The discovery and structure-activity relationships of 2-(piperidin-3-yl)-1H-benzimidazoles as selective, CNS penetrating H1-antihistamines for insomnia

Article abstract of DOI:10.1016/j.bmcl.2010.03.027

A series of 2-(3-aminopiperidine)-benzimidazoles were identified as selective H₁-antihistamines for evaluation as potential sedative hypnotics. Representative compounds showed improved hERG selectivity over a previously identified 2-aminobenzimidazole series. While hERG activity could be modulated via manipulation of the benzimidazole N1 substituent, this approach led to a reduction in CNS exposure for the more selective compounds. One example, 9q, retained a suitable selectivity profile with CNS exposure equivalent to known centrally active H₁-antihistamines.

Full text of DOI:10.1016/j.bmcl.2010.03.027

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2916–2919
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery and structure–activity relationships of 2-(piperidin-3-yl)-1H-
benzimidazoles as selective, CNS penetrating H1-antihistamines for insomnia
*
Karine Lavrador-Erb, Satheesh Babu Ravula , Jinghua Yu, Said Zamani-Kord, Wilna J. Moree, Robert E. Petroski,
*
Jianyun Wen, Siobhan Malany, Samuel R. J. Hoare, Ajay Madan, Paul D. Crowe, Graham Beaton
Neurocrine Biosciences, 12780 El Camino Real, San Diego, CA 92130, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 2-(3-aminopiperidine)-benzimidazoles were identified as selective H1-antihistamines for eval-
uation as potential sedative hypnotics. Representative compounds showed improved hERG selectivity
over a previously identified 2-aminobenzimidazole series. While hERG activity could be modulated via
manipulation of the benzimidazole N1 substituent, this approach led to a reduction in CNS exposure
for the more selective compounds. One example, 9q, retained a suitable selectivity profile with CNS expo-
sure equivalent to known centrally active H1-antihistamines.
Received 21 January 2010
Revised 3 March 2010
Accepted 5 March 2010
Available online 10 March 2010
Keywords:
H1-Antihistamines
H1-Antagonist
Insomnia, hERG
CNS
Ó 2010 Elsevier Ltd. All rights reserved.
Benzimidazole
SAR
Sedative
Selective
Insomnia is one of the most common CNS disorders particularly
in industrialized nations.1 Sleep disorders have significant eco-
nomic impact on both managed care2 and nations’ workforces as
a consequence of higher work absenteeism and decreased job per-
formance.3 We have been interested in the discovery of selective,
centrally acting H1-antihistamines for development of an effective
insomnia therapeutic that is free of the issues exhibited by over-
the-counter (OTC) sedating antihistamines. These include selectiv-
ity for muscarinic receptors, a property thought to contribute to
undesirable side effects such as dry mouth, blurred vision, consti-
pation, tachycardia, urinary retention.4 Next-day impairment after
bedtime use of these antihistamines is also common and has been
attributed to long plasma half-lives and protracted CNS exposure.5
As described in our previous publication,6 we used the core of the
known H1-antihistamine mizolastine7 to identify the sedating com-
pound, 1 (Fig. 1). This compound and other basic analogs in this ser-
ies were, however, determined to be potenthERG channelinhibitors,
as has been observed for other H1-antihistamine series.8 Optimiza-
tion of this class for the hERG liability resulted in the 2-aminobenz-
imidazole 2 (Fig. 1), a relatively selective antihistamine with
acceptable blood–brain barrier (BBB) penetration (30 mpk oral dose
at 4 h: brain concentration 262 ng/g; B/P ratio of 2.3).6 However, 2
displayed a less than optimal profile with high metabolic clearance
and poor solubility (<0.01 mg/mL at pH 7.4).9 In an effort to improve
properties of the 2-aminobenzimidazole series while maintaining
selectivity, replacement of the 4-aminopiperidine moiety in 1 by
alternative diamines was examined, using chemistry previously de-
scribed.6 Compounds were assessed for selectivity versus mono-
amine receptors as well as inhibition of CYP2D6, CYP3A4 enzymes
N
N
N
N
N
N
N
N
N
O
N
N
N
N
N
N
N
N
F
F
1
2
Figure 1. Early 2-aminobenzimidazole H1-antihistamine leads.
O
4
3
* Corresponding authors. Tel.: +1 619 200 5689 (S.B.R.); tel.: +1 858 337 1801 (G.B.).
E-mail addresses: satheesh.ravula@gmail.com (S.B. Ravula), beaton.graham@
gmail.com (G. Beaton).
O
Figure 2. 3-Aminopyrrolidin-1-yl-2-aminobenzimidazoles.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.03.027

K. Lavrador-Erb et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2916–2919
2917
Table 1
H1-Antihistamine activity, M1 selectivity, CYP2D6 inhibition, predicted intrinsic clearance and hERG electrophysiology assay results for compounds 1–4
b
Compd
H1 Kia (nM)
M1 Kia lM)
CYP2D6 IC50
(lM)
Pred. Cl Int. (mL/min/kg)
hERG IC50 (nM)
1
2
3
4
3.9 0.3
6.9 0.9
8.4 2.0
4.8 0.6
4.8 0.8
>10
2.5 0.5
1.5 0.1
11.0
5.4
>10
>10
106
242
61
29
809
492
233
21
a
SEM for Ki values derived from dose–response curves generated from triplicate or more data points.
IC50 for CYP3A4 inhibition >5 lM for all compounds.
b
and the hERG channel. From this effort we identified 3-aminopyrr-
olidines 3 and 4 (Fig. 2) with profiles of interest. Data for these exam-
ples is shown in comparison to compounds 1 and 2 (Table 1).
Both compounds 3 and 4 retained selectivity for M1 and
CYP2D6. Interestingly hERG selectivity for these analogs was sig-
nificantly improved compared to compound 1 from the 4-amino-
piperdine series. From our previous studies,5 hERG selectivity
was only slightly influenced by changes to the N1 substituents em-
ployed and suggested that the introduction of a basic moiety with
alternative orientation relative to the benzimidazole core could
facilitate improved hERG selectivity. Both 3 and 4 were also signif-
icantly more stable in a human liver microsome (HLM) assay. This
data was encouraging and offered the potential to access com-
pounds with improved solubility and enhanced bioavailability.
Based on this information we focused our attention for further
studies on the more novel benzimidazole scaffold 9 shown in
Scheme 1. Our intent was to first validate the hypothesis that se-
lected analogs based on 9 were potent H1-antihistamines with
suitable selectivity against a range of key targets. To achieve simi-
lar selectivity to our work in a parallel series that was described
earlier,10 candidate compounds were required to demonstrate high
H1 binding affinity (Ki <10 nM) with at least 100-fold binding selec-
tivity versus the representative muscarinic M1 receptor. Selectivity
of the order of 1000-fold was deemed acceptable for CYP enzyme
inhibition. While our leads from this work demonstrated weak
hERG channel inhibition (hERG IC50/H1 Ki selectivity of 33610),
in vivo characterization indicated an absence of cardiovascular
risks and safety margins significantly higher than previous guide-
lines for the assessment of hERG inhibitors.11 Based on these data
we reasoned that hERG IC50/H1 Ki selectivity of 400 or greater
would be acceptable for candidate compounds. With this accom-
plished we planned to assess stability and in vivo central exposure
for key compounds to determine their candidacy for evaluation as
sleep agents.
Synthesis of these benzimidazole analogs is outlined in
Scheme 1. We started with O-phenylenediamine 5, which was
treated with N-BOC-piperidinecarboxylic acid 6 to give benzimid-
azole 7. N-alkylation of 7 was accomplished using the correspond-
ing alkyl bromide to give the analog 8. Removal of the BOC group
was achieved with TFA to afford the intermediate amine that
was further modified by reductive amination or alkylation to yield
the final benzimidazoles 9a–p.
From preliminary binding data for the racemic compounds 9a–
p, we confirmed that this alternative orientation of the basic center
within the 3-piperidinyl moiety was accommodated in the H1
pharmacophore, all compounds with the exception of 9g showing
potent H1 affinity (Table 2). Within this set both benzylic (9a,b,
9h–l) and extended alkyl (9c–d, 9m–n) substitution was tolerated
at the benzimidazole N1 position. Replacement of the arene with a
heterocycle within the N1 benzyl moiety (9e–f, 9o–p) was also
possible with some reduction in affinity. In addition, substitution
at the piperidine nitrogen was well tolerated for the analogs as-
sessed. Selectivity of this series was evaluated for M1 binding affin-
ity, P450 enzyme inhibition (CYP2D6, CYP3A4) and hERG binding.6
From this assessment, the compounds were determined to be gen-
erally very selective for M1, with only the phenoxyethyl analog, 9c,
showing significant binding affinity. In addition, we were encour-
aged with the observation that the series showed a general selec-
tivity on the basis of hERG binding, similar to that determined
for the original lead, 2 (Ki 3849 nM). 6 Selectivity for CYP2D6 inhi-
bition was, however, reduced. While the compounds 9a and 9e–f
were reasonably selective for CYP2D6, other similar analogs exhib-
ited significant inhibition of the P450 enzyme. Two trends were
noted in the structure–activity relationships for CYP2D6. First, en-
zyme inhibition appeared to increase with increasing hydropho-
bicity at R2, as exemplified for comparisons of 9a with 9h–i, 9d
with 9m and 9f with 9o. In the case of the p-methoxybenzyl series
CYP2D6 inhibition was significant for all comparable samples. The
O
O
N
O
O
a
NH2
NH2
O
N
N
+
HO
N
H
5
6
7
O
N
R2
O
c, d
b
N
N
N
N
7
N
R1
R1
9
8
Scheme 1. Reagents and conditions: (a) neat, 120 °C (65%); (b) R1-CH2-Br, K2CO3, DMF, 80 °C (70–85%); (c) TFA, CH2Cl2, (95%); (d) R2(@O)H, Na(OAc)3BH or R2-X, Et3N, THF
(45–90%).

2918
K. Lavrador-Erb et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2916–2919
Table 2
H1-Antihistamine activity, M1 selectivity, CYP2D6 inhibition and hERG dofetilide assay results for compounds 9a–p
R2
N
N
N
R1
9a-p
R1
R2
H1 Kia (nM)
M1 Kib
(lM)
CYP2D6 IC50 (nM)
hERG Kid (nM)
c
Compd
9a
9b
9c
9d
p-F-Ph
Me
Me
Me
Me
0.5 0.4
1.2 0.3
0.8 0.3
4.6 0.3
2.2 0.7a
9.7 0.2
0.04 0.05
2.60 0.03
3065
465
>10,000
>5000
5291
p-MeOPh
CH2OPh
CH2OEt
NTe
1556
1262
N
9e
9f
Me
Me
Me
14 2b
3.3 0.9
605 63b
4.9 0.1
2.4 0.1
NTe
>10,000
15,989
>10,000
>10,000
638
N
S
O
9g
>10,000
9h
9i
9j
9k
9l
9m
9n
p-F-Ph
p-F-Ph
CH(CH3)2
Cyclohexyl
CH(CH3)2
Cyclohexyl
Tetrahydropyran-4-yl
Cyclohexyl
Tetrahydropyran-4-yl
0.8 0.2
1.9 0.2b
1.8 0.2
4.7 0.5b
5.0 0.9b
5.2 0.9b
5.4 0.9
3.7
618
195
717
122
1901
526
4393
1373
926
2649
1840
>10,000
>10,000
>3000
p-MeOPh
p-MeOPh
p-MeOPh
CH2OEt
CH2OEt
12
13
1
11.9 0.01
>10
>10
14
2
N
9o
9p
Cyclohexyl
3.8 0.5
1.9 0.2
3.2 0.5
480
1183
S
N
Tetrahydropyran-4-yl
10
1
6453
>3000
S
a
b
c
SEM for Ki values derived from dose response curves generated from triplicate or more data points.
Ki values average of two data points.
IC50 for CYP3A4 inhibition >5 lM for all compounds except 9k [IC50 = 3.3 lM].
Ki values were derived from single or duplicate data points.
NT = not tested.
d
e
addition of hydrophobic moieties at the piperidino nitrogen also
increased hERG binding affinity. In all cases, the reduction of
hydrophobicity by substitution of the tetrahydropyranyl moiety
for cyclohexyl showed a general improvement in CYP2D6 IC50
and a decrease in hERG binding affinity. Comparison of analogs
9a with 9c–f suggested a second trend in which CYP2D6 inhibition
could also be influenced by the substitution at N1 of the benzimid-
azole. In particular, replacement of the R1 arene with a heterocycle
was beneficial in providing significant reductions in enzyme
inhibition.
Table 3
H1-Antihistamine activity, M1 selectivity, CYP2D6 inhibition and hERG dofetilide assay results for the enantiomeric compounds 9q–w
R2
N
N
N
R1
9q-w
R1
R2
Stereo-chemistry
H1 Kia (nM)
M1 Kib
(
lM)
CYP2D6 IC50 (lM)
hERG Kid (nM)
c
Compd
9q
9r
9s
9t
p-F-Ph
p-F-Ph
CH2OEt
CH2OEt
Me
Me
Me
Me
R
S
R
S
0.9 0.2
1.10 0.03
2.6 0.4
2.8 0.2
3.5 0.5
1.7 0.1
>10,000
2.3
2.8
15.5
6.6
3860
6515
>10,000
>10,000
32.9 0.1b
N
9u
Me
R
9
2
7
2a
53.3
10,336
N
9v
Me
S
12
1
3.7 0.1
>10
47.5
32.2
>10,000
>3000
9w
CH2OEt
Tetrahydropyran-4-yl
R
6.4 0.3
a
b
c
SEM for Ki values derived from dose response curves generated from triplicate or more data points.
Ki values average of two data points.
None of the compounds showed appreciable inhibition of CYP3A4.
Ki values were derived from single or duplicate data points.
d

K. Lavrador-Erb et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2916–2919
2919
While initial data indicated suitable selectivity profiles were
achievable, CYP2D6 inhibition was identified as a significant po-
tential liability. Preferred substitutions for selectivity at M1 were
p-F benzyl, ethoxyethyl, pyridine-2-ylmethyl and 2-methylthia-
zol-4-ylmethyl with R2 as methyl. Tetrahydropyran-substituted
piperidine analogs were also of potential interest given the ability
of this feature to modulate off-target interactions. Prior to testing
exposure and stability of these compounds a more detailed selec-
tivity study was conducted on the enantiomers of selected key
compounds which were prepared according to Scheme 1 using
the appropriate chiral BOC 3-piperidine carboxylic acid. Data for
compounds 9q–w is summarized in Table 3. From this data, chiral-
ity had little impact on H1 affinity for analogs with R1 as the benzyl
or pyridin-2-ylmethyl motifs. Chiral preference for the R enantio-
mer was observed for the ethoxyethyl substitution, 9s being
approximately 10-fold more potent than 9t. No significant binding
was observed for any of the compounds for either H3 or M3 recep-
tors. In general the more potent compounds were approximately
1000-fold selective for H1 over the other targets assessed. Patch
clamp analysis of 9q indicated a hERG IC50 of 721 nM. This result
was in line with previous observations of trends comparing hERG
binding and electrophysiology data6 and indicated that 9q demon-
strated improved selectivity compared to the lead 2 previously
identified [hERG IC50/H1 Ki = 800 for 9q compared to 117 for 2].
Compound 9w had weaker inhibition in the patch clamp assay
suitable exposure. However, no measurable brain exposure could
be detected at doses up to 30 mg/kg after 4 h. The reduction in
exposure for this and other analogs appears to be dominated by ef-
fects of the benzimidazole R1 substituent, presumably by local
reductions in hydrophobicity and increases in PSA and H-bond
acceptors.13 Calculations of log P and polar surface area14 indicated
that estimates of log P for the compounds in this class were similar
[2.73–4.25] to those of the known sedating antihistamines diphen-
hydramine [3.35], doxepin [3.77] and triprolidine [3.38]. In con-
trast, only polar surface area for 9q [21.1] was similar to that for
the sedating antihistamines [12.5–16.1] with the other analogs
exhibiting significantly higher values [30.2–71.4]. While these R1
substitutions improved in vitro selectivity for H1 over hERG and re-
duced CYP2D6 inhibition compared to 9q, central exposure was
significantly decreased. Nevertheless, 9q retained a suitable selec-
tivity profile and central exposure as a representative of a new
class of selective brain penetrating H1-antihistamines.
In summary, starting from a series of 2-aminobenzimidazoles
we identified a novel class of 2-(piperidin-3-yl)-1H-benzimidaz-
oles as potent and selective H1-antihistamines as potential agents
for the treatment of insomnia. SAR studies within this class indi-
cated that manipulation of the N1 substituent on the benzimid-
azole led to selective compounds with reduced hERG activity
although these analogs lacked CNS penetration. One compound,
9q, retained CNS exposure equivalent to known sedating antihista-
with an IC50 value of 2.1
at 10 M].
lM. 9u was a still weaker inhibitor [39%
mines with
optimization.
a suitable selectivity profile to warrant further
l
In the absence of a high throughput in vivo assessment, suitable
CNS exposure was required of the leading compounds prior to their
evaluation for sedative hypnotic potential in a rat electroencepha-
lography (EEG) model. To assess suitable CNS penetration in poten-
tial lead compounds, representative analogs were evaluated for
their ability to penetrate the BBB in rodents using cassette PK stud-
ies.6 Groups of five compounds including the short-acting brain pe-
netrating antihistamine triprolidine6 (3) were administered (iv) to
rats and brain levels and B/P ratios were determined. Analogs with
CNS penetration similar or better than triprolidine were required
for the candidate compound to be further evaluated in the EEG
model. Cassette data for representatives of the compounds synthe-
sized is summarized in Table 4.
Of the compounds studied, only 9q displayed significant brain
levels that were similar to the sedating antihistamine triprolidine.
Neither the heteroaryl substituted compounds 9u, 9f or 9p nor eth-
oxyethyl analogs 9d (racemate of 9s) or 9n (racemate of 9w)
achieved significant brain exposure greater than 10 ng/g at the iv
dose. These exposures were less than 25% of the relative CNS expo-
sure achieved by the sedating antihistamine control. This data indi-
cated that, despite excellent selectivity profiles, compounds 9s and
9u–w were unsuitable as candidates because of poor CNS penetra-
bility.12 The enantiomer of 9n (9w), was assessed in a discrete PK
study to determine whether increasing oral doses could achieve
Acknowledgments
The authors wish to thank John Harman and Chris DeVore for
analytical support, Dr. Jaimie K. Rueter for solubility data and Dr.
John Saunders, Dr. Paul Conlon, Dr. Wendell Wierenga and Dr. Haig
Bozigian for program support.
References and notes
1. Ohayon, M. M.; Lemoine, P. L’Encephale 2004, 30, 135.
2. Fullerton, D. S. P. Am. J. Manag. Care 2006, 12, S246.
3. Zammit, G. K.; Weiner, J.; Damato, N.; Sillup, G. P.; McMillan, C. A. Sleep 1999,
22, S379.
4. (a) Kubo, N.; Shirakawa, O.; Kuno, T.; Tanaka, C. Jpn. J. Pharmacol. 1987, 43, 277;
(b) Meolie, A. L.; Rosen, C.; Kristo, D.; Kohrman, M.; Gooneratne, N.; Aguillard,
R. N.; Fayle, R.; Troell, R.; Townsend, D.; Claman, D.; Hoban, T.; Mahowald, M.
Clin. Sleep Med. 2005, 1, 173.
5. Kay, G. G.; Plotkin, K. E.; Quig, M. B.; Starbuck, V. N.; Yasuda, S. Am. J. Manag.
Care 1997, 3, 1843.
6. Coon, T.; Moree, W. J.; Li, B.; Yu, J.; Zamani-Kord, S.; Malany, S.; Santos, M. A.;
Hernandez, L. M.; Petroski, R. E.; Sun, A.; Wen, J.; Sullivan, S.; Haelewyn, J.;
Hedrick, M.; Hoare, S. J.; Bradbury, M. J.; Crowe, P. D.; Beaton, G. Bioorg. Med.
Chem. Lett. 2009, 19, 4380. and references cited therein.
7. Benavides, J.; Schoemaker, C.; Dana, C.; Laustre, Y.; Delahaye, M.; Prouteau, M.;
Manoury, P.; Allen, J.; Scatton, B.; Langer, S. Z.; Arbilla, S. Arzneim.-Forsch/Drug
Res. 1995, 45, 551.
8. Aslanian, R.; Piwinski, J. J.; Zhu, X.; Priestley, T.; Sorota, S.; Du, X.-Y.; Zhang, X.-
S.; McLeod, R. L.; West, R. E.; Williams, S. M.; Hey, J. A. Bioorg. Med. Chem. Lett.
2009, 19, 5043.
9. Measured using GLPKa instrumentation (pION Inc.): potentiometric method
using 0.15 M KCl buffer.
10. Moree, W. J.; Li, B.; Jovic, F.; Coon, T.; Yu, J.; Gross, R. S.; Tucci, F. C.; Marinkovic,
D.; Malany, S.; Bradbury, M. J.; Hernandez, L. M.; O’Brien, L.; Wen, J.; Wang, H.;
Hoare, S. R. J.; Petroski, R. E.; Sacaan, A.; Madan, A.; Crowe, P. D.; Beaton, G. J.
Med. Chem. 2009, 52, 5307.
11. Redfern, W. S.; Carlsson, L.; Davis, A. S.; Lynch, W. G.; MacKenzie, I.; Palethorpe,
S.; Siegl, P. K. S.; Strang, I.; Sullivan, A. T.; Wallis, R.; Camm, A. J.; Hammond, T.
G. Cardiovasc. Res. 2003, 58, 32.
12. In some cases, where insufficient amounts of enantiomer were available, the
racemic compound was assessed in the cassette PK experiment. In control
experiments representative enantiomers and their racemates were shown to
exhibit similar CNS penetration properties in cassette studies. No differences
were observed in clearance profile between racemate and enantiomers.
13. Feher, M.; Sourial, E.; Schmidt, J. M. Int. J. Pharm. 2000, 201, 239.
14. Calculated using ACD Labs Software Suite.
Table 4
CNS exposure results for selected analogs compared to triprolidine
Compd
Pred. hCl int.a
(ml/min/kg)
[B]b
ng/g
[P]b
ng/ml
B/Pb
[B]b ng/g
Triprolidine
9q
9u
9d
9n
9f
21.0
23.4
34.1
80.0
32.9
98.3
39.5
4.9
10
7.5
5.4
3.5
14.7
3.5
13.2
3.4
5.6
4.1
2.7
1.4
0.8
2.6
1.0
0.9
46.5
46.5
89.4
37.0
37.0
37.0
9p
a
Predicted based on HLM stability studies.
Cassette dose 1 mg/kg, iv.
b