Discovery of a 1-Methyl-3,4-dihydronaphthalene-Based Sphingosine-1-Phosphate (S1P) Receptor Agonist Ceralifimod (ONO-4641). A S1P1 and S1P5 Selective Agonist for the Treatment of Autoimmune Diseases

Article abstract of DOI:10.1021/acs.jmedchem.7b00785

The discovery of 1-({6-[(2-methoxy-4-propylbenzyl)oxy]-1-methyl-3,4-dihydronaphthalen-2-yl}methyl)azetidine-3-carboxylic acid 13n (ceralifimod, ONO-4641), a sphingosine-1-phosphate (S1P) receptor agonist selective for S1P₁ and S1P₅, is described. While it has been revealed that the modulation of the S1P₁ receptor is an effective way to treat autoimmune diseases such as relapsing-remitting multiple sclerosis (RRMS), it was also reported that activation of the S1P₃ receptor is implicated in some undesirable effects. We carried out a structure-activity relationship (SAR) study of hit compound 6 with an amino acid moiety in the hydrophilic head region. Following identification of a lead compound with a dihydronaphthalene central core by inducing conformational constraint, optimization of the lipophilic tail region led to the discovery of 13n as a clinical candidate that exhibited >30 000-fold selectivity for S1P₁ over S1P₃ and was potent in a peripheral lymphocyte lowering (PLL) test in mice (ED₅₀ = 0.029 mg/kg, 24 h after oral dosing).

Full text of DOI:10.1021/acs.jmedchem.7b00785

Subscriber access provided by READING UNIV
Article
The Discovery of a 1-Methyl-3,4-dihydronaphthalen-based Sphingosine-1-
Phosphate (S1P) Receptor Agonist, Ceralifimod (ONO-4641). A S1P1
and S1P5 Selective Agonist for the Treatment of Autoimmune Diseases
Haruto Kurata, Kensuke Kusumi, Kazuhiro Otsuki, Ryo Suzuki, Masakuni Kurono,
Takaki Komiya, Hiroshi Hagiya, Hirotaka Mizuno, Hiroki Shioya, Takeji Ono, Yuka
Takada, Tatsuo Maeda, Norikazu Matsunaga, Tetsu Kondo, Sachiko Tominaga, Ken-
ici Nunoya, Hidekazu Kiyoshi, Masaharu Komeno, Shinji Nakade, and Hiromu Habashita
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00785 • Publication Date (Web): 09 Nov 2017
Downloaded from http://pubs.acs.org on November 9, 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 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Laboratories
Habashita, Hiromu; ONO PHARMACEUTICAL CO. LTD., MEDICINAL
CHEMISTRY RESEARCH LABORATORIES
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 86
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
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
The Discovery of a 1ꢀMethylꢀ3,4ꢀ
dihydronaphthalenꢀbased Sphingosineꢀ1ꢀPhosphate
(S1P) Receptor Agonist, Ceralifimod (ONOꢀ4641).
A S1P and S1P Selective Agonist for the
1
5
Treatment of Autoimmune Diseases
†
†
†
†
†
Haruto Kurata *, Kensuke Kusumi , Kazuhiro Otsuki , Ryo Suzuki , Masakuni Kurono , Takaki
‡
‡
‡
‡
‡
‡
Komiya , Hiroshi Hagiya , Hirotaka Mizuno , Hiroki Shioya , Takeji Ono , Yuka Takada ,
‡
§
§
§
§
Tatsuo Maeda , Norikazu Matsunaga , Tetsu Kondo , Sachiko Tominaga , Ken-ici Nunoya ,
#
#
‡
†
Hidekazu Kiyoshi , Masaharu Komeno , Shinji Nakade , Hiromu Habashita
†
Medicinal Chemistry Research Laboratories, Ono Pharmaceutical Co., Ltd, 3ꢀ1ꢀ1 Sakurai,
Shimamoto, Mishima, Osaka 618ꢀ8585, Japan
‡
Exploratory Research Laboratories, Ono Pharmaceutical Co., Ltd, 17ꢀ2 Wadai Tsukuba, Ibaragi
300ꢀ4247, Japan
ACS Paragon Plus Environment
1

Page 3 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
§
Pharmacokinetic Research Laboratories, Ono Pharmaceutical Co., Ltd, 17ꢀ2 Wadai Tsukuba,
Ibaragi 300ꢀ4247, Japan
#
Safety Research Laboratories, Ono Pharmaceutical Co., Ltd, 50ꢀ10 Yamagishi, Mikuni, Sakai,
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
Fukui 913ꢀ8538, Japan
KEYWORDS: Ceralifimod, ONOꢀ4641, Sphingosineꢀ1ꢀPhosphate (S1P), Multiple Sclerosis
(MS), Autoimmune Disease
ABSTRACT: The discovery of 1ꢀ({6ꢀ[(2ꢀmethoxyꢀ4ꢀpropylbenzyl)oxy]ꢀ1ꢀmethylꢀ3,4ꢀ
dihydronaphthalenꢀ2ꢀyl}methyl)azetidineꢀ3ꢀcarboxylic acid 13n (ceralifimod, ONOꢀ4641), a
sphingosineꢀ1ꢀphosphate (S1P) receptor agonist selective for S1P and S1P , is described. While
1
5
it has been revealed that the modulation of the S1P receptor is an effective way to treat
1
autoimmune diseases such as relapsingꢀremitting multiple sclerosis (RRMS), it was also reported
that activation of the S1P receptor is implicated in some undesirable effects. We carried out a
3
structureꢀactivity relationship (SAR) study of hit compound 6 with an amino acid moiety in the
hydrophilic head region. Following identification of a lead compound with a dihydronaphthalene
central core by inducing conformational constraint, optimization of the lipophilic tail region led
to the discovery of 13n as a clinical candidate which exhibited >30,000ꢀfold selectivity for S1P1
over S1P and was potent in a peripheral lymphocyte lowering (PLL) test in mice (ED : 0.029
3
50
mg/kg, 24 h after oral dosing).
INTRODUCTION
ACS Paragon Plus Environment
2

Journal of Medicinal Chemistry
Page 4 of 86
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
Sphingosineꢀ1ꢀphosphate (S1P) is a bioactive lysophospholipid that mediates a wide
1
variety of biological responses through the S1P1ꢀ5 receptors . During the course of the clinical
2
development of 1 (fingolimod, FTY720) (Figure 1), which is now on the market as the first
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
orally active medicine for the treatment of relapsingꢀremitting multiple sclerosis (RRMS), it was
2
c
reported that its phosphorylated metabolite 2 is a S1P1,3,4,5 pan agonist and the S1P receptor
1
plays a critical role in sequestering peripheral lymphocytes in the lymph nodes and thymus to
3
prevent lymphocytes from entering into inflammatory tissues . Transient bradycardia was
4
reported as one of the adverse events in the clinical studies of 1 and the S1P receptor was
3
5
thought to be implicated in the heart rate reduction observed in rodents . Based on these
6
observations, many S1P ꢀsparing S1P receptor agonists have been reported and several have
3
1
7
8
9
entered into clinical trials including 3 (siponimod, BAF312) , 4 (ponesimod, ACTꢀ128800) , 2ꢀ
aminoꢀ2ꢀ[2ꢀ[2ꢀchloroꢀ4ꢀ[[3ꢀ(phenylmethoxy)phenyl]thio]phenyl]ethyl]ꢀ1,3ꢀpropanediol
1
0
hydrochloride (KRPꢀ203) , 1ꢀ{5ꢀ[(3R)ꢀ3ꢀaminoꢀ4ꢀhydroxyꢀ3ꢀmethylbutyl]ꢀ1ꢀmethylꢀ1Hꢀpyrrolꢀ
1
1
12
2
ꢀyl}ꢀ4ꢀ(4ꢀmethylphenyl)ꢀ1ꢀbutanone (CSꢀ0777) , 5 (ozanimod, RPC1063) , and (R)ꢀ2ꢀ(7ꢀ(4ꢀ
cyclopentylꢀ3ꢀ(trifluoromethyl)benzyloxy)ꢀ1,2,3,4ꢀtetrahydrocyclopenta[b]indolꢀ3ꢀyl)acetic acid
1
3
(etrasimod, APD334) (Figure 1). Among them, we discovered 1ꢀ({6ꢀ[(2ꢀmethoxyꢀ4ꢀ
propylbenzyl)oxy]ꢀ1ꢀmethylꢀ3,4ꢀdihydronaphthalenꢀ2ꢀyl}methyl)azetidineꢀ3ꢀcarboxylic acid
3n (ceralifimod, ONOꢀ4641) , a sphingosineꢀ1ꢀphosphate (S1P) receptor agonist selective for
S1P and S1P , which showed potential for the treatment of RRMS in phase 2 proof of concept
1
4
1
1
5
studies. It was reported that S1P ꢀsparing S1P receptor agonists including 13n can induce
3
1
7
, 14c
transient bradycardia in humans
. Nevertheless, S1P ꢀsparing S1P agonists should be a safer
3 1
option than a nonselective S1P receptor agonist for the treatment of autoimmune diseases such as
RRMS because the activation of the S1P receptor has been reported to be implicated in
3
ACS Paragon Plus Environment
3

Page 5 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
15
undesirable effects such as bronchoconstriction and blood pressure increase . Herein, we
describe the discovery, synthesis, and detailed structureꢀactivity relationship (SAR) study of 13n.
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 1. Selected examples of S1P agonists in clinical trials and on the market (excerpt from
1
7
the literature )
NH2
OR
OH
F
F
O
CO H
N
2
F
N
1
2
: R = H: fingolimod (FTY720) (as HCl salt)
: R = PO H (Active metabolite)
3: Siponimod (BAF312)
3
2
O
N
O
Cl
O
N
N
OH
S
OH
N
N
H
O
OH
HCl
N
4: Ponesimod (ACTꢀ128800)
5: Ozanimod (RPC1063)
1
6
We first identified hit compound 6 by screening our proprietary library . Compound 6
with a βꢀalanine moiety in the hydrophilic head region and a 5ꢀphenylpentyl moiety in the
2
+
lipophilic tail region showed 26ꢀfold selectivity for S1P over S1P in a Ca influx assay and
1
3
had an ED of 26 mg/kg (4 h after oral dosing) in a peripheral lymphocyte lowering (PLL) test
5
0
in mice. Subsequently, the central region and hydrophilic amino acid region of hit compound 6
were constrained by a double bond and ring closures to investigate conformational preferences
for the S1P receptor and the S1P /S1P selectivity, which led to a selective and potent lead
1
3
1
1
7
compound with a central dihydronaphthalene core in our early work . Further, a SAR study of
ACS Paragon Plus Environment
4

Journal of Medicinal Chemistry
Page 6 of 86
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
the lipophilic tail region revealed that incorporation of both a fluorine at the 4ꢀposition on the
terminal phenyl ring in the 3ꢀphenylpropyloxy moiety linked to the 6ꢀposition of the
dihydronaphthalene core and a methyl group at the 2ꢀposition on the propylene linker improved
the potency toward the S1P receptor, the ED result from the PLL test, and the selectivity for
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
1
50
S1P over S1P . As a result, compound 7 was identified with a preferred (S)ꢀmethyl substituent
1
3
at the aboveꢀmentioned 2ꢀposition and exhibited 26,000ꢀfold selectivity for S1P over S1P and
1
3
1
8
was potent in the PLL test in mice (ED : 0.22 mg/kg, 24 h after oral dosing) (Figure 2). This
5
0
outcome encouraged us to further optimize the lipophilic tail region in the dihydronaphthalene
series leading to the discovery of a clinical candidate that was more potent than 7 in the PLL test
while maintaining or elevating the S1P /S1P selectivity of 7.
3
1
Figure 2. Our Early Work to Identify a Central Dihydronaphthalene Core
F
O
CO H
O
2
(
S)
H
N
N
CO H
2
6a
7b
Hit compound
Ca2+ assay
EC50 (nM)
Ca2+ assay
EC50 (nM)
Mouse PLL
ED50 (mg/kg)
Mouse PLL
ED50 (mg/kg)
V
V
hS1P1
60
hS1P3 Selectivity
p.o. 4hr
26
hS1P1
0.37
hS1P3 Selectivity
9500 26000
p.o. 24hr
0.22
1
4200
26
a: ref. 17
b: ref. 18
RESULTS AND DISCUSSION
ACS Paragon Plus Environment
5

Page 7 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chemistry. In general, amino acid derivatives with the general structure 13 were prepared by
using 6ꢀhydroxylꢀ1ꢀmethylꢀdihydronaphthaleneꢀ2ꢀaldehyde derivative 8 as the key intermediate
(Scheme 1). The intermediate 8 was converted to 6ꢀalkoxyꢀ1ꢀmethylꢀdihydronaphthaleneꢀ2ꢀ
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
aldehyde derivative 11 by the reaction with alcohol 9 (Route A) or alkyl halide 10 derived from
alcohol 9 (Route B) under Mitsunobu and basic conditions, respectively. The aldehyde 11 was
subsequently converted to amino acid derivative with the general structure 13 by reductive
alkylation with 3ꢀazetidine carboxylic acid (Route C) or reductive alkylation with methyl 3ꢀ
azetidine carboxylate followed by saponification of the resulting esters (Route D–E).
Scheme 1. General Synthetic Route for Amino Acid Derivatives 13
A
C
3
R ꢀOH
1
1
R
R
1
9
R
HO
O
R
3
O
R
R
3
CO H
2
R
2
2
R
CHO
2
N
CHO
1
R
D
O
R
B
3
CO Me
E
13
11
R
2
8
2
N
3
R ꢀBr
10
12
Benzyl alcohol derivatives 9f-k with the isoꢀbutyl groups at the 4ꢀposition on the
benzylic phenyl ring were prepared as shown in Scheme 2. Alcohol 9f was readily available from
the reduction of the corresponding carboxylic acid 14 using a BH ·THF complex. The isoꢀbutyl
3
group was introduced to aryl triflate 16 under Fürstner conditions using an isoꢀbutyl Grignard
19
reagent in the presence of a catalytic amount of Fe(acac) . The resulting 1ꢀmethoxyꢀ3ꢀisoꢀ
3
butylbenzene 17 was orthoꢀformylated using tertꢀBuLi/TMEDA in nꢀhexane followed by the
addition of DMF to give 18 in 31% yield, followed by quantitative conversion to alcohol 9g after
ACS Paragon Plus Environment
6

Journal of Medicinal Chemistry
Page 8 of 86
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
reaction with NaBH . Alcohols 9h and 9i were derived from the corresponding methyl 4ꢀ
4
hydroxybenzoates 19h and 19i by reaction with Tf O to yield the aryltriflate derivatives 20h and
2
20i followed by the reaction with the isoꢀbutyl Grinard reagents to yield the methyl 4ꢀisoꢀ
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
butylbenzoate derivatives 21h and 21i and then reaction with LiAlH . Alcohols 9j and 9k were
4
synthesized using the same reaction conditions employed to prepare the alcohols 9h and 9i using
methyl and ethyl 4ꢀchlorobenzoates 23j and 23k, respectively, instead of the aryltriflates as
substrates.
Scheme 2. Synthesis of Benzyl Alcohols 9f–k with isoꢀButyl Substituents at the 4ꢀPosition on
a
the Benzylic Phenyl Ring
a
OH
CO H
2
14
9f
HO
OMe
b
TfO
OMe
c
OMe
d
OMe
CHO
e
OMe
OH
15
16
17
18
9g
R
R
R
R
3
HO
b
c
f
TfO
2
OH
CO Me
CO Me
2
2
CO Me
2
1
9h; R = 3-OMe
20h; R = 3-OMe
20i; R = 2-Cl
21h; R = 3-OMe
21i; R = 2-Cl
9h; R = 3-OMe
9i; R = 2-Cl
19i; R = 2-Cl
R
Cl
Cl
R
R
g
c
f
CO R'
OH
CO H
CO R'
2
2
2
22
23j; R = R' = Me
3k; R = F, R' = Et
24j; R = R' = Me
24k; R = F, R' = Et
9j; R = Me
9k; R = F
2
a
Reagents and conditions: (a) BH ·THF complex, THF, 0°C to r.t., overnight, quantitative yield;
3
(
b) Tf O, Pyridine, DCM, 0°C, 20 min–1 h, 97% to quantitative yield; (c) iꢀBuMgBr/Et O, cat.
2
2
ACS Paragon Plus Environment
7

Page 9 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Fe(acac) , THF, NMP, 0°C to r.t. or r.t., 0.5–3.5 h, 50% to quantitative yield; (d) tertꢀBuLi,
3
TMEDA, nꢀhexane, −78°C to 0°C, 4 h and then DMF, THF, 0°C, 4 h, 31%; (e) NaBH , MeOH,
4
0
°C, 1.5 h, quantitative yield; (f) LiAlH , THF, 0°C, 10 min–1 h, 46–75% in two steps; (g)
4
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
SOCl , MeOH, 0°C to r.t., 2.5 days, quantitative yield.
2
Benzyl alcohol derivatives 9l–p with alkyl substituents other than an isoꢀbutyl group at
the 4ꢀposition on the benzylic phenyl ring were prepared as shown in Scheme 3. Alcohol 9l was
obtained from the reduction of carboxylic acid 25 using a BH ·THF complex in 92% yield.
3
Alcohol 9m with an ethyl substituent at the 4ꢀposition was prepared by using an orthoꢀlithiation
method. The phenol moiety of 3ꢀethylphenol 26m was protected by a methoxymethyl group
using methoxymethyl chloride and potassium carbonate and the resulting methoxymethyl ether
was orthoꢀformylated at the less hindered orthoꢀposition to the methoxymethoxy group in the
presence of tertꢀBuLi followed by DMF to yield the aldehyde derivative 27m in 23% yield.
Compound 27m was converted to alcohol 9m in 74% yield by the deprotection of the
methoxymethyl group under acidic conditions, methylation of the resulting phenol with methyl
iodide and reduction of the formyl group by NaBH . Alcohol 9n with a nꢀpropyl substituent at
4
the 4ꢀposition was prepared using similar procedures as above in which the phenol moiety of 3ꢀ
nꢀpropylphenol 26n was protected by the methoxymethyl group and then the less hindered orthoꢀ
position to the methoxymethyl group was carboxylated using CO gas instead of DMF to yield
2
carboxylic acid derivative 27n. This was converted to alcohol 9n in 75% yield by the
deprotection of the methoxymethyl group under acidic conditions, concomitant methylation of
both the phenol and the carboxylate with methyl iodide and reduction of the ester by LiAlH4.
Alcohol 9o was synthesized by similar procedures to yield 9h and 9i in 22% yield over three
ACS Paragon Plus Environment
8

Journal of Medicinal Chemistry
Page 10 of 86
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
steps using an nꢀbutyl Grignard reagent instead of the isoꢀbutyl Grignard reagents used to
synthesize 9h and 9i. Alcohol 9p with a tertiary alcohol was prepared via intermediate 31
obtained by the reaction of acetone with a benzyl anion species generated in situ from methyl 2ꢀ
methoxyꢀ4ꢀmethylbenzoate 30 using lithium diꢀisopropylamide.
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
Scheme 3. Synthesis of Benzyl Alcohols 9l–p with Substituents other than an isoꢀButyl Group at
a
the 4ꢀPosition on the Benzylic Phenyl Ring
OMe
a
OMe
OH
CO H
2
25
9l
R
OMOM
R
OMe
OH
OH
b
c
R
X
O
2
6m; R = Et
27m; R = Et, X = H
27n; R = n-Pr, X = OH
9m; R = Et
9n; R = n-Pr
26n; R = n-Pr
HO
OMe
OMe
CO Me
d
e
OMe
OH
CO Me
2
2
28
29
9o
OMe
f
OMe
CO Me
e
OMe
OH
OH
OH
CO Me
2
2
30
31
9p
a
Reagents and conditions: (a) BH ·THF complex, THF, 0°C to r.t., 5 h, 92%; (b) i) MOMCl,
3
K CO , DMF, 50°C, overnight for 26m, MOMCl, iꢀPr NEt, THF, 0°C to 100°C, 15 h for 26n, ii)
2
3
2
tertꢀBuLi, nꢀhexane, 0°C, 1 h, and then DMF, 0°C, 15 min, 23% in two steps for 27m or CO gas,
2
ACS Paragon Plus Environment
9

Page 11 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
1
7
1% in two steps for 27n; (c) i) HCl/1,4ꢀdioxane, r.t., overnight, ii) MeI, K CO , DMF, r.t., 3.5–
2
3
5 h, iii) NaBH , MeOH, 0°C to r.t., 1 h, 74% in three steps for 9m or LiAlH , THF, 0°C, 1 h,
4
4
5% in three steps for 9n; (d) i) Tf O, Pyridine, CH Cl /THF, 0°C to r.t., 6 h, ii) nꢀ
2
2
2
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
BuMgBr/THF, cat. Fe(acac) , THF, NMP, 0°C to r.t., 1 h; (e) LiAlH , THF, 0°C, 1 h, 22% in
3
4
three steps for 9o and 17% in two steps for 9p; (f) nꢀBuLi, iꢀPr NH, THF, DMPU, −78°C, 2 h
2
and then acetone, −78°C to 0°C, 1.5 h.
Pyridine methanol derivatives 9q and 9r were prepared as shown in Scheme 4. 2,6ꢀ
Dichloronicotinic acid 32q and 4,6ꢀdichloronicotinic acid 32r were converted to the arylester
derivatives 33q and 33r using a catalytic amount of H SO in MeOH in 78% and 51% yield,
2
4
respectively. Then they were subjected to the aboveꢀmentioned Fe(acac) ꢀcatalyzed coupling
3
reactions with the nꢀpropyl Grignard reagent followed by S Ar reactions with sodium methoxide
N
and reduction by LiAlH to yield the pyridine methanol derivatives 9q and 9r.
4
a
Scheme 4. Synthesis of Pyridine Methanols 9q and 9r
Cl
X
Cl
a
Cl
X
Cl
b
X
OMe
c
X
OMe
OH
Y
Y
Y
Y
CO H
CO Me
CO Me
2
2
2
3
2q; X=N, Y=CH
33q; X=N, Y=CH
33r; X=CH, Y=N
34q; X=N, Y=CH
34r; X=CH, Y=N
9q; X=N, Y=CH
9r; X=CH, Y=N
32r; X=CH, Y=N
a
Reagents and conditions: (a) H SO , MeOH, PhMe, 110°C, 15–18 h, 78% for 33q and 51% for
2
4
3
3r; (b) i) nꢀPrMgBr/Et O, cat. Fe(acac) , THF, NMP, r.t., 2 h for 33q or 0°C to r.t., 2 h for 33r,
2
3
ACS Paragon Plus Environment
10

Journal of Medicinal Chemistry
Page 12 of 86
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
ii) NaOMe, MeOH, r.t., 18 h to 50°C 6 h for 34q or 0°C to 50°C, 2 h for 34r, (c) LiAlH , THF,
4
0°C, 20 min, 10% in three steps for 9q or r.t., 20 min, 19% in three steps for 9r.
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
Dihydronaphthalene intermediates 8x and 8y with methyl substituents at the 5ꢀ and 7ꢀ
positions, respectively, were prepared as described in Scheme 5. 3ꢀMethoxyꢀ2ꢀmethylbenzoic
acid 35x was converted to the corresponding benzyl chloride 36x by reaction with a BH ·THF
3
complex followed by CCl in the presence of PPh in 98% yield. The Cꢀ3 homologation reaction
4
3
of the benzyl chloride 36x was carried out by using ethyl acrylate in the presence of nꢀBu N and
3
20
a catalytic amount of Pd(OAc) . The resulting E/Z mixture of the unsaturated ester derivatives
2
was hydrogenated to the corresponding saturated ester that was saponified to 4ꢀarylbutanoic acid
37x in 30% yield over three steps. The carboxylic acid 37x was converted to an acid chloride in
situ by reaction with oxalyl chloride in the presence of a catalytic amount of DMF followed by
subjection to a FriedelꢀCrafts type cyclization reaction to quantitatively give tetralone 38x. The
tetralone 38x was converted to the dihydronaphthalene derivative 39x by the addition of a methyl
Grignard reagent to the ketone followed by dehydration under acidic conditions. Introduction of
2
1
a formyl group at the 2ꢀposition on the dihydronaphthalene 39x using the Vilsmeier reaction
and deprotection of the methyl group from the methoxy group at the 6ꢀposition on the
dihydronaphthalene core by BBr in CH Cl afforded dihydronaphthalene intermediate 8x in
3
2
2
33% yield from 39x. The other dihydronaphthalene intermediate 8y with a methyl group at the 7ꢀ
position on the core was also prepared by the same procedure as 8x.
a
Scheme 5. Synthesis of Dihydronaphthalene Intermediates 8x and 8y
ACS Paragon Plus Environment
11

Page 13 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
R
O
R
R
a
b
MeO
MeO
MeO
O
R
Cl
3
2
2
2
R
CO H
2
R
R
1
2
3
1
2
1
2
3
3
5x; R = Me, R = R = H
5y; R = H, R = R = Me
36x; R = Me, R = H
36y; R = H, R = Me
37x; R = Me, R = H
37y; R = H, R = Me
1
2
3
1
2
1
2
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
1
1
R
R
R
c
MeO
d
MeO
e
HO
2
2
2
R
R
R
CHO
O
1
2
1
2
1
2
3
3
8x; R = Me, R = H
8y; R = H, R = Me
39x; R = Me, R = H
39y; R = H, R = Me
8x; R = Me, R = H
8y; R = H, R = Me
1
2
1
2
1
2
a
Reagents and conditions: (a) i) BH ·THF complex, THF, 0°C to r.t., 3 h for 35x or LiAlH ,
3
4
THF, 0°C to r.t., 30 min for 35y, ii) PPh , CCl , 50–70°C, 14–16 h, 82–98% in two steps; (b) i)
3
4
CH =CHCO Et, Pd(OAc) , nꢀBu N, 95°C, 14 h for 36x or 130°C, 20 min under microwave
2
2
2
3
irradiation for 36y, ii) H , 10%PdꢀC, MeOH or EtOH, r.t., 3 h, iii) 2 mol/L, NaOHꢀaq, MeOH,
2
THF, r.t., 16 h, 30% in three steps for 37x or EtOH, THF, r.t., 2 h, 68% in three steps for 37y; (c)
(
COCl) , DMF, CH Cl , 0°C to r.t., 20–30 min and then SnCl , PhMe, 0°C to r.t., 20–30 min,
2 2 2 4
67% to quantitative yield; (d) i) MeMgCl/THF, THF, 0°C, 1.5–2 h, ii) 1 mol/L, HClꢀaq, EtOAc,
r.t., 100 min, 86% for 39x or 5 mol/L, HClꢀaq, THF, r.t., 80 min, 96% for 39y; (e) i) POCl3,
DMF, 0°C to r.t., 2–2.5 h, 69%–72%, ii) BBr , CH Cl , −78°C to 0°C, 4 h, 54%–91%.
3
2
2
Amino acid derivatives 13a–u were prepared according to Scheme 6. 6ꢀSubstitutedꢀ1ꢀ
methylꢀ3,4ꢀdihydroꢀ2ꢀnaphthalenecarbaldehydes 11b–f, 11h–k, 11m, and 11o–p were
1
8
synthesized from the reaction of the dihydronaphthalene intermediate 8w with the
corresponding alcohols under Mitsunobu conditions (Alcohols 9b–e were commercially
ACS Paragon Plus Environment
12

Journal of Medicinal Chemistry
Page 14 of 86
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
available and the preparation of 9f, 9h–k, 9m, and 9o–p are described in Schemes 2–4.) (Route
A). Aldehydes 11a, 11g, 11l, 11n, and 11q–r were synthesized from the reaction of the
dihydronaphthalene 8w with the corresponding alkyl bromide under basic conditions (Route B).
The alkyl bromide for 11a was commercially available while the alkyl bromides for 11g, 11l,
11n, and 11q–r were generated from the corresponding alcohols 9g, 9l, 9n, and 9q-r from the
reaction with CBr and PPh in CH Cl or PBr in Et O. All the inꢀhouse synthesized alkyl
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
4
3
2
2
3
2
bromides were used without purification. The resulting aldehydes 11a–r were converted to the
2
2
amino acid derivatives 13a–r using two procedures. One involved the reductive alkylation of
the aldehydes 11a–f, 11h–m, and 11q–r with 3ꢀazetidine carboxylic acid to directly yield the
desired amino acids 13a–f, 13h–m, and 13q–r (Route C). The other employed the reductive
alkylation of the aldehydes 11g and 11n–p with methyl 3ꢀazetidine carboxylate hydrochloride to
the esters 12g and 12n–p followed by the saponification of the esters to the desired amino acid
derivatives 13g and 13n–p (Routes D–E). Amino acid derivatives 13s–u were obtained using the
same procedures as mentioned above starting from dihydronaphthalene 8x or 8y via Route B and
C.
a
Scheme 6. Synthesis of Amino Acid Derivatives 13a–u
ACS Paragon Plus Environment
13

Page 15 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A
C
RꢀOH
9
b-f, h-k,
9m, o-p
HO
O
O
CO H
R
R
2
N
CHO
CHO
O
D
CO Me
R
2
E
B
11a-r
8w
13a-r
N
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
RꢀBr
1
1
0a, g, l, n
0q-r
12g, n-p
R =
a;
b;
c;
g;
d;
h;
*
*
*
*
*
*
*
OMe
OMe
e;
f;
*
Cl
F
OMe
*
i;
j;
k;
o;
l;
*
*
*
OMe
*
OMe
*
OMe
*
OMe
*
m;
n;
p;
OH
N
OMe
*
OMe
*
q;
r;
N
X
OMe
O
1
1
R
X
OMe
1
R
R
C
B
HO
O
R
CO H
2
2
2
R
CHO
R
CHO
2
N
1
2
1
2
1
2
8
8
x; R = Me, R = H
y; R1 = H, R = Me
11s; X = CH, R = Me, R = H
11 t; X = CH, R = H, R = Me
1u; X = N, R = Me, R = H
13s; X = CH, R = Me, R = H
13t; X = CH, R = H, R = Me
13u; X = N, R = Me, R = H
2
1
2
1
2
1
2
1
2
1
a
Reagents and conditions: (Route A) RꢀOH, TMAD, PPh , THF, r.t., 2.5 h to overnight; (Route
3
B) RꢀBr (commercially available or derived from reaction of RꢀOH with CBr , PPh , CH Cl , 0
4
3
2
2
°
C to r.t., 30 min to 1 h or with PBr , Et O, 0°C, 1.5 h), K CO , DMF, 0°C to r.t., 8 h–overnight;
3 2 2 3
(
Route C) 3ꢀazetidine carboxylic acid, NaOH (powder), HC(OMe) , MeOH/THF, 0°C, 3–4 h or
3
0°C to r.t., 5–6 h, and then NaBH , 0°C, 0.5–3 h; (Route D) methyl 3ꢀazetidine carboxylate
4
hydrochloride, NaBH(OAc) , AcOH, DMF, 0°C to r.t., 2 h to overnight; (Route E) NaOHꢀaq,
3
MeOH or MeOH/THF, 0°C to r.t., 0.5–4.5 h.
ACS Paragon Plus Environment
14

Journal of Medicinal Chemistry
Page 16 of 86
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
2
+
Biology.
In vitro potency was evaluated by measuring the intracellular Ca concentration
in CHOꢀK1 cells stably expressing the human S1P (hS1P ) or S1P (hS1P ) receptor as a
1
1
3
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
primary assay to examine SAR. In vivo potency toward PLL was evaluated by determining the
number of lymphocytes 4 or 24 h after oral dosing of the test compounds in mice.
Structure-Activity Relationship Study.
With compound 7 in hand, which exhibited 26,000ꢀ
fold selectivity for S1P over S1P and an ED of 0.22 mg/kg at 24 h after oral dosing in the
1
3
50
PLL test in mice, we aimed to discover compounds more potent than 7 using the PLL test by
undertaking a SAR study on the dihydronaphthalene series. Starting from compound 13v with a
naked phenyl ring at the end of the propylene linker, a structure discovered during the course of
18
identification of compound 7 from our earlier work , we first investigated the effects of
conformational restrictions in the lipophilic tail region (Table 1). Amino acid derivatives 13a,
1
3b, and 13d with naphthalene and tetrahydronaphthalene moieties decreased the potency
toward the S1P receptor by 6 to 32ꢀfold compared with compound 13v. While 13c and 13e
1
bearing indane rings were tolerated in terms of their potency toward the S1P receptor, they were
1
less potent than compound 13v in the PLL test in mice.
Table 1. Structureꢀactivity relationship study of the Effects of Conformational Restrictions in the
Lipophilic Tail Region
ACS Paragon Plus Environment
15

Page 17 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
CO H
R
2
N
Mouse PLL a
ED50 (mg/kg)
Ca2+ Assay a
EC50 (nM)
Selectivity
hS1P1
3.0
hS1P3
5000
hS1P / hS1P
p.o. 4 h
0.40
Compound
R
3
1
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
3v
3a
1700
>190
*
*
1
54
>10000
N.T.
13b
19
>30000
>30000
>1600
>3800
N.T.
> 1
*
13c
13d
7.9
*
*
97
>30000
5500
>310
N.T.
13e
1.7
>3200
> 1
*
N.T.: Not Tested
a
EC and ED values are shown as the mean of at least two and five data points, respectively
5
0
50
(
one experiment with five mice in the peripheral lymphocyte lowering test).
Next, we turned our attention to derivatives with benzyl moieties in the lipophilic tail
region because they provided the opportunity for thorough investigation of the effects of varying
the substituents and substitution patterns on the benzylic aryl ring. A representative of this series
from our library was compound 13f having an isoꢀbutyl substituent at the 4ꢀposition on the
benzylic phenyl ring and less than 50% of maximal activity toward the S1P receptor at 30 ꢀM
3
ACS Paragon Plus Environment
16

Journal of Medicinal Chemistry
Page 18 of 86
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
(
Table 2). To develop a SAR around 13f, the influence of an additional substituent on the
benzylic aryl ring was investigated using the methoxy group as a probe. Compound 13g with a 2ꢀ
methoxy group improved the potency toward the S1P receptor and the PLL test 4 h after oral
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
dosing (ED : 0.0092 mg/kg) by 5ꢀfold and 42ꢀfold, respectively, compared with 13f. While
5
0
compound 13h with a 3ꢀmethoxy group also improved the potency toward the S1P receptor and
1
the PLL test 4 h after oral dosing (ED : 0.034 mg/kg), it was less selective for S1P over S1P
3
5
0
1
(S1P /S1P : 1,300) and less potent than 13g. Further investigation of the substituents at the 2ꢀ
3 1
position revealed that compound 13g with the methoxy group was more potent than compounds
3i, 13j, and 13k with a chlorine atom, a methyl group, and a fluorine atom, respectively, in
1
terms of both its in vitro S1P and in vivo PLL potencies. This encouraging result for 13g
1
prompted us to optimize the substituents at the 4ꢀposition on the benzylic aryl ring.
Table 2. SAR Study of the Effects of Substituents on the Benzylic Aryl Ring
ACS Paragon Plus Environment
17

Page 19 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
O
CO H
R
2
N
Mouse PLLa
ED50 (mg/kg)
Ca2+ Assay a
EC50 (nM)
Selectivity
hS1P1
hS1P3
hS1P / hS1P
Compound
R
3
1
p.o. 4 h
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
3f
5.0
1.0
>30000
>6000
0.39
*
OMe
*
13g
9300
4000
9300
1300
0.0092
OMe
13h
3.0
0.034
*
Cl
*
13i
4.5
13
>30000
>30000
>30000
>6700
>2300
>12000
0.19
0.23
13j
*
F
13k
2.6
0.073
*
a
EC and ED values are shown as the mean of at least two and five data points, respectively
5
0
50
(one experiment with five mice in the peripheral lymphocyte lowering test).
The optimization of the alkyl chain at the 4ꢀposition revealed that methyl (13l), ethyl
(13m), and nꢀpropyl (13n) groups were the optimal length for S1P receptor potency and
1
selectivity for S1P over S1P at concentrations up to 30 ꢀM (Table 3). Removing the terminal
1
3
methyl group from the iso-butyl group of 13g contributed to attenuating the activation of the
ACS Paragon Plus Environment
18

Journal of Medicinal Chemistry
Page 20 of 86
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
S1P receptor without affecting the potency toward the S1P receptor (13n). Interestingly,
3
1
introduction of a tertiary alcohol to the iso-butyl group of 13g to give 13p was found to be
tolerated in terms of both S1P receptor potency and the PLL test, indicating that a specific
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
hydrophilic functionality is tolerated in the lipophilic tail region (ClogP: 4.97 for 13g and 2.91
for 13p). Among them, compound 13n with the most promising S1P /S1P selectivity (>30,000)
3
1
and potency in the PLL test in mice (ED : 0.018 and 0.029 mg/kg 4 and 24 h after oral dosing,
5
0
respectively) was selected for a further SAR study.
Table 3. SAR Study of the Effects of Alkyl Substituents at the 4ꢀPosition on the Benzylic Aryl
Ring
R
OMe
O
CO H
2
N
Mouse PLL a
ED50 (mg/kg)
Ca2+ Assay a
EC50 (nM)
Selectivity
hS1P1
hS1P3
hS1P / hS1P
3
1
p.o. 4 h
p.o. 24 h
Compound
R
13g
*
*
1.0
1.4
9300
9300
0.0092
> 0.1
0.040
N.T.
13l
>30000
>21000
1
3m
0.58
>30000
>30000
>30000
>52000
>30000
>8600
0.026
0.018
0.090
0.30
0.029
0.18
*
*
13n
13o
1.0
3.5
6.6
*
*
1
3p
>30000
>4500
0.16
N.T.
OH
N.T.: Not Tested
ACS Paragon Plus Environment
19

Page 21 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
EC and ED values are shown as the mean of at least two and five data points, respectively
5
0
50
(one experiment with five mice in peripheral lymphocyte lowering test).
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
Following identification of compound 13n, the introduction of a nitrogen atom in the
benzylic aryl ring was investigated (Table 4). While compound 13r with a 5ꢀpyridine ring (Y =
N) decreased the potency toward the S1P receptor by 57ꢀfold compared with 13n, compound
1
1
3q with a 3ꢀpyridine ring (X = N) had comparable potencies to 13n toward the S1P receptor
1
and the PLL test 4 h after oral dosing (ED : 0.011 mg/kg). As the 3ꢀpyridine ring in 13q
5
0
contributed to reducing the ClogP (4.57 for 13n vs 3.97 for 13q), the S1P /S1P selectivity was
3
1
reduced compared with 13n.
Finally, a SAR study of the substituents on the dihydronaphthalene core was undertaken.
While compound 13t with a methyl group at the 7ꢀposition on the core resulted in a 540ꢀfold
decrease in potency toward the S1P receptor, compound 13s with a methyl group at the 5ꢀ
1
position on the core was wellꢀtolerated in terms of both S1P receptor potency and the PLL test
1
(
ED : 0.032 mg/kg 4 h after oral dosing) while maintaining S1P /S1P selectivity (>25,000)
5
0
3
1
(Table 4). A comparison between 13q and 13u bearing a 3ꢀpyridine ring in the benzylic part
revealed that the methyl group at the 5ꢀposition on the core was effective to attenuate the
activation toward the S1P receptor without impairing its potencies toward the S1P receptor and
3
1
in the PLL test. Compounds 13n, 13s, and 13u were subjected to further evaluation.
ACS Paragon Plus Environment
20

Journal of Medicinal Chemistry
Page 22 of 86
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
Table 4. SAR Study of the Effects of Introducing a Nitrogen Atom on the Benzylic Phenyl Ring
and Varying the Methyl Substituents on the Dihydronaphthalene Ring in 13n
X
OMe
O
1
R
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
Y
CO H
2
2
N
R
Ca2+ Assay a
EC50 (nM)
Mouse PLL a
ED50 (mg/kg)
Selectivity
R1
H
R2
H
hS1P1
hS1P3
>30000
6800
hS1P / hS1P1
Compound
X
CH
N
Y
CH
CH
3
p.o. 4 h
0.018
1
3n
3q
1.0
>30000
12000
1
H
H
0.58
0.011
13r
CH
CH
CH
N
N
H
Me
H
H
H
57
1.2
540
1.2
>30000
>30000
>30000
>30000
>530
>25000
>56
>0.10
0.032
N.T.
13s
CH
CH
CH
13t
Me
H
13u
Me
>25000
0.017
N.T.: Not Tested
a
EC and ED values are shown as the mean of at least two and five data points, respectively
5
0
50
(one experiment with five mice in the peripheral lymphocyte lowering test).
Further Evaluation of Selected Compounds.
We
evaluated
three
representative
compounds, 13n, 13s, and 13u, identified from the SAR studies. As summarized in Table 5,
compound 13n was more potent than 13s and 13u, both in the PLL test in mice 24 h after oral
dosing by 2.5ꢀ and 1.7ꢀfold, respectively, and in a cAMP accumulation assay in cells expressing
the S1P receptor by 3.0ꢀ and 2.8ꢀfold, respectively. The activation of the S1P receptor is
5
5
13, 23
believed to contribute to positive effects in neurodegenerative disorders
. Based on these
results, compound 13n was selected for further evaluation.
ACS Paragon Plus Environment
21

Page 23 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 5. In vitro and in vivo Profiles of Three Representative Compounds 13n, 13s, and 13u
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
X
OMe
O
R
CO H
2
N
Ca2+ Assay a
EC50 (nM)
cAMP Assay a
EC50 (nM)
Mouse PLL a
ED50 (mg/kg)
Compound
X
R
H
hS1P
1
hS1P
2
hS1P
3
hS1P
1
hS1P
5
p.o. 24 h
0.029
1
3n
CH
1.0
1.2
>30000
N.T.
>30000
>30000
0.027
0.072
0.33
0.98
1
3s
CH
N
Me
0.072
1
3u
Me
1.2
N.T.
>30000
0.19
0.92
0.048
N.T.: Not Tested
a
EC and ED values are shown as the mean of at least two and five data points, respectively
5
0
50
(
one experiment with five mice in the peripheral lymphocyte lowering test).
The pharmacokinetic properties of 13n were investigated after single intravenous
injection and oral administration in male Crl:CD(SD) rats and monkeys. The total clearance
CL ), oral halfꢀlife (T ) and bioavailability (BA) of 13n were 338 mL/h/kg, 17 h, and 75.9%
(
tot
1/2
and 63.6 mL/h/kg, 29 h, and 66.5%, in rats and monkeys, respectively, at 0.1 mg/kg as shown in
3
Table 6. In liver microsomes, metabolism of [ H]ꢀ13n was slow and similar across species as
3
shown in Table 7. Additionally, the serum protein binding of [ H]ꢀ13n at concentrations of 50 to
5
,000 ng/mL showed no significant interꢀspecies differences as shown in Table 7.
ACS Paragon Plus Environment
22

Journal of Medicinal Chemistry
Page 24 of 86
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
Table 6. Pharmacokinetic Properties of 13n
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
Species
AUCinf
Cmax
CLtot
T1/2
Vss
BA
(%)
Route
i.v.
(Dose)
(ng·h/mL) (ng/mL) (mL/h/kg)
(h)
(mL/kg)
3
02
338
16
7150
(540)
ꢀ
a
(48)
(54)
(1)
Rat
75.9
(
0.1 mg/kg)
2
29
7.47
17
p.o.
i.v.
ꢀ
ꢀ
(14)
(0.54)
(4)
1
640
63.6
30
2660
(480)
ꢀ
a
(
390)
(16.8)
(2)
Monkey
66.5
(0.1 mg/kg)
1
090
28.4
(2.1)
29
p.o.
ꢀ
ꢀ
(60)
(1)
a
Values are the mean from three animals, standard deviation is given in parentheses.
Table 7. In vitro DMPK Properties of 13n
The unchanged form of 13n
b
Species
after 2 hꢀincubation in the
Serum protein binding (%)
a
liver microsomes (%)
Mouse
SD rat
93.3
92.6
96.2 to 97.1
97.7 to 98.2
97.2 to 97.7
97.1 to 97.3
96.8 to 97.1
Lewis rat
not applicable
94.1
Cynomolgus monkey
Human
95.2
a
Values represent the mean of two metabolism experiments.
ACS Paragon Plus Environment
23

Page 25 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
b
Values represent the mean of four protein binding experiments. Concentrations of compound
1
3n were 50, 500, and 5,000 ng/mL.
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 potential efficacy of 13n to treat autoimmune diseases was evaluated using a
2
4
collagenꢀinduced arthritis (CIA) model in female Lewis rats (Figure 3). The clinical scores of
the 13n 0.1 mg/kg group were significantly lower than those of the control group and were
considered to be consistent with the PLL potency of 13n in the female Lewis rats where the
1
4a
maximum effect was induced at a dose of 0.1 mg/kg .
a
Figure 3. Clinical Scores of 13n in the CIA Model
18.0
1
1
5.0
2.0
9.0
6
3
0
.0
.0
.0
**
**
**
***
**
**
0
5
10
15
20
25
30
Days after the first immunization
a
Arthritis was induced by immunizing with bovine type II collagen emulsified in incomplete
Freund’s adjuvant at day 0 and 7 to the female Lewis rats. Arthritis was not induced in the
ACS Paragon Plus Environment
24

Journal of Medicinal Chemistry
Page 26 of 86
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
normal control group. Arthritis efficacy was evaluated by using the arthritis score characterized
by edema and/or erythema in the paws. 13n (0.03 or 0.1 mg/kg) or 0.5% methyl cellulose was
administered orally once daily from day 0 to 27. The arthritis scores were expressed as the mean
value ± standard error for the normal control (open circle, n=5), control (filled circle, n=10), 13n
0.03 mg/kg (square, n=9), 13n 0.1 mg/kg (triangle, n=9). The steel test was used to compare the
scores between the control and 13n groups, with a pꢀvalue of less than 5%; **, p<0.01; ***,
p<0.001.
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
Subsequently the effects of 13n on heart rate and peripheral lymphocyte counts were
examined in monkeys. When compound 13n was orally administered in a single dose to
unrestrained conscious monkeys at dosages of 1, 10 and 100 mg/kg, a significant decrease in
heart rate was observed at 4 to 24 hours after administration at a dosage of 100 mg/kg and the
maximum decrease of heart rate to the preꢀtreatment value was 19% at 4 hours after
administration (Table SI-1). The plasma concentrations of 13n were increased with increasing
dosages (320, 3,570 and 13,500 ng/mL at 4 hours after dosing at 1, 10 and 100 mg/kg,
respectively), and the plasma concentration at 100 mg/kg at 24 hours after dosing was
maintained at 11,100 ng/mL. On the other hand, a significant decrease in peripheral lymphocyte
counts was observed at 4 to 24 hours at dosages of 0.01 and 0.03 mg/kg after oral administration
of 13n (Table SI-2).
Our SAR study was driven by the compound’s ability to activate the human S1P1
receptor and its effect in the mouse PLL test. The amino acid sequences lacking the Nꢀterminal
ACS Paragon Plus Environment
25

Page 27 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
extracellular and Cꢀterminal intracellular domains containing a ligand binding pocket (Ser49 in
TM1 (transmembrane domain) through Thr314 in TM7) are 99.6% identical between the mouse
and rat. Furthermore, the homology between the mouse S1P receptor and the human homolog is
1
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
9
7.4%. Therefore, we considered that any interꢀspecies differences in the potency of 13n toward
33
the S1P receptor would be negligible. In fact, 13n inhibited the binding of [ P]ꢀS1P to the
1
hS1P and rS1P receptors with similar Ki values of 0.626 nM and 0.772 nM, respectively, and
1
1
in the cAMP accumulation assay, 13n activated hS1P and rS1P to the same extent with EC s
1
1
50
1
4a
of 0.0273 nM and 0.0286 nM, respectively . The peripheral lymphocyte count has been used as
8,14b
a pharmacodynamic biomarker to gauge efficacy across species
. Therefore, we used the PLL
potency in mice as the primary indicator to triage compounds and thus, the most potent
compound 13n was evaluated in the rat CIA model. Based on reports that the egress of
3
lymphocytes from the lymph nodes requires the S1P receptor and the PLL effect by S1P
1
1
2
d, 12
agonists are induced by the downꢀregulation of the S1P receptor
, i.e., by blocking the S1P1
1
signal as functional antagonists, we performed PK/PD analysis of 13n tested in the CIA model.
In the female Lewis rats used for the model, the maximum concentrations (Cmax) obtained by
single oral administration of 13n were 5.76 nM and 21.5 nM at 0.03 mg/kg and 0.1 mg/kg,
respectively. From the expression analysis of the human S1P receptor on the plasma membrane
1
of genetically engineered CHO cells, the concentration of 13n required to induce the
1
4a
downregulation of the S1P receptor by approximately 90% (25 nM) was consistent with the
1
Cmax at the efficacious dose of 0.1 mg/kg in the model. We concluded this analysis supports the
1
hypothesis that the S1P agonists induce the PLL effect via the downregulation of the S1P1
receptor resulting in improvement of the clinical scores in the animal model.
ACS Paragon Plus Environment
26

Journal of Medicinal Chemistry
Page 28 of 86
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
Given the selectivity of 13n for S1P as shown by the magnitude of its inhibition of the
1
3
3
specific binding of [ P]ꢀS1P to the human S1P , S1P , S1P , S1P and S1P receptors, with Kis
1
2
3
4
5
1
4a
of 0.626 nM, >5450 nM, >5630 nM, 28.8 nM, and 0.574 nM respectively , we considered
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
compound 13n to be a S1P and S1P dual agonist. Suitable profile including the data on 13n
1
5
1
4a
discussed above, its efficacy in an EAE model in mice and prediction of a favorable human
1
4b
response encouraged us to advance 13n into clinical trials.
CONCLUSIONS
Starting from the hit compound 6 with a simple amino acid structure, 1ꢀ({6ꢀ[(2ꢀmethoxyꢀ
4
ꢀpropylbenzyl)oxy]ꢀ1ꢀmethylꢀ3,4ꢀdihydronaphthalenꢀ2ꢀyl}methyl)azetidineꢀ3ꢀcarboxylic acid
1
3n, a sphingosineꢀ1ꢀphosphate (S1P) receptor agonist selective for S1P and S1P , was
1
5
discovered as a clinical candidate through a detailed SAR study. Facilitated by the convergent
synthetic strategy using the intermediates 8w–y with a central dihydronaphthalene core, the
tolerability of conformational restrictions in the lipophilic tail region was investigated. The
subsequent SAR study focused on the benzyl motif revealed that the methoxy group at the 2ꢀ
position on the benzylic aryl ring improved potency in the PLL test in mice by 42ꢀfold (13f vs
1
3g). Optimization of the alkyl substituents at the 4ꢀposition on the benzylic aryl ring led to the
2+
discovery of 13n which exhibited >30,000ꢀfold selectivity for S1P over S1P in the Ca influx
1
3
assay and was highly potent in the PLL test in mice (ED : 0.029 mg/kg, 24 h after oral dosing).
5
0
The benzyl series allowed us to investigate various substituents and substitution patterns in the
lipophilic tail region compared with the phenylpropyl series as exemplified by 7 previously
identified in our earlier work. A further SAR study revealed hydrophilic groups such as a tertiary
ACS Paragon Plus Environment
27

Page 29 of 86
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
alcohol (13p) and 3ꢀpyridine (13q) in the lipophilic tail region were tolerated in terms of their
potencies toward the S1P receptor and the PLL test while a methyl substituent at the 5ꢀposition
1
on the central dihydronaphthalene core attenuated the activity toward the S1P receptor (13q vs
3
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
3u). Compound 13n was potent in the cAMP accumulation assay not only toward S1P but also
1
S1P , properties which are expected to lead to additional preferred effects for the treatment of
5
RRMS. Suitable profile including pharmacokinetic properties for a onceꢀdaily administration in
rat and monkey, significant efficacies in autoimmune disease models such as the rat CIA model,
little interꢀspecies difference in in vitro DMPK properties, significant difference in the dose
levels to induce lymphopenia and bradycardia in monkeys, and prediction of human response
encouraged us to advance 13n into clinical trials. The directꢀacting S1P agonist 13n may allow
1
more accurate control of its pharmacological activities in vivo compared with proꢀdrugs that
require conversion to phosphorylated active metabolites in vivo.
EXPERIMENTAL SECTION
It should be noted that final compounds 13a-u described herein can induce biological effects at
very low dose in human (See reference 14c).
Chemistry
General.
Unless noted otherwise, all reagents and solvents were used as purchased without
further purification. Analytical samples were homogeneous as confirmed by thin layer
chromatography (TLC) and afforded spectroscopic results consistent with the assigned
ACS Paragon Plus Environment
28