Sublethal Doses of an Anti-erbB2 Antibody Leads to Death byApoptosis in Cardiomyocytes Sensitized by Low ProsenescentDoses of Epirubicin: The Protective Role of Dexrazoxane

Paolo Spallarossa, Paola Altieri, Paolo Pronzato, Concetta Aloi, Giorgio Ghigliotti,Antonio Barsotti, and Claudio BrunelliResearch Center of Cardiovascular Biology, Division of Cardiology, Department of Internal Medicine, University of Genova,Genova, Italy (P.S., P.A., C.A., G.G., A.B., C.B.); and Medical Oncology A, National Institute for Cancer Research, Genova,Italy (P.P.)

Received August 7, 2009; accepted October 16, 2009

ABSTRACTThe cardiotoxic synergism resulting from the sequential treat-ment with anthracyclines and trastuzumab has been attributedto the trastuzumab-induced loss of the erbB2-related functionsthat serve as a salvage pathway against the damaging effectsof anthracyclines. Cellular senescence is a novel mechanism ofcardiotoxicity induced by subapoptotic doses of anthracy-clines. After having identified prosenescent and proapoptoticdoses of epirubicin and rat MAb c-erbB2/Her-2/neu Ab-9 cloneB10 (B10), an anti-erbB2 monoclonal antibody, we investigatedthe effects of the sequential treatment with prosenescent dosesof both drugs on H9c2 cells and neonatal rat cardiomyocytespretreated with or without the cardioprotective agent dexrazox-ane. Cells were analyzed by senescence-associated �-galac-tosidase, single-stranded DNA, annexin/propidium doublestaining, F-actin, and mitochondrial transmembrane potential.ErbB2 expression levels, AKT activation, and the effects of theinhibition of nicotinamide adenine dinucleotide phosphate oxi-

dase [NAD(P)H oxidase] and phosphoinositide-3-OH kinase(PI3K) were also assessed. Data demonstrate that 1) the toxiceffects of epirubicin mainly occur through NAD(P)H oxidaseactivation; 2) the erbB2 overexpression induced by epirubicin isa redox-sensitive mechanism largely dependent on NAD(P)Hoxidase; 3) the loss of erbB2-related functions caused by B10determines marginal cellular changes in untreated cells, butcauses massive death by apoptosis in cells previously exposedto a prosenescent dose of epirubicin, 4) dexrazoxane promotessurvival pathways, as demonstrated by the activation of Aktand the PI3K-dependent erbB2 overexpression; and 5) it alsoprevents epirubicin-induced senescence and renders epirubi-cin-treated cells more resistant to treatment with B10. Dataunderline the importance of NAD(P)H oxidase in epirubicin-induced cardiotoxicity and shed new light on the protectivemechanisms of dexrazoxane.

Epidermal growth factor receptor erbB2 is a transmem-brane glycoprotein receptor with tyrosine kinase activity.Trastuzumab (Damen et al., 2008) is an anti-erbB2 human-ized monoclonal antibody. When added to chemotherapy inpatients with erbB2 overexpressing breast cancer it leads toa significant gain in survival (Slamon et al., 2001). However,the gain in survival occurs at the cost of an increased risk of

hypokinetic cardiomyopathy and heart failure (Bria et al.,2008). Trastuzumab-related cardiotoxicity is not associatedwith severe structural changes and is often reversible afterdiscontinuing therapy (Ewer and Ewer, 2008). However, clin-ical studies have shown that, if women have been cotreatedor pretreated with anthracyclines, both the incidence andseverity of myocardial dysfunction remarkably increase(Slamon et al., 2001; Tan-Chiu et al., 2005). It has beenhypothesized that while activating cardiomyocyte stresspathways, anthracyclines also activate survival pathways,the most important of which is the neuregulin/erbB2 system(Gabrielson et al., 2007; Pentassuglia et al., 2009). At lowdoses of anthracyclines, the survival pathway effects over-

This work was supported by the University of Genova (cofinanziamento diAteneo).

P.S. and P.A. contributed equally to this study.Article, publication date, and citation information can be found at

http://jpet.aspetjournals.org.doi:10.1124/jpet.109.159525.

ABBREVIATIONS: SIPS, stress-induced premature senescence; B10, rat MAb c-erbB2/Her-2/neu Ab-9 clone B10; NAC, N-acetylcysteine;NAD(P)H oxidase, nicotinamide adenine dinucleotide phosphate oxidase; FITC, fluorescein isothiocyanate; DPI, diphenyleneiodonium; PBS,phosphate-buffered saline; PI3K, phosphoinositide-3-OH kinase; LY294002, 2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one; ssDNA, single-stranded DNA; AV/PI, annexinV/propidium iodide; SA-�-gal, senescence-associated �-galactosidase staining; GAPDH, glyceraldehyde 3-phos-phate dehydrogenase; ct, control; Epi, epirubicin; Dox, doxorubicin; Dex, dexrazoxane; ROS, reactive oxygen species.

0022-3565/10/3321-87–96$20.00THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 332, No. 1Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics 159525/3545698JPET 332:87–96, 2010 Printed in U.S.A.

87

whelm the toxic effects, and cardiac dysfunction does notappear, at least for several years. By inhibiting the erbB2receptor, trastuzumab induces a second stress that impairsthis survival pathway, creating an unbalance in favor of thetoxic effects induced by anthracyclines. Thus, Ewer recentlyconcluded that trastuzumab-related cardiac dysfunctionmust be viewed as an anthracycline-related problem, ratherthan a trastuzumab-related problem (Ewer and Tan-Chiu,2007). Accordingly, preventing low-dose anthracycline-in-duced cardiotoxicity may represent a rational approach toprotecting cells from further stress provoked by exposure toan anti-erbB2 antibody.

The most accredited hypothesis for explaining anthracy-cline cardiotoxicity is that anthracyclines induce myocyteloss through oxidative stress and apoptotic cell death (Mennaet al., 2008). It has recently been shown in cardiac myocytesthat when anthracyclines are used at subapoptotic doses,they induce stress-induced premature senescence (SIPS)(Maejima et al., 2008). Senescence is a fundamental cellu-lar program that contributes to the physiology of livingtissues, the aging process, and diseases. The hallmark ofcellular senescence is cell cycle arrest accompanied bymorphological and structural changes including flattenedand enlarged cell shapes and cytoskeleton remodeling.These structural changes are associated with the inabilityto divide and the reduction of telomere length that mayresult in late death (Ben-Porath and Weinberg, 2004).

It has been shown that pretreatment with dexrazoxane(ICRF-187) (Malisza and Hasinoff, 1996), an antioxidant andiron-chelating agent, is effective in preventing many featuresof anthracycline cardiotoxicity (Spallarossa et al., 2006; vanDalen et al., 2008). However, there are no data on whetherdexrazoxane also prevents senescence in cardiac musclecells.

Aims of the present study were, first, to characterize thecell damage induced by an anti-erbB2 monoclonal antibodyin myocardial rat cells previously exposed to subapoptotic,prosenescent concentrations of epirubicin. Second, to assesswhether pretreatment with dexrazoxane is able to preventepirubicin-induced senescence-like phenotype, and to inves-tigate the mechanisms and signal transduction pathwaysthrough which dexrazoxane exerts cardioprotective effects.Third, to assess whether cells that had been protected bydexrazoxane toward epirubicin are more resistant againstthe cardiotoxic effects of an anti-erbB2 monoclonal antibody.

We performed experiments using H9c2 cardiomyoblastsand neonatal cardiomyocytes that represent an extensivelyused model for studying myocardial toxicity induced by an-thracyclines and anti-erbB2 drugs (Spallarossa et al., 2004,2006; Pugatsch et al., 2006; Salvatorelli et al., 2006; Maejimaet al., 2008; Gordon et al., 2009). Although postmitotic adultcardiomyocytes are the most functionally significant cell typein the heart, numerous studies have shown that the heartalso contains a pool of progenitor cells and a population ofimmature, dividing myocytes that are considered the linkbetween progenitor cells and differentiated cardiomyocytes(Chen et al., 2007). Thus, in the normal heart there is aturnover of cardiomyocytes involving the death and the gen-eration of new cardiomyocytes. Because of the enhanced sen-sitivity of dividing cardiomyocytes to anthracyclines, it hasbeen suggested that anthracyclines may inhibit the regener-ative capacity of the heart and, through this mechanism,

impair the self-repairing potential of the heart that ulti-mately leads to late events (Konorev et al., 2008). Neonatalrat cardiomyocytes share some characteristics with the pop-ulation of replicating cells that are present in the adult heart,and that are reminiscent of a fetal/neonatal phenotype (Chenet al., 2007). Thus, our experimental model may be consid-ered a convenient indicator of what might happen to thesecardioregenerative cells when the heart is exposed to anthra-cyclines and an anti-erbB2 antibody.

Materials and MethodsTo carry out the study we chose to use epirubicin and B10, because

epirubicin is becoming the most used anthracycline in the treatmentof breast cancer, and B10 is a rat-anti-erbB2 monoclonal antibodywhose activity has been found to biologically resemble that of tras-tuzumab, the humanized antibody that is used in the clinical setting(Pugatsch et al., 2006). All materials, unless otherwise stated, weresupplied by Sigma-Aldrich (Poole, United Kingdom).

Cell and Culture Conditions. H9c2 rat heart-derived embry-onic myocytes (American Type Culture Collection, Manassas, VA)were cultured as described previously (Spallarossa et al., 2006).Neonatal ventricular myocytes from 2-day-old Sprague-Dawley ratswere purified and cultured as described previously (Antos et al.,2003). Cells were always used at less than 70% of confluence. Theywere incubated with various doses of epirubicin for 3 h or withvarious doses of rat MAb c-ErbB2/Her-2/neu Ab-9 (B10) (Neomark-ers, Fremont, CA) for 24 h to evaluate senescence and apoptosis, andto choose the prosenescent subapoptotic dose of each drug that wasmost suited to perform the study.

Experimental Design. Cells were pretreated for 3 h with orwithout dexrazoxane (20 �M) (Novartis Pharma, Basel, Switzer-land), then they were treated for 3 h with or without epirubicin, andsuccessively they were incubated with or without B10 for 24 h. Toevaluate the mechanisms of cardioprotection and survival signalingpathways, cells were also exposed to the following pretreatments:N-acetylcysteine (NAC, 10 mM) for 90 min; the NAD(P)H oxidaseinhibitor, diphenyleneiodonium (DPI, 20 �M) for 90 min; the phos-phoinositide-3-OH kinase (PI3K) pathway inhibitor, LY294002 (10�M) for 30 min. At the end of the treatments, cells were analyzed forapoptosis (single-stranded DNA, ssDNA), annexin V-fluorescein iso-thiocyanate (FITC)/propidium iodide staining (AV/PI), and mito-chondrial transmembrane potential. AV/PI staining was also ana-lyzed 24 h after the end of the treatments to assess both early andlate apoptosis. Senescence was evaluated 24 h after the end of thetreatments because the rate of senescent cells before this time pointwas too low to allow statistical analysis, whereas, for longer culturetimes, untreated neonatal cardiomyocytes spontaneously underwentreplicative senescence. F-actin was assessed 24 h after the end of thetreatments.

SA-�-Gal Staining. Cells were stained for �-gal activity as de-scribed previously (Dimri et al., 1995). The number of senescence-associated �-galactosidase (SA-�-gal)-positive cells was determinedin 100 randomly chosen low-power fields (�100) and expressed as apercentage of all counted cells.

Detection and Quantitation of Apoptosis by ssDNA Anti-bodies. Cell monolayers grown on glass coverslips were exposed toagents as described earlier so as to detect apoptosis by identifyingssDNA (Tauszig-Delamasure et al., 2007), then they were fixed in100% methanol at �20°C for 24 h. To induce DNA denaturation insitu, cells were heated to 100°C in PBS containing 5 mM MgCl2 for5 min, then immersed in ice-cold water for 10 min. After incubationwith 40% fetal bovine serum in PBS on ice for 15 min, cells wereincubated with a monoclonal antibody to ssDNA (10 mg/ml, ApostainF7-26; MedSystem Diagnostics, Vienna, Austria) for 30 min at roomtemperature, then washed twice in PBS. The last two steps wereperformed using the improved biotin–streptavidin-amplified detec-

88 Spallarossa et al.

tion system. In brief, the cell pellet was incubated with the biotin-ylated Pan-specific secondary antibody (Vector Laboratories, Burlin-game, CA) for 30 min at room temperature. Cells were then washedin PBS, incubated with the FITC-streptavidin for 30 min, and ob-served under a fluorescence microscope.

Annexin V-FITC/Propidium Iodide Staining. Cells were la-beled with Annexin V-FITC and propidium iodide (Spallarossa et al.,2004), and 100 randomly selected fields were counted by use of afluorescence microscope. The number of stained cells was normalizedto the total number of cells as counted by phase microscopy of thesame field. Images from independent fields were counted under afluorescence microscope.

Determination of Mitochondrial Transmembrane Poten-tial. Mitochondrial transmembrane potential was determined by useof the MitoCapture Apoptosis Detection Kit (Furukawa et al., 2002).This kit provides a fluorescence-based method for distinguishingbetween healthy and apoptotic cells by detecting the changes in themitochondrial membrane potential. It uses the MitoCapture Re-agent, a cationic dye that fluoresces differently in healthy cells andin apoptotic cells. In healthy cells, the MitoCapture Reagent accu-mulates and aggregates in the mitochondria, emitting a bright redfluorescence. In apoptotic cells, this reagent cannot aggregate in themitochondria because of the altered mitochondrial membrane poten-tial, and thus it remains in the cytoplasm in its monomeric form andemits green fluorescence. The fluorescent signals can be detectedeasily by fluorescent microscopy using appropriate filters.

F-Actin Detection. Cells were fixed, permeabilized, and labeledsimultaneously in PBS containing 50 �g/ml lysopalmitoyl phosphati-dylcholine, 3.7% formaldehyde, and 5 units/ml of fluorescent phallo-toxin (A-12379 Alexa Fluor 488 phalloidin; Invitrogen, Carlsbad,CA). Cells were rapidly washed three times with PBS and wereviewed by fluorescent microscopy. To quantify the fluorescence andto measure cell area, image analysis was performed by the LeicaQ500 MC Image Analysis System (Leica, Cambridge, UK). Threehundred cells were randomly analyzed for each sample, and theoptical density of the signals was quantitated by a computer. Thevideo image was generated by a charge-coupled device camera con-nected through a frame grabber to a computer. Single images weredigitized for image analysis at 256 gray levels. Imported data werequantitatively analyzed by Q500MC Software-Qwin (Leica). The sin-gle cells were randomly selected by the operators by using the cursor,and positive areas were estimated automatically. Constant opticalthreshold and filter combination were used.

Immunoblotting. Immunoblotting was performed by use of thepreviously described procedure (Spallarossa et al., 2006). All anti-bodies were purchased from Santa Cruz Biotechnology Inc. (SantaCruz, CA). Thirty hours after the beginning of the treatments, cellswere processed to determine the levels of p16INK4A using p16(N-20)and erbB2 using erbB2 antibody C-18. Phosphorylated Akt (pAkt)levels were evaluated at each time point indicated in the experiment.After incubation in horseradish peroxidase secondary antibody, blotswere visualized with enhanced chemiluminescence substrate (GEHealthcare, Little Chalfont, Buckinghamshire, UK), and films werequantified by densitometry with an image analyzer system (Syn-gene, Frederick, MD). Filters were stripped and reprobed with glyc-eraldehyde 3-phosphate dehydrogenase (GAPDH) or total Akt anti-bodies to normalize the amounts of erbB2 and of phosphorylated Akt,respectively.

Statistical Analysis. Data are reported as mean � S.E.M. of fourindependent experiments. Statistical analysis was performed by one-way analysis of variance followed by Bonferroni’s post hoc test.

ResultsAfter having performed preliminary tests on H9c2 cells, we

carried out the study on neonatal rat cardiomyocytes andobtained similar results. Herein, we report the data that

were collected from the experiments on neonatal cardiomyo-cytes. We first performed a dose-effect curve to test the ef-fects of various doses of epirubicin and B10 on senescenceand apoptosis. To assess apoptosis, we chose the ssDNAtechnique, which stains condensed chromatin by using anti-bodies against ssDNA and is considered a definite marker ofapoptosis. To assess senescence, we evaluated SA-�-gal ac-tivity and p16INK4A tumor suppressor protein expression lev-els (Dimri et al., 1995; Sharpless and DePinho, 2004) whichare two widely recognized indicators of cellular senescence.SA-�-gal activity is a manifestation of residual lysosomalactivity at suboptimal pH (pH 6), which becomes detectablebecause of the increased lysosomal content on senescent cells(Gerland et al., 2003).

At low doses, epirubicin and B10 induced a senescence-likephenotype, whereas at high doses they led to apoptosis. Incomparison with doxorubicin, the equicardiotoxic doses ofepirubicin were 2.6-fold higher. Senescent cells showed en-larged volume and flattened morphology, and they seemedvacuolated (Fig. 1).

Because our aim was to create an experimental model inwhich the two agents would exert mild cardiotoxic effectswhen used separately, we chose a concentration of 0.3 �M forepirubicin and 1 mg/ml for B10 to carry out the study, be-cause these are the highest concentrations that can be usedwithout inducing apoptosis. Both epirubicin and B10 inducedsenescence at these concentrations. The study was designedso that cells would have to be pretreated with or withoutdexrazoxane, incubated with epirubicin, B10, or epirubicinfollowed by B10, and then analyzed by several techniques. Inan attempt to investigate whether the mechanism of action ofdexrazoxane is related to its antioxidant properties, we alsorepeated several experiments using the NAC antioxidant andthe NAD(P)H inhibitor, DPI, instead of dexrazoxane.

We observed that the sequential incubation with epirubi-cin, followed by B10, produced a dramatic change in thecellular response program to the stress, which shifted from asenescence-like phenotype induced by epirubicin alone (30%of cells were SA-�-gal positive without signs of apoptosis) tomassive death by apoptosis (65% of cells were ssDNA posi-tive) (Fig. 2). B10 alone produced a very small increase in thenumber of SA-�-gal-positive cells. We also found that theeffects of dexrazoxane pretreatment differed in the varioustreatment groups: dexrazoxane lowered the increase in SA-�-gal activity in epirubicin-treated cells, it resulted in a smallincrease in SA-�-gal activity in solvent control-treated cells,it exerted a mild proapoptotic effect on B10 treated cells, and,most interestingly, it halved the ssDNA positivity rate incells that had been exposed to epirubicin followed by B10. Weevaluated p16INK4A levels as a second marker of senescenceand found that the increase in p16INK4A protein levels in thevarious treatment groups was similar to the increase in SA-�-gal activity (Fig. 3). The effects of pretreatment with NACwere exactly the same as the effects of dexrazoxane in alltreatment groups. Pretreatment with DPI produced the sameeffects as both dexrazoxane and NAC did in cells that hadbeen treated with epirubicin alone. Conversely, pretreatmentwith DPI had no effect on solvent control-treated cells,whereas it determined a small increase in the number ofsenescent cells in B10-treated cells and abolished the apopto-tic effect of the sequential treatment with epirubicin andB10.

Dexrazoxane, Epirubicin, and Anti-erbB2 in Cardiomyocytes 89

We also used MitoCapture to determine whether epirubi-cin or B10 induce a decrease in the mitochondrial membranepotential. In healthy cells the cationic dye accumulates in themitochondria, producing a bright red fluorescence, whereasin cells with altered mitochondrial membrane potential thedye remains in its monomeric form in the cytoplasm, leadingto green fluorescence. We calculated the ratio of red to greenstaining to quantify the results. Fluorescence analysisshowed that neither epirubicin nor B10 produced significantchanges in the red-to-green fluorescence ratio, whereas thesequential application of both produced a severe mitochon-drial perturbation that was counteracted in part by the pre-treatment with dexrazoxane (Fig. 4).

Phalloidin staining for F-actin confirmed that in theseexperimental conditions neither B10 nor epirubicin alonecaused severe damage to cardiomyocytes, but that they didincrease cell size, a typical pattern of the senescence pheno-type, and that they also increased both the length and thedensity of the cytoplasmic actin fibers. Cytoskeletal compo-nents are key regulators of cellular architecture. The en-hanced actin fibers may be interpreted functionally as aframe structure that is needed to support the enlarged cells.Epirubicin, but not B10, also caused a mild myofibrillardisarray. Conversely, when B10 was administered after epi-rubicin, it led to disruption of the actin fibers, thus confirm-ing the presence of apoptosis because cleavage of F-actinfibers is an early indicator of apoptosis (Brown et al.,1997). Itis noteworthy that this phenomenon was almost completely

abolished when cells were pretreated with dexrazoxane be-fore exposure to epirubicin (Fig. 5).

We then analyzed cells for AV/PI. It is known that trans-location of the phospholipid phosphatidylserine membranefrom the inner to the outer leaflet of the plasma membrane isone of the earliest signs of apoptosis and that it precedesother apoptotic processes such as the loss of plasma mem-brane integrity. Consequently, when cells are stained withAV (a Ca2�-dependent phospholipid-binding protein withhigh affinity for phosphatidylserine) and PI (a membrane-impermeable DNA staining), the early-stage apoptosis ischaracterized by an AV(�)/PI(�) cell population, whereas inlate-stage apoptosis many cells become AV(�)/PI(�). Recentstudies indicate that the early loss of plasma membraneintegrity, which is characterized by AV(�)/PI(�), is oftenassociated with the induction of a senescent-like phenotypein cells treated with subapoptotic doses of anthracyclines(Eom et al., 2005). Analysis of AV/PI double staining at 24 hrevealed the following: B10 has very little effect on AV/PIdouble staining; epirubicin induces the early loss of mem-brane integrity in almost 30% of cells, which is not typical ofapoptotic cardiac damage and cannot be interpreted as ne-crosis, because of the loss of plasma membrane integrity thatoccurs in necrosis is preceded by the loss of mitochondrialpotential. The exposure to B10 after epirubicin determinesthe typical features of early-stage apoptosis in more than50% of cells. Again this phenomenon was partially attenu-ated when cells were pretreated with dexrazoxane before

0

10

20

30 ssDNAb-gal

0 0.5 1 3 5

% o

f p

osi

t ive

cel

ls

*

*

*

ba

0

10

20

30

40 ssDNAb-gal

0 0.05 0.1 0.3 0.5 1 3 0.1 1

% o

f p

osi

tive

cel

ls

*

**

*

*

*

*

B10Epi

µµµµmol/L µµµµg/ml

Dox

*

*

* *

c

50µµµµm50µµµµm

EpictFig. 1. Dose-dependent effects of epirubicin and B10 on apoptosis and senescence. Bar graph showing the percentages of ssDNA and SA-�-gal-positivecells after treatment with various doses of epirubicin, doxorubicin (a) or B10 (b). c, representative photographs illustrating the morphological changesand SA-�-gal activity in cardiomyocytes with stress-induced premature senescence. In this picture, cells were treated with and without 0.3 �Mepirubicin, stained for �-gal activity, and counterstained with Giemsa solution (magnification, �400). �, p � 0.05 versus untreated cells.

90 Spallarossa et al.

exposure to epirubicin (Fig. 6). Figure 7 shows the effects ofvarious treatments on erbB2 expression. Treatment withepirubicin increased the expression levels of erbB2 by 80%.Treatment with dexrazoxane, NAC, or DPI significantly at-tenuated such epirubicin-induced increase of erbB2 expres-sion. In contrast, treatment with dexrazoxane or NAC, butnot with DPI, induced an increase of approximately 200% ofthe expression level of erbB2 in cells that had been exposed toepirubicin. Treatment with B10 reduced the expression lev-els of erbB2 by 70%. Neither pretreatment with epirubicin ordexrazoxane, nor pretreatment with dexrazoxane followed byepirubicin influenced the effects of B10 on erbB2 expression.

Both in vivo and ex vivo studies indicate that survivalfactors or transforming events such as erbB2 overexpression/activation activate Akt in a PI3K-dependent manner (Zhou etal., 2000; Kanakry et al., 2007; Tokunaga et al., 2007). In

ct Epi B10

Dex+Epi+ B10 01B+xeD01B+ipE

ct Epi Epi +B10 Dex+ Epi

a

b

0

20

40

60

80

ssDNA+

b-gal+

- + - - - - + + + + - - - + + +- - + - - - + + + + + + + - - -- - - + - - - + - - + - - + - -- - - - + - - - + - - + - - + -- - - - - + - - - + - - + - - +

% o

f c e

lls

EpiB10 Dex

NACDPI

c

*

* * * *

§#

*#*

#

*

*#

§#

* *

§

50µµµµm 50µµµµm

*#

*#

*#

*#

*#

*#

*§

50µµµµm50µµµµm 50µµµµm

50µµµµm

50µµµµm 50µµµµm 50µµµµm 50µµµµm

Fig. 2. Sequential treatment with subapo-ptotic doses of epirubicin and B10 inducesapoptosis: protective effects of dexrazoxane.a, photographs showing that the sequentialexposure to subapoptotic doses of epirubicinand B10 induces massive apoptosis as as-sessed by ssDNA analysis. Note that pre-treatment with dexrazoxane attenuatesapoptosis induced by epirubicin followed byB10, but it results in a small increase in thenumber of ssDNA-positive cells in cells sub-sequently incubated with B10 (magnifica-tion, �200). b, photographs showing theprosenescent effect of subapoptotic doses ofepirubicin. Note that the sequential expo-sure to epirubicin and B10 determines thereduction of SA-�-gal activity, whereas pre-treatment with dexrazoxane attenuates theprosenescent effects of epirubicin (magnifi-cation, �200). c, bar graph showing the per-centages of ssDNA and SA-�-gal-positivecells in cells pretreated with or withoutdexrazoxane, NAC, or DPI, and then incu-bated with epirubicin, B10, or epirubicin fol-lowed by B10. �, p � 0.05 versus untreatedcells; §, p � 0.05 versus Epi � B10; #, p �0.05 versus Epi.

0

100

200

300

400

- - + - + + +

- + - - - + + - - - + + - +

Epi

B10Dex

p16

INK

4A

exp

res s

ion

(%)

*

*

*

*§

*#

Fig. 3. p16INK4A protein levels. Bar graph showing the p16INK4A proteinlevels in cells pretreated with or without dexrazoxane and then incubatedwith epirubicin, B10, or epirubicin followed by B10. The amount ofp16INK4A was normalized to GAPDH. Note that increases in p16INK4A areparallel to those observed in SA-�-gal activity. *, p � 0.05 versus un-treated cells; §, p � 0.05 versus Epi; #, p � 0.05 versus Epi�B10.

Dexrazoxane, Epirubicin, and Anti-erbB2 in Cardiomyocytes 91

keeping with those studies, we found that treatments thatup-regulate erbB2 expression, such as epirubicin or dexra-zoxane, also increase Akt phosphorylation levels, whereaspretreatment with dexrazoxane, which prevents the epirubi-cin-induced erbB2 up-regulation, also prevents epirubicin-induced Akt activation. We then investigated whether PI3Kis involved in the regulation of erbB2 expression. By usingthe PI3K inhibitor, LY294002, we observed that PI3K isinvolved in the overexpression of erbB2 induced by dexrazox-ane or by the sequential administration of dexrazoxane andepirubicin, but not in the overexpression induced by epirubi-cin alone (Fig. 8).

DiscussionData show that both B10 and epirubicin at high doses

induce apoptosis in cardiac myocytes, but at low doses theyinduce SIPS. This result is in agreement with previous stud-ies demonstrating that cells exposed to stress may respondby entering either senescence or apoptosis, and that the typeof cell response to stress depends on cell type and intensity ofthe stress (Eom et al., 2005; Maejima et al., 2008).

Epirubicin Cardiotoxicity Mainly Occurs throughNAD(P)H Oxidase. Compared with doxorubicin, the equi-cardiotoxic doses of epirubicin were 2.6-fold higher. Thisresult is similar to clinical studies that showed that the ratio

is approximately 2:1 (Ewer and Benjamin, 2000). Improvedglucuronidation and systemic elimination of epirubicin arebelieved to decrease cardiotoxicity in cancer patients by di-minishing the deleterious interactions of epirubicin with car-diomyocytes. However, our own data, as well as those ofothers, show that epirubicin is less cardiotoxic than doxoru-bicin, even in cultured cardiomyocytes. Salvatorelli et al.(2006) showed that epirubicin, unlike doxorubicin, undergoessequestration in cytoplasmic acidic organelles and thereforecannot reach the mitochondrial reductases that convert it toROS and generate smaller amounts of toxic alcohol second-ary metabolites. This is why epirubicin was thought to be amuch better partner than doxorubicin for the associationwith other toxic drugs like anti-erbB2 antibodies.

Epirubicin, however, is not devoid of a ROS-mediated toxiceffect. Our study suggests that the ROS-mediated toxic effectof epirubicin mainly occurs through NAD(P)H oxidase, aplasma membrane-associated enzyme.

Prosenescent Doses of Epirubicin Induce ApoptosisWhen erbB2-Related Functions Are Impaired. Eventhough epirubicin does not induce severe cell damage at thedoses we tested, as suggested by the preserved mitochondrialtransmembrane potential, it may act synergistically with thecardiotoxic effects of anti-erbB2 antibodies. In fact, our re-sults indicate that erbB2 plays a key regulatory role in de-

B10ct Epia

02468

1012

- - + + + - + - + +

- - - - +

red

/gr e

en r

ati o

b

Epi B10Dex

*

*§

Epi+B10 Epi+B10+Dex

50µµµµm

50µµµµm 50µµµµm

50µµµµm50µµµµm Fig. 4. Cells scored for intact andmodified mitochondrial membranepotential using MitoCapture probe.a, cells with mitochondrial damageshow diffuse green fluorescence,whereas cells with intact mitochon-drial membrane potential showpunctuated red fluorescence (magni-fication, �400). b, bar graph showingthe ratio of normal (red signal) andcollapsed mitochondrial membranepotential (green signal) in controland in treated cells. Quantization ofdata demonstrates that the loss ofmitochondrial membrane potentialinduced by the sequential treatmentwith epirubicin and B10 is attenu-ated by pretreatment with dexrazox-ane. �, p � 0.05 versus untreatedcells.; §, p � 0.05 versus Epi � B10.

92 Spallarossa et al.

ciding the program of cell response to stress. Cardiomyocytesthat have been exposed to pulsed incubation with low dosesof epirubicin show the activation of a survival pathway char-acterized by erbB2 overexpression and Akt phosphorylationand exhibit SIPS for as long as the erbB2 receptor is active.However, the myocytes undergo early and massive death byapoptosis if they are successively incubated with B10 theanti-erbB2 antibody that internalizes the erbB2 receptor,thus blocking this survival pathway. We used B10 at a con-centration that, by itself, produces only minimal cellulardamage, as documented not only by the low rate of SA-�-galpositive cells and the lack of apoptosis, but also by the main-tenance of the mitochondrial transmembrane potential andF-actin fiber organization. Thus, it stands to reason that theerbB2-related functions gain importance when cells getstressed, and can make the difference when cells are exposedto stresses of mild intensity, such as low doses of epirubicin,because they prevent apoptosis at the cost of a moderate rateof senescence.

Chuang et al. (2007) demonstrated that erbB2 gene expres-sion in human breast cancer cells is repressed by manganesesuperoxide dismutase, a primary antioxidant enzyme. Ourstudy confirms the involvement of a redox-sensitive mecha-nism in the regulation of erbB2 and demonstrates that theerbB2 overexpression induced by epirubicin largely dependson NAD(P)H oxidase. In fact, we observed that DPI is supe-

rior to NAC or dexrazoxane at preventing epirubicin-inducederbB2 overexpression, most likely because DPI abrogates themajor source of ROS, whereas these antioxidants have topassively cross the cell membrane and can only partiallyquench the ROS generated by NAD(P)H oxidase. Thus,NAD(P)H oxidase plays a dual role in treatment with epiru-bicin. On the one hand, it is a major contributor to thecardiotoxic effects of epirubicin, whereas, on the other, ittriggers the activation of survival pathways.

Our data concerning the effects of epirubicin on Akt are inline with a recent study (Gabrielson et al., 2007) showingthat doxorubicin treatment in rats results in an increase inAkt phosphorylation. However, our results are in contrastwith Li et al. (2006), who reported that Akt activation re-mained unchanged in mice after a single treatment withanthracyclines, and with Xiang et al. (2009), who showedthat the cardiomyocytes treated with doxorubicin reveal sig-nificantly decreased Akt activation. We hypothesize thatsuch discrepancies in phospho-Akt modulation may be re-lated to doses and to timing of evaluation. The increase inphospho-Akt suggests an initial survival response to injury,whereas its down-regulation could be associated with a moresevere pattern of cardiotoxicity.

Autophagic Changes in Cardiomyocytes. The aim ofthis study was to assess the synergistic effects of epirubicinand B10 and not to evaluate the fate of epiribicin- or

ct

B10

Epi

Epi+B10+Dex

Epi+B10

a

Are

a (%

)

- + - + + - - + + +

- - - - +

EpiB10Dex

0

100

200

300

400

*

*

*

0

20

40

60

- + - + +

- - + + +

- - - - +

Epi

B10Dex

*

*

*

b

Flu

ore

scen

ce

inte

nsi

ty (

AU

)

c

*§

*§

50µµµµm

50µµµµm 50µµµµm

50µµµµm 50µµµµm

Fig. 5. Phalloidin staining for F-actin filaments. a, photographs showing that epirubicin and B10 alone increase both the length and the density ofthe cytoplasmic actin fiber, whereas the sequential exposure to both agents produces disruption of actin fibers. Note the protective effects ofpretreatment with dexrazoxane (magnification, �400). Bar graphs showing changes in fluorescence intensity (b) and cell size (c). AU, arbitrary units.�, p � 0.05 versus untreated cells; §, p � 0.05 versus Epi � B10.

Dexrazoxane, Epirubicin, and Anti-erbB2 in Cardiomyocytes 93

B10-induced senescent cells. However, the increase in SA-�-gal activity together with the appearance of vacuolatedcells suggests that the abnormal autophagy may be in-volved in the development of a senescent-like phenotype.Three types of autophagic disorders have been described inthe context of heart failure (Takemura et al., 2006): 1)excessive autophagy, as occurs with high doses of anthra-cyclines (Lu et al., 2009); 2) decompensated autophagy, ifautophagy fails because of too much stress-induced over-load; 3) dysfunctional autophagy, if the stress inducesdysfunction of the autophagic process.

We speculate that dysfunctional autophagy dependent onimperfect autophagic turnover due to lipofuscin accumula-tion (Terman et al., 2004) might be involved in SIPS.

Dexrazoxane Renders Cardiomyocytes More Resis-tant to the Sequential Administration of Epirubicinand B10. Using dexrazoxane, we obtained a number of novelresults. First, pretreatment with dexrazoxane can preventthe prosenescent effects of low-dose epirubicin. This is a veryinteresting finding in the light of the growing importanceascribed to the cardiotoxic effects of low-dose anthracyclines.Second, cells that have been protected by dexrazoxaneagainst the toxic effects of low doses of epirubicin are much

more resistant to the stress induced by the subsequent incu-bation with B10. This result confirms the hypothesis thattoxicity of an anti-erbB2 agent could be significantly attenu-ated by preventing the sensitizing effect of anthracyclines.Third, if incubation with dexrazoxane is not followed by anytreatment, there is a mild, yet significant increase in the rateof SIPS coupled with an up-regulation of erbB2 expression,which, unlike what is observed after epirubicin, is not medi-ated by NAD(P)H oxidase and oxidative stress, but occurs viaPI3K. We also found that dexrazoxane activates Akt. Dexra-zoxane is thought to exert cardioprotective effects through itsiron-chelating metabolite that decrease the formation ofROS. However, many ROS scavengers or iron chelators areineffective in preventing Dox cardiotoxicity. Our findingssuggest that the cardioprotective effect of dexrazoxane alsooccurs through iron chelation-independent mechanisms thatinclude the activation of PI3K/Akt and erbB2 survival path-ways. Fourth, if incubation of cells with dexrazoxane or NACis followed by exposure to B10, the subsequent loss of theerbB2 function results in a shift from SIPS to apoptosis,which, however, involves a limited number of cells. Our in-terpretation of the two latter points is as follows. First, wemust bear in mind that some degree of oxidative stress is

B10

ct

Epi+Dex

Epi

0

20

40

60

80A(+)/PI(+)A(+)/PI(-)

% o

f c e

lls

- - + + + +

- + - - + + - - - + - +

Epi+B10+Dex

Epi+B10

Epi

B10Dex

a

b

50µµµµm

50µµµµm 50µµµµm 50µµµµm

50µµµµm 50µµµµmFig. 6. AnnexinV/propidium iodide double staining. a, rep-resentative photographs of AV/PI staining at 24 h intreated and untreated cells (magnification, �200). b, quan-tization of data shows that epirubicin alone increases thepercentage of annexin-positive cells with early loss ofplasma membrane integrity [A(�)/PI(�)], whereas epiru-bicin followed by B10 leads to a dramatic increase in thepercentage of annexin-positive cells with initially pre-served plasma membrane integrity [A(�)/PI(�)]. Note theprotective effects of pretreatments with dexrazoxane.

94 Spallarossa et al.

necessary for cell biology. For example, several gene tran-scription factors require transient oxidation for their func-tion and, most importantly, small quantities of ROS areneeded for cell proliferation (Halliwell et al., 2000). It followsthat antioxidants exert protective effects if cells are exposedto enhanced oxidative stress, but that they may provokedamage if they are not. Data also suggest that the erbB2-related survival pathway might play an important role notonly in anthracycline cardiotoxicity, but also in other condi-tions of cell suffering, such as antioxidant-induced senes-cence. This could explain why the blockade of the erbB2receptor in cells having mild signs of senescence after pre-treatment with dexrazoxane produces mild enhancement ofthe apoptosis rate. This finding is in contrast with the resultsof a recent study showing that pretreatment with NAC pre-

vents cell death induced by an anti-erbB2 antibody (Gordonet al., 2009). It must be highlighted that, although we usedthe anti-erbB2 antibody, B10, at a concentration that doesnot induce cell damage, Gordon et al. (2009) blocked theerbB2 by using the C-18 antibody at a high enough concen-tration to induce cell death. Because it has recently beenfound that erbB2 regulates many antioxidant defense genesand that cell death induced by the erbB2 antibody occursthrough ROS-generating processes (Timolati et al., 2006), itis not surprising that pretreatment with antioxidants wasable to prevent the anti-erbB2-induced cell death in theirexperimental model.

ConclusionsMany women suffer from myocardial dysfunction after

having been treated with anthracycline-trastuzumab regi-

- + - + - + + -

- - - - + + + +

- - + + - - + +

- + - + - +- - + + - -

- - - - + +

GAPDH

0

100

200

300

400

* * * *

*

erbB2

GAPDH

EpiB10Dex

0

100

200

300

400

erbB2

§*

erb

B2

exp

ress

ion

(%)

EpiNACDPI

a

b

erb

B2

exp

ress

ion

(%)

*§

*§

§‡

*

§#

*

Fig. 7. Changes in erbB2 expression levels after various treatments.Quantization of data shows that B10 down-regulates, whereas epirubicinup-regulates erbB2 expression as assessed by Western blot analysis per-formed 24 h later. Note that dexrazoxane and NAC alone increase theerbB2 expression levels, whereas pretreatment with these drugs attenu-ates the epirubicin-induced erbB2 overexpression. The NAD(P)H oxidaseinhibitor, DPI, alone does not affect erbB2 expression, even though itabolishes the epirubicin-induced erbB2 overexpression. Data demon-strate that the epirubicin-induced erbB2 overexpression occurs throughROS-generating NAD(P)H oxidase activation. The amount of erbB2 wasnormalized to GAPDH. �, p � 0.05 versus untreated cells; §, p � 0.05versus Epi; #, p � 0.05 versus Dex; ‡, p � 0.05 versus NAC.

0

2000

4000

6000

8000

10000

12000

14000

0 20m40m

120m3h 3h+20m

3h+40m

3h+120m

Epi

Dex+Epi

Dex

Dex pretreatment

pA

kt (

AU

)

time

a

Epi treatment

0

100

200

300

400

- + + + + - -- - - + + + +

- - + - + - +

GAPDH

erbB2

#

EpiDex

LY

‡

b

erb

B2

ex

pre

ssio

n(%

)

Fig. 8. PI3K/Akt activation is part of the survival pathway induced bydexrazoxane. Data also suggest that if cells are protected by dexrazoxaneagainst the epirubicin-induced damage, epirubicin does not induce erbB2overexpression and Akt activation. a, time-curve analysis showing themodulation of Akt phosphorylation (pAkt) by dexrazoxane and epirubicinand the sequential treatment with dexrazoxane and epirubicin; the valueof pAkt was normalized to Akt. b, bar graph showing that pretreatmentwith Ly294002, a specific PI3K inhibitor, prevents the dexrazoxane-induced erbB2 up-regulation, but not the up-regulation induced by epi-rubicin. AU, arbitrary units; LY, LY294002. Amount of erbB2 was nor-malized to GAPDH. ‡, p � 0.05 versus Epi �Dex; #, p � 0.05 versus Dex.

Dexrazoxane, Epirubicin, and Anti-erbB2 in Cardiomyocytes 95

mens, and many others who have been treated with anthra-cyclines either do not receive trastuzumab at all, or receiveoverly low doses of it because of the risk of cardiac toxicity.Our study represents the pathophysiological experimentalbackground that is needed to plan clinical trials aimed atevaluating whether pretreatment with dexrazoxane duringanthracycline chemotherapy can prevent left ventricular dys-function, which may arise later on during treatment withtrastuzumab.

ReferencesAntos CL, McKinsey TA, Dreitz M, Hollingsworth LM, Zhang CL, Schreiber K, Rindt

H, Gorczynski RJ, and Olson EN (2003) Dose-dependent blockade to cardiomyo-cyte hypertrophy by histone deacetylase inhibitors. J Biol Chem 278:28930–28937.

Ben-Porath I and Weinberg RA (2004) When cells get stressed: an integrative viewof cellular senescence. J Clin Invest 113:8–13.

Bria E, Cuppone F, Fornier M, Nistico C, Carlini P, Milella M, Sperduti I, Terzoli E,Cognetti F, and Giannarelli D (2008) Cardiotoxicity and incidence of brain metas-tases after adjuvant trastuzumab for early breast cancer: the dark side of themoon? A meta-analysis of the randomized trials. Breast Cancer Res Treat 109:231–239.

Brown SB, Bailey K, and Savill J (1997) Actin is cleaved during constitutive apo-ptosis. Biochem J 323:233–237.

Chen X, Wilson RM, Kubo H, Berretta RM, Harris DM, Zhang X, Jaleel N, Mac-Donnell SM, Bearzi C, Tillmanns J, et al. (2007) Adolescent feline heart containsa population of small, proliferative ventricular myocytes with immature physio-logical properties. Circ Res 100:536–544.

Chuang TC, Liu JY, Lin CT, Tang YT, Yeh MH, Chang SC, Li JW, and Kao MC(2007) Human manganese superoxide dismutase suppresses HER2/neu-mediatedbreast cancer malignancy. FEBS Lett 58:4443–4449.

Damen CW, Rosing H, Schellens JH, and Beijnen JH (2008) Quantitative aspects ofthe analysis of the monoclonal antibody trastuzumab using high-performanceliquid chromatography coupled with electrospray mass spectrometry. J PharmBiomed Anal 46:449–455.

Dimri GP, Lee X, Basile G, Acosta M, Scott G, Roskelley C, Medrano EE, LinskensM, Rubelj I, and Pereira-Smith O (1995) A biomarker that identifies senescenthuman cells in culture and in aging skin in vivo. Proc Natl Acad Sci U S A92:9363–9367.

Eom YW, Kim MA, Park SS, Goo MJ, Kwon HJ, Sohn S, Kim WH, Yoon G, and ChoiKS (2005) Two distinct modes of cell death induced by doxorubicin: apoptosis andcell death through mitotic catastrophe accompanied by senescence-like phenotype.Oncogene 24:4765–4777.

Ewer MS and Benjamin RS (2000) Cardiac complications, in Cancer Medicine (BastRC, Kufe DW, Pollock RE, Weichselbaum RR, Holland JF, Frei E, and Gansler TS,eds), 5th ed, pp 2324–2339, Williams & Wilkins, Baltimore.

Ewer SM and Ewer MS (2008) Cardiotoxicity profile of trastuzumab. Drug Saf31:459–467.

Ewer MS and Tan-Chiu E (2007) Reversibility of trastuzumab cardiotoxicity: is theconcept alive and well? J Clin Oncol 25:5532–5533.

Furukawa Y, Nishimura N, Furukawa Y, Satoh M, Endo H, Iwase S, Yamada H,Matsuda M, Kano Y, and Nakamura M (2002) Apaf-1 Is a Mediator of E2F-1-induced Apoptosis. J Biol Chem 42:39760–39768.

Gabrielson K, Bedja D, Pin S, Tsao A, Gama L, Yuan B, and Muratore N (2007) Heatshock protein 90 and ErbB2 in the cardiac response to doxorubicin injury. CancerRes 67:1436–1441.

Gerland LM, Peyrol S, Lallemand C, Branche R, Magaud JP, and Ffrench M (2003)Association of increased autophagic inclusions labeled for beta-galactosidase withfibroblastic aging. Exp Gerontol 38:887–895.

Gordon LI, Burke MA, Singh AT, Prachand S, Lieberman ED, Sun L, Naik TJ,Prasad SV, and Ardehali H (2009) Blockade of the erbB2 receptor induces car-diomyocyte death through mitochondrial and reactive oxygen species-dependentpathways. J Biol Chem 284:2080–2087.

Halliwell B (2000) The antioxidant paradox. Lancet 355:1179–1180.Kanakry CG, Li Z, Nakai Y, Sei Y, and Weinberger DR (2007) Neuregulin-1 regu-

lates cell adhesion via an ErbB2/phosphoinositide-3 kinase/Akt-dependent path-way: potential implications for schizophrenia and cancer. PLoS ONE 2:e1369.

Konorev EA, Vanamala S, and Kalyanaraman B (2008) Differences in doxorubicin-induce apoptotic signaling in adult and immature cardiomyocytes. Free Radic BiolMed 45:1723–1728.

Li L, Takemura G, Li Y, Miyata S, Esaki M, Okada H, Kanamori H, Khai NC,

Maruyama R, Ogino A, et al. (2006) Preventive effect of erythropoietin on cardiacdysfunction in doxorubicin-induced cardiomyopathy. Circulation 113:535–543.

Lu L, Wu W, Yan J, Li X, Yu H, and Yu X (2009) Adriamycin-induced autophagiccardiomyocyte death plays a pathogenic role in a rat model of heart failure IntJ Cardiol 134:82–90.

Maejima Y, Adachi S, Ito H, Hirao K, and Isobe M (2008) Induction of prematuresenescence in cardiomyocytes by doxorubicin as a novel mechanism of myocardialdamage. Aging Cell 7:125–136.

Malisza KL and Hasinoff BB (1996) Inhibition of anthracycline semiquinone forma-tion by ICRF-187 (dexrazoxane) in cells. Free Radic Biol Med 20:905–914.

Menna P, Salvatorelli E, and Minotti G (2008) Cardiotoxicity of antitumor drugs.Chem Res Toxicol 21:978–989.

Pentassuglia L, Graf M, Lane H, Kuramochi Y, Cote G, Timolati F, Sawyer DB,Zuppinger C, and Suter TM (2009) Inhibition of ErbB2 by receptor tyrosine kinaseinhibitors causes myofibrillar structural damage without cell death in adult ratcardiomyocytes. Exp Cell Res 315:1302–1312.

Pugatsch T, Abedat S, Lotan C, and Beeri R (2006) Anti-erbB2 treatment inducescardiotoxicity by interfering with cell survival pathways. Breast Cancer Res 8:R35.

Salvatorelli E, Guarnieri S, Menna P, Liberi G, Calafiore AM, Mariggio MA, Mor-dente A, Gianni L, and Minotti G (2006) Defective one- or two-electron reductionof the anticancer anthracycline epirubicin in human heart. Relative importance ofvesicular sequestration and impaired efficiency of electron addition. J Biol Chem281:10990–11001.

Sharpless NE and DePinho RA (2004) Telomeres, stem cells, senescence, and cancer.J. Clin. Invest 113:160–168.

Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, Fleming T,Eiermann W, Wolter J, Pegram M, et al. (2001) Use of chemotherapy plus amonoclonal antibody against HER2 for metastatic breast cancer that overex-presses HER2. N Engl J Med 344:783–792.

Spallarossa P, Altieri P, Garibaldi S, Ghigliotti G, Barisione C, Manca V, Fabbi P,Ballestrero A, Brunelli C, and Barsotti A (2006) Matrix metalloproteinase-2 and -9are induced differently by doxorubicin in H9c2 cells: The role of MAP kinases andNAD(P)H oxidase. Cardiovasc Res 69:736–745.

Spallarossa P, Garibaldi S, Altieri P, Fabbi P, Manca V, Nasti S, Rossettin P,Ghigliotti G, Ballestrero A, Patrone F, et al. (2004) Carvedilol prevents doxorubi-cin-induced free radical release and apoptosis in cardiomyocytes in vitro. J MolCell Cardiol 37:837–846.

Takemura G, Miyata S, Kawase Y, Okada H, Maruyama R, and Fujiwara H (2006)Autophagic degeneration and death of cardiomyocytes in heart failure. Autophagy2:212–214.

Tan-Chiu E, Yothers G, Romond E, Geyer CE Jr, Ewer M, Keefe D, Shannon RP,Swain SM, Brown A, Fehrenbacher L, et al. (2005) Assessment of cardiac dysfunc-tion in a randomized trial comparing doxorubicin and cyclophosphamide followedby paclitaxel, with or without trastuzumab as adjuvant therapy in node-positive,human epidermal growth factor receptor 2-overexpressing breast cancer. NSABPB-31. J Clin Oncol 23:7811–7819.

Tauszig-Delamasure S, Yu LY, Cabrera JR, Bouzas-Rodriguez J, Mermet-Bouvier C,Guix C, Bordeaux MC, Arumae U, and Mehlen P (2007) The TrkC receptor inducesapoptosis when the dependence receptor notion meets the neurotrophin paradigm.Proc Natl Acad Sci U S A 104:13361–13366.

Terman A, Dalen H, Eaton JW, Neuzil J, and Brunk UT (2004) Aging of cardiacmyocytes in culture: oxidative stress, lipofuscin accumulation, and mitochondrialturnover. Ann N Y Acad Sci 1019:70–77.

Timolati F, Ott D, Pentassuglia L, Giraud MN, Perriard JC, Suter TM, and Zup-pinger C (2006) Neuregulin-1 beta attenuates doxorubicin-induced alterations ofexcitation-contraction coupling and reduces oxidative stress in adult rat cardio-myocytes. J Mol Cell Cardiol 41:845–854.

Tokunaga E, Oki E, Kimura Y, Yamanaka T, Egashira A, Nishida K, Koga T, MoritaM, Kakeji Y, and Maehara Y (2007) Coexistence of the loss of heterozygosity at thePTEN locus and HER2 overexpression enhances the Akt activity thus leading to anegative progesterone receptor expression in breast carcinoma. Breast Cancer ResTreat 101:249–257.

van Dalen EC, Caron HN, Dickinson HO, and Kremer LCM (2008) Cardioprotectiveinterventions for cancer patients receiving anthracyclines (2008) Cochrane Data-base Syst Rev 2:CD003917.

Xiang P, Deng HY, Li K, Huang GY, Chen Y, Tu L, Ng PC, Pong NH, Zhao H, ZhangL, et al. (2009) Dexrazoxane protects against doxorubicin-induced cardiomyopa-thy: upregulation of Akt and Erk phosphorylation in a rat model Cancer Che-mother Pharmacol 63:343–349.

Zhou BP, Hu MC, Miller SA, Yu Z, Xia W, Lin SY, and Hung MC (2000) HER-2/neublocks tumor necrosis factor-induced apoptosis via the Akt/NF-kappaB pathway.J Biol Chem 275:8027–8031.

Address correspondence to: Dr. Paola Altieri, Research Center of Cardio-vascular Biology, University of Genova, Viale Benedetto XV, 6. 16132-Genova.E-mail: Paola.Altieri@unige.it

96 Spallarossa et al.