Dihydropyrazolopyrimidines containing benzimidazoles as KV1.5 potassium channel antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.07.083

Dihydropyrazolopyrimidines with a C6 heterocycle substituent were found to have high potency for block of K_V1.5. Investigation of the substitution in the benzimidazole ring and the substituent in the 5-position of the dihydropyrazolopyrimidine

Full text of DOI:10.1016/j.bmcl.2009.07.083

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5469–5473
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Dihydropyrazolopyrimidines containing benzimidazoles as KV1.5 potassium
channel antagonists
*
John Lloyd , Heather J. Finlay, Karnail Atwal, Alexander Kover, Joseph Prol, Lin Yan, Rao Bhandaru,
Wayne Vaccaro, Tram Huynh, Christine S. Huang, MaryLee Conder, Tonya Jenkins-West, Huabin Sun,
Danshi Li, Paul Levesque
Bristol-Myers Squibb, Pharmaceutical Research and Development, PO Box 5400, Princeton, NJ 08543-5400, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Dihydropyrazolopyrimidines with a C6 heterocycle substituent were found to have high potency for
block of KV1.5. Investigation of the substitution in the benzimidazole ring and the substituent in the 5-
position of the dihydropyrazolopyrimidine ring produced 31a with an IC50 for KV1.5 block of 0.030 lM
without significant block of other cardiac ion channels. This compound also showed good bioavailability
in rats and robust pharmacodynamic effects in a rabbit model.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 12 May 2009
Revised 16 July 2009
Accepted 17 July 2009
Available online 22 July 2009
Keywords:
IKur
KV1.5
Voltage gated potassium channel
Atrial fibrillation
Atrial fibrillation (AF) is a cardiac arrhythmia that affects a large
and growing population. In 1999, over 2 million cases of AF were
reported in the United States and that number increased 2–3-fold
over the previous 15 years.1 The number of cases will likely grow
in the future because the incidence of AF increases with age. It is
predicted that 5.6 million adults in the US will have AF by 2050.2
Although AF is not a fatal condition, the primary mortality risk is
from stroke.3 The inefficient emptying of the atrium caused by
AF results in stasis of blood in the atrium and can lead to thrombus
formation.
Re-polarization of membrane potential in cardiac myocytes by
the outward flow of potassium ions is essential for regulating car-
diac rhythm. The net potassium current is conducted through sev-
eral different potassium channels, including the ultra-rapid (IKur),
rapid (IKr), and slow (IKs) delayed rectifier potassium currents.
The ultra-rapid potassium current (IKur) in humans is functionally
expressed in the atrium but not ventricle.4 Therefore, selectively
blocking this current may lead to a method of treating atrial
arrhythmia without increasing the risk of arrhythmias caused by
prolonging ventricular refractory period.5 IKur is conducted by the
voltage gated potassium channel encoded by KV1.5.6 Drugs that
block this channel have the potential to increase atrial action po-
tential duration and prevent and/or terminate atrial fibrillation
without the risk of ventricular effects.
In our previous paper, we described the discovery of dihydropy-
razolopyrimidines as potent and selective blockers of KV1.5.7 We
learned that amide substitution in the C6-position improved po-
tency and that a wide variety of substituents were tolerated. The
simple phenyl amide (1) and dipropylamide (2) were some of the
more potent compounds in this series. This information suggested
replacement of the amide with a heterocycle could provide potent
and novel blockers of KV1.5 (Fig. 1).
The dihydropyrazolopyrimidine template was synthesized as
recently described.7 The oxazole (4) was synthesized from the acid
(3) by coupling to 2-aminoacetophenone followed by treatment of
the resulting amide with phosphorus oxychloride at 120 °C
(Scheme 1).
The 1,3,4-oxadiazoles (5a,b) were synthesized by coupling of
the appropriate acyl hydrazines to the acid (3) followed by treat-
ment with phosphorus oxychloride (Scheme 2).
Cl
Cl
Cl
N
Cl
N
O
O
7
6
5
N
N
H
N
N
2
N
H
N
H
3
2 IC50 0.070 uM
1
IC50 0.17 uM
* Corresponding author. Tel.: +1 609 818 5327; fax: +1 609 818 3331.
E-mail address: john.lloyd@bms.com (J. Lloyd).
Figure 1. Potent pyrazolodihydropyrimidine KV1.5 blockers.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.07.083

5470
J. Lloyd et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5469–5473
Cl
Cl
Cl
Cl
N
Cl
N
Cl
N
6
R2
7
5
O
N
N
4
a, b
N
R1
N
3
OH
N
a, b
O
N
3
N
H
N
H
N
H
4
Scheme 1. Reagents and conditions: (a) PhCOCH2NH2, PyBroP, CH2Cl2, 39%; (b)
POCl3, 120 °C, 52%.
Compd
7
8
R1
R2
4-F
5-F
6-F
Compd
17
18
19
20
21
22
23
24
R1
R2
6-CN
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
Et
6-CO2CH3
6-CF3
6-SO2CH3
6-SO2NHPh
6-NHCOCH3
6-Cl
9
Cl
10
11
12
13
14
15
16
7-F
Cl
4-Cl
5-Cl
6-Cl
4-CH3
5-CH3
6-CH3
N
N
R
N
i-Pr
n-Pr
6-Cl
6-F
a, b
O
N
3
25
5a R=Ph
5b R=n-Bu
N
H
Scheme 4. Reagents and conditions: (a) 26, PyBroP, CH2Cl2 (30–70%); (b) POCl3,
80–100 °C (25–70%).
Scheme 2. Reagents and conditions: (a) RCONHNH2, PyBroP, CH2Cl2 (30–50%); (b)
POCl3, 110 °C (20–50%).
R2
R2
Synthesis of the benzothiazole (6a) and the benzimidazoles
(6b–d) were similarly accomplished by coupling of either o-phen-
ylenediamine or 2-aminothiophenol to the acid (3) and cyclization
with phosphorus oxychloride (Scheme 3).
These reactions were quite general and the acid (3) could be
coupled to various phenylenediamines (26) with substituents on
either the aromatic ring or nitrogen (Scheme 4). In most cases a
mixture of the two regioisomeric amides was formed. The mix-
tures of amides were cyclized to the substituted benzimidazoles
using phosphorous oxychloride at 80–100 °C in 1–4 h.
a, b
26
H2N
O2N
HN
F
R1
Scheme 5. Reagents and conditions: (a) R1NH2, K2CO3, DMF, 45 °C (30–90%); (b) Zn,
HOAc, MeOH (70–90%).
O
The requisite phenylenediamines (26) were either commer-
cially available or were synthesized by displacement of the appro-
priately substituted 2-fluoronitrobenzene with an amine.
Reduction of the nitro group with zinc and acetic acid in methanol
provided the phenylenediamines (Scheme 5).
OH
O
O
O
O
O
O
a
+
b
O
O
O
O
O
27
O
29
28
Cl
O
In an effort to introduce polar substituents, we also synthesized
compounds with a methoxymethyl group at the 5-position of the
dihydropyrimidine ring system. These compounds were made
using the Biginelli reaction in the same way as the previous com-
pounds (Scheme 6). 4-Keto-5-methoxy ethyl butanoate (28) was
synthesized from condensation of the methoxy acetyl chloride
with Meldrum’s acid (27) followed by decarboxylation with tert-
butanol. The intermediate t-butyl ester (29) was condensed with
3,4-dichlorobenzaldehyde and 3-aminopyrazole using sodium
bicarbonate in DMF to form the dihydropyrazolopyrimidine and
subsequently deprotected to the carboxylic acid (30) using
TMSOTf. This intermediate was condensed with phenylenediam-
ines, then cyclized to form benzimidazoles. In this way we synthe-
sized the fluorobenzimidazole (31). The racemic benzimidazole
was purified by chiral HPLC (Chiracel AD column, 25% isopropanol,
hexane, 0.1% TEA to give the individual enantiomers (31a,b).
Cl
Cl
N
Cl
O
CO2H
O
CHO
c, d
Cl
N
N
O
NH
29
O
N
H
NH2
30
Cl
N
F
N
O
e, f
N
N
N
H
31
Cl
Scheme 6. Reagents and conditions: (a) MeOCH2COCl, pyridine, DCM, 5 °C; (b) t-
butanol, toluene, reflux; (c) NaHCO3, DMF, 70 °C (99%, 3 steps); (d) TMSOTf, DCM,
hexane, (99%); (e) 26, EDCI, CH2Cl2; (f) AcOH, 100 °C (34%, 2 steps).
Cl
N
6a X=S
R=H
6b X=NH R=H
6c X=NMe R=H
6d X=NH R=Me
N
a, b
X
N
R
All the compounds described were tested for their ability to
block the potassium current in mouse fibroblast L929 cells
expressing human KV1.5.8 Our strategy was to investigate hetero-
cyclic amide replacements, particularly, analogs of 1 and 2 with
improved potency or physicochemical properties. Our initial amide
3
N
H
Scheme 3. Reagents and conditions: (a) 2-HX-Ph-NH2, PyBroP, CH2Cl2 (25–70%);
(b) POCl3, 110 °C (20–40%).

J. Lloyd et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5469–5473
5471
replacements were the oxazole (4) and 1,3,4-oxadiazoles (5a,b).
Although these compounds contained the phenyl or alkyl function-
ality found in the amides (1, 2), they lacked potency. We developed
the hypothesis that the conformation of these inactive compounds
was such that the heterocycle and the phenyl substituent would be
orthogonal. This conformation would be in contrast to the confor-
mation of the phenyl amide where the phenyl and the amide
would be expected to be closer to coplanar. We hypothesized that
the solution would be to fuse the phenyl ring to the heterocycle to
force co-planarity. We first synthesized the benzothiazole (6a) and
the benzimidazole (6b) to test this hypothesis. Both compounds
were significantly more potent than the non-fused compounds.
We next replaced the hydrogen on the amine of the benzimidazole
with a methyl (6c) and the potency improved by eightfold. It was
found that the methyl group did not have to be attached to the
nitrogen to improve potency. The benzimidazole with the methyl
substitution on the aromatic ring (6d) was over threefold more po-
tent than the unsubstituted compound (6b) (Table 1). This may
indicate the presence of a hydrophobic binding region that can
be accessed from both positions. It also indicated that the presence
of a hydrogen bond donor in the benzimidazole is not detrimental.
Substitution on the benzimidazole template was easily varied at
either the nitrogen substituent or the ring substituent. We initially
studied substitution on the aromatic ring and synthesized com-
pounds with a halogen (F, Cl) or methyl substituent (Table 2). In
all three cases, substitution in the 5 position was less favored than
other positions. In the case of fluorine substitution, the 6-fluoro
compound (9) was 10-fold more potent than the 5-fluoro analog
(8) and threefold more potent than the unsubstituted compound
(6c).
Because substitution in the 6-position was well tolerated, we
synthesized compounds with a broader range of functional groups
in this position (Table 3). As already mentioned, the fluorine and
chlorine substituents showed similar or improved potency over
the N-methylbenzimidazole (6c). The methyl (16), cyano (17),
and ester (18) substituted compounds resulted in a 2–3-fold loss
Table 1
Dihydropyrazolopyrimidine KV1.5 blockers
Table 2
Aromatic substitution
Cl
Cl
Cl
Cl
6
7
R
5
R
N
N
N
4
N
N
N
N
H
CH3
N
H
a
Compd
R
KV1.5 inhibition IC50
0.17
(lM)
a
Compd
R
KV1.5 inhibition IC50 (lM)
O
O
6c
7
8
H
0.075
0.15
0.29
0.024
0.064
0.071
0.66
0.068
0.11
1
N
H
4-F
5-F
6-F
7-F
4-Cl
5-Cl
9
10
11
12
13
14
15
16
2
0.070
N
O
6-Cl
4-Me
5-Me
6-Me
N
4
Ph
>1.0 (25% inh)
>1.0 (18% inh)
>1.0 (29% inh)
0.74
0.98
0.20
a
Values are means of 2–4 experiments.
N
N
N
5a
5b
6a
Ph
O
Table 3
6-substituted benzimidazoles
N
n-Bu
O
S
Cl
Cl
N
R
N
N
N
N
N
N
N
H
6b
6c
0.51
a
N
H
Compd
R
KV1.5 inhibition IC50 (lM)
6c
9
H
F
0.075
0.024
13
16
17
18
19
20
21
22
Cl
Me
CN
CO2Me
CF3
SO2Me
SO2NHPh
NHCOMe
0.068
0.20
0.22
0.20
0.075
0.150
N
Me
0.54
>1.0 (29% inh)
>1.0 (30% inh)
>1.0 (8% inh)
N
6d
N
Me
H
a
a
Values are means of 2–4 experiments.
Values are means of 2–4 experiments.

5472
J. Lloyd et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5469–5473
Table 4
Table 5
N-Substitution
Ion channel selectivity of 9a and 31a
Cl
Cl
R1
F
Cl
N
Cl
N
9a
31a
R = H
N
N
R = OMe
N
R2
N
N
N
N
H
N
H
R
Compd
R1
R2
KV1.5 inhibition IC50 (lM)
a
Compd
KV1.5 inh
IC50
hERG % inh
INa %inh
ICa % inh
(
lM)
10
l
Ma
10
l
Ma
10 l
Ma
a
13
23
24
9
Cl
Cl
Cl
F
Me
Et
i-Pr
Me
n-Pr
0.068
0.057
0.094
0.024
0.048
9a
31a
0.015
0.030
83
47
37
27
NT
13
25
F
a
Values are means of 2–4 experiments.
a
Values are means of 2–4 experiments.
Table 6
Pharmacokinetic parameters of compound 9a and 31a in rats
in potency. The trifluoromethyl containing compound (19) showed
a fivefold loss of potency and the sulfone (20), sulfonamide (21),
and acetamide (22) were significantly less potent indicating that
polar groups are not well tolerated in the binding site.
9aa
31a
10 (inf)b 20 (po)c
37
a
Dose (
F (%)
t1/2 (h)
Clearance (mL/min/kg)
l
mol/kg)
10 (inf)b 20 (po)c
51
0.57 0.17
35
1.6 0.6
6
7
1.5 0.6
42 4.3
2.1 0.08
With both the chlorine and fluorine substituents in the 6-posi-
tion, we varied the nitrogen substituent. In the 6-chlorine series,
the methyl (13), ethyl (23), and isopropyl (24) substituents were
of similar potency. Likewise in the 6-fluoro series the methyl
(13) and n-propyl (25) compounds were equipotent (Table 4).
Compound 9 with the 6-fluoro substituent on the benzimid-
azole was one of the most potent so we synthesized and tested
the 5-methoxymethyl analog (31). We made this change with
the hope of introducing polar functionality that would improve
physical properties. This compound had similar potency to the
methyl analog (9). The individual enantiomers of both compounds
were separated using chiral chromatography (Chiracel AD column,
1% i-propanol/hexane eluent). The active enantiomers (9a and 31a)
were very potent blockers of KV1.5. Both compounds were tested
for block of hERG, sodium and calcium channels and 31a was found
to have >200-fold selectivity over these other ion channels (Table
5).
The pharmacokinetics of compounds 9a and 31a were investi-
gated in rats (Table 6).9 Compound 9a has intermediate systemic
clearance in rats. Steady-state volume of distribution was greater
than total body water, indicating significant extravascular distribu-
tion. Terminal half-life was 0.57 h in rats. Oral bioavailability (F)
was 51%. Compound 31a showed a longer half-life (1.5 h) with
similar clearance and bioavailability.
9
Vdss (L/kg)
a
Values are means from 3 animals.
inf = intra-arterial infusion for 10 min.
po = oral gavage.
b
c
40
35
30
25
20
15
10
5
0
-5
0.3 mg/kg
1 mg/kg
AERP
3 mg/kg
VERP
10 mg/kg
Because of the acceptable pharmacokinetic profile and ion
channel selectivity, the methoxymethyl compound (31a) was cho-
sen for further in vivo characterization. The pharmacodynamic
activity was tested in a rabbit model which measured the effective
refractory period (ERP) in both atrium and ventricle (Fig. 2).10 Like
humans, rabbits express the IKur current in atrium but not ventri-
cle. The compound was dosed at 0.3, 1.0, 3.0, and 10 mg/kg and
prolonged atrial ERP by >20% at a dose of 3 mg/kg. There was no
effect on ventricular ERP reflecting the selectivity for KV1.5 over
ventricular ion channels.
In conclusion, we have successfully replaced the amides of
known IKur blockers (1, 2) with benzimidazole. We have partially
optimized the substituents on the aromatic ring and nitrogen of
the benzimidazole and the substituent at the 5-position of the
dihydropyrazolopyrimidine ring. We discovered compound 31a
to be a very potent and selective blocker of KV1.5. This compound
also had good oral bioavailability in rats and showed a significant
pharmacodynamic effect in rabbits. For these reasons it was cho-
sen for further preclinical development.
Figure 2. Pharmacodymanic effects of 31a in rabbit.
References and notes
1. Wattigney, W. A.; Mensah, G. A.; Croft, J. B. Circulation 2003, 108, 711–716.
2. Feinberg, W. M.; Blackshear, J. L.; Laupacais, A., et al Arch. Intern. Med. 1995,
155, 469–473; Go, A. E.; Hylek, E. M.; Phillips, K. A.; Chang, Y-C.; Henault, L. E.;
Selby, J. V.; Singer, D. E. J. Am. Med. Assoc. 2001, 285, 2370–2375.
3. Wolf, P. A.; Mitchell, J. B.; Baker, C. S., et al Arch. Intern. Med. 1998, 158, 229–
234.
4. Amos, G. J.; Wettwer, E.; Li, Q.; Himmel, H. M.; Ravens, U. Circulation 1994, 90,
I-581a; Li, G. R.; Feng, J.; Yue, L.; Carrier, M.; Nattel, S. Circ. Res. 1996, 78, 689–
696.
5. Yang, T.; Snyders, D.; Roden, D. J. Cardiovasc. Pharm. 2001, 38, 737–744.
6. Wang, Z.; Fermini, B.; Nattel, S. Circ. Res. 1993, 73, 1061–1076; Feng, J.; Wible,
B.; Li, G. R.; Wang, Z.; Nattel, S. Circ. Res. 1997, 80, 572–579.
7. Vacarro, W.; Huynh, T.; Lloyd, J.; Atwal, K. S.; Finlay, H. J.; Levesque, P. C.;
Conder, M. L.; Jenkins-West, T.; Shi, H.; Sun, L. Bioorg. Med. Chem. Lett. 2008, I18,
6381–6385.
8. Snyders, D. J.; Tamkun, M. M.; Bennet, P. B. J. Gen. Physiol. 1993, 101, 513–543.
9. Compound 9a was administered to rats, as a solution in polyethylene glycol
200:ethanol:water (1:1:1). Plasma was prepared from each blood sample by

J. Lloyd et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5469–5473
5473
centrifugation and analyzed by LC/MS. Plasma concentration versus time data
were analyzed by non-compartmental methods. The total plasma clearance,
terminal half-life (t1/2), and the steady state volume of distribution (Vdss) were
calculated after intra-arterial administration. The absolute oral bioavailability
(F, expressed as%) was estimated by taking the ratio of dose-normalized AUC
value after an oral dose to that after an intra-arterial dose.
10. Sun, H.; Lloyd, J.; Shi, H.; Li, D.; Levesque, P. Heart Rhythm. 2008, 5,
S213.