Discovery of potent aryl-substituted 3-[(3-methylpyridine-2-carbonyl) amino]-2,4-dimethyl-benzoic acid EP4 antagonists with improved pharmacokinetic profile

Article abstract of DOI:10.1016/j.bmcl.2015.12.057

Two new series of EP4 antagonists containing a 3-methylaryl-2-carbonyl core have been identified. One series has a 3-substituted-phenyl core, while the other one incorporates a 3-substituted pyridine. Both series led to compounds with potent activity in functional and human whole blood (hWB) assays. In the pyridine series, compound 7a was found to be a highly potent and selective EP4 antagonist, with suitable rat and dog pharmacokinetic profiles.

Full text of DOI:10.1016/j.bmcl.2015.12.057

Accepted Manuscript
Discovery of potent aryl-substituted 3-[(3-methylpyridine-2-carbonyl) ami-
no]-2,4-dimethyl-benzoic acid EP4 antagonists with improved pharmacokinetic
profile
Maria-Jesus Blanco, Tatiana Vetman, Srinivasan Chandrasekhar, Matthew J.
Fisher, Anita Harvey, Mark Chambers, Chaohua Lin, Daniel Mudra, Jennifer
Oskins, Xu-Shan Wang, Xiao-Peng Yu, Alan M. Warshawsky
PII:
S0960-894X(15)30370-X
http://dx.doi.org/10.1016/j.bmcl.2015.12.057
BMCL 23418
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
16 November 2015
16 December 2015
17 December 2015
Please cite this article as: Blanco, M-J., Vetman, T., Chandrasekhar, S., Fisher, M.J., Harvey, A., Chambers, M.,
Lin, C., Mudra, D., Oskins, J., Wang, X-S., Yu, X-P., Warshawsky, A.M., Discovery of potent aryl-substituted 3-
[(3-methylpyridine-2-carbonyl) amino]-2,4-dimethyl-benzoic acid EP4 antagonists with improved pharmacokinetic
profile, Bioorganic & Medicinal Chemistry Letters (2015), doi: http://dx.doi.org/10.1016/j.bmcl.2015.12.057
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
Discovery of potent aryl-substituted 3-[(3-methylpyridine-2-
carbonyl) amino]-2,4-dimethyl-benzoic acid EP4 antagonists with
improved pharmacokinetic profile
Maria-Jesus Blanco,* Tatiana Vetman, Srinivasan Chandrasekhar, Matthew J. Fisher, Anita
Harvey, Mark Chambers, Chaohua Lin, Daniel Mudra, Jennifer Oskins, Xu-Shan Wang, Xiao-
Peng Yu, and Alan M. Warshawsky
6
2
3
7a-b
8a-b

Discovery of potent aryl-substituted 3-[(3-methylpyridine-2-
carbonyl) amino]-2,4-dimethyl-benzoic acid EP4 antagonists with
improved pharmacokinetic profile
Maria-Jesus Blanco,* Tatiana Vetman, Srinivasan Chandrasekhar, Matthew J. Fisher, Anita
Harvey, Mark Chambers, Chaohua Lin, Daniel Mudra, Jennifer Oskins, Xu-Shan Wang, Xiao-
Peng Yu, and Alan M. Warshawsky
Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN 46285
Abstract: Two new series of EP4 antagonists containing a 3-methylaryl-2-carbonyl core have been
identified. One series has a 3-substituted-phenyl core, while the other one incorporates a 3-substituted
pyridine. Both series led to compounds with potent activity in functional and human whole blood (hWB)
assays. In the pyridine series, compound 7a was found to be a highly potent and selective EP4 antagonist,
with suitable rat and dog pharmacokinetic profiles.
Keywords: EP4 receptor antagonist; structure-activity relationship (SAR); Prostaglandin E2; human whole blood
(hWB) assay; inflammation; pain
Prostaglandin E2 (PGE2)1 is an endogenous lipid belonging to the prostaglandin
superfamily. Prostaglandins, including thromboxanes and prostacyclins, are
enzymatically derived from fatty acids and have a variety of robust physiological effects
in the human body. PGE2, an important pro-inflammatory messenger, is formed by
metabolism of arachidonic acid and elicits its biological functions through four EP
subtype receptors EP1-4, all belonging to the G-protein-coupled receptor (GPCR)
superfamily. There are several pathological conditions wherein PGE2 plays a major
role, making EP receptors an appropriate pharmacological target. Because PGE2
predominantly serves as a signaling molecule in inflammation, most approaches to
modulate EP receptors have focused on inflammatory conditions, although EP receptor
modulators might also be beneficial in the treatment of other disorders like oncology2
and treatment of functional bladder disorders.3 The identification of highly selective EP1,
EP3 and EP4 antagonists, has led to pharmacological demonstration that EP4, not EP1
or EP3, is the primary receptor involved in joint inflammation and pain in rodent models
of rheumatoid and osteoarthritis. Since EP4 does not interfere directly with the
biosynthesis of any of the prostanoids, a selective EP4 antagonist might provide
analgesia without the potential side effects observed with NSAIDs and COX-2
inhibitors.4, 5
_____________________
* Corresponding author. Tel.: +1-317-433-3612; e-mail: blanco_maria@lilly.com

Growing interest from the pharmaceutical industry for developing safer
analgesics makes EP4 antagonism an attractive target and has led to the discovery of
several advanced compounds in the past few years. CJ-023,423 (1, Figure 1) was the
first EP4 antagonist evaluated in the clinic for the treatment of signs and symptoms of
osteoarthritis (OA). This compound achieved a positive proof-of-concept in chronic
inflammatory pain (OA pain) in a phase II study in 2007.6 More recently, Merck-Frosst
reported the discovery of a series of naphthalene and quinoline amides and
sulfonylureas that culminated in the discovery of MF-5927 (2, Figure 1). MF-592
demonstrated good functional potency (hEP4 IC50 = 3 nM) and potent inhibition of
PGE2 modulated TNFα release in LPS stimulated human whole blood (hWB) (IC50 = 78
nM). In addition, it showed an excellent oral pharmacokinetic (PK) profile and in vivo
efficacy in a rat chronic adjuvant-induced-arthritis model (ED50 = 0.1 mg/kg/day).
Figure 1. Examples of advanced EP4 antagonists. Clinical candidate CJ-023,423 (1) and MF-592 (2).
1, CJ-023,423
2, MF-592
We have previously reported the discovery of novel quinoline and naphthalene
analogs with potent EP4 antagonist activity.8 From this series compound 3, (Figure 2)
showed excellent hWB activity (IC50 = 6 nM) but suffered from a poor oral PK profile (rat
%F = 6). In contrast, compound 4 (Figure 2), displayed modest hWB activity (IC50 = 243
nM) but excellent oral rat bioavailability (100%). During the course of our research, we
explored replacements for the central domain of the molecules, to identify compounds
with high ligand efficiencies9 (LEAN values),8 potent hWB activities and acceptable
ADME profiles.

Figure 2. Previously reported EP4 antagonists series
Domain 1
Domain 3
Domain 2
3, R = CH2OH
4, R = Cl
X = N, CH
Toward this goal, we synthesized several substituted phenyl- and pyridine-
containing analogs of compounds 3 and 4. Replacement of the naphthalene core of 3
with a 2-methyl-benzene core gave 5a, which maintained the same level of functional
activity (hEP4 IC50 = 2.31 nM) as its bicyclic analog (Figure 3, Table 1). Compound 5a
also showed potent hWB activity (IC50 = 59.5 nM). Introduction of a 2-fluoro-benzene
core led to 5b with activity comparable to the 2-Me analog, (hEP4 IC50 = 4.49 nM and
hWB IC50 72.4 nM). Para-2,5-dimethyl-benzoic acid derivatives, 6a (R=Me) and 6b
(R=F) showed sub-nanomolar functional activity (0.28 and 0.78 nM, respectively) with
hWB IC50 values < 50 nM. While the functional activity is not corrected for human serum
protein binding, the hWB assay is a native tissue assay that represents relevant
physiological conditions.10 Therefore, the hWB assay was important for choosing
molecules to evaluate further.
Replacement of the naphthalene with a 3-methyl-pyridine core gave 7a, with high
potencies in the functional (hEP4 IC50 = 2.36 nM) and hWB (IC50 <40 nM) assays.
Introduction of a trifluoromethyl substituent offered the possibility to modulate electronic
and steric effects on the central ring,11 noting that the molecular volume of CF3
resembles an isopropyl group. Trifluoromethyl analog 7b, demonstrated activities
equivalent to methyl analog 7a, (7b, hEP4 IC50= 1.94 nM) with hWB IC50 = 59 nM.
The para-2,5-dimethyl-benzoic acid derivatives (8a and 8b) showed sub-
nanomolar functional activity (~3-fold improvement over 7a, 7b) that failed to translate to
increases in hWB activity (IC50 >50 nM).
Compound 9 combined the 3-methyl-pyridine central core with the 3-chloro-
benzene substitution in domain 3 as in compound 4. However, the potent functional
activity (hEP4 IC50=1.56 nM) for compound 9 did not translate to potent activity (>600
nM) in the hWB assay.

Overall, the modifications to the central core (domain 2) were favorable. Most of
the compounds showed potent hWB activity, more than 10-fold more potent than clinical
candidate CJ-023,423 (hWB IC50 = 1600 nM).8
With respect to physiochemical properties and ligand efficiency, compounds 5-9
had MW <450, cLogD<3, and the LEAN values were maintained or increased (0.27-
0.33) relative to quinoline/naphthalene analogs. Those attributes were targeted for
potential improvements in pharmacokinetic profiles.
Figure 3. Structures of 5-9 compounds
6
1
5
2
3
2
5a-b
7a-b
9
6a-b
8a-b
Table 1. In vitro functional activity and human whole blood (hWB) data of key analogs.
hEP4 functional
cAMP IC50 (nM)a
LEANb
values
Compd
R
MW
hWB IC50 (nM)a
cLog Dc
3
-
426
430
389
393
389
393
390
444
2.42 + 1.1 (5)
6.2 + 4.5 (3)
243.0 + 116 (6)
59.5 + 37.8 (3)
72.4 + 58.2 (16)
18.7 + 13.6 (6)
47.8 + 15.2 (5)
38.7 + 17.1 (9)
59.0 + 39 (3)
0.28
0.27
0.30
0.29
0.33
0.32
0.30
0.27
2.48
3.85
2.00
1.63
2.08
1.71
1.55
1.92
-
3.62 + 0.7 (3)
2.31 + 2.02 (2)
4.49 + 1.90 (4)
0.28 + 0.13 (5)
0.73 + 0.48 (6)
2.36 + 1.42 (5)
1.94 + 0.93 (4)
4
-Me
-F
5a
5b
6a
6b
7a
7b
-Me
-F
-Me
-CF3

-Me
-CF3
-Me
390
444
395
0.67 + 0.24 (3)
0.67 + 1.55 (3)
1.56 + 0.75 (2)
56.3 + 34.3 (10)
121.0 + 58 (5)
637+ 298 (3)
0.32
0.29
0.32
1.64
2.00
2.93
8a
8b
9
a. Results are expressed as the geometric mean
standard deviation; n = number of independent
determinations. The standard deviation is calculated by the delta method, being SDlogIC50 x geometric
mean x ln(10). For additional details on the assays please see ref 10.
b. LEAN values are based on functional activity and defined as -log(IC50)/number of non-hydrogen atoms.
c. cLog D at pH=7.4, calculated using Marvin and calculator plugin freeware (www.chemaxon.com,
ChemAxon Kft, Budapest, Hungary).
Data to assess solubility and rat pharmacokinetic profiles of key analogs is
presented in Table 2.
Table 2. ADME/PK assessment of key analogs
HTSA
solubilitya
(mg/mL)
Rat iv
clearanceb
(mL/min/kg)
po AUCc
(ng*h/mL)
%F b,c
rat
,
Compd
R
5a
5b
6a
6b
7a
7b
8a
8b
-Me
-F
0.783
0.709
0.031
0.774
0.242
0.029
0.014
0.023
15
30
375
994
7
11-54*
13
-Me
-F
71
88
91
105
12
-Me
-CF3
-Me
-CF3
2.2
7.0
7.0
54
8520
1540
4330
243
22
13
36
16
a- High Throughput Solubility Assay (HTSA), assessment of aqueous solubility at pH=6. Samples prepared in
DMSO were dried for 12 hours. The powder or film was re-dissolved in the solvent at pH 6 and DMSO control
at 2mM target concentration. Samples are stirred for 20 hours and filtered through a 0.7 µm GF filter. The
filtrate was analyzed by HPLC assay for concentration against DMSO standard curve.
b- Dose of 1 mg/kg iv.
c- Dose of 5 mg/kg po.
*Significant variability observed
We noted that small changes in the molecular structure had significant and
unpredictable effects on solubility, clearance and plasma exposure. The compounds
with a phenyl moiety in domain 2 (5a, 5b, 6a, 6b), showed good solubility in general,
although they suffered from poor oral bioavailability and high clearance. The analogs
with a pyridine group in domain 2 (7a, 7b, 8a, 8b) tended to have low solubility (except

7a), but showed lower clearance (except 8b) and acceptable oral bioavailability. Based
on the combination of solubility and rat PK profile, compound 7a was selected for further
characterization.
Compound 7a showed an acceptable pharmacokinetic profile in a dog study with
increased oral bioavailability (48%).
Table 3. Pharmacokinetic profiling for compound 7a 6
iv/po dose
(mg/kg)
po AUC
(ng*h/mL)
po Cmax
(ng/mL)
po T1/2
(h)
iv CL
(mL/min/kg)
Vdss
(L/Kg)
Species
F (%)
rat
dog
1/5
1/5
8520
10900
1640
3820
2.1
1.1
2.2
3.7
0.16
0.97
22
48
Compound 7a was tested in binding assays to assess selectivity against the other EP
receptor subtypes (Table 3). No detectable binding was observed with either EP1, EP2,
or EP3. In addition, compound 7a was found to have no in vitro inhibitory activity against
several CYP enzymes up to 10 µM. Therefore the risk of drug-drug interactions with 7a
was expected to be low. Furthermore, 7a was evaluated at Cerep for broader selectivity
against a panel of 13 receptors, four ion channels, one transporter , and one enzyme.
No significant activity on any of the targets at 10 µM was noted.12 Finally, 7a had no
activity up to 100 µM against hERG in a [3H]-astemizole binding assay.
Table 3. Selectivity data for 7a.
Assay
Resultsa
Ki > 17.5 µM;
hEP1 binding
IC50 > 25 µM (2)
Ki > 18.9 µM;
hEP2 binding
hEP3 binding
IC50>25 µM (5)
Ki > 14 µM;
IC50 > 25 µM (5)
IC50 > 10 µM for all
CYP1A2, CYP2B6, CYP2C19, CYP2C8,
CYP2C9, CYP2D6, CYP3A4 inhibition
aExpressed as geometric mean geometric standard devation, with number of replicates in parentheses.
To establish pain relief efficacy in rats, compound 7a was tested in the
monoiodoacetic acid (MIA) model of joint pain. Functional in vitro activity showed no

significant differences between human and rat EP4 for compound 7a. The injection of
MIA into the knee joint of rats produces an acute inflammatory insult which develops
into chronic degeneration of the tissues in the injected joint and pain. This pain can be
measured via differential weight bearing of the hind legs using an incapacitance tester.13
Efficacy is assessed as by the ability of a test compound to normalize weight distribution
indicating reduction of joint pain. The MIA model has been extensively described in the
literature14 and has been used to demonstrate pain reversal for a variety of mechanisms
with compounds showing efficacy at plasma exposures comparable to clinically effective
human exposures.15
Figure 4 showed the results for the experiment where rats were injected intra-
articularly with 0.3 mg MIA in 50 µl saline into the right knee with 50 µl saline injected
into the left knee on day 0. Twelve days later, the rats were randomized and dosed
orally with either vehicle, 0.3, 1, 3 or 10 mg/kg compound 7a, or the positive control
non-steroidal anti-inflammatory drug (NSAID) diclofenac at 5 mg/kg. Efficacy was
measured using incapacitance testing 30 minutes post dose. Compound 7a dose
dependently inhibited differential weight bearing versus vehicle at 3 and 10 mg/kg as did
the NSAID diclofenac.
As compound 7a inhibits PGE2 actions at the EP4 receptor it would also be expected
to inhibit inflammation. For assessing efficacy versus inflammation the rat adjuvant
induced arthritis (AIA) model was used. This model has been well-described in the
literature16 and has been used to demonstrate efficacy versus inflammation for a variety
of mechanisms and compounds including those that target prostaglandin pathways
such as NSAIDs. Efficacy is measured as the ability of a test compound to inhibit paw
swelling.
Figure 5 showed the results for the experiment where rats had their paw widths
measured and were then inoculated with 0.25 mg of adjuvant (M. Tuberculosis H37 RA)
in 100 µl of mineral oil intradermal at the base of the tail to induce adjuvant disease (day
0). Eleven days post inoculation paw width of both right and left paws was measured
again and the percent change from the day 0 reading calculated. This measurement
along with the body weight of the rats was used to randomize the animals and the rats
were dosed with vehicle and compound 7a at doses of 1, 3 and 10 mg/kg. The
nonsteroidal anti-inflammatory drug (NSAID) diclofenac was included at a dose of 10
mg/kg as a positive control. On day 15 post adjuvant injection the rats were euthanized
without further dosing and paw width measured again. The percent increase in mean
paw thickness from day 11 to day 15 was then calculated as a measure of paw swelling
which is indicative of the amount of inflammation in the paws. Compound 7a
significantly inhibited increases in paw inflammation (swelling) compared to vehicle at
doses of 10 and 30 mg/kg as did the NSAID diclofenac.

Figure 4:
Dose response study with compound 7a in the MIA model with pain measured one hour post dose
The data are presented as mean SEM where group size is n = 5. Statistical comparison to
vehicle: Dunnett’s test (*p<0.05). Vehicle is 10% Acacia in water plus 0.05% antifoam.

Figure 5:
Dose response study with compound 7a in the AIA model to assess inhibition of increase in paw
swelling . The data are presented as mean SEM where group size is n = 8. Statistical
comparison to vehicle: Dunnett’s test (*p<0.05). Vehicle is 10% Acacia in water plus 0.05%
antifoam.
The compounds described in Table 1 were synthesized by the sequence shown
in Scheme 1.10 T3P (n-propanephosphonic acid anhydride)17 mediated coupling of 3-
chloro substituted pyridine (10) or 3-bromo substituted benzene (11) carboxylic acids
with corresponding functionalized aniline derivative led to amides 12 or 13 in good
yields. Suzuki coupling of 3-(hydroxymethyl)-phenyl-boronic acid with 12 or 13 led to
triaryl derivatives 14 and 15. Subsequent ester hydrolysis under basic conditions
provided the final carboxylic acids 5-8.
Scheme 1. General synthetic route to EP4 antagonist compounds.
(a)
(b)
, X = N; , X = C
15
14
, X = N; , X = C
13
12
R1 = CH3, F, or CF3
R1 = CH3, F, or CF3
R3 = CH3 or CH2CH3
, X = N; , X = C
11
10
R1 = CH3, F, or CF3
R2 = Cl or Br
R3 = CH3 or CH2CH3
R2 = Cl or Br
(c)
, X=C, R1=Me; , X=C, R1=F
5b
, X=C, R1=Me; , X=C, R1=F
6b
5a
7a
6a
8a
, X=N, R1=Me; , X=N, R1=CF
, X=N, R1=Me; , X=N, R1=CF
7b
8b
3
3
Reagents and conditions: (a) T3P, CH2Cl2, DIEA, room temperature, 24 h, 28-93%; (b) 1,4-dioxane, [3-
(hydroxymethyl)phenyl]boronic acid, K2CO3, PdCl2(dppf) CH2Cl2, 110 °C, 3h, 39-74%; (c) 1N NaOH, 1:1
THF/MeOH, 30-97%.
In conclusion, a novel series of potent EP4 antagonists have been identified that
demonstrated potent functional activity and robust target engagement in an ex vivo

LPS-stimulated human whole blood (hWB) assay. Furthermore, pharmacokinetic
assessment identified compound 7a with lower rat clearance and efficient oral
exposures. Compound 7a was highly selective over EP1, EP2 and EP3. In vivo efficacy
studies showed comparable effects to NSAID diclofenac in the pain measured MIA
assay and AIA assay to assess inflammation. Further optimization of compounds from
this scaffold and in vivo pharmacology studies will be reported in due course.
Acknowledgements
We would like to thank Jaclyn Barrett for analytical support and solubility data.
Supplementary data
Supplementary data (CEREP panel data) associated with this manuscript can be found,
in the online version, at ……..
References
1. Kawahara, K.; Hohjoh, H.; Inazumi, T.; Tsuchiya, S.; Sugimoto, Y. Biochim Biophys Acta 2015, 414,
1851.
2. (a) Misra, S.; Sharma, K. Curr Drug Targets 2014, 15, 347; (b) Okunishi, K.; DeGraaf, A.J.; Zaslona, Z.;
Peters-Golden, M. FASEB J. 2014, 28, 56.
3. Rahnama'i M.S., van Kerrebroeck P.E., de Wachter S.G., van Koeveringe G.A. Nat Rev Urol. 2012, 9,
283.
4. Hippisley-Cox, J.; Coupland, C. BMJ, 2005, 330, 1366.
5. Hippisley-Cox, J.; Coupland, C.; Logan, R. BMJ, 2005, 331, 1310.
6. Borriello, M.; Stasi, L.P. Pharm Pat Anal. 2013, 2, 387.
7. Burch, J.D.; Farand, J.; Colucci, J.; Sturino, C.; Ducharme, Y.; Friesen, R.W.; Levesque, J.F.; Gagne, S.;
Wrona, M.; Therien, A.G.; Mathieu, M.C.; Denis, D. ; Vigneault, E.; Xu, D.; Clark, P.; Rowland, S.; Han, Y.
Bioorg Med Chem Lett 2011, 21, 1041.
8. (a) Blanco M. J.; Vetman T.; Chandrasekhar S.; Fisher M.J.; Harvey A.; Mudra D.; Wang X.S.; Yu X.P.;
Schiffler M.A.; Warshawsky A.M. Bioorg. Med. Chem. Lett. 2016, 26, 105. (b) LEAN is a metric for ligand
efficiency and is defined as −log(IC50)/number of non-hydrogen atoms.
9. Ligand efficiency is a key parameter for drug discovery. Hopkins, A.L.; Keserue, G.M.; Leeson, P.D.;
Rees, D.C.; Reynolds, C.H. Nat. Rev. Drug Discovery 2014, 13, 105.
10. Synthesis protocols, compound characterization, and assay protocols have been published. Blanco-
Pillado, M.-J.; Manninen, P. R.; Schiffler, M. A.; Vetman, T. N.; Warshawsky, A. M.; York, J. S.
WO2015094912.
11. Hagmann, W.K. J. Med. Chem. 2008, 51, 4359.
12. For more details, please see supplemental data.
13. Benschop, R.J.; Collins, E.C.; Darling, R.J.; Allan, B.W.; Leung, D.; Conner, E.M.; Nelson, J.; Gaynor, B.;
Xu, J.; Wang, X.F.; Lynch, R.A.; Li, B.; McCarty, D.; Nisenbaum, E.S.; Oskins, J.L.; Lin, C.; Johnson, K.W.;
Chambers, M.G. Osteoarthritis Cartilage. 2014, 22, 578.
14. Combe, R.; Bramwell, S.; Field, M.J. Neurosci. Lett. 2004, 370, 236.
15. Goodman s The Pharmacological Basis of Therapeutics; Hardman, L. G., Limbird, L. E., Eds., 10th ed.;
McGraw-Hill: New York, 2001. Appendix Table A-II-1.

16. Hegen, M.; Keith, J.C.; Collins, M.; Nickerson-Nutter, C.L. Ann Rheum Dis 2008, 67, 1505.
17. Dunetz, J.R.; Xiang, Y.; Baldwin, A.; Ringling, J. Org Lett. 2011, 13, 5048.