H. M. Kim et al. / Bioorg. Med. Chem. Lett. 23 (2013) 6410–6414
6413
Table 4 (continued)
Compd
R
hDGAT1 IC50 (nM)
mDGAT1 IC50 (nM)
134
O Cl
N
N
1
7j
N
46
HN
1
7k
3870
180
67
N
N
1
7l
N
N
Cl
1
7m
2269
N
inhibitors (Table 1), suggesting that oxazole amide was the best
moiety in region A to maintain the inhibitory activity.
Then, we shifted our focus on developing SAR in region B. We
discovered that 4-substituted piperidine and piperazine produced
quite potent compounds (Table 2). Ethyl piperidine-4-carboxylate
derivative 4a possessed similar IC50 as 1 while homologated ester
(Scheme 3). Further functionalization of 15 was carried out
by deprotection of the Boc group followed by carbamate forma-
tion to yield 6a. Again, several other reactions including urea
formations and sulfonamide formations were also carried out
on 16 in the preparation of other compounds described in this
communication.
4
b was slightly less potent. Methyl sulfonamide 4c displayed even
To further probe the SAR of this new scaffold, we decided to
have region A fixed as piperidinyloxazole as in 6a and began re-
exploring region B (Table 4). Among the piperazinyl carbamate
compounds (6a, 17a, 17b, and 17c), branched alkyl carbamates
17a and 17b turned out to be more potent in the series (hDGAT1
IC50 41 nM and 79 nM, respectively). In general, less inhibitory
activities were obtained for piperazinyl amide compounds (17d,
17e, and 17f), although SAR trends similar to the piperazinyl carba-
mate series were observed. Contrary to the carbamate and amide
cases, piperazinyl alkyl ureas did not follow a similar trend in
SAR. Thus, branched alkyl ureas 17g and 17h were less active,
while aryl urea 17i was one of the most potent in the series
(hDGAT1 IC50 41 nM). Homopiperazinyl urea 17j was also quite ac-
tive and demonstrated flexibility of the region to accommodate
some small changes. In addition, several aryl piperidine and aryl
piperazine compounds were prepared (17k, 17l, and 17m). In gen-
eral, aryl piperazine compounds exhibited better potencies than
aryl piperidine compounds and aryl piperazine 17m showed the
best potency (hDGAT1 IC50 67 nM) among these types. For selected
cases, mouse DGAT1 potency was measured as well and 17a and
17m showed some selective inhibition favoring human DGAT1
over mouse DGAT1.
less inhibitory activity. However, ethyl carbamate 4d and urea ana-
log 4e showed potent inhibition of human DGAT1 enzymatic
activities.
With newly gained insights on SAR trends in region B, we began
to explore possible modifications in region A to further improve
potency and PK. After extensive efforts to find suitable modifica-
tions, we discovered that the C2-phenyl on the oxazole could be re-
placed with saturated heterocycles such as piperidine without loss
of inhibitory activity (Table 3). Replacement of the phenyl group
with pyrrolidine (compound 5a) or morpholine (compound 5b)
showed significant loss of potency (hDGAT1 IC50 1708 nM and
6
870 nM, respectively). However, piperidine replacement was
more successful, producing potent compounds 5c and 5d (hDGAT1
IC50 66 nM and 93 nM, respectively). As noted in the case of 5b and
5
e, not all six member saturated ring compounds were active and
future SAR optimization potentials were observed. For substituted
piperidines, 2-, 3-, and 4-methylpiperidine analogs 6b, 6c, and 6d
were prepared and compared with unsubstituted piperidine analog
6
a. Parallel to the case of 5c and 5d, 6a and 6b were similar in po-
tency (hDGAT1 IC50 1508 nM and 938 nM, respectively). Further
improvement of the potency was observed for 6c and 6d with 3-
methylpiperidinyloxazole 6c being the most potent (hDGAT1 IC50
As shown for the representative compounds 17a and 17i
(Table 5), PK properties of this series of compounds were quite
acceptable at the hit to lead stage. These compounds have high
Rat AUC and good solubility. As for other properties, their clogP
7
9 nM).
The synthesis of the representative compound 6a commenced
from the combination of ethyl 4,4,4-trifluoro-2-chloroacetoacetate
2
2
23
(7) and urea, which resulted in the formation of 8 (Scheme 1).
values were in the acceptable range (clogP < 5) , even though
Modified Sandmeyer conditions were utilized to convert 8 to 9,
which was a key intermediate for the syntheses of several
molecular weights are slightly higher than desired. The Cyp P450
profiles of the series were also acceptable and so were their hERG
2
-aminooxazole compounds. In the particular case of 6a, 9 was
inhibitions. On the other hand, human hepatocyte clearance for
6
treated with piperidine to provide 10, which was successfully con-
verted to 11 upon basic hydrolysis.
these compounds were high (24.2
l
L/min/10 cells for 17a and
6
16.9 lL/min/10 cells for 17i). Another potential area for improve-
The preparation of the other key intermediate 14 was initiated
by treatment of 2-chloro-5-nitropyridine (12) with 1-Boc
piperazine to provide 13 (Scheme 2). Alternatively, other amines
in place of 1-Boc piperazine were successfully employed for the
syntheses of other compounds in this study. The nitro group of
ment was plasma protein binding as both compounds exhibited
greater than 99% binding. Overall, the 2-aminooxazole amide
series has been shown to possess suitable PK properties as a lead
series.
Selected compounds from this series were tested in a mouse
postprandial triglyceridemia (PPTG) assay to determine in vivo
1
3 was reduced to give aminopyridine 14.
2
4,25
HATU mediated coupling of acid 11 and aminopyridine 14
activity (Table 6).
The compounds were orally dosed (3 mg/
in the presence of catalytic amount DMAP provided 15
kg or 10 mg/kg) to mice fasted overnight and treated with an oral