Discovery of a novel chemotype of potent human ENaC blockers using a bioisostere approach. Part 2: α-Branched quaternary amines

Article abstract of DOI:10.1016/j.bmcl.2012.02.067

We report the synthesis and biological evaluation of a series of novel α-branched pyrazinoyl quaternary amines for their ability to block ion transport via the epithelial sodium channel (ENaC) in human bronchial epithelial cells (HBECs). Compound 12g has an IC₅₀ of 30 nM and is highly efficacious in the Guinea-pig tracheal potential difference (TPD) model of ENaC blockade with an ED₅₀ of 1 μg kg^-1 at 1 h. In addition the SAR results demonstrate for the first time the chiral nature of the binding site of human ENaC. As such, pyrazinoyl quaternary amines represent a promising new class of ENaC blockers for the treatment of cystic fibrosis that are structurally distinct from the pyrazinoyl guanidine chemotype found in prototypical ENaC blockers such as amiloride.

Full text of DOI:10.1016/j.bmcl.2012.02.067

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2877–2879
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a novel chemotype of potent human ENaC blockers using
a bioisostere approach. Part 2: a-Branched quaternary amines
Thomas Hunt a, , Hazel C. Atherton-Watson b, Stephen P. Collingwood a, Kevin J. Coote b, Sarah Czarnecki b,
⇑
Henry Danahay b, Catherine Howsham a, Peter Hunt a, Derek Paisley b, Alice Young b
a Novartis Institutes of Biomedical Research, Global Discovery Chemistry, Wimblehurst Road, Horsham, West Sussex RH12 5AB, United Kingdom
b Novartis Institutes of Biomedical Research, Respiratory Diseases Area, Wimblehurst Road, Horsham, West Sussex RH12 5AB, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
We report the synthesis and biological evaluation of a series of novel a-branched pyrazinoyl quaternary
amines for their ability to block ion transport via the epithelial sodium channel (ENaC) in human
Received 12 December 2011
Revised 19 February 2012
Accepted 20 February 2012
Available online 3 March 2012
bronchial epithelial cells (HBECs). Compound 12g has an IC50 of 30 nM and is highly efficacious in the
Guinea-pig tracheal potential difference (TPD) model of ENaC blockade with an ED50 of 1 l
g kgꢀ1 at
1 h. In addition the SAR results demonstrate for the first time the chiral nature of the binding site of
human ENaC. As such, pyrazinoyl quaternary amines represent a promising new class of ENaC blockers
for the treatment of cystic fibrosis that are structurally distinct from the pyrazinoyl guanidine chemotype
found in prototypical ENaC blockers such as amiloride.
Keywords:
ENaC blocker
Epithelial sodium channel
Quaternary amine
Cystic fibrosis
Amiloride
Ó 2012 Elsevier Ltd. All rights reserved.
Cystic fibrosis (CF) is an autosomal recessive disease that affects
approximately 70,000 people worldwide. The leading cause of
death in CF patients is respiratory failure resulting from chronic
bacterial infection of the lung. Treatment with physiotherapy and
antibacterial agents has improved the management of CF, but the
median life expectancy in the US is still less than 40 years.1 Current
marketed therapeutics only target the infection and inflammation
symptoms of CF, as such there is an unmet medical need for novel
treatments that address the underlying cause of CF.2 The ‘low
volume’ hypothesis suggests that dehydration of the airway sur-
face liquid in the lung leads to mucostasis in the airway lumen
providing an environment for bacterial colonization leading to
the chronic infections seen in CF patients.3
Inhibition of the epithelial sodium channel (ENaC) is widely be-
lieved to be one method of rehydrating the airway lumen and
improving mucociliary clearance (MCC).4 Supporting evidence
comes from patients with pseudohypoaldosteronism type I, where
loss-of-function mutations in ENaC result in pulmonary edema and
increased rates of MCC.5 Aerosolized amiloride (1, see Fig. 1), is a
potassium-sparing diuretic that blocks ENaC and has been shown
to improve MCC,6 but its effectiveness is limited by poor potency
and pharmacokinetic profile.7 Parion Sciences developed amiloride
analogue 552-02 (2, see Fig. 1), which completed Phase II trials as a
nebulized therapy for CF. Compound 2 is reported to have
improved potency at ENaC in HBECs (IC50 8 nM) and improved
in vivo efficacy and duration of action in a sheep model of MCC
when compared to amiloride.8
As part of our work to identify novel blockers of ENaC for the
treatment of CF, we recently reported the identification of a novel
chemotype of pyrazinoyl quaternary amines, exemplified by com-
pound 3 (see Fig. 2), that blocks ENaC function with comparable
activity to amiloride in vitro and in vivo.9 The SAR indicated that
improved potency could be achieved by addition of large lipophilic
tail groups. Herein, we report our efforts to optimize the potency
and in vivo efficacy of this series.
To achieve an order of magnitude improvement in potency we
sought to identify a specific binding interaction rather than relying
on increasing lipophilicity to improve potency. Li et al. have
reported a stereochemical preference between enantiomers of
compound 4 (see Fig. 2) for ENaC blockade in amphibian ENaC with
a ratio of 4:1 in favor of the (S)-enantiomer.10
We began our investigations to establish if human ENaC pos-
sessed a chiral binding site similar to amphibian ENaC by preparing
a series of a-branched quaternary amines as illustrated in Scheme 1.
Chiral 1,2-amino alcohols 5a–k (purchased commercially with >97%
ee) were protected as their Boc derivatives, 6a–k were then con-
verted into 7a–k a Mitsunobu reaction with phthalimide and
diethyl azodicarboxylate (DEAD). Removal of the phthalimide group
provided primary amines 8a–k which underwent an O-(7-aza-
benzotriazol-1-yl)-N,N,N0,N0-tetramethyluronium hexafluorophos-
phate (HATU)-mediated amide coupling reaction with carboxylic
⇑
Corresponding author. Tel.: +44 1403 323576; fax: +44 1403 323837.
E-mail address: tom.hunt@novartis.com (T. Hunt).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.02.067

2878
T. Hunt et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2877–2879
Table 1
Blockade of ENaC by
O
NH
NH2
a
-branched quaternary amines 3, 12a–k and 16 in HBECs
Cl
N
N
N
H
O
R1
H2N
NH2
Cl
H2N
R2
N
N
R2
R3
N
H
Amiloride
1
OH
OH
N+
NH2
O
O
NH
a,b
Compd
R1
R3
HBEC IC50
(lM)
Cl
N
N
Amilorideb
—
H
H
H
—
H
(R)-Me
(S)-Me
Me
(R)-Et
(S)-Et
(R)-nPr
(S)-nPr
(R)-iPr
(S)-iPr
(S)-nBu
(S)-Bn
(S)-nPr
—
0.22 (93)
0.27 (11)
>10 (2)
8.09 (2)
>30 (2)
0.51 (2)
0.75 (2)
0.29 (6)
0.030 (14)
19.51 (2)
2.39 (3)
N
H
NH2
N
H
3c
(CH2)3-4-OMe-Ph
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
552-02
2
H2N
12a
12b
12c
12d
12e
12f
12g
12h
12i
Me
H
H
H
H
H
H
H
H
H
Figure 1. Examples of pyrazinoyl guanidine ENaC blockers.
Br
O
O
Cl
Cl
N
N
N
N
N+
O
N
H
N
H
12j
12k
16
0.059 (2)
0.17 (6)
0.004 (4)
NH2
H2N
NH2
H2N
NH2
4
3
(CH2)3-4-OMe-Ph
a
Mean IC50 data, number in parentheses refers to the number of repetitions.
A description of the assay conditions used to determine IC50’s and the previ-
Figure 2. Examples of non-pyrazinoyl guanidine ENaC blockers.
b
ously reported data for amiloride can be found in Ref. 13.
c
Data for compound 3 has been previously reported using these assay conditions
acid 9 to give Boc-protected amides 10a–k. The resulting amides
were deprotected under acidic conditions to generate the
branched amines 11a–k that were exhaustively methylated with
methyl iodide to generate the series of -branched quaternary
amines 12a–k. This series of -branched quaternary amines were
in Ref. 9b.
a-
a
a
Compound 12g was docked into an ENaC homology model
based on the crystal structure of the related chicken acid-sensing
ion channel shown in Figure 3.11,12
tested in HBECs to investigate the stereochemical SAR with respect
to ENaC blockade, Table 1 summarizes these data. The introduction
of a single methyl group alpha to the quaternary amine afforded
only weakly active compounds (12a and 12b, Table 1). Introducing
a second methyl group was not tolerated in this series of com-
pounds (12c, Table 1). The ethyl analogues (12d and 12e) do give
a significant improvement in the potency of ENaC blockade but dis-
played no stereochemical preference. The npropyl and ipropyl ana-
logues (12f–i) provide the first evidence that the binding site in
human ENaC has a stereochemical preference. Consistent with Li’s
observations,10 the (S)-enantiomers (12g and 12i) were more potent
than the (R)-enantiomers with a ratio of up to 7:1. In this small data
set an optimal inhibition of ENaC is achieved with compound 12g
(R2 is (S)-npropyl), which is approximately 10-fold more potent than
amiloride. As the size of the alpha substituent increases or decreases
from npropyl, or if branching is introduced, there is a trend for
decreasing potency (compounds 12b, e, g and i–k).
Compounds 12a–k were aligned such that the pyrazine moie-
ties overlaid and the more active enantiomer (compound 12g)
was minimized within the transmembrane (TM) environment.
Figure 3 shows that compound 12g contacts the TM domain
(Van der Waals mesh surface) whilst maintaining the ionic and
hydrogen bonding interactions seen with amiloride whereas the
other epimer would direct the hydrophobic substituent more par-
allel to the helical axis and towards the solvent exposed exterior. It
suggests that the npropyl
a hydrophobic region of one of the pore-forming
This model suggests that smaller or larger -substituents would
-helix residue con-
a
-substituent has a key interaction with
-helix residues.
a
a
interact less favorably with the pore-forming
a
sistent with the observed SAR in Table 1.
In vivo efficacy was assessed using the Guinea-pig TPD model
using intratracheal (it) dosing.13 Amiloride and compound 12g
O
1
1
1
1
R
R
R
R
b
c
a
2
HO
2
N
2
2
HO
H2N
R
R
R
R
NH2
5a-k
HNBoc
HNBoc
6a-k
NHBoc
8a-k
O
7a-k
O
Cl
N
OH
NH2
d
H2N
N
9
O
O
O
1
2
1
1
R
R
R
2
Cl
N
Cl
H2N
N
R
Cl
H2N
N
N
f
e
N
H
N
H
2
N
R
R
N+
H
NH2
NHBoc
I-
H2N
N
NH2
N
NH2
NH2
12a-k
11a-k
10a-k
a; R1 = H, R2 = Me
b; R1 = Me, R2 = H
c; R1 = R2 = Me
e; R1 = Et, R2 = H
f; R1 = H, R2 = nPr
g; R1 = nPr, R2 = H
h; R1 = H, R2 = iPr
i: R1 = iPr, R2 = H
j; R1 = nBu, R2 = H
k; R1 = Bn, R2 = H
d; R1 = H, R2 = Et
Scheme 1. Reagents and conditions: (a) (Boc)2O, Et3N, CH2Cl2, rt (99%); (b) phthalimide, PPh3, DEAD, CH2Cl2, rt (39–92%); (c) N2H4ꢁH2O, EtOH/CH2Cl2, rt; (d) HATU, 4-
methylmorpholine, DMF, rt (34–74% over two steps); (e) 4 M HCl in dioxane, rt (65–92%); (f) MeI, K2CO3, MeCN, rt (31–99%).

T. Hunt et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2877–2879
2879
We have previously demonstrated that adding large lipophilic
groups around the quaternary amine provides up to a 20-fold
improvement in potency.9 The modeling of compound 12g sug-
gests that this may also be tolerated in the a-branched quaternary
amine series. To test this hypothesis compound 16 was synthe-
sized as illustrated in Scheme 2. Reductive amination of amine
11g with aldehyde 1414 (formed by oxidation of alcohol 13 with
Dess–Martin Periodane) provides amine 15. As before, quaterniza-
tion is accomplished by exhaustive methylation with methyl io-
dide to give quaternary amine 16.
The addition of this tail group provides an approximate 10-fold
enhancement in in vitro ENaC potency (Table 1, compound 16 vs
12g). This demonstrates that our initial bioisostere hypothesis
can deliver compounds equipotent with the leading amiloride-
based analogues such as compound 2. The importance of the alpha
substituent can be seen by comparing compounds 3 and 16 with an
approximate 50-fold improvement in ENaC blockade suggesting a
key interaction within the binding site has been utilized. It is
important to note that, due to the size and shape differences be-
tween a quaternary amine and a guanidine, the chiral SAR devel-
oped can not be readily transferred on to the pyrazinoyl
guanidine chemotype.
Figure 3. Compound 12g docked into an ENaC homology model.
In summary, we have demonstrated that the binding site of hu-
man ENaC displays a stereochemical preference that can be suc-
cessfully exploited in the quaternary amine series. This led to the
Table 2
ENaC Guinea-pig cross-reactivity and in vivo Guinea-pig TPD efficacy for amiloride
identification of a series of
as compounds 12g and 16 which potently inhibit sodium ion trans-
port via ENaC in HBECs. In compound 12g the (S)-npropyl
-substi-
a-branched quaternary amines such
and compounds 2 and 12g
a
Compd
HBEC IC50
(
lM)
Guinea-pig
Guinea-pig
a
b
a
FRT IC50
(
l
M)
TPD 1 h ED50 (l )
g kgꢀ1
tuent greatly improves potency and in vivo efficacy demonstrating
that a bioisostere approach can deliver a novel class of human
ENaC blockers that are comparable to the leading amiloride-based
analogues in vitro and in vivo.
Amilorideb
0.22 (93)
0.002 (12)
0.030 (14)
0.54 (37)
0.004 (32)
0.030 (4)
16
0.2
1.0
2b
12g
a
Mean IC50 data, number in parentheses refers to the number of repetitions.
A description of these assay conditions and the previously reported data for
amiloride and compound 2 can be found in Ref. 13.
b
References and notes
1. Storey, S.; Wald, G. Nat. Rev. Drug Disc. 2008, 7, 555.
2. Clunes, M. T.; Boucher, R. Curr. Opin. Pharmacol. 2008, 8, 292.
3. Boucher, R. C. J. Intern. Med. 2007, 261, 5.
4. Mall, M. A. Exp. Physiol. 2009, 94, 171.
5. Kerem, E.; Bistritzer, T.; Hanukoglu, A. N. Eng. J. Med. 1999, 341, 156.
6. (a) Köhler, D.; App, E.; Schmitz-Schumann, M.; Würtemberger, G.; Matthys, H.
Eur. J. Respir. Dis. Suppl. 1986, 146, 319; (b) App, E. M.; King, M.; Helfesrieder, R.;
Köhler, D.; Matthys, H. Am. Rev. Resp. Dis. 1990, 141, 605.
OMe
a
OMe
HO
O
13
14
O
7. Hoffmann, T.; Senier, I.; Bittner, P.; Hüls, G.; Schwandt, H. J.; Lindemann, H. J.
Aerosol Med. 1997, 10, 147.
O
Cl
H2N
N
N
H
8. Hirsh, A. J.; Zhang, J.; Zamurs, A.; Fleegle, J.; Thelin, W. R.; Caldwell, R. A.;
Sabater, J. R.; Abraham, W. M.; Donowitz, M.; Cha, B.; Johnson, K. B.; St. George,
J. A.; Johnson, M. R.; Boucher, R. C. J. Pharmacol. Exp. Ther. 2008, 325, 77.
9. (a) Collingwood, S. P.; Devereux, N. J.; Howsham, C.; Hunt, P.; Hunt, T. A.
International Patent, WO 150 137 A2, 2009.; (b) Hunt, T.; Danahay, H.;
Atherton-Watson, H. C.; Axford, J.; Collingwood, S. P.; Coote, K. J.; Cox, B.;
Czarnecki, S.; Devereux, N.; Howsham, C.; Hunt, P.; Paddock, V.; Paisley, D.;
Young, A. Bioorg. Med. Chem. Lett. 2012, 22, 929.
b
Cl
N
OMe
NH2
N
H
N
NH2
HN
11g
H2N
N
NH2
15
O
c
10. Li, J. H.; Shaw, A.; Kau, S. T. J. Pharmacol. Exp. Ther. 1993, 267, 1081.
11. (a) Jasti, J.; Furukawa, H.; Gonzales, E. B.; Gouaux, E. Nature 2007, 449, 316; (b)
Stockand, J. D.; Staruschenko, A.; Pochynyuk, O.; Booth, R. E.; Silverthorn, D. U.
IUBMB Life 2008, 60, 620.
Cl
N
N
OMe
N
N+
H
H2N
NH2
CF3COO-
12. The homology model for ENaC was created using the modified chicken ASIC
16
crystal structure (PDB code 2QTS) as a template. The sequences of Human
a, b
and subunits were aligned with the 2QTS sequence using the alignment
c
Scheme 2. Reagents and conditions: (a) Dess–Martin Periodane, CH2Cl2, rt (90%);
(b) 11g, NaB(OAc)3H, CH2Cl2, reflux (69%); (c) MeI, K2CO3, MeCN, 80 °C (99%).
methodology within MolSoft’s ICM Pro software with some manual alterations
in the regions where the homology was low due to the presence of large amino
acid insertions in the ENaC sequences. The SAR around the amiloride pyrazine
ring is known to be tight and so this was positioned into the transmembrane
region of the model in such a way that hydrogen bonding could be maximized,
the chloro substituent was directed towards a gap in the TM bundle (to
account for the ability to replace this group by hydrophobic replacements such
as a phenyl ring) and this directed the basic guanidine towards the exterior of
the TM bundle (which was likely considering the Parion derivatives such as
both show excellent cross-reactivity to Guinea-pig ENaC (using
Fischer Rat Thyroid (FRT) cells transiently infected with Guinea-
pig ENaC).13 Compound 3 was the first example of a quaternary
amine that was tested in this model and gave promising results
compound 2) and also close to GLU68 on the
a subunit which could act as a
with an ED50 of 44 l . Pleasingly, when the more potent
g kgꢀ1 9b
counterion to the guanidine.
ENaC blocker, compound 12g, was dosed it showed a significant
13. Coote, K. J.; Atherton, H.; Sugar, R.; Burrows, R.; Smith, N. J.; Schlaeppi, J.-M.;
Groot-Kormelink, P. J.; Gosling, M.; Danahay, H. Br. J. Pharmacol. 2008, 155,
1025.
improvement in efficacy and was 16-fold more efficacious than
amiloride with an ED50 of 1 l
g kgꢀ1 and is comparable to com-
14. Evans, W. C.; Walker, N. J. Chem. Soc. 1947, 1571.
pound 2 (Table 2).