ACS Medicinal Chemistry Letters
Letter
Table 3. Physicochemical Properties and ADMET Profiles of Compounds 8, 10, and 11
a
b
Log D
compound (pH = 7.4)
solubility (μg/mL)
MDCK Papp
MS (m/
CYP direct
c
hERG
e
Amesf
assays
rat hepatocyte
toxicity (IC )
−
6
d
(JP1/JP2)
(10 cm/s)
mon/h)
inhibition
MBI
inhibition
50
g
g
8
1
1
−0.8
−0.1
0.1
590/590
620/600
>650/>650
2.7
0.8
0.6
47/95/99
21/79/98
14/86/95
0/0/0/2/0/1
7/4/8/5/4/10
0/0/0/0/0/1
82
94
90
5/6/5/5
−6/−9/−6/1
4/4/8/16
N.T.
N.T.
0
1
negative
>300 μM
g
g
N.T.
N.T.
a
b
JP1/JP2: Japanese Pharmacopoeia first/second test fluid (pH = 1.2/6.8). The percentage (%) of the tested compound remaining after 0.5 h of
c
incubation with mouse/monkey/human liver microsomes (0.5 mg/mL). The percentage (%) inhibition of 1A2/2C8/2C9/2C19/2D6/3A4 at 10
μM. The percentage (%) remaining at a concentration of 100 μM of compound reacted with CYP3A4 probe substrates after 30 min of
preincubation in human liver microsomes. The percentage (%) inhibition at 1/3/10/30 μM. Ames assays were performed up to 1000 μg/well
with or without S9 using Salmonella TA98 and TA100. Not tested.
d
e
f
g
Table 4. Mouse Pharmacokinetics Profiles of Compounds 8, 10, and 11
a
b
PO (10 mg/kg; n = 2)
IV (1 mg/kg; n = 2)
compound
Cmax (μM)
AUCinf (μM·h)
Tmax (h)
BA (%)
C0 (μM)
AUCinf (μM·h)
T1/2 (h)
Vd (L/kg)
ss
CLinf (mL/min/kg)
8
1
1
0.04
0.44
0.51
0.64
12.0
3.6
3.75
1.5
1.00
4.5
96
11
2.4
1.5
1.3
1.4
1.2
3.4
21.1
11.9
13.9
36.5
31.1
14.7
39.4
43.8
15.0
0
1
a
b
Dosing vehicle: 0.5% methylcellulose (MC) solution. Dosing vehicle: DMA/Tween80/saline = 10/10/80.
Table 5. Monkey Pharmacokinetics Profile of Compound 10
a
b
PO (5 mg/kg; n = 2)
IV (0.5 mg/kg; n = 2)
compound
Cmax (μM)
AUCinf (μM·h)
Tmax (h)
BA (%)
38
C0 (μM)
AUCinf (μM·h)
T1/2 (h)
Vd (L)
CLinf (mL/min/kg)
77.7
ss
1
0
0.11
1.36
3.00
0.86
0.35
22.7
109
a
b
Dosing vehicle: 0.5% MC solution. Dosing vehicle: DMA/Tween80/saline = 10/10/80
directed toward modifications of N-substituents at the
tetrahydroazepine ring. When the substituent on the N of the
tetrahydroazepine ring was transformed, the ethyl group (11)
retained activity, while the methyl group (10) showed a 3-fold
decrease in activity and the isopropyl group (12) showed a 2-
fold decrease in activity. Replacing the cyclobutyl group of
compound 10 with cyclopentyl (13) and cyclohexyl (14) groups
conferred much lower potency (IC = 32 and 172 nM for G9a,
activity, likely due to low steric repulsion. Further, the activity of
methoxy derivatives 21 and 23 was greatly decreased, likely due
to steric repulsion with Y1154. In contrast, the hydroxy group in
compounds 20 and 22 appeared to form hydrogen bonds with
the main chain carbonyl group of Y1154, and especially the
hydrogen bond angle of compound 22 would be better than
compound 20, which led to retain the activity.
In the comparison of the cocrystal structures of compound 10
and the previously reported compound 6 (Figure 3), the
tetrahydroazepine moiety of each inhibitor was found to be
located at approximately the same position and was hydrogen-
bonded to the side chain of Leu1086. In contrast, the helix
(1074−1079) was slightly off, and in the case of compound 10, it
was closer to the inhibitor. Considering the difference of the
molecular size of each inhibitor, it seems that the conserved
interaction of the tetrahydroazepine moiety with Leu1086
controls the location of the inhibitor in the active sites of
enzymes. As a result, the interaction patterns with the helix
(1074−1079) are different between the inhibitors. Namely, the
primary amino group of compound 10 characteristically forms
hydrogen bonds with the side chains of Asp1078 and Asp1074,
whereas the amino moiety of compound 6 forms hydrogen
bonds with only the side chain of Asp 1078, not Asp1074. In
addition, with respect to the interaction with Asp1088,
aminopyrimidine (6) forms a hydrogen bond and salt-bridge,
while the indole nitrogen of compound 10 only forms a salt-
bridge with Asp1088.
50
respectively). These potency results corresponded with those of
1
8
the A-366 series, suggesting that these derivatives bind to the
same G9a/GLP protein site as A-366. Likewise, the tetrahy-
dropyranyl group derivative 15 resulted in a loss of potency.
Next, to obtain further SARs and improve the activity, a methyl
group was introduced at the α position of the tetrahydroazepine
ring (16−19), which led to a decrease in activity. In addition, the
introduction of a hydroxy or methoxy group at the β position of
the tetrahydroazepine ring (20−23) resulted in reduced activity.
The description of these activity results based on the cocrystal
structure is described later.
To determine the binding mode of the newly synthesized
inhibitors, we solved the X-ray cocrystal structure of compound
1
0 with G9a (Figure 2). The cocrystal structure of compound 10
with G9a revealed the same binding site as that previously
18
reported for A-366; thus, the mechanism of action of inhibitor
0 was expected to show a noncompetitive inhibition pattern
1
against the cofactor S-adenosyl methionine. The hydrogen
bonding patterns revealed interactions between the side chain of
Leu1086, the amino group of tetrahydroazepine, and the side
chains of Asp1078, Asp1088, Asp1074, and the amidine group.
These findings are similar to those found for A-366.
Next, the activity results of compounds 16−23 were examined
based on this cocrystal structure. Compounds 16, 17, and 19
showed a significant reduction in activity, likely due to steric
repulsion with Y1154. Only compound 18 showed retained
The physicochemical properties and ADMET profiles of
representative compounds 8, 10, and 11 are summarized in
Table 3. These compounds showed low log D values, high
solubility, and high metabolic stability in both monkey and
human liver microsomes. These compounds did not exhibit
CYP inhibitory activity at a concentration of 10 μM, and they
also did not exhibit mechanism-based inhibition (MBI) of
1
24
ACS Med. Chem. Lett. 2021, 12, 121−128