Minimum structure requirement of immunomodulatory glycolipids for predominant Th2 cytokine induction and the discovery of non-linear phytosphingosine analogs

Article abstract of DOI:10.1016/j.bmcl.2007.02.081

Analogs of immunomodulatory glycolipid OCH (2) were prepared and minimum structure requirement to exhibit equivalent profiles was disclosed. Analogs bearing non-linear hydrocarbon chain in the phytosphingosine moiety (18, 19) were shown for the first time

Full text of DOI:10.1016/j.bmcl.2007.02.081

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2781–2784
Minimum structure requirement of immunomodulatory glycolipids
for predominant Th2 cytokine induction and the discovery of
non-linear phytosphingosine analogs
Tetsuya Toba,a Kenji Murata,a Kyoko Nakanishi,a Bitoku Takahashi,a
Naohiro Takemoto,a Minako Akabane,a Takashi Nakatsuka,a Seiichi Imajo,a
Takashi Yamamura,b Sachiko Miyakeb and Hirokazu Annouraa,*
aBiomedical Research Laboratories, Daiichi Asubio Pharma Co., Ltd, Japan
bNational Institute of Neuroscience, National Center for Neuroscience and Psychiatry, Japan
Received 1 December 2006; revised 22 February 2007; accepted 26 February 2007
Available online 3 March 2007
Abstract—Analogs of immunomodulatory glycolipid OCH (2) were prepared and minimum structure requirement to exhibit equiv-
alent profiles was disclosed. Analogs bearing non-linear hydrocarbon chain in the phytosphingosine moiety (18, 19) were shown for
the first time to possess comparable cytokine inducing profile to 2. Molecular modeling of 2/hCD1d complex based on the crystal
structure of a-GalCer (1)/hCD1d complex is also described.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Natural killer T (NKT) cells are potent producers of
immunoregulatory cytokines, and are restricted to gly-
colipid antigens presented by CD1d, a glycoprotein
strucurally and functionally related to non-classical
major histocompatibility complex (MHC) class I.1 The
glycolipids, an a-galactosylceramide named KRN7000
12 and an altered analog termed OCH 2 possessing a
shorter phytosphingosine side chain,3 have been identi-
fied as NKT cell ligands (Fig. 1). It was recently shown
that a isoglobotrihexosylceramide (iGb3 3), though not
yet purified and characterized in all mammalian species,
may be an endogenous ligand of CD1d.4 The X-ray
crystallographic structures of mouse (m) CD1d5 and hu-
man (h) CD1d in complex with 16 indicated that the
complete occupation of the binding groove of CD1d
by 1 contributes to the sustained stimulation of NKT
cells to induce robust immunological response, while
altered analogs such as 2 with short phytosphingosine
chain may result in short duration of stimulation and
Figure 1. Known glycocylceramides as ligands of CD1d.
cause differential polarization of NKT cells.7 Compound
2 was shown to induce a predominant production of a
key immunomodulatory Th2 cytokine interleukin-4
(IL-4) over proinflammatory Th1 cytokine interferon-c
(IFN-c), while 1 induced both Th1/Th2 cytokines. Only
compound 2 but not 1 is significantly effective in animal
models of Th1-mediated autoimmune diseases such as
experimental autoimmune encephalomyelitis (EAE)
and collagen induced arthritis (CIA).3,8
Keywords: OCH; Th2 cytokine; Non-linear phytosphingosine analogs.
*
Corresponding author. Tel.: +81 75 962 8188; fax: +81 75 962
6448; e-mail: annoura.hirokazu.wk@asubio.co.jp
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.02.081

2782
T. Toba et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2781–2784
Savage et al. have reported the influence of the chain
lengths on the cytokine releasing profile in the context
of IL-4/IFN-c ratio, which presented three compounds
of short phytosphingosine chain including 2.9 Quite
recently, Wong et al. have reported a structure–activity
relationship (SAR) of fatty acyl chain analogs of 1 in
which some showed more potent Th1 cytokine response
than 1.10 As part of our efforts to obtain more potent
ligand for the enhancement of Th2 response, a series
of analogs based on 2 with altered ceramide moiety
was prepared and assayed in vitro for cytokine
production.11 In this letter, the SAR of derivatives of
2 including non-linear hydrocarbon chain analogs is
disclosed and minimum structure requirement to exhibit
equivalent profiles is discussed.
mice, which were incubated with 100 ng/ml of glycolipids
for 72 h and the cytokines in the culture supernatant were
measured by ELISA.13
Influence of the chain length of the fatty acid was exam-
ined first (R1, Table 1). The chain length of the phyto-
sphingosine moiety was fixed to that of 2. For
unambiguous understanding, the length of the alkyl
chain will be given as the number of carbon atoms
(Cn) in the R moiety. Acyl chains shorter than C19 (9)
showed only a weak cytokine production. As the chain
became longer the cytokine release increased and for
chains longer than C25 there was a marked increase in
IFN-c prduction that predominated IL-4 (13, 14). For
reference, 1 showed 128% release of IL-4 and 569% of
IFN-c relative to 2 in this assay.
The analogs were prepared by the versatile method
developed by our group12 (Scheme 1). The moieties of
sphingosine base substituents were introduced to the
known epoxide 4 by means of nucleophilic addition of
alkyl or aryl lithium reagents or corresponding magne-
sium bromides. After regioselective mesylation of the
axial hydroxyl group, compound 5 was subjected to ben-
zylidene cleavage and azidation, after which secondary
hydroxyl groups were protected to provide isopropylid-
ene acetal 6. Glycosidation with tetra-O-benzyl-a-D-
galactosyl bromide in the presence of tetrabutylammonium
bromide gave 7, whose azido group was reduced to an
amine and acylated with suitable carboxylic acids.
Finally, all the protective groups were removed to give
the desired analogs.
Our interest was next focused on the phytosphingosine
moiety. This is where it makes 2, a mere truncated ana-
log of 1, a completely different switch of the NKT cell
signal. We have prepared analogs altered in the phyto-
sphingosine chain (R2, Table 2). In our hands chain
length of C1–C4 (15–17 and 2) showed similar profiles.
Table 1. Dependency of cytokine production on fatty acyl chain
lengtha
The analogs were evaluated for their ability to induce
IL-4 and IFN-c relative to 2. IL-4 and IFN-c secretion
were assessed with spleen cells prepared from C57BL/6
Compound
R1
n
IL-4 (%)b
IFN-c (%)b
9
10
11
2
–(CH2)18CH3
–(CH2)20CH3
–(CH2)21CH3
–(CH2)22CH3
–(CH2)23CH3
–(CH2)24CH3
–(CH2)26CH3
–(CH2)24CH3
3
3
3
3
3
3
3
6
51
9
46
103
100
154
129
178
128
93
100
103
504
761
569
12
13
14
1
12
a At 100 ng/ml.
b Compared to 2 at 100 ng/ml.
Table 2. Dependency of cytokine production on phytosphingosine
chain modificationa
Compound
R2
IL-4 (%)b
IFN-c (%)b
15
16
17
2
–CH3
102
105
115
100
81
96
Scheme 1. Reagents and conditions: (a) R2Li or R2MgBr, CuI, THF,
ꢀ40 ꢁC, 93–98%; (b) MsCl, pyridine, ꢀ40 ꢁC, 34–93%; (c) H2, cat.
Pd(OH)2, EtOH, rt, or 6 N HCl, MeOH, rt, 69–100%; (d) NaN3,
DMF, 95 ꢁC, 27–66%; (e) cat. p-TsOH, 2,2-dimethoxypropane, rt, 65–
75%; (f) tetra-O-benzyl-a-D-galactosyl bromide, n-Bu4NBr, MS4A,
DMF–toluene (1:2.5), rt, 21–68%; (g) H2, Lindlar catalyst, EtOH, rt;
(h) R1CO2H, EDCIÆHCl, HOBt, i-Pr2NEt, DMF–CH2Cl2 (1:3.5),
40 ꢁC, 51–100% (2 steps); (i) HCl–dioxane, rt, or 80% AcOH, 80 ꢁC; (j)
H2, cat. Pd(OH)2, MeOH–CHCl3 (3:1), rt, 40 ꢁC, 52–91% (2 steps).
–CH2CH3
–(CH2)2CH3
–(CH2)3CH3
112
100
18
19
98
74
211
284
a At 100 ng/ml.
b Compared to 2 at 100 ng/ml.

T. Toba et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2781–2784
2783
Although the mechanism of action at the molecular level
is unclear,14 this result is inconsistent with Savage’s
report in which compound 15 showed about twofold
increase in IL-4 production without a change in
IFN-c.9 Analogs bearing cyclic chain (18) or aromatic
ring (19) were also prepared. To our knowledge, these
are the first examples of non-linear hydrocarbon chain
analogs of the phytosphingosine moiety that show
similar cytokine inducing profile to 2. As can be seen
from the side view of the X-ray structure of 1/hCD1d
complex6, the phytosphingosine chain of 1 is bent and
enters the C0 pocket beyond ꢁC6. There is a wide space
before entering the C0 pocket, and we assume analogs
15–19, which in length do not reach the C0 pocket,
showed similar profiles to 2.15 Compound 19 with an
aromatic ring appears to show slight increase in both
cytokines compared to aliphatic analogs 2 and 15–18.
This is suggestive of aromatic interaction(s) with
residues such as Phe77 and Trp131 in juxtaposition,
although the possible interaction of the phenyl ring with
the T cell receptor of NKT cells should not be omitted in
this three-component interaction.
Figure 2. Optimized result of 2 (green)/hCD1d complex (stereo view).
X-ray structure of 1 (gray) is superimposed.
In order to explain the above results of short phyto-
sphingosine analogs, we performed a molecular model-
ing of 2/hCD1d complex based on the crystal structure
of 1/hCD1d complex6 utilizing Maestro16 program.
Both acyl and phytosphingosine chains of 1 bound to
hCD1d in the crystal structure were truncated in silico
to correspond to 2, and the complex thus obtained
was further optimized17. Contrary to our expectation,
the optimized structure of 2 in the complex had only a
subtle, insignificant difference from 1 (Fig. 2).
Figure 3. Side view of the optimized structure of 2 (green)/hCD1d
complex. X-ray structure of 1 (red) is superimposed.
As can be seen from Figure 3, the phytosphingosine
chain of 2 has the precise length to reach the entrance
of the C0 pocket. It can well be assumed from the model
that C1–C5 segment of the phytosphingosine chain con-
tributes little to the interaction with CD1d, and this
might result in the short duration of 2 in complex with
CD1d.7 We have calculated the binding energy of
CD1d and the ligands, whose acyl chain length was fixed
to that of 2 and phytosphingosine chain altered incre-
mentally from C2 to C9.19 The binding energy increased
linearly as the number of the methylene unit increased
(data not shown), and no clear changing point was
observed. It may be assumed that ‘‘static’’ energy
calculation does not reflect the dynamics and the
probability of the phytosphingosine chain to reside
outside of the pocket. Dynamic simulation regarding
the movement of the short chain is now in progress.
Acknowledgments
Authors thank Dr. Y. Tanaka, Ms. M. Goto, Ms. K.
Yamaoka, and Ms. M. Nakajima for performing cyto-
kine assays. Ms. J. Futamura is acknowledged for initial
contribution to this program. Dr. T. Nishihara is
acknowledged for support and encouragement through-
out this study.
References and notes
1. For reviews: (a) Porcelli, S. A.; Modlin, R. L. Annu. Rev.
Immunol. 1999, 17, 297; (b) Hong, S.; Scherer, D. C.;
Singh, N.; Mendiratta, S. K.; Serizawa, I.; Koezuka, Y.;
Van Kaer, L. Immunol. Rev. 1999, 169, 31; (c) Sidobre, S.;
Kronenberg, M. J. Immunol. Methods 2002, 268, 107; (d)
Wilson, M. T.; Van Kaer, L. Curr. Pharm. Des. 2003, 9,
201; (e) Cox, T. Philos. Trans. R. Soc. Lond. B 2003, 358,
967; (f) Brigl, M.; Brenner, M. B. Annu. Rev. Immunol.
2004, 22, 817.
2. (a) Morita, M.; Motoki, K.; Akimoto, K.; Natori, T.;
Sakai, T.; Sawa, E.; Yamaji, K.; Koezuka, Y.; Kobayashi,
E.; Fukushima, H. J. Med. Chem. 1995, 38, 2176; (b)
Morita, M.; Sawa, E.; Yamaji, K.; Sakai, T.; Natori, T.;
Koezuka, Y.; Fukushima, H.; Akimoto, K. Biosci.
Biotechnol. Biochem. 1996, 60, 288; (c) Takikawa, H.;
Muto, S.; Mori, K. Tetrahedron 1998, 54, 3141; (d)
Graziani, A.; Passacantilli, P.; Piancatelli, G.; Tani, S.
Tetrahedron: Asymmetry 2000, 11, 3921.
Several analogs related to 1 have been prepared to
date, and many of them are equipotent to or even more
potent than 1 in the aspect of IFN-c secretion. In this
study, the minimum structure requirement to exhibit
equivalent profiles to 2 was disclosed, and also first
examples of non-linear hydrocarbon chain phytosphin-
gosine analogs that show similar cytokine inducing
profile to 2 were discovered. The possibility of this
alteration allows for synthetic access to the future ana-
logs improved in their pharmacological and physico-
chemical properties.

2784
T. Toba et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2781–2784
3. (a) Miyamoto, K.; Miyake, S.; Yamamura, T. Nature
10% fetal bovine serum (purchased from GIBCO),
5 · 10 ꢀ5 M 2-mercaptoehtanol (purchased from GIBCO),
1 mM pyruvate (purchased from SIGMA), and 25 mM
HEPES (purchased from SIGMA). The cells (5 · 105 cells/
well) were stimulated with glycolipid derivatives at a
concentration of 100 ng/ml for 72 h at 37 ꢁC in a 96-well
flat bottomed plate (purchased from IWAKI), and the
concentration of IL-4 and IFN-c in the culture superna-
tant was measured by ELISA (BD Pharmingen EIA Kit).
OCH 2 was always included in the assay as a control and
the cytokine release were expressed as relative to that of 2
for the mean of at least three experiments.
2001, 413, 531; (b) Yamamura, T.; Miyake, S. PCT Int.
Appl. WO 2003/016326, 2003; Chem. Abstr. 2003, 138,
205292m; (c) Yamamura, T.; Miyamoto, K.; Illes, Z.; Pal,
E.; Araki, M.; Miyake, S. Curr. Top. Med. Chem. 2004, 4,
561; (d) Oki, S.; Chiba, A.; Yamamura, T.; Miyake, S.
J. Clin. Invest. 2004, 113, 1631.
4. Zhou, D.; Mattner, J.; Cantu, C., III; Schrantz, N.; Yin,
N.; Gao, Y.; Sagiv, Y.; Hudspeth, K.; Wu, Y.-P.;
Yamashita, T.; Teneberg, S.; Wang, D.; Proia, R. L.;
Levery, S. B.; Savage, P. B.; Teyton, L.; Bendelac, A.
Science 2004, 306, 1786.
5. Zeng, Z.-H.; Castano, A. R.; Segelke, B. W.; Stura, E. A.;
Peterson, P. A.; Wilson, I. A. Science 1997, 277, 339.
6. Koch, M.; Stronge, V. S.; Shepherd, D.; Gadola, S. D.;
Mathew, B.; Ritter, G.; Fersht, A. R.; Besra, G. S.;
Schmidt, R. R.; Jones, E. Y.; Cerundolo, V. Nat. Immunol.
2005, 6, 819.
14. From the disclosed information in Ref. 9, the only
difference in the experimental condition is the stimulation
time, 72 h for us versus 60 h for Savage’s group.
15. An analog bearing –(CH2)4CH3 for R2 was prepared but
with an acyl chain shorter by one methylene unit
[–NHCO(CH2)21CH3]. The cytokine induction by this
compound was 97% for IL-4 and 113% for IFN-c, relative
to 2.
7. Oki, S.; Chiba, A.; Yamamura, T.; Miyake, S. J. Clin.
Invest. 2004, 113, 1631.
8. Chiba, A.; Oki, S.; Miyamto, K.; Hashimoto, H.;
Yamamura, T.; Miyake, S. Arthritis Rheum. 2004, 50, 305.
9. Goff, R. D.; Gao, Y.; Mattner, J.; Zhou, D.; Yin, N.;
Cantu, C., III; Teyton, L.; Bendelac, A.; Savage, P. B.
J. Am. Chem. Soc. 2004, 126, 13602.
10. Fujio, M.; Wu, D.; Garcia-Navarro, R.; Ho, D. D.; Tsuji,
M.; Wong, C.-H. J. Am. Chem. Soc. 2006, 128, 9022.
11. (a) Annoura, H.; Murata, K.; Yamamura, T. PCT Int.
Appl. WO2004/072091, 2004; Chem. Abstr. 2004, 141,
225769n; (b) Pacifichem 2005, Honolulu, Hawaii, Dec.
15–20, 2005, ID#378.
16. Maestro Ver. 7.5.112; Schro¨dinger, LLC: New York.
17. Optimization of the complex was performed stepwise as
follows, utilizing Macromodel Ver. 9.018 [force field
OPLS2005/solv.Water] (convergence threshold .05 kJ/
˚
mol/A): (i) main chain, Asp80, Asp151, Thr154 and ligand
fixed, (ii) main chain, Asp80(Od1, Od2), Asp151(Od1, Od2),
Thr154(Oc) and oxygen atoms of the ligand fixed, (iii)
main chain, distances among selected ligand atoms and
Asp80, Asp151, Thr154 fixed, (iv) main chain fixed, (v)
optimization of all the atoms.
18. Macromodel Ver. 9.0 (Schro¨dinger, LLC); Mohamadi, F.;
Richards, N. G. J.; Guida, W. C.; Liskcamp, R.; Lipton,
M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C.
J. Comput. Chem. 1990, 11, 440.
12. Murata, K.; Toba, T.; Nakanishi, K.; Takahashi, B.;
Yamamura, T.; Miyake, S.; Annoura, H. J. Org. Chem.
2005, 70, 2398.
13. Splenocytes were prepared from the spleens of C57BL/6
mice (6–8 weeks old, female) and suspended in a
RPMI1640 medium (purchased from Nacalai) containing
19. The optimization was performed for each analog
according to the same procedure as for 2 described in
Ref. 17.