Impairment of myocardial contractility by anticancer anthracyclines: Role of secondary alcohol metabolites and evidence of reduced toxicity by a novel disaccharide analogue

Article abstract of DOI:10.1038/sj.bjp.0704369

1 The anticancer anthracycline doxorubicin (DOX) causes cardiotoxicity. Enzymatic reduction of a side chain carbonyl group converts DOX to a secondary alcohol metabolite that has been implicated in cardiotoxicity. We therefore monitored negative inotropism, assessed as inhibition of post-rest contractions, in rat right ventricle strips exposed to DOX or to analogues forming fewer amounts of their alcohol metabolites (epirubicin, EPI, and the novel disaccharide anthracycline Men 10755). 2 Thirty μM EPI exhibited higher uptake than equimolar DOX, but formed comparable amounts of alcohol metabolite due to its resistance to carbonyl reduction. MEN 10755 exhibited also an impaired uptake, and consequently formed the lowest levels of alcohol metabolite. Accordingly, DOX and EPI inhibited post-rest contractions by ～40-50%, whereas MEN 10755 inhibited by ～6%. 3 One hundred μM EPI exhibited the same uptake as equimolar DOX, but formed ～50% less alcohol metabolite. One hundred μM MEN 10755 still exhibited the lowest uptake, forming ～60% less alcohol metabolite than EPI. Under these conditions DOX inhibited post-rest contractions by 88%. EPI and MEN 10755 were ～18% (P<0.05) or ～80% (P<0.001) less inhibitory than DOX, respectively. 4 The negative inotropism of 30-100 μM DOX, EPI, or MEN 10755 correlated with cellular levels of both alcohol metabolites (r=0.88, P<0.0001) and carbonyl anthracyclines (r=0.79, P<0.0001). Nonetheless, multiple comparisons showed that alcohol metabolites were ～20-40 times more effective than carbonyl anthracyclines in inhibiting contractility. The negative inotropism of MEN 10755 was therefore increased by chemical procedures, like side chain valeryl esterification, that facilitated its uptake and conversion to alcohol metabolite but not its retention in a carbonyl form. 5 These results demonstrate that secondary alcohol metabolites are important mediators of cardiotoxicity. A combination of reduced uptake and limited conversion to alcohol metabolite formation might therefore render MEN 10755 more cardiac tolerable than DOX and EPI.

Full text of DOI:10.1038/sj.bjp.0704369

British Journal of Pharmacology (2001) 134, 1271 ± 1278
ã
2001 Nature Publishing Group All rights reserved 0007 ± 1188/01 $15.00
www.nature.com/bjp
Impairment of myocardial contractility by anticancer
anthracyclines: role of secondary alcohol metabolites and evidence
of reduced toxicity by a novel disaccharide analogue
2
1
1
*,1Giorgio Minotti, Massimo Parlani, Emanuela Salvatorelli, Pierantonio Menna,
2
2
2
2Amalia Cipollone, Fabio Animati, Carlo A. Maggi & Stefano Manzini
1Department of Drug Sciences, G. D'Annunzio University School of Pharmacy, Via dei Vestini, 66013 Chieti, Italy and
2Menarini Ricerche S.p.A., Via Tito Speri 10, 00040 Pomezia (Rome); Italy
1
The anticancer anthracycline doxorubicin (DOX) causes cardiotoxicity. Enzymatic reduction of a
side chain carbonyl group converts DOX to a secondary alcohol metabolite that has been implicated
in cardiotoxicity. We therefore monitored negative inotropism, assessed as inhibition of post-rest
contractions, in rat right ventricle strips exposed to DOX or to analogues forming fewer amounts of
their alcohol metabolites (epirubicin, EPI, and the novel disaccharide anthracycline MEN 10755).
2
Thirty mM EPI exhibited higher uptake than equimolar DOX, but formed comparable amounts
of alcohol metabolite due to its resistance to carbonyl reduction. MEN 10755 exhibited also an
impaired uptake, and consequently formed the lowest levels of alcohol metabolite. Accordingly,
DOX and EPI inhibited post-rest contractions by *40 ± 50%, whereas MEN 10755 inhibited by
*6%.
3
One hundred mM EPI exhibited the same uptake as equimolar DOX, but formed *50% less
alcohol metabolite. One hundred mM MEN 10755 still exhibited the lowest uptake, forming *60%
less alcohol metabolite than EPI. Under these conditions DOX inhibited post-rest contractions by
88%. EPI and MEN 10755 were *18% (P50.05) or *80% (P50.001) less inhibitory than DOX,
respectively.
4
The negative inotropism of 30 ± 100 mM DOX, EPI, or MEN 10755 correlated with cellular levels
of both alcohol metabolites (r=0.88, P50.0001) and carbonyl anthracyclines (r=0.79, P50.0001).
Nonetheless, multiple comparisons showed that alcohol metabolites were *20 ± 40 times more
e€ective than carbonyl anthracyclines in inhibiting contractility. The negative inotropism of MEN
10755 was therefore increased by chemical procedures, like side chain valeryl esteri®cation, that
facilitated its uptake and conversion to alcohol metabolite but not its retention in a carbonyl form.
5
These results demonstrate that secondary alcohol metabolites are important mediators of
cardiotoxicity. A combination of reduced uptake and limited conversion to alcohol metabolite
formation might therefore render MEN 10755 more cardiac tolerable than DOX and EPI.
British Journal of Pharmacology (2001) 134, 1271 ± 1278
Keywords: Doxorubicin; epirubicin; MEN 10755; secondary alcohol metabolites; cardiotoxicity
Abbreviations: Anthracycline-ol, anthracycline secondary alcohol metabolite; DOX, doxorubicin (7-[(3-amino-2,3,6-trideoxy-a-L-
lyxo-hexopyranosyl)]doxorubicinone); DOXol, doxorubicinol (C-13 hydroxydoxorubicin); EPI, epirubicin (7-[(3-
amino-2,3,6-trideoxy-a-L-arabino-hexopyranosyl)]doxorubicinone); MEN 10755, (7-[(2,6-dideoxy-4-O-[3-amino-
2,3,6-trideoxy-a-L-lyxo-hexopyranosyl)]-a-L-lyxohexopyranosyl]-4-demethoxydoxorubicinone); MEN 11609, (7-
[(2,6-dideoxy-4-O-[3-amino-2,3,6-trideoxy-a-L-lyxo-hexopyranosyl)]-4-demethoxy-14-O-valeryl-doxorubicinone)
Introduction
The anthracycline doxorubicin (DOX) is active against
several human tumours but also causes severe cardiac
toxicity, resulting in heart failure (Weiss, 1992). DOX is
composed of aglyconic and sugar moieties. The aglycone,
called doxorubicinone, is composed of a quinone-containing
tetracyclic ring, and of a short side chain with a carbonyl
group at C-13 and a primary alcohol at C-14; the sugar,
called daunosamine, is attached by a glycosidic bond to C-7
of the tetracyclic ring, and consists of a 3-amino-2,3,6-
trideoxy-L-fucosyl moiety (Figure 1). Two-electron reduction
of the side chain carbonyl group (7CO-CH2OH?
7CHOH-CH2OH), mediated by cytoplasmic aldo/keto or
carbonyl reductases, converts DOX to a secondary alcohol
metabolite (doxorubicinol, DOXol) that was shown to
inactivate cytoplasmic aconitase, an important post-transcrip-
tional regulator of iron homeostasis (Minotti et al., 1998).
DOXol was also shown to be 10 ± 50 times more potent than
DOX at inhibiting calcium and proton pumps in isolated
sarcoplasmic reticulum or mitochondria (Olson & Mushlin,
1990). Other in vitro studies have questioned the greater
toxicity of DOXol vs DOX, and have shown that DOX may
be superior to DOXol in inducing inappropriate activation of
the calcium release channel (ryanodine receptor) of sarco-
plasmic reticulum (Zucchi et al., 2000), or in producing
oxygen free radicals through one-electron redox cycling of the
*Author for correspondence; E-mail: gminotti@unich.it

1272
G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity
Figure 1 Structures of the anthracyclines used in this study. The arrows indicate epimerization at C-4 in daunosamine (EPI); lack
of a methoxy group at C-4 in the aglycone, and intercalation of 2,6-dideoxy-L-fucose between the aglycone and daunosamine (MEN
10755 and MEN 11609); valeryl esteri®cation at C-14 (MEN 11609).
quinone moiety (Gervasi et al., 1986). Attempts to
recapitulate the di€erent actions of DOX and DOXol in a
unifying picture are made dicult by conceptual and
technical problems. The myocardial levels of unmetabolized
DOX often exceed those of DOXol several-fold (Mushlin et
al., 1993), making it dicult to establish whether DOXol
reaches sucient levels for competing or cooperating with
DOX in impairing contractility. Inhibitors of aldo/keto or
carbonyl reductases, like phenobarbital or rutin-type ¯avo-
noids, have been reported to mitigate cardiotoxicity induced
by DOX in rats or mice (Behnia & Boroujerdi, 1999; van
Acker et al., 1995); however, the decisive role of inhibiting
DOXol formation may have been confounded by several
other e€ects of barbiturates and ¯avonoids, including
interferences with free radicals formation and/or reactivity
(Almaas et al., 2000; Pietta, 2000). Other problems are posed
by the chemical characteristics of DOXol, which is more
wild type controls, these animals also presented higher
cardiac levels of the anthracycline (Forrest et al., 2000). This
raises the possibility that cardiotoxicity was enhanced not
only by a facilitated formation and action of DOXol, but
also by a reduced elimination of the anthracycline molecule
from the heart.
Recent studies in cytosolic fractions from human myo-
cardial samples have allowed us to demonstrate that certain
analogues of DOX may be inherently resistant to the action
of NADPH-dependent carbonyl reductases (Minotti et al.,
2000). In the case of epirubicin (EPI), an approved
anthracycline, the formation of its alcohol metabolite EPIol
was impaired by axial-to-equatorial epimerization of the
hydroxyl group at C-4 in daunosamine. In the case of MEN
10755, an investigational anthracycline, alcohol metabolite
formation was diminished by the lack of a methoxy group at
C-4 in the aglycone and by the presence of a disaccharide
moiety, obtained through intercalation of 2,6-dideoxy-L-
fucose between the aglycone and daunosamine (Figure 1).
Based upon these ®ndings we reasoned that comparing DOX
to EPI or MEN 10755 might be an alternative approach to
probing the role of secondary alcohol metabolites as
mediators of cardiotoxicity. We therefore designed experi-
ments in which rat isolated heart was exposed to DOX and
EPI or MEN 10755, or its valeryl ester MEN 11609 (Figure
1), and the decline in contractility (measured as inhibition of
post-rest contraction) was correlated to the cellular levels
attained by each anthracycline in its carbonyl or secondary
alcohol forms.
polar than DOX and consequently exhibits
a reduced
di€usion from extracellular ¯uids inside cardiomyocytes
(Danesi et al., 1988). This factor limits the interpretation of
studies in which isolated cardiac preparations were exposed
to DOXol to see how it compared to DOX in impairing
contractility (Mushlin et al., 1993). Increased polarity also
reduces the cardiac elimination of DOXol as compared to
DOX, causing a greater accumulation of the anthracycline
molecule inside cardiomyocytes. This factor became evident
in studies of transgenic mice bearing cardiac-restricted
overexpression of carbonyl reductases. While producing
DOXol and developing cardiotoxicity more rapidly than
British Journal of Pharmacology vol 134 (6)

G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity 1273
Methods
mined based on the sum ((parent anthracycline)+(secondary
alcohol metabolite)).
Anthracycline-dependent inhibition of myocardial
contractility
Anthracycline metabolism in human heart
Right ventricle strips (*10 mm long and 2 mm wide) from
male Sprague-Dawley rats (*400 ± 450 g) were mounted in a
5 ml organ bath containing 95% O2-5% CO2 Krebs solution,
378C. These preparations were subjected to continuous
electrical-®eld stimulation with square wave pulses of thresh-
old voltage at 2.5 Hz, using a pair of platinum electrodes.
Resting tension was adjusted to 0.5 g and the isometric force
of contraction was continuously recorded by using an
isometric force transducer and a Basile 7050 polygraph
recorder (Varese, Italy). After 60 min equilibration, electrical
stimulation was stopped and resumed 30 s later. The
amplitude of the ®rst twitch recorded under such conditions
is hereafter de®ned as post-rest contraction (Hagane et al.,
1988). Reproducible post-rest contractions were obtained at
5 min intervals; thereafter, anthracyclines were included and
their e€ects on the amplitude of post-rest contractions were
determined up to 120 min, a time which was long enough to
unravel the e€ects of both carbonyl anthracyclines and
alcohol metabolites (Mushlin et al., 1993), while not
impairing contractility in control samples (evidenced by a
98% amplitude of post rest-contractions at the end of the
experiments). DOX, EPI and MEN 10755 were used at 30
and 100 mM; MEN 11609 was used at 30 mM due to solubility
limit. The e€ects of anthracyclines were quanti®ed based
upon reductions in the amplitude of post-rest contractions,
and were expressed as per cent of inhibition relative to
controls. At the end of experiments, the ventricle strips were
frozen and stored at 7808C until assays for drug content and
composition.
Where indicated, NADPH-dependent conversion of anthra-
cyclines to secondary alcohol metabolites was assayed in
cytosolic fractions of human myocardial samples, precisely as
reported (Licata et al., 2000; Minotti et al., 2000). Values
were expressed as nmol alcohol metabolite mg prot.71 4 h71
.
All myocardial samples derived from the lateral aspect of
excluded right atrium of patients undergoing aorto-coronary
bypass grafting (Minotti et al., 1995). The specimens were
routinely disposed of by the surgeons during cannulation
procedures; therefore, patients were not subjected to any
unjusti®ed or ethically unacceptable loss of tissue.
Statistical analyses
All values were means+s.e. of 3 ± 5 experiments. Di€erences
between two sets of data were analysed by one-tailed
unpaired Student's t-test; di€erences between 42 sets of
data were analysed by one-way ANOVA followed by
Bonferroni's test for multiple comparisons. Correlations
between inhibition of post-rest contractions and anthracycline
content and composition were determined by one-tailed
nonparametric Spearman analysis. Di€erences and correla-
tions were considered to be signi®cant when P50.05.
Drugs
DOX and EPI were obtained through the courtesy of Dr
Antonino Suarato (Chemistry Department, Pharmacia-Up-
john, Milan, Italy). MEN 10755 and MEN 11609 were
synthesized at the Chemistry Department of Menarini
Ricerche S.p.A. (Pomezia, Rome, Italy). Alcohol metabolites
were prepared by us after NaBH4 reduction of the side chain
carbonyl group (Minotti et al., 1995). All other chemicals
were from Sigma (St. Louis, MO, U.S.A.).
Assays for anthracyclines
Ventricle strips were washed in 2 ml of ice-cold 0.3 M NaCl,
pH 7.0, homogenized and extracted with a 4 fold excess of
(1 : 1) CHCl3/CH3OH. The organic phases were recovered,
pooled, and dried under vacuum. Dry residues were
dissolved in 200 ml of methanol and cleared of gross
particulates by low speed centrifugation. The supernatants
were vacuum-dried and residues were suspended in 20 ± 50 ml
of methanol for analysis by two dimensional TLC on
0.25 mm (20620 cm) Silica Gel F254 plates (Merck,
Darmstadt, Germany) (Minotti et al., 1995; 2000). Mobile
phases and Rf values were: CHCl3/CH3OH/CH3COOH/H2O
(80 : 20 : 14 : 6) in either dimension for separation of DOX
(0.61) and DOXol (0.43); CHCl3/CH3OH/CH3COOH
(80 : 20 : 12) in either dimension for separation of EPI
(0.50) and EPIol (0.27); CHCl3/CH3OH/CH3COOH/H2O
(80 : 20 : 20 : 6) in either dimension for separation of MEN
10755 (0.49), MEN 10755ol (0.35), MEN 11609 (0.72) and
MEN 11609ol (0.59). After identi®cation by co-chromato-
graphy against authentic standards, parent anthracyclines
and secondary alcohol metabolites were eluted and quanti-
®ed by ¯uorescence spectroscopy (Minotti et al., 1995).
Values were expressed as nmol g71 (wet weight). Other
metabolites, like 7-hydroxy- or deoxy- aglycones, usually
were absent or detected at negligible levels; therefore, unless
otherwise indicated, total anthracycline uptake was deter-
Results
Anthracycline uptake, metabolism, and negative
inotropism
DOX, EPI and MEN 10755 were evaluated at 30 and 100 mM,
concentrations that previous studies have shown to be
appropriate for characterizing the pharmacological behaviour
of anthracyclines in isolated cardiac preparations (Hagane et
al., 1988; Temma et al., 1993). Table 1 shows results obtained
at 30 mM. At this concentration DOX exhibited noticeable
levels of uptake, retention in its carbonyl form, and
conversion to DOXol. EPI exhibited signi®cantly higher
levels of uptake and retention than DOX, but the formation
of EPIol was similar to that of DOXol. This suggested that
the higher uptake and retention of EPI were balanced by the
intrinsic resistance of this anthracycline to the action of
carbonyl- or aldo/keto- reductases. While exhibiting di€erent
uptake and retention but comparable conversion to alcohol
metabolites, DOX and EPI proved similar in inhibiting post-
rest contractions, an established index of contractility that
British Journal of Pharmacology vol 134 (6)

1274
G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity
Table 1 Anthracycline content, composition, and negative
inotropism in rat right ventricle strips: studies at 30 mM
Table 2 Anthracycline content, composition, and negative
inotropism in rat right ventricle strips: studies at 100 mM
Carbonyl
anthracycline metabolite
Alcohol
Inhibition of
post-rest
formation contractions
Carbonyl
anthracycline metabolite
Alcohol
Inhibition of
post-rest
formation contractions
Uptake
(nmol g71
retention
(nmol g71
Uptake
(nmol g71
retention
(nmol g71
Drug
)
)
(nmol g71
)
(%)
Drug
)
)
(nmol g71
)
(%)
DOX
EPI
7.7+2{
15.4+4{{
4.8+0.9*
12.8+2**
0.9+0.2
2.9+0.8}
2.5+0.3}}
0.8+0.1
43+6{
50+4{{
6.5+2
DOX
EPI
MEN 10755
103+7{
107+5{{
37+4
98+7*
104+5**
36+4
5.4+0.5}
2.6+0.3}}
1+0.1{{
88+1{
73+3{{
18+1
MEN 10755 1.7+0.1
All values were determined at 2 h and analysed by one-way
ANOVA followed by Bonferroni's test for multiple compar-
isons. {P50.05 vs EPI and P50.05 vs MEN 10755;
{{P50.001 vs MEN 10755. *P50.01 vs EPI and P50.05
vs MEN 10755; **P50.001 vs MEN 10755. }P40.05 vs
EPI and P50.05 vs MEN 10755; }}P50.05 vs MEN 10755.
{P40.05 vs EPI and P50.001 vs MEN 10755; {{P50.001
vs MEN 10755.
All values were determined at 2 h and analysed by one-way
ANOVA followed by Bonferroni's test for multiple compar-
isons. {P40.05 vs EPI and P50.001 vs MEN 10755;
{{P50.001 vs MEN 10755. *P40.05 vs EPI and P50.001
vs MEN 10755; **P50.001 vs MEN 10755. }P50.01 vs
EPI and P50.001 vs MEN 10755; }}P50.05 vs MEN
10755. {P50.05 vs EPI and P50.001 vs MEN 10755;
{{P50.001 vs MEN 10755.
re¯ects the amount of calcium available in the sarcoplasmic
reticulum (Hagane et al., 1988). This suggested that alcohol
metabolites were important determinants of negative inotrop-
ism induced by anthracyclines. Interestingly, MEN 10755 was
found to produce even less alcohol metabolite than EPI, in
spite of the fact that MEN 10755 and EPI exhibited the same
resistance to carbonyl reduction when assessed in cell-free
systems (Minotti et al., 2000). The reduced formation of
MEN 10755ol vs EPIol was attributable to an impaired
uptake of MEN 10755 in the ventricle strips, probably
re¯ecting sterical interferences caused by the presence of a
disaccharide moiety in the anthracycline molecule. Because of
multiple reductions in uptake, retention, and alcohol
metabolite formation, MEN 10755 inhibited post-rest con-
tractions by as little as *6% as compared to 43 or 50% in
the case of DOX or EPI, respectively. Table 2 shows results
obtained at 100 mM. This concentration allowed DOX to
reach much higher levels of uptake, retention in its carbonyl
form, and conversion to DOXol. EPI was able to reach the
same levels of uptake and retention, but failed to produce
signi®cantly more EPIol than it did at 30 mM, in keeping with
its resistance to the action of carbonyl- aldo/keto- reductases.
Therefore, EPIol was signi®cantly lower than DOXol when
EPI and DOX were assessed at 100 mM. EPI was also slightly
but signi®cantly less e€ective than DOX in inhibiting post-
rest contractions, con®rming that alcohol metabolites were
important determinants of negative inotropism. Under
comparable conditions MEN 10755 still exhibited signi®-
cantly lower levels of uptake, retention, and conversion to
alcohol metabolite, and consequently inhibited post-rest
contractions much less e€ectively than DOX or EPI.
Collectively, the three anthracyclines impaired myocardial
contractility by an order of ecacy that was DOX=EPI 4
MEN 10755 at 30 mM, or DOX 4 EPI 4 MEN 10755 at
100 mM. This pattern of cardiotoxicity seemed to re¯ect the
levels of alcohol metabolites produced by each anthracycline
at the two concentrations. However, inhibition of post-rest
contractions by 30 ± 100 mM DOX, EPI, or MEN 10755, was
shown to correlate not only with anthracycline uptake and
conversion to alcohol metabolites, but also with cellular
retention of anthracyclines in their carbonyl forms (Figure 2).
This demonstrated that both carbonyl anthracyclines and
secondary alcohol metabolites were involved in impairing
contractility.
Role of secondary alcohol metabolites
We designed experiments to elucidate the role of alcohol
metabolites. For this purpose we used MEN 11609, an
analogue obtained by esterifying the side chain C-14 primary
alcohol of MEN 10755 with a valeryl residue (cfr. Figure 1).
When assessed in human cardiac cytosol, a model which we
have developed for screening the structural determinants of
anthracycline carbonyl reduction (Licata et al., 2000; Minotti
et al., 2000), 30 mM MEN 11609 was similar to equimolar
EPI or MEN 10755 in producing less alcohol metabolite than
DOX (nmol mg protein71 4 h71: DOXol, 0.47+0.06; EPIol,
0.23+0.04; MEN 10755ol, 0.25+0.03; MEN 11609ol,
0.27+0.05; n=3, P50.05 for EPIol, MEN 10755ol and
MEN 11609ol vs DOXol). These ®ndings showed that the
presence of a valeryl moiety did not a€ect the metabolic
behaviour of MEN 11609 as compared to authentic MEN
10755. However, the lipophilic character of the valeryl
residue was expected to improve di€usion of MEN 11609
in cardiomyocytes, favouring its greater conversion to alcohol
metabolite as compared to MEN 10755. Increased lipophi-
licity was expected to facilitate also the di€usion of
unmetabolized MEN 11609 from cardiomyocytes back to
the extracellular environment, while leaving the more polar
alcohol metabolite inside the cell. MEN 11609 was therefore
exploited to increase anthracycline uptake and alcohol
metabolite formation while also preventing retention of the
carbonyl form of the drug, making it possible to highlight the
role of alcohol metabolites in impairing contractility.
Incubation of ventricle strips with MEN 11609 resulted in
formation of cellular levels of both unmodi®ed drug and
MEN 11609ol (3.9+0.3 and 2.1+0.5 nmol g71 2 h71
,
respectively; n=4). Minor amounts of MEN 11609 were
recovered also in the form of MEN 10755 and MEN 10755ol
(1.2+0.2 and 0.6+0.1 nmol g71 2 h71, respectively, n=4),
re¯ecting the activity of esterases which removed the valeryl
moiety. Whereas direct measurements of uptake were
precluded by continuous equilibration of MEN 11609 across
the plasma membrane of cardiomyocytes, carbonyl anthracy-
cline retention could be determined from the sum ((MEN
11609)+(MEN 10755)), and alcohol metabolite formation
could be determined from the sum ((MEN 11609ol)+(MEN
10755ol)). These values were compared to those determined
for authentic MEN 10755. As shown in Table 3, 30 mM MEN
British Journal of Pharmacology vol 134 (6)

G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity 1275
between 30 mM MEN 11609 and equimolar MEN 10755
showed that the cardiotoxicity of an anthracycline could be
enhanced by facilitating its conversion to alcohol metabolite
while not promoting overt increase in retention of its
carbonyl form. This demonstrated that alcohol metabolites
were independent mediators of anthracycline-induced cardi-
otoxicity. On the other hand, comparisons between 30 mM
MEN 11609 and 100 MEN 10755 showed that the
cardiotoxicity of an anthracycline did not decrease, but
actually tended to increase 2 fold when a loss in retention of
*31 nmol of carbonyl anthracycline g71 was balanced by a
net increase of as little as *1.7 nmol of alcohol metabolite
g71 (cfr. Table 3). This demonstrated that alcohol metabolites
were *20 ± 40 times more e€ective than their parent
compounds in impairing contractility. Accordingly, the
negative inotropism of 30 mM MEN 11609 vs 30 or 100 mM
MEN 10755 was shown to correlate in a highly signi®cant
manner with the levels of alcohol metabolites but not of
carbonyl anthracyclines (Figure 3).
Role of carbonyl anthracyclines
To characterize the role of carbonyl anthracyclines we re-
analysed experiments in which the drugs produced comparable
cellular levels of their alcohol metabolites but exhibited
di€erent degrees of inhibition of post-rest contractions. One
opportunity was o€ered by comparisons between 30 and
100 mM MEN 10755. As shown in Figure 4, these concentra-
tions of MEN 10755 resulted in very similar levels of
formation of MEN 10755ol inside cardiomyocytes (0.8+0.1
vs 1+0.1 nmol g71 2 h71), yet 100 mM MEN 10755 was
signi®cantly more e€ective in inhibiting post-rest contractions,
at least when assessed by unpaired Student's t-test for limited
comparisons between two groups of data (P50.001). Because
100 mM MEN 10755 also exhibited a much higher level of
retention in its carbonyl form, these observations suggested
that carbonyl anthracyclines were able to induce cardiotoxicity
in addition to that caused by alcohol metabolites. Nonetheless,
it was worth noting that the enhanced negative inotropism of
100 vs 30 mM MEN 10755 required the accumulation of as
many as *36 nmol of carbonyl anthracycline over *1 nmol
of alcohol metabolite, con®rming that carbonyl anthracyclines
were much less toxic than alcohol metabolites. Additional
information was obtained by multiple comparisons within a
broader panel of anthracyclines (30 mM MEN 11609, 30 mM
DOX, and 30 or 100 mM EPI) (Figure 5). These anthracyclines
produced fully comparable levels of alcohol metabolites
(range: 2.5 ± 2.9 nmol g71 2 h71). In the case of 30 mM MEN
11609 and DOX, comparable levels of alcohol metabolites
coincided with comparable levels of retention of carbonyl
anthracyclines (5.0+0.4 and 4.8+0.9 nmol g71 2 h71), and of
inhibition of post-rest contractions (35+8 vs 43+6%,
P40.05). Thirty mM EPI exhibited levels of carbonyl
anthracycline retention that were *2.5 times higher than
those of 30 mM MEN 10755 or DOX (12.8+1.6 nmol g71
Figure 2 Inhibition of post-rest contractions by DOX, EPI or MEN
10755: Correlations with uptake, carbonyl anthracycline retention,
and formation of alcohol metabolites. Values were taken from
individual experiments at 30 or 100 mM anthracyclines.
11609 produced cellular levels of carbonyl anthracycline
which were *5 times higher than those produced by 30 mM
MEN, but *7 times lower than those produced by 100 mM
MEN 10755. Such di€erences were statistically signi®cant in
the case of 30 mM MEN 11609 vs 100 mM MEN 10755, but
not in the case of 30 mM MEN 11609 vs 30 mM MEN 10055.
At the same time 30 mM MEN 11609 produced cellular levels
of its alcohol metabolite that exceeded *3 times, in a
statistically signi®cant manner, those produced by either 30
or 100 mM MEN 10755. Such greater levels of conversion to
alcohol metabolite, but not of retention in its carbonyl form,
enabled 30 mM MEN 11609 to inhibit post-rest contractions
*5 or *2 times more e€ectively than 30 or 100 mM MEN
10755, respectively, although the increase in negative
inotropism of 30 mM MEN 11609 vs 100 mM MEN 10755
did not approach a level of statistical signi®cance when
analysed within multiple comparisons (cfr. Table 3). The
implications of these ®ndings were several fold. Comparisons
2 h71), but these levels were insucient to produce
a
signi®cantly increased inhibition of post-rest contractions
(50+4%, P40.05 vs 30 mM MEN 11609 or DOX). Only
100 mM EPI proved to induce a signi®cant exacerbation of
negative inotropism (P50.01 vs the other anthracyclines), but
its enhanced toxicity seemed to require cardiac retention of as
many as 104+5 nmol of carbonyl anthracycline g71 2 h71
.
British Journal of Pharmacology vol 134 (6)

1276
G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity
Table 3 Comparisons between MEN 10755 and MEN 11609
Carbonyl
anthracycline
retention
Alcohol
metabolite
formation
Inhibition of
post-rest
contractions
(%)
Drug
mM
(nmol g71
)
(nmol g71
)
MEN 11609
MEN 10755
30
30a
5+0.4{
0.9+0.2{{
36+4
2.7+0.5*
0.8+0.1**
1+0.1
35+8}
6.5+2}}
18+1
MEN 10755 100b
All values were determined at 2 h and analysed by one-way
ANOVA followed by Bonferroni's test for multiple compar-
isons. a,bValues taken from Tables 1 and 2, respectively.
{P40.05 vs 30 mM MEN 10755 and P50.001 vs 100 mM
MEN 10755; {{P50.001 vs 100 mM MEN 10755. *P50.05
vs 30 mM MEN 10755 and P50.05 vs 100 mM MEN 10755;
**P40.05 vs 100 mM MEN 10755. }P50.05 vs 30 mM MEN
10755, and P40.05 vs 100 mM MEN 10755; }}P40.05 vs
100 mM MEN 10755.
These observations con®rmed that an accumulation of
carbonyl anthracyclines over secondary alcohol metabolites
resulted in an increased inhibition of post-rest contractions.
However, the observation that such greater negative inotrop-
ism required the accumulation of *100 nmol of carbonyl
anthracycline over *2.5 nmol of secondary alcohol metabo-
lite also con®rmed that the former was approximately *40
times less toxic than the latter.
Figure 3 Inhibition of post-rest contractions by MEN 11609 or
MEN 10755: Correlations with carbonyl anthracycline retention, and
formation of alcohol metabolites. Values were taken from individual
experiments at 30 or 100 mM anthracyclines. See also text and Table 3.
Discussion
This study lends support to the concept that secondary alcohol
metabolites may be important determinants of cardiotoxicity
induced by anthracyclines. This is indicated by several lines of
evidence. First, we have shown that 30 mM EPI exhibited higher
levels of uptake and retention than equimolar DOX in ventricle
strips, yet it did not induce a higher inhibition of post rest-
contractions. This was due to the fact that EPI was inherently
resistant to the action of carbonyl- or aldo/keto-reductases, and
consequently produced as much metabolite as DOX in spite of
its improved availability to cellular reductases (cfr. Table 1).
Second, we have shown that 100 mM DOX and EPI exhibited
comparable levels of uptake and retention, but DOX was
signi®cantly more e€ective than EPI at inhibiting post-rest
contractions. This was due to the fact that raising anthracy-
clines from 30 to 100 mM enhanced the conversion of DOX to
DOXol but had no e€ect in increasing conversion of EPI to
EPIol (cfr. Table 2). Third, we have shown that MEN 10755,
which is similar to EPI in its intrinsic resistance to alcohol
metabolite formation, was less toxic than EPI at both 30 and
100 mM. This was due to the fact that MEN 10755 exhibited
also an impaired uptake in the ventricle strips, a factor which
diminished formation of MEN 10755ol below the levels
attained by EPIol (cfr. Tables 1 and 2). The cardiotoxicity of
MEN 10755 was therefore enhanced by chemical modi®ca-
tions, like valeryl esteri®cation in MEN 11609, which increased
its uptake and conversion to alcohol metabolite. Our results
also demonstrate that the cardiotoxicity of secondary alcohol
metabolites cannot be attributed to increased polarity and
consequent retention of a larger pool of the anthracycline
molecule inside cardiomyocytes. In fact, the greater negative
inotropism of 30 mM MEN 11609 vs 30 or 100 mM MEN 10755
was observed under conditions in which a facilitated formation
Figure 4 E€ects of increasing carbonyl anthracyclines over secondary
alcohol metabolites on inhibition of post-rest contractions: Compar-
isons between 30 and 100 mM MEN 10755. Based on data in Tables 1
and 2, and analysed by unpaired Student's t-test. Values without
vertical bars have s.e. within symbols. See also text for explanations.
of the alcohol metabolite was accompanied by either marginal
increase or signi®cant decrease in retention of the carbonyl
form of the anthracycline, respectively, making negative
inotropism eventually correlate with the levels of alcohol
metabolites but not of their parent compounds (cfr. Table 3 and
Figure 3).
British Journal of Pharmacology vol 134 (6)

G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity 1277
inducing cardiotoxicity. The di€erential actions of secondary
alcohol metabolites in cancer cells or cardiomyocytes are
therefore exploitable for designing anthracyclines which form
fewer amounts of such metabolites and consequently exhibit
equal or improved antitumour activity, while also inducing
less severe cardiac toxicity. Clinical studies have con®rmed
that EPI, characterized by an intrinsic resistance to
conversion to its alcohol metabolite, retains good antitumour
activity but causes heart failure only after administration of
higher cumulative doses than DOX (Weiss, 1992). Similarly,
preclinical studies have shown that MEN 10755, whose
conversion to MEN 10755ol is limited also by a reduced
uptake in cardiac tissue, induces less severe cardiotoxicity
when given chronically to rats (Cirillo et al., 2000), but
exhibits greater activity than DOX in tumour cell lines or
human tumour xenografts (Arcamone et al., 1997; Pratesi et
al., 1998). Recent studies in athymic mice xenografted with
human ovarian cancer and treated with a single i.v. dose of
anthracyclines have shown that MEN 10755 exhibited a
reduced uptake not only in normal tissues but also in the
tumour, yet it retained greater therapeutic ecacy than DOX
(Gonzalez-Paz et al., 2001). This suggests that MEN 10755 is
intrinsically superior to DOX in inducing topoisomerase II
inhibition and apoptosis in cancer cells (Arcamone et al.,
1997; Pratesi et al., 1998), consistent with its impaired
conversion to an alcohol metabolite exhibiting reduced
activities in these settings.
Figure 5 E€ects of increasing carbonyl anthracyclines over second-
ary alcohol metabolites on inhibition of post-rest contractions:
Comparisons between 30 mM MEN 11609 or DOX, and 30 ±
100 mM EPI. Based on data in Tables 1 ± 3, and analysed by one-
way ANOVA followed by Bonferroni's test for multiple comparisons.
Values without vertical bars have s.e. within symbols. See also text
for explanations.
Anthracycline-induced cardiotoxicity is a multifactorial
process in which also the carbonyl forms of the drugs can
contribute to impair contractility. This information has been
obtained by comparing groups of anthracyclines which
produced comparable cellular levels of their alcohol metabo-
lites but exhibited di€erent negative inotropism and retention
in their carbonyl forms (cfr. Figures 4 ± 5). However, a
quantitative analysis of such comparisons suggests that
carbonyl anthracyclines may be *40 times less active than
their alcohol metabolites in inhibiting post-rest contractions.
The same approximation was reached in the experiments in
which MEN 11609 was exploited for the opposite purpose of
increasing the levels of alcohol metabolite formation while
not facilitating retention of the carbonyl anthracycline.
Understanding that anthracycline-induced cardiotoxicity is a
multifactorial process, involving carbonyl and alcohol
species, has several important implications. On the one hand,
this picture helps to explain how the negative inotropism of
30 ± 100 mM DOX, EPI, and MEN 10755 correlates with the
cardiac levels of both carbonyl anthracyclines and secondary
alcohol metabolites. On the other hand, the much greater
toxicity of alcohol metabolites vs carbonyl anthracyclines
helps to explain how the metabolites always contribute to
impair contractility, even in the presence of a *20 ± 40 fold
excess of their parent compounds (e.g., in ventricle strips
exposed to 100 mM anthracyclines) (cfr. Table 2 and Figure
2).
In conclusion, we have shown that secondary alcohol
metabolites are important mediators of cardiotoxicity in-
duced by anthracyclines, and that a unique combination of
reduced uptake and alcohol metabolite formation might
improve the cardiac safety of MEN 10755 in comparison to
DOX or EPI. According to in vitro studies, the cardiotoxicity
of MEN 10755 would be further reduced by the fact that
disaccharide secondary alcohol metabolites are less reactive
than DOXol or EPIol toward cytoplasmic aconitase (Minotti
et al., 2000). This peculiar aspect does not seem to have
surfaced in the present study; in fact, comparisons between
30 mM DOX, EPI, and MEN 11609, which produced
comparable cellular levels of alcohol metabolites within a
narrow range of retention of their carbonyl forms, clearly
showed that the three anthracyclines were similar in
inhibiting post-rest contractions (cfr. Figure 5). The reduced
reactivity of disaccharide metabolites might nonetheless
surface, and prove useful to further prevent cardiotoxicity,
if anthracyclines were given chronically and cytoplasmic
aconitase were needed to counteract alterations of iron
homeostasis induced by the presence of cancer and
in¯ammatory reactions (Torti & Torti, 1994; Minotti et al.,
1996). Whereas the mechanisms of alcohol metabolite-
induced toxicity remain open to debate (Minotti et al.,
1999), the results described in this study, and previous
evidence for a higher antitumour activity of MEN 10755 vs
DOX, anticipate that MEN 10755 might exhibit a better
therapeutic index than other approved anthracyclines.
An intriguing feature of secondary alcohol metabolites is
that they do not always contribute to the antitumour activity
of anthracyclines (Ku€el et al., 1992); in some cases, an
excessive conversion to secondary alcohol metabolites
actually diminishes the antitumour ecacy of anthracyclines
(Gonzalez et al., 1995; Ax et al., 2000). Here we have
demonstrated that secondary alcohol metabolites may be
several fold more e€ective than their parent anthracyclines in
This work was supported by AIRC; MURST COFIN '99; and
MURST ``Center of Excellence in Aging at the University of
Chieti'' (to G. Minotti).
British Journal of Pharmacology vol 134 (6)

1278
G. Minotti et al
Anthracyclines, secondary alcohol metabolites, and cardiotoxicity
References
ALMAAS, R., SAUGSTAD, O.D., PLEASURE, D. & ROOTWELT, T.
(2000). E€ect of barbiturates on hydroxyl radicals, lipid
peroxidation, and hypoxic cell death in human NT2-N neurons.
Anesthesiology, 92, 764 ± 774.
ARCAMONE, F., ANIMATI, F., BERETTONI, M., BIGIONI, M.,
CAPRANICO, G., CASAZZA, A.M., CASERINI, C., CIPOLLONE,
A., DE CESARE, M., FRANCIOTTI, M., LOMBARDI, P., MADAMI,
A., MANZINI, S., MONTEAGUDO, E., POLIZZI, D., PRATESI, G.,
RIGHETTI, S.C., SALVATORE, C., SUPINO, R. & ZUNINO, F.
(1997). Doxorubicin disaccharide analogue: apoptosis-related
improvement of ecacy in vivo. J. Natl. Cancer Inst., 89, 1217 ±
1223.
AX, W., SOLDAN, M., KOCH, L. & MASER, E. (2000). Development of
daunorubicin resistance in tumour cells by induction of carbonyl
reduction. Biochem. Pharmacol., 59, 293 ± 300.
BEHNIA, K. & BOROUJERDI, M. (1999). Inhibition of aldo-keto
reductases by phenobarbital alters metabolism, pharmacoki-
netics and toxicity of doxorubicin in rats. J. Pharm. Pharmacol.,
51, 1275 ± 1282.
CIRILLO, R., SACCO, G., VENTURELLA, S., BRIGHTWELL, J.,
GIACHETTI, A. & MANZINI, S. (2000). Comparison of doxor-
ubicin- and MEN 10755-induced long term progressive cardio-
toxicity in the rat. J. Cardiov. Pharmacol., 35, 100 ± 108.
DANESI, R., PAPARELLI, A., BERNARDINI, N. & DEL TACCA, M.
(1988). Cyto¯uorescence localization and disposition of doxor-
ubicin and doxorubicinol in rat cardiac tissue. Eur. J. Cancer
Clin. Oncol., 24, 1123 ± 1131.
FORREST, G.L., GONZALEZ, B., TSENG, W., LI, X. & MANN, J.
(2000). Human carbonyl reductase overexpression in the heart
advances the development of doxorubicin-induced cardiotoxicity
in transgenic mice. Cancer Res., 60, 5158 ± 5164.
MINOTTI, G., CAVALIERE, A.F., MORDENTE, A., ROSSI, M.,
SCHIAVELLO, R., ZAMPARELLI, R.
& POSSATI, G.F. (1995).
Secondary alcohol metabolites mediate iron delocalization in
cytosolic fractions of myocardial biopsies exposed to anticancer
anthracyclines. J. Clin. Invest., 95, 1595 ± 1605.
MINOTTI, G., LICATA, S., SAPONIERO, A., MENNA, P., CALAFIORE,
A.M., DI GIAMMARCO, G., LIBERI, G., ANIMATI, F., CIPOL-
LONE, A., MANZINI, S. & MAGGI, C.A. (2000). Anthracycline
metabolism and toxicity in human myocardium: comparisons
between doxorubicin, epirubicin, and
a
a novel disaccharide
analogue with reduced level of formation and [4Fe-4S]
reactivity of its secondary alcohol metabolite. Chem. Res.
Toxicol., 13, 1336 ± 1341.
MINOTTI, G., MANCUSO, S., FRUSTACI, A., MORDENTE, A.,
SANTINI, S.A., CALAFIORE, A.M., LIBERI, G. & GENTILONI, N.
(1996). Paradoxical inhibition of cardiac lipid peroxidation in
cancer patients treated with doxorubicin. J. Clin. Invest., 96,
650 ± 661.
MINOTTI, G., RECALCATI, S., LIBERI, G., CALAFIORE, A.M.,
MANCUSO, C., PREZIOSI, P. & CAIRO, G. (1998). The secondary
alcohol metabolite of doxorubicin irreversibly inactivates
aconitase/iron regulatory protein-1 in cytosolic fractions from
human myocardium. FASEB J., 12, 541 ± 551.
MUSHLIN, P.S., CUSACK, B.J., BOUCEK, R.J. JR, ANDREJUK, T., LI,
& OLSON, R.D. (1993). Time-related increases in cardiac
X.
concentrations of doxorubicinol could interact with doxorubicin
to depress myocardial contractile function. Br. J. Pharmacol.,
110, 975 ± 982.
OLSON, R.D. & MUSHLIN, P.S. (1990). Doxorubicin cardiotoxicity:
analysis of prevailing hypotheses. FASEB J., 4, 3076 ± 3086.
PIETTA, P.G. (2000). Flavonoids as antioxidants, J. Nat. Prod., 63,
1035 ± 1042.
GERVASI, P.G., AGRILLO, M.R., CITTI, L., DANESI, R.
& DEL
TACCA, M. (1986). Superoxide anion production by adriamycinol
from cardiac sarcosomes and by mitochondrial NADH dehy-
drogenase. Anticancer Res., 6, 1231 ± 1235.
PRATESI, G., DE CESARE, M., CASERINI, C, PEREGO, P., DAL BO, L.,
POLIZZI, D., SUPINO, R., BIGIONI, M., MANZINI, S., IAFRATE,
E., SALVATORE, C., CASAZZA, A., ARCAMONE, F. & ZUNINO, F.
(1998). Improved ecacy and enlarged spectrum of activity of a
novel anthracycline disaccharide analogue of doxorubicin
against human tumor xenografts. Clin. Cancer Res., 4, 2833 ±
2839.
TEMMA, K., AKERA, T., CHUGUN, A., KONDO, H., HAGANE, K. &
HIRANO, S. (1993). Comparison of cardiac actions of doxor-
ubicin, pirarubicin and aclarubicin in isolated guinea-pig heart.
Eur. J. Pharmacol., 234, 173 ± 181.
GONZALEZ, B., AKMAN, S., DOROSHOW, J., RIVERA, H., KAPLAN,
W.D. & FORREST, G.L. (1995). Protection against daunorubicin
cytotoxicity by expression of a cloned human carbonyl reductase
cDNA in K562 leukemia cells. Cancer Res., 55, 4646 ± 4650.
GONZALEZ-PAZ, O., POLIZZI, D., DE CESARE, M., ZUNINO, F.,
BIGIONI, M., MAGGI, C.A., MANZINI, S. & PRATESI, G. (2001).
Tissue distribution, antitumour activity and in vivo apoptosis
induction by MEN 10755 in nude mice. Eur. J. Cancer, 37, 431 ±
437.
HAGANE, K., AKERA, T. & BERLIN, J.R. (1988). Doxorubicin :
Mechanism of cardiodepressant actions in guinea pigs. J.
Pharmacol. Exptl. Ther., 246, 655 ± 661.
KUFFEL, M.J., REID, J.M. & AMES, M.M. (1992). Anthracyclines and
their C-13 alcohol metabolites: growth inhibition and DNA
damage following incubation with tumor cell lines. Cancer
Chemother. Pharmacol., 30, 51 ± 57.
TORTI, F.M. & TORTI, S.V. (1994). Cytokines, iron homeostasis, and
cancer. Adv. Exp. Med. Biol., 354, 161 ± 170.
VAN ACKER, S.A., KRAMER, K., GRIMBERGEN, J.A., VAN DEN BERG,
D.J., VAN DER VIJGH, W.J. & BAST, A. (1995). Monohydrox-
yethylrutoside as protector against chronic doxorubicin-induced
cardiotoxicity. Br. J. Pharmacol., 115, 1260 ± 1264.
WEISS, R.B. (1992). The anthracyclines: will we ever ®nd a better
doxorubicin? Sem. Oncol., 19, 670 ± 686.
LICATA, S., SAPONIERO, A., MORDENTE, A. & MINOTTI, G. (2000).
Doxorubicin metabolism and toxicity in human myocardium:
role of cytoplasmic deglycosidation and carbonyl reduction.
Chem. Res. Toxicol., 13, 414 ± 420.
ZUCCHI, R., YU, G., GHELARDONI, S., RONCA, F. & RONCA-
TESTONI, S. (2000). E€ect of MEN 10755, a new disaccharide
analogue of doxorubicin, on sarcoplasmic reticulum Ca(2+)
handling and contractile function in rat heart. Br. J. Pharmacol.,
131, 342 ± 348.
MINOTTI, G., CAIRO, G.
& MONTI, E. (1999). Role of iron in
anthracycline cardiotoxicity : new tunes for an old song? FASEB
J., 13, 199 ± 212.
(Received August 7, 2001
Accepted September 4, 2001)
British Journal of Pharmacology vol 134 (6)