Histamine-3 receptor inverse agonists for the treatment of obesity: Validation of the target and identification of novel series

Article abstract of DOI:10.2533/chimia.2009.275

Obesity is a major risk factor for the development of conditions such as hypertension, hyperglycemia, dyslipidemia, coronary artery disease and cancer. Several pieces of evidence, including data in primates, have demonstrated the beneficial effects of his

Full text of DOI:10.2533/chimia.2009.275

La u r e a t e s : sCs Pr iz e s a n d sCs Me t t L e r -to L e d o aw a r d s 2008
CHIMIA 2009, 63, No. 5 275
doi:10.2533/chimia.2009.275
Chimia 63 (2009) 275–278 © Schweizerische Chemische Gesellschaft
Histamine-3 Receptor Inverse Agonists for
the Treatment of Obesity: Validation of the
Target and Identification of Novel Series
Pascale David Pierson, Christian Freichel, Silvia Gatti-Mac Arthur, Cornelia Hertel,
Jörg Huwyler, Peter Mohr, Toshito Nakagawa, Matthias Nettekoven, Jean-Marc Plancher§*,
Susanne Raab, Hans Richter, Olivier Roche, Rosa María Rodríguez Sarmiento, Monique Schmitt,
Franz Schuler, Tadakatsu Takahashi, Sven Taylor, Christoph Ullmer, and Ruby Wiegand
§SCS-Mettler Toledo Award Winner (Oral Presentation)
Abstract: Obesity is a major risk factor for the development of conditions such as hypertension, hyperglycemia,
dyslipidemia, coronary artery disease and cancer. Several pieces of evidence, including data in primates, have
demonstrated the beneficial effects of histamine-3 receptor (H3R) inverse agonists in the regulation of food intake
and body weight. A pharmacophore model based on selected published H3R ligands and validated by previ-
ous investigations, was used to identify the 5-oxy-2-carboxamide-indole core as a novel series of H3R inverse
agonists. Extensive structure–activity relationship (SAR) investigations were rewarded by the identification of
several compounds reversing (R)-α-methyl-histamine-induced water intake increase and reducing food intake/
body weight in rodent models of obesity. Among those compounds, (4,4-difluoro-piperidin-1-yl)-[1-isopropyl-5-
(1-isopropyl-piperidin-4-yloxy)-1H-indol-2-yl]-methanone, selected as a lead compound, was exhibiting a prom-
ising profile, including excellent pharmacokinetic properties, good in vitro safety profile and high efficacy in a
chronic rodent model of obesity.
Keywords: hERG · Histamine-3 receptor · Inverse agonist · Obesity · Pharmacophore model
1. Introduction
sleep disorders and metabolic disorders.[4] characterized by a basic sp nitrogen (yel-
Among the latter, we decided to concen- low circle in Fig. 1), acting3as the main an-
The histamine receptor family, compris- trate on obesity, a dramatically increasing chor, most probably through a salt bridge
ing the four subtypes 1–4,[1] has already burden with tremendous needs for novel formation with Asp3.32. At a distance of
received considerable attention, and led and efficient therapies.
4–5 Å, an electron-rich entity, typically an
to several blockbusters in the inflamma- H3R is highly constitutively active; ether or a carbonyl oxygen (red circle in
tion and gastric secretion fields.[2] Within therefore, most of the antagonists should Fig. 1) is often observed across different
this family, the histamine-3 receptor (H R) be properly classified as inverse agonists. chemical series. Optionally, a second ba-
acts both as an autoreceptor, controlling3the In term of mechanism, H R inverse ago- sic nitrogen atom can be present at around
synthesis and release of histamine, and also nists/antagonists suppress 3food intake via 10–11 Å from the first one, most probably
as a heteroreceptor, controlling other cru- increased histamine release and conse- interacting with Glu5.46. However, some
cial neurotransmitters (e.g. acetylcholine, quently, increased activation of postsynap- concerns about risks for brain accumula-
norephedrine, serotonin, dopamine).[3]
As a consequence of this rich pharma- activation is restricted to the CNS, where avoid this second basic nitrogen. In con-
cology, three distinct therapeutic areas can the H R is mainly expressed. trast, we postulated the importance of a
potentially be envisaged: central nervous
In3deed, in recent years, some potent second electron rich entity, probably in-
tic histamine-1 receptor (H1R). This H1R tion and phospholipidosis prompted us to
system (CNS) disorders (e.g. Alzheimer, and selective H3R antagonists/inverse ago- teracting with Thr6.52. Interestingly, this
schizophrenia, anxiety, cognitive deficit), nists have demonstrated efficacy in several threonine is only present in the H R.
models of obesity, including in primates.
Compared to other scaffolds3 initially
However, some of these compounds also investigated,[6–8] the indole series appeared
showed clear limitations mainly in terms to have the widest potential, including the
of safety (e.g. hERG affinity/cardiovascu- option of easy access to a third exit vec-
lar liabilities, genotoxicity, CNS side ef- tor on the indole nitrogen – so far poorly
fects).[5]
explored – and an excellent chemical trac-
tability.
*Correspondence: Dr. J.-M. Plancher
F. Hoffmann-La Roche Ltd
Pharmaceuticals Division
B092/4.18C
CH-4070 Basel
Tel.: + 41 61 68 86 725
Fax: +41 61 68 88 367
2. Pharmacophore Model
3. Early SAR and Chemistry
The availability of many published
structures of potent H R ligands has fa-
Capitalizing on our previous investiga-
cilitated the constructi3on of an accurate tions,[7,8] the identification of H3R inverse
pharmacophore model.[6,7] This model is agonists proved to be straightforward by
E-mail: jean-marc.plancher@roche.com

276 CHIMIA 2009, 63, No. 5
La u r e a t e s : sCs Pr iz e s a n d sCs Me t t L e r -to L e d o aw a r d s 2008
F
F
Glu5.46
(only in H3R, H4R)
O
O
O
N
N
O
4-5 Å
10-11 Å
N
N
N
H
O
N
H
Compound 1 (logD: 0.09)
In vitro profile
Compound 2 (log D: 1.2)
In vitro profile
hH3R Ki: 5 nM
hH3R Ki: 2 nM
Micros. Cl h / r: <1 / 12 µL/min/mg
In vivo pharmacokinetic (rat, iv.)
Cl: 50 mL/min/kg
Vss: 3.5 L/kg
Half-life: 1.1 h
hH1R: -9% at 3 µM
hH2R: -11% at 3 µM
hH4R: 10% at 3 µM
Micros. Cl h / r: 5 / 9 µL/min/mg
In vivo pharmacokinetic (rat, iv.)
Cl: 18 mL/min/kg
Asp3.32
(All Histamine R)
In vivo efficacy
Dipsogenia: 92% at 30 mg/kg
Vss: 6 L/kg
Half-life: 4 h
In vivo efficacy
Dipsogenia: 60% at 1 mg/kg
Subchronic model: -12% at 10 mg/kg bid
In vitro safety
Basic center
Acceptor / Electron rich
Aromatic
Thr6.52
(only in H3R)
hERG inhib.: 78% at 3 µM
Fig. 1. 5-Oxy-2-carboxamide indole docked into the H3R
pharmacophore model.
Fig. 2. Selected examples for the influence of the amide substituent on
the metabolic stability.
(t1/2 = 4 h) while maintaining an excellent
affinity (hH3R Ki = 2 nM). It was moreover
extremely selective over the three other
histamine isoforms (Fig. 2).
Effect of compound 2 at 10 mg/kg bid on
food intake of male SD rats on high fat diet
g
(means SEM; n= 7/group)
To our delight, this profile translated
into a robust efficacy in the sub-chronic
model of obesity, in line with our hypoth-
esis on the critical importance of the half-
life (Fig. 3).
24.0
22.0
F
*
F
20.0
Compound 2
However, compound 2 demonstrated
some inhibitory potential for the human
ether-a-go-go related gene (hERG). hERG
inhibition is considered to be an in vitro
predictor of potential cardiovascular liabil-
ities. Taking into account the late failure
of some competitors’ compounds because
of cardiovascular-related issues,[5] the 78%
inhibition value of the hERG channel at 10
µM was therefore judged suboptimal (Fig.
2).
As far as hERG inhibition is concerned,
a basic nitrogen is often suspected as the
usual culprit, although definitely not the
only one.[10]
O
N
O
18.0
16.0
14.0
12.0
10.0
N
N
H
**
Vehicle
Compound 2 at 10 mg/kg bid
***
Fig. 3. Efficacy of
compound 2 in a
subchronic model of
obesity at 10 mg/kg
bid.
0
1
2
3
days of treatment
combining a 1-isopropyl-piperidin-4-oxy fat diet (HFD, 42% of energy intake as fat),
Therefore the relationship between
substituent on the 5-position of the indole and their body weight was recorded. To our pKa, hH R affinity and hERG inhibition
with a cyclic amide on the 2-position. Be- disappointment, compound 1 did not show was stud3ied by varying the needle at the
ing potent (hH R Ki = 5 nM) and meta- any significant effect in this sub-chronic 5-position. Compound 3, an analog of the
bolically stable3in microsomal preparation model. Although the compound was able original compound 1 decorated with an
(micros. Cl human/rat = <1/12 µl/min/mg), to permeate the CNS (brain to plasma ra- additional N-2,2,2-trifluoroethyl group
compound 1 was an excellent candidate for tio = 0.3), pharmacokinetics disappoint- side-chain, exhibited a pKa of 9.7. In the
the first line in vivo screening experiment: ingly revealed a fast metabolism (Cl tot. = same series, replacement of the N-isopro-
the dipsogenia assay.[9] This experiment 50 ml/min/kg) and a short half-life (t1/2 = 1 pyl by a N-cyclopropyl reduced the pKa
is based on an antagonist/inverse agonist h). We hypothesized that a longer half-life by about 2 log units (compound 4: pKa =
reversion of the water-intake induced by might be required to cover the whole ac- 7.5) without a major change on the hH3R
the H3R selective agonist (R)-α-methyl- tive period of the animal (i.e. 12 h during affinity. On the other hand, using the linear
histamine (RAMH) in rat.
In this in vivo functional assay, com- compounds with such a property.
pound 1 reversed the RAMH-induced wa- The replacement of the morpholine by be advantageously replaced by a carbonyl
the dark phase) and started searching for propyloxy linker, it was demonstrated that
the phenolic ether in compound 5 could
ter intake by 92% at 30 mg/kg. Therefore, the more lipophilic piperidine, shielded by group (compound 6). We decided to com-
we decided to further profile this promis- two fluorine atoms, led simultaneously to bine both modifications – a low pKa and a
ing compound in a disease-relevant model. a reduction of the in vivo clearance and an carbonyl group – into a unique needle. As
Compound 1 was applied at a dose of 30 enlarged volume of distribution, both lead- a result, the carbonylpiperazine 7 featured
mg/kg bid (twice daily) over three days to ing towards an improved half-life. Com- a very low pKa value (pKa = 6.8) while
Sprague Dawley (SD) rats fed with a high pound 2 exhibited a promising half-life keeping an excellent binding affinity (hH3R

La u r e a t e s : sCs Pr iz e s a n d sCs Me t t L e r -to L e d o aw a r d s 2008
CHIMIA 2009, 63, No. 5 277
Ki = 7 nM). The hERG inhibition was also
reduced (Fig. 4). Although, thanks to this
exercise it became clear that H3R affinity
was independent of the basicity within a
broad pKa range (ca. 6.8–10.0), the rela-
tionship between pKa and hERG inhibi-
tion remained elusive.
O
O
N
N
O
Compound 3 (R: CH2CF3)
hH3R Ki: 3 nM
Micros. Cl h/r: <1 / 13 µL/min/mg
Compound 4 (R: CH2CF3)
hH3R Ki: 2 nM
Micros. Cl h/r: 11 / 143 µL/min/mg
O
N
Generic core
N
pKa: 9.7
pKa: 7.5
N
O
N
R
O
Compound 7 (R: CH2CF3)
hH3R Ki: 7 nM
Micros. Cl h/r: 10 / 16 µL/min/mg
N
O
N
In order to gain a better understand-
ing, we have used ROSARA, the Roche
Automatic SAR Analyser.[11] ROSARA is
an intuitive and user friendly proprietary
tool which democratizes the use of partial
least square analysis. It is designed to al-
low a fast study of the putative correlations
between a defined compound property (in
this case: hERG inhibition, converted to
pIC50) and numerical descriptors. These
descriptors may be both measured (e.g.
LYSA[12] solubility, logD, hH R Ki, pKa,
etc.) or calculated (e.g. clogP,3 PSA, etc.).
In our case, ROSARA was able to generate
a valuable hERG inhibition model (regres-
sion coefficient = 0.736) by selecting only
two in vitro descriptors: the percentage of
membrane-bound compound in the Paral-
lel Artificial Membrane Permeation Assay
(PAMPA)[13] and the logD value (measured
lipophilicity at pH= 7.4). As some small
molecule hERG inhibitors are assumed to
interact with this channel through the mem-
brane section, this correlation did indeed
makesense. Interestingly, althoughthepKa
spanned over more than three log units in
the training set, no correlation with hERG
affinity was observed. The model has been
validated with two new compounds from
the same series (blue squares, Fig. 5).
The independence of pKa and hERG
inhibition encouraged us to reinvestigate
compounds having rather strong basicity
(pKa = 9–10). Among them, compound 9
(Fig. 6) attracted our attention by its strik-
ing combination of a high H3R binding
affinity (hH3R Ki = 7 nM), low clearance
in vivo (rat Cl tot. = 10 ml.min–1.kg–1), ex-
tended half life (t = 11 h in rat), rather low
hERG inhibition1(/242% hERG inhibition at
10 µM) and, eventually, excellent perfor-
mance in the 3-day obesity model (–17%
reduction of food intake at 10 mg/kg bid.
compared to control).
Compound 5 (R: H)
hH3R Ki: 71 nM
Micros. Cl h/r: <1 / 6 µL/min/mg
Compound 6 (R: H)
hH3R Ki: 30 nM
Micros. Cl h/r: <1 / 14 µL/min/mg
pKa: 6.8
hERG inhib.: 33% at 10 µM
pKa: 10.0
pKa: 9.9
Fig. 4. Selected examples exploring the side chain basicity.
5.6
Training set:
ROSARA: Roche Automatic SAR Analyser
- Intuitive and user friendly proprietary tool
- Democratizing the use of partial least square analysis
- Designed to study potential correlation between a
defined compound property (i.e. hERG inhibition) and
numerous descriptors (both measured or calculated
values).
5.4
5.2
5.0
4.8
4.6
4.4
4.2
4.0
3.8
3.6
Validation set:
Line of Identity:
F
F
O
N
O
N
N
Compound 9
pKa: 9.9
O
O
N
N
N
N
F
O
O
Compound 7
pKa: 6.8
F
Application:
H
F
N
N
-hERG pIC50 = f (Membrane fraction in PAMPA, logD)
N
-Regression coefficient: 0.736
N
H
O
Compound 8
pKa: 9.9
-Prediction used to prioritize compound profiling
5.0
5.4
3.8
4.0
4.2
4.4
4.6
4.8
5.2
Predicted pIC50
Fig. 5. Correlation of experimental vs. calculated hERG pIC50, using Roche Automatic SAR
Analyser (ROSARA).
Compound 9: synthesis
Compound 9: profile
2-PrOMs 1.75 eq.
Cs2CO3 1.75 eq.
CH3CN, reflux, 18h
In vitro Pharmacology
Human H3R Ki: 7 nM
BnO
BnO
OEt
O
OEt
O
Human H3R EC50 (GTPγS): 5 nM (inv. ago.)
crystallisation
89%
N
H
N
Rat H3R Ki: 5 nM
Compound 11
Compound 10
Commercially
available
Human H1R: -3% at 3 µM
Human H2R: -2% at 3 µM
Human H4R: 1% at 3 µM
LiOH·H2O 1.1 eq.
THF/ H2O/ MeOH 3/1.5/1
80°C , 2h
precipitation
95%
Pd/C 10% 0.03 eq.
H2 1 Atm.
In vitro pharmacokinetic
AcOEt/EtOH 1.4/1
R.T., 2.5h
HO
BnO
OH
O
OH
Micros. Cl h/r: 4 / 12 µL/min/mg
no purification
quant.
N
N
O
In vivo pharmacokinetics (rat, i.v.)
Clearance: 10 mL/min/kg
Vss: 9 L/kg
Compound 13
EDC·HCl 2 eq.
Compound 12
4,4'-DFP.HCl 2 eq.
Hunig's base 1 eq.
DMF, R.T., 1h
Half-life: 11 h
1-Isopropyl-piperidin-4-ol
PPh3 1.2 eq.
F
precipitation
85%
F
In vivo Efficacy
F
F
Dipsogenia: -80% at 3 mg/kg
3-Days food intake: -17% at 10 mg/kg bid
DtBAD 1.2 eq.
THF, R.T., 24h
O
HO
N
N
Chromatography
recrystallisation
60%
N
In vitro Safety
O
N
N
O
hERG inhibition: 42% at 10 µM
Compound 9
Compound 14
Fig. 6. Profile and chemical synthesis of compound 9.
Effect of compound 9 on body weight development
The synthesis of compound 9 was
straightforward: commercially available
5-benzyloxy-1H-indole-2-carboxylic ac-
id ethyl ester 10 was N-alkylated with iso-
propyl methane sulfonate. Both the acidic
and the phenolic functionalities were lib-
erated by saponification with lithium hy-
droxide and hydrogenolysis, respectively,
followed by an amine coupling reaction.
The ether was obtained by applying a
Mitsunobu reaction. This end-game reac-
tion required extensive optimization in
order to minimize the elimination of the
activated alcohol as a side-reaction. The
overall yield was satisfactory (43% over
five steps).
g
in male SD rats on HFD (Means SEM; n: 7-8/group)
560
540
520
500
480
460
440
420
400
Control citric acid
Compound 9 at 5 mg/kg bid
Compound 9 at 10 mg/kg bid
Compound 9 at 20 mg/kg bid
Fig. 7. Chronic
application of
compound 9 at 5
(yellow), 10 (orange)
and 20 (red) mg/kg
bid in a Diet-Induced
Obesity rat model.
treatment
wash out
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 19 20 23 25 27 30 days

278 CHIMIA 2009, 63, No. 5
La u r e a t e s : sCs Pr iz e s a n d sCs Me t t L e r -to L e d o aw a r d s 2008
[1] M. E. Parsons, C. R. Ganellin, Br. J. Pharmacol. [11] H. Fischer, M. Kansy, PCT Int. Appl. US
Applied in a chronic model of obesity
(DIO model in male SD rats) over 17 days,
compound 9 led to an impressive and dose-
dependent slower progression of body
weight gain in all groups (–4, –10, –13%
respectively), statistically significant for
the 10 and 20 mg/kg groups (Fig. 7).
2006, 147, S127.
2007027632 A1, 2007.
[2] a) S. J. Hill, C. R. Ganellin, H. Timmerman, J. [12] LYophilisation Solubility Assay (LYSA):
C. Schwartz, N. P. Shankley, J. M. Young, W.
Schunack, R. Levi, H. L. Haas, Pharmacol.
Rev. 1997, 49, 253; b) S. J. Offord, Clin.
solubility was measured from lyophilized
DMSO stock solutions spectrophotometrically
at pH = 6.5 in a 0.05M phosphate buffer.
Pharmacol. Ther. 2002, 71, Abst WPIII-27 P83; [13] M. Kansy, F. Senner, K. Gubernator, J. Med.
c) M. Harada, M. Terai, H. Maeno, Biochem.
Pharmacol. 1983, 32, 1635.
Chem. 1998, 41, 1007.
[14] a) M. Nettekoven, J.-M. Plancher, H. Richter,
O. Roche, S. Taylor, PCT Int. Appl. WO
2007115938A1, 2007; b) M. Nettekoven,
J.-M. Plancher, O. Roche, T. Takahashi, S.
Taylor, PCT Int. Appl. WO2007080140, 2007;
c) M. Nettekoven, J.-M. Plancher, H. Richter,
O. Roche, S. Taylor, U.S. Pat. Appl. Publ. US
2007123525 A1, 2007; d) M. Nettekoven,
J.-M. Plancher, H. Richter, O. Roche, R. M.
Rodriguez Sarmiento, S. Taylor, U.S. Pat.
Appl. Publ. US 2007123526 A1, 2007; e) M.
Nettekoven, J.-M. Plancher, H. Richter, O.
Roche, V. Runtz-Schmitt, S. Taylor, U.S. Pat.
Appl. Publ. US 2007123515 A1, 2007; f) M. H.
Nettekoven, J.-M. Plancher, O. Roche, R. M.
Rodriguez-Sarmiento, U.S. Pat. Appl. Publ. US
2006160855 A1, 2006.
[3] a) J.-M. Arrang, M. Garbarg, J.-C. Schwartz,
Nature 1983, 302, 832; b) R. Leurs, P. Blandina,
C. Tedford, H. Timmerman, Trends Pharmacol.
Sci. 1998, 19, 177; c) R. Leurs, H. Timmerman,
in ‘The Histamine H3-Receptor: A Target for
New Drugs’, Ed. Elsevier, Amsterdam, 1998; d)
H. Stark, J.-M. Arrang, X. Ligneau, M. Garbarg,
C. R. Ganellin, J. C. Schwartz, W. Schunack, in
‘The Histamine H3-receptor and its Ligands in
Progress in Medicinal Chemistry’, Ed. Elsevier,
Amsterdam, 2001.
[4] a) T. A. Esbenshade, G. B. Fox, M. D. Cowart,
Mol. Interventions 2006, 6, 77; b) T. A.
Esbenshade, K. E. Browman, R. S. Bitner, M.
Strakhova, M. D. Cowart, J. D. Brioni, Brit. J.
Pharmacol. 2008, 1.
4. Conclusion
The 5-indole-2-carboxamide-indole
scaffold was identified as a valid chemistry
entry point by the use of a pharmacophore
model. Aseriesofderivativeswasidentified
as novel class of H3R inverse agonists.[14]
After SAR development, extensive mul-
tidimensional optimization, and careful
evaluation of the underlying drivers for the
H3R binding affinity and hERG inhibition,
compound 9 proved to combine suitable
overall characteristics for advancement. In
the indirect acute pharmacodynamic dip-
sogenia model, compound 9 reversed the
water intake induced by the H R selective
agonist RAMH. Furthermore, 3in a 17-day
Diet-Induced Obesity (DIO) rat model,
compound 9 demonstrated a robust and
dose dependant reduction in body weight
gain. On the basis of its efficacy, metabolic
stability and safety profile, compound 9
was selected for further evaluation of its
potential for the treatment of obesity.
[5] A. A. Hancock, Bio. Pharm. 2006, 71, 1103.
[6] O. Roche, R. M. Rodriguez Sarmiento, Bio.
Med. Chem. Lett. 2007, 17, 3670.
[7] O. Roche, M. Nettekoven, W. Vifian, R. M.
Rodriguez Sarmiento, Bio. Med. Chem. Lett.
2008, 18, 4377.
[8] M. Nettekoven, J.-M. Plancher, H. Richter, O.
Roche, R. M Rodriguez Sarmiento, S. Taylor,
Bio. Med. Chem. Lett. in press.
[9] a) G. B. Fox, J. B. Pan, T. A. Esbenshade, R.
S. Bitner, A. L. Nikkel, T. Miller, C. H. Kang,
Y. L. Bennani, L. A. Black, R. Faghih, A. A.
Hancock, M. W. Decker, Pharm. Bio. & Behav.
2002, 72, 741.
(b) X. Ligneau, J.-S. Lin, G. Vanni-Mercier, M.
Jouvet, J. L. Muir, C. R. Ganellin, H. Stark, S.
Elz, W. Schunack, J.-C. Schwartz, J. Pharm.
Exp. Ther. 1998, 287, 658.
Acknowledgment
Authors wish to thank Dr. Holger Fischer
for his help in the use of Roche Structure
Activity Relationship Analyser (ROSARA).
[10] a) L. Jia, H. Sun, Bio. Med. Chem. 2008, 16,
6252; b) E. Raschi, V. Vasina, E. Poluzzi, F. De
Ponti, Pharmacol. Res. 2008, 57, 181; c) M. J.
Waring, C. Johnstone, Bio. Med. Chem. Lett.
2007, 17, 1759; d) A. M. Aronov, J. Med. Chem.
2006, 49, 6917.
Received: December 19, 2008