Inhibition of sphingosine 1-phosphate lyase for the treatment of rheumatoid arthritis: Discovery of (E)-1-(4-((1 R,2 S,3 R)-1,2,3,4-tetrahydroxybutyl)-1 H -imidazol-2-yl)ethanone oxime (LX2931) and (1 R,2 S,3 R)-1-(2-(Isoxazol-3-yl)-1 H -imidazol-4-yl)butane-1,2,3,4-tetraol (LX2932)

Article abstract of DOI:10.1021/jm101183p

Sphingosine 1-phosphate lyase (S1PL) has been characterized as a novel target for the treatment of autoimmune disorders using genetic and pharmacological methods. Medicinal chemistry efforts targeting S1PL by direct in vivo evaluation of synthetic analogues of 2-acetyl-4(5)-(1(R),2(S),3(R),4- tetrahydroxybutyl)-imidazole (THI, 1) led to the discovery of 2 (LX2931) and 4 (LX2932). The immunological phenotypes observed in S1PL deficient mice were recapitulated by oral administration of 2 or 4. Oral dosing of 2 or 4 yielded a dose-dependent decrease in circulating lymphocyte numbers in multiple species and showed a therapeutic effect in rodent models of rheumatoid arthritis (RA). Phase I clinical trials indicated that 2, the first clinically studied inhibitor of S1PL, produced a dose-dependent and reversible reduction of circulating lymphocytes and was well tolerated at dose levels of up to 180 mg daily. Phase II evaluation of 2 in patients with active rheumatoid arthritis is currently underway.

Full text of DOI:10.1021/jm101183p

8650 J. Med. Chem. 2010, 53, 8650–8662
DOI: 10.1021/jm101183p
Inhibition of Sphingosine 1-Phosphate Lyase for the Treatment of Rheumatoid Arthritis: Discovery
of (E)-1-(4-((1R,2S,3R)-1,2,3,4-Tetrahydroxybutyl)-1H-imidazol-2-yl)ethanone Oxime (LX2931)
and (1R,2S,3R)-1-(2-(Isoxazol-3-yl)-1H-imidazol-4-yl)butane-1,2,3,4-tetraol (LX2932)
Jeffrey T. Bagdanoff,† Michael S. Donoviel,‡ Amr Nouraldeen,‡ Marianne Carlsen,† Theodore C. Jessop,† James Tarver,†
Saadat Aleem,† Li Dong,† Haiming Zhang,† Lakmal Boteju,† Jill Hazelwood,‡ Jack Yan,† Mark Bednarz,† Suman Layek,†
Iris B. Owusu,‡ Suma Gopinathan,‡ Liam Moran,‡ Zhong Lai,† Jeff Kramer,‡ S. David Kimball,† Padmaja Yalamanchili,†
William E. Heydorn,† Kenny S. Frazier,‡ Barbara Brooks,‡ Philip Brown,‡ Alan Wilson,‡ William K. Sonnenburg,‡
Alan Main,† Kenneth G. Carson,† Tamas Oravecz,‡ and David J. Augeri*,†
†Lexicon Pharmaceuticals, Inc., 350 Carter Road, Princeton, New Jersey 08540, United States, and ‡Lexicon Pharmaceuticals, Inc.,
8800 Technology Forest Place, The Woodlands, Texas 77381, United States
Received September 10, 2010
Sphingosine 1-phosphate lyase (S1PL) has been characterized as a novel target for the treatment of
autoimmune disorders using genetic and pharmacological methods. Medicinal chemistry efforts targeting
S1PL by direct in vivo evaluation of synthetic analogues of 2-acetyl-4(5)-(1(R),2(S),3(R),4-tetrahydroxy-
butyl)-imidazole (THI, 1) led to the discovery of 2 (LX2931) and 4 (LX2932). The immunological
phenotypes observed in S1PL deficient mice were recapitulated by oral administration of 2 or 4. Oral
dosing of 2 or 4 yielded a dose-dependent decrease in circulating lymphocyte numbers in multiple species
and showed a therapeutic effect in rodent models of rheumatoid arthritis (RA). Phase I clinical trials
indicated that 2, the first clinically studied inhibitor of S1PL, produced a dose-dependent and reversible
reduction of circulating lymphocytes and was well tolerated at dose levels of up to 180 mg daily. Phase II
evaluation of 2 in patients with active rheumatoid arthritis is currently underway.
Introduction
including T- and B-cell trafficking.2,3 This approach serves
to mimic the endogenous effects of S1P and was led by the
discovery of Fingolimod (FTY720),4 recently approved for
the treatment of multiple sclerosis by the FDA.
Research dedicated to the study of sphingosine signaling
pathways identified sphingosine 1-phosphate lyase (S1PLa)
as an important enzyme involved in the regulation of the
immune system.1 Human S1PL is an intracellular protein
composed of 568 amino acids and bears 91% similarity to
its mouse homologue. The N-terminus is anchored in the
membrane-rich endoplasmic reticulum (ER). The C-terminus
contains the catalytic site that extends into the cytosol and
performs an irreversible retro-aldol degradation of sphingo-
sine 1-phosphate (S1P), a key signaling lipid present in all
mammalian cells. S1P can serve both as an extracellular signal
and as a secondary messenger in biochemical pathways that
regulate cell differentiation and apoptosis as well as the
vascular and immune systems.2 S1P is an extracellular ligand
for a family of five G-protein-coupled receptors (GPCR)
termed S1P1-S1P5, which are expressed in many tissues.3
To date, the majority of the research aimed at pharmacologi-
cal intervention of the S1P signaling pathway has focused
on synthetic agonists of S1P receptors (S1PR), in particular
S1P1, which regulates numerous aspects of immune function
A functional consequence of S1P or Fingolimod binding to
and activation of S1P1 is receptor internalization and intra-
cellular degradation.5 This so-called functional antagonism of
S1P1 eventually leads to inhibition of lymphocyte egress from
secondary immune tissues, resulting in peripheral lymphopenia
and immunosuppression. Inhibition of S1PL increases tissue
S1P content that leads to similar immunomodulatory
effects.1c,6 The mechanism likely involves the action of both
intracellular and extracellular S1P and may not be limited to
functional antagonism of S1PR or inhibition of lymphocyte
recirculation. A somewhat surprising result of the analysis of
mice and rats expressing reduced levels of S1PL activity was
that significant increases in tissue S1P concentrations were
observed predominantly in lymphoid tissues and that mice
expressing less than 10% wild type S1PL exhibited lympho-
penia without other overt physiological effects.1a The relative
tissue specificity of S1PL inhibition suggested that targeting
this enzyme may provide an attractive alternative approach to
modulation of S1P receptors by synthetic ligands that pre-
sumably achieve systemic circulation and wide tissue distribu-
tion following administration.
*To whom correspondence should be addressed. Phone: (609) 466-
5526. Fax: (609) 466-6079. E-mail: daugeri@lexpharma.com.
aAbbreviations: S1P, sphingosine 1-phosphate; S1PL, sphingosine
1-phosphate lyase; KO, knockout; S1P1, sphingosine 1-phosphate recep-
tor 1; S1PR, sphingosine 1-phosphate receptor; GPCR, G-protein-
coupled receptor; CBC, complete blood count; SAR, structure-activity
relationship; ADME, absorption, distribution, metabolism, excretion;
CIA, collagen induced arthritis; RA, rheumatoid arthritis; ER, endoplas-
mic reticulum; NK cells, natural killer cells; ECG, electrocardiogram;
AUC, area under curve; PK-PD, pharmacokinetic-pharmacodynamic.
S1PL-deficient mice showed resistance to various inflamma-
tory and autoimmune challenges, indicating that pharmacolog-
ical inhibition of S1PL represents a novel therapeutic strategy
for the treatment of autoimmune disorders.1a,7 We recently
reported a series of small molecules that, when administered
r
pubs.acs.org/jmc
Published on Web 11/22/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24 8651
Scheme 1a
a Conditions: (a) 1.0 equiv HCl, Ph3CONH2, dioxane:water (2:1); (b) 2.0 M HCl in dioxane; (c) 2,2-dimethoxypropane, DCM, p-TsOH-H2O; (d) n-
BuLi, -78 °C, THF, then DMF; (e) (CF3CO)2O, TEA, DCM; (f) dioxane:1 N aq HCl (1:1), 50 °C; (g) NaBH4, MeOH, 0 °C; (h) Ph3P, I2, TEA, THF.
Scheme 2a
a Conditions: (a) Bn2NH, EtOH, HOAc, 80 °C; (b) H2, Pd-C, EtOH, HOAc; (c) R-nitrile, MeONa, MeOH, then HOAc.
orally to mice, recapitulated the immune phenotype of S1PL
deficiency.8 Herein, we describe the research effort that led to
the identification of compounds 2 and 4. Compound 2 is the
first clinically studied inhibitor of S1PL for the treatment of
rheumatoid arthritis.
investigated, synthetic studies in our own laboratories invol-
ving deuterated solvents have elucidated certain mechanistic
nuances of the venerable Buchi cyclization.11
Results and Discussion
During the course of synthetic investigations with 1, we
noticed the formation of aldol side products that result from
enolization of the ketone, even under mildly acidic conditions
(Scheme 3). Upon concentration of various HPLC purified
ketone containing analogues from a mixture of acetonitrile
and water buffered with ammonium acetate, the aldol adduct
20 was routinely detected by LCMS and NMR in variable but
significant quantities. By contrast, oxime 2 was found to be
more stable under conditions that proved problematic for the
ketone present in 1. The relative reactivity of the ketone of 1
versus 2 is illustrated by Scheme 3. Treatment of 1 with dilute
DCl in excess D2O provides deuterated intermediate d8-1 due
to facile enolization of the ketone. However, after trapping
the reactive ketone as oxime d3-2, no enolization was ob-
served under protic aqueous conditions. No loss of deuterium
content, dimerization, erosion of oxime geometry, or other
decomposition was observed upon standing in 1.0 M aqueous
HCl at 40 °C for up to two weeks. In contrast to ketone 1,
compound 2 displayed enhanced solubility properties as the
free base and showed extended solubility as the HCl salt; the
HCl salt was readily formulated in neutral water.
We have previously determined that peak lymphopenia
occurred between 12 and 24 h following oral administration
of 1 to mice.8 Accordingly, 18 h after oral administration of a
30 mg/kg dose of 2, the resulting lymphocyte reduction in
peripheral blood was measured by complete blood cell count
(CBC) analysis. As shown in Table 1, regiochemically pure
E-oxime 2 lowered peripheral lymphocyte counts by 60% as
compared to vehicle control. To further clarify the effect of
the structural transformation on in vivo activity, the phar-
macokinetic profile was obtained from a separate PK study.
Chemistry
As shown in Scheme 1, a number of investigational ana-
logues were generated from 2-acetyl-4(5)-(1(R),2(S),3(R),4-
tetrahydroxybutyl)-imidazole (THI, 1). Condensation of 1
with O-trityl hydroxylamine followed by removal of the trityl
protecting group afforded 2 exclusively as the configuration-
ally pure E-oxime. Ketalization of the tetraol side chain prior
to selective deprotection of the trityl group under anhydrous
acidic conditions provided bis-ketal 3. Subsequent deproto-
nation of oxime 3 and alkylation of the aza-enolate with N,N-
dimethylformamide provided isoxazole 4 after aromatization
and deprotection of the tetraol side chain. Alternatively,
reduction of the intermediate N,N-dimethylformamide addi-
tion product followed by cyclization furnished dihydroiso-
xazole 5 after deprotection of the tetraol side chain.
A method for rapidly generating variously substituted C2-
heterocycles was adapted from the Amadori rearrangement9
and Buchi cyclization10 methods (Scheme 2). In the Amadori
rearrangement, reaction of D-glucose with dibenzylamine
under mildly acidic conditions provided, after hydrogenolysis
of the benzyl groups, D-fructosamine acetate 6. Subsequent
Buchi condensation of the aza-sugar with imidates generated
from nitriles and sodium methoxide provided heterobicyclic
tetraols7-19. We have previously reported synthetic methods
for alteringtheside chainand substitutingthe coreheterocycle
present in this chemotype represented by1.8 Inthatdisclosure,
it was determined that good chemical yields in the Buchi
cyclization were obtained only when fructosamine HOAc,
3
the aza-sugar derived from naturally occurring D-glucose, was
utilized. While this phenomenon has not been extensively

8652 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24
Bagdanoff et al.
Scheme 3a
a Conditions: (a) NH4OAc, MeCN, water; (b) 1.0 M DCl in D2O; (c) Ph3CONH2, 1.0 M DCl in D2O, dioxane; (d) 4.0 M HCl in dioxane, water.
Table 1. SAR of C-2 Substituted Imidazoles Containing sp2-Hybridized Heteroatoms
a Reported as % change in circulating lymphocyte numbers relative to vehicle control 18 h after oral delivery of 30 mg/kg dose of HCl salt to mice
(n = 5). b AUCt finfinity following 10 mg/kg oral dose to mice (n = 4). For AUC data for 2 and 4, see Table 2 for complete data (including SD) from
multiple experiments.
0
Table 2. Pharmacokinetic Parameters of 2 and 4 in Micea
b
c
d
compd
T1/2
Cmax
Tmax
Vsse
Clf
AUCg
%Fh
2
4
1.6 ( 0.3
2.9 ( 0.2
3.6 ( 1.2
6.2 ( 2.8
0.45 ( 0.1
2.1 ( 1.4
0.9 ( 0.4
0.54 ( 0.4
7.6 ( 1.1
6.8 ( 0.7
8.8 ( 1.6
21 ( 6.7
6.8 ( 1.5
2.4 ( 5.7
a All values calculated from 10 mg/kg po unless otherwise stated. b Half-life in (h). c Maximum exposure in (μM). d Time at maximum exposure (h).
e Volume of distribution in L/kg (calculated from 1 mg/kg IV dose to mice. f Clearance in (mL/min/kg). Calculated from 1 mg/kg IV dose to mice (n = 4).
g Exposure calculated as t0finfinity in (μM h) following 10 mg/kg po dose to mice (n = 5). h Bioavailability calculated from 1 mg/kg IV dose and 10 mg/
kg po dose. Compound 2 values calculated from n = 3 experiments (both 1 mg/kg IV and 10 mg/kg po). Compound 4 values determined from n = 2
experiments (both 1 mg/kg IV and 10 mg/kg po).
3
The enhanced pharmacology observed with 2 relative to 1 was
observed despite a reduction in overall exposure. This ob-
servation indicated that the enhanced lymphopenia observed
with 2 is likely a consequence of increased intrinsic potency
against S1PL.
That oxime 2 generated substantial lymphopenia suggested
the possibility that other analogues displaying a sp2-hybri-
dized heteroatom at this position might also provide potent
analogues. We considered that heterocycles containing an
appropriately oriented CdN bond within a heterocyclic ring
could provide pharmacologically active compounds. To test
this hypothesis, isoxazole 4, the aromatic analogue of 2, was
orally administered to mice at 30 mg/kg and resulted in a
dramatic reduction in circulating lymphocytes. Pharmaco-
kinetic evaluation of 4 showed up to a 3-fold enhancement in
exposure over oxime 2. The more highly saturated dihydro-
isoxazole 5 showed less activity than the aromatic congener.
Regioisomeric isoxazole 8 proved to be an inactive com-
pound. Investigation into the effect of the isoxazole substitu-
tion revealed that addition of alkyl groups at the 5-position
detracted from activity. As is apparent from methyl isoxazole 9,
cyclopropyl isoxazole 10, isopropyl isoxazole 11, and phenyl
isoxazole 12, potency generally decreased with increasing
steric bulk vicinal to the isoxazole oxygen. Because this series
of compounds displayed comparable levels of exposure, the
effect on potency probably relates to a disfavored steric

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24 8653
Table 3. Pharmacokinetic Parameters of Compound 2 in Multiple
Species
speciesa
Dcse
Tmax
Cmax
AUCe
Vssf
CLg
%F h
b
c
d
rat
dog
10
50
2.0
2.0
2.3
2.0
2.7
79.5
16.9
3
18.7
323.0
70.1
0.6
0.3
0.5
10.2
3.5
26.0
32.1
16.4
monkey
human
30
4.8
∼1
19.0
a Male Sprague-Dawley rats; Beagle breed dogs; cynomolgous mon-
keys; healthy human volunteers. b Oral dose in mg/kg. c Time at maxi-
mum exposure (h). d Maximum exposure in (μM). e Exposure calculated
as t0finfinity in ( μM h) following indicated dose. f Volume of distribu-
3
tion in L/kg. g Clearance in (mL/min/kg). Calculated from 1 mg/kg IV
dose h Bioavailability calculated from 1 mg/kg IV dose and indicated
oral dose.
interaction in this region. Introduction of electron withdraw-
ing groups at the 5-position, as in trifluoromethyl isoxazole 13,
significantly reduced in vivo potency. Introduction of an elec-
tron donating group, as presented by ether 14, also proved
fruitless. Probing for a potential donor-acceptor interaction
vicinal to the isoxazole heterocycle, alcohol 15 and amine 16
were also studied. In both cases, diminished activity suggested
that no such interaction was realized.
While substitution at the 5-position of the pendant iso-
xazole heterocycle uniformly reduced potency, the effect of
substitution at the 4-position was variable. Comparison of
isoxazole 17 with isoxazole 9 indicates that the acetyl group
reduces the degree of lymphopenia after oral dosing at
30 mg/kg. Nitro-arene 18 exhibited modest in vivo activity
despite a slight decrease in exposure. 4-Substitued alkyl 19
triggered notable lymphopenia with good exposure. Electron
withdrawing groups at the 4-position tended to reduce po-
tency, while alkyl substitution was generally tolerated.
Figure 1. Pharmacokinetic and pharmacodynamic relationship of
1 in mice. (A) Pharmacokinetic analysis of 1 following 100 mg/kg
PO dose to male C57BL/6 albino mice (n = 4). (B) Lymphocyte
levels now normalized to vehicle control, *p < 0.05; single 100 mg/kg
dose was given to mice (n = 5). Change in circulating peripheral
lymphocytes is shown for mice treated with 1 at time points follow-
ing oral gavage. Peak lymphopenia observed with 1 occurs over ten
hours past peak plasma levels.
pharmacology, as assessed by reduced peripheral circulating
lymphocyte numbers. The nadir in blood lymphocytes oc-
curred 12-24 h following administration of 1, following
the peak plasma levels of 1 by more than 10 h (Figure 1).8,12
Data for compounds 2 and 4 was in agreement with this
observation. In fact, at the 18 h time point, when peripheral
lymphopenia was at or near its maximum, 2 or 4 con-
centration in plasma was no longer detectable. Small polar
molecules like 2 often achieve tissue concentrations by para-
cellular absorption.13 Intracellular concentrations of drug
are likely AUC driven, and the exposure achieved by in vivo
administration is sufficient to inhibit S1PL activity in immune
tissues that results in a gradual increase of S1P levels within
these tissues.8 Following prolonged S1PL inhibition, the
increased S1P concentration (i.e., the S1P gradient) triggers
receptor internalization which limits lymphocyte egress from
secondary lymphoid tissue and therefore acts locally to lower
levels of lymphocytes in peripheral circulation. In addition,
the delayed pharmacodynamic effect of S1PL inhibition is
consistent with both the intracellular location of S1PL and
its slow turnover rate for S1P decomposition.14 By contrast,
administration of the direct S1PR agonists like FTY-720
leads to a more rapid onset of pharmacology.15
To determine the relationship between S1P tissue concen-
tration and the pharmacodynamic profile of 2, a compre-
hensive analysis of tissue S1P content and peripheral blood
lymphocyte levels was performed in Sprague-Dawley rats
(Figure 2). Consistent with our dosing protocol in mice, 2
(30 mg/kg) or vehicle was orally administered daily for 3 days
and multiple tissues were assayed for S1P content 18 h
after the final oral administration of 2. The tissues in-
cluded thymus, spleen, jejunum, lymph node, heart, eye,
liver, kidney, and lung. CBC readings at 18 h taken immedi-
ately before tissue collection correlated with reduced levels of
Owing to its superior in vivo activity following oral admin-
istration to mice, excellent solubility, and overall favorable
physical properties, including pharmacokinetic properties,
oxime 2 was further profiled. Following its discovery, iso-
xazole 4 was similarly profiled. Compounds 2 and 4 are both
low molecular weight, polar compounds (MW 245, 255,
respectively, PSA ∼ 140, AlogP ∼ -1.5), and their HCl salts
were both easily formulated in neutral water. As shown in
Table 2, a more detailed pharmacokinetic analysis revealed
that both compounds show low clearance and a low volume
of distribution. Additionally, they were characterized by a
short half-life and quickly reached maximum concentrations
in blood. Plasma concentrations of 2 and 4 were sufficiently
sustained, and the AUC values were 8.8 μM h and 21.0 μM h,
3
3
respectively, at 10 mg/kg oral dose. Owing to their PK/PD
profile and performance against multiple animal models of
arthritis (vide infra), compound 2 was nominated for clinical
development. Subsequently, isoxazole 4 demonstrated out-
standing properties and was nominated for clinical develop-
ment as a backup compound to 2.
To facilitate advancement of2 to clinical evaluation, a more
detailed pharmacokinetic profile was obtained in multiple
species (Table 3). Compound 2 was characterized by low
clearance and low volume of distribution following pharma-
cokinetic studies in rodents, dogs, monkeys, and, ultimately,
humans. Maximum plasma concentrations occurred at ap-
proximately 2 h following oral administration. Exposure
generally increased with higher species, showing the highest
exposure in humans.
Pharmacodynamic Aspects of S1PL Inhibition. The PK-
PD relationship for 1 demonstrated a significant time delay
between oral administration of the compound and onset of

8654 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24
Bagdanoff et al.
Figure 2. The lymphoid system is the primary target of compound 2 activity. Male Sprague-Dawley rats were treated with 2 or vehicle (n = 10
per tissue type, n = 10 for CBC). Group dosed orally with 2 at 30 mg/kg qd for 3 days. Blood and tissues were harvested 18 h after final dose.
Blood lymphocyte counts were measured by CBC analysis.
peripheralcirculatinglymphocytes. Thenadir of lymphopenia
was concomitant with increased S1Pconcentration (Figure 2).
S1P levels in thymus, spleen, and jejunum increased approxi-
mately 80-, 24-, and 45-fold, respectively, in rats treated with 2
over tissues from untreated rats. By contrast, several other
nonlymphoid tissues, including the heart and eye, displayed
no increase in S1P levels. Cardiac and ophthalmic tissues were
of particular interest because adverse pharmacology is man-
ifested in these tissues following administration of a S1PR
agonist.4 The observed tissue-selective increase in endogenous
S1P resulting from S1PL inhibition distinguishes this target
from synthetic S1P receptor agonists that have the capacity
to modulate S1P receptors in a wide variety of tissues.12 The
lymphoid-tissue selective increase in S1P in response to S1PL
inhibition is likely due to differential expression profiles of S1PL
and other S1P regulating enzymes (i.e., sphingosine kinases,
S1P phosphatase, S1P transporter) across the spectrum of tissue
types.21b For example, while S1PL is widely expressed, S1P
homeostasis in lymphoid tissues may be particularly sensitive
to inhibition of S1PL activity due to reduced representation
of mechanisms that decrease S1P levels. Furthermore, that a
localized increase in S1P concentration does not lead to sub-
stantial increase in systemic S1P levels suggests that S1P
exerts a relatively local effect primarily in lymphoid tissues.
Another mechanistic nuance in this signaling pathway
relates to the relationship between 1 and S1PL. In accord
with previous studies with 1, compounds 2 and 4 showed no
direct inhibition of S1PL when tested in our in vitro assay. To
date, in vitro assay conditions required to determine direct
S1PL inhibition by 1 or analogues thereof, remain elusive.6,8
Previous mechanistic studies with 1 revealed that oral ad-
ministration led to maximum lymphopenia (12-24 h after
oral dose, measured in blood by CBC) and correlated with a
concomitant decrease in S1PL activity and enhanced S1P
levels in lymphoid tissue.8 This suggests that the established
in vivo efficacy is potentially due to a more complex scenario
than a simple binary ligand-protein interaction and may
result from a higher-order complex. Described here, mecha-
nistic studies in rat with 2 showed enhanced S1P levels mainly
in lymphoid tissue at peak lymphopenia (Figure 2). Accord-
ingly, we performed comparative analysis of analogues of 1
directly by in vivo analysis using CBC measurements of
peripheral lymphocyte levels at 18 h.
most heavily relied on in vitro methods during the last two
decades. However, direct in vivo testing has been of tremen-
dous benefit to the drug discovery process historically, from
aspirin16 and penicillin17 to the development of modern
imaging and X-ray contrast agents.18 A more recent example
of in vivo screening methods in modern drug discovery can
be found in the S1P signaling pathway itself. In 1995, Adachi
andco-workers19 applied ex vivoand in vivoscreening tools to
the SAR of myriocin,20 a serine palmitoyltransferase inhibitor
with potent immunosuppressant activity.21 Continued in vivo
screening of analogues, without application of an in vitro
screening tool, led to the discovery of FTY720 (fingolimod).22
Interestingly, in vitro analysis of this analog revealed no
inhibition of serine palmitoyltransferase. Subsequently, it
was discovered that in vivo phosphorylation of fingolimod
was necessary for its immunosuppressant activity. Ultimately
it was determined that phosphorylated fingolimod is a direct
S1P receptor agonist. Since myriocin and fingolimod act on
different biological targets, in vitro screening alone could not
have predicted the striking invivo potency of fingolimod. Oral
administration of these analogs, in vivo pharmacology experi-
ments, and ex vivo analysis of blood samples by CBC were
therefore crucial to the discovery of fingolimod.
Pharmacology. Mice with reduced levels of S1PL activity
due to genetic modification or treatment with 1 present an
altered pattern of lymphocyte trafficking and changes in
tissue S1P concentration,1a analogous to manipulation of the
S1P/S1PR axis by fingolimod.23 Importantly, the physiolog-
ical effects of inhibiting S1PL activity by more than 90% is
limited to the immune system.1 Oral dosing with 2 or 4
recapitulated the changes in lymphocyte trafficking and tissue
S1P concentrations observed in genetically modified mice
expressing reduced levels of S1PL activity. These results sug-
gested that immune modulation achieved by treatment with
compounds 2 or 4 may be of therapeutic benefit in inflamma-
tory and autoimmune diseases. Dose escalation studies indi-
cated that single oral doses of 30-100 mg/kg 2 administered to
mice induced 40-60% lymphopenia, while lower doses have a
minimal to nonstatistically significant effect (Figure 3).
The effect of 2 and 4 on disease development and progres-
sion was assessed in a collagen-induced arthritis (CIA) model
of rheumatoid arthritis (RA) in mice, using prophylactic and
therapeutic dosing regimes. In vivo models of collagen-
induced arthritis using prophylactic dosing of 100 mg/kg 2
resulted in a significant delay in the onset of arthritis; treated
Direct in vivo evaluation of potential drug candidates is
less common to the current drug discovery paradigm that has

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24 8655
Figure 3. Compound 2 treatment induces up to 60% depletion of
circulating lymphocytes. Mice were treated with the indicated single
oral doses of 2, and blood lymphocyte counts were measured 18 h
after dosing. Data were pooled from three independent experi-
ments, giving similar results and represent 8-10 mice each cohort.
Data are presented as mean ( SEM; *p < 0.05.
mice had only a minimal increase in clinical scores and joint
swelling (Figure 4A). As expected from its effect on T-cell
trafficking, mice treated with 2 showed an approximately
40-60% decrease in the number of circulating T cells and an
increased percentage of CD4 and CD8 single positive cells in
the thymus at the conclusion of the experiment (Figure 4B).
Microscopic analysis of tarsal sections showed that treat-
ment with 2 significantly reduced synovial hyperplasia, in-
flammation, erosion, and exudate formation (Figure 4C).
Statistical analysis of histopathology scores confirmed that
treatment with 2 effectively reduced joint damage, hyperpla-
sia, and immune cell infiltration compared to vehicle control.
The effect of 2 and 4 was further assessed in CIA using a
therapeutic dosing regime whereby 2, 4, or vehicle control was
administered once 50% of the animals within the group became
symptomatic. Using this dosing paradigm, both 2 and 4 dosed
at 30 mg/kg gave significant relief from disease as measured by
clinical scores and joint swelling, compared to vehicle control
(Figure 5A). Therapeutic treatment with 2 or 4 at a dose which
induces minimal lymphopenia (5 mg/kg, Figure 3) also yielded
a trend of reduced disease activity (Figure 5A).
The disease-modifying efficacy observed in the presence of
relatively mild lymphopenia during both prophylactic and
therapeutic treatments indicates that the potent anti-inflam-
matory effect of S1PL inhibition cannot be explained solely
by the effect on the levels of circulating lymphocytes.
It is important to note that a 40% decrease in circulating
lymphocyte numbers alone is not expected to affect mouse
inflammatory challenge models. Several other gene knock-
out lines have been analyzed at Lexicon that exhibited sig-
nificant lymphopenia without resistance to inflammatory
challenge (data not shown). While alteration of lymphocyte
trafficking likely plays a significant role in immunosuppres-
sion via S1PL inhibition, the increased S1P levels may also
affect other biological processes, in addition to cell migra-
tion, contributing to the observed anti-inflammatory effect.
These processes may include the inflammatory response of
cell types as diverse as endothelial cells,24 macrophages,25
and NK cells.26 Importantly, the disease-modifying effect of
2 and 4 was not due to general immunosuppression because
the antibody responses to the injected CII antigen were not
blocked by treatment with compound (Figure 5B). Taken
together, the results suggest that inhibition of S1PL activ-
ity can achieve significant anti-inflammatory effect without
hampering the immune response completely. Of note, the
Figure 4. Compound 2 prevents arthritis development in the mouse
collagen-induced arthritis model. (A) Average values for clinical
scores (left panel) and joint swelling (right panel) taken from the
hind limbs for vehicle control (n = 16) and compound 2 (n = 13)
arms are shown with vertical bars representing SEM. X axis values
represent days after collagen immunization. Daily oral dosing was
initiated on day -1; the vehicle control mice were given 10 mL/kg
water, and the compound 2 cohort was given 100 mg/kg compound
in 10 mL/kg water. *p < 0.05. (B) Analysis of CD4 and CD8 single
positive T cells in blood and thymus as indicated. Bars represent
average values, black for vehicle control, and white for 2; SEM is
indicated by the vertical lines. Blood data were obtained from
13-16 animals per cohort; two animals per arm were used for
thymus staining; numbers above bars indicate p values. (C) H&E
staining of representative tarsal joint sections from control (left) and
2 treated (right) animals shown at 10ꢀ magnification. Regions of
synovial hyperplasia, inflammation, and erosion are highlighted by
small arrows (minimal in 2 sample).
murine CIA results were further supported by analysis of
efficacy of compounds 2 and 4 in the rat adjuvant-induced
model of RA, a particularly aggressive model of RA.27
Telemetry Studies. Given that bradycardia was a clinical
observation associated with fingolimod,4a the effect of 2
administration on heart rate was assessed in multiple species.
Assessment of cardiovascular effects of 2 in dogs and pri-
mates was based on measurements of electrocardiogram
(ECG) and hemodynamic parameters (blood pressure and
heart rate). Cynomolgus monkeys and dogs were adminis-
tered single ascending doses up to 1000 mg/kg and 500 mg/kg,
respectively. Compound 2 did not affect heart rate at the
highest dose studied, in either dog or in primate (Figure 6).
These data correlate with the observation that inhibition
of S1PL affects S1P concentration primarily in lymphatic
tissues.
Phase I Clinical Trial Data for Compound 2. On the basis
of the preclinical data demonstrating a favorable safety

8656 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24
Bagdanoff et al.
Figure 5. Therapeutic dosing of 2 and 4 reduces severity of cia symptoms in mice. (A) Average values for joint swelling (left panel) and clinical
scores (right panel) taken from the hind limbs for vehicle control (circles) and 2 (triangles) arms are shown with vertical bars representing SEM.
X-axis represent days after collagen immunization. Daily oral dosing was initiated on day 20 (arrow), Vehicle control mice were given 10 mL/kg
water; 2 was given at the indicated doses po in 10 mL/kg water. n = 9-10 each cohort; *P value <0.05. (B) Serum anti-CII Ab titers of mice
from the study presented in (A) were measured by ELISA at the indicated time points after initial collagen immunization.
Figure 6. Telemetry studies of compound 2 in primate and dog. Ten male Cynomolgus monkeys (Macaca fascicularis) were given a single
administration of 1000 mg/kg using a Latin square design by oral gavage. Electrocardiogram (ECG), blood pressure, and heart rate were used
to monitor cardiovascular and hemodynamic parameters. Telemetry data collection began 90 min prior to dosing and continued at the
indicated time points for 24 h. Four female Beagle dogs were given a single dose administration of 500 mg/kg using a Latin square design by oral
gavage. Data collection commenced 24 h prior to dose administration. One-minute tracings of ECGs were obtained at approximately 15 min
prior to dosing and at time intervals shown.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24 8657
Experimental Procedures
Animal Handling and in Vivo Methods. All procedures were
conducted according to applicable standard operating proce-
dures and with the approval of the Lexicon Pharmaceuticals,
Inc. Institutional Animal Care and Usage Committee. For PK
analysis, the tested analogues were formulated in sterile water
and dosed to 129/SvEvBrd ꢀ C57BL/6 F1 mice. The pH of
intravenous dosing solution was adjusted to between 5 and 6 and
between 3 and 4 for oral administration. The intravenous dose
volume was 3.7 mL/kg body weight, and the oral gavage volume
was 5 mL/kg body weight at 1 and 10 mg/kg body weight,
respectively. Animals were observed regularly throughout the
day on days in which they were dosed, and these clinical
observations were recorded. For the purpose of exposure, blood
was collected using saphenous vein bleeds in unanaesthetized
animals, except for terminal bleeds, which were performed using
cardiac puncture immediately following euthanasia. Blood
samples were collected at 0.08, 0.25, 0.5, 1, 3, and 6 h post
intravenous dosing and at 0.25, 0.5, 1, 2, 4, 6, and 24 h post oral
administrations. An aliquot of 25 μL of blood in K2-EDTA
tubes was spun following each time point, and 10 μL of the
resulting plasma was transferred to fresh polypropylene tubes
for exposure assessment by liquid chromatography-mass spec-
trometry according to the following experimental LC/MS pro-
tocol: Compound 2 levels in plasma were determined using a
Thermo Finnigan TSQ Quantum Ultra AM, Thermo Finnigan
Sureyor MS LC Pump, and CTC Analytics HTC PAL auto-
sampler. A Phenomenex Luna phenyl-hexyl, 4.6 mm ꢀ50 mm
column was used for chromatography. The mobile phases used
in the gradient chromatography method were: A, water with
0.1% formic acid; B, acetonitrile with 0.1% formic acid. The
gradient had the following time profile: 0 min, 95%A; 0.5 min,
95% A; 1 min, 5%A; 2.5 min, 5%A; 2.51 min, 95%A; 3 min,
95%A. The chromatographic flow rate was 0.7 mL/min and
the injection volume was 15 μL. Positive ion ESI LC/MS/MS
analysis was used for detection of the analyte. The multiple
reaction monitoring transition for 2 detection was 246 f 210 m/z,
collision energy was 14 eV, and the tube lens energy was 77 V.
Determination of sphingosine 1-phosphate levels were made
according to protocol published in detail in the supporting
online material by Schwab and Cyster.6
Cardiovascular Evaluation of 2: Administration by Oral Ga-
vage to Telemetry-Instrumented Conscious Nonhuman Primate.
Ten male Cynomolgus monkeys (Macaca fascicularis) were
assigned to study and given a single oral administration of 2
dosages using a Latin square design. Assessment of cardiovas-
cular effects was based on measurements of electrocardiogram
(ECG) and hemodynamic parameters (blood pressure and heart
rate). The general health of the animals was assessed based on
body weights, clinical signs, and intra-abdominal body tem-
perature. Telemetry data were collected for at least 90 min prior
to dosing and continuously for at least 24 h postdose. Quanti-
tative ECG analysis was completed for baseline and at time
intervals post dose indicated in Figure 6.
Cardiovascular Effects of 2 Administered to Conscious Tele-
metered Dogs. Four female Beagle dogs were assigned to study
and given a single administration of all dosages using a Latin
square design. Animals were administered various single doses and
dosed via oral gavages. The data collection commenced approxi-
mately 24 h prior to dose administration and continued for
approximately 24 h post dose. Data was collected continuously.
One-minute tracings of the ECGs were obtained at approximately
15 min prior to dosing and at intervals indicated in Figure 6.
Collagen-Induced Arthritis. For experiments utilizing prophy-
lactic dosing of test compounds a synovial autoimmune response
was induced in 129/SvEvBrd ꢀ C57BL/6 F1 mice as previously
described.28 Studies utilizing the therapeutic dosing regime were
performed using DBA1/j mice and CFA containing a reduced
concentration of Mycobacterium tuberculosis (4 mg/mL).
Figure 7. Phase 1a clinical trial data: lymphocyte counts are de-
creased after treatment with compound 2, with recovery 48 h after
nadir. (A) Dose-dependent reduction in lymphocyte counts ob-
served after single dose phase 1a trial: 159 patients total in single
and multiple ascending dose studies; 120 active dosing regimens and
39 placebo regimens, participated in randomized, double-blinded
study. (B) Rapid reduction of blood lymphocytes from a single
125 mg dose (N = 6 patients) followed by a return to baseline in 48 h
after nadir.
and efficacy profile of LX2931, phase 1 clinical trials were
initiated to determine safety in human subjects. As a surro-
gate biomarker of S1PL inhibition, blood lymphocyte pop-
ulations were determined by CBC analysis. As shown in
Figure 7A, a single ascending dose provided a clear dose-
responsive relationship, resulting in up to a ∼50% decrease
in peripheral lymphocytes at the highest administered dose
of 180 mg. The compound was generally well tolerated in
all dose groups. Importantly, the induced reduction in cir-
culating blood lymphocytes (N = 6, 125 mg single dose) was
reversible (Figure 7B); lymphocyte populations rebounded
to predose levels by 48 h after nadir.
Conclusion
S1PL appears to play a more significant role in controlling
the S1P gradient in lymphoid tissues than in other tissues. This
local effect has the systemic impact of lowering levels of
lymphocytes in peripheral circulation. Increased S1P levels
after inhibition of S1PL may also affect secondary mecha-
nisms that contribute to anti-inflammatory activity. Explora-
tion of the SAR of 1 and identification of the pendant 2-acetyl
group was crucial to the discovery of 2. Recognition of the
importance of the oxime was the basis for the rational design
of the isosteric isoxazole contained within 4. Compound 2 has
shown robust efficacy in multiple disease models of rheu-
matoid arthritis. Peripheral lymphocyte levels have been
similarly reduced by oral administration of 2 across multi-
ple species studied, from mouse to man. Phase II trials
are currently underway in patients with active rheumatoid
arthritis. Results on these studies will be reported as they
become available.

8658 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24
Bagdanoff et al.
Detection of Anti Type II Collagen (CII) Antibodies. Total
CII-specific IgG antibody levels were determined in preimmu-
nization and day 35 sera by ELISA. Briefly, 96-well Nunc
Maxisorp flat-bottomed plates (Fisher, Pittsburgh, PA) were
coated with 100 μL/well of 2 μg/mL chicken collagen II (Sigma,
St Louis, MO) in PBS and incubated overnight at 4 °C. Plates
were subsequently washed four times with 100 μL of PBS/Tween
per well and then blocked for 1 h with 100 μL of PBS containing
10% FBS. Then 100 μL of a 1:25000 dilution of sera was loaded
into wells and incubated for 1 h at room temperature. Wells
were washed as before and 100 μL of a 1:5000 dilution of
horseradish peroxidase-conjugated goat antimouse total IgG
(HþL) (Southern Biotech, Birmingham, AL) was added to each
well and incubated at RT for an hour. Plates were washed a final
time and peroxidase activity was assessed using the 3,30,5,50-
tetramethylbenzidine (TMB) liquid substrate system for ELISA
per manufacturer’s instructions (Sigma, St Louis, MO). OD450 was
determined using a Vmax kinetic microplate reader (Molecular
Devices, Sunnyvale, CA). The concentration of CII-specific anti-
body was expressed as the OD450 of the sample minus the OD450 of
the blank.
Chemical Methods. All reactions were conducted under a
static atmosphere of argon or nitrogen and stirred magnetically
unless otherwise noted. Reagents, starting materials, and sol-
vents were purchased from commercial suppliers and used as
received. Flash column chromatography was carried out using
prepacked silica gel columns from Biotage or ISCO or by slurry
preparation using EMD Silica Gel 60 (particle size 0.040-
0.063 mm). 1H and 13C NMR spectra were collected on Bruker
ARX300, DRX400, or DPX400, or Varian Mercury 400 MHz
NMR spectrometers. Chemical shifts are reported in parts per
million (ppm) relative to tetramethylsilane in δ-units, and
coupling constants (J-values) are given in hertz (Hz). Data are
reported in the following format: chemical shift, multiplicity,
coupling constants, and assignment. Reactions were monitored
by TLC using 0.25 mm E. Merck silica gel plates (60 F254) and
were visualized with UV light. Analytical HPLC spectra were
collected on Shimadzu HPLC systems equipped with a UV
detector measuring absorbance at 220 and 254 nm. Purity was
determined to be g95% purity for all compounds contained
within this manuscript using a Shimadzu analytical HPLC by
the method specified in the experimental procedure. Mass
spectra were obtained on Waters ZQ or ZMD LCMS systems
equipped with an autosampler, and ELSD detector, a UV
detector measuring absorbance at 220 and 254 nm, and a mass
detector. High resolution mass spectra were obtained on a
Waters LCT Premier XE Micromass MS Technologies instru-
ment equipped with an autosampler. Elemental analysis was
conducted by Robertson Microlit Laboratory, Madison, NJ.
(E)-1-(4-((1R,2S,3R)-1,2,3,4-Tetrahydroxybutyl)-1H-imida-
zol-2-yl)ethanone Oxime (2). To a flask charged with 1 (4.88 g,
21.20 mmol) was added water (25 mL) and 1 N aqueous HCl
(21.2 mL, 21.2 mmol). After all solids dissolved, a solution of
trityl hydroxylamine (7.00 g, 25.44 mmol) in dioxane (55 mL)
was added and the reaction was maintained at 50 °C for 4 h. At
completion, the reaction was cooled to room temperature and
the solution was adjusted to pH = 7 by addition of 1N aqueous
NaOH . The neutralized solution was then concentrated to a
plastic mass, which was purified by flash chromatography on
silica gel [10% MeOH/1% NH4OH (5 wt % solution in water)
in DCM] to provide the trityl-ether as clear plastic. The plastic
mass was azeotroped with hexane and concentration provided
a white foam, which could be dried under vacuum to a solid.
To a vigorously stirred, room temperature solution of the
obtained intermediate trityl oxime-ether (4.8 g, 10 mmol) in
dioxane (90 mL) was added a solution of HCl in dioxane (4.0 M,
60 mL). After a few minutes, a white precipitant was observed,
and stirring was continued for a total of 30 min before filtering
over a fritted glass filter and rinsing the cake with dioxane and
ether. The cake was redissolved in water (200 mL), sonicated for
5 min then cooled to 0 °C, treated with Celite (5 g) and filtered
over a fritted glass filter. The aqueous solution was concentrated
to dryness then triturated with methanol:diethyl ether (30 mL:
60 mL) to provide the E-oxime as an analytically pure white
powder (3.8 g, 79% yield, 2 steps). 1H NMR (400 MHz, 1 M DCl
in D2O) δ 7.47 (d, J = 0.50 Hz, 1 H), 5.20 (dd, J = 2.26 Hz, 0.75,
1H), 3.85 (dd, J = 13.05, 2.76 Hz, 1H), 3.82 (dd, J = 13.05, 2.51
Hz, 1H), 3.76 (dd, J = 8.53, 2.26 Hz, 1H), 3.65-3.72 (m,
1 H) 2.30 (s, 3H). 13C NMR (101 MHz, 1.0 M DCl in D2O)
δ146.85, 144.03, 138.02, 119.94, 75.59, 73.60, 67.49, 65.77, 13.5.
MS (EI) m/z: 246 [M þ H]þ. HRMS calcd for C9H15N3O5
[MH]þ 246.1090, found 246.1082; mp 101 °C. HPLC (Zorbax
4.6 mm ꢀ 150 mm, 2-40% [methanol/water/TFA (95/5/0.1)]:
water, 6 min gradient) tR = 1.15 min, 100% integrated area.
(1R,2S,3R)-1-(2-(Isoxazol-3-yl)-1H-imidazol-4-yl)butane-
1,2,3,4-tetraol (4). A solution of trityl hydroxylamine (15.15 g,
55 mmol) in dioxane (120 mL) was added to a solution of 1-HCl
(13.30 g, 50 mmol) in water (60 mL), and the reaction was
maintained at room temperature with vigorous stirring for 4 h.
At completion, the reaction was quenched with excess triethyl-
amine (10 mL) and then concentrated and flashed over silica gel
(5-10% MeOH/DCM) to provide an intermediate plastic mass
that was azeotroped with toluene (2 ꢀ 200 mL). The plastic mass
was redissolved in 1,1-dichloroethane (250 mL) and 2,2-di-
methoxypropane (250 mL) and then treated with p-TsOH-
H2O (1.28 g, 7.72 mmol) and heated to 60 °C for 16 h. At com-
pletion, the reaction was diluted with dichloromethane (500 mL)
and washed with satd aq NaHCO3 (2 ꢀ 200 mL), water (200 mL),
and brine (200 mL) and then dried over MgSO4 and concen-
trated. The resulting plastic mass was azeotroped with toluene
(200 mL) and then redissolved in anhydrous dioxane (300 mL)
and treated with 4 N HCl (150 mL) in anhydrous dioxane. After
maintaining at room temperature for 1 h, the reaction was
concentrated, then flashed over silica gel (3-10% MeOH:
DCM eluent) to provide 2 (7.87 g, 48% yield, 3 steps).
A solution of 2 (1.65 g, 5.08 mmol) in THF (50 mL) was
cooled to -15 °C and then treated with a 1.6 M solution of
n-BuLi in hexanes (18.5 mL, 29.60 mmol). The reaction was
maintained for 30 min and then quenched with anhydrous DMF
(3.92 mL, 50.8 mmol). The reaction was allowed to warm to
room temperature for 30 min and then quenched with satd aq
NH4Cl (50 mL) and diluted with EtOAc (200 mL). The layers
were separated, and the organics were washed with water
(100 mL) and then brine (50 mL) before drying over MgSO4
and concentrating. The resulting crude material was purified by
flash chromatography over silica gel (3-8% MeOH/DCM
eluent) to provide a plastic mass (1.10 g, 3.12 mmol). The
material was redissolved in THF (10 mL) and then cooled to
0 °C and treated sequentially with pyridine (0.95 mL, 12.48
mmol) and trifluoroacetic acid anhydride (0.52 mL, 3.74 mmol).
The reaction was then warmed to 50 °C for 16 h and then
concentrated and flashed over silica gel (5-8% EtOAc:hexanes
eluent). The resulting material was dissolved in dioxane (10 mL)
and treated with 1 M aq HCl and heated to 50 °C for 2 h, then
concentrated. The solid was dissolved in water (10 mL) and then
lyophilized to provide 4 as a white powder (321 mg, 22% yield, 3
steps). 1H NMR (400 MHz, 1.0 M DCl in D2O) δ 8.70 (d, J =
1.8 Hz, 1 H), 7.41 (s, 1 H), 6.88 (d, J = 1.5 Hz, 1 H), 5.07 (s, 1 H),
3.72-3.43 (m, 4 H). 13C NMR (101 MHz, 1.0 M DCl in D2O) δ
ppm 162.23, 148.64, 136.24, 133.98, 117.83, 103.47, 72.59, 70.51,
64.48, 62.73. MS (EI) m/z: 256 [M þ H]þ. HRMS calcd for
C10H13N3O4S [M þ H]þ 256.0933, found 256.0919. HPLC
(Zorbax C8 4.6 mm ꢀ 150 mm, 5% [methanol/water/TFA
(95/5/0.1)]:water, 5 min isocratic) tR = 1.04 min, 100% inte-
grated area.
General Procedure A: (1R,2S,3R)-1-(2-(Isoxazol-3-yl)-1H-
imidazol-4-yl)butane-1,2,3,4-tetraol (4). To a room temperature
solution of the nitrile29 (80.0 g, 850 mmol) in MeOH (1700 mL)
was added MeONa solution in MeOH (25% w/w, 110 mL, 510
mmol). After 1 h, the imidate formation was deemed complete

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24 8659
by 1H NMR analysis of a reaction aliquot in MeOH-d4. At this
time, fructosamine-HOAc (202.4 g, 850 mmol) was added to
the reaction in one portion. After 18 h, another portion of
MeONa solution in MeOH (25% w/w, 92.0 mL, 425 mmol) was
added to the heterogeneous reaction mixture. The reaction was
maintained at room temperature with vigorous stirring for 24 h,
then diluted with water (500 mL) and acidified to pH = 2 with
conc aq HCl. The reaction was concentrated and then redis-
solved in water (800 mL) before adjusting to pH = 10 with 10 M
aq NaOH. The precipitate was collected by filtration and rinsed
with water. The precipitate was acidified with water and conc
HCl to pH∼3, and then the freely soluble solution was concen-
trated to dryness. The resulting solid was azeotroped with
toluene (2 ꢀ 1 L) to remove trace water. The solid was ∼98%
pure at this stage but was purified further by dissolving in a
minimum volume of hot MeOH and then crystallized after
addition of diethyl ether and cooling. Filtration and drying on
high vacuum afforded the HCl salt of 4 as a white crystalline
powder (240 g, 97% yield). All analytical data was identical with
the material generated by the initial medicinal chemistry route.
(1R,2S,3R)-1-(2-(5-Phenylisoxazol-3-yl)-1H-imidazol-4-yl)-
butane-1,2,3,4-tetraol (12). Prepared according to the general
procedure A from 5-phenylisoxazole-3-carbonitrile (170 mg,
1.00 mmol) and fructosamine-HOAc (238 mg, 1.00 mmol) to
provide 12 as the HCl salt (327 mg, 89% yield). 1H NMR (300
MHz, 1.0 M DCl inD2O) δ ppm 7.74-7.83 (m, 2 H), 7.49 (s, 1 H),
7.39-7.46 (m, 3 H), 7.15 (s, 1 H), 5.12-5.17 (m, 1 H), 3.55-
3.79 (m, 4 H). MS (EI) m/z: 332 [M þ H]þ. HRMS calcd for C16-
H17N3O5 [M þ H]þ 332.1246, found 332.1259. HPLC (Zorbax
4.6 mm ꢀ 150 mm, 10-90% [methanol/water/TFA (95/5/0.1)]:
water, 6 min gradient) tR = 4.30 min, 98% integrated area.
(1R,2S,3R)-1-(2-(3-Methylisoxazol-5-yl)-1H-imidazol-4-yl)-
butane-1,2,3,4-tetraol (8). Prepared according to the general
procedure A from 3-methylisoxazole-5-carbonitrile (324 mg,
3.00 mmol) and fructosamine-HOAc (714 mg, 3.00 mmol) to
provide 8 as the HCl salt (610 mg, 66% yield). 1H NMR (400
MHz, 1.0 M DCl in D2O) δ ppm 7.38 (d, J = 1.0 Hz, 1 H), 6.86
(s, 1 H), 5.04 (dd, J = 2.0, 0.8 Hz, 1 H), 3.52-3.66 (m, 3 H),
3.42-3.50 (m, 1 H), 2.16 (s, 3 H). 13C NMR (101 MHz, 1.0 M
DCl in D2O) δ ppm 162.25, 153.36, 136.28, 132.20, 117.85,
107.89, 72.62, 70.52, 64.49, 62.73, 10.36. MS (EI) m/z: 270 [M þ
H]þ. HRMS calcd for C11H15N3O5 [M þ H]þ 270.1090, found
270.1081. HPLC (Luna Phenyl-hexyl 4.6 mm ꢀ 50 mm, 5%
[10 mmol aq NH4OAc:MeCN, isocratic) tR = 0.73 min, 100%
integrated area.
(1R,2S,3R)-1-(2-(5-(Trifluoromethyl)isoxazol-3-yl)-1H-imid-
azol-4-yl)butane-1,2,3,4-tetraol (13). Prepared according to
the general procedure A from 5-(trifluoromethyl)isoxazole-3-
carbonitrile (405 mg, 2.5 mmol) and fructosamine-HOAc
(595 mg, 2.5 mmol) to provide 13 as the HCl salt (718 mg, 80%
yield). 1H NMR (300 MHz, 1.0 M DCl in D2O) δ ppm 7.47 (d,
J = 0.8 Hz, 1 H), 7.45 (d, J = 0.8 Hz, 1 H), 5.09 (d, J = 1.0 Hz,
1 H), 3.67-3.76 (m, 3 H), 3.52-3.60 (m, 1 H). MS (EI) m/z: 324
[M þ H]þ. HRMS calcd for C11H13F3N3O5 [M þ H]þ 324.0807,
found 324.0815. HPLC (Sunfire C18 4.6 mm ꢀ 50 mm, 6-30%
[methanol/water/TFA (95/5/0.1)]:water with 0.1% TFA, 5 min
gradient) tR = 1.07 min, 100% integrated area.
(1R,2S,3R)-1-(2-(5-Isopropylisoxazol-3-yl)-1H-imidazol-4-yl)-
butane-1,2,3,4-tetraol (11). Prepared according to the general
procedure A from 5-isopropylisoxazole-3-carbonitrile (489 mg,
3.57 mmol) and fructosamine-HOAc (850 mg, 3.57 mmol) to
1
provide 11 as the HCl salt (1.03 g, 87% yield). H NMR (400
MHz, 1.0 M DCl in D2O) δ ppm 7.00 (d, J = 0.5 Hz, 1 H), 6.17
(d, J = 1.0 Hz, 1 H), 4.67 (dd, J = 1.9, 0.9 Hz, 1 H), 3.16-3.31
(m, 3 H), 3.04-3.14 (m, 1 H), 2.51-2.68 (m, 1 H), 0.67-0.77 (m,
6 H). MS (EI) m/z: 298 [M þ H]þ. HRMS calcd for C13H19N3O5
[M þ H]þ 298.1403, found 298.1410. HPLC (Sunfire C18 4.6 mm
ꢀ 50 mm, 10-90% 10 mM aq NH4OAc:MeCN, 2 min gradient)
tR = 1.47 min, 97% integrated area.
(1R,2S,3R)-1-(2-(4,5-Dihydroisoxazol-3-yl)-1H-imidazol-4-yl)-
butane-1,2,3,4-tetraol (5). To a -78 °C solution of the oxime 2
(2.40 g, 7.37 mmol) in THF (100 mL) was added a 1.6 M solution
of n-BuLi in hexanes (27.6 mL, 44.2 mmol). The reaction was
maintained at -78 °C for 15 min and then treated with DMF
(4.5 mL, 59.0 mmol) and warmed to room temperature. After 2 h,
the reaction was quenched with satd aq NH4Cl solution (50 mL),
then diluted with EtOAc (200 mL). The layers were separated and
the organics were washed with water (100 mL) and brine (50 mL),
dried over MgSO4, and concentrated under vacuum to provide a
white solid (1.63 g) which was redissolved in MeOH (45 mL). The
solution was cooled to 0 °C and then treated portionwise with
NaBH4 (700 mg, 18.4 mmol) and then warmed to room tempera-
ture for 1 h. The reaction mixture was then concentrated under
vacuum, and the resulting paste was dissolved in water (100 mL)
and extracted with EtOAc (3 ꢀ 100 mL). The combined organics
were washed with brine (50 mL), dried over MgSO4, and con-
centrated and then flashed over silica gel (80% EtOAc:hexanes
eluent, Rf = 0.15) to provide a white solid (557 mgs). The total
material was redissolved in THF (10 mL) and added to a 0 °C
solution of PPh3 (1.23 g, 4.71 mmol), iodine (1.04 g, 4.08 mmol),
and triethylamine (0.66 mL, 4.71 mmol). The reaction was
maintainedatroom temperatureovernight andthenconcentrated
and purified by flash chromatography over silica gel (3-6%
MeOH:DCM eluent) toprovide a whitesolid which was dissolved
in dioxane, treated with 1 N aq HCl, and heated to 50 °C for 2 h.
The reaction was then concentrated and purified by reverse phase
preparative HPLC to provide the isoxazoline 5 (224 mgs, 12%
yield, 4 steps) as a white powder. 1H NMR (400 MHz, 1.0 M DCl
in D2O) δ ppm 7.49 (s, 1 H), 5.17 (d, J = 2.3 Hz, 1 H), 4.61 (t, J =
1.0 Hz, 2 H), 3.67-3.82 (m, 3 H), 3.59-3.67 (m, 1 H), 3.46 (t, J =
1.0 Hz, 2 H). 13C NMR (100 MHz, 1.0 M DCl in D2O) δ ppm
146.88, 136.70, 135.53, 118.30, 73.09, 71.97, 71.02, 65.05, 63.21,
34.09. MS (EI) m/z: 258 [M þ H]þ. HRMS calcd for C10H16N3O5
[M þ H]þ 258.1090, found 258.1073. HPLC (Sunfire C18
1-(5-Methyl-3-(4-((1R,2S,3R)-1,2,3,4-tetrahydroxybutyl)-
1H-imidazol-2-yl)isoxazol-4-yl)ethanone (17). Prepared accord-
ing to the general procedure A from 4-acetyl-5-methylisoxazole-
3-carbonitrile (410 mg, 3.01 mmol) and fructosamine-HOAc
(720 g, 3.01 mmol) to provide 17 as the HCl salt (81 mg, 23%
yield). 1H NMR (400 MHz, 1.0 M DCl in D2O) δ ppm 7.57 (d,
J = 0.5 Hz, 1 H), 5.21 (d, J = 1.3 Hz, 1 H), 3.74-3.82 (m, 2 H),
3.72 (d, J = 1.8Hz, 1 H), 3.57-3.66 (m, 1 H), 2.83 (s, 3 H), 2.58 (s,
3 H). MS (EI) m/z: 312 [M þ H]þ. HRMS calcd for C13H18N3O6
[M þ H]þ 312.1196, found 312.1181. HPLC (Sunfire C18
4.6 mm ꢀ 50 mm, 5% [methanol/water/TFA (95/5/0.1)]:water,
8 min isocratic) tR = 3.03 min, 100% integrated area.
(1R,2S,3R)-1-(2-(5-Methylisoxazol-3-yl)-1H-imidazol-4-yl)-
butane-1,2,3,4-tetraol (9). Prepared according to the general
procedure A from 5-methylisoxazole-3-carbonitrile (216 mg, 2.0
mmol) and fructosamine-HOAc (476 mg, 2.0 mmol) to provide
9 as the HCl salt (403 mg, 75% yield). 1H NMR (400 MHz, 1.0 M
DCl in D2O) δ ppm 7.54(s, 1 H), 6.69 (s, 1 H), 5.20 (d, J = 1.0 Hz,
4.6 mm ꢀ 50 mm, 10 mM aq NH4OAc, 8 min isocratic) tR
=
3.41 min, 100% integrated area.
(1R,2S,3R)-1-(2-(5-Ethoxyisoxazol-3-yl)-1H-imidazol-4-yl)-
butane-1,2,3,4-tetraol (14). Prepared according to the general
procedure A from 5-ethoxyisoxazole-3-carbonitrile (2.42 g,
17.52 mmol) and fructosamine-HOAc (4.19 g, 17.52 mmol)
to provide 14 as the HCl salt (5.16 g, 88% yield). 1H NMR (400
MHz, 1.0 M DCl in D2O) δ ppm 7.44 (s, 1 H), 5.87 (s, 1 H), 5.11
(d, J = 1.3 Hz, 3 H), 4.27 (q, J = 7.1 Hz, 2 H), 3.49-3.78 (m,
4 H), 1.29 (t, J = 7.1 Hz, 3 H). 13C NMR (101 MHz, 1.0 M DCl
in D2O) δ175.06, 150.95, 136.20, 134.30, 117.72, 77.07, 72.62,
1 H), 3.69-3.86 (m, 3 H), 3.56-3.69 (m, 1 H), 2.49 (s, 3 H). 13
C
NMR (100 MHz, 1.0 M DCl in D2O) δ ppm 174.29, 150.01,
136.55, 134.96, 118.13, 100.98, 73.11, 71.01, 65.00, 63.20, 11.91.
MS (EI) m/z: 270 [M þ H]þ. HRMS calcd for C11H16N3O5 [M þ
H]þ 270.1090, found 270.1108. HPLC (HILIC Silica 4.6 mm ꢀ
50 mm, 100-40% [methanol/water/TFA (95/5/0.1)]:water with
0.1% TFA, 3 min gradient) tR = 0.30 min, 100% integrated area.

8660 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24
Bagdanoff et al.
70.51, 70.45, 64.51, 62.74, 13.48. MS (EI) m/z: 300 [M þ H]þ.
HRMS calcd for C10H13N3O4S [M þ H]þ 300.1196, found
300.1185. HPLC (Luna Phenyl-hexyl 4.6 mm ꢀ 50 mm, 5%
10 mM aq NH4OAc, 3 min isocratic) tR = 2.49 min, 99%
integrated area.
Acknowledgment. We thank the analytical group at Lex-
icon Pharmaceuticals in Princeton, NJ, for their assistance in
the characterization of compounds and the animal handling
and vivarium operation groups in The Woodlands, TX, for
their maintenance of animal resources.
(1R,2S,3R)-1-(2-(5-Cyclopropylisoxazol-3-yl)-1H-imidazol-
4-yl)butane-1,2,3,4-tetraol (10). Prepared according to the gen-
eral procedure A from 5-cyclopropylisoxazole-3-carbonitrile
(68mg, 0.50 mmol) and fructosamine-HOAc (120 mg, 0.5 mmol)
Supporting Information Available: Elemental analysis, 1H and
13C NMR for compounds 2 and 4. This material is available free
of charge via the Internet at http://pubs.acs.org.
1
to provide 10 as the HCl salt (86 mg, 52% yield). H NMR
(400 MHz, 1.0 M DCl in D2O) δ ppm 7.01 (s, 1 H), 6.03-6.07 (m,
1 H), 4.67 (s, 1 H), 3.17-3.30 (m, 3 H), 3.06-3.15 (m, 1 H),
1.56-1.66 (m, 1 H), 0.52-0.60 (m, 2 H), 0.42 (dd, J = 4.8, 2.5 Hz,
2 H). 13C NMR (101 MHz, 1.0 M DCl in D2O) δ ppm 179.11,
149.27, 135.85, 134.29, 117.77, 97.42, 72.59, 70.62, 64.50, 62.79,
8.70, 7.61. MS (EI) m/z: [MH]þ 296 HRMScalcd for C13H18N3O5
[M þ H]þ 296.1246, found 296.1259. HPLC (Sunfire C18
4.6 mm ꢀ 50 mm, 5-90% [methanol/water/TFA (95/5/0.1)]:0.1%
aq TFA, 2 min gradient) tR = 0.59 min, 99% integrated area.
(1R,2S,3R)-1-(2-(5-Methyl-4-nitroisoxazol-3-yl)-1H-imidazol-
4-yl)butane-1,2,3,4-tetraol (18). Prepared according to the gen-
eral procedure A from 5-methyl-4-nitroisoxazole-3-carbonitrile
(1.00 g, 6.54 mmol) and fructosamine-HOAc (1.55 g, 6.54
mmol) to provide 18 as the HCl salt (76 mg, 3% yield). 1H
NMR (400 MHz, 1.0 M DCl in D2O) δ ppm 7.31 (s, 1 H), 4.93 (s,
1 H), 3.63-3.81 (m, 3 H), 3.49-3.57 (m, 1 H), 2.78 (s, 3 H). MS
(EI) m/z: 315 [M þ H]þ. HRMS calcd for C11H15N7O7 [M þ H]þ
315.0941, found 315.0952. HPLC (Sunfire C18 4.6 mm ꢀ 50 mm,
5-90% [methanol/water/TFA (95/5/0.1)]:0.1% aq TFA, 2 min
gradient) tR = 0.54 min, 98% integrated area.
References
(1) (a) Vogel, P.; Donoviel, M. S.; Read, R.; Hansen, G. M.;
Hazelwood, J.; Anderson, S. J.; Sun, W.; Swaffield, J.; Oravecz,
T. Incomplete Inhibition of Sphingosine 1-Phosphate Lyase Mod-
ulates Immune System Function yet Prevents Early Lethality and
Non-Lymphoid Lesions. PloS ONE 2009, 4 (1), e4112. (b) Fyrst, H.;
Saba, J. D. Sphingosine-1-phosphate lyase in development and disease:
Sphingosine metabolism takes flight. Biochim. Biophys. Acta 2008,
1781, 448–458. (c) For a review, see: Kumar, A.; Saba, J. D. Lyase to
live by: sphingosine phosphate lyase as a therapeutic target. Expert
Opin. Ther. Targets 2009, 13, 1–13. (d) Weber, C.; Krueger, A.; M€unk,
€
A.; Bode, C.; Van Veldhoven, P. P.; Graler, M. H. Discontinued
postnatal thymocyte development in sphingosine 1-phosphate-lyase-
deficient mice. J. Immunol. 2009, 7, 4292–4301. (e) Bektas, M.;
Allende, M. L.; Lee, B. G.; Chen, W; Amar, M. J.; Remaley, A. T.;
Saba, J. D.; Proia, R. L. Sphingosine 1-phosphate lyase deficiency
disrupts lipid homeostatis in liver. J. Biol. Chem. 2010, 10880–10889.
(2) (a) Reiss, U.; Oskouian, B.; Zhou, J.; Gupta, V.; Sooriyakumaran,
P.; Kelly, S.; Wang, E.; Merrill, A. H.; Saba, J. D. Sphingosine-
phosphate lyase enhances stress-induced ceramide generation and
apoptosis. J. Biol. Chem. 2004, 279, 1281–1290. (b) Moore, A. N.;
Kampfl, A. W.; Zhao, X.; Hayes, R. L.; Dash, P. K. Sphingosine
1-phosphate induces apoptosis of cultured hippocampal neurons that
requires protein phosphatases and activator protein-1 complexes. Neu-
roscience 1999, 94, 405–415. (c) Maceyka, M. H.; Sankala, N. C.; Hait,
H.; Le Stunff, H.; Liu, R.; Toman, C.; Collier, M.; Zhang, L. S.; Satin,
A. H.; Merrill, S.; Milstien, S.; Spiegel, S. SphK1 and SphK2,
sphingosine kinase isoenzymes with opposing functions in sphingolipid
metabolism. J. Biol. Chem. 2005, 280, 37118–37129. (d) Pyne, S.;
Pyne, N. J. Sphingosine 1-phosphate signalling in mammalian cells.
J. Biochem. 2000, 349, 385–402. (e) McVerry, B.; Garcia, J. In vitro and
in vivo modulation of vascular barrier integrity by sphingosine 1-phos-
phate: mechanistic insights. Cell. Signalling 2005, 131–139. (f) Cyster,
J. G. Chemokines, sphingosine-1-phosphate, and cell migration in
secondary lymphoid organs. Annu. Rev. Immunol. 2005, 23, 127–
159. (g) Hait, N. C.; Oskeritzian, C. A.; Paugh, S. W.; Milstien, S.;
Spiegel, S. Sphingosine kinases, sphingosine 1-phosphate, apoptosis
and diseases. Biochim. Biophys. Acta 2006, 1758, 2016–2026.
(h) Rosen, H.; Goetzl, E. J. Sphingosine 1-phosphate and its receptors:
an autocrine and paracrine network. Nature Rev. Immunol. 2005, 5,
560–570. (i) Saba, J. D.; Hla, T. Point-counterpoint of sphingosine
1-phosphate metabolism. Circ. Res. 2004, 94, 7240–7234.
(1R,2S,3R)-1-(2-(5-(Hydroxymethyl)isoxazol-3-yl)-1H-imida-
zol-4-yl)butane-1,2,3,4-tetraol (15). Prepared according to the
general procedure A from 5-(hydroxymethyl)isoxazole-3-carbo-
nitrile (244 mg, 1.17 mmol) and fructosamine-HOAc (279 g,
1.17 mmol) to provide 15 as the HCl salt (217 mg, 68% yield). 1H
NMR (400 MHz, 1.0 M DCl in D2O) δ 7.23 (s, 1 H), 6.63 (s,
1 H), 4.88 (s, 1 H), 4.46 (s, 2 H), 3.36-3.53 (m, 3 H), 3.32
(s, 1 H). 13C NMR (75 MHz, 1.0 M DCl in D2O) δ ppm 174.40,
149.52, 136.17, 134.15, 117.93, 101.23, 72.66, 70.59, 64.55,
62.83, 54.83. MS (EI) m/z: [MH]þ 286. HRMS calcd for C11
-
H16N3O6 [M þ H]þ 286.1039, found 286.1056. HPLC (YMC-
Pack ODS-A 4.6 mm ꢀ 50 mm, 10-95% [methanol/water/
TFA (95/5/0.1)]:0.1% aq TFA, 4 min gradient) tR = 0.44 min,
98% integrated area.
(1R,2S,3R)-1-(2-(5-((Dimethylamino)methyl)isoxazol-3-yl)-
1H-imidazol-4-yl)butane-1,2,3,4-tetraol (16). Prepared according
to the general procedure A from 5-((dimethylamino)methyl)-
isoxazole-3-carbonitrile (215 mg, 1.42 mmol) and fructosamine-
HOAc (339 mg, 1.42 mmol) to provide 16 as the HCl salt
(126 mg, 25% yield). 1H NMR (400 MHz, 1.0 M DCl in D2O)
δ 7.27 (s, 1 H), 7.05 (s, 1 H), 4.90 (s, 1 H), 4.38 (s, 2 H), 3.39-3.51
(m, 3 H), 3.27-3.35 (m, 1 H). 13C NMR (75 MHz, 1.0 M DCl in
D2O) δ ppm 163.87, 150.26, 136.50, 133.35, 118.23, 107.31, 72.63,
70.57, 64.55, 62.81, 50.39, 43.05. MS (EI) m/z: 313 [M þ H]þ.
HRMS calcd for C13H21N4O5 [Mþ H]þ 313.1512, found 313.1499.
HPLC (YMC-Pack ODS-A 4.6 mm ꢀ 50 mm, 10-95%
[methanol/water/TFA (95/5/0.1)]:0.1% aq TFA, 4 min gradient)
(3) For a general review, see: (a) Marsolais, D.; Rosen, H. Chemical
modulators of sphingosine-1-phosphate receptors as barrier or-
iented therapeutic molecules. Nature Rev. Drug Discovery 2009, 8,
297–307. (b) Goetzl, E. J.; Wang, W.; McGiffert, C.; Huang, M.-C.;
Graeler, M. H. Sphingosine 1-phosphate and its G protein-coupled
receptors constitute
a multifunctional immunoregulatory system.
J. Cell. Biochem. 2004, 92, 1104–1114. (c) Sanchez, T; Hla, T. Struc-
tural and functional characteristics of S1P receptors. J. Cell. Biochem.
2004, 92, 913–922. (d) Goetzl, E. J.; Liao, J.-J.; Huang, M.-C.
Regulation of the roles of sphingosine 1-phosphate and its type 1 G
protein-coupled receptor in T cell immunity and autoimmunity. Bio-
chem. Biophys Acta 2008, 1781, 503–507. (e) Goetzl, E. J.; Rosen, H.
Regulation of immunity by lysosphigolipids and their G protein-
coupled receptors. J. Clin. Invest. 2004, 114, 1531–1537.
tR = 0.25 min, 100% integrated area.
(1R,2S,3R)-1-(2-(4-Ethylisoxazol-3-yl)-1H-imidazol-4-yl)-
(4) (a) Cohen, J. A.; Barkhof, F.; Comi, G.; Hartung, H.-P.; Khatri,
B. O.; Montalban, X.; Pelletier, J.; Capra, R.; Gallo, P.; Izquierdo,
G.; Tiel-Wilck, K.; de Vera, A.; Jin, J.; Stites, T.; Wu, S.; Aradhye,
S.; Kappos, L. Oral fingolimod or intramuscular interferon for
relapsing multiple sclerosis. N. Engl. J. Med. 2010, 362, 402–415.
(b)Adachi, K.;Kohara, T.;Nakao, N.;Masafumi, A.;Chiba, K.; Mishina,
T.; Sasaki, S.; Fujita, T. Design, Synthesis, and Structure-Activity
Relationships of 2-Substituted-2-Amino-1,3-Propanediols: Discovery
of a Novel Immunosuppressant, FTY720. Bioorg. Med. Chem. Lett.
1995, 5, 853–856. (c) Chi, H.; Flavell, R. A. Cutting Edge: Regulation
of T Cell Trafficking and Primary Immune Responses by Sphingosine
1-Phosphate Receptor 1. J. Immunol. 2005, 174, 2485–2488. (d) Wang,
W.; Huang, M.-C.; Goetzl, E. J. Type 1 Sphingosine 1-Phosphate G
butane-1,2,3,4-tetraol (19). Prepared according to the general
procedure A from 4-ethylisoxazole-3-carbonitrile (366 mg,
3.00 mmol) and fructosamine-HOAc (714 mg, 3.00 mmol) to
provide 19 as the HCl salt (200 mg, 21% yield). 1H NMR (400
MHz, 1.0 M DCl in D2O) δ ppm 8.68 (s, 1 H), 7.58 (s, 1 H),
5.22 (d, J = 1.5 Hz, 1 H), 3.78 (s, 3 H), 3.58-3.67 (m, 1 H),
2.58 (q, J = 7.5 Hz, 2 H), 1.13 (t, J = 7.5 Hz, 3 H). MS (EI)
m/z: 284 [M þ H]þ. HRMS calcd for C12H18N3O5 [M þ H]þ
284.1246, found 284.1259. HPLC (Sunfire C18 4.6 mm ꢀ
50 mm, 10-90% MeCN:10 mM aq NH4OAc, 2 min gradient)
tR = 0.77 min, 98% integrated area.

Article
Protein-Coupled Receptor (S1P1) Mediation of Enhanced IL-4 Gen-
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24 8661
and further internal study with a fluorogenic substrate, in addition
to radioactively-labeled S-1-P, also supported slow catalytic turn-
over rates for S-1-P Lyase. For references, see: Van Veldhoven,
P. P. Sphingosine-1-phosphate Lyase. Methods Enzymol. 1999,
311, 244–254. (b) Van Veldhoven, P. P.; Mannaerts, G. P. Sphingo-
sine-Phosphate Lyase. Adv. Lipid Res. 1993, 26, 69–98. For fluoro-
genic substrate, see (c) Bedia, C.; Camacho, L.; Casas, J.; Abad, J. L.;
Delgado, A.; Van Veldhoven, P. P.; Fabrias, G. Synthesis of a Fluoro-
genic Analogue of Sphingosine-1-Phosphate and Its Use to Determine
Sphingosine-1-Phosphate Lyase Activity. ChemBioChem 2009, 10,
820–822.
eration by CD4 Tcells from S1P1 Transgenic Mice. J. Immunol. 2007,
178, 4885–4890.
(5) (a) For a comprehensive treatment, see: Brinkmann, V. Sphingo-
sine 1-phosphate receptors in health and disease: mechanistic
insights from gene deletion studies and reverse pharmacology.
Pharmacol. Ther. 2007, 115, 84–105. (b) Maceyka, M.; Spiegel, S.
Sphingosine-1-phosphate receptors. Handb. Cell Signaling 2004, 2,
247–251. (c) Oo, M. L.; Thangada, S.; Wu, M. T.; Liu, C. H.;
Macdonald, T. L.; Lynch, K. R.; Lin, C. Y.; Hla, T. Immunosuppressive
and anti-angiogenic sphingosine 1-phosphase receptor-1 agonists un-
duce ubiquitinylation and proteasomal degredation of the receptor.
J. Biol. Chem. 2007, 282, 9082–9089. (d) Parrill, A. L.; Wang, D.;
Bautista, D. L.; Van Brocklyn, J. R.; Lorincz, Z.; Fischer, D. J.; Baker,
D. L.; Liliom, K.; Spiegel, S.; Tigyi, G. Identification of edg1 receptor
residues that recognize sphingosine 1-phosphate. J. Biol. Chem. 2000,
275, 39379–39384. (e) LaMontagne, K.; Littlewood-Evans, A.;
Schnell, C.; O'Reilly, T.; Wyder, L.; Sanchez, T.; Probst, B.; Butler,
J.; Wood, A.; Liau, G.; Billy, E.; Theuer, A.; Hla, T.; Wood, J.
Antagonism of sphingosine-1-phosphate receptors by FTY720 inhibits
angiogenesis and tumor vascularization. Cancer Res. 2006, 66, 221–
231. (f) Maceyka, M.; Milstein, S.; Spiegel, S. Sphingosine-1-phos-
phate: the Swiss army knife of sphingolipid signaling. J. Lipid Res.
2009, S272–S276.
(6) Schwab, S. R.; Pereira, J. P.; Matloubian, M.; Xu, Y.; Huang, Y.;
Cyster, J. G. Lymphocyte sequestration through S1P lyase inhibi-
tion and disruption of S1P gradients. Science 2005, 309, 1735–1739.
(7) (a) Bandhuvula, P.; Fyrst, H.; Saba, J. D. A rapid fluorescence
assay for sphingosine-1-phosphate lyase enzyme activity. J. Lipid
Res. 2007, 48, 2769–2778.
(8) Bagdanoff, J. T.; Donoviel, M. S.; Nouraldeen, A.; Tarver, J.; Fu,
Q; Carlsen, M.; Jessop, T.; Zhang, H.; Hazelwood, H.; Nguyen, H.;
Baugh, S. D. P.; Gardyan, M; Terranova, K. M.; Barbosa, J.; Yan,
J.; Bednarz, M.; Layek, S.; Taylor, J.; Digeorge-Foushee, A. M.;
Gopinathan, S.; Bruce, D.; Smith, T.; Moran, L.; O’Neill, E.;
Kramer, J.; Lai, Z.; Kimball, S. D.; Liu, Q.; Sun, W.; Yu, S.;
Swaffield, J.; Wilson, A.; Main, A.; Carson, K. G.; Oravecz, T.;
Augeri, D. J. Inhibition of Sphingosine-1-Phosphate Lyase (S1PL)
for the Treatment of Autoimmune Disorders. J. Med. Chem. 2009,
52, 3941–3953.
(9) (a) Hodge, J. E.; Rist, C. E. Amadori rearrangement under new
conditions and its significance for nonenzymatic browning reac-
tions. J. Am. Chem. Soc. 1953, 75, 316–322. (b) Hodge, J. E.; Fisher,
B. E. In Methods in Carbohydrate Chemistry; Whistler, R. L.,
Wolfrom, M. L., Bemiller, J. N., Eds.; Academic Press: New York,
1963; Vol. II, pp 99-103.
(10) Buchi, G.; Halweg, K. M. A Convenient Synthesis of 2-Acetyl-
4(5)-(1(R),2(S),3(R),4-tetrahydroxybutyl)-imidazole. J. Org.
Chem. 1985, 50, 1134–1136.
(11) Conducting the reaction in methanol-d4 reveals that deuterium is
incorporated into the core imidazole ring. On the basis of this
observation and in situ NMR experiments, a reasonable mecha-
nism involves formation of the imidiate B from ethoxyacrylonitrile
A. Subsequent condensation with fructosamine acetate 6 provides
C, which equilibrates to deuterated intermediate D in the presence
of excess methanol-d4. Intramolecular cyclization to E preceeds
aromatization and enol-ether hydrolysis to deuterated THI analo-
gue d1-1.
(15) (a) Internal findings using FTY-720 as a positive control, unpub-
lished results. (b) Phosphorylation of FTY720 is a facile process
and synthetic agonist binds to S1PR on the cell surface of lympho-
cytes. For related published results, see: Foster, C. A.; Howard,
L. M.; Schweitzer, A.; Persohn, E.; Hiestand, P. C.; Balatoni, B.;
Reuschel, R.; Beerli, C.; Schwartz, M.; Billich, A. J. Pharmacol.
Exp. Ther. 2007, 323, 469–476.
(16) (a) Queneau, P. The saga of aspirin: centuries-old ancestors of an
old lady who doesn’t deserve to die. Therapie 2001, 56, 723–726.
(b) Miner, J.; Hoffhines, A. The discovery of aspirin's antithrombotic
effects. Texas Heart Inst. J. 2007, 34, 179–186. (c) Mahdi, J. G.;
Mahdi, A. J.; Madhi, A. J.; Bowen, I. D. The historical analysis of
aspirin discovery, its relation to the willow tree and antiproliferative and
anticancer potential. Cell Proliferation 2006, 39, 147–155.
(17) (a) Bentley, R. The development of penicillin: Genesis of a famous
antibiotic. Perspect. Biol. Med. 2005, 48, 444–452. (b) Diggins, F. W.
The true history of the discovery of penicillin, with refutation of the
misinformation in the literature. Br. J. Biomed. Sci. 1999, 56, 83–93.
(18) For a general review, see: Masataka, Y. Iodine and Living Body.
Kagaku Keizai 2008, 55, 75–80.
(19) (a) Fujita, T.; Yoneta, M.; Hirose, R.; Sasaki, S.; Inoue, K.; Kiuchi,
M; Hirase, S.; Adachi, K.; Arita, M.; Chiba, K. Simple compounds,
2-alkyl-2-amino-1,3-propanediols have potent immunosuppres-
sive activity. Bioorg. Med. Chem. Lett. 1995, 5, 847–852. (b) Adachi,
K.; Kohara, T.; Nakao, N.; Masafumi, A.; Chiba, K.; Mishina, T.;
Sasaki, S.; Fujita, T. Design, Synthesis, and Structure-Activity Rela-
tionships of 2-Substituted-2-Amino-1,3-Propanediols: Discovery of a
Novel Immunosuppressant, FTY720. Bioorg. Med. Chem. Lett. 1995,
5, 853–856. (c) Kiuchi, M.; Adachi, K.; Kohara, T.; Teshima, K.;
Masubuchi, Y.; Mishina, T.; Fujita, T. Synthesis and biological evalua-
tion of 2,2-disubstituted 2-aminoethanols: analogues of FTY720.
Bioorg. Med. Chem. Lett. 1998, 8, 101–106.
(20) (a) Bagli, J. F.; Kluepfel, D. Elucidation of Structure and Stereo-
chemistry of Myriocin. A Novel Antifungal Antibiotic. J. Org.
Chem. 1973, 38, 1253–1260. (b) Bagli, J. F.; Kluepfel, D.; St.-Jacques,
M. Edlucidation of Structure and Stereochemistry of Myriocin. A Novel
Antifungal Antibiotic. J. Org. Chem. 1973, 38, 1253–1260.
(21) (a) Miyake, Y.; Kozutsumi, Y.; Nakamura, S.; Fujita, T.; Kawasaki,
T. Serine palmitoyltransferase is the primary target of a sphingosine-
like immunosuppressant, ISP-1/ myriocin. Biochem. Biophys. Res.
Commun. 1995, 211, 396–403. (b) Chen, J. K.; Lane, W. S.; Schreiber,
S. L. The identification of myriocin-binding proteins. Chem. Biol. 1999,
6, 221–235.
(22) Brinkmann, V.; Davis, M. D.; Heise, C. E.; Albert, R.; Cottens, S.;
Hof, R.; Bruns, C.; Prieschl, E.; Baumruker, T.; Hiestand, P.;
Foster, C. A.; Zollinger, M.; Lynch, K. R. The immune modulator
FTY720 targets sphingosine 1-phosphate receptors. J. Biol. Chem.
2002, 277, 21453–21457.
(23) (a) Rivera, J.; Proia, R. L.; Olivera, A. The alliance of sphingosine-
1-phosphate and its receptors in immunity. Nature Rev. Immunol.
2008, 8, 753–763. (b) Billich, A.; Baumruker, T.; Sphingolipid Meta-
bolizing Enzymes as Novel Theraputic Targets. In Lipids in Health and
Disease; Quinn, P. J., Wang, X., Eds.; Springer Science Business Media
BV: New York, 2008; pp487-522.
(24) (a) Peng, X.; Hassoun, P. M.; Sammani, S.; McVerry, B. J.; Burne,
M. J.; Rabb, H.; Pearse, D.; Tuder, R. M.; Garcia, J. G. N.
Protective effects of sphingosine 1-phosphate in murine endotox-
in-induced inflammatory lung injury. Am. J. Respir. Crit. Care
Med. 2004, 169, 1245–1251. (b) Nofer, J. R.; van der Giet, M.; Tolle,
M.; Wolinska, I.; von Wnuck Lipinski, K.; Baba, H. A.; Tietge, U. J.;
Godecke, A.; Ishii, I.; Kleuser, B.; Schafers, F.; Fobker, M; Zidek, W.;
Assmann, G.; Chun, J.; Levkau, B. HDL induces NO-dependent
vasorelaxation via the lysophospholipid receptor S1P3. J. Clin. Invest.
2004, 113, 569–581. (c) Van der Giet, M.; Tolle, M.; Kleuser, B.
Relevance and potential of sphingosine-1-phosphate in vascular in-
flammatory disease. Biol. Chem. 2008, 389, 1381–1390.
(25) Hughes, J. E.; Srinivasan, S.; Lynch, K. R.; Proia, R. L.; Ferdek, P.;
Hedrick, C. C. Sphingosine-1-phosphate induces an antiinflamma-
tory phenotype in macrophages. Circ. Res. 2008, 102, 950–958.
(26) Rolin, J.; Sand, K. L.; Knudsen, E.; Maghazachi, A. A. FTY720
and SEW2871 reverse the inhibitory effect of S1P on natural killer
(12) Yu, X. Q.; Kramer, J.; Moran, L; O’Neill, E.; Nouraldeen, A.;
Oravecz, T.; Wilson, A. Pharmacokinetic and Pharmacodynamic
modeling of 2-acetyl-4(5)-tetrahydroxyimidazole-induced periph-
eral lymphocyte sequestration through increasing lymphoid sphin-
gosine 1-phosphate. Xenobiotica 2010, 40, 350–356.
(13) Kondoh, M.; Yagi, K. Progress in absorption enhancers based on
tight junction. Expert Opin. Drug Delivery 2007, 4, 275–286.
(14) (a) Van Veldhoven determined the catalytic turnover of S1PL
in mammalian tissues and cultured cells to be between 20 and
100 pmol/min/mg protein. Internal unpublished results with full
length protein including N-terminal membrane-binding region
normally bound to endoplasmic reticulum supported these findings

8662 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 24
Bagdanoff et al.
cell mediated lysis of K562 tumor cells and dendritic cells but not on
cytokine release. Cancer Immunol. Immunother. 2010, 59, 575–586.
(27) (a) Augeri, D. J. Inhibition of Sphingosine-1-Phosphate Lyase for
the Treatment of Autoimmune Disorders. First Disclosure of
LX2931/LX2932. 239th ACS National Meeting, San Francisco,
CA, March 21-25, 2010, MEDI-36. (b) Fleischmann, R.; Frazier,
K.; Freiman, J.; Brooks, B.; Donoviel, M.; Oravecz, T.; Augeri, D.;
Heydorn, W.; Kelly, M.; Brown, P. Co-Administration of the Oral
S1P-Lyase Inhibitor LX2931 with Methotrexate was Well Toler-
ated Over 14 Days in Patients with Stable Rheumatoid Arthritis.
Meeting of the American College of Rheumatology, Philadelphia,
PA, October 18, 2009, abstract 426.
reaction mixture indicated complete consumption of the starting
material. The reaction was concentrated, and then the resulting oil
was redissolved in chloroform (1 L). The solids were filtered off,
and the liquor was concentrated. The resulting oil was azeotroped
with methanol (2 ꢀ 1 L) to chase off trace chloroform. The crude
material was used immediately in the next transformation. The oil
was redissolved in methanol (1.5 L) and then treated with 25% w/w
NaOMe (33.0 mL, 0.15 mol). After 1 h, 1H NMR of an aliquot of
the reaction mixture indicated complete conversion. The reaction
mixture was then treated with a 7N solution of ammonia in MeOH.
After 1 h, the reaction was concentrated to dryness. The resulting
dark solid mass was pulverized with a mortar and pestle and then
slurried in water (500 mL) and filtered to remove to provide, after
rigorous drying, an orange solid which was used directly in the next
reaction. The dried solid was slurried in DCM (1.5 L) and cooled to
0 °C and treated with pyridine (303.0 mL, 3.75 mol) while main-
taining the internal temperature of the reaction below 10 °C. To the
cooled solution, trifluoroacetic acid anhydride was added slowly
through an addition funnel, again while carefully monitoring the
internal temperature of the reaction. Once the addition was com-
plete, the reaction was allowed to warm to room temperature for an
hour before quenching with water (1 L). The aqueous layer was
adjusted to pH=2, and then the biphasic mixture was agitated.
The aqueous layer was readjusted to pH = 2. The layers were
separated, and the organics were washed with water (2 ꢀ 500 mL)
and brine (500 mL) and then dried over MgSO4. The resulting
crude liquid was distilled (bp 81-83 °C, 31 tor) to provide the nitrile
(82.0 g, 58% yield, 4 steps) as a colorless liquid. 1H NMR (300
MHz, CDCl3) δ ppm 8.70 (d, J = 1.5 Hz, 1 H), 6.77 (d, J = 1.2 Hz,
1 H). 13C NMR (101 MHz, CDCl3) δ ppm 160.92, 138.92, 109.81,
107.24.
(28) See experimental section of : Salojin, K.; Owusu, R. B.; Millerchip,
K. A.; Potter, M.; Platt, K. A.; Oravecz, T. Essential Role of
MAPK Phosphatase-1 in the Negative Control of Innate Immune
Responses. J. Immunol. 2006, 176, 1899–1907.
(29) Isoxazole-3-carbonitrile synthesized by following procedure: Ethyl
vinyl ether (75 mL, 1.35 mol) was added dropwise over 30 min to
neat ethyl cholorooxoacetate cooled to 0 °C. The cooling bath was
left in place, and the reaction was allowed to gradually warm to
room temperature overnight. The excess ethyl vinyl ether was
removed by rotovap, and then crude reaction was distilled rapidly
(bp 102-104 °C, 9 Torr) to provide the vinylogous ester (68.8 g,
59% yield) as a clear liquid. Note: failure to initiate distillation
within a few minutes of heating will result in polymerization of
crude material. Therefore, it is recommended that the heating bath
is ramped quickly above ∼150 °C while externally heating a well-
insulated distillation head with a heat gun. To a room temperature
solution of the resulting vinylogous eneone (260.0 g, 1.51 mol) in
MeOH (1.5 L) was added hydroxylamine hydrochloride (110.0 g,
1.59 mol) in one portion. After 2 h, 1H NMR of an aliquot of the