Synthesis of a new family of glycolipidic nitrones as potential antioxidant drugs for neurodegenerative disorders

Article abstract of DOI:10.1016/S0960-894X(02)01079-X

This paper deals with the synthesis of a novel series of amphiphilic glycosylated spin-traps derived from α-Phenyl-N-tert-butyl nitrone (PBN) and an initial characterization of their anti-caspase-3 activity. Preliminary investigation of their anti-apoptos

Full text of DOI:10.1016/S0960-894X(02)01079-X

Bioorganic & Medicinal Chemistry Letters 13 (2003) 859–862
Synthesis of a New Family of Glycolipidic Nitrones as Potential
Antioxidant Drugs for Neurodegenerative Disorders
Gregory Durand,a Ange Polidori,a,* Jean-Pierre Sallesb and Bernard Puccia,*
a
` ´ ´
Laboratoire de Chimie BioOrganique et des Systemes Moleculaire Vectoriels, Faculte des Sciences, 33 rue Louis Pasteur,
84000 Avignon, France
bTargeting System Pharma, 830 chemin de vergon, 13510 Eguilles, France
Received 15 October 2002; revised 16 December 2002; accepted 16 December 2002
Abstract—This paper deals with the synthesis of a novel series of amphiphilic glycosylated spin-traps derived from a-Phenyl-N-tert-
butyl nitrone (PBN) and an initial characterization of their anti-caspase-3 activity. Preliminary investigation of their anti-apoptosis
effect showed they dramatically inhibit the activity of caspase-3 in cultured neuronal cells following induction of apoptosis by
hydrogen peroxide.
# 2003 Elsevier Science Ltd. All rights reserved.
Oxygen free radicals are produced in numerous and
varied pathological states such Alzheimer’s disease,1
chronic neurodegenerative disorders,2 ischemia-reperfu-
sion or stroke.3 Agents that can trap these free radicals
to form more stable adducts may be useful as ther-
apeutic agents. a-Phenyl-N-tert-butyl nitrone (PBN) has
been used extensively for decades in electron spin reso-
nance spectroscopy as an efficient spin-trap. More
recently, it has been shown that PBN can minimize oxi-
dative damage in various biological processes4 and dis-
ease models5 in which free radical species have been
implicated. For example, PBN has been used to protect
against a variety of agents that promote the formation
of brain lesions by disrupting mitochondrial ener-
getics.6,7
In this context, a previous work described the synthesis
of a new compound, glycosylated amphiphilic nitrone
PBN (Compound A).11 We showed that this compound
exhibits the same trapping activity as PBN (using the
RPE technique) and is efficiently recognized by galac-
tose-specific yeast lectins.
Moreover, it appears that the amphiphilic character of
compound A enhances its membrane-crossing ability.12
Whereas PBN shows no protective activity, this com-
pound readily diminishes SOD production and apopto-
sis in cultured skin fibroblasts with an isolated complex
V deficiency of the respiratory chain (NARP mutation).
However, such a structure was relatively complex and
not easy to prepare in large quantities. To simplify the
preparation of these amphiphilic compounds, while
maintaining their spin-trap properties, we chose to graft
a hydrophobic moiety onto the t-butyl group of PBN
and a hydrophilic head onto its aromatic group.
With the ultimate goal of improving its bio-availability,
several studies have focussed on the measurement of the
hydrophilic-lipophilic balance (HLB) of PBN, sub-
sequent to mainly small modifications of either the aro-
matic or N-terminal portion of PBN.8,9 These studies
showed that amphipathic spin-traps exhibit better
membrane-crossing ability than PBN itself and may
provide protection against free radicals in both the
cytosol and in cell membranes.10
*Corresponding authors. Tel.: +33-4-90-144442; fax: +33-4-90-
144449; e-mail: bernard.pucci@univ-avignon.fr (B. Pucci);
yE-mail: ange.polidori@univ-avignon.fr (A. Polidori).
Here we report a new, efficient and convergent synthetic
route to amphiphilic PBN analogues. To provide a
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(02)01079-X

860
G. Durand et al. / Bioorg. Med. Chem. Lett. 13 (2003) 859–862
molecule with significant hydrophilicity, a lactobion-
amide group was grafted onto the aromatic portion of
PBN. The hydrophobic moiety was an alkyl or per-
fluoroalkyl chain linked to the tert-butyl group through
a thioether, a carbamate or an amide bond. In order to
maintain good water solubility of the nitrone analogues,
we chose to graft only short hydrocarbon chains of 6 to
8 carbon atoms. Furthermore, in order to minimize the
detergent properties (and thus the cytotoxicity), essen-
tially due to the hydrocarbon chain of these substrates,
while maintaining their amphiphilic properties, we also
used a perfluorocarbon chain as the hydrophobic moi-
ety. In this new amphiphilic PBN family, the nitronyl
function is thus located at the heart of the surfactant,
between the hydrophobic tail and the polar head. Such
a structure allows us to modulate either the nature and
the length of the lipophilic portion or the hydrophilic
head easily and to determine the impact of the HLB of
these traps on their biological activity.
strong basic conditions used for this reaction (sodium
methylate) could explain the difference in yield observed
between the hydrocarbon (97%) and the per-
fluorocarbon (58%) series.
To prepare the amide derivatives, using the Vaultier et
al. synthetic pathway,14 12 was first converted into the
amine compound 13 by the Staudinger reaction. It is
noteworthy that the substitution kinetics of the tosyl
groups by sodium azide was increased by ultrasonic
activation. Condensation of octanoyl chloride at À20 ꢀC
with 13, in a mixture of CH2Cl2 and triethylamine
(TEA) (3 equiv), led to the hydrocarbon amide deriva-
tive 17 with a yield of 77%. The perfluorocarbon
analogue 18 was obtained in good yield (90%) by reacting
13 in dry CH2Cl2 with the corresponding perfluorocarbon
carboxylic acid in the presence of dicyclohexyl-
carbodiimide(DCC)/hydroxybenzotriazole (HOBt) as a
catalyst.
The retrosynthetic strategy (Scheme 1) is based on two
key steps: (i) the synthesis of substituted long chain N-
tert-butyl hydroxylamines from 2-methyl-2-nitropropa-
nol and (ii) the condensation of these hydroxylamines
on N-4-formyl benzyl lactobionamide 6 to obtain nitro-
nyl function. The synthesis of this synthon 6 has been
previously described.13
Reduction of the nitro group of compounds 14–17
under mild conditions, using Kende and Mendoza’s
method,15 led to the hydroxylamines 7–11 in good yield
(Scheme 3, Table 1).
Nitrones were finally obtained following our well-
established procedure by condensation of the sub-
The five steps leading to 6 were achieved with a 26%
overall yield using the 4-cyanobenzaldehyde pathway.
Hydrophobic chains were grafted onto the 2-methyl-2-
nitropropanol through three different bonds: carba-
mate, amide or thioether bonds. First, the carbamate
derivative 14 was obtained in quantitative yield by
reaction of hexyl isocyanate on 2-methyl-2-nitropropa-
nol, in the presence of diazabicyclooctane (DABCO) as
a catalyst, in refluxing toluene under an atmosphere of
argon (Scheme 2). To obtain the thioether compounds
15–16, a tosylation reaction of the hydroxyl function of
2-methyl-2-nitropropanol was first carried out by treat-
ment with tosyl chloride and pyridine in dry CH2Cl2.
Compound 12 was isolated with a yield of 83% after
recrystallization in AcOEt/hexane. Substitution of the
tosyl group in dry DMF by the hydro or per-
fluorocarbon thiol led to thioether derivatives 15–16. A
partial degradation of the perfluorocarbon thiol in the
Scheme 2. (a) 1.3 Equiv hexyl isocyanate, toluene, DABCO, reflux, 3
h, 100%; (b) 1.2equiv TsCl, pyridine, CH Cl2, 48 h, 83%; (c) 2equiv
2
octanethiol, tBuOK, DMF, 50 ꢀC, 4 h, 97% for 15 or 2equiv 1H, 1H,
2H, 2H perfluorooctanethiol, MeONa, same conditions, 58% for 16;
(d) NaN3, DMF, ultrasounds, 30 min, 85%; (e) PPh3, THF, 2h, then
NaOH 2N, 24 h, 82%; (f) octanoyl chloride, TEA, DMAP, À15 ꢀC, 2
h, 77% for 17 or 2H, 2H, 3H, 3H perfluorononanoic acid, DCC/
HOBt, 36 h, 90% for 18.
Scheme 3. Controlled reduction of nitroalkanes to alkyl hydroxyl-
amines (a) 4 equiv SmI2, THF/MeOH, 15 min.
Scheme 1. Retrosynthetic strategy.

G. Durand et al. / Bioorg. Med. Chem. Lett. 13 (2003) 859–862
Table 1. Reduction yields of nitro compounds Table 2. Physico-chemical data of nitrones 1–5
861
Hydroxylamines
X–CH2–CH2–R
Yields (%)a
Compd
X-CH2-CH2-R
Yield %a
CMCc
gd
7
8
9
10
11
O–CO–NH–C6H13
S–C8H17
S–CH2–CH2–C6F13
NH–CO–C7H15
NH–CO–CH2–CH2–C6F13
73 (100)b
70 (88)b
85 (98)b
100
1
2
3
4
5
O–CO–NH–C6H13
S–C8H17
S–CH2–CH2–C6F13
NH–CO–C7H15
NH–CO–CH2–CH2–C6F13
65
68
56
63
No CMC
0.41
0.035
No CMC
0.228
—
30
20
—
20
50 (70)b
44 (70)b
aAfter purification by flash chromatography.
bNon-reduced nitro compounds were recovered and re-used to com-
plete the reaction.
aCoupling reaction yields, measured after purification on silica gel and
on Sephadex resin.
bCalculated with the residual starting benzaldehyde.
cCMC=critical micellar concentration, in mM.
dg=limit surface tension, in mN/m.
stituted N-tert-butyl hydroxylamine on compound 6 in
dry THF (Scheme 4). In order to avoid the possible
oxidation of hydroxylamine via a free radical process,
the reaction was carried out in the dark under argon
atmosphere and small amounts of hydroxylamine were
added to the mixture every three days. The coupling
reaction was slow and regular addition of molecular
sieves increased the kinetics of the condensation reac-
tion. After 15 days of reaction, acetylated amphiphilic
nitrones were isolated in good yields, after chromato-
graphy on silica gel and filtration on Sephadex LH-20,
with small amounts of starting aldehyde (less than
5%).16 Hydrolysis of the acetyl groups using the Zem-
plen method with catalytic amounts of MeONa in dry
MeOH, led to the production of amphiphilic nitrones 1–
5 in quantitative yields. All compounds were purified by
C18 reversed-phase HPLC (eluant methanol–water) and
were obtained as a pure white powder without a resi-
dual ESR signal.
its CMC (0.41 mM). This toxicity could thus be due to
the detergent property of this amphiphilic compound,
which could lead to the dissolving of cell membranes at
this concentration.
The ability of compound 2 to reduce the level of cas-
pase-3 activity in cultured neuronal cells was then eval-
uated and compared to that of PBN and DMPO, two
commercially available nitrones. The level of caspase-3
activity, an enzyme that plays a crucial role in apopto-
sis, is a good quantitative parameter for this phenom-
enon in neurons.19 (Fig. 1)
The data presented in Figure 2 shows that nitrone 2 at
0.2mM level dramatically inhibits the activity of cas-
pase-3 in cultured neuronal cells following exposure to
0.05 mM hydrogen peroxide. Furthermore, the activity
of this nitrone at 200 mM appears much greater than
those observed with commercially available nitrones
PBN (a-phenyl-N-tert-butyl nitrone) and DMPO (5,5-
dimethylpyrroline-N-oxyde). However, at lower con-
centration (10 and 100 mM), these last spin-traps exhibit
a better efficiency. Further experiments with the other
amphiphilic spin-traps should allow us to explain this
behavior.
The main physico-chemical properties of these amphi-
philic spin-traps is reported in Table 2. The critical
micellar concentration (CMC) was determined by sur-
face tension measurement. As expected, per-
fluorocarbon nitrone 3 exhibited lower CMC than its
hydrocarbon analogue 2. Due to their relatively short
tail nitrones 1 and 4 presented no CMC, and were
probably less detergent than compound 2.
These preliminary results clearly show the ability of
amphiphilic nitrone 2 to decrease the apoptosis in
Preliminary biological assays were carried out with
amphiphilic nitrone 2.17 The cytotoxicity of compound
2 in cultured neuronal cells was specified using the MTT
viability assay (Fig. 1).18 The cytotoxicity of this com-
pound appeared at a concentration of 400 mM, close to
Scheme 4. (a) 2Equiv of hydroxylamines, THF, 55 ꢀC, molecular
sieves 4 A, dark, 15 days; (b) MeONa, MeOH, 1 h, quantitative
yields.
Figure 1. Cytotoxicity of compound 2 in cultured neuronal cells. MTT
staining after 20 h exposure to different concentration of nitrone 2.
Values are expressed as % control value and are mean Æ SD (N=4–6).

862
G. Durand et al. / Bioorg. Med. Chem. Lett. 13 (2003) 859–862
7. Schulz, J. B.; Henshaw, D. R.; Matthews, R. T.; Beal, M. F.
Exp. Neurol. 1995, 132, 279.
8. Janzen, E. G.; Zawalski, R. C. J. Org. Chem. 1978, 43,
1900. Janzen, E. G.; Dudley, R. L.; Shetty, R. V. J. Am. Chem.
Soc. 1979, 101, 243.
9. Janzen, E. G.; West, M. S.; Kotake, Y.; Dubosc, C. M. J.
Biochem. Biophys. Meth. 1996, 32, 183.
10. Thomas, C. E.; Ohlweiler, D. F.; Taylor, V. L.; Schmidt,
C. J. J. Neurochem. 1997, 68, 1173.
11. Ouari, O.; Polidori, A.; Pucci, B.; Tordo, P.; Chalier, F. J.
Org. Chem. 1999, 64, 3554.
12. Partially published in Geromel, V.; Kadhom, N.; Cebalos-
Picot, I.; Ouari, O.; Polidori, A.; Munnich, A.; Rotig, A.;
Rustin, P. Hum. Mol. Genet. 2001, 10, 1221.
13. Ouari, O.; Chalier, F.; Bonaly, R.; Pucci, B.; Tordo, P. J.
Chem. Soc. Perkin Trans. 2 1998, 2299.
14. Vaultier, M.; Knouzi, N.; Carrie, R. Tetrahedron Lett.
1983, 24, 763.
Figure 2. Caspase-3 activities (kit Sigma CASP3C) in cultured neuro-
nal cells after 20 h exposure to variable concentrations of nitrone 2,
PBN and DMPO (after 8 days of cells culture). After incubation with
nitrones, apoptosis was induced by the addition of 0.05 mM hydrogen
peroxide. Values are mean Æ SD of four experiments.
15. Kende, A. S.; Mendoza, J. S. Tetrahedron Lett. 1991, 32,
1699.
1
16. Determined by H NMR analysis.
17. Nitrone 2 analysis: Rf: 0.52(ethyl acetate/methanol/
water 7:2:1). Mp: 115 ꢀC (dec). [a]D=+17.2(05.2c, 1,
CH3OH). 1H NMR (250 MHz, CD3OD) d 8.28 (2H, d,
J=8.25 Hz, H arom.), 7.82 (1H, s, CH=N(O)), 7.42 (2H,
d, J=8.25 Hz, H arom.), 4.65–4.35 (4H, m, CH2-NH, H-10,
H-2), 4.25 (1H, m, H-3), 4.00–3.35 (10H, m, H-4, H-5,
CH2-OH, H-40, H-50, H-30, H-20), 3.01 (2H, s, CIV–CH2–S),
2.43 (2H, t, J=7.3 Hz, CH2–S), 1.61 (6H, s, CH3 tert-butyl),
1.44 (2H, m, CH2–CH2–S), 1.30–1.10 (10H, m, CH2), 0.87
(3H, t, J=6.9 Hz, CH3 chain). 13C NMR (62.86 MHz,
cultured neuronal cells at a threshold of concentration.
Further analysis of all compounds of this new spin traps
family is currently underway in order to determine the
impact of both the nature of the hydrophobic chain
(hydro or perfluorocarbon), and that of the bond
between the hydrophobic chain and the nitrone moiety,
on their antiapoptosis activity.
CD3OD)
(CH=N(O)), 131.1 (CH arom.), 130.6 (CIV arom.), 128.3 (CH
d
175.3 (CO–NH), 143.4 (CIV arom.), 136.0
0
arom.), 105.8 (CH-10), 83.3 (CH-4), 77.2(CH-5 ), 74.8 (CIV),
References and Notes
74.6 (CH-30 or CH-20), 74.1 (CH-2), 73.2 (CH-5), 72.8 (CH-30
or CH-20), 72.5 (CH-3), 70.4 (CH-40), 63.8, 62.7 (CH2-6, CH2-
60), 43.5, 43.0 (CH2–NH, CH2–S), 34.2(CH 2–S), 32.9, 31.0,
30.3, 30.2, 29.7 (CH2), 26.0 (CH3 tert-butyl), 23.7 (CH2), 14.4
(CH3). UV (MeOH, nm): lmax=298.8. FABMS (m-nitro-
benzyl alcohol matrix): m/z=729 [M+K]+; 713 [M+Na]+.
Elemental analysis C32H54N2O12S, 3H2O (744.9): calcd C,
51.60; H, 8.12; N, 3.76; S, 4.30; found C, 51.83; H=7.66;
N=3.70; S=4.25.
1. Richardson, J. S. Ann. N.Y. Acad. Sci. 1993, 695, 73.
2. Ames, B. N.; Shigenaga, M. K.; Hagen, T. M. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 7915.
3. Kontos, H. A. Chem. Biol. Int. 1989, 72, 229.
4. Sack, C. A.; Socci, D. J.; Crandall, B. M.; Arendash, G. W.
Neurosci. Lett. 1996, 205, 181.
5. Yue, T. L.; Gu, J. L.; Lysko, P. G.; Cheng, H. Y.; Barone,
F. C.; Feuerstein, G. Brain Res. 1992, 574, 193.
6. Schulz, J. B.; Henshaw, D. R.; Siwek, D.; Jenkins, B. G.;
Ferrante, R. J.; Cipollini, P. B.; Kowall, N. W.; Rosen, B. R.;
Beal, M. F. J. Neurochem. 1995, 64, 2239.
18. Hansen, M. B.; Nielsen, S. E.; Berg, K. J. Immunol. Meth.
1989, 119, 203.
19. Marks, N.; Berg, M. J.; Guidotti, A.; Saito, M. J. Neu-
rosci. Res. 1998, 52, 334.