Discovery and Characterization of 2-Acylaminoimidazole Microsomal Prostaglandin e Synthase-1 Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.5b01249

As part of a program aimed at the discovery of antinociceptive therapy for inflammatory conditions, a screening hit was found to inhibit microsomal prostaglandin E synthase-1 (mPGES-1) with an IC₅₀ of 17.4 μM. Structural information was used to improve enzyme potency by over 1000-fold. Addition of an appropriate substituent alleviated time-dependent cytochrome P450 3A4 (CYP3A4) inhibition. Further structure-activity relationship (SAR) studies led to 8, which had desirable potency (IC₅₀ = 12 nM in an ex vivo human whole blood (HWB) assay) and absorption, distribution, metabolism, and excretion (ADME) properties. Studies on the formulation of 8 identified 8·H₃PO₄ as suitable for clinical development. Omission of a lipophilic portion of the compound led to 26, a readily orally bioavailable inhibitor with potency in HWB comparable to celecoxib. Furthermore, 26 was selective for mPGES-1 inhibition versus other mechanisms in the prostanoid pathway. These factors led to the selection of 26 as a second clinical candidate.

Full text of DOI:10.1021/acs.jmedchem.5b01249

Subscriber access provided by UNIV OF NEBRASKA - LINCOLN
Article
Discovery and characterization of 2-acylaminoimidazole mPGES-1 inhibitors
Matthew Allen Schiffler, Stephen Antonysamy, Shobha N Bhattachar, Kristina M Campanale,
Srinivasan Chandrasekhar, Bradley Condon, Prashant V Desai, Matthew J. Fisher, Christopher
Groshong, Anita K Harvey, Michael J Hickey, Norman E Hughes, Scott A Jones, Euibong Jemes
Kim, Steven L. Kuklish, John G. Luz, Bryan H Norman, Richard E Rathmell, John R Rizzo,
Thomas W Seng, Stefan J Thibodeaux, Timothy A Woods, Jeremy S York, and Xiao-Peng Yu
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01249 • Publication Date (Web): 10 Dec 2015
Downloaded from http://pubs.acs.org on December 12, 2015
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 free 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 accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 37
Journal of Medicinal Chemistry
1
2
3
4
5
6
Discovery and characterization of 2-acylaminoimidazole mPGES-1 inhibitors
7
8
9
Matthew A. Schiffler,a,* Stephen Antonysamy,b Shobha N. Bhattachar,a Kristina M. Campanale,a Sriniꢀ
vasan Chandrasekhar,a,† Bradley Condon,b Prashant V. Desai,a Matthew J. Fisher,a Christopher Groꢀ
shong,b Anita Harvey,a,† Michael J. Hickey,b Norman E. Hughes,a Scott A. Jones,a Euibong J. Kim,a,‡
Steven L. Kuklish,a John G. Luz,b Bryan H. Norman,a Richard E. Rathmell,c John R. Rizzo,a Thomas W.
Seng,a Stefan J. Thibodeaux,a,§ Timothy A. Woods,a Jeremy S. York,a and XiaoꢀPeng Yua
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aLilly Research Laboratories, A Division of Eli Lilly and Company, Indianapolis, Indiana 46285, USA
bEli Lilly Biotechnology Center, San Diego, CA 92121, USA
cLilly Research Laboratories, A Division of Eli Lilly and Company, Windlesham, Surrey GU20 6PH,
United Kingdom
KEYWORDS: mPGESꢀ1, aminoimidazole, clinical candidate
ABSTRACT: As part of a program aimed at the discovery of antiꢀnociceptive therapy for inflammatory
conditions, a screening hit was found to inhibit mPGESꢀ1 with an IC50 of 17.4 µM. Structural inforꢀ
mation was used to improve enzyme potency by over 1000ꢀfold. Addition of an appropriate substituent
alleviated timeꢀdependent CYP3A4 inhibition. Further SAR studies led to 8, which had desirable poꢀ
tency (IC50 = 12 nM in an ex vivo human whole blood (HWB) assay) and ADME properties. Studies
on the formulation of 8 identified 8•H3PO4 as suitable for clinical development. Omission of a lipoꢀ
philic portion of the compound led to 26, a readily orally bioavailable inhibitor with potency in HWB
comparable to celecoxib. Furthermore, 26 was selective for mPGESꢀ1 inhibition versus other mechaꢀ
nisms in the prostanoid pathway. These factors led to the selection of 26 as a second clinical candidate.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 37
INTRODUCTION
1
2
3
4
5
6
7
8
9
For millennia, it has been known that willow bark possesses analgesic properties.1,2 However, only in
the 19th century was the active component, salicylic acid, identified,3 and only much later did related
research reveal that reduction in systemic PGE2 levels correlated with an analgesic effect.4–6 Therapies
indirectly modulating PGE2 through inhibition of COX enzymes include such wellꢀknown and widelyꢀ
available drugs as aspirin,7 naproxen, and ibuprofen.8 These drugs nonꢀselectively inhibit both COXꢀ1,
a constitutive enzyme; and COXꢀ2, which is induced during the inflammatory response.9 In recent
years, selective COXꢀ2 inhibitors have entered the market as treatments for pain.10,11
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
However, in addition to suppressing PGE2 production, COXꢀ2 inhibition disrupts the downstream balꢀ
ance of TXA2 and PGI2 (Figure 1), which is thought to lead to an increased risk of thrombotic events.12
Such observations have, in fact, led to the withdrawal of COXꢀ2 inhibitors such as rofecoxib and valꢀ
decoxib from the market.13 Therefore, inhibition of targets further downstream in the prostaglandin
pathway might be desirable.14 One such approach would entail direct inhibition of a PGE2 synthase.
Figure 1. Prostaglandin synthesis pathway.
Microsomal prostaglandin E synthaseꢀ1 (mPGESꢀ1) is a 17ꢀkDa membraneꢀbound, glutathioneꢀ
dependent isomerase which, like COXꢀ2, is induced during the inflammatory response.15 It is one of
three enzymes known to convert PGH2 to PGE2; the other two (mPGESꢀ2 and cPGES) are constitutively
ACS Paragon Plus Environment

Page 3 of 37
Journal of Medicinal Chemistry
expressed enzymes,16 and evidence has emerged that mPGESꢀ2 does not in fact produce PGE2 in vivo.17
Therefore, it has been hypothesized that selective inhibition of mPGESꢀ1 could alleviate the nociception
that accompanies inflammation while limiting the disruption of normal PGE2 synthesis.18 Indeed, much
effort has been made to discover mPGESꢀ1 inhibitors.16–23 Chart 1 illustrates some examples of such
compounds. Imidazophenanthrolines such as MFꢀ63 (1),20 aryl indoles such as 2,21 and oxicams such as
PFꢀ9184 (3)22 have each been shown to inhibit mPGESꢀ1. In addition, an mPGESꢀ1 inhibitor known as
GRCꢀ27864 has entered clinical trials for inflammatory pain.24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For the above reasons, we set out to identify an inhibitor of mPGESꢀ1. We tested potency in a microꢀ
somal prep enzyme inhibition assay, and we measured PGE2 and PGI2 by EIA in hILꢀ1βꢀstimulated
A549 cells to further evaluate potency and selectivity against other enzymes in the prostaglandin pathꢀ
way.25 It has been shown for COXꢀ2 inhibitors that the IC80 in an ex vivo LPSꢀstimulated human whole
blood (HWB) assay (determined by measuring inhibition of PGE2 response) correlates with efficacious
in vivo therapeutic concentrations.26 Therefore, we elected to evaluate our compounds by an analogous
assay, using PGE2 as a biomarker. Additionally in this assay, we measured TXB2 and PGF2α as selecꢀ
tivity readouts.27
Celecoxib has been reported28 to have a HWB IC50 of 540 ± 70 nM.29 In humans, a 200 mg oral dose of
celecoxib results in a plasma drug concentration exceeding this IC50 value for 8–12 hours.30 Our aim
was to find a selective mPGESꢀ1 inhibitor amenable to a dosing regimen that would produce a similar or
superior relationship of exposure to HWB potency.
Chart 1. Examples of known mPGES-1 Inhibitors.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
In a screen of compounds from the Lilly collection, 4 (Chart 2) was identified in the enzyme assay as
an mPGESꢀ1 inhibitor (IC50 = 17.4 µM ×/÷ 1.25 (n = 2)).31 The xꢀray coꢀcrystal structure of 4 with
mPGESꢀ132 revealed a hydrogen bond between the benzamide nitrogen and His53, and showed that the
C(5)ꢀC(6) region of the A ring projected toward a cleft in the enzyme surface (Figure 2).
Chart 2. Initial screening hit.
ACS Paragon Plus Environment

Page 5 of 37
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Xꢀray coꢀcrystal structures of mPGESꢀ1 with 4 (A) and 5 (B). Ligands, key side chains, and
glutathione cofactors are highlighted with thick bonds. Some atoms omitted for clarity. Hꢀbond disꢀ
tances given in Å.
Table 1. SAR of mPGES-1 Inhibitors 5–10 compared to reference compounds 1 and 2.
X
Y
R1 R2
mPGESꢀ1 A549
comp
HWB
HWB
Microsoꢀ Microsoꢀ clog
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 37
a
d
IC50
PGE2
PGE2
PGE2
mal
mal
Pc
1
2
3
4
5
6
7
8
a
a
a
IC50
IC50
IC80
CYP3A4 CYP3A4
inh. @ 10 TDI @ 10
µM (%) µM (%)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5.3 nM
76 nM
676 nM 3.65 µM
n/a n/a n/a n/a ×/÷ 1.51 ×/÷ 2.16 ×/÷ 1.96 ×/÷ 2.23
32
46
44
22
0
NT
NT
75
5.58
10.0
4.38
3.32
3.45
1
2
5
6
7
(239)
(234)
(61)
NT
(52)
NT
19.5 nM 1230 nM
n/a n/a n/a n/a ×/÷ 1.84 ×/÷ 1.21
(318)
(12)
19 nM
338 nM 994 nM
Cl CH H iꢀPr <3 nM (2) ×/÷ 1.59 ×/÷ 1.65 ×/÷ 1.57
(2)
(6)
(6)
37 nM
29 nM
146 nM
Cl
N
H
iꢀPr <3 nM (3) ×/÷ 1.04 ×/÷ 2.06 ×/÷ 2.46
72
(2)
(5)
(5)
39 nM
91 nM
480 nM
Cl N CH3 iꢀPr <3 nM (5) ×/÷ 1.43 ×/÷ 1.58 ×/÷ 2.35
27
(2)
(5)
(5)
0.9 nM
12 nM
12 nM
69 nM ×/÷
1.72 (11)
CF3 N CH3 iꢀPr ×/÷ 1.80 ×/÷ 1.55 ×/÷ 1.94
0
0
23
15
3.72
3.18
8
9
(10)b
(4)
(11)
CF3 N CH3 Et
10.3 nM
98 nM
374 nM 991 nM
ACS Paragon Plus Environment

Page 7 of 37
Journal of Medicinal Chemistry
×/÷ 1.53 ×/÷ 1.08 ×/÷ 1.49 ×/÷ 1.52
1
2
3
(5)
(2)
(3)
(3)
4
5
6
7
8
74 nM
CF3 N CH3 CH3 ×/÷ 1.17
(2)
720 nM
(1)
NT
NT
19
8
2.48
10
9
10
11
12
13
14
15
a Values are reported as geometric means with geometric standard deviations. Number of replicates is
in parentheses.
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
b Twelve values of <3 nM were not factored into this calculation.
c Calculated using Marvin and calculator plugin freeware (www.chemaxon.com, ChemAxon Kft, Buꢀ
dapest, Hungary).
Subsequent SAR studies revealed that an imidazole was a preferred replacement for the thiazole. We
then sought to leverage the C(5)ꢀC(6) region to gain additional potency. Incorporation of a C(5) substitꢀ
uent led to 5 (Table 1), which contained a pendant amide. It was notable that in the xꢀray structure of 5,
Gln36 and Arg52 had folded toward the ligand to participate in hydrogenꢀbonding interactions that were
absent in 4.
Despite the strong enzyme potency of 5, the IC50 shifted by at least two orders of magnitude between
the enzyme and the HWB assay. We supposed that decreasing the lipophilicity would attenuate this
shift. Indeed, substitution with a nitrogen atom at an appropriate location in the A ring (6) lessened the
enzymeꢀtoꢀHWB potency shift by approximately an order of magnitude.
Compounds 5 and 6 had sufficient activity in HWB to be of further interest. In contrast to celecoxib,
these analogues suppressed neither PGI2 formation in the A549 assay nor TXB2 or PGF2α production in
HWB.33 Additionally, 8 was at least 10,000ꢀfold selective against COXꢀ1 and COXꢀ2 (no inhibition of
either enzyme observed at 10 µM).34 Taken together, these biochemical outcomes suggested selective
inhibition of mPGESꢀ1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 37
While neither 5 nor 6 exhibited >50% inhibition at 10 ꢁM in in vitro microsomal CYP inhibition asꢀ
says, they did display in vitro TDI of CYP3A4 at 10 µM. This indicated the potential for DDIs over
prolonged dosing, a risk we wished to mitigate.35 It was found that adding a substituent to the imidazole
would lower the potential for TDI (see analogues 7–10.)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Although 7 seemed suitable in many respects, we embarked on further studies aimed at maximizing
the activity in HWB. Ultimately, replacement of the distal chlorine with a trifluoromethyl group to furꢀ
nish 8 improved the HWB potency by nearly an order of magnitude. Lower homologues of the isobuꢀ
tyramide (9, 10) were less potent. Therefore, our interest turned to additional studies of the potency, seꢀ
lectivity, and ADME properties of 8.
Against hERG, 8 had a measurable IC50 and Ki (82.6 µM and 41.3 µM, respectively), but the HWB
IC80 of 69 nM meant that the therapeutic plasma concentration of 8 was likely to be such that a suffiꢀ
cient margin should exist between hERG activity and mPGESꢀ1 activity to afford low risk of QT proꢀ
longation.36 Upon testing against a CEREP panel of receptors and ion channels, 8 did not display any
activity which would be deemed to present an undue risk.37
Compound 8 was active in dog whole blood (DWB IC50 = 139 nM ×/÷ 1.76 and IC80 = 703 nM ×/÷
2.37 (n = 16)). However, in a rat mPGESꢀ1 enzyme assay, 8 had an IC50 of 33.9 µM ×/÷ 1.94 (n = 3).37
This drastic species selectivity was similar to known observations with other classes of mPGESꢀ1 inꢀ
hibitors.38
ACS Paragon Plus Environment

Page 9 of 37
Journal of Medicinal Chemistry
Figure 3. PK of 8 in dog after 2 mg/kg po dosing (solid diamonds, n = 3) and 0.5 mg/kg iv dosing (holꢀ
low diamonds, n = 3). Diamonds indicate geometric means, and error bars represent geometric standard
deviations of measured values.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Timeꢀdependent exposure profile of 8 and 8•H3PO4 in a dynamic dissolution test in artificial
stomach (A) and artificial duodenum (B).
Inhibitor 8 was subjected to a dog PK experiment (Figure 3). The bioavailability was found to be 37%,
and at an oral dose of 2 mg/kg, a plasma concentration exceeding 200 nM was sustained for 24 h. This,
taken with the known correlation between ex vivo IC80 and efficacious plasma concentration,25 gave us
preꢀclinical evidence even in the absence of a per se in vivo efficacy study that a clinically relevant exꢀ
posure could be achieved at an acceptably low dose.39
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 37
However, the slow absorption and low FasSIF solubility of 8 (0.007 mg/mL) prompted us to evaluate
enablement options for the compound in order to facilitate dose escalation studies to assess safety withꢀ
out reaching a barrier of doseꢀlimited exposure due to dissolutionꢀlimited absorption. While measured
thermodynamic solubility in FaSSIF is a valuable initial indicator of the oral absorption potential of
compounds, it can sometimes be misleading as it is a result of extended hours of mixing in that medium
and not necessarily representative of the in vivo gastrointestinal environment. For basic compounds that
have a conjugate acid pKa of approximately 4.5 or greater, it is sometimes possible to enhance dissoluꢀ
tion rate through salt formation, and then briefly sustain metastable supersaturation at the site of absorpꢀ
tion.40 Accordingly, depending on the extent and duration of supersaturation, absorption can be signifiꢀ
cantly improved.41
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
This approach works best for compounds that do not rapidly crystallize and/or precipitate under bioꢀ
relevant conditions. From a dynamic dissolution test that mimicked the transfer of orally dosed comꢀ
pound from the stomach to the duodenum,42 it was evident that the phosphate salt of 8 had a tendency to
dissolve faster than the free base and remain amorphous for an extended period of time (Figure 4). Inꢀ
deed, in vivo PK studies in beagle dog after a 60 mg/kg PO dose showed that 8•H3PO4 was more rapidꢀ
ly absorbed than the free base, resulting in higher plasma concentrations and reduced variability at every
time point through 24 h (Figure 5). Furthermore, this salt form had acceptable solid state properties.
The melting point (190.4 °C), hygroscopicity (<0.5% weight gain when subjected to RH values up to
95%), and physical stability (stable for 2 weeks at 70 °C/75% RH and for 4 weeks at 40 °C/75% RH) all
were suitable for development. Therefore, 8•H3PO4 was selected as a clinical candidate.
ACS Paragon Plus Environment

Page 11 of 37
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Figure 5. PK of 8 (solid diamonds, n = 6) and 8•H3PO4 (hollow diamonds, n = 6) in dog after 60
mg/kg po dosing. Diamonds indicate geometric means, and error bars represent geometric standard deꢀ
viations of measured values.
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The synthesis of 8•H3PO4 was achieved as shown in Scheme 1. Aminoimidazole 12 was prepared by
bromination of ketone 13 to give 14,43 which was cyclized with 2ꢀaminopyrimidine to furnish 15. This
was then subjected to the action of hydroxylamine,44 which cleaved the pyrimidine to reveal 12. The
core of 16 was constructed by a modified Hantzsch pyridine synthesis. Condensation of 1745 with triꢀ
ethylorthoformate provided 18, which was warmed in the presence of 3ꢀdimethylaminopropꢀ2ꢀ
enenitrile46 to give pyridine 19 after acidꢀpromoted elimination of dimethylamine. Reduction of the niꢀ
trile, followed by acylation of the resulting amine and in situ saponification of the ester, gave 16.
a
Scheme 1. Synthesis of 8•H3PO4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (a) Br2, AcOH, 92%; (b) 2ꢀaminopyrimidine, NaHCO3, iꢀPrOH, 80 °C,
22%; (c) NH2OH, EtOH, 82 °C, 95%; (d) (EtO)3CH, Ac2O, 90–100 °C, 86%; (e) (CH3)2NCH=CHCN,
60–65 °C, then NH4OAc, 60–65 °C, 85%; (f) H2, Pd/C, Boc2O, EtOH, Et3N, then HCl, 1,4ꢀdioxane,
MTBE, 66%; (g) iꢀPrCOCl, Et3N, toluene, then LiOH(aq), 70%; (h) T3P, NMM, EtOAc, 65–75 °C,
64%; (i) H3PO4, CH3CN, H2O, 79%.
To complete the synthesis, 12 and 16 were coupled using T3P to furnish 8. Generally with unprotected
2ꢀaminoimidazoles, we observed a kinetic preference for acylation of either one of the ring nitrogens
versus the exocyclic nitrogen, so heating was required to induce isomerization to the N(2)ꢀacylated
product. Lastly, 8 was treated with H3PO4 to give the phosphate salt.
While 8•H3PO4 was suitable for clinical development, we sought a second molecule with improved
physical properties.47 We hypothesized that omitting the largely exposed lipophilic ring of 8 would imꢀ
ACS Paragon Plus Environment

Page 13 of 37
Journal of Medicinal Chemistry
prove solubility, bioavailability, and/or absorption while maintaining the critical hydrogen bonding inꢀ
teractions with the enzyme (Figure 6A). Therefore, we prepared 20 (Table 2). This analogue had only
singleꢀdigit micromolar potency in our enzyme assay, but it efficiently translated this activity to the ex
vivo HWB PGE2 assay. Furthermore, 20 did not demonstrate significant activity in a CYP3A4 TDI asꢀ
say (20% inhibition after 30 min at 10 ꢁM). Although 20 contained an unsubstituted imidazole, a
known risk factor for CYP inhibition,48 we reasoned that the lower molecular weight and lipophilicity of
20 vis-à-vis analogues 5–10 likely limited the affinity of the former for CYP enzymes.49
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Xꢀray coꢀcrystal structure of mPGESꢀ1 with 8. Ligand, key side chains, and glutathione coꢀ
factor highlighted with thick bonds. Some atoms omitted for clarity. Hꢀbond distances given in Å.
In rats, 20 was found to be orally bioavailable (F = 100%) and rapidly absorbed (tmax,po = 15 min) with
low clearance (CLiv = 26.1 mL/(min*kg)) and Vdss (2.27 L/kg). The efficient translation of enzyme acꢀ
tivity to HWB, combined with the PK properties of 20, encouraged us to pursue an SAR study of these
smaller analogues (Table 1).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 37
Increasing the bulk of the amide (21, 22) resulted in a slight potency improvement. In a study of substiꢀ
tution at the C(2) position (23–26), halogen atoms proved to be optimal. On the 2ꢀchloro system, furꢀ
ther increasing the bulk of the amide provided no potency advantage (27). Scaling the benzylic substitꢀ
uent back to an isobutyramide (28) reduced potency in the A549 IC50 and HWB IC80 assays, suggesting
along with the comparison of 20 and 22 that a pivalamide was optimal. Substitution at the C(4) position
tended to decrease potency (29, 30).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. SAR of Unsubstituted 2-Aminoimidazole mPGES-1 Inhibitors.
comp
d
A549 PGE2
HWB PGE2
clo
a
a
X
Y
H
R
mPGESꢀ1 IC50
HWB PGE2 IC50
gPb
a
a
IC50
IC80
CF2
H
2.46 µM ×/÷
1.13 (3)
3.15 ꢁM ×/÷
1.60 (2)
4.61 µM ×/÷ 1.46
(5)
31.0 ꢁM ×/÷
5.19 (5)
2.2
1
i-Pr
20
21
(S)ꢀ
sꢀ
CF2
H
1.12 µM ×/÷
1.05 (2)
8.86 ꢁM ×/÷
2.18 (2)
10.5 µM ×/÷ 1.08
(2)
21.1 ꢁM ×/÷
1.08 (2)
2.6
5
H
Bu
CF2
H
tꢀ
0.676 µM ×/÷
1.18 (2)
3.25 ꢁM ×/÷
1.43 (2)
1.43 µM ×/÷ 2.08
(2)
33.7 ꢁM ×/÷
1.16 (2)
2.7
6
H
H
H
22
23
24
Bu
tꢀ
1.08 µM ×/÷
1.07 (2)
3.23 ꢁM ×/÷
1.50 (2)
2.14 µM ×/÷ 1.64
(3)
10.5 ꢁM ×/÷
1.93 (3)
3.2
5
CF3
Bu
tꢀ
3.11 µM ×/÷
1.35 (2)
4.62 ꢁM ×/÷
2.72 (4)
2.43 µM ×/÷ 2.47
(3)
9.08 ꢁM ×/÷
1.31 (3)
2.8
9
CH3
Bu
ACS Paragon Plus Environment

Page 15 of 37
Journal of Medicinal Chemistry
tꢀ
0.198 µM ×/÷
0.284 ꢁM ×/÷
1.52 (2)
0.607 µM ×/÷
2.39 ꢁM ×/÷
1.43 (3)
3.1
4
1
2
3
4
5
6
7
8
Br
Cl
Cl
Cl
Cl
Cl
H
H
H
H
25
Bu
1.26 (3)
1.96 (3)
tꢀ
0.241 µM ×/÷
1.38 (6)
0.870 ꢁM ×/÷
1.55 (5)
0.744 µM ×/÷
1.46 (6)
2.44 ꢁM ×/÷
1.80 (6)
2.9
8
26
27
28
29
30
Bu
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tꢀ
0.385 µM ×/÷
1.4 (2)
1.23 ꢁM ×/÷
1.10 (2)
0.853 µM ×/÷
1.50 (3)
9.37 ꢁM ×/÷
1.87 (3)
3.4
1
Am
0.885 µM ×/÷
1.32 (2)
2.77 ꢁM ×/÷
1.56 (3)
0.792 µM ×/÷
1.51 (2)
8.80 ꢁM ×/÷
1.34 (2)
2.4
2
i-Pr
i-Pr
i-Pr
CH
2.32 µM ×/÷
1.14 (2)
8.85 ꢁM ×/÷
1.13 (2)
13.0 µM ×/÷ 1.69
(3)
132 ꢁM ×/÷
2.13 (3)
2.9
4
3
3.67 µM ×/÷
1.03 (2)
9.53 µM ×/÷ 1.89
(3)
57.4 ꢁM ×/÷
4.29 (3)
2.5
7
F
>10 ꢁM (2)
a Values are reported as geometric means with geometric standard deviations. Number of replicates is
in parentheses.
b Calculated using Marvin and calculator plugin freeware (www.chemaxon.com, ChemAxon Kft, Buꢀ
dapest, Hungary).
None of the analogues 20–30 decreased PGI2 production in A549 cells up to 10 ꢁM or PGF2α or TXB2
production in HWB up to 30 ꢁM, suggesting that these compounds did not have COX inhibitory activiꢀ
ty.33 Indeed, compound 26 was inactive against COXꢀ1 and COXꢀ2 (9% inhibition at 100 µM and 16%
at 30 µM, respectively)).34 Therefore, it was concluded that the PGE2 reduction exhibited by 26 resulted
from selective mPGESꢀ1 inhibition.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 37
Scheme 2. Synthesis of 24.a
1
2
3
4
5
6
7
8
9
aReagents and conditions: (a) tꢀBuCONHCH2OH, H2SO4, rt, then chromatographic separation, 27% of
32 and 28% of 33; (b) 2ꢀaminoimidazole monosulfate, BOP, Et3N, DMF, 60 °C, 87%.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For this SAR study, most cores which were not commercially available could be prepared by Tscherꢀ
niacꢀEinhorn alkylation.50,51 For example, oꢀtoluic acid (31, Scheme 2) was treated with Nꢀ
(hydroxymethyl)pivalamide in the presence of sulfuric acid to furnish a separable mixture of 32 and 33,
and then 33 was carried forward to 24. Although this chemistry was not regioselective, it enabled facile
exploration of C(2) and C(4) substituents by providing rapid access to the requisite functionalized benꢀ
zoic acids.
Of the analogues in Table 2, 25 and 26 had the lowest HWB IC80 values. Between these two comꢀ
pounds, 26 had the lower molecular weight and clogP, so it was selected for further evaluation.
Against hERG, 26 had no significant measurable activity (–4% and 14% inhibition at 10 µM and 100
µM, respectively). However, the decreased mPGES potency of 26 and the high hERG IC50 of 8 preꢀ
vented a meaningful comparison between 8 and 26 in this regard. However, on the basis of these data,
26 was not expected to carry a significant QT prolongation risk. Compound 26 did not display any sigꢀ
nificant activity against a CEREP panel of receptors and ion channels.37
Similarly to 8, compound 26 was active in dog whole blood (DWB IC50 = 0.555 ꢁM ×/÷ 1.90 and IC80
= 1.84 ꢁM ×/÷ 2.06 (n = 5)) but had substantially less potency in the rat mPGESꢀ1 assay (IC50 >1000
ꢁM) than in the human enzyme assay. With the potential to use the dog as an onꢀtarget species, we studꢀ
ied the PK properties of 26 in dog (Figure 6).
ACS Paragon Plus Environment

Page 17 of 37
Journal of Medicinal Chemistry
In dogs, it was found that 26 was readily orally bioavailable when administered as an aqueous suspenꢀ
1
2
3
sion of crystalline material (F = 60% ×/÷ 1.17 after 0.5 mg/kg po and iv dosing, Figure 7A).
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. PK of 26 in dog (A) after 0.5 mg/kg po dosing (solid diamonds, n = 3) and 0.5 mg/kg iv dosꢀ
ing (hollow diamonds, n = 3) and (B) after 2 mg/kg po dosing (n = 3). Diamonds indicate geometric
means, and error bars represent geometric standard deviations of measured values.
After a 2 mg/kg po dose (Figure 7B), plasma concentrations exceeding 1 ꢁM were achieved in less
than 15 minutes and sustained for at least 8 hours. Compound 26 exhibited relatively little exposure
variability between animals; the geometric SD for plasma concentration was ≤ 2.0 at all timepoints in
both dog PK studies. The desirable PK properties of 26, along with its potency in HWB and selectivity
for mPGESꢀ1, informed us in our selection of 26 as a clinical candidate.37 It was notable that the supeꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 37
rior physicochemical properties and reduced exposure variability of 26 compared to 8 sufficiently comꢀ
pensated for reduced potency to make 26 a viable molecule to test in the clinic.47
1
2
3
4
5
6
Scheme 3. Synthesis of 26.a
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (a) tꢀBuCOCl, DIEA, CH2Cl2, rt, 1 h; (b) LiOH, H2O, 1,4ꢀdioxane, rt, 1 h,
then HCl(aq), 97% (2 steps); (c) 2ꢀaminoimidazole monosulfate, TBTU, DIEA, DMF, 80 °C, 18 h, 83%.
A largerꢀscale synthesis of 26 proceeded straightforwardly from commercially available 34 (Scheme
3).52 Acylation of the amine with pivaloyl chloride furnished 35. Saponification of the crude ester gave
36, which was coupled with 2ꢀaminoimidazole with heating to furnish 26 in ca. 81% yield over three
steps. We viewed the simplicity of this synthesis as a potential advantage for the development of 26.
CONCLUSION
SAR efforts guided by structural insights yielded two clinical candidates, 8•H3PO4 and 26, with difꢀ
ferent profiles. While 8•H3PO4 was significantly more potent in the enzyme and HWB assays, 26 had
an improved pharmacokinetic profile in dog while still exhibiting a HWB IC50 for PGE2 suppression
comparable to that of celecoxib. Based on the known PK/PD relationship with COXꢀ2 inhibitors, this
pair of compounds was expected to offer insight into the validity of mPGESꢀ1 inhibition as a mechaꢀ
nism for analgesia in human patients.
EXPERIMENTAL SECTION
General Methods. All reagents and anhydrous solvents were obtained from commercial sources and
1
used without further purification unless noted otherwise. H NMR spectra were recorded on a Bruker
300 MHz spectrometer, a Bruker 400 MHz spectrometer, a Bruker 500 MHz spectrometer, or a Varian
1
400 MHz spectrometer. H NMR chemical shifts are reported in ppm with the solvent as the internal
ACS Paragon Plus Environment

Page 19 of 37
Journal of Medicinal Chemistry
standard (DMSOꢀd5 2.49 ppm, CHCl3 7.26 ppm). 13C NMR chemical shifts are reported in ppm with
the solvent as the internal standard (DMSOꢀd6 39.52 ppm). 19F NMR chemical shifts are reported in
ppm relative to an external standard of CFCl3. 31P NMR chemical shifts are reported in ppm relative to
an external standard of 85% H3PO4.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compounds were analyzed for purity by HPLC and HPLCꢀMS, and unless otherwise stated, purities
of synthesized compounds were all found to be >95% by the following HPLC method:
Agilent 1100 HPLC with VWD detectors, Waters Exterra MS 4.6 mm × 150 mm × 3.5 µm C18 colꢀ
umn; Eluent: 0.1% formic acid in a gradient of 5% to 100% CH3CN in water over 3.75 min; Flow rate
1.0 mL/min; Column temp. 50 °C; λ 300 nm and 214 nm.
2-bromo-1-(4-(trifluoromethyl)phenyl)-1-propanone (14). At room temperature, a mixture of 4ꢀ
(trifluoromethyl)propiophenone (13, 100 g, 494 mmol) and glacial acetic acid (200 mL) was treated
with a solution of bromine (79 g, 494 mmol) in glacial acetic acid (200 mL) over the course of 60
minutes. After the end of the addition, the mixture was stirred for 1.5 hours, then was quenched with
cold water (<5 °C, 1.2 L) and stirred for 3 hours at between 0 °C and 5 °C. The resulting slurry was filꢀ
tered and washed with water (10 °C–15 °C, 1 L), and dried under reduced pressure at between 25 °C
1
and 30 °C for 15 hours to furnish 14 (128.1 g, 92% yield). H NMR (400 MHz, CDCl3): δ 8.13 (d, J =
8.3 Hz, 2H), 7.76 (d, J = 8.3 Hz, 2H), 5.26 (q, J = 6.7 Hz, 1H), 1.93 (d, J = 6.3 Hz, 2H).
2-methyl-3-(4-(trifluoromethyl)phenyl)imidazo[1,2-a]pyrimidine (15). A mixture of 14 (100 g,
356 mmol), 2ꢀaminopyrimidine (33.9 g, 356 mmol), NaHCO3 (59.8 g, 811 mmol), and toluene (500
mL) was heated to between 90 °C and 100 °C for 24 hours with stirring. The mixture was cooled to beꢀ
tween 40 °C and 45 °C, and the product was isolated by vacuum distillation. The distillate was cooled
to between 25 °C and 30 °C, treated with water (1 L), and stirred for 4 h. The solids were collected by
filtration, washed with a 9:1 (v/v) mixture of hexanes and MTBE (200 mL), and dried in a vacuum oven
1
at between 45 °C and 50 °C to furnish 15 (35.4 g, 35% yield). H NMR (400 MHz, DMSOꢀd6): δ 8.86
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 37
(dd, J = 6.9, 1.9 Hz, 1H), 8.56 (dd, J = 4.2, 2.0 Hz, 1H), 8.07 (d, J = 8.8 Hz, 2H), 7.84 (d, J = 8.7 Hz,
2H), 7.11 (dd, J = 6.9, 4.0 Hz, 1H), 2.69 (s, 3H). LC/MS (ESI+): (m/z) 278 (M+H)+.
1
2
3
4
5
6
7
8
4-methyl-5-(4-(trifluoromethyl)phenyl)-1H-imidazol-2-amine (12). A mixture of 15 (180 g, 649
mmol), EtOH (1.4 L), and hydroxylamine (159 mL, 2.60 mol) was stirred at 82 °C for 48 hours, then
cooled and concentrated to dryness. The residue was diluted with CH2Cl2 (1.5 L) and washed with waꢀ
ter (2 × 500 mL) and saturated aqueous NaCl (500 mL). The organic phase was dried over MgSO4, filꢀ
tered, and concentrated under reduced pressure to give a yellow gum. This was subjected to flash
chromatography on silica gel, eluting with a gradient of 0% to 25% (2% NH3 in CH3OH) in CH2Cl2, to
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
furnish 12 as a yellow foam (150.3 g, 96% yield). H NMR (300 MHz, CDCl3): δ 7.42 (ABq, JAB = 8.9
Hz, ꢂδAB = 0.058, 4H), 6.48 (v br s, 3H), 2.16 (s, 3H). LC/MS (ESI+): (m/z) 242 (M+H)+.
2-(difluoromethyl)-5-(isobutyramidomethyl)-N-(4-methyl-5-(4-(trifluoromethyl)phenyl)-1H-
imidazol-2-yl)nicotinamide (8). A mixture of 16 (230 g, 845 mmol), 12 (281 g, 845 mmol), NMM
(428 g, 4.22 mol), and EtOAc (2.5 L) was stirred between 15 °C and 25 °C for 1 h, then treated with
propylphosphonic anhydride (50% w/w in EtOAc, 1.88 kg, 2.95 mol) over the course of 17 minutes,
maintaining the internal temperature below 40 °C. The addition vessel was rinsed with EtOAc (415
mL), which was added to the reaction vessel. The mixture was stirred for one additional hour, then was
warmed to between 65 °C and 75 °C with stirring overnight. The mixture was cooled to between 15 °C
and 25 °C and diluted with EtOAc, then washed with water (2 × 2 L), 1 N aqueous HCl (2 × 2 L), satuꢀ
rated aqueous NaCl (2 × 2 L), 10% (w/w) aqueous Na2CO3 (2 × 2 L), then deionized water (2 × 2 L).
The organic phase was concentrated under reduced pressure until solids appeared (at a volume of ca. 4
L), then absolute EtOH (3 L) was added, and the mixture was concentrated to a volume of ca. 3 L. Abꢀ
solute EtOH (1 L) was added, and the mixture was stirred at between 65 °C and 75 °C for 30 min, then
at between 15 °C and 25 °C for 63 hours. The slurry was cooled to between –5 °C and –15 °C with stirꢀ
ring for two hours, then the solids were isolated by filtration, rinsed with cold (–10 °C) absolute EtOH
1
(420 mL), and dried in a 50 °C vacuum oven for two nights to furnish 8 (268 g, 64% yield). H NMR
ACS Paragon Plus Environment

Page 21 of 37
Journal of Medicinal Chemistry
(400 MHz, DMSOꢀd6): δ 12.08 (br s, 2H), 8.65 (s, 1H), 8.40 (t, J = 5.6 Hz, 1H), 8.06 (s, 1H), 7.81 (A of
ABq, d, J = 8.1 Hz, 2H), 7.73 (B of ABq, d, J = 8.1 Hz, 2H), 7.35 (v br t, J = 54 Hz, 1H), 4.38 (d, J =
5.6 Hz, 2H), 2.46 (s, 3H), 2.44 (septet, J = 6.8 Hz, 1H), 1.04 (d, J = 6.8 Hz, 6H). 13C{1H} NMR (100
MHz, DMSOꢀd6): δ 176.4, 165.5, 149.4, 147.7 (t, J = 22.0 Hz), 141.1 (br s), 138.1 (br s), 137.1, 135.8,
130.7 (br s), 125.9 (q, J = 31.9 Hz), 125.7, 125.4 (q, J = 3.7 Hz), 124.5 (q, J = 271.4 Hz), 122.3 (br s),
111.5 (t, J = 238.4 Hz), 34.0, 19.5, 11.2. 19F{1H} NMR (376.5 MHz, DMSOꢀd6): δ –60.7, –115.0. 31P
NMR (400 MHz, DMSOꢀd6): no signal. LC/MS (ESI+): (m/z) 496 (M+H)+, 518 (M+Na)+, 534 (M+K)+.
HRMS calcd for C23H23F5N5O2 (M+H)+: 496.1772, found 496.1771. Anal. calcd. for C23H22F5N5O2: C,
55.75; H, 4.48; N, 14.14; found: C, 55.53; H, 4.43; N, 14.11.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2-(difluoromethyl)-5-(isobutyramidomethyl)-N-(4-methyl-5-(4-(trifluoromethyl)phenyl)-1H-
imidazol-2-yl)nicotinamide phosphoric acid salt (8•H3PO4). To a mixture of 8 (250 g, 505 mmol) in
CH3CN (5 L) was added a solution of 85% H3PO4 (110.7 g, 960 mmol) in deionized water (1 L). The
mixture was heated to between 60 °C and 70 °C, then filtered through a 1.2 ꢁM filter capsule. The filꢀ
trate was heated to between 55 °C and 65 °C, then was stirred at ambient temperature overnight. The
mixture was cooled to 5 °C and stirred at that temperature for 2 hours. The resulting solids were colꢀ
lected by filtration and rinsed with cold (0 °C–10 °C) CH3CN (2 × 475 mL), followed by cold (0 °C–10
°C) deionized water (2 × 475 mL). The wet solids were slurried with deionized water (2.5 L) at room
temperature for 2 hours, then collected by filtration and rinsed with the filtrate (3×), followed by deionꢀ
ized water (1.25 L). The solids were dried at 110 °C under reduced pressure to furnish 8•H3PO4 as a
1
pale yellow solid (237 g, 79% yield). H NMR (400 MHz, DMSOꢀd6): δ 8.65 (d, J = 1.8 Hz, 1H), 8.42
(t, J = 5.8 Hz, 1H), 8.06 (s, 1H), 7.81 (A of ABq, d, J = 8.1 Hz, 2H), 7.73 (B of ABq, d, J = 8.1 Hz, 2H),
7.35 (t, J = 54.1 Hz, 1H), 4.38 (d, J = 5.6 Hz, 2H), 2.46 (s, 3H), 2.45 (septet, J = 6.9 Hz, 1H), 1.04 (d, J
= 6.9 Hz, 6H). 13C{1H} NMR (100 MHz, DMSOꢀd6): δ 176.5, 165.5, 149.5, 147.7 (t, J = 22.0 Hz),
141.1, 137.9, 137.2, 135.8, 130.7 (t, J = 3.7 Hz), 127.6, 126.1, 125.9 (q, J = 31.9 Hz), 125.4 (q, J = 3.7
Hz), 124.5 (q, J = 271.4 Hz), 122.4, 111.5 (t, J = 238.4 Hz), 34.0, 19.5, 11.2. 19F{1H} NMR (376.5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 37
MHz, DMSOꢀd6): δ –60.7, –115.0. 31P NMR (400 MHz, DMSOꢀd6): δ –0.7. HRMS calcd for
C23H23F5N5O2 (M)+: 496.1772, found 496.1755. Anal. calcd. for C23H24F5N5O6P: C, 46.55; H, 4.25; F,
16.01; N, 11.80; found: C, 46.63; H, 4.22; F, 16.30; N, 11.91.
1
2
3
4
5
6
7
8
9
2-chloro-5-(pivalamidomethyl)benzoic acid (36). A thermally controlled reactor was charged with
methyl 5ꢀ(aminomethyl)ꢀ2ꢀchlorobenzoate hydrochloride (34, 700 g, 2.96 mol), dichloromethane (4.90
L), and N,Nꢀdiisopropylethylamine (1.45 L, 8.30 mol). Pivaloyl chloride (436 mL, 3.56 mol) was added
at such a rate that the internal temperature did not rise above 21 °C. The resulting solution was stirred at
room temperature for 1 h. Then, the mixture was quenched with saturated aqueous NaHCO3 (10 L) and
extracted with dichloromethane (2 × 1000 mL). The combined organic extracts were dried over sodium
sulfate and filtered. The filtrate was concentrated under reduced pressure to furnish crude 35 as a strawꢀ
colored oil. This was dissolved in 1,4ꢀdioxane (5.25 L) and treated with water (3.15 L) and LiOH•H2O
(408 g, 9.62 mol) and stirred at room temperature for 1 h. The mixture was then treated with 5.0 M
aqueous HCl (3 L) (Caution! Exotherm!) and concentrated under reduced pressure to remove ca. 6 L of
volatile material. The resulting suspension was cooled to 10 °C, and the solids were isolated by filtraꢀ
tion, washed with water (8 L), and dried on the sinter for 3 h. The solids were then dried in a 40 °C
vacuum oven for 48 hours to furnish 36 as a freeꢀflowing white solid (775 g, 97% yield over two steps).
1H NMR (400 MHz, DMSOꢀd6): δ 13.32 (s, 1H), 8.13 (t, J = 5.8 Hz, 1H), 7.62 (d, J = 2.4 Hz, 1H), 7.47
(d, J = 8.3 Hz, 1H), 7.34 (dd, J = 8.3, 2.0 Hz, 1H), 4.25 (d, J = 5.8 Hz, 2H), 1.11 (s, 9H). LC/MS
(ESI+): (m/z) (35Cl/37Cl) 270/272 (M+H)+.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2-chloro-N-(1H-imidazol-2-yl)-5-(pivalamidomethyl)benzamide (26). A thermally controlled reacꢀ
tion vessel containing 36 (420 g, 1.56 mol), 2ꢀaminoimidazole monosulfate (349 g, 1.87 mol), and DMF
(2.94 L) was treated with oꢀbenzotriazolꢀ1ꢀylꢀN,N,N’,N’ꢀtetramethyluronium tetrafluoroborate (722 g,
2.18 mol) and then DIEA (869 mL, 4.98 mol) and heated to an internal temperature of 80 °C for 18 h.
Then, the reaction mixture was cooled to 10 °C and added to ice water (14 L) and stirred for 90 min.
The resulting solids were collected by filtration and air dried for 24 h, then dried further in a 50 °C vacꢀ
ACS Paragon Plus Environment

Page 23 of 37
Journal of Medicinal Chemistry
uum oven for 24 h. The resulting solid material was slurried in isopropanol (7 L), heated to an internal
temperature of 73 °C for 3 h, cooled to 10 °C, and the solids were isolated by filtration. The resulting
material was air dried for 24 h, then dried further in a 55 °C vacuum oven for 48 h to furnish 26 as a fiꢀ
1
2
3
4
5
6
7
8
1
ne, offꢀwhite solid (435 g, 83% yield). H NMR (400 MHz, DMSOꢀd6): δ 11.83 (br s, 2H), 8.12 (t, J =
9
6.0 Hz, 1H), 7.46 (d, J = 8.3 Hz, 1H), 7.39 (d, J = 1.7 Hz, 1H), 7.30 (dd, J = 8.3, 2.0 Hz, 1H), 6.72 (br s,
2H), 4.26 (d, J = 5.9 Hz, 2H), 1.11 (s, 9H). 13C{1H} NMR (100 MHz, DMSOꢀd6): δ 177.6, 165.6,
141.3, 139.5, 135.7, 129.6, 129.5, 128.1, 127.6, 118.1 (v br s), 41.4, 38.0, 27.4. LC/MS (ESI+): (m/z)
(35Cl/37Cl) 335/337 (M+H)+. HRMS calcd. for C16H20ClN4O2 (M+H)+: 335.1275, found 335.1274.
Anal. calcd. for C16H19ClN4O2: C, 57.40; H, 5.72; N, 16.73; Cl, 10.59; O, 9.56; found: C, 57.27; H,
5.64; N, 16.79; Cl, 10.59, O, 9.57.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Biological Assays.
mPGES-1 Enzyme Inhibition Assays. Human or rat mPGESꢀ1 (Invitrogen™ (Cat #97002RG, clone
ID 6374722)) was subcloned into pcDNA™3.1 and transiently expressed in 293E cells. Microsomes
were prepared from cell pellets based on published methods.10,11 Pellets were brought up in homogeniꢀ
zation buffer (15 mM TrisꢀHCl, pH 8.0; 0.25 M sucrose; 0.1 mM EDTA; 1 mM glutathione) and soniꢀ
cated 5 × 30 seconds on ice. Homogenate was centrifuged at 5000g for 10 minutes at 4 °C. The superꢀ
natant fraction was decanted, loaded into Beckman QuickꢀSeal® tubes, and centrifuged at 150000g for
90 minutes at 4 °C. The supernatant fraction was discarded by decantation, and the pellets were resusꢀ
pended in assay buffer (10 mM sodium phosphate (pH 7.0), 10% glycerol, and 2.5 mM glutathione
Complete Protese Inhibitor Cocktail (Roche)). Protein concentration was determined using the Pierce
Coomassie Plus™ reagent.
For the enzyme assay, the microsomes were diluted into assay buffer and 14 ꢁL/well was added to 384
well plates. Compound dilution plates (Nunc® Catalog #249944) were generated on a Tecan Evo™, and
4 ꢁL/well was added to the assay plates. PGH2 was diluted into assay buffer immediately prior to use,
and 7 ꢁL/well was added. Final concentrations were 6.52 ꢁg/mL of microsomes and 1.67 ꢁM PGH2.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 37
After a 2.5 minute incubation at room temperature, 5 ꢁL/well of (1 mg/mL of SnCl2 in 0.5 N HCl(aq))
was added to stop the reaction. 5 ꢁL of the reaction was transferred to a 384 well plate containing 0.1%
aqueous formic acid (45 ꢁL) for Mass Spec dilution, then the plates were stored at –20 °C. The plates
were analyzed for PGE2 using standard LC/MS/MS analysis (Biocius Lifesciences, Wakefield, MA,
USA).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cell-Based Assay. Human epithelial lung carcinoma cell line A549 was purchased from ATCC (CCLꢀ
185) and was maintained in Kaighn’s F12 + 10% FBS, 5% CO2. For the assay, cells were plated at
40,000/well in 96ꢀwell Falcon plates, 24 hours prior to treatment. Compounds were diluted in DMSO
and were added at 1 µL /well, n=2, to give seven concentrations each. Cells were pretreated 30 minutes
at 37 °C, 5% CO2. rhILꢀ1β (R&D Systems) was added to give 0.2 ng/mL final. The treatment period
was ~18 hours. The conditioned medium was assayed for levels of PGE2 and 6ꢀketo PGF1α by EIA
(Cayman). The IC50s were calculated using Graphpad Prism nonlinear regression sigmoidal dose reꢀ
sponse curve fitting. Data are the geometric mean ×/÷ geometric standard deviation of the indicated
number of determinations.
Human Whole Blood Assay. With informed consent, blood was collected from normal volunteer doꢀ
nors into sodium heparin vacutainer tubes. Donors had not taken NSAIDs, aspirin, celecoxib, or glucoꢀ
corticoids within two weeks of the donation. All tubes from each donor were pooled into 250 mL Cornꢀ
ing conical centrifuge tubes and 436.5 ꢁL/well was distributed into deep well polypropylene plates.
Test compounds were diluted in DMSO to 100× the final concentration, and 4.5 ꢁL/well in duplicate or
triplicate was added to give 7 point curves. The blood was pretreated at 37 °C, 5% CO2, in a humidified
atmosphere, loosely covered with a silicone cap mat, for 30 minutes, after which 9 ꢁL/well of a solution
of 5 mg/mL of lipopolysaccharide (LPS) (Sigma 0111:B4) in 1 mg/mL bovine serum albumin
(BSA)/PBS was added to give a final LPS concentration of 100 ꢁg/mL. The plates were incubated for
20–24 hours, loosely covered, at 37 °C, 5% CO2, in a humidified atmosphere, on an orbital shaker at
approximately 100 rpm. The plates were tightly sealed with silicone cap mats and chilled on ice for apꢀ
ACS Paragon Plus Environment

Page 25 of 37
Journal of Medicinal Chemistry
proximately 1 hour. Then, the plates were centrifuged at 1800g for 10 minutes at 4 °C in an Eppendorf
5810R centrifuge. Plasma was removed from the cell layer using the Rainin L200 with sterile filtered
tips and transferred to vꢀbottom polypropylene plates. 100 ꢁL was quantitatively transferred to Costar
cluster tubes blocks; and 400 ꢁL/well of the methanol stop reagent and of each of the internal standards
PGE2ꢀd4, PGF2αꢀd4, and TXB2βꢀd4 were added. Samples were vortexed for 5 minutes and placed at –20
°C for at least one hour. Samples were centrifuged for 10 minutes at 4000 rpm in an Eppendorf 5810R.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Solid phase extraction was performed using Waters HLB 30 mg/bed 96 well plates on a vacuum maniꢀ
fold: 1) the matrix was washed with methanol (1 mL), followed by 0.1% formic acid in water (1 mL); 2)
400 µL sample was applied along with 0.1% formic acid in water (900 µL) and allowed to bind for 5
minutes; 3) the matrix was washed with 0.1% formic acid in water (600 µL), followed by 80/20 waꢀ
ter/methanol (600 µL); 4) the products were eluted with EtOAc (2 × 500 ꢁL); 5) the samples were dried
under nitrogen and reconstituted in 75/25 water/acetonitrile with 0.1% formic acid (50 µL). The prodꢀ
ucts were analyzed by LCꢀMS/MS.
Dog Whole Blood Assay. Dog whole blood was obtained from Covance (Greenfield, Indiana). All
tubes from each donor were pooled and kept on ice. To 96ꢀwell Falcon 3072 plates, 98 ꢁL of blood was
added per well. Test compounds were diluted in DMSO to 100× the final concentration, and 1 ꢁL/well
in duplicate or triplicate was added to each well to give 7 point curves. The blood was pretreated with
compounds at 37 °C, 5% CO2, in a humidified atmosphere for 30 minutes, after which 1 ꢁL/well of a
solution of 0.1 mg/mL of lipopolysaccharide (LPS) (Sigma, serotype 0111:B4) in 0.1% bovine serum
albumin (BSA)/PBS was added to give a final LPS concentration of 1 ꢁg/mL. The plates were incubatꢀ
ed for 5 hours at 37 °C, 5% CO2, in a humidified atmosphere. After incubation, the plates were chilled
on ice for approximately 10 min. Then, the plates were centrifuged at 1800g for 10 minutes at 4 °C in
an Eppendorf 5810R centrifuge. Plasma was removed from the cell layer and transferred to vꢀbottom
polypropylene plates. Plasma samples were assayed for PGE2 at 1:50 dilution using PGE2 Express EIA
assay kit (Cayman Chemical). The plates were read at A412 on a plate reader (Molecular Devices Versaꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 37
max) using SOFTmaxPRO (v. 4.3.1) software. IC50 values were calculated using Graphpad Prism nonꢀ
linear regression, sigmoidal dose response curve fitting.
1
2
3
4
5
6
7
8
CYP 3A4 Inhibition Assay. Incubation samples were prepared by adding 150 ꢁL of a solution of miꢀ
crosomes (protein concentration 66.8 ꢁg/mL, adjusted to pH 7.4 with phosphate buffer) and 2 ꢁL of a 1
mM (90:10 CH3CN:DMSO) solution of test compound, and mixed well. Samples were preꢀincubated
for approximately 10 minutes at approximately 37 °C. Following the preꢀincubation period, the reaction
was initiated with the addition of 50 ꢁL of a solution of NADPH (2 mM), containing marker substrates
bufuralol (40 ꢁM), midazolam (20 ꢁM), and diclofenac (40 ꢁM) in MilliꢀQ water. Final incubation volꢀ
ume was 200 ꢁL, which contained 10 ꢁM test compound, 10 ꢁM bufuralol, 5 ꢁM midazolam, 10 ꢁM
diclofenac, 0.5 mM NADPH, and 0.05 mg/mL protein.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The reaction mixture was incubated for 3 minutes at approximately 37 °C, after which a 70 ꢁL aliquot
was taken from the reaction mixture and quenched by addition to 60 ꢁL of an acetonitrile solution conꢀ
taining 1ꢀhydroxybufuralol maleateꢀd9 (0.12 ꢁM), αꢀhydroxymidazolamꢀd4 (0.17 ꢁM) and [13C6]ꢀ4’ꢀ
hydroxydiclofenac (0.19 ꢁM). The reaction mixture was centrifuged at approximately 4000 rpm for 15
minutes at approximately 5 °C. Following centrifugation, the samples were analyzed by LCꢀMS/MS
using the following method:
LC Column: Javelin Betasil C18 (20 mm × 2 mm); Mobile Phase A: 98.5:1:0.5 MilliꢀQ H2O : CH3OH:
1 M NH4HCO3 (v/v); Mobile Phase B: 99:1 CH3OH: 1 M NH4HCO3 (v/v). Stepwise gradient from
98% Mobile Phase A/2% Mobile Phase B to 2% Mobile Phase A/98% Mobile Phase B over 0.85
minutes.
MS/MS method: Collision gas: 10 psi; Curtain gas: 40 psi; Ion source gas 1: 55 psi; Ion source gas 2:
70 psi; IonSpray voltage: 1500 V; Source temperature: 700 °C; Entrance potential: 10 V. Samples were
analyzed for metabolites and internal standards with their respective mass transitions 1ꢀOHꢀbufuralol
(278.1/186.1), 1ꢀOHꢀbufuralolꢀd9 (287.1/186.1), 1ꢀOHꢀmidazolam (342.1/203.1), 1ꢀOHꢀmidazolamꢀd4
(346.1/203.1), 4ꢀOHꢀdiclofenac (312.2/231.2), and 4ꢀOHꢀdiclofenacꢀd6 (318.2/231.2). Area ratios (meꢀ
ACS Paragon Plus Environment

Page 27 of 37
Journal of Medicinal Chemistry
tabolite area / stable label metabolite internal standard) were determined for incubations with inhibitor
and compared with incubations containing solvent only (vehicle control). Percent inhibition was calcuꢀ
lated by 100 – (area ratio with test compound/area ratio vehicle control × 100).
1
2
3
4
5
6
7
8
9
CYP 3A4 Time-Dependent Inhibition Assay. Incubation samples were prepared by adding 178 ꢁL
of a solution of microsomes (protein concentration 56.18 ꢁg/mL, adjusted to pH 7.4 with phosphate
buffer) and mixing well. These were preꢀincubated for approximately 12 minutes at approximately 37
°C. Following the preꢀincubation period, the inactivation reaction was initiated with the addition of 2
ꢁL of a 1 mM (90:10 CH3CN:DMSO) solution of inhibitor and 20 ꢁL of a solution of NADPH (10 mM)
in a pH 7.4 aqueous (100 mM) phosphate buffer to give a final concentration of 10 ꢁM test compound, 1
mM NADPH, and 0.5 mg/mL microsomes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Separately, two activity plates (one each for a 0ꢀminute and a 30ꢀminute time point) were prepared
containing 95 ꢁL of a solution of NADPH (1 mM) and midazolam (100 ꢁM) in a pH 7.4 aqueous (100
mM) phosphate buffer, and preꢀincubated at 37 °C for approximately 12 minutes. These were called
Plate 0 and Plate 30, respectively.
Immediately upon preparation of the inactivation sample, a 5 ꢁL aliquot was taken and added to Plate
0 and incubated at 37 °C for 1 minute. This was quenched by the addition of 50 ꢁL of an acetonitrile
solution of αꢀhydroxymidazolamꢀd4 (0.17 ꢁM).
After incubating for 30 minutes, a 5 ꢁL aliquot of the inactivation sample was taken and added to
Plate 30 and incubated at 37 °C for 1 minute. This was quenched by the addition of 50 ꢁL of an aceꢀ
tonitrile solution of αꢀhydroxymidazolamꢀd4 (0.17 ꢁM).
The reaction mixtures were centrifuged at approximately 4000 rpm for 15 minutes at approximately 5
°C. Following centrifugation, the samples were analyzed by LCꢀMS/MS using the following method:
LC Column: Javelin Betasil C18 (20 mm × 2 mm); Mobile Phase A: 95:5 MilliꢀQ H2O : 1 M
NH4HCO3 (v/v); Mobile Phase B: 95:5 CH3OH: 1 M NH4HCO3 (v/v). Gradient: 0.3 min at 95:5 Mobile
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 37
Phase A:Mobile Phase B; ramp to 5:95 Mobile Phase A:Mobile Phase B over 1.0 minute; hold at 5:95
Mobile Phase A:Mobile Phase B for 0.5 minutes.
1
2
3
4
5
6
7
8
MS method: MRM at 342.1/203.1 (1ꢀOHꢀmidazolam) and 346.1/203.1 (αꢀhydroxymidazolamꢀd4) usꢀ
ing TurboIon spray under positive ion conditions. Nebulizer gas: 12 psi; Curtain gas: 8 psi; IonSpray
voltage: 2500 V; Source temperature: 500 °C; Declustering potential: 26 V; Focusing potential: 100 V;
Entrance potential: 4 V; Dwell time 200 ms. Area ratios (metabolite area / stable label metabolite interꢀ
nal standard) were determined for 30 minute inactivation incubations with inhibitor and compared with
incubations containing solvent only (vehicle control). Percent inhibition was calculated by 100 – (area
ratio 30 minute inactivation/area ratio 0 minute inactivation × 100).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Solubility Assays. The test compound (2 mg) was treated with the solubility medium (FasSIF or 0.1N
aqueous HCl, 1.0 mL) and agitated at 25 rpm at room temperature for 24 h. The resulting slurry was
centrifuged at 14000 rpm (16873 rcf) with a 0.45 ꢁm centrifugal filter for 2 min. The pH of the filtrate
was measured, and the concentration of dissolved drug was measured by HPLC analysis.
In vivo studies.
The pharmacokinetic and plasma exposure animal experiments are protocol driven studies. All proceꢀ
dures in the protocols are in compliance with the U.S. Department of Agriculture (USDA) Animal Welꢀ
fare Act (9CFR Parts 1, 2, and 3); the Guide for the Care and Use of Laboratory Animals from the Instiꢀ
tute of Laboratory Animal Resources, National Academy Press, Washington, D.C., 1996; and the Naꢀ
tional Institutes of Health, Office of Laboratory Animal Welfare.
Pharmacokinetic: Male Beagle dogs (9.2 to 12.7 kg, N = 3/dose group) were administered test comꢀ
pounds as an intravenous (bolus) solution and oral (gavage) suspension. Blood (approximately 1
mL/timepoint) was collected from the jugular vein at 0.08 (IV only), 0.25, 0.5, 1, 2, 4, 8, 12, and 24
hours post dose into tubes containing EDTA antiꢀcoagulant. Blood samples were centrifuged (3000
RPM for 10 minutes at 4°C), plasma was removed and stored at approximately –60 °C until analyzed.
ACS Paragon Plus Environment

Page 29 of 37
Journal of Medicinal Chemistry
Plasma concentrations of test compounds were determined using an LC/MSꢀMS bioanalytical method
(test compound concentration curve range = 1 to 5000 ng/mL).
1
2
3
4
5
6
7
8
Plasma exposure (free base vs phosphate salt comparison): Male Beagle dogs (9.1 to 12.7 kg, N =
6/dose group) were administered test compounds as oral (gavage) suspensions. The phosphate salt dose
was corrected to give the dose as free base equivalents. Blood (approximately 1 mL/timepoint) was colꢀ
lected from the jugular vein at 0.25, 0.5, 1, 2, 4, 8, 12, 24 and 48 hours post dose into tubes containing
EDTA antiꢀcoagulant. Blood samples were centrifuged (3000 RPM for 10 minutes at 4°C), plasma was
removed and stored at approximately –20 °C until analyzed. Plasma concentrations of test compounds
were determined using an LC/MSꢀMS bioanalytical method (test compound, as free base, concentration
curve range = 1 to 20000 ng/mL).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ASSOCIATED CONTENT
Supporting Information
Procedures for the preparation of 5–7, 9–10, 20–25 and 27–30, crystallization and structural determinaꢀ
tion, A549 assay PGI2 data, HWB assay TXB2 and PGF2α data, and CEREP panel results. This material
is available free of charge via the Internet at http://pubs.acs.org.
Accession codes
Coordinates and structure factors are available from the Protein Data Bank with accession codes 5BQG,
5BQH, and 5BQI for mPGESꢀ1/4, mPGESꢀ1/5, and mPGESꢀ1/8 binary complexes, respectively.
AUTHOR INFORMATION
Corresponding Author
* (M.A.S.) Eꢀmail: schifflerma@lilly.com; Tel: +1ꢀ317ꢀ433ꢀ4994.
Present Addresses
† Retired.
ACS Paragon Plus Environment