A Focused Library of Psychotropic Analogues with Neuroprotective and Neuroregenerative Potential

Article abstract of DOI:10.1021/acschemneuro.8b00242

Overcoming the lack of effective treatments and the continuous clinical trial failures in neurodegenerative drug discovery might require a shift from the prevailing paradigm targeting pathogenesis to the one targeting simultaneously neuroprotection and neuroregeneration. In the studies reported herein, we sought to identify small molecules that might exert neuroprotective and neuroregenerative potential as tools against neurodegenerative diseases. In doing so, we started from the reported neuroprotective/neuroregenerative mechanisms of psychotropic drugs featuring a tricyclic alkylamine scaffold. Thus, we designed a focused-chemical library of 36 entries aimed at exploring the structural requirements for efficient neuroprotective/neuroregenerative cellular activity, without the manifestation of toxicity. To this aim, we developed a synthetic protocol, which overcame the limited applicability of previously reported procedures. Next, we evaluated the synthesized compounds through a phenotypic screening pipeline, based on primary neuronal systems. Phenothiazine 2Bc showed improved neuroregenerative and neuroprotective properties with respect to reference drug desipramine (2Aa). Importantly, we have also shown that 2Bc outperformed currently available drugs in cell models of Alzheimer's and Parkinson's diseases and attenuates microglial activation by reducing iNOS expression.

Full text of DOI:10.1021/acschemneuro.8b00242

Subscriber access provided by University of Sunderland
Article
A focused library of psychotropic analogs with
neuroprotective and neuroregenerative potential
Elisa Uliassi, Luis Emiliano Peña-Altamira, Aixa V. Morales, Francesca Massenzio, Sabrina Petralla, Michele
Rossi, Marinella Roberti, Loreto Martínez-González, Ana Martínez, Barbara Monti, and Maria Laura Bolognesi
ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.8b00242 • Publication Date (Web): 25 Sep 2018
Downloaded from http://pubs.acs.org on September 27, 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 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
A focused library of psychotropic analogs with
neuroprotective and neuroregenerative potential
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a, ‡
a, ‡
b
a
Elisa Uliassi , Luis Emiliano Peña-Altamira , Aixa V. Morales , Francesca Massenzio ,
a
a
a
c
Sabrina Petralla , Michele Rossi , Marinella Roberti , Loreto Martinez Gonzalez , Ana
c
a
a
Martinez , Barbara Monti , and Maria Laura Bolognesi*
a
Department of Pharmacy and Biotechnology, Alma Mater Studiorum ꢀ University of Bologna,
Via Belmeloro 6, Bologna, Italy;
b
Department of Cellular, Molecular and Developmental Neurobiology, Cajal Institute, Consejo
Superior de Investigaciones Científicas (CSIC), Madrid, Spain;
c
Department of Chemical and Physical Biology, Centro de Investigaciones BiológicasꢀCSIC,
Madrid, Spain.
ACS Paragon Plus Environment

ACS Chemical Neuroscience
Page 2 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
ABSTRACT
Overcoming the lack of effective treatments and the continuous clinical trial failures in
neurodegenerative drug discovery might require a shift from the prevailing paradigm targeting
pathogenesis to the one targeting simultaneously neuroprotection and neuroregeneration. In the
studies reported herein, we sought to identify small molecules that might exert neuroprotective
and neuroregenerative potential as tools against neurodegenerative diseases. In doing so, we
started from the reported neuroprotective/neuroregenerative mechanisms of psychotropic drugs
featuring a tricyclic alkylamine scaffold. Thus, we designed a focusedꢀchemical library of 36
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
entries
aimed
at
exploring
the
structural
requirements
for
efficient
neuroprotective/neuroregenerative cellular activity, without the manifestation of toxicity. To this
aim, we developed a synthetic protocol, which overcame the limited applicability of previously
reported procedures. Next, we evaluated the synthesized compounds through a phenotypic
screening pipeline, based on primary neuronal systems. Phenothiazine 2Bc showed improved
neuroregenerative and neuroprotective properties with respect to reference drug desipramine
(2Aa). Importantly, we have also shown that 2Bc outperformed currently available drugs in cell
models of Alzheimer's and Parkinson's diseases and attenuates microglial activation by reducing
iNOS expression.
Keywords: neurodegenerative diseases, psychotropic agents, phenotypic screening,
phenothiazines, neuroprotection, neural regeneration
ACS Paragon Plus Environment
2

Page 3 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Introduction
The progress in understanding processes and mechanisms involved in neurodegenerative
diseases, such as Alzheimer’s disease (AD), has been impressive in the last decade. Despite that,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
no effective drug or drug candidate has been identified in the wake of that understanding. The
current discouraging record of translation has been attributed to various factors, spanning from
1
the lack of sufficient target engagement to the presence of serious toxic effects. Furthermore, the
1
multifactorial etiology and complex pathophysiology of AD, together with the absence of
2
diagnostic methods that allow detecting the disease earlier, has likely contributed to AD drug
1
failures. Closely related to this is the paucity of validated drug targets, which has been
exacerbated by the recent Phase III failures of investigational drugs designed to reduce amyloidꢀ
3
β (Aβ) plaque formation. On the other hand, the marketed drugs (three acetylcholinesterase
inhibitors and one NꢀmethylꢀDꢀaspartate (NMDA) antagonist) were not generated from recent
molecular insights, but from previous neurotransmitter system research. Unfortunately, they
4
show limited ability to modify disease course, offering only transient benefit to patients. This
situation is not different from that of Parkinson’s disease (PD), where similar symptomatic
5
treatments are available.
These frustrating aspects fueled many researchers to question whether the viability of
neurodegenerative targets for treating AD and related dementia might yield diseaseꢀmodifying
6
drugs. Indeed, in addition to the classical “positive” neurodegeneration lesions, such as Aβ
plaques and neurofibrillary tangles, characteristic losses of neurons and synaptic elements are
7
core “negative” features. Particularly, a significant reduction of neuronal density caused by cell
7
death in specific hippocampal areas is one of the major hallmarks of AD progression. This
suggests the need for restorative intervention and prompts the pharmaceutical community to
ACS Paragon Plus Environment
3

ACS Chemical Neuroscience
Page 4 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
explore the emerging possibility of targeting neuroprotection and neuroregeneration to revert
6
, 8ꢀ10
neuronal death.
In principle, activating endogenous neuroprotective pathways and sustaining
brain neurogenesis could be more effective than blocking Aβꢀ and tauꢀmediated toxic cascades
6
, 8ꢀ10
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
leading to neurodegeneration.
On this basis, small molecules might prevent death and
11,12
restore function to damaged neurons by directly targeting neuroprotective mechanisms.
Thus,
the prospect of developing new chemical entities capable of modulating regenerative processes
in vivo, has attracted much interest from the neurodegenerative diseases drug discovery
community. However, due to the complexity of the involved processes, a distinct molecular
target for drug discovery endeavors has not been prioritized yet. Conversely, phenotypic
approaches, i.e., to find compounds that modulate key phenotypes of dementia, without
requirement for the detailed mechanistic knowledge that is often lacking, has gained a new
13
momentum.
With these concepts in mind, we were interested in small molecules that, by crossing the
blood–brain barrier (BBB) and gaining access to the central nervous system (CNS), might
promote a neuroprotective and neuroregenerative phenotype in diseased neuronal cells, while
showing no toxic effect. Indeed, in addition to lack of efficacy, low BBB penetration and liver
1
4
toxicity are still frequent reasons for AD clinical trial discontinuation. Thus, following a
chemocentric approach, we looked for chemotypes possessing these distinct features.
We were illuminated by the reported neuroprotective/neuroregenerative mechanisms of
1
5
psychotropic drugs. Particularly, the effect on neurogenesis of tricyclic antidepressants (TCA)
16, 17
has been widely recognized.
It has been demonstrated that chronic TCA treatment increases
the number of hippocampal granule cells and protects neurons from further damage, or possibly
1
6, 17
even reverse the damage that has already occurred.
Thus, imipramine (1Aa in Figure 1) has
ACS Paragon Plus Environment
4

Page 5 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
18
been proposed as a hippocampal neurogenic agent, although it is hard to relate this effect to its
1
9
complex pharmacodynamic mechanisms. Moreover, 1Aa has been shown as a neuroprotectant
2
0
against glutamateꢀinduced toxicity in primary rat cortical cells. Similarly, phenothiazine (PTZ)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
antipsychotic drugs, such as chlorpromazine (1Ab in Figure 1), have displayed neuroprotective
21
effects by protecting mitochondria and preventing apoptosis, and by promoting neuronal
2
2
regeneration and survival. Knowingly, TCA and PTZ share similar chemical motifs: i.e., a
tricyclic moiety connected to an amino functionality through an alkyl linker. More generally,
they all belong to the arylalkylamine class, which represents one of the oldest and most
frequently encountered structural motifs among CNSꢀactive, naturally occurring or synthetic
2
3
small molecules. Thus, they might be considered “privileged” for the synthesis of a hypothesisꢀ
24
driven library of potential neuroprotective/neuroregenerative small molecules.
On this basis, we report the results of an in vitro phenotypic evaluation of a focused chemical
library (subsets 1ꢀ3 in Figure 1) aimed at exploring the structural requirements for efficient
neuroprotective/neuroregenerative cellular activity, without the manifestation of toxicity. The
library has been built around the selected tricyclic alkylamine scaffold (Figure 1) (i) by
diversifying the tricyclic portion (aꢀf), (ii) by changing the methylene units of the linker (A and
B), (iii) by varying the amino group from a tertiary (subset 1) to a secondary (2), or primary
amine (3). As illustrated in Figure 1, we selected three tricyclic scaffolds. They were the
dibenzoazepine a of 1Aa, as well as the 10,11ꢀdihydroꢀdibenzazepine e of the mood stabilizer
carbamazepine (1 in Figure S1), which has been similarly shown to induce neuronal
25
differentiation and neuroprotection. We also included differently substituted PTZs (bꢀd), which
26
have been shown to exert neuroprotective effects through different mechanisms. The choice of
the carbazole (CBZ) moiety f was motivated by the fact that it belongs to aminopropyl carbazole
ACS Paragon Plus Environment
5

ACS Chemical Neuroscience
Page 6 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
P7C3 (2 in Figure S1), with outstanding neuroregenerative and neuroprotective profile in various
27
neurodegenerative conditions. We came up with the design of a combinatorial library of 36
2
8
entries.
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Design strategy to subsets 1ꢀ3.
Once the library has been synthesized, a cellular pipeline has been set up to screen subsets
1A(B)a-f, 2A(B)a-f, and 3A(B)a-f in a targetꢀagnostic fashion. Anticipating our results, the
preliminary evaluation of viability profiles in primary neurons and liver cells was performed.
ACS Paragon Plus Environment
6

Page 7 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Fourteen out thirtyꢀsix initial compounds were further progressed to the neuroprotection assay.
To ascertain the neuroregenerative potential, neurosphere (NS) assay was subsequently carried
out. As final step, the neuroprotective and neuroregenerative hit identified (2Bc) was tested in
cellular models of neurodegenerative diseases and neuroinflammation.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chemistry
The reported synthetic route for TCA and PTZ are obsolete and hazardous procedures,
29
optimized as industrial processes more than half a century ago. In particular, a challenging step
is the functionalization of the aromatic nitrogen, which involved Nꢀalkylation by using strong
2
9
alkali metals or tedious multistep routes exploiting toxic reagents. These undesirable
characteristics result in limited applicability to expedite broad SAR studies. To overcome this
challenge, we turned our attention to a oneꢀpot reductive Nꢀalkylation procedure previously
3
0
reported by Righi. It was extended to the direct Nꢀalkylation of aꢀf with acetals 7 and 8, in the
presence of triethylsilane (TES), as nontoxic and stable reducing agent (Scheme 2).
a
Scheme 1. Synthesis of acetal intermediates 7 and 8.
a
Reagents and conditions: a. 5, NaBH , NiSO Ac O, MeOH, rt (56%); b. 6, TFAA, TEA, THF,
4
4,
2
N , rt (75%).
2
ACS Paragon Plus Environment
7

ACS Chemical Neuroscience
Page 8 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
To obtain acetamide 7 from nitrile 5, we slightly modified a procedure involving the use of a
large excess of NaBH (7 eq.), NiSO as catalyst, and acetic anhydride as trapping agent, to
4
4
prevent dimerization side reaction, often occurring during the reduction of nitriles to primary
3
1
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
amine (Scheme 1). The synthesis of 8 was accomplished by reacting commercially available
amine 6 with trifluoroacetic anhydride (TFAA).
With the acetals 7 and 8 in hands, the reductive Nꢀalkylation reaction of tricyclic aromatic
3
0
amines a-f was performed (Scheme 2). This provided alkylꢀamide derivatives 4 in very good
yields (62%ꢀ95%). Probably, the lower yields for 4A(a-f) (62%ꢀ82%), compared to those for
4
B(a-f) (70%ꢀ95%), reflects the lower reactivity of acetamide 7 with respect to
trifluoroacetamide 8.
a
Scheme 2. Synthesis of target subsets 1ꢀ3.
a
Reagents and conditions: a. 7 or 8, TES, TFA, DCM, rt (62%ꢀ95%); b. 4A(a-f), KOH,
MeOH/H O, MW, 150 W, 160 °C (72%ꢀ98%); c. 4B(a-f), K CO , MeOH/H O, rt (quantitative);
2
2
3
2
d. 3A(a-c) or 3B(a-c), HCOOH, HCOH, H O, MW, 50 W, 100 °C (63%ꢀ74%); e. 3A(b-d) or
2
ACS Paragon Plus Environment
8

Page 9 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
3
B(b-d), HCOOH, HCOH, H O, 80 °C (45%ꢀ84%); f. HCOH, K10ꢀclay, NaBH , MeOH, MW,
2
4
5
0 W, 80 °C (24%ꢀ44%).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The subsequent hydrolysis of 4 to afford primary amines of series 3 involved two different
procedures. In the first one, treatment of 4A(a-f) under microwave (MW) irradiation with a
strong base (KOH) gave the primary amines 3A(a-f) in good to excellent yield (72%ꢀ98%). It
should be noted that under conventional reflux conditions, dramatic differences have been
observed in terms of both yields (45%ꢀ67%) and decreased reaction time (2 days vs. 90 min).
Conversely, hydrolysis of 4B(a-f) under milder reaction conditions (K CO ) led to primary
2
3
amines 3B(a-f) in quantitative yield.
To synthesize tertiary amines of subset 1, we developed a MWꢀassisted Eschweiler–Clarke
procedure, which led to consistent increases in yields (63%ꢀ74%) and shortening of reaction time
(from 10 h to 30 min), compared to conventional heating protocol. However, MW irradiation
failed in the case of PTZ 3A(b-d) and 3B(b-d), which displayed an excessive reactivity under
these conditions. Thus, conventional heating (80 °C) for 8ꢀ10 h provided 1A(b-d) and 1B(b-d) in
satisfactory to good yields (45%ꢀ84%). Finally, primary amines of series 3 were converted to the
corresponding secondary amines 2 by exploiting a modified version of a MWꢀassisted reductive
3
2
amination. This was based on the use of paraformaldehyde and acidic clay montmorillonite Kꢀ
10 as green solid catalyst. Further optimization of reaction conditions led to subset 2 in
1
13
acceptable yields (24%ꢀ44%). All the final compounds were characterized by Hꢀ and CꢀNMR,
MS analyses and their purity confirmed to be ≥ 95% by analytical HPLC.
ACS Paragon Plus Environment
9

ACS Chemical Neuroscience
Page 10 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
Results and Discussion
With the 36 library compounds in hand, we were interested to implement a phenotypic
screening pipeline aimed to identify those more effective in counteracting the regenerative
failure typical of injured neurons and, at the same time, enhancing neurogenesis. Although most
of the synthesized compounds have been reported as antidepressants, antipsychotics and
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
33ꢀ42
antihistaminics,
to our knowledge, this is the first study allowing for their direct comparison
and potentially enabling SAR exploration and identification of key features for the design of new
neuroprotective/neuroregenerative compounds.
To meet this goal, we looked for a screening pipeline tailored to our academic setting and
taking into account the ethical constraints related to the manipulation of primary cultures.
Recently, the use of in vitro toxicity screens to facilitate prioritization of compounds for further
4
3
testing was emphasized. On this basis, we thought that cytotoxicity determination as primary
test might improve effectiveness of our screening in a timely and costꢀeffective manner. Thus,
the preliminary evaluation of viability profiles in primary neurons and liver cells was performed,
to further progress to the neuroprotective and neuroregenerative assay only those molecules
devoid of toxicity. In particular, the unbiased phenotypic screening pipeline depicted in Figure 2
was defined, whose main goals are: (i) to make a preliminary screening of molecules for
neuronal and hepatic viabilities; (ii) to test the neuroprotective properties of the less toxic
+
compounds against neurodegeneration caused by low serum and K deprivation in rat primary
neurons; iii) to assess the most promising compounds for their neurogenic activity in adult mouse
NS; iv) to assess BBB permeability; v) to validate the neuroprotective and neuroregenerative
identified hit in established cellular models of neurodegenerative diseases and
neuroinflammation.
ACS Paragon Plus Environment
10

Page 11 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Schematic drawing of the phenotypic pipeline used to screen subsets 1ꢀ3.
(i) First, we tested all subsets 1ꢀ3 on primary rat cerebellar granule neurons (CGNs), in order to
proceed only with those molecules showing no neurotoxicity. Primary cells provide higherꢀ
quality models than immortalized cell lines, as they form synapses and incorporate significant
neuromodulatory and trophic inputs. Particularly, CGNs are considered a reliable model for
studying
cellular
and
molecular
mechanisms
of
survival/apoptosis
and
4
4
neurodegeneration/neuroprotection. These primary neuronal cultures have the advantage to be
almost pure neuronal cells. Furthermore, being CGNs prepared from rat pups at 7 days from
birth, their preparation has a higher yield than other primary neuronal cultures. This has an
important ethical implication, in agreement with the reduction requirements of the EU and Italian
laws on the use of animals for experimental purposes. Thus, differentiated CGNs were exposed
to 1ꢀ3 for 24 h at clinically relevant concentrations varying from 0 to 50 ꢁM (i.e., 5, 25, 50 ꢁM).
Cell viability was measured by MTT, which is one of the most widely used assay to screen drug
cytotoxicity. Although it is worth noting that MTT reduction is a marker reflecting viable cell
ACS Paragon Plus Environment
11

ACS Chemical Neuroscience
Page 12 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
45
metabolism and not specifically cell proliferation, MTT assay remains a widely accepted
approach for measuring cell viability/drug cytotoxicity, as indicated by a huge amount of
4
6, 47
literature starting from the late ‘50s.
Analysis for data statistical significance was performed
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
though oneꢀway ANOVA followed by Bonferroni’s postꢀhoc comparison test (Figure 3). Antiꢀ
AD drug galantamine (3 in Figure S1) was used as positive control (Figure S2A).
ACS Paragon Plus Environment
12

Page 13 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
Figure 3. Toxicity of 1ꢀ3 to differentiated CGNs after 24 h treatment. The red line marks
cytotoxicity of 50%. Results are expressed as percentage of controls and are the mean ± SE of at
least 3 different experiments, each run in triplicate. ***p<0.001; **p<0.01; *p<0.05 compared to
control conditions (0 ꢁM), Bonferroni’s test after ANOVA.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
As expected, being these compounds mainly derived by FDAꢀapproved or investigational
drugs, they showed a general low toxicity at 5 µM concentration. Most of them showed toxicity
only at higher concentrations (25 and 50 ꢁM). In particular, the most toxic subset was 3A,
featuring a primary amino group and a linker of three methylenes. As a general trend, we
experienced a lower neurotoxicity for tricyclic systems c (unsubstituted PTZ) and f (carbazole).
In addition, we noted that CGNs treated with 1Af, 2Aa, and 2Bc showed slight increased MTT
values compared to untreated cells. This might be indicative of increased cellular metabolic
activity.
In addition to neurotoxicity, the assessment of hepatotoxicity would be of critical importance
48
for the drugꢀlikeness of the library. Indeed, both TCA and PTZ drugs can induce liver toxicity.
Furthermore, aging, comorbidity and subsequent polytherapy significantly contribute to increase
the risk of pharmacological side effects (i.e., liver injury) and drug interactions in many AD
4
9
patients. Experiments were performed in human hepatoma cell line (HepG2) using those
compounds (28 out of the starting 36) that show no significant neurotoxicity after 24 h treatment
(CGNs viability > 50% at 25 µM). Hepatotoxic antiꢀAD drug tacrine (4 in Figure S1) was used
as positive control (Figure S2B). Thus, 1A(a-c), 1A(e-f), 1B(b-f), 2A(a-b), 2A(d-f), 2B(a-f),
3
Aa, 3Ac, 3Ae, 3Ba, 3Bc, and 3B(f-e) were incubated at 5, 25, and 50 ꢁM for 24 h. 1Aa, 1B(d-
ACS Paragon Plus Environment
13

ACS Chemical Neuroscience
Page 14 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
e), 2Ab, 2Ad, 2Bb, 2Bd, 2Bf, 3Aa, 3Bc, and 3Bf negatively affected hepatic viability (viability
<
60% at 50 µM), whereas no dramatic variation was found for 1A(b-c), 1A(e-f), 1B(b-c), 1Bf,
2Aa, 2Ae, 2Af, 2Ba, 2Bc, 2Be, 3Ac, 3Ae, 3Ba, and 3Be, as evaluated through MTT assay and
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
statistical analysis though oneꢀway ANOVA followed by Bonferroni’s postꢀhoc comparison test
48
(Figure S3). The obtained data for 1Ab are in agreement with its reported hepatotoxicity.
To confirm nonꢀtoxic compounds 1A(e-f), 1B(b-c), 1Bf, 2Aa, 2A(e-f), 2Ba, 2Bc, 2Be, 3Ba, and
Be (CGNs viability > 50% and HepG2 viability > 60% at 50 µM) and reference compound 1Ab
3
as true positives, a secondary (nonꢀmetabolic) viability assay was used to test neurotoxicity in
CGNs. To this aim, condensed (apoptotic) nuclei counting after 24 h treatment following
staining with Hoechst dye, which binds DNA, has been paralleled to the MTT protocol for
assessing cell number in combination with cellular function. The condensed or healthy/total
nuclei counting after Hoechst staining has been previously demonstrated to be a good approach
to test apoptosis in CGNs, by comparing this method with MTT and LDH assay, as well as
5
0
TUNEL staining. Gratifyingly, no toxicity was observed by all the tested compounds at 5 µM
concentration, except for CBZꢀderived tertiary amine 1Af, which showed a statistically
significant decrease (75%) of CGNs viability (Figure 4). Thus, these data further confirm the
selected chemotypes. In addition, to rule out any direct interference between PTZ and the MTT
salt, we have also checked that 2Bc with no cell control was not affecting the MTT assay (data
not shown).
ACS Paragon Plus Environment
14

Page 15 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Toxicity of selected compounds to differentiated CGNs after 24 h treatment. Results
are expressed as percentage of healthy nuclei on the total nuclei number after counting following
Hoechst staining and are the mean ± SE of at least 2 different experiments, each run in duplicate.
*p<0.05 compared to control conditions (0 ꢁM), Bonferroni’s test after ANOVA. Scale bar = 20
µm; white arrows point to condensed nuclei.
Taken these data together, 1A(e-f), 1B(b-c), 1Bf, 2Aa (desimipramine), 2A(e-f), 2Ba, 2Bc,
2Be, 3Ba, and 3Be, as the less neurotoxic and hepatotoxic derivatives, were progressed further
ACS Paragon Plus Environment
15

ACS Chemical Neuroscience
Page 16 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
into the screening cascade to assess their neuroprotection profile, in comparison with reference
drug 1Ab.
(ii) To this end, we tested the ability of 1Ab, 1A(e-f), 1B(b-c), 1Bf, 2Aa, 2A(e-f), 2Ba, 2Bc,
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Be, 3Ba, and 3Be to support survival of primary CGN cultures, in which oxidative stress and
+
neuronal death had been induced by serum and K withdrawal for 48 h, by following the method
5
1
developed by Miller et al with minor modifications. Since this insult leads to partial ATP
+
depletion, K loss, and ultimately apoptosis, it represents a useful model of neuronal injury, as
51
+
well as neuronal senescence. The apoptotic effect of serum/K deprivation in CGNs has been
5
1, 52
clearly demonstrated through many biochemical, molecular and genetic approaches
and
5
3, 54
MTT assay has been widely used to evaluate CGNs viability after this treatment.
As shown
in Figure 5, deprived CGNs showed significantly decreased viability, whereas the addition of
PTZ 1Bb and 2Bc at 1 and 5 ꢁM was able to rescue neurons in a doseꢀdependent manner.
Conversely, no doseꢀdependence was observed for compounds 1Ab, 1Ae, 1Bc, 2Aa, 2Ba, 2Be,
and 3Be; in spite of a weak neuroprotective effect at the lowest concentration (1 µM), reduced
CGN viability at higher concentrations was displayed. On the other hand, 1Af, 1Bf, 2Ae, 2Af,
and 2Ba showed no significant changes in cell viability at all tested concentrations. CBZꢀbased
compounds (subgroup f), carrying linkers of two (subgroup B) or three methylenes (subgroup
A), appeared ineffective in rescuing neurons in this model, whereas only marginal
neuroprotection was elicited by dibenzoazepine (subgroup a) and 10,11ꢀdihydroꢀdibenzazepine
(subgroup e) derivatives, bearing both linkers of two and three methylene units. Thus, a linker
length of two methylenes (subgroup B) and the presence of a PTZ (both unsubstituted c and 2ꢀ
chloroꢀsubstituted b) as in 1Bb and 2Bc seem important features for neuroprotection.
Remarkably, compounds 1Bb and 2Bc increased MTT values even in the absence of insult at 5
ACS Paragon Plus Environment
16

Page 17 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
µM (Figure 3), pointing to a possible effect related to an increased metabolic activity. Thus, the
+
possibility that serum starvation/K depletion doesn’t simply lower metabolic rate, which is then
“rescued” by 1Bb and 2Bc cannot be ruled out. On the other side, it should be noted that 1Af,
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
which similarly increased MTT viability in the neurotoxicity assay, displayed no sign of
+
neuroprotection on serum/K deprivation assay. All in all, the collected data are in agreement
2
6, 55
with the literature purporting the neuroprotective effects of PTZ scaffold.
ACS Paragon Plus Environment
17

ACS Chemical Neuroscience
Page 18 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
Figure 5. Neuroprotection of 1Ab, 1A(e-f), 1B(b-c), 1Bf, 2Aa, 2A(e-f), 2Ba, 2Bc, 2Be, 3Ba,
+
and 3Be at different concentrations (0ꢀ10 µM) on 48 h serum/K deprivationꢀinduced neuronal
+
death in CGNs. Black bars represent control condition, whereas white bars serum/K deprivation
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
in absence (0 ꢁM) or presence of increasing concentrations (1, 5 and 10 ꢁM) of the tested
compound. Results are the mean ± S.E. of at least 3 different experiments, each run in triplicate.
§
§§
§§
§
*
**p<0.001; **p<0.01; *p<0.05 compared to control conditions, p<0.001; p<0.01; p<0.05
+
compared to serum/K deprivation (0 ꢁM) conditions Bonferroni’s test after ANOVA.
(iii) In addition to identifying molecules capable of rescuing neuronal death associated with
+
serum/K deprivation, we looked for compounds capable of enhancing neuronal regeneration.
The ability to simultaneously stimulate neural progenitor proliferation and provide
5
6
neuroprotection has been proposed as a useful trait for AD drug candidates. To address the
neurogenic potential, the largest germinal region in the adult mammalian brain, the
subventricular zone (SVZ), has been selected as NS model system, since the rate of progenitor
proliferation has been differently affected by neurodegenerative diseases, making it a potential
57
target for therapeutic intervention. Furthermore, NS production from SVZ guarantees a higher
yield compared to other neurogenic areas and, therefore, a reduction on the use of animals, as
required by the Italian and EU law. Thus, NS were derived from the SVZ of adult mice and the
NS formation capacity of the tested compounds was evaluated by quantifying NS size and
5
8
number. In addition to the most neuroprotective compounds 1Bb and 2Bc, we tested reference
drugs TCA 2Aa and PTZ 1Ab, because of their known neurogenic activity (see above). We also
included CBZ derivative 2Af, which, similar to 2Aa, carries a linker of three methylenes and a
secondary amine, in order to assess the neurogenic potential of CBZ scaffold. Thus, NS
ACS Paragon Plus Environment
18

Page 19 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
proliferation assays were set up after 15 days (Figure S4) and then, NS cultures were plated and
after 30 min incubated for 4 days with 1Ab, 1Bb, 2Aa, 2Af, and 2Bc at 1, 2.5 and 5 ꢁM
concentrations. After 4 days, compounds 2Bc and 2Af (at 1 ꢁM) increased the average NS size
by 34% and 40%, compared to control (Figure 6A), while 1Ab at the same and higher
concentrations significantly reduces NS size. Conversely, no significant effect was observed for
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Aa and 1Bb. However, at 5 ꢁM concentration, all compounds, but 1Bb, decreased average NS
size. Thus, it appears that higher concentration of 1Ab, 2Aa, 2Af, and 2Bc negatively affects NS
formation because of toxic effects. Particularly, at 5 µM, 2Bc and 2Af induced a marked
reduction of average size down to 22% and 19%, respectively, whereas 1Bb showed an increase
of average NS size up to 122%.
ACS Paragon Plus Environment
19

ACS Chemical Neuroscience
Page 20 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
Figure 6. Effect of 1Ab, 1Bb, 2Aa, 2Af, and 2Bc on the proliferation and of mice primary NS
by analyzing size (A) and number (B). Each bar represents the percentage of control and is the
mean ± SE of at least 3 different experiments, each run in six replicates. ***p<0.001; **p<0.01;
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
*
p<0.05 compared to control conditions, Bonferroni’s test after ANOVA.
Regarding the average NS number (Figure 6B), no differences were observed at 1 ꢁM
concentration in comparison to control, except for 1Bb, which decreased NS number of
approximately 30% at all tested concentrations. In contrast, at 5 ꢁM concentration, a higher
average number of NS with 2Bc (149%), 2Af (157%) and, to a lesser extent, 1Ab (119%) was
determined.
The collected results suggest that, at 1 µM, 2Bc and 2Af have higher NS formation capacity than
59, 60
2
Aa and 1Ab, for which a regenerative profile has been claimed.
(iv) Despite the high potential as BBB permeable compounds (see also their compliance with
FAFꢀdrugs3 prediction in SI), we confirmed the CNS penetration ability of 2Af and 2Bc in a
61
PAMPAꢀBBB test. Starting from a validation study on known CNS permeable drugs and
comparing the obtained permeability values with the corresponding literature data, we verified
their PAMPAꢀBBB permeability (Table S1).
(v) Encouraged by the promising profile of 2Bc in terms of both neuroprotective and
neuroregenerative properties, we sought to evaluate its therapeutic potential. We tested 2Bc in
established cellular models of neurodegenerative diseases, mimicking distinct aspects of
neurodegeneration (i.e., Aβꢀ, 6ꢀhydroxydopamine (6ꢀOHDA)ꢀ, and glutamateꢀinduced
4
4
neurotoxicity and neuronal death), in comparison with AD and PD approved drugs. This would
ACS Paragon Plus Environment
20

Page 21 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
allow verifying whether the identified compound, with the preꢀdefined neuroprotective and
neuroregenerative profile, could outperform currently available drugs. Even in this case, CGNs
have been used, as that the neurotoxic effect of Aβ, 6ꢀOHDA and glutamate on this neuronal
6
2, 63
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
model is well known and characterized since many years.
pretreated (2 h) with serial concentrations of 2Bc or AD drug galantamine (3 in Figure S1) (5–25
M) in the presence of toxic Aβ1–42 (25 ꢁM) for 24 h and viability was evaluated through MTT
To this end, primary CGNs were
ꢁ
assay (Figure 7A). As expected, viability of CGNs exposed to Aβ was reduced to 80% compared
to the control. Notably, a 2hꢀpretreatment of CGNs with 2Bc (5–25 ꢁM) reversed Aβꢀinduced
neuronal death with a maximal effect obtained at 5 ꢁM (viability increased to 129%).
Conversely, such effect was not observed for 3 at the same concentration. This result was
confirmed by counting of condensed/total nuclei after Hoechst staining (Figure 7B, C), where
2
Bc (5 ꢁM) was able to significantly reverse the neurotoxic effect of Aβ toxicity; even in this
case, no neuroprotective effect was observed for 3. Interestingly, 2Bc was not able to block Aβꢀ
induced toxicity upon coꢀtreatment with Aβ (data not shown).
44
The protective effect of 2Bc was further assessed in the 6ꢀOHDA cellular model of PD.
Primary CGNs were pretreated (2 h) with serial concentrations of 2Bc or antiꢀPD dopaminergic
drug apomorphine (5 in Figure S1) (5–25 ꢁM) in the presence of toxic 6ꢀOHDA (20 ꢁM) for 24
h and viability was evaluated through MTT assay. CGNs were found to be sensitive to 6ꢀOHDA,
which significantly decreased cell viability by 20% (Figure 7D). Again, preꢀtreatment (2 h) with
2
Bc prevented neuronal death in a concentrationꢀdependent inverse manner, with a maximal
effect observed at 5 µM. Gratifyingly, 2Bc showed neuroprotective effects comparable to those
elicited by apomorphine (5 in Figure S1) at 5 µM, whereas, 2Bc turned out to be more effective
at 10 and 25 µM. Also in this case, coꢀtreatment experiments failed to rescue 6ꢀOHDAꢀinduced
ACS Paragon Plus Environment
21

ACS Chemical Neuroscience
Page 22 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
neuronal death (data not shown). Counting of condensed/total nuclei after Hoechst staining
confirmed this result (Figure 7E, F); in fact 2Bc at 5 ꢁM concentration was able to reverse the
neurotoxic effect of 6ꢀOHDA, while no significant neuroprotective effect was observed for 5.
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Neuroprotective effects of 2Bc on Aβꢀ and 6ꢀOHDAꢀinduced neurotoxicity in CGNs.
Cells were pretreated (2 h) at various concentrations of the tested compounds (from 5 to 25 µM
ACS Paragon Plus Environment
22

Page 23 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
for MTT assay and 5 µM for Hoechst staining and cell counting) in the presence of Aβ1–42 (25
ꢁ
M) (A, B, C) or 6ꢀOHDA (20 ꢁM) (D, E, F) for 24 h, and viability was evaluated through MTT
assay (A, D) and cell counting results (B, E) after Hoechst staining (C, F). MTT results (A, D)
are expressed as percentage of controls and each bar is the mean ± S.E. of 2 different
experiments, each run in quadruplicate; cell counting results (B, E) are expressed as percentage
of apoptotic (condensed) on the total nuclei number after Hoechst staining and are the mean ± SE
of 2 different experiments, each run in duplicate. ***p<0.01, **p<0.05, *p<0.1 compared to
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
###
##
control conditions; p<0.01; p<0.05 compared to Aβ or 6ꢀOHDA conditions, Student's t test.
Scale bar = 20 µm.
To have a wider picture of the neuroprotective capability of 2Bc, we considered excitotoxicity
4
4
as another valuable model to mimic neurodegeneration. We investigated and compared the
neuroprotection of 2Bc and memantine (6 in Figure S1), as the marketed NMDA receptor
antagonist, against glutamateꢀinduced excitotoxicity. CGNs were pretreated with 2Bc for 2 h
before adding 100 ꢁM glutamate/10 µM glycine insult. As shown in Figure S5, 2 h pretreatment
with 2Bc at all tested concentrations resulted ineffective in rescuing CGNs against glutamateꢀ
induced excitotoxicity, whereas 5 was protective in a doseꢀdependent manner.
Interestingly, in this case, both coꢀ and preꢀ treatments with 2Bc were ineffective in reversing
glutamateꢀinduced neurotoxicity. This might indicate that the underlying mechanisms to the
observed neuroprotective effects of 2Bc is specific and does not involve the glutamatergic
system.
Interestingly, the neuroprotective effect of 2Bc on Aβꢀ and 6ꢀOHDAꢀinduced neurodegeneration
in CGNs is completely due to a neuronal effect, as this neuronal culture is almost completely
ACS Paragon Plus Environment
23

ACS Chemical Neuroscience
Page 24 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
pure, with no glial cells growing. However, in in vivo condition, glial cells, especially microglia,
strongly affect neuronal survival/death through their activation. In fact, microglia is among the
most potent modulators of CNS repair/regeneration. Particularly, microglial polarization, i.e. the
shift from a neurotoxic (M1) to a neuroprotective phenotype (M2), is crucial to make the brain
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
64
microenvironment permissive to neuroregeneration. To this end, we tested whether 2Bc could
trigger microglia M1/M2 switch by evaluating the inducible nitric oxide synthase (iNOS), and
triggering receptor expressed on myeloid cells 2 (TREM2) profiles expression and nitrite
65
production upon inflammatory stimuli with lipopolysaccharide (LPS) (100 ng/mL) (Figure 8).
In LPSꢀtreated primary cultured glial cells, we observed a substantial induction of the M1 marker
iNOS, which was doseꢀdependently reduced by treatment with 2Bc. However, 2Bc did not
counteract nitrite production. Encouragingly, in the case of classical M2 marker, 2Bc did not
affect TREM2 expression at all the tested concentrations. Given such a profile, 2Bc (at 5 ꢁM)
resulted as a peculiar modulator of gliaꢀderived inflammation able, in principle, to revert the proꢀ
inflammatory phenotype and sustain and favor the endogenous neuroregenerative processes.
Therefore, 2Bc could determine neuroprotection either directly, through its effect on neuronal
survival, or indirectly through microglial phenotypic shift. This result is of great therapeutic
relevance as abnormal or prolonged glial cell activation is one of the main contributors to the
66
pathogenesis and progression of neurodegenerative disorders.
ACS Paragon Plus Environment
24

Page 25 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. Immunomodulatory effects of 2Bc in glial cells were evaluated through Western blot
analysis of iNOS and TREM2 expression as microglial polarization markers (A), and relative
densitometries of iNOS (B) and TREM2 (D) expressions, as well as the indirect extent of
released NO trough nitrite measurement in the medium (C). GAPDH was used as loading
control. Densitometric results are expressed as percentage of LPS only and are the mean ± SE of
three different experiments. ***p<0.001 compared to LPS condition; Bonferroni’s test after
ANOVA.
ACS Paragon Plus Environment
25

ACS Chemical Neuroscience
Page 26 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
Conclusions
Based on the reported neuroprotective/neuroregenerative properties of psychotropic drugs, a
new hypothesisꢀdriven chemical library of 36 tricyclic arylalkylamines has been designed.
Thanks to the development of a new versatile procedure, the library has been synthesized and
investigated in a phenotypic screening pipeline aimed at selecting entries displaying
simultaneously neuroprotective and neuroregenerative effects, while devoid of neurotoxicity and
hepatotoxicity. By using primary neurons and NS cultures, we have identified PTZ 2Bc, which
functions as a neuroprotective and neuroregenerative phenotypic hit. Remarkably, it outperforms
drug 2Aa, a wellꢀknown neurogenic agent, and displays immunomodulatory effects on
microglia. Importantly, modulation of the microglial proinflammatoryꢀantiꢀinflammatory balance
by 2Bc provides an additional effect, which could be useful to further improve neuroprotection
and neuroregeneration. Furthermore, its higher efficacy with respect to currently available drugs
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
3 and 5) in two cell models of AD and PD seems to corroborate the starting idea that targeting
neuroprotection/neuroregeneration rather than neurodegeneration might be a reasonable strategy
for neurodegenerative drug discovery. In addition, these preliminary findings appear to validate
the tricyclic arylalkylamine structure, not only for the wellꢀknown psychotropic effects, but as a
privileged chemotype against neurodegeneration. Future studies aim to confirm the
neuroprotective and neuroregenerative potential of 2Bc in clinically relevant human cell models
of neurodegeneration (i.e. patientꢀderived induced pluripotent stem cells), which would allow to
capture those neuropathological phenotypes of neuronal and microglial impairments, highly
suited for advancing compound 2Bc to the next hitꢀtoꢀlead stage. In parallel, to validate the PTZ
scaffold as privileged neuroprotective structure for the development of new small molecules
ACS Paragon Plus Environment
26

Page 27 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
against neurodegenerative diseases, novel PTZꢀbased compounds endowed with additional and
beneficial properties are being developed and will be reported in due course.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Methods
Chemistry
General chemical methods
All the commercial available reagents and solvents were used purchased from SigmaꢀAldrich
(
Italy), TCI chemicals (Europe) and Alfa Aesar (Germany) without purification. Column
chromatography purifications were performed under "flash conditions" using SigmaꢀAldrich
silica gel grade 9385, 60 Å, 230ꢀ400 mesh. CEM Discover SP focused microwave reactor was
used for microwaveꢀmediated reactions. Thin layer chromatography (TLC) separations were
performed on 0.20 mm silica gel 60 F254 plates (Merck, Germany), which were visualized by
exposure to ultraviolet light (254 and 366 nm), iodine and potassium permanganate stains.
Reactions involving generation or consumption of amine were visualized by using bromocresol
green (0.04% in EtOH made blue by NaOH) following heating of the plate. Compounds were
named following IUPAC rules as applied by ChemDraw Professional 15.0. NMR experiments
1
13
were run on Varian VXR 400 (400 MHz for H; 100 MHz for C). Spectra were acquired at 300
1
13
K, using CDCl and CD OD as solvents. Chemical shifts for H and C spectra were recorded
3
3
parts per million (ppm) using the residual nonꢀdeuterated solvent as the internal standard. Data
are reported as follows: chemical shift (ppm), multiplicity (indicated as: s, singlet; br s, broad
singlet; exch, exchangeable proton with D O; d, doublet; t, triplet; q, quartet; p, pentet; m,
2
multiplet and combinations thereof), coupling constants (J) in Hertz (Hz) and integrated
ACS Paragon Plus Environment
27

ACS Chemical Neuroscience
Page 28 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
intensity. The molecular mass was determined by lowꢀresolution mass spectra ESIꢀMS, recorded
on a Waters ZQ 4000 apparatus. Target compounds were converted into the corresponding
hydrochloride salts and were assessed for purity (≥95%) by analytical RPꢀHPLC (see Supporting
Information and Table S2 for experimental details). All compounds, with the exception of 2Bd
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
33ꢀ42
and 3Bd, were previously synthesized using different methodologies for other applications.
General procedure for the preparation of tertiary amine (subset 1). Method A: To a
suspension of primary amines 3 (1 mmol) in H O (4 mL), 37% aq. formaldehyde (10 mmol) and
2
9
0% aq. formic acid (5 mmol) solutions were sequentially added in a pressureꢀtight tube cooled
to 0ꢂ°C. After warming up at rt, the mixture was submitted to MW irradiation for 1 h at 100 °C,
with an irradiation power of 50 W. The reaction mixture was neutralized with NaHCO , followed
3
by extraction with DCM (3 x 15 mL). The collected organic extracts were dried over Na SO ,
2
4
and evaporated in vacuum to give crude products, which were purified by column
chromatography on silica gel with an eluent of DCM/MeOH/33% aq. NH solution (9.5: 0.5:
3
0
.05) to give 1Aa, 1Ae-f and 1Ba, 1Be-f. Method B: To a suspension of primary amines 3 (1
mmol) in H O (4 mL), 37% aq. formaldehyde (10 mmol) and 90% aq. formic acid (5 mmol)
2
solutions were sequentially added in a pressure tube cooled to 0ꢂ°C. After 5 min, the reaction was
allowed to reach rt, and then was warmed at 80 °C for 8 h. The reaction mixture was neutralized
with NaHCO , followed by extraction with DCM (3 x 15 mL). The collected organic extracts
3
were dried over Na SO , and evaporated in vacuum to give crude products, which were purified
2
4
by column chromatography on silica gel with an eluent of DCM/MeOH/33% aq. NH solution
3
(9.5: 0.5: 0.05) to give 1Ab-d and 1Bb-d.
ACS Paragon Plus Environment
28

Page 29 of 58
ACS Chemical Neuroscience
1
2
3
4
5
6
7
8
9
3
-(10,11-Dihydro-5H-dibenzo[b,f]azepin-5-yl)-N,N-dimethylpropan-1-amine (1Aa).
The title compound 1Aa was obtained as pale yellow oil by using Method A. Yield 61%.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
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
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
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
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
H NMR (400 MHz, CDCl ) δ 7.16 – 6.93 (m, 6H), 6.97 – 6.87 (m, 2H), 3.76 (t, J = 6.9 Hz, 3H),
3
13
3
.15 (s, 4H), 2.31 (t, J = 7.8 Hz, 2H), 2.14 (s, 6H), 1.72 (p, J = 7.1 Hz, 2H). C NMR (100 MHz,
CDCl ) δ 148.28, 134.19, 129.74, 126.32, 122.36, 119.97, 57.63, 48.82, 45.45, 32.20, 26.11.
3
+
+
MS(ESI): m/z C H N ; calcd. [M]: 280.42, found [M + H] : 281, [M + Na] : 303.
1
9
24
2
3
-(2-Chloro-10H-phenothiazin-10-yl)-N,N-dimethylpropan-1-amine (1Ab).
The title compound 1Ab was obtained as yellow oil by using Method B. Yield 66%.
1
H NMR (400 MHz, CDCl ) δ 7.20 – 7.11 (m, 2H), 7.03 – 7.01 (m, 1H), 6.95 – 6.87 (m, 4H),
3
13
3
.88 (t, J = 7.7 Hz, 2H), 2.40 (t, J = 7.0 Hz, 2H), 2.22 (s, 6H), 1.97 – 1.90 (m, 2H). C NMR
(100 MHz, CDCl ) δ 146.52, 144.51, 133.21, 127.84, 127.47, 124.82, 123.53, 122.83, 122.20,
3
115.85, 115.78, 57.00, 45.56, 45.45, 25.06. MS(ESI): m/z C H ClN S; calcd. [M]: 318.86,
1
7
19
2
+
+
found [M + H] : 319, [M + Na] : 341.
N,N-Dimethyl-3-(10H-phenothiazin-10-yl)propan-1-amine (1Ac).
The title compound 1Ac was obtained as yellow oil by using Method B. Yield 74%.
1
H NMR (400 MHz, CDCl ) δ 7.15 – 7.11 (m, 4H), 6.91 – 6.86 (m, 4H), 3.92 – 3.88 (m, 2H),
3
1
3
2
.40 (t, J = 7.1 Hz, 2H), 2.19 (s, 6H), 1.92 (p, J = 7.0 Hz, 2H). C NMR (100 MHz, CDCl ) δ
3
1
50.35, 145.24, 127.42, 127.17, 125.13, 122.38, 115.49, 57.14, 45.54, 45.36, 25.21. MS(ESI):
+
+
m/z C H N S; calcd. [M]: 284.42, found [M + H] : 285, [M + Na] : 307.
17
20
2
ACS Paragon Plus Environment
29