Lead optimization of 2-(piperidin-3-yl)-1H-benzimidazoles: Identification of 2-morpholin- and 2-thiomorpholin-2-yl-1H-benzimidazoles as selective and CNS penetrating H1-antihistamines for insomnia

Article abstract of DOI:10.1016/j.bmcl.2011.10.115

The structure-activity relationships of 2-(piperidin-3-yl)-1H- benzimidazoles, 2-morpholine and 2-thiomorpholin-2-yl-1H-benzimidazoles are described. In the lead optimization process, the pK_a and/or log P of benzimidazole analogs were reduced e

Full text of DOI:10.1016/j.bmcl.2011.10.115

Bioorganic & Medicinal Chemistry Letters 22 (2012) 421–426
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Lead optimization of 2-(piperidin-3-yl)-1H-benzimidazoles: Identification
of 2-morpholin- and 2-thiomorpholin-2-yl-1H-benzimidazoles as selective
and CNS penetrating H1-antihistamines for insomnia
⇑
Satheesh Babu Ravula , Jinghua Yu, Joe A. Tran, Melissa Arellano, Fabio C. Tucci, Wilna J. Moree,
Bin-Feng Li, 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:
The structure–activity relationships of 2-(piperidin-3-yl)-1H-benzimidazoles, 2-morpholine and 2-thi-
omorpholin-2-yl-1H-benzimidazoles are described. In the lead optimization process, the pKa and/or logP
of benzimidazole analogs were reduced either by attachment of polar substituents to the piperidine
nitrogen or incorporation of heteroatoms into the piperidine heterocycle. Compounds 9a and 9b in the
morpholine series and 10g in the thiomorpholine series demonstrated improved selectivity and CNS pro-
files compared to lead compound 2 and these are potential candidates for evaluation as sedative
hypnotics.
Received 3 August 2011
Revised 26 October 2011
Accepted 31 October 2011
Available online 15 November 2011
Keywords:
Lead optimization
Structure–activity relationship (SAR)
Benzimidazole
Ó 2011 Elsevier Ltd. All rights reserved.
H1-antihistamines
Insomnia
CNS penetration
Among CNS disorders, insomnia has significant economic im-
pact in industrialized nations.1 Several over-the-counter (OTC)
antihistamines, while sedating, either show a lack of selectivity to-
wards muscarinic receptors and/or prolonged CNS exposure2 that
result in side effects including dry mouth, blurred vision, constipa-
tion, tachycardia, urinary retention, memory deficits and next-day
impairment. As described in our previous publications,3,4 we have
been interested in the discovery of selective, centrally acting H1-
antihistamines for insomnia that are free from the issues exhibited
by their OTC counterparts.
demonstrated improvements in both selectivity profile for the
hERG channel [hERG IC50/H1 Ki 117 for 1 compared to 800 for 2]
and stability. From our initial SAR, a change in orientation of the
basic center in 2 was a key step in establishing the required H1
affinity and selectivity.
In this Letter we describe the synthesis and SAR studies around
the benzimidazole lead 2 in a series of analogs represented by the
general structure 4. Our objective was to further characterize this
compound class for improved selectivity while maintaining CNS
exposure and identify suitable leads for in vivo evaluation. From
our previous work,4 while variation of the R1 substituent with het-
erocycles afforded a number of selective analogs, the resultant in-
crease in polar surface area reduced CNS exposure to unacceptable
levels. Previous exploration of the R2 substituent was limited. In
our initial approach we explored the effect of addition of more po-
lar groups onto the piperidine ring nitrogen that would reduce pKa
and/or logP via electron withdrawing effects, a method that was
consistent with previously described optimization strategies.5 In-
deed, the addition of polar surface to the R2 substitution was pre-
In a previous study, we identified the piperidine lead 1 as a high
affinity and selective H1-antihistamine.3 This compound exhibited
limitations with both poor metabolic stability and solubility.
Further optimization of compound 1 resulted in the discovery
of (R)-1-(4-Fluoro-benzyl)-2-(1-methyl-piperidin-3-yl)-1H-
benzimidazole,
2 as a
high affinity H1-antihistamine.4 Both
compounds 1 and 2 showed central exposure comparable to the
known sedating antihistamine triprolidine 3 in cassette PK studies
(Fig. 1, Table 1).3,4 However, when compared to 1, compound 2
viously shown to have
a positive impact on P450 enzyme
inhibition.4 From this effort we sought to evaluate the possibility
of identifying additional analogs of 2 that maintained CNS expo-
sure and further improved selectivity. For this study the R1 group
was maintained as p-F-benzyl which had been previously
⇑
Corresponding authors. Tel.: +1 619 200 5689 (S.B.Ravula); tel.: +1 858 337
1801 (G. Beaton).
E-mail addresses: satheesh.ravula@gmail.com (S.B. Ravula), beaton.graham@
gmail.com (G. Beaton).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.10.115

422
S. B. Ravula et al. / Bioorg. Med. Chem. Lett. 22 (2012) 421–426
enantiomers was achieved by chiral SFC when chiral acids were
not used for the synthesis of the benzimidazole.
CH3
N
N
N
N
N
N
N
N
In order to evaluate effects of the proposed substitutions, com-
pounds were tested in a series of assays including the primary his-
tamine H1 receptor binding assay.3 For initial selectivity,
compounds were subsequently tested for hERG affinity using a
3H-dofetolide binding assay as an alternative to whole cell patch
clamp electrophysiology. Inhibition of CYP2D6 and CYP3A4 was
also investigated at an early stage. Key examples were subse-
quently assessed for receptor selectivity using muscarinic M1
receptor binding affinity. As recent data7–9 has indicated the poten-
tial importance of metabolic stability in reducing the variability in
PK profile of sedating H1-antihistamines, stability of the com-
pounds was also assessed using human liver microsomes (HLM).
Profiles for a series of 2-(piperidin-3-yl)-1H-benzimidazole ana-
logs compared to 2 and its racemate (2a) are shown in Table 2. Addi-
tion of several polar groups including alcohols, amides and amide
bio-isosteric heterocycles resulted in compounds with acceptable
(0.5–24.6 nM) H1 binding affinity (4a–f), but only the alcohol (4f)
showed comparable affinity to 2a. In the dofetilide binding assay,
inhibitionby 4a–f was notsignificant at the highest assay concentra-
tion tested indicating that incorporation of these polar features did
not necessarily increase hERG affinity. Our initial observation, link-
ing polar surface at the R2 position to CYP2D6 inhibition, was con-
firmed. Compounds 4a, 4e and 4f showed a similar selectivity to
2a with 4e and 4f showing minimal inhibition of the enzyme. An in-
crease in hydrophobicity in the analogs 4b and 4c showed a marked
increase in CYP2D6 inhibition. Increasing polar surface for the mor-
pholine analog 4d again reduced enzyme inhibition although H1
affinity was reduced by greater than an order of magnitude com-
pared to 2a. Next, profiles for the enantiomers of key compounds
were compared to 2 in the p-F-benzyl series. Compounds 4h and
4i exhibited greater compounds 4h and 4i exhibited greater H1 affin-
ity H1 affinity [Ki values for the corresponding enantiomers were
CH3
F
F
2
1
R2
N
N
N
N
N
X
R1
CH3
4
3
Fig. 1. Our previous leads (1,2), triprolidine (3) and current general structure (4).
Table 1
H1-antihistamine affinity, M1 affinity, CYP2D6 inhibition, predicted systemic clear-
ance, hERG electrophysiology assay results, and cassette PK data
Compd H1 Kia
M1 CYP2D6b
Pred. Sys. hERG B/P
[B]c
Ratio (ng/g)
4 h
[P]c
(ng/ml)
4 h
(nM)
Kia
m)
IC50
(l
M) Cl (mL/
IC50
(nM)
(l
min/kg)
4 h
1
2
6.9 0.9 >10 5.4
0.9 0.2 2.8 2.3
18.5
10.2
809
721
2.1
2.7
24.9
39.5
11.9
14.7
a
SEM for Ki values derived from dose response curves generated from triplicate
or more data points.
b
IC50 for CYP3A4 inhibition >5 lM for all compounds.
Cassette dose 1 mg/kg, iv.
c
0.7 lM and 3.7 lM, respectively]. The analog 4 h, in comparison to
2, was less selective for CYP2D6 largely as a consequence of reduced
binding affinity for H1. hERG binding data for this compound indi-
cated similar or slightly higher affinity for the channel suggesting
that reductions in calculated pKa and logP [6.9 and 3.8 compared
to 9.1 and 4.5 for 2]11 had little effect on hERG SAR. Compound 4i,
while displaying significantly less affinity than 2, had comparable
selectivity for these targets. Interestingly, compound 4g, the desm-
ethyl analog of 2, retained significant affinity for H1. While some-
what less selective for M1, this compound appeared to show
improved selectivity for hERG in accord with our previous observa-
identified as an optimal side chain. We also chose to examine the
phenoxyethyl substitution for R1 as this feature has appeared in re-
lated compounds with CNS effects.6
The synthesis of analogs for this communication is described in
Scheme 1 and relies on formation of the benzimidazole core, 7,
from coupling of a suitably protected amino acid, 6, with O-pheny-
lene diamine, 5. Introduction of the appropriate arene-containing
moiety R1 was accomplished by alkylation of 7. Boc removal fol-
lowed by alkylation or reductive amination was accomplished as
the last step to give the final products, 4. Further separation of
O
O
N
O
O
N
a
NH2
NH2
O
N
+
X
HO
N
H
X
6
5
7
O
N
R2
O
b
c, d, e
N
N
N
N
N
7
X
X
4
9
X = CH2
X = O
R1
R1
8
10 X= S
Scheme 1. Reagents and conditions: (a) EtOH, 120 °C (40–65%); (b) R1CH2Br, K2CO3, DMF, 80 °C (60–85%); (c) TFA, CH2Cl2, (80–95%), (d) R2-(@O)H, Na(OAc)3BH or R2X, Et3N,
THF (45–90%); (e) separation of enantiomers by chiral SFC when chiral acids were not used.

S. B. Ravula et al. / Bioorg. Med. Chem. Lett. 22 (2012) 421–426
423
Table 2
H1-antihistamine affinity, CYP2D6 inhibition, metabolic stability, M1 affinity and hERG dofetolide binding assay results for compounds 2a and 4a–n
R2
N
N
N
R1
R1
R2
Stereo chemistry
( )
H1 Kia (nM)
CYP2D6 IC50
2.5
(lM)
Pred. Sys. Cld (mL/min/kg)
9.7
M1 Kib
(lM)
hERG Kie
(lM)
c
Compd
2a
p-F-Ph
Me
0.5 0.4
2.9
>10
N
4a
4b
p-F-Ph
p-F-Ph
( )
8.9 1.6
11.3
16.3
N.T.
N.T.
>3
>3
O
N
( )
6.3 0.8
1.9
19.3
O
N
4c
p-F-Ph
p-F-Ph
( )
4.9 1.2
1.4
19.6
N.T.
N.T.
N.T.
>3
>3
O
O
N
4d
( )
16.6 0.6
>20
O
N
4e
4f
p-F-Ph
p-F-Ph
( )
( )
8.1 1.6
1.0 0.0
18.8
4.4
17.8
11.5
N.T.
N.T.
>3
>3
O
OH
4g
2
p-F-Ph
p-F-Ph
H
Me
(+)
(+)
1.2 0.2
0.9 0.2
N.T.
2.3
3.0
10.2
0.48
2.8
>10
3.9
N
4h
4i
p-F-Ph
(+)
5.7 1.3
2.4
14.4
>10
2.6
>3
O
N
p-F-Ph
(+)
13.1 1.7
32.6
17.9
>10
O
4j
4k
CH2OPh
CH2OPh
H
Me
( )
( )
1.0 0.0
0.8 0.3
N.T.
N.T.
N.T.
N.T.
0.43
0.04
>3
5.9
N
4l
CH2OPh
CH2OPh
CH2OPh
( )
( )
(+)
11.9 1.5
24.6 5.7
4.9 0.9
N.T.
18.1
8.7
15.1
N.T.
17.9
>10
N.T.
>10
N.T.
>3
O
N
4m
4n
O
N
>3
O
a
b
c
SEM for Ki values derived from dose response curves generated from triplicate or more data points.
Ki values average of 2 data points.
No compounds showed significant inhibition of CYP3A4 at the highest concentration tested (10
Predicted based on HLM stability studies.
lM).
d
e
Ki Values were derived from single or duplicate data points.

424
S. B. Ravula et al. / Bioorg. Med. Chem. Lett. 22 (2012) 421–426
Table 3
H1-antihistamine activity, CYP2D6 inhibition, CYP3A4 inhibition, hERG electrophysiology assay results, M1 selectivity and metabolic stability data for morpholine/thiomorpholine
compounds 9 and 10
CH3
N
NH
N
N
NH
N
N
N
N
S
O
R1
R1
F
10a-h
2
9a-d
Compd
R1
Stereo chemistry
H1 Kia (nM)
CYP2D6 IC50
(l
M)
CYP3A4 IC50
(
lM)
hERG IC50
(lM)
M1 Kib
(lM)
Pred. Sys. Clc (mL/min/kg)
2
9a
9b
9c
p-F-Ph
p-Me-Ph
p-F-Ph
3,4-diMe-Ph
3,4-diMe-Ph
p-Me-Ph
p-Me-Ph
p-Me-Ph
p-F-Ph
p-F-Ph
p-F-Ph
p-MeO-Ph
p-MeO-Ph
(+)
( )
( )
(+)
(-)
( )
(+)
(-)
( )
(+)
(-)
(+)
(-)
0.9 0.2
1.0 0.1
0.8 0.0
1.9 0.2
365 26.6
1.5 0.3
1.3 0.3
63 7.4
4.0 1.4
1.3 0.2
95 4.4
1.4 0.0
92 74.7
2.5
>20
>20
>20
4.9
7.6
7.5
7.8
9.5
16.7
15.6
10.6
5.8
0.72
2.3
3.4
2.8
10.2
9.4
9.5
28.3
>20
4.9
5.8
3.7
3.6
4.3
15.1
15.0
6.6
N.T.
>10
>10
>10
N.T
0.93
N.T.
N.T.
0.34
0.23
>10
N.T.
2.1
2.6
9d
N.T.
N.T.
0.93
N.T.
N.T.
0.90
N.T.
1.5
N.T.
N.T
10.1
N.T.
N.T.
6.4
N.T.
10.5
N.T.
10a
10b
10c
10d
10e
10f
10g
10h
3.3
7.0
15.8
N.T.
a
b
c
SEM for Ki values derived from dose response curves generated from triplicate or more data points.
Ki values average of 2 data points.
Predicted based on HLM stability studies.
tion that secondary amines were less active than their tertiary coun-
terparts in channel inhibition.3 HLM data for the analogs 4a–c, 4e–f,
4h–i all indicated significantly higher clearance compared to the
methyl analogs 2, 2a and desmethyl compound 4g, the latter being
most stable in the assay. In-house metabolite identification studies
on compound 4i indicated that the des-alkyl compound 4g was
the major metabolite (69%) produced from HLM incubations.10 This
suggested the possibility that N-dealkylation was a key step in bio-
transformation of this class of compounds such that 4g would be an
active metabolite.
Similar profiling results were observed for the phenoxyethyl
series 4j–n with the exception that the methyl analog 4k showed
little selectivity for M1. This selectivity was improved in alkyl
substituents containing more polar functions though reductions
in H1 affinity and poor metabolic stability were dominant factors
in compound profiles. The inhibition of hERG channel in a patch
clamp electrophysiology assay was also examined for 4n. This
compound exhibited an IC50 of 450 nM in this assay, indicating
the compound had similar or if anything slightly higher channel
Profiling data from Table 2 and the metabolite identification
studies suggested some approaches to identifying compounds with
improved profiles. The secondary amine 4g was a high affinity
H1-antihistamine with promising hERG selectivity although some
further optimization was necessary to improve selectivity versus
muscarinic receptors. A second approach suggested itself, involv-
ing the introduction of a heteroatom into the piperidine ring.
Morpholines or thiomorpholines would be expected to exhibit
lower pKa in the six-membered heterocycle and in principle
contribute to an improvement in selectivity profile for hERG.
Several variations of the R1 benzimidazole substitution were
evaluated to examine effects to binding at M1 receptors. These
compounds were prepared using the chemistry described in
Scheme 1 from either the morpholine or thiomorpholine amino
acid (6; X@O or S). Profile data for a series of morpholine and
thiomorpholine analogs are summarized in Table 3.
Inspection of the morpholine series of analogs 9a–d indicated
excellent selectivity profiles for the anti-targets examined namely
CYP2D6, hERG channel as now measured in an electrophysiology
assay [none of the compounds showed any significant inhibiton
blocking activity compared to
2 [previously measured at
721 nM4], an observation similar to that for compound 4h. In con-
clusion, the introduction of polar surface to the R2 substituent of
this core did provide some compounds with sufficient H1 affinity
and reasonable selectivity, particularly in the area of CYP2D6 inhi-
bition. However, selectivity for hERG was reduced compared to 2
[hERG IC50/H1 Ki 90 for 4n compared to 800 for 2]. Overall, poor
metabolic stability of the compound class was a primary issue
for the series and none of the compounds identified showed any
obvious advantages to 2. Predictive pharmacokinetics is also an
important consideration in the selection of lead compounds for
use in insomnia.7–9 The discovery that the des-alkyl compound
4g was a high affinity H1-antihistamine and likely a major metab-
olite of this N-alkylpiperidine class of compounds was a complicat-
ing factor in the use of 2 or its analogs in a sleep indication.
in the hERG binding assay at a test concentration of 3 lM], and
M1 receptor. In the case of the enantiomeric examples 9c and 9d,
a chiral preference for H1 binding was demonstrated for the (+)-
enantiomer in contrast to the original piperidine series.4 For the
corresponding thiomorpholine derivatives, 10a–h, the compounds
were somewhat less selective particularly with respect to M1
receptor. Nevertheless selectivity for both CYP2D6 and hERG
approached 1000-fold. M1 selectivity appeared to be influenced
by the benzyl substituent with the p-methoxybenzyl analog 10g
showing the best profile. In similar fashion to the morpholine
series, the (+)-enantiomers of the thiomorpholines 10b, 10e and
10g had high affinity for H1 receptor. Notably all the compounds
examined in the HLM assay exhibited comparable stability to the
original starting points 2 (and 4g).

S. B. Ravula et al. / Bioorg. Med. Chem. Lett. 22 (2012) 421–426
425
Table 4
CNS exposure results for selected analogs compared to triprolidine (3)a
CH3
N
NH
NH
NH
NH
N
N
N
N
N
N
N
N
N
N
O
S
O
F
F
F
H3C
F
9a
4g
9b
2
10e
Compd
[B] (ng/g) 4 h
[P] (ng/ml) 4 h
B/P ratio 4 h
[B] (ng/g) 4 h
46.5
139
69.9
139
139
2
39.5
41
83.1
185
159
14.7
16.6
42
32
20.5
2.7
2.5
2.3
5.9
7.9
4g
9a
9b
10e
a
Cassette dose 1 mg/kg, iv.
As with our previous studies,3,4 attempts to modulate hERG
been a key determinant for the viability of chemical series as
potential sleep agents.3,4 To assess suitable CNS penetration in po-
tential lead compounds, representative analogs were evaluated for
their ability to penetrate the blood brain barrier in rodents using
cassette PK studies.3 Groups of five compounds including the short
acting brain penetrating antihistamine triprolidine6 were adminis-
tered (IV) to rats and brain levels and B/P ratios were determined.
Analogs with CNS penetration similar to triprolidine were consid-
ered likely to be brain penetrating. Cassette data for representa-
tives of the compounds synthesized is summarized in Table 4.
Overall, representatives of morpholine (9a, 9b) and thiomorph-
o-
line (10e) classes achieved excellent brain levels in comparison to
the sedating antihistamine triprolidine and the original piperidines
(2, 4g). These compounds appeared to demonstrate significantly
higher CNS exposure than 4g. This factor may be a consequence
of the reduction in pKa for the heterocyclic analogs compared to
the piperidine.12 CNS exposure on oral dosing was confirmed for
the more selective compound 10g. At an oral dose of 30 mg/kg in
Sprague Dawley rats, 10g exhibited a measured plasma exposure
of 260 ng/mL and brain levels of 380 ng/g at 4h. These exposures
were comparable to those seen in previous studies3 (e.g., 11 and
1, Table 4) and should enable the characterization of sleep proper-
selectivity through reductions in pKa and/or logP provided
relatively small or no improvements. The principle source of the
selectivity for this series appears to arise from the orientation of
the basic nitrogen in the heterocycle attached to the benzimidazole.4
Considering this point and our previous observations for selected N-
alkylpiperidines, where similar reductions of pKa and logP provided
minimal effects on hERG activity, the possibility cannot be excluded
that the incremental improvements for hERG selectivity are simply a
result of subtle structural changes to the heterocyclic core. In any
event, the combination of modifications including the morpholine
or thiomorpholine ring and the appropriate R1 substitution to this
benzimidazole class provided compounds with more optimal
in vitro selectivity profiles [selectivity for hERG >1100 for 9c and
10g compared to ꢀ800 for 2 as estimated from hERG IC50/H1 Ki].
From the data presented here, best in vitro profiles were
obtained from modifications to the piperidine heterocycle of 2 as
opposed to nitrogen substitutions of the piperidine ring, the key
factor differentiating the two approaches being metabolic stability.
Having established a positive selectivity profile for the classes
represented by compounds 9a–d and 10a–h, we next determined
whether any of these compounds were capable of demonstrating
CNS exposure in an in vivo setting. CNS exposure has previously
Table 5
Profile of 10g compared to 11 and 1
CH3
N
NH
N
N
N
N
N
N
N
N
N
N
O
CH3
S
F
F
O
11
1
10g
Compd
H1 Kia (nM)
Selectivity over M1,M3, 5HT2a, H3
hERG IC50 (nM)
Pred. Sys. Cl (ml/min/kg)
B/P ratio 4 h
[B] 4 h (ng/g)
11
1
10g
3.9 0.8
6.9 0.9
1.4 0.3
>1000
>1000
>1000
29
809
1500
16.8
18.5
10.5
10a
94a
2.3b
1.4b
262b
380b
a
Dose 10 mg/kg, p.o.
Dose 30 mg/kg, p.o.
b

426
S. B. Ravula et al. / Bioorg. Med. Chem. Lett. 22 (2012) 421–426
ties for this and other representatives from the series in EEG/EMG
models.
References and Notes
1. (a) Ohayon, M. M.; Lemoine, P. L’Encephale 2004, 30, 135; (b) Zammit, G. K.;
Weiner, J.; Damato, N.; Sillup, G. P.; McMillan, C. A. Sleep 1999, 22, S379.
2. (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.
3. Coon, T.; Moree, W. J.; Li, B.-F.; 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.
4. Lavarador-Erb, K.; Ravula, S. B.; Yu, J.; Zamani-Kord, S.; Moree, W. J.; Petroski, R.
E.; Wen, J.; Malany, S.; Hoare, S. R. J.; Madan, A.; Crowe, P. D.; Beaton, G. Bioorg.
Med. Chem. Lett. 2010, 20, 2916–2919. and references cited there in.
5. For a review see Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. Methods
and Principles in Medicinal Chemistry (2008), 38 (Antitargets: Prediction and
Prevention of Drug Side Effects), pp. 423–455.
6. Iemura, R.; Ohtaka, H. Chem. Pharm. Bull. 1989, 37(4), 967.
7. Moree, W. J.; Li, B.-F.; Jovic, F.; Coon, T.; Yu, Y.; Tucci, F.; Marinkovic, D.; Gross,
R. S.; 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.
Table 5 demonstrates the progression of leads in this study
compared to our previous work.3,4 Our first lead compound 11,
exhibited excellent CNS exposure and efficacy in an EEG/EMG
model of sleep.3 However, this analog was a potent hERG channel
blocker precluding further development. Optimization of the 4-
aminopiperidine for the hERG liability was possible by removal
of the basic amine and replacement with the pyrazole resulting
in compound 1.3 However, the properties of this molecule were
sub-optimal including poor metabolic stability and solubility.
Changing the orientation of the basic amine was a key factor in
improving H1 affinity, selectivity for hERG channel and metabolic
stability.4 Strategies to reduce pKa and/or logP in this series pro-
vided minimal improvements to hERG selectivity. In the case of
secondary amines identified during this study as exemplified by
10g, reductions in these parameters resulted in an improvement
in CNS exposure. (Table 5).
8. Moree, W. J.; Jovic, F.; Coon, T.; Yu, J.; Li, B.-F.; Tucci, F.; Malany, S.; Bradbury, M.
J.; Hernandez, L. M.; O’Brien, Z.; Wen, J.; Wang, H.; Hoare, S. R.; Petroski, R. E.;
Sacaan, A.; Madan, A.; Crowe, P. D.; Beaton, G. Bioorg. Med. Chem. Lett. 2010, 20,
2316.
9. Madan, A.; O’Brien, Z.; Wen, J.; O’Brien, C.; Farber, R. H.; Beaton, G.; Crowe, P.
D.; Oosterhuis, B.; Garner, R. C.; Lappin, G.; Bozigian, H. P. Br. J. Clin. Pharmacol.
2009, 67, 288.
In summary, our SAR studies around 2 resulted in the identifi-
cation of several compounds with improved selectivity for H1
receptor over a selection of anti-targets including CYP2D6, hERG
channel and the M1 receptor. Best overall profiles were obtained
for a selection of benzimidazole morpholines and thiomorpholines
which demonstrated reasonable predicted metabolic stability in
addition to in vitro selectivity. A reduction in pKa for these
compounds was presumed to be responsible for the increased
CNS exposure relative to the initial starting point. The excellent
CNS exposure demonstrated for morpholine analogs 9a and 9b in
cassette PK studies and thiomorpholine 10g from oral dosing affor-
ded a series of candidates for further in vivo evaluation in EEG/
EMG models.
10. For metabolic ID study, test compounds at 50 lM concentration were
incubated in Human Liver Microsomes at 0.5 mg/mL protein for 120 min
with shaking at 37 °C then checked by LCMS and methods were published as
Supplementary data in: Li, B.-F.; Moree, W. J.; Yu, J.; Coon, T.; Zamani-Kord, S.;
Malany, S.; Jalali, Kayvon.; Wen, J.; Wang, H.; Yang, C.; Hoare, S. J.; Petroski, R.
E.; Madan, A.; Crowe, P. D.; Beaton, G. Bioorg. Med. Chem. Lett. 2010, 20, 2629.
11. Calculated using ACD Labs Software Suite.
12. pKa and logP values were measured using GLPKa instrumentation from
pION Inc. Values were measured for 9c: pKa 7.4; logP 3.4 (calculated pKa
8.2; logP 4.8) and a representative thiomorpholine [R1 3,4-dimethylphenyl]
with S stereochemistry: 1-(3,4-dimethyl-benzyl)-2-morpholin-2-yl-1H-benzoi
midazole]: pKa 7.5; logP 3.3 (calculated pKa 8.3; logP 4.9).
Acknowledgments
The authors wish to thank John Harman and Chris DeVore for
analytical support and Dr. John Saunders, Dr. Paul Conlon and Dr.
Haig Bozigian for program support.