Discovery and optimization of novel fatty acid transport protein 1 (FATP1) inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.05.130

The discovery, optimization and structure-activity relationship of novel FATP1 inhibitors have been described. The detailed SAR studies of each moiety of the inhibitors combined with metabolite analysis led to the identification of the potent inhibitors 1

Full text of DOI:10.1016/j.bmcl.2012.05.130

Bioorganic & Medicinal Chemistry Letters 22 (2012) 5067–5070
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and optimization of novel fatty acid transport
protein 1 (FATP1) inhibitors
Tetsuyoshi Matsufuji a, Mika Ikeda b, Asuka Naito b, Masakazu Hirouchi c, Hideo Takakusa c,
Shoichi Kanda b, Masanori Izumi b, Jun Harada b, Tsuyoshi Shinozuka a,
⇑
a Lead Discovery & Optimization Research Laboratories I, Daiichi Sankyo Co., Ltd, 1-2-58 Hiromachi, Shinagawa-ku, Tokyo 140-8710, Japan
b Cardiovascular Metabolics Research Laboratories, Daiichi Sankyo Co., Ltd, 1-2-58 Hiromachi, Shinagawa-ku, Tokyo 140-8710, Japan
c Drug Metabolism & Pharmacokinetics Research Laboratories, Daiichi Sankyo Co., Ltd, 1-2-58 Hiromachi, Shinagawa-ku, Tokyo 140-8710, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery, optimization and structure–activity relationship of novel FATP1 inhibitors have been
described. The detailed SAR studies of each moiety of the inhibitors combined with metabolite analysis
led to the identification of the potent inhibitors 11p and 11q with improved blood stability.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 10 May 2012
Revised 30 May 2012
Accepted 31 May 2012
Available online 19 June 2012
Keywords:
Fatty acid transport protein
Acyl-CoA synthetase
High-throughput screening
Insulin resistance
Triazole
Fatty acid transport protein 1 (FATP1) is known as a transmem-
brane protein with Acyl-CoA synthase activity that is highly ex-
pressed in a skeletal muscle as well as in adipose tissue, brown
adipose tissue and heart.1 One of the important roles of FATP1 is
thought to be the uptake and metabolism of a fatty acid by con-
verting corresponding Acyl-CoA with an insulin signal. The previ-
ous report of a FATP1 deficient mouse showed the improvement
of high fat-induced insulin resistance as well as the reduction of
intramuscular accumulation of fatty acyl-CoA, suggesting that the
inhibition of FATP1 is an attractive therapeutic target for insulin
resistance.2 Herein we report the first discovery and structure–
activity relationship of novel triazole derivatives as FATP1 inhibi-
tors with improved blood stability.3
By high-throughput screening of our corporate library, several
FATP1 inhibitors were identified. Since the transportation of a fatty
acid by FATP1 is known to involve acyl-CoA synthesis, measuring
the inhibition of acyl-CoA synthetase (ACS) led to the identification
of structural related FATP1 inhibitors, triazole 1 and 2 (Fig. 1) with
moderate human in vitro activity. Unfortunately, both compounds
lack in vitro activity against mouse, and high blood stability cannot
be expected because of the existence of several metabolically
unstable functional groups. Thus, we initiated the derivatization
of 1 and 2 to improve in vitro activities against human and mouse
FATP1s avoiding metabolically labile functional groups.
Replacement of benzylthio group of 2 with aniline moiety of 1
led to the identification of 3a with improved human and mouse
in vitro activities by over ten times (Table 1). The inhibitory activ-
ities against human and mouse FATP1s are evaluated by measuring
acyl-CoA synthetase activities of human and mouse FATP1s. As
metabolically unstable ester moiety still remained in 3a, our efforts
were focused on modifying the substituent of the left phenyl ring
in compound 3a as indicated in Table 1. Meta derivative 3b proved
to decrease in vitro activities against both species. Methyl ester 3c
showed the substantial loss of in vitro activity against mouse,
whereas t-butyl ester 3d and carboxylic acid 3e lost the potencies
against both species. These results suggest that the proper size of
O
O
O
O
O
HN
S
N
N
S
S
N
H
N
H
N
N
N
N
1
2
humanFATP1 IC50 = 6.6 µM
mouseFATP1 IC50 >10 µM
humanFATP1 IC50 = 7.6 µM
mouseFATP1 IC50 >10 µM
⇑
Corresponding author.
E-mail address: shinozuka.tsuyoshi.s5@daiichisankyo.co.jp (T. Shinozuka).
Figure 1. HTS hits 1 and 2 with FATP1 IC50 values.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.05.130

5068
T. Matsufuji et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5067–5070
Table 1
Table 2
IC50 values of FATP1 inhibitors 1–3
IC50 values of FATP1 inhibitors 3p, 6–11
3
4
2
a
a
Compd
R
Human
FATP1 IC50
Mouse
FATP1 IC50
(lM)
Blood
stabilityb
(%)
Compd
R
Human FATP1 IC50
Mouse FATP1 IC50
M)
a
a
(lM)
(l
(lM)
1
2
6.6
7.6
0.33
>1
0.54
>1
>1
>1
>1
>1
>10
>10
1.1
3a
3b
3c
3d
3e
3f
3g
3h
3i
4-CO2Et
3-CO2Et
4-CO2Me
4-CO2t-Bu
4-CO2H
4-CONHMe
4-CONMe2
4-CONHEt
H
2.9
3p
6
0.40
0.40
1.0
1.5
28
NDb
>10
>10
>10
>10
>10
>10
>10
>10
>10
5.1
NDc
>1
>1
>1
>1
3j
3k
3l
4-Cl
4-OMe
4-SO2Me
4-CN
7
8
0.58
>1
2.0
1.9
3m
>1
3n
>1
>10
NDc
3o
3p
3q
3r
4-COMe
4-COEt
4-COn-Pr
4-COPh
>1
NDc
1.0
1.1
>10
0.40
0.23
>1
>10
a
Inhibition of recombinant human or mouse acyl-CoA synthetase activity of
FATP1 in an assay using 10
M [1-14C] oleic acid as a substrate in 100 mM Tris–HCl
NDc
l
9
>1
2.9
(pH 7.4), 30 mM NaCl, 0.5 mM ATP, 5 mM MgCl2, 0.5 mM coenzyme A, 0.05% Triton-
X 100, 2 mM dithiothreitol. The IC50 values represent the average of at least n = 2.
b
Not determined. 3c Inhibits 56% of FATP1 ACS activity at 10
Not determined. 3o Inhibits 63% of FATP1 ACS activity at 10
l
l
M.
M.
c
10
>1
>10
NDc
8.9
0
lipophilic substituent is required in this position to retain in vitro
potencies against human and mouse FATP1s. In addition, other
kinds of carbonyl derivatives such as amides 3f–h lost the activity.
The replacement of ester function to a hydrogen atom, chloride,
methoxy and methylsulfonyl groups resulted in the loss of
in vitro activities against both species. Ethyl tetorazole 3n, known
as ester bioisostere, is not active at all. Therefore, synthesis of sev-
eral ketone derivatives was decided after the identification of
methyl ketone 3o with a moderate mouse in vitro activity. Among
the ketone derivatives, benzophenone 3r lost the activities,
whereas ethyl ketone 3p and n-propyl ketone 3q retained fair hu-
man and mouse in vitro activities.
11a
11b
0.19
0.68
0.36
0.84
a
Inhibition of recombinant human or mouse acyl-CoA synthetase activity of
FATP1 in an assay using 10
M [1-14C] oleic acid as a substrate in 100 mM Tris–HCl
l
(pH 7.4), 30 mM NaCl, 0.5 mM ATP, 5 mM MgCl2, 0.5 mM coenzyme A, 0.05% Triton-
X 100, 2 mM dithiothreitol. The IC50 values represent the average of at least n = 2.
b
Remaining (%) of the test compound after 2 h incubation with rat blood.
Not determined.
c
Exceedingly high in vivo clearance of ethyl ketone 3p was ob-
served after intravenous administration to mouse, and high clear-
ance found to be mediated in large part by blood instability: the
blood stability of ethyl ketone 3p was 28%. The blood stability
was evaluated by measuring the remaining (%) of the parent com-
pound after 2 h incubation with rat blood. To understand the cause
of instability, the metabolites were analyzed. Ethyl ketone 3p
decomposed within 5 min after intravenous administration to
mouse. LC–MS/MS analysis of the blood sample indicated the exis-
tence of two main metabolites, carboxylic acid 4 and secondary
alcohol 5 (Fig. 2). This analysis clearly indicated that amide and ke-
tone moieties are metabolically unstable. Then we focused on the
replacement of these moieties as indicated in Table 2 and 3.
The replacement of amide moiety is summarized in Table 2.
Pyridine 6 retained in vitro potencies. Although, N-methyl amide
7 retained in vitro activities without the improvement of blood
stability, N-ethyl derivative 8 lost the potencies. As N-methyl
derivative retained in vitro activities, tetrahydroquinoline 9 and
dihydroindole 10 with ester functional group were prepared
because of the synthetic feasibility, and these compounds showed
reduced activities. Another fused pattern was then examined to
modify the amide bond in triazole 3p. This modification led to
the identification of benzoxazoles 11. In benzoxazole 11a, we
achieved not only replacement of the amide bond, but also an in-
crease in in vitro potencies against both species by several times.
Still the blood stability of benzoxazoles 11a is moderate, and
OH
O
O
HN
HN
N
N
S
S
HO
N
H
N
N
N
N
4
5
Figure 2. Two main metabolites of triazole 3p in rodent.

T. Matsufuji et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5067–5070
5069
Table 3
IC50 values of FATP1 inhibitors 11
H
N
HN
a
HS
N
H2N
N
H
6
5
O
N
N
12
13
O
R1
R1
b
Cl
N
NH2
H
15
Compd
R
n
Human FATP1
Mouse FATP1
IC50 (lM)
Blood
stabilityb
(%)
a
a
IC50
(
lM)
14
O
R1
c
HN
11a
11b
11c
11d
11e
11f
5-COEt
5-CO2Et
6-COEt
5-COi-Pr
5-COEt
5-CH(OH)Et
5-
C(@NHOH)Et
5-
C(@NHOMe)Et
H
5-OMe
5-Me
5-Et
5-F
5-Cl
0
0
0
0
2
0
0
0.19
0.36
0.84
1.2
0.52
11.9
3.7
8.9
0
S
N
N
0.68
0.90
0.60
1.67
>1
0.26
H
N
N
35
42
44
80
100
3a-d
O
11g
0.20
O
X
d
HN
S
N
3d
13
NDc
N
H
11h
0
>1
>10
N
N
11i
11j
11k
11l
11m
11n
11o
11p
11q
0
0
0
0
0
0
0
0
0
>1
>1
>1
>1
>1
>1
>1
>10
2.7
>10
1.8
7.8
2.3
5.4
0.36
0.091
NDc
NDc
NDc
NDc
NDc
92
3e: X = OH
e
3f-h: X = NR2R3
O
HN
f, g
e
S
N
6-10
NDc
NDc
100
HO
6-Cl
5-CF3
5-OCF3
N
N
0.36
0.10
4
a
Inhibition of recombinant human or mouse acyl-CoA synthetase activity of
FATP1 in an assay using 10
M [1-14C] oleic acid as a substrate in 100 mM Tris–HCl
(pH 7.4), 30 mM NaCl, 0.5 mM ATP, 5 mM MgCl2, 0.5 mM coenzyme A, 0.05% Triton-
Scheme 1. Reagents and conditions: (a) EtNCS, NaOH, THF, H2O, reflux, 57%; (b)
chloroacetyl chloride, KI, Et3N, THF, reflux; (c) 13, NaOAc, EtOH, reflux; (d) HCl,
EtOAc; (e) corresponding amine, WSCDÁHCl, HOBtÁH2O, Et3N, CH2Cl2; (f) t-butyl
bromoacetate, NaOAc, EtOH, reflux, 68%; (g) TFA, CH2Cl2, 53%.
l
X 100, 2 mM dithiothreitol. The IC50 values represent the average of at least n = 2.
b
Remaining (%) of the test compound after 2 h incubation with rat blood.
Not determined.
c
provided benzylchloride 18, whereas benzylchloride 22 was pre-
pared from 2-benzoxazolinone 19. Other benzoxazole derivatives
except for 11e–h were prepared in a similar manner.6,7 Sulfone
11e and alcohol 11f were synthesized via N-Boc protected interme-
diate 25, which was prepared from N-Boc aniline 23 in a similar
manner as described in Scheme 1. Then, oxidation of sulfur or
reduction of ketone in 25 followed by deprotection of the Boc
group provided sulfone 11e and alcohol 11f, respectively. Oxime
11g and 11h were prepared using the usual manner.
Trifluoromethyl derivatives 11p and trifluoromethoxy derivative
11q were selected for the evaluation of pharmacokinetics in rat.
Pharmacokinetic (PK) parameters of 11a, 11p and 11q were deter-
mined by orally (10 mg/kg) and intravenously (1 mg/kg) adminis-
tered rats. As expected, the very high clearance (378 mL/min/kg)
of ethyl ketone 11a was improved by the replacement of the ketone
group; and the clearance of 11p and 11q is 56.0, 52.0 mL/min/kg,
respectively. Similarly, Cmax and AUC values were improved: ethyl
ketone 11a was not observed in rat plasma, whereas a moderate
amount of 11p and 11q was observed in rat plasma.
obviously ethyl ketone moiety causes this blood liability. Therefore
further optimization of this position was required.
The synthesis of triazoles 3–10 is shown in Scheme 1. After syn-
thesizing thiol 13 in the usual manner,4 S-alkylation of versatile
thiol 13 with various chlorides 15, which were prepared from cor-
responding aniline and chloroacetyl chloride provided triazoles
3a–d. Amides 3f–h were synthesized by amidation of carboxylic
acid 3f prepared from t-butyl ester 3d. Triazoles 6–10 were synthe-
sized by amidation of carboxylic acid 4.
The replacement of the carbonyl group on benzoxazole ring in 11
was examined as shown in Table 3. Regioisomer 11c and isopropyl
ketone 11d retained human and mouse in vitro activities without
improving blood stability. The blood stability of sulfone 11e was
not improved. The metabolite analysis of 11a indicated the exis-
tence of metabolite 11f (data not shown). This racemic alcohol 11f
decreased in vitro potencies by over ten times with the improve-
ment of blood stability. Oxime 11g retained in vitro potencies with-
out blood instability, whereas methyl oxime 11h lost in vitro
activities. In addition, methoxy or alkyl derivatives lost the activi-
ties. Although the introduction of fluoro substituent resulted in
the reduction of mouse in vitro activity, a moderate in vitro activity
against mouse was observed in chloro derivative 11n. This encour-
aged us to synthesize and evaluate trifluoromethyl derivative 11p,
which possesses a good in vitro activities against human and mouse
FATP1s. Similarly, with retaining the excellent blood stability, triflu-
oromethoxy derivative 11q showed good human and mouse in vitro
In summary, we described the discovery and optimization
of triazole derivatives as the first FATP1 inhibitors. By the
Table 4
Rat PK parameters of 11a, 11p and 11q
a
a
a
Compd Tmax
Cmax
g/mL)
AUCa
(hr ⁄
mL)
T1/2
(h)
Clb (mL/
min/kg)
Vdb
(L/kg)
%F
(h)
(
l
l
g/
11a
11p
11q
LOQc
0.33
0.33
LOQc
0.22
0.19
LOQc
0.28
0.20
LOQc 378
2.69
1.86
1.73
NA
9.2
5.4
1.23
1.02
56.0
52.0
activities of 0.10 and 0.091 lM, respectively Table 4.
Benzoxazole derivatives 11 were synthesized by S-alkylation of
versatile thiol 13 with benzylchlorides as shown in Scheme 2. Re-
gio selective syntheses of 5-acyl benzoxazole 11a and 6-acyl benz-
oxazole 11e were accomplished using the Henichart method.5
Friedel-Cfafts acylation of amide 16, followed by ring closure
a
Average of two rats dosed at 10 mg/kg po in DMA/PG/(20%HPbCD/saline):
10/10/80.
b
Average of two rats dosed at 1 mg/kg iv in DMA/PG/(20%HPbCD/saline): 10/10/
80.
c
Under limit of quantification.

5070
T. Matsufuji et al. / Bioorg. Med. Chem. Lett. 22 (2012) 5067–5070
OH
OH
resistance model mice is underway and to be reported in the near
future.
O
b
Cl
NHR
N
H
O
References and notes
15: R = H
a
17
16: R = COCH2Cl
1. (a) Schaffer, J. E.; Lodish, H. F. Cell 1994, 79, 427; (b) Abumrad, N.; Coburn, C.;
Ibrahimi, A. Biochem. Biophys. Acta 1999, 1441, 4; (c) Bonen, A.; Miskovic, D.;
Kiens, B. Can. J. Appl. Physiol. 1999, 24, 515; (d) Kazantzis, M.; Stahl, A. Biochem.
Biophys. Acta 2012, 1821, 852.
2. (a) Kim, J. K.; Gimeno, R. E.; Higashimori, T.; Kim, H.-J.; Choi1, H.; Punreddy, S.;
Mozell, R. L.; Tan, G.; Stricker-Krongrad, A.; Hirsch, D. J.; Fillmore, J. J.; Liu, Z.-X.;
Dong, J.; Cline, G.; Stahl, A.; Lodish, H. F.; Shulman, G. I. J. Clin. Invest. 2004, 113,
756; (b) Wu, Q.; Ortegon, A. M.; Tsang, B.; Doege, H.; Feingold, K. R.; Stahl, A.
Mol. Cell. Biol. 2006, 26, 3455; (c) Wu, Q.; Kazantzis, M.; Doege, H.; Ortegon, A.
M.; Tsang, B.; Falcon, A.; Stahl, A. Diabetes 2006, 55, 3229.
3. Fatty acid transport protein 4 inhibitors are known. See: Blackburn, C.; Guan, B.;
Brown, J.; Cullis, C.; Condon, S. M.; Jenkins, T. J.; Peluso, S.; Ye, Y.; Gimeno, R. E.;
Punreddy, S.; Sun, Y.; Wu, H.; Hubbard, B.; Kaushik, V.; Tummino, P.; Sanchetti,
P.; Sun, D. Y.; Daniels, T.; Tozzo, E.; Balanic, S. K.; Ramana, P. Bioorg. Med. Chem.
Lett. 2006, 16, 3504.
4. (a) Shivarama, H. B.; Venkatramana, U. K. Farmaco 1992, 47, 305; (b) Abdel-Aal,
M. T.; El-Sayed, W. A.; Abdel Aleem, A. H.; El Ashry, E. S. H. Pharmazie 2003, 58,
788.
O
c
d
11a
N
Cl
O
18
O
OH
NHR
O
O
e, f
N
H
20: R = H
19
a
21: R = COCH2Cl
O
O
c
d
11c
N
Cl
5. Aichaoui, H.; Lesieur, I.; Henichart, J.-P. Synthesis 1990, 679.
6. The synthesis of 11p was performed as follows: Step 1. Synthesis of 4-ethyl-5-
[(phenylamino)methyl]-4H-1,2,4-triazole-3-thiol
(13)
A
mixture
of
22
phenylamino-acetic acid hydrazide (3.56 g, 21.6 mmol), ethyl isothiocyanate
(1.88 mL, 21.6 mmol) and 2 M NaOH (5.0 mL, 10.0 mmol) in THF(10 mL) was
stirred at reflux for 2 h. 1 M HCl was added to the cooled reaction mixture until
the mixture became pH 6. The mixture was extracted with CH2Cl2 (10 mL). The
organic layers were dried (Na2SO4), concentrated and purified by column
chromatography (hexane/EtOAc = 1:1) to provide 13 (4.23 g, 84%) as a white
solid. 1H NMR (500 MHz, CDCl3) d 1.39 (3H, t, J = 6.8 Hz), 4.08 (1H, brs), 4.15 (2H,
q, J = 7.3 Hz), 4.39 (2H, d, J = 5.9 Hz), 6.72 (2H, d, J = 8.8 Hz), 6.84 (1H, t,
J = 7.3 Hz), 7.23 (2H, t, J = 8.3 Hz), 10.28 (1H, brs). Step 2. Synthesis of N-{[4-
ethyl-5-({[5-(trifluoromethyl)-1,3-benzoxazol-2-yl]methyl}sulfanyl)-4H-1,2,
4-triazol-3-yl]methyl}aniline (11p) A mixture of 13 (0.026 g, 0.11 mmol), 2-
chloromethyl-5-trifluoromethyl-1,3-benzoxazole (0.024 g, 0.10 mmol), sodium
acetate (0.018 g, 0.31 mmol) in ethanol (5.0 mL) was stirred at reflux for 1 h. The
cooled reaction mixture was concentrated and purified by column
BOC
g, h, i
N
j
N
HS
NHBoc
N
N
24
23
O
BOC
O
k, l or m, l
N
N
S
11e or 11f
N
N
N
chromatography (hexane/EtOAc = 1:1) to provide 11p (0.033 g, 74%) as
a
25
white solid. 1H NMR (500 MHz, CDCl3) d 1.28 (3H, t, J = 7.4 Hz), 3.97 (2H, q,
J = 7.2 Hz), 4.44 (2H, s), 4.71 (2H, s), 6.71 (2H, d, J = 8.2 Hz), 6.78 (1H, t,
J = 7.0 Hz), 7.19 (2H, dd, J = 8.6, 7.4 Hz), 7.50 (1H, d, J = 8.6 Hz), 7.58 (1H, d,
J = 8.6 Hz), 7.92 (1H, s).; Anal. Calcd for C20H18F3N5OS: C, 55.42; H, 4.19; N,
16.16. Found: C, 55.20; H, 4.29; N, 15.92.
Scheme 2. Reagents and conditions: (a) Chloroacetyl chloride, Et3N, CH2Cl2; (b)
propionyl chloride, AlCl3, DMF; (c) PPTS, xylene, 150 °C; (d) 13, AcONa, EtOH, reflux;
(e) propionic acid, polyphospholic acid, 14%; (f) NaOH, H2O, reflux, 50%; (g) methyl
bromoacetate, NaOt-Bu, THF, reflux, 61%; (h) hydrazine monohydrate, EtOH, reflux,
91%; (i) EtNCS, NaOH, THF, H2O, reflux, 52%; (j) 22, NaOAc, EtOH, reflux; (k) mCPBA,
CH2Cl2, 73%; (l) HCl, EtOAc; (m) NaBH4, MeOH, 0 °C, 98%.
7. The synthesis of N-{[4-ethyl-5-({[5-(trifluoromethoxy)-1,3-benzoxazol-2-
yl]methyl}sulfanyl)-4H-1,2,4-triazol-3-yl]methyl}aniline (11q):
A mixture of
13 (0.276 g, 1.18 mmol), 2-chloromethyl-5-trifluoromethoxy-1,3-benzoxazole
(0.270 g, 1.07 mmol), sodium acetate (0.187 g, 3.22 mmol) in ethanol (10 mL)
was stirred at reflux for 1 h. The cooled reaction mixture was concentrated and
purified by column chromatography (hexane/EtOAc = 1:1) to provide 11q
optimization of several positions, benzoxazoles 11p and 11q were
identified with the improvement of both human and mouse
in vitro potencies as well as blood stability. Our strategy includes
removing each metabolic site which resulted in highly metaboli-
cally stable compounds. The in vivo evaluation with insulin
(0.44 g, 57%) as
a d 1.35 (3H, t,
white solid. 1H NMR (500 MHz, CDCl3)
J = 7.3 Hz), 4.04 (2H, q, J = 7.2 Hz), 4.51 (2H, s), 4.76 (2H, s), 6.78 (2H, d,
J = 8.3 Hz), 6.85 (1H, t, J = 7.3 Hz), 7.25 (3H, t, J = 7.8 Hz), 7.46 (1H, d, J = 9.3 Hz),
7.57 (1H, s); Anal. Calcd for C20H18N5O2S: C, 53.45; H, 4.04; N, 15.58. Found: C,
53.37; H, 4.05; N, 15.56.