6620
P. J. Skinner et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6619–6622
Table 1. GPR109b agonist activity of selected 4-substituted amino-3-
nitro benzoic acids (5)a
Until recently, the only other reported ligand for the
GPR109b receptor, in addition to the very weak agonist
niacin, was acifran (2)9 (EC50 = 4.2 lM),1,10 an agent
previously shown to elevate HDL in rodents and hu-
mans.11 However, acifran lacks selectivity over
GPR109a (EC50 = 1.3 lM).1,10 Further studies on some-
what larger series of acifran analogs have shown a sim-
ilar lack of selectivity between the two receptors, despite
improvements in potency.12,13 We recently reported the
discovery of a series of 1-alkyl-benzotriazole-5-carbox-
ylic acids (3) as the first potent and selective agonists
of GPR109b.14 We herein describe the identification
and initial SAR around two related series of selective
agonist ligands for the GPR109b receptor which may
be useful tools with which to further explore the func-
tion of this receptor.
Compound R1
R2
GPR109b pEC50 (n)
5a
5b
5c
5d
5e
5f
Me
Et
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
5.72 0.11 (4)
6.64 0.22 (4)
7.53 0.13 (4)
7.00 0.31 (3)
6.80 0.19 (4)
6.55 0.13 (4)
6.59 0.19 (4)
6.82 0.07 (4)
7.24 0.12 (4)
6.51 0.09 (4)
6.70 0.15 (4)
6.51 0.10 (4)
6.44 0.14 (4)
5.61 0.10 (4)
5.51 0.23 (4)
6.23 0.18 (4)
5.85 0.15 (4)
6.51 0.14 (4)
5.69 0.23 (4)
n-Pr
Allyl
n-Bu
n-Pentyl
5g
5h
5i
i-Pr
2-Butyl
3-Pentyl
c-Pr
5j
5k
5l
c-Bu
c-Pentyl
5n
5o
5p
5q
5r
5s
5t
–CH(Me)CH(Me)2
Ph
Bz
–CH2CH2OMe
The 1-alkyl-benzotriazole-5-carboxylic acids (3) de-
scribed in our previous work14 were synthesized in three
steps from 4-fluoro-3-nitrobenzoic acid (4) via substitu-
tion of the aryl fluoride with a series of primary amines
to give the associated 4-substituted amino-3-nitroben-
zoic acids (5) (Scheme 1). Catalytic reduction to the
3,4-diamine, and formation of the benzotriazole via con-
densation with an azide source afforded the desired ben-
zotriazoles. However, during the course of this
investigation we discovered that the intermediate 4-ami-
no-3-nitrobenzoic acids also displayed significant activ-
ity at GPR109b while retaining selectivity against
GPR109a. As a result of this discovery, a wider range
of 4-substituted amino-3-nitrobenzoic acids were syn-
thesized and tested in a GPR109b cAMP whole cell
assay. Seventeen compounds were found to display ago-
nist responses with pEC50 greater than 6. Of these, four
compounds displayed agonist responses with pEC50 of 7
or greater. 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. As may
be seen from Table 1, the best activity was observed
for small linear alkyl groups, such as n-ethyl (5b) and
n-butyl (5e), with the greatest potency observed for the
n-propyl derivative (5c). Substitution a to the amine
was tolerated, albeit with modest reductions in poten-
cies, and with the greatest potency observed for the 3-
pentyl (5i) analog. Cyclic substituents (5j–l) were toler-
ated, but generally resulted in reduced potency com-
pared to the equivalent non-cyclized analogs.
Substitution b to the amine was also tolerated. Bis-N-
substituted analogs were tolerated with modest reduc-
tions in potency relative to the mono-N-substituted ana-
logs, with the exception of the pyrrolidine analog (5x)
CHMeCH2OMe
–CH2CH-[O(CH2)3]–
Me
Me
5u
5v
5w
5x
n-Pr 7.28 0.18 (3)
Et 6.20 0.09 (4)
n-Pr 6.92 0.07 (4)
Et
n-Pr
–[(CH2)4]–
<5 (4)
Errors are log SD. Compounds that showed no response are des-
ignated NA (not active). Compounds displaying only a weak response
at high concentration are designated <5. Accurate pEC50 values for
these compounds were not determined.
a Activities were measured from 30 pM to 100 lM and are provided as
the negative log of the molar value of EC50
.
which was essentially inactive. None of the compounds
prepared showed any activity at GPR109a at concentra-
tions up to 30 lM.
It was clear from this SAR that in addition to a modest
improvement in potency, the 4-substituted amino-3-
nitrobenzoic acid series allowed a greater range of sub-
stitution than the 1-alkyl-benzotriazole-5-carboxylic
acid series investigated earlier.14 The nitro benzene moi-
ety takes up a smaller footprint than the benzotriazole
scaffold, and we attributed this wider range of active
compounds to an increase in available space within the
binding site. We further hypothesized that a pyridyl ring
could make a suitable electronic isosteric replacement of
the nitro aryl moiety and avoid putative metabolic lia-
bilities in addition to providing a smaller footprint.
Thus, we decided to investigate a series of 6-amino
substituted nicotinic acids (6) as possible alternative
ligands for the GPR109b receptor.
The 6-amino substituted nicotinic acids were readily
obtained in a single step from 6-chloro-nicotinic acid
(7) (Scheme 2). Substitution of the chloro-nicotinic acid
by the relevant primary or secondary amine was carried
out in 2-propanol under microwave irradiation, albeit in
very low yield. The biological activity of each member of
the series was measured using the cAMP whole cell as-
say. Thirteen compounds were found to display agonist
responses with pEC50 greater than 6, and of these, four
displayed agonist responses with a pEC50 greater than 7.
Again, in each case the compounds were able to fully
Scheme 1. Reaction and conditions: (i) R1R2NH, H2O, NaHCO3,
150 °C, 20 min lW.