and these basal levels remained unchanged
after treatment with the tested compounds (white bars, Figure 6). As expected, LPS stimulation induced an
increased release of IL-1β and TNF-α (taken as 100%). Not only was this effect significantly suppressed by
15
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 16 of 41
1
2
3
4
5
cur, as we previously showed 33, 43, but also by its prenylated analogue 1, starting from the concentration of
5 μM.
6
7
8
9
a)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
+ LPS
+ LPS
cur
1
b)
+ LPS
+ LPS
cur
1
Figure 6. Effects of cur and analogue 1 on cytokine release from LPS-stimulated cortical microglia cells. a) IL-1β and b)
TNF-α release. Data are mean ± SEM (standard error of mean) (n=3). * p<0.05, ** p<0.01 and *** p<0.001 vs LPS-
stimulated cultures, Kruskal-Wallis followed by post-hoc Dunn's test.
Evaluation of anti-oxidant properties. In addition to the anti-inflammatory activity, we then
investigated the potential of compounds 1 and 2 as antioxidants. We evaluated in SH-SY5Y neuroblastoma
cells their scavenger ability when co-incubated with H2O2, using cur as reference. In comparison to
untreated cells, (grey line, Figure 7a)), the intracellular DCFH-fluorescence intensity in H2O2-treated cells
significantly increased (black line, Figure 7a)). Treatment with cur and compounds 1 and 2 reduced H2O2-
induced intracellular ROS production, albeit to a different extent. At any time tested, cur showed the
strongest scavenger activity when compared to analogues 1 and 2. In particular, the presence of both
vanillin moieties (as for cur) appears to be important for this anti-oxidant activity. Compound 2, lacking 4-
hydroxy,3-methoxy group on both aromatic rings, was less effective than compound 1, where one of the
two vanillin functions of curcumin is preserved.
16
ACS Paragon Plus Environment

Page 17 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Further studies were carried out on the mechanism of action of 1 and 2 on the Nrf2 cellular pathway, for
investigating their ability to modulate the expression of the Nrf2 transcription factor 44. Recent evidence
further showed that Nrf2 activation suppresses inflammation through redox control 45. Therefore, we
investigated the activation of Nrf2 pathway by analyzing its translocation into nucleus. A concentration of 5
µM of cur and 1 induced Nrf2 nuclear translocation, whereas 2 did not show to be an Nrf2-inducer. The
asymmetrical analogue 1, at the concentration of 10 μM, did not produce statistically significant results in
our experimental setting, although we could assume an increasing trend as compared to control (Figure 7
b)).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
These data suggest that 1 may modulate anti-oxidant response by acting through different pathways,
besides Nrf2 activation. Evidence from literature supports the ability of different natural products, including
polyphenols in general, to interact with selected miRNAs, thus targeting multiple genes and showing
pleiotropic activity 46. Hence, the differential Nrf2 activation by 10 μM 1 may be putatively related to the
modulation of the levels of certain miRNAs associated with the Nrf2 signaling pathway. On the basis of
these observations, the investigation of miRNA modulation could potentially be important and will be next
further evaluated to better understand the underlying anti-oxidant mechanism of natural products and
their derivatives.
a)
b)
17
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 18 of 41
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Anti-oxidant activity of cur and analogues 1 and 2. a) Reduction of H2O2-induced intracellular ROS
production: fluorescence intensity for all compounds is significant at any time from 30 to 270 min with p<0.001 vs
H2O2, with exception of compound 2 at the concentration of 5 M. For compound 2 the significance is as follows:
p<0.05 at 90 min; p<0.01 at 120 min and 270 min; p<0.001 at 150, 180, 210 and 240 min. Dunnett’s multiple
comparison test; b) Activation of Nrf2 pathway: SH-SY5H cells were treated for 3 hours with compounds. Results are
shown as ratio Nrf2/lamin ± SEM; *p < 0.05, **p < 0.01, ****p<0.0001 versus control; Dunnett’s multiple comparison
test.
CONCLUSIONS
The curcumin scaffold was rationally modified and a small set of structurally related compounds was
obtained and tested against Aβ42 oligomer and fibril formation, inflammation and oxidative stress, i.e.
copathogenic factors related to AD. To this aim, derivatives were obtained in which i) a prenyloxy function
was introduced as substituent in one or both the cur side aryl functions, and ii) the cur keto-enol central
linker was simplified. Two couples of analogues 1-2, and 3-4, respectively were obtained.
The new compounds were firstly evaluated for their anti-amyloid effects, leading to identify cur and
analogue 1, both endowed with the hepta-trien-3-one linker, as the most effective agents in slowing down
the formation of toxic Aβ oligomers. Interestingly, CE and TEM analyses showed how 1 potently inhibits
oligomers building up and, contrary to cur, it does not lead to the formation of a dense network of fibrils.
Due to its remarkably lower toxicity on microglia cells if compared to 2 and even to cur, analogue 1 was
then selected to investigate its anti-inflammatory effect on microglia activation. In particular, at 5 µM
concentration, a suppression of pro-inflammatory cytokines release was detected. This analogue also
showed an anti-oxidant potential, with particular reference to the capability to induce, at 5 µM, Nrf2
nuclear translocation. By contrast, the symmetrical curcuminoid 2, in comparison to 1, showed a
remarkably weaker anti-oxidant activity and a lower efficacy in contrasting oligomer formation. Notably,
fibrils were not observed in presence of compound 2 and this suggests that the two prenyloxy substituents
may serve as β-sheet breakers by intercalating in the amyloid fibrillar core. The anti-amyloid effects of 1
and 2, compared to those of cur, were evidently corroborated by molecular docking and SMD simulations
18
ACS Paragon Plus Environment

Page 19 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
7
which revealed the growing role of the hydrophobic interactions when simulating Aβ structures of
increasing complexity. The simplified counterparts 3 and 4 failed to block both the formation of large Aβ
assemblies and the deposition of fibrils.
8
9
Taken together, these data underlined the pivotal role of the curcumin scaffold in eliciting anti-
amilodogenic effects, in particular when both the 4-hydroxy,3-methoxy and prenyloxy aryl substitution
patterns are present. Indeed, among the newly synthesized compounds, 1 turned out to be a very
promising anti-AD tool, endowed with a better biological profile with respect to cur. Due to its multipotent
behaviour it could be a lead compound, worthy further optimization to obtain effective AD-modifying drug
candidates.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
Chemicals and Materials. Synthetic Aβ42 was purchased from Bachem (Bubendorf, Switzerland) as
lyophilized powder and stored at -20°C. 1,1,1,3,3,3-Hexafluoropropan-2-ol (HFIP), acetonitrile (ACN),
dimethylsulfoxide (DMSO) and sodium carbonate (Na2CO3) were from Sigma-Aldrich (St. Louis, MO, USA).
Ethanol 96° was supplied by Carlo Erba (Cornaredo, Italy). Sodium hydroxide (NaOH) and sodium dodecyl
sulphate (SDS) were provided by Merck (Darmstadt, Germany). Na2HPO4 and NaH2PO4, supplied by Sigma-
Aldrich, were used for the preparation of the background electrolyte (BGE) in the CE analyses. BGE
solutions were prepared daily using Millipore Direct-QTM deionized water (Bedford, MA, USA) and filtered
on 0.45 µm Sartorius membrane filters (Göttingen, Germany). Uncoated fused-silica capillary was from
Polymicro Technologies (Phoenix, AZ, USA).
All cell culture reagents, culture medium and chemicals were purchased from Sigma-Aldrich.
Curcumin was purchased from Sigma Aldrich (analytical standard, purity ≥98%).
All samples containing curcumin and curcumin-derivatives were kept protected from light throughout all
experiments.
Chemistry. General Procedures. Starting materials, unless otherwise specified in the Experimental Section,
were used as high-grade commercial products. Solvents were of analytical grade. Melting points were
1
determined in open glass capillaries, using a Büchi apparatus and are uncorrected. H-NMR and 13C-NMR
spectra were recorded on Varian INOVA spectrometer operating at 400 MHz. Chemical shifts are reported
as parts per million (ppm  value). Standard abbreviations indicating spin multiplicities are given as follows:
s (singlet), d (doublet), dd (doublet of doublet), t (triplet), br (broad), q (quartet) or m (multiplet). High-
resolution mass spectra (HRMS) were acquired with an LTQ Orbitrap XL instrument (Thermo Fisher
Scientific, Rodano, Milano. Resolution was set at 30000 (FWHM at 400 m/z) and scan range was 250-1000
m/z with a target of 5 × 105 ions per scan. Instrument control and data analysis/elaboration were provided
19
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 20 of 41
1
2
3
4
5
6
7
8
9
by the software Xcalibur (version 2.07, ThermoFisher Scientific, Rodano, MI, Italy). Chromatographic
separations were performed on silica gel columns using the flash method (Kieselgel 40, 0.040-0.063 mm,
Merck). Reactions were followed by thin layer chromatography (TLC) on precoated silica gel plates (Merck
Silica Gel 60 F254) and then visualized with a UV lamp. Purities of all tested compounds used in the
biological assays were determined by HPLC using the area percentage method on the UV trace recorded at
254 nm. The analyses were performed under reversed-phase conditions on a Phenomenex Luna 5µm C18
column (150 × 4.60 mm), by using a ternary mixture of 0.1 % H3PO4/MeCN/MeOH (40:40:20, v/v) as mobile
phase, flow rate: 0.7 mL / min. A liquid chromatograph PU-1587 UV model equipped with a 20 μL loop
valve (Jasco Europe, Italy) was employed. All compounds were found to have >95% purity, as confirmed by
NMR spectra (see Supplementary material). Compounds were named using Chem-BioDraw Ultra 14.0
IUPAC name algorithm developed by CambridgeSoft Corporation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4-(3,3-dimethylallyloxy)benzaldehyde (7). 4-hydroxybenzaldehyde (1.00 g, 8.19 mmol) and 3,3-
dimethylallyl bromide (1.14 mL, 9.83 mmol) were allowed to react according to the general procedure of
the Williamson reaction for 8 h to give the crude product that was purified by crystallization from PE;
yellow oil, 83 % yield. 1H-NMR (400 MHz, CDCl3):  1.76 (s, 3H, CH3), 1.81 (s, 3H, CH3), 4.60 (d, 2H, J = 6.8 Hz,
OCH2), 5.45 (t, 1H, J = 6.8 Hz, CH=), 7.00 (d, 2H, J = 8.8 Hz, H-3’ and H-5’), 7.83 (d, 2H, J = 8.8 Hz, H-2’ and H-
6’), 9.88 (s, 1H, CHO).
Pabon reaction: general procedure for the synthesis of compounds 1, 2, and 5. To a stirred solution of
pentane-2,4-dione (1.00 mmol) in EtOAc (1.0 mL), B2O3 (1.0 molar equiv) was added, and the suspension
was stirred for 30 min at 80 °C before addition of a solution of the appropriate aldehyde/s, (0.9 molar equiv
for monoaryl or 1.8 molar equiv for bi-aryl curcumin derivatives), tri-n-butyl borate (2.0 molar equiv for
monoaryl or 4.0 molar equiv for bi-aryl curcumin derivatives), in EtOAc (0.5 mL). The reaction mixture was
stirred at 80 °C for 30 min, then a solution of n-butylammine (0.4 molar equiv in 1.0 mL of EtOAc) was
added over a period of 15 min. The mixture was heated to 80 °C for 8 h and then, after cooling to r.t., it was
acidified with 0.5 N HCl (30 mL) and then stirred at 80 °C for 30 min. The organic phase was separated and
the aqueous layer was extracted with EtOAc (3 x 10 mL). The combined organic layers were sequentially
washed with saturated aqueous NaHCO3 and brine, dried over Na2SO4, filtered, and concentrated under
reduced pressure. The crude residue was purified by flash column chromatography using a mixture of
petroleum ether / ethyl acetate (PE / EtOAc) as eluent, followed by crystallization from suitable solvent.
3Z,5E-4-hydroxy-6-(4-hydroxy-3-methoxyphenyl)hexa-3,5-dien-2-one (5) 28. Reaction of pentane-2,4-dione
(0.50 g, 5.00 mmol), B2O3 (0.45 g, 5.00 mmol), and vanillin (0.69 g, 4.5 mmol), in EtOAc (7.5 mL), gave a
crude product that was purified by flash chromatography (PE / EtOAc, 9.75:0.25), yellow powder, 55% yield,
1
mp 144-146 °C (EtOH). H NMR (400 MHz, CDCl3):  2.16 (s, 3H, CH3), 3.94 (s, 3H, OCH3), 5.40 (br, 1H, OH),
20
ACS Paragon Plus Environment

Page 21 of 41
ACS Chemical Neuroscience
1
2
3
4
5
5.63 (s, 1H, keto-enol-CH), 6.33 (d, J = 16.0 Hz, 1H, CH=CH), 6.92 (d, J = 8.0 Hz, 1H, H-5), 7.02 (d, J = 1.8 Hz,
1H, H-2), 7.09 (dd, J = 1.8, 8.0 Hz, 1H, H-6), 7.53 (d, J = 16.0 Hz, 1H, CH=CH).
6
7
8
9
(1E,4Z,6E)-5-hydroxy-7-(4-hydroxy-3-methoxyphenyl)-1-(4-((3-methylbut-2-en-1-yl)oxy)phenyl)hepta-
1,4,6-trien-3-one (1). Reaction of intermediate 5 (0.23 g, 1.00 mmol) and 7 (0.17 g, 0.9 mmol) gave a crude
product that was purified by flash chromatography (PE/EtOAc, 8.5:1.5) and further crystallization from
EtOH affording 1 as orange solid, 41 % yield, mp 144-146 °C. 1H-NMR (CDCl3):  1.77 (s, 3H, CH3), 1.82 (s, 3H,
CH3), 3.96 (s, 3H, OCH3), 4.56 (d, J = 6.8 Hz, 2H, OCH2), 5.50 (t, J = 6.8 Hz, 1H, CH=C), 5.80 (s, 1H, keto-enol-
CH), 5.85 (br s, 1H, OH), 6.49 (d, J = 15.6 Hz, 1H, CH=CH), 6.50 (d, J = 15.6 Hz, 1H, CH=CH), 6.94 (d, J = 8.0 Hz,
2H, H-3’and H-5’), 6.95 (d, J = 8.0 Hz, 1H, H-5), 7.07 (s, 1H, H-2), 7.13 (d, J = 8.0 Hz, 1H, H-6), 7.51 (d, J = 8.8
Hz, 2H, H-2’ and H-6’), 7.60 (d, J = 15.6 Hz, 1H, CH=CH), 7.63 (d, J = 15.6 Hz, 1H, CH=CH). 13C-NMR (CDCl3): 
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18.4, 26.0, 56.1, 65.1, 101.4, 109.7, 115.0, 115.2 (2C), 119.4, 122.0, 121.9, 123.1, 127.8, 127.9, 129.9 (2C),
14+
138.9, 140.4, 140.6, 146.9, 148.0, 160.8, 183.4, 183.5. HRMS (ESI) calcd. for C25H27O5
407.18530; found
407.18533 (Δ = 0.1 ppm).
(1E,4Z,6E)-5-hydroxy-1,7-bis(4-((3-methylbut-2-en-1-yl)oxy)phenyl)hepta-1,4,6-trien-3-one (2). Reaction
of pentane-2,4-dione (0.15 mL, 1.46 mmol) and 7 (0.5 g, 2.63 mmol) gave the crude product that was
purified by flash chromatography (PE/EtOAc, 9.75:0.25) and further crystallization from EtOH, affording 2 as
1
yellow solid, 45 % yield, mp 165-167 °C. H-NMR (400 MHz, CDCl3):  1.76 (s, 6H, CH3), 1.81 (s, 6H, CH3),
4.56 (d, J = 6.8 Hz, 4H, OCH2), 5.50 (t, J = 6.8 Hz, 2H, CH=C), 5.78 (s, 1H, keto-enol-CH), 6.51 (d, J = 15.6 Hz,
2H, CH=CH), 6.93 (d, J = 8.4 Hz, 4H, H-3’and H-5’), 7.51 (d, J = 8.4 Hz, 4H, H-2’ and H-6’), 7.63 (d, J = 15.6 Hz,
2H, CH=CH). 13C-NMR (100 MHz, CDCl3):  18.4 (2C), 26.0 (2C), 65.1 (2C), 101.5, 115.2 (4C), 119.4 (2C), 121.9
14+
(2C), 127.9 (2C), 129.9 (4C), 138.8 (2C), 140.3 (2), 160.8 (2C), 183.5 (2C). HRMS (ESI) calcd. for C29H33O4
445.23734; found 445.23553 (Δ = –1.6 ppm).
Claisen–Schmidt Reaction: general procedure for the synthesis of compounds 3, 4, and 6. To a solution of
ketone (1.0 mmol) and the selected aldehyde (1.1 mmol or 2.2 mmol) in EtOH (10 mL), a KOH aqueous
solution (50% p/v, 1 mL) was added dropwise. The reaction mixture was stirred overnight at room
temperature, then diluted with H2O and acidified with aqueous 6N HCl. The separated solid was collected
by vacuum filtration and purified by flash chromatography or by crystallization.
(E)-4-(4-hydroxy-3-methoxyphenyl)but-3-en-2-one (6). Reaction of propan-2-one (0.25 g, 5.00 mmol),
vanillin (0.98 g, 5.5 mmol), in EtOH (5.0 mL), gave a crude product that was purified by crystallization from
EtOH to give 6 as white solid, 75% yield, mp 95-98 °C. 1H NMR (400 MHz, CDCl3):  2.37 (s, 3H, CH3), 3.94 (s,
3H, OCH3), 5.99 (br, 1H, OH), 6.59 (d, J = 16.0 Hz, 1H, CH=CH), 6.94 (d, J = 8.4 Hz, 1H, H-6), 7.06 (d, J = 2.0
Hz, 1H, H-2), 7.09 (dd, J = 2.0, 8.4 Hz, 1H, H-5), 7.46 (d, J = 16.0 Hz, 1H, CH=CH).
(1E,4E)-1-(4-hydroxy-3-methoxyphenyl)-5-(4-((3-methylbut-2-en-1-yl)oxy)phenyl)penta-1,4-dien-3-one
(3). Reaction of intermediate 6 (0.19 g, 1.00 mmol) and 7 (0.21 g, 1.1 mmol), gave the crude product that
21
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 22 of 41
1
2
3
4
5
6
7
8
9
was purified by flash chromatography (PE/EtOAc, 7:3) and further crystallization from EtOH, affording 3 as
pale orange solid 85% yield, mp 101-113 °C. 1H NMR (400 MHz, CDCl3):  1.77 (s, 3H, CH3), 1.82 (s, 3H, CH3),
3.96 (s, 3H, OCH3), 4.56 (d, J = 6.8 Hz, 2H, OCH2), 5.50 (t, J = 6.8 Hz, 1H, CH=C), 5.97 (br, 1H, OH), 6.91-7.00
(m, 5H, CH=CH, H-5, H-3’, H-5’), 7.12 (d, J = 1.6 Hz, 1H, H-2), 7.18 (dd, J = 2.0, 8.4 Hz, 1H, H-6), 7.57 (d, J =
8.8 Hz, 2H, H-2’ and H-6’), 7.67 (d, J = 15.2 Hz, 1H, CH=CH), 7.71 (d, J = 15.2 Hz, 1H, CH=CH). 13C NMR (100
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MHz, CDCl3):  18.4, 26.0, 56.1, 65.1, 109.9, 115.0, 115.2 (2C), 119.3, 123.2, 123.5, 123.8, 127.6, 130.2 (2C),
14+
138.9, 142.9, 143.2, 147.0, 148.3, 161.1, 189.0. HRMS (ESI) calcd. for C23H25O4
365.17474; found
365.17471 (Δ = –0.1 ppm).
(1E,4E)-1,5-bis(4-((3-methylbut-2-en-1-yl)oxy)phenyl)penta-1,4-dien-3-one (4) 47. Reaction of propan-2-
one (0.25 g, 5.00 mmol), vanillin (1.95 g, 11.0 mmol), in EtOH (5.0 mL), gave a crude product that was
1
purified by crystallization from EtOH to give 4 as yellow solid, 87% yield, mp 74-76 °C. H NMR (400 MHz,
CDCl3):  1.77 (s, 6H, CH3), 1.82 (s, 6H, CH3), 4.56 (d, J = 6.8 Hz, 4H, OCH2), 5.50 (t, J = 6.8 Hz, 2H, CH=C),
6.94-6.98 (m, 6H, CH=CH, H-3, H-5),7.56 (d, J = 8.8 Hz, 4H, H-2 and H-6), 7.70 (d, J = 16.0 Hz, 2H, CH=). 13C
NMR (CDCl3)  18.3 (2C), 25.9 (2C), 65.0 (2C), 115.2 (4C), 119.3 (2C), 123.5 (2C), 127.6 (2C), 130.1 (4C),
138.8 (2C), 142.8 (2C), 161.0 (2C), 189.0. HRMS (ESI) calcd. for C27H31O3 14+ 403.22677; found 403.22675 (Δ =
0.0 ppm).
Capillary electrophoresis. Aβ42 peptide was solubilized by following the procedure described in 39. Briefly,
lyophilized Aβ42 was dissolved in HFIP and then the solvent was left to evaporate after an appropriate
incubation time. The Aβ42 aliquots were redissolved in a basic mixture (ACN/300 µM Na2CO3/250 mM
NaOH, 48.3:48.3:3.4, v/v/v) to obtain 500 µM Aβ42. This solution was then diluted to the operative
concentration (100 µM Aβ42 control peptide) with 20 mM phosphate buffer pH=7.4, with or without small
molecules.
Stock solutions of cur and of curcumin-based analogues (1.53 mM) were prepared in pure ethanol.
For co-incubation studies, 500 µM Aβ42 peptide (in the basic mixture) was resuspended in an appropriately
diluted compound solution, so as to keep the peptide concentration at 100 µM and obtain different
peptide/compound ratios: 1:100, 1:10, 1:4, 1:2 for cur and 1; 1:10, 1:4, 1:2 for 2 and 1:2 for 3 and 4. The
final percentage of ethanol was equal or lower than 3.26%.
The aggregation process of Aβ42 in the presence or not of cur and of curcumin-based analogues was
monitored by an Agilent Technologies 3D CE system with built-in diode-array detector (Waldbronn,
Germany), following the analytical method reported in 34. For the separation, a fused silica capillary
(Polymicro Technologies, Phoenix, AZ, USA) of 33 cm (24.5 cm, effective length) was employed. The
background electrolyte (BGE, 80 mM Na2HPO4/NaH2PO4 (pH 7.4)) was prepared daily and filtered on 0.45
µm membrane filters.The injection of the samples was carried out by applying a pressure of 30 mbar for 3
s. The capillary was thermostatted with circulating air at 25°C and separations were carried out at 12 kV
22
ACS Paragon Plus Environment

Page 23 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
(operative current: 75–78 µA) with the anode at the sample injection end. The acquisition wavelength was
200 nm. Oligomeric species were identified on the basis of effective mobilities (μeff), which are calculated
by subtracting the contribute of the electrosmotic flow (μEOF) from the apparent mobility (μapp).
Electrosmotic flow is measured as a perturbation of the baseline due to the sample solvent mixture.
Semiquantitative analyses were performed based on the normalized area % 34
.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Transmission Electron Microscopy. Amyloid fibril indentification was carried out by using a JEOL JEM
1400-Plus electron microscope (Peabody, MA, USA) operating at 80 kV. When no more peaks are detected
by CE, precipitated samples were prepared as follows: Aβ42 sample suspensions with or without cur and
analogues 1-4 were diluted at 10 µM with 20 mM Na2HPO4/NaH2PO4 and then 10 µL of diluted suspensions
were left to sediment on carbon-coated Formvar nickel grids (200 mesh) (Electron Microscopy Sciences,
Washington, PA,USA). After 15 minutes the excess of sample was drained off by means of a filter paper.
The negative staining was performed with 10 µL of 2%w/v uranyl acetate solution (Electron Microscopy
Sciences).
Docking studies. In order to rationalize the different activity of curcumin and the here proposed
derivatives, molecular docking simulations were performed involving the resolved Aβ42 peptide (PDB Id:
1IYT), which was simulated in its monomeric, dimeric and tetrameric forms. Calculations also included the
resolved amyloid fibril (PDB Id: 2MXU). The ligands were simulated in their keto-enolic form and their
conformational space was explored by combining MonteCarlo simulations and PM7-based semi-empirical
minimizations as previously described 48. In detail, the amyloid monomer was minimized by keeping fixed
the backbone atoms to preserve the resolved folding and utilized to build the corresponding dimer by
following the computational procedure as described by Rao and co-workers 49. Briefly, two optimized
monomers were initially aligned at an average distance between the backbone atoms of the two
monomers of about 10 Å and then the monomers were progressively approached and optimized to reach a
final average distance of about 5 Å. The resolved fibril structure underwent the same refinement protocol
as described for monomer before docking simulations. Thus, SMD simulations involved Aβ42 in its
monomeric, dimeric and fibril forms and were performed by using PLANTS and by including the entire
amyloid structure in its searches 50. For each simulated ligand, 20 poses were generated and ranked by
ChemPLP score with a speed equal to 1. All obtained poses were then minimized. The best dimeric
complexes were finally utilized to generate the corresponding tetramers by docking on them a second
dimer structure. Such a dimer-dimer docking was performed using PatchDock by adopting its default
parameters and the so generated 20 docking results were finally refined by FireDock 51. The so produced
best tetrameric complexes were finally minimized by keeping fixed the backbone atoms. The computed
tetrameric and fibril complexes for 1, 2 and cur were then neutralized and inserted into a 80 Å side box of
23
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 24 of 41
1
2
3
4
5
water. After a preliminary minimization, the systems underwent 3 ns SMD simulations with the same
characteristics as described in 52
.
6
7
8
9
Inhibition of cytokines release. All animal-related procedures complied with the ARRIVE guidelines and
were performed in accordance with EU guidelines for the care and use of laboratory animals and those of
the Italian Ministry of Health (D.Lg. 26/2014) and were approved by the Institutional Review Board for
Animal Research (Organismo Preposto al Benessere Animale, OPBA) of the University of Padua and by the
Italian Ministry of Health (Protocol number 958/2016-PR). One-day old Sprague-Dawley rat pups (CD strain)
were rapidly decapitated, minimizing suffering, discomfort or stress. Primary microglia cells were isolated
from mixed glial cell cultures prepared from cerebral cortex, as previously described 53. Briefly, upon
reaching confluence (7–10 days after isolation) microglia adhering to the astroglial monolayer were
dislodged by shaking (200 r.p.m. for 1 h at 37°C), resuspended in high-glucose Dulbecco’s Modified Eagle’s
Medium (DMEM) supplemented with 2 mM L-glutamine, 10% heat-inactivated fetal bovine serum (FBS),
100 U/mL penicillin, 100 μg/mL streptomycin and 50 μg/mL gentamicin and plated on uncoated plastic
wells at a density of 1.25 × 105 cells/cm2. Cells were allowed to adhere for 45 min and then washed to
remove non-adhering cells. After a 24 h incubation period, the medium was replaced with serum-free
medium containing the agents under study. Purity of the cultures was confirmed by immunocytochemistry
using a primary polyclonal antibody against ionized calcium binding adaptor molecule 1 (Iba1, 1:800, Wako
Chemicals USA Inc., Richmond, VA, USA). Ninety-seven per cent of the cells were Iba1 immunopositive.
Cells were maintained at 37°C in a humidified atmosphere containing 5% CO2/95% air.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Microglia were pre-treated for 1 hour with non-cytotoxic concentrations of cur and 1 and then stimulated
with 100 ng/mL Ultra-Pure LPS-EB for an additional 16 hours. At the end of incubation, culture medium was
collected and the levels of IL-1β and TNF-α released by microglia were assayed using a commercially
available ELISA kits (Antigenix America, Huntington Station, NY, USA), according to the manufacturer’s
instructions. Cytokine concentrations (pg/mL) in the medium were determined by reference to standard
curves obtained with known amounts of IL-1β and TNF-α and the results expressed as percentage relative
to LPS-stimulated cultures. Data were analyzed using GraphPad software, version 3.03 (GraphPad Software,
Inc., La Jolla, CA, USA) and expressed as mean ± SEM of at least 3 independent experiments. As raw values
varied between experiments and the variability could obscure the treatment effect, data were expressed as
percentage LPS treatment, taken as baseline of each independent experiment. Data were analyzed by
means of Kruskal-Wallis one-way analysis of variance followed by post-hoc Dunn's test for multiple
comparisons vs LPS treatment. Significance level was set at p < 0.05. Additional details are provided in the
figure legends, where appropriate.
Measurement of intracellular ROS. The fluorescent probe 2,7-dichlorodihydrofluorescein diacetate
(DCFH-DA) (Merck KGaA, Darmstadt, Germany) was used as a specific marker for quantitative intracellular
24
ACS Paragon Plus Environment

Page 25 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
ROS formation. Cells (2x104 cells per well) were loaded with 25 μM DCFH-DA for 45 min. After
centrifugation DCFH-DA was removed and cells were exposed to 5 and 10 μM of cur, compounds 1 and 2
and 300 μM H2O2. ROS levels were determined from 0 to 270 min using Synergy HT multidetection
microplate reader (BioTek) with excitation and emission wavelengths of 485 and 530 nm, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Immunodetection of Nrf2. The expression of Nrf2 in nuclear cell lysates was assessed using Western blot
analysis. Cell monolayers were washed twice with ice-cold PBS, harvested and subsequently homogenized
15 times using a glass-glass dounce homogenizer in ice-cold fractionation buffer (20 mM Tris/HCl pH 7.4, 2
mM EDTA, 0.5 mM EGTA, 0.32 M sucrose, 50 mM β-mercaptoethanol). The homogenate was centrifuged at
300×g for 5 min to obtain the nuclear fraction. An aliquot of the nuclear extract was used for protein
quantification, whereas the remaining extract was prepared for Western blot by mixing the nuclear cell
lysate with 2X sample buffer (125 mM Tris-HCl pH 6.8, 4% SDS, 20% glycerol, 6% β-mercaptoethanol, 0.1%
bromophenol blue) and then denaturing at 95°C for 5 min. Equivalent amounts of nuclear extracted
proteins were loaded into a SDS-PAGE gel, electrophoresed under reducing conditions, transferred to a
PVDF membrane (Merck KGaA, Darmstadt Germany) and then blocked for 1 hour with 5% w/v BSA in Tris-
buffered saline containing 0.1% Tween 20 (TBS-T). Membranes were immunoblotted with rabbit anti-
human Nrf2 (1:2000) diluted in 5% w/v BSA in TBS-T. Detection was carried out by incubation with
horseradish peroxidase conjugated goat anti-rabbit IgG (1:5000 dilution in 5% w/v BSA in TBS-T) for 1 hour
at room temperature. Membranes were then washed three times with TBS-T and proteins of interest were
visualized using an enhanced chemiluminescent reagent (Pierce, Rockford, IL, USA). Lamin A/C was
performed as a control for gel loading.
SUPPORTING INFORMATION
Cell viability data, further capillary electrophoresis data, further molecular docking results and steered
molecular dynamics data, 1H NMR spectra, 13C NMR spectra, 2D-COSY spectra.
ABBREVIATIONS
AD, Alzheimer’s disease; Aβ, amyloid-beta; Aβ42, Aβ1-42; CE, capillary electrophoresis; AUC, area under the
curve; DMSO, dimethylsulfoxide; cur, curcumin; HFIP, 1,1,1,3,3,3-Hexafluoropropan-2-ol; LPS,
lipopolysaccharide; SMD, steered molecular dynamics; MTT, 3-(4,5-dimethylthiazol-2-ly)-2,5-
diphenyltetrazolium bromide; ROS, reactive oxygen species; TEM, transmission electron microscopy;
25
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 26 of 41
1
2
3
4
AUTHOR INFORMATION
5
6
Corresponding author: Prof. Ersilia De Lorenzi, Department of Drug Sciences, University of Pavia, Viale
7
8
Taramelli 12, 27100, Pavia, Italy; tel +39 0382 987747; fax +39 0382 422975; email ersidelo@unipv.it.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Author contributions: E. De Lorenzi conceived the project, supervised all contributions and finalized the
draft manuscript. F. Bisceglia performed CE studies and data analysis, TEM data interpretation, figures and
manuscript preparation. F. Seghetti synthesized curcumin analogues and performed NMR data. M. Serra
assisted data analysis and interpretation, the revision of final version of the manuscript and NMR results.
M. Zusso conceived and carried out the data on microglia, wrote the draft text and prepared the figures of
the relative section. S. Gervasoni performed molecular modelling studies and assisted data discussion. L.
Verga performed TEM data and elaborated the relative images. G. Vistoli conceived and supervised the
molecular modelling studies and wrote the draft of the relative section. C. Lanni conceived and supervised
the experiments on anti-oxidant activity and cell viability on neuroblastoma cells, wrote the draft of the
relative section. M. Catanzaro carried out the cell viability experiments on neuroblastoma cells and the
data on anti-oxidant activity. F. Belluti conceived the design of curcumin analogues, supervised synthesis
and NMR characterization, drafted the manuscript.
The authors declare that none of them has competing financial interests in this work.
Funding sources:
University of Padua, Italy (Progetto di Ateneo, grant n. CPDA144389/14 to M.Z.); University of Pavia, Italy
(PhD student grant to F.Bi); University of Bologna (PhD student grant to F.S.).
Conflict of interest: The authors declare that none of them has competing financial interests in this work.
ACKNOWLEDGEMENTS
We thank Dr. Luca Regazzoni (University of Milan, Italy) for providing the HRMS data.
26
ACS Paragon Plus Environment

Page 27 of 41
ACS Chemical Neuroscience
1
2
3
REFERENCES
4
5
6
7
8
1. Winblad, B., Amouyel, P., Andrieu, S., Ballard, C., Brayne, C., Brodaty, H., Cedazo-Minguez, A., Dubois, B.,
Edvardsson, D., Feldman, H., Fratiglioni, L., Frisoni, G. B., Gauthier, S., Georges, J., Graff, C., Iqbal, K.,
Jessen, F., Johansson, G., Jonsson, L., Kivipelto, M., Knapp, M., Mangialasche, F., Melis, R., Nordberg, A.,
Rikkert, M. O., Qiu, C., Sakmar, T. P., Scheltens, P., Schneider, L. S., Sperling, R., Tjernberg, L. O.,
Waldemar, G., Wimo, A., and Zetterberg, H. (2016) Defeating Alzheimer's disease and other dementias:
a priority for European science and society, Lancet Neurol 15, 455-532.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2. Chiti, F., and Dobson, C. M. (2017) Protein Misfolding, Amyloid Formation, and Human Disease: A
Summary of Progress Over the Last Decade, Annu Rev Biochem 86, 35.1-35.42.
3. Calsolaro, V., and Edison, P. (2016) Neuroinflammation in Alzheimer's disease: Current evidence and
future directions, Alzheimers Dement 12, 719-732.
4. Querfurth, H. W., and LaFerla, F. M. (2010) Alzheimer's disease, N Engl J Med 362, 329-344.
5. Younkin, S. G. (1995) Evidence that A beta 42 is the real culprit in Alzheimer's disease, Ann Neurol 37,
287-288.
6. Da Mesquita, S., Ferreira, A. C., Sousa, J. C., Correia-Neves, M., Sousa, N., and Marques, F. (2016) Insights
on the pathophysiology of Alzheimer's disease: The crosstalk between amyloid pathology,
neuroinflammation and the peripheral immune system, Neurosci Biobehav Rev 68, 547-562.
7. Maji, S. K., Ogorzalek Loo, R. R., Inayathullah, M., Spring, S. M., Vollers, S. S., Condron, M. M., Bitan, G.,
Loo, J. A., and Teplow, D. B. (2009) Amino acid position-specific contributions to amyloid beta-protein
oligomerization, J Biol Chem 284, 23580-23591.
8. Liao, M. Q., Tzeng, Y. J., Chang, L. Y., Huang, H. B., Lin, T. H., Chyan, C. L., and Chen, Y. C. (2007) The
correlation between neurotoxicity, aggregative ability and secondary structure studied by sequence
truncated Abeta peptides, FEBS Lett 581, 1161-1165.
9. Tjernberg, L. O., Lilliehook, C., Callaway, D. J., Naslund, J., Hahne, S., Thyberg, J., Terenius, L., and
Nordstedt, C. (1997) Controlling amyloid beta-peptide fibril formation with protease-stable ligands, J
Biol Chem 272, 12601-12605.
10. Belluti, F., Rampa, A., Gobbi, S., and Bisi, A. (2013) Small-molecule inhibitors/modulators of amyloid-
beta peptide aggregation and toxicity for the treatment of Alzheimer's disease: a patent review (2010 -
2012), Expert Opin Ther Pat 23, 581-596.
11. Cleary, J. P., Walsh, D. M., Hofmeister, J. J., Shankar, G. M., Kuskowski, M. A., Selkoe, D. J., and Ashe, K.
H. (2005) Natural oligomers of the amyloid-beta protein specifically disrupt cognitive function, Nat
Neurosci 8, 79-84.
27
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 28 of 41
1
2
3
4
5
6
12. Walsh, D. M., Klyubin, I., Fadeeva, J. V., Cullen, W. K., Anwyl, R., Wolfe, M. S., Rowan, M. J., and Selkoe,
D. J. (2002) Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term
potentiation in vivo, Nature 416, 535-539.
7
8
9
13. Benilova, I., Karran, E., and De Strooper, B. (2012) The toxic Abeta oligomer and Alzheimer's disease: an
emperor in need of clothes, Nat Neurosci 15, 349-357.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14. Lee, S. J., Nam, E., Lee, H. J., Savelieff, M. G., and Lim, M. H. (2017) Towards an understanding of
amyloid-beta oligomers: characterization, toxicity mechanisms, and inhibitors, Chem Soc Rev 46, 310-
323.
15. Leon, R., Garcia, A. G., and Marco-Contelles, J. (2013) Recent advances in the multitarget-directed
ligands approach for the treatment of Alzheimer's disease, Med Res Rev 33, 139-189.
16. Di Martino, R. M. C., Bisi, A., Rampa, A., Gobbi, S., and Belluti, F. (2017) Recent progress on curcumin-
based therapeutics: a patent review (2012-2016). Part II: curcumin derivatives in cancer and
neurodegeneration, Expert Opin Ther Pat 27, 953-965.
17. Esatbeyoglu, T., Huebbe, P., Ernst, I. M., Chin, D., Wagner, A. E., and Rimbach, G. (2012) Curcumin--from
molecule to biological function, Angew Chem Int Ed Engl 51, 5308-5332.
18. Prasad, S., Gupta, S. C., Tyagi, A. K., and Aggarwal, B. B. (2014) Curcumin, a component of golden spice:
from bedside to bench and back, Biotechnol Adv 32, 1053-1064.
19. Masuda, Y., Fukuchi, M., Yatagawa, T., Tada, M., Takeda, K., Irie, K., Akagi, K., Monobe, Y., Imazawa, T.,
and Takegoshi, K. (2011) Solid-state NMR analysis of interaction sites of curcumin and 42-residue
amyloid beta-protein fibrils, Bioorg Med Chem 19, 5967-5974.
20. Zhao, L. N., Chiu, S. W., Benoit, J., Chew, L. Y., and Mu, Y. (2012) The effect of curcumin on the stability
of Abeta dimers, J Phys Chem B 116, 7428-7435.
21. Zhang, X., Tian, Y., Li, Z., Tian, X., Sun, H., Liu, H., Moore, A., and Ran, C. (2013) Design and synthesis of
curcumin analogues for in vivo fluorescence imaging and inhibiting copper-induced cross-linking of
amyloid beta species in Alzheimer's disease, J Am Chem Soc 135, 16397-16409.
22. Orlando, R. A., Gonzales, A. M., Royer, R. E., Deck, L. M., and Vander Jagt, D. L. (2012) A chemical analog
of curcumin as an improved inhibitor of amyloid Abeta oligomerization, PLoS One 7, e31869.
23. Reinke, A. A., and Gestwicki, J. E. (2007) Structure-activity relationships of amyloid beta-aggregation
inhibitors based on curcumin: influence of linker length and flexibility, Chem Biol Drug Des 70, 206-215.
24. Nelson, K. M., Dahlin, J. L., Bisson, J., Graham, J., Pauli, G. F., and Walters, M. A. (2017) The Essential
Medicinal Chemistry of Curcumin, J Med Chem 60, 1620-1637.
25. Baell, J., and Walters, M. A. (2014) Chemistry: Chemical con artists foil drug discovery, Nature 513, 481-
483.
28
ACS Paragon Plus Environment

Page 29 of 41
ACS Chemical Neuroscience
1
2
3
4
5
6
7
26. Jasial, S., Hu, Y., and Bajorath, J. (2017) How Frequently Are Pan-Assay Interference Compounds Active?
Large-Scale Analysis of Screening Data Reveals Diverse Activity Profiles, Low Global Hit Frequency, and
Many Consistently Inactive Compounds, J Med Chem 60, 3879-3886.
8
9
27. Chojnacki, J. E., Liu, K., Yan, X., Toldo, S., Selden, T., Estrada, M., Rodriguez-Franco, M. I., Halquist, M. S.,
Ye, D., and Zhang, S. (2014) Discovery of 5-(4-hydroxyphenyl)-3-oxo-pentanoic acid [2-(5-methoxy-1H-
indol-3-yl)-ethyl]-amide as a neuroprotectant for Alzheimer's disease by hybridization of curcumin and
melatonin, ACS Chem Neurosci 5, 690-699.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28. Di Martino, R. M., De Simone, A., Andrisano, V., Bisignano, P., Bisi, A., Gobbi, S., Rampa, A., Fato, R.,
Bergamini, C., Perez, D. I., Martinez, A., Bottegoni, G., Cavalli, A., and Belluti, F. (2016) Versatility of the
Curcumin Scaffold: Discovery of Potent and Balanced Dual BACE-1 and GSK-3beta Inhibitors, J Med
Chem 59, 531-544.
29. Kuzuyama, T., Noel, J. P., and Richard, S. B. (2005) Structural basis for the promiscuous biosynthetic
prenylation of aromatic natural products, Nature 435, 983-987.
30. Brezani, V., Smejkal, K., Hosek, J., and Tomasova, V. (2018) Anti-inflammatory Natural Prenylated
Phenolic Compounds - Potential Lead Substances, Curr Med Chem 25, 1094-1159.
31. Marchiani, A., Rozzo, C., Fadda, A., Delogu, G., and Ruzza, P. (2014) Curcumin and curcumin-like
molecules: from spice to drugs, Curr Med Chem 21, 204-222.
32. Pabon, H. J. J. A synthesis of curcumin and related compounds. Recl Trav Chim Pay B 83, 379-386.
33. Mercanti, G., Ragazzi, E., Toffano, G., Giusti, P., and Zusso, M. (2014) Phosphatidylserine and curcumin
act synergistically to down-regulate release of interleukin-1beta from lipopolysaccharide-stimulated
cortical primary microglial cells, CNS Neurol Disord Drug Targets 13, 792-800.
34. Bisceglia, F., Natalello, A., Serafini, M. M., Colombo, R., Verga, L., Lanni, C., and De Lorenzi, E. (2018) An
integrated strategy to correlate aggregation state, structure and toxicity of Ass 1-42 oligomers, Talanta
188, 17-26.
35. Sabella, S., Quaglia, M., Lanni, C., Racchi, M., Govoni, S., Caccialanza, G., Calligaro, A., Bellotti, V., and De
Lorenzi, E. (2004) Capillary electrophoresis studies on the aggregation process of beta-amyloid 1-42 and
1-40 peptides, Electrophoresis 25, 3186-3194.
36. Colombo, R., Carotti, A., Catto, M., Racchi, M., Lanni, C., Verga, L., Caccialanza, G., and De Lorenzi, E.
(2009) CE can identify small molecules that selectively target soluble oligomers of amyloid beta protein
and display antifibrillogenic activity, Electrophoresis 30, 1418-1429.
37. Butini, S., Brindisi, M., Brogi, S., Maramai, S., Guarino, E., Panico, A., Saxena, A., Chauhan, V., Colombo,
R., Verga, L., De Lorenzi, E., Bartolini, M., Andrisano, V., Novellino, E., Campiani, G., and Gemma, S.
(2013) Multifunctional cholinesterase and amyloid Beta fibrillization modulators. Synthesis and
biological investigation, ACS Med Chem Lett 4, 1178-1182.
29
ACS Paragon Plus Environment