S. Han et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3034–3038
3035
O
GPR119 receptors. Ultimately, the 5-(methylsulfonyl)pyrazine
(12) was selected as the preferred aryloxy group for further SAR
investigation.
MeO2S
MeO2S
MeO2S
N
N
N
O
N
N
H
O
We next turned our SAR effort toward exploration of the piperi-
dine substitution (Table 2). Piperidine motifs incorporating carba-
mates (12, 16–25), amides (26–28), fluorinated alkyl groups (29–
33), and heteroaromatic rings (34–37) were prepared. These were
introduced in the final steps of the synthesis as illustrated in
Scheme 1. Trifluoromethyl analogues (21–23), in particular, dis-
played significantly improved potency relative to our clinical com-
pound (1) and were stable in human, rat, and mouse liver
microsomes. Amide 26 attained good potency and had a low shift
in the serum shift assay, but suffered from significantly reduced
microsomal stability in mouse. Tertiary amines (29–33) comprised
of fluorinated alkyls led to a reduction in GPR119 activity and
reduced microsomal stability (e.g., 30). Based on the combination
of in vitro potency and microsomal stability, a handful of com-
pounds were selected for further consideration. Compound 22 pos-
sessed a similar potency and PK profile relative to 21, however this
enantiomer was less effective in increasing glucose excursion in an
oGTT experiment performed in 129SVE mice. The exceedingly
potent hexafluoro analogue (23) was removed from further consid-
eration as it impaired physical mobility of rodents in a dose depen-
dent manner. Additionally, we were concerned that the activated O-
hexafluoroisopropyl (HFIP) carbamate may form a covalent bond
with serine hydrolases.11 Oxadiazole analogues 34–36 showed
reduced GIP release in normoglycemic 129SVE mice relative to 21,
despite their overall good potency, stability, and pharmacokinetic
properties. For the reasons described above, 21 selected for further
analysis.
CH3
OCH3
1
(APD597, partial Agonist)
O
N
N
N
O
N
O
O
CH3
F
2 (Full Agonist)
O
O
N
O
O
3 (trans/cis-1,3-dioxy)
(trans/cis-1,4-dioxy)
4
Figure 1. Design of a novel dioxycyclohexane GPR119 series.
Table 1
Aryloxy SAR
O
R1
O
R
N
O
O
Compd
R
R1
HTRF cAMP (nM)
c logPc
Starting from commercially available 1,4-dioxaspiro[4.5]decan-
8-ol, 21 was prepared by the synthetic sequence illustrated in
Scheme 1. Aromatic nucleophilic substitution of 1,4-dioxas-
piro[4.5]decan-8-ol, followed by an acid catalyzed deprotection
efficiently gave 39. A cis- enriched mixture of 40 (cis/trans = ꢀ3/2)
was obtained from the NaBH4 reduction, which was separated effi-
ciently to afford isomerically pure (>98%) cis and trans material on
>100 g scale. The separation was carried out by an initial crystal-
lization, followed by trituration with acetone. Stepwise reduction
of the pyridinium salt of 40 subsequently offered the key interme-
diate, trans-4-((1-methylpiperidin-4-yl)oxy)cyclohexanol (41). The
final intermediate 44 was straightforwardly prepared by nucle-
ophilic substitution, copper promoted sulfonylation, and de-
methylation. Intermediate 44 was then converted to 21 via a stan-
dard carbamate formation.
More extensive characterization entailed expanded in vitro
screening and pharmacokinetic analysis. Compound 21 was found
to maintain agonist activity at canine, monkey, and mouse GPR119
receptors (6.1 nM, 144%; 3.4 nM, 73%; 20.5 nM, 130%, respec-
tively). It was stable in liver microsomes across species (human,
rat, mouse half-life (t1/2) >60 min), and also demonstrated good
stability in human hepatocytes (t1/2 >120 min). It was highly bound
to both human and rat plasma proteins (human 97.5%, rat 96.0%),
but less so for mouse (93.5%). In a subsequent Sprague-Dawley
rat PK study, 21 exhibited a low systemic clearance and low
steady-state volume of distribution, while possessing excellent
systemic exposure, terminal half-life, and oral bioavailability
(Table 3). In an abbreviated CNS study, 21 demonstrated a brain
to plasma (b/p) ratio of approximately 1 (brain 820 ng/g, plasma
841 ng/mL) at 2 h post-dose, indicating virtually unrestricted
blood-brain barrier (BBB) penetration.
hGPR119
EC50 (Emax
rGPR119
EC50 (Emax
a
a
)
)
1
2
—
—
—
—
46.0 (75)
13.8 (94)
421.0 (89)
169.0 (93)
3.52
2.45
SO2Me
SO2Me
4
CH3 16.1 (115)
94.0 (120)
2.91
2.91
cis-4
CH3 266.0 (120)
1390.0 (121)
61.0 (119)
F
8.5 (113),
CH3
5
6
2.89
3.09
227.0b
SO2Me
F
CH3 52.0 (106)
167.0 (109)
SO2Me
CN
8.6 (116),
CH3
7
8
9
45.0 (106)
110.0 (108)
201.0 (125)
3.88
2.46
2.46
334.0b
N
CH3 47.0 (111)
SO2Me
SO2Me
N
H
H
13.0 (101)
11.0 (108)
N
10
11
12
89.0 (114)
400.0 (111)
64.0 (110)
2.96
2.84
2.07
SO2Me
N
CH3 262.0 (121)
CON(CH3)2
SO2Me
N
N
N
5.9 (110),
CH3
N
N
N
94.0b
SO2i-Pr
13
14
CH3 18.5 (109)
54.0 (128)
2.91
8.3 (110),
CH3
CN
28.4 (100)
2.19
1.21
144.0b
SO2Me
15
CH3 651.0 (97)
1160.0 (86)
N N
a
Emax are the mean of three or more replicates, and relative to our reference
standard, AR231453.
b
Serum shift assay (Ref. 10).
ChemBioDraw Ultra 12.0.
c
In vivo GPR119 activation and corresponding glucose control
was assessed for 21 in normoglycemic 129SVE mice after oral
administration (Figs. 2 and 3a). Since GIP release is expected upon
activation of GPR119, plasma GIP levels were measured 45 min
post compound administration (Fig. 2). Two different doses of 21
compounds prepared, 12 exhibited the smallest potency shift in
the presence of serum,10 correlating well with the order of c logP
(12 < 14 < 5 < 7). Notably, this new series was associated with a
very high intrinsic activity (Emax >100) at both human and rat