X. Du et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6218–6223
6219
H
N
B
and six-membered heterocyclic D-rings were tolerated with good
potency and efficacy, such as imidazole, pyrazole, pyridine and
pyrimidine. However, all of them displayed high turnover in rat
O
H
N
B
O
O
S
O
NH2
H
N
D
N
H
9
A
N
N
H
and human microsomal stability assays.
N
A
N
C
The benzyl-like nature of the carbon between the amino group
and the heterocyclic D-ring may be one explanation as to why N-
dealkylation was so facile. To mitigate this effect, the linker was
elongated to two carbons (compound 13 in Table 2). The potency
and efficacy were both significantly lowered, while no improve-
ment in microsomal stability was seen. The heteroaromatic D-ring
was also replaced with saturated heterocycles as in compounds
14–17. Although the microsomal stabilities of compound 14 and
17 were improved, possibly due to the presence of polar groups
such as a sulfone or amide, none of the compounds was sufficiently
potent.
C
1
2
h-GPR142 EC50 (Emax
4.8 µM (113%)
)
h-GPR142 EC50 (Emax
93 nM (120%)
)
Scheme 1. Increased potency through N-substitution.
phenylpropanoic acid 9 was converted to the acid fluoride with
cyanuric fluoride and coupled with compound 8 to form an amide
bond. Subsequent deprotection of the acetate group generated (R)-
2-hydroxy-N-(6-methoxy-[3,40-bipyridin]-5-yl)-3-phenylpropana-
mide 10. Compound 10 was converted to its nosylate upon
treatment with 4-nitrobenzene-1-sulfonyl chloride (nosyl chlo-
ride). Nucleophilic displacement of the nosylate with various
amines followed by deprotection under acidic conditions gener-
ated the final GPR142 agonists.
Compounds 25 and 26 were made through alkylation of the B-
ring pyridinone nitrogen of compound 22 with either tert-butyl 2-
bromoacetate or 2-chloro-N,N-dimethylethanamine (Scheme 3).
Synthesis of compound 35 followed the synthetic route shown
in Scheme 4. 2-Chloroisonicotinic acid 28 was refluxed in thionyl
chloride to generate the acid chloride and then coupled with
hydrazinecarbothioamide to form 2-(2-chloroisonicotinoyl)hydra-
zinecarbothioamide 29. Acidic cyclization of compound 29 using
PPA generated thiadiazole 30. The chloro group in compound 30
was then converted to a methyl amino group using a SNAr reaction.
The amino group connected to thiadiazole in compound 31 was
coupled to Boc-protected phenylalanine to generate compound
32. The Boc group in the resulting compound 32 was removed with
TFA. Reductive amination of compound 33 with thiazole-4-carbal-
dehyde led to compound 35. The other compounds in Table 5, i.e.,
34, 36, and 37, were prepared by the same route starting from the
appropriate heterocyclic carboxylic acid.
We then turned our attention to adding a substituent to the
benzyl-like carbon to prevent the N-dealkylation. Methyl substitu-
tion in compound 18 (Table 3) was well tolerated with good effi-
cacy and only
a two-fold decrease in potency. For mono-
substitution, the stereochemistry significantly influenced potency
with one diastereomer being more potent as evident in the diaste-
reomeric pair of compounds 18 and 19. Compounds with ethyl
substitution (20), and dimethyl substitution (21) 10 displayed a fur-
ther decline in potency and significant lowering in efficacy, indicat-
ing that there is a steric limitation in this position. The exception
was the cyclopropyl substitution on the benzylic position. Com-
pound 22 not only was more potent than the parent pyridylmethyl
compound 6, but also displayed significantly improved stability
(<10% turnover in human microsomes and 60% turnover in rat
microsomes) as well as a great efficacy (127%). Replacing the pyr-
idine D-ring in compound 22 by pyrimidine (23) or methyl thiazole
(24) also resulted in a combination of good potency, efficacy and
improved microsomal stability compared to compound 6. How-
ever, compound 22 displayed the best overall profile. The dramatic
difference between cyclopropyl- and dimethyl- substitution pat-
tern could be due to the more planar nature of the cyclopropyl ring
and reduced steric demand through distinct bond angles. As a re-
sult, the cyclopropyl ring can orient itself much differently from
the dimethyl group.
Our initial lead optimization efforts aimed at exploring different
heterocyclic D-rings to see whether the microsomal stabilities of
the agonists could be improved while maintaining the potency
and efficacy. The results are listed in Table 1. A variety of five-
The pharmacokinetic profile of compound 22 was evaluated in
rats (Table 6). Though compound 22 showed improved microsomal
stability, it nonetheless displayed high in vivo clearance. The
O
O
NH2
B(OH)2
N
NH2
N
a
+
NO2
N
Br
N
8
O
S
O
O
OH
OAc
H
O
N
H
N
O
N
HO
N
b
c
N
d
O
O
O
9
10
N
e
11
R
R
HN
O
O
N
HN
H
H
N
N
N
HN
O
O
7, 17, 18, 21-24, 27
N
12
Scheme 2. Synthesis of GPR142 agonists. a. Pd2(dba)3, X-Phos, K3PO4, BuOH, 3 h, 110 °C, 74%; b. (i) cyanuric fluoride, pyridine, DCM, ꢁ20 °C–10 °C; then 8, DIPEA, DCM; (ii)
K2CO3, 88% for 2 steps; c. 4-nitrobenzene-1-sulfonyl chloride, Et3N, 69%; d. 1-(pyridin-2-yl)cyclopropanamine or other primary amines RNH2, DMF, 100 °C, 30–59%; e.
dioxane, water, HCl, 50 °C, 41–90%.