Discovery of Trifluoromethyl Glycol Carbamates as Potent and Selective Covalent Monoacylglycerol Lipase (MAGL) Inhibitors for Treatment of Neuroinflammation

Article abstract of DOI:10.1021/acs.jmedchem.8b00070

Monoacylglycerol lipase (MAGL) inhibition provides a potential treatment approach to neuroinflammation through modulation of both the endocannabinoid pathway and arachidonoyl signaling in the central nervous system (CNS). Herein we report the discovery of compound 15 (PF-06795071), a potent and selective covalent MAGL inhibitor, featuring a novel trifluoromethyl glycol leaving group that confers significant physicochemical property improvements as compared with earlier inhibitor series with more lipophilic leaving groups. The design strategy focused on identifying an optimized leaving group that delivers MAGL potency, serine hydrolase selectivity, and CNS exposure while simultaneously reducing log D, improving solubility, and minimizing chemical lability. Compound 15 achieves excellent CNS exposure, extended 2-AG elevation effect in vivo, and decreased brain inflammatory markers in response to an inflammatory challenge.

Full text of DOI:10.1021/acs.jmedchem.8b00070

Subscriber access provided by UNIV OF SCIENCES PHILADELPHIA
Article
Discovery of Trifluoromethyl Glycol Carbamates as
Potent and Selective Covalent Monoacylglycerol Lipase
MAGL) Inhibitors for Treatment of Neuroinflammation
(
Laura A. McAllister, Christopher R. Butler, Scot Mente, Steven V. O'Neil, Kari R. Fonseca,
Justin R. Piro, Julie A. Cianfrogna, Timothy L. Foley, Adam M. Gilbert, Anthony R. Harris,
Christopher J. Helal, Douglas S. Johnson, Justin I. Montgomery, Deane M Nason, Stephen
Noell, Jayvardhan Pandit, Bruce N. Rogers, Tarek A. Samad, Christopher L. Shaffer,
Rafael Guimaraes da Silva, Daniel P. Uccello, Damien Webb, and Michael A. Brodney
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00070 • Publication Date (Web): 02 Mar 2018
Downloaded from http://pubs.acs.org on March 3, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discovery of Trifluoromethyl Glycol Carbamates as
Potent and Selective Covalent Monoacylglycerol Lipase
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
MAGL) Inhibitors for Treatment of Neuroinflammation
a
a
a
b
a
Laura A. McAllister* , Christopher R. Butler , Scot Mente , Steven V. O'Neil , Kari R. Fonseca , Justin
a
b
b
b
b
R. Piro , Julie A. Cianfrogna , Timothy L. Foley , Adam M. Gilbert , Anthony R. Harris , Christopher J.
b
a
b
b
b
Helal , Douglas S. Johnson , Justin I. Montgomery , Deane M. Nason , Stephen Noell , Jayvardhan
b
a
a
a
b
Pandit , Bruce N. Rogers Tarek A. Samad , Christopher L. Shaffer , Rafael G. da Silva Daniel P.
,
b
a
a
Uccello , Damien Webb , and Michael A. Brodney .
a
Pfizer Worldwide Research and Development, 610 Main St, Cambridge, MA, 02139, United
b
States; Pfizer Worldwide Research and Development, 445 Eastern Point Road, Groton, CT,
06340, United States.
KEYWORDS: Monoacylglycerol Lipase, MAGL, neuroinflammation, covalent inhibitors
ABSTRACT Monoacylglycerol lipase (MAGL) inhibition provides a potential treatment
approach to neuroinflammation through modulation of both the endocannabinoid pathway, and
arachidonoyl signaling in the central nervous system (CNS). Herein we report the discovery of
compound 15 (PF-06795071), a potent and selective covalent MAGL inhibitor, featuring a novel
trifluoromethyl glycol leaving group which confers significant physicochemical property
improvements as compared with earlier inhibitor series with more lipophilic leaving groups. The
design strategy focused on identifying an optimized leaving group that delivers MAGL potency,
1
ACS Paragon Plus Environment

Page 3 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
serine hydrolase selectivity, and CNS exposure; while simultaneously, reducing LogD,
improving solubility, and minimizing chemical lability. Compound 15 achieves excellent CNS
exposure, extended 2-AG elevation effect in vivo, and decreased brain inflammatory markers in
response to an inflammatory challenge.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Introduction
Monoacylglycerol lipase (MAGL) is a serine hydrolase which regulates the endocannabinoid
1-6
pathway and arachidonoyl signaling in the central nervous system (CNS) (Figure 1a).
Emerging biological data suggests that inhibiting MAGL may be a promising therapeutic
approach to multiple diseases, particularly those in which neuroinflammation plays a central
7-10
role. Neuroinflammation occurs as a consequence of disease or injury resulting in activation of
resident immune cells (microglia and astrocytes), production of pro-inflammatory cytokines and
11
chemokines, and deterioration of the blood brain barrier. Persistent neuroinflammation can
12
result in neuronal death and ultimately contribute to neurodegeneration. Inhibition of MAGL
1
3
has therefore been proposed as a potential therapeutic approach for Alzheimer’s disease (AD),
1
4
15
16-19
Parkinson’s disease (PD), multiple Sclerosis (MS), pain and acute brain injury.
MAGL is highly expressed throughout the body, including the brain, liver and lungs. In the
brain, MAGL has a central role in controlling levels of the endocannabinoid 2-
arachidonylglycerol (2-AG), and the pro-inflammatory eicosanoid precursor arachidonic acid
20
(
AA). MAGL catalyzes the transformation of 2-AG to AA (Figure 1b), which can be further
converted to pro-inflammatory eicosanoids. MAGL serves as the primary regulator of AA in the
brain, controlling >85% of 2-AG metabolism, in contrast to the periphery, where AA is primarily
derived from phospholipids via cytosolic phospholipase A2 (cPLA2). Arachidonic acid (AA) is
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
also utilized by COX enzymes for the synthesis of inflammatory prostaglandins and
thromboxanes, thus lowering AA levels results in further anti-inflammatory effects. Moreover,
MAGL is located presynaptically at axon terminals, and co-localized with the cannabinoid
receptor 1 (CB1). As a potent CB1/2 agonist, 2-AG activates the endocannabinoid system,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
21
which has been reported to have beneficial impact on inflammation. Therefore, this dual role in
controlling neuroinflammation through central 2-AG and AA regulation supports wide ranging
therapeutic potential for MAGL inhibitors.
Figure 1. (a) 2-AG plays a central role in endocannabinoid pathway signaling and arachidonyl
inflammatory cascade in the brain. (b) MAGL catalyzes the hydrolytic conversion of 2-
arachidonyl glycerol (2-AG) to arachidonic acid (AA).
MAGL is a membrane bound serine hydrolase, tethered intracellularly by an anchoring α-
2
2
helical region. The substrate binding site comprises a large hydrophobic tunnel which forms the
acyl-chain binding region (ACB pocket), accommodating the long aliphatic chain of the natural
3
ACS Paragon Plus Environment

Page 5 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substrate 2-AG. At the head of the ACB pocket is a catalytic triad of amino acids (Ser122,
His269 and Asp239). The triad activates Ser122, which nucleophilically attacks the carbonyl of
2-AG. An oxyanion hole, comprised of two backbone amide residues in this region, stabilizes
the developing negative charge on the carbonyl in the transition state en route to the tetrahedral
intermediate during catalysis. Beyond the catalytic triad is the cytoplasmic access channel (CA
channel), that consists of more polar residues, ultimately leading to the cytoplasm. The polar
residues of this channel complement the hydrophilic glyceryl moiety of 2-AG, and upon
hydrolysis, the expelled glycerol molecule presumably exits the enzyme via this channel.
A covalent inhibition approach is one strategy to deliver sustained inhibition of MAGL at a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
feasible human dose. This approach was deemed to be the most viable for achieving good
MAGL potency, and the high levels of sustained inhibition that were targeted for efficacy in
neuroinflammatory conditions. Reversible inhibitors of MAGL with sub-micromolar potency
2
4-25
have been recently reported.
However, a reversible profile is desirable when the objective is
to achieve more transient MAGL blockade. In addition, achieving the desired MAGL potency in
a desirable property space has proved challenging with reversible inhibitors. Several series of
covalent MAGL inhibitors have emerged in the last decade, and some representative examples
2
6-36
are shown in Figure 2a.
A typical feature of these inhibitors is an amine core linking to a
lipophilic tail region, needed to complement the lipophilic ACB pocket. The amine core is
capped with a carbamate or urea head group featuring a leaving group, which is generally an
2
2
alcohol or heterocycle. This moiety presumably binds in the CA channel of the active site, or
“leaving group pocket” prior to covalent modification. Upon binding of the inhibitor to the
MAGL active site, Ser122 attacks the carbonyl carbon, triggering leaving group departure. This
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
results in a stable adduct with the core and tail region of the inhibitor molecule, rendering the
enzyme incapable of further degradation of 2-AG (Figure 2b).
The hydrophobic nature of a large section of the MAGL binding pocket demands significant
lipophilicity for binding and inhibition. To complement the highly lipophilic nature of the ACB
pocket, many potent inhibitors reside in a highly lipophilic space (cLogP > 4). As a result, these
compounds often have poor solubility, making evaluation of in vivo efficacy and further
development challenging. Moreover, metabolic clearance also increases with lipophilicity,
which leads to higher requisite doses to achieve efficacy. Molecules in these higher cLogP
ranges are also considered non-ideal because of their higher propensity for associated off-target
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
7
promiscuity and safety risks.
A
common motif employed in reported covalent MAGL inhibitors is the
hexafluoroisopropanol (HFIP) leaving group (Figure 2a, compounds 1-3). This group confers a
favorable potency and selectivity profile, providing a balance between the levels of reactivity
required for reaction with Ser122 at the MAGL active site, with low enough inherent chemical
lability to avoid promiscuous serine hydrolase inhibition. However, a major disadvantage of this
moiety is its high lipophilicity, which exacerbates the already requisite lipophilic regions of the
ACB pocket. This has the potential to result in compounds with high clogP, which often
correlates with high clearance and poor solubility.
The development of MAGL inhibitors that could treat neuroinflammation resulting from acute
traumatic or non-traumatic brain injury would potentially require delivery via intravenous (IV)
route of administration. To that end, our goal was to develop highly soluble MAGL inhibitors for
the treatment of these acute neurological conditions. Incorporation of the HFIP moiety would
increase the challenge of achieving such a profile, with the resulting compounds having
5
ACS Paragon Plus Environment

Page 7 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
38
unfavorably high clogP, poor solubility, high clearance, and poor lipophilic efficiency (LipE).
We sought to identify a new series of brain penetrant MAGL inhibitors with improved solubility,
and thus more amenable to IV formulation. The requirement to reduce LogD to improve
solubility needed to be balanced with the requirement to avoid transporter liabilities (PgP and
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
39
BCRP) to maintain good CNS penetration. Since the MAGL pocket contains more polar
residues in the region corresponding to the leaving group binding pocket (the CA channel), the
design strategy focused on modifying the leaving group to effectively balance these properties. A
further objective in the design of the leaving group was to engender potency from a balance of
target-specific binding efficacy and reactivity, rather than simple non-specific chemical
40
reactivity. Minimizing inherent chemical reactivity of the carbamate is one approach to avoid
indiscriminate reactivity with other off-target proteins, and to optimize selectivity for MAGL
versus other related serine hydrolases (e.g. fatty acid amide hydrolase (FAAH)). A further
consideration associated with covalent approaches, which is simultaneously addressed by
minimizing chemical reactivity, is the desire to mitigate the risk of idiosyncratic adverse drug
reactions (IADRs). It has been reported that minimizing total daily covalent body burden (via
low dose), and reducing the intrinsic electrophilic activity may contribute to reducing IADR
4
1-44
safety risks in covalent drugs.
Herein we report the identification of compound 15 (PF-06795071), a covalent MAGL
inhibitor featuring a novel, stereodefined, trifluoromethyl glycol leaving group. This compound
shows a comparable potency and selectivity profile to compounds containing the more lipophilic
and chemically labile HFIP leaving group. This compound shows significantly improved
solubility compared with earlier leads, improving the feasibility of our objective of a MAGL
inhibitor suitable for IV dosing. Compound 15 shows excellent selectivity against FAAH and
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
other serine hydrolases, demonstrates good brain penetration, and a clean in vitro safety profile.
Furthermore, compound 15 is efficacious in sustaining elevation of 2-AG levels in vivo, and in
reducing markers of neuroinflammation in animal models.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. (a) Representative examples of covalent MAGL Inhibitors with the
heaxfluoroisopropanol (HFIP) and other leaving groups. (b) MAGL inactivation by covalent
inhibitors.
7
ACS Paragon Plus Environment

Page 9 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Results and Discussion
We previously reported a series of pyrazole azetidine analogs such as compound 7 (Figure 3),
45
displaying excellent MAGL potency (MAGL IC = 0.3 nM) and a good selectivity profile. An
5
0
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
expectation of irreversible MAGL inhibitors is that one should observe an extended in vivo
pharmacodynamic (PD) effect of MAGL inhibition (raised 2-AG levels). Upon testing analogs
from this azetidine series in mouse in vivo PK/PD studies, we observed only a transient elevation
in 2-AG. Cellular activity based protein profiling (ABPP) studies confirmed that this class of
compounds covalently inhibited MAGL transiently and led us to speculate that the azetidine
adducts formed at the MAGL site are rapidly hydrolyzed from the enzyme. We observed that
analogs containing larger ring cores, such as piperidines, resulted in improved MAGL adduct
stability and thus prolonged PD effects.
Recent efforts to optimize the core system of these inhibitors resulted in the discovery of
compounds based on a [3.1.0] bicyclic core system, such as compound 8 (Figure 3). Compound
8
retains MAGL potency (MAGL IC = 1.4 nM) and has excellent FAAH selectivity (FAAH
50
IC = 4.3 µM, Table 1). Despite a promising pharmacology profile, compound 8 resides in a
5
0
high LogD space, and has poor kinetic solubility (< 1µM). To address these challenges, we
systematically explored leaving groups with the goal of improving drug-like properties. To this
end, a series of analogs that varied the alcohol leaving group in lower LogD space was prepared
(Table 1).
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Design of MAGL Inhibitors with improved adduct stability and lower LogD
Synthesis of MAGL Inhibitor Leaving Group Analogs. Synthesis of analogs based upon
the [3.1.0] system core with various leaving groups was achieved using the general strategy
outlined in Scheme 1. Beginning with Boc-protected 3-azabicyclo[3.1.0]hexane-6-carboxylic
acid 20, the carboxylic acid group was converted to the corresponding methyl ketone 21 via
Weinreb amide formation and Grignard addition. Subsequent conversion to pyrazole 22,
followed by copper-catalyzed N-arylation with 4-fluorophenyl boronic acid afforded Boc-
protected N-aryl pyrazole compound 23. Removal of the Boc protecting group provided the key
amine intermediate 24. The reaction of 24 with an activated carbonate or chloroformate derived
from an appropriate alcohol (corresponding to the desired leaving group) afforded the analogs in
Table 1 and 2.
9
ACS Paragon Plus Environment

Page 11 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of carbamate analogs in Table 1 and 2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
o
(
a) i. NHMeOMe.HCl, EDCI, HOBT, DIPEA, CH Cl 0 C; ii. MeMgBr, THF, 0 C, 99%
2 2,
o
o
over 2 steps; (b) i. DMF-DMA, DMF, 110 C; ii. NH NH , EtOH, 80 C, 16 h, 94%; (c) 4-F
2
2
o
phenylboronic acid, Cu(OAc) , pyridine, 4Å MS, CH Cl , 30 C, 18h, 60%; (d) CF CO H, rt.;
2
2
2
3
2
(e) chloroformates or activated carbonates .
In vitro Assays. MAGL potency was evaluated in vitro by measuring inhibition of turnover of
a fluorogenic substrate (7-HCA, 7-hydroxycoumarinyl arachidonate) by recombinant MAGL,
46
following 30 minute incubation with inhibitors. While this method allows for rank ordering of
potency, results are dependent on pre-incubation time. kinact/K ratios were measured for more
I
accurate characterization of irreversible inhibition; a greater kinact/K value indicates a more
I
47
effective inhibitor. kinact/KI ratio accounts for binding efficiency of the entire inhibitor
molecule in the pre-covalent state (within the K parameter), as well as the rate of the covalent
I
attachment step (within kinact). kinact/K values for MAGL inhibitors shown in Table 1 correlate
I
well with respective IC50 measurements. FAAH inhibition potency was assessed using
10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
recombinant FAAH and a fluorescent substrate (arachidonoyl-AMC), following a 30 minute pre-
incubation with compounds.
Structure Activity Relationships. Table 1 shows the impact of combining various leaving
groups with compounds based on a [3.1.0] bicyclic core system on MAGL inhibition profile and
properties. As a reference, several analogs containing the literature alcohol leaving groups were
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
evaluated (compounds 8-10). The calculated pK of the leaving group alcohol was used as an
a
indicator of relative chemical reactivity of the groups. Compound 8 with hexafluoroisopropanol
(
HFIP) as the leaving group showed excellent MAGL potency (IC = 1 nM) and selectivity
50
versus FAAH (IC = 4.3 µM). In contrast, while the p-nitrophenol-derived analog 9 is potent at
5
0
MAGL (IC50 = 24 nM), it lacks selectivity over FAAH (IC50 = 18 nM). The N-
hydroxysuccinimidyl carbamate 10 is significantly less potent despite greater reactivity to HFIP.
A potential explanation may be less favorable binding of this group in the CA pocket.
Using the HFIP alcohol as a starting point, small modifications to reduce LogD were designed
and prepared. Replacement of one trifluoromethyl group with a simple methyl in compound 11
resulted in a significant potency reduction (MAGL IC = 242 nM), likely attributable to reduced
5
0
reactivity of this group as suggested by the pK . Interestingly, acceptable FAAH selectivity
a
levels were largely maintained (4300 fold for 8 vs > 95 fold for 11). Removal of the branched
alkyl group altogether resulted in a simple trifluoroethyl leaving group in compound 12 that
exhibited significantly reduced MAGL potency, and interestingly resulted in significant erosion
of FAAH selectivity. Comparison of compound 11 and 12 would suggest that branching in the
leaving group pocket is an important structural feature for maintaining selectivity over FAAH.
This is consistent with an earlier report on the affinity of anandamide analogs for rat anandamide
48
amino hydrolase, where branched analogs disfavored binding. Encouragingly, compound 13
11
ACS Paragon Plus Environment

Page 13 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
which contains a tetra-fluorinated leaving group, corresponding to one LogD unit reduction in
lipophilicity relative to 11, regained potency at MAGL (IC = 26 nM) and maintained FAAH
5
0
selectivity (73 fold).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
With these results in hand, and inspired by the structure of the natural substrate for MAGL (2-
AG), we designed compound 14, in which one highly lipophilic trifluoromethyl group is
replaced with a more polar hydroxyl group, resulting in a LogD reduction of almost two units
(Figure 4). It was postulated the polar “OH” group should be tolerated in this region of the CA
pocket, since the glycerol moiety in 2-AG can bind in this region. Gratifyingly, we observed that
racemic glycol analog 14 retained excellent potency at MAGL (IC50 = 6 nM), excellent
38
selectivity against FAAH (>2000-fold), and improved lipophilic efficiency (LipE = 4.3) as
compared with the HFIP analog 8 (LipE 3.4).
12
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 1. Leaving Group SAR
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Leaving
Group
pKa
MAGL
k
/K
FAAH
d
f
inact
I
Compound
-OR1
a
-1 -1 b
c
LogD
LipE
IC (nM)
(M s )
IC (µM)
g
5
0
50
4
7
-
-
5.6
16173
>11
5.1
5.1
3.1
4.4
7.2
9.8
0.33
1
650964
41708
0.66
8
9
4.3
5.6
3.4
3.4
9.8
7.2
e
24
46
2879
156
0.018
4.2
1
0
e
3.6
2.6
4.7
7.8
1
1
1
1
1
2
3
4
242
445
26
ND
103
>23
0.97
1.9
4.8
4.2
4.3
3.9
1.8
2.2
3.3
4.3
12.5
12.5
11.2
11.7
1578
12966
6
14
a
IC values measured from MAGL inhibition assay. Activity measured after 30 min pre-
5
0
incubation with compound. Values represent the geometric mean of at least three experiments.
b
c
9
5% confidence intervals are provided in the supporting information.
kinact/K ratio. IC
I 50
values measured from FAAH inhibition assay. Values represent the geometric mean of at least
d
three experiments. Measured eLogD values. eLogD is an HPLC based method for measuring
49
e
50
f
LogD . SFLogD measured at pH 7.4 . Lipophilic efficiency (LipE) defined as MAGL pIC -
5
0
g
LogD. Calculated pK of alcohol leaving group (ACD labs).
a
1
3
ACS Paragon Plus Environment

Page 15 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Design of glycol moiety in compound 14 inspired by 2-AG, the natural substrate of
MAGL.
Encouraged by the promising profile of compound 14, closely related analogs of 14 were
synthesized and profiled, as shown in Table 2. Separating enantiomers afforded the more active
compound 15, which had the expected improvement in potency (MAGL IC 3 = nM, k /K =
5
0
inact
I
-1 -1
29,724 M s ) consistent with a positive binding interaction with the pocket. Enantiomer 16 was
less potent (MAGL IC = 890 nM), which reinforced the suggestion that the MAGL inhibition
5
0
did not predominantly result from leaving group reactivity alone (wherein the enantiomers would
exhibit the same potency), but rather by making a positive binding interaction in the pocket.
Capping the hydroxyl group with a methyl group (compound 17), or homologating the glycol to
position the hydroxyl further from the carbamate (compound 18) results in a significant potency
drop. Modifying the position of trifluoromethyl substituent on the glycol (compound 19) also
diminishes MAGL activity. The glycol leaving group in compound 15 was an intriguing
advance, because it attained similar levels of MAGL inhibition to compound 8, despite the glycol
leaving group (pK = 11.7) being significantly less reactive than the hexafluoroisopropanol in 8
a
14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
pK = 9.8). We postulated that the hydroxyl group therefore made a favorable binding
a
interaction in the leaving group pocket in the pre-covalent binding step, relative to the HFIP.
It was postulated that the pK of the alcohol could be used as a leading indicator of potency, as
a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
it is a reflection of inherent reactivity. To understand the relationship between MAGL pIC50
versus the pK of the leaving groups, a number of analogs in the series were plotted (Figure 5).
a
For most analogs, MAGL potency (pIC ) correlates with pK of the leaving group, with more
50
a
acidic leaving groups having greater potency. This trend indicates that electrophilicity of the
carbamate carbonyl (correlating with pK ) is the major driving force for potency, and suggests no
a
specific interaction (beneficial or detrimental) of the molecule with the leaving group pocket. In
the case of active glycol enantiomer 15, MAGL inhibition potency is greater than would be
expected for this pK range.
a
In addition to desirable MAGL potency, compound 15 retains good selectivity over FAAH
(>1000-fold), good lipophilic efficiency (LipE = 4.7), and is in a more desirable property space
(
LogD = 3.8) than its HFIP analog (8, LogD = 5.6). Encouragingly, 15 had an apparent kinetic
solubility of 115 µM at pH 7.4, which was a considerable enhancement compared with
compound 8 with apparent kinetic solubility of <1 µM at the same pH. A crystalline form of
compound 15 had a thermodynamic solubility of 25 µM at pH 6.5 (after 24 h). Furthermore, we
were encouraged that compound 15 could potentially have low in vivo clearance as oxidative
turnover in human liver microsomes is low (HLM Cl int < 8 ml/min/kg).
15
ACS Paragon Plus Environment

Page 17 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. SAR of Trifluoromethyl Glycol Related Analogs
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
MAGL a
k
/K
FAAH
d
e
HLM CLint f
(mL/min/kg)
inact
I
Compound
-OR1
-1 -1 b
LogD LipE
c
IC50 (nM)
(M s )
IC (µM)
5
0
15
3
29724
3.1
3.8
3.9
4.2
4.2
4.0
4.7
2.2
2.3
2.0
2.9
<8
<8
1
1
1
1
6
7
8
9
890
354
582
132
79
110
4573
244
26
g
1.94
>12
>30
nd
<12
9.4
a
IC values measured from MAGL inhibition assay. Activity measured after 30 min pre-
5
0
incubation with compound. Values represent the geometric mean of at least three experiments.
95% confidence intervals are provided in the supporting information.
b
c
kinact/K ratio. IC50
I
values measured from FAAH inhibition assay. Values represent the geometric mean of at least
d
three experiments. Measured eLogD values. eLogD is an HPLC based method for measuring
e
f
LogD. Lipophilic efficiency (LipE) defined as MAGL pIC -LogD. Scaled intrinsic clearance
5
0
g
in human liver microsomes. nd = not determined
16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. MAGL pIC of inhibitors from the [3.1.0] series vs pKa of alcohol leaving group.
5
0
Asymmetric Synthesis of Compound 15. In order to further profile compound 15, an efficient
asymmetric synthetic route to the glycol leaving group was developed. This synthetic route is
outlined in Schemes 2 and 3. The initial synthesis of PMB-protected chiral glycol 27 was
achieved by nucleophilic addition of p-methyoxylbenzyl alkoxide to commercially available
racemic epoxide 25 followed by chiral chromatography (Scheme 2). Alternatively, racemic
epoxide 25 can be resolved using a Jacobsen hydrolytic kinetic resolution reaction, to give the
17
ACS Paragon Plus Environment

Page 19 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
51
desired enantioenriched (R) epoxide (26) in >98% ee. Ring opening of chiral epoxide 26 with
p-methyoxylbenzyl alcohol affords the desired enantiomer of 27 without the need for chiral
chromatography.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Synthetic route to protected chiral glycol intermediate
(a) (S,S)-(Salen)Co(III)acetate complex, 0.45 equiv. H O, then distillation (>98% ee (R)); (b)
p-methoxybenzyl alcohol, NaHMDS, THF, 41% (25 to 27).
2
With chiral glycol intermediate in hand, compound 15 can be synthesized as outlined in
Scheme 3. PMB protected alcohol 27 can be converted to the NHS activated carbonate 28 as
shown in Scheme 3, by activation with N,N'-disuccinimyl carbonate. Coupling the NHS
activated carbonate intermediate 28 to the pyrazole amine 24 and subsequent PMB deprotection
provided desired compound 15 in excellent yield.
18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 3. Synthesis of compound 15
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
(a) N,N'-disuccinimyl carbonate, NEt , CH Cl , 18h, 25 C (in situ); (b) 28, CH Cl , rt, 18h,
3 2 2 2 2
6
1% over 2 steps; (c) CF CO H, rt., 4h, 65%.
3 2
X-Ray Co-crystal structures and MAGL Docking Models. Irreversible modification of
MAGL by this series of inhibitors was confirmed through x-ray co-crystal structures. Figure 6
illustrates a MAGL co-crystal structure obtained following treatment with NHS carbamate 10.
The structure confirmed covalent modification of Ser122 in the active site by the inhibitor and
demonstrates that the [3.1.0] core with aryl pyrazole tail of the inhibitor complements the shape
of the lipophilic MAGL acyl chain binding (ACB) pocket. Although we did not attempt to co-
crystalize compound 15 itself with MAGL, we inferred covalent modification of MAGL by
compound 15 from its time-dependent inhibition profile in vitro assay systems, and assume an
identical adduct is formed to that with compound 10.
19
ACS Paragon Plus Environment

Page 21 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. X-ray co-crystal structure of compound 10 bound to MAGL. Crystal structure shows a
post-covalent modification snapshot of binding in the MAGL pocket, with leaving group no
longer attached (PDB Accession code: 6BQ0).
Unfortunately, co-crystal structures were not helpful in prospectively guiding structure-based
design of the leaving group, as the leaving group has already departed in the snapshot of the
protein in a post-covalent modification state. Therefore, in order to understand binding and drive
leaving group design, molecular modeling efforts focusing on the construction of a “pre-
covalent” MAGL-ligand complex were pursued. This was accomplished via modeling of
compound 15 into the x-ray structure with the catalytic Ser122 repositioned to that of the
published non-covalent MAGL-ligand complex (PDB: 3PE6). The approximate positioning of
20
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
52
the leaving group was then found via docking and subsequent minimization in this MAGL
active site.
Docking of compound 15 into the MAGL site with Ser122 placed in the pre-covalent state is
shown in Figure 7a and 7b (rotated view). The docked “pre-covalent” compound 15 is shown in
green sticks, and can be compared to the X-ray of the post-covalent state shown in purple lines.
Here the carbonyl (facing away in Figure 7a, visible in Figure 7b) is oriented toward the two
backbone NH donors of Ala51 and Met123, which forms the oxyanion hole. In these docking
poses, the carbonyl-NH distance is closer (2.1 Å) for the Ala51 interaction compared to the
Met123 (3.1 Å) due to the steric blocking of Ser122. Presumably, as the reaction proceeds to the
transition state, the distance from the Ser122 oxygen to the ligand carbon will decrease, and the
second oxyanion interaction distance will also shorten. In the structure of the covalent ligand,
these distances are shortened to about 2.0 Å for both of these interactions.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
An additional prediction is that the terminal glycerol -OH can form a polar interaction with
His269, which forms, along with Arg57 and Tyr194 the polar residues of the CA channel via
which glycerol molecules (or other leaving groups) exit the protein. The CF moiety does not
3
appear to contribute to any favorable binding interactions. Rather, we believe its primary
contribution to MAGL activity is manifest in the reactivity component, via the inductive effect of
electronegative fluorines on the C=O bond.
The other half of the “pre-covalent” ligand occupies a space in the ACB pocket which is
similar in orientation, but twisted about 60 degrees relative to the covalent ligand. Due to the
change in position of Ser122, it is not possible to fit the carbonyl effectively in the oxyanion
hole, while simultaneously preserving the exact position of the ‘core’ piece of the molecule to
that of the covalent x-ray. It is our hypothesis that active MAGL ligands must be able to
21
ACS Paragon Plus Environment

Page 23 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
effectively fit into multiple conformations in the ACB area in order to not hinder this rotation as
the reaction proceeds. That the ACB pocket is lined with predominately hydrophobic residues (6
Leucine residues, 1 each Met, Phe and Ile) likely reflects the need for this pocket to favor non-
specific binding, thus keeping energy wells for twisting shallow. This is further supported by the
assortment of X-ray structures which show that a variety of binding poses can be accommodated.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Docking model of compound 15 binding to MAGL in the pre-covalent state (green
sticks). A superimposed X-ray structure of the post-covalent modification state is shown (purple
lines- PDB Accession code: 6BQ0) (a) Shows predicted interaction of the leaving group “OH”
with His239 of the catalytic triad. (b) Rotated view shows the predicted interaction of carbonyl
group with backbone “NH” groups from Ala57 and Met123.
In vitro selectivity and in vitro safety profile. Serine hydrolase selectivity of compound 15
53
was tested in an in vitro panel of selected serine hydrolases (Figure 8). Compound 15 exhibits
22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a comparably clean selectivity profile compared to analogs containing previously described
leaving groups, such as HFIP (compound 8) and NHS (compound 10) analogs. Compound 15
shows complete inhibition of MAGL without significant inhibition of other serine hydrolases,
with the exception of carboxyesterase 1(CES1), which is inhibited at 80% when tested at 10 µM.
In contrast, HFIP analog 8 shows complete inhibition of MAGL, but also ABHD6 (95%) and
CES1 (58%) at the same dose. NHS carbamate analog 10 had weaker off-target activity at a
number of serine hydrolases, including the carboxyesterase enzymes (CES1 and CES2),
ABHD6, FAAH and FAAH2.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. Heat map showing serine hydrolase inhibition profile for MAGL inhibitors dosed at
1
0µM against a panel of 42 human serine hydrolase enzymes (% Inhibition). A full list of serine
hydrolase enzymes tested and method is provided in SI, table S2.
23
ACS Paragon Plus Environment

Page 25 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compound 15 was also profiled for general off-target pharmacology in a
binding/functional panel comprising a number of enzymes, transporters, receptors, and ion
channel targets (CEREP). At a concentration of 10 µM, compound 15 hit only a single target,
showing functional activity at the cannabinoid 1 receptor (EC50 = 1.5 µM). Follow up
radioligand binding studies demonstrated a lack of activity at CB1, suggesting that the compound
does not engage CB1 directly. An indirect functional CB1 agonism effect would be expected for
a MAGL inhibitor due to the expected increase in 2-AG (a CB1 agonist). In addition, absence of
binding at the hERG channel (IC50 >30 µM), also suggests a low risk of cardiovascular QT
prolongation, further increasing the confidence in the safety profile of compound 15.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In vitro metabolism and pharmacokinetic properties of compound 15. With the desired in
vitro pharmacology and safety profile for compound 15, the in vitro ADME properties were
evaluated to assess suitability for in vivo studies (Table 3). In vitro screening data indicated high
passive permeability, as measured by apparent passive permeability (P ) in the Madin-Darby
app
-6
canine kidney low efflux cell line (MDCKII-LE; RRCK cell permeability assay = 21 x 10
cm/s). In addition, compound 15 was assessed for efflux via human P-glycoprotein (MDCKII-
LE transfected with human MDR1) and mouse breast cancer-resistant protein (MDCKII-LE
54
tranfected with mBCRP) using well-established transwell assay procedures. The transporter
efflux ratios were in the low range indicating minimal interaction with either transporter;
increasing the probability of achieving sufficient brain penetration.
Compound 15 was evaluated in human in vitro clearance screening assays, including human
liver microsomes and cryopreserved human hepatocytes (HHEPs) monitoring for substrate
55
depletion. In human liver microsomes (HLM), turnover of the compound was not observed
24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
HLM scaled Cl < 8 mL/min/kg); however, in HHEPs the turnover observed was significantly
int
higher (HHEPs scaled Clint = 52 mL/min/kg). Taken together, the data suggested that
cytochrome P450 (CYP)-mediated oxidative clearance was minimal with substantial
involvement of a non-CYP-mediated clearance pathway. The hydroxyl group in compound 15 is
a potential site for direct conjugation by enzymes such as uridine 5’-diphospho-
glucuronosyltransferase (UGT) and/or sulfotransferase (SULT). To further explore the clearance
pathways, compound 15 was incubated in HHEPs in the presence of aminobenzotriazole (ABT)
and tienilic acid which result in pan-CYP inhibition. In the presence of the CYP inhibitors the
turnover of compound 15 was only partially inhibited as compared with the control incubation
without CYP inhibitors, indicating that the predominant clearance metabolism was non-CYP
mediated with a minor oxidative component. The biotransformation of compound 15 was
performed using HHEPs and the major metabolite identified was a direct glucuronide conjugate,
supporting the hypothesis that compound 15 is cleared predominantly through direct
glucuronidation on the hydroxyl of the leaving group. Compound 15 was also evaluated for
competitive inhibition of five CYP enzymes (CYP3A4, 2D6, 2C9, 2C8, and 1A2) in screening
assays using relevant probe substrates. The observed CYP inhibition was < 10% at 3 µM for
each of the isoforms, indicating a low risk for drug-drug interactions mediated through
competitive inhibition.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
25
ACS Paragon Plus Environment

Page 27 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Summary of chemical properties and in vitro ADME characteristics of compound 15
/
ꢀꢁꢂꢃꢄꢅꢆ tꢇꢈꢉꢁꢇꢊꢃꢁꢋ
Lꢌ ꢍꢃꢊꢇꢈ !5a9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
.
ꢒ
ꢊ
ww/Y tꢊꢑꢑ ꢕ+ ,- ꢃꢆꢗꢌ"
&,
,ꢓ&
,ꢓ#
6%
aí
t{!
#
$$
ꢏ
a5w, ꢀ00ꢍꢎ+ ꢅꢊꢐꢋꢁ
ꢒ%
ꢃ
ꢆꢘ/wt ꢀ00ꢍꢎ+ ꢅꢊꢐꢋꢁ
ꢀ
[ꢁꢂ5
#
&
ꢓ%
ꢓ'
ꢇ
I[a /[ꢋꢉꢐ2ꢌꢃꢊꢍꢀꢇ ꢕꢆ[ꢗꢆꢋꢉꢗ3ꢂ"
ꢃ[ꢁꢂt
ꢀ
II9t /[ꢋꢉꢐ2ꢌꢃꢊꢍꢀꢇ ꢕꢆ[ꢗꢆꢋꢉꢗ3ꢂ"
II9t /[ꢋꢉꢐ2ꢌꢃꢊꢍꢀꢇ ꢕꢆ[ꢗꢆꢋꢉꢗ3ꢂ"ꢕꢙ!ꢘÇ"
ꢔ&
Çꢄꢀꢅꢆꢁꢇꢈꢉꢊꢆꢋꢃ
ꢌꢁꢍꢎꢏꢋꢍꢋꢐꢈ ꢑI ꢒꢓꢔ ꢕꢖa"
&
ꢔ
0
#'
a
b
-
Permeability in the MDCKII-LE cell line. Ratio of permeability, measured as a rate in 1 x10
6
cm/sec in and out (BA/AB) in the MDCKII-LE cell line transfected with MDR1 (human P-
c
-6
glycoprotein). Ratio of permeability, measured as a rate in 1 x 10 cm/sec in and out (BA/AB)
in the MDCKII-LE cell line transfected with mouse BCRP. Scaled intrinsic clearance in human
liver microsomes. Scaled intrinsic clearance in human hepatocytes. Scaled intrinsic clearance
in human hepatocytes in the presence of CYP inhibitors. PSA = polar surface area
d
e
f
With an intriguing in vitro profile, in vivo pharmacokinetic (PK) properties were assessed. As
we were targeting an intravenous therapy, IV PK studies in rat, dog, and non-human primate
(nhp) were conducted (Table 4). Following IV administration, compound 15 exhibited a
moderate to high plasma clearance (CL ) and a moderate volume of distribution (V ) cross-
p
d
species. The resulting half-life was short, ranging from ~1-2 h, although sufficient for in vivo
evaluation. Compound 15 was dosed subcutaneously to mice followed by collection of plasma
and brain samples in a time course manner (Table 4). No impairment to brain penetration was
observed, as evidenced by the AUC(0-Tlast)-derived unbound brain to unbound plasma ratio
(
C /C ) of ~1.4. We therefore concluded that compound 15 had an adequate PK profile to test
b,u p,u
central MAGL inhibition efficacy in vivo.
2
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 4. In vivo pharmacokinetics of compound 15
a
Administration
CL
(mL/min/kg)
p
d
e
Species
b
c
V (L/kg) T (h)
d
C /C
C /C
b,u pu
1/2
b
p
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Route
rat
IV
IV
IV
SC
61
3.1
1.6
1.5
nd
1.3
1.7
1.7
nd
nd
nd
nd
2.4
nd
nd
nd
1.4
dog
nhp
21
16
nd
mouse
a
Rat, dog, and nhp studies conducted at 1 mg/kg (n = 2 per species), mouse study conducted at
.0 mg/kg. IV = intravenous, SC = subcutaneous. Plasma clearance. AUC
b
c
d
1
-derived ratio
(
0-Tlast)
e
of the drug concentration in mouse brain to the drug concentration in plasma. AUC(0-Tlast)-
derived ratio of the unbound drug concentration in mouse brain to the unbound drug
concentration in plasma (mouse fraction unbound in plasma, fu,p = 0.0135; mouse fraction
unbound in brain, fu, b = 0.00775). nd = not determined.
In vivo PK/PD and Efficacy of compound 15. Compound 15 was tested in a wild type mouse
PK/PD model to understand the relationship between 2-AG and AA changes with compound
exposure in the brain. Mice (male C57Bl6) were administered a 1 mg/kg subcutaneous dose of
compound 15, and then brain levels of 2-AG and arachidonic acid were monitored over a 24 h
time course. We observed the expected increase in brain 2-AG levels, which persisted for 8 h
post dose, compared to vehicle treated mice (Figure 9a). In addition, a concomitant decrease in
brain arachidonic acid level was observed compared to vehicle treated mice over a similar time
period (Figure 9b). Excellent total brain exposure of compound 15 was achieved, with greater
than one micromolar concentration at Cmax (Figure 9c). Unbound brain concentrations of
compound 15 above the IC (3 nM) were achieved (Figure 9d) until approximately 2 hours post-
50
dose. Gratifyingly, as is expected from a covalent inhibitor, the desired PD response was
observed for several hours beyond the PK time course.
27
ACS Paragon Plus Environment

Page 29 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 9 Effect of compound 15 following subcutaneous administration of a 1 mg/kg dose in
mice. (a) brain 2-AG levels (b) brain AA levels (c) Total plasma and brain concentrations of
compound 15 (d) Free plasma and brain concentrations of compound 15.
Additionally, the effect of compound 15 in reducing brain levels of inflammatory mediators
was tested in a mouse model of inflammation (Figure 10). A high dose of systemic
lipopolysaccharide (LPS) was used to induce a sepsis or encephalitis-like state. As expected, 2-
AG and AA levels were not impacted by LPS challenge, however, increases in brain PGE2, IL-
28
ACS Paragon Plus Environment