4
208
P. J. Skinner et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4207–4209
reported ligand for the GPR109b receptor, in addition to the very
1
0
11
weak agonist niacin, was acifran (2) (EC50 = 4.2
lM). Acifran
1
2
has been shown to elevate HDL in rodents and humans, however
acifran lacks selectivity over GPR109a (EC50 = 1.3
11
l
M). Additional
acifran analogs have shown a similar lack of selectivity between
the two receptors, despite improvements in potency.1
1,13
We
initially reported the discovery of a series of 1-alkyl-benzotria-
zole-5-carboxylic acids (3) as the first potent and selective agonists
4
of GPR109b. Subsequently we also reported a series of 6-aminon-
icotinic acids (4) as isosteres of the 4-substituted amino-3-nitro-
benzoic acids (5) that were active intermediates on the route to
1
4
3.
These two series of compounds showed improved potency
compared to the benzotriazole series. We hypothesized that
-aminopyrazole-3-carboxylic acid may provide a reasonable
5
replacement for the 6-aminonicotinic acid moiety and thus
decided to investigate this motif as potential ligands for GPR109b
(Fig. 1).
Mono- or symmetrically di-substituted 5-aminopyrazole-3-
carboxylic acids (6a–r) were synthesized in 4 steps from 5-nitropy-
razole -3-carboxylic acid (7) (Scheme 1). Protection of 5-nitropyraz-
ole-3-carboxylic acid (7) as the ethyl ester (8) and subsequent
reduction under a hydrogen atmosphere with palladium on carbon
gave the intermediate 5-aminopyrazole-3-carboxylic acid ethyl
ester (9). Reductive amination with an appropriate aldehyde pro-
vided the mono- or symmetrically di-substituted 5-aminopyra-
zole-3-carboxylic acid ethyl ester (10), which was subsequently
hydrolyzed under acidic or basic conditions to give the desired
mono- or symmetrically di-substituted 5-aminopyrazole-3-carbox-
ylic acid (6).
Scheme 1. Reaction conditions: (i) AcCl, EtOH, 80 °C, 18 h; (ii) Pd/C (10%), EtOH, H
(1 atm), 25 °C, 18 h; (iii) RCHO, NaBH(OAc) , DCE, 60 °C, 18 h (mono addition) or
170 °C 20 min lW (di addition); (iv) 1:5:1 MeOH/THF/1 M (aq) LiOH or HCl.
2
3
observed, a lack of potency against GPR109a prevented accurate
measurement of selectivity for 6k and 6n, however 6o displayed
an agonist response against GPR109a around 3000 times less than
that for GPR109b (Table 1).
Interestingly, the incorporation of additional functionality onto
the thiophenyl or benzyl moieties led to complete, or near
Agonist dose responses for Gi-coupled GPR109a and GPR109b
were generated using a cAMP Homogenous Time-Resolved Fluores-
cence (HTRF) assay in CHO stable cell lines. Positive controls were
0
complete, loss of potency. In addition, replacement of the 3 -thio-
0
phenyl moieties with the smaller 3 -furanyl moieties also led to
defined from the amount of cAMP generated by 5
stimulated cells with a GPR109a or GPR109b agonist, namely niacin
1) (pEC50 = 7.57) or 6-(allylamino)nicotinic acid (pEC50 = 7.38)
respectively. Negative controls were defined as cAMP generated by
M forskolin stimulated cells.
It was immediately obvious that although some mono-func-
tionalized amines (6a, 6b) displayed weak agonist responses
pEC50 = 5.35, 5.17), a significant increase in potency was observed
l
M forskolin
near complete loss of potency.
0
In an attempt to investigate if the 3 thiophenylmethyl moiety
(
was optimal for both amine substitution positions we prepared a
small series of non-symmetrically disubstituted amines (6s–w) in
three steps from ethyl 5-amino-1H-pyrazole-3-carboxylate (9)
(Scheme 2). Reductive amination with a single equivalent of thio-
phene-3-carbaldehyde gave ethyl 5-(thiophen-3-ylmethylamino)-
5
l
(
1H-pyrazole-3-carboxylate (11) which was then subjected a
for some of the symmetric disubstituted amines (6d–r). Specifi-
cally the dibenzyl (6j, pEC50 = 6.07) and bis-2 -thiophenylmethyl
second orthogonal reductive amination to provide the non-sym-
metrically di-substituted amine ester (10s–w). Subsequent hydro-
lysis then gave the desired non-symmetrically di-substituted
amines (6s–w). However, the only examples which displayed ago-
nist responses with pEC50 greater than 6 were the benzyl (6s,
0
(
6n, pEC50 = 6.59) analogs displayed agonist responses with pEC50
0
greater than 6, and the di-3 thiophenylmethyl (6o, pEC50 = 8.50)
analog displayed the greatest agonist response we had hitherto
observed. In each case the compounds were able to fully reverse
the cAMP elevating effect of forskolin in stably transfected CHO-
K1 cells, suggesting that they are likely to be full agonists of the
receptor. In addition, significant selectivity over GPR109a were
0
pEC50 = 7.84) and 2 thiophenylmethyl (6u, pEC50 = 6.85) analogs
with neither showing improved potencies over the symmetrically
0
di-3 thiophenylmethyl analog (6o).
In summary, a series of N,N-difunctionalized 5-aminopyra-
zole-3-carboxylic acids were prepared that displayed excellent
in vitro agonist activity at GPR109b in a whole cell cAMP assay.
The observed activity was highest for the dibenzyl amine, and
for analogs utilizing thiophene as an isosteric replacement for
the phenyl moiety. Optimal potency was observed for the
0
0
di-3 thiophenylmethyl amine. Replacement of one of these 3 -
0
0
0
thiophenyl moieties with benzyl, 2 thiophene or 2 - or 3 -furan
led to a decrease in potency, All active substances displayed a
full agonist effect as compared to niacin (1) or 6-(allylamino)nic-
otinic acid. None of the compounds prepared displayed any
significant activity against the closely related receptor GPR109a,
with the exception of the monobenzyl analog (6c) which was
inactive at GPR109b and which may prove to be a very interest-
ing starting point for the further elaboration of new GPR109a
agonists. We believe that the further development of selective
GPR109b agonists, such as those described herein, is essential
to further explore the therapeutic utility of this receptor.
Figure 1. Ligands for GPR109a and GPR109b.