Polypharmacological profile of 1,2-dihydro-2-oxo-pyridine-3-carboxamides in the endocannabinoid system

Article abstract of DOI:10.1016/j.ejmech.2018.05.019

The endocannabinoid system (ECS) represents one of the major neuromodulatory systems involved in different physiological and pathological processes. Multi-target compounds exert their activities by acting via multiple mechanisms of action and represent a promising pharmacological modulation of the ECS. In this work we report 4-substituted and 4,5-disubstituted 1,2-dihydro-2-oxo-pyridine-3-carboxamide derivatives with a broad spectrum of affinity and functional activity towards both cannabinoid receptors and additional effects on the main components of the ECS. In particular compound B3 showed high affinity for CB1R (K_i = 23.1 nM, partial agonist) and CB2R (K_i = 6.9 nM, inverse agonist) and also significant inhibitory activity (IC₅₀ = 70 nM) on FAAH with moderate inhibition of ABHD12 (IC₅₀ = 2.5 μΜ). Compounds B4, B5 and B6 that act as full agonists at CB1R and as partial agonists (B5 and B6) or antagonist (B4) at CB2R, exhibited an additional multi-target property by inhibiting anandamide uptake with sub-micromolar IC₅₀ values (0.28–0.62 μΜ). The best derivatives showed cytotoxic activity on U937 lymphoblastoid cells. Finally, molecular docking analysis carried out on the three-dimensional structures of CB1R and CB2R and of FAAH allowed to rationalize the structure-activity relationships of this series of compounds.

Full text of DOI:10.1016/j.ejmech.2018.05.019

Accepted Manuscript
Polypharmacological profile of 1,2-dihydro-2-oxo-pyridine-3-carboxamides in the
endocannabinoid system
Andrea Chicca, Chiara Arena, Simone Bertini, Francesca Gado, Elena Ciaglia, Mario
Abate, Maria Digiacomo, Margherita Lapillo, Giulio Poli, Maurizio Bifulco, Marco
Macchia, Tiziano Tuccinardi, Jürg Gertsch, Clementina Manera
PII:
S0223-5234(18)30423-9
10.1016/j.ejmech.2018.05.019
EJMECH 10430
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 19 January 2018
Revised Date: 11 May 2018
Accepted Date: 12 May 2018
Please cite this article as: A. Chicca, C. Arena, S. Bertini, F. Gado, E. Ciaglia, M. Abate, M. Digiacomo,
M. Lapillo, G. Poli, M. Bifulco, M. Macchia, T. Tuccinardi, Jü. Gertsch, C. Manera, Polypharmacological
profile of 1,2-dihydro-2-oxo-pyridine-3-carboxamides in the endocannabinoid system, European Journal
of Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.05.019.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Polypharmacological Profile of 1,2-Dihydro-2-Oxo-Pyridine-3-Carboxamides
in the Endocannabinoid System
b,1
,1
d
d
a
a
a
Andrea Chicca, Chiara Arena, Simone Bertini, Francesca Gado, Elena Ciaglia, Mario Abate,
e
c
a
a
a
Maria Digiacomo, Margherita Lapillo, Giulio Poli, Maurizio Bifulco, Marco Macchia, Tiziano
,
f
b
,*
a
a
Tuccinardi, Jürg Gertsch, Clementina Manera
a
b
Department of Pharmacy, University of Pisa, 56126 Pisa. Institute of Biochemistry and
Molecular Medicine, NCCR TransCure, University of Bern, Bühlstrasse 28, CH-3012 Bern,
c
Switzerland. Department of Molecular Medicine and Medical Biotechnology, University of Naples
d
Federico II, 80131 Naples, Italy. Department of Medicine, Surgery and Dentistry "Scuola Medica
e
Salernitana", University of Salerno, 84081 Baronissi, Salerno, Italy Department of Biotechnology,
g
Chemistry and Pharmacy, University of Siena, via Aldo Moro 2, 53019 Siena, Italy; Sbarro
Institute for Cancer Research and Molecular Medicine, Center for Biotechnology, College of
Science and Technology, Temple University, Philadelphia, PA, USA.
Corresponding Author
*Clementina Manera: Phone: +39 050 2219548. Fax: +39 050 2210680. e-mail:
clementina.manera@farm.unipi.it.
ORCID
Clementina Manera: 0000-0002-7379-5743
1
Author Contributions
A. C. and C. A. contributed equally.
Notes
The authors declare no competing financial interest.
RECEIVED DATE (to be automatically inserted after your manuscript is accepted if
required according to the journal that you are submitting your paper to)
1

ACCEPTED MANUSCRIPT
Highlights
•
•
•
We synthesized 1,2-dihydropyridine-2-oxo-3-carboxamide with multi-activity on ECS
Some derivatives showed cytotoxic activity on U937 lymphoblastoid cell
Molecular docking on 3D structures of CBRs and of FAAH suggested their binding mode
2

ACCEPTED MANUSCRIPT
Graphical abstract
CH
N
3
O
O
ABHD12
ABHD6
CH
3
O
O
R
1
N
H
R
1
N
H
CB1
receptor
CB2
receptor
CB1
receptor
CB2
receptor
N
F
F
AEA
uptake
FAAH
3

ACCEPTED MANUSCRIPT
ABSTRACT
The endocannabinoid system (ECS) represents one of the major neuromodulatory systems involved
in different physiological and pathological processes. Multi-target compounds exert their activities
by acting via multiple mechanisms of action and represent a promising pharmacological modulation
of the ECS. In this work we report 4-substituted and 4,5-disubstituted 1,2-dihydro-2-oxo-pyridine-
3-carboxamide derivatives with a broad spectrum of affinity and functional activity towards both
cannabinoid receptors and additional effects on the main components of the ECS. In particular
compound B3 showed high affinity for CB1R (K = 23.1 nM, partial agonist) and CB2R (K = 6.9
i
i
nM, inverse agonist) and also significant inhibitory activity (IC50 = 70 nM) on FAAH with
moderate inhibition of ABHD12 (IC50 = 2.5 µΜ). Compounds B4, B5 and B6 that act as full
agonists at CB1R and as partial agonists (B5 and B6) or antagonist (B4) at CB2R, exhibited an
additional multi-target property by inhibiting anandamide uptake with sub-micromolar IC values
5
0
(0.28−0.62 µΜ). The best derivatives showed cytotoxic activity on U937 lymphoblastoid cells.
Finally, molecular docking analysis carried out on the three-dimensional structures of CB1R and
CB2R and of FAAH allowed to rationalize the structure-activity relationships of this series of
compounds.
Keywords: Endocannabinoid system; Cannabinoid receptors; Polypharmacology; Molecular
docking; U251MG glioblastoma cell line; U937 lymphoblastoid cells;
4

ACCEPTED MANUSCRIPT
1
. Introduction
The endocannabinoid system (ECS) consists of two G protein-coupled receptors, the type-1
CB1R) and type-2 (CB2R) cannabinoid receptor, a class of lipids mediators called
(
endocannabinoids (ECs) which are produced on-demand from membrane phospholipid precursors
and several enzymes involved in the biosynthesis and degradation of ECs. So far, the most well-
studied ECs are N-arachidonoylethanolamine (anandamide, AEA), which is a member of the large
family of N-acylethanolamines (NAEs) and 2-arachidonoylglycerol (2-AG) that belongs to the
monoacylglycerol family [1]. The biological activities of these lipid mediators are terminated upon
cellular re-uptake and subsequent metabolism. The main endocannabinoid degrading enzymes are
fatty acid amide hydrolase (FAAH) [2] for AEA and monoacylglycerol lipase (MAGL) [3] for 2-
AG. Two serine hydrolases α,β-hydrolase domain-6 (ABHD6) and α,β-hydrolase domain-12
(ABHD12) were recently identified as complementary 2-AG-degrading enzymes in the brain [4,5].
The ECS is involved in different physiological and pathological processes including cancer,
appetite, memory, neuropathic and inflammatory pain, obesity, neuroprotection and
neurodegenerative diseases [6]. Numerous CB1R and CB2R ligands have been developed and
tested in vitro and in vivo in the context of drug discovery, but none of them reached an advanced
stage of clinical development due to central or peripheral side effects that are mainly associated to
chronic activation or blocking of CB1R [7-9]. Alternative strategies to modulate the ECS are
focused on the increase of AEA and 2-AG levels by inhibiting their enzymatic degradation thus
preserving the beneficial effects derived from the direct activation of cannabinoid receptors (CBRs)
by their endogenous ligands [10]. These efforts led to the generation of many potent and selective
inhibitors of the main enzymes involved in AEA and 2-AG metabolism (FAAH, MAGL, ABHDs)
[
11-13]. However in preclinical animal models, the repeated administrations of covalent MAGL
inhibitors rapidly induced CB1R desensitization in the brain [14], while chronic ablation of FAAH
5

ACCEPTED MANUSCRIPT
activity may be associated to potential side effects on the cardiovascular and metabolic system
[
15,16].
Recent findings highlighted the emerging role of each target of the ECS to control symptoms and
progression of several diseases [17,18]. Multi-target compounds exert pro-cannabinoid activities by
means of more than one mechanism of action [19]. Indeed, it is reasonable to assume that the
simultaneous modulation of more targets within the ECS may offer a safer and more effective
pharmacological strategy to tackle the complexity of the ECS [19]. However, a proper evaluation of
the different multi-target approaches as compared to the classic modulation of a single target needs
to be further investigated. Therefore, the development of different multi-target compounds could get
a foothold to better explore the polypharmacology of the ECS and to assess its potential
pharmacological advantages for the treatment of different pathologies.
Previously, in a research program aimed at obtaining CB2R selective ligands [20-23], a series of
1,2-dihydro-2-oxo-pyridine-3-carboxamide derivatives with general structure A, (Figure 1) was
developed [24,25].
CH3
O
O
R1
N
H
N
R2
B
O
O
R1
R3
N
H
N
R2
O
R1
A
N
H
N
O
R2
C
Figure 1. General structure of 1,2-dihydro-2-oxo-pyridine-3-carboxamide derivatives A, B and C.
6

ACCEPTED MANUSCRIPT
These compounds showed a broad spectrum of affinity for both CB1R and CB2R. Furthermore,
the functional activity of this series is controlled by the presence of a substituent at the position 5 of
the 2-oxo-pyridine nucleus [25]. The significant affinity and selectivity towards CB2R displayed by
some derivatives of A make this new scaffold useful as initial building block for designing new
CB2R ligands. For this reason and to further investigate the structure-activity relationships (SAR)
of this series of compounds, in this study the central scaffold A was modified by insertion of a
methyl group (B1-B4, Figure 1) or an aryl moiety (compounds C1-C6, Figure 1) at the 4-position
of the 1,2-dihydro-2-oxo-pyridine ring to obtain 4-substituted or 4,5-disubstituted-1,2-dihydro-2-
oxo-pyridine-3-carboxamide derivatives (Table 1). The new compounds are characterized by the
presence of a p-fluorobenzyl moiety at the N-1 position of the 1,2-dihydro-2-oxo-pyridine ring (R1)
and a N-cycloheptyl carboxamide group at the 3-position that were chosen on the basis of the best
results obtained with the previous series. Furthermore, the substituent in the N-1 position was
shifted to the oxygen in position 2 of the heterocyclic nucleus (B5 and B6, Table 1).
The new 1,2-dihydro-2-oxo-pyridine-3-carboxamide derivatives B and C were evaluated for their
binding affinities (K values) for both CB1R and CB2R. Subsequently, the compounds which
i
showed the most potent CBRs affinity values were also investigated for their functional activity
3
5
using a [ S]GTPγS assay. All the synthesized compounds were tested on the main targets of the
ECS (MAGL, FAAH, ABHDs and AEA cellular uptake). The best compounds were also tested for
their capability to inhibit viability of U251MG glioblastoma cell line and that of U937
lymphoblastoid cells. Finally, to rationalize the experimentally observed SAR, a molecular docking
analysis of the synthetized compounds was carried out into the three-dimensional structure of CB1R
and CB2R and of FAAH.
7

ACCEPTED MANUSCRIPT
2. Results and discussion
2.1 Chemistry
The synthesis of 1,2-dihydro-2-oxopyridine derivatives B1-B4 and C1-C6 and 2-substituted
pyridines B5 and B6 was accomplished as depicted in Schemes 1-4. As reported in Scheme 1, 1,2-
dihydro-4-methyl-2-oxo-pyridine-3-carbonitrile (1) synthesized as reported in literature [21] from
4,4-dimethoxyl-2-butanone with cyanoacetamide, was hydrolyzed to the corresponding carboxylic
acid derivative 2 by heating at 120 °C with 50% H SO (w/v). Treatment of 2 with concentrated
2
4
H SO in MeOH, carried out under microwave irradiation at 100 °C for 50 minutes or under
2
4
conventional heating at 85 °C for 12 h, gave the corresponding methyl ester 3. The treatment of 3
with cycloheptylamine under microwave irradiation at 130 °C for 30 minutes or under conventional
heating at 150 °C for 48 h provided the carboxamide derivative 4. Finally, compound 4 was treated
with CsF in anhydrous DMF and then with p-fluorobenzyl chloride to afford the desired N-
alkylated derivative B1 together to the O-substituted derivative B5, analogously to previously
reported for compounds A [24,25]. The two structural isomers were purified by flash
chromatography.
Scheme 1. Synthesis of the 1,2-dihydro-4-methyl-2-oxo-pyridine B1 and 2-substituted pyridine B5.
8

ACCEPTED MANUSCRIPT
CH3
NH2
O
CN
O
O
OCH3
O
i
ii
OCH3
CN
OCH3
H N
2
NC
N
H
OCH3
1
iii
CH3
O
O
CH3
O
O
CH3
O
v
OH
N
H
OCH3
iv
N
H
N
O
N
H
H
4
3
2
vi
O
O
CH3
N
O
O
CH3
N
H
N
H
N
F
F
B1
B5
Reagents and conditions: (i) NH OAc, AcOH, toluene, 116 °C, 8 h; (ii) conc. H SO , 50 °C, 3 h;
4
2
4
(
iii) 50% H SO , 120 °C, 8 h; (iv) MeOH, conc. H SO ., MW, 110 °C, 50 min, (200 W, 100 psi) or
2 4 2 4
85 °C, 12 h; (v) cycloheptylamine, MW, 130 °C, 30 min, (140 W, 100 psi) or 150 °C, 48 h; (vi) 1)
CsF, dry DMF, rt, 1 h; 2) p-fluorobenzyl chloride, rt, 12 h.
The synthesis of the 4,5-disubstituted derivatives B2-B4 and B6 was carried out starting from
compound 4 (Scheme 2), which was treated with a solution of bromine in CHCl affording the
3
corresponding 5-bromo derivative 5. The desired N-alkylated derivative B2 was obtained by
treatment of 5 with p-fluorobenzyl chloride via a procedure similar to that previously used for
compound B1. Also in this case, the O-substituted derivative B6 was obtained. The 5-aryl
derivatives B3 and B4 were obtained starting from the 5-bromo derivative B2 via a Suzuki cross-
9

ACCEPTED MANUSCRIPT
coupling reaction with the suitable boronic acids in dry toluene by generating in situ Pd(PPh ) as
3
4
the catalyst and with anhydrous potassium carbonate as the base, refluxing at 100 °C, for 12 h.
Scheme 2. Synthesis of 4-methyl-1,2-dihydro-2-oxo-pyridine B2-B4 and 2-substituted pyridine B6.
CH3
O
O
CH3
O
O
Br
Br
CH3
O
O
CH3
O
O
N
H
N
Br
H
i
ii
N
N
+
H
H
N
N
N
H
N
H
4
5
B2
B6
F
F
iii
R
CH3
O
N
H
N
O
B3: R = H
B4: R = OCH
F
3
Reagents and conditions: (i) Br , CHCl , rt, 12 h; (ii) anhydrous DMF, CsF, rt, 1 h, p-
2
3
fluorobenzyl chloride, rt, 12 h; (iii) 1) anhydrous toluene, PPh , Pd(OAc) , rt, 15 min; 2) K CO ,
3
2
2
3
suitable arylboronic acid, 100 °C, 12 h.
Initially, the synthesis of 4-aryl substituted compounds C1-C6 (Table 1) was planned as
described in Scheme 3, starting from acetophenone or p-methoxyacetophenone which by
Knoevenagel condensation with ethyl cyanoacetate afforded the ethyl 2-cyano-3-arylbut-2-enoate
derivatives 6 or 7 respectively [27]. The reaction of compounds 6 or 7 with dimethylformamide-
dimethylacetal (DMF-DMA), at room temperature gave the enaminonitriles 8 and 9, respectively
[
27]. Finally, the cyclization of the enaminonitriles 8 or 9 with p-flurobenzylamine, at 90 °C led to
the desired 3-cyano-2-pyridones 10 or 11 (see Supporting Information for experimental details).
1
0

ACCEPTED MANUSCRIPT
Unfortunately, several attempts made to hydrolyze the 3-cyano-2-pyridones to the corresponding
carboxylic acid derivatives have not been successful.
Scheme 3. Attempt for the synthesis of 4-aryl-1,2-dihydro-2-oxo-pyridine.
NC
COOEt
NC
COOEt
O
O
i
ii
NC
NMe
2
OEt
NC
R
R
R
8
,9
6,7
O
iii
N
F
10,11
R
1
0: R = H
1
1: R = OCH3
O
HOOC
N
F
R
Reagents and conditions: (i) TiCl , pyridine, CH Cl , 0 °C to rt., 12 h; (ii) DMF-DMA, solvent-
4
2
2
free, 24 h; (iii) p-fluorobenzylamine, solvent-free, 90 °C, 2 h.
For this reason, we focused the attention on a recent procedure to obtain the 2-pyridone scaffold
with an aromatic substituent in position 4 [28]. As described in Scheme 4, N-alkylation of the
commercial available 2-hydroxynicotinic acid with p-fluorobenzyl bromide in presence of NaH,
followed by hydrolysis with 10% NaOH at reflux and then acidification, afforded the carboxylic
acid derivative 12. The condensation reaction of 12 with cycloheptylamine in presence of TBTU,
and N,N-diisopropylethyl amine (DIPEA), gave the desired carboxamide derivative 13. Compound
13 was treated with the proper Grignard’s reagent (phenylmagnesium bromide or p-
1
1

ACCEPTED MANUSCRIPT
methoxyphenylmagnesium bromide) using copper(I) iodide and lithium bromide in dry THF, at -40
°
C to obtain the 4-substituted 1,2-dihydro-2-oxo-pyridines C1 and C2. These compounds were then
treated with a solution of bromine in CHCl to afford the desired compounds C3 and C4. To
3
obtained compounds C5 and C6, derivatives C3 and C4 were submitted to a Suzuki cross-coupling
reaction with the suitable arylboronic acid, in dry toluene by generating in situ Pd(PPh ) as the
3
4
catalyst and with anhydrous potassium carbonate as the base, refluxing at 100 °C.
Scheme 4. Synthesis of 4-aryl-1,2-dihydro-2-oxo-pyridines C1-C6.
R1
O
O
O
OH
O
O
N
H
i
ii
iii
OH
N
O
N
H
N
O
N
OH
N
F
1
2
13
F
C1: R = H
1
C2: R = OCH
1
3
F
iv
R1
R1
R2
v
O
O
O
O
Br
N
N
H
H
N
N
C5: R = H, R = H
1
2
C3: R = H
1
C6: R = H. R = OCH
1
2
3
C4: R = OCH
1 3
F
F
Reagents and conditions: (i) 1) NaH, p-fluorobenzyl bromide, dry DMF, 50 °C, 12 h; 2) 10%
NaOH, reflux, 100 °C, 4 h; (ii) 1) anhydrous DMF, DIPEA, TBTU, 0 °C, 30 min, 2)
cycloheptylamine, 0 °C to rt, 12 h; (iii) CuI, LiBr, phenylmagnesium bromide or p-
1
2

ACCEPTED MANUSCRIPT
methoxyphenylmagnesium bromide, THF, -40 °C, 6 h; (iv) Br , CHCl , rt, 12 h; (v) 1) anhydrous
2
3
toluene, PPh , Pd(OAc) , rt, 15 min; 2) K CO , suitable arylboronic acid, 100 °C, 12 h.
3
2
2
3
2
2
.2. Biological evaluation
.2.1 CB1R and CB2R affinity
The binding affinities (K values) of target compounds were evaluated in competitive radioligand
i
3
displacement assays against [ H]CP-55,940 using membrane preparations obtained in-house from
stable transfected CHO-hCB and CHO-hCB cells. The results are summarized in Table 1.
1
2
Table 1. Data of 1,2-dihydro-2-oxopyridine- and pyridine-3-carboxamide derivatives B1-B6 and
a
C1-C6.
K value (mean ± SD,
i
IC (mean ± SD, µM)
50
nM)
MAG
L
ABHD1
2
AEA
uptake
Cmpd
B1
R1
R2
CB1R
CB2R
FAAH
ABHD6
H
CH3
11.7 ± 3.6
5.9 ± 1.5
8.8 ± 2.6
>10
>10
>10
1.8 ± 0.5
1
3

ACCEPTED MANUSCRIPT
B2
B3
B4
Br
CH3
CH3
CH3
2.3 ± 1.6
23.1 ± 4.2
30.7 ± 9.5
3.5 ± 4.6)
6.9 ± 3.8
7.9 ± 2.1
2.3 ± 0.9
0.07 ± 0.06
1.3 ± 1.1
>10
>10
>10
>10
>10
>10
2.1 ± 1.2 0.93 ± 0.33
phenyl
2.5 ± 1.1
>10
n.d.
0.28 ± 0.17
>10
p-OCH phenyl
3
C1
C2
C3
C4
C5
C6
H
H
Phenyl
>10000
>10000
>10000
>10000
>10000
>10000
187 ± 65
>10000
437 ± 117
>10000
>10000
>10000
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
>10
p-OCH phenyl
>10
>10
3
Br
Phenyl
3.0 ± 0.5
>10
>10
>10
Br
p-OCH phenyl
>10
>10
3
phenyl
Phenyl
Phenyl
CH3
>10
>10
>10
p-OCH phenyl
>10
>10
>10
3
B5
B6
H
131 ±26
113 ± 63
46.3 ± 19.2
606 ± 305
>10
>10
>10
>10
>10
>10
>10
0.58 ± 0.19
0.62 ± 0.12
Br
CH3
0.53 ± 0.21
a
Results are expressed as mean ± SD calculated from at least 3 experiments each performed in
triplicate.
The results reported in Table 1 indicate that the pyridine-3-carboxamide derivatives B1-B6, C1
and C3, showed nanomolar or submicromolar binding affinities to CB2R and/or CB1R, unlike their
analogs C2 and C4-C6. Generally, all ligands tend to show higher affinity for CB2R, with a
selectivity factor in the range of 2 to 54. This behavior is in agreement with that previously reported
for A derivatives [24,25]. The only exceptions were the 5-bromo-4-methyl-derivatives B2 and B6
that showed a K (CB1R)/K (CB2R) ratio of 0.65 and 0.18, respectively. (R)-WIN55,212 was used as
i
i
positive control for both CB1R and CB2R binding experiments and showed K value of 2.5 ± 0.9
i
nM for CB1R and 1.6 ± 0.5 nM for CB2R.
The compounds B possess a broad spectrum of affinity towards both CBR subtypes, with Ki
values ranging from 2.3 nM (B2) to 131 nM (B5) for CB1R and from 3.5 nM (B2) to 606 nM (B6)
for CB2R. The results showed that the insertion of a small substituent, like the methyl group at the
4-position of the 1,2-dihydro-2-oxo-pyridine ring (compounds B1-B4), does not alter the binding
1
4

ACCEPTED MANUSCRIPT
properties at both CBRs as also shown by the previous reported 5-monosubstituted A derivatives
[
25]. Indeed, these compounds generally showed K values in the low nanomolar range (Table 1).
i
On the contrary, the insertion of bulkier substituents, like a phenyl or a p-methoxyphenyl group
C1-C6), led to a significant decrease of the affinity towards both CBR subtypes independently of
(
the type of substituent at position 5 of the pyridine core. Moreover, when the substituent at the 1-
position was shifted to the oxygen at the 2-position of the pyridine ring (compound B5 and B6), the
binding affinity for both CBRs decreased by a factor 10 for the 4-methyl-2-substituted derivative
B5 as compared to the corresponding 1-substituted compound B1, and by a factor 50 for CB1R and
200 for CB2R in the case of 5-bromo-4-methyl-2-substituted derivative B6, as compared to the 5-
bromo-4-methyl-1-substituted derivative B2 (Table 1). These data are in agreement with the results
previously reported on the binding affinity of this chemical scaffold [24,25].
Altogether, the results obtained for these new compounds suggest that either the type or the
position of the substituent on the pyridine ring play a key role for the binding interactions to the
cannabinoid receptors.
2.2.2. CB1R and CB2R functional activity
Compounds that showed the best results in terms of binding properties on CBRs were tested for
3
5
their functional activity, using the [ S]GTPγS assay. The results are summarized in Table 2 and
showed that the new compounds work as full agonists at CB1R, with the exception of B3 which
behaved as partial agonist. Compounds B1, B2, B5 and B6 behaved as partial agonist at CB2R,
compound C3 as full agonist, compounds B3 and C1 acted as inverse agonists and compound B4 as
an antagonist. These data indicate that while the combination of substituents in positions 4 and 5 do
not modify the agonist behavior at CB1R, they have impacts on CB2R activation. In agreement
with what we previously observed with a series of 4-unsubstituted 1,2-dihydro-2-oxo-pyridine-3-
carboxamides, the type of substituent in position 5 (R1) dictates the functional behavior at CB2R
1
5

ACCEPTED MANUSCRIPT
generating agonism, antagonism and inverse agonism [25]. The presence of a methyl group in
position 4 (R2) combined with a bulky substituent in position 5 leads to antagonism (B4) or inverse
agonism (B3) at CB2R. The increase of the substituent size in position 5 in combination with a
phenyl group in position 4 led to a significant loss of CB2R binding activity (H, K = 187 nM; Br, K
i
i
=
437 nM; phenyl and p-methoxyphenyl, K > 10 µM). Interestingly, while the unsubstituted 4-
i
phenyl-derivative (C1) behaved as an inverse agonist, the 5-bromo-4-phenyl derivative (C3)
showed an opposite effect, thus behaving as CB2R full agonist (Table 2). These results suggest that
the nature of the substituent in position 5 of the pyridine ring is responsible for the functional switch
in accordance to what was previously reported [25], and that the combination with the substituents
in position 4 can also play a role in determining the activity of the ligand at CB2R.
CB2R pharmacology is quite versatile and both agonists and inverse agonists showed beneficial
effects in vitro and in vivo [29,30]. CB2Rs have a relatively high ligand-independent (constitutive)
activation which makes certain classic receptor antagonists/inverse agonists (e.g. AM630) behaving
as agonists and vice versa (e.g. AM1241) depending on the conditions [31,32]. Furthermore, many
of the CB2R ligands, including the endocannabinoids AEA and 2-AG behave as functional selective
(biased) ligands at CB2R, thus activating only certain signaling pathways, while being inactive or
even hindering other pathways [33]. In our experiments we measured the recruitment of G-proteins
35
(
[ S]GTPγS assay) as functional assay and compound B3 and B4 behaved as inverse agonist and
antagonist at CB2R, respectively. Additional studies should be performed to fully characterize the
receptor pharmacological profile of this new series of 4-substituted and 4,5-disubstituted-1,2-
dihydro-2-oxo-pyridine-3-carboxamide derivatives and may highlight different behaviors at other
CB1R- and CB2R-mediated signaling pathways (e.g. cAMP production, β-arrestin, ion channels).
1
6

ACCEPTED MANUSCRIPT
3
5
Table 2. Summary of functional activity ([ S]GTPγS assay) of compounds B1-6 and C1-6 at
CB1R and CB2R. Data are express as mean ± SD obtained from at least three experiments
performed in triplicate.
CB1R
CB2R
EC value
Emax
mean ± SD) (% of control)
27 ± 56
Full Ago)
98 ± 35
Full Ago)
31 ± 42
Part Ago)
05 ± 26
EC value
(mean ± SD)
Emax
(% of control)
128 ± 31
(Part Ago)
134 ± 28
(Part Ago)
39 ± 15
(Inv Ago)
99 ± 6
(Antago)
134 ± 35
50
50
Cmpd
B1
R1
H
R2
(
2
CH3
CH3
CH3
CH3
69 ± 15 nM
1.9 ± 2.5 nM
(
(
(
(
1
B2
Br
1.4 ± 1.5 nM
237 ± 44 nM
1.1 ± 1.8 nM
12 ± 10 nM
n.c.
1
B3
phenyl
2
B4
p-OCH phenyl
96 ± 38 nM
n.d.
3
Full Ago)
n.d.
C1
C2
C3
C4
C5
C6
B5
H
H
Phenyl
501 ± 143 nM
n.d.
(
Part Ago)
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
p-OCH phenyl
3
176 ± 31
Full Ago)
n.d.
Br
Phenyl
316 ± 96 nM
n.d.
(
Br
p-OCH phenyl
3
n.d.
n.d.
n.d.
n.d.
phenyl
Phenyl
Phenyl
CH3
p-OCH phenyl
3
2
18 ± 28
127 ± 25
(Part Ago)
H
229 ± 62 nM
31 ± 20 nM
(
Full Ago)
1
7

1
93 ± 30
125 ± 28
(Part Ago)
196 ± 21
(Full Ago)
n.d.
B6
CP55940
SR1
Br
-
C
A
H
C3 CEPTE18
D
2 ±
M
51
A
n
N
M
USCRIPT
687 ± 183 nM
(
Full Ago)
12 ± 33
Full Ago)
7 ± 19
2
-
-
-
5.5 ± 3.9 nM
5.5 ± 3.9 nM
n.d.
(
2
-
36 ± 12 nM
n.d.
(
Inv Ago)
n.d.
35 ± 12
AM630
-
167 ± 82 nM
(
Inv Ago)
a
35
In brackets the behavior on CB1R and CB2R is reported from the [ S]GTPγS assay: Part Ago:
partial agonist; Full Ago: full agonist; Inv Ago: inverse agonist; Antago: antagonist. n.d.: not
determined; n.c.: not calculated.
2.2.3. Polypharmacology in the ECS.
In order to further characterize the pharmacology of these new compounds, we tested their effects
on the main components of the ECS. Our results identified interesting polypharmacological
properties among the tested compounds. The p-fluorobenzyl N-1 substituted derivatives bearing a
methyl group in position 4 (B1-B4) inhibited the main AEA hydrolytic enzyme fatty acid amide
hydrolase (FAAH) (Table 1), with increasing potencies in correlation with the presence of a bulkier
substituent in position 5. Indeed, while the unsubstituted 4-methyl-deriativative B1 inhibited FAAH
with an IC value of 8.8 µM, the 4-methyl-5-phenyl derivative B3 blocked AEA hydrolysis with an
5
0
IC value of 70 nM (see Fig. 2). The 4-methyl derivatives where the p-fluorobenzyl group was
5
0
shifted to the oxygen in position 2 of the heterocyclic nucleus (B5 and B6) were completely inactive
on FAAH (IC > 10 µM) (Table 1).
5
0
In our small compound library, the potency of 1,2-dihydro-2-oxo-pyridine-3-carboxamides to
inhibit FAAH was associated with the combined presence of a methyl group in position 4, a bulky
substituent in position 5, and the p-fluorobenzyl group in the N-1 position of the 1,2-dihydro-2-oxo-
pyridine ring (Table 1, Fig. 2).
1
8

ACCEPTED MANUSCRIPT
O
R1
NH
N
O
1
1
2
0
8
6
4
2
0
F
H
Br
OCH3
G BB 13 8b GB
B
58b
2
G
B
B
4
66
B3
GB61
Potency
Figure 2. Impact of the substituent in position 5 to the potency for FAAH inhibition.
None of the compounds showed significant effects on the main 2-AG hydrolytic enzyme MAGL,
while B2, B3, B6, and C3 exhibited a moderate low micromolar inhibition of the minor 2-AG
hydrolytic enzymes ABDH6 and ABHD12. Compound C3 selectively inhibited ABHD6 without
affecting ABHD12, while B2 and B3 inhibited ABHD12 without exerting any effects on ABHD6
up to 10 µM. The 4-methyl-5-bromo derivative B6 resulted the most potent ABHD6 inhibitor (IC50
value of 530 nM). Interestingly, its analog B2 bearing the p-fluorobenzyl substituent at the N-1
position of the 1,2-dihydro-2-oxo-pyridine ring was inactive up to 10 µM. The unsubstituted 4-
methyl derivatives B1 and B5 were both inactive. Altogether, these data suggest that 1,2-dihydro-2-
oxo-pyridine-3-carboxamides are versatile molecules that can selectively inhibit the
endocannabinoid degrading enzymes FAAH, ABHD6 and ABHD12. Finally, we also investigated
the effects of our compounds on the inhibition of AEA cellular uptake. Compounds B1, B2 and B4
1
9

ACCEPTED MANUSCRIPT
inhibited AEA uptake with a moderate potency. Nonetheless, due to their significant activity as
FAAH inhibitors, a partial FAAH-dependent contribution in AEA uptake inhibition cannot be ruled
out. Indeed, due to the tight interplay between FAAH inhibition and AEA uptake [34-36], it is
impossible to undoubtedly differentiate between FAAH-dependent and FAAH-independent AEA
uptake inhibition in our assay format. Therefore, we have not tested the most potent FAAH
inhibitor B3 for AEA uptake. However, compounds B5 and B6 showed sub-micromolar potencies
in inhibiting AEA cell uptake without affecting FAAH up to 10 µM, thus representing potential
selective AEA uptake inhibitors as compared to FAAH.
Altogether, our results show that the 1,2-dihydro-2-oxo-pyridine-3-carboxamide is a flexible
scaffold that can be further exploited to design multi-target compounds in the ECS. Interestingly,
the combinations of substituents that generate CBRs ligands also lead to the interaction with
endocannabinoid degrading enzymes and/or AEA cell uptake. In agreement, compounds that were
inactive at CBRs did not exhibit any effect on the other targets (Table 1,2). Our results thus suggest
that the combination of a bulky substituent in position 4 with any substituent in position 5 have
detrimental effects for the interactions with ECS proteins measured here. The presence of a methyl
group in position 4 generated potent CB2R ligands poorly selective over CB1R, which also
inhibited FAAH (Table 1 and Fig. 3). This type of polypharmacology was recently reported for a
class of functionalized beta-caryophyllene derivatives [37]. Interestingly, the presence of a bromine
atom or phenyl group in position 5 led to the moderate inhibition of ABHD12 which was selective
over the other 2-AG hydrolytic enzymes ABHD6 and MAGL (Table 1 and Fig. 3). This
combination of pharmacological effects in vivo could lead to an increase of AEA levels combined
with a mild and cell-specific augmentation of 2-AG concentrations, which was shown to avoid
CB1R tolerance in brain [36,38,39]. Other multi-target properties were shown by the 5-bromo-4-
phenyl derivative C3, which behaved as moderately potent CB2R agonist and ABHD6 inhibitor
(Table 1,2). Interestingly, compound B5 and B6 bearing the p-fluorobenzyl on the oxygen in
2
0

ACCEPTED MANUSCRIPT
position 2 of the heterocyclic nucleus rather than on the N-1 of the 1,2-dihydro-2-oxo-pyridine ring
exhibited two additional multi-target properties. Beyond the moderate interaction with CB1R and
CB2R, B5 inhibited AEA uptake with a sub-micromolar IC value without affecting FAAH up to
5
0
10 µM, similarly to compound B6. The latter showed also a selective inhibition of ABHD6 without
affecting the other degrading enzymes up to 10 µM (Table 1, Fig. 3). The corresponding N-
substituted compounds B1 and B2 inhibited AEA uptake and FAAH with similar potencies. Despite
the small number of compounds tested, our data suggest that the shift of the p-fluorobenzyl group
from nitrogen to the oxygen in position 2 of the heterocyclic ring generates relatively potent AEA
uptake inhibitors which are at least 20-times more selective over FAAH.
O
O
O
O
R1
R1
NH
NH
N
N
H
F
F
3
H CO
Br
GB
H
Br
B6
G
B
B
4
66
GB
B
5
3
8b
B
6
2
1
GB3
8
B5
GB58a
ABHD12
ABHD6
CB1
receptor
CB2
receptor
CB1
receptor
CB2
receptor
CB1
receptor
CB2
receptor
CB1
receptor
CB2
receptor
AEA
uptake
AEA
uptake
FAAH
FAAH
Figure 3. Summary of the most interesting polypharmacology properties observed with 1,2-
dihydro-2-oxo-pyridine-3-carboxamides.
Overall, our study showed that the introduction of a small substituent as methyl in position 4 of
the 1,2-dihydro-2-oxo-pyridine ring does not significantly affect the binding affinity and functional
behavior at CB2R of 1,2-dihydro-2-oxo-pyridine-3-carboxamides as compared to the unsubstituted
2
1

ACCEPTED MANUSCRIPT
analogs [25]. However, the new molecules exhibited unexpected and interestingly
polypharmacology effects on endocannabinoid degrading enzymes and AEA cell uptake, suggesting
interesting SAR properties of this scaffold which need to be further investigated (Fig. 3).
2.2.4. Cell viability assay
The 4-methyl derivatives B1-B6, that showed the best results in terms of binding properties on
both CB1R and CB2R were tested to assess whether the observed activity on CBRs, would result
into their different biological behavior. The selective CB2R inverse agonist 4-phenyl substituted
compound C1 and the CB2R neutral antagonist N-Cycloheptyl-5-(4-methoxyphenyl)-1-(4-
fluorobenzyl)-pyridin-2(1H)-on-3-carboxamide (MC21) previously reported by us [25] were also
assessed. First, we tested the compounds for their capability to inhibit viability of U251MG
glioblastoma cell line, which mainly expresses CB1R [40], and that of U937 lymphoblastoid cells
where, on the contrary, CB2R is predominantly expressed [41]. To this end, U251MG and U937
cells were incubated with the compounds. After 48 hours, the viable cell number was determined by
MTT assay. Overall, the results showed that exposure to the compounds in a 0.6-40 µM
concentration range led to a significant reduction in the number of viable U937 cells (Fig. 4, lower
panel) but not of U251 cells for which a certain increase in cell proliferation was instead detected
(Fig. 4, upper panel) in response to B1, B2, B5 and B6 treatment. As U937 cells exhibited CB2R
but not CB1R, these results suggest that even though to a different extent, the cytotoxic activity of
all tested compounds was mediated through the CB2R but not CB1R.
2
2

ACCEPTED MANUSCRIPT
Figure 4. Effect of MC21, B1-B6 and C1 on cellular viability of human GBM (U251MG) and
human lymphoblastoid U937 cells. U251MG cells (upper panel) and U937 cells (lower panel) were
cultured for 48h in the presence of the indicated concentrations (0–40 µM) of the individual
substances, before the assessment of MTT assay. Results are expressed as means ± SD of five
independent experiments performed in triplicate and reported as percentage vs. the untreated control
Then, the collected data were analyzed for statistical significance using ANOVA followed by
Bonferroni correction for multiple comparisons. (ANOVA *P<0.05, **P<0.01 and ***P<0.001 vs.
control).
We recently reported the new concept of CB1R behaving as a modulator both of cell proliferation
and immune escape in glioma [42]. In particular, the use of the CB1R inverse agonist SR141761
2
3

ACCEPTED MANUSCRIPT
suggested that CB1R blockage mediates both an anti-proliferative action and an immunomodulatory
capacity, in terms of up-regulation of immunogenic signals on glioma cell surface [42]. Taking all
these in account, we asked whether the residual CB1R activity of tested compounds might induce
the expression of the NKG2D ligand MICA/B on the surface of U251MG glioma cells. Indeed,
while the anti-proliferative action is related to CB1R-blocking, we cannot exclude that the
immunogenic potential might be related at least in part to CB1R activation. So, by FACS analysis
we monitored the cell surface expression of the NKG2D ligand MICA/B on untreated and
compounds treated U251 glioma cells after 24 h of incubation to exclude that the possibility of
observed effects were linked to the stress of dying cells. We specifically found that the incubation
with CB1R agonists B2, B4 and B6, resulted in a significant up-regulation of MICA/B expression
on U251 cells in terms of Mean Fluorescence Intensity (MFI) (Fig. 5). These results could help
clarifying some aspects of the effects shown by these new ECS modulators, which may be able to
exert both cytostatic effects through CB2R and immunogenic activity mainly through CB1R
modulation.
2
4

ACCEPTED MANUSCRIPT
Figure 5. Effect of MC21, B1-B6 and C1 on cell surface expression of MICA/B on U251 glioma
cells. Representative example for cytofluorimetric histogram profiles of MICA/B protein levels at
the U251 cell surface of control (black long dash dot lanes) or cells treated with compound for 24 h
(empty grey profiles) (upper panel). The mean fluorescence intensity (MFI) of U251 positive cells
for MICA/B was calculated based on 3 independent experiments. Bar graphs report mean values ±
SD (lower panel). Then, the collected data were analyzed for statistical significance using ANOVA
followed by Bonferroni correction for multiple comparisons (ANOVA, (*P < 0.05, ***P < 0.001
compared with untreated cells).
2
5

ACCEPTED MANUSCRIPT
2.3. In silico ADME profiling
In order to get a preliminary picture of the compounds’ ADME profile, some of the most relevant
physicochemical properties were predicted for the most active ligands. For instance, the software
QikProp from Schrodinger suite [43] was used to calculate the octanol/water partition coefficient
(expressed as logP), which is commonly used as an indicator of molecular lipophilicity, the aqueous
solubility (expressed as logS) and the apparent permeability of the ligands through Caco-2 and
MDCK (Madin-Darby canine kidney) cells, which are considered to be a good mimic of the gut-
blood barrier and the blood-brain barrier, respectively [44]. Such properties are reported in Table 3,
together with the Lipinski and Jorgensen rules violations.
As expected, all ligands showed high values of logP, and low values of logS, which are both in
agreement with the substantial lipophilic character of the compounds that can be deduced by their
chemical structure. Accordingly, very high permeability to both Caco-2 and MDCK cells was
calculated for all ligands. Compound B1 was unsurprisingly predicted to be the most soluble and
the least lipophilic of the series, bearing only a methyl group and a hydrogen in position 4 and 5 of
the oxopyridine core, respectively. Interestingly, compound B5, which is a structural isomer of B1,
presented similar values of logP and logS but a higher permeability. The presence of a bromine
atom in position 5 of the core determines an increased lipophilicity and a decreased solubility in B2
and B6, with respect to their non-brominated analogues B1 and B5. Interestingly, the presence of
the bromine substituent seems to confer a particular boost to the MDCK apparent permeability of
B2 and B6 and this is a highly desirable effect considering that the two ligands showed the best
activity in inducing the expression of MICA/B on the surface of U251MG glioma cells. Compounds
B3 and B4, where the 4-methyl group was replaced by an aromatic substituent, showed the highest
lipophilicity and the lowest solubility values. Overall, the physicochemical properties predicted for
the compounds should be suitable for the activity of the ligands on the central nervous system.
However, although all compounds have logP values in the range of 95% of known drugs, a lower
2
6

ACCEPTED MANUSCRIPT
lipophilicity would be desirable for a good bioavailability and a low toxicity risk. In fact, all ligands
but B1 violated a Lipinski rule, having a logP value higher than 5. Moreover, derivatives with
higher water solubility would be required in order perform in vivo studies on this series of
compounds. Indeed, five ligands (B2-B4, B6 and C1), violated a Jorgensen rule, presenting a logS
value lower than -5.7, while B3, B4 and C1 showed solubility values that fall outside the range of
95% of known drugs (-6.5 < logS < 0.5). Based on these considerations, the series of compounds
could be expanded introducing derivatives bearing more hydrophilic substituents, such as the
morpholine ring of the recently reported CB2 ligand VL15 [45], with the aim of improving ADME
profile and developing better candidates for in vivo studies.
Table 3. Predicted lipophilicity (logP), water solubility (logS), Caco-2 and MDCK cell apparent
permeability, as well as rules violation of compounds B1-B6 and C1.
a
b
Caco-2
nm/s)
MDCK
(nm/s)
RoF
RoT
Cmpd
B1
logP
4.7
logS
-5.6
-6.4
-7.5
(
violations
violations
2664.5
2677.4
2727.3
2573.3
6182.6
2638.9
0
0
B2
5.2
1
1
1
1
1
1
B3
6.3
B4
B5
B6
C1
6.3
5.1
5.6
6.0
-7.6
-5.4
-6.1
-6.7
2723.2
4379.0
4304.2
3140.9
2634.5
4281.1
9848.6
3073.9
1
1
1
0
1
1
a
b
Lipinski rule of Five, Jorgensen rule of Three.
2.4. Molecular Modeling Studies
2
7

ACCEPTED MANUSCRIPT
To rationalize the structure-activity relationships (SARs) experimentally observed, a molecular
docking analysis of the synthetized compounds was carried out into the three-dimensional structure
of CB1R and CB2R.
Recently, the crystal structure of CB1R in complex with the antagonist AM6538 has been
determined (PDB code: 5TGZ) [46]. To date, the CB2R subtype has not been crystallized yet. Thus,
a homology model of this target has been developed by using CB1R structure as template, since the
two receptors are endowed with a good sequence similarity (44% of amino acid sequence identity).
Starting from the alignment shown in Figure S1 (Supporting information), an initial model of CB2R
was created and subjected to a simulated annealing protocol by means of the software Modeller
[
47]. The model was inserted into a membrane bilayer and water environment and subjected to 100
ns of MD simulation with AMBER14 [48]. Before the simulation, geometry optimization of the
CB2R model has been performed with a position constraint of 10 kcal/(mol⋅Å) applied to the
protein α carbons, employing a total of 20’000 steps of energy minimization. The same analysis was
applied to the CB1R-AM6538 complex in order to validate the procedure. The synthesized
compounds were docked using AUTODOCK 4.2 into the CB1R and CB2R. For each ligand, a total
of 200 different docking poses were generated and clustered with an RMSD threshold of 2.0 Å. For
each compound, the top-scored pose was considered. Figure 6A shows the docking results obtained
for compound B2 into the CB1R and CB2R binding sites. Among all the compounds of this series,
compound B2 revealed the highest CB1R and CB2R affinity. In the CB1R model the compound is
localized between the transmembrane domain (TM) 2, 3, 6 and 7 and shows a large number of
lipophilic interactions. The p-fluorobenzyl substituent is directed towards the intracellular side of
the receptor and interacts with V196 and L387 whereas the cycloheptyl group is inserted into a
lipophilic cleft mainly delimited by T197, F268, W279, W356, L359 and M363. Finally, the 5-
bromo-4-methyl-2-oxopyridine central core of the molecule shows additional lipophilic interactions
with F102, M103, F268 and F379. The analysis of the docking results into the CB2R highlights that
2
8

ACCEPTED MANUSCRIPT
this compound shows the same binding mode and the same interaction present in the CB1R (see
Figure 6B). This result is in agreement with the analysis of the non-conserved residues around the
compound. By applying a threshold of 4 Å around the hypothesized binding disposition of B2, there
are only three non-conserved residues and their mutation does not affect the interactions of the
compound, as the CB1R G166, L193 and S199 are changed in the CB2R with A83, I110 and T116,
respectively (see Figure S1).
Figure 6. Top-scored poses of compound B2 docked into CB2R (A) and CB1R (B).
As shown in Figure S2, the replacement of the 5-bromine atom with a bulkier group such as the
phenyl ring as for compound B3, does not determine important modifications in the binding mode
of the compound into both receptors as the phenyl ring fulfils the binding site cavity near the TM1,
2, 7 region (see Figure S2). The higher affinity for CB2R with respect to CB1R showed by this
compound (3-fold selectivity) could be probably due to the interaction of the phenyl ring with the
2
9