Agonist lead identification for the high affinity niacin receptor GPR109a

Article abstract of DOI:10.1016/j.bmcl.2007.06.028

A strategy for lead identification of new agonists of GPR109a, starting from known compounds shown to activate the receptor, is described. Early compound triage led to the formulation of a binding hypothesis and eventually to our focus on a series of pyrazole acid derivatives. Further elaboration of these compounds provided a series of 5,5-fused pyrazoles to be used as lead compounds for further optimization.

Full text of DOI:10.1016/j.bmcl.2007.06.028

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4914–4919
Agonist lead identification for the high affinity
niacin receptor GPR109a
Tawfik Gharbaoui,a Philip J. Skinner,a Young-Jun Shin,a Claudia Averbuj,a
Jae-Kyu Jung,a Benjamin R. Johnson,a Tracy Duong,a Marc Decaire,a Jane Uy,a
Martin C. Cherrier,a Peter J. Webb,a Susan Y. Tamura,a Ning Zou,a Nathalie Rodriguez,a
P. Douglas Boatman,a Carleton R. Sage,a Andrew Lindstrom,a Jerry Xu,a
Thomas O. Schrader,a Brian M. Smith,a Ruoping Chen,b Jeremy G. Richman,b
Daniel T. Connolly,b Steven L. Colletti,c James R. Tatac and Graeme Semplea,*
aDepartment of Medicinal Chemistry, Arena Pharmaceuticals, 6166 Nancy Ridge Drive, San Diego, CA 92121, USA
bDepartment of Discovery Biology, Arena Pharmaceuticals, 6166 Nancy Ridge Drive, San Diego, CA 92121, USA
cDepartment of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ, USA
Received 9 May 2007; revised 6 June 2007; accepted 7 June 2007
Available online 10 June 2007
Abstract—A strategy for lead identification of new agonists of GPR109a, starting from known compounds shown to activate the
receptor, is described. Early compound triage led to the formulation of a binding hypothesis and eventually to our focus on a series
of pyrazole acid derivatives. Further elaboration of these compounds provided a series of 5,5-fused pyrazoles to be used as lead
compounds for further optimization.
Ó 2007 Elsevier Ltd. All rights reserved.
Niacin (nicotinic acid) is a water-soluble vitamin that at
high doses in humans, favourably modulates multiple
serum lipid and lipoprotein parameters that are cardio-
vascular risk factors.1 Niacin increases high density lipo-
protein (HDL) to a greater extent than any other
currently used drug and it has been used for many years
for the prevention and treatment of cardiovascular dis-
ease.2 Recent mechanistic investigations have shown
that niacin interacts with, and may exert its beneficial ef-
fects on lipids via, a G-protein coupled receptor (GPCR)
expressed on adipocytes.3 It is postulated that the resul-
tant decrease in intracellular cAMP leads to inhibition
of lipolysis via negative modulation of intracellular
lipase activity, thereby decreasing plasma free fatty acid
(FFA) levels. A further hypothesis suggests that this
pathway may ultimately result in elevation of HDL.4
Two Gi-coupled orphan GPCRs that share 95% identity
and which are both expressed in human adipocytes have
been identified as possible molecular targets for niacin.5
GPR109a (HM74a) is the human orthologue of the pre-
viously described rodent receptor (PUMA-G)6 whereas
GPR109b (HM74) differs from GPR109a and PUMA-
G mainly in the intracellular C-terminal tail portion
and has no rodent equivalent.7 Niacin has been reported
to activate GPR109a with an EC50 of 250 nM in a
GTPcS assay and displaces H-niacin from GPR109a
expressing CHO cell membranes with an IC50 of
81 nM.5 It is a much weaker ligand for GPR109b.
3
The use of niacin as a therapeutic is limited by a number
of associated side-effects, in particular an uncomfortable
cutaneous flushing response which, although tolerance
may be acquired, has a negative impact on patient com-
pliance. The development of novel agonists of the niacin
receptor that retain the beneficial effect on atherosclero-
sis but with no flushing side-effect would clearly be of
value. In this report we describe our early investigations
into the identification of appropriate chemical starting
points for the discovery of such compounds.
Following our in-house discovery of potential GPCR
targets for niacin,8 we established high throughput
cAMP assays using the stably expressed human
Keywords: GPR109a; HM74a; Niacin receptor; Free fatty acids; Lead
identification.
*
Corresponding author. E-mail: GSemple@arenapharm.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.06.028

T. Gharbaoui et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4914–4919
4915
GPR109a and GPR109b receptors in CHO cells.9 We
initially used these assays to examine whether structur-
ally related small acid analogues of nicotinic acid had
any activity at the human receptors in a similar manner
to studies with the rodent receptors using rat adipocytes
or spleen membranes and mouse macrophages.3,10
From these early data we formulated a preliminary
binding hypothesis in which the acid portion of the ago-
nist molecules would interact with Arg-111, essentially
the only available basic residue in the TM domains,
on TM3; a common site of interaction for small mole-
cule agonists with GPCRs. The suggestion that Arg-
111 is involved in the critical binding interaction is
now well supported by compelling mutagenesis and
modeling data.7 In addition to this key bidentate inter-
action, all of the ligands identified for the receptor thus
far had included a hydrogen bond acceptor moiety in
the 3-position of the aromatic group relative to the car-
boxylate function. Other molecules that have been
shown to activate GPR109a such as acipimox and
acifran14 also contain these two important pharmaco-
phore features. With this in mind, we selected an
extended range of 6-membered ring acids for further
testing all of which were commercially available in either
the acid or ester form (Table 1). As can be seen from the
data, addition of a second nitrogen to the ring resulted
in at least a 20-fold loss in activity for GPR109a. To fur-
ther solidify the requirement for a 1,3-orientation of the
carboxylate and the H-bond acceptor nitrogen, we also
confirmed that pyrimidine 2-carboxylate and pyrimidine
4-carboxylate had no activity. In contrast to the classical
5-membered ring analogues described above, where a
methyl substituent was required for good activity, sub-
stitution on the pyridine ring was poorly tolerated which
we attributed to a non-favourable steric interaction.
Even the small fluoro-substituent was only tolerated in
the 5-position and interestingly this compound has also
been previously reported to be antilipolytic in the dog.15
The same authors also reported on a small number of
other heterocycles that had moderate antilipolytic activ-
ity in the same model from a large selection of 188 com-
pounds.16 Other halo-substitutions in the 5-position
Our initial investigation of commercially available pyri-
dine acids showed that as expected, only the 3-pyridyl
carboxylate had any activity at the human cloned
GPR109a receptor (EC50 = 0.12 lM, n = 216; 95% CI
0.109–0.126 lM, in the cAMP assay; reported
3
Ki = 33 nM in rat spleen membrane H-niacin binding
assay3) whereas the 2- and 4-pyridyl carboxylates were
inactive. Nicotinic acid had no activity at the closely
related GPR109b in this assay platform at concentrations
up to 30 lM. Similar to the observations of Lorenzen
et al. in rat adipocytes and spleen membranes, we found
that homologation of niacin to 3-pyridyl acetic acid
resulted in a significant loss of potency (EC50 = 2.8 lM,
n = 2; reported Ki = 403 nM in rat spleen binding
assay3) and further homologation abolished all activity.
We hypothesized early on that, based on their in vivo
activity and structural similarity to niacin, several com-
pounds with well documented activity in acutely lower-
ing free fatty acids in rats (and in some cases, human),
were likely to be agonists of GPR109a. Thus, a number
of 5-membered heterocyclic acids were examined includ-
ing 13 and 19 (Table 2). The former compound was
investigated by Upjohn in the 1960s11 and in our hands
it showed good activity at GPR109a in the cAMP assay
and perhaps more surprisingly in view of the high
homology between the receptors, essentially no activity
(up to 100 lM) at GPR109b. Interestingly, the com-
pound lacking the 5-methyl substituent, that is, pyra-
zole-3-carboxylate, was an order of magnitude less
potent (EC50 = 4.2 lM, n = 2; Ki = 594 nM in rat spleen
binding assay12). A similar observation was also made
with the isoxazole analogues. Compound 19, investi-
gated in the 1960s by workers at Pfizer,13 was an effec-
tive agonist of GPR109a (Table 2) and again had no
effect at GPR109b. In this case, the isoxazole analogue
lacking the methyl group had barely measurable activity
with an EC50 for GPR109a in excess of 30 lM. The iso-
meric isoxazole (5-methylisoxazole-3-carboxylate) also
had no effect.
Table 1. GPR109a agonist activity for
heterocyclic acids
a series of 6-membered
O
A1
A2
A3
OH
A4
N
Compound A1 A2
A3 A4 GPR109a EC50a, lM (n)
1, Niacin
2
CH CH
CH CH
CH CH
CH CH 0.12 (216)
CH
N
N n.e.
CH 23 (1)
A comparison of the data from our cAMP assay with
those from the rat spleen membrane binding assay
showed that we had retained approximately the same
rank order of potency, providing good validation of
our assay, but with an apparent 3- or more fold right-
ward shift in the dose–response curves. To determine
whether this was a species difference or more a reflection
of the different types of readout from the two assay plat-
forms, we also established cAMP assays in CHO cells
using the cloned rat and mouse receptors. The potency
of each of these early compounds tested was essentially
identical across species giving us a good correlation
between rodent and human receptor activity in the
cAMP assays despite the modest difference compared
to the endogenous receptor assays.
3
4
CH
N
N
CH
CH CH 2.9 (2)
CH CH 6.3 (3)
CH CF n.e.
5
6
CH CH
CH CH
CH CF
CF CH
CH CCl
CH CBr
7
CF CH n.e.
CH CH 5.3 (5)
CH CH n.e.
8
9
10
11
12
CH CH 7.6 (2)
CH CH 13.6 (2)
CH CCH3 CH CH 16 (1)
n.e., less than 50% activity observed at any concentration up to 30 lM.
a EC50 determination from the cAMP assay.9 Values are means of
multiple determinations. The number of experiments is listed in
parentheses. The standard deviations where calculable were <30% of
the mean.

4916
T. Gharbaoui et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4914–4919
NH2
O
O
(i)
(iii)
(ii)
Si
O
OH
18
CH3CN
HN
HN
S
O
NH2
(iv)
(v)
N
OH
NH2OH.HCl
HO
O
N
O
O
N
20
O
O
R
(vii)
(vi)
OH
R
R
NH
NH
N
N
Scheme 1. Preparation of 5-membered heterocyclic acids. Reagents and conditions: (i) a—(trimethylsilyl)methyl triflate, 80 °C; b—thiophenol,
benzene, rt; (ii) ethyl propiolate, AgF, MeCN, rt; (iii) aq NaOH, 80 °C, 1 h; (iv) MeCN, Na2CO3, H2O, EtOH, reflux 45 min; (v) ethyl 2-chloro-2-
oxoacetate, 120 °C, microwave heating; (vi) a—(CO2Et)2, EtOH, KOtBu (or EtONa); b—H2NNH2; (vii) 2.5 M NaOH or LiOH.
were not tolerated and no activity was seen with any
other commercially available pyridine 3-carboxylate
tested. We thus concluded that niacin and the related
diazine acids would not make good chemical starting
points to identify more potent analogues due to the
apparent lack of scope for further expansion.
in the ring resulted in a reduction in activity as did incor-
poration of additional nitrogen atoms.
With these data in hand we elected to focus further on
the pyrazole series both for reasons of the relative ease
of synthesis of pyrazoles and that pursuing the thiazole
class may be limited by the ring only having one avail-
able position for substitution. The known in vivo activ-
ity of pyrazole analogues was also attractive. In addition
to 5-methylpyrazole carboxylate, the antilipolytic effects
of pyrazole acids incorporating longer side chains in the
5-position have been reported.17 The in vitro activity of
series of pyrazoles with a wider variety of substitutents
at the 5-position has also been described, with a number
of the compounds prepared showing partial agonist
activity in a GTPcS functional assay in rat spleen and
adipocytes.12 In that study, the most potent activity
was seen with straight chain and branched alkyl substit-
uents of 3 or 4 carbon atoms. We prepared compounds
of this type to compare their activities in our human
cloned receptor system and in addition designed a num-
ber of new analogues with hydrogen bond donors and
acceptors in the side chain to try and solicit an addi-
tional binding interaction with the receptor to further
increase potency. The compounds were in general, pre-
pared by conventional pyrazole synthesis via a keto ester
intermediate and cyclization with hydrazine followed by
ester hydrolysis, as outlined in Scheme 1. (A number of
analogues required more specific synthetic methods
including synthesis of the starting ketone as detailed in
the supplementary material.) As can be seen from the
data in Table 3, a number of such modifications were
well tolerated, compared to the simple carbon chain
compound of the same length, but did not significantly
improve potency in the cAMP assay. The most potent
compound in this series was still 13. Extension of the
carbon chain to as few as 3 carbons decreased activity,
but no further loss in activity was observed until the car-
bon chain was extended to 6 carbons. These data are
somewhat different to those previously reported for the
binding affinity of these compounds (Ki = 0.143 and
0.072 lM for compounds 23 and 26, respectively, in
We next turned our attention to a series of 5-membered
heterocyclic acids, again with our focus on those hetero-
cycles containing a nitrogen atom in the 3-position rela-
tive to the carboxylate to function as the hydrogen bond
acceptor to comply with our binding hypothesis. Again,
each of these heterocycles was commercially available as
either the ester or the free acid except 18 and 20 which
were synthesized as outlined in Scheme 1. As can be seen
from Table 2, the 1,3-orientation of the nitrogen relative
to the carboxylate appears to be required but is not suf-
ficient for activity at the receptor. The most potent com-
pounds in this series were 13 and the methyl thiazole
(14) acids. Incorporation of a sulfur in other positions
Table 2. GPR109a agonist activity for
heterocyclic acids
a
series of 5-membered
O
A3
OH
A2
A1
N
(H)
a
GPR109a EC50 (n)
Compound
A1
A2
A3
13
14
15
16
17
18
19
20
21
NH
CH
CH
CH
N
CCH3
CCH3
S
CH
S
0.46 (11)
0.5 (3)
1.5 (4)
n.e.
CH
N
S
S
CH
CH
CH
N
n.e.
n.e.
CH
O
CCH3
CCH3
CCH3
NCH3
1.3 (12)
n.e.
>20 (3)
O
CH
CH
n.e., less than 50% activity observed at any concentration up to 30 lM.
a EC50 determination from the cAMP assay.9, lM Values are means of
multiple determinations. The number of experiments is listed in
parentheses. The standard deviations where calculable were <30% of
the mean.
3
the H-niacin rat spleen binding assay12) and indeed,
our data are more comparable to the functional activity
(GTPcS binding) reported in the same study. In our

T. Gharbaoui et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4914–4919
4917
Table 3. GPR109a agonist activity for a series of 5-substituted
pyrazole acids
lin in CHO cells overexpressing the cloned human
GPR109a receptor. In addition, the efficacy of 44 in this
assay was comparable to both niacin and b-hydroxy
butyrate, which has been proposed to be a physiologi-
cally relevant ligand for the receptor,18 suggesting that
44 is a full agonist of the receptor under these condi-
tions. It is possible that the differences in the apparent
efficacy of compounds in the two systems may be due
to the higher level of expression of the receptor in the
cloned receptor assay versus the endogenously expressed
receptor used in the rat tissue assays. The only clear par-
tial agonists that we identified (showing that our assay
system was indeed capable of detecting a partial agonist
effect) were compounds 45 and 46 which showed efficacy
values of around 70–90% relative to niacin, although a
definitive assignment of partial agonism was con-
founded by the weak potency observed. All the other
compounds prepared which had measurable activity
and complete dose–response curves, had efficacy values
in the range of 95–100% both in the cloned human
and (where tested) in the cloned rat receptor assay.
O
OH
R
NH
N
a
GPR109a EC50 (n)
Compound
R
13
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
Me
Et
0.46 (11)
0.8 (6)
3.9 (3)
8.3 (8)
1.0 (2)
1.4 (5)
16.2 (5)
n.e.
n-Pr
i-Pr
c-Pr
n-Bu
c-Bu
CH(CH3)CH2CH3
CH2OCH3
CH2SCH3
CH2S(O)CH3
CH2S(O)2CH3
CH2NHCH3
CH2N(CH3)2
CH2CH2OCH3
CHO
8.7 (2)
3.7 (2)
4.9 (2)
n.e.
n.e.
n.e.
12.2 (2)
6.1 (3)
2.5 (3)
n.e.
In addition to the 5-substituted pyrazole series, we pre-
pared a number of 4,5-dialkyl substituted analogues. We
CH2OH
COCH3
n-Pentyl
n-Hexyl
t-Bu
1.8 (2)
5.6 (5)
n.e.
Table 4. GPR109a agonist activity for a series of bicyclic pyrazole
acids
Ph
CH2Ph
n.e.
O
8.9 (3)
2.8 (8)
6.2 (15)
10.5 (6)
A3
A2
CH2(3–F–Ph)
CH2(3–Cl–Ph)
CH2(3–Br–Ph)
A1
OH
NH
A2
N
n.e., less than 50% activity observed at any concentration up to 30 lM.
a EC50 determination from the cAMP assay.9, lM Values are means of
multiple determinations. The number of experiments is listed in
parentheses. The standard deviations where calculable were <30% of
the mean.
Compound A1
A3
GPR109a
EC50a, lM (n)
47
48
49
50
51
52
53
54
55
CH2
–CH2CH2–
CH2
CH2
CH2
CH2
CH2 0.86 (2)
CH2 n.e.
CH(CH3)
CH(CH3)2
CH2 8.3 (3)
CH2 n.e.
hands, branching of the side chain was not well toler-
ated, although interestingly we did observe that cyclo-
propyl analogue 25, was more active than the
isopropyl compound 24. However, the s-butyl 28 and
cyclobutyl 27 analogues were significantly less potent
than 13. Incorporation of ether or thioether groups or
a primary alcohol into the side chain was tolerated but
did not offer any improvement in activity. Oxidation
of the thioether to the sulfoxide, 31, was also tolerated,
but further oxidation to the sulfone 32 was not. The
incorporation of basic groups along the side chain was
not tolerated in any of the positions tried, further
emphasizing the lipophilic nature of the binding pocket.
Although only phenyl is shown in the table, a number of
other aromatic analogues were prepared including
substituted phenyl, thiophene, furan, pyrazole and pyr-
role analogues in multiple orientations, but all were
devoid of activity. The benzyl analogues, when substituted
in the 3-position with halogen atoms, did retain some
activity, with the 3-fluoro analogue 44 being the most
potent. It has been reported that this compound and a
number of other 5-substituted pyrazole-3-carboxylates
act as partial agonists in the rat adipocytes and spleen
GTPcS assay system.12 However, in our hands 44 was
able to fully reverse the cAMP elevating effect of forska-
CH(CH2CH2CH3) CH2
CH2 6.6 (2)
CH2 n.e.
CH(OCH2CH3)
CH(Ph)
CH2
CH2
CH2
CH
CH2 12 (3)
CH2 n.e.
CH2 n.e.
CH(SCH3)
CH2
(CH3)
O
56
57
58
59
60
61
62
63
64
65
66
67
CH2
CH2 2.7 (2)
CH2 n.e.
–CH2CH2–
CHCH3
CH(CH2CH3)
CH(CH2CH2CH3)
CH2
O
O
CH2 2.5 (5)
CH2 8.7 (1)
CH2 2.1 (2)
CH2 5.6 (4)
CH2 n.e.
O
O
S
CH(CH3)
CH2
CH2
S
SO2
CH2
NH
CH2 n.e.
S 2.2 (5)
CH2 n.e.
CH2
CH2
NCH2Ph CH2 n.e.
@CH
CH2 0.4 (2)b
CH@
n.e., less than 50% activity observed at any concentration up to 30 lM.
a EC50 determination from the cAMP assay.9 Values are means of
multiple determinations. The number of experiments is listed in
parentheses. The standard deviations where calculable were <30% of
the mean.
b Exists as a 1:1 mixture of 2 compounds; 1,4-dihydrocyclopen-
ta[c]pyrazole-3-carboxylic acid and 1,6-dihydrocyclopenta[c]pyra-
zole-3-carboxylic acid.

4918
T. Gharbaoui et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4914–4919
O
making use of a recombinant human receptor GPR109a
and a cAMP assay. Data with reported compounds in
our assay matched well with literature data from previ-
ously described endogenous rodent receptor assays. This
was further confirmed by testing of selected analogues in
the cloned mouse and rat receptor cAMP assays. That
these compounds have similar activity across species is
of significance for future in vivo studies. Furthermore,
an early exploration of the available heterocyclic chem-
ical space allowed us to formulate a binding hypothesis
which focused our efforts around aromatic carboxylates
with a hydrogen bond acceptor function in the 3-posi-
tion of the aromatic ring. This spatial arrangement
was shown to be required, but not sufficient for receptor
activity. From this information we were able eventually
to focus on a series of pyrazole acid derivatives. Further
elaboration of these compounds provided a series of 5,5-
fused pyrazoles which were suitable lead compounds for
further optimization.
O
O
, (i)
(ii)
O
N
N
H
(iii)
O
O
OH
67
(iv)
O
N
N
N
N
H
H
Scheme 2. Synthesis of Compound 67. Reagents and conditions: (i)
BF3ÆEt2O, reflux; (ii) a—(CO2Et)2, EtOH, KOtBu; b—H2NNH2; (iii)
Ph2O, 250 °C; (iv) 2.5 M NaOH.
observed however, that activity could only be retained if
the two alkyl or substituted alkyl groups were cyclized
to form a bicyclic ring system resulting in compounds
of the type shown in Table 4. All of these compounds
were prepared via our conventional synthesis outlined
in Scheme 1, with the exception of 63, which was pre-
pared by oxidation of the cyclic thioether 61 prior to
ester hydrolysis, 65, which was prepared via debenzylation
of 64, and 67. This latter compound was prepared using
a Diels–Alder and retro Diels–Alder protection strategy
(Scheme 2) as the 3,4-cyclohexeneone starting material
was prone to isomerization and/or polymerization and
we were unable to induce it to react using our conven-
tional procedure. Even following this sequence, the iso-
lated product 67 was found to exist as a mixture of
double bond isomers.
References and notes
1. Tavintharan, S.; Kashyap, M. L. Curr. Atheroscler. Rep.
2001, 3, 74.
2. Carlson, L. A. J. Int. Med. 2005, 258, 94.
3. Lorenzen, A.; Stannek, C.; Lang, H.; Andrianov, V.;
Kalvinsh, I.; Schwabe, U. Mol. Pharmacol. 2001, 59,
349.
4. Offermanns, S. Trends Pharmacol. Sci. 2006, 27, 384.
5. Wise, A.; Foord, S. M.; Fraser, N. J.; Barnes, A. A.;
Elshourbagy, N.; Eilert, M.; Ignar, D. M.; Murdock, P.
R.; Steplewski, K.; Green, A.; Brown, A. J.; Dowell, S. J.;
Szekeres, P. G.; Hassall, D. G.; Marshall, F. H.; Wilson,
S.; Pike, N. B. J. Biol. Chem. 2003, 278, 9869.
6. Tunaru, S.; Kero, J.; Schaub, A.; Wufka, C.; Blaukat, A.;
Pfeffer, K.; Offermanns, S. Nat. Med. 2003, 9, 352.
7. Tunaru, S.; La¨ttig, J.; Kero, J.; Krause, G.; Offermans, S.
Mol. Pharmacol. 2005, 68, 1271.
8. Semple, G. Niacin Receptor Agonists. MEDI-224, 233rd
A.C.S. National Meeting, Chicago, IL, Mar. 25–29, 2007.
9. Semple, G.; Skinner, P. J.; Cherrier, M. C.; Webb, P. J.;
Sage, C. R.; Tamura, S. Y.; Chen, R.; Richman, J. G.;
Connolly, D. T. J. Med. Chem. 2006, 49, 1227, Briefly,
GPR109a and GPR109b receptors with an N-terminal HA
tag were cloned into pcDNA3.1 (Invitrogen) and stable
cell lines were generated by G418 selection. Positive clones
were selected by anti-HA immunostaining. Compound
potencies were determined by the 96-well adenylyl cyclase
activation FlashPlateÒ assay from Perkin-Elmer as
described by the manufacturer. The cAMP assay was
optimized for the appropriate receptor stable clone, 5 lM
forskolin was used for stimulation and 50,000 cells were
used for each well. Positive controls were defined as cAMP
generated by cells without forskolin stimulation and
negative controls were defined as cAMP generated by
cells with 5 lM forskolin stimulation.
The bicyclic ring was relatively tolerant to the incorpo-
ration of heteroatoms with the exception of a basic
nitrogen group and the somewhat bulky sulfone.
Increasing the size of the ring from 5 to 6 atoms how-
ever, resulted in complete loss of agonist activity.
Methyl substitution was poorly tolerated in the 5-posi-
tion but several alkyl substitutions were tolerated in
the 6-position. Compounds 67 and 47 were among the
most potent compounds prepared thus far, despite the
former being a mixture of isomers, with activity compa-
rable to 13. The observation that the partially saturated
ring of the bicyclic system could tolerate substitution,
suggested to us that there was scope for further improve-
ments in activity leading us to select this 5,5-fused pyra-
zole series as our lead series of choice for further
optimization. We favoured this bicyclic series over the
simple 5-substituted pyrazoles, as the latter had some-
what less desirable physicochemical and pharmacoki-
netic properties, particularly as the lipophilicity of the
5-substituent was increased. Further optimization of this
lead series will be detailed in future communications.
10. Lorenzen, A.; Stannek, C.; Burmeister, A.; Kalvinsh, I.;
Schwabe, U. Biochem. Pharmacol. 2002, 64, 645.
11. Kupiecki, F. P.; Marshall, N. B. J. Pharmacol. Exp. Ther.
1968, 160, 166.
In summary, we have outlined our strategy for lead
identification of new agonists of GPR109a starting from
known small molecule compounds that were shown to
activate the receptor. In doing so, we have presented
the first comprehensive SAR study of such molecules
12. van Herk, T.; Brussee, J.; van den Nieuwendijk, A. M. C.
H.; van der Klein, P. A. M.; Ijzerman, A. P.; Stannek, C.;
Burmeister, A.; Lorenzen, A. J. Med. Chem. 2003, 46,
3945.
13. Pereira, J. N.; Holland, G. F. Experientia 1967, 23, 905.

T. Gharbaoui et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4914–4919
4919
14. Jung, J.-K.; Johnson, B. R.; Duong, T.; Decaire, M.; Uy,
J.; Gharbaoui, T.; Boatman, P. D.; Sage, C. R.; Chen, R.;
Richman, J. G.; Connolly, D. T.; Semple, G. J. Med.
Chem. 2007, 50, 1445.
15. Carlson, L. A.; Hedbom, C.; Helgstrand, E.; Sjo¨berg, B.;
Stjernstro¨m, N. E. Adv. Exp. Med. Biol. 1969, 4, 85.
16. Carlson, L. A.; Hedbom, C.; Helgstrand, E.; Sjo¨berg, B.;
Stjernstro¨m, N. E. Acta Pharm. Suec. 1972, 9, 289.
17. Seki, K.; Isegawa, J.; Fukuda, M.; Ohki, M. Chem.
Pharm. Bull. 1984, 32, 1568.
18. Taggart, A. K.; Kero, J.; Gan, X.; Cai, T. Q.;
Cheng, K.; Ippolito, M.; Ren, N.; Kaplan, R.; Wu,
K.; Wu, T. J.; Jin, L.; Liaw, C.; Chen, R.;
Richman, J.; Connolly, D.; Offermanns, S.; Wright,
S. D.; Waters, M. G. J. Biol. Chem. 2005, 280,
26649.