Anthranilic acid replacements in a niacin receptor agonist

Article abstract of DOI:10.1016/j.bmcl.2010.04.001

Niacin is an effective drug for raising HDL cholesterol. However, niacin must be taken in large doses and significant side effects are often observed, including facial flushing, loss of glucose tolerance, and liver toxicity. An anthranilic acid was identified as an agonist of the niacin receptor. In order to improve efficacy and provide structural diversity, replacements for the anthranilic acid were investigated and several compounds with improved properties were identified.

Full text of DOI:10.1016/j.bmcl.2010.04.001

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3426–3430
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Anthranilic acid replacements in a niacin receptor agonist
a
a
a
a
Darby Schmidt a, , Abigail Smenton , Subharekha Raghavan , Hong Shen , Fa-Xiang Ding ,
Ester Carballo-Jane b, Silvi Luell b, Tanya Ciecko b, Tom G. Holt a, Michael Wolff a, Andrew Taggart b,
Larissa Wilsie b, Mihajlo Krsmanovic b, Ning Ren b, Daniel Blom b, Kang Cheng b, Peggy E. McCann b,
M. Gerard Waters a, James Tata a, Steven Colletti a
*
a Department of Medicinal Chemistry, Merck & Co., Rahway, NJ 07065, United States
b Department of Cardiovascular Disease, Merck & Co., Rahway, NJ 07065, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Niacin is an effective drug for raising HDL cholesterol. However, niacin must be taken in large doses and
significant side effects are often observed, including facial flushing, loss of glucose tolerance, and liver
toxicity. An anthranilic acid was identified as an agonist of the niacin receptor. In order to improve effi-
cacy and provide structural diversity, replacements for the anthranilic acid were investigated and several
compounds with improved properties were identified.
Received 23 February 2010
Revised 31 March 2010
Accepted 2 April 2010
Available online 8 April 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Niacin receptor
HDL
Atherosclerosis
Coronary disease
GPR109A
Anthranilic acid
Low HDL cholesterol levels are a risk factor for atherosclerotic
coronary disease, a leading cause of death in the developed world.
However, current strategies for raising HDL are limited.1,2 Exercise
and weight loss can afford modest HDL increases (3–9%). In addi-
tion, several pharmaceuticals also provide modest HDL increases.
For example, statins show a 5–15% HDL increase and fibrates show
a 10–25% increase.3 Niacin, a drug in use since the 1950s, shows
the most dramatic HDL elevation (20–35%), but its use has been
limited by side effects. Patients on niacin therapy experience in-
tense facial reddening called flushing, which causes low patient
compliance with treatment. Additionally, glucose tolerance is re-
duced in patients taking niacin, limiting its use in diabetic and pre-
diabetic patients. Liver toxicity can be a problem as well.4
Early efforts to improve upon niacin met with little success be-
cause its mechanism of action was not known. Recent renewed
interest in HDL raising drugs spurred efforts to identify niacin’s cel-
lular target, and a Gi coupled GPCR, GPR109A, also referred to as
HM74A or 19CD, was identified in fat cells. In vitro, this receptor
showed a high affinity for niacin.5 Animal studies involving
GPR109A were also undertaken. While it is not possible to show
HDL effects in mice because of their naturally high HDL levels, a
decrease in free fatty acids can be used as a surrogate marker for
HDL changes. With niacin treatment, a decrease in free fatty acid
is seen in humans, which is then followed by HDL changes. Upon
niacin treatment, GPR109A knockout mice did not show a decrease
in free fatty acids that is seen in wild type mice, further supporting
GPR109A as the likely target of niacin. Unfortunately, niacin recep-
tor knockout mice also failed to show a niacin induced flush, indi-
cating that flushing was in fact mediated through the niacin
receptor.
We felt that cloning of the receptor would allow for the identi-
fication of a novel GPR109A agonist. Furthermore, it was felt that
pharmacokinetics would be the key to reduced flushing and im-
proved HDL levels. Niacin itself is rapidly metabolized resulting
in a very short half life in man. Thus, a high dose of niacin (2–
4 g/day) must be taken by patients to achieve the maximum HDL
elevation (35%). A niacin receptor agonist with lower clearance
and a longer half life may show less severe flushing at an effica-
cious dose. We believed that a longer half life may also increase
the HDL levels achieved. The short half life of niacin led us to ques-
tion whether full 24 h coverage of the niacin receptor was being
achieved using the current treatment regime. Although the recep-
tor may not be fully activated, HDL increases of 25–35% are seen.
Perhaps, even greater changes in HDL could be obtained by com-
pounds with longer occupancy of the receptor.
Cloning of the niacin receptor allowed for high throughput
screening, and a structurally distinct molecule that activated
* Corresponding author.
E-mail address: darbyrschmidt@verizon.net (D. Schmidt).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.04.001

D. Schmidt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3426–3430
3427
GPR109A was identified. Further optimization led to the identifica-
tion of compound 6 (Table 1).6 We sought to modify the cyclohex-
enyl ring of 6 with the goal of improving the potency and
pharmacokinetic properties. Alkyl and aryl substitution in the
C4-, 5-, and 6-position were explored. This Letter discusses our
SAR efforts at the C4-position.
As shown in Scheme 1, compounds containing a C4 substituted
cyclohexenyl ring were synthesized from the corresponding ketone
1. Deprotonation with LiHMDS followed by quenching of the anion
with Mander’s reagent provided ketoester 2. Treatment of 2 with
sodium hydride at À78 °C followed by Comins reagent gave vinyl
triflate 3 as a single regioisomer.7 Palladium catalyzed coupling
of vinyl triflate 3 with amide 4 gave 5a. Hydrolysis of the ester pro-
vided the desired acid 5b.
All compounds were evaluated in a binding assay and a func-
tional assay.8 For the binding assay, membrane preparations were
made from CHO-KI cells stably expressing the niacin receptor. Dis-
placement of radiolabeled niacin from the membrane preparation
was measured in the absence and presence of human serum. For
the functional assay, the same membrane preparations were incu-
bated with compound and then
[ cS. Retention of
32S]GTP
[
32S]GTP
cS on the membrane preparation was measured to deter-
mine activation of the G proteins coupled to the niacin receptor,
generating EC50 values. In this assay, the full response to niacin
was defined as 100% and was used as a reference for other
compounds.
We first synthesized a series of C-4 alkyl substituted cyclohex-
enyl compounds. A methyl group in the C4-position, resulted in a
Table 1
Binding and functional activity of compounds for niacin receptor
H
N
OH
R
O
OH
O
Compound
R =
h-3H-Niacin IC50 (nM)
hGTP
c
S EC50 (nM)
Serum shift fold change
1
NIACIN
H
140
8
1000
76
6
7
8
Me
Pr
43
8
430
10
>400
9
Bu
6
30
10
11
12
Ph
34
420
7% at 25
92
nd
5300
OH
CN
F
lM
13
11
31
>400
>400
F
14
15
7.4
8.6
28
130
F
F
16
17
3
11
>400
F
F
52
410
100
90
N
N
F
18
19
16
18
64
94
N
F3C
Me
N
N
20
21
14
480
200
Me
N
N
120
2650
>400
F
N
NH
22
36
137
O
O
OH
OH
Compounds containing chiral centers are racemic unless otherwise indicated. Binding to niacin receptor determined by displacement of [3H]niacin. Functional assays were
run using binding of hGTP S to membrane preparations. Data are an average of two independent titrations having calculated standard errors below 15%. The assay-to-assay
c
variation was generally 2-fold.

3428
D. Schmidt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3426–3430
O
O
OTf
The 2-fluoro 3-pyridine compound, 17, had comparable binding
H
N
R2
CO2Me
CO2Me
activity to compound 10, but a fourfold decrease in functional
activity with a 100-fold serum shift. The 4-fluoro 3-pyridine com-
pounds 18 had activity comparable to compound 10, but had only
a 90-fold serum shift. The methyl pyrazole analogs 20 and 21 were
also synthesized. Compound 20 had good binding activity (compa-
rable to compound 10), but poor functional activity and large ser-
um shift (200-fold), while the regioisomer, 21, had poor activity
and >400-fold serum shift.
Having identified substituted cyclohexenyl compounds with
good activity and an improved serum shift, we investigated the
pharmacokinetic properties of this series. Along with the naph-
thol sidechain, we had also identified a biphenol side chain, B,
as a potent group in the anthranilic acid series. When this side
chain was coupled with the 2-fluoropyridine cyclohexenyl group
to provide 22, a threefold improvement in functional activity over
compound 17 was obtained. Therefore, compound 22 was se-
lected to be evaluated for its pharmacokinetics in mice (Table 2).
It had a modest half life (1.8 h) and low clearance (5.4 mL/min/
Kg) with poor oral bioavailability (12%). This profile was compa-
rable to the reference anthranilic acid analog 23, indicating that
the substituted cyclohexenyl ring had a relatively neutral effect
on pharmacokinetics.
Next, we combined the substituted cyclohexenyl group with
an oxadiazole hydroxypyridine side chain C (Scheme 1) that
had previously demonstrated reduced serum shift and good
in vitro activity in the anthranilic acid series.8 This allowed us
to explore whether the combination of the substituted cyclohex-
enyl head group and the oxadiazole hydroxypyridine side chain
would have additive effects on potency and/or better pharmaco-
kinetics. However, the substitutions that had improved potency
in combination with side chain A had no effect on compounds
with side chain C (Table 3, compounds 25–27). The propyl com-
pound 25, 3-fluorophenyl compound 26, and 4-fluoro-3-pyridine
compound 27 had binding and functional activity comparable to
the unsubstituted cyclohexenyl compound 24 and had no
improvement in serum shift. Interestingly, disubstituted cyclo-
hexenyl compound 28 had greater activity than the unsubstituted
cyclohexenyl compound 24, achieving an EC50 of 6 nM in the
functional assay.
O
a
b
c
O
+
H2N
R2
OR3
R1
R1
1
R1
R1
O
2
3
4
5a R3 =Me
5b R3 =H
d
R2=
OH
A
B
OH
F
N
N
N
N
OH
O
R4
O
N
N
D
C
R4=H, NH2
Scheme 1. Reagents and conditions: (a) LiHMDS CNCO2Me, THF, À78 °C to rt, 90%;
(b) NaH, Comins reagent, THF, À78 °C to rt, 90%; (c) Pd2dba3, Xanthphos, Cs2CO3,
dioxane, 90 °C; (d) LiOH, MeOH/THF, 5–70%.
fivefold decrease in binding and functional activity (compound 7,
Table 1), but larger alkyl chains, including propyl and butyl (com-
pounds 8 and 9), showed similar binding activity compared to 6,
along with a 2–7-fold improvement in the functional assay. Polar
substituents such as OH and CN (11 and 12) resulted in a 50-fold
decrease in binding activity.
Installation of a phenyl group (compound 10) in the C4-position
of the cyclohexene resulted in a fourfold decrease in binding activ-
ity but similar functional activity compared to compound 6 (Ta-
ble 1). We then explored substitution on the phenyl ring to
determine if further enhancements in potency could be achieved.
The 2- and 3-fluoro phenyl substituted analogs (13 and14) had a
threefold improvement in the binding and functional assay com-
pared to 10. The 4-fluoro phenyl substituted analog 15, on the
other hand, showed no improvement compared to 10. The 2,5-
difluorophenyl compound 16 was synthesized next and had a 9–
10-fold improvement in binding activity and functional activity
compared to compound 10.
While compounds 8, 13, 14, and 16 had excellent activity in
both the binding and functional assays in the absence of serum,
in the presence of serum the activity decreased by 400-fold. As this
was a concern, we sought to replace the phenyl ring with more po-
lar groups with the goal of reducing the serum shift. Thus, a series
of fluorinated 3-pyridines were synthesized (compounds 17–19).
We next sought to determine the effect of the substituted cyclo-
hexenyl groups in combination with side chain C on pharmacoki-
netics (Table 2). Therefore, we decided to investigate the
pharmacokinetics of compound 29, the mono fluoro analog of com-
pound 28. Compound 29 had comparable clearance to the unsub-
Table 2
Pharmacokinetics of niacin receptor agonists
Compound
Clp (mL/min/kg)
Vdss (L/kg)
t1/2 (h)
AUC (
lM h kg/mg)
Cmax
(
lM)
F (%)
Mouse
R
NH
OH
O
N
F
22
23
5.4
8.7
0.42
0.78
1.8
2.4
6.8
4.0
1.2
4.0
12
35
O
OH
O
OH
24
29
31
35
2.0
5.6
0.5
2.3
0.21
0.76
1.4
1.3
3.9
14
3.0
0.16
82
3.9
0.3
16
13
2
87
44
Rat
2.4
4.8
13
3.8

D. Schmidt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3426–3430
3429
Table 3
tional assay. Therefore, for the next round of compounds, we
sought a side chain with excellent pharmacokinetics that was in
need of improvement in the binding and functional activity (Ta-
ble 4). Side chain D, with a fluoropyridine in place of the
hydroxypyridine moiety, conferred excellent pharmacokinetics,
but poor potency and large serum shifts (>400-fold). Compound
31 with an unsubstituted cyclohexenyl group and an amine in
the linker region was evaluated for its pharmacokinetics in mice
and had low clearance (0.5 mL/min/Kg), a long half life (14 h),
and good bioavailability (87%). As hoped, the combination of side-
chain D with several substituted cyclohexenyl groups provided
dramatic improvements in potency. The 3-fluorophenyl com-
pounds 32 and 33 showed twofold improvements in functional
activity and the 3,5-difluorophenyl compound 34 and 2,3,5-triflu-
orocompounds 36 and 37 provided 20- and 10-fold improve-
ments in the functional assay compared to compounds 30 and
31. In addition, compounds 34 and 35 showed improved serum
shift. On the other hand, a heterocyclic substituted head group
exemplified by fluoropyridine 38, as well as the disubstituted
cyclohexenyl group in 39, were poorly tolerated in combination
with side chain D.
Binding and functional activity of compounds for niacin receptor
H
N
O
N
N
R
N
O
OH
OH
O
Compound R =
h-3H-
Niacin EC50
hGTPcS
IC50
(nM)
(nM)
NIACIN
140
4
1000
21
24
25
H
Pr
4
17
F
26
27
7
9
27
98
N
F
N
H
N
O
N
F
F
OH
N
28
29
12
8
6
O
OH
Compound 35 was selected for pharmacokinetic studies in a rat
(Table 2). Gratifyingly, compound 35 had low clearance (2.3 mL/
min/Kg), good oral AUC (13.2
lM h), good bioavailability (44%),
and a long half life (4.8 h). In compound 35, we have identified a
potent GPR109A agonist with good pharmacokinetics.
F
O
N
H
N
O
N
F
OH
N
42
O
OH
O
In summary, the effects of substituted cyclohexenyl groups on
potency and pharmacokinetics were investigated. The substituted
cyclohexenyl groups had a neutral or slightly detrimental effect
on pharmacokinetics, but showed significant potency improve-
ments. This allowed for the use of a side chain with excellent phar-
See legend of Table 1.
stituted compound 24 and a longer half life (3.9 h), but also had
much lower oral AUC and bioavailability (%F = 2).
macokinetics that was hindered by
a lack of potency. The
combination of side chain D with a difluoro substituted phenyl
cyclohexenyl head group provided compound 35, which had the
desired potency and pharmacokinetic properties.
Although the results with the oxadiazole hydroxypyridine side
chain C were disappointing, they did indicate that a substituted
cyclohexenyl could provide significant improvements in the func-
Table 4
Binding and functional activity of compounds for niacin receptor
H
N
R2
O
N
R1
N
N
O
F
OH
O
Compound
R1
=
R2
=
h-3H-Niacin IC50 (nM)
hGTP
c
S EC50 (nM)
Serum shift fold change
NIACIN
30
31
140
35
32
1000
490
460
H
H
NH2
>400
>400
32
33
H
NH2
57
85
230
250
F
34
35
H
NH2
19
28
22
100
30
30
F
F
F
F
36
37
H
NH2
9
15
45
46
F
N
38
H
H
280
6
7400
75
F
F
F
H
N
O
N
39
N
O
N
F
OH
F
O
See legend of Table 1.

3430
D. Schmidt et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3426–3430
6. Raghavan, S.; Tria, G. S.; Shen, H.; Ding, F.; Taggart, A.; Ren, N.; Wilsie, L.;
References and notes
Krsmanovic, M.; Holt, T.; Wolff, M.; Waters, M. G.; Hammond, M.; Tata, J.; Colletti,
S. Bioorg. Med. Chem. Lett. 2008, 18, 3163; Shen, H.; Ding, F.; Raghavan, S.; Deng, Q.;
Luell, S.; Forrest, M.; Carballo-Jane, E.; Wilsie, L.; Krsmanovic, M.; Taggart, A.; Wu,
K.; Wu, T.; Cheng, K.; Ren, N.; Cai, T.; Chen, Q.; Wang, J.; Wolff, M.; Tong, X.; Holt, T.;
Waters, M. G.;Hammond, M.; Tata, J.;Colletti, S. J. Med. Chem. 2010, 53, 2666; Shen,
H.; Ding, F.; Deng, Q.; Wilsie, L.; Krsmanovic, M.; Taggart, A.; Carballo-Jane, E.; Ren,
N.; Cai, T.; Wu, T.; Wu, K.; Cheng, K.; Chen, Q.; Wolff, M.; Tong, X.; Holt, T.; Waters,
M. G.; Hammond, M.; Tata, J.; Colletti, S. J. Med. Chem. 2009, 52, 2587; Shen, H.;
Colletti, S. L. Expert Opin. Ther. Patents 2009, 19, 957.
1. Knopp, R. Am. J. Cardiol. 2000, 86, 51; Etchason, J. A.; Miller, T. D.; Squires, R. W.;
Allison, T. G. Mayo Clin. Proc. 1991, 66, 23.
2. Barter, P.; Kastelein, J.; Nunn, A.; Hobbs, R. Atherosclerosis 2003, 168, 195. and
references therein.
3. Ashen, M.; Blumenthal, R. N. Eng. J. Med. 2005, 353, 1252.
4. Etchason, J. A.; Miller, T. D.; Squires, R. W.; Allison, T. G. Mayo Clin. Proc. 1991, 66,
23.
5. Tunaru, S.; Kero, J.; Schaub, A.; Wufka, C.; Blaukat, A.; Pfeffer, K.; Offermanns, S.
Nat. Med. 2003, 9, 352; Soga, T.; Kamohara, M.; Takasaki, J.; Matsumoto, S.; Saito,
T.; Ohishi, T.; Hiyama, H.; Matsuo, A.; Matsushime, H.; Furuichi, K. Biochem.
Biophys. Res. Commun. 2003, 303, 364; 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.
7. Comins, D.; Dehghani, A. Tetrahedron Lett. 1992, 33, 6299; Watanabe, N.; Tanino,
K.; Kuwajima, I. Tetrahedron Lett. 1999, 40, 8133.
8. Shen, H.; Ding, F.; Luell, S.; Forrest, M.; Carballo, J.; Wu, K.; Wu, T.; Cheng, K.;
Wilsie, L.; Krsmanovic, M.; Taggart, A.; Ren, N.; Cai, T.; Deng, Q.; Wang, J.; Wolff,
M.; Tong, X.; Holt, T.; Waters, G. M.; Hammond, M.; Tata, J.; Colletti, S. J. Med.
Chem. 2007, 50, 6303.