W. Wen et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4562–4566
4565
some representative 3-cyclopropyl, N-phenyl analogs 9a–d, where
the nature of the 3-substitutent modulates the mode of
pharmacology.
In sharp contrast, additional analogs of the isopropyl derivative
6c were exclusively either GIRK inhibitors, or inactive, and many
displayed a slight preference for GIRK1/4 over GIRK1/2 (Fig. 4).
Interestingly, the optimal phenyl substituents in the expanded
6b library (analogs 9a–d), were not optimal for the 6c scaffold
(9e–h), yet substituents in the 3-position were uniformly more
active.
A library combining the 3-cyclopropyl moiety with the N-ben-
zyl group, functionalities that both imparted GIRK activation, pro-
vided highly potent and efficacious dual GIRK1/2 and GIRK1/4
activators (Fig. 5). The SAR within this series was shallow, with vir-
tually all substituents on the aryl ring affording submicromolar,
dual GIRK1/2 and GIRK1/4 activators.
The corresponding 3-isopropyl, N-benzyl congener library affor-
ded fewer active compounds, and with the exception of a few low
efficacy (<10%) GIRK1/2 preferring activators, the majority of these
analogs were either GIRK1/2 selective or dual GIRK1/2 and GIRK1/4
inhibitors (Fig. 6).
O
O
N
N
N
N
F
N
H
N
H
N
H
N
H
9q
9p
GIRK1/2 IC50
GIRK1/4 IC50
=
=
0.45 µM, 75%
0.18 µM, 72%
GIRK1/2 IC50 = 0.52 µM, 83%
GIRK1/4 IC50 = 0.25 µM, 47%
F
O
O
N
N
N
N
N
N
H
F3C
N
H
N
H
H
9s
9r
GIRK1/2 IC50 = 0.64 µM, 80%
GIRK1/4 IC50 = 0.27 µM, 54%
GIRK1/2 IC50 = 0.54 µM, 88%
GIRK1/4 IC50 = 0.25 µM, 45%
F
O
N
N
O
N
N
Cl
N
H
N
H
Cl
N
H
N
H
The cyclopropylmethyl derivative 6g was a dual GIRK1/2 and
GIRK1/4 inhibitor, and replacement of the N-phenyl with an
N-benzyl group provided only weak to inactive analogs. However,
further evaluation of alternate ureas within the N-phenyl 6g series
provided a number of sub-micromolar dual GIRK1/2 and GIRK1/4
inhibitors (Fig. 7). Here, SAR lacked texture, with virtually any sub-
stituents on the urea phenyl ring providing GIRK inhibitors of com-
parable potency.
9u
9t
GIRK1/2 IC50 = 0.63 µM, 76%
GIRK1/4 IC50 = 0.34 µM, 72%
GIRK1/2 IC50 = 0.81 µM, 87%
GIRK1/4 IC50 = 0.31 µM, 70%
Figure 7. Selected analogs 9 with 3-cyclopropyl, N-phenyl motif afford highly
potent dual GIRK1/2 and GIRK1/4 inhibitors.
Finally, these library efforts highlighted the impact of slight
structural variations leading to dramatic changes in GIRK channel
selectivity, mode of GIRK pharmacology, and in some cases, an or-
der of magnitude gain or loss of GIRK activity. Therefore, we pre-
pared an expanded library around the 3-cylobutyl analog 6d to
determine if GIRK activity could be achieved with alternate func-
tionality. While >90% of these analogs were inactive, several
emerged that displayed activity as GIRK inhibitors, with a slight
preference for GIRK1/4 (Fig. 8). As seen earlier, substituents in
the 3-position are preferred, and groups in the 4-position generally
abolish GIRK1/4 inhibition.
O
O
N
N
N
N
F3C
N
H
N
H
Cl
N
H
N
H
9s
9r
GIRK1/2 IC50
= 1.3 µM, 65%
GIRK1/2 IC50
=
1.4 µM, 64%
GIRK1/4 IC50 = 0.69 µM, 84%
GIRK1/4 IC50 = 0.48 µM, 89%
In general, these analogs possess high clogPs (>4), but experi-
mental logP values range from 3.2 to 3.9. The majority of GIRK
ligands reported herein display moderate protein binding in rat
(1–3% free) and human (1–4% free), moderate intrinsic clearance
F
O
O
N
N
N
N
Cl
N
H
N
H
Br
N
H
N
H
(CLINT <40 mL/min/kg) and very clean CYP profiles (>20 lM versus
3A4, 2D6, 2C9 and 1A2). Studies probing in vivo PK and CNS expo-
sure are in progress.
9u
9t
GIRK1/2 IC50
= 1.9 µM, 70%
GIRK1/2 IC50
GIRK1/4 IC50 = 0.72 µM, 69%
= 1.9 µM, 51%
In summary, we detailed a multi-dimensional SAR campaign
based on a potent, efficacious and selective GIRK1/2 activator
(ꢀ10-fold versus GIRK1/4 and inactive on GIRK2/3) ML297. Further
chemical optimization through an iterative parallel synthesis effort
identified multiple ‘molecular switches’ that modulated the mode
of pharmacology from activator to inhibitor, as well as engendering
varying selectivity profiles for GIRK1/2 and GIRK1/4. Importantly,
these compounds were all inactive on nonGIRK1 containing GIRK
channels. However, SAR was challenging as subtle structural mod-
ifications had large effects on both mode of pharmacology and
GIRK1/2 and GIRK1/4 channel selectivity. Despite the optimization
challenges, this effort afforded potent and selective GIRK inhibi-
tors, activators with improved potency/efficacy, and a valuable
set of tool compounds to further dissect the roles of GIRK channels
in various pathological states. Detailed molecular pharmacology
studies are underway (e.g., developing mutants that exchange var-
ious domains between GIRK1/2 with GIRK 2/3) to understand the
mode/site of binding of these novel GIRK ligands and attempt to
GIRK1/4 IC50 = 0.61 µM, 69%
Figure 8. Selected analogs 9 with 3-cyclobutyl, N-phenyl motif afford GIRK1/4
preferring inhibitors.
and assess if these would either be additive or if additional ‘molec-
ular switches’12–14 be discovered (Scheme 2).
First, we elected to employ the N-phenyl and N-benzyl moieties
in analogs 5, with the cyclopropyl, isopropyl and cyclopropyl
methyl groups found in active analogs 6, and explored a diverse ar-
ray of isocyanates to deliver 120 analogs 9 from a matrix library
(2 Â 3 Â 20). These new analogs displayed
a wide range of
GIRK1/2 and GIRK1/4 selectivity profiles, as well as promiscuous
modulation in the mode of pharmacology from activator to inhib-
itor. Whereas the cyclopropyl group in 6b afforded a selective
GIRK1/2 partial activator, other substitutions on the phenyl ring
of the urea afforded both weak activators (EC50s 0.9–5
lM, 13–
65%) and potent, GIRK1/2 selective inhibitors. Figure 3 highlights
Wen, Wandong
