Utilization of an Active Site Mutant Receptor for the Identification of Potent and Selective Atypical 5-HT2C Receptor Agonists

Article abstract of DOI:10.1021/acs.jmedchem.7b00385

Agonism of the 5-HT_2C receptor represents one of the most well-studied and clinically proven mechanisms for pharmacological weight reduction. Selectivity over the closely related 5-HT_2A and 5-HT_2B receptors is critical as their activation has been shown to lead to undesirable side effects and major safety concerns. In this communication, we report the development of a new screening paradigm that utilizes an active site mutant D134A (D3.32) 5-HT_2C receptor to identify atypical agonist structures. We additionally report the discovery and optimization of a novel class of nonbasic heterocyclic amide agonists of 5-HT_2C. SAR investigations around the screening hits provided a diverse set of potent agonists at 5-HT_2C with high selectivity over the related 5-HT_2A and 5-HT_2B receptor subtypes. Further optimization through replacement of the amide with a variety of five- and six-membered heterocycles led to the identification of 6-(1-ethyl-3-(quinolin-8-yl)-1H-pyrazol-5-yl)pyridazin-3-amine (69). Oral administration of 69 to rats reduced food intake in an ad libitum feeding model, which could be completely reversed by a selective 5-HT_2C antagonist.

Full text of DOI:10.1021/acs.jmedchem.7b00385

Subscriber access provided by EAST TENNESSEE STATE UNIV
Article
Utilization of an active site mutant receptor for the identification
of potent and selective atypical 5-HT2C receptor agonists
Joseph Carpenter, Ying Wang, Gang Wu, Jianxin Feng, Xiang-Yang Ye, Christian L.
Morales, Matthias Broekema, Karen A Rossi, Keith J. Miller, Brian J. Murphy, Ginger
Yingzhe Wu, Sarah E. Malmstrom, Anthony V. Azzara, Philip M. Sher, John M. Fevig,
Andrew Alt, Robert L. Bertekap, Mary Jane Cullen, Timothy M. Harper, Kimberly A. Foster,
Chiuwa Emily Luk, Qian Xiang, Mary F Grubb, Jeffrey A. Robl, and Dean A. Wacker
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00385 • Publication Date (Web): 21 Jun 2017
Downloaded from http://pubs.acs.org on June 22, 2017
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 94
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Utilization of an active site mutant receptor for the identification
of potent and selective atypical 5ꢀHT receptor agonists
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
C
a*
a*
a
a
Joseph Carpenter , Ying Wang , Gang Wu , Jianxin Feng , Xiang-Yang Ye, Christian L.
a
a
a
a
Morales , Matthias Broekema , Karen A. Rossi , Keith, J. Miller, Brian J. Murphy , Ginger Wu,
a
Sarah E. Malmstrom, Anthony V. Azzara , Philip M. Sher, John M. Fevig, Andrew Alt, Robert L.
a
a
a
Bertekap Jr. , Mary Jane Cullen, Timothy M. Harper, Kimberly Foster , Emily Luk , Qian
a
a
a*
Xiang , Mary F. Grubb , Jeffrey A. Robl, and Dean A. Wacker
a
Departments of Discovery Chemistry, Discovery Biology, Lead Evaluation, Computer-Assisted Drug Design,
Discovery Toxicology, and Pharmaceutical Candidate Optimization, Bristol-Myers Squibb, Pharmaceutical
Research Institute, P.O. Box 5400, Princeton, New Jersey 08543-5400
KEYWORDS: Obesity, serotonin, 5ꢀHT , 5ꢀHT , 5ꢀHT , agonist, heterocycle, weight loss,
2
C
2B
2A
safety, mutant receptor, allosteric
ABSTRACT: Agonism of the 5ꢀHT receptor represents one of the most wellꢀstudied and
2
C
clinicallyꢀproven mechanisms for pharmacological weight reduction. Selectivity over the closely
related 5ꢀHT and 5ꢀHT receptors is critical as their activation has been shown to lead to
2
A
2B
undesirable sideꢀeffects and major safety concerns. In this communication, we report the
development of a new screening paradigm which utilizes an active site mutant D134A (D3.32) 5ꢀ
HT receptor to identify atypical agonist structures. We additionally report the discovery and
2
C
optimization of a novel class of nonꢀbasic heterocyclic amide agonists of 5ꢀHT . SAR
2C
investigations around the screening hits provided a diverse set of potent agonists at 5ꢀHT with
2
C
1
ACS Paragon Plus Environment

Page 3 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
high selectivity over the related 5ꢀHT and 5ꢀHT receptor subtypes. Further optimization
2
A
2B
through replacement of the amide with a variety of 5ꢀ and 6ꢀmembered heterocycles led to the
identification of 6ꢀ(1ꢀethylꢀ3ꢀ(quinolinꢀ8ꢀyl)ꢀ1Hꢀpyrazolꢀ5ꢀyl)pyridazinꢀ3ꢀamine (69). Oral
administration of 69 to rats reduced food intake in an ad libitum feeding model, which could be
completely reversed by a selective 5ꢀHT antagonist.
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
C
Introduction. The everꢀincreasing prevalence of obesity represents a great burden on global
human health. As a major risk factor for type II diabetes, coronary heart disease, heart failure and
stroke, obesity and its coꢀmorbidities account for a growing number of worldwide deaths each
1
ꢀ3
year. In 2014, it was estimated that 13% of the world’s adult population were considered obese
4
and 42 million children under the age of 5 were obese in 2013. Conceptually, the prevention and
treatment of obesity can be accomplished through diet and exercise, but unfortunately, many
people are unable to attain significant and lasting reductions in body weight through these
methods alone. Drugꢀbased therapy may provide an opportunity for these patients to achieve
clinically beneficial weight loss and concomitant reductions in incidence of associated coꢀ
5
morbidities.
The neurotransmitter serotonin (5ꢀhydroxytryptamine or 5ꢀHT, Figure 1) is known to
play a critical role in the regulation of numerous neurological functions including mood, sexual
6, 7
desire and function, appetite, sleep, memory, learning, and temperature regulation. There are
14 serotonin receptor subtypes, all of which are G proteinꢀcoupled receptors (GPCR) with the
exception of the ligandꢀgated ion channel 5ꢀHT . Serotonin’s effects on appetite and feeding
3
have been shown to be largely controlled through the expression and activation of the 5ꢀHT2C
8
receptor. Lorcaserin (Figure 1) elicits its response through activation of the 5ꢀHT receptor,
2
C
which remains among the most wellꢀstudied and proven mechanisms for pharmacological weight
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
9, 10
reductions.
The 2012 FDA approval of lorcaserin was an important milestone for the
treatment of obesity, as it represents the first regulatory approval of a weight loss drug in over a
decade.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. 5ꢀHT agonist structures.
2
C
Despite lorcaserin’s approval, the search for highly selective 5ꢀHT agonists remains an
2
C
attractive area of research, as agonism of the closely related 5ꢀHT and 5ꢀHT receptors can
2
A
2B
1
1ꢀ16
lead to hallucinations and potentially fatal fibrotic cardiac valvulopathy, respectively.
The
discovery and development of novel agonist structures may enable greater efficacy with a
differentiated safety and tolerability profile through better selectivity versus the 2A and 2B
receptor subtypes. A major hindrance to developing selective agonists is the highlyꢀconserved
nature of the serotonin binding site across all receptor subtypes. Aromatic amino acids (Phe, Tyr,
Trp) form what has been coined the aromatic box, which interact with the indole portion of
serotonin while a conserved Asp residue serves to anchor the agonist through formation of a tight
1
7
saltꢀbridge with serotonin’s basic nitrogen. It should not be surprising then, that typical 5ꢀHT
2C
agonists (Figure 1) consist of a hydrophobic aryl or heteroaryl ring system tethered to a basic
18ꢀ24
nitrogen that is capable of mimicking the interactions observed with serotonin.
These
3
ACS Paragon Plus Environment

Page 5 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
common structural features are typically considered indispensable for active site binding and
receptor activation, and likely contribute to poor receptor subtype selectivity. In addition to the
adverse events such as valvular hypertrophy and CNSꢀbased events mentioned previously, the
amphiphilic nature of these common structures has also been implicated in various offꢀtarget
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
5ꢀ28
toxicities such as ion channel inhibition and phospholipidosis.
Allosteric modulation of GPCR’s is emerging as a promising new strategy for the
identification of novel chemical entities for the treatment of a variety of central nervous system
2
9
disorders. In 2003, Dinh and coworkers reported on the discovery of a positive allosteric
modulator (PAM) of 5ꢀHT which enhanced both the binding and functional potency of
2
C
serotonin in several mammalian cell lines with no analogous effect on the related 2A and 2B
3
0
receptors. In principle, a positive allosteric modulator of 5ꢀHT could act to enhance the in
2
C
vivo activity of endogenous serotonin which is known to be elevated in the fed state, and could
lead to enhanced satiety, an overall reduction of food intake and concomitant weight reduction.
Moreover, because allosteric compounds would not bind to the serotonin site, they would likely
have a structure quite divergent from that of classical 5ꢀHT agonists. In turn, we reasoned that
2
C
these atypical structures could provide an opportunity to not only obtain high selectivity over
closely related serotonin subtypes, but also avoid common aforementioned toxicities associated
with the typical 5ꢀHT pharmacophore.
2
C
Lead Identification. Given the possibilities presented by an allosteric modulator, we embarked
on a journey to identify novel positive allosteric modulators of 5ꢀHT . In order to realize this
2
C
outcome, we screened the BMS compound library in PAM mode using serotonin as the agonist.
Unfortunately, no positive allosteric modulators were found, but we did identify nearly 50,000
compounds that exhibited 5ꢀHT agonism. Although our original goal of identifying a positive
2
C
4
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
allosteric modulator was not achieved, we were intrigued by the number of atypical structures
that did not conform to the classic 5ꢀHT agonist pharmacophore. From this set of screening
2C
hits, we considered the possibility of identifying a completely unique series of compounds that
could allow for a deviation from the known selectivity and toxicity concerns that plague the
field. After eliminating all of the known 5ꢀHT agonists and any additional compounds with
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
C
high structural similarity to known 5ꢀHT agonists (i.e., compounds that contained an aromatic
2
C
ring with a 2ꢀ3 atom tether to a basic amine moiety), we were left with approximately 18,000
compounds that did not fit the typical basic amine containing pharmacophore. From this set, 68
unique highꢀpotency compounds with differentiated chemical structures were chosen for
purification and retesting. To our disappointment, none of the compounds retained any 5ꢀHT2C
potency following purification, which was likely attributed to removal of contaminants in the
deck sample. Four representative examples of screening hits are given in Figure 2. As indicated
in red, these compounds possess embedded substructures that are wellꢀknown 5ꢀHT agonists
2C
that could be remnants of synthetic impurities or liberated during deterioration of the parent
molecule upon longꢀterm storage in DMSO. Although these results were discouraging, we still
wondered if within the remaining screening hits there were compounds where the 5ꢀHT2C
potency was attributed to the parent molecule and not a possible impurity or degradant.
5
ACS Paragon Plus Environment

Page 7 of 94
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 2. Screening hits with embedded 5ꢀHT agonist subꢀstructures indicated in red.
2
C
Purification and retesting of 18,000 compounds was not feasible, so an alternative
approach to evaluating these hits was needed. We reasoned that this selection process could be
accomplished by siteꢀselective mutagenesis of the 5ꢀHT receptor active site. Our hypothesis
2C
was that while the residues that form the aromatic box are necessary for the receptor’s agonist
activity, the active site aspartic acid (Asp 134, D3.32) was only important for amine binding and
could be mutated without impacting receptor function. This mutant receptor should not respond
to compounds whose activity was driven by amine binding eliminating any samples
contaminated with classical 5ꢀHT agonist impurities. This approach was not without significant
2
C
risk as mutation of Asp 134 could lead to a misfolded protein, or a functionally inactive receptor,
even if the protein was able to adopt the proper tertiary structure. The desired D134A mutant was
prepared and found to have no functional activity in the presence of serotonin or several
additional known 5ꢀHT agonists. However, upon screening against our compound library
2
C
several hits were obtained which were agonists at both the wildꢀtype and mutant receptor, but
they were not active in a parental cell line that lacks the 5ꢀHT receptor. The screening hits were
2C
6
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
also tested for PAM activity in the presence of serotonin and were found to have no impact on
serotonin’s binding or functional response.
Among the high throughput screening hits identified, two thiazoleꢀbased leads 1 and 2
were found to give promising 5ꢀHT agonism in both the wildꢀtype and D134A mutant
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
C
receptors, while being completely devoid of functional activity at the 5ꢀHT and 5ꢀHT
2B
2
A
receptors (Figure 3). Furthermore, these leads could be divided into three simple fragments (aryl
ring, heterocyclic core, and carboxamide) that should allow for rapid SAR evaluation of each
segment. These features, together with their low molecular weight and polar surface area
represented an attractive starting point for the optimization of a novel chemical series for brainꢀ
penetrant 5ꢀHT agonists.
2C
Figure 3. Novel thiazoleꢀbased leads from HTS.
Initial investigations focused on amide SAR of 1, as we were wary of the hydrophobic
aryl ring system coupled with a basic amine functionality present in 2 which was reminiscent of
the traditional 5ꢀHT agonist pharmacophore we hoped to avoid due to potential offꢀtarget
2
C
toxicity. For thiazole compound 1, over 100 amide variations were prepared (alkyl, heteroalkyl,
aryl, heteroaryl) resulting in mostly inactive compounds with only a few close analogs showing
similar levels of potency or receptor activation (Table 1, 3ꢀ6). Similarly, attempts to modify the
dichlorophenyl moiety with differentially substituted monoꢀ and bicyclic aryl and heteroaryls led
7
ACS Paragon Plus Environment

Page 9 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
to compounds with diminished 5ꢀHT potency (7ꢀ13). We next turned our efforts to the
2
C
evaluation of the thiazole core (14ꢀ17). Morpholine amides of several isomeric and methyl
substituted thiazoles were prepared and although most of the core replacements resulted in
significant reductions in potency, this exercise led to identification of compound 17 with
promising 5ꢀHT potency of 28 nM and full intrinsic activity. Unfortunately, as we had seen
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
C
with modifications to 1, we were again met with disappointment as the improved 5ꢀHT2C
potency of 17 did not translate broadly to more potent derivatives upon further modification
(data not shown). Variation of the amide revealed narrow SAR, and given an in vitro human
halfꢀlife of less than two minutes, this thiazole isomer series was not progressed.
Table 1. Modifications of lead thiazole 1.
5ꢀHT2C
Compound
1
Structure
a
b
EC50 (nM)
IA
119/128
3100/3700
320/733
679/840
129/315
0.8
0.3
0.5
0.7
0.8
3
4
5
6
8
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
8
9
4600 ± 1300
0.1
ꢀꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
>10000
1272/2402
>10000
0.3
ꢀꢀ
1
0
1
2
3
4
1
1
1
1
501/705
>10000
0.7
ꢀꢀ
766/1168
>10000
0.6
ꢀꢀ
15
16
>10000
ꢀꢀ
1520 ± 634
0.6
9
ACS Paragon Plus Environment

Page 11 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
7
18/38
0.8
a
EC50 values were calculated from doseꢀresponse curves. Functional screenings were carried out in HEK293E
cells expressing the human 5ꢀHT2C receptor. EC50 determination experiments were performed at least in triplicate
and the values are presented as Mean ± SD unless indicated otherwise or by inclusion of two values. Positive control
was mCPP which gave 5ꢀHT , EC = 15 ± 4 nM. All compounds were functionally inactive (EC50 > 10 ꢀM) at 5ꢀ
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
C
50
b
HT2A and 5ꢀHT . Intrinsic activity (IA) for all compounds as compared to serotonin at 3 µM (defined as IA = 1).
2B
After our unsuccessful attempt to identify a tractable lead from 1 we shifted our efforts to
modification of screening hit 2 (Table 2). We were pleased to find that the morpholine amide
analog 18 retained the 5ꢀHT potency of 2 without the basic amine functionality we had hoped
2
C
to avoid. In an attempt to gain potency, we inserted the thiazole core from 17 to give 19, but
despite the apparent similarities of leads 1 and 2, the SAR was found to diverge and this
substitution resulted in a loss of potency. This variance was also exemplified by 13 (Table 1), as
incorporation of the 1ꢀnaphthyl core from screening hit 2 into 1 resulted in a nearly 8ꢀfold loss of
potency. Removal of the 4ꢀmethyl substituent (20, Table 2), or swapping the location of nitrogen
and sulfur (21) were also detrimental to 5ꢀHT potency, and thus 18 was chosen for further
2
C
optimization.
Table 2. Modification of thiazole lead 2.
5
ꢀHT2C
Compound
Structure
a
b
EC50 (nM) IA
2
43/123
0.8
1
1
8
9
95/115
0.8
0.6
729/1623
10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
20
21
1522/2271
0.6
0.3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2780 ± 788
a
EC50 values were calculated from doseꢀresponse curves. Functional screenings were carried out in HEK293E cells
expressing the human 5ꢀHT , 5ꢀHT , or 5ꢀHT receptor. EC50 determination experiments were performed at least
in triplicate and the values are presented as Mean ± SD unless indicated otherwise or by inclusion of two values.
Positive control was mCPP which gave 5ꢀHT , EC = 15 ± 4 nM. All compounds were functionally inactive (EC50
2A
2B
2C
2C
50
b
>
(
10 ꢀM) at 5ꢀHT2A and 5ꢀHT2B. Intrinsic activity (IA) for all compounds as compared to serotonin at 3 µM
defined as IA = 1).
Chemistry. A general synthetic approach to analogs of thiazole lead 2 is outlined in Scheme 1 as
exemplified by three key compounds 31, 41, and 47. For thiazole 31, synthesis commenced with
condensation of naphthyleneꢀ1ꢀthiocarboxamide (A) with ethyl 2ꢀchloroꢀ3ꢀoxobutanoate (B) to
give the thiazole ester intermediate C. The carboxylic acid, obtained via saponification,
underwent facile amide coupling with piperidinꢀ4ꢀylmethanol to give thiazole amide 31 in the
threeꢀstep process depicted in Scheme 1 line (i). Pyrrole 41 (line ii) was prepared through an
initial Suzuki coupling between naphthaleneꢀ1ꢀboronic acid (D) and commercially available or
readily synthesized ethyl 5ꢀbromoꢀ2ꢀmethylꢀ1Hꢀpyrroleꢀ3ꢀcarboxylate (E), followed by
saponification of F and amide coupling to yield the final product. For the synthesis of pyrazole
47 (line iii) we employed a condensation of ethyl 4ꢀ(naphthalenꢀ1ꢀyl)ꢀ2,4ꢀdioxobutanoate (G)
with hydrazine and subsequent alkylation with iodomethane to give the Nꢀmethyl pyrazole ester
intermediate (H). As with previous examples, final compound 47 was readily obtained by
hydrolysis and amide coupling.
As described in line (iv) commercially available 1ꢀethylꢀ3ꢀnitroꢀ1Hꢀpyrazole (I) was converted
to the corresponding pinacol boronic ester prior to Suzuki coupling with bisꢀBoc protected 6ꢀ
chloropyridazinꢀ3ꢀamine (J). During optimization of this coupling, we found that the more
11
ACS Paragon Plus Environment

Page 13 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reactive 1,1’ꢀbis(diꢀtertꢀbutylphosphino)ferrocene (dtbpf) ligand was vastly superior to the more
commonly used 1,1’ꢀbis(diphenylphosphino)ferrocene (dppf) ligand at forging the biaryl union.
Reduction of the nitro group and conversion of the resultant amino pyrazole (K) to the
corresponding bromo pyrazole under Sandmeyer conditions set up a penultimate Suzuki
coupling with quinolineꢀ8ꢀboronic acid (L). Deprotection of the Boc protecting groups was
facilitated with TFA to give compound 69.
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 1. Synthesis of key compounds 31, 41, 47, and 69.
O
O
S
O
O
S
N
b, c
S
N
(i)
a
O
N
NH2
O
OH
Cl
A
B
C
31
O
O
O
B(OH)2
Br
O
O
N
d, e
b, c
(ii)
OH
N
N
N
H
H
H
41
D
E
F
O
O
O
O
O
f, g
O
b, c
N
(iii)
O
OH
N
N
N
N
G
H
47
N(Boc)2
NH2
O N
2
N(Boc)2
B(OH)2
(iv)
N
N
H2N
N
N
N
h, i, j
k, e, l
N
N
N
N
N
N N
N
Cl
J
K
L
N
I
6
9
Reagents: (a) EtOH, 80 °C, sealed tube. (b) LiOH, THF/H O/MeOH, rt. (c) Piperidinꢀ4ꢀylmethanol, T P, CH Cl ,
2
3
2
2
Et N, rt. (d) NBS, DMF, rt. (e) PdCl (dppf) , Cs CO , dioxane, 90ꢀ100 °C, sealed tube. (f) Hydrazine, AcOH, 90 °C.
3
2
2
2
3
(
g) CH I, Cs CO , MeCN, rt. (h) LDA, isopropyl pinacol borate, −78 °C to rt, THF. (i) PdCl (dtbpf) , Cs CO ,
3
2
3
2
2
2
3
dioxane, 80 °C, sealed tube. (j) H , Pd/C, MeOH, rt. (k) tBuONO, Cu(I)Br, LiBr, acetonitrile, 50 °C. (l) TFA, rt.
2
Lead Optimization. To our satisfaction, we discovered that further SAR investigation of 18
gave rise to a diverse set of compounds that exhibited appreciable 5ꢀHT potency while being
2
C
completely devoid of functional activity at 5ꢀHT and 5ꢀHT (Table 3). Although the primary
2
A
2B
12
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
carboxamide 22 was inactive, we found that with the appropriate amide substitution we were
able to retain and even improve the fullꢀagonist activity and potency of 18. For example, simple
acyclic and cyclic aliphatic amides 23ꢀ24 were functionally active with an EC as low as 24 nM
50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
in the case of piperidine amide 24. Further amide SAR revealed that tertiary and cyclic amides
were preferred, as secondary amides typically lacked potency or gave only partial 5ꢀHT2C
agonism (25 vs. 26). Results were also sensitive to steric bulk as exemplified by the complete
loss of potency for 28 vs. 27. Although we were pleased with the potency attainable in this
series, most compounds tested suffered from poor metabolic stability when incubated with
human and rat liver microsomes. In an effort to improve metabolic stability, heteroatoms were
introduced into the amide chain with mixed results (25ꢀ32). Hydroxylation of the amide side
chains (27, 29ꢀ31) not only improved microsomal stability, but also retained or increased 5ꢀHT2C
potency, with the hydroxymethyl piperidine amide 31 being optimal. We were somewhat
surprised to find the decrease in potency for piperizine amide 32 given the presence of a basic
secondary amine and structural similarity to traditional 5ꢀHT agonists. Importantly, this may
2
C
suggest that 32 acts through an alternative mode of binding and receptor activation than that of
serotonin. Nitrogenꢀcontaining heterocycles 33ꢀ36 were synthesized with the intent of further
improving microsomal stability, and although several were wellꢀtolerated, they did not improve
stability or provide a clear advantage relative to the nonꢀheterocycles in terms of potency.
Table 3. Amide SAR: 5ꢀHT EC , and microsomal stability.
2
C
50
5
ꢀHT2C
Metstab:
Compound
R
Human/Rat
a
b
EC50 (nM)
IA
c
%
Remaining
13
ACS Paragon Plus Environment

Page 15 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
2
8
2
95/115
0.8
50/6
>10000
59/72
ꢀꢀ
ND
ND
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
23
1.0
2
2
2
2
2
4
5
6
7
8
24 ± 6
14 ± 1
>10000
30/50
1.0
0.9
ꢀꢀ
24/1
28/7
ND
0.9
ꢀꢀ
67/22
28/7
>10000
2
9
47 ± 18
17/25
0.9
0.9
84/34
50/8
N
30
OH
3
3
3
1
2
3
6 ± 3
1040 ± 1320
13/19
1.1
0.4
1.0
80/24
ND
ND
3
4
68 ± 38
38/39
1.0
0.7
70/15
<1/1
35
14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
6
12/15
0.8
5/4
a
EC50 values were calculated from doseꢀresponse curves. Functional screenings were carried out in HEK293E
cells expressing the human 5ꢀHT , 5ꢀHT , or 5ꢀHT receptor. EC50 determination experiments were performed at
least in triplicate and the values are presented as Mean ± SD unless indicated otherwise or by inclusion of two
values. Positive control was mCPP which gave 5ꢀHT , EC = 15 ± 4 nM. All compounds were functionally
inactive (EC50 > 10 ꢀM) at 5ꢀHT and 5ꢀHT . Intrinsic activity (IA) for all compounds as compared to serotonin
at 3 µM (defined as IA = 1). Compounds were incubated with human or rat liver microsomes at 5 mM. Values
represent % parent remaining after 10 minutes. ND = not determined.
2
A
2B
2C
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2C
50
b
2A
2B
c
At this juncture, we chose to evaluate 5ꢀmembered heterocyclic replacements for the
thiazole core with the hope of identifying a molecule that would retain 5ꢀHT potency and
2
C
selectivity while further improving metabolic stability. Holding the naphthyl ring system and
hydroxymethyl piperidine amide of 31 constant, we systematically modified the core heterocycle
structure (Table 4). Our anticipation, based on previous observations with different thiazole
isomers, was that changing the effective ring size and geometry by moving or replacing the
sulfur atom would be detrimental to 5ꢀHT activity, but again the SAR was mostly
2
C
unpredictable. Cores containing one heteroatom were generally tolerated, but as observed with
thiazoles, the potency could vary dramatically between isomeric structures. Both thiophene
isomers 37 and 38 retained 5ꢀHT potency (20 nM, 83 nM, respectively), but gave no
2
C
improvement in terms of microsomal stability. However, furan 39 was >250 fold less potent than
isomer 31. Pyrrole 41 (12 nM) was the most potent single nitrogenꢀcontaining heterocycle with
isomers 40 and 42 losing 3ꢀ30 fold potency at 5ꢀHT . Of the cores containing multiple
2
C
heteroatoms, oxazoles, imidazoles, triazoles and tetrazoles (43-46, 49-50) were not ideal giving
~16ꢀ700 fold loss in potency relative to thiazole 31. Pyrazoles 47 and 48 on the other hand were
potent (6, 16 nM), and as a result of this exercise we had several potent, selective 5ꢀHT2C
agonists within this series that we chose to advance for further in vitro and in vivo profiling.
Table 4. Core replacement of thiazole 31.
15
ACS Paragon Plus Environment

Page 17 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
ꢀHT2C
5ꢀHT
5ꢀHT
Metstab
Human/Rat
% Remaining
2B
2A
a
a
Compound
31
Core
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
a
b
EC50
nM)
b
EC50
(nM)
b
EC50 (nM)
IA
IA
IA
c
(
6.0 ± 3
1.0
1.0
0.8
>10000
>10000
>10000
ꢀꢀ
>10000
ꢀꢀ
80/24
87/20
66/23
3
3
7
8
20 ± 14
47/119
ꢀꢀ
ꢀꢀ
>5000
ꢀꢀ
ꢀꢀ
>10000
39
1032/2000
41/45
0.7
1.0
>10000
>10000
ꢀꢀ
ꢀꢀ
>10000
>10000
ꢀꢀ
ꢀꢀ
ꢀꢀ
4
4
4
0
1
2
76/57
12 ± 7
0.9
>10000
ꢀꢀ
>10000
ꢀꢀ
76/44
369/433
63/140
0.7
0.9
>10000
>10000
ꢀꢀ
ꢀꢀ
>10000
>10000
ꢀꢀ
ꢀꢀ
67/18
33/24
43
4
4
4
5
381/575
1.0
0.7
>10000
>10000
ꢀꢀ
ꢀꢀ
>10000
>10000
ꢀꢀ
ꢀꢀ
62/16
ꢀꢀ
1804/3525
4
4
6
7
1870 ± 422
6 ± 4
0.4
1.0
>10000
>10000
ꢀꢀ
ꢀꢀ
>10000
ꢀꢀ
ꢀꢀ
d
75 ± 62
0.1
92/37
16
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
4
5
8
9
0
16 ± 9
334/385
1.0
1.0
0.3
>10000
>10000
>10000
ꢀꢀ
ꢀꢀ
ꢀꢀ
>10000
>10000
>10000
ꢀꢀ
ꢀꢀ
ꢀꢀ
88/51
58/24
ꢀꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4100/>10000
a
EC50 values were calculated from doseꢀresponse curves. Functional screenings were carried out in HEK293E cells
expressing the human 5ꢀHT , 5ꢀHT , or 5ꢀHT receptor. EC50 determination experiments were performed at least
in triplicate and the values are presented as Mean ± SD unless indicated otherwise or by inclusion of two values.
Positive control was mCPP which gave 5ꢀHT , EC = 15 ± 4 nM; 5ꢀHT , EC = 287 ± 94 nM; 5ꢀHT , EC
290 ± 110 nM. Intrinsic activity for all compounds as compared to serotonin at 3 µM (defined as 1). Compounds
were incubated with human or rat liver microsomes at 5mM. Values represent % parent remaining after 10 minutes.
EC50 values represent the average of multiple test occasions, several of which indicated no functional 5ꢀHT2A
2A
2B
2C
=
50
2C
50
2B
50
2A
b
c
d
activity. ND = not determined.
Given the promising metabolic stability in human liver microsomes and the higher overall
plasma and brain exposures observed for compound 47 (Table 7), the pyrazole core represented
an attractive lead for additional SAR investigations. With the intent of further increasing
lipophilicity via changes to the polar amide moiety, we prepared a set of heterocyclic amide
isosteres. Various 5ꢀ or 6ꢀmembered heteroaryl amide replacements were investigated (Table 5).
Nꢀmethyl imidazole 51 (73 nM) was a promising initial result, and we found that by changing
the core pyrazole alkyl substituent to ethyl to give 52, we gained approximately twoꢀfold potency
at 5ꢀHT (32 nM). Unfortunately these compounds suffered from poor microsomal metabolic
2
C
stability. Introduction of the isomeric imidazole 53 or pyrazole 54 resulted in a substantial loss
of functional activity and potency (4120 nM, 3500 nM respectively), which was only partially
restored in the case of triazole 55 or tetrazole 56 (400 nM, 450 nM respectively). With respect to
sixꢀmembered heterocyclic amide replacements, pyridine 57 and pyrimidine 58 lacked sufficient
5
ꢀHT potency for advancement, whereas pyridazines 59 and 60 exhibited good potency,
2C
intrinsic activity and selectivity for 5ꢀHT .
2
C
17
ACS Paragon Plus Environment

Page 19 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 5. Amide Isostere SAR: 5ꢀHT EC , and microsomal stability.
2
C
50
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
5
ꢀHT2C
5ꢀHT2B
5ꢀHT2A
Metstab:
Human
a
Compound
R1
R2
a
b
EC50
b
a
b
EC50 (nM)
IA
IA
EC50 (nM)
IA
%
(nM)
c
Remaining
5
5
1
2
Me
88/58
0.9
0.8
>10000
ꢀꢀ
93, 126
0.2
ꢀꢀ
15
5
Et
Et
37/27
>10000
>10000
ꢀꢀ
ꢀꢀ
>10000
>10000
5
3
3850/4391
0.2
ꢀꢀ
ND
5
5
4
5
Et
Et
3460/>5000
190/616
0.3
0.7
>10000
>10000
ꢀꢀ
ꢀꢀ
>10000
>10000
ꢀꢀ
ꢀꢀ
ND
4
5
6
Et
450 ± 357
0.6
>5000
ꢀꢀ
1800, >2500
0.2
4
5
5
7
8
Et
Et
576/613
0.8
0.4
>5000
>5000
ꢀꢀ
ꢀꢀ
>5000
>5000
ꢀꢀ
ꢀꢀ
ND
ND
1697/1903
5
6
9
Me
Et
40/25
1.2
1.2
>10000
>10000
ꢀꢀ
ꢀꢀ
>5000
ꢀꢀ
12
17
d
0
15 ± 7
884 ± 820
0.2
a
EC50 values were calculated from doseꢀresponse curves. Functional screenings were carried out in HEK293E
cells expressing the human 5ꢀHT , 5ꢀHT , or 5ꢀHT receptor. EC50 determination experiments were performed at
2A
2B
2C
18
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
least in triplicate and the values are presented as Mean ± SD unless indicated otherwise or by inclusion of two
values. Positive control was mCPP which gave 5ꢀHT , EC = 15 ± 4 nM; 5ꢀHT , EC = 287 ± 94 nM; 5ꢀHT ,
2C
50
2B
50
2A
b
c
EC50 = 290 ± 110 nM. Intrinsic activity for all compounds as compared to serotonin at 3 µM (defined as 1).
Compounds were incubated with human or rat liver microsomes at 5mM. Values represent % parent remaining after
d
1
5
0 minutes. EC50 values represent the average of multiple test occasions, several of which indicated no functional
ꢀHT2A activity. ND = not determined.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Having established that the pyridazine ring was optimal for 5ꢀHT potency and activity, detailed
2
C
metabolite profiling of 60 revealed the naphthalene ring to be particularly susceptible to
oxidative metabolism, and thus further SAR exploration focused on its stabilization via
incorporation of a nitrogen atom into the ring system (Table 6). To this end, we prepared a series
of compounds with a single nitrogen atom in each position around the naphthalene ring with the
intent of blocking the major site of oxidative degradation. While analogs 61ꢀ63 exhibited ~20ꢀ30
fold diminished 5ꢀHT potency, and analogs 64ꢀ65 were >100 fold less potent, we were pleased
2C
to find that isoquinoline 66 at 28 nM was comparable to 60. Moreover, quinoline 67 gave a
marked improvement in 5ꢀHT potency (2 nM) but remained highly prone to oxidative
2
C
metabolism. We again turned to metabolite identification to determine what region of 67 was
subject to oxidative degradation and found that the liability had now shifted to the pyridazine.
Introduction of an amino substituent at the 3ꢀposition on the pyridazine was found to partially
mitigate its metabolism and retain or even improve 5ꢀHT potency for compounds 68 and 69
2
C
relative to the unsubstituted analogs (7 nM, 79% remaining, 5 nM, 61% remaining
3
1
respectively).
Table 6. Amide Isostere SAR: 5ꢀHT , 5ꢀHT , 5ꢀHT EC , and microsomal stability.
2
C
2B
2A
50
Compound
R1
R2
5ꢀHT2C
5ꢀHT2B
5ꢀHT2A
Metstab:
19
ACS Paragon Plus Environment

Page 21 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Human
a
a
b
EC50
b
a
b
EC50 (nM) IA
IA
EC50 (nM)
IA
%
(nM)
c
Remaining
6
1
H
H
291/518
0.8
>10000
>10000
ꢀꢀ
>10000
>10000
ꢀꢀ
ND
ND
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
62
530 ± 144
0.7
0.6
ꢀꢀ
ꢀꢀ
ꢀꢀ
ꢀꢀ
6
3
H
H
315/396
>10000
>10000
>10000
>10000
ND
ND
N
64
4528/4731
0.3
ꢀꢀ
ꢀꢀ
N
6
6
6
5
6
7
H
H
1600 ± 804 0.7
>5000
ꢀꢀ
ꢀꢀ
>5000
ꢀꢀ
ꢀꢀ
ND
16
8
13/43
2 ± 0
0.7
1.0
1.1
0.8
>10000
>10000
d
d
H
6 ± 3
0.2
ꢀꢀ
151 ± 58
0.2
ꢀꢀ
e
68
NH2
NH2
7
>5000
>10000
79
61
d
d
6
9
5 ± 3
48 ± 41
0.4
635 ± 862
0.2
a
EC50 values were calculated from doseꢀresponse curves. Functional screenings were carried out in HEK293E
cells expressing the human 5ꢀHT , 5ꢀHT , or 5ꢀHT receptor. EC50 determination experiments were performed at
2A
2B
2C
least in triplicate and the values are presented as Mean ± SD unless indicated otherwise or by inclusion of two
values. Positive control was mCPP which gave 5ꢀHT , EC = 15 ± 4 nM; 5ꢀHT , EC = 287 ± 94 nM; 5ꢀHT ,
2C
50
2B
50
2A
b
c
EC50 = 290 ± 110 nM. Intrinsic activity for all compounds as compared to serotonin at 3 µM (defined as 1).
Compounds were incubated with human or rat liver microsomes at 5mM. Values represent % parent remaining after
10 minutes. EC values represent the average of multiple test occasions, several of which indicated no functional
d
5
0
e
5ꢀHT2A activity. EC value represents results from a single experiment. ND = not determined.
50
Pharmacokinetics. Compounds 31, 41, 47 and 69 were selected for further in vitro profiling and
rat pharmacokinetic studies, the data for which is given in Table 7. These studies were designed
20
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
to assess the permeability of each compound along with relative exposure of the compounds in
the brain versus the plasma. In general, the compounds had excellent pH independent
permeability as measured by a parallel artificial membrane permeability assay (PAMPA), in
addition to Cacoꢀ2 ratios that suggested good permeability with a low probability of transporter
mediated efflux. In a preliminary experiment, compound 31 was administered orally (P.O.) to
rats at a dose of 30 mg/kg and the brain and plasma concentrations were determined 7 hours post
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
administration. Given the low in vitro rat microsomal stability of 31 (T = 6 min), we were not
½
surprised to observe low exposures (736 nM plasma, 35 nM brain, b/p = 0.05) with a short halfꢀ
life. In an attempt to circumvent degradation of the compound through oxidative firstꢀpass
metabolism, compound 31 was subsequently administered intravenously (I.V.) at 1 mg/kg and its
brain and plasma concentrations were determined at 1 hour post dose. Despite the modified
dosing protocol, low plasma and brain concentrations were again observed for 31 (569 nM, 76
nM respectively, b/p = 0.15). Taken together, the low brain to plasma concentrations were
surprising considering the high PAMPA permeability and lack of Cacoꢀ2 efflux. In an attempt to
understand possible reasons for low brain exposure, several follow up studies were conducted
utilizing Pꢀglycoprotein and breast cancer resistance protein (Pgp, BCRP) transporter knockout
mice and specially engineered cellular systems that overexpress these membrane transporters.
These experiments ultimately revealed that compound 31 was a substrate of rodent Pgp and
BCRP transporters, but exhibited no significant in vitro affinity for human Pgp or BCRP
transporters.
Given to our lack of high throughput rodent Pgp and BCRP assays, and with no clear
advantage to I.V. administration, we chose to move forward with oral administration of the
remaining compounds as the means to determine if high rodent brain exposure would be feasible
21
ACS Paragon Plus Environment

Page 23 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
within our current series of heterocyclic 5ꢀHT agonists. Compounds 41 and 47 were
2
C
subsequently administered to rats according to the original oral dosing regimen (30 mg/kg P.O.)
and, as with compound 31, we observed reasonable plasma concentrations for compounds 41 and
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
4
7, but relative brain penetration was again low (b/p = 0.3, 0.2 respectively). On the other hand,
when compound 69 was administered orally at 30 mg/kg, it exhibited higher plasma and brain
exposures (7.5 ꢀM, 3.9 ꢀM respectively) with an increased brain to plasma ratio (0.9) relative to
compounds 31, 41, and 47.
Table 7. In vitro and in vivo profiling of selected 5ꢀHT agonists.
2
C
a, b
Cacoꢀ2
nm/s)
Efflux ratio
BA/AB)
Protein
binding
(% bound)
human/rat
Exposure (nM)
PAMPA
(nm/s)
Microsomal
(
B/Pa
Compound
T ½ (min)
human/rat
ratio
Plasma
Brain
pH 5.5/7.4
607/739
624/492
588/613
727/889
(
c
c
3
4
4
6
1
1
7
9
0.5 (175/329)
1.3 (219/172)
0.5 (168/323)
0.7 (528/777)
99.5/99.6
98.3/92.5
97.7/92.4
93.8/92.1
21/6
41/11
36/8
736
35
0.05
0.2
d
d
1137
202
e
e
2476
721
0.3
f
f
16/25
7543
3920
0.9
a
The value represents the average of three male animals. Plasma and brain concentrations were determined after the
indicated time. The brain sample were diluted (1:2(vol: weight)) with water before homogenization. The dilution
factor is 3. The compound was dosed orally at 30 mg/kg. Dosing vehicle: 0.15% docusate sodium; 2% polyvinyl
pyrrolidinone ꢀ K; 97.85% water. Plasma and brain concentration were determined at 7 h. Plasma and brain
concentration were determined at 1 h Plasma and brain concentration were determined at 2 h. Plasma and brain
concentration were determined at 3 h.
b
c
d
e
f
Based on its potency, and overall pharmacokinetic properties, compound 69 was chosen
for evaluation in a rodent efficacy model (Figure 4). In an acute feeding model, ~225 gram male
Sprague−Dawley rats were orally administered 69 one hour prior to free access to food and
3
2
water. In this study, 69 showed a doseꢀdependent reduction in food pellet consumption.
Specifically, the 3 and 10 mg/kg dose groups exhibited reduced food intake for the first 4 hours
and the 30 mg/kg group showed reduced food intake over the entire sampling period. The effect
observed at the 30 mg/kg dose was equivalent to the cannabinoid typeꢀ1 (CB1) receptor blocker,
rimonabant, which was used as a positive control. Moreover, there was no effect on locomotor
22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
activity as assessed by light beam breaks, and no reduction in water intake. In order to determine
if the observed food intake reduction was driven by 5ꢀHT agonism, a subsequent study was
2C
conducted with 69 either alone or in combination with 70 (SBꢀ243213, Figure 5), a selective 5ꢀ
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
33
HT antagonist. Indeed the feeding effects of 69 were completely reversed by coꢀ
2
C
administration of 70 suggesting a 5ꢀHT mechanismꢀbased reduction in food intake (Figure 5).
2
C
The modest reduction of feeding observed at late time points with the 70 alone and 70 + 69 may
3
3
be due to the anxiolytic properties of 70.
a
Figure 4. Rat adꢀlibitum feeding model with 5ꢀHT agonist 69.
2
C
a
The values represent the average of six male animals. Cumulative food intake was analyzed via repeatedꢀmeasure
betweenꢀgroup analysis of variance using StatView software (Scientific Computing, Cary, North Carolina). Time
points with overall significance were further analyzed with Bonferroni postꢀhoc tests to determine betweenꢀgroup
significance.
23
ACS Paragon Plus Environment

Page 25 of 94
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
a
Figure 5. Reversal in the rat adꢀlibitum feeding model with 5ꢀHT antagonist 70.
2
C
a
The values represent the average of six male animals. Cumulative food intake was analyzed via repeatedꢀmeasure
betweenꢀgroup analysis of variance using StatView software (Scientific Computing, Cary, North Carolina). Time
points with overall significance were further analyzed with Bonferroni postꢀhoc tests to determine betweenꢀgroup
significance.
Conclusion. A new screening protocol utilizing an active site mutant receptor was developed for
the identification of novel 5ꢀHT receptor agonists. This protocol successfully enabled the
2
C
discovery of a series of heterocycles which exhibited good potency at 5ꢀHT . Further, with few
2
C
exceptions, these compounds were found to have excellent subtype functional selectivity against
the closely related 5ꢀHT and 5ꢀHT receptors. These results, combined with promising in vitro
2
B
2A
and in vivo profiles, make the compounds described attractive candidates for further
investigation. Moreover, these compounds represent a significant structural divergence from
what has typically been considered a requisite for agonism of the 5ꢀHT receptor, namely
2
C
exclusion of the basic amine that is proposed to interact with the active site Asp134 residue. This
unique structural series may additionally lead to clinical candidates with improved efficacy,
tolerability and toxicity profiles compared to presently known 5ꢀHT agonists.
2
C
Experimental Section.
24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Chemistry. All nonꢀaqueous reactions were carried out under an argon or nitrogen atmosphere at
room temperature, unless otherwise noted. All reagents and solvents were purchased from
commercial sources and were used without further purification or distillation, unless otherwise
stated. Organic solutions were concentrated under reduced pressure on a Büchi rotary evaporator
using a water bath. Chromatographic purification of products was accomplished using Teledyne
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Isco ICN 60 32ꢀ64 mesh SiO preꢀpacked columns. Thinꢀlayer chromatography (TLC) was
2
performed on Silicycle 0.25 mm SiO Fꢀ254 plates. Visualization of the developed
2
chromatogram was performed by fluorescence quenching or by anisaldehyde, ceric ammonium
1
13
molybdate, or potassium permanganate stain. H and C NMR spectra were recorded on a
Bruker 500 (500 MHz and 125 MHz) unless otherwise noted. Chemical shifts (δ) are reported
from tetramethylsilane with the solvent resonance as the internal standard (CDCl : δ 7.26, C D :
3
6
6
δ 7.15, CD OD: δ 4.78, 3.31: DMSOꢀd : δ 2.50 CD CN: δ 1.94). Data are reported as follows:
3
6
3
chemical shift (δ), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, h = heptet, br =
broad, m = multiplet), integration, coupling constants (Hz), and assignment. Unless noted
1
otherwise, the reported H NMR signals were assigned using standard NMR techniques or by a
1
direct comparison to the H NMR spectra of corresponding starting materials. Analytical highꢀ
pressure liquid chromatography (HPLC) and LCMS analyses were conducted using Shimadzu
LCꢀ10AS pumps and a SPDꢀ10AV UV-vis detector set at 220 nm with the MS detection
performed with a Micromass Platform LC spectrometer. All compounds were found to be ≥95%
pure by HPLC analysis unless otherwise noted.
General procedure for amide coupling: General method A.
Triethylamine (3 equiv.) and 2,4,6ꢀtripropylꢀ1,3,5,2,4,6ꢀtrioxatriphosphinane 2,4,6ꢀ
trioxide (T3P) (2 equiv.) were added to a stirring suspension of the appropriate carboxylic acid (1
25
ACS Paragon Plus Environment

Page 27 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
equiv.) in an appropriate solvent (0.1M) at room temperature. After 5 minutes, the desired amine
(1.5 equiv.) was added and the mixture was stirred at room temperature until LCMS analysis
indicated completion. The crude reaction mixture was then loaded directly onto an Isco SiO2
cartridge for purification. Alternatively, the products could be purified via preparative HPLC.
The desired fractions were combined and concentrated to give the designated amide products.
General procedure for Suzuki coupling: General method B.
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
A degassed mixture of the appropriate heteroaryl halides (2 equiv.), 1ꢀethylꢀ3ꢀ
(naphthalenꢀ1ꢀyl)ꢀ5ꢀ(4,4,5,5ꢀtetramethylꢀ1,3,2ꢀdioxaborolanꢀ2ꢀyl)ꢀ1Hꢀpyrazole (1 equiv.),
tetrakis(triphenylphosphine)palladium(0) (0.15 equiv.), and K PO (3 equiv.) in Nꢀmethylꢀ2ꢀ
3
4
pyrrolidinone (0.3M) was heated to 95ꢀ100 °C until LCMS analysis indicated completion. The
reaction mixture was cooled to room temperature and then purified by preparative HPLC
(Phenomenex Luna 21.2 × 100 mm column; mobile phase A = 10:90 methanol: water with 0.1%
trifluoroacetic acid; mobile phase B = 90:10 methanol: water with 0.1% trifluoroacetic acid or
Waters XBridge C18, 19 × 100 mm column; mobile phase A = 5:95 acetonitrile: water with 10ꢀ
mM ammonium acetate; mobile phase B = 95:5 acetonitrile: water with 10ꢀmM ammonium
acetate). Alternatively, the crude reaction mixtures could be purified by Isco SiO flash
2
chromatography. The desired fractions were combined and concentrated to give the designated
products.
(2-(2,3-Dichlorophenyl)thiazol-4-yl)(morpholino)methanone (1). Step 1. Lithium
hydroxide monohydrate (0.5 g, 12 mmol) was added to a stirring solution of ethyl 2ꢀ
bromothiazoleꢀ4ꢀcarboxylate (0.6 g, 2.4 mmol) in tetrahydrofuran (23 mL), water (4.6 mL) and
methanol (2.3 mL). After 1.5 hours the reaction mixture was diluted with ethyl acetate and
acidified to ~pH 2 with 1.0 N HCl. The organic layer was washed with brine, dried over Na SO ,
2
4
26
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 94
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
filtered and the solvent removed in vacuo to give 2ꢀbromothiazoleꢀ4ꢀcarboxylic acid (0.5 g, 2.4
mmol, 100% yield) as a paleꢀyellow solid. The product was used without further purification.
+
1
MS (ESI) m/z: 207.7 [M+H] ; H NMR (400MHz, DMSOꢀd ) δ 13.29 (br. s., 1H), 8.45 (s, 1H).
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Step 2. To a solution of 2ꢀbromothiazoleꢀ4ꢀcarboxylic acid (330 mg, 1.6 mmol) and
morpholine (140 mg, 1.6 mmol) in dichloromethane (6 mL) was added Hunig's Base (0.8 mL,
4
.8 mmol) followed by 2,4,6ꢀtripropylꢀ1,3,5,2,4,6ꢀtrioxatriphosphinane 2,4,6ꢀtrioxide (T3P),
0% solution in ethyl acetate (1.5 mL, 2.9 mmol). The resulting solution was stirred under argon
5
at room temperature for 2 hours until complete by HPLC analysis. The reaction was loaded onto
a SiO flash column via a solid cartridge was purified by Isco flash chromatography (0ꢀ75%
2
ethyl acetate/hexanes) to afford (2ꢀbromothiazolꢀ4ꢀyl)(morpholino)methanone (380 mg, 1.4
+
1
mmol, 87% yield) as an offꢀwhite solid. MS (ESI) m/z: 279.0 [M+H] ; H NMR (400MHz,
CDCl ) δ 7.87 (s, 1H), 3.85 (br. s., 2H), 3.67 (br. s., 6H).
3
Step 3. To a solution of (2ꢀbromothiazolꢀ4ꢀyl)(morpholino)methanone (50 mg, 0.18
mmol) and (2,3ꢀdichlorophenyl)boronic acid (41 mg, 0.22 mmol) in tetrahydrofuran (4 mL) and
water (1 mL) was added K CO (62 mg, 0.45 mmol). The mixture was degassed with a stream
2
3
of argon for several minutes. Pd(dppf)Cl (7.4 mg, 9.0 µmol) was then added and the reaction
2
mixture was heated with stirring to 80 °C in a sealed vial. After 2 hours, the reaction vessel was
removed from heat and stirred overnight. The reaction mixture was diluted with ethyl acetate,
washed with saturated aqueous NaHCO , washed with brine, dried over Na SO and
3
2
4
concentrated in vacuo to afford an oil. The material was taken up in minimal dichloromethane
and purified on SiO via Isco flash chromatography (0ꢀ100% ethyl acetate/hexanes, 15 min linear
2
gradient, Isco 12 g column). The desired fractions were combined and concentrated to give 1 (25
+
mg, 0.07 mmol, 38% yield) as a paleꢀyellow solid. MS (ESI) mass calculated for [M+H]
27
ACS Paragon Plus Environment

Page 29 of 94
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
(
C H Cl N O S) requires m/z 343.0, found m/z 343.1; H NMR (500MHz, CDCl ) δ 8.07 (s,
1
4
13
2
2
2
3
1H), 7.94 (dd, J=8.0, 1.7 Hz, 1H), 7.51 (dd, J=8.0, 1.4 Hz, 1H), 7.26 (t, J=7.8 Hz, 1H), 3.98 (br.
s., 2H), 3.81ꢀ3.64 (m, 6H).
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
(
4-Methyl-2-(naphthalen-1-yl)thiazol-5-yl)(4-methylpiperazin-1-yl)methanone, TFA
(2). Step 1. Ethyl 2ꢀchloroꢀ3ꢀoxobutanoate (2.7 mL, 19 mmol) was added to a solution of
naphthaleneꢀ1ꢀcarbothioamide (3.5 g, 19 mmol) in ethanol (19 mL). The reaction vessel was
sealed and heated to 80 °C for 19 hours and then cooled to room temperature. The excess solvent
was removed in vacuo and the residue was purified on SiO via Isco flash chromatography (0ꢀ
2
2
0% ethyl acetate/hex, Isco 120 g column) to give ethyl 4ꢀmethylꢀ2ꢀ(naphthalenꢀ1ꢀyl)thiazoleꢀ5ꢀ
+
1
carboxylate (4.6 g, 15 mmol, 82% yield) as a waxy white solid. MS (ESI) m/z: 298.1 [M+H] ; H
NMR (400 MHz, CDCl ) δ 8.77 (d, J=8.53 Hz, 1H), 7.98 (d, J=8.28 Hz, 1H), 7.89ꢀ7.95 (m, 1H),
3
7
.85 (dd, J=1.13, 7.15 Hz, 1H), 7.49ꢀ7.66 (m, 3H), 4.40 (q, J=7.19 Hz, 2H), 2.89 (s, 3H), 1.42 (t,
J=7.15 Hz, 3H).
Step 2. Lithium hydroxide monohydrate (1.3 g, 31 mmol) was added to a rapidly stirring
suspension of ethyl 4ꢀmethylꢀ2ꢀ(naphthalenꢀ1ꢀyl)thiazoleꢀ5ꢀcarboxylate (4.6 g, 16 mmol) in
tetrahydrofuran (118 mL), water (24 mL) and methanol (12 mL). The reaction mixture was
stirred at room temperature for 14 hours and was then diluted with ethyl acetate and acidified to
pH 3 with 10% aqueous citric acid. The organic phase was washed with brine and the combined
aqueous layers were further extracted with ethyl acetate (2x100 mL). The organic layers were
combined, dried over Na SO , filtered and concentrated in vacuo to give 4ꢀmethylꢀ2ꢀ
2
4
(naphthalenꢀ1ꢀyl)thiazoleꢀ5ꢀcarboxylic acid. The product was used subsequently without further
1
purification. H NMR (400 MHz, DMSOꢀd ) δ 13.43 (br. s., 1H), 8.81 (d, J=8.28 Hz, 1H), 8.14
6
(d, J=8.28 Hz, 1H), 8.02ꢀ8.08 (m, 1H), 7.96 (dd, J=1.25, 7.28 Hz, 1H), 7.58ꢀ7.73 (m, 3H), 2.78
28
ACS Paragon Plus Environment