C-Alkyl 5-membered ring imino sugars as new potent cytotoxic glucosylceramide synthase inhibitors

Article abstract of DOI:10.1039/b613460b

The stereoselective preparation of C-alkyl 5-membered ring imino sugars, their biological evaluation with regard to glucosyleramide synthase (GCS) inhibition, and cytotoxicity in a murine melanoma B16 cell model, were investigated. It was found that epoxy

Full text of DOI:10.1039/b613460b

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
C-Alkyl 5-membered ring imino sugars as new potent cytotoxic
glucosylceramide synthase inhibitors
Vanessa Faugeroux,a Yves Ge´nisson,*a Nathalie Andrieu-Abadie,b Sandra Colie´,b Thierry Levadeb and
Michel Baltasa
Received 15th September 2006, Accepted 19th October 2006
First published as an Advance Article on the web 2nd November 2006
DOI: 10.1039/b613460b
The stereoselective preparation of novel C-alkyl 5-membered
ring imino sugars and their biological evaluation with regard
to GCS inhibition and cytotoxicity in a murine melanoma
model are reported.
Glucosylceramide synthase (GCS, EC 2.4.1.80) catalyses the
transfer of glucose to the C-1 hydroxyl of ceramide, the first step
in the biosynthesis of glucosylceramide-based glycosphingolipids
(GSLs). Slowing glucosylceramide production already consti-
tutes a relevant approach against lysosomal storage disorders
characterized by inherited deficiencies in GSL catabolism.1 This
strategy, the so called substrate deprivation therapy, resulted in
R
the commercialisation of Zavescaꢀ (N-butyl-1-deoxynojirimycin)
for the treatment of mild to moderate type I Gaucher disease.
Notably, GSLs have been involved in many cellular processes
Fig. 1 Structure of two GCS inhibitors.
including cell–cell communication, cell adhesion, differentiation,
proliferation and oncogenic transformation.2 Overexpression of
an attempt to enhance GCS inhibition through better ceramide
GCS observed in cancer cells has been related to the ineffec-
mimicry.10 Potent b-glucocerebrosidase inhibitors based on a
tive host immune response and to tumour progression and/or
6-alkyl isofagomine scaffold have also been described by Fan.11
metastasis.3 Moreover, by reducing the accumulation of the pro-
Interestingly, the potential of 5-membered ring imino sugars in the
apoptotic ceramide, GCS could prevent cancer cell death in
context of sphingolipid metabolism remains almost unexplored.12
response to some chemotherapeutic agents.4 Conversely, GCS
We anticipated that a pyrrolidine-based imino sugar could
inhibition may resensitize multidrug resistant (MDR) cancer cells
represent a suitable molecular scaffold to elaborate GCS in-
to antineoplastic agents, suggesting that an elevated GCS activity
hibitor candidates. In the course of the development of a
participates in a new MDR mechanism.5 These observations have
stereoselective route towards imino sugars, we gained access to
led to the concept that GCS may represent a key enzymatic target
the versatile epoxypyrrolidine intermediate 2, which appeared
to be a suitable precursor of novel C-alkyl 5-membered ring
in anti-cancer chemotherapy.6
Known synthetic low molecular mass GCS inhibitors belong
imino sugars (Scheme 1).13 We decided to take advantage of a
to two main classes. The “P drug” family encompasses highly
regioselective epoxide opening reaction to achieve branching of the
lipophilic D-threo sphingolipid analogues based on a 1-phenyl-2-
aliphatic chain in the 4-position, yielding all-trans trisubstituted
pyrrolidines. Further oxidative manipulation of the vinyl moiety
would allow access to a proper imino sugar framework through
alkoylamino-3-amino-1-propanol pattern (Fig. 1).7 The N-butyl-
1-deoxynojirimycin (NB-DNJ), and analogues thereof, represent
the second category of GCS inhibitors.8 Derived from imino
formation of an hydroxymethyl group at C-2. The structural and
sugars, these non-cytotoxic derivatives typically retain their glu-
stereochemical homology between the aminodiol moiety of the
cosidase inhibition potency, which represents a limitation to the
therapeutic use of NB-DNJ.
Whereas the N-alkyl imino sugars such as NB-DNJ have
been extensively studied, examples of their C-alkyl congeners are
less common and concern 1-substituted derivatives.9 Compain
and Martin have prepared series of N-alkyl-1-deoxynojirimycin
analogues bearing additional C-branched lipophilic chains in
aLaboratoire de Synthe`se et Physicochimie de Mole´cules d’Inte´reˆt
Biologique, UMR 5068CNRS/Universite´ Paul Sabatier, 118 route
de Narbonne, 31062, Toulouse Cedex 9, France. E-mail: genisson@
chimie.ups-tlse.fr; Fax: (+33) (0)5 61 55 82 45
bInstitut Louis Bugnard/INSERM U.466 CHU Rangueil, BP 84225, 31432,
Toulouse Cedex 4, France
Scheme 1 Structure of the ceramide, of the GCS inhibitor candidate 1
and of its synthetic precursor.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 4437–4439 | 4437
©

View Article Online
targeted pyrrolidines and that of the sphingosine backbone of
ceramide further prompted us to initiate this study.
We thus report here the stereoselective preparation and the bio-
logical evaluation in regard with GCS inhibition and cytotoxicity
of C-alkyl 5-membered ring imino sugars of type 1.
Introduction of the n-alkyl residue was achieved via treatment of
epoxide 2 with the corresponding Gilman-like cuprate (Scheme 2).
Remarkably, the expected adducts were isolated as a unique regio-
and diastereoisomer. Thus a methyl, n-butyl and n-octyl residue
could be readily branched in 60 to 77% yields. In order to prepare
analogues lacking the lipophilic chain, unsubstituted intermediate
6 was also prepared in 95% yield by totally regiocontrolled
C-4 hydride delivery. This level of regioselectivity is worth noting
with regard to the relatively modest steric bulk generated by
the vinyl moiety. Setting up of the hydroxymethyl appendage
was smoothly accomplished using a metal-free oxidative cleavage.
Thus, ozonolysis of olefins 4–6 was carried out after protection
of the amino group by salification. Immediate reduction of the
resulting sensitive amino aldehydes afforded the expected inter-
mediates 7–9 in 52 to 70% overall yields. A final hydrogenolysis
gave the targeted pyrrolidines 10–12 in good yields. Treatment of
intermediate 4 and 6 under same conditions delivered derivatives
13 and 14 bearing an ethyl residue in place of the hydroxymethyl
group.
Scheme 3 Synthesis of the tested piperidines. Reagents and conditions:
˚
(a) BnBr, NaH, NaI, 4 A MS, DMF, R = Me, 85%; R = n-butyl, 84%;
(b) (i) HCl, MeOH, 0 ◦C; (ii) O3, MeOH, −78 ◦C; (iii) NaBH4, MeOH,
−78 ◦C to −10 ◦C, 15, 63%; 16, 63%; (c) (i) TFAA, p-dioxane; (ii) Et3N,
reflux; (iii) NaOH 2.5 M, R = Me, 91% brsm at 85% conversion; R =
n-butyl, 78% brsm at 86% conversion; (d) 7–8 bar H2, Pd/C, conc. HCl,
MeOH, 17, 71%; 18, 78%.
this event, intermediates 3 and 4 were first benzylated before being
converted to primary alcohols 15 and 16 via ozonolysis. These
3,4-disubstituted precursors initially proved to be poorly reactive
when treated with trifluoroacetic anhydride and Et3N. However,
after careful optimisation of the reaction conditions, the expected
enlarged products could be secured in good yields. In this respect,
the use of p-dioxane as solvent proved decisive. The deprotected
piperidines 17 and 18 were finally obtained after hydrogenolysis
of the benzyl groups.
5-Membered ring imino sugars 8 and 10–14 were tested on
a mouse melanoma B16 cell line (Fig. 2).15 Indeed, melanoma is
largely considered to be a radiation- and chemotherapy-refractory
neoplasm. As observed after incubating intact living cells with
a fluorescent ceramide substrate, not only pyrrolidine 10 (at
50 lM) exerted a 34% GCS inhibition but also this potency was
comparable to that of NB-DNJ under the same conditions (49%
inhibition).16 In order to distinguish the contribution to activity
of the different structural features of imino sugar 10, we then
tested a series of analogues. The primary hydroxyl group appeared
crucial as its replacement by a methyl residue proved unfavourable,
giving around 6% inhibition for 13. The observed drop in potency
(16% inhibition) for the C-4 unsubstituted compound 12 clearly
indicated the importance of the lipophilic portion of inhibitor 10.
Compound 14, lacking both the primary hydroxyl group and the
Scheme 2 Synthesis of the tested pyrrolidines. Reagents and conditions:
(a) Me2CuLi, or n-Bu2CuLi, or n-Oct2CuLi, Et2O, −20 ◦C, 3, 77%; 4, 63%;
5, 60%; (b) LiAlH4, Et2O, 0 ◦C, 6, 95%; (c) (i) HCl, MeOH, 0 ◦C; (ii) O3,
MeOH, −78 ◦C; (iii) NaBH4, MeOH, −78 to −10 ◦C, 7, 70%; 8, 52%;
9, 68%; (d) 7–8 bar H2, Pd/C, conc. HCl, MeOH, 10, 90%; 11, 52%; 12,
quant. brsm at 82% conversion; (e) 7–8 bars H2, Pd/C, MeOH, 13, quant.;
14, 40%.
In order to highlight the eventual relevance of the pyrrolidinic
scaffold of our GCS inhibitor candidates, we decided to compare
them to their direct 6-membered ring analogues. Having in hands
2-hydroxymethyl pyrrolidines, we thus focused our attention on
the stereospecific ring expansion reaction developed by Cossy.14
Although precedents are scarce, such a process should allow
access to all-trans 3-alkyl-4,5-dihydroxy piperidines (Scheme 3).
Interestingly, these structures could also be viewed as relevant
2-deoxy-2-alkyl DNJ analogues. Indeed, it has been proposed,
on the basis of Butters’ ceramide mimicry model, that an
alkyl chain at the 2-position of NB-DNJ would simulate the
hydrophobic portion of the sphingosine framework (Fig. 1).10 In
Fig. 2 Glucosylceramide synthase inhibition. Melanoma cells were
pre-incubated overnight with the indicated compounds (10 lM for
8 and 11, 50 lM for 10, 12–14 and NB-DNJ). NBD-C6-ceramide
(5 lM) was added and after 2 h lipids were extracted. The extent
of NBD-C6-glucosylceramide formation was then assessed. Data are
expressed as a percentage of the values of B16 untreated cells.
4438 | Org. Biomol. Chem., 2006, 4, 4437–4439
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
C-alkyl chain, was also shown to be inactive. Notably, introduction
of an n-octyl residue, instead of an n-butyl, was accompanied by a
dramatic enhancement in potency, reaching 73% GCS inhibition
(at only 10 lM) for 11. Finally, the detrimental influence of the
substitution at the nitrogen atom was suggested by the comparison
of 11 with its N-benzyl analogue 8 (10% inhibition at 10 lM).
Remarkably, the pyrrolidinic framework of these C-alkyl
imino sugars revealed a determinant for their activity, since the
piperidines 17 and 18 only displayed a marginal GCS inhibition.
We next investigated the potential cytotoxic effect of our new
GCS inhibitors on the B16 cell line (Fig. 3). Compound 10
exhibited dose-dependent cytotoxicity, reaching 50% activity at
25 lM. Additional findings, namely caspase activation17 and an
elevation of intracellular ceramide concentration (not shown),18
suggest that derivative 10 triggered a ceramide-mediated apoptotic
cell death. Notably, cytotoxicity was raised to 70% at 5 lM with
the more potent inhibitor 11.
new prospects for the development of chemotherapeutic anti-
cancer agents potentially acting through alterations of ceramide
metabolism.20
References
1 T. D. Butters, R. A. Dwek and F. M. Platt, Curr. Top. Med. Chem.,
2003, 3, 561–574.
2 A. Huwiler, T. Kolter, J. Pfeilschifter and K. Sandhoff, Biochim.
Biophys. Acta, 2000, 1485, 63–99.
3 N. S. Radin, Biochem. Pharmacol., 1999, 57, 589–595.
4 Y. Y. Liu, T. Y. Han, A. E. Giuliano and M. C. Cabot, J. Biol. Chem.,
1999, 274, 1140–1146; R. J. Bleicher and M. C. Cabot, Biochim. Biophys.
Acta, 2002, 1585, 172–178.
5 V. Gouaze, Y. Y. Liu, C. S. Prickett, J. Y. Yu, A. E. Giuliano and M. C.
Cabot, Cancer Res., 2005, 65, 3861–3867.
6 C. P. Reynolds, B. J. Maurer and R. N. Kolesnick, Cancer Lett., 2004,
206, 169–180.
7 A. Abe, S. R. Wild, L. Lee and J. A. Shayman, Curr. Drug Metab., 2001,
2, 331–338.
8 R. A. Dwek, T. D. Butters, F. M. Platt and N. Zitzmann, Nat. Rev.
Drug Discovery, 2002, 1, 65–75.
9 For recent reports see: J.-Y. Goujon, D. Gueyrard, P. Compain, O. R.
Martin, K. Ikeda, A. Kato and N. Asano, Bioorg. Med. Chem., 2005,
13, 2313–2324; L. A. Augustin, J. Fantini and D. R. Mootoo, Bioorg.
Med. Chem., 2006, 14, 1182–1188; M. S. M. Pearson, R. O. Saad, T.
Dintinger, H. Amri, M. Mathe´-Allainmat and J. Lebreton, Bioorg. Med.
Chem. Lett., 2006, 16, 3262–3267.
10 C. Boucheron, V. Desvergnes, P. Compain, O. R. Martin, A. Lavi,
M. Mackeen, M. Wormald, R. Dwek and T. D. Butters, Tetrahedron:
Asymmetry, 2005, 16, 1747–1756.
11 X. Zhu, K. A. Sheth, S. Li, H.-H. Chang and J.-Q. Fan, Angew. Chem.,
Int. Ed., 2005, 44, 7450–7453.
12 Davis reported inhibition of GCS by racemic (2S*,3S*,4S*)-2-ethyl-
3,4-dihydroxy pyrrolidine: T. M. Chapman, S. Courtney, P. Hay and
B. G. Davis, Chem.–Eur. J., 2003, 9, 3397–3414. We prepared Davis’
pyrrolidines in an enantiomerically pure form from epoxypyrrolidine
2 and observed that they both displayed a ca. 20% GCS inhibition (at
50 mM) in the B16 melanoma cell line.
Fig. 3 Sensitivity of melanoma cells to compounds 10 or 11. Melanoma
B16 cells were seeded in 24-well plates and at 50% confluency, compounds
10 or 11 were added at the indicated concentrations. After 48 h, cell viability
was estimated by assessing the cellular MTT conversion capacity. Data are
expressed as a percentage of the values of untreated cells. Shown are the
means SE (n = 2–5).
13 T. Ayad, Y. Ge´nisson, M. Baltas and L. Gorrichon, SYNLETT, 2001,
6, 866–868; T. Ayad, Y. Ge´nisson, S. Broussy, M. Baltas and L.
Gorrichon, Eur. J. Org. Chem., 2003, 2903–2910; T. Ayad, Y. Ge´nisson,
M. Baltas and L. Gorrichon, Chem. Commun., 2003, 582–583; T. Ayad,
The present results showed that imino sugars 10 and 11
displayed an activity profile contrasting with that of the non-
cytotoxic deoxynojirimycin-like GCS inhibitors.19 This might be
related to the branching position of the aliphatic residue on the
imino sugar scaffold, which has been shown to influence the
inhibition of sphingolipid glucosidic processing. This structural
feature might somehow broaden their inhibition pattern and hence
modulate their overall activity on cancer cells.
We thus described a straightforward and enantioselective access
to 4-alkyl pyrrolidine-based imino sugars, relying on a regio-
and stereocontrolled epoxide opening reaction, as well as their
conversion into six-membered ring imino sugars.
The novel C-alkyl 5-membered ring imino sugars behaved as
potent GCS inhibitors, displaying furthermore a marked cytotoxic
behaviour in a murine melanoma model. The close resemblance
between these structures and the sphingosine backbone of sphin-
golipids suggests that they might act as ceramide mimics. Studies
aiming at elucidation of their mode of enzymatic interaction
are currently pursued. The present work should also open up
V. Faugeroux, Y. G
e´nisson, C. Andre´, M. Baltas and L. Gorrichon,
J. Org. Chem., 2004, 69, 8775–8779; T. Ayad, Y. Ge´nisson and M.
Baltas, Org. Biomol. Chem., 2005, 3, 2626–2631.
14 J. Cossy, C. Dumas and D. Gomez Pardo, Eur. J. Org. Chem., 1999,
1693–1699; J. Cossy, Chem. Rec., 2005, 5, 70–80 and references cited
therein. To our knowledge, synthetic applications making use of
3,4-disubstituted pyrrolidines rely on the use of chlorine anions as
nucleophilesto give 3-chloropiperidines.
15 For a precedent on the influence of an imino sugar-based GCS inhibitor
on melanoma tumor growth, see: M. Weiss, S. Hettmer, P. Smith and
S. Ladisch, Cancer Res., 2003, 63, 3654–3658.
16 A. D. Tepper, S. H. Diks, W. J. van Blitterswijk and J. Borst, J. Biol.
Chem., 2000, 275, 34810–34817.
17 Executioner caspase activity, as measured by the cleavage of the
fluorogenic tetrapeptide substrate DEVD-AMC, peaked at 6 h. C.
Tardy, H. Autefage, V. Garcia, T. Levade and N. Andrieu-Abadie,
J. Biol. Chem., 2004, 279, 52914–52923.
18 Conversion of ceramide into sphingomyelin was also monitored and
none of the tested imino sugars displayed inhibition of sphingomyelin
synthase.
19 The non-cytotoxic behaviour of NB-DNJ was confirmed experimen-
tally on B16 cell line.
20 B. Se´gui, N. Andrieu-Abadie, J.-P. Jaffre´zou, H. Benoist and T. Levade,
Biochim. Biophys. Acta, 10.1016/j.bbamem.2006.05.024.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 4437–4439 | 4439
©