Piperidine and piperazine inhibitors of fatty acid amide hydrolase targeting excitotoxic pathology

Article abstract of DOI:10.1016/j.bmc.2019.115096

FAAH inhibitors offer safety advantages by augmenting the anandamide levels “on demand” to promote neuroprotective mechanisms without the adverse psychotropic effects usually seen with direct and chronic activation of the CB1 receptor. FAAH is an enzyme implicated in the hydrolysis of the endocannabinoid N-arachidonoylethanolamine (AEA), which is a partial agonist of the CB1 receptor. Herein, we report the discovery of a new series of highly potent and selective carbamate FAAH inhibitors and their evaluation for neuroprotection. The new inhibitors showed potent nanomolar inhibitory activity against human recombinant and purified rat FAAH, were selective (>1000-fold) against serine hydrolases MGL and ABHD6 and lacked any affinity for the cannabinoid receptors CB1 and CB2. Evaluation of FAAH inhibitors 9 and 31 using the in vitro competitive activity-based protein profiling (ABPP) assay confirmed that both inhibitors were highly selective for FAAH in the brain, since none of the other FP-reactive serine hydrolases in this tissue were inhibited by these agents. Our design strategy followed a traditional SAR approach and was supported by molecular modeling studies based on known FAAH cocrystal structures. To rationally design new molecules that are irreversibly bound to FAAH, we have constructed “precovalent” FAAH-ligand complexes to identify good binding geometries of the ligands within the binding pocket of FAAH and then calculated covalent docking poses to select compounds for synthesis. FAAH inhibitors 9 and 31 were evaluated for neuroprotection in rat hippocampal slice cultures. In the brain tissue, both inhibitors displayed protection against synaptic deterioration produced by kainic acid-induced excitotoxicity. Thus, the resultant compounds produced through rational design are providing early leads for developing therapeutics against seizure-related damage associated with a variety of disorders.

Full text of DOI:10.1016/j.bmc.2019.115096

Journal Pre-proofs
Piperidine and piperazine inhibitors of fatty acid amide hydrolase targeting ex-
citotoxic pathology
Manjunath Lamani, Michael S. Malamas, Shrouq I. Farah, Vidyanand G.
Shukla, Michael F. Almeida, Catherine M. Weerts, Joseph Anderson, JodiAnne
T. Wood, Karen L.G. Farizatto, Ben A. Bahr, Alexandros Makriyannis
PII:
S0968-0896(19)31023-5
https://doi.org/10.1016/j.bmc.2019.115096
BMC 115096
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
19 June 2019
7 September 2019
8 September 2019
Please cite this article as: Lamani, M., Malamas, M.S., Farah, S.I., Shukla, V.G., Almeida, M.F., Weerts, C.M.,
Anderson, J., Wood, J.T., Farizatto, K.L.G., Bahr, B.A., Makriyannis, A., Piperidine and piperazine inhibitors of
fatty acid amide hydrolase targeting excitotoxic pathology, Bioorganic & Medicinal Chemistry (2019), doi: https://
doi.org/10.1016/j.bmc.2019.115096
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will
undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing
this version to give early visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
©
2019 Published by Elsevier Ltd.

Piperidine and piperazine inhibitors of fatty acid amide hydrolase targeting excitotoxic
pathology
1
*1
1
1
Manjunath Lamani , Michael S. Malamas , Shrouq I. Farah , Vidyanand G. Shukla , Michael
2
1
1
1
F. Almeida , Catherine M. Weerts , Joseph Anderson , JodiAnne T. Wood , Karen L.G.
2
2
1
Farizatto , Ben A. Bahr and Alexandros Makriyannis .
1Center for Drug Discovery and Departments of Chemistry and Chemical Biology and
2
Pharmaceutical Sciences, Northeastern University, Boston, MA 02115, USA; Biotechnology
Research and Training Center, University of North Carolina Pembroke, Pembroke, NC 28372,
USA.
Abstract
FAAH inhibitors offer safety advantages by augmenting the anandamide levels “on demand” to
promote neuroprotective mechanisms without the adverse psychotropic effects usually seen with
direct and chronic activation of the CB1 receptor. FAAH is an enzyme implicated in the hydrolysis
of the endocannabinoid N-arachidonoylethanolamine (AEA), which is a partial agonist of the CB1
receptor. Herein, we report the discovery of a new series of highly potent and selective carbamate
FAAH inhibitors and the evaluation for neuroprotection. The new inhibitors showed potent
nanomolar inhibitory activity against human recombinant and purified rat FAAH, were selective
(>1000-fold) against serine hydrolases MGL and ABHD6 and lacked any affinity for the
cannabinoid receptors CB1 and CB2. Evaluation of FAAH inhibitors 9 and 31 using the in vitro
competitive activity-based protein profiling (ABPP) assay confirmed that both inhibitors were
highly selective for FAAH in the brain, since none of the other FP-reactive serine hydrolases
1

in this tissue were inhibited by these agents. Our design strategy followed a traditional SAR
approach and was supported by molecular modeling studies based on known FAAH cocrystal
structures. To rationally design new molecules that are irreversibly bound to FAAH, we have
constructed “precovalent” FAAH-ligand complexes to identify good binding geometries of the
ligands within the binding pocket of FAAH and then calculated covalent docking poses to select
compounds for synthesis. FAAH inhibitors 9 and 31 were evaluated for neuroprotection in rat
hippocampal slice cultures. In the brain tissue, both inhibitors displayed protection against synaptic
deterioration produced by kainic acid-induced excitotoxicity. Thus, the resultant compounds
produced through rational design are providing early leads for developing therapeutics against
seizure-related damage associated with a variety of disorders.
Keywords: Fatty acid amide hydrolase (FAAH); FAAH inhibitor; monoacylglycerol lipase
(MGL); N-arachidonoylethanolamine; cannabinoid CB1 receptor; endocannabinoid system;
hippocampus; neuroprotection.
1
. Introduction
1
Fatty acid amide hydrolase (FAAH) is an integral membrane-bound enzyme that primarily
degrades the N-acylethanolamines (NAEs) family of signaling biolipids, including N-
arachidonoylethanolamine (anandamide, AEA), palmitoylethanolamine (PEA), and
oleoylethanolamine (OEA). AEA was identified as an endogenous partial agonist of the
2
cannabinoid receptors CB1 and CB2 , two members of the endocannabinoid system (eCB). The
eCB consists of the known G-protein-coupled cannabinoid receptors CB1 and CB2,3, 4 the
endogenous lipid ligands arachidonoylethanolamine (anandamide, AEA)2 and 2-
2

5
arachidonoylglycerol (2-AG), as well as their primary hydrolytic enzymes FAAH and
6
monoacylglycerol lipase (MGL) . The CB1 receptors are widely distributed throughout the brain
7
and mediate the majority of the cannabinoid effects in the central nervous system (CNS). In
8
contrast, the CB2 receptors are primarily found in cells of immune origin, including microglia,
9
although low-level expression has been reported in healthy neurons. There is an abundance of
evidence to support that dysfunction of the eCB system leads to a wide range of neuropathological
disorders, such as obesity, immunological dysfunction, metabolic syndromes, psychiatric
conditions, epilepsy, and neurodegenerative diseases, including Parkinson’s disease and
Alzheimer’s disease10-13
.
Activation of CB1 receptors can protect against hyperexcitability induced neuronal damage.14-18
Hence, FAAH inhibition and increased AEA levels offer the opportunity to protect neurons against
excitotoxic damage. Moreover, activation of the CB1 receptors with exogenous natural and
synthetic cannabinoids found to exert neuroprotective functions in several models of
neurotoxicity19, 20 and disruption of CB1 receptors signaling caused an increase in excitotoxic
vulnerability and seizure susceptibility, as well as attenuation of neuronal maintenance.17, 21, 22
Notably, studies have shown that neuronal depolarization triggered by excitotoxic kainic acid (KA)
in animals resulted in marked increase of anandamide levels leading to neuroprotection,23-25 while
the tissue concentrations of 2-AG remained unchanged. In this paper, we discover new FAAH
inhibitors and study their role in neuroprotection and epilepsy.
We hypothesize that the use of FAAH inhibitors will lead to “on demand” augmentation of
anandamide as a response to sustained neuronal depolarization, as seen to occur with seizure
activity, and contribute to the activation of CB1 receptors and trigger protective mechanisms
against a neurotoxic insult.26-28 Recent findings have reinforced the role of FAAH inhibitors in
3

neuroprotection by reducing hippocampal excitability29 and seizure severity.14 Furthermore,
FAAH inhibitors offer safety advantages over CB1 agonists in neuroprotection by increasing the
levels of AEA “on-demand” to activate CB1 receptors, while lacking the adverse psychotropic
3
0
effects usually seen with direct and persistent stimulation of CB1 receptors .
Herein, we report a new series of highly selective FAAH inhibitors and their assessment to protect
against KA-induced excitotoxic damage as a potential treatment for epileptic seizures as well as
other disorders involving excitotoxicity in vulnerable brain regions like the hippocampus. The
FAAH inhibitors were evaluated for neuroprotection against (1) molecular indicators of pathology
in the excitotoxic hippocampal slice model, and (2) pathogenic indicators after intrahippocampal
injection of excitotoxin. Our results indicate that after excitotoxic insults, FAAH inhibition with 9
O
H
and 31 protects against neural compromise in
hippocampal slices and reduces seizure
activity and seizure-associated damage in
rats. These findings provide further evidence
for the therapeutic potential of FAAH
inhibition against excitotoxic brain injury.
N
H2N
O
N
O
N
O
N
O
OL-135
URB597
N
O
CF3
N
N
O
O
N
N
N
H
H
N
N
H
O
O
O
SA-47
PF-04457845
O
N
N
O
O
F
F
Cl
N
N
N
N
H
H
Cl
N
N
O
JNJ-40355003
JNJ-42165579
O
S
N
N
NH
O
S
N
N
O
N
S
N
N
O
S
N
1
(Abbott)
2 (Amgen)
1
.1
Known FAAH Inhibitors
Figure 1. Known FAAH inhibitors
Many FAAH inhibitors have been reported31, 32 during the last decade, mostly represented by three
distinct structural classes. These inhibitors are grouped as follow: (1) covalent reversible
3
3
inhibitors, such as -ketoheterocycle OL-135 ; (2) covalent irreversible inhibitors exemplified
3
4
35
36
by aryl-carbamates (e.g. URB597 , SA-47 ) and aryl-ureas (e.g. PF-04457845 , JNJ-40355003,
JNJ-42165579) and (3) non-covalent reversible inhibitors exemplified by benzothiazole 1
4

3
7
38
(
Abbott) and benzothiophene 2 (Amgen) . To date, only a few FAAH inhibitors have entered
clinical trials for evaluation in various diseases. Pfizer’s PF-04457845 a highly potent and
selective inhibitor with a robust druggability profile was evaluated in Phase II clinical trials
(NCT00981357) as an analgesic in patients with osteoarthritis pain but was found to be
3
9
ineffective . Also, PF-04457845 is currently under investigation for central-related pathologies
for treating post-traumatic stress disorders (PTSD; NCT02216097), cannabis related withdrawal
symptoms (NCT03386487), cannabinoid receptor augmentation on the facilitation of fear
conditioning (NCT01665573) and Tourette’s syndrome (NCT02134080). Furthermore, other
companies as well are studying FAAH inhibitors in clinical
trials including Vernalis V158866 for neuropathic pain
following spinal cord injury (NCT01748695) and Sanofi-
Aventis SSR411298 as a treatment for major depressive
disorder in elderly patients (NCT00822744).
C
B
A
O
ABP channel
N
O
R2
CP channel
X
X = C, N
MAC channel
R1
Figure 2. Expolre regional hydrophilic
paterns of ligand substitution to improve
druggability
In an effort to synthesize selective and potent inhibitors of FAAH for neuroprotection, we have
examined series of carbamates derived from piperidine and piperazines motifs (Fig. 2).
2
2
. Chemistry
.1 Structural features of FAAH in inhibitor design
FAAH is a membrane bound serine hydrolase enzyme defined with a highly conserved primary
sequence rich in serine and glycine residues and characterized by a unique Ser241-Ser217-Lys142
4
0
catalytic triad in which Ser241 plays a critical catalytic role . Human FAAH shares ~82%
sequence identity with rat FAAH, however there are six amino acids differences within the
substrate-binding pockets of hFAAH and rFAAH. To better explore the hFAAH structure in drug
5

design, a “humanized” rat FAAH (h/rFAAH) model was developed41 where in the resolved
h/rFAAH structure in complex with inhibitor PF-750 (2.75 Å, PDB code: 2VYA) six amino acids
of the active site were mutated into those of the human FAAH protein sequence (namely
Leu192Phe, Phe194Tyr, Ala377Trp, Ser435Asn, Ile491Val, and Val495Met). This discovery
facilitated the design of FAAH inhibitors with equipotencies for either species, alleviating
discrepancies in inhibitor activity between rodent (preclinical model) and human (clinical
development). The structural insights of the resolved FAAH structures revealed the presence of
three distinct channels responsible for ligand-protein interactions known as (i) the membrane
access channel (MAC); (ii) the acyl-chain binding pocket (ABP); and (iii) the cytosolic port (CP).42
This ligand-protein interaction arrangement allowed the design of FAAH inhibitors with diverse
physicochemical properties (ClogP, tPSA, LipE) taking advantage of the distinct “make-up” of
those regions. 30, 31 Molecular recognition of inhibition mechanism within these channels produced
several series of FAAH inhibitors that interacted with the protein in a covalent or a noncovalent
4
2
manner. Furthermore, the resolution of the apo structure of rFAAH revealed that an open
conformation is the genuine native conformation for the enzyme residue, while the closed
conformation is likely triggered by the rearrangement of flexible residue side chains of the
binding pocket upon ligand binding. Many designed FAAH inhibitors (Fig. 1) contain
carbamate or urea pharmacophores which covalently bind to catalytic Ser241 forming a covalent
carbamate adduct intermediate as an approach to achieve sustained FAAH inhibition and
prolonged pharmacodynamics (PD) effects. However, this newly formed carbamate adduct, in a
second phase hydrolysis process is converted back to the active enzyme, while in the process a
structurally modified ligand is released as the hydrolytic product. This interconversion process
“
intrinsic reversibility” can be exploited by modifications of the ligand’s size (length and
6

bulkiness) and generate analogs with “tunable” adduct residence time (). This unique reversible
mode of FAAH inhibition represents a powerful mechanism for abrogating enzyme activity, and
it is considered the most pharmacologically tractable mechanism of choice for enzyme inhibition.
This strategy on targeting “covalent - reversible” drugs offers the advantage of “tuning-up” a
desirable pharmacological effect of a molecular target and mimic the long-lasting activity of
irreversible inhibitors, with dissociation rates in vivo that approach the rate of target degradation
and re-synthesis.43, 44 Also, this strategy offers an additional advantage in minimizing potential
“
drug-drug” interactions, since the drug could rapidly clear from the circulation (short PK effect),
while still maintains a lasting target engagement effect (long PD effect). In contrast, fully
irreversible adduct formation (suicidal inhibitors) have raised serious safety concerns in drug
development because of the relationship between covalent drug binding and the potential of
immunogenic-driven allergic reactions and idiosyncratic drug toxicity.45, 46
2
.2 Design of new carbamate FAAH inhibitors
Our design strategy followed a traditional SAR approach and was supported by molecular
modeling studies based on FAAH cocrystal structures. Examination of FAAH-inhibitor co-
crystals41, 42, 47-50 revealed that the FAAH active-site possesses high flexibility allowing several
residues lining the binding cavity to attain conformations “friendly” to the interactions with
structurally diverse ligands. Taking advantage of these observations, we explored carbamate
FAAH inhibitors seeking compounds with high potency and selectivity with equipotency for both
human- and rat-FAAH. An emphasis was given to identify new molecules with increased polarity
(lower ClogP) and enhanced water solubility suitable as oral drugs. To that end, we have explored
structural features based on existing FAAH templates in the literature or generated in our own lab
7

(Fig. 2). To support the design of new FAAH inhibitors, we have applied computational methods
supported by the available X-ray crystallographic data for FAAH41, 42, 47, 51. To that end:
1
. We have incorporated polar moieties (i.e. morpholine, dioxo-thiomorpholine) at the
pharmacophoric region (C; Fig. 2) predicted to occupy the cytoplasmic channel (CP) of the
binding pocket of FAAH, which represents a hydrophilic region adjacent to the catalytic site
and contributes to key interactions with the polar head groups of substrates (i.e. ethanolamide
moiety of AEA) and
inhibitors. Those polar
moieties counter the
hydrophilic residues at
the CP channel targeting
hydrogen bonding and
Fig. 3 Induced-fit docking of 9 with humanized FAAH crystal structure. A
Connolly surface shown as electrostatic potential (red-white-blue) was used to represent
electrostatic interactions
the shape of the binding pocket. Catalytic residue Ser241 and Phe432, a key residue of
the MAC channel, are shown in green. Key residues forming the binding pocket shown
in wireframe. The ligand is shown in gray. Atoms are colored by atom type. Hydrogen
bonds are indicated by yellow dashed lines, and the distance of catalytic Ser241 from the
carbamate group in magenta dashed line.
between
ligand
and
protein to enhance the
ligand’s affinity. Even though this part of the molecule is departed upon formation of the
covalent adduct between ligand-protein (carbamylation step), still it plays a critical role in
forming a favorable “precovalent” pose between ligand and protein that leads to a nucleophilic
attack of the catalytic Ser241 of FAAH to the carbamate moiety of the ligand to form a covalent
adduct.
2
. We have added polar groups (i.e. alkoxy, cyclic-amines) at the distal aromatic region of the
molecule (A; Fig. 2) predicting to occupy the ABP and MAC channels. This region of the
binding pocket is primarily hydrophobic in nature, but a few residues with polar characteristics
8

are present at the junction of
these two channels allowing
polar groups of the ligand to
interact
favorably
(i.e.
hydrogen bonding) with the
Fig. 4 Covalent docking of 9 with humanized FAAH crystal structure with the leaving
protein and increase ligand group no longer attached. A Connolly surface shown as electrostatic potential (red-
white-blue) was used to represent the shape of the binding pocket. Catalytic residue
Ser241 and Phe432, a key residue of the MAC channel, are shown in green. Key
residues forming the binding pocket shown in wireframe. The ligand is shown in gray.
Atoms are colored by atom type. Hydrogen bonds are indicated by yellow dashed lines.
affinity.
3
. We have introduced bulky
groups (i.e. substituted-, bridged-piperazines) with relative and absolute configuration as well
as conformational restriction within individual analogs between the distal aromatic region and
the carbamate pharmacophore (B; Fig. 2) to influence ligand selectivity and increase residence
time of the newly formed covalent adducts between ligands and protein. We postulated that
the bulkiness of the molecule close to the vicinity of the carbamylation adduct would retard its
hydrolysis rate to restore the free enzyme (second phase hydrolysis), while prolonging the
inactivation period of the enzyme and extending the pharmacodynamic effect of the drug.
.3 Molecular modeling: Crystallographic and biochemical studies have confirmed that
2
carbamate- or urea-based FAAH inhibitors bind covalently to catalytic Ser241 forming a
carbamate adduct intermediate.41, 42, 51, 52 However, direct structural methods assessing the binding
mode of covalently bound drugs that form adducts with the protein by displacing a portion of the
ligand is not practical, since the displaced group has been departed at the postcovalent state.
Therefore, to gain insight of the interactions of the intact molecule with the protein to design new
covalent inhibitors, we have constructed “precovalent” FAAH-ligand complexes and minimized
them in the binding channel. This allowed us to pre-position the ligand within the binding pocket
9

of FAAH to determine if it is aligned into
binding site with acceptable binding
geometry. For molecules with good fit
into the binding pocket, we have
calculated covalent docking poses by
employing Schrodinger methods. Fig. 3
illustrates the induced-fit docking of
analog 9 (entry 9; Table 1) within the
h/rFAAH structure (PDB code: 2VYA)
Fig. 5 Overlapped covalent docking poses of 9 (green) and PF-3845 (magenta) with
humanized FAAH crystal structure. A Connolly surface shown as electrostatic potential
(red-white-blue) was used to represent the shape of the binding pocket. Phe432, a key
residue of the MAC channel, is shown in green. Key residues forming the binding pocket
are shown in thin tube representation (gray). Atoms are colored by atom type. Hydrogen
bonds are indicated by yellow dashed lines, and π-stacking interactions shown in blue
dashed lines.
in the “precovalent” state orientation, and Fig. 4 illustrates the covalent adduct formed between 9
and the catalytic Ser241 of FAAH. In the “precovalent” pose (Fig. 3), 9 orients within the FAAH
binding pocket by making key hydrogen bonding interactions with residues (Ile238, Gly239,
Ser241) of the CP channel, positioning the carbamate group of the ligand at a close vicinity (~ 3
Å) to the catalytic Ser241 suitable for adduct formation (carbamylation step). The azetidine ring
at the distal end of the acyl chain-binding pocket is close to Phe432, a key residue of the MAC
channel, which undergoes a shift toward the ABP channel expanding its regional access. The
MAC channel has been suggested to serve as the “gate” for lipid substrates to enter FAAH from
5
3
the cell membrane. A similar ligand orientation was maintained after the covalent bond formation
with the catalytic Ser241 (Fig. 4). Two key hydrogen bonding interactions between 9 and residues
Ile238 and Gly239 of the CP channel are shown in Fig. 4. We have overlapped 9 and PF-3845
(Fig. 5) and found that both orient similarly within the FAAH binding pocket. During our modeling
studies, we have postulated that FAAH ligands are able to effectively occupy into multiple
conformations within the binding pocket to facilitate ligand adaptation for covalent bond
1
0

formation. In support of this hypothesis, known FAAH-inhibitor cocrystal structures have revealed
that the FAAH active site has remarkable flexibility, where the ABP channel is lined-up with
predominately hydrophobic residues, reflecting the need of this pocket to favor nonspecific
binding and keep energy wells shallow for ligand adaptation. We have routinely determined the
lower energy conformations of the ligand in the absence of protein and then compared it to the low
energy conformer generated during “precovalent” docking to determine structures with minimal
energy penalty for synthetic consideration. Structure-activity relationship studies have produced
many compounds, and in this paper, we outline selected analogs in Tables 1 and 2.
Scheme 1
F
F
F
F
Y
a
Y
O
2
.4 Synthesis
HO
HO
NH2
e
O
O
N
1
HO
N
F
5
X
4
a-4e
X
The compounds needed to
delineate the SAR for this study
were prepared according to
Methods A-C.
b,c,d
NH
Y
3a: X = SO , Y = C
2
b: X = NMe, Y = C
c: X = SO , Y =N
d: X = CH , Y = C
f
Y = N
R
3
3
3
3
c
Y
Br
2
O
2
Y = C
2
e: X = O; Y = C
N
O
N
X
R
6a-6e
Reagents and conditions: (a) Divinyl sulfone, boric acid, glycerol (cat.), water,
reflux, 3 h; (b) benzyl bromide, K CO , DMF; (c) thiomorpholine 1,1-dioxide
2
3
t
o
or1-methylpiperazine, Pd (dba) , BINAP, NaO Bu,toluene 110 C; (d) H , Pd/C,
2
3
2
EtOH; (b); (e) 3 bis(pentafluorophenyl) carbonate, N,N-diisopropylethylamine,
0 oC to rt, 2h; (f) 4, N,N-diisopropylethylamine, 0 C to rt, 2h.
o
2
.4.1 Synthesis of piperidine
analogs
Method A
The piperidine-carbamates (Table 1) were prepared according to schemes 1 and 2. Aniline 1 was
treated with divinyl sulfone and boric acid to afford 4-(3-hydroxyphenyl) thiomorpholine 1,1-
dioxide 3a (X = SO , Y C)). Buchwald-Hartwig palladium cross-coupling methodology was
2
employed to prepare phenol 3b (X = NMe, Y = C). 3-Bromophenol was coupled with 1-
1
1

methylpiperazine in the presence of Pd (dba) , BINAP and NaOtBu to afford 3b. The pyridinol 3c
2
3
(
X = SO , Y =N) was prepared from 5-bromopyridin-3-ol upon protection of the hydroxyl group
2
with benzyl bromide/potassium carbonate, followed by Buchwald-Hartwig palladium cross-
coupling with thiomorpholine 1,1-dioxide and then removal of the benzyl group with H / Pd/C to
2
afford 3c. Piperidine-phenol 3d (X = CH Y = C) and morpholine-phenol 3e (X = O; Y = C) were
2
,
commercially available. Treatment
of phenols 3a-3e with
Scheme 2
Boc
c
Boc
R = Br
R = pyrolidine
N
N
N
a
R = OH
b
R
R
R
8a: R = pyrolidine
b: R = O(CH ) F
bis(pentafluorophenyl) carbonate
in the presence of N, N-
diisopropylethylamine produced
carbonates 4a-4e. Reaction of
7
=O(CH2)2F
8
2
2
R = Br, OH
O
TFA
N
O
N
d
X
R
R
9a: R = pyrolidine
b: R = O(CH ) F
10a: R = pyrolidine
10b: R = O(CH2)2F
9
2
2
Reagents and conditions: (a) Pd (dba) , 2-(di-tert-butylphosphino)biphenyl,
2
3
o
carbonates
4a-4e
with
an
sodium tert-butoxide, pyrrolidine, toluene, 80 C, 12h; (b)1-bromo-2-fluoroethane,
o
K CO , CH CN, 80 C, 16h; (c) TFA, CH Cl 6h, rt; (d) 3-(1,1-
2
3
3
2
2
dioxidothiomorpholino)phenyl (perfluorophenyl) carbonate, N,N-
appropriately
substituted
o
diisopropylethylamine, 0 C to rt, 2h.
piperidine 5 in the presence of N, N-diisopropylethylamine afforded carbamates 6a-6e.
Method B
Functionalization of the terminal aryl region of 7 was accomplished according to scheme 2. Aryl
bromide 7 (R = Br) was treated with tris(dibenzylideneacetone) dipalladium(0)/2,2'
bis(diphenylphosphino)-1,1'-binaphthyl and a cyclic amine (i.e. pyrrolidine) to produce aryl-
piperidine 8a. Also, alkylation of phenol 7 (R = OH) with an alkyl halide (i.e. 1-bromo-2-
fluoroethane) in the presence of base (i.e. K CO ) produced 8b (R = OCH CH F). BOC
2
3
2
2
deprotection of 8a, 8b with trifluoroacetic acid produced amines 9a, 9b, which were converted to
final carbamates 10a, 10b as above. (Method A).
1
2

2
.4.2 Synthesis of piperazine analogs
Method C
The piperazine-carbamates (Table 2) were prepared according to scheme 3. Buchwald-Hartwig
5
4
palladium cross-coupling methodology was employed to prepare the aryl-piperazines 12. Aryl
halide 11 was treated with commercially BOC protected piperazines (i.e. tert-butyl (R)-2-
Scheme 3
Y
methylpiperazine-1-carboxylate)
chiral substituted-piperazines to prepare
analogs (28-38; Table 2).
or
N
X
b
a
N
R1
R2
Y = BOC
R1
12
11
Y = BOC, H
X = Br, I
O
O
NH
N
N
Tris(dibenzylideneacetone)
dipalladium(0)/2,2'-
c
N
N
X
R2
R2
R1
R
1
4
1
3
Reagents and conditions: (a) i. Pd (dba) , 2-(di-tert-butylphosphino) biphenyl,
2
3
bis(diphenylphosphino)-1,1'-binaphthyl
o
sodium tert-butoxide, tert-butyl (R)-2-methylpiperazine-1-carboxylate, 80 C,
1
2h; ii. (R)-2-methylpiperazine toluene, 2,2'-bis(diphenylphosphino)-1,1'-
o
was utilized as the coupling catalyst binaphthyl and (S)-2-methylpiperazine, 80 C, 12h; (b) TFA, CH Cl , 6h, rt; (c)
2
2
o
3
(scheme 1), N,N-diisopropylethylamine, 0 C to rt, 2h
when
Y
=
BOC
or
2,2'-
bis(diphenylphosphino)-1,1'-binaphthyl and (S)-2-methylpiperazine (BINAP when Y = H to
afford aryl-piperazine 12. Deprotection of the BOC group of 12 (Y= BOC) with trifluoroacetic
acid
produced
amine.TFA
13.
Amine
13
was
treated
with
3-(1,1-
dioxidothiomorpholino)phenyl(perfluorophenyl) carbonate 3 (scheme 1) in the presence of N, N-
diisopropylethylamine to afford piperazine-carbamate 14.
3
. Results and Discussion
1
3

All synthesized compounds were tested for their ability to inhibit human-recombinant FAAH
5
5
56
(
hFAAH) and purified rat FAAH (rFAAH) using the fluorogenic substrate arachidonoyl 7-
amino-4-methylcoumarin amide (AAMCA)56, 57. We used both species hFAAH and rFAAH to
assess potential inhibitory differences, which have been previously identified by various structural
chemotypes. Following similar fluorometric procedures, all compounds were also tested against
hydrolytic enzymes recombinant human monoacylglycerol lipase (hMGL)58 and human
alpha/beta-hydrolase domain containing 6 (hABHD6)59 for selectivity. Furthermore, selected
6
0
compounds were evaluated for their ability to bind to CB1 and CB2 receptors using rat brain or
HEK293 cell membranes expressing mouse CB2 (mCB2) or human CB2 (hCB2),61-63 respectively,
3
62-64
via competition-equilibrium binding using [ H]CP-55,940
to minimize their involvement due
to cross-reactivity of potential common pharmacophoric features.
Our design strategy was primed to optimize structural features of template shown in Fig. 2. As we
discussed above, computational methods supported by available X-ray crystallographic data for
FAAH and molecular modeling studies applied to pre-evaluate proposed synthetic targets. In
addition, in silico chemoinformatics calculations (i.e., MW, LE, LipE, ClogP, and tPSA) guided
the choice and priority for compounds to be synthesized to introduce desirable pharmacokinetic
characteristics. Full assessment of the in vitro potency and selectivity as well ADME compound
profiling (i.e., plasma stability, microsomal stability) was performed for selected molecules to
achieve compounds with suitable druggable properties for evaluation in preclinical models of
neuroprotection. A selection of compounds of our structure-activity relationship (SAR) studies are
shown in Tables 1-2. We have initially sub-divided our working template (Fig. 2) into regions A
to C with the ligand arranged to occupy the three distinct binging channels of FAAH.
1
4

Region A: We have identified that phenyl-dioxothiomorpholine carbamate 1 (Table 1) inhibited
both species (rat, human) of FAAH at low nanomolar concentrations and displayed excellent
selectivity against MGL. Computational calculations positioned the phenyl-dioxothiomorpholine
moiety deep into CP channel of the binding pocket of FAAH making several key hydrogen
bonding interactions and van der Waals hydrophobic contacts (i.e. Fig. 3). Keeping constant the
phenyl-dioxothiomorpholine moiety, we examined the effects of the substituents at the distal
phenyl moiety occupying the ABP/MAC channels of the binding pocket of FAAH. We have
introduced a variety of groups at region A with polar characteristics to improve the water solubility
of the parent compound and target favorable hydrogen-bonding interactions with the protein to
further improve ligand affinity. Introduction of electronegative fluorine or polar alkoxy groups
(entries 2-4) were beneficial on improving about 2-3-fold rFAAH potency (entries 2-4 vs 1). Both
the fluorine nucleus and the trifluoromethoxy group were also introduced to retard any potential
oxidative cytochrome P450 metabolic instability of this part of the molecule. Addition of a linear
nitrile group (entry 5) did not affect ligand affinity for rFAAH, but it decreased affinity ~3-fold
for hFAAH. The longer fluoroethoxy group (entry 6) was similar to 1 for rFAAH but showed 3-
fold drop in hFAAH potency. Rotationally constraining the phenyl group of 7 by adding a methyl
group at the ortho-position resulted in 2-fold drop in potency (7 vs 1). Next, we introduced polar
amines and heterocycles (entries 8-13) to further lower ClogP and improve water solubility.
Amines (8-10) were similar to 1 for rFAAH, but about 2-3-fold better in hFAAH. In contrast,
heterocycle pyridyl 11 and oxadiazole 12 were significantly >10-fold less potent than 1. Addition
of a hydroxyl group at the position-4 of the piperidine nucleus (13) also markedly reduced ligand
potency (13 vs 1). The truncated oxetaneyl analog 14 was inactive for the target.
1
5

Region C: A significant effort was launched to examine groups positioned at the vicinity of the
CP channel of FAAH’s binding pocket. This part of the molecule (phenoxy-dioxothiomorpholine)
departs during adduct formation, but it plays a pivotal role during the initial “preconvalent” pose
of the ligand by influencing ligand adaptation within the binding pocket in a suitable orientation
for the nucleophilic reaction to occur between the ligand and the catalytic residue Ser241. To that
end, we have maintained the same 6-membered cyclic amine motif and altered the basicity of the
group. A morpholine group (analog 15) was similar in potency to that of the analogous
dioxothiomorpholine 2. The piperidine and piperazine analogs (17 & 18) showed similar potency
for rFAAH with a small 3-4-fold decrease against hFAAH. To further increase the hydrophilicity
of this region of the ligand, we have introduced a pyridine nucleus (entry 19), which was well
tolerated and exhibited high potency for rat and human FAAH, with IC values of 5 and 7 nM,
5
0
respectively. Lastly, we have converted the carbamate pharmacophore to the analogous urea
molecule 20 (20 vs 1), since many known FAAH inhibitors carry a urea-carbamate pharmacophore
(examples shown in Fig. 1). Surprisingly, the urea analog 20 was found to be inactive. This was a
bit unexpected considering the close resemblance of these two scaffolds and the structural
similarities they share with other reported urea-carbamate FAAH inhibitors (Fig. 1). Modeling
calculations were unable to explain such a dramatic shift in FAAH potency. We postulated that
the formation of a covalent adduct between the catalytic Ser241 and the urea pharmacophore of
2
0 didn’t occur.
Table 1. Piperidine-carbamates FAAH inhibitors as neuroprotective drugs
O
N
O
N
X
Ar1
1
6

a
IC nM
5
0
Compd
Ar1
X
rFAAH
hFAAH
hMGL ClogPc
1
2
3
4
5
6
7
8
9
Ph
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
6.0±01b
2.3±0.2
2.61±0.05
4.0±0.1
9.9±0.1
5.5+0.2
10.3±1.1
6.9±0.3
5.8±0.2
6.3±0.3
75.5±8.1
8.0±0.2
10.0±0.7
6.3±1.4
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
>1000
2.62
2.76
2.54
3.65
2.05
2.79
3.07
2.78
2.34
2.90
1.12
4-FPh,
4-OMePh
4-OCF Ph
3.7±0.3
3
4-CNPh
22.7±1.9
21.2±0.9
15.9±1.1
2.4±0.4
4-OCH CH F-Ph
2
2
2-MePh
4-N,N-dimethyl Ph
4-azetidyl-Ph
4-pyrrolidyl-Ph
4-pyridyl
7.9±1.0
1
1
0
1
3.5±2.2
131.6±11.1
3
-methyl-1,2,4-
oxadiazolyl
1
1
2
3
SO2
SO2
120±6.5
147.4±11.4
>1000
>1000
0.01
0.98
N
Ph
78.1±14.3 172.2±10.3
HO
1
1
1
1
1
4
5
6
7
8
oxetaneyl
4-FPh
SO2
O
NAd
NA
>1000
>1000
>1000
>1000
>1000
-0.3
3.73
3.75
5.10
4.23
2.93±0.06
14.7±1.1
6.8±0.2
9.4±0.5
16.3±1.3
31.0±7.2
40±4.2
4-N,N-dimethyl-Ph
4-FPh
O
CH
4-FPh
N-CH3
33.2±2.1
N
O
1
2
9
0
N
O
N
7.1±0.5
NA
4.9±0.7
NA
>1000
NA
2.21
1.75
SO2
4
F-Ph
O
N
N
N
H
SO2
Ph
1
7

a
Inhibition data for h/rFAAH and hMGL were determined using throughput
b
fluorescent assays. IC ± SD values were determined from triplicate experiments.
5
0
c
IC values were calculated using Prism software (GraphPad). ClogP values were
5
0
calculated using ChemDraw, version 18.2; dNA=not active at 1 uM.
Region B: To further expand the breadth of our SAR, we focused on region B of the molecule
(Fig. 2) which potentially influences the carbamylation step between ligand and protein. This
region can also affect the second phase hydrolysis of the formed adduct to regenerate the active
enzyme. Piperazine 21 was almost equipotent to piperidine 1, while piperazines 22 and 23 were a
bit 3-4-fold weaker than the analogous substituted piperidines 2 and 5. Both pyridyl and
pyridazinyl analogs 24 and 25 were markedly weaker for FAAH. The more flexible benzyl analogs
2
6 and 27 showed good activity for both hFAAH and rFAAH with IC values of about 2 nM.
50
Next, we introduced bulkier substituents at the piperazine moiety of relative and absolute
configuration as well as conformational restriction to potentially influence both potency and ligand
selectivity. Methyl substituted piperazines (28, 29, 31, 33, 34) were similar or better in potency for
FAAH when compared to parent compound 1. The methyl substitution displayed high
enantiospecificity for FAAH, with the (R)-enantiomer being the active species (for example 29 vs
3
0). Both, the dimethyl-substituted piperazine 36 and the spiro-cyclopropane 37 were weakly
active at 1 uM concentration. Unpredictably, the bridged piperazines 38 and 39 were quite different
with 38-(R, R)-enantiomer being active for FAAH, while 39-(S, S)-enantiomer was markedly 20-
fold less potent. Molecular modeling studies (not shown) indicated that steric bulkiness of the
ligand at region B can restrict its “preconvalent” pose alignment within the binding pocket of
FAAH, and only small groups, such as methyl, with the correct conformational geometry were
able to form favorable ligand poses within the binding pocket to facilitate the formation of the
1
8

covalent adduct between the ligand and the catalytic residue Ser241 of FAAH. Lastly, we
synthesized piperazine-urea 40 and found to be inactive for FAAH, similarly to the urea-piperidine
2
0 analog.
T able 2. Piperazine carbamates FAAH inhibitors as neuroprotective drugs
R1
O
N
O
N
X
N
Ar1
a
IC nM
5
0
Compd
Ar1
Ph
R1
X
rFAAH
hFAAH hMGL ClogPc
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
H
H
H
H
H
H
H
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
SO2
O
8.0±0.1b 15.0±0.5 >1000
12.5±0.3 20.5±2.1 >1000
62.7±3.5 89.3±8.3 >1000
1.93
2.24
1.76
0.98
0.03
4.11
4.32
2.45
2.45
2.45
2.76
2.76
2.47
3.43
3.69
4-F Ph
4-CN Ph
2-pyridyl
4-pyridazinyl
3-Ph-benzyl
3-PhO-benzyl
Ph
102.3±6.5
NAe
NDd
NA
>1000
>1000
>1000
>1000
>1000
2.1±0.1
2.2±0.5
5.5±0.5
1.9±0.1
2.3±0.3
2.6±0.1
(RS)-Me
(R)-Me
(S)-Me
(R)-Me
(S)-Me
(R)-Me
(R)-Me
Ph
5.7±0.2 17.7 ±3.4 >1000
Ph
NA
NA
ND
4-F Ph
6.3±0.5
11.7±2.6 >1000
4-F Ph
40.5±1.4 72.9±4.5 >1000
4-OMe
4-OMe
4-OMe
7.4±0.3
5.2±0.9
NA
11.1±1.0 >1000
4.1±1.8
NA
>1000
ND
(S)-CHMe2
SO2
O
3
5% @
uM
3
3
6
7
N
O
N
N
ND
ND
ND
ND
3.28
2.56
N
N
SO2
SO2
1
F
F
O
54% @
N
O
1
9

1
uM
O
O
N
N
O
O
N
3
8
17.1±0.9 20.3±2.4 >1000
2.0
N
SO2
F
6
7% @
uM
N
3
4
9
0
341±12
NA
>1000
NA
2.0
N
SO2
1
F
O
NA
1.29
N
N
N
H
SO2
N
Ph
a
Inhibition data for h/rFAAH and hMGL were determined using throughput fluorescent assays.
IC ± SD values were determined from triplicate experiments. IC values were calculated
b
5
0
50
c
using Prism software (GraphPad). ClogP values were calculated using ChemDraw, version
d
e
1
8.2; ND = not determined; NA=not active at 1 uM.
4
. Evaluation of FAAH inhibitors using the in vitro competitive activity-based protein
profiling
We have assessed FAAH inhibitors 9 and 31 in the activity-based protein profiling (ABPP) assay
6
5
to further evaluate their selectivity. Gel-based ABPP analysis was performed using membrane
homogenates prepared from rat brain tissue and the rhodamine-tagged fluorophosphonate (FP-Rh)
probe that is typically used to profile the serine hydrolase superfamily6, 51, 66. The clinical candidate
FAAH inhibitor PF04457845 was used as the positive control. Protein homogenates (10 mg/mL)
were incubated with compounds 9, 31 (1 and 10 µM), PF04457845 (10 µM) or DMSO for 30 min
at room temperature. The tissues were then treated with 10 µM FP-Rh probe for 45 min at room
temperature. The reaction was quenched using 2x SDS-PAGE loading buffer and separated with
®
SDS-PAGE (12% acrylamide). Fluorescence was detected using Amersham Imager 600 with the
green epi light (520 nm) as the light source. Both inhibitors 9 and 31 completely inhibited FAAH
at 1 and 10 µM without significant inhibition of other serine hydrolases. In our experiments, all
2
0

compounds appeared to share a clean selectivity profile, comparable to that of PF04457845 (Fig.
). These ABPP studies confirmed that 9 and 31 were highly selective for FAAH in the brain, since
none of the other FP-reactive serine hydrolases in this tissue were inhibited by these agents.
ADME Properties
6
5
.
We have evaluated the ADME properties of several
potent (IC <10 nM) and selective (>1000-fold)
5
0
carbamates FAAH inhibitors. Prepared carbamates were
found to be stable (t > 60 min) in human, rat and mouse
1
/2
plasma and in gastric fluids. The microsomal stability
t , min) of both classes of piperidine and piperazine
(
1
/2
carbamates in human, rat and mouse liver microsomal
preparations ranged between 15 to 25 min, with human
microsomal preparations being the most stable species.
We have also evaluated several carbamates in cassette
pharmacokinetics experiments (dose: 2 mg/kg, iv, at 15
min) to rank-order them for brain penetrability.
Figure 6. Gel-based ABPP analysis
was performed using membrane
homogenates prepared from rat brain
tissue and the rhodamine-tagged
fluorophosphonate (FP-Rh) probe.
FAAH inhibitors 9 and 31 completely
inhibited FAAH at 1 and 10 µM
without significant inhibition of other
serine hydrolases. Representative
additional brain serine hydrolases are
designated based on previous ABPP
Based on the findings (Table 3), the tested analogs were
brain permeable by comparing brain and plasma levels
after iv administration at 15 min.
6
studies .
Table 3. Cassette pharmacokinetics experiments of brain permeability
8
6
Compd
5 min IV brain
2
15
1
.11 ± 0.21
0.909 ± 0.158
1
1.23 ± 0.18
0.508 ± 0.034
(
µg/mL)
5 min IV plasma
µg/mL)
0
.746 ± 0.121
0.388 ± 0.062
1
0.474 ± 0.066 0.506 ± 0.056
(
Cassette pharmacokinetics at 2 mg/kg, iv. The compounds were brain permeable by
comparing brain and plasma levels after iv administration at 15 min.
2
1

6
.
Pharmacology
Excitotoxic damage can occur from epilepsy and other seizure disorders, stroke, birth trauma,
traumatic brain injury, fever (especially in children), genetic conditions, and exposure to
convulsive environmental toxins and nerve agents. Effective treatments must be administered
immediately or several hours after an excitotoxic insult to reduce brain deterioration and cognitive
defects, the latter correlating best with synaptic decline in many disorders67, 68. FAAH inhibitors
represent an advantageous pharmacological approach of modulating repair signaling linked to the
“
on-demand action” of endocannabinoids that are elevated in response to neurodegenerative
insults. We tested inhibitors 9 and 31 for their neuroprotective ability by assessing synaptic
markers known to be vulnerable to excitotoxic insults in the hippocampal slice models14, 69, 70 as
well as in different in vivo models of excitotoxicity and related seizure induction71-73
.
Hippocampal slice cultures prepared from rats at postnatal days 12 and 13 were used for the
experiments. The cultured slices express compensatory responses to injury and survival signaling
pathways that are similar to those found in the adult brain. To monitor protection by the carbamate
inhibitors, we applied particular focus on an AMPA receptor subunit (GluR1) as a vulnerable
synaptic component in the KA insult slice model since previous studies found reductions and
altered functionality of AMPA receptor measures correlating with seizure effects after KA
2
2

exposure69, 71, 74 and after soman-induced seizures . The KA insult caused significant reductions
72
A
rat brain
B
C
cultured slice
CA1
CA2
DG
CA3
GluR1
CA1
synaptophysin
hippocampus
CA1
CA2
CA2
DG
DG
CA1
DG
4
00-µm
slices
CA3
CA3
CA3
D
1
h
2 h
24 h
pre-treatment
veh or drug)
KA insult +
veh or drug
post-insult
veh or drug
harvest / analysis
(
E veh
KA +
KA AM11987
F
KA +
KA AM11994
KA +
KA AM11994
G
2
1
1
0000
5000
0000
veh
**
*

GluR1 
SNP


syn IIa 
syn IIb 
syn IIa
syn IIb
5
000
0

KA
-
-
-
+
-
-
+
+
-
+
-
+
AM11994
AM11987

actin 
actin
Figure 7. Synaptic protection mediated by FAAH inhibitors in the hippocampal slice model of
kainic acid (KA)-induced excitotoxicity. Hippocampal tissue from 12-day-old rats was removed and
transverse slices of 400-µm thickness prepared (A), the slices being maintained in six-well culture
plates and exhibiting intact neuronal subfields (B). Stable neuronal density determined by Nissl as
well as pre- and postsynaptic marker staining was consistently found after several weeks in culture (C,
width of view-fields: 1.5 mm). The experimental design includes control slices treated with vehicle
only, as well as a 1-h pre-treatment step with vehicle (veh, 0.01% DMSO), FAAH inhibitor compound
9
(AM11994), or compound 31 (AM11987) before a 2-h KA insult with continued presence of vehicle
or inhibitor (D). After washout of KA, the cultured slices were incubated with vehicle or FAAH
inhibitor for 24 h, followed by harvesting of the brain tissue for subsequent analyses. In order to
evaluate potential synaptic protection by compounds 31 (E) and 9 (F), equal protein aliquots of
homogenized samples from the different treatment groups were assessed by immunoblot for GluR1,
synapsin IIa and IIb (syn IIa and IIb), synaptophysin (SNP), and actin. Integrated optical density
measures for GluR1 are shown as means ± SEM (G). Unpaired t-tests compared to KA insult alone:
**p<0.01 (two-tailed), *p=0.0292 (one-tailed).
in the postsynaptic AMPA receptor subunit GluR1 and in the synaptic vesicle-associated synapsin
IIa and synapsin IIb. Rat hippocampal slice cultures were pre-treated with FAAH inhibitor or with
vehicle (0.1% DMSO) for 1 h. The cultures then were subjected to control treatment (NT) or to an
2
3

excitotoxic insult with 60 μM kainic acid (KA) for 2 h in the presence of vehicle or the
corresponding FAAH inhibitors used for the pre-treatments. After a washout step, the cultured
slices were then incubated with vehicle or correspondence FAAH inhibitor for 24 h. Then, slices
vehicle
KA insult
KA + AM11994
sp
sp
sp
sp
sr
sr
sr
sp
sp
sr
sr
sr
Figure 8. Pre- and postsynaptic protection by FAAH inhibitor 9 (AM11994) in
the excitotoxic hippocampal slice. Mature slice cultures were treated with vehicle
only, or pre-treated with vehicle or FAAH inhibitor compound 9 (AM11994) for 1 h
before the 2-h KA insult step. After washout of KA, slices were incubated with vehicle
or compound 9 for the 24-h post-insult period, then fixed and double immunostained
with anti-GluR1 (red) and anti-synaptophysin (green). Size bars: 100 μm. DG, dentate
gyrus; sp, stratum pyramidale; sr, stratum radiatum.
were harvested into groups of 7-9, sonicated, and equal protein aliquots assessed for AMPA
receptor subunit GluR1, synaptophysin, and synapsin IIb by immunoblot (Fig. 7). FAAH inhibitors
afforded neuronal protection in the hippocampus after KA-induced excitotoxicity with 9 as well
as with 31. Pre- and postsynaptic protection was evident in the dense neuropil of the hippocampal
subfield CA1 (Fig. 8). Due to the protective results produced by 9 in the slice model, it is
noteworthy that initial in vivo studies were conducted with 9 to assess its effects on KA-induced
seizures, and preliminary results found an expected reduction in seizure severity scores
(unpublished observation).
2
4

7
. Conclusions
We have discovered new series of potent and selective carbamate FAAH inhibitors. FAAH is an
enzyme implicated in the hydrolysis of endocannabinoid N-arachidonoylethanolamine (AEA)
which is a partial agonist of CB1 receptors. CB1 receptor activation reduces excessive
glutamatergic signaling and excitotoxic progression, thereby protecting hippocampal cells from
cytoskeletal and synaptic damage. FAAH inhibitors offer safety advantages by augmenting the
anandamide levels “on demand” to promote CB1 protective neuroprotective mechanisms without
the adverse psychotropic effects usually seen with direct and chronic activation of the CB1
receptor. The new inhibitors showed potent nanomolar inhibitory activity against recombinant
human and purified rat FAAH, and were selective (>1000-fold) against serine hydrolases MGL
and ABHD6, while lacked of any affinity for the cannabinoid receptors CB1 and CB2. Evaluation
of FAAH inhibitors 9 and 31 using the in vitro competitive activity-based protein profiling (ABPP)
assay confirmed that both inhibitors were highly selective for FAAH in the brain, since none of
the other FP-reactive serine hydrolases in this tissue were inhibited by these agents.
Our design strategy followed a traditional SAR approach and was supported by molecular
modeling studies based on FAAH cocrystal structures. To rationally design new molecules that
are irreversibly bound to FAAH, we have constructed “precovalent” FAAH-ligand complexes to
identify good ligand binding geometries within the binding pocket of FAAH and then calculated
covalent docking poses by employing Schrodinger methods to select compounds for synthesis.
FAAH inhibitors 9 and 31 were evaluated for neuroprotection in rat hippocampal slice cultures. In
the brain tissue, both inhibitors displayed protection against synaptic deterioration produced by
kainic acid-induced excitotoxicity. Thus, the resultant compounds produced through rational
2
5

design are providing early leads for developing therapeutics against seizure-related damage
associated with a variety of disorders.
7
. Supplementary data
Experimental details for the preparation of all compounds described in Tables 1 and 2 have been
provided.
Acknowledgments. The authors thank Heather Romine and Matthew Montilus for excellent
assistance. This work was supported by the Alzheimer’s Drug Discovery Foundation (ADDF),
Boston, MA [grant number ADDF # 20150503].
8
8
.
Experimental Section
.1 Chemistry
Proton nuclear magnetic resonance spectra were obtained on a VARIAN 400 spectrometer at 500
MHz. Spectra are given in ppm () and coupling constants, J values, are reported in hertz. Splitting
patterns are designated as follows: s, singlet; brs, broad singlet; d, doublet; t, triplet; q, quartet; m,
multiplet. Tetramethylsilane was used as an internal reference standard. Mass spectra were
obtained on a Waters Micromass ZQ spectrometer. HPLC techniques and high-resolution mass
spectrometry were used to determine compound purity. All reagents and solvents were obtained
from commercial suppliers and used without further purification. All non-aqueous reactions were
carried out in oven-dried glassware under an atmosphere of dried argon or nitrogen. The reactions
were monitored by thin layer chromatography (TLC plates F254, Merck) or LC-MS analysis. All
2
6

products, unless otherwise noted, were purified by flash chromatography using a Biotage Isolera
purification system with pre-packed silica cartridges. Purity of all final products was > 96% as
determined by LC-MS using the following protocol. Mobile Phase A = water, B = acetonitrile
solvent gradient 95/5 to 5/95 A:B in 11 min; flow rate 1.5 mL/min; Waters XTerra MS C8 column
(4.6 × 50 mm) with UV detection at 190-400 nm wavelength. The following abbreviations are
used: CDCl , deuterated chloroform EtOAc, ethyl acetate; CH Cl , dichlorometane; MgSO ,
3
;
2
2
4
magnesium sulfate; THF, tetrahydrofuran; NH Cl, ammonium chloride; MeOH, methanol;
4
NaHCO sodium bicarbonate; TFA, trifluoroacetic acid; LC-MS, liquid chromatography with
3
,
mass spectrometry; MS, mass spectrometry NMR, nuclear magnetic resonance spectrometry;
;
TLC, thin layer chromatography.
Method A (Scheme 1)
8.1.1 4-(3-hydroxyphenyl)thiomorpholine 1,1-dioxide (4; X = SO2)
Step a) Water (30 mL), boric acid (30 mol %), and glycerol (6 drops) were added to a stirred
solution of 3-amino phenol 1 (3.2 g, 29.32 mmol) and divinyl sulfone (3.46 g, 29.32 mmol) and
the mixture was refluxed for 3 h. Then, the reaction mixture was extracted with ethyl acetate, and
the organic layer was washed three times with water and dried over MgSO . The solvents were
4
removed under vacuum and the crude product was washed with hot petroleum ether to give 4-(3-
1
hydroxyphenyl)thiomorpholine 1,1-dioxide 6.21 g, as white solid. H NMR (500 M Hz, CDCl ) δ
3
7
3
.26 (t, 1H, J = 8.0 Hz); 6.48 (dd, 1H, J = 8.0 Hz, J = 2.5 Hz); 6.40-6.37 (m, 3H); 5.15 (bs, 1H,);
.85 (t, 2H, J = 5.0 Hz); 3.08 (t, 2H, J = 5.0 Hz).
2
7

8
.1.2 3-(Benzyloxy)-5-bromopyridine
Step b) Potassium carbonate (1.38 g, 10 mmol) was added to a stirred solution of 5-bromopyridin-
-ol (0.860 g, 5 mmol) in DMF (20 mL). Benzyl bromide (1.28 g, 1.43 mL, 7.5 mmol) was added
3
and the resulting reaction mixture was stirred at 80 °C for 6 hours. The reaction mixture was cooled
to room temperature and partitioned between diethyl ether (100 mL) and water (50 mL). The
organic phase was separated, dried over Na SO and evaporated. The residue was purified by flash
2
4
chromatography on silica (Biotage; eluting solvents Hex: EtOAc 1/2 ratio) (1.06 g, 80% yield).
1
H NMR (500 MHz, CDCl ) δ 8.30 (dd, J=4.4, 2.4 Hz, 2H), 7.45-7.33 (m, 6H), 5.08 (s, 2H).
3
8
.1.3 4-(5-(Benzyloxy)pyridin-3-yl)thiomorpholine 1,1-dioxide
Step c). To a stirred solution of 3-(benzyloxy)-5-bromopyridine (1 g, 3.7 mmol) in toluene (25
mL), were added tris(dibenzylideneacetone)dipalladium(0) (169 mg, 0.185 mmol), 2,2'-
bis(diphenylphosphino)-1,1'-binaphthyl (230 mg, 0.37 mmol) and sodium tert-butoxide (710 mg,
7
.4) and the flask was purged with argon. Thiomorpholine 1,1-dioxide (766 mg, 5.68) was added
o
to the mixture and heated to 80 C for 14 hours. After the completion of the reaction (monitored
by TLC) the solvent was removed under vacuum. Water was added (10 mL) and the mixture was
extracted with ethyl acetate (3x30 mL). The combined organic layer was dried over Na SO and
2
4
concentrated under reduced pressure. The residue was purified on silica gel (Biotage; eluting
1
solvents DCM: MeOH 10/1 ratio) to obtain red oil (720 mg, 60% yield); H NMR (500 MHz,
CDCl ) δ 7.96 (m, 2H), 7.41-7.34 (m, 5H), 6.75 (t, J = 2.2 Hz, 1H), 5.10 (s, 2H), 3.85 – 3.83 (m,
3
4
H), 3.08 – 3.06 (m, 4H).
8
.1.4 4-(5-Hydroxypyridin-3-yl)thiomorpholine 1,1-dioxide (3; X = SO , Y = N)
2
2
8

Step d). 10% Palladium-carbon (50 mg) was added to an ethanol (20 ml) solution containing 4-(5-
benzyloxy)pyridin-3-yl)thiomorpholine 1,1-dioxide (500 mg), The resulting reaction mixture was
(
stirred at room temperature in presence H atmosphere for 24 hours. The reaction mixture was
2
filtered through celite bed the bed was washed with ethanol. The filtrate was concentrated under
reduced pressure to afford 4-(5-hydroxypyridin-3-yl)thiomorpholine 1,1-dioxide as colorless
1
liquid ( 350 mg, 98% yield); H NMR (500 MHz, CDCl ) δ 7.87 (d, J = 1.9 Hz, 1H), 7.79 (d, J =
3
2
.3 Hz, 1H), 6.93 (d, J = 2.1 Hz, 4H), 3.83 (d, J = 4.8 Hz, 18H), 3.09 – 3.05 (m, 4H).
8
.1.5 Morpholinophenyl (perfluorophenyl) carbonate (3; X = O; Y = C)
It was prepared according to step d) from 3-bromophenol and 1-methylpiperazine in the presence
1
of Pd (dba) , BINAP and NaOtBu. H NMR (500 MHz, CDCl ) δ 7.29-7.24 (m, 2H),) 6.83-6.81
2
3
3
(m, 1H), 6.77-6.75 (m, 2H), 3.86-3.84 (m, 4H), 3.19-3.17 (m, 4H).
8
.1.6 3-(1,1-Dioxidothiomorpholino)phenyl (perfluorophenyl) carbonate (3; X = SO2)
o
Step e). To a cold (0 C) solution of 4-(3-hydroxyphenyl)thiomorpholine 1,1-dioxide (227 mg, 1
mmol), and N, N-diisopropylethylamine (0.26 mL, 1.5 mmol) in dichloromethane (15 mL) was
added bis(pentafluorophenyl) carbonate (433 mg, 1.1 mmol). The reaction mixture was gradually
allowed to come to room temperature and stirred for 2 h. Then, the reaction was diluted in
dichloromethane (25 mL) and washed with water and brine. The organic extracts were dried over
anhydrous Na SO . The solvents were removed under vacuum and the residue was purified on
2
4
silica gel (Biotage; eluting solvents hexanes: EtOAc 4/1 ratio) to afford 3-(1,1-
dioxidothiomorpholino)phenyl (perfluorophenyl) carbonate as colorless solid (400 mg, 91%
2
9