Fragment-Based Approach to the Development of an Orally Bioavailable Lactam Inhibitor of Lipoprotein-Associated Phospholipase A2 (Lp-PLA2)

Article abstract of DOI:10.1021/acs.jmedchem.6b01427

Lp-PLA₂ has been explored as a target for a number of inflammation associated diseases, including cardiovascular disease and dementia. This article describes the discovery of a new fragment derived chemotype that interacts with the active site of Lp-PLA₂. The starting fragment hit was discovered through an X-ray fragment screen and showed no activity in the bioassay (IC₅₀ > 1 mM). The fragment hit was optimized using a variety of structure-based drug design techniques, including virtual screening, fragment merging, and improvement of shape complementarity. A novel series of Lp-PLA₂ inhibitors was generated with low lipophilicity and a promising pharmacokinetic profile.

Full text of DOI:10.1021/acs.jmedchem.6b01427

Subscriber access provided by United Arab Emirates University | Libraries Deanship
Article
A Fragment-Based Approach to the Development of an Orally Bioavailable
2
Lactam Inhibitor of Lipoprotein-Associated Phospholipase A2 (Lp-PLA)
Alison J.-A. Woolford, Philip J. Day, Véronique Bénéton, Valerio Berdini, Joseph E. Coyle,
Yann Dudit, Pascal Grondin, Pascal Huet, Lydia Y. W. Lee, Eric S. Manas, Rachel L.
McMenamin, Christopher W Murray, Lee W. Page, Vipulkumar K. Patel, Florent Potvain,
Sharna J. Rich, Yingxia Sang, Don O. Somers, Lionel Trottet, Zehong Wan, and Xiaomin Zhang
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01427 • Publication Date (Web): 10 Nov 2016
Downloaded from http://pubs.acs.org on November 10, 2016
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 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A FragmentꢀBased Approach to the Development of an Orally Bioavailable Lactam
2
Inhibitor of LipoproteinꢀAssociated Phospholipase A2 (LpꢀPLA )
,†,#
†,#
†
∥
Alison J.ꢀA. Woolford,* Philip J. Day, Véronique Bénéton, Valerio Berdini, Joseph E.
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
†
†
§
∥
∥
∥
Coyle, Yann Dudit, Pascal Grondin, Pascal Huet, Lydia Y. W. Lee, Eric S. Manas,
†
†
†
,‡
Rachel L. McMenamin, Christopher W. Murray, Lee W. Page, Vipulkumar K. Patel,*
†
‡
∥
⊥
∥
Florent Potvain, Sharna J. Rich, Yingxia Sang, Don O. Somers, Lionel Trottet, Zehong
Wan, and Xiaomin Zhang.⊥
⊥
†
Astex Pharmaceuticals, 436 Cambridge Science Park, Milton Road, Cambridge CB4 0QA,
United Kingdom.
‡
GlaxoSmithKline, Gunnels Wood Road, Stevenage SG1 2NY, United Kingdom.
§
GlaxoSmithKline, 1250 South Collegeville Road, Collegeville, Pennsylvania 19426, United
States.
∥Centre de Recherches Francois Hyafil, GlaxoSmithKline, 25–27 Avenue du Québec, Les
Ulis, France.
⊥Neurodegeneration DPU, GlaxoSmithKline, 898 Halei Road, Zhangjiang HiꢀTech Park,
Pudong, Shanghai 201203, China.
2
ABSTRACT: LpꢀPLA has been explored as a target for a number of inflammation
associated diseases, including cardiovascular disease and dementia. This article describes the
discovery of a new fragment derived chemotype that interacts with the active site of Lpꢀ
PLA
2
. The starting fragment hit was discovered through an Xꢀray fragment screen and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
showed no activity in the bioassay (IC50 >1mM). The fragment hit was optimised using a
variety of structureꢀbased drug design techniques, including virtual screening, fragment
merging and improvement of shape complementarity. A novel series of LpꢀPLA
was generated with low lipophilicity and a promising pharmacokinetic profile.
2
inhibitors
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
KEYWORDS: LpꢀPLA , fragmentꢀbased drug discovery, structure guided optimisation.
INTRODUCTION
LpꢀPLA is a member of the phospholipase A2 superfamily, which specifically cleave the snꢀ
2
1
2 ester bond of glycerophospholipids. LpꢀPLA is found in the plasma and circulates in
2
2
complex to high and low density lipoproteins. The biological substrate(s) for this lipase are
unknown but studies examining hydrolysis of a variety of snꢀ2 acyl chain lengths have shown
3
that LpꢀPLA
2
hydrolyses oxidized and truncated phospholipids. Cleavage of oxidatively
damaged phospholipid substrates releases proꢀinflammatory factors such as
4
lysophosphatidylcholine and oxidized nonꢀesterified fatty acids, which can lead to
5
inflammation in the arteries and cardiovascular disease (CVD). Thus, LpꢀPLA
2
is a proꢀ
have been shown to
is now an established biomarker. Other diseases that have
inflammatory, lipidꢀmodifying, enzyme. The plasma levels of LpꢀPLA
2
correlate with CVD and LpꢀPLA
2
6
7
8
elevated levels of LpꢀPLA include dementia, diabetic macular edema and prostate cancer.
2
9
Inhibitors of LpꢀPLA
2
have been investigated in the clinic for atherosclerosis and
10
Alzheimer’s disease. For example, darapladib (1) is an orally bioavailable and very potent
11
2
inhibitor of LpꢀPLA with IC50 = 49 pM in a biochemical assay. A secondary ‘plasma’
assay measures inhibition in whole plasma, which approximates the physiological
12ꢀ13
environment of the enzyme, and darapladib has an IC50 = 35 nM.
Darapladib has
ACS Paragon Plus Environment

Page 3 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
undergone two phase three trials for cardiovascular disease but has been discontinued
because it did not meet the primary end points. Several other classes of inhibitors have been
14
reported in the literature but none have yet progressed to the clinic.
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
Et N
2
N
O
N
O
S
CF3
N
F
Darapladib (1)
1
5
16
The crystal structure of apo LpꢀPLA
2
, and its complex with darapladib, have previously
been reported. LpꢀPLA exhibits the classic lipase α/βꢀhydrolase fold with a catalytic triad
2
comprised of Ser273, His351 and Asp296 adjacent to an oxyanion hole formed by the
backbone NH’s of both Leu153 and Phe274. A water molecule occupies the oxyanion hole in
2
apo LpꢀPLA but is displaced by the pyrimidone carbonyl of darapladib which forms the
same two Hꢀbonds to Leu153 and Phe274 (Figure 1a). The biꢀaryl rings in darapladib stack
upon the fluorophenyl in a hydrophobic collapse and together they fill the lipophilic groove
16
marked by the residues Phe110 and Phe357. The (2ꢀaminoethyl)diethylamine moiety is
mainly solvent exposed and is important for the solubility and pharmacokinetic properties of
17
the molecule.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(a)
(b)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
Figure 1. (a) Xꢀray crystal structure darapladib (1) bound to LpꢀPLA . The Connolly surface
of darapladib is shown in solid purple and the protein as a gray mesh. (b) Superposition of the
Xꢀray crystal structures of uracil 2 (cyan), hydantoin 3 (magenta) and sulfonamide 4 (green)
bound to LpꢀPLA . Selected residues are shown for clarity. The hydrogen bond interactions
2
are shown as dotted lines.
ACS Paragon Plus Environment

Page 5 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(a) Uracil 2
(b) Hydantoin 3
(c) Sulfonamide 4
Me
O
Me
NH
Me
O
S
Me
NH
N
O
O
O
Me
O
N
N
Me
Me
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
O
CN
Figure 2. Xꢀray crystal structures of compounds 2–4 bound to LpꢀPLA
2
; uracil 2 (cyan) is a
fragment hit that binds adjacent to the catalytic residues Ser273, His351, and Asp296;
hydantoin 3 (magenta) is a fragment hit that is bound in the oxyanion hole; and sulfonamide 4
16a
(
green) is derived from a fragment hit. The Connolly surface of the protein is shown as a
gray mesh. Selected residues are shown for clarity. The hydrogen bond interactions are
shown as dotted lines and water is shown as gray spheres.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
a, 15
PLA2ꢀVIIB is a closely related intracellular phospholipase.
PLA and PLA2ꢀVIIB are highly conserved but in the adjoining groove there are a number of
differences including Phe110 (LpꢀPLA )/Tyr65 (PLA2ꢀVIIB), Gln352 (LpꢀPLA )/Arg315
PLA2ꢀVIIB) and Ala355 (LpꢀPLA )/Thr318 (PLA2ꢀVIIB). Darapladib (1) possesses a 1200ꢀ
The catalytic sites of Lpꢀ
2
2
2
(
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
fold selectivity for LpꢀPLA over PLA2ꢀVIIB, and makes van der Waals contacts with the
2
1
6
sidechains of these residues.
One of the goals for this project was to identify novel chemotypes that had comparable
activity in plasma assay to darapladib. We reasoned that if we were able to achieve a
significant decrease in the drop off between the LpꢀPLA biochemical and plasma assays, the
2
lead compound does not need to be picomolar in the biochemical assay. Therefore, the lead
compound could have lower molecular weight (MW) and more reasonable lipophilicity than
1
8
darapladib (MW = 667, clogP = 8.3). Literature has shown that aspiring to keep the
physicochemical properties within guidelines is desirable (i.e. MW <500 and clogP <5)
1
9
through association with increased success in the clinic. However, optimisation against the
plasma potency should not be performed without a parallel consideration of in vivo
20
parameters and we therefore looked to achieve good pharmacokinetic properties.
Additionally, we wished to retain greater than 100ꢀfold selectivity against PLA2ꢀVIIB. We
chose to use fragment based technology to meet these goals, as this approach can facilitate
tighter control of physicochemical properties.
16a
Our fragment screen methodology and output have been reported previously.
Crystal
structures for 50 fragment hits were obtained, and were observed to bind throughout the
entire length of the canyonꢀlike groove of LpꢀPLA occupying a total length of ~24 Å.
2
ACS Paragon Plus Environment

Page 7 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
The Xꢀray crystal structures of three compounds (2–4) bound to LpꢀPLA are superimposed
in Figure 1b (and shown separately in Figure 2). These compounds occupy different regions
16a
of the binding site and recapitulate many of the key binding interactions of darapladib.
2
1
Hydantoin 3 (IC50 >1 mM) was perceived to be an attractive starting point. It is a small
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
fragment (heavy atom count = 10), that is synthetically accessible with excellent vectors for
21
21
exploring the binding pocket. Uracil 2 (IC50 ~1 mM) and sulfonamide 4 (IC50 ~1 mM)
showed potential areas to gain binding affinity, for example hydrogen bonds to Gln352
sidechain), Leu153 (backbone carbonyl), Leu369 (backbone carbonyl), Phe357 (backbone
(
NH), and lipophilic/πꢀstacking interactions with the indolyl of Trp298. This letter will
disclose how we evolved hydantoin 3, using insights from uracil 2 and sulfonamide 4, to
2
generate novel low MW inhibitors of LpꢀPLA with a promising PK profile.
RESULTS AND DISCUSSION
Hydantoin Optimisation. Hydantoin 3 (clogP = 0.03) did not show any activity in the
biochemical assay, and was only discovered through the Xꢀray crystallographic screen. The
crystal structure shows that the fragment occupies the oxyanion hole of LpꢀPLA
2
with two
hydrogen bonds to the backbone NH of both Leu153 and Phe274 (Figure 2b).
The overall strategy to develop hydantoin 3 into a potent inhibitor is illustrated in Figure 3. It
began by overlaying the crystal structures of hydantoin 3 and sulfonamide 4 bound to Lpꢀ
2
PLA (Figure 1b). These two compounds map out the very bottom of the pocket and a narrow
channel is present between the NH of hydantoin 3 and phenyl ring of sulfonamide 4. Merging
these motifs with a suitable linker should simultaneously increase potency and lipophilicity of
the hydantoin template. The predicted clash between the sulfonamide oxygen and hydantoin
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
carbonyl precluded directly linking the two fragments. We employed a virtual screening (VS)
strategy to investigate a variety of linkers to grow from hydantoin 3 to areas of known
potency (such as regions occupied by uracil 2, sulfonamide 4 and darapladib).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Schematic design process. Moieties within the compounds are coloured according
a
b
to the original compound they are derived from. rhLpꢀPLA
2
PED6 assay. rhLpꢀPLA
2
Thioꢀ
PAF assay.
ACS Paragon Plus Environment

Page 9 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Substructure screening identified ~16,000 hydantoins from inꢀhouse and commercially
available compounds. This library was filtered on molecular weight (MW <300), and docking
2
runs were performed against inꢀhouse LpꢀPLA crystal structures using the Astex proprietary
22
version of Gold software. A total of 33 hits were subsequently screened using Xꢀray
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
crystallography and biochemical assays. This method quickly identified hydantoin 5 with an
2
1
IC = 14 ꢁM . The crystal structure of hydantoin 5 bound in LpꢀPLA is shown in Figure 4.
5
0
2
The hydantoin forms the expected two Hꢀbonds in the oxyanion hole and the ethoxy linker
places the phenyl ring into a deep pocket terminated by residue Phe357. The cyclohexyl ring
sits in a predominantly lipophilic groove surrounded by the sidechains of Leu153, Trp298
and Phe322.
O
O
N
F
N
O
H
5
2
Figure 4. Xꢀray crystal structure of hydantoin 5 bound in LpꢀPLA . The Connolly surface of
the ligand is shown in solid orange and the protein as a gray mesh. Selected residues are
shown for clarity. The hydrogen bond interactions are shown as dotted lines.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
3
Although hydantoin 5 was a reasonably ligand efficient lead with LE = 0.30, we decided to
strip back the cyclohexyl ring and optimize the hydantoin motif (Table 1). Hydantoin 6 was
designed to retain the
Nꢀmethyl (from hydantoin 3) and the ethoxy linker (from hydantoin 5),
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
21
and was only weakly active at IC ~1 mM . Addition of the gemꢀdimethyl in hydantoin 7
50
2
1
gave a ~50ꢀfold increase in potency to IC50 = 20 ꢁM (LE = 0.34) and demonstrated that the
branched nature of the gemꢀdimethyl was important. Incorporation of the 4ꢀfluoro in
hydantoin 8 gave a modest 2ꢀfold increase in affinity and retained the LE at 0.34. Systematic
exploration of the hydantoin ring by removing the Nꢀmethyl (9), and then replacing the amide
carbonyl with methylene (10), did not yield an improvement in LE over hydantoin 8.
However, replacing the NH by methylene to give γꢀlactam 11, resulted in a more ligand
efficient template (LE = 0.38). Ring expansion was also tolerated and yielded δꢀlactam 12
21
with similar potency (IC50 = 10 ꢁM , LE = 0.36).
Table 1. StructureꢀBased Optimization of Hydantoin 5
rhLpꢀPLA
2
rhLpꢀPLA
2
rhPLA2ꢀ
VIIB IC50
(µM)
1
2
3
ThioꢀPafb
Cpd
R
R
R
MW/clogP
306 / 3.3
234 / 1.3
PED6 IC50
a
IC50 (µM) /
LE
c
(µM) / LE
5
6
ꢀF
ꢀH
ꢀH
14 / 0.30
ꢀ
ꢀ
ꢀ
O
N
ꢀH
~1000 / ~0.24
ꢀ
N
O
Me
ACS Paragon Plus Environment

Page 11 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
rhLpꢀPLA
2
rhLpꢀPLA
2
rhPLA2ꢀ
VIIB IC50
(µM)
1
2
3
ThioꢀPafb
Cpd
7
R
R
R
MW/clogP
262 / 2.3
280 / 2.6
PED6 IC50
a
IC50 (µM) /
LE
c
(
µM) / LE
ꢀH
ꢀF
ꢀH
ꢀH
20 / 0.34
ꢀ
ꢀ
ꢀ
ꢀ
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
8
9
10 / 0.34
44 / 0.31
ꢀF
ꢀF
ꢀH
ꢀH
266 / 2.3
252 / 2.6
ꢀ
ꢀ
>100
>100
1
1
0
1
39 / 0.33
11 / 0.38
10 / 0.36
ꢀF
ꢀF
ꢀF
ꢀH
ꢀH
251 / 2.8
265 / 3.4
276 / 2.5
ꢀ
>100
ꢀ
12
13
ꢀ
ꢀCN
3.0 / 0.38
15 / 0.33
>100
a
All assay details are described in Supporting Information. Recombinant human LpꢀPLA
2
2
PED6 assay. The main LpꢀPLA bioassay used a ThioꢀPaf substrate but the assay format was
unsuitable for weakly binding compounds (~50 ꢁM or weaker). A PED6 fluorogenic
substrate was employed for the weaker inhibitors. These two primary bioassays correlated
b
well across many compounds (data not shown). Recombinant human LpꢀPLA ThioꢀPAF
2
c
assay. Recombinant human PLA2ꢀVIIB ThioꢀPAF assay.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The crystal structure of sulfonamide 4 bound in LpꢀPLA
pocket formed by Phe357, Ala355, the sidechain of Leu159 and a water molecule (Figure
c). The cyano also forms a long Hꢀbond to the backbone NH of Phe357. When the Xꢀray
2
indicates that the cyano fills a small
2
crystal structures of sulfonamide 4 or hydantoin 5 bound to LpꢀPLA are overlaid, both the
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
phenyl rings reside in the same position. We therefore substituted a cyano in the equivalent
21
position on the lactam series leading to the design of lactam 13, which had an IC50 = 3.0 ꢁM
and an ITC K = 2.0 ꢁM. This compound was potent enough to be characterized in the main
d
1
1
ThioꢀPaf biochemical assay and had an IC50 = 15 ꢁM (see Table 1, footnote a).
The crystal structure of lactam 13 in complex with LpꢀPLA (Figure 5) gave the expected
2
binding mode and the cyano fills the small sub pocket. The gemꢀdimethyl motif in lactam 13
provides excellent complementarity to the pocket and completely fills the groove between the
sidechains of Leu153, Trp298 and Phe322. The ether oxygen in 13 is important to orientate
the phenyl ring into the pocket lined with Phe110, Ala355 and Phe357, and also forms a close
contact with the sidechain of Gln352. The residues Phe110, Gln352 and Ala355 are
significant because they are replaced in PLA2ꢀVIIB by Tyr, Arg and Thr, respectively.
Lactam 13 shows good selectivity over PLA2ꢀVIIB and this template became our lead for
further optimisation.
ACS Paragon Plus Environment

Page 13 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
Me
Me
N
F
O
CN
1
3
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
Figure 5. Xꢀray crystal structure of lactam 13 in complex with LpꢀPLA . The Connolly
surface of the ligand is shown as an orange mesh. Selected residues are shown for clarity. The
hydrogen bond interactions are shown as dotted lines and a water molecule is shown as a gray
sphere.
Growth of lactam 13. The gemꢀdimethyl in lactam 13 makes contact with the indolyl of
Trp298 and fills a lipophilic space between the sidechains of Leu153 and Phe322. Uracil 2
occupies a similar region and stacks on the indolyl of Trp298. We therefore designed lactam
(
±)ꢀ14 to mimic this interaction by substituting a phenyl ring from the Cꢀ4 position on the
lactam (Table 2), and designed further analogues with an additional methyl (enantiomers (S)ꢀ
5 and (R)ꢀ15).
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. StructureꢀBased Optimization of Lactam 13.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
rhLpꢀPLA
2
rhPLA2ꢀ
VIIB IC
1
ThioꢀPaf
IC50 (µM) /
LE
Cpd
R
MW/clogP
276 / 2.5
324 / 3.3
338 / 3.5
338 / 3.5
a
5
0
b
(µM)
1
3
15 / 0.33
0.86 / 0.34
1.6 / 0.32
1.0 / 0.33
>100
>100
>100
>100
(
±)ꢀ14
c
(
S)ꢀ15
c
(
R)ꢀ15
a
All assay details are described in Supporting Information. Recombinant human LpꢀPLA2
b
c
ThioꢀPAF assay. Recombinant human PLA2ꢀVIIB ThioꢀPAF assay. Chirality of the
stereocentre has been confirmed from the Xꢀray crystal structure of the ligand bound to Lpꢀ
2
PLA .
Lactams (±)ꢀ14, (S)ꢀ15 and (R)ꢀ15 all successfully gave an increase in potency of ~15ꢀfold
and retained LE. Interestingly, the three analogues were approximately equipotent, and in
agreement with available ITC measurements (lactam (S)ꢀ15: K
d
= 0.45 ꢁM and lactam (R)ꢀ
1
5: K = 0.53 ꢁM). Figure 6 shows the superimposed binding modes of (S)ꢀ15 and (R)ꢀ15 in
d
complex with LpꢀPLA . The unsubstituted phenyl rings both form edgeꢀtoꢀface stacking
2
ACS Paragon Plus Environment

Page 15 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
interactions with the indolyl of Trp298, and the multiple equipotent orientations of the phenyl
rings are possible because the groove above Trp298 is wide and lipophilic. Addition of
substituents to the phenyl ring or replacement by heterocycles (such as pyridine), did not
significantly increase the potency (data not shown) and we sought a better solution to gain
potency in this region.
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. Superposition of the Xꢀray crystal structures of lactams (S)ꢀ15 (green) and (R)ꢀ15
2
(orange), each in complex with LpꢀPLA . The Connolly surfaces of the ligands are shown as
a mesh. Selected residues are shown for clarity. The hydrogen bond interactions are shown as
dotted lines.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Alongside aromatic substitutions (lactams (±)ꢀ14, (S)ꢀ15 and (R)ꢀ15), we also tried to grow
from the template using alkyl groups. A series of compounds were synthesized where we
introduced methyl groups to the gemꢀdimethyl moiety in lactam 13 in order to explore the
groove above Trp298 (Table 3). Lactam (±)ꢀ16 gave an 8ꢀfold increase in potency to an IC50
= 2.0 ꢁM. Separation of the enantiomers gave lactams (S)ꢀ16 and (R)ꢀ16. These enantiomers
had a low eudysmic ratio (ratio of these potencies) of 2, and again showed the pocket could
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
accommodate both isomers. Removing the lactam Cꢀ4 methyl to generate the
7 gave an equipotent analogue and slightly increased the LE. This LE was maintained by
further substitution of the ꢀpropyl to give the ꢀbutyl (±)ꢀ18 (IC = 0.98 ꢁM, LE = 0.37). The
crystal structure of (±)ꢀ18 in LpꢀPLA showed the binding mode was consistent with previous
iꢀpropyl in (±)ꢀ
1
i
t
50
2
analogues and the
tꢀbutyl sits in the Leu153, Trp298 and Phe322 groove (Figure 7).
ACS Paragon Plus Environment

Page 17 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. StructureꢀBased Optimization of Lactam 13
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
A:
B:
Plasma IC50
rhPLA2ꢀ
VIIB IC
(µM)
1
rhLpꢀPLA
ThioꢀPaf IC50 (µM) / LE
2
Cpd
R
MW / clogP
b
B / A
5
0
c
a
(
µM) / LE
1
3
276 / 2.5
304 / 3.4
304 / 3.4
304 / 3.4
290 / 2.9
304 / 3.3
345 / 2.1
15 / 0.33
ꢀ
ꢀ
>100
>100
>100
>100
>100
>100
>100
(
±)ꢀ16
2.0 / 0.35
2.3 / 0.35
5.4 / 0.33
1.9 / 0.37
ꢀ
ꢀ
d
(
S)ꢀ16
6.2 / 0.32
13 / 0.30
ꢀ
2.7
2.4
ꢀ
d
(
R)ꢀ16
(
±)ꢀ17
(
±)ꢀ18
0.98 / 0.37
0.62 / 0.34
5.1 / 0.33
5.2
1.4
e
(
±)ꢀ19
0.87 / 0.33
a
All assay details are described in Supporting Information. Recombinant human LpꢀPLA
2
b
c
ThioꢀPAF assay. LpꢀPLA
2
in whole human plasma ThioꢀPAF assay. Recombinant human
d
PLA2ꢀVIIB ThioꢀPAF assay. Chirality of the stereocentre has been confirmed from the Xꢀ
e
ray crystal structure of the ligand bound to LpꢀPLA
2
. Formate salt.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Me
Me
Me
O
N
F
O
CN
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(±)ꢀ18
2
Figure 7. Xꢀray crystal structure of LpꢀPLA in complex with the (S)ꢀenantiomer of (±)ꢀ18.
The Connolly surface of the ligand is shown in solid orange and the protein as a gray mesh.
Selected residues are shown for clarity. The hydrogen bond interactions are shown as dotted
lines.
We felt the interactions between the tꢀbutyl in (±)ꢀ18 with the sidechains of Trp298, Leu153
and Phe322 might be improved by cyclising into a ring. We initially chose a piperidine ring
where the NH group would point towards solvent and allow modulation of physicochemical
properties of the molecule. The basic moiety in lactam (±)ꢀ19 was tolerated and had an IC50
.62 ꢁM.
=
0
Lactams 13–19 were selective over PLA2ꢀVIIB (Table 3) and all had IC50 >100 ꢁM. These
compounds also showed good activity in the plasma assay, and the drop off between the Lpꢀ
PLA
2
assay and the plasma assay remained extremely low (≤5ꢀfold). This represented a
ACS Paragon Plus Environment

Page 19 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
significant improvement over darapladib (710ꢀfold ratio, Table 4). We attributed this result to
the relatively low lipophilicity of this series (clogP ≤3.4), and the chromLogD7.4 of lactam
(±)ꢀ19 was measured to be 1.5. Lactam (±)ꢀ19 became our lead template for further
exploration and assessment of the PK profile of the series.
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
Optimisation of (±)ꢀ19. The secondary amine (±)ꢀ19 had a potency of IC50 = 0.62 ꢁM and
comparable activity in the plasma assay. The amino group of the piperidine is too far away
from the indolyl face of Trp298 to from a productive cationꢀπ interaction and so we explored
alternatives. Exchanging the amine for an ether gave lactam (±)ꢀ20, which is an uncharged
analogue at physiological pH, and showed a 4ꢀfold increase in potency in the LpꢀPLA
2
bioassay. Acylation of the amine (±)ꢀ19 to give amide (±)ꢀ21 did not yield further potency,
but interestingly, methylation to the tertiary amine (±)ꢀ22 improved the potency by 8ꢀfold to
IC = 0.075 ꢁM. The enantiomers where chirally separated to give (S)ꢀ22 and (R)ꢀ22, which
5
0
had IC50 of 0.072 ꢁM and 2.7 µM respectively. The eudysmic ratio is now 38, and this is
significantly higher than the ratio of ~2 for the previous pairs (i.e. (S)ꢀ15/(R)ꢀ15, and (S)ꢀ
16/(R)ꢀ16). The (R)ꢀenantiomer of 22 is the distomer (the less active isomer), and although
the compound is accommodated in the pocket, the interactions with the protein are inferior
with respect to the (S)ꢀenantiomer of 22.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 4. StructureꢀBased Optimization of Lactam (±)ꢀ19
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A:
B:
MW / clogP
/ Chrom
LogD7.4
rhPLA2ꢀ
4
3
rhLpꢀPLA
2
Plasma IC50
Cpd
R
R
b
B / A VIIB IC
5
0
ThioꢀPaf IC50 (µM) / LE
c
a
(µM)
(
µM) / LE
d
345 / 2.1 /
1.5
(
±)ꢀ19
ꢀCN
ꢀCN
ꢀCN
ꢀCN
0.62 / 0.34
0.16 / 0.37
0.25 / 0.32
0.075 / 0.37
0.87 / 0.33
0.26 / 0.36
0.66 / 0.30
0.27 / 0.34
1.4
1.6
2.6
3.6
>100
>100
>100
>10
3
46 / 2.1 /
.7
(±)ꢀ20
3
387 / 1.2 /
.9
(
±)ꢀ21
±)ꢀ22
2
(
359 / 2.6 / ꢀ
359 / 2.6 / ꢀ
e
(
S)ꢀ22
ꢀCN
ꢀCN
ꢀCN
ꢀCN
ꢀCN
0.072 / 0.37
2.7 / 0.29
0.20 / 0.35
7.9 / 0.27
2.8
2.9
0.6
0.3
1.5
8.0
e
359 / 2.6 /
1.7
(
R)ꢀ22
>100
>100
>100
>100
394 / 0.75 /
2.8
(
±)ꢀ23
0.18 / 0.34
0.12 / 0.35
4.3 / 0.27
0.11 / 0.35
0.032 / 0.38
6.3 / 0.26
f
394 / 0.75 /
3.0
(
S)ꢀ23
O
O
Me
394 / 0.75 /
f
(
R)ꢀ23
S
2.8
ACS Paragon Plus Environment

Page 21 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A:
B:
MW / clogP
/ Chrom
LogD7.4
rhPLA2ꢀ
4
3
rhLpꢀPLA
ThioꢀPaf IC
2
Plasma IC50
Cpd
R
R
b
B / A VIIB IC50
(µM) / LE
c
5
0
a
(µM)
(
µM) / LE
3
87 / 1.2 /
.4
(
±)ꢀ24
ꢀF
ꢀF
0.16 / 0.36
0.086 / 0.37
3.9 / 0.28
0.29 / 0.34
0.10 / 0.37
7.9 / 0.27
1.8
1.2
2.0
>100
>100
>100
3
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
f
387 / 1.2 /
3.4
(
S)ꢀ24
(
±)ꢀ25
ꢀCN 380 / 0.49 / ꢀ
3
69 / 1.1 /
.1
(±)ꢀ26
1
ꢀH
0.76 / 0.33
1.5 / 0.32
2.0
>100
0.063
3
667 / 8.3 /
6.3
0.000049 /
0.30
Darapladib
0.035 / 0.22
710
a
All assay details are described in Supporting Information. Recombinant human LpꢀPLA
2
b
c
2
ThioꢀPAF assay. LpꢀPLA in whole human plasma ThioꢀPAF assay. Recombinant human
d
e
PLA2ꢀVIIB ThioꢀPAF assay. Formate salt. Chirality of the stereocentre is inferred from the
f
Xꢀray crystal structure of the eutomer from the racemic mixture bound to LpꢀPLA . Chirality
2
of the stereocentre has been confirmed from the Xꢀray crystal structure of the ligand bound to
LpꢀPLA
2
, and agrees with the crystal structure of the eutomer from the racemic mixture
bound to LpꢀPLA
2
.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Next, we chose to exchange the secondary amine in (±)ꢀ19 for a sulfone ((±)ꢀ23). We
rationalized that the electron withdrawing capability of the sulfone may polarize the adjacent
CH’s which are in contact with the electron rich indolyl of Trp298 and improve the
interaction. Sulfone (±)ꢀ23 showed a small increase in potency (3ꢀfold) in the LpꢀPLA assay,
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
but a larger increase in potency in the plasma assay to IC50 = 110 nM (8ꢀfold). The
enantiomers (S)ꢀ23 and (R)ꢀ23 also showed a separation in potency in the LpꢀPLA
eudysmic ratio of 36) and the eutomer (more active isomer) (S)ꢀ23 had an IC50 = 120 nM.
Lactam (S)ꢀ23 is equipotent to darapladib in the plasma assay but with a vastly improved LE
LE = 0.38 for (S)ꢀ23, and LE = 0.22 for darapladib). Turning our attention back to the
phenyl ring we found that the cyano could be replaced by an equipotent fluorine ((±)ꢀ24 and
S)ꢀ24).
2
assay
(
(
(
Throughout the optimisation, selectivity over PLA2ꢀVIIB was retained (Table 4) and the ratio
between plasma assay and the LpꢀPLA ThioꢀPaf assay remained low (<4). Overall, the LE
2
was maintained at 0.35–0.37 for this set of compounds (in the LpꢀPLA ThioꢀPaf assay).
2
An important feature of the compounds in Table 4 is the quaternary methyl group on the sixꢀ
membered aliphatic ring. Comparison of (±)ꢀ23 with the desꢀmethyl sulfone (±)ꢀ25 shows the
methyl group engenders a 22ꢀfold increase in potency. QM calculations exploring the
torsional landscape around the bond joining the two saturated rings in (±)ꢀ23 and (±)ꢀ25 were
performed, and showed both molecules possessed the same lowest energy conformation that
is also observed in the bioactive conformation. Thus, the potency gain is presumably through
significantly improved ligandꢀprotein complimentary in the Leu153, Trp298 and Phe322
groove. This methyl was only discovered by moving away from aromaticity and careful
stepwise optimisation of the gemꢀdimethyl moiety (13) into the tꢀbutyl ((±)ꢀ18). By way of
ACS Paragon Plus Environment

Page 23 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
comparison, the cyano contributes a useful but less spectacular 4ꢀfold increase in potency
(comparing (±)ꢀ23 and (±)ꢀ26). Contributions of different parts of sulfone (±)ꢀ23 to potency
24
are shown in Figure 8 together with their group efficiency (GE). Each moiety analysed has
a good contribution to the potency, with the least group efficient moiety being the sulfone
which occupies a region that is more solvent exposed.
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
Figure 8. Group efficiency (GE = kcal mol per nonꢀhydrogen atom) schematic for
compound (±)ꢀ23 based on pairwise comparisons (see Supporting Information). GE ≥0.3 is
2
4
considered desirable.
2
Figure 9 shows the crystal structure of (S)ꢀ23 bound in LpꢀPLA with the expected binding
mode. Figure 10a superimposes the Xꢀray crystal structures of (S)ꢀ23 in complex LpꢀPLA2
with compounds 2–4, and Figure 10b superimposes the Xꢀray crystal structures of (S)ꢀ23 and
darapladib. Together these figures show that compounds 2–4, (S)ꢀ23 and darapladib (1) have
common pharmacophore elements, such as the carbonyl that occupies the oxyanion hole and
a phenyl ring that resides in the deep lipophilic pocket beside Phe357. Additionally, sulfone
(S)ꢀ23 has efficiently utilized the space along the bottom of the pocket between Trp298 and
Phe357 that was contacted by compounds 2–4; in particular, we exploited the groove above
Trp298 (which was not utilized by darapladib) to improve potency and simultaneously
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
modulate the physicochemical properties of the ligand. Finally, we avoided using areas that
were less likely to yield potency such as the region occupied by the (2ꢀ
aminoethyl)diethylamine moiety in darapladib.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
Me
O
S
N
O
F
O
CN
(S)ꢀ23
2
Figure 9. Xꢀray crystal structure of LpꢀPLA in complex with lactam (S)ꢀ23. The Connolly
surface of the ligand is shown in solid orange and the protein as a gray mesh. Selected
residues are shown for clarity. Hydrogen bond interactions are represented as dotted lines.
(a)
(b)
Figure 10. (a) Superposition of the Xꢀray crystal structures of uracil 2 (cyan), hydantoin 3
(magenta) and sulfonamide 4 (green) (only showing the Connolly surface as a mesh), and
ACS Paragon Plus Environment

Page 25 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(S)ꢀ23 (orange), bound in LpꢀPLA
orange) and darapladib (purple) bound in LpꢀPLA
clarity.
2
. (b) Superposition of the crystal structures of (S)ꢀ23
(
2
. Only selected residues are shown for
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
PK Discussion. Lactams (±)ꢀ19, (±)ꢀ22, (S)ꢀ23 and (S)ꢀ24 are all low MW (<400 MW)
compounds, with solubilities >240 µM in the CLND assay (Table 5); in particular (S)ꢀ23 was
measured by quantitative NMR to be fully soluble at 3.5 mM. The plasma protein binding
(PPB) in both rat and dog are extremely low for these compounds, and helps to explain why
the drop off between the LpꢀPLA assay and the plasma assay is so small (Table 4). The
2
series does not show significant inhibition of the CYP450’s 1A2, 3A4, 2C9, 2C19 and 2D6,
as exemplified by CYP450 3A4 data in Table 5, where the compounds have an IC50 >50 µM.
However, there is a difference between the compounds in the permeability assay where the
secondary amine (±)ꢀ19 is poor, but the tertiary amine (±)ꢀ22 and sulfones (S)ꢀ23 and (S)ꢀ24
perform better. Assessment of the stability of (±)ꢀ23 in rat blood showed the compound was
100% stable.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 5: In Vitro Profile for Key Compounds
CYP450
3A4
CLND
Solubility PPB PPB
Rat Dog
Pampa
Permeability
(nm/s)
MW / clogP /
ChromLogD7.4
Cpd
Structure
a
a
IC50b
(µM)
(%) (%)
(
µM)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
c
e
(
±)ꢀ19
345 / 2.1 / 1.5
242
19
ꢀ
> 50
< 3
d
c
e
(
±)ꢀ22
S)ꢀ23
S)ꢀ24
359 / 2.6 / 1.7
> 396
> 509
29
44
25
> 50
>50
>50
210
e
(
(
394 / 0.75 / 3.0
387 / 1.2 / 3.4
20
36
e
> 450
8
30
ꢀ
ꢀ
ꢀ
170
f
1
Darapladib
667 / 8.3 / 6.3
ꢀ
c
230
a
b
d
3
–6 replicates tested at 10 µM. Inhibition of CYP3A4, ≥2 replicates. HCl salt. (R)ꢀ
e
f
enantiomer tested. Measured at pH = 7.4. Measured at pH = 7.05.
ACS Paragon Plus Environment

Page 27 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The secondary amine (±)ꢀ19 is a low in vivo clearance compound in rat (Table 6) with a
reasonable volume of distribution (Vss) and halfꢀlife (T1/2); this is in part attributed to the
basic centre in the molecule. However, amine (±)ꢀ19 has poor permeability in the PAMPA
assay which translates into an extremely low bioavailability. In contrast, the tertiary amine
(±)ꢀ22 is a moderate clearance compound (Cl = 24 mL/min/kg), and its superior permeability
leads to good oral bioavailability (%F = 52%).
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
Sulfone (S)ꢀ23 has a low intrinsic clearance of 1.1 mL/min/g in rat liver microsomes.
Unfortunately, this does not translate in vivo where it shows a much higher clearance of 67
mL/min/kg (Table 6). However, modulation of the rat clearance is possible as demonstrated
by sulfone (±)ꢀ24.
When assessed in dog, the PK profile of (±)ꢀ22 and (S)ꢀ23 look similar, and are both low
clearance compounds (8.6 and 11 mL/min/kg, respectively). (±)ꢀ22 shows good oral
bioavailability (%F = 76%), although has a shorter halfꢀlife than desired. Overall, the series
shows a promising profile and provides a suitable starting point for further lead optimization.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 6: In Vivo Profile of Key Compounds in Rat and Dog
Rat
Dog
T1/2
Cl
mL/min
%F
oral
(%)
Cl
(mL/min
/kg)
%F
oral
(%)
Cpd
Structure
T1/2
Vss
Vss
(hr) (L/kg)
(
(hr) (L/kg)
/
kg)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
a
(
±)ꢀ19
±)ꢀ22
15
2.4
2.0
3.2
1.2
< 5
52
ꢀ
ꢀ
8.6
11
ꢀ
ꢀ
ꢀ
ꢀ
76
ꢀ
(
24
67
22
1.8
3.2
1.3
ꢀ
2.5
1.3
ꢀ
(S)ꢀ23
0.34
(
±)ꢀ24
0.35 0.63
47
ꢀ
a
In rat, HCl salt.
Synthesis. Synthesis of sulfone (±)ꢀ23 is laid out in Schemes 1 and 2, and followed the
convergent strategy which involved a late stage alkylation of 2ꢀpyrroldinone (±)ꢀ30 with
methanesulfonate 32.
Construction of 2ꢀpyrroldinone (±)ꢀ30 was central to this synthesis, and was generated in four
steps from commercially available thianeꢀ4ꢀcarbaldehyde (Scheme 1). Thianeꢀ4ꢀcarbaldehyde
was methylated with MeI and NaH to generate the quaternary carbon center in 27. This was
followed by a Horner−Wadsworth−Emmons reaction with triethyl phosphonoacetate and tꢀ
BuOK to give (E)ꢀunsaturated ester 28. Nitromethane was reacted with the conjugated ester
through a Michael addition to generate γꢀnitro ester (±)ꢀ29. The nitro group was then
selectively reduced with NiCl
2 4
and NaBH , and in situ cyclisation of the resulting γꢀamino
ester directly afforded 2ꢀpyrrolidone (±)ꢀ30.
ACS Paragon Plus Environment

Page 29 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Scheme 1. Synthesis of lactam Intermediate (±)ꢀ31
2
7
28
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
(
±)ꢀ29
(±)ꢀ30
a
Reagents and conditions: (a) MeI, tꢀBuOK, THF, 0 °C, 2 h, 94%; (b) Triethyl
phosphonoacetate, tꢀBuOK, THF, 0 °C, 3 h, 96%; (c) Nitromethane, 1,1,3,3ꢀ
tetramethylguanidine, 100 °C (microwave), 4 h, 43%; (d) NiCl ·6H O, NaBH , MeOH, 0 °C,
.5 h, 65%.
2
2
4
1
Methanesulfonate 32 was prepared in two steps from the commercially available 2ꢀfluoroꢀ5ꢀ
hydroxybenzonitrile (Scheme 2). 2ꢀFluoroꢀ5ꢀhydroxybenzonitrile was coupled with ethylene
2 3
carbonate in the presence of K CO to give alcohol 31 in a high yield. This was treated with
methanesulfonyl chloride and triethylamine to give methanesulfonate 32. Alkylation of 2ꢀ
pyrrolidine (±)ꢀ30 with 32 proceeded smoothly at elevated temperatures to afford sulfide (±)ꢀ
3
3 in 44% yield. Finally, oxidation of (±)ꢀ33 with mCPBA gave sulfone (±)ꢀ23.
ACS Paragon Plus Environment