Novel antioxidant agents deriving from molecular combinations of vitamins C and E analogues: 3,4-Dihydroxy-5(R)-[2(R,S)-(6-hydroxy-2,5,7,8-tetramethyl-chroman-2(R,S)-yl-methyl)-[1,3]dioxolan-4(S)-yl]-5H-furan-2-one and 3-O-octadecyl derivatives

Article abstract of DOI:10.1016/S0968-0896(00)00205-4

Molecular combinations of two antioxidants (i.e., ascorbic acid and the pharmacophore of α-tocopherol), namely the 2,3-dihydroxy-2,3-enono-1,4-lactone and the chromane residues, have been designed and tested for their radical scavenging activities. When evaluated for their capability to inhibit malondialdehyde (MDA) production in rat liver microsomal membranes, the 3,4-dihydroxy-5R-2(R,S)-(6-hydroxy-2,5,7,8-tetramethylchroman-2(R,S)yl-methyl)-1,3]dioxolan-4S-yl]-5H-furan-2-one (11a-d), exhibited an interesting activity. In particular the 5R,2R,2R,4S and 5R,2R,2S,4S isomers (11c,d) displayed a potent antioxidant effect compared to the respective synthetic α-tocopherol analogue (5) and natural α-tocopherol or ascorbic acid, used alone or in combination. Moreover, the mixture of stereoisomers 11a-d also proved to be effective in preventing damage induced by reperfusion on isolated rabbit heart, in particular at the higher concentration of 300 μM. In view of these results our study represents a new approach to potential therapeutic agents for applications in pathological events in which a free radical damage is involved. Design, synthesis and preliminary biological activity are discussed. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(00)00205-4

Bioorganic & Medicinal Chemistry 8 (2000) 2791±2801
Novel Antioxidant Agents Deriving From Molecular
Combinations of Vitamins C and E Analogues:
3,4-Dihydroxy-5(R)-[2(R,S)-(6-hydroxy-2,5,7,8-tetramethyl-
chroman-2(R,S)-yl-methyl)-[1,3]dioxolan-4(S)-yl]-5H-furan-
2-one and 3-O-Octadecyl Derivatives
Stefano Manfredini, a,* Silvia Vertuani, a Barbara Manfredi, b Giuseppe Rossoni, b
Gabriella Calviello c and Paola Palozza c
a
Á
Dipartimento di Scienze Farmaceutiche, Universita di Ferrara, Via Fossato di Mortara 17-19; I-44100 Ferrara, Italy
Á
Á
Istituto di Patologia Generale, Universita Cattolica di Roma, Largo F. Vito, 1; I-00168 Roma, Italy
b
Dipartimento di Farmacologia, Chemoterapia e Tossicologia Medica, Universita di Milano, Via Vanvitelli 32; I-20129 Milano, Italy
c
Received 27 January 2000; accepted 31 July 2000
AbstractÐMolecular combinations of two antioxidants (i.e., ascorbic acid and the pharmacophore of a-tocopherol), namely the
2,3-dihydroxy-2,3-enono-1,4-lactone and the chromane residues, have been designed and tested for their radical scavenging
activities. When evaluated for their capability to inhibit malondialdehyde (MDA) production in rat liver microsomal membranes,
the 3,4-dihydroxy-5R-2(R,S)-(6-hydroxy-2,5,7,8-tetramethylchroman-2(R,S)yl-methyl)-1,3]dioxolan-4S-yl]-5H-furan-2-one (11a±d),
exhibited an interesting activity. In particular the 5R,2R,2R,4S and 5R,2R,2S,4S isomers (11c,d) displayed a potent antioxidant
e€ect compared to the respective synthetic a-tocopherol analogue (5) and natural a-tocopherol or ascorbic acid, used alone or in
combination. Moreover, the mixture of stereoisomers 11a±d also proved to be e€ective in preventing damage induced by reperfu-
sion on isolated rabbit heart, in particular at the higher concentration of 300 mM. In view of these results our study represents a new
approach to potential therapeutic agents for applications in pathological events in which a free radical damage is involved. Design,
synthesis and preliminary biological activity are discussed. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
The importance of vitamins E and C in disease preven-
tion has become widely recognized and is supported by
several clinical and epidemiological studies.4 a-toco-
pherol (vitamin E) provides the most important line of
defence against lipid peroxidation of cell membranes,
and some of its e€ects are mediated by interaction with
speci®c and saturable binding sites on cell membranes.
However, this activity is strongly dependent on the
regeneration of vitamin E by ascorbic acid (vitamin C)
and glutathione present in the cytosol. Electron spin
resonance studies (ESR) have demonstrated the inter-
action between vitamin C and E during the free radical
mediated oxidative stress, thus indicating the role of
ascorbic acid as the terminal small-molecule antioxidant
in biological systems.5 Therefore, ascorbic acid itself is
not able to prevent lipid peroxidation.6 It is well known
that dietary supplementations of these vitamins are
scarcely active, due to their low stability and bioavail-
ability: oxygen-derived free radicals have short half-life
and are involved in fast rate reactions (i.e., free radical
chain reactions). Thus, the principal requisite that a
radical scavenger must possess is to be available, in the
The area of free radical biology and medicine is devel-
oping faster since the discovery, in recent years, of the
involvement of free radicals in oxidative tissue injury
and a growing number of diseases. The production of
free radicals and various reactive oxygen species (ROS),
derives from oxidative injuries related to pollution,
radiation, food constituents and is also a function of the
body's normal defence system and metabolic reactions.1
The involvement of free radicals in the pathology of
human diseases (i.e., intestinal diseases, atherosclerosis,
reperfusion injuries, cardiac diseases, neurodegeneration,
respiratory disorders, in¯ammation, diabetes, cancer and
aging) has been recognized by a number of authors.2
Potential therapeutic interventions might include natural
antioxidants or synthetic pharmacological agents with
antioxidant activity.2,3
*Corresponding author. Tel.: +39-532-291-292; fax: +39-532-291-
296; e-mail: mv9@unife.it
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00205-4

2792
S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
Table 1. Antioxidant eciency of the studied compounds 3, 5, 11, 12
and l-ascorbic acid in rat liver microsomal membranes initiated with
AAPHa
correct delivery of the active moieties to the oxidation site.
Taking this into account and continuing our e€orts
toward the study of radical scavenging agents,10,11 we
have now designed new analogues of these vitamins in
the attempt to overcome the above cited drawbacks. In
particular, we have studied molecular combinations of
the pharmacophores of the two vitamins (1), namely the
chromane and the 2,3-dihydroxy-2,3-enono-1,4-lactone
rings. The title compounds were synthesized and investi-
gated for their antioxidant activity by evaluation of their
capability to inhibit malondialdehyde (MDA) pro-
duction in rat liver microsomal membranes (Table 1).
Moreover, in view of the high relevance of myocardial
infarct among cause of death, the compounds (11a±d),
emerging as promising from the primary antioxidant
test were further evaluated for their capability to reduce
ischemia-reperfusion damage.
Compound
Concentration
(mM)
Malondialdehyde
inhibition (%)
5
10
25
100
15
20
22
3
10
25
100
8
20
24
11a,b (cis)
11c,d (trans)
10
25
100
20
48
93
2.5
10
54
77
25
100
80
100
Our study, initiated by a careful examination of the lit-
erature, led us to select, as suitable candidates for anti-
oxidant and conjugable moieties, the following molecules:
(i) the(R,S)-6-hydroxy-(2,5,7,8-tetramethylchroman-2-yl)
acetaldehyde (5),12,13 (ii) the ascorbic acid itself (2) and
(iii) the 2-O-octadecyl-ascorbic acid (3), which has
shown, together with important antioxidant properties,
an increased stability by comparison to ascorbic
acid.6,14
12
10
25
100
8
18
27
l-ascorbic acid
10
25
100
47
91
93
aMembranes were incubated with 25 mM AAPH for 15 min. Values
are the means of three di€erent experiments. Each mean has a stan-
dard error less than 10%.
appropriate concentration, at the place where the
radical-promoted degradation of the cellular compo-
nents initiate (i.e., membrane or nucleus and cyto-
plasm). Up to date, attempts have been made to develop
analogues of these vitamins with improved pharmaco-
kinetics (i.e., esters),2,7 and a few conjugates between
vitamin C and E have been reported.8,9 However, the
strategy generally employed for the design of these
compounds makes use of the more accessible acid
hydroxy residues for the conjugation, namely at the 2
and 3 position. Because these functional groups are
responsible for the scavenging activity, these com-
pounds act as merely pro-drugs that require activation
in the body of the pro-vitamin and do not warrant a
Chemistry
Synthesis
After having evaluated several functionalities, the retro-
synthetic analysis suggested us to select an aldehyde on
the chromanyl fragment, in order to employ acetalization
of the 5- and 6-hydroxy groups on ascorbic acid, for the
conjugation of the two fragments. This strategy allowed
us to use, for the conjugation process, functional groups
of the molecules not involved in the radical scavenging
process.
Scheme 1. (i) DIBAL-H, hexane, 70 ꢀC; (ii) benzylbromide, K2CO3, DMF, 25 ꢀC; (iii) TsOH, toluene, re¯ux; (iv) H2, Pd/C, EtOAc, 25 ꢀC.

S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
2793
isomers respectively, with regards to the relative stereo-
chemistry of the protons at the C-2 and C-4 of the
dioxolane ring.
Protective benzyl groups, at positions 3 and/or 2 on the
ascorbic acid moiety and at position 6 on the chromanyl
residue, were concomitantly removed by hydrogenation
in the presence of 5% Pd/C to give compounds 11a±d
and 12a±d (Fig. 1).
Structure assignments were also conducted on the base
of our previous ®ndings16 and of quantum-mechanics
calculations. Brie¯y, the structure of a,b and c,d isomers of
compound 11 were optimized and subjected to a systematic
conformational search. The structures having a maxi-
mum energy di€erence of 20 kcal/mol were further opti-
mized until convergence. The geometry was optimized
both with a molecular mechanics force ®eld (MAXI-
MIN2) and with a semi-empirical molecular orbital
method (AM1). The conformations having closer dis-
tances between C5 proton at the furanone ring and C3
proton and C2 methyl at the chromanyl residue were
studied with ab initio Hartree±Fock methods. Molecular
modeling con®rmed the up®eld shift of about 0.06 ppm
observed in the 11c,d (trans) isomers for the C-5 proton
at the furanone ring, that might be explained on the base
of a preferred conformation in which the C-5 proton was
in the shielding region of the chromanyl ring. In parti-
cular, the 11c,d isomers show a closer contact of the C5
proton at the furanone ring with the methyl at the C2 of the
cromanyl residue, about 4.19 A, versus 6.43 A, as compared
to the corresponding 11a,b (cis) isomers. Although mol-
ecular modeling calculations have permitted the struc-
ture assignments for the a,b and c,d couple of isomers,
because no crystalline structure suitable for X-ray spectro-
scopy could be obtained, it was impossible to assign the
absolute con®guration of the single diastereoisomers.
Title compounds were synthesized (Scheme 1) in two
simple steps from compound 6, prepared by reduction
of (R,S)-6-ethyl-(2,5,7,8-tetramethylchroman-2-yl)acetate
(4)12 with diisobutylaluminum hydride (DIBAH) to give
the corresponding aldeyde 5, which was protected at the
hydroxy function in position 6 by benzylation to a€ord
6. Compound 6 was next acetalized, in the presence of
p-toluenesulfonic acid (TsOH), as catalyst, with 2,3-di-
O-benzyl-ascorbic acid (7)14 or with 2-O-octadecyl-3-O-
benzyl-ascorbic acid (8).15 Acetalisation proceeded
smoothly to give in a 1:1:1:1 mixture of the four expected
diastereoisomers, namely 3,4-dibenzyloxy-5R-2(R,S)-(6-
hydroxy-2,5,7,8-tetramethylchroman-2(R,S)-yl-methyl)-
1,3]dioxolan-4S-yl]-5H-furan-2-one (9a±d) and 4-O-benzyl-
oxy-5R-2(R,S)-(6-benzyloxy-2,5,7,8-tetramethylchroman-
2(R,S)-yl-methyl)-1,3]dioxolan-4S-yl]-3-octadecyloxy-5H-
furan-2-one (10a±d). Compounds 10a±d could be isolated
only as inseparable mixtures of the two couple of dia-
stereoisomers: a,b and c,d. The composition of the mixture
was estimated, by 1H NMR analysis, as the relative
intensity of the H-C4 proton, on the furanone ring, in
the a,b/c,d isomers. These data were also corroborated
by NOESY and NOE-di€erence experiments, indicating
a shielding e€ect, of the methyl at C2 of the chromanyl
moiety on the H-C5 at the furanone ring, in the case of
the c,d isomers. In particular the 5R,2S,2S,4S and
5R,2S,2R,4S (a,b) and the 5R,2R,2R,4S and 5R,2R,
2S,4S (c,d) isomers may be considered as cis and trans
Biology
Prevention of lipid peroxidation
The above prepared compounds were evaluated for
their ability to inhibit peroxyl radical dependent lipo-
Figure 1. Absolute con®guration of stereogenic centers of compounds 9±12.

2794
S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
Figure 2. Inhibition of AAPH-induced malondialdehyde production
by di€erent concentrations of a-tocopherol in rat liver microsomal
membranes. Values are the meansÆSEM of 10 di€erent experiments.
Each mean has a standard error less than 10%.
Figure 4. Inhibition of AAPH-induced malondialdehyde production
by di€erent concentrations of compound 3, in rat liver microsomal
membranes. Values are the meansÆSEM of 10 di€erent experiments.
Each mean has a standard error less than 10%.
Figure 3. Inhibition of AAPH-induced malondialdehyde production
by di€erent concentrations of l-ascorbic acid in rat liver microsomal
membranes. Values are the meansÆSEM of 10 di€erent experiments.
Each mean has a standard error less than 10%.
Figure 5. Inhibition of AAPH-induced malondialdehyde production
by di€erent concentrations of compound 5, in rat liver microsomal
membranes. Values are the meansÆSEM of 10 di€erent experiments.
Each mean has a standard error less than 10%.
peroxidation in rat liver microsomal membranes.
Microsomes were prepared from Winstar rats as pre-
viously reported by some of us17 and microsomal proteins
were determined by the Bio-Rad method. Lipid perox-
idation was initiated by 2,20-azobis(2-amidinopropane)-
dihydrochloride (AAPH), which produces peroxyl
radicals by thermal decomposition. Reactions were carried
out at 37 ꢀC with 25 mM AAPH for 15 min and lipid
peroxidation was measured by the appearance of malon-
dialdehyde (MDA) formation.18 All compounds were
used at the concentration of 5 mM. Similar results were
also obtained using antioxidant concentrations ranging
from 2.5 to 100 mM (data not shown).
these conditions, we have compared the antioxidant
ability of the natural antioxidants, a-tocopherol (Fig. 2)
and l-ascorbic acid (Fig. 3) to that of the above described
synthetic compounds, including 3 (Fig. 4), 5 (Fig. 5),
11a±d (Fig. 6) and 12a±d (Fig. 7). It is evident that all
synthetic compounds, were able to act as antioxidants,
inhibiting AAPH-induced MDA production. In par-
ticular, compounds 11a,b and 11c,d, which are the
molecular combination between the a-tocopherol ana-
logue 5 and ascorbic acid, exhibited greater antioxidant
eciency (IC50 12 and 7 mM) than the natural anti-
oxidants, a-tocopherol (IC50 37 mM) and ascorbic acid
(IC50 of 14 mM). Moreover, this compound was a more
powerful antioxidant than the synthetic analogue of
ascorbic acid, 3, and comparable to the other synthetic
analogue of a-tocopherol, 5.
In the absence of antioxidants, the formation of MDA
in microsomal membranes treated with 25 mM AAPH
for 15 min, was 5.40Æ0.5 nmols/mg protein. Under

S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
2795
The con®guration at the dioxolane ring and chromane
ring also in¯uenced the antioxidants properties of this
compound, as shown by the di€erent IC50 values of
compounds 11a,b and 11c,d. Finally, 11c,d exhibited a
greater antioxidant eciency than the respective 12,
con®rming that the substition at position 2 of the
ascorbyl residue is important for the antioxidant prop-
erties of these compounds, indeed speci®c enzymes, not
present in our preliminary assay, may be responsible of
the reported activity of 3.14
The mechanism at the base of the improved antioxidant
eciency of compounds 11c,d, as compared to other
natural and synthetic antioxidants here tested, is cur-
rently under investigation. We may advance the
hypothesis that the concomitant presence of the ascor-
bic acid moiety and a-tocopheryl residue may facilitate
a radical trapping activity in both hydrophilic and lipo-
philic compartments by improving the antioxidant
potential through concomitant scavenging of lipoper-
oxyl radicals (preferred by tocopherol) and reactive
oxygen species (ROS, preferred by ascorbic acid).
Moreover, it may induce increased stability of the com-
pound by limiting its oxidation:19 the ascorbic acid moiety
on a-tocopheryl residue may reduce the production of
tocopheroxyl radicals, which are usually formed during
the radical scavenging activity and, consequently, may
inhibit oxidation induced by these radicals.
Figure 6. Inhibition of AAPH-induced malondialdehyde production
by di€erent concentrations of compounds, 11a,b and 11c,d in rat liver
microsomal membranes. Values are the meansÆSEM of 10 di€erent
experiments. Each mean has a standard error less than 10%.
In order to verify the in¯uence of possible metabolic
transformation, and to compare its activity with that of
simple physical mixtures of the corresponding anti-
oxidant moieties, stereoisomers 11c,d were also com-
pared to associations of the natural a-tocopherol and
ascorbic acid and of the respective synthetic analogues 5
and 3. As shown in Fig. 7, although 3 was the less active
among the test compounds (compare Fig. 4 with Figs 2,
3 and 5±7) no signi®cant di€erences in AAPH-induced
MDA inhibition were found among the compounds
11c,d, and the association of both 3 and 5. This occur-
rence may be explained on the basis of the high activity
displayed by 5 itself (Fig. 5) which represent an unex-
pected result of this study. The lack of signi®cative
Figure 7. Inhibition of AAPH-induced malondialdehyde production
by di€erent concentrations of compounds 12a,b and 12c,d in rat liver
microsomal membranes. Values are the meansÆSEM of 10 di€erent
experiments. Each mean has a standard error less than 10%.
Figure 8. AAPH-induced MDA production in the absence (control) or in the presence of a-tocopherol+ascorbic acid, compound 5+compound 3
or compounds 11c,d, in rat liver microsomal membranes. Values are the means SEM of three di€erent experiments. The value with asterisk are
signi®cantly di€erent from control values (p<0.05).

2796
S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
increasing in potencies may also be explained by a
reduced mobility of compounds 11c,d in aqueous phase
or within the membrane bilayer with respect to a-toco-
pherol and ascorbic acid.
antioxidants such as a-tocopherol o€er protection
against reperfusion injury and, in view of the relevance
of ischemic diseases, there is an increasing interest in the
research of new agents. In view of this evidence, we thus
decided to investigate the potential of our compounds in
preventing damages induced by reperfusion on isolated
rabbit heart.
Pharmacology
Ischemia-reperfusion in isolated rabbit heart
When the perfusion ¯ow rate of electrically paced iso-
volumic left rabbit heart preparations was reduced from
20 to 1 mL/min for a period of 40 min, a progressive
increase in LVEDP values (from 4.1Æ0.5 to 72.2Æ6.8
mmHg; p<0.001) occurred suggesting that an ischemic
phenomenon was taking place. In fact, ventricular
function was impaired since the recovery of post-
ischemic LVDP was low and after 20 min of reperfu-
sion, only 38% of the preischemic strength of heart
contractility was restored (38.9Æ4.2 versus 102.5Æ8.5
mmHg; p<0.001). At this time, CPP considerably aug-
mented over the basal values (from 52.5Æ5.1 to
96.6Æ6.3 mmHg; p<0.001) suggesting an increase in
coronary vascular resistance owing to the sti€ness of the
heart (Figs 8 and 9). Furthermore, the partial functional
recovery of the hearts during reperfusion was accom-
panied by a consistent release of CK into the cardiac
perfusates. In fact, peak concentration of CK was
increased 9.1-fold compared to preischemic values
(from 49Æ6 to 448Æ40 mU/min/g wet tissue; p<0.001)
(Fig. 10).
Ischemic tissues, i.e., heart after myocardial infarct, are
vulnerable to damage when rapidly reperfused with
oxygen-rich blood in order to reduce the extent of the
injury; this occurrence is known as reperfusion damage.
The further damage is due to the presence of oxygen-
derived free radicals and reactive oxygen species present
in the reperfusion medium. Many classes of drugs and
When the hearts were perfused for 30 min in the pre-
ischemic period with graded concentrations of 11a±d
(100±300 mM),
a
myocardial protection against
mechanical changes due to ischemia-reperfusion was
signi®cantly evident at the higher concentration of 300
mM. In fact, the increase in LVDP values observed in
untreated preparations at the end of the ischemic period
(from 5.8Æ0.5 to 34.1Æ2.5 mmHg; p<0.05) was sig-
Figure 9. E€ects of compounds 11a±d in perfused rabbit hearts sub-
jected to ischemia-reperfusion: trend of left ventricular end diastolic
pressure (LVEDP) and coronary perfusion pressure (CPP). Data are
meanÆSEM of ®ve di€erent heart preparations per group. Compound
11 (100±300 mM) was perfused through the hearts for a period of 30
min before reduction of coronary ¯ow. The areas under the curve
(AUCs) related to LVEDP are: a, 3046Æ284; b, 2577Æ191; c,
2021Æ177. The AUCs related to CPP (increase in mmHg over the
preischemic values) are: a, 716Æ83; b, 593Æ49; c, 424Æ23; with statis-
tical di€erences for both LVEDP and CPP: a versus b, p>0.05; a ver-
sus c, p<0.05. The AUCs was evaluated according to the trapezoid
method: in ordinate, LVEDP or CPP in mmHg; in abscissa, time from
0 to 60 min (for LVEDP) or from 40 to 60 min (for CPP).
Figure 10. E€ects of compounds 11a±d on left ventricular developed
pressure (LVDP) in perfused rabbit heart subjected to ischemia-reper-
fusion. Data are meanÆSEM of 5 di€erent heart preparations per
group. Compounds 11a±d (100±300 mM) were perfused through the
hearts for a period of 30 min before reduction of coronary ¯ow. The
areas under the curve (AUCs) related to LVDP are: a, 593Æ42; b,
672Æ57, c, 835Æ52; with statistical di€erences for LVDP: a versus b,
p>0.05; a versus c, p<0.05. The AUCs was evaluated according to
the trapezoid method: in ordinate, LVDP in mmHg; in abscissa, time
from 40 to 60 min.

S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
2797
In conclusion, some new molecules, based on the concept
of the molecular combination of cooperative antioxidants,
have been prepared. In a peroxyl radical-dependent lipo-
peroxidation assay, these compounds were all inhibitors
of MDA formation. However, the synthetic compounds
11c,d, which are the 5R,2R,2R,4S and 5R,2R,2S,4S iso-
mers of the molecular combination of an a-tocopherol-
analogue and l-ascorbic acid, resulted in interesting anti-
oxidants with a potency comparable, if not slightly
superior, to the respective synthetic a-tocopherol analogue
(5) and natural a-tocopherol or l-ascorbic acid, used
alone or in combination. Moreover, the major advantages
of this compound derives from its capability to trap free
radicals simultaneously in both hydrophilic and lipo-
philic environments. In view of these data 11a±d were
also investigated for their ability to prevent damage
induced by reperfusion on isolated rabbit heart result-
ing e€ective at the higher concentration of 300 mM.
Taken together, these results are of signi®cance for
possible therapeutic applications in pathological events
in which free radical damage is involved and for the
possible extension of our approach to other synergistic
antioxidants.
Figure 11. Kinetic pro®le of creatine kinase (CK) activity released in
perfusates of rabbit hearts subjected to ischemia-reperfusion. Data are
meanÆSEM of ®ve di€erent heart preparations per group. Compound
11a±d (100±300 mM) was perfused through the hearts for a period of
30 min before reduction of coronary ¯ow. The areas under the curve
(AUCs) related to CK are: a, 5081Æ408; b, 4553Æ254, c, 3220Æ272;
with statistical di€erences for CK: a versus b, p>0.05; a versus c,
p<0.01. The AUCs was evaluated according to the trapezoid method:
in ordinate, CK in mU/min/g wet tissue; in abscissa, time from 40 to
60 min.
ni®cantly less in 11a±d in infused hearts and at the end
of reperfusion period LVDP reached 57% of the pre-
ischemic values (57.5Æ5.8 versus 101.4Æ9.3 mmHg;
p<0.01). At the same time, the increase in CPP in
untreated hearts was diminished (from 96.6Æ6.3 to
77.5Æ6.9 mmHg; p<0.05) indicating a reduction of
coronary vascular resistance (Figs 9 and 10). In keeping
with these results, the kinetic pro®le of CK released in
the cardiac e‚uent was signi®cantly di€erent from that
observed in untreated preparations. At the peak of the
concentration, CK was increased 6.2-fold (from 50Æ5 to
312Æ30 mU/min/g wet tissue; p<0.001) (Fig. 11).
Experimental
Materials
Compounds 4,13 715 and 814 were prepared following
and adapting reported procedures.
General
Melting points were determined by a Ko¯er apparatus
and are uncorrected. Reaction courses were routinely
monitored by thin layer chromatography (TLC) on
silica gel precoated Durasil-25 UV254 Machery±Nagel
plates with detection under 254 nm UV lamp and/or by
spraying the plates with 10% H2SO4/MeOH and/or
with 5% KMnO4/H2O solutions and heating. Nuclear
magnetic resonance (1H NMR) spectra were determined
in DMSO-d6, CDCl3 solutions with a Bruker AC-200
spectrometer and chemical shift are presented in ppm
from internal tetramethylsilane as a standard. Matrix-
assisted laser desorption ionization time-of-¯ight
(MALDI-TOF) were obtained using the HPG2025A
mass spectrometer. HPLC analysis was performed on
Beckman System Gold with a Beckman ultrasphere
ODS column (5 mm; 4.6Â250 mm). Analytical determi-
nations were carried out by two solvent systems:
(method 1) A=water, B=acetonitrile, a linear gradient
was initiated after injection of a sample and run from
100% A to 100% B in a 25 min. All compounds showed
a purity higher than 99% following analytical HPLC
monitored at 220 and 254 nm. Column chromatography
was performed with Macherey±Nagel 70±230 mesh
silica gel. Purity of ®nal compounds was also assessed
by high resolution mass and combustion analyses.
Where not di€erently stated, microanalyses were in
agreement with calculated values (sdÆ0.4%).
Stability
In order to investigate if the observed activity may be
related to a possible hydrolysis of our molecules, stabi-
lity of the mixture of stereosiomers 11a±d was evaluated
in water at 37 ꢀC, at pH ranging from 3 to 8. Com-
pounds 11a±d proved stable, with half-lives ranging
from 1 to 3 days (Table 2). The obtained half-lives con-
®rm that they act per se and not after transformation
into aldehyde 5 and l-ascorbic acid.
Table 2. In vitro hydrolysis of stereoisomers 11a±d at di€erent pH;
t1=2 37 ꢀC
Compound
Product
In vitro hydrolysis
t1=2 37 ꢀC, ha
pH
11a±d
5, ascorbate
3
5
7
8
24
38
72
60
aHPLC analysis; t1=2 is the time required for 50% hydrolysis to the
corresponding 5 and ascorbic acid.

2798
S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
(R,S)-6-Hydroxy-(2,5,7,8-tetramethylchroman-2-yl)-acet-
aldehyde (5). A solution of 6-ethyl-(2,5,7,8-tetramethyl-
chroman-2-yl)acetate (4) (2.5 g, 7.8 mmol) in 130 mL of
anhydrous hexane, was cooled under argon atmosphere,
in a dry ice/ethanol bath to 70 ꢀC. To the resulting
suspension was added, dropwise, a solution of diisobutyl±
aluminum hydride in hexane (25%, 13 mL, 23.4 mmol).
The reaction was monitored by TLC (hexane:Et2O 7:3).
After 15 min, the reaction was quenched at 70 ꢀC with
methanol and water, and worked-up with ether, washed
with 2M HCl, dried over anhydrous Na2SO4 and the
solvent was removed under reduced pressure. The crude
product was puri®ed by silica gel chromatography
(hexane:Et2O 7:3) to give 1.55 g of white solid; yield
82%. Mp 82±85 ꢀC (Et2O/hexane). 1H NMR (CDCl3): d
1.32 (s, 3H, CH3-C2); 1.79±1.87 (m, 4H, ArCH2CH2);
2.03, 2.05, 2.09 (3s, 9H, 3 CH3 Ar); 2.4±2.7 (m, 4H,
ArCH2CH2, CH2COH); 9.85 (dd, 1H, J=6 and 2.4 Hz,
CHO). MALDI-TOF MS: m/z 248.7 (M)+, C15H20O3
requires 248.03. Anal. C, H.
1H, H-C2 dioxolane); 5.14 (s, 2H, CH2-Ph at C4 fur-
anone); 7.1±7.5 (m, 15H, Ph). MALDI-TOF MS: m/z
676.9 (M)+, C42H44O8 requires 676.79.
9b oil; 320 mg, yield 14%; TLC (hexane:Et2O 6:4, 0.34);
1H NMR (CDCl3): d 1.21 (s, 3H, CH3-C2 benzopyran);
1.7 (m, 2H, ArCH2CH2); 1.9 (dd, 2H, CH2, J=5 Hz,
J=4.4); 2.01, 2.08, 2.13 (3s, 9H, 3 CH3 Ar); 2.5 (t, 2H,
ArCH2CH2); 3.6±3.95 (m, 2H, H-C5 dioxolane); 4.2 (m,
1H, H-C4 dioxolane); 4.47±4.49 (d, 1H, H-C5 furanone,
J=4 Hz); 4.46 (d, 2H, CH2-Ph, benzopyran); 5.01 (d,
2H, CH2-Ph at C3 furanone); 5.09 (m, 1H, H-C2 diox-
olane); 5.13 (s, 2H, CH2-Ph at C4 furanone); 7.1±7.5 (m,
15H, Ph). MALDI-TOF MS: m/z 677.8 (M+H)+,
C42H44O8 requires 676.79.
9c foam; 280 mg, yield 12%; TLC (hexane:Et2O 6:4,
1
0.32); H NMR (CDCl3): d 1.21 (s, 3H, CH3-C2 benzo-
pyran); 1.7 (m, 2H, ArCH2CH2); 1.9 (dd, 2H, CH2,
J=5 Hz, J=4.4); 2.01, 2.08, 2.13 (3s, 9H, 3 CH3 Ar);
2.5 (t, 2H, ArCH2CH2); 3.6±3.95 (m, 2H, H-C5 dioxo-
lane); 4.2 (m, 1H, H-C4 dioxolane); 4.48±4.50 (d, 1H,
H-C5 furanone, J=3.4 Hz); 4.6 (d, 2H, CH2-Ph benzo-
pyran); 5.01 (d, 2H, CH2-Ph at C3 furanone); 5.10 (m,
1H, H-C2 dioxolane); 5.14 (s, 2H, CH2-Ph C4 fur-
anone); 7.1±7.5 (m, 15H, Ph). MALDI-TOF MS: m/z
676.8 (M)+, C42H44O8 requires 676.79.
(R,S)-6-Benzyloxy-(2,5,7,8-tetramethylchroman-2-yl)-acet-
aldehyde (6). A mixture of 5 (2.2 g, 8.8 mmol), anhy-
drous K2CO3 (1.28 g, 9.2 mmol), benzyl bromide (1.2
mL, 10.8 mmol), and 20 mL of anhydrous DMF was
stirred for 22 h at room temperature. The obtained sus-
pension was poured into H2O, worked-up with ether,
dried over anhydrous Na2SO4 and the solvent was
removed under reduced pressure. The crude product
was puri®ed by silica gel chromatography (hexane:Et2O
9d foam; 370 mg, yield 16%; TLC (hexane:Et2O 6:4,
1
0.30); H NMR (CDCl3): d 1.21 (s, 3H, CH3-C2 benzo-
1
9:1) to give 1.3 g of yellow oil; yield 43%. H NMR
pyran); 1.6±1.8 (m, 2H, ArCH2CH2); 1.9 (dd, 2H, CH2,
J=5 Hz, J=4.4); 2.01, 2.08, 2.13 (3s, 9H, 3 CH3 Ar);
2.5 (t, 2H, ArCH2CH2); 3.6-3.95 (m, 2H, H-C5 dioxo-
lane); 4.2 (m, 1H, H-C4 dioxolane); 4.50±4.52 (d 1H, H-
C5 furanone, J=3.2 Hz); 4.6 (d, 2H, CH2-Ph benzo-
pyran); 5.01 (d, 2H, CH2-Ph al C3 furanone); 5.09 (m,
1H, H-C2 dioxolane); 5.14 (s, 2H, CH2-Ph at C4 diox-
olane); 7.1±7.5 (m, 15H, Ph). MALDI-TOF MS: m/z
676.9 (M)+, C42H44O8 requires 676.79.
(CDCl3): d 1.32 (s, 3H, CH3-C2); 1.79±1.87 (m, 4H,
ArCH2CH2); 2.03, 2.05, 2.09 (3s, 9H, 3 CH3 Ar); 2.48±
2.61 (m, 4H, ArCH2CH2, CH2COH); 4.73 (s, 2H,
CH2Ph); 7.40±7.56 (m, 5H, Ph); 9.85 (dd, 1H, CHO,
J=6 and 2.4 Hz). MALDI-TOF MS: m/z 338.7 (M)+,
C22H26O3 requires 338.44.
3,4-Dibenzyloxy-5R-2S-(6-benzyloxy-2,5,7,8-tetramethyl-
chroman-2(R and S)yl-methyl)-1,3]dioxolan-4S-yl]-5H-
furan-2-one (9a,b) and 3,4-dibenzyloxy-5R-2R-(6-benzyl-
oxy-2,5,7,8-tetramethylchroman-2(R and S)yl-methyl)-
1,3]dioxolan-4S-yl]-5H-furan-2-one (9c,d). A mixture of
2,3-O-benzyl-ascorbic acid (7) (1.2 g, 3.3 mmol), 6 (2.3
g, 6.8 mmol), p-toluenesulfonic acid (150 mg, 0.78
mmol) in 90 mL of benzene, was re¯uxed in a Dean±
Stark apparatus for 8 h. The reaction was then poured
into H2O, worked-up with EtOAc, dried over anhy-
drous Na2SO4 and the solvent was removed under
reduced pressure, to give a 1:1:1:1 mixture of the
four diastereoisomers (HPLC analysis). The crude
product was puri®ed by silica gel chromatography
(hexane: Et2O 6:4) to give compounds 9a±d, overall
yield 55%.
3,4-Dihydroxy-5R-2S-(6-hydroxy-2,5,7,8-tetramethyl-chro-
man-2(R and S)yl-methyl)-1,3]dioxolan-4S-yl]-5H-furan-2-
one (11a,b) and 3,4-dihydroxy-5R-2R-(6-hydroxy-oxy-
2,5,7,8-tetramethyl-chroman-2(R,S)yl-methyl)-1,3]di-oxo-
lan-4S-yl]-5H-furan-2-one (11c,d). The four diastere-
isomers were deprotected by the following general
procedure: each of the above obtained 9a±d (200 mg,
0.29 mmol each-one) were dissolved in EtOAc (10 mL)
and hydrogenated in presence of 10% Pd/C (50 mg)
for 3 h, at atmospheric pressure. The catalyst was
then removed by ®ltration, and the ®ltrate was con-
centrated under reduced pressure. The residue was
puri®ed by silica gel chromatography (CH2Cl2:MeOH
9:1) to give the expected 11a±d in good yield (90±
93%).
9a oil; 230 mg, yield 10%; TLC (hexane:Et2O 6:4, 0.35);
1H NMR (CDCl3): d 1.21 (s, 3H, CH3-C2 benzopyran);
1.7 (m, 2H, ArCH2CH2 benzopyran); 1.9 (dd, 2H, CH2,
J=5 Hz, J=4.4); 2.01, 2.08, 2.13 (3s, 9H, 3 CH3 Ar);
2.5 (t, 2H, ArCH2CH2); 3.6±3.95 (m, 2H, H-C5 dioxo-
lane); 4.2 (m, 1H, H-C4 dioxolane); 4.49±4.50 (d, 1H,
H-C5 furanone, J=2.4 Hz); 4.47 (d, 2H, CH2-Ph, ben-
zopyran); 5.01(d, 2H, CH2-Ph at C3 furanone); 5.09 (m,
1
11a oil; 106 mg, yield 90%; H NMR (DMSO-d6): d
1.21 (s, 3H, CH3-C2 benzopyran); 1.7 (m, 2H,
ArCH2CH2); 1.85 (dd, 2H, CH2, J=4 Hz, J=4.4); 1.95,
2.0, 2.03 (3s, 9H, 3 CH3 Ar); 2.49 (t, 2H, ArCH2CH2);
3.9±4.1 (m, 2H, H-C5 dioxolane); 4.24 (m, 1H, H-C4
dioxolane); 4.79 (d, 1H, H-C5 furanone, J=4 Hz); 5.02
(t, 1H, H-C2 dioxolane); 7.40 (s, 1H, OH 6-benzo-

S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
2799
1
pyran); 8.5 (sbr, 1H, OH at C3 furanone); 11.35 (br, 1H,
OH, at C4 furanone). MALDI-TOF MS: m/z 406.8
(M)+, C21H26O8 requires 406.4. Anal. C, H.
10c,d: foam; 310 mg, yield 46%; H NMR (CDCl3): d
0.80 (m, 3H, CH3 alkyl); 1.20 (s, 3H, CH3-C2 benzo-
pyran); 1.22±1.26 (m, 32H, alkyl); 1.6±1.8 (m, 2H,
ArCH2CH2); 1.9 (m, 2H, CH2); 2.01, 2.08, 2.14 (3s, 9H,
3 CH3 Ar); 2.5 (t, 2H, ArCH2CH2); 3.8±4.0 (m, 4H,
-OCH2 alkyl, CH2-C5 dioxolane); 4.1±4.3 (m, 1H, H-C4
dioxolane); 4.50 (c) and 4.52 (d) (dÂ2, 1H, H-C5 fur-
anone, J=4 Hz); 4.61 (d, 2H, CH2-Ph benzopyran);
5.08±5.10 (m, 1H, H-C2 dioxolane); 5.4 (s, 2H, CH2-
Ph at C4 furanone); 7.1±7.4 (m, 10H, Ph). MALDI-
TOF MS: m/z 840.2 (M+H)+, C53H74O8 requires
839.14.
1
11b oil; 108 mg, yield 92%; H NMR (DMSO-d6): d
1.21 (s, 3H, CH3-C2 benzopyran); 1.7 (m, 2H,
ArCH2CH2); 1.8 (dd, 2H, CH2, J=4 Hz, J=4.4); 1.95,
2.0, 2.03 (3s, 9H, 3 CH3 Ar); 2.49 (t, 2H, ArCH2CH2);
3.9±4.1 (m, 2H, H-C5 dioxolane); 4.24 (m, 1H, H-C4
dioxolane); 4.77 (d, 1H, H-C5 furanone, J=4Hz); 5.02
(t, 1H, H-C2 dioxolane); 7.40 (s, 1H, OH 6 benzo-
pyran); 8.5 (sbr, 1H, OH at C3 furanone); 11.35 (br, 1H,
OH at C4 furanone). MALDI-TOF MS: m/z 406.9
(M)+, C21H26O8 requires 406.4. Anal. C, H.
4-O-Hydroxy-5R-2S-(6-hydroxy-2,5,7,8-tetramethyl-
chroman-2(R,S)-yl-methyl)-1,3]dioxolan-4S-yl]-3-octadecyl-
oxy-5H-furan-2-one (12a,b) and 4-O-hydroxy-5R-2R-(6-
hydroxy-2,5,7,8-tetramethylchroman-2(R,S)-yl-methyl)-
1,3]dioxolan-4S-yl]-3-octadecyloxy-5H-furan-2-one (12c,d).
Compounds 10a,b (200 mg, 0.24 mmol) and 10c,d (280
mg, 0.33 mmol) were deprotected by the procedure
described above for compounds 9a±d. After the usual
work up and silica gel chromatography puri®cation
(CH2Cl2:MeOH 9:1), the expected 12a,b and 12c,d were
obtained in good yields as inseparable mixtures of the
of two diastereoisomer.
1
11c foam; 109 mg, yield 93%; H NMR (DMSO-d6): d
1.21 (s, 3H, CH3-C2 benzopyran); 1.7 (m, 2H,
ArCH2CH2); 1.8 (dd, 2H, CH2, J=4 Hz, J=4.4); 1.95,
2.0, 2.03 (3s, 9H, 3 CH3 Ar); 2.49 (t, 2H, ArCH2CH2);
3.9±4.1 (m, 2H, H-C5 dioxolane); 4.24 (m, 1H, H-C4
dioxolane); 4.73 (d, 1H, H-C5 furanone, J=4Hz); 5.02
(t, 1H, H-C2 dioxolane); 7.42 (s, 1H, OH benzopyran);
8.5 (sbr, 1H, OH at C3 furanone); 11.35 (br, 1H, OH, at
C4 furanone). MALDI-TOF MS: m/z 406.9 (M)+,
C21H26O8 requires 406.4. Anal. C, H.
1
1
11d foam; 106 mg, yield 90%; H NMR (DMSO-d6): d
12a,b: oil; 142 mg, yield 90%; H NMR (DMSO-d6): d
1.21 (s, 3H, CH3-C2 benzopyran); 1.7 (m, 2H,
ArCH2CH2); 1.8 (dd, 2H, CH2, J=4 Hz, J=4.4.); 1.95,
2.0, 2.03 (3s, 9H, 3 CH2 Ar); 2.49 (t, 2H, ArCH2CH2);
3.9±4.1 (m, 2H, H-C5 dioxolane); 4.24 (m, 1H, H-C4
dioxolane); 4.72 (d, 1H, H-C5 furanone, J=4Hz); 5.02
(t, 1H, H-C2 dioxolane); 7.42 (s, 1H, OH benzopyran);
8.5 (sbr, 1H, OH at C3 furanone); 11.35 (br, 1H, OH, at
C4 furanone). MALDI-TOF MS: m/z 406.9 (M)+,
C21H26O8 requires 406.4. Anal. C, H.
0.84 (t, 3H, J=6 Hz, CH3, alkyl); 1.20 (s, 3H, CH3-C2
benzopyran); 1.22±1.26 (m, 32H, CH2 alkyl); 1.6±1.8
(m, 2H, Ar-CH2-CH2); 1.9 (dd, 2H, J=5 and 4.4 Hz,
-CH2-); 1.95, 1.99, 2.03 (3 s, 9H, 3 CH3 Ar.); 2.44±2.5
(m, 2H, Ar-CH2-CH2-); 3.83±3.92 (m, 4H, CH2-C5
dioxolane, -OCH2 alkyl); 4.31±4.38 (m, 1H, -CH, H-C4
dioxolane); 4.82 (a) and 4.84 (b) (d, 1H, J=2.2 Hz, H-
C5 furanone); 4.96±5.01 (m, 1H, H-C2 dioxolane); 7.40
(sbr, 1H, OH benzopyran); 10.9 (sbr, 1H, OH at C4
furanone). MALDI-TOF MS: m/z 659.8 (M+H)+,
C39H62O8 requires 658.9. Anal. C, H.
4-O-Benzyloxy-5R-2S-(6-benzyloxy-2,5,7,8-tetramethyl-
chroman-2(R,S)yl-methyl)-1,3]dioxolan-4S-yl]-3-octadecyl-
oxy-5H-furan-2-one (10a,b)- and 4-O-benzyloxy-5R-2R-
(6 - benzyloxy - 2,5,7,8 - tetramethylchroman - 2(R,S) - yl -
methyl)-1,3]dioxolan-4S-yl]-3-octadecyloxy-5H-furan-2-one
(10c,d). 2-O-Octadecyl-3-O-benzyl-ascorbic acid (8)
(414 mg, 0.8 mmol) was reacted with 6 (540 mg,
1.6 mmol) in the same conditions described above for
9a±d. After work up the solvent was removed under
reduced pressure to give a 1:1:1:1 mixture of the four
diastereoisomers. The crude product was puri®ed by
silica gel chromatography (hexane:Et2O 6:4) to give
10a,b and 10c,d as inseparable mixtures of the couple of
two diastereo-isomers.
1
12c,d: foam; 148 mg, yield 94%; H NMR (DMSO-d6):
d 0.84 (t, 3H, J=6 Hz, CH3, alkyl); 1.20 (s, 3H, CH2-C2
benzopyran); 1.22±1.26 (m, 32H, CH2, alkyl); 1.6±1.8
(m, 2H, Ar-CH2-CH2); 1.9 (dd, 2H, J=5 and 4.4 Hz,
-CH2-); 1.95, 1.99, 2.03 (3 s, 9H, 3 CH3 Ar); 2.44±2.5
(m, 2H, Ar-CH2-CH2); 3.83±3.92 (m, 4H, CH2-C5
dioxolane, -OCH2 CH2 alkyl); 4.31±4.38 (m, 1H, -CH-,
H-C4 dioxolane); 4.78 (d, 1H, J=2.2 Hz, H-C5 fur-
anone); 4.99±5.01 (m, 1H, -CH-, H-C2 dioxolane); 7.42
(sbr, 1H, OH benzopyran); 11.1 (sbr, 1H, OH at C4
furanone). MALDI-TOF MS: m/z 658.9 (M)+,
C39H62O8 requires 658.9. Anal. C, H.
10a,b: oil; 230 mg, yield 34%; 1H NMR (CDCl3): d 0.80
(m, 3H, CH3); 1.2 (s, 3H, CH3-C2 benzopyran); 1.22±
1.26 (m, 32H, alkyl); 1.6±1.8 (m, 2H, ArCH2CH2); 1.9
(m, 2H, CH2); 2.01, 2.08, 2.14 (3s, 9H, 3 CH3 Ar); 2.5 (t,
2H, ArCH2CH2); 3.8±4.0 (m, 4H, -OCH2 alkyl, CH2-C5
dioxolane); 4.1±4.3 (m, 1H, H-C4 dioxolane); 4.48 (a)
and 4.49 (b) (dÂ2, 1H, J=4 Hz, H-C5 furanone,); 4.61
(d, 2H, CH2±Ph benzopyran); 5.06±5.08 (m, 1H, H-C2
dioxolane); 5.4 (s, 2H, CH2-Ph at C4 furanone); 7.1±7.4
(m, 10H, Ph). MALDI-TOF MS: m/z 839.8 (M)+,
C53H74O8 requires 839.14.
Chemical stability
Solutions of compounds 11a±d were prepared dissolving
an aliquot of compounds in water at pH 3, 5 and 8 to
give a ®nal concentration of about 10 3 M. The solution
was maintained at 37 ꢀC and aliquots were withdrawn
every 2 h for the initial 12 h of incubation and succes-
sively every 12 h for 4 days. The disappearance of the
compounds was monitored by HPLC analysis using
the method 1. Conjugates half-lives were from 1 to 3
days.

2800
S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
Molecular modeling
Pharmacology
Ischemia-reperfusion in isolated rabbit heart
Computational studies have been conducted on a Sili-
con Graphics Indigo 2 and O2 work stations using the
SYBYL20 and SPARTAN21 software packages. A sys-
tematic conformational search was conducted on
compounds 11(a,b) cis and 11(c,d) trans with SYBYL
and all the conformers obtained were energy minimized
using MAXIMIN2 (Tripos force ®eld, Powell
method22). Semiempirical quantum-mechanic calcula-
tions were performed using the MOPAC (AM1) pro-
gram23 under SYBYL. The conformations having
closer distances between C5 proton at the furanone
ring and C3 proton and C2 methyl at the chromanyl
residue were studied with ab initio Hartree±Fock
methods,24 geometry was optimized with STO-3G
basic set to determine the position of the three ring
systems.
All experimental protocols were approved by the Animal
Care Committee of the University of Milan and were in
accordance with the Italian guidelines for the laboratory
animals which conform with the European Commu-
nities Directive of November 1986 (86/609/EEC).
Male New Zealand White rabbits (BMG-Allevamento,
Cividate al Piano, BG, Italy) weighing 2.2±2.5 kg were
used for these experiments. The hearts were excised and
perfused retrogradely at 37 ꢀC through the aorta as
previously described by Henry et al.27 and slightly
modi®ed by Berti et al.28 The perfusion medium (Krebs
Henseleit) contained (in mM): NaCl 118, KCl 2.8,
KH2PO4 1.2, CaCl2 2.5, MgSO4 1.2, NaHCO3 25 and
glucose 5.5. The pH of the perfusate was 7.4 after a
period of equilibration with a 5% CO2 and 95% O2 gas
mixture, and the rate of perfusion was maintained at 20
mL/min with a roller pump (Minipuls 3, Gilson, Vil-
liers-Le Bel, France). Coronary perfusion pressure
(CPP) and left ventricular pressure (LVP) were mea-
sured with two HP-1280C pressure transducers (Hew-
lett-Packard, Waltham, MA, USA) connected with a
Hewlett-Packard dynograph (HP-7754A). LVP was
recorded with a polyethylene catheter (with a small latex
balloon on the top) inserted in the left ventricular cav-
ity. The balloon was ®lled slowly with saline until left
ventricular end-diastolic pressure (LVEDP) stabilised in
the range of 4±6 mmHg. Left ventricular developed
pressure (LVDP; peak left ventricular systolic pressure
minus LVEDP) was also evaluated. The hearts were
electrically paced at a frequency of 180 beats/min with
rectangular impulses (1 ms duration, voltage 10% above
threshold) by a Grass stimulator (mod. S-88; Grass
Instr., Quincy, MA, USA). CPP of 55±60 mmHg was
obtained with a ¯ow of 20 mL/min (preischemic period).
Ischemia was induced by reducing the ¯ow rate to 1 mL/
min for 40 min (ischemic period; CPP=12±14 mmHg).
A normal ¯ow rate (20 mL/min) was then restored, and
the perfusion was continued for 20 min (reperfusion
period). The compound under investigation, 11a±d was
tested in groups of ®ve hearts each. In particular, 11a±d
(100±300 mM) was perfused through the hearts for a
period of 30 min before reduction of coronary ¯ow.
Biology
Preparation of microsomal membranes
Microsomal membranes were isolated from adult male
Wistar rats as previously reported.25 Microsomal pro-
tein were determined by Bradford method.
Lipid peroxidation. Lipid peroxidation was initiated by
2-20-azobis(2-amidinopropane)dihydrochloride (AAPH)
as previously described.10 The reaction was carried out
with 25 mM AAPH at 37 ꢀC under air for 15 min. Lipid
peroxidation was measured by the appearance of mal-
ondialdeyde (MDA) formation. MDA was extracted
and analyzed by HPLC as indicated.26 Brie¯y, aliquots
of 1 mL of membrane suspension were mixed with 0.5
mL of TBA solution (two parts 0.4% TBA in 0.2 M
HCl and one part distilled water) and 0.07 mL of 0.2
BHT in 95% ethanol. Samples were then placed in a
90 ꢀC water bath for 45 min. After incubation, the
TBA±MDA adduct was extracted with isobutyl alcohol.
The isobutanol extract was mixed with methanol (2:1)
prior to injection in HPLC system. The column was
packed with Supelcosil LC-18 material, 3 mm particle
size, in a 15 cmÂ4.6 mm cartridge format (Supelco,
Bellefonte, PA). A 2 cm cartridge precolumn containing
5 mm LC-18 Supelcosil packing was used. The mobile
phase was a 1:1 (v:v) mixture of methanol and double
distilled water, with the addition of tetrabutyl ammo-
nium dihydrogen phosphate (0.06%, w:v), as an ion
pairing reagent. The TBA±MDA adduct was detected
by a ¯uorimeter set at an excitation wavelength of
515 nm and an emission wavelength of 550 nm. At a
¯ow rate of 1 mL/min, the retention time of TBA±
MDA adduct was 5 min. MDA concentration was
calculated from a calibration curve generated from a
peak height of the MDA standard, prepared by acid
hydrolysis of 1,1,3,3-tetramethoxypropane.27 The con-
centration necessary to inhibit MDA formation by
50% (IC50) was calculated from linear plots (con-
centration versus percent of control) of at least three
di€erent concentrations.
Creatine kinase determination
The heart perfusates were collected every 150 s in an ice-
cooled beaker before ¯ow reduction and during reper-
fusion and the activity of creatine kinase (CK) was
evaluated according to the method of Bergmeyer et al.29
by a speci®c kit (Amersham-Italia, MI, Italy). The
amount of the enzyme was determined on a Lambda16-
spectrophotometer (Perkin±Elmer Italia, Monza, MI,
Italy) and expressed as mU/min/g/wet tissue.
Data analysis
In all experiments, di€erences between control and
treated groups were analyzed for statistical signi®cance
using a one-way analysis of variance (ANOVA) and

S. Manfredini et al. / Bioorg. Med. Chem. 8 (2000) 2791±2801
2801
Tukey±Kramer test for multiple comparisons. A value
of p<0.05 was considered signi®cant. In all ®gures,
results are expressed as meanÆstandard error of the
mean (SEM). The areas under the curve (AUCs) were
assessed by using a computerized program Microcal
Origin.30
10. Palozza, P.; Krinsky, N. Archiv. Biochem. Biophis. 1992,
297, 184.
11. Manfredini, S.; Baraldi, P.; Bazzanini, R.; Simoni, D.; Bal-
zarini, J.; De Clercq, E. Bioorg. Med. Chem. Lett. 1997, 7, 473.
12. Scott, J.; Bizzarro, F.; Parrish, D.; Saucy, G. Helv. Chim.
Acta 1976, 59, 290.
13. Scott, J.; Cort, W.; Harley, H.; Parrish, D.; Saucy, G. Am.
Oil. Chem. Soc. 1974, 51, 200.
14. Kato, K.; Terao, S.; Shimamoto, N.; Hirata, M. J. Med.
Chem. 1988, 31, 793.
15. Kulkarni, M.; Shankar, R. Tetrahedron 1996, 52, 1293.
16. Manfredini, S.; Guarneri, M.; Ferroni, R.; Simoni, D.;
Zaidi, J. H.; Grana, E.; Boselli, C. Drug Des. Dis. 1995, 13, 1.
17. Bartoli, G.; Giannattasio, B.; Palozza, P.; Cittadini, A.
Biochem. Biophys. Acta 1988, 966, 214.
18. Fitting of data point was obtained by using the Cricket
Graph III curve ®tting, Computer Associates International,
Inc.: New York, USA.
Acknowledgements
This work was supported by the University of Ferrara,
Italy.
References and Notes
19. Niki, E.; Noguchi, N.; Tsuchihashi, H.; Gotoh, N. Am. J.
Clin. Nutr. 1995, 62, 1322S.
1. Scott, G. Chemistry in Britain 1995, 11, 879.
2. Andersson, C.; Halleberg, A.; Hogberg, T. In Advances in
Drug Research; Testa, B. A., Meyers, U. A., Eds.; Academic:
London, 1997; pp 64±180.
20. SYBYL Molecular Modeling System, version 6.2, Tripos
Inc.: St. Louis, MO, USA
21. SPARTAN, version 5.0. Wavefunction, Inc.: Irvine, CA,
USA.
3. Sies, H. In Oxidative Processes and Antioxidants; Paoletti,
R., Samuelsson, B., Catapano, A., Poli, A., Rinetti, M., Eds.;
Raven: New York, 1994; pp 211±216.
22. Powell, M. J. D. Ed. by Rabinowitz, Gordon & Breach:
New York, 1970; Chapters 6 and 7.
23. Stewart, J. J. P. 2 IV, 19
4. Packer, L. In Oxidative Processes and Antioxidants; Pao-
letti, R., Samuelsson, B., Catapano, A., Poli, A., Rinetti, M.,
Eds.; Raven: New York, 1994; p 153.
24. Herhe, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A.;
Wiley: New York, 1986.
25. Bradford, M. M. Anal. Chem. 1976, 72, 248.
26. Tatum, V.; Changoir, C.; Chow, C. Lipids 1990, 25, 226.
27. Henry, P. D.; Shuchleib, R.; Davies, J.; Weiss, E. S.;
Sobel, B. E. Am. J. Physiol. 1977, 8, H667.
28. Berti, F.; Rossoni, G.; Magni, F.; Caruso, D.; Omini, C.;
Puglisi, L.; Galli, G. J. Cardiovasc. Pharmacol. 1988, 12, 438.
29. Bergmeyer, H. U.; Rich, W.; Butter, H.; Schmidt, E.;
Hillman, G.; Kreuz, F. H.; Stamm, D.; Lang, H.; Szasz, G.;
Laue, D. Z. Klin. Chem. 1970, 8, 658.
5. May, J. M. FASEB J. 1999, 13, 995.
6. Grisar, J.; Marciniak, G.; Bolkenius, F.; Verne-Mismer, J.;
Wagner, R. J. Med. Chem. 1995, 38, 2880.
7. Maxwell, S. Drugs 1995, 49, 345.
8. Morizaki, K.; Fukuoka, T.; Sasaki, M. Jpn. Kokai Tokkyo
Koho JP 05,331,166; CA 1994, 121, 109569k.
9. Rosneau, T.; Habicher, W. Chem. Pharm. Bull. 1997, 45, 1080.
30. Microcal Software Inc.: Northampton, MA, USA.