Discovery of AAT-008, a novel, potent, and selective prostaglandin EP4 receptor antagonist

Article abstract of DOI:10.1016/j.bmcl.2017.01.067

Starting from acylsufonamide HTS hit 2, a novel series of para-N-acylaminomethylbenzoic acids was identified and developed as selective prostaglandin EP4 receptor antagonists. Structural modifications on lead compound 4a were explored with the aim of impr

Full text of DOI:10.1016/j.bmcl.2017.01.067

Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of AAT-008, a novel, potent, and selective prostaglandin EP4
receptor antagonist
Yoshiyuki Okumura a, ,1, Tatsuya Yamagishi b,1, Seiji Nukui c, Kazunari Nakao d
⇑
a AskAt Inc., 4F Ito Bldg., 4-11-28 Meieki Nakamura-ku, Nagoya 450-0002, Japan
b Discovery Research, RaQualia Pharma Inc, 8F Daiwa Meieki Bldg., 1-21-19 Meieki Minami, Nakamura-ku, Nagoya 450-0003, Japan
c INC Research, 16F Shinagawa East One Tower, 2-16-1 Konan, Minato-ku, Tokyo 108-0075, Japan
d Discovery Research, Mochida Pharmaceutical Co., Ltd., 722 Uenohara, Jimba, Gotemba, Shizuoka 412-8524, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Starting from acylsufonamide HTS hit 2, a novel series of para-N-acylaminomethylbenzoic acids was
identified and developed as selective prostaglandin EP4 receptor antagonists. Structural modifications
on lead compound 4a were explored with the aim of improving potency, physicochemical properties,
and animal PK predictive of QD (once a day) dosing regimen in human. These efforts led to the discovery
of the clinical candidate AAT-008 (4j), which exhibited significantly improved pharmacological profiles
over grapiprant (1).
Received 3 October 2016
Revised 17 January 2017
Accepted 24 January 2017
Available online xxxx
Keywords:
Ó 2017 Elsevier Ltd. All rights reserved.
EP4 receptor antagonist
Structure–activity relationship (SAR)
Prostaglandin E2
Inflammation
Pain
Autoimmune diseases
Cancer
Clinical candidate
Prostaglandin E2 (PGE2) is a pro-inflammatory mediator gener-
ated from arachidonic acid by the action of cyclooxygenase (COX)
isoenzymes under inflammatory conditions. Four PGE2 receptor
subtypes (EP1, EP2, EP3, and EP4) responsible for different pharma-
cological properties have been cloned and classified.1,2 The EP4
subtype, a G-protein-coupled receptor (GPCR), stimulates cyclic
adenosine monophosphate (cAMP) production3 and is distributed
in a wide variety of tissues suggesting an important role of EP4
receptor in PGE2-mediated biological events such as inflamma-
tion,4 pain,5 and cancer.6–10 Selective blockade of the PGE2 signal-
ing through the EP4 receptor pathway represents an attractive
approach to discover novel analgesic, immunomodulating, and
antineoplastic agents. Analgesic potentials are clearly supported
by the reduction of pain and inflammation in EP4 receptor knock
out/knock down animals and the similar results using EP4 receptor
selective antagonists in animal studies.11,12 Since the PGE2-EP4 sig-
nal blockade does not affect actions of the other subtype PGE2
receptors as well as other prostanoids, selective EP4 antagonists
might provide analgesic effects without adverse events observed
with nonsteroidal anti-inflammatory drugs (NSAIDs) or cyclooxy-
genase-2 (COX-2) inhibitors,13–15 and are expected to be a new
therapy for the treatment of both acute and chronic inflammatory
pain. In addition, selective EP4 receptor antagonists are also
expected to offer an attractive therapeutic approach for autoim-
mune diseases such as inflammatory bowel disease (IBD), rheuma-
toid arthritis (RA), and multiple sclerosis (MS)16–18 through the
inhibition of interleukin-23 (IL-23) production and suppression of
T helper 1 (Th1) and T helper 17 (Th17). Moreover, the recent
reports suggest that the EP4 receptor which is expressed in certain
types of cancer promotes tumor cell proliferation and metastasis.19
Therefore, selective antagonism of the EP4 receptor might have sig-
nificant clinical potential for the treatment of colorectal, breast,
prostate, lung, gastric, bladder, head and neck, hepatocellular,
pancreatic, and ovarian cancers. Herein we report novel and selec-
tive EP4 receptor antagonists of benzoic acid with the nicotinamide
or benzamide scaffolds.
Grapiprant (Fig. 1) is a selective antagonist for prostaglandin E2
(PGE2) receptor subtype 4 (EP4) identified as a clinical candidate
for the treatment of inflammatory pain associated with
osteoarthritis (OA). It is currently under development for use in
humans20 and dogs.21 The projected dosing regimen of grapiprant
⇑
Corresponding author.
E-mail address: yoshiyuki.okumura@askat.co.jp (Y. Okumura).
1
These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.bmcl.2017.01.067
0960-894X/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067

2
Y. Okumura et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
transformation24,25 of the N-acyl phenylsulfonamide moiety to
the corresponding carboxy group was well tolerated and resulted
in the identification of compound 3 with reduced molecular
weight. Compound 3 displayed modest EP4 selective functional
antagonism (IC50: 370 nM) and good stability in human liver
microsomes (HLM) (T1/2 > 120 min), however it did not have
acceptable physicochemical properties (typically for lead com-
pounds, solubility in phosphate-buffered saline (PBS) > 10 lM;
molecular weight <400; ALogP < 4) (Fig. 2). Next, with further
reduction of MW and lipophilicity (ALogP and cLogD) in mind, a
structural similarity search by using a simplified pharmacophore
query as shown in Fig. 2 led to identify ca. 1000 compounds. The
subsequent filtering of these compounds by applying Lipinski’s
rule of 526,27 and eliminating compounds with toxicophore28–30
narrowed down to 50 hit compounds. These 50 compounds were
evaluated in a series of assays of functional EP4 receptor antago-
nism, binding selectivity against EP receptors, HLM stability,
human cytochrome P450 inhibition, aqueous solubility, and mem-
brane permeability. As a result, compound 4a was shown to be
superior to the others, demonstrating EP4 selective functional
antagonism (IC50: 575 nM), binding affinity for EP4 receptor (Ki:
73 nM), stability in HLM (T1/2 > 120 min), no notable CYP 450 inhi-
bition, high Caco-2 cell permeability (Papp: 25 Â 10À6 cm/s),31 good
Fig. 1. Selective EP4 antagonist, grapiprant under development.
for humans is 50–100 mg, PO, BID (twice a day) based on the phar-
macology and pharmacokinetic (PK) data in clinical studies.
In order to identify another clinical candidate with improved
efficacy, safety, and pharmacokinetic profiles with potential for
QD (once a day) dosing, we embarked on a back-up discovery pro-
ject to discover the second generation of grapiprant (1).22 The tar-
get profiles of this candidate are summarized as follows. The back-
up candidate should demonstrate improved potency over grapipr-
ant, i.e.; (1) antagonize PGE2-mediated cAMP elevation in transfec-
tants expressing human EP4 receptor-with a pA2 > 8.6; (2) exhibit
selectivity of >100 over other PG receptor subtypes; (3) clearly
demonstrate improved oral activity over grapiprant in key acute
and chronic inflammatory pain models; i.e., carrageenan induced
mechanical hyperalgesia and complete Freund’s adjuvant (CFA)
induced weight bearing deficit in the rat; (4) demonstrate a PK pro-
file in animals predictive of QD dosing in human. From a medicinal
chemistry perspective, the back-up compound should have
improved physicochemical characteristics; lower molecular
weight and lower lipophilicity while improving efficacy and PK
profile. In order to meet the above criteria, we pursued an alterna-
tive core structure different from the sulfonylurea since the struc-
ture–activity relationships (SAR) studies around grapiprant
revealed compounds consisting of the sulfonylurea core had extre-
mely low volume of distribution and high to moderate clearance in
rats and other experimental animals.
solubility in PBS (>10 lM), lower MW (MW: 384.33), and accept-
able lipophilicity (ALogP: 3.61, cLogD: 2.14). The hit-to-lead efforts
resulted in the identification of compound 4a that has a core struc-
ture of para-N-acylaminomethylbenzoic acid (Fig. 2). The mole-
cules containing carboxylic acid often have undesirable metabolic
instability, limited permeability, and potential toxicities. Despite
the drawbacks of the carboxylic acid functional group, 4a exhibited
selective EP4 functional antagonism with a suitable metabolic pro-
file in vitro and good physicochemical properties. Thus, we envis-
aged that the optimization efforts around 4a would provide the
backup candidate that meets the target profiles.
The optimization of the lead compound (4a) was initiated and
the initial key SAR and the results of the structural modifications
of lead compound 4a are summarized in Fig. 3. Replacement of
the carboxylic acid moiety with other functional groups led to loss
of functional activity against EP4 receptor. Although the corre-
sponding tetrazole exhibited slightly increased functional activity,
the tetrazole analog was a substrate for efflux pumps and a strong
inhibitor of CYP3A4. Shifting the carboxylic acid moiety from para-
to meta-position showed loss of intrinsic activity. The benzoic acid
moiety was not replaceable by nicotinic acid, cyclohexanecar-
boxylic acid, or 4-thiazolecarboxylic acid. Moreover, modifications
of the amide moiety by exchanging the nitrogen with methylene,
reduction of the amide carbonyl, or N-methylation of the nitrogen
High throughput screening (HTS) of the Pfizer compound library
using a human EP4 functional assay measuring PGE2-induced
cAMP formation in HEK-293 cells expressing human EP4 receptor
and the subsequent verification in a membrane binding assay
using [3H]PGE2 resulted in the identification of N-acyl sulfonamide
2 as a moderately potent EP4 antagonist (IC50: 302 nM) without
EP4 selectivity over other subtypes (Fig. 2). HTS hit 2 has high
molecular weight (MW: 610.08) and is quite lipophilic (ALogP:
5.70, cLogD: 5.80).23 Further screening of HTS hit 2 were initiated
by using the in-house compound libraries. First, a bioisosteric
Fig. 2. Hit to lead: genesis of benzoic acid EP4 antagonists.
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067

Y. Okumura et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
3
Fig. 3. Lead optimization: initial SAR study results.
Fig. 4. Lead development: completion of optimization strategy for lead compound 4a.
also reduced intrinsic activity. These early SAR studies of the para-
(aminomethyl)benzoic acid moiety of lead compound 4a showed
that the carboxylic acid group at the para-position was a key ele-
ment for the EP4 antagonist activity. A limited class of substituents
on the phenoxypyridine moiety of lead compound 4a, maintained
or enhanced the potencies when satisfied regiochemical require-
ments. The introduction of a substituent to the 4-position of pyri-
dine ring caused loss of intrinsic activity. Additionally, it was found
that the aliphaticoxy groups were not suitable surrogates for the
phenoxy group on the 2-position of the pyridine of 4a. Even though
the aliphaticoxy substituted analogs retained intrinsic activity,
their metabolic stabilities in HLM were decreased.
In order to develop detailed SAR around lead compound 4a for
improving its profile, structural modifications were performed in
three portions of lead compound 4a: the benzylic position of
para-aminomethylbenzoic acid moiety for Portion A, the central
ring core with substituents for Portion B, and the substitution pat-
tern of the phenoxy group for Portion C (Fig. 4). Furthermore, this
strategy allowed us to easily construct the molecule in a concerted
manner for effective SAR studies.32 We aimed to improve not only
pharmacological properties, but also physiochemical properties
such as MW < 450 and cLogD < 3,33 although the compounds with
a MW over 450 and a cLogD over 3 were prepared for exploring
SAR involving the selective EP4 antagonist.
conditions. The SNAr reaction of the benzoates derived from ben-
zoic acids 6 (A = CH) required more severe reaction conditions than
that of the nicotinates derived from nicotinic acids 5 (A = N). The
target benzoic acids 4 and 11 were prepared by reaction of car-
boxylic acids 7 or 8 with para-(methoxycarbonyl)benzylamine
using EDCI/HOBt condensation conditions and subsequent hydrol-
ysis of the methyl ester or t-butyl ester intermediates 9 or 10
thereby obtained.
The human EP4 binding and functional activities of representa-
tive compounds are presented in Table 1. Switching the aromatic
ring in portion B from pyridine to benzene was tolerated (4a vs
11a). Replacement of 5-fluoro substituent on the pyridine ring
moiety of lead compound 4a with 5-chloro substituent showed
about 2-fold increase in EP4 binding and functional activity (4a
vs 4b). The 5-chlorobenzene analogs were almost equipotent to
the corresponding 5-chloropyridine analogs (4b vs 11b) although
it is concerned that benzene 11b has an evidently high cLogD value
by 0.6 units relative to pyridine 4b (3.21 vs 2.60). Substitution of
hydrophilic functionality such as carboxamides, sulfones and sul-
fonamides on the benzene ring was also attempted, however none
of them exhibited sufficient EP4 antagonistic activity at all (data
not shown). 5-Methylpyridine and 6-methylpyridine analogs (4c
and 4d) displayed substantially decreased EP4 functional activity.
Incorporation of a methyl group at the benzylic position in portion
A led to a marked improvement in EP4 binding and functional
activity in a stereochemical dependent manner (4e, 4f, and 11c).
The (S)-methyl isomers were dramatically more active than the
corresponding antipodes (4f vs 4g). The geminal methyl groups
or ethyl group at the same position decreased the binding affinity
(4h and 4i). Key compounds exhibiting EP4 antagonist activities
were selected and tested to assess selectivity against the other
EP receptor subtypes. These compounds did not show any signifi-
cant binding to EP1 and EP3 at the highest concentration tested
Our optimization studies on lead compound 4a started by
exploring portions A and B. The synthetic route to the target com-
pounds from easily available materials is outlined in Scheme 1.
Starting from ortho-halo-substituted aromatic acids 5 or 6, the cor-
responding methyl esters were formed, which were then converted
to ortho-phenoxy-substituted aromatic acids 7 or 8 with nucle-
ophilic aromatic substitution (SNAr) reaction followed by alkali
hydrolysis. Depending on the aromatic ring (A = CH or N) of
starting materials 5 or 6, the phenol was reacted under different
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067

4
Y. Okumura et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
Scheme 1. R1, R2, R3, R4, and R5 fragments are given in Tables 1 and 2; Reagents and conditions: (a) H2SO4, MeOH reflux, 58–99%; (b) NaH, ‘‘phenol”, DMF, 120 °C, 65–87% or
K2CO3, ‘‘phenol”, toluene, reflux, 77–99% or Cu, CuI, K2CO3, ‘‘phenol”, DMF, reflux, 68–92%; (c) NaOH-H2O, MeOH, r.t., 80–99%; (d) EDCI, HOBt, ‘‘benzylamine”, Et3N, CH2Cl2, r.t.,
73–93%; (e) NaOH-H2O, MeOH, r.t., 60–91% or TFA, CH2Cl2, r.t., 79–93%.
Table 1
Optimization around portion A & B of para-N-acylaminomethylbenzoic acid.
Compd
R1, R2
R3
R4
MW
cLogD
hEP4 functional
cAMP IC50 (nM)
Binding IC50 (nM)
Caco-2a A-B
Papp (Â10À6 cm sÀ1
)
hEP4
hEP1
hEP2
hEP3
4a
H, H
H, H
H, H
H, H
H, H
H, H
(S)-Me
(S)-Me
(S)-Me
(R)-Me
Me, Me
rac-Et
F
F
H
H
H
H
H
Me
H
H
H
H
H
H
384.33
383.35
400.79
399.80
380.37
380.37
398.36
414.81
413.83
414.81
412.39
412.39
2.14
2.75
2.60
3.21
2.42
2.21
2.54
2.99
3.61
2.99
3.07
3.21
575
317
254
101
2940
>5000
143
25.9
38.2
1210
N.D.
N.D.
104
124
37.9
21.6
N.D.
N.D.
28.7
9.68
6.59
N.D.
545
181
>20,000
>20,000
>20,000
N.D.
N.D.
N.D.
>20,000
236
>20,000
>20,000
>20,000
N.D.
N.D.
N.D.
25.3
36.3
20.1
51.4
N.D.
N.D.
30.2
30.3
32.4
25.3
36.3
20.1
11a
4b
11b
4c
4d
4e
4f
11c
4g
Cl
Cl
Me
H
F
Cl
Cl
Cl
F
1290
N.D.
N.D.
N.D.
N.D.
4360
709
N.D.
N.D.
N.D.
N.D.
N.D.
>20,000
>20,000
N.D.
N.D.
N.D.
>20,000
>20,000
N.D.
N.D.
N.D.
4h
4i
F
N.D. = Not determined.
a
In vitro permeability (Papp, apical to basolateral) at pH 6.5/7.4 determined in Caco-2 cells.
We utilized Scheme 2 for an alternative synthesis of nicoti-
namide scaffold 12, which is quite effective for derivatization of
portion C. Thus nicotinic acid 14 was coupled with tert-butyl 4-
[(1S)-1-aminoethyl]benzoate using EDCI/HOBt condition to give
2-chloro-nicotinamide 15. Nucleophilic displacement of 2-chloro-
nicotinamide 15 with appropriately substituted phenols afforded
2-phenoxy-nicotinamides 9. A final acid hydrolysis of tert-butyl
ester moiety in intermediates 9 under acidic conditions furnished
respective benzoic acids 4.
A SAR of the representative analogs of portion C are shown in
Table 2. Substitution at the meta-position of the benzene ring in
portion C of nicotinamide scaffold (12) with a fluorine or chlorine
atom enhanced the functional activity as compared with the corre-
sponding ortho- or para-positions (4j vs 4k or 4f, 4l vs 4m or 4n).
Therefore, the meta-substitution was explored in more details. The
introduction of small and electron-withdrawing substituents such
as fluoro, chloro, cyano, and trifluoromethyl at the meta-position of
the benzene ring was found favorable (4j, 4l, 4o, and 4p). Among
Fig. 5. Scaffolds of EP4 antagonist.
(20
l
M). Although some compounds were found to have weak
binding affinity to EP2, most of compounds (4f and 11c) showed
high selectivity to EP4, including compounds with candidate cre-
dentials. In the Caco-2 assay, compound 4f and 11c exhibited suf-
ficiently high permeability.
Based upon these findings on portions A and B, novel scaffolds
of EP4 antagonist, 12 and 13 (R1 & R2 = (S)-Me; R3 = Cl; R4 = H),
shown in Fig. 5, were identified. We next focused on optimizing
portion C of these scaffolds.
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067

Y. Okumura et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
5
Scheme 2. R5 given in Table 2; Reagents and conditions: (a) EDCI, HOBt, ‘‘benzylamine”, Et3N, CH2Cl2, r.t., 70%; (b) BEMP, ‘‘phenol”, toluene, 110 °C, 72–93%;34 (c) TFA, CH2Cl2,
r.t., 61–86%
Table 2
Optimization around portion C of para-N-acylaminomethylbenzoic acid.
Compd
R5
MW
cLogD
hEP4 functional
cAMP IC50 (nM)
Binding IC50 (nM)
Caco-2a A-B
Papp (Â10À6 cm sÀ1
)
hEP4
hEP1
hEP2
hEP3
4j
4k
4f
4l
4m
4n
4o
4p
4q
3-F
2-F
4-F
414.81
414.81
414.81
431.27
431.27
431.27
421.83
464.82
426.85
442.92
480.82
415.80
432.26
413.83
430.28
413.83
2.99
2.99
2.99
3.45
3.45
3.45
2.67
3.73
2.77
3.33
4.91
1.84
2.30
3.61
4.06
3.61
16.3
61.4
25.9
21.8
93.5
94.3
36.2
60.4
340
2.40
N.D.
9.68
5.02
N.D.
N.D.
13.4
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
8.76
N.D.
6.59
>20,000
N.D.
>20,000
N.D.
N.D.
N.D.
>20,000
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
>20,000
N.D.
>20,000
1890
N.D.
4360
N.D.
N.D.
N.D.
3320
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
1620
N.D.
709
>20,000
N.D.
>20,000
N.D.
N.D.
N.D.
>20,000
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
>20,000
N.D.
>20,000
20.1
N.D.
30.3
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
41.9
N.D.
32.4
3-Cl
2-Cl
4-Cl
3-CN
3-CF3
3-OMe
3-SMe
3-OCF3
F
Cl
3-F
3-Cl
4-F
4r
4s
400
1790
175
16ab
16bb
11d
11e
11c
137
30.6
41.1
38.2
N.D. = Not determined.
a
In vitro permeability (Papp, apical to basolateral) at pH 6.5/7.4 determined in Caco-2 cells.
This compound was prepared according to the reaction sequence shown in Scheme 1 by using ‘‘5-substituted-3-hydroxypyridine” instead of ‘‘phenol”.
b
the four analogs, meta-fluoro analog 4j was the most active com-
pound. Although di-fluoro substituted analogs such as 3,4-di-flu-
oro-, 2,3-di-fluoro-, 2,5-di-fluoro-, and 3,5-di-fluoro-analog
exhibited IC50 values of less than 50 nM in the functional assay
(data not shown), these analogs did not surpass the functional
activity of 4j. On the other hand, the substitution with electron-
donating groups such as methoxy or methylthio at the meta-posi-
tion appeared to be less favorable (4q and 4r). Trifluoromethoxy
group at the meta-position was not tolerated at all (4s). Replace-
ment of the phenyl group in portion C by 3-pyridyl group
decreased the functional activity (16a and 16b). Incorporation of
hydrophilic functionality such as carboxamides, sulfones, and sul-
fonamides on the benzene ring also caused substantial decrease in
intrinsic activity (data not shown). A similar tendency in the SAR
studies of the analogs with benzamide scaffold (13) were observed,
however approximately 2-fold less potent than the corresponding
nicotinamide scaffold analogs (4j vs 11d, 4l vs 11e) with the
exception of 11c, which displayed equivalent in vitro activity to
4f. Key compounds (4j, 4o, and 11d) that exhibited potent EP4
functional activity were evaluated for their selectivity against the
other EP receptor subtypes. These potential compounds showed
sufficient EP4 selectivity. Furthermore, compounds 4j, 4o, and
11d had high permeability in the Caco-2 assay.
Based on these SAR studies, six compounds which exhibit IC50
values <40 nM in the human EP4 functional assay were selected
for further characterization. Table 3 shows comparative data of
the six compounds. All the compounds showed high metabolic sta-
bility in HLM (T1/2 > 120 min). Although the intrinsic activity of 3-
cyano analog 4o was sufficient to progress to the next stage, 4o
was found to be positive in the in vitro micronucleus test35 and suf-
fered from poor membrane permeability and high efflux ratio in
the Caco-2 assay at equivalent pH condition (Papp
A to B:
0.644 Â 10À6 cm/s, Papp B to A: 25.8 Â 10À6 cm/s, Efflux Ratio:
40).36 It was found that the 3-chloro analog 4l was not a pure
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067

6
Y. Okumura et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
Table 3
Comparative profiles of potential six compounds.
Compd
R5
hEP4 functional
cAMP IC50 (nM)
hEP4 binding IC50 (nM)
HLM T1/2 (min)
hEP4 functional cAMP pA2
CIMHa MED (mg/kg)
4f
4-F
3-F
3-Cl
3-CN
4-F
25.9
16.3
21.8
36.2
38.2
30.6
9.68
2.40
5.02
13.4
6.59
8.76
>120
>120
>120
>120
>120
>120
8.61
8.97
N.D.
N.D.
8.75
8.90
10
1
N.D.
N.D.
10
4j (AAT-008)
4l
4o
11c
11d
3-F
10
N.D. = Not determined.
a
CIMH = carrageenan induced mechanical hyperalgesia.
As shown in Table 4, AAT-008 (4j) selectively antagonized the
prostaglandin receptor EP4. AAT-008 inhibited [3H]-PGE2 binding
to human EP4 receptor with a Ki value of 0.97 nM and had more
than 1000-fold selectivity over other prostaglandin receptors, such
as EP1, EP2, EP3, DP, FP, IP, and TP. AAT-008 showed potent binding
affinity for human, rat, and dog EP4 receptors, with Ki values of
0.97, 6.1, and 38 nM, respectively. In the human EP4 receptor
highly expressed in HEK293 cells, AAT-008 significantly sup-
pressed PGE2-induced elevation of intracellular cAMP level in a
PGE2-competitive fashion with a pA2 value of 8.97 (1.1 nM). Com-
pared with grapiprant (1), AAT-008 was approximately 4-fold
more potent in the functional assay.
The in vitro ADME profile of AAT-008 (4j) was very promising,
with high stability in HLM. The pre-clinical pharmacokinetic prop-
erties of AAT-008 were also assessed in rats (Sprague-Dawley,
male), dogs (beagle, male), and monkeys (cynomolgus, male). The
experimentally determined parameters are summarized in Table 5.
AAT-008 was absorbed into the systemic circulation following oral
administration to rats, dogs, and monkeys with a Tmax of 0.438–
8.00 h. In all the pre-clinical species, AAT-008 showed good oral
bioavailability in rats, dogs, and monkeys with 73.6%, 80.6%, and
73.3%, respectively. AAT-008 had a pharmacokinetic profile with
low clearance (CLtotal) and moderate volume of distribution (Vdss).
After intravenous injection of AAT-008 at the dose of 1 mg/kg, the
plasma elimination half-life (T1/2) values are 4.59 h in rats, 8.43 h
in dogs, and 14.5 h in monkeys.
Table 4
Comparative in vitro profiles of AAT-008 (4j) and grapiprant (1).
Assay
AAT-008
(4j)
Grapiprant
(1)
In vitro binding to prostaglandin receptors
Recombinant human EP1: Kia (nM)
Recombinant human EP2: Kia (nM)
Recombinant human EP3: Kia (nM)
Recombinant human EP4: Kib (nM)
Recombinant human DP: Kia (nM)
Recombinant human FP: Kia (nM)
Recombinant human IP: Kia (nM)
Human TP (platelet): Kib (nM)
>6000
1836
>6000
>5000
>5000
>5000
0.97 0.09 13
4
>6000
>6000
>6000
>6000
6.1 0.3
2926
>5000
>5000
>5000
Recombinant rat EP4: Kic (nM)
20
24
1
3
Recombinant dog EP4: Kic (nM)
38
1
In vitro inhibition of functional activity
PGE2-induced cAMP in human EP4
transfectant: pA2
8.97 0.06 8.32 0.03
8.34 0.05 8.19 0.16d
c
PGE2-induced cAMP in rat EP4 transfectant:
c
pA2
PGE2-induced cAMP in mouse EP4
transfectant: pA2
9.00
8.11
b
a
b
c
N = 2.
N = 1.
Values are the Mean SE (N = 3).
N = 4.
d
antagonist but a partial agonist, therefore it is assumed that the
functional antagonism of 4l (IC50: 21.8 nM) was caused by its par-
tial agonism. The remaining four compounds (4f, 4j, 11c, and 11d)
were subsequently re-evaluated for their in vitro antagonistic
potencies (pA2 values) against human EP4 receptor and oral activ-
ity in the rat carrageenan induced mechanical hyperalgesia model.
Among them, 4j showed the highest potency which met the
backup candidate criteria (pA2: 8.97), and was selected for preclin-
ical development with code AAT-008.
Moreover, AAT-008 (4j) had a significantly improved pharmaco-
logical profile over grapiprant, as was demonstrated in acute and
chronic inflammatory pain models in rats (Table 6). The oral dosing
of AAT-008 reduced carrageenan induced mechanical hyperalgesia
in rats in a dose dependent manner with an MED 1 mg/kg, PO at 1-
h post dosing (vs 30 mg/kg, PO for grapiprant). In a model of
chronic inflammatory pain, CFA induced weight bearing deficit in
the rat, AAT-008 exhibited an analgesic effect in a dose dependent
Table 5
IV and PO pharmacokinetics of AAT-008 (4j).a
Species
Rat
Route
Cmax (ng/mL)
Tmax (h)
AUC0-inf (ng h/mL)
T1/2 (h)
CLtotal (mL/min/kg)
Vdss (L/kg)
F%
IV
PO
IV
PO
IV
PO
–
–
9620 1180
7090 3330
7490 1510
6080 1590
3390 1120
2460 879
4.59 1.05
3.71 0.39
8.43 2.47
7.95 2.88
14.5 5.1
21.9 8.1
1.75 0.24
–
2.30 0.50
–
5.31 1.60
–
0.593 0.104
–
1.21 0.33
–
4.70 0.87
–
–
994 244
–
1450 320
–
76.0 17.2
0.750 0.289
–
0.438 0.125
–
8.00 0.00
73.6 34.5
–
80.6 10.6
–
73.3 14.1
Dog
Monkey
Values are the Mean SD (N = 4).
a
Compound was administered at a dose of 1 mg/kg (rats, dogs, and monkeys) intravenously (IV) or orally (PO).
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067

Y. Okumura et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
7
Table 6
Comparative in vivo profiles of AAT-008 (4j) and grapiprant (1).
References
1. Narumiya S, Sugimoto Y, Ushikubi F. Physiol Rev. 1999;79:1193.
2. Kawahara K, Hohjoh H, Inazumi T, Tsuchiya S, Sugimoto Y. Biochim Biophys Acta.
2015;1851:414.
Assay
AAT-008 Grapiprant
(4j)
(1)
3. Ichikawa A, Sugimoto Y, Negishi MJ. Lipid Mediators Cell Signall. 1996;14:83.
4. McCoy JM, Wicks JR, Audoly LP. J Clin Invest. 2002;110:651.
5. St.-Jacques B, Ma W. Pain. 2013;154:313.
6. Mutoh M, Watanabe K, Kitamura T, et al. Cancer Res. 2002;62:28.
7. Fujino H, Regan JW. Trends Pharmacol Sci. 2003;24:335.
8. Yang L, Huang Y, Porta R, et al. Cancer Res. 2006;66:9665.
9. Kundu N, Ma X, Kochel T, et al. Breast Cancer Res Treat. 2014;143:19.
10. Majumder M, Xin X, Liu L, Girish GV, Lala PK. Cancer Sci. 2014;105:1142.
11. Lin C-R, Amaya F, Barrett L, et al. J Pharmacol Exp Ther. 2006;319:1096.
12. Nakao K, Murase A, Ohshiro H, et al. J Pharmacol Exp Ther. 2007;322:686.
13. Hippisley-Cox J, Coupland C. BMJ. 2005;330:1366.
14. Hippisley-Cox J, Coupland C, Logan R. BMJ. 2005;331:1310.
15. Zimecki M. Postepy Hig Med Dosw (Online). 2012;66:287.
16. Cua DJ, Sherlock J, Chen Y, et al. Nature. 2003;421:744.
17. Murphy CA, Langrish CL, Chen Y, et al. J Exp Med. 2003;198:1951.
18. Yen D, Cheung J, Scheerens H, et al. J Clin Invest. 2006;116:1310.
19. Yokoyama U, Iwatsubo K, Umemura M, Fujita T, Ishikawa Y. The prostanoid EP4
receptor and its signaling pathway. Pharmacol Rev. 2013;65:1010.
20. For more information, please see the AskAt Inc. <http://askat-inc.com/>.
21. For information on the development for dogs, please see the Aratana
Therapeutics, Inc. <http://www.aratana.com/therapeutics/osteoarthritis-pain/>.
22. All authors were employed by Pfizer at the time this work was carried out.
Intellectual properties regarding this work are all transferred to AskAt Inc.
23. ALogP (calculated partition coefficients, atom-based logP) and cLogD
(calculated distribution coefficients at pH 7.4) were determined using
Pipeline Pilot (http://accelrys.com/). The values of grapiprant (1) were
calculated as follows: MW 491.61, ALogP 4.55, and cLogD 4.71.
24. Patani GA, LaVoie EJ. Chem Rev. 1996;96:3147.
In vivo effects: acute and chronic inflammatory pain
Carrageenan-induced mechanical hyperalgesia in 1 (n = 3)
rats: MED (mg/kg, PO)
CFA-induced weight bearing deficit in rats: MED 1 (n = 3)
(mg/kg, PO)
30 (n = 2)
20 (n = 2)
manner with an MED of 1 mg/kg, PO (vs 20 mg/kg, PO for
grapiprant). AAT-008 exhibited good passive permeability and
good metabolic stability in HLM (T1/2 > 120 min) or hepatocytes
(T1/2 > 360 min). Inhibitory effects of AAT-008 (1 lM) on CYP1A,
CYP2C9, CYP2C19, CYP2D6, and CYP3A activities were determined
to be less than 5%.
In conclusion, a novel series of para-N-acylaminomethylbenzoic
acid-based selective prostaglandin EP4 receptor antagonists has
been identified, starting from N-acyl sulfonamide HTS hit 2, fol-
lowed by hit-to-lead optimization using in-house compound
library. Lead compound 4a was sequentially optimized by dividing
it into three parts. The most significant improvement was achieved
when a (S)-methyl substituent was introduced at the benzylic posi-
tion. Eventually, 4j (AAT-008) was identified as a development can-
didate, having excellent potency, selectivity for EP4, and metabolic
stability. Superior potency of AAT-008 to grapiprant was demon-
strated in different in vivo models. AAT-008 was stable in human
liver microsomes or hepatocytes and was predicted to be well
absorbed in humans. AAT-008 was selected as a candidate for pre-
clinical development and subsequent clinical studies.20 The pro-
jected efficacious oral dosing regimen of AAT-008 for OA pain is
1–20 mg QD.
25. Meanwell NA. J Med Chem. 2011;54:2529.
26. Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. Adv Drug Del Rev. 1997;23:3.
27. Lipinski CA. Drug Discovery Today. 2004;1:337.
28. Smith GF. Prog Med Chem. 2011;50:1.
29. Williams DP, Naisbitt DJ. Curr Opin Drug Discov Devel. 2002;5:104.
30. Blagg J. Structural alerts for toxicity. In: Abraham DJ, Rotella DP, eds. Burger’s
Medicinal Chemistry, Drug Discovery and Development. 7th ed. NJ: John Wiley &
Sons Inc.; 2010:301–334.
31. Human colon carcinoma cell line (Caco-2) permeability assay was performed at
constant pH conditions (apical pH of 6.5 and basolateral pH of 7.4).
32. Compound synthesis and characterization and assay protocols have been
published: Yamagishi T, Okumura Y, Nukui S, Nakao K. PCT Int Appl. 2005.
WO2005021508.
Acknowledgements
33. Waring MJ. Expert Opin Drug Discov. 2010;5:235.
The authors would like to acknowledge EP4 receptor antagonist
project teams at PGRD, Nagoya.
34. SNAr reaction using organic base is essential to afford the desired target
products with better yield instead of using inorganic base to avoid producing
byproducts, BEMP: 2-tert-Butylimino-2-diethylamino-1,3-dimethylperhydro-
1,3,2-diazaphosphorine (CAS No.: 98015-45-3).
35. Matsuoka A, Yamazaki N, Suzuki T, Hayashi M, Sofuni T. Mutat Res.
1992;272:223.
A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2017.01.
067.
36. Arturson P, Karlsson J. Biochem Biophys Res Commun. 1991;175:880.
Please cite this article in press as: Okumura Y., et al. Bioorg. Med. Chem. Lett. (2017), http://dx.doi.org/10.1016/j.bmcl.2017.01.067