Discovery of novel prostaglandin analogs of PGE2 as potent and selective EP2 and EP4 receptor agonists

Article abstract of DOI:10.1016/j.bmcl.2007.05.025

Analogs of PGE₂ with introduction of diene groups at the ω-side chain have been synthesized and evaluated for their binding affinity for EP₂ and EP₄ receptors. An optimized analog (compound 9b) showed high potency and sele

Full text of DOI:10.1016/j.bmcl.2007.05.025

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4323–4327
Discovery of novel prostaglandin analogs of PGE2 as potent
and selective EP2 and EP4 receptor agonists
Yufang Xiao,a,* Gian Luca Araldi,a, Zhong Zhao,a,à Nadia Brugger,a
Srinivasa Karra,a David Fischerb and Elizabeth Palmerb
aDepartment of Medicinal Chemistry, EMD Serono Research Institute, Inc., Rockland, MA 02370, USA
bDepartment of Lead Discovery, EMD Serono Research Institute, Inc., Rockland, MA 02370, USA
Received 20 March 2007; revised 4 May 2007; accepted 9 May 2007
Available online 16 May 2007
Abstract—Analogs of PGE2 with introduction of diene groups at the x-side chain have been synthesized and evaluated for their
binding affinity for EP2 and EP4 receptors. An optimized analog (compound 9b) showed high potency and selectivity for the
EP4 receptor over other known receptors.
Ó 2007 Elsevier Ltd. All rights reserved.
Prostaglandins are lipid mediators derived from arachi-
donic acid.1 Prostaglandin E2 (PGE2, Fig. 1) is the most
well-known prostanoid derivative and exhibits a broad
range of biological actions in diverse tissues through
the binding to specific receptors present in the plasma
membranes. It is well known that prostaglandin E2
(PGE2) is a potent agonist of the four subtypes of
PGE receptors, designated EP1, EP2, EP3, and EP4,
which mediate a wide variety of biological activities,
including bronchodilation, fertility, bone resorption,
and inflammation.1 Of these four receptors, three are in-
volved in modulation of cAMP levels.2 EP1 receptor is
involved in regulating intracellular calcium levels. Acti-
vation of EP3 receptor results in a reduction of intracel-
lular cAMP level, and EP2 receptor and EP4 receptor
increase the intracellular cAMP level, which is linked
to the treatment of infertility. EP2 and EP4 receptor ago-
nists have been proved to be beneficial for the treatment
of preterm labor by suppressing uterine contraction and
inducing oophorus maturation required for fertilization
during and after ovulation.1b
PGE2 easily undergoes acid- or base-catalyzed elimina-
tion of the 11-hydroxy to give the more stable a,b-unsat-
urated ketone system. PGE2, metabolized by enzymes,
undergoes rapid oxidation of the 15-hydroxy group to
a ketone, b-oxidation of the carboxylic acid chain to
generate acetic acid and the dinor PG acid, and x-oxida-
tion at C-20 to produce the 20-hydroxy and carboxylic
acid.3 The metabolites of PGE2, though inactive against
the PGE receptors, produce numerous side effects.3
PGE2 also shows no selectivity for the EP1À4 receptors
(in-house binding affinity data shown in Table 1). Until
now, efforts to improve the selectivity and chemical sta-
bility of PGE2 have been focused on only two general
O
CO2H
CO2H
O
N
HO
HO
PGE2
HO
1a
CO2H
O
Keyword: Prostaglandin.
*
N
Corresponding author. Tel.: +1 781 681 2816; fax: +1 781 681
R
2939; e-mail: yufang.xiao@emdserono.com
Present address: Forest Laboratories, 45 Adams Ave., Hauppague,
NY 11788, USA.
1b
Figure 1. PGE2 and prostaglandin diene derivative.
à
Present address: Genzyme Corporation, 153 Second Ave., Walth-
man, MA 02451, USA.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.05.025

4324
Y. Xiao et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4323–4327
Table 1. c-Lactam diene derivatives
R
h-EP1
Ki (nM)
h-EP2 Ki
(nM)
h-EP2
EC50 (nM)
h-EP3
Ki (nM)
h-EP4
Ki (nM)
h-EP4
EC50 (nM)
CO2H
O
N
R
1a
PGE2
>10,000
9.1
125.5
4.9
16
9000
0.33
2.5
0.79
0.002
9a
9b
16
17
18
19
20
21
22
10,000
10,000
134
60
222
35
10,000
10,000
181
4
15
0.1
2
88
197
25
1850
802
29
45
4
443
145
24
23
21
3
173
60
117
29
749
>10,000
0.5
chemical modifications,4 i.e., Replacement of the a-alke-
nyl side chain with the more chemically stable phenyl-
ethyl group and substitution of c-lactam ring for the
11-hydroxy cyclopentanone (structure 1a in Fig. 1).4
benzoic acid has been reported.4c–f Synthesis of the tri-
phenylphosphonium bromide salt 6 started from com-
mercially available methyl 3-oxoheptanoate. Treatment
of 3-oxoheptanoate 2 with sodium hydride, followed
by quenching with diethyl chlorophosphate, provided
the only one isomer (Z) of enol phosphate 3 in 77%
yield.5 The regioselective coupling of this enol phos-
phate 3 with 3 equiv of lithium dimethylcuprate at
À35 °C smoothly introduced a methyl group in the de-
sired product 4 in 77% yield.5 Reduction of ester 4 with
DIBALH offered the allylic alcohol, which was con-
verted into bromide 5 with treatment of the resulting
alcohol with phosphorus tribromide in 69% yield in 2
steps. Reflux of bromide 5 with triphenylphosphine in
toluene provided the triphenylphosphonium bromide
salt 6 in 84% yield.6 Treatment of triphenylphosphoni-
um bromide salt 6 with n-butyllithium (2.5 N) at 0 °C
generated the phosphorus yield to react with the alde-
hyde 7, to produce inseparable diene 8a and 8b in a
1:1 ratio of the two isomers (8a: 13E,15E and 8b:
13Z,15E). After saponification of ester 8a and 8b into
Our goal was to develop PGE2 analogs with different
selectivity and high potency for EP2 receptor or EP4
receptor. This allowed an investigation of the new
PGE2 analogs in the treatment of infertility. In this re-
port, we describe the synthesis and structure–activity
relationship of a novel PGE2 analog of general struc-
ture 1b consisting of a phenylethyl moiety in the a-side-
chain, a c-lactam ring as a backbone, and a third,
unprecedented modification: use of a diene as the x-
side chain. The diene moiety was selected to circum-
vent oxidation of the hydroxy group at the x-side
chain.
Synthesis of a representative c-lactam diene of general
formula 1b is shown in Scheme 1. Construction of the
key intermediate, c-lactam aldehyde 7, from 4-formyl

Y. Xiao et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4323–4327
4325
OP(O)(OEt)2
b
a
b
a
O
I
CO2Me
Me3Sn
CO2Me
CO2Me
79%
72%
77%
77%
3
2
12
11
10
CO2Me
e
CO2Me
O
c,d
O
Br
c
d
N
N
84%
5
69%
Br
Br
4
81%
65%
CHO
CO2Me
O
f
7
13
N
PPh3Br +
85%
CHO
CO2Me
CO2Me
O
O
6
7
d
N
N
Br
CO2Me
O
13-61%
CO2Me
O
g
14
N
8b
+
N
95%
Scheme 2. Reagents and conditions: (a) i—Cp2ZrCl2, AlMe3, DCM,
À10 °C; ii—I2, THF; (b) Me3SnSnMe3, Pd(PPh3)4, THF; (c) CBr4,
PPh3, DCM; (d) Bu3SnH, Pd(PPh3)4, toluene; (e) Pd2(dba)3CHCl3,
PPh3, reflux, THF, 3 h.
8a
8b
CO2H
CO2H
O
O
N
N
+
a
b
I
(OH)2B
75%
9b
68%
9a
15
11
Scheme 1. Reagents and conditions: (a) NaH, ClP(O)(OEt)2, THF,
0 °C; (b) MeMgCl, CuI, MeLi, THF, À35 °C; (c) DIBALH, ether,
À78 °C; (d) PBr3, petroleum ether 0 °C; (e) PPh3, PhMe, refux; (f) n-
BuLi (2.5 N), THF, 0 °C.; (g) NaOH (1 N), MeOH/H2O, then reverse
phase HPLC.
CO2Me
O
N
8b
carboxylic acid with aqueous NaOH (1 N) in MeOH/
H2O, the two carboxylic acid isomers (9a and 9b)7 were
separated through preparative reverse phase HPLC.
Scheme 3. Reagents and conditions: (a) BuLi, B(OiPr)3, THF, À78 °C;
(b) Pd(PPh3)4, Ag2CO3, KOH, THF, 2 h, rt.
bromide 14 in 81% yield.9 The key cross coupling reac-
tion of vinyl tin 12 and vinyl bromide 14 was carried out
between PPh3 and Pd2(dba)3CHCl3 in THF, yielding
13–61% of the desired product 8b (from 8 mmol scale
to 0.6 mmol scale).10
The original synthetic route, as shown in Scheme 1,
involving the Wittig reaction of the aldehyde 7 with
the triphenylphosphonium bromide salt 6 unfortunately
provided a 1:1 ratio of two isomers and the final product
was inconveniently separated through reverse phase
HPLC. In order to avoid the inconvenient separation
and to quickly scale-up compound 9b for the animal
toxic study, we sought a regioselective formation of cis
C@C double bond at C13–C14. Preliminary attempts
to explore the utility of different bases (BuLi, KHMDS),
solvents (THF, PhMe), and temperatures (À78 °C, 0 °C,
ambient temperature) failed to improve the regioselec-
tivity of Wittig reaction.
Because of the low yield at Stille cross-coupling condi-
tions on the large scale (over 8 mmol), the Suzuki cross
coupling reaction was investigated next. Vinyl boronic
acid 15 was obtained from iodide 11 by an exchange
of iodo with lithium, and then trapping with triisopropyl
borate in a 75% yield.11 Only the desired product 8b was
obtained by this method in 27–68% yield (14–1.4 mmol
scale).12
Our efforts to selectively form the cis C@C double bond
at C13–C14 then shifted from the Wittig reaction to the
cross-coupling reactions, such as Stille reaction (in
Scheme 2) and Suzuki cross coupling reaction
(in Scheme 3).
Receptor binding assays were performed on membranes
prepared from HEK293 expressing the EP1, EP2, EP3,
and EP4 receptors. A 100 ll reaction mixture containing
20 lg of membrane was mixed with [5,6,8,11,12,
14,15(n)-3H]prostaglandin E2([3H]PGE2) (Perkin-El-
mer), along with increasing concentrations of test com-
pounds in a final concentration of 1% DMSO. The
compounds were diluted in 25 mM MES, 10 mM
MgCl2, 1 mM EDTA, pH 6 (binding buffer), containing
4% DMSO. [3H]PGE2 was added at concentrations
equal to previously determined Kd values for EP1
Stille coupling partner 12 was prepared from 1-pentyne
in 2 steps with an overall yield of 57%.8 Conversion of
the aldehyde 7 with carbon tetrabromide and triphenyl-
phosphine smoothly furnished vinyl dibromide 13 in
65% yield. Palladium-catalyzed stereoselective hydrog-
enolysis of 13 with n-Bu3SnH yielded the desired Z-vinyl

4326
Y. Xiao et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4323–4327
(3 nM), EP2 (8 nM), EP3 (2 nM), and EP4 (2 nM). For
the EP2 and EP4 receptor membrane reactions, 500 lg
of wheatgerm agglutinin SPA (Scintillation Proximity
Assay, Amersham) beads was added to the wells. All
reaction mixtures were incubated at room temperature,
with shaking, for 1 h. The EP2 and EP4 reactions were
quantitated on a Topcount reader, whereas the EP1
and EP3 reactions were terminated by filtration through
glass fiber (GF/C) unifilter plates (Whatman, catalog #
7700-4301) that were previously soaked in 0.5% PEI
(polyethylinamine, Sigma). Unifilter plate wells were
then washed 4 times with 200 ll of binding buffer and
dried for 30 min at 50 °C. After sealing the bottom of
the plates, 100 ll of scintillation cocktail (Ultima goldTM
XR, Packard # 6013119) was added in the wells and fil-
ters were incubated for 1 h at rt and radioactivity
remaining on filters was measured using a Topcount
plate counter (Packard).
might be responsible for the improvement of the selec-
tivity for EP4 receptor over the other receptors (1a vs
PGE2, 9b vs PGE2).4b,d The diene analogs (9b vs 1a)
without the 15-hydroxy group still maintain the similar
selectivity. It can be explained that the 15-hydroxy
group or diene at x-side chain residing in the hydropho-
bic pocket might have different flexibility and the differ-
ent binding motifs.4d
The rat pharmacokinetics data showed 9b underwent
slow degradation. The iv clearance rate of 9b was mea-
sured to be 0.15 L/kg/h, which was a remarkable de-
crease in clearance rate in comparison with 1a (iv in
rat, CL: 2.5 L/kg/h, half-life time 0.78 h).4d This slow
clearance rate resulted in its prolonged half-life time
(7.7 h). Its bioavailability was calculated to be 72.6%
(rat, dosing with 1 mg/kg of 9b). We also studied the
utility of 9b in animal models of the ovulation induction,
Asthma, and Ulcerative Colitis. Preliminary result of
diene 9b showed very good efficacy in an in vivo mice
ovulation induction model (CD-1 adult mice) with
ED50 = 1 mg/kg when administered orally.4e The other
discoveries will be reported in due course.
Production of cAMP in response to prostanoid com-
pounds was measured in HEK293 cells transfected with
EP2 or EP4 receptor, respectively. The cells were plated
at a density of 20,000 cells/well in 96-well plates, one day
prior to the assay. Stimulation was carried out in assay
buffer (phenol red-free DMEM/F12, containing 0.1%
BSA, 0.1 mM isobutylmethyl-xanthine, and 1% penicil-
lin–streptomycin) for 60 min with increasing doses of
test molecules. Following stimulation, cells were lysed
and cAMP in the lysate was measured using a cAMP
chemiluminescent assay kit (Tropix, Bedford, MA,
USA) as per manufacturers’ instructions.
Conclusions. Analogs of PGE2, wherein the hydroxy
cyclopentanone ring has been replaced by a c-lactam
and a 15-hydroxy alkenyl group of the x-chain side
has been substituted by a diene, were found to be potent
and selective agonist of the EP2 and EP4 receptors. In
particular, it was determined that the C-15 hydroxy
group present at the x-side chain was not necessary.
Furthermore, we found that incorporation of a methyl
group at C-16 and cis C@C bond geometry at C13–
C14 of the x-side chain played important roles in the
improvement of potency and selectivity for the specific
subtypes of the EP2 and EP4 receptors. The rat PK data
exhibited that compound 9b had more stability during
enzyme metabolization.
The binding affinity data for the selected compounds
listed in Table 1 revealed that compounds 9a (the
13E,15E isomer) and 9b (the 13Z,15E isomer) were cose-
lective for EP2 and EP4 receptors. Notably, Compound
9b exhibited greater selectivity and potency for the EP4
receptor (Ki = 4 nM, EC50 = 0.1 nM) than the EP2
receptor (Ki = 60 nM, EC50 = 35 nM). Similarly, isomer
18 (13Z,15E) showed selectivity for the EP4 receptor,
while its isomer, compound 19 (13E,15E), showed selec-
tivity for EP2 receptor. It means that the cis C@C dou-
ble bond geometry at C13–C14 is a major factor
contributing to improving the selectivity for EP4 recep-
tor over the EP2 receptor. Compound 18, a methyl
group at C-15, showed decreased binding affinity and
potency for the EP4 receptor by 100-fold, in comparison
with compound 9b, a methyl group at C-16. It indicates
that having a methyl group in the C-16 plays a crucial
role in improving the EP4 activity. Compound 16, which
lacks a methyl group at either C-15 or C-16, almost re-
tains EP2 and EP4 receptor activity, but showed de-
creased selectivity. Introduction of a shorter alkyl
chain at C-16 (compounds 17 and 20) sharply decreased
the binding affinity and potency for the EP2 receptor.
The replacement of an n-alkyl chain at C-16, with cyclo-
alkyl and simple phenyl groups, (e.g., compounds 21
and 22) led to a decrease in EP2 and EP4 activity as well.
Acknowledgments
We gratefully acknowledge Dr. Ben Askew and Dr. Les-
ley Liu-Bujalski for proofreading.
References and notes
1. (a) Funk, C. D. Science 2001, 294, 1871; (b) Narumiya, S.;
Sugimoto, Y.; Ushikubi, F. Pharmacol. Rev. 1999, 79,
1193; (c) Regan, J. Life Sci. 2003, 74, 143; (d) Kobayashi,
T.; Narumiya, S. Prostaglandins Other Lipid Mediat. 2002,
68–69, 557; (e) Sugimoto, Y.; Narumiya, S.; Ichikawa, A.
Prog. Lipid Res. 2000, 39, 289; (f) Boie, Y.; Stocco, R.;
Sawyer, N.; Slipetz, D.; Ungrin, M.; Neuschafer-Rube, F.;
Puschel, G.; Metters, K.; Abranovitz, M. Eur. J. Phar-
macol. 1997, 340, 227.
2. Breyer, R. M.; Bagdassarian, C. K.; Myers, S. A.; Breyer,
M. D. Annu. Rev. Pharmacol. Toxicol. 2001, 41, 661.
3. (a) Yamamoto, S. Compr. Nat. Prod. Chem. 1999, 1, 255;
(b) Goto, M. P.; Goldman, A. S. Handb. Exp. Pharmacol.
1997, 124, 433 (Drug Toxicity in Embryonic Development
I); (c) Collins, P. W.; Djuric, S. W. Chem. Rev. 1993, 93,
1533.
Compound 1a wherein the hydroxy cyclopentanone ring
has been replaced by a lactam was found to exhibit po-
tent and selective agonist activity at EP4 receptor and
lack of the 11-hydroxy group in these lactam analogs

Y. Xiao et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4323–4327
4327
4. (a) Cameron, K. O.; Lefker, B. A.; Crawford, D. T.;
DaSilva-Jardine, P.; DeNinno, S. L.; Gilbert, S.; Grasser,
W. A.; Ke, H. Z.; Lu, B.; Owen, T. A.; Paralkar, V. M.;
Thompson, D. D.; Tjoa, C. M.; Zamistoski, M. P.;
Abstracts of Papers, 224th National Meeting of the
American Chemical Society, Boston, MA, 2002; (b) Billot,
X.; Chateauneuf, A.; Chauret, N.; Denis, D.; Greig, G.;
Mathieu, M.-C.; Metters, K. M.; Slipetz, D. M.; Young,
R. N. Bioorg. Med. Chem. Lett. 2003, 13, 1129; (c) Zhao,
Z.; Bao, B.; Brugger, N.; Fischer, D.; Giachetti, C.;
Golzio, L.; Karra, S.; Marinelli, P.; McKenna, S.; Palmer,
E.; Reddy, A.; Xiao, Y.; Araldi, G. Abstracts of Papers,
228th National Meeting of the American Chemical Soci-
ety, Boston, MA, 2004; (d) Elworthy, T. R.; Brill, E. R.;
Chiou, S.-S.; Chu, F.; Harris, J. R.; Hendricks, R. T.;
Huang, J.; Kim, W.; Lach, L. K.; Mirzadegan, T.; Yee, C.;
Walker, K. A. M. J. Med. Chem. 2004, 47, 6124; (e)
Brugger, N.; Bao, B.; Fischer, D.; Karra, S.; Mckenna, S.;
Palmer, E.; Xiao, Y.; Zhao, Z.; Araldi, G. L. Abstracts of
Papers, 230th National Meeting of the American Chem-
ical Society, Washington, DC, 2005.; (f) Zhao, Z.; Araldi,
G. L.; Reddy, P. A.; Xiao, Y.; Liao, Y.; Karra, S.;
Brugger, N.; Fischer, D.; Palmer, E. Bioorg. Med. Chem.
Lett. submitted for publication.
5. Sum, F.-W.; Weiler, L. Can. J. Chem. 1979, 57, 1431.
6. Yoda, H.; Oguchi, T.; Takabe, K. Tetrahedron: Asymme-
try 1996, 7, 2113.
7. Compounds 9a and 9b were stable up to 1 week at acidic (pH
1), neutral (pH 7.4), and basic (pH 12) conditions. No
isomerization or decomposition was observed during syn-
thesis and other handling for binding and functional assays.
8. (a) Negishi, E.; Van, H.; David, E.; King, A.; Okukado,
N. Synthesis 1979, 501; (b) Bellina, F.; Carpita, A.;
Fontana, E.; Rossi, R. Tetrahedron 1994, 50, 5189.
9. (a) Uenishi, J.; Kawahama, R.; Shiga, Y.; Yonemitsu, O.;
Tsuji, J. Tetrahedron Lett. 1996, 37, 6759; (b) Uenishi, J.;
Kawahama, R.; Yonemitsu, O.; Tsuji, J. J. Org. Chem.
1996, 61, 5716; (c) Uenishi, J.; Kawahama, R.; Yonemitsu,
O.; Tsuji, J. J. Org. Chem. 1998, 63, 8965.
10. Milstein, D.; Stlle, J. K. J. Am. Chem. Soc. 1979, 101,
4992.
11. Uenishi, J.; Matsui, K.; Wada, A. Tetrahedron Lett. 2003,
44, 3093.
12. Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457.