S-linked ganglioside analogues for use in conjugate vaccines

Article abstract of DOI:10.1021/ol036460p

(Equation presented) Glycosidase resistant thioglycoside precursors of the melanoma-associated ganglioside GM₂ have been synthesized starting from lactose. Syntheses of several analogues of ganglioside GM₃ and a positional isomer hav

Full text of DOI:10.1021/ol036460p

ORGANIC
LETTERS
2004
Vol. 6, No. 6
897-900
S-Linked Ganglioside Analogues for Use
in Conjugate Vaccines
Jamie R. Rich* and David R. Bundle
Alberta Ingenuity Centre for Carbohydrate Science, Department of Chemistry,
UniVersity of Alberta, Edmonton, AB T6G 2G2, Canada
jrich@ualberta.ca
Received December 18, 2003
ABSTRACT
Glycosidase resistant thioglycoside precursors of the melanoma-associated ganglioside GM have been synthesized starting from lactose.
2
Syntheses of several analogues of ganglioside GM and a positional isomer have been developed. These compounds contain thio-linked sialic
3
acid residues and a modified ceramide aglycon functionalized for coupling to proteins, surfaces, or matrices. The hydrolytic stability of these
oligosaccharides enhances the immunogenicity of the corresponding conjugate vaccines by ensuring their integrity in the acidic compartments
of antigen processing cells.
The overexpression of glycolipid and glycoprotein structures
on the surface of cancerous cells makes these structures
attractive targets for use in the active immunotherapy of
tumors.1 Gangliosides are a class of N-acetyl neuraminic acid
(Neu5Ac) containing glycolipid molecules that are expressed
at greatly elevated concentrations in melanomas and other
cancers of neuroectodermal origin.2 Low levels of immune
response to vaccines containing naturally occurring ganglio-
side antigens have thus far proven prohibitive to the
development of a viable anticancer vaccine.1b,c There is
therefore a need for improved design of ganglioside-based
melanoma vaccines that have the capability to induce a strong
humoral and if possible also a cell-mediated immune
response, since recent preclinical and clinical studies with
poorly immunogenic semisynthetic GM2 vaccines linked to
KLH enjoyed only limited success.
Glycosidase-resistant thioglycoside analogues of the mela-
noma-associated ganglioside GM2 are attractive candidates
as components of improved vaccines since their much slower
rates of either chemical or enzymatic hydrolysis3 would
render these antigens more persistent. Conformational studies
of S-linked oligosaccharides (thiooligosaccharides) suggest
that although the flexibility about the anomeric linkage is
increased for these molecules, they sample similar confor-
mational space to their natural O-linked counterparts.4 We
propose, based on preliminary immunochemical evidence,5
that the inclusion of thiooligosaccharides in conjugate vaccine
(3) (a) Defaye, J.; Gelas, J. In Studies in Natural Products Chemistry;
Atta-ur-Rahman, Ed.; Elsevier: Amsterdam, The Netherlands, 1991; pp
315-357. (b) Driguez, H. ChemBioChem 2001, 2, 311-318. (c) Fair-
weather, J. K.; Driguez, H. In Carbohydrates in Chemistry and Biology;
Ernst, B., Hart, G. W., Sinay¨, P., Eds.; Wiley-VCH: Toronto, Canada, 2000;
Vol. 1, pp 531-564. (d) Driguez, H. Top. Curr. Chem. 1997, 187, 85-
116.
(1) (a) Hakomori, S. Cancer Res. 1985, 45, 2405-2414. (b) Allen, J.
R.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2000, 39, 836-863. (c) Bitton,
R. J.; Guthmann, M. D.; Gabri, M. R.; Carnero, A. J.; Alonso, D. F.;
Fainboim, L.; Gomez, D. E. Oncol. Rep. 2002, 9, 267-276.
(2) (a) Tsuchida, T.; Saxton, R. E.; Morton, D. L.; Irie, R. F. J. Natl.
Cancer Inst. 1987, 78, 45-54. (b) Hamilton, W. B.; Helling, F.; Lloyd, K.
O.; Livingston, P. O. Intl. J. Cancer 1993, 53, 566-573.
(4) (a) Geyer, A.; Hummel, G.; Eisele, T.; Reinhardt, S.; Schmidt, R. R.
Chem. Eur. J. 1996, 2, 981-988. (b) Aguilera, B.; Jime´nez-Barbero, J.;
Ferna´ndez-Mayoralas, A. Carbohydr. Res. 1998, 308, 19-52. (c) Weimar,
T.; Kreis, U. C.; Andrews, J. S.; Pinto, B. M. Carbohydr. Res. 1999, 315,
222-233. (d) Montero, E.; Vallmitjana, M.; Pe´rez-Pons, J. A.; Querol, E.;
Jime´nez-Barbero, J.; Can˜ada, F. J. FEBS Lett. 1998, 421, 243-248.
(5) Yu, H. N. Doctoral Thesis, University of Alberta, 2002.
10.1021/ol036460p CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/18/2004

preparations will result in the production of antibodies, which
cross-react with gangliosides on the tumor cell surface.
Owing to their metabolic stability S-linked gangliosides
should serve as antigens with extended in vivo stability, and
thus they are one part of our strategy to enhance the
immunogenicity of carbohydrate-based anticancer vaccine
preparations.6
Scheme 1 a
Ganglioside GM3 (1) is the most commonly expressed
ganglioside structure in human cells, and is a biosynthetic
precursor to tumor-associated gangliosides of higher com-
plexity such as GM2 and GD3. We report here the synthesis
of GM3 analogues 2 and 3 as well as their positional isomer
Figure 1. GM3 (1) and GM3 analogues 2 and 3.
4 (Scheme 1) for all of which the terminal Neu5Ac residue
is attached via a thioglycosidic linkage. These structures bear
a modified ceramide aglycon that has been designed to
facilitate coupling of the neoglycolipids to carrier protein
while maintaining the stereochemical integrity and relevant
structural features found in naturally occurring ceramides.7
GM3 analogues 2 and 3 and their precursors are poised for
chemical and enzymatic elaboration into analogues of more
complex gangliosides.
Before embarking on a synthesis of a sulfur-containing
GM3 oligosaccharide we sought to identify suitable chem-
istries for the assembly of our target. To this end, we
envisioned the construction of an isomer (4) that would serve
as an appropriate model system. It was expected that the
formation of the thioglycosidic linkage at the primary
position would proceed most readily. The synthesis of an R
2-6 linked thiooligosaccharide is outlined in Scheme 1.
Hydrolysis of the benzylidene acetal 58 afforded the
crystalline diol, which after selective sulfonylation of the
primary alcohol yielded triflate 6. Displacement of the triflate
with a Neu5Ac thiolate, generated in situ from thioacetate
79 by treatment with diethylamine,10 yielded the trisaccharide
in 86% yield.11 Acetylation of this mixture followed by
deprotection8 of the anomeric position yielded the hemi-
a Reagents and conditions: (a) 80% AcOH, 50 °C, 9 h, 86%;
(b) Tf2O, pyridine, DCM, -25 °C, 1 h, 72%; (c) Et2NH, 7, DMF,
-25 °C, 2 h, 86%; (d) Ac2O, pyridine, rt, 12 h, 96%; (e) TFA/
toluene (1:1), rt, 5 h; (f) Cl3CCN, DBU, DCM, 0 °C, 5 h; (g)
(2S,3R)-2-azido-3-O-benzoyl-4-pentene-1,3-diol, BF3Et2O, Drierite,
-10 to 20 °C, 2.5 h, 71%; (h) (i) PPh3, pyridine/H2O, 50 °C, 4 h,
(ii) N-(octadecanoyloxy)succinimide, 50 °C, 5 h, 88%; (i) (i)
NaOMe, MeOH, rt, 24 h, (ii) NaOH, rt, 24 h, 84%.
acetal.12 Conversion to the trichloroacetimidate13 and cou-
pling with (2S,3R)-2-azido-O-benzoyl-4-pentene-1,3-diol7
afforded the glycoside 10. This aglycon represents a truncated
version of azidosphingosine reported by Schmidt et al.14 and
was designed to allow (a) N-acylation to form ceramide-
type structures and (b) opportunities for coupling of the
glycolipid to protein via either the amine or olefin. The azide
in 10 was reduced with triphenylphosphine, and the amine
acylated. Removal of the acetates by transesterification and
saponification of the remaining methyl ester gave thio-
sialoside 4.
(11) Alternatively, 8 could be obtained directly from the diol without
purification of the sulfonate ester. Interestingly, some acetyl transfer to the
4′-OH (presumably originating from sulfur) was also observed in this case
and 9 (7%) accompanied the formation of 8 (73%).
(12) Hasegawa, A.; Morita, M.; Ito, Y.; Ishida, H.; Kiso, M. J. Carbohydr.
Chem. 1990, 9, 369-392.
(13) (a) Schmidt, R. R.; Michel, J. Angew. Chem., Int. Ed. 1980, 19,
731-732. (b) Schmidt, R. R.; Kinzy, W. AdV. Carbohydr. Chem. Biochem.
1994, 50, 21-123.
(14) (a) Schmidt, R. R.; Zimmermann, P. Tetrahedron Lett. 1986, 27,
481-484. (b) Schmidt, R. R.; Zimmermann, P. Angew. Chem., Int. Ed.
Engl. 1986, 25, 725-726. (c) Zimmermann, P.; Bommer, R.; Bar, T.;
Schmidt, R. R. J. Carbohydr. Chem. 1988, 7, 435-452.
(6) Metabolically stable glycosides have recently been proposed for use
in vaccines: (a) Kuberan, B.; Sikkander, S. A.; Tomiyama, H.; Linhardt,
R. J. Angew. Chem., Int. Ed. 2003, 42, 2073-2075. (b) Bousquet, E.;
Spadaro, A.; Pappalardo, M. S.; Bernardini, R.; Romeo, R.; Panza, L.;
Ronsisvalle, G. J. Carbohydr. Chem. 2000, 19, 527-541.
(7) Ling, C.-C.; Bundle, D. R. Unpublished results.
(8) Jansson, K.; Ahlfors, S.; Frejd, T.; Kihlberg, J.; Magnusson, G.;
Dahmen, J.; Noori, G.; Stenvall, K. J. Org. Chem. 1988, 53, 5629-5647.
(9) Hasegawa, A.; Nakamura, J.; Kiso, M. J. Carbohydr. Chem. 1986,
5, 11-19.
(10) Bennett, S.; Von Itzstein, M.; Kiefel, M. J. Carbohydr. Res. 1994,
259, 293-299.
898
Org. Lett., Vol. 6, No. 6, 2004

There are several reports of the synthesis of sulfur-
containing Neu5Ac(R2-3)Gal linkages in the literature.15
Owing to the apparent difficulties encountered in the
installation of sulfur at this position,15d,16 we began with
synthesis of the disaccharide (Neu5Ac-S-(R2-3)Gal) rather
than the trisaccharide. The protected thiodisaccharide 18 was
obtained as outlined in Scheme 2. Galactoside 1117 was
Scheme 3 a
Scheme 2 a
a Reagents and conditions: (a) for 12: (i) benzaldehyde dimethyl
acetal, p-TSA, MeCN, rt, 25 min, (ii) BzCl, pyridine, rt, 12 h, 76%;
(b) for 13: (i) benzaldehyde dimethyl acetal, p-TSA, MeCN, rt,
25 min, (ii) NaH, BnBr, rt, 3 h, 71%; (c) CAN, MeCN/H2O (19:
1), rt, 30 min, 85% (12), 77% (13); (d) (i) Tf2O, pyridine, DCM,
-20 °C, 1 h, (ii) Bu4N+-NO2, MeCN, rt, 12 h, 72% (14), 40%
(15); (e) (i) Tf2O, pyridine, DCM, -20 °C to rt, 50 min, (ii) KSAc,
DMF, 60 °C, 12 h, 79%; (f) (i) hydrazinium acetate, DMF, rt, 90
min, (ii) 17, NaH, DMF, rt, 12 h, 67%.
a Reagents and conditions: (a) benzaldehyde dimethyl acetal,
DMF, p-TSA, rt, 13 h; (b) BzCl, pyridine, rt, 12 h, 70%; (c) CAN,
MeCN/H2O, rt, 40 min, 86%; (d) (i) Tf2O, pyridine, DCM, -40
°C, 70 min, (ii) Bu4N+-NO2, DMF, rt, 14 h, 91%; (e) (i) Tf2O,
pyridine, DCM, -65 °C to rt over 30 min, then rt for 1 h, (ii)
KSAc, DMF, 70 °C, 1 h, 79%; (f) hydrazinium acetate, DMF, rt,
1.5 h, then 17, NaH, DMF, Kryptofix-21, rt, 12 h, 65%; (g) 80%
AcOH, 45 °C, 48 h, 81%; (h) Ac2O, pyridine, rt, 12 h, 95%; (i)
toluene/TFA, rt, 2 h, 96%; (j) (i) Cl3CCN, DBU, DCM, rt, 1.5 h,
(ii) (2S,3R)-2-azido-3-O-benzoyl-4-pentene-1,3-diol, BF3Et2O, Drier-
ite, -15 °C, 24 h, 80%; (k) (i) PPh3, pyridine/H2O, 40 °C, 7 h, (ii)
N-(octadecanoyloxy)succinimide, 40 °C, 12 h, 71%; (l) (i) PPh3,
pyridine/H2O, 45 °C, 3 h, (ii) N-(S-acetyl-16-mercaptohexade-
canoyloxy)succinimide, 45 °C, 5 h, 74%; (m) (i) NaOMe, MeOH,
rt, 24 h, (ii) NaOH, rt, 24 h, 76% for 27, 82% for 28.
transformed to the 4,6-O-benzylidene acetal and the remain-
ing hydroxyl protected as a benzoate ester or benzyl ether.
Removal of the p-methoxybenzyl ether yielded 1218 or 13,
which after triflation could be converted to the gulosides 14
(72%) or 15 (40%) by displacement with tetrabutylammo-
nium nitrite.19 Installation of the sulfur atom was effected
in 79% yield by treatment of the triflate formed from alcohol
14 with potassium thioacetate. The benzyl-protected guloside
proved to be a poor substrate for the triflation-inversion
sequence and thus its further elaboration to the disaccharide
was abandoned. Removal of the thioester with hydrazine
acetate provided the thiol, which following treatment with
sodium hydride underwent reaction with 17 to afford the
disaccharide 18.
route identified for assembly of the thiodisaccharide. 3′-O-
p-Methoxybenzyl lactoside 2120 underwent benzylidene
acetal formation followed by benzoylation to give 22.
Following selective deprotection of the 3′-hydroxyl, triflation
of the alcohol and displacement with nitrite gave the axial
epimer. A second triflation-inversion sequence yielded
thioacetate 24.21 After reaction of the thiolate with Neu5Ac
The synthesis of a 3′-thiolactoside (Scheme 3) for the
construction of the GM3 trisaccharide followed closely the
(18) Previous synthesis: Lowary, T. L.; Eichler, E.; Bundle, D. R. J.
Org. Chem. 1995, 60, 7316-7327.
(15) (a) Kanie, O.; Nakamura, J.; Itoh, Y.; Kiso, M.; Hasegawa, A. J.
Carbohydr. Chem. 1987, 6, 117-128. (b) Eisele, T.; Schmidt, R. R. Liebigs
Ann. Recl. 1997, 865-872. (c) Wilson, J. C.; Kiefel, M. J.; Angus, D. I.;
von Itzstein, M. Org. Lett. 1999, 1, 443-446. (d) Turnbull, W. B.; Field,
R. A. J. Chem. Soc., Perkin Trans. 1 2000, 1859-1866.
(16) Marcaurelle, L. A.; Bertozzi, C. R. J. Am. Chem. Soc. 2001, 123,
1587-1595.
(19) (a) Lattrell, R.; Lohaus, G. Justus Leibigs Ann. Chem. 1974, 901-
920. (b) Albert, R.; Dax, K.; Link, R. W.; Stutz, A. E. Carbohydr. Res.
1983, 118, C5-C6.
(20) Hotta, K.; Ishida, H.; Kiso, M.; Hasegawa, A. J. Carbohydr. Chem.
1994, 13, 175-191.
(21) Efforts to employ the anomeric thiolate of sialic acid as a nucleophile
(see Scheme 1) to produce disaccharide 18 or trisaccharide 25 met with
failure, only 2,3-elimination products were obtained.
(17) Hasegawa, A.; Ando, T.; Kameyama, A.; Kiso, M. J. Carbohydr.
Chem. 1992, 11, 645-658.
Org. Lett., Vol. 6, No. 6, 2004
899

glycosyl chloride 17, the benzylidene acetal was replaced
by acetates, and the trisaccharide was converted to its
hemiacetal. Formation of the trichloroacetimidate and cou-
pling with (2S,3R)-2-azido-3-O-benzoyl-4-pentene-1,3-diol7
afforded the benzoylated azido alcohol 26. At this stage,
azide reduction, N-acylation with an activated octadecanoate,
and deprotection gave a GM3 analogue suitable for further
elaboration. For example, derivatives such as 27 will undergo
radical addition of cysteamine to the terminal olefin.22
Alternatively, 26 could be converted to 28, which bears a
thiol at the N-acyl chain terminus.23 These two strategies
allow for coupling of the thioganglioside analogue to carrier
protein via either the sphingosine base or hydrocarbon
portions of the aglycon.
A recent report24 describing the first synthesis of a 3′-thio
lactoside from lactose suggests that a 4,6-O-benzylidene
acetal is important for successful gulo to galacto inter-
conversions with KSAc and confirms our concurrent find-
ings. In our hands, only cyclic acetal (benzylidene or
isopropylidene)-protected (ga)lactosides undergo smooth
inversion with KSAc. To circumvent this protecting group
requirement we have demonstrated that tetrabutylammonium
thioacetate can be employed as a nucleophile when esters
are used as protecting groups as shown for the conversion
of 29 to 30 in Scheme 4.25 However, we found that acetal-
Scheme 5
have observed two side reactions as outlined in Scheme 5.
Acyl migration was frequently observed during the synthesis
of 14 and the analogous disaccharide, and could be sup-
pressed by conducting the reaction at lower temperatures.
While acyl migration (see 31) appears to be a general
feature26 of similarly protected galactosides, a second product
that we have tentatively assigned as 3227 was observed less
frequently, and only with the substrate shown in Scheme 5.
We are currently investigating the basis and scope of this
reaction.
In conclusion, we have developed syntheses of several
novel thioglycoside analogues of the ganglioside GM3. These
compounds are functionalized so as to allow for their facile
incorporation into carbohydrate-protein conjugates for use
as conjugate vaccines. GM3-type molecules 27 and 28 will
be elaborated by chemoenzymatic means to analogues of
higher gangliosides such as GM2, also for inclusion in
antimelanoma vaccine preparations. Similarly, intermediates
such as 18 and 25 will serve as advanced synthetic precursors
to analogues of gangliosides GM2, GM1, and GD1a. It is
expected that the enhanced metabolic stability of these
antigens will render them more immunogenic and that
antibodies raised in this manner will cross react with their
naturally occurring O-glycosides.
Scheme 4
Acknowledgment. We thank Dr. Chang-Chun Ling and
Mr. Graham Murphy for helpful discussions and Dr. Angie
Morales-Izquierdo for acquiring mass spectra. J.R. thanks
the Alberta Heritage Foundation for Medical Research
(AHFMR) and Natural Sciences and Engineering Research
Council of Canada (NSERC) for graduate fellowships. This
work was funded by the Alberta Ingenuity Centre for
Carbohydrate Science and the Canadian Institutes of Health
Research (CIHR).
protected 3-thio(ga)lactosides are superior nucleophiles for
reaction with 17 when compared to 3-thio-4,6-diesters.
The conversion of galacto-configured alcohols such as 12
and 23 to the corresponding gulo alcohols via displacement
with nitrite ion also bears consideration. We and others24
have found that inversion of configuration at C-3 of both
galactosides and gulosides precludes the use of a benzyl
protecting group at O-2. While the benzoate ester proves
superior in this regard, careful temperature optimization and
control are often required to realize acceptable yields. We
Supporting Information Available: Experimental pro-
cedures and 1H NMR and other spectral data. This material
is available free of charge via the Internet at http://pubs.acs.org.
(22) Liu, L. Doctoral Thesis. University of Alberta, 2003.
(23) (a) Ohlsson, J.; Magnusson, G. Tetrahedron Lett. 1999, 40, 2011-
2014. (b) Ohlsson, J.; Magnusson, G. Tetrahedron 2000, 56, 9975-9984.
(24) Liakatos, A.; Kiefel, M. J.; von Itzstein, M. Org. Lett. 2003, 5,
4365-4368.
(25) Rich, J. R.; Szpacenko, A.; Palcic, M. M.; Bundle, D. R. Angew.
Chem., Int. Ed. 2004, 43, 613-615.
OL036460P
(26) This reaction has been observed for â-octyl, â-allyl, and â-8-
methoxycarbonyloctyl galactosides and a variety of lactosides.
(27) The proposed structural assignment of 32 is based on analysis of
IR, HRMS, and H and 13C NMR data, including TROESY spectra.
1
900
Org. Lett., Vol. 6, No. 6, 2004