Discovery of (3-(4-(2-Oxa-6-azaspiro[3.3]heptan-6-ylmethyl)phenoxy)azetidin-1-yl)(5-(4-methoxyphenyl)-1,3,4-oxadiazol-2-yl)methanone (AZD1979), a Melanin Concentrating Hormone Receptor 1 (MCHr1) Antagonist with Favorable Physicochemical Properties

Article abstract of DOI:10.1021/acs.jmedchem.5b01654

A novel series of melanin concentrating hormone receptor 1 (MCHr1) antagonists were the starting point for a drug discovery program that culminated in the discovery of 103 (AZD1979). The lead optimization program was conducted with a focus on reducing lipophilicity and understanding the physicochemical properties governing CNS exposure and undesired off-target pharmacology such as hERG interactions. An integrated approach was taken where the key assay was ex vivo receptor occupancy in mice. The candidate compound 103 displayed appropriate lipophilicity for a CNS indication and showed excellent permeability with no efflux. Preclinical GLP toxicology and safety pharmacology studies were without major findings and 103 was taken into clinical trials.

Full text of DOI:10.1021/acs.jmedchem.5b01654

Subscriber access provided by NEW MEXICO STATE UNIV
Article
Discovery of (3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-
ylmethyl)phenoxy)azetidin-1-yl)(5-(4-methoxyphenyl)-1,3,4-oxadiazol-2-
yl)methanone (AZD1979), a melanin concentrating hormone receptor
1 (MCHr1) antagonist with favourable physicochemical properties
Anders Johansson, Christian Löfberg, Madeleine Antonsson, Sverker von
Unge, Martin Hayes, Robert Judkins, Karolina Ploj, Lambertus Benthem, Daniel
Lindén, Peter Brodin, Marie Wennerberg, Marléne Fredenwall, Lanna Li, Joachim
Persson, Rolf Bergman, Anna Pettersen, Peter Gennemark, and Anders Hogner
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01654 • Publication Date (Web): 07 Jan 2016
Downloaded from http://pubs.acs.org on January 12, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of (3ꢀ(4ꢀ(2ꢀoxaꢀ6ꢀazaspiro[3.3]heptanꢀ6ꢀ
ylmethyl)phenoxy)azetidinꢀ1ꢀyl)(5ꢀ(4ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
methoxyphenyl)ꢀ1,3,4ꢀoxadiazolꢀ2ꢀyl)methanone
(AZD1979), a melanin concentrating hormone
receptor 1 (MCHr1) antagonist with favourable
physicochemical properties
Anders Johansson*, Christian Löfberg, Madeleine Antonsson, Sverker von Unge, Martin A.
Hayes, Robert Judkins, Karolina Ploj, Lambertus Benthem, Daniel Lindén, Peter Brodin, Marie
Wennerberg, Marléne Fredenwall, Lanna Li, Joachim Persson, Rolf Bergman, Anna Pettersen#,
Peter Gennemark and Anders Hogner
Cardiovascular & Metabolic Diseases iMed and #Global Medicines Development, AstraZeneca
Gothenburg, 431 83 Mölndal, Sweden.
ABSTRACT: A novel series of melanin concentrating hormone receptor 1 (MCHr1) antagonists
were the starting point for a drug discovery program that culminated in the discovery of 103
(AZD1979). The lead optimisation program was conducted with a focus on reducing
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 48
1
2
3
4
5
6
7
8
9
lipophilicity and understanding the physicochemical properties governing CNS exposure and
undesired offꢀtarget pharmacology, such as hERG interactions. An integrated approach was
taken where the key assay was ex vivo receptor occupancy in mice. The candidate compound 103
displayed appropriate lipophilicity for a CNS indication and showed excellent permeability with
no efflux. Preclinical GLP toxicology and safety pharmacology studies were without major
findings and 103 was taken into clinical trials.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Introduction
Obesity is one of the underlying causes of several serious conditions such as type II diabetes,
coronary heart disease, high blood pressure, liver disease and cancer.1 Over the last 15 years,
great efforts have been devoted to finding effective, safe and easy to use weight loss
medications2 and much of the attention has been directed towards appetite control.3 The cyclic
19 amino acid neuropeptide MCH is associated with feeding behavior.4 It activates the MCH
receptor1 (MCHr1) located in the hypothalamus. The receptor is also known as the human
orphan Gꢀprotein coupled receptor SLCꢀ1 or GPR24.5 Mice lacking MCH are lean and
hypophagic.6 It has been shown independently by several groups that MCHr1 knockout (KO)
mice are resistant to dietary induced obesity (DIO) mainly due to elevated motor activity and
energy expenditure.7 Small molecule MCHr1 antagonists have been discovered and the effect
observed on body weight is in accordance with those for KO animals.8 However, while
experimental data in rodents are supportive, human translation remains unconfirmed.9 In nonꢀ
rodents, the picture is complicated by expression of a second MCH receptor, MCHr2,10 for
which a role has yet to be defined.11 Only very limited data is available in the literature on food
intake/weight loss in MCHr2 expressing species12 and only recently a selective MCHr2
ACS Paragon Plus Environment
2

Page 3 of 48
Journal of Medicinal Chemistry
1
2
antagonist was disclosed.13 Although the main attention of MCHr1 antagonists has been drawn to
their potential for the treatment of obesity, other biological effects are also worth mentioning.
For instance, MCHr1 antagonists have been investigated preclinically as a treatment option for
depression and anxiety disorders.14 Furthermore, an MCHr1 antagonist has been shown to
alleviate steatohepatitis15 and MCHr1 KO mice are resistant to seizures.16
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The established connection between the MCHr1 system and weight control in rodents has
stimulated extensive efforts in the drug discovery of small molecules as MCH receptor
antagonists.17 Despite being an attractive target for indications such as obesity and
depression/anxiety, a major hurdle has been to identify a chemical series with attractive
physicochemical properties. A large number of published MCHr1 antagonists are lipophilic
bases. Not surprisingly, one known liability for many programs has been to circumvent
interactions with the hERG channel.18 However, not only is it necessary to achieve a large
enough separation between MCHr1 and hERG, but this also needs to be accomplished without
impairing necessary CNS exposure.19 This presents quite a challenge.20
Results and discussion
Following an earlier MCHr1 program at AstraZeneca,21 a new high throughput screen (HTS)
looking for compounds binding to MCHr1 was run on a subset of the AstraZeneca compound
collection. The HTS was run looking for functional antagonists using a high throughput FLIPR
assay.22 Special attention was paid to include all compounds added to the collection since the
previous HTS in the earlier program. Amongst the hits, one series was identified which was
considered to possess promising physicochemical properties. To guide continued chemistry
efforts a functional primary assay was preferred and a [35S]GTPγS assay was developed for
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 48
1
2
MCHr1 and used throughout the program.23 The lead identification chemistry culminated in
compound 1, which was the starting point for the lead optimisation program. A summary of the
key properties of 1 is given in Table 1. An analysis of the physicochemical property space of
known MCHr1 antagonists was made. Physicochemical properties were calculated for all
publicly disclosed MCHr1 antagonists, including compounds from patent applications derived
from the online database GOSTAR.24 This gave an average clogP of 5.2 and an average polar
surface area (PSA) of 54.25 The physicochemical properties of 1 contrasted to the
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
physicochemical property space of published MCHr antagonists and suggested it to be a good
starting point. Additionally, and even more importantly, this compound and the series it
represented had high passive permeability as assessed in Caco2 cells and exhibited no
Table 1. In vitro data for compound 1
IC50
MCH
IC50
MCH
Caco
ABf
(Efflux
ratio)
IC50
CYP
min
IC50
hERG
(µM)
Solubilityb
(µM)
pKa
(calc)
logDc
(ClogP)
HLM
RLM
d
PSA
63
LLEe
4.3
Clintg Clintg
GTPγS bind.
(nM)a (nM)
(3A4)
3.4
(2.9)
43
(0.37)
20
11
5.5
70
9.8
5.2
13
17
Functional potency.23 b Solubility at pH 7.4 from DMSO solution. LogD7.4 hplc
Polar
a
c
d
Surface Area e LLE calculated from pIC50 MCH GTPγS minus logD. f 10ꢀ6 cm/s g µl/min/mg
ACS Paragon Plus Environment
4

Page 5 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
efflux. This suggested a lack of interaction with efflux transporter proteins and the potential to
achieve high exposure in the CNS. A deeper understanding of target engagement in the CNS can
be attained by measuring receptor occupancy in the regions of interest in the brain. Although the
receptor expression level of MCHr1 is quite low both in both rodents and humans,26 ligands that
are useful for measuring ex vivo receptor occupancy (RO) have been reported.27 Our RO model
was based on the use of [3H](S)ꢀmethyl 3ꢀ(3ꢀ(4ꢀ(3ꢀacetamidophenyl)piperidinꢀ1ꢀ
yl)propylcarbamoyl)ꢀ4ꢀ(3,4ꢀdifluorophenyl)ꢀ6ꢀ(methoxymethyl)ꢀ2ꢀoxoꢀ1,2,3,4ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tetrahydropyrimidineꢀ5ꢀcarboxylate ([3H]ꢀSNAPꢀ7941) as radioligand.28 After po administration
of 50 µmol/kg of 1 to mice a MCHr1 receptor occupancy of more than 90% could be detected,
thus providing evidence of compound exposure and target coverage in the CNS. The lead
compound and the series it represents contained an amide bond, which was not completely stable
to amidase hydrolysis when measuring rat and mouse plasma protein binding data using a
standard method. However, by shortening the equilibration time in the assay this could be
overcome. Hydrolytic cleavage was never observed when analysing plasma samples from in vivo
studies. Further in vivo assessment of the compound included a 21 day weight loss study in diet
induced obese (DIO) mice, where a dose related decrease in body weight was observed.
However, the compound contained the threeꢀpoint hERG pharmacophore29 with two aromatic
rings and a basic nitrogen and showed an IC50 of 5 µM on the hERG channel. Although no acute
CV findings were detected in vivo in anaesthetized guinea pigs, a reduction in hERG was
considered an improvement parameter for the lead optimisation program. A move to nonꢀbasic
compounds can be a strategy for removing hERG interactions.30 However, in this series, without
exception, replacing the basic N with amide functionalities resulted in loss or severe decline in
MCHr1 potency. This change also moved compounds into more lipophilic physicochemical
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 48
1
2
3
4
5
6
7
8
9
space, where the most lipophilic variants in some cases displayed hERG activity in the low micro
molar range, despite the absence of a basic N. The physicochemical properties of 1 are a good
starting point, with a logD of 3.4, a solubility of 70 µM and an attractive permeability profile.
Whilst the basic N appeared to be an integral part of the lead, we hypothesized that it should be
possible to modify chemistry in the series to achieve a wide enough margin over the hERG
channel by driving down lipophilicity to minimize hERG interactions and other potential off
target pharmacology.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Chemistry
The lead series was divided into three distinct regions as shown in Figure 1. SAR investigations
took advantage of the synthetic flexibility in terms of selection of last step coupling to provide
the final compound. Final compounds were prepared either by an amide formation between the
Figure 1. (A) Series regions and coupling strategies for MCH1r lead series. (B) Low energy
conformation of 1 used as 3Dꢀhypothesis for SAR analysis.
secondary N of the linker region and an eastern region ester/acid or the final step was a reductive
amination of a benzylic aldehyde, by which the western region amine was introduced last. A
ACS Paragon Plus Environment
6

Page 7 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
third option was to form the ether bond as the last step by a nucleophilic displacement reaction.
This was less frequently used in the lead optimization program, but could be applied with good
atom economy when preparing larger batches for toxicological studies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Preparation of final compounds using amide formation as the last step started with preparation of
western amine fragments as shown in Scheme 1. Reductive amination of a 4ꢀ
hydroxybenzaldehyde with the selected amine using Na(CH3CO2)3BH and triethylamine in
Scheme 1a Synthesis of building blocks 4, 6, 9, 12, 15, 18, 21, 25, 28, 31, 34, 37.
aReagents and conditions: (i) Amino compound, Na(CH3CO2)BH, TEA, DCM, rt (81ꢀ96%); (ii)
Ph3P, tertꢀbutyl 4ꢀhydroxypiperidineꢀ1ꢀcarboxylate, DIAD, DCM, 0°C → rt (77ꢀ93%); (iii) The
NꢀBoc deprotection was accomplished by different techniques for different compounds.
Synthetic details are presented in the Experimental section; (iv) NaH or Cs2CO3, tertꢀbutyl 3ꢀ
[(methylsulfonyl)oxy]ꢀazetidineꢀ1ꢀcarboxylate, DMF, 80ꢀ90°C (55ꢀ99%). Compound 22 was
synthesized using 2ꢀfluoroꢀ4ꢀmethoxybenzaldehyde during a reductive amination step followed
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 48
1
2
3
4
5
6
7
8
9
by an ether hydrolysis.; (v) The NꢀBoc deprotection was accomplished by different techniques
for different compounds. Synthetic details are presented in the Experimental section. Compound
34 was synthesized in a reversed fashion, i.e. (i) Mitsunobu reaction (52%); (ii) Reductive
amination (76%); (iii) NꢀBoc deprotection (67%).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
dichloromethane provided the pꢀhydroxy benzylamine products in good yield. Substitutions in
the phenyl region were generally well tolerated and the method could be used for a variety of
amines, from dimethyl amine and pyrrolidine to the less nucleophilic morpholine and piperazine.
The western amine fragment was then connected via formation of an ether to a piperidine or
azetidine linker. Two different methods were utilized to accomplish this ether formation.
Piperidine building blocks were prepared using a Mitsunobu reaction between the pꢀhydroxy
benzylamine and 4ꢀhydroxy Bocꢀpiperidine followed by Bocꢀdeprotection. Azetidine ethers were
prepared by Oꢀalkylation of the pꢀhydroxy benzylamine by Bocꢀprotected azetidinꢀ3ꢀmesylate, in
the presence of NaH or Cs2CO3, followed by acidic hydrolysis of the Bocꢀgroup. The reaction
order could also be reversed, as exemplified by 34, where the Mitsunobu reaction between Bocꢀ
piperidinꢀ4ꢀol and 4ꢀhydroxybenzaldehyde was performed first, followed by reductive amination
Scheme 2a Conditions for amide formation.
ACS Paragon Plus Environment
8

Page 9 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (i) R=OCHF2, 4ꢀ(4ꢀ(Pyrrolidinꢀ1ꢀylmethyl)phenoxy)ꢀpiperidine (4),
EtOH, reflux 20h, 37%. (ii) R=H, 1ꢀ(4ꢀ(Azetidinꢀ3ꢀyloxy)ꢀ3ꢀmethylphenyl)ꢀN,Nꢀ
dimethylmethanamine (18), neat, 120 °C, 3h, 57%. (iii) R=MeO, 6ꢀ(4ꢀ(azetidinꢀ3ꢀyloxy)benzyl)ꢀ
2ꢀoxaꢀ6ꢀazaspiro[3.3]heptane (37), catalytic NaCN, MeOH, rt, 2h, 65%. The structures of 60, 81
and 103 are shown in Table 2, Table 4 and Table 6, respectively
and deprotection. Amide formation could be accomplished in several ways. However, since the
1,3,4ꢀoxadiazole carboxylic acids decarboxylate spontaneously at room temperature, standard
coupling reagents could not be used and therefore coupling reactions were performed with ethyl
esters. The different conditions used are summarized in Scheme 2. The reaction conditions
utilized initially (Scheme 2, i) worked well for R= pꢀCl or other electron withdrawing
substituents. However, for electron rich 1,3,4ꢀoxadiazole esters, where R= pꢀMeO, these
conditions suffered from long reaction times along with significant byꢀproduct formation. For
example, considerable amounts of 2ꢀphenylꢀ1,3,4ꢀoxadiazole was formed by hydrolysis of the
ester, followed by decarboxylation. To avoid these problems an alternative protocol was
developed (Scheme 2, ii). In this protocol, the amine building block was allowed to react with
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 48
1
2
3
4
5
6
7
8
9
the 1,3,4ꢀoxadiazole ester under solventꢀfree conditions (neat) at 120 °C for 2ꢀ3h. This procedure
provided rapid and straightforward access to final compounds with a cleaner reaction profile,
improved reaction times and easier purification. Often, the final compounds were crystallized
directly from the crude reaction mixture. Reaction conditions were not optimized for each
individual compound. However, a further improved method was developed (Scheme 2, iii). The
mild NaCN catalysed coupling of the amine to the 1,3,4ꢀoxadiazole phenyl ester at room
temperature.31 This method became the preferred method as the program developed. Many other
heterocycles in the eastern region were stable as carboxylic acids and therefore classic amide
coupling conditions were employed (Scheme 3). It is worth noting that the 1,2,4ꢀoxadiazole
phenyl amides also were prepared by coupling to corresponding methyl esters.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3a Synthesis of analogues 63-67.
aReagents and conditions: (i) For compounds 63, 65 and 66: 4, EDC, DMAP, DCM, rt (65ꢀ72%);
(ii) For compounds 64 and 67: 4, NaCN (cat), MeOH, rt (52ꢀ83%). For detailed structural
information of compound 63-67, see Table 3.
For synthesis of piperidine linked western pyridines, the synthetic routes were modified.
Preparation of 46 started with reductive amination of 6ꢀchloronicotinaldehyde with pyrrolidine
using Na(CH3CO2)3BH and triethylamine in dichloromethane (Scheme 4). Nucleophilic
ACS Paragon Plus Environment
10

Page 11 of 48
Journal of Medicinal Chemistry
1
2
3
4
displacement of Cl with Bocꢀprotected 4ꢀhydroxypiperidine in the presence of NaH in DMSO
5
6
Scheme 4a Synthesis of pyridine analogue 46.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
aReagents and conditions: (i) Pyrrolidine, Na(CH3CO2)3BH, TEA, DCM, rt (79%); (ii) tertꢀbutyl
4ꢀhydroxypiperidineꢀ1ꢀcarboxylate, NaH, DMSO, 70°C (57%); (iii) TFA, DCM, rt (92%); (iv)
Ethyl 5ꢀ(4ꢀmethoxyphenyl)ꢀ1,3,4ꢀoxadiazoleꢀ2ꢀcarboxylate, EtOH, reflux (48%). Compound 46
is shown in Table 4.
and subsequent deprotection using TFA in dichloromethane gave intermediate 45. Amide
formation with the pꢀmethoxyphenyl oxadiazole ester in refluxing ethanol afforded 46.
Synthesis of the pyridine analogue 51 started with formation of the acid chloride of 5ꢀ
hydroxypicolinic acid, using thionyl chloride. Treatment with pyrrolidine gave the corresponding
amide (Scheme 5). Subsequent reduction of the amide to amine was accomplished using
(CH3)2S.BH3 in THF. Mitsunobu coupling followed by deprotection of the Bocꢀgroup gave 50,
which was finally converted to 51 by treatment with pꢀmethoxyphenyl oxadiazole ester under
solventꢀfree conditions at 120 °C.
Scheme 5a Synthesis of pyridine analogue 51.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 48
1
2
aReagents and conditions: (i) SOCl2, DCM, reflux 3 h, then pyrrolidine, rt (25%); (ii)
(CH3)2S.BH3, THF, 0°C → rt, then MeOH; (iii) Ph3P, tertꢀbutyl 4ꢀhydroxypiperidineꢀ1ꢀ
carboxylate, DIAD, THF, 0°C → rt (35%, two steps); (iv) TFA, DCM, rt (46%); (v) Ethyl 5ꢀ(4ꢀ
methoxyphenyl)ꢀ1,3,4ꢀoxadiazoleꢀ2ꢀcarboxylate, neat, 120°C (38%). Compound 51 is shown in
Table 4.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The linear synthetic sequence culminating in a reductive amination reaction to form the final
compounds was employed most widely throughout the program, as the largest number of
variations was investigated in the western part of the molecule and this route introduces diversity
in this part of the molecule as the last step (Scheme 6). The common aldehyde building blocks
were prepared starting from the desired phenylic 1,3,4ꢀoxadiazole ester, which was treated with
3ꢀhydroxyazetidine, catalytic amounts of NaCN and triethylamine in methanol to form the
corresponding 3ꢀhydroxyazetidine amides in good yield. The free hydroxyl group was readily
converted to a mesylate, under standard conditions, to afford 53 and 56 in high yield.
Scheme 6a Synthesis of building blocks 54, 57.
aReagents and conditions: (i) NaCN, 3ꢀhydroxyazetidinium chloride, TEA, MeOH, rt (77ꢀ87%);
(ii) TEA, MsCl, DCM, 0°C → rt (77ꢀ99%); (iii) 4ꢀhydroxybenzaldehyde, Cs2CO3, DMA, 110°C
(54ꢀ75%).
Finally, the aldehyde intermediates 54 and 57 were obtained by Oꢀalkylation of 4ꢀ
hydroxybenzaldehyde with 53 and 56 in the presence of Cs2CO3 in DMA at 110 °C. These
ACS Paragon Plus Environment
12

Page 13 of 48
Journal of Medicinal Chemistry
1
2
3
4
western aldehyde building blocks were then converted into a variety of target benzylic amines
5
6
via a reductive amination step as shown in Scheme 7.
7
8
9
Scheme 7a Synthesis of analogues 89, 90, 94, 97, 98, 99, 100, 101, 102, 103 by a reductive
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
amination as the final step.
aReagents and conditions: (i) Na(CH3CO2)3BH, TEA, DCM, 0 °C → rt (17ꢀ76%). The
structures of the target benzylic amines are shown in Tables 5 and 6.
Structure activity relationships and discovery of the clinical candidate 103 (AZD1979)
The reported MCHr1 pharmacophore with a planar elongated structure containing a hydrophobic
aromatic moiety, a hydrogen bond acceptor, and a positively charged site appeared to be relevant
for the chemical series.18, 32 However, a more detailed hypothesis of a putative bioactive
conformation of 1 arose from a conformational analysis of a set of published MCH1r antagonist
structures.33 Lowꢀenergy conformations of each molecule were generated, and subsequently pairꢀ
wise shapeꢀbased alignments were performed.34 This resulted in a number of molecular
alignments, where visual inspection favored the alignment shown in Figure 1b, since it allowed
the rationalization of series specific SAR within the chemical series. This molecular alignment
served as an initial working 3Dꢀhypothesis for SAR analysis and the 3D model was in line with
the characteristics of the reported MCHr1 pharmacophore.
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 48
1
2
3
4
Table 2. Eastern phenyl substitutions
5
6
7
8
9
Caco
ABd
(Efflux
ratio)
IC50
MCH
IC50
hERG
(µM)
Solubilityb
(µM)
logDc
(ClogP)
HLM
Structure
Compound
LLEf
5.2
GTPγSa
(nM)
Clinte
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.2
(2.2)
55
(0.28)
43
28
11.0
11.0
4.2
>100
>100
ND
S
S
58
59
60
61
62
2.3
(2.2)
46
(0.23)
5.3
4.9
5.0
3.3
2.9
(2.7)
50
(0.21)
15
15
S
2.4
(2.3)
52
(0.3)
44
6.2
>100
ND
2.7
(2.7)
59
(0.34)
1000
5.3
13
Functional potency.23 b Solubility at pH 7.4 from DMSO solution. LogD7.4 hplc 10ꢀ6 cm/s
a
c
e
d
µl/min/mg f LLE calculated from pIC50 MCH GTPγS minus logD. ND Not determined S Stable
to assay conditions (100% parent compound remaining after incubation)
Initially, attention was turned to the eastern region of the molecules. An early assessment of the
series included replacements for the pꢀCl substituent, which gratifyingly proved not to be
essential for MCHr1 potency. As shown in Table 2, MCHr1 potency does not vary significantly
between paraꢀsubstituted phenyls in the eastern region. However, a marked drop in potency
could be observed for other substitution patterns. In line with the 3D working hypothesis, oꢀClꢀ
phenyl 6 forces the phenyl group out of plane. This drastically affects the MCHr1 potency and
indicates that the MCHr1 site is complementary to the planar shape of the distal phenyl ring.
ACS Paragon Plus Environment
14

Page 15 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Permeability and efflux characteristics, as measured in Caco2 cells, did not differ with
modifications in the eastern phenyl region, as all variants showed high Caco2 permeability with
no signs of efflux. Furthermore, metabolic stability in vitro in both rat and human liver
microsomes was improved when the Clꢀsubstituent was replaced. For example, phenyl derivative
58 and pꢀMeOꢀphenyl derivative 59 were both completely stable in both rat and human liver
microsomal incubations. Additionally, 59 did not show any sign of CYP450 inhibition against
any of the CYP450’s measured (CYP1A2, CYP3A4, CYP2C8, CYP2C9, CYP2C19, and
CYP2D6). Within this set of compounds, pꢀphenyl substitution showed small variations in hERG
potency. However, when taking measured logD into account and calculating lipophilic ligand
efficiency (LLE), the unsubstituted phenyl derivative 5 and pꢀMeOꢀphenyl derivative 6 were
considered most promising. Both 58 and 59 gave a small decrease in hERG potency and, in
combination with the highest LLE values, made these variations attractive for further
exploration. Hence, it was reasoned that these would provide the lowest baseline lipophilicity
and unsubstituted phenyl and pꢀMeOꢀphenyl were chosen as eastern substituents for SAR
expansion of the other regions of the molecule.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Eastern region heterocycles
IC50
IC50
MCH
GTPγSa
(nM)
Solubilityb
(µM)
logDc
(ClogP)
HLM
Structure
Compound
hERG
(µM)
LLEe
4.2
Clintd
3.0
(3.6)
63
61
3.3
7.5
198
991
S
6
2.6
(2.8)
64
150
4.2
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 48
1
2
3
4
5
6
7
8
9
2.2
(3.5)
65
66
67
I
7.4
3.0
1.8
930
754
7
S
S
NA
2.9
3.9
3.0
(3.5)
1200
75
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3.2
(2.8)
<0.005
a Functional potency.23 b Solubility at pH 7.4 from DMSO solution. c LogD7.4 hplc d µl/min/mg e
LLE calculated from pIC50 MCH GTPγS minus logD. S Stable to assay conditions (100% parent
compound remaining after incubation) I Inactive
In the heterocyclic eastern region, several five membered ring heterocycles were investigated
(Table 3). However, the 1,3,4ꢀoxadiazole 58 proved to best combine physicochemical properties
and MCHr1 potency. Displaying the lowest lipophilicity of the heterocycles investigated, the
1,3,4ꢀoxadiazole 58 also gave the greatest separation between MCH potency and hERG activity.
In addition, the LLE was the highest of the heterocycles examined. There is a remarkable
variation in solubility and measured logD between the 1,3,4ꢀoxadiazole and one of the two
isomeric 1,2,4ꢀoxadiazoles, 67.35 The difference in activity on the hERG channel between 58 and
67 is also noteworthy. Changing the shape and electrostatic properties by substitution on the 5ꢀ
membered heterocyclic rings in 65 and 66 were not beneficial for MCHr1 potency, again
emphasizing the importance of the planar shape complementarity towards this part of the MCHr1
cavity.
Whilst the chemical series showed great potential for further optimization, an effective screening
cascade was required to direct the chemistry efforts. In the first wave of the screening cascade,
ACS Paragon Plus Environment
16

Page 17 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MCHr1 potency was assessed using a functional GTPγS assay. The first wave of basic data also
included parallel generation of data on physicochemical properties, hERG activity and
fundamental DMPK properties, e.g. metabolic stability and CYP450 inhibition. As in vitro
Caco2 permeability data did not provide sufficient guidance to funnel compounds through the
screening cascade, the ex vivo receptor occupancy assay in mice was introduced as a routine
screening model.27c Compounds that progressed into the receptor occupancy (RO) assay were
also investigated in a binding assay on mouse MCHr1 for correlation with the free RO IC50
values. Although in vitro permeability data did not differ substantially within the chemical series,
large variations in the degree of RO were observed. Free concentrations of compound in brain or
cerebrospinal fluid were not measured on a routine basis, but instead RO values were related to
unbound plasma concentrations to drive the chemistry effort. Mice were administered po a fixed
dose of compound and euthanized after different time points ranging from 15 minutes to 24
hours after dosing. The RO experimental model was used with a 10 day cycle time and the data
(ex vivo RO vs. unbound plasma concentration) were analyzed by fitting a simple Emax model to
the data, resulting in an estimate of apparent in vivo IC50. The RO data served two purposes. It
provided an understanding of the difference in the level of target engagement between different
compounds and guided dose setting in the longer in vivo disease model experiments, such as
weight loss measurements in chronic (3 weeks) treatment in dietꢀinduced obese (DIO) mice. The
receptor occupancy assay can be described as the gate keeper assay in the program, offering both
ranking between compounds and dose setting for longer in vivo studies, as well as providing
guidance for the design of molecules with improved properties.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The focus to move chemistry to a lower lipophilicity range prompted the investigation of
azetidines in the linker region. According to the 3D working hypothesis, the azetidines would
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 48
1
2
3
4
5
6
7
8
9
still adopt a planar distal eastern region and also position the basic N in a similar position to
compound 1. Indeed, the piperidine linker could be contracted to an azetidine, which combined a
general improvement in MCHr1 potency with lower lipophilicity. However, despite the shift
towards lower lipophilicity with the azetidine linker, there was not a general shift towards
decreased interactions with the hERG channel. Whilst there was no actual gain in the MCHr1 to
hERG ratio between matched piperidineꢀazetidine pairs, the focus of the chemistry efforts now
turned to the western region of the molecule and the lower baseline lipophilicity provided by
moving to an azetidine linker gave a greater degree of freedom for exploring the western region.
In addition, there was a better correlation between measured in vitro binding data and RO for the
azetidine linked compounds. Although the piperidine linkers have been revisited by matched pair
synthesis throughout the program, the focus of the chemistry effort was shifted towards the
azetidine linker. Pyrolidine linkers were also investigated as an alternative, but did not offer any
improvement over piperidine or azetidine linkers. In fact, both CYP inhibition and hERG
potency went in the wrong direction for both enantiomers of the pyrolidine rings. Functional
potency as measured in the MCHr1 GTPγs assay was in the same range as for the azetidines and
piperidines.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As suggested by the 3D working hypothesis, the western region of the molecules proved to be
the most tolerant to structural modifications and, hence, became the region to which the majority
of chemistry modification was dedicated. Initially, we were interested in reducing the size of the
western region and it was possible to replace the pyrrolidine ring with dimethyl amines, without
eroding MCHr1 potency (compound 77, Table 4). In a sense, this can be regarded as the
minimized lead structure. However, this did not result in an increased margin over hERG and
Table 4. Western modifications
ACS Paragon Plus Environment
18

Page 19 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
IC50
Caco
ABe
(Efflux
ratio)
IC50
MCH
IC50
RO,
MCH
bind
Mu
IC50
HLM
RLM
Structure
Compound
hERG logDc
(µM)
pKad
9.9
GTPγSa
(nM)
freeb
(nM)
Clintf
Clintf
(nM)
35
(0.37)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
76
77
78
46
51
79
80
81
34
29
6
4
14
ND
ND
ND
9
5.5
21
1.8
21
17
13
24
S
40
53
67
(0.29)
13
ND
ND
ND
15
14
9
1.55 8.9g
200
140
220
22
>33
>33
25
1.8
2.3
1.8
2.2
2
8.5
9.2
9.1
8.6
9.3
9.3
ND
145
28
49
(0.25)
31
(0.39)
S
62
(0.32)
15
42
22
S
218
149
54
54
(0.30)
23
1
10
62
(0.32)
21
5h
8.6
1.9
Functional potency.23 b RO receptor occupancy. IC50 relates to free plasma levels. LogD7.4
a
c
hplc Experimental pKa36 10ꢀ6 cm/s µl/min/mg calculated value Human plasma protein
binding used to calculate IC50 RO, free. ND Not determined S Stable to assay conditions (100%
parent compound remaining after incubation)
d
f
g
h
e
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 48
1
2
3
4
5
6
7
8
9
demethylation of the western amines became an additional metabolic route, primarily in rodents,
which made the evaluation of these compounds complex. Substitutions in the western phenyl
region were investigated for two reasons: to tune pKa and to investigate restrictions of this
flexible part of the molecule. We speculated that both of these effects individually or combined
could be a way to increase the margin between MCHr1 and hERG activity. A decrease in pKa
would reduce the affinity towards the hERG channel and conformational restrictions could
potentially increase the population of MCHr1 preferred conformations. Introduction of a
pyridine ring in the aromatic western region reduced pKa of the western amines, most notably in
78. The lower pKa also led to a reduced hERG activity, but also resulted in lower MCHr1
potency. With other changes in the western phenyl region, there was no clear correlation between
hERG potency and pKa, but a majority of compounds with a reduced pKa showed a lower hERG
activity. RO IC50 correlated well with measured in vitro binding IC50 in mice, indicating a lack of
CNS efflux. However, it is noteworthy that methylation of the western phenyl region in some
cases showed a marked increase in potency in the RO model, as compared to the in vitro binding
value. This effect was consistently observed in methylated western phenyls such as in 80. The
increase in potency could be translated to in vivo experiments and RO IC50 was used to set the
dose for these studies. The main hurdle with these structural modifications proved to be
combining metabolic stability with an increased margin to hERG activity. For example,
removing the methoxy substituent in the eastern region as in compound 81 resulted in an
increased metabolic stability, but this modification made the compounds more active on the
hERG channel.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cyclic amine substituents with lower pKa were then investigated (Table 5). Introduction of
piperazine rings showed lower levels of hERG interaction (e.g. as in compounds 83 and 84), but
ACS Paragon Plus Environment
20

Page 21 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
did not give the required MCHr1 potency. Although not consistent for all western piperazines,
metabolic stability worsened in some cases with these modifications. This effect was more
pronounced for morpholine substituents, as shown with compounds 85 and 86. The logD
increase observed with the lower pKa morpholines completely eroded metabolic stability.
Introduction of a hydrogen bond donor (HBD) in the western region in the form of hydroxy
substituted pyrolidines resulted in compounds with good MCHr1 potency, a lower level of hERG
interaction and acceptable metabolic stability (Table 5, compounds 87, 88, 89, 90 and 91).
However, despite Caco2 efflux values indicating a low degree of efflux, the introduction of an
HBD had a detrimental effect on RO. Attempts were made to mask the hydroxyl substituent by
geminal or vicinal methylation, but this was unsuccessful.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Even if cutꢀoff values for physicochemical properties such as logD or pKa, could not be firmly
Table 5. Western modifications: pKa modifications
IC50
IC50
MCH
IC50
RO,
Caco
ABe
(Efflux
ratio)
MCH
bind
Mu
IC50
hERG
(µM)
logDc
(ClogP)
HLM
Structure
Compound
pKad
8.5
8.4
GTPγSa
(nM)
freeb
(nM)
Clintf
(nM)
2.5
(2.0)
56
(0.29)
70
40
42g
25g
ND
#
15
S
82
83
F
O
N
2.2
(2.7)
46
(0.35)
O
N
N
>33
36
O
N
N
O
2.1
(2.8)
46
(0.35)
84
85
93
71
ND
ND
ND
>33
16
8.4
6.5
20
72
3.9
(1.5)
56
(0.29)
40g
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 48
1
2
3
O
O
4
5
6
7
8
9
3.3
(2.1)
51
(0.47)
16g
ND
39g
6
ND
9.8
31
7.5
8.9
8.9
9.0
9.0
8.9
332
S
O
N
N
86
87
88
89
90
91
27
50
26
22
23
25
N
O
O
N
O
1.9
(1.0)
31
(0.9)
>100
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
36
(0.63)
h
293
24
S
(1.0)
1.45
(1.9)
34
(0.26)
146
35
>33
26
S
1.6
(2.0)
42
(0.78)
9
2.8
19
2.6
(2.0)
34
(0.26)
39g
91
10
h
Functional potency.23 b RO receptor occupancy. IC50 relates to free plasma levels. c LogD7.4
a
hplc Experimental pKa36 e 10ꢀ6 cm/s f µl/min/mg g Human binding data h Human plasma protein
binding used to calculate IC50 RO, free. ND Not determined S Stable to assay conditions (100%
parent compound remaining after incubation) # No EC50 curve was generated for this compound
(83): 90% RO free @ 688 nM.
d
established, it was hypothesized that a logD below 2.5 should be targeted for metabolic stability
and that a pKa below 9 would combine acceptable MCHr1 potency with reduced hERG
interactions. Since introduction of a single HBD had a detrimental effect on RO pointing towards
efflux and leading to poor CNS exposure, an additional design criterion was established to not
incorporate an HBD. A set of compounds designed along these criteria are shown in Table 6. As
design parameters, clogP and pKa were considered separately, rather than calculating a
ACS Paragon Plus Environment
22

Page 23 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
composite logD. Certainly, a direct comparison between measured logD and the rather artificial
clogP was not expected to correlate perfectly, nevertheless measured lipophilicity was markedly
higher for the open oxetanes derivatives 95 and 96 despite good agreement between calculated
and measured pKa. In line with earlier observations, metabolic turnover numbers for 95 and 96
Table 6. Western modifications: spiro compounds and analogues
IC50
IC50
MCH
IC50
RO,
pKa
expd
(pKa
calc)
Caco
ABe
(Efflux
ratio)
MCH
bind
Mu
IC50
hERG
(µM)
logDc
(clogP)
HLM
Structure
Compound
GTPγSa
(nM)
freeb
(nM)
Clintf
(nM)
2.3
(2.5)
8.5
(7.8)
19
30
35
34
45
20
20
16
ND
9
ND
8.5
16
14
25
12
16
16
13
27h
21h
24h
51
110
117
37
74
14
2
94
95
90%
@ 0.5
nMg
4.1
(1.7)
8.0
(8.0)
4
8.5
(8.2)
96
14
11
15
7
ND
ND
(1.7)
O
N
O
N
N
2.4
(2.0)
8.1
(8.0)
51
(0.34)
O
O
97
N
O
35%
@
3.9
(2.3)
6.0
(7.9)
24h
98
5nMg
2.2
(2.1)
8.6
(8.6)
60
(0.3)
99
2
ND
2
2.1
(2.0)
8.6
(8.6)
7
22h
100
101
2.7
(2.3)
8.4
(7.7)
41
(0.3)
3
29
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 48
1
2
3
4
5
6
7
8
9
2.6
(2.2)
8.4
(7.7)
102
103
20
27
7
ND
4
8.9
22
31h
6
2.0
(2.1)
8.2
(8.0)
61
(0.46)
11
13
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1.8
(2.1)
8.0
(8.0)
79
(0.26)
104
23
11
15
28
S
Functional potency.23 b RO receptor occupancy. IC50 relates to free plasma levels. LogD7.4
a
c
d
f
g
37
Experimental pKa36 e 10ꢀ6 cm/s µl/min/mg Incomplete curves Calculated values.
h
hplc
Probability of efflux were low for all these compounds. S Stable to assay conditions (100%
parent compound remaining after incubation) ND Not determined
were high. The homomorpholine derivative 97 did offer an improvement in properties over
morpholines, but metabolic turnover rate was still too high. Our primary attention was turned to
spirocyclic amines as substituents in the western part of the molecule. Spirocyclic building
blocks have seen an increasing interest in the literature recently.38 For the 6ꢀoxaꢀ2ꢀ
azaspiro[3.4]octanes (99 and 100) and the 2ꢀoxaꢀ6ꢀazaspiro[3.3]heptanes (103 and 104), the
measured pKa values mirrored well the predicted values and, also, measured lipophilicity was in
agreement with clogP. It was pleasing to see the targeted physicochemical properties translated
into improved metabolic stability coupled with a wider MCHr1 potency margin over the hERG
channel. For other spirocycles, such as 2ꢀoxaꢀ5ꢀazaspiro[3.4]octane (98) and 5ꢀoxaꢀ2ꢀ
azaspiro[3.4]octanes (101 and 102) the measured physicochemical properties were more distant
from the predicted values. None of these compounds showed sufficiently improved metabolic
stability combined with an increased margin to hERG. It was also interesting to note that
structurally related open ring analogues such as 95 and 96 were markedly more lipophilic for a
ACS Paragon Plus Environment
24

Page 25 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
given pKa value and in none of these compounds could metabolic stability be combined with an
increased margin over hERG potency. Although the metabolic turnover rate was somewhat
higher for 99 and 103 than for the desꢀmethoxy analogues 100 and 104, the major metabolic
route was shifted from the spirocycle to demethylation of the methoxy substituent in the eastern
region of the molecules. It was speculated that metabolism in the strained spirocycle could be
associated with formation of reactive metabolites and it was considered advantageous to direct
metabolism to other parts of the molecule. A more detailed account of metabolite identification
and metabolism pathways in the strained spirocyclic compounds is in preparation and will be
reported. Turning to CNS properties, RO IC50 values for these spirocyclic amines suggested
good target coverage in the CNS. Mouse binding data in vitro confirmed the potency observed in
the RO assay, indicating a lack of efflux. In general, brain/plasma ratios were not measured in
this program. The majority of compounds tested in the receptor occupancy model showed good
agreement between in vivo and in vitro binding. Due to this correlation, in vitro potency could
confidently be used to select compounds for testing in the DIO mouse model. Large deviations
from the in vitroꢀin vivo correlation were attributed to efflux at the bloodꢀbrain barrier, presence
of active metabolites or slow offꢀrate from the receptor in vivo. For example, compounds 80 and
95 displayed a markedly higher potency in the RO assay. The major metabolic pathway for these
were demethylation of the eastern methoxy substituent leading to loss of MCHr1 potency, which
speaks against active metabolites and suggests slow offꢀrate of the parent compounds in vivo.
This shift towards higher potency was consistent for a small number of compounds, but this has
not been confirmed in binding kinetics in vitro.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Structural overlay of Xꢀray structure of 103 (green carbons) in comparison with the
low energy conformation of 1 (gray carbons) used as 3Dꢀhypothesis for SAR analysis.
At this stage, the Xꢀray structure of compound 103 was solved (Supplementary information).
Interestingly, this low energy conformation was in relatively good agreement with our 3D
hypothesis, described by an elongated structure with overlapping pharmacophore elements
(Figure 2). Although neither of these structures may be treated as the “true” bioactive
conformation, the 3D information served as a good basis for further idea generation and
interpretation of the SAR observed.
The results from the initial ex vivo receptor occupancy experiments were used to guide the dose
setting in 3 week DIO mice studies. For most compounds in the series, a bid dosing regime was
needed to keep the RO above a significant level for 24 hours. During the chemistry program, a
trough RO of 70% was aimed for as a working hypothesis. At the end of the study, brain MCHr1
RO was determined ex vivo by autoradiography using the RO screening protocol. RO
measurements at other time points were used together with plasma exposure profiles to model
the RO over time. On the basis of the overall promising profile of 103, this compound was
further evaluated in a repeated dose DIO study in mice over 21 days. Based on the RO data, three
different doses were selected aiming for a spread in RO profile. The aim was to vary the trough
RO that would translate into different degrees of weight loss. In the DIO study, 103 was found to
give a robust decrease in body weight when administered 50 or 20 ꢀmol/kg twice daily via oral
ACS Paragon Plus Environment
26

Page 27 of 48
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Effects of 103 on body weight in diet-induced obese mice.
Group housed female C57BL/6J mice were given free access to a cafeteria diet and randomized
into experimental groups based on cageꢀaverage body weights and dosed with 103 (5, 20 or 50
ꢀmol/kg bid at 8 am and 3 pm) or vehicle via oral gavage for 20 days. A) Plasmaꢀexposure
observations on the last study day and twoꢀcompartment pharmacokinetic model fit (ka=0.27 hꢀ1,
Cl=1.1 L/h/kg, V1=0.25 L/kg, Q=0.92 L/h/kg, V2=1.8 L/kg; relative bioavailability 1.7 for the 5
ꢀmol/kg group and 1 for the other groups). B) Receptorꢀoccupancy observations on the last study
day and fits of a receptorꢀkinetic model with elementary reactions that was driven by the plasma
concentration (KD=0.18 µM, koff =0.68 hꢀ1, Rtot =100). Effects on (C) body weight and (D)
percentage vehicle adjusted body weight change (n = 5ꢀ6 per group). Data represent mean ±
s.e.m. *P<0.05, **P<0.01, ***P<0.001 onwards vs. vehicle, one sided tꢀtests.
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 48
1
2
3
4
5
6
7
8
9
gavage, while 5 ꢀmol/kg twice daily had a minor effect compared to vehicle dosed DIO mice
(Figure 3). Thus, RO 24 hours after the first dose represents the trough value. For the two higher
doses of 103, trough RO levels were high enough to adequately inhibit MCHr1. It has been
established in other studies that 103 had no effect on food intake or body weight in DIO mice
lacking the MCHr1. These results confirmed that the observed weight loss was connected to the
MCHr1 mechanism and not an effect related to taste of the drug or some toxic effect. Thus, 103
binds to MCHr1 in the brain and produces a RO driven body weight loss in DIO mice. The RO is
driven in a predictable way by free plasma exposure, as 103 is not a Pgp substrate and shows no
signs of efflux.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Compound 103 was extensively profiled from a DMPK, off target and safety pharmacology
perspective. Additional pharmacokinetic data are summarized in Table 7. For this compound,
and for the series in general, there was a good correlation between in vitro and in vivo clearance.
No reversible or time dependent CYP450 inhibition was observed and in vitro safety
pharmacology from broad panel screening was considered to be low risk.39 There was no
evidence of reactive metabolite formation from GSH or KCN trapping studies in HLM.
Table 7. Pharmacokinetic properties of 103.
Hepatocyte
Ppbb
(% free)
Vss
(L/kg)
t
(h)
F
(%)
½
CLc
9
a
Species
Rat
CLint
10
36
26
2.0
5.4
NA
1.8
92
18
NA
5
37
1.7
NA
24
11
Dog
NA
Human
ACS Paragon Plus Environment
28

Page 29 of 48
Journal of Medicinal Chemistry
1
2
µl/min/106 cells b plasma protein binding c ml/min/kg NA Not applicable
3
4
a
5
6
7
8
9
Additionally, 103 was inactive in an MCHr2 binding assay. Cardiovascular safety studies in vivo
in either anaesthetized guinea pigs or dogs equipped with a telemetry system did not produce any
noticeable treatment related effects on blood pressure, heart rate or QT up to a free plasma
concentration of 3µM. Subsequently, compound 103 was evaluated in one month rat and dog
GLP safety studies without any treatment related findings up to a dose of 100 mg/kg. The
margins established in these studies together with the rest of the preclinical data package gave
support for progression of 10340 into clinical studies, but the study was terminated after the study
stopping criteria were reached.41 Results from these clinical trials, as well as detailed accounts of
human PK and the basis for dose setting in these studies will be disclosed in future publications.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusion
Lead optimization on a series of novel MCHr1 antagonists resulted in the discovery of the
clinical candidate 103. The lead series fits well into the MCHr1 pharmacophore model, but
occupies a markedly different physicochemical property space with reduced lipophilicity as
compared to other published MCHr1 antagonists. The starting point 1 displayed favourable
physicochemical properties and a clear focus throughout the chemistry program was to further
reduce lipophilicity whilst retaining CNS properties and MCHr1 RO in the brain.
The use of an ex vivo RO model in mice was the key to providing fast and high quality data. In
addition to ranking between compounds, the RO assay was also used to set doses for in vivo
disease model studies, as well as guiding subsequent rounds of compound design. The candidate
compound 103 has excellent DMPK properties and gave robust effects on weight loss when
ACS Paragon Plus Environment
29