Spiro-1-benzofuranpiperidinylalkanoic acids as a novel and selective sphingosine S1P5 receptor agonist chemotype

Article abstract of DOI:10.1016/j.bmcl.2017.12.018

The synthesis and SAR of a novel class of spirobenzofuranpiperidinyl-derived alkanoic acids 6–34 as sphingosine S1P₅ receptor agonists are described. The target compounds generally elicit high S1P₅ receptor agonistic potencies and in general are selective against both S1P₁ and S1P₃ receptor subtypes. The key compound 32 shows a high bioavailability of 73% and a CNS/plasma ratio of 0.8 after oral administration in rats.

Full text of DOI:10.1016/j.bmcl.2017.12.018

Accepted Manuscript
Spiro-1-benzofuranpiperidinylalkanoic acids as a novel and selective sphingo-
sine S1P5 receptor agonist chemotype
Axel R. Stoit, Jos H.M. Lange, Hein K.A.C. Coolen, Annemieke Rensink, Adri
van den Hoogenband, Arnold P. den Hartog, Sjoerd van Schaik, Chris G. Kruse
PII:
S0960-894X(17)31178-2
https://doi.org/10.1016/j.bmcl.2017.12.018
BMCL 25474
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Accepted Date:
29 September 2017
8 December 2017
Please cite this article as: Stoit, A.R., Lange, J.H.M., Coolen, H.K.A., Rensink, A., van den Hoogenband, A., den
Hartog, A.P., van Schaik, S., Kruse, C.G., Spiro-1-benzofuranpiperidinylalkanoic acids as a novel and selective
sphingosine S1P5 receptor agonist chemotype, Bioorganic & Medicinal Chemistry Letters (2017), doi: https://
doi.org/10.1016/j.bmcl.2017.12.018
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Spiro-1-benzofuranpiperidinylalkanoic acids as a novel and selective sphingosine S1P5
receptor agonist chemotype
Axel R. Stoit, Jos H. M. Lange, Hein K. A. C. Coolen, Annemieke Rensink, Adri van den
Hoogenband, Arnold P. den Hartog, Sjoerd van Schaik and Chris G. Kruse
Abbott Healthcare Products B.V. (formerly Solvay Pharmaceuticals B.V.), C. J. van
Houtenlaan 36, 1381 CP Weesp, The Netherlands

Direct correspondence to Dr. J.H.M. Lange, Kraton Chemical B.V., Transistorstraat 16, 1322 CE Almere, The Netherlands. Telephone: +31
(0)365462800. E-mail: jos.lange@kraton.com

Abstract. The synthesis and SAR of a novel class of spirobenzofuranpiperidinyl-derived
alkanoic acids 6-34 as sphingosine S1P5 receptor agonists are described. The target
compounds generally elicit high S1P5 receptor agonistic potencies and in general are
selective against both S1P1 and S1P3 receptor subtypes. The key compound 32 shows a
high bioavailability of 73% and a CNS/plasma ratio of 0.8 after oral administration in rats.
Keywords: S1P5 Receptor agonists; Subtype selectivity; Spirocyclic scaffold; Molecular
modelling; Homology model; Oral bioavailability; Turbidimetric aqueous solubility;
Microsomal stability; Membrane permeation
2

Sphingosine-1-phosphate (S1P, 1) is a bioactive lipid with important functions in multiple
cellular signaling systems1. S1P affects the central nervous system,2 cardiovascular system
and immune system and has been implicated in a broad range of diseases3 (Figure 1). S1P
also activates the G protein-coupled sphingosine receptors S1P1 - S1P5. The approval of
the S1P receptor agonist fingolimod (2, FTY720, Gilenya®) in 2010 for the treatment of
relapsing multiple sclerosis has intensified sphingosine research efforts.4 Compound 2 acts
as a pro-drug by phosphorylation into the active (S)-Fingolimod-phosphate (3) which is an
analogue of 1. Compound 2 was reported5 to act (via 3) on four of the five S1P receptor
subtypes (excluding S1P2) and to lead to lymphopenia. More S1P receptor subtype
selective ligands with better pharmacokinetic profiles and fewer side effects will be
required4,6 to further elucidate the physiological background and therapeutic roles of S1P1-
5. Several selective S1P1/5 receptor dual agonists have been discovered such as siponimod,
ozanimod, ceralifimod, AMG3697 and GSK2018682 for the treatment of multiple
sclerosis and other autoimmune and inflammatory disorders.4,8
The S1P5 receptor is most highly expressed in the central nervous system, particularly on
oligodendrocytes and brain endothelium. The S1P5 receptor was shown to mediate the
immune quiescence of the human endothelium barrier.9 The number of orally available,
selective S1P5 receptor agonists is very limited. The 2H-phthalazin-1-one analogue 4 was
reported10 by Novartis as a selective and orally active S1P5 receptor agonist. Recently, A-
971432 was disclosed11 as a highly selective S1P5 receptor agonist. A-971432 exhibited
excellent plasma and CNS exposure after oral dosing in several preclinical species and
reversed lipid accumulation as well as age-related cognitive decline in rodents. Hanessian
reported12 the chiral pyrrolidine derivative 5 which acted as an agonist on both S1P4 and
S1P5, being devoid of activity at S1P1 and S1P3. Compound 5 can be regarded as a
3

constrained azacyclic analogue of 3 and has a relatively poor pharmacokinetic profile. It
serves in the present study as a molecular modeling tool.
Herein, a set of novel spirocyclic benzofuranpiperidinylalkanoic acids 6-34 and the
phosphate analogue 35 are disclosed as a new selective S1P5 receptor agonist chemotype.
Sphingosine-1-phosphate (1) Fingolimod (2)
(S)-Fingolimod phosphate (3)
4
A-971432
5
Figure 1. Chemical structures of selected S1P receptor agonists.
The synthesis of the target 2H-spiro(1-benzofuran-3,4’-piperidine) compounds 6-21 is
depicted in Scheme 1. 2-Bromo-5-methoxyphenol 36 was coupled to the protected
tetrahydropyridinemethanol derivative13 37 in a Mitsunobu reaction to give 38 in 63%
yield. Radical spirocyclization of 38 in the presence of tributyltin hydride and a catalytic
amount of the radical initiator AIBN provided 39 in a chemical yield of 68%. The
methoxy group in 39 was removed under strongly acidic conditions to furnish 40 in
quantitative yield. Reductive debenzylation of 40 efficiently led to 41 which was
converted into 42 by reaction with tert-butylacrylate in the presence of diisopropylamine
as organic base. Mitsunobu coupling of 42 with a set of appropriate alcohols gave the tert-
butyl esters 43-58, respectively. The target compounds 6-21 were obtained from 43-58 by
acidic removal of their tert-butyl protective group, thereby liberating the carboxylic acid
moiety.
4

Ph
O
Br
Br
N
N
a
N
b
+
OH
O
HO
.HCl
O
O
O
36
37
38
39
CO2-t-Bu
CO2-t-Bu
Ph
O
H
N
N
N
N
d
c
e
f
39
O
O
HO
O
HO
HO
RO
40
41
42
43-58
Compound
R
Compound
R
CO2H
6 .HCl
phenylpropyl
14 .HCl
2-Cl-6-CF3-benzyl
7 .HCl
8 .TFA
9 .HCl
benzyl
15 .HCl
16 .HCl
17 .HCl
2-Cl-6-Et-benzyl
2-Cl-6-i-Pr-benzyl
2-Cl-6-c-Pr-benzyl
N
2-Cl-benzyl
3-Cl-benzyl
g
43-58
10 .HCl
11 .HCl
12 .HCl
13 .HCl
2,6-Cl2-benzyl
2,5-Cl2-benzyl
2,3-Cl2-benzyl
18 .HCl
19 .HCl
20 .HCl
21
2-Cl-6-OCF3-benzyl
3,5-Cl2-pyridin-4-ylmethyl
2,6-Cl2-4-Me-benzyl
2,6-(C2H5)2-benzyl
O
RO
2-Cl-6-F-benzyl
6-21
Scheme 1. Reagents and conditions: (a) PPh3, DIAD, THF, < 10 oC, 2h followed by rt, 16 h (63%). (b) n-
Bu3SnH, AIBN (cat.), benzene, N2, reflux, 16 h (68%). (c) 48% HBr, AcOH, reflux, 24 h (quantitative
yield). (d) H2, Pd(OH)2, 4N HCl, MeOH, rt, 48 h (68%). (e) tert-butylacrylate, (i-Pr)2NH, MeOH, reflux, 16
h (91%). (f) R-OH, PPh3, DIAD, CH2Cl2, rt, 16 h (50-90%). (g) 4M HCl, 1,4-dioxane, 50 oC, 16 h (80-
95%).
The synthesis of the fluoro analog 22 is depicted in Scheme 2. The aromatic difluoride14
59 was coupled to 4-pyridinemethanol 60 to furnish 61. Benzylation of 61, followed by
partial reduction of the in situ formed positively charged quaternary pyridine ring by
sodium borohydride in methanol at low temperature gave 62. Microwave-assisted radical
spirocyclization of 62 in the presence of tributyltin hydride and a catalytic amount of
radical initiator provided 63 in a chemical yield of 72%. Subsequent reductive
debenzylation by ammonium formate in the presence of Pd(OH)2 gave a high yield of 64
which was analogously converted as described earlier for 6-21 to the fluorinated target
compound 22 via the intermediates 65 and 66.
5

H
Ph
Br
F
Ph
Ph
N
Ph
O
O
O
O
N
e
O
f
c,d
a,b
F
F
F
59
O
HO
+
Br
Br
O
F
O
Ph
N
F
HO
63
64
N
N
60
Ph
Ph = phenyl
61
62
CO2H
CO2-t-Bu
CO2-t-Bu
N
N
N
g
h
i
64
Cl
Cl
O
O
Cl
O
.HCl
O
Cl
O
HO
F
F
F
65
66
22
Scheme 2. Reagents and conditions: (a) reaction of 60 with NaH, NMP, rt, 30 min. (b) Addition of 59, 100
oC, 15 min (83%). (c) BnBr, acetone, 40 oC, 16 h. (d) NaBH4, MeOH, -10 oC→rt, 4 h (63%). (e) n-Bu3SnH,
2,2’-azobis(2-methylpropionitrile) (cat.), benzene, microwave, SiC, 175 oC, 1 h (72%). (f) ammonium
formate, Pd(OH)2, MeOH, 60 oC, 16 h (96%). (g) tert-butylacrylate, (i-Pr)2NH, MeOH, reflux, 16 h (82%).
(h) R-OH, PPh3, DIAD, CH2Cl2, rt, 16 h (66%). (i) 4M HCl, 1,4-dioxane, 50 oC, 16 h (85%).
The synthesis of the target compounds 23-26 with a modified carboxylic acid chain is
shown in Scheme 3. The synthesis route proceeds analogously to the reaction sequence of
scheme 1, viz. reaction of the spirocycle 41 with either a bromoalkylcarboxylic acid tert-
butyl ester (leading to 67-68) or tert-butylmethacrylate (leading to 69) in the presence of a
base, followed by a Mitsunobu coupling and acidic hydrolysis of the formed tert-butyl
esters 70-72, respectively, to afford 23-25. The analogue 26 was made analogously to 25
by applying (2-chloro-6-ethylphenyl)methanol in the Mitsunobu reaction instead of (2,6-
dichlorophenyl)methanol.
CO2-t-Bu
CO2-t-Bu
CO2H
CO2H
n
n
n
N
N
R
R
N
R
N
a
b
c
Cl
O
Cl
O
Cl
O
41
.HCl
O
O
O
HO
O
Cl
Cl
67-69
70-72
23 n=0, .HCl
26
24 n=2, R = H
25 n=1, R = Me (.HCl salt)
6

Scheme 3. Reagents and conditions: (a) for 23: BrCH2CO2-t-Bu, DIPEA, CH3CN, 65 oC 16 h (89 %); for
24: Br(CH2)3CO2-t-Bu, K2CO3, CH3CN, 65 oC, 16 h (75 %); for 25: H2C=C(CH3)-CO2-t-Bu, DBU, DMF,
140 oC, 16 h, sealed flask (47 %). (b) 2,6-Cl2-C6H4CH2OH, PPh3, DIAD, CH2Cl2, rt, 16 h (90-98 %). (c) 4M
HCl, 1,4-dioxane, rt, 48 h (23: 86%, 24: 87%, 25: 80%).
The synthesis of the target compounds 27 and 28 is shown in Scheme 4. The phenolic
hydroxy group of 42 was converted in its triflate 73. A hetero cross coupling reaction
between 73 and potassium thioacetate15 mediated by Pd2(dba)3/CypF-t-Bu smoothly gave
74. After liberation of the thiol group under mild basic conditions and coupling to the
appropriate arylmethyl bromides, 75 and 76 were obtained, respectively. Acidic ester
hydrolysis of 75 and 76 led to the target compounds 27 and 28. Compound 75 was
converted to the corresponding hydrochloride salt by brief treatment with 4M HCl in 1,4-
dioxane and its sulfanyl moiety was subsequently oxidized by oxone into a sulfonyl group.
Target compound 29 was obtained by acidic ester hydrolysis of 77.
CO2-t-Bu
CO2-t-Bu
CO2H
N
N
N
a
e
c,d
Cl
S
Cl
S
42
O
O
O
.HCl
R
R
R
75 R = Cl
76 R = Et
73 R = OTf
74 R = SAc
27 R = Cl
28 R = Et
b
CO2H
CO2-t-Bu
N
N
h
f,g
Cl
Cl
S
75
.HCl
.HCl
O
O
Cl
S
Cl
O
O
O
O
77
29
Scheme 4. Reagents and conditions: (a) Trifluoromethanesulfonimide, DMAP, Et3N, CHCl3, 60 oC, 16 h
(93%). (b) KSAc, Pd2(dba)3, CyPF-t-Bu, toluene, 110 oC 24 h (80%). (c) NaOH pellets, EtOH, rt, 30 min.
(d) 2-Cl-6-R-C6H3-CH2Br, EtOH, 0 oC, 50 min (73%). (e) 4M HCl, 1,4-dioxane, rt, 48 h (87%). (f) 4M
HCl, 1,4-dioxane, r.t., 5 min. (g) 3 Mol equiv. potassium peroxymonosulfate (Oxone), H2O, rt, 2 h (72%).
(h) 4M HCl, 1,4-dioxane, rt, 48 h (76%).
The synthesis of the target compounds 30-32 is depicted in Scheme 5. The Sonogashira16
coupling of two arylacetylenes with the aryltriflate 73 gave 78 and 79, respectively in high
7

yields. The alkynes 78 and 79 were hydrolyzed under acidic conditions to afford the target
compounds 30 and 31, respectively. The triflate moiety of 73 was successively converted
into a tributylstannyl group (80) and bromo substituent (81). Subsequent treatment with
the appropriate arylethenyl-BF3K reagent17-19 in a toluene/water mixture at elevated
temperature led to 82, which was further hydrolyzed into target compound 32.
CO2-t-Bu
CO2H
N
N
b
O
O
.HCl
a
73
73
78 R = H
79 R = Cl
30 R = H
31 R = Cl
R
R
CO2R
CO2-t-Bu
N
N
c
e
Cl
O
O
R
Cl
80 R = Bu3Sn
81 R = Br
82 R = t-Bu
32 R = H (.HCl salt)
f
d
Scheme 5. Reagents and conditions: (a) 2-Cl-phenylacetylene or phenylacetylene, Pd(OAc)2, PPh3,
K3PO4.H2O, DMSO, 80 oC, 1 h (72-90%). (b) 4M HCl, 1,4-dioxane, 60 oC, 5 h (70-90%). (c) Sn2(n-Bu)6,
(PPh3)4Pd, LiCl, 1,4-dioxane, 108 oC, 72 h (61%). (d) NBS, THF, 0 oC, 1.5 h (57%). (e) 2-[(E)-2-(2,6-Cl2-
C6H3)ethenyl]BF3K, DPPF, Cs2CO3, toluene/water = 1/3, 100 oC, 16 h (67%). (f) 4M HCl, 1,4-dioxane, rt,
48 h (93%).
The 2H-spiro(1-benzofuran-3,3’-piperidine)-6-ol 83 was prepared in several steps
analogously to the procedures described in scheme 2 by starting from 3-pyridinemethanol
and 4-(benzyloxy)-1-bromo-2-fluorobenzene and was converted into 84 by reaction with
tert-butylacrylate in the presence of a base. Mitsunobu coupling of 84 with two
appropriate alcohols afforded 85 and 86, respectively. These compounds were hydrolyzed
into the target compounds 33 and 34.
8

CO2-t-Bu
b
CO2-t-Bu
c
CO2H
NH
N
N
N
Cl
O
a
Cl
O
.HCl
O
O
O
HO
O
HO
R
R
83
84
85 R = Cl
86 R = Et
33 R = Cl
34 R = Et
Scheme 6. Reagents and conditions: (a) tert-butylacrylate, (i-Pr)2NH, MeOH, reflux, 16 h (84%). (b)
ArCH2OH, PPh3, DIAD, CH2Cl2, rt, 16 h (85: 36%, 86: 72%). (c) 4M HCl, 1,4-dioxane, 50 oC, 2 h (73-
82%).
The synthesis of the spirocyclic phosphate 35 is outlined in scheme 7. The key
intermediate 41 was Boc-protected and the resulting 87 was O-benzylated in quantitative
yield according to the Mitsunobu procedure into 88. After Boc deprotection, the formed
89 was coupled with a THP-protected ethyloxy moiety to give 90. THP-deprotection
under acidic conditions led to the alcohol 91 which was isolated as its hydrochloride salt
and subsequently reacted with di-tert-butyl N,N-diisopropylphosphoramidite, followed by
treatment with tert-butylhydroperoxide to furnish the di-tert-butyl phosphonate 92, which
was hydrolyzed with mineral acid into the phosphate 35.
O
O
R
O
P
OH
t-Boc
OH
H
N
N
N
N
i
c
d
a
41
Cl
O
Cl
O
Cl
O
.HCl
O
O
O
O
HO
Cl
Cl
Cl
87 R = H
88 R = 2,6-Cl2-C6H3
89
90 R = OTHP
91 R = H (.HCl salt)
92 R = -P(=O)(O-t-Bu)2
35
b
e,f
g,h
Scheme 7. Reagents and conditions: (a) Boc2O, CH2Cl2, rt, 16 h (81%). (b) 2,6-Cl2-C6H3-CH2OH, PPh3,
DIAD, CH2Cl2, rt, 16 h (100%). (c) 1M HCl, EtOH, 50 oC, 2 h (59%). (d) 2-(2-chloroethoxy)tetrahydro-2H-
pyran, K2CO3, NaI (cat.), DMF, 100 oC, 16 h (50%). (e) p-toluenesulfonic acid, MeOH, 50 oC, 1 h. (f) 1M
HCl, EtOH, rt, 1 h (8%). (g) di-tert-butyl N,N-diisopropylphosphoramidite, tetrazole, DMF/CH3CN, rt, 2h.
(h) t-BuOOH, nonane, rt, 30 min (62%). (i) 4N HCl, rt, 16 h (81%).
The pharmacological results of the pro-drug fingolimod (2) and the target compounds 6-
35 are given in Table 1. They were primarily evaluated at the human S1P5 receptor and
also profiled at hS1P1 and hS1P3 receptor subtypes. Functional in vitro human S1P assays
9

were applied, based on aequorin-derived luminescence triggered by S1P receptor-
mediated changes in intracellular calcium levels.
Table 1. Pharmacological in vitro results of fingolimod (2) and the target compounds 6-
35.
Compound pEC50
(S1P5),a
pEC50
pEC50
Compound pEC50
(S1P5),a
pEC50
pEC50
(S1P1),b
(S1P3),c
5.6 ± 0.5
(S1P1),b
(S1P3),c
5.5 ± 0.2 < 5.0d
2
6
7
8
9
5.7 ± 0.2 5.3 ± 0.1 5.8 ± 0.3
7.7 ± 0.1
8.2 ± 0.2
7.5 ± 0.2
7.8 ± 0.2
5.8 ± 0.1 5.8 ± 0.1
6.0 ± 0.1 < 5.0
21
22
23
24
6.0 ± 0.1 < 4.5
6.9 ± 0.1d n. d.e
6.5 ± 0.1 < 5.0
< 5.0
< 4.5
< 5.0
5.2 ± 0.1 < 6.0
5.3 ± 0.1 5.9 ± 0.1
6.8 ± 0.2 5.2 ± 0.2 < 5.0
6.7 ± 0.2 4.9 ± 0.2d 6.0 ± 0.2
6.7 ± 0.2 5.8 ± 0.2 < 5.0
7.2 ± 0.2 5.2 ± 0.1 < 5.0
6.9 ± 0.2 6.5 ± 0.2 < 5.0
7.2 ± 0.2
7.8 ± 0.1
7.9 ± 0.1
8.6 ± 0.1
6.1 ± 0.1
7.4 ± 0.2
7.6 ± 0.2
7.9 ± 0.1
7.0 ± 0.1
7.5 ± 0.1
7.7 ± 0.1
5.9 ± 0.1 < 5.0
6.5 ± 0.1 < 5.0
10
11
12
13
14
15
16
17
18
19
20
25
26
27
28
29
30
31
32
33
34
35
< 5.0d
5.4 ± 0.2d
6.9 ± 0.1 < 6.0
< 4.5
< 4.5
< 4.5
< 4.5
< 4.5
< 4.5
< 5.0
< 5.0
< 5.0
< 5.0
< 5.0
< 5.0
7.8 ± 0.1 n. d.e
< 5.0
7.6 ± 0.1 5.9 ± 0.1 < 5.0
7.8 ± 0.1 5.7 ± 0.1 5.7 ± 0.1
7.7 ± 0.1 5.4 ± 0.1 < 5.0
6.5 ± 0.2 < 4.5
< 5.0
8.1 ± 0.3 6.5 ± 0.1 < 5.0
6.6 ± 0.1 < 5.0
aAgonism at the human S1P5 receptor (stably transfected in CHO cells, Aequorin luminescence read-out),
expressed as pEC50 ± SEM values. The values represent the mean result based on at least three independent
experiments, unless otherwise noted. bAgonism at the S1P1 receptor (stably transfected in CHO cells,
Aequorin luminescence read-out), expressed as pEC50 ± SEM values. cAgonism at the S1P3 receptor (stably
10

transfected in CHO cells, Aequorin luminescence read-out), expressed as pEC50 ± SEM values. dResult of
duplicate measurement. en. d.: Not determined.
In general, compounds 6-34 can be regarded as rigidified analogs of the S1P receptor
agonists 1 and 3, wherein the polar phosphate group is replaced by a relatively less polar
carboxylic acid moiety. A-971432 also complies with this general S1P agonist
pharmacophore, thereby deviating from 4, which is lacking a polar phosphate or
carboxylic acid moiety. In this context, our spirocyclic phosphate derivative 35 constitutes
a spirocyclic analog of 1 wherein the polar phosphate group is retained, like in
Hanessian’s rigidified molecular probe 5.
The flexible phenylpropoxy analog 6 showed comparable S1P5 receptor agonist potency
as compared to the pro-drug fingolimod (2), but lacked selectivity against the S1P1 and
S1P3 receptor subtypes. Replacement of the phenylpropoxy group in 6 by the smaller and
less flexible benzyloxy group (7) led to retained S1P5 receptor agonism in combination
with S1P1 and S1P3 selectivity. This fortuitous finding prompted substitution efforts at the
phenyl moiety of the benzyloxy group of 7 in order to increase the S1P5 receptor agonist
potency in this series. The incorporation of a 2-chloro substituent (8) increased the S1P5
potency almost tenfold and turned out to be preferable over 3-Cl substitution (9). The
related dichlorophenyl substituted analogs 10-12 did not lead to increased S1P5 agonism.
Replacement of one of the chloro atoms in 10 by a fluoro atom (13) or a CF3 moiety (14)
gave only small effects on S1P5 agonism. However, a significant improvement on S1P5
agonism was observed for the substitution by an ethyl group (15: pEC50 = 7.8 ± 0.1).
Alternative replacements of one of the chloro atoms in 10 by an isopropyl- (16),
cyclopropyl- (17), or -OCF3 group (18) led to increased S1P5 agonistic potency but a
phenyl (10) to pyridyl (19) ring modification had not much impact. The presence of an
additional 4-methyl group (20) led to an approximate tenfold increase in S1P5 receptor
11

agonism. The 2,6-diethylphenyl substitution pattern (21) was also well-tolerated by the
S1P5 receptor. Interestingly, the presence of an addition fluoro substituent in the central
aromatic region (22) resulted in improved S1P5 receptor agonism (pEC50 = 8.2 ± 0.2).
In the compounds 23-26 the SAR of the terminal alkylcarboxylic acid chain was
investigated in more detail. Interestingly, both truncation (23) and elongation (24) of the
chain led to increased S1P5 receptor agonism, whereas branching of the chain (25 and 26)
had not much impact.
Replacement of the phenolic oxygen atom in 10 and 21 by a sulphur atom (27 and 28) led
in both cases to improved S1P5 receptor agonism, thereby rendering 28 (pEC50 = 8.6 ±
0.1) as the most potent S1P5 receptor agonist in our series. Contrarily, the corresponding
more polar sulfonyl derivative 29 was almost 100 times less potent at this lipophilic
GPCR. The arylethynyl congeners 30 and 31 were significantly more potent than their
arylmethoxy counterparts 7 and 8, respectively. The arylethenyl variant 32 (pEC50 = 7.9 ±
0.1) was tenfold more S1P5 agonistic than its arylmethoxy analogue 10. Interestingly, the
phospate 35 was found almost tenfold more S1P5 agonistic than its carboxylic acid
counterpart 10.
The regioisomeric congeners 33 and 34 of 10 and 21, respectively showed high S1P5
receptor agonism. Interestingly, 34 elicited considerable higher selectivity over the S1P5
and S1P5 receptor subtypes than 21.
From Table 1 it becomes clear that fingolimod (2) has no S1P5/S1P3 receptor selectivity in
contrast with our more rigid spirocyclic series, apart from the more flexible
phenylpropoxy analogue 6. The majority of our target compounds elicited high S1P5/S1P3
receptor subtype selectivities. In particular, 20 and 22 elicited over 1000-fold S1P5/S1P3
selectivities, whereas 13, 15-18, 23, 25-28, and 30-35 were found over 100-fold
S1P5/S1P3 selective.
12

The target compounds 6-35 were also S1P5/S1P1 selective to varying degrees. In
particular, compounds 27, 30, 31, 32, and 34 are of interest since they were found
approximately 1000 fold S1P5/S1P1 selective.
The key compound 32 was also tested for agonism at the human S1P2 and S1P4 receptors.
The compound was devoid of S1P2 agonism (pEC50 < 5.0), but acted as an agonist at the
S1P4 receptor (pEC50 = 7.1 ± 0.1).
Table 2. Turbidimetric aqueous solubility data and in vitro human microsomal stability
data of fingolimod (2) and representative target compounds.
Compound
Aqueous
Microsomal Microsomal Compound
Aqueous
Microsomal
Microsomal
stabilityc
Solubilitya
stabilityb
(Clint)
stabilityc
Solubilitya stabilityb
(Clint)
(t )
½
(t )
½
2
1/6.5
n. d.d
9.1
n. d.c
153
447
505
671
26
27
28
29
30
100/> 100 6.1
100/> 100 21
228
65
8
30/> 100
100/> 100
30/> 100
30/100
10
13
16
3.1
30/ 100
42
33
2.7
100/> 100 33
129
167
2.1
100/100
30/ >100
8.3
19
21
22
> 100/> 100 13.9
100
32
33
34
0.76
1820
454
100/> 100
30/100
0.92
6.5
1510
213
100/> 100 3.1
100/> 100 3.4
409
aKinetic (turbidimetric) solubility, determined at Cyprotex, expressed as the minimal/ maximal value in μM.
b Human microsomal stability at 10-6 M, determined at Cyprotex, expressed as the intrinsic clearance (Clint)
in μl/min/mg protein. cHuman microsomal stability at 10-6 M, determined at Cyprotex, expressed as the half-
life (t ), in minutes. dn. d.: Not determined.
½
The aqueous solubility of a compound constitutes an important factor in determining its
absorption from the gastrointestinal tract and ultimately its oral bioavailability. The
turbidimetric solubility of fifteen representative target compounds was determined and
13

compared with fingolimod (2). The turbidimetric solubility of 2 was only modest.
Contrarily, all tested target compounds exhibited much higher values (Table 2). This was
according to our expectation since the presence of the ionisable nitrogen atom and
carboxylic acid moiety in our target compounds are important features to ameliorate
aqueous solubility.
Human microsomal stability studies were conducted to determine the in vitro intrinsic
clearance and half-life of the fifteen target compounds. In general, the compounds were
metabolically stable in vitro, with the exception of 19 and 27-29. The presence of the
metabolically vulnerable pyridyl ring in 19 and the sulphur atom in 27 and 28 may
account for their reduced in vitro metabolic stability. The compounds 21 and 32 were
found the most stable in this assay.
Since the Caco-2 permeability assay is an established method20 to predict human intestinal
permeability and to investigate drug efflux phenomena, fingolimod (2) and the target
compounds 10, 13 and 19 were tested herein. All three compounds showed satisfactory
apparent permeability coefficient values (Papp(A-B)) which ranged from to 9-21 x 10-6
cm/s. On the contrary, the observed (Papp(A-B)) value for 2 was only 0.53 x 10-6 cm/s,
which is an indicator for poor intestinal epithelial cell permeability. Interestingly,
compound 2 appeared to be subject of active transmembrane transport processes since the
observed Asymmetry Index (AI = Papp(B-A)/Papp(A-B)) value was 4.1. The AI values of
10 and 13 were 1.0 and 0.7, respectively which indicates that active transport is not a
dominant factor. On the contrary, the more polar pyridyl congener 19 showed a higher AI
value of 2.7.
In agreement with its favourable turbidimetric aqueous solubility and microsomal
metabolic stability, the key compound 32 showed a good oral bioavailability (73%) in rat
14

after oral administration and also elicited a favourable CNS/plasma ratio of 0.8, which
turned out to be significantly higher than the CNS/plasma ratio of its analogue 10 (0.2).
Homology models for the S1P1 and S1P5 receptors were built using the RCSB Protein
Data Bank crystal structure at atomic resolution of 2.2 Å of bovine rhodopsin21 (1U19).
Compound 3 was manually docked into the S1P1 receptor in a comparable way as reported
earlier.22 An important difference is the adaptation of the rotameric state of W6.48
(Trp268) to the trans-configuration, which is thought to be associated with the agonistic
state of the receptor.23-25 Compound 3 is bound via a double salt bridge with its basic
nitrogen atom and phosphate group to the vicinal Glu-Arg couple on helix 3. The aromatic
core has multiple -stacking interactions with a cluster of aromatic residues. The tail is
situated in a deep lipophilic pocket between helix 5 and 6 which is characteristic26 for the
GPCR lipid receptor cluster (CB, LPA and S1P).
Figure 2. Compound 32 docked in the homology model of the S1P5 receptor.
15

The binding pockets of the S1P-receptors are highly conserved. The marginal differences
in interacting residues between the S1P1 and S1P5 receptors cannot account for the
selectivity as observed for compound 5. Significant dissimilarities, however, are found in
the second extracellular loop. Especially of interest is the presence of an additional Arg in
de EC2-loop of the S1P5 receptor. In addition, it appeared not to be possible to dock 5 in
the same way as compound 3, having a double salt bridge with the Arg-Glu couple on
helix 3. Due to the rigid configuration of the substituents on the chiral pyrrolidine ring, the
phosphate group points towards the loop region. Therefore, it was hypothesized that the
phosphate group of 5 interacts in the S1P5 receptor with Arg178 and its protonated basic
nitrogen with Glu112, which on itself is also bridged to Arg111. To this purpose the EC2-
loop of our S1P5-model was remodelled in such a way that Arg178 could have an optimal
interaction with the phosphate group of 5. Figure 2 shows a similar docking pose for
compound 32 in our modified S1P5 model. Besides an identical salt bridge pattern as
described above, the benzofuran ring is stacked via several -interactions with Trp264,
Phe196 and Phe116, respectively. The 2-(2,6-dichlorophenyl)ethenyl-tail is positioned in a
lipophilic pocket delimited by the Val123, Leu126, Ala200, Gly203, Ile204, Leu205,
Ile208, Val256, Leu257, Val261 and Leu271 residues.
In conclusion, a new spirocyclic chemotype is disclosed which elicited high S1P5 receptor
agonistic potencies and showed varying degrees of both S1P1 and S1P3 receptor subtype
selectivities. In general, the target compounds featured high turbidimetric aqueous
solubilities as well as high in vitro human microsomal stabilities and membrane
permeation. The potent and selective key compound 32 was found orally available (73%)
in rat after oral administration in combination with a favourable CNS/plasma ratio of 0.8.
16

Supporting information available
Selected analytical and synthetic data for compound 32.
References and notes
1.
2.
Marsolais D, Rosen H. Nat Rev Drug Discov. 2009;8:297-307.
Dev KK, Mullershausen F, Mattes H, Kuhn RR, Bilbe G, Hoyer D, Mir A. Pharmacol Ther. 2008;
117:77-93.
3.
4.
5.
Kunkel GT, Maceyka M, Milstien S, Spiegel S. Nat Rev Drug Discov. 2013;12:688-702.
Park S, Im D. Biomol Ther. 2017;25:80-90.
Brinkmann V, Davis MD, Heise CE, Albert R, Cottens S, Hof R, Bruns C, Prieschl E, Baumruker T,
Hiestand P, Foster CA, Zollinger M, Lynch KR. J Biol Chem. 2002;277:21453-21457.
Chew WS, Wang W, Herr DR. Pharmacol Res. 2016;113:521-532.
6.
7.
Cee VJ, Frohn M, Lanman BA, Golden J, Muller K, Neira S, Pickrell A, Arnett H, Buys J, Gore A,
Fiorino M, Horner M, Itano A, Lee MR, McElvain M, Middleton S, Schrag M, Rivenzon-Segal D,
Vargas HM, Xu H, Xu Y, Zhang X, Siu J, Wong M, Bürli RW. ACS Med Chem Lett. 2011;2:107-
112.
8.
9.
O’Sullivan C, Schubart A, Mir AK, Dev KK. J Neuroinflammation. 2016;13:31.
Van Doorn R, Lopes Pinheiro MA, Kooij G, Lakeman K, van het Hof B, van der Pol SMA, Geerts D,
van Horssen J, van der Valk P, van der Kam E, Ronken E, Reijerkerk A, de Vries HE. J
Neuroinflammation. 2012;9:133.
10.
11.
Mattes H, Dev KK, Bouhelal R, Barske C, Gasparini F, Guerini D, Mir AK, Orain D, Osinde M,
Picard A, Dubois C, Tasdelen E, Haessig S. ChemMedChem. 2010;5:1693-1696.
Hobson A, Harris C, van der Kam E, Turner S, Abibi A, Aguirre A, Bousquet P, Kebede T,
Konopacki D, Gintant G, Kim Y, Larson K, Maull J, Moore N, Shi D, Shrestha A, Tang X, Zhang P,
Sarris K. J Med Chem. 2015;58:9154−9170.
12.
13.
Hanessian S, Charron G, Billich A, Guerini D. Bioorg Med Chem Lett. 2007;17:491-494.
Gijsen HJM, De Cleyn MJA, Love CJ, Surkyn M, Van Brand SFA, Verdonck MGC, Moens L,
Cuypers J, Bosmans JRMA. Tetrahedron. 2008;64:2456-2464.
14.
Stensrud KF, Heger D, Sebej P, Wirz J, Givens RS. Photochem Photobiol Sci. 2008;7:614-624.
17

15.
Van den Hoogenband A, Lange JHM, Bronger RPJ, Stoit AR, Terpstra JW. Tetrahedron
Lett. 2010;51:6877-6881.
16.
17.
18.
19.
20.
21.
22.
Chinchilla R, Nájera C. Chem Rev. 2007;107:874-922.
Darses S, Genet JP. Chem Rev. 2008;108:288-325.
Willis MC, Brace GN, Findlay TJK, Holmes IP. Adv Synth Catal. 2006;348:851-856.
Molander GA, Bernardi CR. J Org Chem. 2002;67:8424-8429.
Van de Waterbeemd H, Smith DA, Beaumont K, Walker DK. J Med Chem. 2001;44:1313-1333.
Okada T, Sugihara M, Bondar AN, Elstner M, Entel P, Buss V. J Mol Biol. 2004;342:571-583.
Pham TT, Fells JI, Osborne DA, North EJ, Naor MM, Parrill AL. J Mol Graph Model.
2008;26:1189-1201.
23.
24.
Ruprecht JJ, Mielke T, Vogel R, Villa C, Schertler GFX. EMBO Journal. 2004;23:3609-3620.
Pardo L, Deupi X, Govaerts C, Campillo M. 3-D Structure of G Protein-Coupled Receptors. In:
Rognan D. ed. Ligand Design for G Protein-Coupled Receptors. Weinheim, Germany: Wiley-VCH;
2006, 183-203.
25.
26.
Jongejan A, Bruysters M, Ballesteros JA, Haaksma E, Bakker RA, Pardo L, Leurs R. Nature Chem
Biol. 2005;1:98-103.
Surgand J, Rodrigo J, Kellenberger E, Rognan D. Proteins. 2006;62:509-538.
18

Spiro-1-benzofuranpiperidinylalkanoic acids as a novel and selective sphingosine
S1P5 receptor agonist chemotype
Axel R. Stoit, Jos H. M. Lange, Hein K. A. C. Coolen, Annemieke Rensink, Adri van den
Hoogenband, Arnold P. den Hartog, Sjoerd van Schaik and Chris G. Kruse
CO2H
N
Cl
O
.HCl
Cl
32
The potent and orally available S1P5 receptor agonist 32 showed high S1P5 receptor
selectivities against the S1P1-3 receptor subtypes.

Direct correspondence to Dr. J.H.M. Lange, Kraton Chemical B.V., Transistorstraat 16, 1322 CE Almere, The Netherlands. Telephone: +31
(0)365462800. E-mail: jos.lange@kraton.com
19