(The FASEB Journal. 1999;13:199-212.)
© 1999 FASEB
Role of iron in anthracycline cardiotoxicity: new tunes for an old song?
GIORGIO MINOTTIa , 1 ,
GAETANO CAIROb and
ELENA MONTIc
Departments of
a Pharmacology and Pharmacognosy, G. D'Annunzio University School of Pharmacy, Chieti;
b General Pathology, University of Milan School of Medicine, Milan; and
c Structural and Functional Biology, University of Insubria School of Sciences, Varese, Italy
 |
ABSTRACT
|
|---|
The clinical use of anticancer anthracyclines is limited by the
development of a distinctive and life-threatening form of
cardiomyopathy upon chronic treatment. Commonly accepted mechanistic
hypotheses have assigned a pivotal role to iron, which would act as a
catalyst for free radical reactions and oxidative stress. Although
perhaps involved in acute aspects of anthracycline cardiotoxicity, the
role of free radical-based mechanisms in long-term effects has been
challenged on both experimental and clinical grounds, and alternative
hypotheses independent of iron and free radicals have flourished. More
recently, studies of the role of C-13 hydroxy metabolites of
anthracyclines have provided new perspectives on the role of iron in
the cardiotoxicity of these drugs, showing that such metabolites can
impair intracellular iron handling and homeostasis. The present review
applies a multisided approach to the critical evaluation of various
hypotheses proposed over the last decade for the role of iron in
anthracycline-induced cardiotoxicity. The main goal of the authors is
to build a unifying pattern that would both account for hitherto
unexplained experimental observations and help design novel and more
rational strategies toward a much-needed improvement in the therapeutic
index of anthracyclines.Minotti, G., Cairo, G., Monti, E. Role of
iron in anthracycline cardiotoxicity: new tunes for an old song?
Key Words: mechanisms cardioprotection metallothionein DOX
 |
BACKGROUND
|
|---|
SINCE ITS ISOLATION from Streptomyces
peucetius some 30 years ago, doxorubicin (DOX; former generic
name, adriamycin)2 has proved remarkably useful in the
clinical management of carcinomas, sarcomas, and lymphomas.
Unfortunately, long-term treatment with DOX and other extractive or
pharmaceutically engineered anthracyclines is limited by acute and
chronic cardiotoxicity. Whereas the acute toxicity presents with
transient and clinically manageable arrhythmias and hypotension, the
chronic toxicity can insidiously evolve into congestive heart
failure, which is refractory to standard inotropic medications
(1)
. Chronic cardiomyopathy and heart failure may develop
any time after completion of anthracycline regimens, and occur more
frequently in patients given cumulative doses of DOX above 550
mg/m2 (1)
.
Accordingto the prevailing hypothesis, the cardiotoxicity of anthracyclines is
mediated by mechanisms that are distinct from those underlying the
antitumor effects of these drugs, i.e., DNA intercalation and
interference with the catalytic cycle of DNA topoisomerase II
(2)
. A major role in the development of cardiotoxicity has
been assigned to iron, presumably because this metal can catalyze free
radical reactions that overrule the antioxidant defenses of
cardiomyocytes. Elegant reviews and editorials in this
3-5)
and other journals (2
,
6-8
) have contributed to the popularity of such
hypothesis; however, there have also been reports showing that the
`iron and free radical hypothesis' of cardiotoxicity does not
withstand scrutiny according to some biochemical or pharmacological
criteria (4
, 9
, 10
). Hence, some
investigators have proposed mechanisms of cardiotoxicity that are
independent altogether of both iron and free radicals (4)
.
In an attempt to bridge the two extremes of this field, other studies
have maintained a role for iron but not for free radicals, suggesting
that anthracycline cardiotoxicity reflects disturbances in iron
homeostasis within cardiomyocytes rather than the outcome of
iron-catalyzed free radical injury (11
, 12
).
The aim of this review is to weigh biochemical and pharmacological
arguments for and against the aforementioned hypotheses.
 |
DOES IRON MEDIATE ANTHRACYCLINE CARDIOTOXICITY? STUDIES WITH
CHELATORS SEEM TO SAY YES
|
|---|
The possible involvement of iron in anthracycline-induced
cardiotoxicity first emerged in the middle 1970s, when a series of
nonpolar derivatives of EDTA, including a bis-ketopiperazine
codenamed ICRF-187 and subsequently given the nonproprietary name of
dexrazoxanewere shown to prevent the cardiac lesions and dysfunction
induced by anthracyclines in both isolated heart models and whole
animals (13)
. A role for iron has been maintained ever
since and has been made increasingly appealing by several pieces of
evidence. For example, electron paramagnetic studies have shown that
mitochondrial, nuclear, and microsomal reductases catalyze one-electron
addition to the quinone moiety of the tetracycline ring of
anthracyclines. This results in the formation of a semiquinone free
radical that readily regenerates the parent quinone by reducing
molecular oxygen to superoxide anion (O2·-) and its
dismutation product hydrogen peroxide (H2O2),
increasing these species above the physiologic levels generated during
normal aerobic metabolism (6)
. Both
O2·- and H2O2 have their
inherent toxicity, but cell damage ensues more rapidly when they react
with low molecular weight (low mol wt) iron (14)
. To
prevent toxic interactions of iron with physiologic or pathologic
sources of O2·- and H2O2,
cells have evolved ferritin, a 24 subunit protein in which iron is
safely stored as polynuclear ferric oxohydroxide (15)
.
Such a protective mechanism of iron segregation can be overcome by
O2·-, which is small enough to penetrate the
transprotein channels of ferritin and has a reduction potential lower
than that of ferritin-bound iron. It follows that
O2·- has both sterical and thermodynamic requisites
to reductively release iron from ferritin (16)
.
Anthracycline semiquinones are too large to penetrate ferritin;
however, they can release iron by virtue of electron-tunneling
mechanisms, which are relayed by redox sites located on the surface and
transprotein channels of ferritin (16)
. This latter
mechanism can actually be more efficient than that mediated by
O2·-; in fact, chemical studies have shown little or
no interference by superoxide dismutase (SOD) with the ability of DOX
to delocalize Fe(II) (17
, 18
). DOX can also
form iron complexes that redox-cycle between ferrous and ferric forms
by donating electrons to oxygen or by drawing reducing equivalents from
either glutathione or the side chain C-14 primary alcohol of the
anthracycline itself, respectively (3)
. These iron
complexes probably contribute to the redox effect of DOX on
cardiomyocytes.
On the whole, one-electron redox cycling of anthracyclines increases
the intracellular levels of both low mol wt Fe(II) and
H2O2, setting the stage for the formation of
more damaging species. On the one hand, low mol wt Fe(II) can decompose
H2O2 into highly reactive hydroxyl radicals
( · OH) [Fe(II) + H2O2
Fe(III) +
OH- + · OH]; on the other hand, Fe(II) itself can
be converted by H2O2 into ferryl
ions (FeIV=O) having the same reactivity as · OH
[Fe(II) + H2O2
FeIV=O +
H2O] (16)
. Both · OH and
FeIV=O can eventually cause cardiac dysfunction by
oxidizing lipids, proteins, and DNA, perhaps in concert with the redox
activity of DOXiron complexes (Fig. 1
).

View larger version (28K):
[in this window]
[in a new window]
|
Figure 1. The `iron and free radical hypothesis of cardiotoxicity'.
F/FH2, oxidized/reduced flavoproteins (e.g., NADH
dehydrogenase, NADPH cytochrome P450 reductase); LMW Fe(II), low
molecular weight Fe(II); · OH, hydroxyl radical;
FeIV=O, ferryl ion; DOX Fe, doxorubiciniron complex.
See text for details.
|
|
In most cells, the chances for Fe(II) to form · OH or
FeIV=O are kept to a minimum by the presence of detoxifying
enzymes that concur in reducing O2·- and
H2O2 to water (SOD, catalase, and glutathione
peroxidase). Consistent with the `iron and free radical hypothesis',
cardiomyocytes are ill equipped with these enzymes and thus may provide
the ideal scenario for the formation of · OH or
FeIV=O (5
, 6
). Moreover, the
earliest ultrastructural changes induced by DOX involve mitochondria
and sarcoplasmic reticulum, as would be expected based on the presence
in these organelles of NADH dehydrogenase or NADPH-cytochrome P450
reductase, respectively, which reduce DOX to semiquinone free radical
(4
, 6
). Dexrazoxane can protect against
oxidative damage by entering cardiomyocytes and by undergoing stepwise
hydrolysis of the two piperazine rings. This process releases one- and
two- open ring by-products that chelate free iron or divert it from
potentially toxic interactions with DOX (19
,
20
). Initial studies in laboratory animals showed that
dexrazoxane had unique pharmacokinetic, pharmacodynamic, and
toxicologic properties. First, it did not interfere with anthracycline
distribution, metabolism, or excretion (21)
, nor did it
appear to mitigate the antitumor potency of anthracyclines
(22)
. Second, it imposed relatively limited damage on
mitotically active tissues such as bone marrow, testes, and
gastrointestinal epithelia (23)
. Finally, it protected
against both acute and chronic cardiotoxicity in all the animal models
tested, ranging from mice to dogs or swine (8)
. These
findings were largely confirmed in clinical trials showing that
dexrazoxane could decrease the incidence or severity of anthracycline
cardiotoxicity without affecting objective tumor response
(24)
. At longer follow-up, dexrazoxane also allowed
patients to receive significantly greater cumulative doses of DOX
without increasing its cardiotoxicity, the only drawback being a
moderate to severe exacerbation of myelosuppression (25)
.
However, more recent studies have raised the possibility that
dexrazoxane decreases tumor response in breast cancer patients treated
with DOX (26)
. These clinical findings appear to validate
in vitro studies showing that dexrazoxane antagonizes the
apoptotic and antiproliferative effects of DOX (27
,
28
), presumably by interfering with
anthracyclinetopoisomerase II interactions (29)
. Such
interference is mediated by intact dexrazoxane but not by its
hydrolysis products, indicating that a mechanism(s) other than iron
chelation and prevention of free radical events may be involved
(30)
.
 |
DOES IRON MEDIATE ANTHRACYCLINE CARDIOTOXICITY BY CATALYZING
FREE RADICAL REACTIONS? STUDIES WITH ANTIOXIDANTS YIELD INCONCLUSIVE
EVIDENCE
|
|---|
The `iron and free radical hypothesis' of cardiotoxicity has
been tested with a variety of compounds endowed with antioxidant
properties. Thus, laboratory animals have been treated with vitamin E,
N-acetylcysteine, or coenzyme Q10, a nutritional cofactor
with the dual advantage of acting as an antioxidant while participating
in the mitochondrial electron transport chain involved in ATP
production. All these compounds have yielded encouraging results in
murine models of acute or chronic anthracycline cardiotoxicity, as
indicated by suppression of biochemical indices of oxidative damage,
e.g., lipid peroxidation (8)
. Excellent results have
likewise been obtained with probucol, a compound initially developed as
a lipid-lowering drug. Due to the presence of two phenolic groups in
its molecule, probucol inhibits lipid peroxidation and affords
protection in a mouse model of chronic cardiotoxicity without affecting
the antitumor efficacy of the anthracycline (31)
.
Moreover, probucol can provide protection by increasing intracardiac
levels of SOD, catalase, and glutathione peroxidase, both per se and in
combination with DOX (32)
.
One additional strategy to probe and possibly counteract the role of
free radicals in small animal models of anthracycline cardiotoxicity
has relied on the use of low mol wt spin trapping agents such as
N-tert-butyl-
-phenylnitrone (PBN),
-(4-pyridyl-1-oxide)-N-tert-butylnitrone (POBN), and
5,5-dimethyl-1-pyrroline N-oxide (DMPO). The rationale for using these
compounds has been twofold. The ability of spin traps to form stable
adducts with free radicals in intact tissues or cell-free systems, thus
allowing their detection by electron paramagnetic spectroscopy,
strongly suggested that they might prevent anthracycline cardiotoxicity
if free radicals were involved in the process. On the other hand, spin
traps can intercept both oxygen- and carbon-centered radicals, and thus
may prove to be more versatile and effective protectants than other
antioxidants. As expected from these biochemical premises, spin traps
were found to protect atria and whole heart preparations from
DOX-induced mechanical dysfunction (33)
. The degree of
protection achieved has been found to correlate with the lipophilicity
of these compounds, being highest with PBN, followed by POBN and DMPO
(CHCl3/H2O partition coefficients at 24°C:
199.0, 1.19, and 0.13, respectively) (34)
. Comparative
studies of DMPO vs. PBN have also shown that the former distributes
almost exclusively within the cytosolic compartment, whereas PBN can
also be found in mitochondria. The higher efficacy of PBN, subsequently
confirmed by in vivo studies (35)
, could thus
be explained by its ability to achieve substantial levels at the very
sites where DOX bioactivation initiates a cascade of free radical
reactions.
Research groups have recently readdressed the role of free radicals by
evaluating the in vivo cardiotoxicity of DOX in transgenic
mice overexpressing antioxidant defense systems. Such studies focused
mainly on myocardial overexpression of catalase (36)
and
mitochondrial MnSOD (37)
or the thiol-rich nonenzymatic
protein metallothionein (MT), which has been shown to effectively
scavenge · OH and to act as an iron chelator in vitro
(38, and references therein). The transgenic mouse model was designed
to elude the shortcomings due to poor cell penetration of exogenously
administered proteins and to guarantee high catalase, MnSOD, and MT
activity inside the cardiomyocytes. All three approaches
proved successful in reducing acute myocardial damage by DOX, as
assessed by ultrastructural lesions, intracellular enzyme release, and
functional impairment. Although the transgenic approach may be
impractical for clinical purposes, it clearly indicates that stable
increases in intramyocardial levels of antioxidant proteins may
effectively protect the heart from DOX, at least in small animal models
of anthracycline cardiotoxicity.
Whereas the effects of antioxidants in rats and mice or in in
vitro models of cardiotoxicity have lent credibility to the free
radical hypothesis, similar studies in large animals have produced
contrasting notions, especially when chronic anthracycline regimens
were adopted. Although not all of the cardioprotective strategies
outlined above have been tested in large animal models of anthracycline
cardiotoxicity, carefully designed experiments with N-acetylcysteine or
vitamin E have shown that neither compound would prevent or
significantly reduce cardiac lesions induced by chronic treatment of
dogs with DOX (23
, 39
). Similar negative
results have been obtained in clinical trials in which patients were
given vitamin E (40)
or N-acetylcysteine (41)
prior to and/or concomitant with DOX. In none of these trials could
antioxidants prevent cardiac dysfunction associated with multiple doses
of the anthracycline. In the case of vitamin E, an intensified regimen
was started as early as a week prior to DOX treatment, increasing the
serum levels of this antioxidant by six- to eightfold. Such vigorous
and pharmacokinetically optimized treatment was nonetheless
insufficient to prevent heart failure and ultrastructural changes at
endomyocardial biopsy (40)
. To better appraise these
negative findings, one should keep in mind that lesser increases in
vitamin E plasma levels are sufficient to prevent or decrease the
incidence of other cardiac events that presumably involve oxidative
stress, such as nonfatal myocardial infarction in patients with
coronary atherosclerosis (42)
, arrhythmias and
reinfarction after aortocoronary bypass (43)
, and fatal
infarction in smokers (44)
.
In summary, chelating iron with dexrazoxane mitigates both acute and
chronic cardiotoxicity, and such protection can be seen in patients as
well as in all the animal models tested. In contrast, antioxidant
interventions afford protection in small animals, but not in patients
or large animals at risk for chronic cardiomyopathy. This discrepancy
raises a potentially critical issue that needs to be addressed from
mechanistic and pathophysiologic viewpoints.
 |
DO ANTHRACYCLINES GENERATE TOO MANY RADICALS FOR THEM TO
MEDIATE CARDIOTOXICITY?
|
|---|
To explain the lack of protection by antioxidants, one might argue
that free radical processes are more a consequence than a cause of
tissue damage; hence, antioxidants would interfere with second-line
reactions having little relevance to the etiology of cardiomyopathy.
This issue has been brought into focus by several authors, especially
with regard to lipid peroxidation (45)
. Another
possibility is that anthracyclines form too many radicals, setting the
stage for radicalradical interactions that eventually extinguish
oxidative stress. For example, semiquinone free radicals might
terminate lipid peroxidation by preferentially interacting with lipid
radical species rather than with oxygen or iron (46)
.
Unproductive reactions may likewise ensue after
H2O2 formation or low mol wt Fe(II)
delocalization by semiquinones. Stoichiometric considerations indicate
that the simultaneous redox coupling of DOX semiquinone free radicals
with molecular oxygen and either ferritin or other newly identified
microsomal iron proteins generates a large excess of
H2O2 over Fe(II) (18)
. Whereas
Fe(II) can undoubtedly decompose some H2O2
to · OH under these conditions, the vast majority of
H2O2 would remain available to interact
with · OH and convert it into the less damaging
O2·- [H2O2 + · OH
H2O + O2·- + H+]
(47)
. Substantial amounts of · OH could thus be
diverted from reacting with lipids or other targets. One last
possibility is that oxidative damage proceeds through reactive species
that are different from · OH and cannot be formed when
anthracyclines generate a stoichiometric excess of
H2O2 over Fe(II). In particular, early studies
by Aust and co-workers (48
, 49
), and later
observations by several other investigators (reviewed in ref 16
), have
suggested that oxidative damage could be mediated by oxygen-bridged
Fe(II)-Fe(III) complexes or equally reactive perferryl ions
[Fe(II)O2/Fe(III)O2·-]. Both these
species are formed when some Fe(II) oxidizes with oxygen or
less than stoichiometric amounts of H2O2,
yieldingFe(II):Fe(III) ratios that impart unusual reactivity to
these ironoxygen complexes (16)
. Neither complex would
therefore be formed when semiquinone free radicals generate
H2O2 in excess of Fe(II), causing the oxidation
of too much Fe(II) and shifting the Fe(II)/Fe(III) ratio
toward the unreactive Fe(III) (16)
. Studies in microsomal
systems (50)
or cell cultures (51)
exposed to
DOX suggest that this may be the case. In these systems, lipid
peroxidation and other indices of damage were significantly enhanced by
addition of catalase or by genetic manipulations leading to catalase
overexpression, respectively, as would be expected if excess
H2O2 were negating moderate oxidation of Fe(II)
with oxygen and consequent formation of optimal Fe(II)/Fe(III) ratios.
Keeping these mechanisms in mind, one might conclude that antioxidants
cannot protect against cardiotoxicity for the very reason that
anthracyclines possibly prevent, rather than induce, oxidative damage.
Studies in isolated heart models or in whole animals (reviewed in ref 4
) appear to validate these conclusions, showing that cardiotoxicity
may sometimes develop in the absence of lipid peroxidation or other
indices of oxidative damage. Such a possibility is more difficult to
prove in patients, as both ethical and practical constraints preclude
biochemical measurements of oxidative stress in fine needle myocardial
biopsies. To overcome these problems, cardiac lipid peroxidation has
been assessed by collecting blood samples from the coronary sinus and
by measuring the plasma levels of lipid conjugated dienes and
hydroperoxides before and after DOX infusions. These studies have shown
that DOX does not increase but actually decreases the myocardial
release of conjugated dienes and hydroperoxides, apparently confirming
that a free radical-generating drug can paradoxically abolish oxidative
damage (10)
.
These new findings call for a critical reassessment of previous studies
supporting the oxidative nature of anthracycline cardiotoxicity,
especially regarding the transgenic mice models. For instance, one
would expect overexpression of MnSOD to increase the dismutation of
O2·- without significantly affecting iron
delocalization by semiquinone free radicals, thus increasing the
availability of H2O2 for reaction
with Fe(II) and formation of · OH or FeIV=O.
Nevertheless, cardiac damage is halted by MnSOD overexpression
(37)
, suggesting that 1) neither · OH
nor FeIV=O contribute to toxicity, and 2) tissue
damage probably is mediated by ironoxygen complexes, which become
redox-inactive when excess dismutation of O2·- to
H2O2 shifts Fe(II)/Fe(III) ratios toward
unreactive Fe(III). Similar considerations can be extended to the
catalase studies. In fact, cardiac lipid peroxidation was suppressed at
catalase expression levels approaching values 100-fold over the
wild-type; however, higher levels of catalase expression did not
prevent, but actually enhanced, lipid peroxidation (36)
.
While indicating that catalase overexpression provides a very narrow
window of protection, these studies indirectly confirm that free
radical injury is promoted when Fe(II) is `protected' from excessive
oxidation by H2O2, which presumably occurs at
very high levels of enzyme expression. On balance, the results obtained
with SOD- or catalase-overexpressing mice are open to debate and do not
provide unequivocal evidence that anthracycline cardiotoxicity is
indeed mediated by oxidative processes.
 |
IS ANTHRACYCLINE CARDIOTOXICITY INDEPENDENT OF IRON AND FREE
RADICALS?
|
|---|
The last decade has witnessed the identification of cardiotoxicity
mechanisms that do not implicate free radicals as causative agents.
Several studies along this line of research have shown that DOX can
activate signal transduction pathways that disrupt the cardiac gene
expression program. DOX decreases the intracardiac levels of mRNAs for
several important proteins and enzymes 52-55)
, which may
account for both morphologic and functional features of chronic
cardiotoxicity (Table 1
).The effects of DOX on mRNAs are independent of free radicals
(56)
and are considerably less evident at extracardiac
sites, possibly because such effects are mediated by transcriptional
regulatory proteins expressed selectively in cardiomyocytes (e.g., the
so-called cardiac adriamycin-responsive protein/CARP)
(57)
.
Alternative mechanisms of cardiotoxicity have implicated DOX
metabolites other than semiquinone free radicals. This is the case for
doxorubicinol (DOXol), a hydroxy metabolite formed upon two-electron
reduction of the C-13 carbonyl group in the side chain of
DOX:
DOXol is formed by cytosolic NADPH-dependent enzymes sharing
similarities with the multigene family of carbonyl and aldo-keto
reductases (58)
. From a pharmacokinetic standpoint, DOXol
accumulates in the heart of rats receiving multiple DOX injections;
such accumulation reflects intramyocardial DOX metabolism rather than
metabolite uptake from the bloodstream (4)
. This has
suggested that the delayed cardiotoxicity induced by chronic treatment
with DOX in laboratory animals may reflect a unique tendency of
cardiomyocytes to form and retain DOXol. This also seems to occur in
DOX-treated patients, as limited studies in postmortem tissues have
demonstrated that DOXol can be detected within cardiomyocytes well
after completion of anthracycline treatments (59)
.
Experiments in ex vivo samples, obtained from patients
undergoing aortocoronary bypass grafting, have confirmed that human
cardiomyocytes can reduce DOX to DOXol by virtue of cytosolic aldo-keto
reductases (11)
. From a pharmacodynamic viewpoint, the
C-13 secondary alcohol moiety of DOXol is redox-inactive toward oxygen
and thus is unable to initiate a free radical cascade. Nonetheless,
DOXol can potently and directly inhibit the
Ca2+-Mg2+ ATPase of sarcoplasmic reticulum, the
f0-f1 proton pump of mitochondria, and the
Na+-K+ ATPase and
Na+-Ca2+ exchanger of sarcolemma
(60
, 61
). DOXol can therefore affect
myocardial energy metabolism, ionic gradients, and Ca2+
movements, ultimately impairing cardiac contraction and relaxation.
By rationalizing anthracycline cardiotoxicity as a free
radical-independent disease, the two mechanisms described above may
explain why antioxidants cannot protect patients or large animals. At
the same time, these mechanisms are difficult to reconcile with the
protective efficacy of dexrazoxane (8)
or with the ability
of iron to potentiate anthracycline toxicity in cultured cardiomyocytes
(62)
, unless we assume that iron mediates DOX disruption
of gene expression or DOXol inhibition of key enzymes. Inasmuch as
these two possibilities remain unverified at the present time, one can
only conclude that anthracycline cardiotoxicity is a multifactorial
process involving both iron-independent and iron-mediated processes. It
would also appear that the role of iron cannot be confined to the
promotion of free radical events, as not only dexrazoxane, but also
antioxidants would protect patients or large animals if this were the
case. To better appraise the role of iron, one should probably address
the following questions. Can anthracyclines perturb the mechanisms
whereby iron is handled as an essential cofactor rather than as a free
radical catalyst? Further, can dexrazoxane antagonize the effect of DOX
on iron homeostasis and related cell functions?
 |
THE IRP-IRE MACHINERY: A KEY MECHANISM OF IRON HOMEOSTASIS AND
A NOVEL TARGET OF ANTHRACYCLINE CARDIOTOXICITY
|
|---|
Cellular iron homeostasis relies on the concerted action of the
transferrin receptor (TfR) and ferritin. The former mediates iron
uptake by internalizing iron-laden transferrin (63)
; the
latter stores iron in excess of the metabolic requirements of the cell
(64)
. Both TfR and ferritin can be regulated at a
transcriptional level (65)
, but a major regulatory
mechanism occurs posttranscriptionally 66-68)
. This
involves interaction of iron regulatory protein (IRP) -1 with specific
motifs in the 5' untranslated region of ferritin mRNA and in the 3' end
of TfR mRNA. These are the so-called iron-responsive elements (IREs),
highly conserved sequences with the potential to assume a stemloop
structure. When cells need iron, IRP-1 binds to IREs in TfR mRNA,
protecting the transcripts from degradation (68)
; at the
same time, IRP-1 interacts with IREs in ferritin mRNA and blocks
translation by sterically preventing the binding of translation
initiation factors (69)
. These divergent but coordinate
actions of IRP-1 cause iron uptake by TfR to prevail over iron
sequestration by ferritin, thus yielding a pool of low mol wt iron
readily available to meet metabolic requests. Conversely, high iron
levels decrease the IRE binding activity of IRP-1, leading to efficient
translation of ferritin mRNA and increased degradation of TfR mRNA.
Thus, IRP-1 links ferritin and TfR levels to cellular iron status.
Definition of the structure of IRP-1 provided a clue to understanding
how this protein can sense iron levels. IRP-1 is the cytosolic
counterpart of mitochondrial aconitase, the enzyme that isomerizes
citrate into isocitrate in the Krebs cycle by virtue of a catalytic
[4Fe-4S] cluster (70)
. When cells are iron depleted, the
[4Fe-4S] cluster is disassembled and the cytosolic aconitase switches
to IRP-1; when cells are iron repleted, the cluster is reconstituted
and IRP-1 switches back to aconitase 66-68)
(Fig. 2
).The -SH groups of several cysteine residues have been implicated in
RNA binding or cluster iron coordination, but only Cys437
appears to participate in both processes (71)
.
Anthracyclines have the potential to disrupt this simple but highly
efficient mechanism of iron homeostasis. In fact, recent studies from
our laboratories have shown that reconstitution of DOXol with cytosolic
fractions from human myocardial samples causes delocalization of low
mol weight Fe(II) from the [4Fe-4S] cluster of aconitase
(12)
. This reaction proceeds via redox mechanisms, which
regenerate DOX and form DOX-Fe(II) complexes as ultimate by-products;
the latter can, in turn, inhibit IRP-1 and prevent its physiologic
switch to aconitase (12)
. Sulfydryl reductants like
ß-mercaptoethanol cannot rescue either IRP-1 or aconitase activities,
and this implies that DOXFe(II) complexes can irreversibly modify
Cys437 to sulfinic/sulfonic derivatives that cannot be
converted back to -SH (12)
. Although in vitro
(68
, 72
) and in vivo
(73)
studies have implicated O2·- and
H2O2 as pathophysiologic modulators
of aconitase/IRP-1, neither SOD nor catalase appear to protect this
protein from the action of DOXol (12)
. These findings
confirm that C-13 metabolites can cause cardiac damage by virtue of
oxyradical-independent mechanisms.
Besides IRP-1, cells may contain an IRP-2 with similar binding affinity
and specificity to consensus IREs (68
, 74
).
However, IRP-2 cannot form [4Fe-4S] clusters and hence cannot sense
iron levels by switching between aconitase and IRE binding activities,
as does IRP-1. The mechanism whereby IRP-2 contributes to iron
homeostasis is regulated by its ubiquitination and proteasome-mediated
degradation (75)
. When iron is abundant, IRP-2 is
degraded; when iron is scarce, IRP-2 is protected from degradation and
can bind to IREs, leading to up-regulation or down-regulation of TfR or
ferritin mRNA translation, respectively. A unique 73 amino acid
sequence serves as the degradation domain of IRP-2 (75)
.
Currently available techniques do not always allow the separation of
IRP-1 from IRP-2 in human samples; hence, we do not know whether DOXol
inactivation of aconitase/IRP-1 is accompanied by a simultaneous loss
of IRP-2. However, binding of IRP-2 to mRNAs is mediated by cysteine
residues that are particularly susceptible to chemical modifications,
similar to the cysteines of IRP-1 (76)
. It is therefore
conceivable that the DOXFe(II) complexes that inactivate IRP-1 can do
the same with IRP-2.
What are the pathologic consequences of disrupting the IRP/IRE
machinery? Broadly speaking, one can expect that cardiomyocytes become
unable to sense iron levels and to coordinate iron trafficking between
transferrin receptor, ferritin, and the different sites of iron
metabolic use. Disruption of the IRP/IRE machinery might also affect
posttranscriptional processes and metabolic events not immediately
related to the TfRferritin pathway. This is suggested by the presence
of IRE motifs in the 5' regions of mRNAs for mitochondrial aconitase,
Drosophila succinate dehydrogenase, and erythroid
-aminolevulinate synthase (68)
, as well as in the 3'
region of mRNA for mammalian proton-coupled metal ion transporters
(e.g., DCT1/Nramp2) (77)
. The effect of IRP-1
modifications on cellular iron homeostasis has indeed been demonstrated
by several studies. Cells overexpressing iron-insensitive IRP-1 mutants
cannot withstand iron supplementation (78)
, and
inappropriately high IRP activity has been been found in duodenal
(79)
or reticuloendothelial cells (80)
from
iron overload-prone hemochromatosis patients. In contrast, low IRP
activity has been found in monocytes from patients with inflammation,
providing a mechanistic link between enhanced iron sequestration in the
reticuloendothelial system, reduced iron availability for
hematopoiesis, and persistent anemia in chronic diseases
(81)
.
By disrupting cardiac iron homeostasis, the decline in IRP activity
would also lead to inadequate reconstitution of several enzymes and
oxygen binding proteins that function by virtue of this metal
(cytochromes, catalase, lipoxygenases, myoglobin, etc.). Additional
consequences of DOXol/IRP interactions might be expected after
misplacement of iron ions at cellular sites critical to the
contractionrelaxation cycle. This is best exemplified by recent
studies showing that chronic anthracycline treatment desensitizes
sarcoplasmic reticulum to Ca2+-induced Ca2+
release, indicating a dysfunction of the
Ca2+-gated/ryanodine receptor-2 Ca2+ release
channel (82)
. Whereas anthracycline treatment definitely
decreases the levels of channel mRNA (55)
and
immunoreactive proteins (83)
, a corollary mechanism for
channel dysfunction has been identified in its sterical occupancy by
iron ions. In this context, Fe(II) is remarkably more inhibitory than
Fe(III), presumably because it has greater affinity for the cysteine
residues lining the channel (84)
. Collectively, these
studies suggest that Ca2+ channels and IRPs share similar
mechanisms of inactivation by Fe(II) and foretell that changes in the
IRP/IRE machinery may lead to anomalous iron movements and sterical
association with cell constituents other than storage proteins or
iron-dependent enzymes.
The `IRP/IRE hypothesis' of cardiotoxicity has two additional
convincing pros. First, the cardiotoxicity scores of structurally
distinct anthracyclines correlate better with the intracardiac
formation of C-13 metabolites than with systemic metabolism and
disposition (11)
. This suggests that the cardiotoxic
potential of anthracyclines might be reliably predicted from their
affinity for myocardial aldo-keto reductases, regardless of
drug-to-drug variability in extracardiac pharmacokinetics. Second, the
`IRE-IRP hypothesis' may help clarify the role of dexrazoxane from
novel standpoints. Dexrazoxane might protect cardiomyocytes by
competing with DOX for iron and preventing formation of those
drugiron complexes that inactivate aconitase/IRP-1 and, perhaps,
IRP-2, as previously hinted. In so doing, dexrazoxane can mitigate
disturbances in iron movements, metabolic use, and sterical association
with proteins or enzymes. In either case, dexrazoxane would act on
homeostatic processes rather than free radical reactions, possibly
explaining why antioxidant interventions are less protective than
chelation therapy against chronic cardiotoxicity. The IRE/IRP
hypothesis of cardiotoxicity nonetheless has its drawbacks. For
example, there have been sporadic reports of delayed cardiotoxic
effects in laboratory animals after a single DOX administration, in
spite of the fact that DOXol accumulates to a minimal extent under
these conditions and eventually disappears within a week after drug
treatment (85)
. Again, studies in laboratory animals have
shown that 5-iminodaunorubicin can form a C-13 hydroxy metabolite and
yet is less cardiotoxic than other anthracyclines (86
,
87
). Inasmuch as the C-13 derivatives from structurally
distinct anthracyclines have been shown to retain comparable iron
reactivity (11)
, it is unlikely that the reduced
cardiotoxicity of 5-iminodaunorubicin may be attributed to decreased
interactions of its hydroxymetabolites with aconitase/IRPs or other
targets. What is peculiar to 5-iminodaunorubicin is that its quinone
moiety has been chemically modified so as to generate much less
O2·- and H2O2
(86
, 87
). This suggests that anthracyclines
can be made less cardiotoxic by decreasing generation of oxyradicals
rather than of C-13 metabolites, in the face of conceptual efforts to
identify mechanisms alternative to oxidative stress! Although the
clinical use of 5-iminodaunorubicin has been precluded by severe
extracardiac toxicity, preclinical evidence accumulated with this drug
clearly argues against the DOXol and IRE/IRP hypothesis.
 |
ROLE OF IRON IN ANTHRACYCLINE CARDIOTOXICITY: IS THERE ROOM
FOR A UNIFYING HYPOTHESIS?
|
|---|
To elucidate the role of iron in anthracycline cardiotoxicity, we
have reviewed mechanisms that either implicate this metal as a catalyst
of oxidative damage or envision possible disturbances of its metabolic
use. Neither mechanism can satisfactorily encompass the experimental
and clinical aspects of anthracycline cardiotoxicity, and the reason
for this may be twofold. First, anthracycline cardiotoxicity is a
multifactorial event in which iron-based reactions are accompanied by
equally important iron-independent processes. Second, the involvement
of iron in oxidative damage or metabolic disturbances must be preceded
by one- or two-electron reduction of anthracyclines to semiquinone free
radicals or C-13 hydroxy derivatives, respectively. Inasmuch as these
metabolites are formed by different reductases at different cell sites,
the way iron intervenes in cardiotoxicity may vary with time and
reflect more than one mode of action.
A key to a unifying hypothesis might be to separate the mechanism(s) of
acute cardiotoxicity from those of chronic cardiomyopathy and to see
whether the former can be preparatory to the latter. As for the acute
cardiotoxicity, biochemical indices of oxidative damage and effective
protection by both antioxidants and dexrazoxane
indicate that, with the possible exception of few experimental
conditions (4)
, iron-catalyzed free radical injury is
important in this setting. As mentioned, this is more difficult to
prove in patients, who respond to acute DOX treatments with a
paradoxical inhibition of cardiac lipid peroxidation rather than with
stimulation (10)
. This is a unique situation, perhaps
because cancer patients present with spontaneously up-regulated
prooxidant tone and react to DOX metabolism and
H2O2 formation with unproductive processes that
annihilate · OH or ironoxygen complexes (cf. Section IV
and
ref 10
).
As anthracyclines are given chronically, depletion of antioxidant
defenses such as glutathione and related enzymes (4)
and
the consequent accumulation of H2O2 may
likewise divert · OH from oxidative damage or inactivate
ironoxygen complexes, thereby terminating free radical injury. The
transition from acute to chronic effects is therefore accompanied by a
decline of the protective efficacy of antioxidants but not of
dexrazoxane, perhaps because this is the time when iron moves from the
free radical arena to the homeostatic disorders associated with DOXol
formation and consequent inactivation of the IRE/IRP machinery
(Fig. 3
).Can this hypothesis incorporate a mechanistic link between acute and
chronic cardiotoxicity? We propose that it can. In fact, recent studies
in smooth muscle cells have shown that aldo-keto reductase induction
may represent a common response in tissues exposed to
H2O2 and certain by-products of lipid
peroxidation (e.g., 4-hydroxy-nonenal) (88)
. Whereas the
induction of these enzymes enables cells to detoxify lipid aldehydes
formed during the acute phase of cardiotoxicity, it also paves the road
to time-delayed reduction of DOX to DOXol and consequent inactivation
of IRPs. Thus, chronic cardiotoxicity might be `primed' by events
occurring at earlier stages, such as the redox cycling of
anthracyclines with oxygen (yielding H2O2) and
the promotion of iron-catalyzed damage (causing lipid peroxidation)
(cf. Fig. 3
). Direct evidence for anthracycline induction of aldo-keto
reductases is lacking, but the observation that intracardiac levels of
DOXol increase with time and repeated dosing of DOX strongly indicates
that this may be the case.

View larger version (39K):
[in this window]
[in a new window]
|
Figure 3. A unifying hypothesis for the role of iron in anthracycline
cardiotoxicity and possible interplay with iron-independent mechanisms.
Shaded and white boxes indicate acute and chronic events, respectively.
Multiple arrows and encircled `plus' symbols indicate potentiation
between independent cardiotoxicity mechanisms.
|
|
Our hypothesis may also help address two additional questions. The
first question is: Can iron-independent mechanisms exacerbate the
iron-based ones? Again, we propose that it can. For example, free
radicals generated during the acute phase of cardiotoxicity might
damage nucleic acids without causing overt cardiac dysfunction until
myocytes fail to repair or replace important enzymes and proteins with
low turnover rates (35)
. The ability of DOX to
concomitantly disrupt cardiac gene expression in an iron-independent
fashion might definitely exacerbate this pathway of cardiotoxicity.
Likewise, the ability of DOXol to directly inhibit constitutive or de
novo synthesized enzymes would amplify the consequences of free
radical-mediated DNA damage and free radical-independent impairment of
gene expression (Fig. 3)
.
The second question to be answered is: Can our unifying hypothesis
explain the decreased cardiotoxicity of 5-iminodaunorubicin and
reconcile this with the DOXol and IRE/IRP story? We propose that
positive answers can be given with respect to both acute and chronic
cardiotoxicity mechanisms. Due to chemical modifications at the quinone
moiety, 5-iminodaunorubicin cannot reduce oxygen nor can it delocalize
iron from cellular stores; therefore, it clearly fails to promote iron-
and free radical-dependent acute toxicity. Because of its lower redox
activity, 5-iminodaunorubicin might similarly fail to link the acute
toxicity to the free radical-independent phase of cardiac damage, as
the latter must be `prepared' by H2O2 or
lipid aldehyde induction of the aldo-keto reductases that form C-13
metabolites. The lower toxicity of 5-iminodaunorubicin can thus be
reconciled with both free radical-independent and -dependent
mechanisms.
 |
NOVEL HYPOTHESES CAN HELP DESIGN BETTER PROTECTIVE STRATEGIES
|
|---|
As mentioned, the clinical usefulness of dexrazoxane is limited by
moderate to severe exacerbation of myelotoxicity as well as by possible
interference with the antitumor efficacy of DOX. Therefore, the need
for high-dose and long-term anthracycline regimens has justified the
continuing search for alternative cardioprotectants to be used in place
of, or in combination with, dexrazoxane. To fit into our unifying
hypothesis, `good protectants' should scavenge both oxygen- and
lipid-derived species in order to prevent the induction of mechanisms
possibly increasing DOXol formation and linking the acute and chronic
phases of cardiotoxicity (cf. Fig. 3
). Moreover, `good protectants'
should directly shield IRPs and other cell constituents from the
damaging effects of DOXol-dependent reactions. Vitamin E and
N-acetylcysteine probably have too narrow a spectrum of activity to
fulfill these requirements, as they act preferentially on either lipid-
or oxygen-derived species, respectively. This may provide one more clue
to the lack of efficacy of these compounds against chronic
cardiotoxicity in patients or large animals.
A `good protectant' has been identified in amifostine (WR2721), a
phosphorylated aminothiol that exposes a hyperreactive sulfydryl moiety
(WR1065) upon dephosphorylation by membrane-bound alkaline phosphatase
(89)
. Amifostine dephosphorylation occurs in normal
tissues more effectively than in cancer foci, presumably because of
1) higher alkaline phosphatase expression in the capillary
endothelium of well-perfused normal tissues, and 2) impaired
alkaline phosphatase activity in malignant tissues, reflecting the low
pH values found within hypoxic tumor masses (8)
.
Amifostine can thus afford selective protection by shielding cellular
thiols from ionizing radiations (90)
or alkylating
antineoplastic agents (91)
. Drug metabolism studies have
indicated that WR1065 can be recovered from cells or whole organs in
the form of symmetrical disulfides and both protein and nonprotein
mixed disulfides, similar to other thiol-based antioxidants
(92)
. However, WR1065 can also be recovered in a form of
sulfinic/sulfonic derivatives (92)
, the same that probably
form in IRP-1 when DOX-Fe(II) damages critical cysteines
(12)
. This indicates that WR1065 might have the potential
to protect IRPs and perhaps other proteins from oxidation of critical
SH residues. In addition, amifostine is characterized by selective
uptake and retention in the heart and has the ability to scavenge
oxygen free radicals produced by DOX (93)
, leaving the
antitumor efficacy of the anthracycline virtually unchanged. Based on
these premises and other encouraging results in experimental models
(8)
, amifostine can be expected to mitigate both acute and
chronic cardiotoxicity, proving more versatile than N-acetylcysteine or
other thiol-based protectants.
Another cardioprotective approach that would fit well into our
unifying hypothesis of cardiotoxicity is based on low mol wt nitroxyl
radicals (nitroxides, >N=O
). Due to their chemical stability toward
most diamagnetic molecules, nitroxides have long been used in the
experimental settings as biophysical probes, spin labels, or contrast
agents in nuclear magnetic resonance and EPR imaging (94)
.
More recently, nitroxides, particularly piperidine derivatives, have
been found to protect cells (95)
, organs
(96)
, and whole animals (97
, 98
)
in a number of experimental models of oxidative stress. Such protective
effects have been attributed to the ability of the nitroxyl moiety to
dismutate O2·-, detoxify semiquinone free radicals,
reduce FeIV=O, and terminate radicalchain reactions by
direct interaction with alkyl, alkoxyl, and alkylperoxyl radicals (100,
and references therein). Recent studies suggest that some of these
antioxidant effects might be mediated at least in part by the
hydroxylamine (>N-OH) resulting from one-electron reduction of
nitroxyl (101)
.
The ability of nitroxides to intercept such a wide array of free
radicals suggests that they might prove better protectants than
N-acetylcysteine or vitamin E, as also mentioned for spin traps.
Accordingly, the piperidine nitroxide TEMPOL has been reported to
afford remarkable protection in an in vitro model of
anthracycline-induced acute cardiotoxicity (102)
. Two
additional considerations may concur to explain this effect. First,
substantial amounts of TEMPOL can be detected in subcellular
compartments that are critical sites of DOX metabolism and oxidative
damage (e.g., mitochondria and sarcoplasmic reticulum)
(102)
. Second, nitroxides can oxidize Fe(II) to Fe(III)
(94
, 102
), thereby decreasing the formation
of · OH or equivalent iron complexes from reaction of Fe(II)
with H2O2 or oxygen (cf. Section IV
). Besides
hampering the oxidative damage underlying acute cardiotoxicity,
nitroxide oxidation of Fe(II) might interfere with processes occurring
during the chronic phase of cardiac damage, such as DOX-Fe(II)
inactivation of IRPs and consequent misplacement of unregulated Fe(II)
at the Ca2+ release channel of sarcoplasmic reticulum or
other cellular sites (cf. Section VI
). Thus, the ability to oxidize
Fe(II) might enable nitroxides to protect against both acute and
chronic events of cardiotoxicity.
Recent studies with TEMPOL have shown that protective interventions
with nitroxides may have two additional advantages. First, nitroxides
are remarkably more toxic to neoplastic cells than to their normal
counterparts; second, TEMPOL is toxic to both wild-type and multidrug
resistant tumor cells (103)
. Inasmuch as nitroxides per se
cannot interfere with DOX-induced double strand breaks and cell
survival (104)
, these findings suggest that such compounds
can improve the therapeutic index of anthracyclines not only by
protecting cardiac tissues, but also by imposing their own toxicity on
cancer cells. This unexpected observation in cancer cells has been
tentatively explained by one-electron oxidation of the nitroxyl moiety
of TEMPOL to an oxo-ammonium cation species (>N+-O) with
the potential to oxidize critical cellular components
(100)
. The in vivo efficacy of nitroxides might
be limited by their too rapid reduction to the corresponding
hydroxylamines, which are probably less effective as
antioxidants and almost completely devoid of cytotoxic effects on tumor
cells. This problem might be solved by adoption of slow-release devices
and/or by the recent development of a polynitroxylated form of human
serum albumin that reoxidizes hydroxylamine, shifting the equilibrium
in favor of nitroxides (105)
.
One last approach to cardiac protection might be to decrease the
formation of oxyradicals or C-13 metabolites, thus interfering with
either or both steps in our tentative scheme (cf. Fig. 3
). Lessons from
5-iminodaunorubicin indicate that chemical modifications of the quinone
moiety decrease the formation of oxyradicals and attenuate
cardiotoxicity; however, extracardiac toxicity can unpredictably occur
after these manipulations. Attempts to modify the yield of C-13
metabolites have not been pursued so far, but should be given priority
in both academic and industrial environments for at least two different
reasons. First, C-13 metabolites do not contribute significantly to the
antitumor activity of anthracyclines (106)
. Drugs
inhibiting aldo-keto reductases or anthracyclines forming fewer C-13
metabolites should therefore prove useful to prevent cardiac toxicity
without affecting tumor response. Second, our scheme suggests that C-13
metabolites are the ultimate and most critical mediators of chronic
cardiotoxicity. Decreasing the formation of these metabolites would
generate a bottleneck in the process of cardiotoxicity, irrespective of
prior contributions from oxyradicals and related oxidative events.
 |
CONCLUSIONS
|
|---|
In this review, we have shown that the role of iron in
anthracycline cardiotoxicity may be less obvious than is usually
believed. Most popular hypotheses and less-established mechanisms of
iron-dependent toxicity have been tentatively reconciled in a unifying
framework; clearly, several aspects of this picture are speculative and
call for validation by both preclinical and clinical studies. This
limitation notwithstanding, we propose that looking at iron from
different, but not mutually exclusive, perspectives may provide new
routes to design better protectants and/or less toxic
anthracyclines.
Note added in proof: For a different picture of the
same topic, the reader may want to consult two reviews that appeared
while our article was in press [Myers, C. E. (1998) The role of iron
in doxorubicin-induced cardiomyopathy, Semin. Oncol. 25
(Suppl. 10), 1014; Singal, P. K., and Iliskovic, N. (1998)
Doxorubicin-induced cardiomyopathy N. Engl. J. Med. 339,
900905].
 |
ACKNOWLEDGMENTS
|
|---|
Work performed in the authors' laboratories was funded by
MURST (Special Project ex 40% `New assessment approaches in
toxicology') and CNR (grant no. 96.03344.CT04). The authors thank
Profs. A. Bernelli-Zazzera, F. Piccinini, P. Preziosi, and Drs. M.
Gariboldi, A. Mordente, and S. Recalcati for support and assistance.
 |
FOOTNOTES
|
|---|
1 Correspondence: G. D'Annunzio University School of Pharmacy, Institute of Drug Sciences, Department of Pharmacology and Pharmacognosy, Via dei Vestini, 66013 Chieti, Italy. E-mail: gminotti{at}unich.it 
2 Abbreviations: DOX, doxorubicin; DOXol, doxorubicinol;
O2·-, superoxide anion;
H2O2, hydrogen peroxide; low mol wt, low
molecular weight; · OH, hydroxyl radical; FeIV=O,
ferryl ion; SOD, superoxide dismutase; PBN,
N-tert-butyl-
-phenylnitrone; POBN,
-(4-pyridyl-1-oxide)-N-tert-butylnitrone; DMPO,
5,5-dimethyl-1-pyrroline N-oxide; MT, metallothionein; TfR, transferrin
receptor; IRP, iron regulatory protein; IREs, iron-responsive elements;
TEMPOL, 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl.
 |
REFERENCES
|
|---|
-
Weiss R. B.. The anthracyclineswill we ever find a better doxorubicin?. Semin. Oncol. 1992;19:670-686.[Medline]
-
Keizer H. G., Pinedo H. M., Schuurhuis G. J., Joenje H.. Doxorubicin (adriamycin)a critical review of free radical-dependent mechanisms of cytotoxicity. Pharmacol. Ther. 1990;47:219-231.[Medline]
-
Myers C. E., Gianni L., Zweier J., Muindi J., Sinha B. K., Eliot H.. Role of iron in adriamycin biochemistry. Federation Proc 1986;45:2792-2797.[Medline]
-
Olson R. D., Mushlin P. S.. Doxorubicin cardiotoxicityanalysis of prevailing hypotheses. FASEB J 1990;4:3076-3086.[Abstract]
-
Singal P. K., Iliskovic N., Li T., Kumar D.. Adriamycin cardiomyopathypathophysiology and prevention. FASEB J 1997;11:931-936.[Abstract]
-
Powis G.. Free radical formation by antitumor quinones. Free Rad. Biol. Med. 1989;6:63-101.[Medline]
-
Doroshow J. H.. Doxorubicin-induced cardiotoxicity. N. Engl. J. Med. 1991;324:843-845.[Medline]
-
Dorr R. T.. Cytoprotective agents for anthracyclines. Semin. Oncol. 1996;23(Suppl. 8):23-34.[Medline]
-
Praet M., Ruysschaert J. M.. In vivo and in vitro mitochondrial damages induced by adriamycin and derivatives. Biochim. Biophys. Acta 1993;1149:79-85.[Medline]
-
Minotti G., Mancuso C., Frustaci A., Mordente A., Santini S. A., Calafiore A. M., Liberi G., Gentiloni N.. Paradoxical inhibition of cardiac lipid peroxidation in cancer patients treated with doxorubicin. J. Clin. Invest. 1996;98:650-661.[Medline]
-
Minotti G., Cavaliere A. F., Mordente A., Rossi M., Schiavello R., Zamparelli R., Possati G. F.. Secondary alcohol metabolites mediate iron delocalization in cytosolic fractions of myocardial biopsies exposed to anticancer anthracyclines. J. Clin. Invest. 1995;95:1595-1605.
-
Minotti G., Recalcati S., Liberi G., Calafiore A. M., Mancuso C., Preziosi P., Cairo G.. The secondary alcohol metabolite of doxorubicin irreversibly inactivates aconitase/iron regulatory protein-1 in cytosolic fractions from human myocardium. FASEB J 1998;12:541-551.[Abstract/Free Full Text]
-
Herman E. H., Ardalan B., Bier C., Warawdekar V., Krop S.. Reduction of daunorubicin lethality and myocardial and cellular alterations by pretreatment with ICRF-187 in Syrian golden hamsters. Cancer Treat. Rep. 1979;63:89-92.[Medline]
-
Ryan T. P., Aust S. D.. The role of iron in oxygen-mediated toxicities. Crit. Rev. Toxicol. 1992;22:119-141.[Medline]
-
Theil E. C.. Ferritinstructure, function, and regulation. Theil E. C. Eichorn G. L. Marzili L. eds. Iron Binding Proteins without Cofactors or Sulfur Clusters 1983:1-38 Elsevier New York. .
-
Minotti G.. Sources and role of iron in lipid peroxidation. Chem. Res. Toxicol. 1993;6:134-146.[Medline]
-
Thomas C. E., Aust S. D.. Release of iron from ferritin by cardiotoxic anthracycline antibiotics. Arch. Biochem. Biophys. 1986;248:684-689.[Medline]
-
Minotti G.. Adriamycin dependent release of iron from microsomal membranes. Arch. Biochem. Biophys. 1989;268:398-403.[Medline]
-
Buss J. L., Hasinoff B. B.. The one-ring open hydrolysis product intermediates of the cardioprotective agent ICRF-187 (dexrazoxane) displace iron from iron-anthracyclines complexes. Agents Actions 1993;40:86-95.[Medline]
-
Buss J. L., Hasinoff B. B.. Ferrous ion strongly promotes the ring opening of the hydrolysis intermediates of the antioxidant cardioprotective agent dexrazoxane (ICRF-187). Arch. Biochem. Biophys. 1995;317:121-127.[Medline]
-
Baldwin J. R., Phillips B. A., Overmyer S. K., Hatfield N. Z., Narang P. K.. Influence of the cardioprotective agent dexrazoxane on doxorubicin pharmacokinetics in the dog. Cancer Chemother. Pharmacol. 1992;30:433-438.[Medline]
-
Green M. D., Alderton P., Gross J., Muggia F. M., Speyer J. L.. Evidence of the selective alteration of anthracycline activity due to modulation by ICR-187 (ADR-529). Pharmacol. Ther. 1990;48:61-69.[Medline]
-
Herman E. H., Ferrans V. J., Myers C. E., van Vleet J. F.. Comparison of the effectiveness of (±)-1,2-bis(2,5-deoxopiperanizyl-1-yl)-propane (ICRF-187) and N-acetylcysteine in preventing chronic doxorubicin cardiotoxicity in beagles. Cancer Res 1985;45:276-281.[Abstract/Free Full Text]
-
Speyer J. L., Green M. D., Kramer E., Rey M., Sanger J., Ward C., Dubin N., Ferrans V., Stecy P., Zeleniuch-Jacquotte A., Wernz J., Feit F., Slater W., Blum R., Muggia M.. Protective effect of the bispiperazine ICRF-187 against doxorubicin induced cardiac toxicity in women with advanced breast cancer. N. Engl. J. Med. 1988;319:745-752.[Abstract]
-
Speyer J. L., Green M. D., Zeleniuch-Jacquotte A., Wernz J., Rey M., Sanger J., Kramer E., Ferrans V., Hochster H., Meyers M., Blum R., Feit F., Attubato M, Burrows M., Muggia M.. ICRF-187 permits longer treatment with doxorubicin in women with breast cancer. J. Clin. Oncol. 1992;10:117-127.[Abstract]
-
Swain S. M., Whaley F. S., Gerber M.C., Weisberg S., York M., Spicer D., Jones S. E., Wadler S., Desai A., Vogel C., Speyer J., Mittelman A., Reddy S., Pendergrass K., Velez-Garcia E., Ewer M. S., Bianchine J. R., Gams R. A.. Cardioprotection with dexrazoxane for doxorubicin-containing therapy in advanced breast cancer. J. Clin. Oncol. 1997;15:1318-1332.[Abstract/Free Full Text]
-
Zhang J., Clark J. R., Jr, Herman E. H., Ferrans V. J.. Doxorubicin-induced apoptosis in spontaneously hypertensive ratsdifferential effects in heart, kidney and intestine, and inhibition by ICRF-187. J. Mol. Cell. Cardiol. 1996;28:1931-1943.[Medline]
-
Hasinoff B. B., Yalowich J. C., Ling Y., Buss J. L.. The effect of dexrazoxane (ICRF-187) on doxorubicin- and daunorubicin-mediated growth inhibition of Chinese hamster ovary cells. Anticancer Drugs 1996;7:558-567.[Medline]
-
Hasinoff B. B., Creighton A.M., Kozlowska H., Thampatty P., Allan W. P., Yalowich J. C.. Mitindomide is a catalytic inhibitor of DNA topoisomerase II that acts at the bisdioxopiperazine binding site. Mol. Pharmacol. 1997;52:839-845.[Abstract/Free Full Text]
-
Hasinoff B. B., Kuschak T. I., Fattman C. L., Yalowich J. C.. The one-ring open hydrolysis intermediates of the cardioprotective agent dexrazoxane (ICRF-187) do not inhibit the growth of Chinese hamster ovary cells or the catalytic activity of DNA topoisomerase II. Anticancer Drugs 1998;5:465-471.
-
Siveski-Iliskovic N., Hill M., Chow D. A., Singal K.. Probucol protects against adriamycin cardiomyopathy without interfering with its antitumor effect. Circulation 1995;91:10-15.[Abstract/Free Full Text]
-
Siveski-Iliskovic N., Kaul N., Singal K.. Probucol promotes endogenous antioxidants and provides protection against adriamycin-induced cardiomyopathy in rats. Circulation 1994;89:2829-2835.[Abstract/Free Full Text]
-
Monti E., Paracchini L., Perletti G., Piccinini F.. Protective effects of spin-trapping agents on adriamycin-induced cardiotoxicity in isolated rat atria. Free Rad. Res. Commun. 1991;14:41-45.[Medline]
-
Cova D., De Angelis L., Monti E., Piccinini F.. Subcellular localization of two spin trapping agents in rat heartpossible explanation for their different protective effects against doxorubicin-induced cardiotoxicity. Free Rad. Res. Commun. 1992;15:353-366.[Medline]
-
Monti E., Prosperi E., Supino R., Bottiroli G.. Free radical-dependent DNA lesions are involved in the delayed cardiotoxicity induced by adriamycin in the rat. Anticancer Res 1995;15:193-198.[Medline]
-
Kang Y. J., Chen Y., Epstein P. N.. Suppression of doxorubicin cardiotoxicity by overexpression of catalase in the heart of transgenic mice. J. Biol. Chem. 1996;271:12610-12616.[Abstract/Free Full Text]
-
Yen H. C., Oberley T. D., Vichitbandha S., Ho Y. S., St. Clair D. K.. The protective role of manganese superoxide dismutase against adriamycin-induced acute cardiac toxicity in transgenic mice. J. Clin. Invest. 1997;98:1253-1260.[Medline]
-
Kang Y. J., Chen Y., Yu A., Voss-McCowan M., Epstein P. N.. Overexpression of metallothionein in the heart of transgenic mice suppresses doxorubicin cardiotoxicity. J. Clin. Invest. 1997;100:1501-1506.[Medline]
-
van Vleet J. F., Ferrans J. V., Weirich W. E.. Cardiac disease induced by chronic adriamycin administration in dogs and an elevation of vitamin E and selenium and cardioprotectants. Am. J. Pathol. 1980;99:13-22.[Abstract]
-
Legha S. S., Wang Y. M., Mackay B., Ewer M., Hortobagy G. N., Benjamin S. R., Ali M. K.. Clinical and pharmacologic investigation of the effects of
-tocopherol on adriamycin cardiotoxicity. Ann. N.Y. Acad. Sci. 1982;393:411-418.[Abstract]
-
Myers C., Bonow R., Palmieri S., Jenkins J., Corden B., Locker G., Doroshow J., Epstein S. A.. A randomized controlled trial assessing the prevention of doxorubicin cardiomyopathy by N-acetylcysteine. Semin. Oncol. 1983;10(Suppl.):53-55.[Medline]
-
Stephens N. G., Parsons A., Schofield P. M., Kelly F., Cheeseman K., Mitchinson M. J.. Randomised controlled trial of vitamin E in patients with coronary diseaseCambridge Heart Antioxidant Study. Lancet 1995;347:781-786.
-
Sisto T., Paajanen H., Metsa-Ketela T., Harmoinen A., Nordback I., Tarkka M.. Pretreatment with antioxidants and allopurinol diminishes cardiac onset events in coronary artery bypass grafting. Ann. Thorac. Surg. 1995;59:1519-1523.[Abstract/Free Full Text]
-
Virtamo J., Rapola J. M., Ripatti S., Heinonen O. P., Taylor P. R., Albanes D., Huttunen J. K.. Effect of vitamin E and beta carotene on the incidence of primary nonfatal myocardial infarction and fatal coronary heart disease. Arch. Int. Med. 1998;158:668-675.[Abstract/Free Full Text]
-
Brattin W. J., Glende E. A., Jr, Recknagel R. O.. Pathological mechanisms in carbon tetrachloride hepatotoxicity. J. Free Rad. Biol. Med. 1985;1:27-36.[Medline]
-
Talcott R. R., Smith M. T., Giannini D. D.. Inhibition of microsomal lipid peroxidation by naphtoqu