Identification of ABX-1431, a Selective Inhibitor of Monoacylglycerol Lipase and Clinical Candidate for Treatment of Neurological Disorders

Article abstract of DOI:10.1021/acs.jmedchem.8b00951

The serine hydrolase monoacylglycerol lipase (MGLL) converts the endogenous cannabinoid receptor agonist 2-arachidonoylglycerol (2-AG) and other monoacylglycerols into fatty acids and glycerol. Genetic or pharmacological inactivation of MGLL leads to elevation in 2-AG in the central nervous system and corresponding reductions in arachidonic acid and eicosanoids, producing antinociceptive, anxiolytic, and antineuroinflammatory effects without inducing the full spectrum of psychoactive effects of direct cannabinoid receptor agonists. Here, we report the optimization of hexafluoroisopropyl carbamate-based irreversible inhibitors of MGLL, culminating in a highly potent, selective, and orally available, CNS-penetrant MGLL inhibitor, 28 (ABX-1431). Activity-based protein profiling experiments verify the exquisite selectivity of 28 for MGLL versus other members of the serine hydrolase class. In vivo, 28 inhibits MGLL activity in rodent brain (ED₅₀ = 0.5-1.4 mg/kg), increases brain 2-AG concentrations, and suppresses pain behavior in the rat formalin pain model. ABX-1431 (28) is currently under evaluation in human clinical trials.

Full text of DOI:10.1021/acs.jmedchem.8b00951

Subscriber access provided by - Access paid by the | UCSB Libraries
Article
Identification of ABX-1431, a Selective Inhibitor of Monoacylglycerol
Lipase and Clinical Candidate for Treatment of Neurological Disorders
Justin S. Cisar, Olivia D. Weber, Jason R. Clapper, Jacqueline Blankman, Cassandra L. Henry, Gabriel M.
Simon, Jessica P. Alexander, Todd K. Jones, R. Alan B. Ezekowitz, Gary P. O'Neill, and Cheryl A Grice
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00951 • Publication Date (Web): 01 Aug 2018
Downloaded from http://pubs.acs.org on August 1, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification of ABXꢀ1431, a Selective Inhibitor of
Monoacylglycerol Lipase and Clinical Candidate for
Treatment of Neurological Disorders
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
1
1
Justin S. Cisar , Olivia D. Weber , Jason R. Clapper , Jacqueline L. Blankman , Cassandra L.
1
,2
1
1
1
Henry , Gabriel M. Simon , Jessica P. Alexander , Todd K. Jones , R. Alan B. Ezekowitz , Gary
1
1*
P. O’Neill and Cheryl A. Grice
1
Abide Therapeutics, 10835 Road to the Cure, Suite 250, San Diego, CA 92121, USA
Vividion Therapeutics, 3565 General Atomics Ct, Suite 100, San Diego, CA 92121, USA.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT. The serine hydrolase monoacylglycerol lipase (MGLL) converts the endogenous
cannabinoid receptor agonist 2ꢀarachidonoylglycerol (2ꢀAG) and other monoacylglycerols into
fatty acids and glycerol. Genetic or pharmacological inactivation of MGLL leads to elevation in
2ꢀAG in the central nervous system and corresponding reductions in arachidonic acid and
eicosanoids, producing antiꢀnociceptive, anxiolytic, and antiꢀneuroinflammatory effects without
inducing the full spectrum of psychoactive effects of direct cannabinoid receptor agonists. Here,
we report the optimization of hexafluoroisopropyl carbamateꢀbased irreversible inhibitors of
MGLL, culminating in a highly potent, selective, and orally available, CNSꢀpenetrant MGLL
inhibitor, 28 (ABXꢀ1431). Activityꢀbased protein profiling experiments verify the exquisite
selectivity of 28 for MGLL versus other members of the serine hydrolase class. In vivo, 28
inhibits MGLL activity in rodent brain (ED50 = 0.5 – 1.4 mg/kg), increases brain 2ꢀAG
concentrations, and suppresses pain behavior in the rat formalin pain model. ABXꢀ1431 (28) is
currently under evaluation in human clinical trials.
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 (MGLL, also known as MAG lipase, MAGL and MGL) is a serine
hydrolase enzyme responsible for controlling the content and signaling of the endogenous
1ꢀ4
cannabinoid 2ꢀarachidonoylglycerol (2ꢀAG) in the central nervous system (Figure 1A).
5
MGLL also regulates 2ꢀAG and other MAG lipids in peripheral tissues. In select tissues, such as
the brain, MGLLꢀmediated metabolism of 2ꢀAG serves as an important source of arachidonic
6
acid and contributes to subsequent biosynthesis of proinflammatory prostanoids. Thus, MGLL
2
ACS Paragon Plus Environment

Page 3 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
is a critical point of regulation of both the endocannabinoid and eicosanoid signaling pathways in
the CNS and select peripheral tissues.
Pharmacological inactivation of MGLL in animals increases 2ꢀAG content of brain 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
peripheral tissues and has been found to produce antinociceptive, anxiolytic, and antiꢀ
3, 7ꢀ9
inflammatory effects that are dependent on CB1 and/or CB2 cannabinoid receptors.
MGLL
inhibitors also promote CB1/2ꢀindependent effects on neuroinflammation that may reflect
1
0ꢀ11
reductions in proinflammatory prostanoid signaling in models of traumatic brain injury,
6, 12ꢀ13
14
neurodegeneration,
and status epilepticus. Importantly, MGLLꢀmediated modulation of
endocannabinoid signaling in animals does not produce the full spectrum of neurobehavioral
9
5,
effects observed with ∆ ꢀtetrahydrocannabinol (THC) and other exocannabinoid CB1 agonists.
15
Thus, MGLL inhibitors have potential to produce therapeutic effects through both enhanced
CB1/2 signaling and reduced proinflammatory prostanoid action and may demonstrate
1
6ꢀ17
significantly reduced side effects relative to direct CB1 receptor agonists.
Several academic and industrial groups have reported MGLL inhibitors that act via
18ꢀ22
23ꢀ30
reversible
or irreversible mechanisms.
Irreversible MGLL inhibitors (Figure 1B) such as
JZL184 (1) carbamoylate the catalytic serine nucleophile resulting in persistent enzyme
5
inactivation. A distinct advantage of irreversible over reversible inhibitors is that
pharmacodynamic effects may be maintained for extended periods with a low dose of
administered drug, since enzyme activity only returns in vivo following the synthesis of new
3
1ꢀ32
protein (vs. recovery shortly after compound clearance).
Additionally, MGLL and other
mechanistically related serine hydrolase enzymes can be assayed using activityꢀbased protein
3
3ꢀ35
profiling (ABPP) methods,
which allow rapid assessment of compound selectivity in vitro
3
6
and in living systems. As will be described below, we have leveraged ABPP as a core
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 97
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
screening technology to identify MGLL inhibitors with high potency and selectivity that would
provide robust pharmacological responses at low compound exposures.
The carbamate substructure is present in a number of selective covalent MGLL inhibitors (2 –
36ꢀ38
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
7
, Figure 1B) that are active both in vitro and in vivo.
At the outset of this program, both Nꢀ
39
24
hydroxysuccinimidyl (NHS) carbamates and Oꢀhexafluoroisopropyl (HFIP) carbamates were
4
0ꢀ
reported to inhibit MGLL in vitro and in vivo and demonstrated efficacy in preclinical models.
4
1
39
While the NHS carbamates such as 4 have good proteomeꢀwide selectivity, we observed
aqueous instability due to hydrolysis of the imide carbonyl, thus posing a potential challenge for
2
4
clinical development. As previously noted, the HFIP carbamate constitutes a reactive group of
tempered electrophilicity (pKa value of hexafluoroisopropanol is ~9.3) and the
hexafluoroisopropanol component structurally resembles glycerol in the natural MAG substrates
of MGLL. In our experience, the HFIP carbamate has excellent aqueous chemical stability over a
broad pH range (typically >95% stable after 24 hrs at pH 2 – 8). Finally, the safety of
hexafluoroisopropanol, which is presumably liberated upon MGLL carbamoylation, is supported
by the clinical experience of the widely used anesthetic sevofluorane, of which
4
2
hexafluoroisopropanol is a circulating metabolite. Thus, we focused our efforts on elaboration
4
3ꢀ
of the HFIP carbamate substructure to optimize both MGLL potency and drugꢀlike properties.
45
At the outset of these studies, the reported covalent MGLL inhibitors were limited by their
2
9
selectivity and/or drugꢀlike properties. More recent reports from Pfizer have demonstrated a
series of MGLL inhibitors with high potency and selectivity and good physicochemical
3
0
properties suitable for intravenous administration. Herein, we describe the discovery of 28
4
3
(ABXꢀ1431), a highly potent, selective, and CNSꢀpenetrant MGLL inhibitor with excellent
4
ACS Paragon Plus Environment

Page 5 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
drugꢀlike properties suitable for onceꢀperꢀday oral administration. Our lead optimization strategy
was implemented using ABPP to measure the potency and selectivity of candidate MGLL
inhibitors directly in native brain tissue proteomes. ABPP was also used to measure brain target
engagement and selectivity in animals, which served to rapidly identify compounds with
desirable CNS activity. Finally, we show that the optimized product of these efforts, 28, has low
nanomolar potency for MGLL and is active at low doses in an animal model of pain.
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 1. (A) Enzymatic reaction catalyzed by MGLL and (B) representative covalent MGLL
inhibitors.
Results and Discussion
Our goal was to develop a MGLL inhibitor that was highly potent and selective, orally active,
CNS penetrant, and suitable for clinical use. While previously reported tool compounds and
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 97
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
inhibitors lacked at least one of these critical properties, several HFIP carbamates such as
2
4
KML29 (2) exhibited high potency and selectivity for MGLL. In general, the HFIP carbamate
has high serine hydrolase selectivity; however, common offꢀtargets include the serine hydrolases
4
6
47
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
fatty acid amide hydrolase (FAAH), α/β hydrolase domain containing 6 (ABHD6),
48
29, 49
phospholipase A2 group VII (PLA2G7), and the carboxyesterases (CES).
In order to
optimize both MGLL potency and serine hydrolase selectivity simultaneously, our initial
50
screening analysis included competitive gelꢀbased ABPP methods in multiple human
proteomes. Potency (IC ) measurements in native lysates derived from the human PC3 cell line
5
0
4
9
using the HFIP carbamate ABPP probe JW912 provided initial information about potency for
MGLL and selectivity assessments for the offꢀtargets ABHD6 and PLA2G7. Additional in vitro
selectivity analysis was carried out in human postmortem prefrontal cortex proteome using a
5
1
fluorophosphonate activity probe (FPꢀRh), which enabled detection of many additional serine
hydrolases enzymes, including FAAH.
MGLL potency and offꢀtarget selectivity was first established for symmetrical biaryl analogs
24
of 9, a previously reported MGLL inhibitor (
Table 1). The symmetrical biaryl compounds in Table 1 included both phenyl (10) as well as
heterocyclic (11 – 13) analogs. Analog 10 (Table 1) maintained potency and selectivity;
however, the compound was not further pursued due to its high lipophilicity. The more
hydrophilic biaryl analogs 11–13 generally maintained MGLL potency but displayed low
selectivity against ABHD6 and PLA2G7. Thus, replacements for the biaryl motif were explored
to optimize selectivity.
Table 1. Symmetrical biaryl analogs 9 – 13 and in vitro potency for MGLL and the off-
targets ABHD6 and PLA2G7.
ACS Paragon Plus Environment
6

Page 7 of 97
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
a
b
IC s determined using ABPP with a 30ꢀmin inhibitor preincubation in PC3 lysate. cLogP
50
5
2
values calculated using XLogP (Dotmatics).
Given their reduced lipophilicity, 1,3,4ꢀsubstituted pyrazoles (14 – 18) linked to the piperazine
HFIP carbamate were examined as replacements for the biaryl substituent (Table 2). Phenyl
substituted pyrazoles were tolerated with respect to MGLL potency (14, 15, 17, 18). Replacing
phenyl with an Nꢀacetylpiperidinyl moiety (16), however, resulted in reduction in MGLL
potency. While the phenyl substituted pyrazole motif was effective in potently inhibiting MGLL
and reducing lipophilicity, it offered little advantage over the biaryl analogs with respect to
ABHD6 and PLA2G7 offꢀtarget selectivity.
Table 2. Pyrazole analogs and in vitro potency for MGLL and the off-targets ABHD6 and
PLA2G7.
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 97
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
a
b
IC s determined using ABPP with a 30ꢀmin inhibitor preincubation in PC3 lysate. cLogP
5
0
5
2
values calculated using XLogP (Dotmatics).
The impact of removing one of the aryl groups of the biaryl analogs (Table 3) was evaluated,
resulting in analogs exemplified by compound 19, which maintained MGLL potency but was
less selective against ABHD6. Importantly, introduction of a morpholine ring at R2 (20) resulted
in an improved selectivity profile. In analogs 21 – 23, the position of the chlorine substituent of
2
0 was varied. Chlorine was preferred at the R4 or R3 positions, with both analogs 20 and 23
displaying high MGLL potency and offꢀtarget selectivity. Analysis of in vitro metabolic stability
in human liver microsomes indicated that the overall profile of 20 was superior to that of 23.
ACS Paragon Plus Environment
8

Page 9 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Impact of R2–R4 variation on in vitro potency for MGLL, the off-targets ABHD6
and PLA2G7, and the intrinsic clearance values in human liver microsomes.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
IC s determined using ABPP with a 30ꢀmin inhibitor preincubation in PC3 lysate. Cl
50
int
(µL/min/mg protein) in human liver microsomes.
Based on the properties of 20, we investigated additional R2/R4ꢀsubstituted analogs to
optimize selectivity and drugꢀlike properties further, as shown in Table 4. Morpholine analogs
0 (R4 = Cl) and 24 (R4= CF ) afforded similar MGLL and selectivity profiles; however, the
2
3
trifluoromethyl analog 24 possessed improved human liver microsome stability. Analogs
containing morpholine (20, 24) and piperidine (25, 26) at the R4 position were all potent against
human MGLL (PC3 lysates); however, 25 and 26 (R4 = piperidine) displayed significantly lower
rodent MGLL potency (mouse brain homogenates) as compared to 20 and 24 (R4 = morpholine).
Pyrrolidine analog 28 (R4 = CF ) displayed high potency and selectivity for MGLL in human
3
PC3 lysates and rodent MGLL brain homogenates. Furthermore, 28 demonstrated improved
microsome stability as compared to 20 and 24. Interestingly, other R4 substituents such as
chlorine (27) and bromine (29) gave similar MGLL potency and selectivity; however, the polarꢀ
aprotic sulfone moiety (30) significantly reduced selectivity against ABHD6 and PLA2G7.
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 97
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
Introduction of an alkyne (31) or proton (32) at the R4 position resulted in a reduction of MGLL
potency.
In addition to assessing in vitro potency and selectivity, the in vivo target engagement of
36
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
inhibitors following oral dosing in mice was routinely measured using ABPP methods (Table
). Mice were administered test compounds (5 mg/kg, po) and whole brains were collected from
4
the euthanized animals four hours postꢀdose for target engagement analysis by gelꢀbased ABPP
using the FPꢀRh activity probe. Assessment of target engagement under these conditions reflects
the cumulative impact of the compound across multiple parameters including oral absorption,
brain penetrance, pharmacokinetics, and the molecular interaction of the compound and enzyme
in the target tissue. Analogs 24 and 28 showed high potency in vivo, with both molecules
demonstrating complete MGLL inhibition in the mouse brain 4 hours following a single 5 mg/kg
oral gavage (po) dose.
Table 4. Impact of R2 and R4 variation on potency for MGLL, off-targets, in vivo target
engagement, and microsome stability.
ACS Paragon Plus Environment
10

Page 11 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
b
Determined using ABPP with a 30ꢀmin inhibitor preincubation in PC3 lysate. Brain target
c
engagement measured 4 hrs following 5 mg/kg dose po using ABPP. Cl (µL/min/mg protein).
int
A broader survey of substituents at R2 while preserving either a trifluoromethyl or chlorine
group at R4 (Table 5) was conducted. MGLL potency was maintained over a broad range of R2
substitutions including elaborated morpholine (33 – 36) and pyrrolidine (37 – 41) functionalities.
Among the allowed substituents were larger ring systems including bridged morpholine analogs
33 – 35, and fused pyrrolopyrimidine analog 36. Elaborated pyrrolidines 37 – 41 were also
tolerated at the R2 position, including highly potent fused ring (R2= hexahydroꢀ1Hꢀfuro[3,4ꢀ
c]pyrrole, 39) and spirocyclic (R2 = 2ꢀmethylꢀ8ꢀoxaꢀ2ꢀazaspiro[4.5]decane, 41) analogs.
Together, in vitro human potency, human in vitro selectivity, and in vivo target engagement data
provided a means to prioritize compounds. Analogs such as 36 and 39 had favorable in vitro
MGLL potency (MGLL IC50 = 11, 2, 4 nM, respectively) but submaximal in vivo MGLL
engagement in the brain and were thus deprioritized.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
Introduction of a F fluorine atom would provide PET tracer candidates that could be used to
18
image MGLL activity distribution in the CNS and enable drug target occupancy studies. Fꢀ
based radiotracers should have longer halfꢀlives and sensitivity over the previously reported
1
1
53ꢀ55
MGLL radiotracers, all of which contain the C radioisotope.
The R2 position was selected
for fluorine incorporation, as the SAR demonstrated a broad range of tolerated substitution
patterns. Out of several candidates, 42 was prepared and found to be potent and selective against
MGLL and active in the in vivo target engagement assay in rodents. Additional results involving
the successful preclinical and clinical application of 42 as a MGLL PET tracer will be described
elsewhere.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 97
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 5. Impact of R2 variation on potency for MGLL, off-targets, and in vivo MGLL
target engagement.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
Determined using ABPP with a 30ꢀmin inhibitor preincubation in PC3 lysate. Brain target
engagement measured 4 hrs following 5 mg/kg dose po using ABPP.
43
44
Together, these medicinal chemistry efforts identified 28 (ABXꢀ1431) and 34 (ABDꢀ1970)
as inhibitors warranting further consideration. ABDꢀ1970 (34) was extensively characterized and
used as a tool to probe MGLL pharmacology, and its activity in rodents and human biological
systems will be described in a separate publication. ABXꢀ1431 (28) was selected as the lead
compound for clinical development and is described further herein.
ACS Paragon Plus Environment
12

Page 13 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABPP was used to profile 28 in multiple of human and animal proteomes using gelꢀbased and
massꢀspec based technologies. Using gelꢀbased ABPP, 28 was assayed in human cell line lysates
(Table 4, Supporting Information Figure S1) and human postꢀmortem prefrontal cortex
homogenates (Figure 2D) following a 30ꢀminute inhibitor incubation time. Overall, 28 was
determined to be a potent human MGLL inhibitor (average IC50 = 0.014 µM) with >100ꢀfold
selectivity against ABHD6 and >200ꢀfold selectivity against PLA2G7 (Table 4, Supporting
Information Table S1). Treatment of intact human PC3 cells with 28 following a 30ꢀminute
inhibitor incubation time caused concentration dependent inhibition of MGLL activity with an
IC value of 0.0022 µM (Supporting information, Table S1), which is ~6ꢀfold more potent than
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
5
0
that observed in vitro. In the cellꢀbased assay, >100ꢀfold selectivity for MGLL over ABHD6
IC50 = 0.253 µM) and PLA2G7 (IC50 = 494 µM) was maintained (Supporting information,
(
Table S1). In mouse, rat, and dog brain proteomes, 28 was found to be selective and potent for
MGLL as judged by gelꢀbased ABPP (Figure 2A – C, Supporting Information Table S1).
Additional analysis of 28 selectivity was carried out using MSꢀbased ABPP methods in human
postꢀmortem prefrontal cortex homogenates and shown to be highly selective against a panel of
3
7 serine hydrolases (Figure 2E).
The kinetics for inactivation of human MGLL by 28 were determined by measuring the
kinact/K , a time independent measure of inhibition. The average k /K value of 28 was 20,200
i
inact
i
ꢀ1 ꢀ1
M s against purified human MGLL using a fluorescent substrate assay.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 97
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. In vitro potency and selectivity for 28. AꢀD) gelꢀbased ABPP in mouse, rat, dog, and
human brain membrane homogenates. Brain membrane homogenates were treated with 28
(0.001ꢀ10 µM, 30 min, 37ºC) and subsequently labeled with FPꢀRh (1 µM, 30 min, room
temperature), quenched in SDSꢀPAGE loading buffer, and analyzed by inꢀgel fluorescence
scanning. Human ABHD6 was not visible by inꢀgel fluorescence scanning. E) MSꢀbased ABPP:
Human prefrontal cortex total homogenates were treated with 10 µM 28, labeled with FPꢀbiotin.
and enriched with avidin beads. Trypsinized peptides were analyzed by LC MS/MS.
ACS Paragon Plus Environment
14

Page 15 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Carbamate electrophiles are expected to inhibit MGLL via covalent labeling of the activeꢀsite
serine nucleophile (Ser122) and thus afford a durable carbamoylated inhibitor–enzyme adduct
5
that is catalytically inactive (Figure 3A). To confirm the covalent mechanism of action, the
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
interaction of 28 with MGLL was characterized via a mass spectrometryꢀbased assay. Following
incubation of recombinant human MGLL with 28 and subsequent tryptic digestion, analysis by
high resolution nanoLCꢀMS/MS revealed an adduct corresponding to the peptide mass of the
predicted carbamate adduct of 28 with Ser122 (Figure 3B). This adduct was only observed in the
presence of inhibitor and not in DMSOꢀtreated samples. Furthermore, DMSOꢀtreated samples
contained greater levels of the unmodified Ser122ꢀcontaining peptide than those treated with 28.
These data indicate that 28 inhibits MGLL via carbamoylation of the catalytic nucleophile
Ser122. A biochemical assay was employed to measure the in vitro stability of the 28–MGLL
adduct. Accordingly, the MGLL–28 covalent adduct was prepared by incubation of recombinant
human MGLL and 28 (1 µM). Exogenous compound was then removed via repeated washing,
and recovery of enzyme activity was monitored over 24 hours (Supporting Methods and Table
S). Over this timeframe, MGLL activity regained less than three percent activity, suggesting that
the carbamoylated inhibitor–MGLL adduct is stable and nonhydrolyzable for at least 24 hours.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 97
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. Mechanism of 28 inhibition of MGLL. (A) Proposed reaction of 28 and MGLL and the
resulting carbamoylated activeꢀsite serine nucleophile (Ser 122). (B) Extracted ion
chromatograms (EICs) showing the unmodified (top) and 28ꢀmodified (bottom) activeꢀsite
tryptic peptide in DMSOꢀtreated (red) and 28ꢀtreated (blue) samples. The 28ꢀmodified adduct is
calculated to increase the parent peptide mass by 339.15585 amu. The mass of the baseꢀpeak was
used for each EIC trace: m/z 1050.967 and 1118.799, respectively.
Pharmacokinetic analysis in rats and dogs of 28 (Figure 4, Table 6) indicated low to moderate
systemic clearance, moderate volume of distribution, and high oral bioavailability (64% in rat,
5
7% in dog). While 28 was stable in human and dog plasma, it was not stable in rat plasma
ACS Paragon Plus Environment
16

Page 17 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
Supporting Methods and Table S2). This result is thought to be a consequence of increased
5
6ꢀ57
expression of the CES enzymes,
which are known to metabolize ester and carbamate
xenobiotics, and are more abundant in rodent plasma than other higher mammal species
5
7
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
including dogs, primates, and humans. Thus, in vivo studies that required measurement of 28 in
rat blood, where degradation could occur exꢀvivo, were conducted using a modified procedure
that stabilizes 28 during sample preparation (see Experimental Section).
Additional selectivity analysis revealed that 28 has low propensity to inhibit human
recombinant CYP enzymes (IC > 50 ꢁM for CYP1A2, CYP2C9, CYP2C19, CYP3A4/5, IC =
5
0
50
6
.5 ꢁM for CYP2D6) and the hERG channel (IC20 ~ 7 µM). Analysis in various transporter
assays indicated that 28 is not an inhibitor or a substrate for Pꢀgp, BCRP, and OCT2 at 10 µM.
Finally, 28 (10 µM) did not display significant binding or activity (defined as >50% of control
ligand binding or activity) against a panel of 95 targets including enzymes, receptors,
transporters and ion channels.
A
Rat
B
Dog
10
1
10
1
IV - 1 mg/kg
PO - 5 mg/kg
IV - 1 mg/kg
PO - 1 mg/kg
0
.1
0.01
.001
0.1
0.01
0.001
0
0
2
4
8
12
16
20
24
0
2
4
8
12
16
20
24
Time (h)
Time (h)
Figure 4. Concentration of 28 vs time profiles following a single iv or po administration in rat
(A) and dog (B).
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 97
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 6. Pharmacokinetic parameters of 28 in rat and dog.
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
To determine the in vivo MGLL potency of 28, the compound was administered by oral
gavage to mice (Figure 5A–C). Four hours after a single oral dose, 28 doseꢀdependently
inhibited mouse brain MGLL activity with ED = 1.4 mg/kg (Figure 5B). At a dose ~20ꢀfold
5
0
higher (32 mg/kg), 28 demonstrated high selectivity, with the only other serine hydrolases
inhibited being ABHD6 (ED ~ 32 mg/kg) (Figure 5A) and the carboxylesterease CES1c (ED
50
5
0
~
32 mg/kg, Supplemental Figure S2), a rodentꢀspecific CES isoform with no direct human
homologue. Based on comparison of the ED values, 28 displays ~23ꢀfold selectivity for MGLL
50
versus ABHD6 and CES1c. As expected, inhibition of brain MGLL activity resulted in doseꢀ
dependent increases in brain 2ꢀAG concentrations (Figure 5C).
To enable in vivo proteomeꢀwide selectivity assessment of the 28 chemotype, an analog of 28,
31, containing an alkyne at the R4 position of the benzyl ring, was used. The alkyne group serves
as a latent affinity handle suitable for conjugation to reporter tags by copperꢀcatalyzed azideꢀ
58
alkyne cycloaddition (CuAAC or click chemistry). In vitro, compound 31 inhibited mouse
MGLL with slightly reduced potency (mouse MGLL IC = 0.159 µM, Table 4) relative to 28
50
(mouse MGLL IC50 = 0.027 µM, Table 4). Following oral administration of compound 31 to
mice (25 mg/kg, po), the only probeꢀlabeled proteins, as determined by gelꢀbased ABPP, were
MGLL and ABHD6, demonstrating the high proteomeꢀwide selectivity of the 28 chemotype
(Figure 5D).
ACS Paragon Plus Environment
18

Page 19 of 97
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 5. ABXꢀ1431 (28) is a potent and selective inhibitor of MGLL in mouse brain in vivo. (A
ꢀ
C) Mice were administered vehicle or 28 (0.5 – 32 mg/kg, po, n = 3) and after 4 h, sacrificed,
brains collected and prepared for (A and B) gelꢀbased ABPP with the serine hydrolase directed
probe FPꢀRh or (C) 2ꢀAG quantification by LCMS/MS. (D) Click probe (31) analysis in mouse
brain 4 h following a single oral 25 mg/kg dose. *** p < 0.001 vs. vehicle.
Further characterization of 28 revealed similar in vivo activity when the compound was
administered orally to rats (Figure 6). The ED of 28 for inhibition of brain MGLL activity was
5
0
0
.5 mg/kg (Figure 6A and B), whereas a 6ꢀfold higher dose was required to produce substantial
inhibition of ABHD6 (ED50 = 2.9 mg/kg). Similar to the mouse, the rodentꢀspecific CES1c
isoform was inhibited by 28 (ED = 1.2 mg/kg, Supplemental Figure S3). Inhibition of MGLL
5
0
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 97
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
activity in the brain in rats by 28 corresponded to a doseꢀdependent elevation of brain 2ꢀAG
tissue concentrations (Figure 6C).
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. ABXꢀ1431 (28) is a potent and selective inhibitor of MGLL in rat brain in vivo and
produces antinociceptive effects in rats. (A – C) Rats were administered vehicle or 28 and after 4
h, sacrificed, brains collected and prepared for (A and B) gelꢀbased ABPP with the serine
hydrolase directed probe FPꢀRh or (C) 2ꢀAG quantification by LCMS/MS. (D) A) Single doses
of 28 (3 mg/kg, PO) was administered 4 hours prior to formalin (2.5%, 50 ꢀL, intraplantar),
reduced pain behavior measured as total paw licking duration (n=10 rats per group, *p < 0.05,
*
*p < 0.01 vs. respective vehicle).
To evaluate the potential antinociceptive effects of 28, the compound was evaluated in the
intraplantar formalin paw test in rats. The formalin paw test is a model of pain induced by local
injection of formalin (2.5%, 50 µL) into one hind paw, producing an early response (0ꢀ5 min)
ACS Paragon Plus Environment
20

Page 21 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
due to direct stimulation of nociceptors and a delayed response (10ꢀ30 min) driven by
59
inflammation and central sensitization. Pain behavior in these studies was quantified by total
time the rats spent licking the affected paw. A single 3 mg/kg oral dose of 28 administered to rats
four hours prior to formalin injection significantly reduced formalinꢀevoked paw licking duration
in both phases (Figure 6D), demonstrating the antinociceptive properties of 28. The formalin
model evaluates effects of a compound on spontaneous pain, thus effects on paw licking can be
due to antinociceptive mechanisms or sedation. To differentiate between these possibilities, the
effects of 28 on spontaneous locomotor activity was evaluated in rats that were not administered
formalin. In support of an antinociceptive effect, single 10 mg/kg oral administration of 28 did
not alter locomotor activity in rat (vehicle locomotor counts = 53.0 ± 3.7; 28 locomotor counts =
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
5
4.9 ± 6.6; n = 10 animals per group). The doses of 28 used in these studies produced near
complete MGLL inhibition and maximal elevation of brain 2ꢀAG content; however, previously
published studies of MGLL inhibitors suggest that partial MGLL inhibition and submaximal
elevation of brain 2ꢀAG is sufficient to produce antinociceptive effects in mice following either
single or multiple day dosing. Characterization of the relationship between MGLL inhibition
resulting from treatment with 28 and antinociceptive effects will be described elsewhere.
Chemistry
The synthesis of 28 was carried out in four steps with 38% overall yield (Scheme 1). The
sequence began with S Ar reaction between aldehyde 28a and pyrrolidine 28b to give 2ꢀ
N
pyrroldinebenzaldehyde 28c. Reductive amination of 28c with Bocꢀpiperazine provided 28d.
Subsequent Boc deprotection and carbamoylation with hexafluoroisopropanol and triphosgene
and diisopropylethylamine yielded 28. Additional analogs (20 – 29) were prepared using this
approach, and a general scheme is shown in (Scheme 2).
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 97
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 1. Preparation of 28 (ABXꢀ1431).
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. General strategy for preparation of aldehyde intermediates and benzyl substituted
HFIP carbamates.
ACS Paragon Plus Environment
22

Page 23 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 3. Preparation of 10 – 13.
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
Preparation of the biaryl analogs 10 – 13 (Scheme 3) began with synthesis of biarylmethanol
intermediates 10b – 13b. Intermediates 10b and 11b were formed via nucleophillic addition of
two equivalents of aryl magnesium or lithium reagents to ethyl formate. Reduction of
commercially available ketone 12a yielded 12b. Biaryl methanol 13b was produced by treatment
of 13a with nꢀbutyl lithium and subsequent reaction with oxazoleꢀ4ꢀcarbaldehyde. The
biarylmethanol intermediates 10b – 13b were subsequently activated using methanesulfonyl
chloride or thionyl chloride and reacted with Bocꢀpiperazine to yield intermediates 10c – 13c.
Protecting group removal and carbamoylation with hexafluoroisopropanol gave final products 10
–
13.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 97
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 4. Preparation of 14 – 18.
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
Preparation of pyrazole analogs 14 – 18 are shown in Scheme 4. For the preparation of
4, 15, 17, and 18, pyrazole carbaldehydes 14c – 18c were combined with Bocꢀpiperazine under
1
reductive amination conditions and subsequently deprotected and carbamoylated to give the
desired HFIP carbamates. The synthesis of 16 required an alternative route and began with
reaction of the methyl hydrazine with 16a and subsequent cyclization with Nꢀ(chloromethylene)ꢀ
Nꢀmethylmethanaminium chloride to give pyrazole carbaldehyde 16c. Reductive amination of
16c with Bocꢀpiperazine resulted in 16d. The distal pyridine moiety was alkylated, reduced, and
acylated to give 16f. Finally, Bocꢀdeprotection of 16f and carbamoylation with
hexafluoroisopropanol gave 16.
ACS Paragon Plus Environment
24

Page 25 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 5. Preparation of 30 and 31.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The preparation of analogs 30 and 31 (Scheme 5) required slight deviations from the general
routes described above. 30d was subjected to sulfonation conditions to produce 30e. Subsequent
Bocꢀdeprotection of 30e and carbamoylation with hexafluoroisopropanol afforded 30. Analog 31
was produced by incorporation of a Sonogashira coupling to unite 29c and TMSꢀacetylene. The
resulting product, 31f, was fully deprotected to yield 31g. The penultimate intermediate 31g was
carbamoylated with hexafluoroisopropanol to give final compound 31.
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 97
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 6. Preparation of 32, 40, and 41.
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 alternative approach to HFIP carbamate analogs was used to prepare analogs 32, 40, and 42
Scheme 6). In these cases, key common intermediate 32b was prepared by reaction of Bocꢀ
(
piperazine and bisꢀHFIP carbonate and followed by subsequent TFAꢀmediated Bocꢀdeprotection.
32b could be reacted with benzaldehydes 32c, 40c, and 42c to give either a final product (32) or
an advanced intermediate such as 40d or 42d. The compounds 40d and 42d could be treated with
TFA to remove the distal Boc protecting group and subjected to alkylation conditions to give
final products 40 and 42.
Conclusion
While the beneficial pharmacology of direct CB1 and CB2 agonists is well recognized,
these agents are hindered by numerous side effects. Significant preclinical evidence suggests that
ACS Paragon Plus Environment
26

Page 27 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MGLL inhibitors that elevate the endogenous CB receptors’ agonist 2ꢀAG have the potential to
produce the beneficial effects of direct CB agonists while achieving a superior sideꢀeffect profile.
Our detailed optimization of the HFIP carbamate motif, achieved by integrating ABPP screening
methods that efficiently determine MGLL potency and selectivity both in vitro and in vivo, led to
discovery of 28. ABXꢀ1431 (28) is a potent and selective MGLL inhibitor with favorable ADME
properties and is currently being evaluated in clinical trials.
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
Experimental Section
Chemistry General Methods. All commercially available chemicals were obtained from
Aldrich, Acros, Fisher, Fluka, Maybridge or the like and were used without further purification,
except where noted. Anhydrous solvents and ovenꢀdried glassware were used for synthetic
transformations sensitive to moisture and/or oxygen. All reactions are typically carried out under
an inert nitrogen atmosphere using ovenꢀbaked glassware unless otherwise noted. Flash
chromatography is performed using 230ꢀ400 mesh silica gel 60 using Isco Combiflash
instruments. NMR spectra were generated on either Bruker 300 or Bruker 400 instruments.
Chemical shifts are typically recorded in ppm relative to tetramethylsilane (TMS) with
multiplicities given as s (singlet), bs (broad singlet), d (doublet), t (triplet), dt (double of triplets),
q (quadruplet), qd (quadruplet of doublets), hept (heptuplet), m (multiplet). Chemical purities
were >95% for all final compounds as assessed by LC/MS analysis using a Shimadzu 2020 or an
Agilent 1200 LCMS with detection at 210 and 254 nm. The column used was an xBridge 50mm
x 3mm using a 0 to 100% acetonitrile in water (with 0.05% TFA additive) gradient over 1 min at
1
ml/min at 40 °C. Alternatively an ACE 2.1x100mm C18 column was used with a 0 to 100%
acetonitrile in water (with 0.1% formic acid additive) gradient over 1 min at 1 ml/min at 25 °C.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 97
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
bis(4-chloro-2-methylphenyl)methanol (10b). A round bottom flask was charged with a 4ꢀ
chloroꢀ2ꢀmethylphenyl magnesium bromide solution (10 mL of a 0.5 M THF solution, 5 mmol)
and THF (50 mL). The solution was cooled to –78 °C and an ethyl formate solution (200 µL,
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
1
.50 mmol, in 10 mL THF) was added drop wise. The reaction was allowed to stir at ꢀ78 °C for
5 min and allowed to warm to rt slowly and stirred for 18 hrs. The reaction was diluted in ethyl
acetate and washed with brine (3 x 100 mL). The organics were dried over sodium sulfate,
filtered, and concentrated under reduced pressure. The residue was chromatographed on a silica
gel column (100 %hexanes to 15 % ethyl acetate in hexanes) to yield 10b (643 mg, 46%) as a
1
clear crystalline solid. H NMR 400 MHz (CDCl3) δ 7.25 – 7.14 (m, 6H), 6.07 (s, 1H), 2.26 (s,
+
6H), 1.57 (s, 2H). LCMS (ESI, m/z): 263 [M+H] .
tert-butyl 4-(bis(oxazol-4-yl)methyl)piperazine-1-carboxylate (10c). A round bottom flask
was charged with 10b (200 mg, 0.711 mmol) and DCM (7 mL). Thionyl chloride (100 µL,
1.37 mmol) was added and the reaction was stirred at rt for 24 h. The reaction was concentrated.
Acetonitrile was added and the reaction was concentrated two times. Acetonitrile (6 mL), tertꢀ
butyl piperazineꢀ1ꢀcarboxylate (200 mg, 1.07 mmol) and K CO (200 mg, 1.42 mmol) were
2
3
added and the reaction was heated to 80 °C for 4 h followed by 120 °C for 18 h. The reaction
was poured into brine and extracted with ethyl acetate (2 x 100 mL). The residue was
chromatographed on a silica gel column (100% hexanes to 20% ethyl acetate) and yielded 10c
1
(
184 mg, 57%) as a clear oil. H NMR 400 MHz (CDCl3) δ 7.49 (d, J = 8.4 Hz, 2H), 7.16 (d, J =
8.4 Hz, 2H), 7.11 (s, 2H), 4.65 (s, 1H), 3.44 – 3.37 (m, 4H), 2.38 (s, 4H), 2.29 (s, 6H), 1.46 (s,
+
9
H). LCMS (ESI, m/z): 471 [M+H] .
1
-(bis(4-chloro-2-methylphenyl)methyl)piperazine (10d). A round bottom flask was
charged with 10c (184 mg, 409 mmol), DCM (5 mL), and NMM (90 µL, 0.819 mmol). The
ACS Paragon Plus Environment
28

Page 29 of 97
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reaction was cooled to 0 °C and TMSI (70 µL, 0.491 mmol) was added dropwise. After 15 min
at 4 °C the reaction was quenched with saturated sodium carbonate and extracted with DCM (3 x
50 mL). The organics were dried over sodium sulfate, filtered, and concentrated. The residue
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
was chromatographed on a silica gel column (100% DCM to 6% 2M NH in MeOH ) and 10d
3
1
(
143 mg, 70%). H NMR 400 MHz (CDCl3) δ 7.47 (dd, J = 8.4, 1.1 Hz, 2H), 7.16 – 7.11 (m,
1
H), 7.11 – 7.06 (m, 2H), 4.64 (s, 1H), 2.88 – 2.80 (m, 4H), 2.44 – 2.33 (m, 5H), 2.28 (s, 7H).
+
LCMS (ESI, m/z): 349 [M+H] .
1
,1,1,3,3,3-hexafluoropropan-2-yl 4-(bis(4-chloro-2-methylphenyl)methyl)-piperazine-1-
carboxylate (10). The title compound was prepared from 10d as described for the preparation of
1
28. (18 mg, 59%). H NMR 400 MHz (CDCl ) δ 7.47 (dd, J = 8.4, 2.1 Hz, 2H), 7.22 – 7.16 (m,
3
2
4
H), 7.16 – 7.10 (m, 2H), 5.83 – 5.70 (m, 1H), 4.70 (s, 1H), 3.58 – 3.50 (m, 4H), 2.51 – 2.40 (m,
+
H), 2.30 (s, 6H). LCMS (ESI, m/z): 263 [bis(4ꢀchloroꢀ2ꢀmethylphenyl)methane cation]
.
bis(1-methyl-1H-indazol-5-yl)methanol (11b). A round bottom flask was charged with 5ꢀ
bromoꢀ1ꢀmethylꢀ1Hꢀindazole (300 mg, 1.42 mmol) and THF (45 mL). The solution was cooled
to ꢀ78 °C and an nꢀbutyllithium solution (2.3 M in THF, 680 µL, 1.56 mmol) was added
dropwise. After 30 min, a solution of ethyl formate (57 µL, 0.697 mmol, in 10 mL THF) was
added dropwise, and the reaction was stirred at ꢀ78 °C for 10 min and rt for 3h. The reaction was
quenched with saturated NH Cl and extracted with ethyl acetate (3 x 50 mL). The organics were
4
dried over sodium sulfate, filtered, and concentrated under reduced pressure. The residue was
chromatographed on a silica gel column (100% DCM to 10% MeOH in DCM) to yield 11b
1
(
134 mg, 32%) as a brown oil. H NMR 400 MHz (CDCl3) δ 7.90 (s, 2H), 7.77 (s, 2H), 7.39
(dd, J = 8.7, 1.2 Hz, 2H), 7.31 (d, J = 8.7 Hz, 2H), 6.07 (s, 1H), 4.02 (s, 7H). LCMS (ESI, m/z):
+
293 [M+H] .
ACS Paragon Plus Environment
29