Synthesis and antioxidant efficiency of a new amphiphilic spin-trap derived from PBN and lipoic acid

Article abstract of DOI:10.1016/S0960-894X(03)00545-6

The synthesis of a new amphiphilic antioxidant called PBNLP and derived from both α-phenyl-N-tert-butyl nitrone (PBN) and lipoic acid was described. Grafting a lactobionamide moiety onto the aromatic group of the PBN provided the water solubility of this compound. In vitro preliminary biological evaluations of its antioxidant capacity were performed using the KRL biological test based on free radical-induced hemolysis. The PBNLP induces a protection of erythrocytes against exogenous free radicals higher than that measured with lipoic acid or PBN alone or with lipoic acid or PBN derivatives in admixtures.

Full text of DOI:10.1016/S0960-894X(03)00545-6

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2673–2676
Synthesis and Antioxidant Efficiency of a New Amphiphilic
Spin-Trap Derived from PBN and Lipoic Acid
G. Durand,a A. Polidori,a,* J. P. Salles,b M. Prost,c P. Durandc and B. 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
´
Centre Europeen de Recherches et d’Analyses, 3 rue des mardors, 21560 Couternon, France
c
Received 31 March 2003;revised 26 May 2003;accepted 28 May 2003
Abstract—The synthesis of a new amphiphilic antioxidant called PBNLP and derived from both a-phenyl-N-tert-butyl nitrone
(PBN) and lipoic acid was described. Grafting a lactobionamide moiety onto the aromatic group of the PBN provided the water
solubility of this compound. In vitro preliminary biological evaluations of its antioxidant capacity were performed using the KRL
biological test based on free radical-induced hemolysis. The PBNLP induces a protection of erythrocytes against exogenous free
radicals higher than that measured with lipoic acid or PBN alone or with lipoic acid or PBN derivatives in admixtures.
# 2003 Elsevier Ltd. All rights reserved.
The generation of Reactive Oxygen Species (ROS) is
associated with numerous and various pathological
states such Alzheimer’s disease,1 chronic neurodegen-
erative disorders,2 ischemia-reperfusion or stroke.3
Floyd et al.4,5 were the first to show that nitrone-based
free radical spin traps, especially a-phenyl-N-tert-butyl
nitrone (PBN), exhibit neuroprotective activity. In a
previous paper,6 we have described the synthesis of a
novel series of amphiphilic glycosylated nitrones derived
from PBN. Preliminary investigations on their anti-
apoptotic effect in cultured neuronal cells showed that
they drastically inhibit caspase-3 activity following
H2O2-induced apoptosis. Furthermore, we have under-
lined that the hydrophilic lipophilic balance and the
amphiphilic properties of these molecules plays a key
role on their protective activity. In the present paper, we
reported the synthesis of an amphiphilic hybrid of PBN
and lipoic acid. The latter and its reduced form, dihy-
drolipoic acid, which are natural metabolic substances,
appeared like ideal antioxidants for the treatment of
brain and neurodegenerative disorders in which free
radical processes are involved.7,8 Synergistic interactions
between lipoate and other natural antioxidants like
vitamin C and E that resulted from lipoate recycling,
were previously described.9ꢀ14 In order to improve these
synergetic effects, lipoate was recently connected to
Trolox1, a water-soluble analogue of vitamin E, which
is a strong inhibitor of lipid peroxidation.14 In a similar
way, Harnett et al.15 have synthesized a hybrid of lipo-
ate and of a nitric oxide synthase inhibitor, capable of
protecting neuronal cells against glutamate toxicity.
Thereafter, we decided to prepare a new amphiphilic
hybrid of PBN and lipoate, called PBNLP. The nitronyl
function is inserted into the heart of the surfactant
between the hydrophobic tail and the polar head
(Scheme 1). Whereas the hydrophobic part is made up
of lipoate group, a lactobionamide moiety linked to the
para position of PBN aromatic group forms the polar
head and gives the water solubility to the whole mole-
cule. To evaluate the biological effect of this new
amphiphilic molecule PBNLP and to specify the possi-
ble beneficial effect of the chemical association of water-
soluble derivatives of PBN, lactobioamide PBN
(LAMPBN) and of lipoic acid, lipoylglytris (LGT)
(Scheme 1, see16,17 for synthesis processes), we have
assessed the antioxidant efficiency of PBNLP as compared
to specific antioxidant capacities of PBN, lipoic acid,
LAMPBN and LGT and to the antioxidant effectiveness
of LAMPBN and LGT or lipoate in admixtures.
*Corresponding author. Tel.:+33-490-144-442;fax: +33-490-
144-449;e-mail: ange.polidori@univ-avignon.fr;bernard.pucci@
univ-avignon.fr
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00545-6

2674
G. Durand et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2673–2676
Scheme 1. Structure of PBN and lipoate derivatives.
In the chosen PBNLP synthetic route (Scheme 2) the
grafting of the lipoate group has to be carried out after
the nitronyl function preparation. Indeed, the nitronyl
function is currently obtained by binding the benzalde-
hyde group to a hydroxylamine group, which can be
easily prepared by reduction of a nitro function with the
Kagan reagent, SmI2. Nevertheless, this type of reduc-
tion is prevented by the presence of the lipoate group,
whose disulfide group could oxidize the Kagan reagent
and hence provide a dihydrolipoate function. Taking
into account such a reaction, we chose to synthesize first
the nitronyl function and then to graft the lipoate moi-
ety. Therefore, we used the 2-methyl-2-nitropropanol as
starting material. According to the fact that the final
opening of acetylated lactobionamide using the Zem-
plen method could induce an ester group hydrolysis, we
decided to link the lipoate moiety to the t-butyl nitrone
by a carbamate bond. However, Harnett et al.15 have
already observed the formation of an oxadiazoline-5-
one through a cycloaddition reaction between an isocy-
anate group and the nitronyl function. Coupling trials
with decyl isocyanate and t-butyl nitrone have shown
that such a cycloaddition reaction was mainly achieved.
To avoid these problems, we have inserted a g-amino-
butyric acid (GABA) as spacer arm between the nitrone
function and the lipoate group. First, the Boc-GABA
was bound to the hydroxyl group of the 2-nitro-2-
methyl propanol through an urethane bond. The reac-
tion was performed in toluene at 60 ꢁC in the presence
of diphenyl phosphoryl azide (DPPA) and diaminobi-
cyclooctane (DABCO). The obtained compound 1 was
reduced in a very good yield (92%) with Kagan reagent.
The hydroxylamine 2 was then grafted onto the aldehyde
3 in a THF/acetic acid mixture (9/1). This very slow
reaction was achieved in 8 days. The nitrone 4 was iso-
lated in 68% yield after purification on silica gel column.
Scheme 2. Synthesis of PBNLP.
densed on the resulting amino group in 70% yield. The
synthesized compound was then purified by silica gel
column chromatography and size exclusion chromato-
graphy on Sephadex LH20 column. Hydrolysis of acetyl
groups with catalytic amounts of MeONa in dry MeOH
according to the Zemplen method led to glycosylated
nitrones in quantitative yield. PBNLP compound was
carefully purified by C18 reverse phase HPLC (mobile
phase;methanol–water) and was obtained pure as a
white powder. It was next fully characterized by 1H, 13
C
NMR, DEPT and HMQC NMR sequences. Mass
spectrometry (positive FAB) allowed the observation of
the characteristic fragmentations of PBNLP hydro-
phobic N terminal part.18
The overall antioxidant effectiveness of PBN and lipo-
ate derivatives was evaluated in a dynamic way using
the KRL biological test based on free radical-induced
hemolysis (Laboratoires Spiral, France).19,20 Diluted
blood samples without and in the presence of different
doses of compounds were submitted to organic free
radicals produced in the aqueous phase at 37 ꢁC under
air atmosphere from the thermal decomposition of a 50
mM solution of 2,20-azobis(2-amidinopropane) dihy-
drochloride (AAPH). Hemolysis was recorded by turbi-
dimetry using a 96-well microplate reader. Results were
expressed as the time that is required to reach 50% of
After hydrolysis of the t-butyloxy group of compound 4
with a trifluoroacetic acid/CH2Cl2 mixture at room
temperature, succinimidyl ester of lipoic acid was con-

G. Durand et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2673–2676
2675
more active than its individual constituents, the water-
soluble derivatives of PBN and lipoate that unexpectedly
suggested a negative interaction between LAMPBN and
LGT. Moreover, at low concentrations below 50 mM
this new compound presented an antiradical capacity
higher than Trolox1 although at more than 200 mM, it
presented a lower activity than Trolox1, which totally
prevented hemolysis probably by scavenging radicals in
the aqueous phase as soon they are produced (Fig. 1).
In good agreement with our previous results on the
protective effect of PBN derived amphiphilic com-
pounds on H2O2-mediated cell apoptosis,6 the hydro-
philic lipophilic balance and thereby the amphiphilic
character of PBNLP, which could act in the extra- as in
the intra-cellular medium, may play a determinant role
on the higher improvement of cellular resistance against
radicals as compared to unmodified PBN and lipoic
acid alone. To verify this assumption and the possible
negative interaction between LAMPBN and LGT that
would have limited PBNLP beneficial effects, we have
Figure 1. Dose–response studies of the antiradical efficiency of the
new amphiphilic compound PBNLP ( ) as compared to antiradical
efficiencies of lipoic acid (&), LGT (~), PBN (*), LAMPBN (~)
and Trolox1 (&). Results are expressed as percentages of compound-
induced increase in the control half-hemolysis time (HT50). They are
meansÆstandard deviations (SD) of three separated experiments.
ꢂ
compared
PBNLP
antiradical
efficiency
with
LAMPBN/LGT mixtures one. The experimental activ-
ities of PBNLP and mixtures at 1.33/0.67, 1/1 and 0.67/
1.33 molar ratios were put together with their theore-
tical activities, which were calculated by summing
experimental HT50 increases obtained with molar
equivalents of pure LAMPBN and LGT (Table 1). As
regards the mixtures, no additional antioxidant effect
was observed whatever the concentration used. By con-
trast, our results showed a strong antagonistic effect
between the two PBNLP constituents that was also
observed between LAMPBN and lipoate (results not
shown). Thus, LAMPBN would be unable to recycle
lipoate in contrast to that was previously described with
natural antioxidants.9ꢀ14 According to the enhanced
antioxidant efficiencies of mixtures that were noted
when increasing LGT (or lipoate) molar proportion at
the expense of LAMPBN molar proportion, this antag-
onism might result from the counteracted effect of
LAMPBN on lipoic acid.
maximal hemolysis (half-hemolysis time, HT50 in min),
which refers to the whole blood resistance to free radical
attack. The overall antiradical efficiency was defined as
the percentage of the compound-mediated HT50 increase
as compared to control. Preliminary cytotoxic assays
without AAPH addition showed that in the tested con-
centration range from 0 to 500 mM, lipoic acid, LGT,
PBN, LAMPBN and PBNLP exhibited no cytolytic
activities towards blood samples.
At concentrations below 100 mM, lipoic acid showed
antioxidant efficiency higher than PBNLP that could
result from the hydrophobic character of lipoic acid
favouring its insertion and/or its diffusion across cell
membranes and hence temporarily improving the cell
resistance to free radical attacks (Fig. 1).
However, in the whole concentration range the new
amphiphilic compound PBNLP exhibited a higher glo-
bal protective activity towards free radical aggression of
blood samples as compared to unmodified PBN and
lipoate alone. Nevertheless, PBNLP was only slightly
Furthermore, we found that despite this antagonistic
action, the PBNLP-mediated enhancement of cellular
resistance to radical aggression is significantly higher
Table 1. Experimental and theoretical antiradical capacity of the new amphiphilic molecule PBNLP as compared to LAMPBN and LGT mixtures
Compd
Conc.
50 mM
100 mM
200 mM
500 mM
Mol. ratio
% HT50
% HT50
% HT50
% HT50
% HT50
% HT50
% HT50
% HT50
Exp.
49.0
41.5
44.6
56.6
Theo.
74.8
70.1
74.8
76.8
Exp.
82.9
47.4
51.6
60.8
Theo.
139.1
127.8
139.1
132.1
Exp.
135.2
49.8
53.7
62.0
Theo.
212.7
201.1
212.7
188.4
Exp.
167.2
86.3
86.5
88.6
Theo.
272.0
281.1
272.0
247.5
PBNLP
1
LAMPBN/LGT
LAMPBN/LGT
LAMPBN/LGT
1.33/0.67
1/1
0.67/1.33
Note. Experimental (Exp.) and theoretical (Theo.) results are expressed as average percentages (three separated experiments) of compound-induced
increase in the control half-hemolysis time (% HT50). Experimental antiradical efficiencies are raw percentages obtained or percentages extrapolated
from dose–response curves. Theoretical antiradical capacity of one PBNLP mole is the sum of experimental capacities of one LAMPBN mole and of
one LGT mole. Theoretical antiradical capacity of mixtures that provides 1.33, 1 or 0.67 LAMPBN moles and 0.67, 1 or 1.33 LGT moles is the sum
of experimental capacities of 1.33, 1 or 0.67 LAMPBN moles and of 0.67, 1 or 1.33 LGT moles, respectively. According to quantification limits of
the KRL hemolysis test, percentage differences between experimental and theoretical HT50 increases less than 5% are considered as non-significantly
different at P <0.05.

2676
G. Durand et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2673–2676
than the antioxidant protection induced by mixtures of
LAMPBN and LGT at concentrations above 50 mM
(Table 1) and by mixtures of LAMPBN and lipoate in
the whole concentration range from 0 to 500 mM
(Results not shown). Nevertheless, we observed that the
overall experimental antioxidant efficiency of one mole
of PBNLP kept lower than the theoretical antioxidant
capacities of LAMPBN and LGT mixtures indepen-
dently of the molar ratio (Table 1). Thus, the combi-
nation of water-soluble derivatives of PBN and lipoic
acid in an amphiphilic molecule PBNLP did not
induced synergistic interactions between these two
compounds, but it markedly diminished their antag-
onistic actions highlighted by the free radical-mediated
hemolysis test.
References and Notes
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. Floyd, R. A. FASEB J. 1990, 4, 2587.
5. Floyd, R. A.;Hensley, K.;Forster, M. J.;Kelleher-Anders-
son, J. A.;Wood, P. L. Mech. Aging Dev. 2002, 123, 1021.
6. Durand, G.;Polidori, A.;Salles, J.-P.;Pucci, B.
Med. Chem. Lett. 2003, 13, 859.
7. Packer, L.;Witt, E. H.;Tritschler, H. J. Free Rad. Biol.
Med. 1995, 19, 227.
8. Packer, L.;Tritschler, H. J.;Wessel, K. Free Rad. Biol.
Med. 1997, 22, 359.
9. Rosenberg, H. R.;Culik, R. Arch. Biochem. Biophys. 1959,
80, 86.
Bioorg.
10. Haramacki, N.;Assadnazari, H.;Zimmer, G.;Schepkin,
V.;Packer, L. Biochem. Mol. Biol. Int. 1995, 37, 591.
11. Thurich, T.;Bereiter-Hahn, J.;Schneider, M.;Zimmer, G.
Arzneim.-Forsch./Drug Res. 1998, 48, 13.
12. Coombes, J. S.;Powers, S. K.;Demirel, J.;Jessup, H. K.;
Vincent, K. L.;Hamilton, K. L.;Naito, H.;Shanely, R. A.;
Sen, C. K.;Packer, L.;Ji, L. L. Acta Physiol. Scand. 2000, 169,
261.
13. Lu, C.;Liu, Y. Arch. Biochem. Biophys. 2002, 406, 78.
14. Koufaki, M.;Calogeropoulou, T.;Detsi, A.;Roditis, A.;
Kourounakis, A. P.;Papazafiri, P.;Tsiakitzis, K.;Gaitanaki,
C.;Beis, I.;Kourounakis, P. N. J. Med. Chem. 2001, 44, 4300.
15. Harnett, J. J.;Auguet, M.;Viossat, I.;Dolo, C.;Bigg, D.;
Chabrier, P.-E. Bioorg. Med. Chem. Lett. 2002, 12, 1439.
16. Ouari, O.;Chalier, F.;Bonaly, R.;Tordo, P.;Pucci, B.
Chem. Soc. Perkin Trans. 2 1998, 2299.
17. Neuburger, M.;Polidori, A.;Faure, M.;Pietre, E.;Bour-
It is noteworthy that the measured antiradical efficiency
using this hemolysis test reflected the overall antioxidant
behavior of the tested compounds including exogenous
radical scavenger activity but also metal ion chelator
properties, protective actions on the intrinsic anti-
oxidant defense system and regenerative effects on anti-
oxidant molecules involved in the regulation of the
intra- and extra-cellular redox status. Thus, numerous
chemical mechanisms at cell cytoplasmic, cell mem-
branes and/or extra-cellular levels may be involved in
the antagonism that occurred between LAMPBN and
LGT or lipoate. Therefore, further studies are needed to
thoroughly investigate specific chemical mechanisms
underlying antagonistic antioxidant effects between the
water-soluble derivatives of PBN and lipoate that are
profoundly reduced by their chemical association in
PBNLP.
J.
guignon, J.;Douce, R.;Pucci, B. Eur. J. Biochem. 2000, 267,
2882.
18. PBNLP analysis: mp 105 ꢁC (dec). 1H NMR (250 MHz,
DMSO-d6): d 8.29 (2H, d, J=8.2 Hz.), 8.26 (1H, m), 7.77 (1H,
s), 7.73 (1H, m), 7.32 (2H, d, J=8.2 Hz), 7.15 (1H, m), 4.50–
4.00 (7H, m), 3.85–3.25 (11H, m), 3.25–2.85 (6H, m), 2.41 (1H,
m), 2.04 (2H, t, J=6.9 Hz), 1.86 (1H, m), 1.75–1.20 (14H, m).
13C NMR (62.86 MHz, DMSO-d6): d 173.1, 172.4 (CO-NH),
156.3 (O-CO-NH), 142.0 (CIV arom.), 130.7 (CH¼N(O)),
130.2 (CIV arom.), 128.9, 127.2 (CH arom.), 105.1 (CH-10),
83.4 (CH-Lac.), 76.2 (CH-Lac.), 73.7 (CH-Lac.), 72.6 (CH-
Lac.), 72.5 (CIV), 71.9 (CH-Lac.), 71.6 (CH-Lac.), 71.1 (CH-
Lac.), 68.7 (CH-Lac.), 68.1 (CH2-O-CO-NH), 62.8, 61.1 (CH2-
OH), 56.6 (CH-S), 42.2 (CH2-NH-Lac.), 40.2 (CHS-C (CH2-S
and CH2-NH), 36.5 (CH2-NH), 35.7 (CH2-CO), 34.6 (CH2-
CH-S), 29.9, 28.8, 25.5 (CH2 lipoate and CH2 GABA), 23.6
(CH3). UV (MeOH, nm): lmax=297.4. FABMS (m-nitrobenzyl
alcohol matrix): m/z=851[M+H]+;446 [C 19H28NO11]+;289
[C12H21N2O2S2]+. Elemental analysis C36H58N4O15S2, 1H2O
(869): calcd C, 49.76;H, 6.96;N, 6.45;S, 7.38;Found C,
49.39;H, 6.97;N, 6.31;S, 7.06.
All together, these findings pointed out that the chemi-
cal association of water-soluble derivatives of PBN and
lipoic acid has a strong beneficial effect on blood resis-
tance to free radical attack as compared to these deri-
vatives in admixtures and to PBN and lipoic acid alone.
Moreover, it suggested that the PBNLP beneficial effect
is partly governed by the amphiphilic character of the
molecule that probably increased its bioavailability and
by the reduction of negative chemical interactions
between the two PBNLP constituents LAMPBN and
LGT. These preliminary encouraging results on the in
vitro protection of blood against oxidative stress
emphasized the benefit of the synthesis of new amphi-
philic antioxidant molecules such as PBNLP to improve
their biological effects. In vivo studies are now needed
to further confirm the beneficial effect of PBNLP and to
investigate the therapeutic benefit of modified anti-
oxidant molecules as compared to unmodified mole-
cules. The use of the free radical-mediated hemolysis
test could also be helpful for this purpose.
19. Prost, M. U.S. Patent 5, 135, 850, 1992.
20. Kumagai, J.;Kawaura, T.;Miyazaki, T.;Prost, M.;Prost,
E.;Watanabe, M.;Quetin-Leclerc, J.
2002, 66, 17.
Radiat. Phys. Chem.