A new class of histamine H3 receptor antagonists derived from ligand based design

Article abstract of DOI:10.1016/j.bmcl.2007.04.056

Design and synthesis of highly potent and selective non-imidazole inverse agonists for the histamine H₃ receptor is described. The study validates a new pharmacophore model based on the merging of two previously described models. It also demonstrates that the removal of the basic center potentially interacting with ASP3.32 and common to both models leads to loss of activity, whereas the replacement of the second basic center by an acceptor retains the potency.

Full text of DOI:10.1016/j.bmcl.2007.04.056

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3670–3675
A new class of histamine H3 receptor antagonists derived
from ligand based design
*
´
´
Olivier Roche and Rosa Marıa Rodrıguez Sarmiento
F. Hoffmann-La Roche Ltd, Pharmaceutical Research Basel, Discovery Chemistry, CH-4070 Basel, Switzerland
Received 12 March 2007; revised 13 April 2007; accepted 15 April 2007
Available online 25 April 2007
Abstract—Design and synthesis of highly potent and selective non-imidazole inverse agonists for the histamine H3 receptor is
described. The study validates a new pharmacophore model based on the merging of two previously described models. It also dem-
onstrates that the removal of the basic center potentially interacting with ASP3.32 and common to both models leads to loss of
activity, whereas the replacement of the second basic center by an acceptor retains the potency.
Ó 2007 Elsevier Ltd. All rights reserved.
Histamine is a central modulator of physiological pro-
cesses acting both in the central nervous system (CNS)
and the periphery through four distinct receptors (H1–
H4) known to date.1,2 Histamine H1 and H2 receptor
antagonists are already used in the clinic to treat allergic
asthma and excess gastric acid production, respectively.
The newly discovered H4 receptor seems to have a role
in regulating inflammatory responses.3 The H3 receptor,
mainly located in the CNS, is a presynaptic autoreceptor
that does not only modulate the production and the
release of histamine from histaminergic neurons4 but
also regulates the release of other neurotransmitters in
both the central and peripheral nervous systems.5–7 An-
other feature of the H3 receptor is its high constitutive
activity.8,9 H3 antagonists/inverse agonists have been
primarily studied in cognitive models for Alzheimer’s
disease, attention-deficit disorder, and schizophre-
nia.10,11 Pharmacological data also suggest a potential
role for H3 antagonists and/or inverse agonists in the
control of feeding, appetite, and body weight, and
support the role of H3 receptor in obesity.12,13
drug metabolizing enzymes, the imidazole-based H3
antagonists were discarded.14 As it has been shown that
basic amine-based molecules have different binding
modes than imidazole-based ones, a large part of the
modeling information cannot be used.15–19 The focus
has then been on the basic amine-containing compounds
that are postulated to interact with the highly conserved
ASP3.32.16 During the manual alignment of the mole-
cules, we can observe two distinct pharmacophore distri-
butions (Fig. 1). Each model consists of four
pharmacophoric features. Three of the features are
shared by both models, namely a distal positive charge,
an electron rich position, and a central aromatic ring.
The fourth feature can be either a second basic amine
exemplified with a potent Johnson & Johnson com-
pound (model 1)20 or another aromatic contained in
many Abbott molecules coming from A. Hancock group
like A-349821 (model 2),21 but also later ABT-239.22
Both types of molecules have been extensively described
separately and, at the time of the study, no molecules
have been disclosed combining all the features of the
combined pharmacophore. Nevertheless, since then,
research teams at Johnson & Johnson have published
various chemotypes supporting our pharmacophore
hypothesis.23–25 Our study describes the rational
approach followed at Roche to generate new chemo-
types validating our assumption. Considering the large
overlapping of the pharmacophores, we assume a com-
mon binding mode, which will be driven by the interac-
tion between the common basic amine and the
conserved aspartate in helix 3 (ASP3.32). Thus, we
An analysis of known low nanomolar H3 antagonists
published in patents and literature has been made in
order to derive a three-dimensional pharmacophore
model. Since imidazole-containing compounds could
have potential liability toward cytochrome-P450 (CYP)
Keywords: Histamine H3 receptor; SAR; Pharmacophore model; De
novo design; Skelgen.
*
Corresponding author. Tel.: +41 61 6883585; fax: +41 61 6886459;
e-mail: rosa_maria.rodriguez_sarmiento@roche.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.056

´
O. Roche, R. M. Rodrıguez Sarmiento / Bioorg. Med. Chem. Lett. 17 (2007) 3670–3675
3671
Model 2
Model 1
O
N
O
N
N
O
A-349821
hH3 pKi = 9.31
rH3 pKi = 8.78
JNJ-5207852
hH3 pKi = 9.24
hH3 pA2 = 9.84
N
O
Figure 1. Two pharmacophore models derived from known antagonists of the histamine H3 receptors. The pharmacophoric features are displayed as
dashed circles: blue, positive charge (basic amine); red, electron rich; cyan, aromatic/lipophilic.
designed a new pharmacophore model consisting of five
features: two aromatics, two positively charged moieties,
and one electron rich feature. We included the electron
rich feature because our entire set of reference molecules
contains an ether linker at the same position, but the
available SAR does not allow us to be sure it is fully
mandatory. To provide starting points for the chemistry
effort, we use the de novo program Skelgen.26–28 The
program automatically builds new molecules from frag-
ment libraries that match the three-dimensional phar-
macophoric constraints. It provides hundred of
possible molecules that have been manually clustered.
This leads to four templates, which share the same phar-
macophores but have different connectivity patterns
(Fig. 2). The chemical synthesis has been developed
around the connective pattern of template two in order
to validate the new pharmacophore.
Synthesis of 4-amino quinazoline of type 1 was better
accomplished as indicated in Scheme 1, starting with
the O-alkylation of commercially available nitro phenol
(3) with benzyl chloride. The reductive amination of the
Template 1
Template 4
Template 2
Template 3
Figure 2. Projection scheme of the various connectivity patterns between the pharmacophore features suggested by Skelgen. Blue, positive charge
(basic amine); red, electron rich; cyan, aromatic/lipophilic.
O
O
O
O
O
O
O
Bn
N
O
H
Bn
Bn
H
O
N
O
H
O
(a)
(b, c)
(d)
N+
O
N+
O
N+
O
N+
O
3
4
5
6
(e)
O
R
N
O
O
(g)
(h, i)
Bn
N
Bn
R= OCH2Ar
R= Br(CH2)3COOEt
R=MeI
R=Cl(CH2)3NR1R2
N
H2N
N
N
N
Cl
N
N
N
8
7
1
9
N
N
OH
R1
Cl
O
(f)
N
O
N
N
N
R2
N
(k)
(h,j)
N
N
N
N
1
N
N
Scheme 1. Reagents and conditions: (a) BnCl (excess), K2CO3, DMF, 110 °C, 4 h, 100%; (b) KMnO4, acetone/water (1:1), 5 h, 98%; (c) HBTU,
dimethylamine hydrochloride, TEA, DMF, 3 h, 89%; (d) BH3ÆTHF, THF, 70 °C, 8 h, 90% (complex); (e) Pt–C 10%, H2, EtOAc, 18 h, 80%; (f)
POCl3, 100 °C, microwave, 15 min; (g) 8, EtOH, DIPEA, reflux, 12 h, 85%; (h) Pd (10% C), EtOH, H2, rt, overnight, 83%; (i) RX, K2CO3, DMF,
18 h, 60 °C, 40–10% or two-step procedure with (j) Br(CH2)3Cl, Cs2CO3, acetone, rt, 5 h, 93%; (k) Corresponding amine, EtOH, reflux, 5 h, 65–90%.

´
O. Roche, R. M. Rodrıguez Sarmiento / Bioorg. Med. Chem. Lett. 17 (2007) 3670–3675
3672
Table 1. Binding data on human histamine receptors H1, H2, H3 for the compounds of series 1
X
O
R
N
Y
N
N
Compound
X
R
H
Y
hH3 Ki (nM)
hH3% inhibitiona
hH1% inhibitiona
hH2% inhibitiona
ABT-23922
JNJ-520785220
1a
2.4
1.4
H
H
H
H
H
H
O
N
N
N
N
N
N
O
11.6%
31%
O
O
1b
1c
O
28%
O
Cl
1d
1e
49%
18
À15%
À2%
N
1f
0.3
—
À2.5%
2.4%
N
18a
1021
N
a Percentage of inhibition measured at 10 lM.
Cl
O
O
O
O
O
N
O
O
O
O
OH
(a)
(b)
O
O
O
O
O
N
O
O
N
HO
HO
or
Br
R1
11
12
13
14
or
N
R1
(c,d)
O
O
O
O
N
O
N
O
R
O
O
O
N
R
O
O
N
N
O
N
(e,f)
O
O
+
or
N
N
R1
15
R1
17
16
18
(h, i)
O
O
R
H
N
N
R= OCH2Ar
R= Br(CH2)3COOEt
R=MeI
R=Cl(CH2)3NR1R2
O
O
(j)
(g)
2
19
R1
R1
Scheme 2. Reagents and conditions: (a) DHP, PPTS, CH2Cl2, 12 h, 97%; (b) 4-chloro-quinazoline 8, Cs2CO3, DMF, 90 °C, 4 h, 90% or BrCH2Ar,
Cs2CO3, 40 °C, acetone, 4 h, 100%; (c) LiOH, THF, H2O, 40 °C, 18 h, 100%; (d) CDI, HN(CH3)2ÆHCl, DMF, 60 °C, or HBTU, Et3N,
HN(CH3)2ÆHCl, DMF, 2 h, 90–70%; (e) pTsOH, MeOH, 60 °C, 8 h, 84%; (f) K2CO3, RX, DMF, 100 °C, 2 h, 90–70%; (g) BH3ÆTHF THF, 90 °C, 4 h,
63%; (h) LiAlH4, THF, 40 °C, 18 h, 70%; (i) pTsOH, MeOH, 60 °C, 8 h, 69%; (j) RX, K2CO3, DMF, 100 °C, 3 h, 20–70%.

´
O. Roche, R. M. Rodrıguez Sarmiento / Bioorg. Med. Chem. Lett. 17 (2007) 3670–3675
3673
a-nitroaldehyde (4) gave low yields and an alternative
route via the reduction of the corresponding amide
was followed. The aldehyde (4) was oxidized to the acid
to obtain via HBTU coupling the N,N-dimethyl amide
(5). Reduction of the amide was performed with bor-
ane–THF complex to obtain the amine (6) as a borane
complex. After reduction of the nitro group in the pres-
ence of the benzyl group with platinum as catalyst, the
compound 7 was obtained free of borane. Alkylation
of the nitrogen with 4-chloro-quinazoline (8) (obtained
from the treatment of commercially available quinazo-
lin-4-ol with POCl3) was performed in ethanol. The
use of DIPEA allows the reaction to proceed rapidly
and more cleanly toward 9. Removal of the benzyl
group gives the phenol that can react with a range of
alkylating agents in the presence of K2CO3 to obtain
compounds of type 1 described in Table 1. In cases
where direct alkylation did not provide high yields, a
two-step procedure can be used instead with a first alkyl-
ation with 1-bromo-3-chloropropane using Cs2CO3 to
obtain the O-alkylchloride that can be heated with the
corresponding amine to prepare 1. Alternative routes
with the use of 4-fluorophenol or 2,5-dihydroxybenzal-
dehyde gave undesired byproducts and lower yields.
In the case of compounds of type 2 (4-oxo quinazolines
or benzyloxy compounds), the most successful synthetic
route starts with the regioselective alkylation of the
commercially available 2,5-dihydroxy-benzoic acid (11)
with THP under acidic conditions to obtain 12.29 Alkyl-
ation with 4-chloro-quinazoline (8) or with the corre-
sponding alkylbromide with Cs2CO3 gives 13 or 14,
respectively. Hydrolysis of the ester using LiOH and
amide formation with HNMe2ÆHCl using CDI gives
Table 2. Binding data on human histamine receptors H1, H2, H3 for the compounds of series 2
X
O
N
R
O
R1
Compound
X
H
R
R1
hH3 Ki (nM)
hH3% inhibitiona
hH1% inhibitiona
hH2% inhibitiona
2a
Cl
Cl
150
N
N
2b
H
88
2c
2d
H
H
Cl
Cl
135
2.4
N
N
—
—
17a
17b
17c
17d
17e
17f
17g
O
O
O
O
O
O
O
Cl
33%
17%
F
Cl
47
8%
N
O
Cl
O
O
Cl
1100
1810
1393
12
N
N
Cl
OMe
OMe
73.4%
N
N
a Percentage of inhibition measured at 10 lM.

´
O. Roche, R. M. Rodrıguez Sarmiento / Bioorg. Med. Chem. Lett. 17 (2007) 3670–3675
3674
the amides 15 or 16. In both cases removal of the THP
and O-alkylation with an appropriate alkylating agent
gave compounds 17 or 18. Reduction of the amide
group on 17 or 18 to the corresponding amines using
borane–tetrahydrofuran complex gives reasonable yields
except on compounds for type 17 when a benzyl or
methyl group is present (R = OCH2Ar and R = Me).
An alternative way to obtain the desired amine 2 was
through the reduction of compound 15, using LiAlH4.
Removal of the THP group under acidic conditions
gives the phenol 19 that can be alkylated with K2CO3
and the corresponding halogenated group. The com-
pounds synthesized by this route (Scheme 2) are exem-
plified in Table 2.
work possible and so enjoyable, especially Dr. Silvia.
Gatti McArthur, Sandra Grall-Ulsemer, Regine Denu,
Dr. Harald Mauser, Dr. Wolfgang Guba, and Dr. M.
Nettekoven.
References and notes
1. Sneader, W. Drug News Perspect. 2001, 14, 618.
2. Hough, L. B. Mol. Pharmacol. 2001, 59, 415.
3. Thurmond, R. L.; Desai, P. J.; Dunford, P. J.; Fung-
Leung, W.-P.; Hofstra, C. L.; Jiang, W.; Nguyen, S.;
Riley, J. P.; Sun, S.; Williams, K. N.; Edwards, J. P.;
Karlsson, L. J. Pharmacol. Exp. Ther. 2004, 309, 404.
4. Arrang, J. M.; Garbarg, M.; Schwartz, J. C. Neuroscience
1987, 23, 149.
The H3 activities presented in Tables 1 and 2 are binding
data measured by displacement of [3H](R)-a-methylhis-
tamine (RAMH).30 For selected compounds, human
H1 and H2 subtype affinity has been measured using
[3H]pyrilamine and [3H]tiotidine, respectively (see
Tables 1 and 2).31 The SAR explored by the molecules
shown in these tables validate the new pharmacophore
model since compounds 1e, 1f, and 2a–2d carrying 2
basic centers and a lipophilic substituent altogether are
highly active at the H3 receptor. When we remove the
distal basic amine in compounds 1a–1d, potentially
interacting with ASP3.32, the activity at the H3 receptor
is completely lost. For the molecules 18a, 17b, 17e, and
17g where the proximal basic center has been eliminated
by generating the amide, part of activity can be con-
served. Especially, compounds 17b and 17g have a Ki
of 42 and 12 nM at the H3 subtype, respectively. From
our results, it is likely that both the proximal basic nitro-
gen and the carbonyl oxygen of the amide bond interact
with the same residue. A potential candidate residue for
this type of interaction could be Glu5.46, which can be
both a hydrogen bond donor (carbonyl, 17g) and accep-
tor (basic amine, 2d).
5. Schlicker, E.; Kathmann, M. In The Histamine H3
Receptor; a Target for New Drugs; Leurs, R., Timmerman,
H., Eds., 1st ed.; Elservier Science B.V.: Amsterdam, 1998;
pp 13–26.
6. Stark, H. Expert Opin. Ther. Pat. 2003, 13, 851.
7. Leurs, R.; Blandina, P.; Tedford, C.; Timmerman, H.
Trends Pharmacol. Sci. 1998, 19, 177.
8. Rouleau, A.; Ligneau, X.; Tardivel-Lacombe, J.; Morisset,
S.; Gbahou, F.; Schwartz, J. C.; Arrang, J. M. Brit.
J. Pharmacol. 2002, 135, 383.
9. Takahashi, K.; Tokita, S.; Kotani, . J. Pharmacol. Exp.
Ther. 2003, 307, 213.
10. Witkin, J. M.; Nelson, D. L. Pharmacol. Therapeut. 2004,
103, 1.
11. Alguacil, L. F.; Perez-Garcia, C. Curr. Drug Targets CNS
Neurol. Disord. 2003, 2, 303.
12. Hancock, A. A.; Brune, M. E. Expert Opin. Invest. Drugs
2005, 14, 223.
13. Hancock, A. A. Curr. Opin. Invest. Drugs 2003, 4, 1190.
14. Yang, R.; Hey, J. A.; Aslanian, R.; Rizzo, C. A.
Pharmacology 2002, 66, 128.
15. Rivara, S.; Mor, M.; Bordi, F.; Silva, C.; Zuliani, V.;
Vacondio, F.; Morini, G.; Plazzi, P. V.; Carrupt, P.-A.;
Testa, B. Drug Des. Discov. 2003, 18, 65.
16. Yao, B. B.; Hutchins, C. W.; Carr, T. L.; Cassar, S.;
Masters, J. N.; Bennani, Y. L.; Esbenshade, T. A.;
Hancock, A. A. Neuropharmacology 2003, 44, 773.
17. Stark, H.; Sippl, W.; Ligneau, X.; Arrang, J. M.; Ganellin,
C. R.; Schwartz, J. C.; Schunack, W. Bioorg. Med. Chem.
Lett. 2001, 11, 951.
18. De Esch, I. J.; Mills, J. E.; Perkins, T. D.; Romeo, G.;
Hoffmann, M.; Wieland, K.; Leurs, R.; Menge, W. M.;
Nederkoorn, P. H.; Dean, P. M.; Timmerman, H. J. Med.
Chem. 2001, 44, 1666.
19. De Esch, I. J. P.; Timmerman, H.; Menge, W. M. P. B.;
Nederkoorn, P. H. J. Arch. Pharm. (Weinheim, Germany)
2000, 333, 254.
20. Apodaca, R.; Dvorak, C. A.; Xiao, W.; Barbier, A. J.;
Boggs, J. D.; Wilson, S. J.; Lovenberg, T. W.; Carruthers,
N. I. J. Med. Chem. 2003, 46, 3938.
21. Faghih, R.; Dwight, W.; Bao Pan, J.; Fox, G. B.; Krueger,
K. M.; Esbenshade, T. A.; McVey, J. M.; Marsh, K.;
Bennani, Y. L.; Hancock, A. A. Bioorg. Med. Chem. Lett.
2003, 13, 1325.
22. Cowart, M.; Pratt, J. K.; Stewart, A. O.; Bennani, Y. L.;
Esbenshade, T. A.; Hancock, A. A. Bioorg. Med. Chem.
Lett. 2004, 14, 689.
From the potent compounds 1e, 1f, 17b tested against
human H1 and H2 receptors we confirmed that our series
are selective for the H3 subtype. The most promising
compound 1f has been further profiled in terms of func-
tionality and species selectivity. It has been tested in a
GTPcS functional assay and was found to be a potent
inverse agonist with an EC50 of 0.2 nM. Moreover, 1f
also displays a high affinity toward the rat H3 receptor
with a Ki of 9.8 nM qualifying the molecule for further
in vivo experiments. Preliminary safety data show that
1f does not block the hERG channel.32
In conclusion, applying a new pharmacophore model,
we have discovered highly potent and selective inverse
agonists for both human and rat H3 receptors. In this
study, we have exemplified only one of the connection
patterns proposed by Skelgen. Based on these promising
results, other chemotypes will be explored.
23. Keith, J. M.; Gomez, L. A.; Letavic, M. A.; Ly, K. S.;
Jablonowski, J. A.; Seierstad, M.; Barbier, A. J.; Wilson,
S. J.; Boggs, J. D.; Fraser, I. C.; Mazur, C.; Lovenberg, T.
W.; Carruthers, N. I. Bioorg. Med. Chem. Lett. 2007, 17,
702.
Acknowledgments
It is with real pleasure that we wish to thank all our col-
laborators whose contributions made the described

´
O. Roche, R. M. Rodrıguez Sarmiento / Bioorg. Med. Chem. Lett. 17 (2007) 3670–3675
3675
24. Letavic, M. A.; Keith, J. M.; Jablonowski, J. A.; Stocking,
E. M.; Gomez, L. A.; Ly, K. S.; Miller, J. M.; Barbier, A.
J.; Bonaventure, P.; Boggs, J. D.; Wilson, S. J.; Miller, K.
L.; Lord, B.; McAllister, H. M.; Tognarelli, D. J.; Wu, J.;
Abad, M. C.; Schubert, C.; Lovenberg, T. W.; Carruthers,
N. I. Bioorg. Med. Chem. Lett. 2007, 17, 1047.
25. Letavic, M. A.; Keith, J. M.; Ly, K.; Barbier, A. J.; Boggs,
J. D.; Wilson, S. J.; Lord, B.; Lovenberg, T. W.;
Carruthers, N. I. Bioorg. Med. Chem. Lett. 2007, 17, 2566.
26. Dean, P. M.; Firth-Clark, S.; Harris, W.; Kirton, S. B.;
Todorov, N. P. Expert Opin. Drug Disc. 2006, 1, 179.
27. Lloyd, D. G.; Buenemann, C. L.; Todorov, N. P.;
Manallack, D. T.; Dean, P. M. J. Med. Chem. 2004, 47,
493.
30. Saturation binding experiments were performed using
HR3-CHO membranes prepared as described in: (a)
Takahashi, K.; Tokita, S.; Kotani, H. J. Pharmacol.
Exp. Ther. 2003, 307, 213; All compounds were tested at a
single concentration in duplicate. Compounds that
showed an inhibition of [3H]RAMH by more than 50%
were tested again to determine IC50 in a serial dilution
experiment. Ki’s were calculated from IC50 based on
Cheng–Prusoff equation: (b) Cheng, Y.; Prusoff, W. H.
Biochem. Pharmacol. 1973, 22, 3099; (c) For a more
detailed description see: Gatti, S.; Hertel, C.; Nettekoven,
M.; Plancher, J.-M.; Raab, S.; Roche, O.; Rodriguez-
Sarmiento, R.-M. PCT Int. Appl. 2005, WO2005117865.
31. Membranes and protocols originated from Euroscreen.
For more information, see: characterization of histamin-
ergic receptors Current Protocols in Pharmacology, Ed. J.
Wiley & Sons, 2006; Vol. 1, Chapter 19, pp 1–19.
28. Stahl, M.; Todorov Nikolay, P.; James, T.; Mauser, H.;
Boehm, H.-J.; Dean Philip, M. J. Comput. Aided Mol.
Des. 2002, 16, 459.
29. Tori, M.; Hamaguchi, T.; Sagawa, K.; Sono, M.; Asak-
awa, Y. J. Org. Chem. 1996, 61, 5362.
32. Compound 1f showed 30% inhibition at 10 lM concen-
tration at the hERG channel (IC50 > 10 lM ).