The molecular mechanisms underlying doxorubicin (DOX)-induced cardiomyopathy include alterations in cardiomyocytes' oxidative stress status and in gene expression. Although such alterations have been reported during in vivo DOX treatment of animals, it remains to be clarified whether they persist after treatment cessation. To address this question, rats were injected with either saline (1 ml/kg/day i.p; control) or DOX (1 mg/kg/day i.p.) for 10 days, and 70 days later cardiac functional parameters were evaluated in vivo by left ventricular catheterization. Hearts were also harvested for histological analyses as well as measurements of oxidative stress parameters by various techniques and gene expression by quantitative polymerase chain reaction of markers of cardiac pathological remodeling, namely atrial natriuretic factor, myosin heavy chain β, vascular endothelial growth factor A (VEGF-A), and sarcoplasmic reticulum Ca+2 ATPase. Compared with controls, DOX-treated rats displayed marked alterations in most parameters even 2 months after cessation of treatment. These included 1) lower left ventricular contractility (+dP/dt), 2) increased levels of plasma and myocardial oxidative stress markers, namely thiobarbituric acid reactive substances or dihydroethidium fluorescence, and 3) markedly altered transcript levels for all measured markers of cardiac remodeling, except VEGF-A. These changes correlated significantly with +dP/dt values assessed in the two groups of animals. In conclusion, this study demonstrated that as many as 2 months after cessation of DOX treatment cardiac alterations persisted, reflecting increased oxidative stress and pathological remodeling, the latter being linked to the development of contractile dysfunction.
Doxorubicin (DOX) is an anthracycline antineoplasic drug used in the treatment of a wide range of solid tumors and leukemia in children and adults. Despite its good therapeutic results, the clinical use of doxorubicin during chemotherapy is limited by the development of acute, subacute, early chronic, or late chronic cardiotoxicity (Kremer et al., 2001). Cardiac alterations become more obvious several years after the end of the treatment and may occur as dilated cardiomyopathy, leading to congestive heart failure (Giantris et al., 1998; Scully and Lipshultz, 2007).
The different lines of evidence have provided putative mechanisms, but the molecular mechanisms involved in chronic anthracycline cardiotoxicity remain a major topic of discussion. The antitumor activity of DOX is probably distinct from the mechanisms of its cardiotoxicity. Several aspects of this phenomenon, such as apoptosis, alteration of iron, and calcium homeostasis have been described, but the exact mechanism is not yet been fully understood. Oxidative stress is believed to be an important pathway in the cardiac side effects of anthracycline therapy (Tokarska-Schlattner et al., 2006). It is widely known that superoxide anion free radical (O2̇̄) is generated during the “redox cycling” of anthracycline or after oxido-reduction processes taking place inside the anthracycline-iron complex. Then O2̇̄ is transformed into hydroxyl radical (HO•) in the presence of transition metals or forms peroxinitrite (ONOO−) in the presence of nitric oxide (•NO). These oxidants induce cellular injury (Delemasure et al., 2007; Simůnek et al., 2009). The molecular mechanisms involved in DOX-induced cardiomyopathy have not yet been fully identified, but it is likely that they are associated with alterations in the structure and gene expression of cardiac myocytes. In this way, it has been reported that DOX induced effects on L-type calcium channels (Campbell et al., 1996), sarcoplasmic reticulum Ca2+ ATPase (SERCA) (Arai et al., 2000), ryanodine receptors (Wang and Korth, 1995), and Na+/Ca+ exchanger activity (Goldhaber, 1996). On the other hand, it has also been reported that some pathological cardiac dysfunctions such as hypertrophy were associated with the up-regulation of fetal genes such as atrial natriuretic factor (ANF) and isoforms of myosin heavy chain (MHC) (Ritter and Neyses, 2003; van den Bosch et al., 2006). Early alterations in heart gene expression profiles associated with cardiotoxicity have been studied in vitro and in vivo (Thompson et al., 2010). Using microarrays, Berthiaume and Wallace (2007) studied global gene expression changes occurring in rats' hearts 5 weeks after subchronic DOX treatment. The major findings of this study were a decrease in the expression of genes governing fatty acid metabolism. But to our knowledge, there have been no specific studies on oxidative stress and gene expression long after the end of in vivo treatment with DOX in an animal model.
In this context, the aim of our work was to clarify the role of oxidative stress and the expression of certain genes coding for proteins likely to be implicated in DOX-induced cardiac injury 2 months after cessation of treatment and determine the functional and cellular parameters in the plasma and myocardium.
Materials and Methods
Animals and Experimental Protocol
The local ethics committee approved the experimental protocol, and the investigators complied with authorization 6007 from the French government, which agrees with the Guide of Care and Use of Laboratory Animals published by the National Institutes for Health (Institute of Laboratory Animal Resources, 1996).
For the purpose of our study, male Wistar rats (Charles River Laboratories, L'Abresle, France), 300 to 350 g at the beginning of the experiment, were divided into two groups.
In the control group (n = 8), rats received saline solution: 1 ml/kg/day i.p. for 10 days, day 0 (D0) to D9.
In the DOX-treated group (n = 8), rats were injected intraperitoneally with 1 mg/kg/day DOX (Pfizer, Paris, France) for 10 days, from D0 to D9. A total dose of 10 mg/kg doxorubicin was administered over the 10-day period.
The body weight of the rats was measured daily during the period of treatment (D0-D9), then every 10 days. Moreover, food and water consumption were recorded. Plasma concentrations of lipid peroxides were measured.
At the end of the study, 60 days after the end of the treatment (D70) heart functional parameters were evaluated in vivo by left ventricular catheterization.
Heart Functional Parameters Measured In Vivo
The rats were anesthetized with sodium thiopental (60 mg/kg i.p.) and heparinized (500 IU/kg). Once the thorax had been shaved, a catheter connected to a pressure transducer was inserted into the left ventricle through the chest wall to measure the heart rate (HR) and left-ventricular pressures [left-ventricular end-diastolic pressure, left-ventricular systolic pressure, and left-ventricular developed pressure (LVDP; left-ventricular systolic pressure − left-ventricular end-diastolic pressure)] during the first minute of cardiac catheterization. Left-ventricle contractility was expressed as +dP/dt and left-ventricle relaxation as −dP/dt.
At the end of the study, blood was taken by cardiac puncture and centrifuged, and the plasma was immediately frozen in liquid nitrogen. The hearts were excised, cut into segments, and frozen.
The hearts of three controls and three DOXO-treated rats were harvested for histological analyses. The heart samples were snap-frozen in isopentane then in OCT compound. Five-micrometer-thick sections were cut using a microtome (CM3050S; Leica, Wetzlar, Germany) and stained using the Sirius Red protocol (0.1% Direct Red 80 in saturated picric acid solution), which is specific to fibrillar collagens (collagen I and III). Randomly selected fields were evaluated under a polarized-light microscope (Eclipse 600; Nikon, Champigny-Sur-Marne, France) connected to a digital video camera (Tri CCD; Sony, Paris, France). The quantity of collagen was measured using Visilog software (Noesis, Saint Aubin, France) and expressed as the ratio of the mean area of collagen to the total area.
In the presence of superoxide, ethidine, a fluorescent compound, is formed from dihydroethidium (DHE) and thus allows quantification of superoxide production. Frozen heart tissues were fixed for 10 min in acetone. Slides were incubated in a light-protected humidified chamber at room temperature with DHE (5 μM) for 5 min. The slides were immediately analyzed with a computer-based digitizing image system (Microvision, Paris, France) using a fluorescent microscope (Eclipse 600; Nikon) connected to a video camera (TriCCD; Sony). Fluorescence was detected at 590 nm and carried forward to the nuclear number. Results were expressed in fluorescence intensity/nuclear number.
Determination of Thiobarbituric Acid Reactive Substances.
Plasma lipid peroxides were measured by using a colorimetric reaction with thiobarbituric acid. Trichloroacetic acid/thiobarbituric acid/hydrochloric acid solution (1.5 ml) was added to 500 μl of plasma. The color of the thiobarbituric acid pigment was developed in a water bath at 100°C for 15 min. After cooling with ice to room temperature, 1 ml of 70% trichloroacetic acid was added. After 90 min, the tubes were centrifuged, and the color of TBARS layers was measured at 553 nm. The absorbance values were compared with a standard curve. Results were expressed in μmoles/gram of proteins; plasma protein was determined according to the Bradford method (Ghibu et al., 2009). Plasma TBARS were assessed only at the end of the study.
Plasma Ascorbyl Free Radical Determinations by Electron Spin Resonance Spectroscopy.
Thirty five microliters of plasma samples were inserted into a quartz capillary tube, which was placed in a high-sensitivity cavity to analyze them at room temperature with a Bruker (Wissembourg, France) EMX-100 X-band spectrometer. The following parameters were selected for optimal detection of AFR (Vergely et al., 1998): modulation frequency, 100 kHz; amplitude modulation, 0.8 G; microwave power, 40 mW; microwave frequency, 8.5 GHz; conversion time, 40 ms; time constant, 327 ms; scan time, 41 s; gain, 5 × 105; number of scans, 6. The height of AFR signal intensity was measured and expressed in arbitrary units.
Determination of Cardiac Thiobarbituric Acid Reactive Substances.
Heart lipid peroxides were measured by using a colorimetric reaction with thiobarbituric acid. The hearts were homogenized in ice-cold phosphate-buffered saline (0.05 M, pH 7). Then, 1.5 ml of trichloroacetic acid/thiobarbituric acid/hydrochloric acid solution was added (see previously). Results were expressed in micromole/gram of cardiac tissue.
mRNA Levels Assessed by Quantitative Polymerase Chain Reaction.
Freeze-clamped hearts were used to assess mRNA levels for selected marker genes (listed in Table 1) that reflect cardiac remodeling, namely ANF, myosin heavy chain β (β-MHC), vascular endothelial growth factor (VEGF), and SERCA2a, as well as a marker of oxidative stress-induced adaptive response, namely the transcription factor Nrf2 gene, heme oxygenase-1 (HMOX). Gene-specific primer pairs were designed (Beacon Designer 5.0 program; Bio-Rad Laboratories, St. Laurent, QC, Canada) on the basis of rat sequences available in GenBank. Total RNA was extracted with QIAGEN (Valencia, CA) RNeasy Plus columns, quantified by an Agilent Technologies (Santa Clara, CA) BioAnalyzer 2100, and assessed by reverse transcription, followed by real-time quantitative polymerase chain reaction as described previously (Lauzier et al., 2011). Cycling was achieved in a MX3005p cycler (Stratagene, Mississauga, ONT, Canada) under conditions 95°C for 10 min and 46 cycles at 94°C for 15 s, 60°C for 30 s, and 72°C for 30 s.
All data are expressed as means ± S.E.M. To compare the groups at 2 months after the end of treatment, statistical analyses were performed with the one-factor analysis of variance (ANOVA) test (SigmaStat; Systat Software, Inc., San Jose, CA). ANOVA was followed, if necessary, by a Newman-Keuls test. To compare the evolution of parameters throughout the period of study (periods of treatment and observation) we used a two-factor repeated-measures ANOVA test (SigmaStat). Significance was established at a value of P < 0.05.
Body and Heart Weights and Ratio of Heart Weight to Body Weight
In the rats, DOX induced a significant loss of body weight starting from the third day of treatment. After interruption of treatment, there was a trend toward recovery of body weight with kinetics identical to those of control rats.
Two months (D70) after the end of the treatment, the hearts of DOX-treated rats weighed significantly less than control hearts. The heart-to-body-weight ratio, which estimates cardiac hypertrophy, was significantly (P < 0.05) increased in the DOX group (Table 2).
Cardiac Parameters Evaluated In Vivo by Left Ventricle Catheterization
HR was not affected by DOX treatment at D70 (Table 3). However, LVDP, +dP/dT, and −dP/dT were modified 2 months after the end of treatment; animals treated with DOX had significantly lower +dP/dT, −dP/dT, and LVDP than controls. These deleterious effects were observed in the left ventricle. The effects in the right ventricle were not significant (data not shown).
Sirius Red staining of heart sections showed apparent collagen type I and III deposition in the left ventricle (Figs. 1 and 2). The most consistently and severely affected areas occurred around the branches of the coronary vessels. The main histological alterations in vessels are shown in Fig. 1A. DOX increased quantities of collagen in the arterial adventitia: DOX (293 ± 98%) compared with controls (124 ± 55%) group (p < 0.05) (Fig. 2A). However, no difference was found between the DOX and control group for collagen content in the subepicardial regions of the left-ventricular myocardium (Figs. 1B and 2B). In contrast, there were focal morphological changes that could be related to the formation of edema.
Heart Tissue Superoxide Anion Production.
Two months after the end of the treatment, the production of O2̇̄ in heart tissue was significantly (p < 0.05) higher in the group of rats treated with DOX than in the control group (Table 4).
Plasma Concentrations of Thiobarbituric Acid Reactive Substances.
Two months after the end of the treatment, the plasma level of TBARS was significantly higher in the DOX group than in the control group (Table 4).
Ascorbate Radicals and AFR/Ascorbate Ratio in Plasma.
There was a significant decrease in plasma concentrations of ascorbate in the group of rats treated with DOX, associated with a significant increase in the AFR to ascorbate ratio (Table 4).
Heart Tissue TBARS.
Two months after the end of the treatment, levels of TBARS in the hearts of rats from the DOX group were significant higher (+50%; P < 0.05) than those in control hearts (Table 4).
Cardiac Transcript Levels for Selected Gene Markers of Cardiac Remodeling.
Compared with control hearts, the hearts from DOX-treated rats displayed markedly altered mRNA levels for markers of cardiac remodeling (Fig. 3), namely ANF and β-MHC (increased 4.8- and 3.2-fold, respectively), and SERCA2a (decreased; −25%) (p < 0.001). There was no difference, however, for VEGF-A. It is noteworthy that there were significant correlations between cardiac gene expression for ANF, β-MHC, and SERCA2a and +dP/dT for the two groups of animals (Fig. 4).
In rats, the administration of a cumulative dose of 10 mg/kg DOX induced a decrease in body weight, which was associated with reduced food consumption. After the end of the treatment, during the 2 months of the experimental design, body weight and food consumption recovered, but did not reach the values of the control group. In our experiment, the cardiac alterations were more obvious, as revealed by a greater deterioration in heart contractility in the rats treated with DOX. This suggests progressive cardiac dysfunction long after exposure to DOX, as is the case in human patients, who may die months or even years after chemotherapy has been stopped.
Two months after the end of treatment, heart weight in the DOX group was lower than in controls. This phenomenon has been described in other studies (Sacco et al., 2003; Richard et al., 2008; Ghibu et al., 2011) and could be explained by the apoptosis of cardiomyocytes (Bennink et al., 2004; Reeve et al., 2007). Two months after the end of treatment, the heart-to-body-weight ratio, an index of cardiac hypertrophy, was higher in the DOX group and was associated with an increase in the semiquantitative scores from the collagen analysis. These results are in accordance with previous studies (Sanchez-Quintana et al., 1994).
In vivo cardiac functional parameters were impaired 2 months after chronic treatment with DOX, with lower left ventricular contractility (+dP/dt) in DOX hearts, which induced a significant reduction in LVDP. As we have already observed (Richard et al., 2008), the impairment of contractility is a very late event and is only at its beginnings 2 months after a cumulative dose of 10 mg/kg DOX.
Plasma and heart lipid peroxidation was assessed as TBARS concentrations at the time the rats were killed. To get a better evaluation of plasma oxidative stress, we determined the plasma concentration of vitamin C by high-performance liquid chromatography and the plasma ascorbyl free radical levels by electron spin resonance spectroscopy. Two months after the end of the treatment, the plasma concentration of ascorbate was significantly lower in rats treated with doxorubicin, and this was associated with a significant increase in the AFR-to-ascorbate ratio. AFR can be considered as an endogenous indicator of oxidative stress and a terminal paramagnetic product of free-radical transformations in the antioxidant defense system (Vergely et al., 1998). The increase in this ratio could be related to the chronic deleterious effects of DOX long after the end of the treatment. In consequence, this phenomenon could be explained by a decrease in hepatic vitamin C synthesis and an excess production of free radical species, as we reported previously (Richard et al., 2008).
The evaluation of cardiac oxidative stress by TBARs or DHE fluorescence showed significantly higher levels of O2̇̄ production and peroxidation in the DOX group. In our similar short-term study (Richard et al., 2008), we found no evidence of cardiac oxidative stress 8 days after treatment with a cumulative dose of 10 mg/kg doxorubicin, confirming that tissue oxidative stress is a late event in doxorubicin cardiotoxicity.
As shown in this study, DOX induces an increase in collagen I and III content in arterial adventitia. Our results provide direct evidence that advential tissue may have been affected by DOX, this tissue being a source of inflammatory mediators. In contrast to well characterized processes of endothelial lesion, changes in the adventitia during DOX treatment have been neglected. Our results are in agreement with an earlier study in rats receiving DOX that reported marked fibrosis associated with morphological changes that could be related to edema and inflammation (Yagmurca et al., 2003). Most studies support the view that an increase in oxidative stress plays a key role in the pathogenesis, whereas others suggest the release of vasoactive mediators and hormones (Takemura and Fujiwara, 2007). It is well documented in clinical studies that the spectrum of cardiotoxicity with anthracycline agents includes coronary diseases (Swerdlow et al., 2007). Using the isolated perfused heart model, we revealed prolonged deterioration in coronary flow during treatment with anthracyclines (Delemasure et al., 2007).
The molecular mechanisms involved in the deleterious effects of DOX on the heart are complex and remain a matter of controversy (Minotti et al., 2004). Mechanisms that have been suggested include alterations in genes important for the structural integrity and enzymatic function of cardiac and vessel myocytes. These phenomena can lead to inadequate maintenance of contractile function in the heart. In our study, we focused on the possible role of various molecular pathways in relationship with oxidative stress and the development of heart failure. HMOX-1 is one of the three isoforms of heme-oxygenase enzyme that catabolizes the degradation of heme into biliverdin with the production of free iron and carbon monoxide (CO). HMOX-1 is induced by stimuli including agents involved in oxidative stress and serves as a protective gene in a wide range of pathological situations (Ryter et al., 2006). In our experimental conditions, we found no modifications of HMOX-1 in the myocardium of animals treated with DOX, even though there was an increase in oxidative stress. The lack of HMOX-1 induction by DOX is surprising, because HMOX-1 is an inducible defense mechanism activated ubiquitously. However, the essential role of HMOX-1 takes place during the acute stress adaptation and, in our experimental conditions, levels were measured a long time after the end of DOX treatment.
VEGF-A is a pivotal angiogenic factor in most tumors and also seems to be a mediator of angiogenesis in several other disease conditions (Carmeliet and Jain, 2000). Our data demonstrate that VEGF-A is expressed in the heart but that DOX-induced cardiotoxicity was not associated with any change in VEGF-A expression. Our results are in accordance with data obtained on the mesenteric area showing that DOX did not significantly affect angiogenesis, in doses ranging from 1.8 to 10.8 mg/kg/week, even when cotreated with the antioxidant N-acetylcysteine (Albertsson et al., 2006). It is clear from our present data that the severe cardiac disorders induced by DOX did not result in a modified VEGF-A profile.
Pathological cardiac dysfunction has been associated with the up-regulation of fetal genes such as ANF and β-MHC (Izumo et al., 1988; van den Bosch et al., 2006). The important results of our study on DOX-induced cardiotoxicity concern the modifications of mRNA in the genes ANF, β-MHC, and SERCA2a in hearts collected 2 months after the end of the treatment. We observed a very significant increase in ANF and β-MHC expression associated with a significant decrease in SERCA2a. A positive and significant correlation was found between SERCA2a expression and +dP/dt, whereas there was a significant negative correlation between ANF expression and +dP/dt in the hearts of both groups of animals.
The expression of α and β isoforms of MHC genes is developmentally regulated in the myocardium (Lompré et al., 1991). There is growing evidence that the increased expression of β-MHC in the failing ventricle may serve as a compensatory mechanism to increase contractile efficiency by decreasing the tension cost of contraction (Sanbe et al., 2005). It has been reported (de Beer et al., 2000) that chronic DOX treatment significantly increased the ratio of β-MHC to α-MHC in ventricular tissue. Therefore a relationship exists between the decreased contractile performance and the relative increase in β-MHC expression.
Concerning the molecular impact of treatment with DOX on ANF and SERCA2a expression, these effects are related to the progression of cardiac dysfunction. It has been clearly demonstrated that mRNA levels of the sarcoplasmic reticulum calcium pump were lower in failing than in nonfailing hearts (Hasenfuss, 1998). A large body of evidence points to the central role of SERCA2a in the modulation of cardiac relaxation. The levels of these calcium-handling proteins are altered in cardiomyopathies (Kranias and Bers, 2007). Our results are in accordance with the concept that DOX-induced cardiotoxicity depends on the down-regulation of the SERCA pump; SERCA was shown previously to be down-regulated by DOX and oxidative stress induced by hydrogen peroxide (Arai et al., 2000).
Plasma levels of natriuretic peptides increase in patients with severe coronary heart failure and also in patients with asymptomatic left ventricle dysfunction (Ogawa et al., 2002). ANF and B-type natriuretic peptide are also useful markers of ventricular dysfunction in patients undergoing anthracycline therapy (Okumura et al., 2000). To further investigate the markers of DOX-induced cardiac toxicity in our chronic model, we evaluated the RNA expression levels for the fetal gene, ANF, in the myocardium. We observed a substantial increase in ANF mRNA and a negative correlation with the +dP/dt in control animals as well as in DOX-treated animals. Our results are in accordance with the few studies on the subject. Experimental heart failure has been associated with increased ANF mRNA in the left ventricle (Goetze et al., 2006). Because of the strong association between cardiomyocyte hypertrophy and the induction of fetal genes, the expression of ANF and β-MHC has often been used as a marker of hypertrophy and heart failure, but it remains unclear why and how the frequent transition from hypertrophy to failure occurs.
In conclusion, this study confirmed that 2 months after completion of the DOX treatment the state of animals had deteriorated and this was associated with a rise in oxidative stress and a modification in the expression of genes coding for proteins related to the development of cardiac dysfunction.
Participated in research design: Richard, Des Rosiers, Zeller, Cottin, Rochette, and Vergely.
Conducted experiments: Richard, Ghibu, Delemasure-Chalumeau, and Des Rosiers.
Contributed new reagents or analytic tools: Guilland and Zeller.
Performed data analysis: Richard, Delemasure-Chalumeau, Rochette, and Vergely.
Wrote or contributed to the writing of the manuscript: Richard, Des Rosiers, Rochette, and Vergely.
This work was supported by grants from the Conseil Régional de Bourgogne, the Association de Cardiologie de Bourgogne, and the Ligue Contre le Cancer.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
- atrial natriuretic factor
- ascorbyl free radical
- myosin heavy chain
- left ventricular contractility
- left ventricular relaxation
- heme oxygenase-1
- heart rate
- left ventricular developed pressure
- sarcoplasmic reticulum Ca2+ ATPase
- thiobarbituric acid reactive substances
- vascular endothelial growth factor
- analysis of variance
- day number.
- Received July 8, 2011.
- Accepted September 19, 2011.
- Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics