Discovery of a novel series of selective HCN1 blockers

Article abstract of DOI:10.1016/j.bmcl.2011.07.051

The discovery of a series of novel, potent, and selective blockers of the cyclic nucleotide-modulated channel HCN1 is disclosed. Here we report an SAR study around a series of selective blockers of the HCN1 channel. Utilization of a high-throughput VIPR assay led to the identification of a novel series of 2,2-disubstituted indane derivatives, which had moderate selectivity and potency at HCN1. Optimization of this hit led to the identification of the potent, 1,1-disubstituted cyclohexane HCN1 blocker, 2-ethoxy-N-((1-(4- isopropylpiperazin-1-yl)cyclohexyl)methyl)benzamide. The work leading to the discovery of this compound is described herein.

Full text of DOI:10.1016/j.bmcl.2011.07.051

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5197–5201
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a novel series of selective HCN1 blockers
⇑
Kelly J. McClure , Michael Maher, Nancy Wu, Sandra R. Chaplan, William A. Eckert III,
Dong H. Lee, Alan D. Wickenden, Michelle Hermann, Brett Allison,
Natalie Hawryluk, J. Guy Breitenbucher, Cheryl A. Grice
Johnson & Johnson Pharmaceutical Research & Development, L.L.C., 3210 Merryfield Row, San Diego, CA 92121, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery of a series of novel, potent, and selective blockers of the cyclic nucleotide-modulated chan-
nel HCN1 is disclosed. Here we report an SAR study around a series of selective blockers of the HCN1
channel. Utilization of a high-throughput VIPR assay led to the identification of a novel series of 2,2-
disubstituted indane derivatives, which had moderate selectivity and potency at HCN1. Optimization
of this hit led to the identification of the potent, 1,1-disubstituted cyclohexane HCN1 blocker, 2-eth-
oxy-N-((1-(4-isopropylpiperazin-1-yl)cyclohexyl)methyl)benzamide. The work leading to the discovery
of this compound is described herein.
Received 6 May 2011
Revised 11 July 2011
Accepted 13 July 2011
Available online 23 July 2011
Keywords:
HCN1
HCN4
Ó 2011 Elsevier Ltd. All rights reserved.
Ion channel
VIPR assay
Strecker reaction
Spared nerve injury model
Analgesic
Bradycardia
Neuropathic pain
Patients suffering from peripheral nerve damage frequently
experience a variety of painful, spontaneous and evoked sensa-
tions. These sensations include sensitivity to light touch (tactile
allodynia), pressure and thermal stimuli. Discharges emanating
from both the site of injury and the dorsal root ganglion (DRG)
are now believed to be responsible for the initiation and mainte-
nance of neuropathic pain syndromes.1 Recent findings suggest
that a family of hyperpolarization-activated, cation-nonselective,
cyclic nucleotide-modulated (HCN) channels contributes to the
abnormal spontaneous discharges emanating from the DRG of
damaged nerves and the increased firing rate with evoked stim-
uli.1,3,4 Because HCN1 is expressed at much higher levels than
the other subtypes in DRG neurons,2 this channel has been pro-
posed as a novel target for pharmacological intervention of pain.1
A recent review discusses other clinical indications and potential
side effects of HCN inhibitors.15
(Chung model) and in a model of partial sciatic nerve ligation (Selt-
zer model).1,3,5 More recently, ZD-7288 has been shown to inhibit
spontaneous pain and tactile allodynia resulting from acute insults
such as mild thermal burn and skin incision.5,6 Ivabradine was ap-
proved by the European Medicines Agency in 2005 for the treat-
ment of stable angina pectoris and has been shown to be of
equal efficacy to the beta-blocker atenolol and calcium channel
antagonist amlodipine in treating this disease.7 While ZD-7288
and ivabradine are potent blockers of HCN1, a lack of selectivity
over other HCN channels has prevented determination of the abso-
lute effectiveness of an HCN1 blocker in treating neuropathic pain.
The activity of these compounds at HCN4 is problematic for a
non-cardiac indication, because blockade of this channel results
in bradycardia due to its critical role in sinoatrial node action po-
tential.8,9 An important advance in the search for subtype selectiv-
ity was recently demonstrated in a series of phenylalkylamines.10
One of the compounds (4, Fig. 1) in this study showed HCN1 selec-
tivity over HCN4 (>150-fold) and HCN2 (ꢀ30-fold).
Most of the HCN inhibitors reported in the literature are equi-
potent at the four HCN subtypes. Examples include ZD-7288 (1),
Zatebradine (2), and Ivabradine (3) (Fig. 1). ZD-7288 has been
shown to inhibit injury-induced ectopic neuronal discharges in iso-
lated spinal nerves and in damaged nerves in situ.1,3,4 ZD-7288 also
exhibits anti-allodynic effects in the rat spinal nerve ligation model
In an effort to identify a novel series of HCN isoform selective
blockers, we used high-throughput functional assays13 to screen
a proprietary library of compounds and identified an indane with
a modest degree of HCN1/HCN4 selectivity (5, Fig. 1, pIC50
HCN1 = 5.9, pIC50 HCN4 = 4.7). Medicinal chemistry efforts on this
compound by our laboratory led to the identification of several po-
tent HCN1 blockers having at least 10-fold selectivity for HCN1
⇑
Corresponding author.
E-mail address: kmcclure@its.jnj.com (K.J. McClure).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.07.051

5198
K. J. McClure et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5197–5201
OMe
OMe
O
H
N
N
N
N
N
N+
MeO
MeO
Zatebradine
2
ZD-7288
1
OMe
OMe
N
O
N
N
N
MeO
MeO
HN
Ivabradine
3
O
OMe
5
MeO
OMe
O
N
MeO
MeO
R
N
N
4
O
OMe
OMe
Figure 1. Structures of nonselective HCN1 blockers ZD-7288 (1), Zatebradine (2), Ivabradine (3) and an HCN1 selective indane (4).
over HCN4. The routes used to prepare these analogs and their
associated biological data are reported below.
with electron withdrawing groups, such as chlorine or fluorine
(examples 11f and 11h), yielded only weak HCN1 blockers. While
the data set is limited, it also appears that the presence of a basic
nitrogen is critical for HCN1 potency as the two piperidine deriva-
Preparation of the indane analogs shown in Table 1 began with
a Strecker reaction between indanone, the desired amine and
TMSCN to give the nitrile intermediates (7) in good yield (Scheme
1).11 Reduction of the nitrile to the primary amine was problem-
atic, and depending on the nature of the diamine, led to differing
amounts of the retro-Strecker by-product (9) and desired primary
amine (8). In many cases, use of LAH as the reducing agent led to
significant amounts of the retro-Strecker by-product (9). Forma-
tion of this unwanted by-product could be minimized in some of
those cases by using a mixture of LAH/H2SO4 as the reductant.12
Following isolation, the primary amine was acylated with the de-
sired benzoyl chloride (10) to afford the target amides (11).
Biological activity of the analogs shown in Tables 1–3 was mea-
sured using a high-throughput, functional, VIPR assay.13 While
only moderate HCN1 potency was obtained with this series, it be-
came evident that a substituted piperazine in combination with an
alkoxy substituent in the 2-position of the aromatic ring (examples
5, 11a, 11b, and 11i) were required for HCN1 potency. Moving the
2-alkoxy group to either the 3 or 4-position (examples 11d and
11e), or dealkylation to give a free hydroyl group (example 11g),
resulted in a significant loss in HCN1 potency. Incorporation of
an electron withdrawing group, such as the trifluoromethoxy,
(example 11c) also led to a reduction in HCN1 potency. These re-
sults led us to postulate that the oxygen of the alkoxy substituent
could be making a direct contribution to binding acting as a hydro-
gen bond acceptor. In addition, replacement of the 2-alkoxy group
Table 1
pIC50 data for indanes at hHCN1 and hHCN4
X
N
R
HN
O
Compds
X
R
hHCN1 (pIC50
)
hHCN4 (pIC50)
ZD-7288
5
—
—
6.8
5.9
6.4
6.2
4.6
4.2
4.0
4.7
<3.7
4.6
6.2
4.3
4.8
7.5
4.7
5.3
N-Me
N-Me
N-Me
N-Me
N-Me
N-Me
N-Me
N-Me
N-Me
N-iPr
CH2
2-OMe
2-OEt
2-OBn
2-OCF3
3-OMe
4-OMe
2-Cl
2-OH
2-F
2-OEt
2-OMe
2-OEt
11a
11b
11c
11d
11e
11f
11g
11h
11i
5.6
<3.7
<3.7
<3.7
3.8
<3.7
<3.7
5.5
11j
11k
<3.7
4.4
CH2
pIC50 values are the mean of three experiments.

K. J. McClure et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5197–5201
5199
O
Cl
NR2
NR2
CN
b
a
+
O
NH2
8
+
7
6
R
10
NR2
9
NR2
c
R
HN
O
11
Scheme 1. Reagents and conditions: (a) TMSCN, ZnI2, Et2O, NHR2, MeOH, 80 °C, 8 h, 50–90%; (b) LAH, Et2O, THF, 2 h, reflux, 0–70% or LAH, H2SO4, Et2O, 5 h, reflux, 0–75%; (c)
Pyridine, 50 °C, 24 h, 15–65%.
Table 2
Table 3
pIC50 data for cyclohexane core analogs at hHCN1 and hHCN4
pIC50 data for cyclohexane core replacements at hHCN1 and hHCN4
X
X
N
N
Y
R
R
HN
HN
O
O
Compds
X
R
hHCN1 (pIC50
)
hHCN4 (pIC50
)
Compds
Y
X
R
hHCN1
hHCN4
(pIC50
)
(pIC50
)
12a
12b
12c
12d
12e
12f
12g
12h
12i
12j
12k
12l
12m
12n
12o
12p
N-Me
N-Me
N-Me
N-Me
N-iPr
N-iPr
N-iPr
N-iPr
2-OMe
2-OEt
2-Me
2-Cl
2-OEt
2-OMe
2-OiPr
2-Et
2-OEt
2-OEt
2-OEt
2-OEt
2-OEt
2-OEt
2-OMe
2-OEt
6.4
6.9
4.2
<3.7
7.8
7.5
6.6
<3.7
7.5
7.3
7.1
6.7
6.4
4.8
<3.7
<3.7
5.3
5.9
<3.7
4.5
6.5
6.1
5.8
4.1
6.8
6.5
6.1
6.1
5.4
4.4
4.3
4.0
N-
Me
2-
OMe
13a
5.8
5.2
N-
Me
13b
13c
13d
2-OEt
2-OEt
6.4
6.5
6.2
5.3
5.4
5.3
N-
iPr
N-Et
O
H
N-Cyclopropyl
N-Cyclobutyl
N-Cyclopentyl
NH
CH2
O
N-
iPr
2-
OMe
N
N-
Me
2-
OMe
13e
13f
5.6
4.7
N-Phenyl
N-
Me
2-
OMe
<3.7
<3.7
pIC50 values are the mean of three experiments.
H
pIC50 values are the mean of three experiments.
tives 11j and 11k, were devoid of HCN1 activity. Efforts to deter-
mine if the proximal piperazine nitrogen was required for HCN1
affinity were not pursued.
analog (12j) exhibited good potency in the HCN1 VIPR assay
despite the piperazine nitrogen being roughly 100-fold less basic
than other analogs such as methyl (12b) and isopropyl piperazine
(12h). The importance of the basicity of the distal nitrogen was
probed further by preparation of acylated and sulfonylated ana-
logs, which resulted in a significant loss of HCN1 potency (data
not shown); however, as mentioned previously, a full investigation
of the distal nitrogen for HCN1 blocking potential was not under-
taken. A metabolite i.d. study of the N-Me piperazine analog 11a,
where the parent compound was incubated with human liver
microsomes, showed the primary route of metabolism to be
N-demethylation leading to the unsubstituted piperazine metabo-
lite. Based on this result, the analogous cyclohexane analog 12m
was prepared with the hope that it would possess a higher degree
of metabolic stability in human liver microsomes relative to the
parent methyl piperazine analog 12b. In fact, 12m did exhibit an
Replacement of the indane ring with cylcohexane was accom-
plished by substitution of indanone with cyclohexanone in Scheme
1 yielding compounds 12a–p (Table 2). This led to a significant in-
crease in potency at HCN1, and the 10-fold selectivity over HCN4
was maintained in most cases. As with the indane analogs, an alk-
oxy group in the 2-position of the aryl ring was required for HCN1
potency. When this alkoxy group was replaced with other simple
substituents, a significant loss in HCN1 potency was observed
(12c, 12d, and 12h). Small alkyl substituted piperazine rings led,
in some cases, to very potent HCN1 blockers with the best combi-
nation being an isopropyl piperazine group and a 2-ethoxy phenyl
group (example 12e). Small, cycloalkyl substituted piperazines also
yielded potent analogs (examples 12j and 12k); however, as the
size of the cycloalkyl group increased the resulting HCN1 potency
diminished (example 12l). It is worth noting that the cyclopropyl

5200
K. J. McClure et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5197–5201
a 150
H
H2N
SNI ED50=3mg/kg
ECG EC20=1.3 mg/kg
N
N
a
N
125
100
75
50
25
0
N
OMe O
N
12f
Scheme 2. Reagents and conditions: (a) 2-Methoxybenzoyl chloride, pyridine,
50 °C, 24 h, 72%.
improved metabolic stability profile as measured by human liver
microsomes, having a t1/2 of 235 min versus a t1/2 of 22 min for
12b. Unfortunately, this structural modification did lead to a loss
in HCN1 potency relative to 12b and 12e. In the case of the cyclo-
hexane analogs, replacement of the substituted piperazine group
with a simple piperidine ring (example 12n) led to a complete loss
of HCN1 potency, paralleling similar results in the indane series.
Not all substitution on the piperazine ring was tolerated. Attach-
ment of a phenyl ring to the terminal nitrogen led to a complete
loss of HCN1 potency (example 12p), further supporting the theory
that the presence of a basic nitrogen in this region is important,
and contributes to the potency of the series.
Other cores besides indane and cyclohexane were also investi-
gated (Table 3). With the exception of compounds 13e (compound
obtained from original screening deck) and 13f, all analogs were
prepared using the route outlined in Scheme 1. Compound 13f
was prepared by acylation of the commercially available
N-methyl-N0-aminoethyl piperazine (Scheme 2).
1
10
ZD7288 (mg/kg i.p.)
150
125
100
75
b
SNI ED50=6mg/kg
ECG EC20=25 mg/kg
50
25
As can be seen from the data in Table 3, all other cores investi-
gated offered no advantage over the cyclohexane core. The 6,
6-bicyclic system (example 13a) displayed similar potency to the
indane core, while the cyclopentane core (example 13b) resulted
in about a half log loss in activity relative to its cyclohexane coun-
terpart (example 12b, Table 2). The tetrahydropyran derivative
(example 13c) and phenyl piperidine analog (example 13d) main-
tained some HCN1 potency, but were both significantly less active
than their corresponding cyclohexane analogs (example 12d and
12e, Table 2). Removing some of the conformational constraint in
the molecule by converting the quaternary carbon atom to a ter-
tiary carbon (Example 13e) had a negative impact on HCN1/
HCN4 potency. Decreasing the conformational restriction between
the benzamide and piperazine groups further by converting the
quaternary carbon to a secondary carbon (13f), led to a complete
loss in HCN1 potency, highlighting the importance of the spatial
relationship of these two groups.
Based upon these results, compound 12m was chosen for fur-
ther characterization. Inhibition of heterologous human HCN2
and HCN3 were measured in the functional assay13 (pIC50 = 5.3
and 5.5, respectively). Experiments were performed in vivo to test
the effectiveness of this novel series of HCN1 blockers to alleviate
nerve injury-induced tactile allodynia in a spared nerve injury
(SNI) model14 and to evaluate their effects on heart rate, using
ZD-7288 for comparison. Figure 2a demonstrates the analgesic effi-
cacy of ZD-7288 (ED50 = 3 mg/kg) and the complete lack of thera-
peutic window with respect to heart rate reduction; significant
0
1
10
100
11m (mg/kg i.p.)
Figure 2. Comparison of the effects ZD-7288 (a) and 12m (b) on tactile allodynia
and bradycardia in mouse.
HCN4 displayed by some of these compounds in a VIPR assay was
substantiated by a similar or slightly greater separation in analge-
sic versus bradycardic potencies. Furthermore, the selectivity pro-
file of these compounds has allowed for testing in a SNI model that
clearly showed a contribution of the HCN1 channel to nerve in-
duced allodynia. Compound 12m represents a novel tool com-
pound for further exploration of the role of HCN subtypes in vivo.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.07.051.
References and notes
bradycardia is evident below the analgesic threshold (ED20
1.3 mg/kg). In contrast, the advantage of compound 12m (Table 2),
is clearly seen in Fig. 2b, where analgesic effects (ED50 = 6 mg/kg)
are substantially separated from bradycardic effects (ED20
25 mg/kg). Thus, the selective HCN1 blocker, 12m provides a ther-
apeutic window for the treatment of neuropathic pain, whereas the
nonselective blocker, ZD-7288, cannot be administered in a con-
centration that would alleviate neuropathic pain without affecting
heart rate.
=
1. Chaplan, S. R.; Guo, H. Q.; Lee, H. D.; Luo, L.; Liu, C.; Kuei, C.; Velumian, A. A.;
Butler, M. P.; Brown, S. M.; Dubin, A. E. J. Neurosci. 2003, 23, 1169.
2. Moosmang, S. J.; Stieber, J.; Zong, X.; Biel, M.; Hofmann, F.; Ludwig, A. Eur. J.
Biochem. 2001, 268, 1646.
3. Lee, D. H.; Chang, L., et al J. Pain 2005, 6, 417.
4. Sun, Q.; Xing, G. G., et al Brain Res. 2005, 1032, 63.
5. Dalle, C.; Eisenach, J. C. Reg. Anesth. Pain Med. 2005, 30, 243.
6. Luo, L.; Chang, L., et al Neuroscience 2007, 144, 1477.
7. Tardiff, J.; Ford, I.; Tendera, M.; Bourassa, M. G.; Fox, K. Eur. Heart J. 2005, 26,
2529.
=
8. Milanesi, R.; Baurscotti, M.; Gnecchi-Ruscone, T.; DiFrancesco, D. N. Engl. J. Med.
2006, 354, 151.
9. Herrmann, S.; Stieber, J.; Stockl, G.; Hofmann, F.; Ludwig, A. EMBO 2007, 26,
4423.
In summary, we disclose a series of novel and selective HCN1
blockers. The 10-fold or higher in vitro selectivity for HCN1 versus

K. J. McClure et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5197–5201
5201
10. Melchiorre, M.; Del Lungo, M.; Guandalini, E.; Martini, E.; Dei, S.; Manetti, D.;
Scapecchi, S.; Teodori, E.; Sartiani, L.; Mugelli, A.; Cerbai, E.; Romanelli, M. N. J.
Med. Chem. 2010, 33, 6993.
12. Yoon, N. M.; Brown, H. C. J. Am. Chem. Soc. 1968, 90, 2927.
13. Maher, M. P.; Nyan-Tsz, W.; Hong-Qing, G.; Dubin, A. E.; Chaplan, S. R.;
Wickenden, A. D. Comb. Chem. High Throughput Screen. 2009, 12, 64.
11. Backer, R. T.; Collado Cano, I.; De Frutos-Garcia, O.; Doecke, C. W.; Fisher, M. J.;
Kuklish, S. L.; Mancuso, V.; Martinelli, M. J.; Mullaney, J. T.; Ornstein, P. L.; Xie,
C. WO 2003061660 (Pct Int. App. 2003061660), 2003.
14. Shields, S. D.; Eckert, W. A.; Basbaum, A. I. J. Pain 2003, 4, 465.
15. Wickenden, A. D.; Maher, M. P.; Chaplan, S. R. Curr. Pharm. Des. 2009, 15, 2149.