Structure based design of novel irreversible FAAH inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.07.101

Fatty acid amide hydrolase (FAAH) has attracted significant attention due to its promise as an analgesic target. This has resulted in the discovery of numerous chemical classes as inhibitors of this potential therapeutic target. In this paper we disclose

Full text of DOI:10.1016/j.bmcl.2009.07.101

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5970–5974
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure based design of novel irreversible FAAH inhibitors
*
Jane L. Wang , Scott J. Bowen, Barbara A. Schweitzer, Heather M. Madsen, Joseph McDonald,
Matthew J. Pelc, Ruth E. Tenbrink, David Beidler, Atli Thorarensen
Pfizer Global Research and Development, 700 Chesterfield PKWY West, Chesterfield, MO 63017, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Fatty acid amide hydrolase (FAAH) has attracted significant attention due to its promise as an analgesic
target. This has resulted in the discovery of numerous chemical classes as inhibitors of this potential ther-
apeutic target. In this paper we disclose a new series of novel FAAH irreversible azetidine urea inhibitors.
In general these compounds illustrate potent activity against the rat FAAH enzyme. Our SAR studies
allowed us to optimize this series resulting in the identification of compounds 13 which were potent
inhibitors of both human and rat enzyme. This series of compounds illustrated good hydrolase selectivity
along with good PK properties.
Received 7 July 2009
Revised 17 July 2009
Accepted 22 July 2009
Available online 25 July 2009
Keywords:
Fatty acid amide hydrolase
FAAH
Ó 2009 Elsevier Ltd. All rights reserved.
Endocannabinoid anandamide
Inflammation
Irreversible FAAH inhibitors
The endocannabinoid system has been under intense scrutiny
due to its multiple therapeutic implications.1 Fatty acid amide
hydrolase (FAAH) has been identified as an important regulator
of signaling molecules in this complex pathway.2 FAAH is an inte-
gral membrane protein that hydrolyzes bioactive amides, from the
endocannabinoid anandamide and agonists of the peroxisome
proliferator-activated receptors, such as N-oleoylethanolamine
and N-palmitoylethanolamine, to free fatty acid and ethanolamine.
Genetic or pharmacological inactivation of FAAH leads to analgesic,
anti-inflammatory, anxiolytic, and anti-depressant phenotypes in
rodents. This occurs without showing the undesirable side effects
observed with direct cannabinoid receptor agonists. This indicates
that FAAH represents an attractive therapeutic target for the treat-
ment of pain, inflammation, and other central nervous system dis-
orders. Several reversible and irreversible FAAH inhibitors have
been reported including electrophilic ketones (e.g., OL-135) and
carbamates (e.g., URB597), Figure 1.3 A series of publications from
Pfizer have detailed the optimization of piperidine/piperazine urea
analogs resulting in the identification of PF-3845.4 PF-3845 has
illustrated potent in vivo analgesic effects correlating with inhibi-
tion of FAAH. These ureas inhibit FAAH by covalently modifying
the enzyme’s active site serine nucleophile and are completely
selective for FAAH relative to other mammalian serine hydrolases.
Optimization of irreversible inhibitors requires an alternative ap-
proach compared to optimization of reversible inhibitors.5 In a
seminal report from Pfizer, Ahn described the use of the second
order rate constant kinact/Ki as the appropriate way to describe
inhibitor potency.4a Unlike IC50 values, kinact/Ki values do not
change with variable pre-incubation times and have been de-
scribed as the best measure of potency for irreversible inhibitors.
In our ongoing chemistry program aimed at identifying potent
inhibitors of FAAH, it was apparent that the unique properties pos-
sessed by PF-3845 (kinact/Ki = 14,310 (Mꢀ1 sꢀ1)) had to be incorpo-
rated in any new analog design. Our desire in expanding the
chemical space of FAAH inhibitors was to explore what variations
were tolerated as replacement of the piperidine ring system. The
piperidine ring system was described as core structural motif in
publications and patents from several companies. In this Letter,
we describe the synthesis and structure–activity relationships
(SAR) of a series of irreversible azetidine urea FAAH inhibitors.
* Corresponding author. Tel.: +1 636 247 5489; fax: +1 636 247 2086.
E-mail addresses: jane.l.wang@pfizer.com, j147wang@yahoo.com (J.L. Wang).
Figure 1. Structures of FAAH inhibitors.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.07.101

J. L. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5970–5974
5971
The availability of an X-ray structure of the piperidine analog,
PF-3845, complexed to FAAH allowed us to visualize the structural
requirements needed to design a replacement for the core piperid-
inyl moiety, while maintaining other key interactions. Covalently
adducted FAAH-ligands were driven by key constraints. The posi-
tion of the carbonyl moiety was critical because it made hydrogen
bonds to several amide protons (residues 238–241) as part of the
stabilization of transition state. At the distal end of the ligand,
the 5-trifluoromethyl-pyridinyl moiety offered favorable potency
and in vivo properties. The HetAr-O-Ph moiety of PF-3845 formed
both aromatic and van der Waals contacts with the protein. Addi-
tionally, PF-3845 potency was likely also enhanced by conforma-
tional restriction which favored the bound conformer.
At the outset, a computational approach was utilized to assist in
the evaluation of various linkers, to insure critical interactions
were preserved, and that ligands were able to adopt a bound con-
formation with minimal strain. Shorter azetidine linkers (e.g., 2-
carbon linker) reduced conformational flexibility, and were not
able to make the same favorable distal interactions found in the
reference structure, PF-3845. The longer linker required conforma-
tional strain to maintain both the distal interactions and position-
ing of the carbonyl moiety. The 3-carbon linker azetidines
maintained critical interactions within the active site similar as
the parent, PF-3845, Figure 2. This docking result convinced us to
explore the 3-carbon linker azetidine as the replacement.
Scheme 1. Reagents and conditions: (a) K-OtBu, THF, ꢀ50 °C; (b) TBAF, THF; (c)
Ar1X, K2CO3, DMF; (d) TFA, CH2Cl2; (d) acetonitrile, DIEA, aryl phenylcarbamate; (f)
SFC separation;8 (g) H2, 10% Pd–C, MeOH.
Table 1
Urea Heads effect on FAAH inhibition
Compounds
Ar2
h kinact/Ki (Mꢀ1 sꢀ1
)
r kinact/Ki (Mꢀ1 sꢀ1
)
clog D
*
Our initial foray was the preparation of the model azetidine 6
with a 3-carbon linker, Scheme 1. Horner–Emmons olefination of
compound 26 with acetaldehyde 17 provided compound 3 in 78%
yield. Compound 3 was then deprotected in 98% yield and the
resulting phenol was reacted with a range of electron deficient aryl
halides forming the biaryl ethers 4a–d. The Boc group was re-
moved with TFA and the corresponding azetidines were treated
with phenyl carbamates, which formed the ureas 5 in 40–86%
yield.4d Hydrogenation of 5 with 10% Pd–C in MeOH obtained the
key analogs 6a–d. The olefin isomers were separated by SFC chro-
matography to provide pure geometrical isomers 5 cis and 5 trans.8
The initial small set of compounds 6a–d probed our design by
incorporating a fixed tail piece while modulating the aryl urea,
Table 1. Selection of the aryl groups was dictated by previous expe-
rience.4d It was gratifying to find that the ureas 6a–d inhibited
FAAH activity. This data validated our design that the azetidine
ring system was a viable replacement for the piperidine. We subse-
quently did further validation of our design by preparing com-
pounds that had a shorter and a longer linker between the
6a
3030
1840
2.32
N
N
O
N
6b
6c
1450
1090
15300
688
3.97
3.54
N
*
*
*
NH
6d
881
5160
3.21
N
azetidine and the aryl group finding that the optimal linker was
a 3-carbon spacer (data not shown).
Since compounds 6a–d were prepared through intermediate 5,
we tested those intermediates as well. The analogs 5 were a mix-
ture of olefin isomers, which displayed activity comparable to 6.
This allowed us to utilize 5 to rapidly explore variation of the tail
and the urea in a library format. First we designed the library con-
taining the tail pieces existing in 6a–d and varied only urea heads.
We found that only a very few aryl ureas were tolerated (data not
shown). The second library focused on variation of the aryl ether
tail pieces and incorporated only the pyridazine and the dimethyl-
isoxazole as the aryl urea head. The SAR revealed that the human
FAAH inhibitory potency was highly sensitive to tail pieces and
only a few aryl groups tolerated, Figure 3a. These inhibitors all
had in common with a 4-substituent in the tail. Although the strik-
ing structural differences in both tether length and size of ring sys-
tem, azetidine analogs shared a very similar pattern, which was
identified for the piperidine ureas.4c We evaluated the compounds
against both human and rat FAAH enzyme and in general, these
azetidine inhibitors were more potent against the rat enzyme, Fig-
ure 3b. The analogs with pyridazine urea head displayed the best
human FAAH inhibition (Fig. 3a, ). On the other hand, analogs
with the dimethyl oxazole compounds showed a preference for
the rat FAAH over human FAAH enzyme (Fig. 3b, ). This clearly
illustrated that the FAAH isozyme selectivity could be modulated
by the appropriate combination of head and tail.
Figure 2. PF-3845 (green, X-ray), superimposed against model of compound 6
(red). Compounds were evaluated for strain using Schrodinger conformational
search, by fixing the ligand covalently to Ser241, performing a conformational
search within the active site using the OPLS force field, then performing
a
conformational search of the free ligand, and evaluating the relative potential
energy difference.

5972
J. L. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5970–5974
Figure 4. Superposition of conformationally optimized (covalently attached to
Ser241) saturated (light green) 3-carbon linker 6a, 5a trans (light blue), 13a trans
(red), and 13a cis (orange).
The olefin 13 was achieved from aldehyde 7 by reacting with
carbon tetrabromide and triphenylphosphine providing the dib-
romoalkene 8, Scheme 2. Selective mono-debromination of 8 was
achieved with NH4Cl, MeOH, and Zn dust to give the bromoalkene
compound 9 in 48% yield. Compound 109 in a mixture of THF and
DMSO was added slowly to the activated Zn10 at temperature be-
low 65 °C. After removal of precipitates by filtration, the active
benzylic-ZnCl reagent solution was then added to the mixture of
9 and Pd catalyst under argon conditions to form 12 as a mixture
of olefin isomers. After deprotection of 12, the intermediate was
then reacted with phenyl carbamate to obtain desired compound
13.
The final olefin 18 was prepared from aldehyde 14 by conver-
sion to the alkene 15, Scheme 3.11 Subsequent hydroboration of al-
kene 15 followed by a Suzuki coupling with 1612 provided the
desired key intermediate 17. Deprotection and urea formation pro-
vided compound 18.
The final analog elected to explore conformational space was
the acetylene analog 21, Scheme 4. This was achieved from benzyl-
chloride 10, which was converted to a benzyl bromide 19. Com-
pound 8 was treated with n-BuLi at ꢀ78 °C, followed by the slow
addition of 19, and the resulting mixture was stirred overnight at
room temperature to afford the key intermediate 20.11 Simple
functional group manipulation provided the desired product 21.
The evaluation of all possible olefin regioisomers suggested that
the location and the cis or trans disposition of the double bond
were important to successful FAAH inhibition, Table 2. The double
bond at position-2 provided analogs 13 with potent FAAH
Figure 3. Illustrates SAR from libraries containing compounds of the generic
structure 5. (a) Tail pieces effect on human FAAH inhibition. (b) Human and rat
enzyme activity correlated with the urea head being isoxazole versus pyridazine.
With demonstration of activity for the 3-carbon linked azeti-
dine series, we sought to identify conformational restraints by fur-
ther optimization of the scaffold, which might further improve the
potency and the properties of these inhibitors. Modeling suggested
that the saturated linker may occupy a predominantly transoid
conformation, Figure 4. This transoid arrangement suggested an
opportunity to explore added rigidity by systematically introduc-
ing cis- and trans-olefins into the linker. Introduction of a double
bond into the 3-carbon linker was ideal because it was a simple
form of restriction, and effectively probes the conformational
requirements for the ligand. Introduction of double bonds was sig-
nificantly easier to access than any ring constrained version.
Scheme 2. Reagents and conditions: (a) PPh3, CBr4, CH2Cl2, yield 77%; (b) NH4Cl,
MeOH, THF, 0 °C 20 min then Zn, yield 48%; (c) Zn, TMSCl, 1,2-dibromoethane,
DMSO, THF; (d) 9, [1,10-bis(diphenyl phosphino) ferrocene]dichloropalla-
dium(II)methylene-chloride complex, CuI, yield 53%; (e) 2.5 equiv 4 M HCl in
dioxane, yield 91%, CH2Cl2; (f) acetonitrile, DIEA, aryl phenylcarbamate, yield, 43%.

J. L. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5970–5974
5973
Table 3
Rat PK profile of selective compounds
Compd #
Cl (mL/min/kg)
V (L/kg)
T1/2 (h)
BA (%)
13 cis
13 trans
1.7
11.1
0.6
1.8
4.1
2.2
73
93
(data not shown). Location of the olefin at position-3 (18) and
propargyl-2 (21) led to poor potency for both human and rat FAAH.
This SAR illustrated that incorporation of conformational con-
straint in the appropriate location in the 3-atom tether resulted
in significant potency improvement compared to non-constrained
analogs (Table 2).
Scheme 3. Reagents and conditions: (a) methyltriphenylphosphonium-bromide, n-
BuLi, THF, ꢀ78 °C, yield 26%; (b) 9-BBN, NaOH, Pd-tetrakis (triphenylphosphine),
yield 91%; (c) 7.3 equiv TFA in CH2Cl2; (d) aryl phenylcarbamate acetonitrile, DIPEA.
The selectivity for other hydrolases was done in a comparable
fashion as previously described.4 The in vitro ABPP proteome pro-
filing was done on both olefin isomers 13. The selectivity profile
observed for PF-3845 was retained in this new azetidine scaffold.
We further evaluated the in vivo PK properties of the olefin 13 (Ta-
ble 3). It was very gratifying to find that the olefin 13 had suitable
PK properties in rat for further rat studies. The olefin 13 demon-
strated low clearance (1.7 and 11.1 <30% rat liver blood flow)
and high bioavailability (73% and 93%), but neither 13 cis nor 13
trans had comparable in vivo13 efficacy to PF-3845, these data sug-
gested that further optimization of azetidine would be required.4c
In summary, we have described the design of a novel scaffold,
which replaced the piperidine of PF-3845 with 3-carbon linked
azetidine. The azetidine compounds were potent inhibitors of both
human and rat FAAH. The incorporation of a conformational
restriction had significant effect on in vitro potency. Finally, the
key properties of PF-3845 of high selectivity against other serine
hydrolases and excellent PK were retained in the azetidine series.
The preparation of ring constrained analogs that replace the olefin
Scheme 4. Reagents and conditions: (a) acetone, LiBr, yield 100%; (b) 8, n-BuLi, THF,
ꢀ78 °C, yield 7%; (c) 2.5 equiv 4 M HCl in dioxane, CH2Cl2; (d) aryl phenylcarba-
mate, acetonitrile, DIEA.
Table 2
Effect of olefin location on FAAH inhibition
in the 3-carbon linker will be the subject of
communication.
a future
Acknowledgments
Compounds
Linker
h kinact/Ki (Mꢀ1 sꢀ1
)
r kinact/Ki (Mꢀ1 sꢀ1
)
clog D
6a
Single
3030
782
2100
9000
9340
621
1840
na
2.32
3.54
3.21
2.02
2.02
2.28
1.64
We would like to thank Prof. Benjamin F. Cravatt for his advice
and extensive discussion on this project. J. Collins for chiral resolu-
tion and S. Yang for 2D NMR determination. In addition, T.K. Noma-
nbhoy at ActivX for ABPP proteome profiles; S. Wene for animal
dosing and sample collection for PK analysis and R. Riley for bioan-
alytical method development, sample analysis, and pharmacoki-
netic analysis. Finally we would like to thank J. Rumsey for PK
parameter assessment and L. Kirkovsky for PK discussions.
5 cis
5 trans
13 cis
13 trans
18
Olefin-1
Olefin-1
Olefin-2
Olefin-2
Olefin-3
Propargyl-2
2310
5390
2380
1310
465
21
861
inhibition for both human and rat enzyme. Surprisingly, the stereo-
chemistry of 13 seemed not to be sensitive to human FAAH inhibi-
tion (kinact/Ki = 9000 and 9340 (Mꢀ1 sꢀ1)), however there was some
impact of the olefin stereochemistry for inhibition of rat FAAH. This
was slightly unexpected, based upon the initial model of the satu-
rated linker. However, the low energy bound conformer of 13 cis
suggested that the cis-olefin was accommodated by way of an axial
disposition relative to the azetidine, rather than equatorial, as sug-
gested by 13 trans (see Fig. 4, red and orange). Trans 5 displayed
similar potency for both human (ꢁkinact/Ki = 2100 (Mꢀ1 sꢀ1)) and
rat (kinact/Ki = 2310 (Mꢀ1 sꢀ1)) FAAH inhibition, while 5 cis was a
poor inhibitor. Comparing olefins at the 1- or 2- position, 5 trans
rigidified the bond attached to the central aromatic ring as well
as the double bond itself. In contrast, 13 trans, by virtue of its posi-
tion as flanked by methylenes rigidified a smaller portion of the li-
gand. The slightly greater flexibility found in 13 trans relative to 5
trans allowed the ligand to adopt a more favored bound conforma-
tion. Conformational searching as described by the legend of Figure
2 suggested that the bound conformer of 5 trans was slightly more
strained than 13 trans, compared to their respective free ligands
References and notes
1. (a) Lambert, D. M.; Fowler, C. J. J. Med. Chem. 2005, 48, 5059; (b) Pacher, P.;
Batkai, S.; Kunos, G. Pharmacol. Rev. 2006, 58, 389; (c) Di Marzo, V. Nat. Rev.
Drug Disc. 2008, 7, 438.
2. Ahn, K.; McKinney, M. K.; Cravatt, B. F. Chem. Rev. 2008, 108, 1687.
3. (a) Seierstad, M.; Breitenbucher, J. G. J. Med. Chem. 2008, 51, 7327; (b) Boger, D.
L.; Miyauchi, H.; Du, W.; Hardouin, C.; Fecik, R. A.; Cheng, H.; Hwang, I.;
Hedrick, M. P.; Leung, D.; Acevedo, O.; Guimaraes, C. R.; Jorgensen, W. L.;
Cravatt, B. F. J. Med. Chem. 2005, 48, 1849.
4. (a) Ahn, K.; Johnson, D. S.; Fitzgerald, L. R.; Liimatta, M.; Arendse, A.; Stevenson,
T.; Lund, E. T.; Nugent, R. A.; Nomanbhoy, T. K.; Alexander, J. P.; Cravatt, B. F.
Biochemistry 2007, 46, 13019; (b) Mileni, M.; Johnson, D. S.; Wang, Z.; Everdeen,
D. S.; Liimatta, M.; Pabst, B.; Bhattacharya, K.; Nugent, R. A.; Kamtekar, S.;
Cravatt, B. F.; Ahn, K.; Stevens, R. C. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 12820;
(c) Ahn, K.; Johnson, D. S.; Mileni, M.; Beidler, D.; Long, J. Z.; McKinney, M. K.;
Weerpana, E.; Sadagopan, N.; Liimatta, M.; Smith, S. E.; Lazerwith, S.; Stiff, C.;
Kamtekar, S.; Bhattacharya, K.; Zhang, Y.; Swaney, S.; Becelaere, K. V.; Stevens,
R. C.; Cravatt, B. F. Chem. Biol. 2009, 16, 411–420; (d) Johnson, D. S.; Ahn, K.;
Kesten, S.; Lazerwith, S. E.; Song, Y.; Morris, M.; Fay, L.; Gregory, T.; Stiff, C.;
Dunbar, J. B.; Liimatta, M.; Beidler, D.; Smith, S.; Nomanbhoy, T. K.; Cravatt, B. F.
Bioorg. Med. Chem. Lett. 2009, 19, 2865.
5. Copeland, R. A. Enzymes: A practical introduction to structure, mechanism, and
data analysis, 2nd ed.; Wiley-VCH: New York, 2000. pp 318–349.
6. Compound 2 was prepared from 3-hydroxy methylbenzoate in four steps.

5974
J. L. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5970–5974
7. Matassa, V. G. WO 9318029 A1, 1993.
11. Acharya, H. P.; Kobayashi, Y. Tetrahedron 2006, 62, 3329.
8. Olefin separation condition: Set up AI 30ꢂ250 mm chiral column elute with
20% EtOH and 80% CO2. Sample dissolved in 20 mL EtOH, 25 mg/mL, 1 mL per
inject. The olefin stereochemistry was assigned by 2D NMR.
9. Compound 10 was prepared as described in WO 08047229, 2008.
10. Corley, E. G.; Conrad, K.; Murry, J. A.; Savarin, C.; Holko, J.; Boice, G. J. Org. Chem.
2004, 69, 5120.
12. Compound 16 was prepared through the similar sequence as compound 9.
13. The Pfizer Institutional Animal Care and Use Committee reviewed and
approved the animal use in these studies. The animal care and use program
is fully accredited by the Association for Assessment and Accreditation of
Laboratory Animal Care, International.