Discovery of a potent and selective agonist of the prostaglandin EP4 receptor

Article abstract of DOI:10.1016/S0960-894X(03)00042-8

Analogues of PGE₂ wherein the hydroxycyclopentanone ring has been replaced by a lactam have been prepared and evaluated as ligands for the EP₄ receptor. An optimized compound (19a) shows high potency and agonist efficacy at the EP

Full text of DOI:10.1016/S0960-894X(03)00042-8

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1129–1132
Discovery of a Potent and Selective Agonist of the
Prostaglandin EP4 Receptor
Xavier Billot,y Anne Chateauneuf, Nathalie Chauret, Danielle Denis, Gillian Greig,
Marie-Claude Mathieu, Kathleen M. Metters, Deborah M. Slipetz
and Robert N. Young,*
´
Merck Frosst Centre for Therapeutic Research, PO Box 1005, Pointe Claire-Dorval, Quebec, Canada H9R 4P8
Received 1 November 2002; accepted 20 December 2002
Abstract—Analogues of PGE2 wherein the hydroxycyclopentanone ring has been replaced by a lactam have been prepared and
evaluated as ligands for the EP4 receptor. An optimized compound (19a) shows high potency and agonist efficacy at the EP4
receptor and is highly selective over the other seven known prostaglandin receptors.
# 2003 Elsevier Science Ltd. All rights reserved.
Prostaglandin E2 (PGE2) is a potent pro-inflammatory
mediator which manifests a variety of activities in the
body through interaction with four distinct subtypes of
receptor (termed EP receptors) namely EP1, EP2, EP3
and EP4.1,2 PGE2 has multiple effects on bone including
stimulation of resorption and formation3 as well as
other systemic effects including induction of diarrhea
and hypotension. PGE2 has been shown to stimulate
bone formation in rats and humans in vivo.4 However,
the multiple side effects of PGE2 have made it unsuitable
for use as a bone-forming therapy. The major pros-
taglandin E receptor on bone cells is EP4 and EP4
receptor knockout mice have shown significant defects
in bone metabolism.5 A recent study utilizing an EP4
receptor antagonist has shown that the in vivo bone
forming effects of PGE2 in rats are effectively blocked by
co-administration of the antagonist.6 Thus the hypoth-
esis has been formed that a selective EP4 receptor ago-
nist could be a useful agent to promote bone growth.7
ing at the EP4 receptor and essentially no activity at any
other receptor in the prostaglandin receptor panel
(entry 1, Table 1). The compound also displayed agon-
ism in a functionally coupled cellular assay (EC50=770
nM). Compound 1 was a mixture of four isomers and
considering general homology with PGE2, it was antici-
pated that activity should reside in one isomer only. Re-
synthesis of 1 as a pair of C-12 isomers (entries 2 and 3)
revealed that the 12-(R)-enantiomer (corresponding to
the natural PGE2 stereochemistry) displayed the large
majority of the activity. Analogues of 1 incorporating a
13,14 trans-double bond (entries 4 and 5) were prepared
and again activity resided essentially in the 12-(R)-
enantiomer (these compounds were still unresolved at
the C-15 center).
Receptor binding assays are available in our laboratory
for all eight of the prostaglandin receptors8 and utilizing
these assays we have screened for selective ligands at the
EP4 receptor. Such screening resulted in the identifica-
tion of compound 19 which exhibited nanomolar bind-
Our goal was to develop an EP4 receptor agonist which
would serve as an in vivo probe in a variety of assays of
bone formation and therefore we sought to identify an
analogue with a degree of metabolic stability which
would be manifested in an improved in vivo half-life.
PGE2 has an extremely short half-life in vivo. As pros-
*Corresponding author. Tel.: +1-514-428-2647; fax: +1-514-428-
2624; e-mail:
yCurrent address: Gemin X Biotechnologies Inc., PO Box 477, Place
du Parc, Montreal, Quebec, Canada H2X 4A5.
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00042-8

1130
Table 1. Binding data for compounds at the prostaglandin receptors
Entry Compd Binding Ki (nM)
EP2 EP3
R. N. Young et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1129–1132
(t1=2=<20 min) and in vitro metabolism studies in rat
hepatocytes indicated significant b-oxidation of the acid
chain. Compounds bearing a b-sulfur atom in an acidic
chain are blocked to b-oxidation and, therefore, the
corresponding b-thio compound (16a, entry 14) was
prepared and was found to be approximately 4-fold less
potent than the corresponding methylene compound
(10b). Homologation of the chain led to a 10-fold loss in
binding activity (entry 15). Examination of the meta-
bolic fate of compound 16a in incubation with rat
hepatocytes indicated the compound was significantly
more metabolically stable than 10b (data not shown). A
major metabolite, however, was observed and was
identified as the corresponding sulfoxide. An alternative
method to block b-oxidation of the acid chain could be
to replace the acid group with an acid equivalent such as
a tetrazole that may not be recognized by the b-oxida-
tion enzymes. Thus, a variety of tetrazolyl analogues of
10b were prepared and evaluated for their potency,
binding and metabolism. The S-tetrazole compound
(entry 16) was less potent than the corresponding
straight chain acid. However, the homologue (entry 17)
retained considerable activity at the EP4 receptor.
Replacement of the acid group in compound 10b with a
tetrazole, however, led to a compound approximately 2-
fold more potent than the parent (entry 18). This com-
pound was resolved into its two diastereoisomers (iso-
meric at C-15) (entries 19 and 20) through
chromatographic separation and virtually all the bind-
ing activity was found to reside with one isomer (19a)
(IC50=1.4 nM; IC50 for PGE2 is 0.7 nM) presumably
having the natural PGE2 stereochemistry at C-15. This
compound was evaluated in rat hepatocytes for meta-
bolic stability and found to be essentially unchanged
after incubation for 2 h. Dosed intravenously in rats,
the compound had a much improved half-life (t1=2=ca 2
h). Examination of bile from rats dosed with 19a indi-
cated that the compound was excreted largely unchan-
ged. Compound 19a was shown to be highly selective
for the EP4 receptor with no significant binding
observed at greater than 14 mM on any of the other
receptors. Compound 19a was a full agonist in the cell
efficacy assay with an EC50 of 2.5ꢀ1.0 nM (n=5). This
is comparable to the EC50 of PGE2 itself
[EC50=3.0ꢀ0.4 nM (n=8)].
EP4
EP1, DP, FP,
IP, TP
1
2
3
4
5
6
7
8
1
6
8
7
35 (n=1)
28 (n=2)
1110 (n=1)
6.0ꢀ0.5 (n=5)
1410 (n=1)
3500 (n=1)
3000
4050
3610
2000
>3000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
>13,000
700
2315
419
6680
>13,000
9
1220
>13,000 >13,000
10a
10b
10c
10d
12
10e
10f
15a
16a
16b
16d
16c
2.6ꢀ0.1 (n=6) >13,000 >13,000
3.6ꢀ0.2 (n=3) >13,000 >13,000
9
102 (n=2)
2100 (n=1)
62 (n=1)
>13,000 2014
>13,000 >13,000
>13,000
>13,000
10
11
12
13
14
15
16
17
18
19
20
10,560
3340
12 (n=1)
443ꢀ48 (n=4) >13,000 >13,000
13ꢀ1 (n=5) >13,000 >13,000
140 (n=1)
243 (n=1)
66 (n=1)
>13,000 >13,000
>13,000 >13,000
>13,000 >13,000
19a+19b 2.5ꢀ0.2 (n=5) >13,000 >13,000
19a
19b
1.2ꢀ0.2 (n=4) >13,000 >13,000
230ꢀ40 (n=3) >13,000 >13,000
taglandin metabolism is known to involve both b-oxi-
dation of the acid chain and terminal oxidation of the
lipophilic chain, it was decided to explore whether these
chains could be modified to inhibit or block the expec-
ted metabolism while maintaining potency at the EP4
receptor and full functional efficacy. Initially, a variety
of compounds were prepared wherein the terminal lipo-
philic group had been replaced by an aryl ring or a
cyclohexyl ring. Introduction of the cyclohexyl ring
adjacent to the hydroxy group at C-15 was accom-
panied by a dramatic drop loss in binding activity (entry
6). Replacement of the C-17 to 20 chain with a phenyl
ring yielded a potent binding compound which retained
its in vitro efficacy in the cell assay (EC50=28 nM) (see
entry 7). Substitution of the aryl ring by a methoxy
methyl group (entry 8) (a substitution shown to be
optimal in other studies on EP4-selective PGE2 analo-
gues)10 yielded a potent compound (10c) but with no
advantage over 10b. Homologation of the chain (entry
9) led to an important drop in activity. A major route of
in vivo metabolism of the prostaglandins involves 15-
dehydrogenation.11 Although such oxidation was not
observed in hepatocyte incubations (with 10b), this type
of oxidation is known to generally take place through
extra-hepatic metabolism in mammals.12 We therefore
prepared the corresponding 15-methyl substituted ana-
logue (10c). The compound, however, was dramatically
less potent (entry 10). In order to block the 16-position
to potential benzylic oxidation, 16,16-distubstituted
analogues (entry 11 and 12) were prepared and were
found to largely retain the binding activity compared to
the 10b. However, these substitutions did not lead to a
significant increase in in vitro metabolic stability and so
it was decided to maintain the terminal methylene aryl
group as in compound 10b and investigate potential
substitution and analogues on the acid-bearing chain.
Compound 3 was synthesized from (R)-pyroglutamic
acid (2) through a series of reactions described in ref 13
(Scheme 1).
The sodium salt of 3 was alkylated with ethyl 7-bromo-
heptanoate under phase-transfer catalysis (88% yield).
After deprotection with HF–Pyridine complex, Dess–
Martin oxidation of the alcohol led to the aldehyde 4
(84% yield). Wittig reaction with the sodium salt of
dimethyl (2-oxoheptyl)phosphonate afforded the inter-
mediate enone (55% yield) which was reduced to the
allylic alcohol 5. Saponification of 5 produced the acid 7
in 94% yield. Catalytic hydrogenation of 5 with PtO2 in
AcOEt followed by saponification with 1 N LiOH gave
the acid 6. Using the methodology decribed above,
compounds 8 and 9 were obtained similar yields starting
from (S)-pyroglutamic acid [(S)-2].
Preliminary in vivo evaluation of compound 10b dosed
intravenously to a rat revealed a relatively short half-life

R. N. Young et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1129–1132
1131
The synthesis of analogues with optimized acid-bearing
chains started with the previously synthesized amide
(R)-3 (Scheme 3). N-Alkylation of the amide (R)-3 fol-
lowed by alcohol deprotection and Dess–Martin oxida-
tion afforded aldehyde 13 in good yield. Wittig reaction
with sodium dimethyl (2-oxo-3-phenylpropyl)phos-
phonate gave the intermediate enone that was immedi-
ately reduced to the allylic alcohol 14 using NaBH4 in
EtOH at ꢁ20 ꢂC. Chlorine displacement using various
alkylsulfides produced compounds 15a–d in good yields.
Compounds 15a and its homologue 15b were saponified
using 1 N LiOH to afford 16a and 16b in good yields
(respectively, 83 and 89%). Sulfide 15c was reacted with
bromoacetonitrile in the presence of nBu4NF to afford
the intermediate nitrile (94% yield), which was then
heated at 120 ꢂC for 3 h with nBu3SnN3 to give the tet-
razole 16c (88% yield). Similarly 15d was transformed
into the tetrazole 16d in good yield (Scheme 4).
The introduction of a terminal tetrazole started with the
N-alkylation of amide 12(R)-3 with NaH and 5-bromo-
heptanenitrile under phase-transfer catalysis followed
by deprotection of the alcohol (84% yield) and Dess–
Martin oxidation (84%) affording the aldehyde 17. A
Wittig reaction of 17 with the sodium salt of dimethyl
Scheme 1. Reagents and conditions: (a) (i) NaH, DMF, 50 ꢂC; (ii)
BrCH2(CH2)5COOEt, nBu4NI, 50 ꢂC, 88%; (b) HF-Py, CH2Cl2, 68%;
(c) Dess–Martin periodinane, CH2Cl2, 84%; (d) (i) NaH, DMA, 0 ꢂC;
(ii) (MeO)2P(O)CH2COC5H11, rt, 55%; (e) NaBH4, EtOH, ꢁ20 ꢂC,
92%; (f) PtO2; AcOEt, H2, 87%; (g) LiOH, H2O, THF, MeOH, 94%.
(2-oxo-3-phenylpropyl)phosphonate
intermediate enone (84%yield), which was reduced with
produced
the
NaBH4 in EtOH at ꢁ20 ꢂC to provide the nitrile 18.
The optimization of the b-chain started with aldehyde
(R)-4 (Scheme 2) which was reacted the appropriate
b-ketophosphonate using NaH as base in THF (yields
from 16 to 72%). The resultant enones were then
reduced to intermediate allylic alcohols using NaBH4 in
EtOH at ꢁ20 ꢂC (yields from 65 to 85%). Ester saponi-
fication afforded compounds 10a–f (yields from 38 to
96%). In the case of compounds 10e and 10f, the
reduction was run using the Luche reaction (NaBH4
ꢂ
. .
and CeCl3 7 H2O in EtOH–H2O at 0 C) in order to
prevent the over-reduction of the double bond. Tertiary
alcohol 12 was synthesized by reaction of enone 11 with
one equivalent of freshly prepared methylcerium
reagent in THF (33%), followed by saponification of
the intermediate ester (66%).
Scheme 3. Reagents and conditions: (a) (i) NaH, DMF, 50 ꢂC; (ii)
Br(CH2)4Cl, Bu4NI, 50 ꢂC, 87%; (b) HF-Py, CH2Cl2, 90%; (c) Dess–
Martin periodinane, CH2Cl2, 95%; (d) (i) NaH, DME, 0 ꢂC; (ii)
(MeOH)2-P(LO)CH2COCH2Ph, rt, 62%; (e) NaBH4, EtOH, ꢁ20 ꢂC,
90%; (f) RSH, MeONa, DMF, Bu4NI, 50 ꢂC; (g); LiOH, H2O–THF–
MeOH, rt; (h) BrCH2CN, nBu4NF, THF, rt 94%; (i) nBu3SnN3, neat,
120 ꢂC, 3 h.
Scheme 2. Reagents and conditions: (a) (i) NaH DME 0 ꢂC; (ii)
(MeO)2-P(O)CH2COR, rt; (b) NaBH4, EtOH, ꢁ20 ꢂC; (c) LiOH,
H2O, THF, MeOH; (d) 1 equiv MeLi, 2 equiv CeCl3, THF, ꢁ78 ꢂC,
33%; (e) LiOH, H2O, THF, MeOH, 66%.

1132
R. N. Young et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1129–1132
group can be replaced by an acid equivalent such as a
tetrazole with comparable or enhanced activity. The
optimized compound (19a) displays a high degree of
selectivity and potency and in vivo half-life much super-
ior to PGE2. This compound and related analogues are
now being studied in a number of biological assays to
determine whether an EP4 selective agonist can serve as a
useful therapeutic for treatment of osteoporosis.
References and Notes
1. Coleman, R. A.; Smith, W. L.; Narumiya, S. Pharmacol.
Rev. 1994, 46, 205.
2. Narumiya, S.; Sugimoto, Y.; Ushikubi, F. Physiol. Rev.
1999, 79, 1193.
3. Pilbeam, C. C.; Harrison, J. R.; Raisz, L. G. In Principles
of Bone Biology; Bilekizian, J. P., Raisz, J. P., Rodan, G. A.,
Eds, Academic: San Diego, 1996; p. 715.
Scheme 4. Reagents and conditions (a) (i) NaH, DMF, 50 ꢂC; (ii)
BrCH2(CH2)5CN, nBu4NI, 50 ꢂC, 84% (b) HF-Py, CH2Cl2, 87%; (c)
Dess–Martin periodinane, CH2Cl2, 88% (d) (i) NaH, DME, 0 ꢂC; (ii)
(MeO)2-P(O)CH2COCH2Ph, rt, 75%; (e) NaBH4, EtOH, ꢁ20 ꢂC; (f)
Bu3SnN3 neat 120 ꢂC, 3 h, 80%.
Finally the tetrazole moiety was introduced by reaction
of 18 with 3 equiv of nBu3SnN3 neat at 120 ꢂC for 3 h
(80% from enone). Reverse-phase HPLC allowed the
separation of both diasterioisomers eluting sequentially
in a ratio (19a/19b) (1:2) in the favor of the putative
15(R)-alcohol 19b.
4. (a) Rodan, G. A. J. Cell. Biochem. (Suppl. 15), 160. (b)
Ueno, K.; Haba, T.; Woodbury, D.; Price, P.; Anderson, R.;
Jee, W. S. S. Bone 1985, 6, 79. (c) Jee, W. S. S.; Ke, H. Z.; Li,
X. J. Bone Mineral 1991, 15, 33. (d) Norrdin, R. W.; Jee, W. S.
S.; High, W. B. Prostaglandins Leukot. Essent. Fatty Acids
1990, 41, 139.
5. Miyaura, C.; Inada, M.; Suzawa, T.; Sugimoto, Y.; Ush-
ikubi, F.; Ichikawa, A.; Narumiya, S.; Suda, T. J. Biol. Chem.
2000, 275, 19819.
Biological Results and Discussion
6. Machwate, M.; Harada, S.; Leu, T.; Seedor, G.; Labelle,
M.; Gallant, M.; Hutchins, S.; Lachance, N.; Sawyer, N.; Sli-
petz, D.; Metters, K. M.; Rodan, S. B.; Young, R.; Rodan,
G. A. Mol. Pharmacol. 2001, 60, 36.
7. Yoshida, K.; Oida, H.; Kobayashi, T.; Maruyama, T.;
Tanaka, M.; Katayama, T.; Yamaguchi, K.; Segi, E.; Tsu-
boyama, T.; Matsushita, M.; Ito, K.; Ito, Y.; Sugimoto, Y.;
Ushikubi, F.; Ohuchida, S.; Kondo, K.; Nakamura, T.; Nar-
umiya, S. Recently published reports of in vivo studies on an
analogue of PGE2 which has selective agonist activity at the
EP4 receptor have shown significant bone formation in rats
which further supports this hypothesis. See: Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 4580.
8. Abramovitz, M.; Adam, M.; Boie, Y.; Carriere, M.-C.;
Denis, D.; Godbout, C.; Lamontagne, S.; Rochette, C.; Saw-
yer, N.; Tremblay, N.; Belley, M.; Gallant, M.; Dufresne, C.;
Gareau, Y.; Ruel, R.; Juteau, H.; Labelle, M.; Ouimet, N.;
Metters, K. M. Biochem. Biophys. Acta 2000, 1483, 285.
9. Smith, R. L.; Lee, T.; Gould, N. P.; Cragoe, E. J., Jr.; Oien,
H. G.; Kuehl, F. A., Jr. J. Med. Chem. 1977, 20, 1292.
10. Maruyama, T.; Asada, M.; Shiraishi, T.; Yoshida, H.;
Maryama, T.; Ohuchida, S.; Nakai, H.; Kondo, K.; Toda, M.
Bioorg. Med. Chem. 2002, 10, 1743.
Receptor binding assays were performed using cell
membranes from HEK293ebna cells recombinantly
expressing the corresponding human prostanoid
cDNA’s.8 EP4 agonist potency and efficacy were eval-
uated utilizing a stable clone of pSV40-EP4 transfected
into HEK293 cells that expresses approximately 50
fmol/mg EP4 receptor. Whole cell cAMP assays were
performed essentially as described in Slipetz et al.14 with
the following modifications. Assays were performed
with cells in suspension in a total of 0.2 mL HBSS con-
taining 2 mM IBMX (phosphodiesterase type IV inhi-
bitor). IBMX and PGE2 or the test compound were
added to the incubation mixture in DMSO to a final
vehicle concentration of 1.8% (v/v) (kept constant in all
samples). The reaction was initiated by the addition of
1ꢃ105 cells per incubation, samples were incubated at
37 ꢂC for 10 min, and the reaction was terminated by
immersing the samples in boiling water for 3 min.
Measurement of cAMP was performed by a [125I]cAMP
scintillation proximity assay. For in vivo and in vitro
metabolism methods, see methods described by Nicoll-
Griffith et al.15
11. Hamberg, M.; Samuelsson, B. J. Biol. Chem. 1971, 246, 6713.
12. Hansen, H. S. Prostaglandins 1976, 12, 647.
13. Yoda, H.; Oguchi, T.; Takabe, K. Tetrahedron: Asym-
metry 1996, 7, 2113.
Analogues of PGE2 wherein the hydroxy cyclopenta-
none ring has been replaced by a lactam were found to
exhibit potent and selective agonism at the EP4 recep-
tor.16 Thus one can conclude that the C-11 hydroxy
group present in PGE2 is not necessary for either binding
or agonism at this receptor. Indeed, one could conclude
that lack of a corresponding hydroxy group in these
lactam analogues might be responsible for a measure of
the selectivity observed for the EP4 receptor over the
other prostaglandin receptors. The natural stereo-
chemistry at C-12 and putatively also at C-15 was found
to be important for activity at the receptor. The carboxyl
14. Slipetz, D.; Buchanan, S.; Mackereth, C.; Brewer, N.;
Pellow, V.; Hao, C.; Adam, M.; Abramovitz, M.; Metters,
K. M. Biochem. Pharmacol. 2001, 62, 997.
15. Nicoll-Griffith, D. A.; Falgueyret, J.-P.; Silva, J. M.;
Morin, P. E.; Trimble, L.; Chan, C.-C.; Clas, S.; Leger, S.;
Wang, Z.; Yergey, J. A.; Riendeau, D. Drug Metab. Disp.
1999, 27, 403.
16. Several patent applications on closely related structures
claimed as EP4 receptor agonists for treatment of osteoporosis
have recently been published. See: (a) Cameron, K. O.; Lefker,
B. A. WO 0242268 A2. (b) Cameron, K. O.; Ke, H.; Lefker, B.
A.; Thompson, D. D. WO 0146140 A1. (c) Maruyama, T.;
Kobayashi, K.; Maruyama, T. WO 0224647 A1.