From virtual to clinical: The discovery of PGN-1531, a novel antagonist of the prostanoid EP4 receptor

Article abstract of DOI:10.1016/j.bmcl.2014.02.068

In this Letter, we present the results of a hit-finding and lead optimization programme against the EP₄ receptor (EP₄R). In a short time period, we were able to discover five structurally diverse series of hit compounds using a combi

Full text of DOI:10.1016/j.bmcl.2014.02.068

Bioorganic & Medicinal Chemistry Letters 24 (2014) 2212–2221
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
From virtual to clinical: The discovery of PGN-1531, a novel
antagonist of the prostanoid EP4 receptor
a
a,
Jon Sutton a, , David E. Clark , Christopher Higgs , Marcel J. de Groot a, Neil V. Harris a, Andrea Taylor a,
⇑
Peter M. Lockey a, Karen Maubach c, Amanda Woodrooffe b,à, Richard J. Davis b,§, Robert A. Coleman b,–
,
Kenneth L. Clark b,k
a Argenta, 8/9 Spire Green Centre, Flex Meadow, Harlow, Essex CM19 5TR, United Kingdom
b Pharmagene plc, 2 Orchard Road, Royston, Hertfordshire SG8 5HD, United Kingdom
c BTG International Ltd, 5 Fleet Place, London EC4M 7RD, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
In this Letter, we present the results of a hit-finding and lead optimization programme against the EP4
receptor (EP4R). In a short time period, we were able to discover five structurally diverse series of hit
compounds using a combination of virtual screening methods. The most favoured hit, compound 6, was
demonstrated to be a competitive antagonist of the EP4R. Compound 73 was identified following
several rounds of optimization, which centred on improving both the primary EP4R affinity and selectivity
against the related EP2R as well as the aqueous solubility. This work culminated in the preparation of
PGN-1531, the sodium salt of 73, which showed a marked improvement in solubility (>10 mg/mL).
PGN-1531 is a potent and selective antagonist at EP4Rs in vitro and in vivo, with the potential to alleviate
the symptoms of migraine that result from cerebral vasodilatation.
Received 8 January 2014
Revised 21 February 2014
Accepted 24 February 2014
Available online 12 March 2014
Keywords:
EP4 receptor antagonist
Hit-finding
Lead optimization
Migraine
Ó 2014 Elsevier Ltd. All rights reserved.
Cerebral vasodilatation
The prostanoid receptor subfamily comprises nine members:
DP, EP1–4, FP, IP, TP and the more recently identified CRTh2.1 The
EP receptors (for which the endogenous ligand is PGE2) have been
cloned and are distinct at both a molecular and pharmacological
level.2 EP4 receptor (EP4R) antagonists have been shown to have
potential in the treatment of inflammatory pain3 and, in particular,
in the treatment of primary headache disorders, which include
migraines, and secondary headache disorders, such as drug-
induced headaches.4 Dilation of the cerebral vasculature and the
subsequent activation of pain-stimulating, perivascular trigeminal
sensory afferent nerves is recognised to play an important role in
the pathophysiology of migraine. A sterile inflammatory response,
associated with activation of cyclooxygenase and the generation of
PGE2, is also implicated in the pathophysiology of migraine. PGE2
levels have been shown to be raised during migraine attacks and
PGE2 contributes to the pain of migraine by directly dilating
cerebral arteries and by stimulating the release of vasoactive/
proinflammatory peptides from the trigeminal nerves. These
effects of PGE2 are mediated in whole or in part by EP4Rs. Thus,
by binding to and preventing the stimulation of EP4Rs, EP4R
antagonists may be used to treat the pain associated with
migraine.
In this Letter, we describe the virtual screening and medicinal
chemistry optimization that led to the discovery of PGN-1531, a
novel, potent and selective EP4R antagonist.
The virtual screening campaign comprised two rounds of
compound selection and acquisition over approximately seven
months. Following a review of the relatively sparse literature
available at the time concerning EP4R antagonists, four compounds
were chosen as the basis for virtual screening (compounds 1–4,
Table 1).
⇑
Corresponding author. Tel.: +44 (0) 1279 645645; fax: +44 (0) 1279 645 646.
E-mail address: jon.sutton@crl.com (J. Sutton).
Current address: Schrödinger, Inc., 120 West 45th Street, 17th Floor, New York,
NY 10036, USA.
à
Current address: Stemgent/Asterand, 2A Orchard Road, Royston, Hertfordshire
SG8 5HD, United Kingdom.
§
Current address: Investment Division, The Wellcome Trust, 215 Euston Road,
Various virtual screening approaches were applied including
2-D and 3-D ligand-based searches together with docking into an
EP4R homology model. Following assessment of compounds
selected from these searches by an experienced medicinal chemist,
a total of 516 compounds was ordered from commercial suppliers.
London NW1 2BE, United Kingdom.
–
Current address: 27 Wodehouse Terrace, Falmouth, Cornwall TR11 3EN, United
Kingdom.
k
Current address: Department of Asthma and Allergy Biology, Respiratory and
Inflammation CEDD, GlaxoSmithKline, Gunnels Wood Road, Stevenage, Hertfordshire
SG1 2NY, United Kingdom.
http://dx.doi.org/10.1016/j.bmcl.2014.02.068
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
2213
Table 1
Table 3
Chemical structures of the compounds used as queries in the first round of database
searches
Summary of EP4R IC50 values obtained in primary screening
IC50 range (
lM)
Number of compounds in IC50 range
Compound
Structure
Refs.
IC50 > 25
5
15
8
10 < IC50 6 25
5 < IC50 6 10
IC50 6 5
O
O
O
O
8
OH
1
5,6
N
Based on the accumulated screening data, five series of com-
pounds were identified as having potential for hit-to-lead explora-
tion. Each series represented a tractable starting point for further
optimisation, possessing low
lM potency, favourable calculated
O
physicochemical properties, generally good LE values and some
early SAR data. The structure of the head compound of each series
is shown in Table 4 together with associated biological data ob-
tained from the screening sample.
From the viewpoint of novelty, the furan ether series, typified
by compound 6, was particularly interesting and so the structure
and activity of this compound were confirmed from a larger batch
of material obtained from the supplier (giving an IC50 value of
2
3
7
8
OH
OH
O
O
N
1.0 lM). These confirmatory results, together with some early
SAR (Table 5), gave sufficient confidence in this series for it to form
the basis of a hit-to-lead chemistry programme. Overall, the early
SAR indicated the requirement for an acidic moiety (see 12) and
a 1–4 biaryl ring system. Compound 6 was found to be selective
CF3
H
O
S
S
N
N
4
9
over both EP1R and EP3R but not EP2R (95% inhibition at 10 lM),
O
O
N
O
therefore the attainment of greater EP4R/EP2R selectivity was one
of the goals of lead optimisation as well as boosting primary po-
tency against the EP4R.
N
Encouragingly, a Scatchard kinetic analysis of compound 6 con-
firmed it as a competitive inhibitor of binding at the EP4R (Fig. 1).
Additionally, the compound behaved as an antagonist in a func-
tional cell-based assay12 (pKb = 5.78 0.39).
The data in Table 5 highlight the requirement for the carboxylic
acid group, which can be rationalised by our EP4R homology
model13,14 (Fig. 2) in terms of the need for a strong interaction with
Arg255 in the EP4R, and also a very specific arrangement of an
approximately planar aromatic ring system at an appropriate dis-
tance from the carboxylate. A further hydrogen-bonding interac-
tion between the ether linker in compound 6 and Tyr184 can also
be observed.
Once the compounds from the first round of virtual screening
had been screened, a second round of searches was carried out.
The purpose of these searches was to exploit the emerging SAR
information acquired from the screening of the first round of com-
pounds and also to take advantage of new information in the liter-
ature concerning a previously undisclosed EP4R antagonist from
Ono (Table 2). As a result of these searches, a further 363 com-
pounds were ordered.
In all, the two rounds of compound selection led to a total of
879 compounds being ordered. In the final analysis, 762 of the
879 compounds (i.e., about 87%) were acquired from 12 suppliers.
Of the 762 compounds, six proved insufficiently soluble in
DMSO to be screened, leaving 756 to be tested in the primary
assay.11 These compounds were screened at a single concentration
Early lead optimization of the furan ether hit 6 centred on var-
iation of the furan core and the use of solid-phase library chemistry
to the probe the SAR of the biaryl moiety, Figure 3.
The emerging early SAR shown in Table 6 indicated that simple
variation of the 5-methyl substituent was tolerated (compounds
15–17). However, chain extension of the carboxylic acid from the
furan core in compounds 18 and 19 led to some reduction in
EP4R affinity, although replacing the acid with a tetrazole bioiso-
stere led to a fivefold improvement in affinity (Ki = 80 nM). Varia-
tion of the furan-ether linker showed a more marked SAR, with
the amide 21 being inactive—although both the homologated ether
and the amine-linked compounds 20 and 22, respectively, were
found to have modest affinity compared to the virtual screening
hit 6. After several rounds of modifications to the original hit com-
pound, we identified a preference for the original hit scaffold, sub-
stituents and linking groups.
With this information in hand, we proceeded to examine varia-
tion of the biaryl system. The solid-supported chemistry utilized to
expand the SAR around the biaryl moiety of compound 6 is shown
in Scheme 1. The commercial furan alcohol 23 was TBDMS-
protected and the core carboxylic acid installed using lithiation
followed by addition of carbon dioxide to give compound 24. The
attachment to 2-chlorotrityl chloride resin was then performed
and the protecting group removed to give the alcohol 25.
of 10
by LC/MS. For 36 compounds showing a sufficient level of inhibi-
tion at 10 M (typically >ca. 40%), IC50 values were determined
using six-point concentration/inhibition curves. The breakdown
of these results is shown in Table 3.
lM. All compounds were also checked for purity and identity
l
Table 2
New antagonist structure
Compound
Structure
Ref.
10
O
OH
5
NH
O

2214
J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
Table 4
Heads of the series identified as having potential for hit-to-lead exploration
Series
Compound
Head of Series
EP4R IC50
(lM)
*LE/LLE
O
O
O
O
OH
Furan ethers
6
2.4
0.35/1.66
N
S
N
N
Pyrazoles
7
8
1.2
0.33/0.16
0.32/2.36
OH
O
N
Phthalimides
0.46
O
O
O
OH
O
ortho-Amides
9
2.6
1.0
0.33/0.94
0.26/2.17
NH OH
O
H
N
H
N
S
para-Amides
10
O
O
O
O
OH
O
*
LE = ligand efficiency index (1.4 ꢀ (ꢁlog IC50)/number of heavy atoms); LLE = ligand lipophilicity index (EP4RpIC50ꢁc Log P).
Table 5
Initial SAR for furan ether series
Compound
Structure
EP4R IC50 (lM)
O
O
O
OH
6
1
O
O
11
12
4.6
O
OH
O
O
O
O
O
15% inh. @ 10 lM
O
O
(X = O) 7% inh. @ 10 lM
(X = C(Me)2) 23% inh. @ 10 lM
13
14
O
OH
X
O
O
10
O
OH

J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
2215
Table 6
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
Linker SAR and core template variation in furan ether series
R2
O
Control
Compound 6
O
X
OH
n
Compound
R2
X
n
EP4R
Ki (
l
M)11
6
Me
H
iPr
CF3
Me
Me
0
0
0
0
1
2
0.42
2
3
15
16
17
18
19
–OCH2–
1.3
61% inhib. @ 10
63% inhib. @ 10
l
M
lM
20
21
22
Me
Me
Me
–OCH2CH2–
–CONH–
–NH–
0
0
0
3.8
>30
1.4
The array compounds 27 were obtained following cleavage from
the resin using TFA/DCM.
Screening data from the biaryl compounds 28–48 (Table 7)
showed that there was a clear preference for the more lipophilic
and/or electron-donating substituents in biphenyl analogues such
as the para-methoxy and para-di-fluoromethoxy compounds (en-
tries 30 and 36, respectively). It is possible that steric effects are
responsible for the difference in activity of the 3,4-methylenedioxy
compound 34 compared to the 3,4-dimethoxy analogue 33, which
shows a fivefold drop in EP4R affinity. A further examination of the
data indicates that compounds in which the distal phenyl ring is
substituted with polar groups are poorly tolerated (entries 40–
42) as well as compounds where the phenyl ring is replaced with
various polar heterocycles (entries 43–47). This presumably re-
flects the highly lipophilic nature of the binding site in this region
of the EP4R. A notable exception to this observation is the para-
methoxy pyridyl analogue 48, which has improved affinity result-
ing from a favourable interaction of the para-methoxy substituent.
A SiteMap15 binding site analysis is shown in Figure 4 for se-
lected compounds 30, 33, 34 and 39, which indicates that the bind-
ing region in the EP4R is highly hydrophobic, with the more potent
analogues placing methyl groups in a hydrophobic area linked by
an oxygen atom in an acceptor area, whereas the various less po-
tent compounds place polar groups in hydrophobic areas. The high
affinity exhibited by the phenyl methyl ketone 39 can be rational-
ized as the compound is able to pick up additional hydrophobic
(methyl group) and hydrogen-bond acceptor (ketone oxygen)
binding interactions.
0.25
Specific /Molar
0.45
0.20
0.30
0.35
0.00 0.05
0.10
0.15
0.40
Figure 1. Scatchard analysis generated from [3H]-PGE2 saturation binding data, in
the absence and presence of
Bmax = 0.4 0.03 nM).
1
lM
of compound
6
(Kd = 1.7 0.2 nM,
Figure 2. Binding mode of compound 6 (grey) showing binding to key residues in
the EP4R homology model.
After the first rounds of early optimisation, the major goals
were to increase EP4R affinity further and to improve EP4R/EP2R
selectivity, which remained poor for most carboxylic acid-contain-
ing analogues (data not shown). To do this, we took advantage of a
recently released structure (at the time) of a potent competitor
compound from Ono10 (5). The Ono compound was found to be
very potent and selective in our hands, (EP4R Ki = 7 nM; >600-fold
selectivity over the related EP2R). A simple FlexS16 overlay of 5
(both R and S enantiomers are shown) with a proposed acylsulf-
onamide target 49, derived from the original hit compound 6, indi-
cated that an acyl sulfonamide group may be able to pick up some
additional interactions similar to those present in the highly EP4R
selective Ono compound (Fig. 5).
Extend the furan
to ether link
Vary ring substituent
at C5
O
Extend the acid with a carbon spacer
replace with acid bioisosteres
/
O
O
HO
6
Vary biaryl moiety
Replace ether link with
amide / amine / thioether link
Gratifyingly, the first acylsulfonamide compound 49 was found
to have greatly improved affinity at EP4R and, very encouragingly,
showed a marked improvement in EP4R/EP2R selectivity profile
compared to the parent carboxylic acid 6 (Table 8). The LLE data
indicate that compound 49 has a comparable value to the Ono
compound 5.
Figure 3. Areas of modification of original hit compound 6.
A Mitsunobu coupling with 4-iodo-phenol gave the ether 26, which
was subsequently used to prepare an array of biaryl analogues
using Suzuki–Miyaura coupling reactions of selected boronic acids.

2216
J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
a, b
c, d
O
O
O
OH
O
TBDMSO
O
HO
HO
O
23
24
25
O
O
e
O
O
f, g
O
O
O
OH
I
R1
26
27
Scheme 1. Synthesis of phenoxymethyl-furan-2-carboxylic acid array: (a) TBDMS-Cl, Imidazole, DMF, RT, 2 h (68%); (b) (i) nBu-Li (2.5 M in hexanes), THF, ꢁ78 °C, 30 min, (ii)
CO2(g) (30%); (c) 2-Chlorotrityl chloride resin (1.3 mmol/g), DCM; (d) TBAF, RT, 18 h; (e) 4-Iodo-phenol, triphenyl phosphine, DIAD, THF, RT, 16 h; (f) R1-B(OH)2, PdCl2(dppf)2,
Cs2CO3, DMF, 100 °C, 1 h (48–100%); (g) 5% TFA, DCM, RT, 20 min (65–85%).
Table 7
Initial SAR from the furan ether array
O
O
O
OH
27
R1
Compound
R1
EP4R
Compound
R1
EP4R
Ki (l
M)11
Ki (l
M)11
~
N
O
6
0.42
38
39
~
1.5
~
HO
HO
28
10
3
0.15
~
H
N
~
~
29
30
40
41
>>10
O
O
~
O
~
0.09
>>10
N
H
O
O
31
32
33
34
35
0.74
1.6
2.8
0.6
2.4
42
43
44
45
46
>>10
>>10
>>10
>10
S
~
O
~
O
N
N
S
~
~
O
N
O
N
~
~
~
O
O
O
N
O
~
F
F
~
~
F
>>10
~
~
O
F
F
36
37
0.02
1
47
48
>>10
1.3
N
O
~
Cl
~
O
N
With these encouraging data in hand for compound 49, we pro-
ceeded to incorporate elements identified from the early lead opti-
misation work into new acylsulfonamide analogues (Table 9,
compound 50). The results from acylsulfonamide compounds
prepared with variation to the distal phenyl ring culminated in
some very potent and selective examples, with the di-fluoromethoxy
acylsulfonamide 52 (Ki = 1.3 nM) exhibiting
a further 20-fold
improvement in EP4R affinity compared to compound 49 (and

J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
2217
R1
*
O
O
O
*
*
*
O
O
O
=
O
O
OH
O
R1
Ki = 0.09 µM
0.6 µM
2.8 µM
0.15 µM
Figure 4. SiteMap binding site profile showing areas preferring hydrophobic (yellow) and hydrogen-bond acceptor (blue) groups.
O
N
H
HO
O
5 Ono Compound
O
O
O
HN
S
O
O
49
x
Proposed follow-up compound
Figure 5. Superimposed FlexS overlay of (R/S)-Ono compound 5 with acyl sulfonamide derivative of lead carboxylic acid, 49.
Table 8
The data in Table 9 highlight the improvement in EP4R affinity
possible when moving from a carboxylic acid group in compound
6 to the acylsulfonamide in compounds typified by example 50,
which can be rationalised in our EP4R homology model using the
related analogue 51 (Fig. 6). As well as compound 51 exhibiting
similar hydrogen bonds to the those in carboxylic acid 6, between
Arg255 and the acylsulfonamide carbonyl group and between the
Comparative selectivity data for Ono compound 5 with furan ether hit 6 and acyl
sulfonamide derivative 49
Compound
EP4R
Ki (nM)
Fold-selectivity
LE/LLE
EP2R
EP3R
5 (Ono)
7
420
25
630
13
400
NT
1000
1000
0.36/2.46
0.35/1.66
0.31/2.57
6
49
ether linker and Tyr184 (see Fig. 2), an additional p–cation interac-
tion with Lys272 and S@O interaction of the acyl sulfonamide moi-
ety with Arg255 can also be seen. Moreover, the para-methoxy
moiety in the biaryl group in 51 is in close proximity to Ser249
and Ser192 and could potentially form a hydrogen bond with
either of these residues if a bridging water molecule (not in model)
were present. The exquisite EP4R potency and EP4R/EP2R selectivity
of the acylsulfonamide compounds, compared to their parent car-
boxylic acids, can be understood by a comparison of residue differ-
ences in the binding region of these receptors (key EP2R residues
are shown in parentheses in Fig. 6). Clearly, the extra binding inter-
actions possible for acylsulfonamide compounds in the EP4R with
NT = Not tested.
some 300-fold improvement over the original carboxylic lead 6). A
further series of analogues was prepared with variation to the acyl-
sulfonamide substituent in compounds bearing the more potent
para-methoxy biphenyl motif (compound 55). The SAR of selected
examples 56–61 indicated that, with the exception of the methyl
analogue 56, a wide degree of substitution was tolerated without
being detrimental to EP4R affinity, and all potent compounds
showed excellent EP4R/EP2R selectivity profiles.

2218
J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
Table 9
Selected acylsulfonamide compounds
O
O
O
O
R3
O
O
O
HN
HN
S
S
O
O
O
R1
50
55
O
Compound
R1
EP4R
EP2R
Compound
R3
EP4R
EP2R
Ki (nM)
Selectivity
Ki (nM)
Selectivity
~
6 (CO2H)
420
13
56
Me
218
NT
~
49
51
25
5
>1000
1000
57
58
nPr
5.4
11
>500
650
~
~
~
O
HO
F
F
N
O
52
53
1.3
3890
59
~
~
1.3
>3000
~
O
O
22
20
>1000
>1000
60
61
3.6
3
1500
S
N
O
54
>1000
~
~
NT = not tested.
both Lys272, Arg255 (a bidentate interaction is now possible) and
Ser192/249, which are not possible in EP2R, account for these
marked differences.
the ortho-tolyl acylsulfonamide substituent to improve primary
EP4R affinity; the solubility of compound 73 was deemed
acceptable for compound progression.
The solubility data for selected acylsulfonamide compounds are
shown in Table 10. These data for compounds with variation in the
distal aromatic ring 50 show that, although very potent EP4R
antagonists can be prepared (see Table 10, compounds 51 and
52), the high lipophilicity has resulted in very poor aqueous solu-
bility. Some improvements in solubility can be seen for the meth-
oxy pyridine-containing analogue 53—although a compromise on
potency has been made. In our experience, further disruption of
the planarity of biaryl rings by incorporation of a suitable ortho
substituent has led to significant improvements in solubility. How-
ever, in this case, these were not realized (see compounds 62 and
63). The solubility data for selected acylsulfonamide compounds
having variation in the acylsulfonamide substituent 55 again
indicated that highly potent examples were possible (dimethyl-
The aqueous solubility of compound 73 was further enhanced
by formation of the sodium salt of the acylsulfonamide, which
led to a marked improvement in solubility (>10 mg/mL). The so-
dium salt of compound 73 was designated PGN1531.
Profiling of PGN1531 was undertaken to confirm that it was a
functional antagonist at the EP4R. To this end, HEK-293 cells were
prepared which stably expressed the human EP4R. PGN1531
showed a concentration-dependent competitive antagonism of
cAMP accumulation mediated by the action of the natural ligand,
PGE2, on the EP4R with a pKb of 7.6.12 Further investigation of
PGN-1531 was undertaken with native EP4Rs in human cerebral
and meningeal arteries. PGN-1531 was found to be a competitive
antagonist of the PGE2-induced vasodilatation of human middle
cerebral (pKB 7.8) and meningeal (pKB 7.6) arteries in vitro, but
had no effect on responses induced by PGE2 on coronary, pulmon-
isoxazole 59 Ki = 1.3 nM; aqueous solubility = 0.2 lg/mL) but
usually at the expense of aqueous solubility. The incorporation of
polar elements failed to improve matters and, in many cases, led
to a reduction in EP4R affinity. In order to identify candidate-
quality molecules, an acceptable compromise had to be found.
In an attempt to boost the aqueous solubility of our potent acy-
lsulphonamide compounds, selected amino-linked analogues were
prepared (Table 11, entries 68, 69 and 72). Compound 68 exhibited
encouraging solubility for an unsubstituted biphenyl analogue and
a sevenfold increase in potency was achieved by optimising the
acyl sulfonamide substituent to give example 69. Ultimately, com-
pound 72 was prepared which was found to have sub-nanomolar
EP4R affinity (0.3 nM) and modest solubility (20 lg/mg). However,
due to concerns over the perceived risk of releasing a biphenyl-
amine moiety either chemically or metabolically in vivo, the
amino-linked compounds were not profiled further. Finally, a
compromise was found with compound 73, which retained the
ether linker, removing any biphenylamine risk, and incorporated
Figure 6. Binding mode of compound 51 (green) showing binding to key residues
in the EP4R homology model—corresponding residues in the EP2R homology model
are shown in white in parentheses.

J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
2219
Table 10
Solubility data for selected acylsulfonamide compounds
O
O
O
O
R3
O
O
O
HN
HN
S
S
O
O
O
R1
50
55
O
Compound
R1
EP4R
Solubilitya 24 h (
lg/mL)
Compound
R3
EP4R
Solubilitya 24 h (
lg/mL)
Ki (nM)
Ki (nM)
~
~
49
51
25
5
2
58
59
HO
11
20
N
O
11
~
1.3
0.2
~
O
F
H2N
F
~
52
1.3
<0.6
64
100
<0.6
O
~
~
~
O
O
53
62
22
21
30
20
65
66
Me2N
115
15
10
5
N
~
N
O
O
N
63
85
8
67
1200
20
~
~
a
Determinations carried out at RT in 0.1 M Phosphate Buffer/0.15 M KCl adjusted to pH 7.4. Sample determinations carried out in triplicate.
Table 11
Optimising solubility and potency
R3
O
N
=
O
O
~
~
~
~
R3
X
HN
S
A
B
C
D
O
O
R1
Compound
R1
Linker (X)
R3
EP4R
Solubilitya 24 h (
lg/mL)
Ki (nM)
68
69
A
B
14
2
300
20
~
–NH–
–O–
O
~
70
D
5
7
F
71
72
–O–
C
B
0.6
0.3
<0.6
20
F
O
~
–NH–
~
O
73
74
–O–
–O–
B
C
3
10
90
N
13
75
–O–
C
24
200
~
O
a
Determinations carried out at RT in 0.1 M Phosphate Buffer/0.15 M KCl adjusted to pH 7.4. Sample determinations carried out in triplicate.
ary or renal arteries in vitro. In further tests, the compound showed
no appreciable affinity at a wide range of other receptors (includ-
ing other prostanoid receptors DP, FP, IP or TP), ion channels, trans-
porters and enzymes (pKi <5). PGN-1531 (1–10 mg kgꢁ1 i.v.) was
also shown to cause a concentration-dependent antagonism of
the PGE2-induced increase in canine carotid blood flow in vivo,
thus demonstrating that an EP4R antagonist is able to reverse the
vasodilatory effect of PGE2 in an in vivo setting.12
The compound profile for PGN-1531 is given in Table 12. The
selection of this compound for clinical development was made fol-
lowing positive efficacy readouts in functional assays, very high
EP4R/EP2R selectivity (high safety margin), excellent bioavailability

2220
J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
Table 12
Compound profile for PGN-1531
Property
PGN-1531
O
O
O
S
N
-
O
H
O
Na+
N
O
EP4R Ki (nM)
EP4R pKB (M) cAMP
3
7.6
EP4R pKB (M) hum. Cerebral artery
EP2R/EP3R selectivity
7.8
>2500-fold
c Log P/c log D/PSA (LE/LLE)
Solubility (g/mL)
4.1/2.2/114 (0.32/4.12)
>10
Caco-2 (nm/sec)
145
Human/Rat Int. Clearance (ml/min/kg)
Oral half-life(h)
6.5 1.4/2.7 1.1
2.3
84
Rat Bioavailability (%)
CYP inhibition (3A4, 2D6, 2C9, 1A2 and 2C19)
CYP induction (CYP 3A4, 1A2)
PPB% (equilibrium dialvsis)
None
None
99.9
O
(a)
(b)
N
O
O
76
Br
N
HO
O
O
(c)
(d)
(e)
(f)
O
O
O
O
HO
TBDMSO
O
23
77
78
N
O
O
O
O
O
O
O
S
N
S
(g)
(h)
N
(i)
-
O
H
O
H
O
Na+
O
N
N
O
O
73
PGN-1531
Scheme 2. Synthesis of PGN-1531: (a) 4-(Benzyloxy)phenylboronic acid, PdCl2dppf, K2CO3(aq), Toluene, 80 °C, 18 h, (84%); (b) H2, 10% Pd/C, EtOH, RT, 1 h, (95%); (c) TBDMS-
Cl, Imidazole, THF, RT, 5 h (93%); (d) (1) nBu-Li (1.3 M in hexanes), THF, ꢁ50 °C, 1 h, (2) Methyl chloroformate (80%); (e) TBAF, RT, 18 h (93%); (f) (1) Ms-Cl, Et3N, DCM, (2) 77,
Cs2CO3, DMF, RT, 18 h (80%); (g) NaOH(aq), EtOH, RT, 16 h (82%); (h) Toluene-2-sulfonamide, EDCI, DMAP, DCM, RT, 18 h (77%); (i) NaHCO3(aq), MeOH (95%).
in the rat, low intrinsic clearance in human hepatocytes and no CYP
inhibition/induction issues.
73. The final stage of the synthesis involved the preparation of
the sodium salt to improve the aqueous solubility—this was
achieved with sodium bicarbonate in MeOH and gave PGN-1531.
The route consists of 11 stages and gave an overall yield
of 27%.
Prior to progressing PGN-1531 into clinical development, a ro-
bust synthetic route had to be found. The formal synthesis of
PGN1531 is given in Scheme 2 and commenced with the prepara-
tion of compound 76. Compound 76 was obtained using traditional
Suzuki–Miyaura cross-coupling chemistry of 4-(benzyloxy)phenyl-
boronic acid with 2-bromo-5-methoxy-pyridine, followed by re-
moval of the benzyl protecting group by hydrogenation. The core
furan template 23 was commercially available and after TBDMS-
protection of the alcohol, lithiation with Bu–Li and quenching with
methyl chloroformate, afforded the furan ester 77. The silyl pro-
tecting group was removed using TBAF and the resulting alcohol
converted to the mesylate under standard conditions. Coupling of
the mesylate with the compound 76, using basic conditions, gave
the intermediate methoxy-pyridyl furan ester 78. The hydrolysis
of the ester and coupling of the resultant acid with ortho-toluene-
sulfonamide using EDCI afforded the acyl sulfonamide compound
Following a successful multiple ascending oral dose phase 1a
study with PGN-1531, additional clinical profiling was undertaken.
In a phase 1b placebo-controlled study, the effects of two doses of
the compound (200/400 mg p.o.) on UVB and capsaicin-induced
neurogenic inflammation and hyperalgesia were assessed. To this
end, the pain response to carbon dioxide laser pulses, applied
directly to the irritated skin, was measured using EEG changes
in somatosensory-evoked potentials (LSEP).17 Gratifyingly, the
peak-to-peak amplitudes in both pain paradigms were significantly
reduced versus placebo, demonstrating a good overall analgesic
response in both neurogenic and inflammatory pain models. These
findings support further development of PGN-1531 for the treat-
ment of migraine.

J. Sutton et al. / Bioorg. Med. Chem. Lett. 24 (2014) 2212–2221
2221
9. Machwate, M.; Harada, S.; Leu, C. T.; Seedor, G.; Labelle, M.; Gallant, M.;
Hutchins, S.; Lachance, N.; Sawyer, N.; Slipetz, D.; Metters, K. M.; Rodan, S. B.;
Young, R.; Rodan, G. A. Mol. Pharmacol. 2001, 60, 36.
In this Letter, we have presented the results of a hit-finding and
lead optimisation programme against the EP4R. In a short time per-
iod, we were able to discover five structurally diverse series of hit
compounds using a combination of virtual screening methods. The
most favoured hit, compound 6, was demonstrated to be a compet-
itive antagonist of the EP4R. Compound 73 was identified following
several rounds of optimization, which centered on improving both
the primary EP4R affinity and selectivity against the related EP2R as
well as the aqueous solubility.18 This work culminated in the prep-
aration of PGN-1531, the sodium salt of 73, which showed marked
improvements in solubility (>10 mg/mL). PGN-1531 is a potent
and selective antagonist at EP4Rs in vitro and in vivo, with the po-
tential to alleviate the symptoms of migraine that result from cere-
bral vasodilatation.
10. Tani, K.; Kobayashi, K.; Maruyama, T. PCT Int. Appl. WO2001/062708 A1, 2001.
11. Primary assays: Membranes were prepared from cells stably transfected with
human EP4 receptor cDNA. In brief, cells were cultured to confluency, scraped
from culture flasks, and centrifuged (800 g, 8 min, 4 °C). Cells were twice
washed in ice cold homogenisation buffer containing 10 mM Tris–HC1, 1 mM
EDTAꢂZNa, 250 mM sucrose, 1 mM PMSF, 0.3 mM indomethacin, pH 7.4,
homogenised and re-centrifuged as before. The supernatant was stored on
ice and pellets re-homogenised and re-spun. Supernatants were pooled and
centrifuged at 40,000g, 10 min, 4 °C. Resultant membrane pellets were stored
at ꢁ80 °C until use. For assay, membranes expressing human EP4, EP3, EP2 or
EP1 receptors were incubated in Millipore (MHVBN45) plates containing assay
buffer, radiolabelled [3H] PGE2 and 0.1–10,000 nM concentrations of
compounds. Incubations were performed at suitable temperatures and for
suitable times to allow equilibrium to be reached. Non-specific binding was
determined in the presence of 10 lM PGE2. Bound and free radiolabel was
separated by vacuum manifold filtration using appropriate wash buffers, and
bound radiolabel was determined by scintillation counting. The affinity or pKi of
each compound for each receptor was calculated from the concentration causing
50% radioligand displacement (IC50) by using the Cheng–Prusoff equation.
12. Maubach, K. A.; Davis, R. J.; Clark, D. E.; Fenton, G.; Lockey, P. M.; Clark, K. L.;
Oxford, A. W.; Hagan, R. M.; Routledge, C.; Coleman, R. A. Br. J. Pharmacol. 2009,
156, 316.
13. (a) Homology modeling carried out using Prime, version 3.1; Schrödinger, LLC:
New York, NY, 2012; (b) Jacobson, M. P.; Pincus, D. L.; Rapp, C. S.; Day, T. J. F.;
Honig, B.; Shaw, D. E.; Friesner, R. A. Proteins: Struct. Funct. Bioinf. 2004, 55, 351;
(c) Jacobson, M. P.; Friesner, R. A.; Xiang, Z.; Honig, B. J. Mol. Biol. 2002, 320, 597.
14. (a) Ligand docking carried out using Glide, version 5.8; Schrödinger, LLC: New
York, NY, 2012; (b) Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.;
Greenwood, J. R.; Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T. J. Med. Chem.
2006, 49, 6177; (c) Halgren, T. A.; Murphy, R. B.; Friesner, R. A.; Beard, H. S.;
Frye, L. L.; Pollard, W. T.; Banks, J. L. J. Med. Chem. 2004, 47, 1750; (d) Friesner, R.
A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.; Klicic, J. J.; Mainz, D. T.; Repasky,
M. P.; Knoll, E. H.; Shaw, D. E.; Shelley, M.; Perry, J. K.; Francis, P.; Shenkin, P. S.
J. Med. Chem. 2004, 47, 1739.
Acknowledgments
The authors wish to thank Mike Podmore and Russell Scammell
for analytical support and George Hynd for assistance in the gener-
ation of early library samples.
References and notes
1. (a) Hata, A. N.; Breyer, R. M. Pharmacol. Ther. 2004, 103, 147; (b) Daniel Flesch ;
Daniel Merk; Christina Lamers; Manfred Schubert-Zsilavecz Expert Opin. Ther.
Pat. 2013, 23, 233.
2. (a) Coleman, R. A. et al The IUPHAR Compendium of Receptor Characterization
and Classification, 2nd ed.; IUPHAR Media: London, UK, 2000. pp 337–353; (b)
Sugimoto, Y.; Narumiya, S. J. Biol. Chem. 2007, 282, 11613.
3. Lin, C.-R.; Amaya, F.; Barrett, L.; Wang, H.; Takada, J.; Samad, T. A.; Woolf, C. J. J.
Pharmacol. Exp. Ther. 2006, 319, 1096.
15. (a) SiteMap, version 2.7; Schrödinger, LLC: New York, NY, 2013; (b) Halgren, T. J.
Chem. Inf. Model. 2009, 49, 377; (c) Halgren, T. Chem. Biol. Drug Des. 2007, 69,
146.
16. (a) FlexS, BioSolveIT GmbH, An der Ziegelei 79, 53757 Sankt Augustin,
Germany.; (b) Lemmen, C.; Lengauer, T.; Klebe, G. J. Med. Chem. 1998, 41, 4502.
17. Master, F. R.; Peachey, M. C.; Graham, C.; Routledge, C. Analgesic effect of the
4. Davis, R. J.; Murdoch, C. E.; Ali, M.; Purbrick, S.; Ravid, R.; Baxter, G. S.; Tilford,
N.; Sheldrick, R. L. G.; Clark, K. L.; Coleman, R. A. Br. J. Pharmacol. 2004, 141, 580.
5. Abramovitz, M.; Adam, M.; Boie, Y.; Carriere, M.-C.; Denis, D.; Godbout, C.;
Lamontagne, S.; Rochette, C.; Sawyer, N.; Tremblay, N. M.; Belley, M.; Gallant,
M.; Dufresne, C.; Gareau, Y.; Ruel, R.; Juteau, H.; Labelle, M.; Ouimet, N.;
Metters, K. M. Biochim. Biophys. Acta 2000, 1483, 285.
novel, potent and selective prostanoid EP4R antagonist, BGC20-1531:
a
6. Clayton, N. M.; Collins, S. D.; Foord, S. M.; Giblin, G. M. P. PCT Int. Appl.
WO2001/010426 A2, 2001.
7. Juteau, H.; Gareau, Y.; Labelle, M.; Lamontagne, S.; Tremblay, N.; Carriere, M.-
C.; Sawyer, N.; Denis, D.; Metters, K. M. Bioorg. Med. Chem. Lett. 2001, 11, 747.
8. Hattori, K.; Tanaka, A.; Kono, Y.; Nakazato, S. PCT Int. Appl. WO2000/018744
A1, 2000.
placebo-controlled study using laser-induced somatosensory-evoked
potentials obtained from UVB and capsaicin irritated skin in healthy
volunteers. Poster presented at 18th International Headache Research
Seminar, 6–8th March 2009, Copenhagen, Denmark.
18. Clark, D.; Clark, K.; Coleman, R.; Davis, R.; Fenton, G.; Harris, N.; Hynd, G.; Newton,
C.; Oxford, A.; Stuttle, K.; Sutton, J. PCT Int. Appl. WO2004/067524 A1, 2004.